(Sozomena 14) David Sider, Dirk Obbink (Eds.) - Doctrine and Doxography - Studies On Heraclitus and Pythagoras-Walter de Gruyter (20

Doctrine and Doxography
Unauthenticated | 178.130.41.255
Download Date | 12/10/13 10:09 PM
Sozomena
Studies in the Recovery of Ancient Texts
Edited
on behalf of the Herculaneum Society
by
Alessandro Barchiesi, Robert Fowler,
Dirk Obbink and Nigel Wilson
Vol. 14
De Gruyter
Doctrine and Doxography
Studies on Heraclitus and Pythagoras
Edited by
David Sider and Dirk Obbink
De Gruyter
ISBN 978-3-11-033116-5
e-ISBN 978-3-11-033137-0
ISSN 1869-6368
Library of Congress Cataloging-in-Publication Data

A CIP catalog record for this book has been applied for at the Library of Congress.
Bibliographic information published by the Deutsche Nationalbibliothek

The Deutsche Nationalbibliothek lists this publication in the Deutsche
Nationalbibliografie; detailed bibliographic data are available in the Internet
at http://dnb.dnb.de.
2013 Walter de Gruyter GmbH, Berlin/Boston

Printing: Hubert & Co. GmbH & Co. KG, Gttingen
Printed on acid-free paper
Printed in Germany
www.degruyter.com
Preface
The essays collected here, all here revised, were first presented in at a
bicontinental conference hosted in July, 2005, by Apostolos Pierris, to
whom we all express our great thanks, not only for the invitation, but
also for the magnificent settings: First on Samos, Pythagoras native
land, and then by chartered boat over to Kusadas, as close to Heraclitus
home town of Ephesus as modern facilities allow. We regret that the
original contributions by Serge Mouraviev, Arnold Hermann and
Dirk Obbink are not present here, and that it has taken so long to
bring these essays to print.
David Sider
Dirk Obbink
Abbreviations
The fragments of Pythagoras, Heraclitus, and the other Presocratics are
given as in DK (14 is DKs number for the former, 59 for the latter).
Frequently cited modern works are abbreviated as indicated below.
Bernab A. Bernab, Poetae Epici Graeci. Testimonia et Fragmenta.
2 vols. Berlin 1996 2004. T & F numbers for Orpheus
and Orphica as in pars 2, fasc. 1 3.
Dilcher, H. R. Dilcher, 1995, Studies in Heraclitus. Hildesheim 1995.
Diels, DG Hermann Diels, Doxographi Graeci. Berlin 1879.
DK H. Diels and W. Kranz, Die Fragmente der Vorsokratiker 6.
Zurich 1952 [later editions are merely photographic
reprints; earlier editions were edited by Diels alone]
FGrHist Fragmente der griechische Historiker, ed. F. Jacoby. Leiden.
FHSG William Fortenbaugh, Pamela Huber, Robert Sharples,
and Dimitri Gutas, Theophrastus of Eresus: Sources for his
Life, Writings, Thought and Influence. 2 vols. Leiden 1992.
Guthrie, HGP W. K. C. Guthrie, A History of Greek Philosophy. Vol. 1
Cambridge 1962.
Kahn, H. C. Kahn, The Art and Thought of Heraclitus. Cambridge
1979.
K-G R. Khner and B. Gerth, Ausfrliche Grammatik der griechi-
schen Sprache. Hannover and Leipzig, 1904.
Kirk, H. G. S. Kirk, Heraclitus: The Cosmic Fragments. Cambridge,
1954.
KRS Kirk, G. S., J. E. Raven, and M. Schofield. The Presocratic
Philosophers. 2nd edn. Cambridge 1983.
Marcovich, H. M. Marcovich, Heraclitus: Greek Text with a Short Com-
mentary Merida 1967. 2nd ed., Sankt Augustin: Academia
Verlag, 2001.
Mouraviev, H. S. Mouraviev, Heraclitea: dition critique complte des t-
moignages sur la vie et loeuvre dHraclite dphse et des ves-
tiges de son livre, cited further by volume. Sankt Augustin
1999 .
Contents
I. Pythagoras
1. Catherine Rowett
Philosophys Numerical Turn: Why the Pythagoreans Interest in
Numbers is Truly Awesome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2. Leonid Zhmud
Pythagorean Communities: From Individuals to a Collective
Portrait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3. Richard McKirahan
Aristotle on the Pythagoreans: His Sources and his Accounts of
Pythagorean Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4. Carl Huffman
Philolaus Critique of Heraclitus . . . . . . . . . . . . . . . . . . . . . . . . . . 121
II. Heraclitus
5. Aryeh Finkelberg
Heraclitus, the Rival of Pythagoras . . . . . . . . . . . . . . . . . . . . . . . 147
6. Herbert Granger
Early Natural Theology: The Purification of the Divine Nature 163
7. Anthony A. Long
Heraclitus on Measure and the Explicit Emergence of Rationality 201
8. Gbor Betegh
On the physical aspect of Heraclitus psychology: With New
Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
9. Roman Dilcher
How Not to Conceive Heraclitean Harmony . . . . . . . . . . . . . . . 263
X Contents
10. Enrique Hlsz

Heraclitus on Logos: Language, Rationality and the Real . . . . . . . 281
11. Daniel W. Graham

Once More unto the Stream . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
12. David Sider

Heraclitus Ethics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 335
General Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
Index Locorum Potiorum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345
1. Philosophys Numerical Turn
Why the Pythagoreans Interest in Numbers is
Truly Awesome
Catherine Rowett
We do not know a great deal about Pythagoreanism in the early days

when Pythagoras was around. However, from what we do know, it
seems reasonable, and not hugely controversial, to suggest that a passion
for numbers was at least an embryonic part of the early heritage, some-
thing that already belongs in the genuine Pythagorean core, and was
certainly already central by what I shall call the second generation (in-
cluding Philolaus). 1 It is not just a feature of the crust of lichen and fun-
gus that grew round the name of Pythagoras in later generations (that is,
the period we regularly call Neopythagoreanism).
1 As regards Pythagoras himself, this is somewhat controversial. Burkert denies

that the stuff about numbers (mainly in Aristotle) has anything to do with
the real Pythagoras. He claims that Aristotle is careful to speak of the so-
called Pythagoreans, and that he means the second generation, the fifth cen-
tury Pythagoreans, not Pythagoras himself. By contrast, Burkert claims, the au-
thentic material about Pythagoras himself is almost exclusively about mysticism
and wonder-working. See Lore and Science in Ancient Pythagoreanism, trans. E. L.
Minar Jr. Cambridge, Mass. 1972. My claim is not that all the elaborate number
theories can be traced to the earliest Pythagoreanism of Pythagorass time, but
that an embryonic interest in harmony, and in the mystical aspects of number,
such as the oath by the tetraktys, can safely be traced back to that period, and
that a more developed interest in number is evident in the second generation.
Burkerts polemic was against F. M. Cornford, Mysticism and science in the
Pythagorean tradition, in A. P. D. Mourelatos (ed.), The Presocratics (Garden
City 1974) 135 60 originally published in 1922 3 (Cornford claimed that
early Pythagoreans had a mystical system that came under criticism from Par-
menides, and that later ones had a pluralist system that he called number atom-
ism which was the object of Zenos attack). C. J. de Vogel, Pythagoras and Early
Pythagoreanism: an Interpretation of Neglected Evidence on the Philosopher Pythagoras
(Assen 1966), published around the same time as Burkerts original publication
date, and not directly addressed to his finished publication, takes a less austere
line than Burkert about science in the first generation period.
4 Catherine Rowett
But what kind of a passion for numbers do we find in the Pythagor-

eans? Is it the kind of enthusiasm that philosophers can safely hail as part
of their authentic heritage? Can we attribute it to the real Pythagoras
(and his early followers) with pride? Or is it an embarrassment, to be
hushed up and omitted from the histories of real philosophy? Were
some of the Pythagoreans guilty of an unfortunate misjudgementper-
haps even some of the greatest thinkers, perhaps Pythagoras himself,
perhaps his followers, even the great Philolaus and Archytas, with the
later Neopythagoreans continuing their strange speculations right into
late antiquity? Did all or some of them fail to see the difference between
respectable mathematics and mystical mumbo-jumbo? Did they get car-
ried away with an inappropriate desire to put numbers where numbers
should not be?
Jonathan Barnes gives a dramatic rendering of the dilemma for the
historian of philosophy by inviting us first to approve and then to con-
demn. First he invites us to approve by sketching a story in which Py-
thagoras figures as a great mathematical hero and astronomer:
These pious offerings portray an impressive figure: Pythagoras, discoverer
and eponym of a celebrated theorem, was a brilliant mathematician; by ap-
plying his mathematical knowledge, he made progress in astronomy and
harmonics, those sister sirens who together compose the music of the
spheres; and finally, seeing mathematics and number at the bottom of
the master sciences, he concocted an elaborate physical and metaphysical
system and propounded a formal, arithmological cosmogony. Pythagoras
was a Greek Newton; and if his intellectual bonnet hummed at times
with an embarrassing swarm of mystico-religious bees, we might reflect
that Sir Isaac Newton devoted the best years of his life to the interpretation
of the number symbolism of the book of Revelations.
If Greek science began in Miletus, it grew up in Italy under the tutelage of
Pythagoras; and it was brought to maturity by Pythagoras school, whose
members, bound in fellowship by custom and ritual, secured the posthu-
mous influence of their masters voice.2
But then he withdraws that story, because (he thinks) it is not true:
What are we to make of this pleasing picture of a Newtonian Pythagoras? It
is, alas, mere fantasy: the shears of scholarship soon strip Pythagoras of his
philosophical fleece.3
2 J. Barnes, The Presocratic Philosophers (London 1982) 100 101.

3 Ibid. 101.
Philosophys Numerical Turn 5
And thirdly he invites us instead to condemn the whole Pythagorean

passion for numbersby implication including the work of Pythagoras
himself which is denigrated by association with late works from the
Neopythagorean period such as the Theologoumena arithmeticae :
What philosophical use did the Pythagoreans make of mathematics? The
cynical will speak dismissively of number mysticism, arithmology, and
other puerilities. And it is undeniable that a great quantity of Pythagorean
number philosophy is a number symbolism of the most jejune and inane
kind. Touching on arithmetic, the Pythagoreans were impressed by
certain properties of the number 10; alas, their impression degenerated
into a sort of mysticism: amazement, the nurse of philosophy, soon has
her milk soured and turns into silly reverence and superstition. Those
with a taste for intellectual folly will have their appetite sated if they go
through the Theologoumena Arithmeticae. That Pythagorean work is a late
compilation; the earliest examples of such symbolism are found in the acous-
mata and probably date from the time of Pythagoras himself: from first to
last the Pythagoreans engaged in arithmology.4
Barnes offers us two options: we could admire Pythagoras if he was, as
suggested, a Newtonian, whose mathematical discoveries were put to
fine use in developing a mathematical astronomy and answers to physics
that sought confirmation in mathematics. Or we could condemn him
(and his followers) if the philosophical use to which they put their num-
ber work was mere mysticism and number symbolism. Since the first
option seems to be ruled out by the lack of sound historical evidence
for the fanciful portrait of the Newtonian Pythagoras, we are left
with the second. And so we condemn.
My task in this paper is to persuade the reader that, notwithstanding
Barness fine rhetoric on the matter, we should nevertheless admire, and
not condemn, the Pythagorean enthusiasm for numbers, tracing its ori-
gins to Pythagoras himself and his immediate followers.5
4 Ibid. 381.
5 The mathematical interests are only one of the many kinds of wisdom attribut-
ed to Pythagoras. C. Huffman, The Pythagorean tradition, in A. A. Long
(ed.), The Cambridge Companion to Early Greek Philosophy (Cambridge 1999),
66 87, 66 7, gives a judicious account of a general difficulty created by the
exaggerated reputation of Pythagoras, and the problematic source materials.
6 Catherine Rowett
Pythagoras and the early Presocratic tradition.
One not particularly original route to defending the Pythagoreans

would be to reject Barness claim that the Newtonian Pythagoras is
mere fantasy. One could try to show that Pythagoras should be reliably
credited with certain significant mathematical discoveries, that his work
on harmonics was of a serious experimental nature, and that it is not im-
probable that he applied these studies to astronomy in a spirit of scien-
tific inquiry, so as to show that his work on numbers was not mere spec-
ulation or mysticism. That is not the task I have set myself here. For not
only is it not a particularly new project, but in any case it is not the case I
want to make, for reasons that I shall go on to explain.
A second rather more interesting project does strike me as worth-
while however. In this first part of my paper I shall not so much seek
to restore Pythagorass credibility as a practitioner of the exact sciences
as to ask why Pythagoras has been downgraded in the contemporary as-
sessments, dismissed as a mystic and wonder-worker, and his followers
dismissed as number mystics, while other Presocratics with rather similar
points to make have been exalted as pioneers of embryonic scientific
and philosophical thought. Is there really so much difference?
I shall consider two comparisons here: first between Anaximanders
notion of the apeiron and Pythagorean discussions of the limited and un-
limited, and secondly between Heraclitus notion of logos and harmony,
and the Pythagoreans theories relating to proportion and harmony.
Other examples could also be used, but these will be sufficient for
this task.6
Let us start with the infinite. First, we should note the appearance of
apeiron as a technical term in Anaximanders cosmology:
Anaximander, son of Praxiades, of Miletus said that the origin and ele-
ment of things is t %peiqom, being the first to use this term of the ori-
gin. Hippolytus Refutatio 1.6
This term is also prominent in certain Pythagorean documents, especial-
ly in Philolaus (chief among the second generation of Pythagoreans).
When this term apeiron occurs in Pythagorean documents (and in Ar-
6 For instance astronomy, in which Philolaus was arguably far more advanced
than his contemporaries such as Democritus, who appears to have rejected or
ignored the advances made by the Pythagoreans. See D. R. Dicks, Early
Greek Astronomy to Aristotle (London 1970) 80 81. Thanks to Carl Huffman
for suggesting this additional example.
istotles reports about them), it invokes a contrast between limit and un-
limited, and leads to a classification of numbers that spills out into reality
more widely.7 For the second generation Pythagoreans, the unlimited is
evidently part of what makes number the basis of the whole of reality.
This then gets taken up into some of the more peculiar bits of Platonic
theory, in particular the so-called unwritten doctrines and the idea of
the one and the indefinite dyad.8
In our modern accounts of Pythagoreanism, we take the Pythagor-
eans to be talking about funny stuff, some kind of weird mathematical
nonsense, when they mention the unlimited or indefinite, and we dis-
miss it as so much garbage, despite the fact that some respected interpre-
tations take Philolauss apeira to be kinds of material stuff or qualities
rather similar to those that appear in other cosmologies of the time.9
Yet when we meet the apeiron in Anaximander we are not so quick
to diagnose something weird or mathematical. We treat it very different-
ly. In Anaximanders case we do not see a primarily mathematical no-
tion being used to create a metaphysical basis for reality, but we read
Anaximander as speaking of an unlimited material: physical stuff. We
read him as a materialist, and for that reason (it seems) the apeiron is de-
fused. It doesnt smack of funny stuff, as it does in the hands of the Py-
thagoreans. It looks instead like a primitive kind of prime matter, some-
thing that Aristotle could look back to, and in which he could trace the
origins of the material cause:
For all things are either an origin or derivative from an origin, but of the
apeiron there is no origin, but it seems to be the origin of the rest and
to encompass all things and control all things, as all those say who do
not make any other causes apart from the apeiron such as mind or friendship.
7 The terminology is widespread, particularly in material from Philolaus. See Phi-

lolaus B 1, 2, 3, 6 etc, and Aristot. Meta. A ch. 5.
8 Aristot. Meta. 987b18 988a1. Details discussed in J. N. Findlay, Plato; The
Written and Unwritten Doctrines (London 1974) ch. 2.
9 Nothing in the extant fragments succeeds in making clear exactly what Philo-
laus has in mind when he refers to limiters and unlimiteds. Jonathan Barnes
(above, n. 2), is among the readers who take the unlimiteds to be material stuffs.
Carl Huffman, Philolaus of Croton, Cambridge 1993, discusses these interpreta-
tions, and mines the fragments for hints (pp. 37 53), and suggests (on the basis
of a fragment from Aristotles lost work on the Pythagoreans) that Philolauss
unlimiteds included time, void, fire and breath; but Aristotles wording hardly
makes that reading secure.
8 Catherine Rowett
And this is what is divine, for it is deathless and indestructible, as Anax-

imander says, and most of the natural philosophers.
Aristotle Physics 203b6 7, 10 15.
But what really is the difference between the notion of apeiron in Anax-
imander and the notion of apeiron in the second generation of Pythagor-
ean thought? Why do we take the one as a great advance in science, and
the other as a kind of metaphysical mistake?
Secondly, we should notice that Anaximander has a universe struc-
tured with concentric circles, which is invoked to explain the positions
and apparent movements of the heavenly bodies relative to the earth.
We do not hear the music of the spheres in Anaximanders universe,
but it is surely improbable that they do not utter sounds, for the circles
that carry the stars have flute-like pipes with breathing holes through
which the fire bursts forth when they are not blocked up.10 It seems cer-
tain that the pressurised fiery vapour escaping from these 1jpmoia_ must
make sounds or notes that reflect the size and diameter of the pipe, rath-
er like the sound of huge pan pipes played across the dark and misty
heavens. How are their notes related? Harmonically or out of tune?
Who can say? But we do knowroughly speakingthat Anaximander
posited some numbers which claim to be the sizes (i. e. diameter) or dis-
tances (i. e. radius) of the circles of the sun and the moon and the stars.11
The numbers seem to form a pattern, a sequence of a geometrical kind,
probably 9, 18, 27, with the earth too having the numerical proportion
3:1 between its height and its diameter.12 In fact, it seems that Anax-
imander was applying a sort of geometrical thinking to his speculations
about the shape and the movements of the heavens.13 Given that all
10 Hippol. Ref 1.6.

11 Hippol. Ref 1.6.
12 The evidence for these numbers in the doxography, and the reconstruction of
the mutilated texts, are discussed by D. OBrien, Anaximanders measure-
ments, CQ 17 (1967) 423 32, and G. Naddaf, Anaximanders measurements
revisited, in Anthony Preus (ed.), Before Plato (Albany 2001) 5 23. See also C.
Kahn, Anaximander and the origins of Greek Cosmology (New York 1960) 61 3,
KRS 133 137, D. L. Couprie, The Discovery of space: Anaximanders as-
tronomy, in id., Robert Hahn, and Gerard Naddaf (eds.), Anaximander in Con-
text (Albany, 2003) 165 254.
13 This is widely agreed. Kahn suggests that the inspiration for Anaximanders
numbers was mathematical rather than mystical; cf. Kahn (above, n. 12) 96
97). A possible link with the geometrical calculations used in architecture has
been explored in detail by R. Hahn, Anaximander and the Architects (Albany
2001); and id., Proportions and numbers in Anaximander and early Greek
must be aware that larger and longer flutes make deeper notes, it would
be surprising if these numerical ratios did not also yield a sense that
lower pitch sounds would issue from the vents of the larger wheels at
the outer spheres of the cosmos.
When Pythagoras, or some Pythagoreans, posit harmonic propor-
tions, based on ratios of numbers, as the explanatory principles in deter-
mining the structures of the heavens, and do this on the basis of discov-
ering that physical sounds can indeed be analysed mathematically, as
systematically related to the size and shape of the physical object that
produces them, we generally dismiss their suggestions as idle specula-
tion.14 When Anaximander engages in some similar, though less well
thought, in D. L. Couprie, Robert Hahn, and Gerard Naddaf, Anaximander in

Context (Albany 2003) 73 163. Much effort has gone into trying to show that,
although the work was speculative rather than experimental, Anaximander was
not just engaged in arithmology. See also W. A. Heidel, The Pythagoreans and
Greek mathematics, in David Furley and R. E. Allen (eds.), Studies in Presocratic
Philosophy (London 1970) 379 . The alternative line that this is a rationalisation
of some mythic or poetic materials is proposed by M. L. West, Early Greek Phi-
losophy and the Orient (Oxford 1971), 94; D. L. Couprie, Anaximanders dis-
covery of space, in A. Preus (ed.), Before Plato (Albany 2001) 23 48, 40 41.
14 It is hard to find references that explicitly present this attitude as strongly as
Barnes does in the passage cited above. Yet I think that the judgement is evi-
dent in the general approach to the study of the Presocratic philosophers and in
the extent to which Pythagorean number theory is marginalized in main-stream
collections of work on Presocratic philosophy. See the evidence presented by P.
Kingsley, Ancient Philosophy, Mystery and Magic: Empedocles and Pythagorean Tra-
dition (Oxford 1995) 317 320, showing how a tradition in intellectual history
has tried to cleanse the early Pythagoreans of the mystical, by implying that it
was a feature of a decadent, late, pseudo-Pythagoreanism, not the true philo-
sophical period of Pythagoras and his early followers. That is not quite the pat-
tern I am seeking to highlight here, but is related. See also G. E. R. Lloyd, Early
Greek Science: Thales to Aristotle (London 1970) 26 7, Secondly many of the
resemblances that the Pythagoreans claimed to find between things and num-
bers were quite fantastic and arbitrary. Obviously while the search for nu-
merical ratios proved fruitful in such fields as the analysis of musical harmonies,
and mathematics itself, it also and more often led to mumbo-jumbo and crude
number mysticism; and D. Furley, The Greek Cosmologists (Cambridge 1987)
58, who dismisses some parts of Pythagorean astronomy as fantasy, though
he does this in the service of a more positive assessment of their particular em-
phasis on form and structure. See Huffman (above, n. 9) 271 2, for a survey of
the bizarre ideas about the inhabitants of the moon, attributed to supposedly
more rational philosophers, that are quietly ignored in modern histories
(while those attributed to Philolaus are taken to discredit the whole of his as-
tronomy as mere fantasy).
10 Catherine Rowett
grounded, fantasy about the relative sizes of the heavenly rings, he is

hailed as a pioneer and an impressive forerunner of mathematical astron-
omy.15 Of course, Anaximander lived a little earlier than Pythagoras, in
the early sixth century rather than the second half of it. But really there
is not a lot of difference in the dates, and we might suppose that by ap-
pealing to harmonic structures, Pythagoras was attempting to put more
plausibility and accuracy into the speculative arithmetic, on the basis of
his discoveries that related audible harmony to numerical proportions
and the sizes of physical objects,16 in place of whatever obscure patterns
lay behind Anaximanders guesswork.
So it seems that when Anaximander indulges in pure numerical fan-
tasy, plucking multiples of 3 and 9 apparently out of the air, he is hailed
as the pioneer who saw for the first time that the plausibility of ones
cosmological theory can be enhanced by showing that it makes mathe-
matical sense. When Pythagoras, or early Pythagoreans, do a more so-
phisticated version of the same thing, selecting harmonic and geometri-
cal sequences in preference to arbitrary patterns, they are accused of
superstitious and fanciful numerological speculation. And yet the reason
for positing harmonic ratios in nature is that harmonic ratios are found
in nature, and are perceptible by us because we are naturally attuned, so
that we find such ratios beautiful, when they occur.
Certainly, to suggest that the heavens manifest a harmonic structure
which we find beautiful is not to engage in empirical science of quite the
sort we are used to. But if we are in the business of speculative astronomy,
rather than empirical astronomy, then Pythagoras (or whatever Pytha-
gorean invented this idea) has at least as good a grounding for his ap-
proach as Anaximander seems to have. And if empirical support is a vir-
15 The importance of this theory is that it is the first attempt at what we may
term a mechanical model of the heavenly bodies in Greek astronomy
Lloyd (above, n. 9) 17; His theory of equilibrium was a brilliant leap into
the realms of the mathematical a prioriKRS 134. Recent fashion has been
rather more low key in its estimate of Anaximander (e. g., The beginnings
of cosmology, in A. A. Long (ed.), The Cambridge Companion to Early Greek
Philosophy (Cambridge 1999) 45 65, 55).
16 See Xenocrates fr. 9 Heinze, apud Porphyry Commentary on Ptolemys Harmonics
30 1 6 Dring.
tue, at least Pythagoras can point to his work on harmony, which evi-
dently does have some empirical support.17
Why, then, do we cite Pythagorean harmony theory as a type of
worthless superstition, but cite Milesian science as an impressive fore-
runner of modern mathematical techniques in empirical science?
Could it be that the idea that something is beautiful, or that there should
be music in it, seems not to be a good reason for supposing it to be true?
But if that is so, we need to examine our preconceptions. For it seems,
first, that we are bringing to the enquiry a prejudice in favour of the idea
that nature is random, disordered or arbitrary, rather than systematic,
displaying patterns and orders at more than one level. Why should it
not be more likely that harmonic patterns figure in the structure of
the heavens? If so, the Pythagoreans have the better approach to the
task than, say, materialists such as the atomists.
Secondly, it seems plausible to suppose, as I have suggested, that
Anaximander too thought that the heavenly bodies uttered a flute-
like whistle. Perhaps he, too, was moved by that thought in composing
his theory about the sizes and shapes of the hoops that circle the earth.
So that even if we do, sadly, start from a post-enlightenment prejudice
in favour of seeing the world as random, meaningless and lacking in
beauty, still there seems to be some inconsistency in our preference
for the speculations of Anaximander over those of Pythagoras. Is that
just because we dont happen to have any texts on the music of the
spheresor rather wheelsin Anaximander?
Moving on from Anaximander to Heraclitus, let us ask a different
question, this time about logos. It has become customary in writing
about Heraclitus to leave the word logos untranslated even when writing
for a Greekless readership.18 Alternatively translators look for a standard
formulaic or non-committal translation (such as account or princi-
ple) 19 in order to avoid giving any specific meaning to the term in
17 Two key texts are Xenocrates fr. 9 Heinze, apud Porphyry Commentary on Ptole-
mys Harmonics 30 1 6 Dring; Aristoxenus fr. 77 Mller, in a scholiast on Plato
Phaedo 108d, Greene p.15.
18 R. Waterfield in the commentary in The First Philosopher, Oxford 2000; KRS;
R. McKirahan, Philosophy before Socrates, Indianapolis 1994.
19 Account in Barnes, Early Greek Philosophy, Harmondsworth 1987; T. M.
Robinson, Heraclitus; Fragments,Toronto 1987; and Kahn, H.; principle in
R. Waterfield in The First Philosophers, Oxford 2000. See also Long in this vol-
ume at n. 19.
12 Catherine Rowett
any particular occurrence.20 These high-minded practices have an unfore-

seen consequence, it seems to me. First they make the term logos stand out
as something like a key concept, inviting us to think that Heraclitus has a
theory of the Logos. This is reinforced by the habit of adding the def-
inite article in the English translation. The definite article is there in some,
but not all, of Heraclitus references to logos,21 but it would not always be
there, typically, in English even where it is normal in Greek (for instance
if b k|cor were translated by a term such as discourse, language, rea-
son, proportion, rationality or logic, we would not put in a def-
inite article at all, on any occasion of use, even if it were there in the
Greek). But because we do not offer one of these translations, but
leave it as logos, and because we therefore add the on every occasion
even where it is missing in the Greek, we produce an unanticipated ef-
fect: the Logos becomes almost imperceptibly hypostasized, until before
we have even observed it happening, we find it occupying, for us, a
place that looks plausibly god-like, and at that point we are sorely tempted
to identify it with the god of B 67. What might well have looked like an
immanent pattern in the behaviour of the world, if it had been properly
translated, now takes on a metaphysical role instead, as a divine entity that
explains or dictates the reciprocal patterns in the world. This seems to
happen, at least in part, because our increasingly entrenched translation
practices, and exegetical practices, irresistibly privilege the term logos
and exalt it to become a term of art.
Aside from the damage that this does to our understanding of Her-
aclitus, it also has a strange effect on our understanding of the relation
between Heraclitus and Pythagoras. Since we now think of Logos as a
kind of god in Heraclitus system, we fail to notice how close is the re-
semblance between Heraclitus interest in proportion and ratios and the
same topics in Pythagoreanism. The failure to translate logos, the adoption
of a special systematic pseudo-translation, and the addition of the defi-
nite article and capital letter as though the logos were an hypostasis or
divinity, mask the links between the notion of logos and the notion of
harmony in Heraclitus thought. Harmony (as in the harmony of op-
20 The motives for both practices are, of course, admirable in their way, in so far as
the translator tries to avoid imposing an interpretation by rendering the term
one way rather than another, or concealing the same term under unrecognisable
variant translations. I am not suggesting that there is a better solution, but rather
that translation is inherently unsatisfactory.
21 Logos occurs with the definite article in B 1, 2, 31b, 50; without in B 39, 45, 72,
87, 108, 115.
posites) is also seen as a key concept in Heraclitus thought, as it is in

Pythagorean thought. So both Pythagorean thought and Heraclitean
thought are constantly playing with the twin notions of ratio and har-
mony, and using these as their main explanatory concepts in natural phi-
losophy. The Pythagorean resonances in Heraclitus would, of course, be
much more apparent if Heraclitus logos were read as ratio or pro-
portion, and the numerical significance of harmony were allowed to
emerge in the context of both opposition (in Heraclitus B 8), of ratio
(in B 49, 79, 82 3), and of measure, and limit, and the unlimited
(in B 120, 94, 30,31 and 45). It is true that Heraclitus uses a wider
range of terms for limit and boundaryincluding owqor (120), t]q-
lata (120), l]tqom (94), as well as the pe_qata of 45so that even in
the Greek we are not alerted very strongly to the link between these
various fragments that are in several ways obsessed with measure and
limits. By contrast, in Pythagorean sources we tend to find a kind of
technical terminologythe %peiqa ja peqa_momta of Philolaus, for in-
stance, and the p]qar ja %peiqom of the first pair in Aristotles table of
Pythagorean opposites. This draws attention to the theme in the Pytha-
goreans more prominently than the variatio in the Heraclitean vocabu-
lary does.22 Still it remains undeniable that Heraclitus uses the idea of
logos or proportion to bring order to the measured processes of the
world, and that he draws connections between patterns of opposition
and the idea of a cosmic harmony (hidden or otherwise).
So while the Pythagoreans attempts to put numbers on the patterns
and proportions that they saw in the cosmos, and to link those numbers
to geometrical and harmonic proportions, are easily dismissed as puerile
fantasies by Barnes (and not just Barnes),23 Heraclitus mysterious logos is
given a definite article, a capital letter, and hailed as an attempt to bring
reason and order into a world of opposition and strife.24 What is the dif-
22 The impression that the Pythagoreans have a systematic technical terminology

may be exaggerated because of the prominence of Philolaus as a prime source
for Presocratic Pythagoreanism.
23 I have used Barnes because he represents an extreme end of a certain kind of
Anglo-Saxon tradition. Not all historians of Presocratic philosophy display
such antagonism towards the Pythagorean school, so that my generalisations
should be taken to have distinguished exceptions. I am happy for the reader
to identify with me or with my opponents as he/she feels most at home.
24 Barnes (above, n. 2) 59, is careful to warn us against taking the logos as a tech-
nical term and the key to Heraclituss secrets, but he goes on (80 81) to allow
that there may be a metaphysical logos doctrine, and to recommend that Heracli-
14 Catherine Rowett
ference? Is it that Heraclitus does not give us the numbers but only hints
tantalisingly at the existence of measures and ratios (the sea returns to
the same measure as was there before it became earth, B 31)? And is it
because the Pythagoreans tell us the numbers (that the number of the
heavenly bodies is ten, for example), but we suspect that their choice
of numbers is motivated by the desire for perfection, not the desire to
save the phenomena? And of course, we ourselves would never counte-
nance a practice of adjusting the observational results to comply with the
predictions of mathematics and theory.
Or is it that (by tradition) we systematically translate out, or interpret
out, the references to harmony and ratio in Heraclitus, so that we dont
see them as number mysticism, because instead we interpret Heraclitus
interest in logos as a kind of monotheismsomething in the tradition of
Xenophanes in which a Zeus-like divinity is rationalised and demythol-
ogised to become an immanent physical regulatory principle in the
world, a guiding principle that sees to it that we dont need to appeal
to weird metaphysical structures or mystical number patterns. It seems
that we try to see in Heraclitus a step on the route from Milesian ma-
terialism to post-enlightenment physics, and we therefore let him
have his logos in the guise of personified reasonexalted, but not tran-
scendent, much as Cartesian theism accounts for the existence and or-
derly functioning of a mathematically regular cosmos by assigning cer-
tain roles to a kind of god figure that is really not an object of cult
or mystery at all.
It is salutary to take a look at what Sextus Empiricus has to say. He
provides a lengthy analysis of the role of reason as a criterion of knowl-
edge in the Presocratic philosophers, an analysis heavily coloured by the
interests and concerns of Hellenistic epistemology. Having dealt with
Anaxagoras (the most physical of the Presocratics) 25 he introduces
the passage on the Pythagoreans thus:
ste b lm )manac|qar joim_r tm k|com 5vg jqit^qiom eWmai oR d Puha-
coqijo tm k|com l]m vasim, oq joim_r d], tm d !p t_m lahgl\tym
peqicim|lemom, jah\peq 5kece ja b Vik|kaor, heyqgtij|m te emta t/r
t_m fkym v}seyr 5weim tim succ]meiam pqr ta}tgm, 1pe_peq rp toO
blo_ou t floiom jatakalb\meshai p]vujem Gm d !qw t/r t_m fkym
tus be placed with the Milesian rational tradition, and sheltered from the pejo-
rative term mystic (which is doubtless reserved for those who fall into the Py-
thagorean mire).
25 Sextus Empiricus Adv. Math. 7.90.
rpost\seyr !qihl|r di ja b jqitr t_m p\mtym k|cor oqj !l]towor m

t/r to}tou dum\leyr jako?to #m !qihl|r.26
Sextus Empiricus Adv Math 7.92 3 (= DK 44 A29) 27
Sextus goes on to provide six or seven further paragraphs of evidence in
support of the claim that mathematical reason is the criterion for the Py-
thagoreans before proceeding to apply the same kind of treatment to
Xenophanes, Parmenides, Empedocles, Democritus and Heraclitus.28
It is no surprise (to us, at least) that Heraclitus too is said to make the
logos a criterion of truth.29 In Heraclitus case it is not cashed out as
mathematical reasonnaturally enough, since that is supposed to be dis-
tinctively Pythagorean. Instead Sextus cites the sequence of fragments
that we standardly use as the prime evidence for the Logos doctrine
in Heraclitus.30 So what do we do when we use this material? We ac-
cept, from this passage, Sextuss account of Heraclitus and his notion
of the logosthis survives almost unscathed into our interpretation of
several major and crucial moves in Presocratic thinking; but we com-
pletely ignore the bits about logos in Pythagoreanism, and its basis in
number and the idea that like is known by like that leads to the idea
that if the universe is numerically ordered, then our understanding of
it will be similarly ordered as a mathematical kind of science.31
That bitthe appeal to a specifically mathematical type of calcula-
tionis not exactly what we find in Heraclitus, although there are (as we
observed) hints of a kind of thinking that invokes proportions and ratios
26 So that Anaxagoras said that the logos generally was the criterion. The Pythagor-
eans also say that it is the logosbut this time not the logos generally but the logos
that is acquired from studies (mathematics?)just as Philolaus also saidand
that given that it contemplates the nature of the universe, it has a certain affinity
with that nature, if like is by nature grasped by like But number was the prin-
ciple of the structure of the universe; hence the logos that is the judge of all
things is not devoid of this power, and would therefore be called number.
27 Text as in Huffman (above, n. 9) 199.
28 S.E. AM 7 92 140.
29 S.E. AM 7.126.
30 That is, B 1 and 2.
31 In defence of this selective use of Sextuss material, one might appeal to the fact
that we do have genuine fragments of Heraclitus in which he uses the term
logos, whereas there is no textual support for attributing that term in that
sense to Philolaus. But we should notice that Sextus is talking in terms that
are alien to the Presocratic discourse throughout, and this applies as much to
his search for a criterion of truth in Heraclitus as it does in Philolaus. On the
anachronism of the terms of the enquiry see Huffman (above, n. 9) 199 201.
16 Catherine Rowett
in Heraclitus too.32 But there is no reason to think that the addition of

mathematics as a criterion of sound understanding of the world is a de-
velopment that we should dismiss as mystical mumbo-jumbo, or despise
as a fairy story, by comparison with the rather vaguer and more general
notion of logos in Heraclitus. On the contrary, we should probably
agree, nowadays, that cosmology requires not just a generic brand of
rational enquiry but more specifically a mathematically trained investiga-
tor, whose criterion for whether he has reached something worthy to
count as knowledge will be whether the mathematics works. So why should
we think of Heraclitus as a model of Presocratic philosophy at its best,
while dismissing Pythagorean theory as an embarrassing disfigurement
of an otherwise pure stream of increasingly rational investigation? Surely
it should be the other way around?
Let me offer a proposal for what lies behind this widespread prefer-
ence for the Heraclitean over the Pythagorean harmony theory. It is
this. It seems to me that Barnes (as well as others who share his judge-
ments) is following a tradition that prefers signs of materialism and re-
ductionism over any kind of metaphysical or teleological picture. It is
a tradition that sees the pure materialist reductionism of the atomists
as the culmination and high point of the Presocratic achievement, and
it assesses the contribution of earlier thinkers by how closely they ap-
proximate to that idealan ideal that is seen as a kind of no-nonsense
physics, even if it has little ambition to provide genuine empirical sup-
port for its speculations.33
It is true that I have suggested that Heraclitus logos gets hypostasized
as The Logos with a capital letter, and in the process takes on a quasi-
god-like role as the governor of cosmic processes. That might suggest that
our admiration for Heraclitus is not because we see him as eliminating
metaphysical and religious entities. But, as I suggested above, despite
the theistic terminology, we tend to conceive of that move as somewhat
reductionist, like Xenophaness theological endeavours. On that reading
32 E.g., B 30, 31, 79, 82, 83.

33 Barnes himself (above, n. 2) 343 4, 76 7, is careful to warn us against over-
enthusiastically assimilating ancient atomism to modern science, and he points
out many ways in which the atomism of Leucippus and Democritus raise prob-
lems that they cannot answer. But these warnings leave untouched the general
sense that the atomists theory is virtuous just in so far as it approximates to the
ideal of modern scienceso that ancient atomism is not quite admirable because it
does not quite live up to that ideal in all respects.
of Heraclitus, God or The Logos, just is the world, when alls said and
done:
God is day, night, winter, summer, war, peace, hunger satiety34
God is the processes that once seemed mysterious; but really (according
to this version) they are not so much mysterious as regular, not unpre-
dictable but reasonable. Reason, not religion, is the way to get control
of the physical events, we understand. That is what we take Heraclitus
to be saying, in his rather obscure and difficult way. So God turns out to
be no more than our picture of how the world works to a predictable
pattern. Harmony theory is then read not as a metaphysical thesis but as
a materialist one.
Conceivably this is correct as an interpretation of Heraclitus. In sup-
port, one might notice the way in which Heraclitus appeals to the no-
tion of measure in connection with specific physical processes that man-
ifest this kind of proportionalitythe notion that we find this logos in
the quantity of sea that you get back when earth has been turned
back into water (B 31), and that there are observable limits to the pas-
sage of the sun across the sky.35 It seems that, no matter how little evi-
dence Heraclitus gives for observing such regularities in nature, he does
mean that the regularities in question are part of nature. 36 Similarly, to
return to Anaximanders picture of the cosmos, those heavenly pipes
seem to be substantial and very concrete material chariot wheels revolv-
ing in the sky, invisible only because they cant be seen for the mist.
Anaximander too is giving us material explanations with numbers on
them, not numerical explanations with no matter to do the work.
34 Heraclitus B 67.
35 This is one traditional interpretation of the claim in B 94 that the sun will not
overstep its measures. The Derveni papyrus (which appears to combine what
we used to know as B 3 and 94) opens the possibility that the measures are the
suns size (a foot across, B 3) rather than its tropics. See G. Betegh, The Derveni
Papyrus: Cosmology, Theology and Interpretation (Cambridge 2004) 10 11. This
alternative is also compatible with the traditional reductionist interpretation
that I am sketching.
36 Here perhaps we assume too readily that the image of the Erinyes in B 3 (oth-
erwise the Furies, ministers of Justice, will find it out) is just a picturesque
metaphor to convey what we take to be a natural constraint on the behaviour
of the sun. Again the Derveni papyrus, with its obsession with the daimonic,
tells us that Heraclitus was not always read so rationalistically in antiquity (see
previous note).
18 Catherine Rowett
Pythagorean mathematics, by contrast, tries to make numbers do the

work. If youre looking for an account of material and efficient causes in
the cosmos, its odd to point to numbers as such, as opposed to applying
numbers to quantities of other things, quantities of some material stuffs
or physical forces that could do the work. So if you conceive the Pre-
socratic project as a project to suggest and improve explanatory factors
that are to be invoked in the interests of a reductionist thesis about how
the world works, numbers as such seem to be the wrong kind of thing.37
This post-enlightenment complaint is closely related to an objection
made long since by Aristotle in Metaphysics N:
oR d Puhac|qeioi di t bqm pokk t_m !qihl_m p\hg rp\qwomta to?r
aQshgto?r s~lasim, eWmai lm !qihlor 1po_gsam t emta, oq wyqistor d],
!kk 1n !qihl_m t emta di t_ d]; fti t p\hg t t_m !qihl_m 1m "qlom_
rp\qwei ja 1m t` oqqam` ja 1m pokko?r %kkoir. 1090a20 25
oR lm owm Puhac|qeioi jat lm t toioOtom oqhem 5mowo_ eQsim, jat l]m-
toi t poie?m 1n !qihl_m t vusij s~lata, 1j l 1w|mtym b\qor lgd jou-
v|tgta 5womta jouv|tgta ja b\qor, 1o_jasi peq %kkou oqqamoO k]ceim
ja syl\tym !kk oq t_m aQshgt_m.38 1090a30 35
Aristotle exonerates the Pythagoreans from any charges to the effect that
they treat numbers as separate from sensible things or as intermediate be-
tween sensible things and forms.39 But this is only because he takes it
that they think that sensible things are numbers, which he thinks is, if
anything, an even more peculiar idea, albeit one that avoids the problem
of duplication of Platonic entities. Aristotle suggests that the Pythagor-
eans adopted this theory because they perceived that there were p\hg of
numbers, such as harmonies, ratios and the like, which can be seen as
numerical characteristics of things, and that these show up all over the
place in astronomy and the other aspects of nature. Seeing those math-
37 See Furley (above, n. 14) 52 3 for a brief discussion of this thought.

38 But the Pythagoreans made things be numbers (because they saw many numer-
ical effects existing in perceptible bodies) but not separate numbers, but things
being constituted of numbers. But why so? Because the numerical effects exist
in harmony in the heavens and in many other things. On the one hand the
Pythagoreans seem not to be liable to any charge on this matter (sc. separating
the mathematicals). But in respect of making physical bodies out of numbers,
making things that have lightness and heaviness out of things that have no heav-
iness or lightness, they seem to be talking about a different heaven and different
bodies, but not the perceptible ones.
39 The Platonists and Speusippus are under attack for a variety of modifications of
the idea of numbers that are separate from aistheta.
ematical phenomena, he supposes, made the Pythagoreans go for the

idea that things actually are numbers.
Although the tendency to exclude Pythagorean speculations from the
serious history of Presocratic philosophy comes from a tradition that is
heavily indebted to Aristotle, and to Aristotles reconstruction of early
cosmology, it does not seem to me that the modern objections to Pytha-
gorean numerology reflect quite the same motivation as Aristotles be-
mused comments in Metaphysics N. Aristotle suggests that numbers
arent the right kind of thing to be the constitutive substance of things
that have mass. Numbers dont have heaviness and lightness, sohe com-
plains you cant make things out of them. But Aristotle does not simply
dismiss the Pythagoreans as stupid or muddled. Rather, he concludes that
in some way they would have to be talking about something else, if their
thoughts are to make any sense. They cant really be talking about things
that have weight and so on at all. They seem to be talking about a differ-
ent heaven and different bodies, but not the perceptible ones.40
Aristotles observations about whether people separate numbers are
pertinent here. In the case of Plato, he says, the theory allows that there
are things that have physical mass, on the one hand, and then, on the
other hand, there are !qihlo_ that are wyqisto_ a separate realm of
things that are not in the same category as the physical things that man-
ifest the numerical effects. It is a two-world view. But among the Py-
thagoreans he does not find a two-world view. They neither separate
numbers, nor make them intermediates.41 Instead he finds that they
have just one world, and since it is a world made of numbers, it should
be the incorporeal one, although it is also the world of perceptible
things. That is why Aristotle says that they seem not to be liable to
any charge of separating numbers from the perceptible things.42 For
there is just one set of things, namely the numbers (which nevertheless
in some way are the perceptible things).43
40 1090a34 5. Aristotle is bemused because they talk as though they are explain-
ing physical bodies, yet the elements they cite suggest that this cant be so. The
word seem in this sentence suggests that he thinks that to make sense of what
they are saying we need to adopt some hypothesis such as this.
41 Cf. Aristotl. Meta. A, 987b27 ja 5ti b lm tor !qihlor paq t aQshgt\, oR d(
!qihlor eWma_ vasim aqt t pq\clata, ja t lahglatij letan to}tym oq
tih]asim.
42 1090a30.
43 Aristotles comment at 1090a30 seems to conflict with the passage in Metaphy-
sics A (987b11) where he tells us that the Pythagoreans have a notion compa-
20 Catherine Rowett
So the Pythagoreans have just the one world, but it is apparently

composed of what we (at least) regard as incorporeals. The result is
that (as Aristotle observes) the Pythagoreans seem to be talking about
a different world and different bodies, 1090a34. If things are made of
numbers, then they are a very funny kind of things.44
So Aristotle resists making the sort of objection that one might ex-
pect from a scientist, complaining that it is nave and stupid to try to cre-
ate bodies out of numbers, or that this is the wrong kind of matter to do
the job. Instead, Aristotle seems to be saying that if their suggestion that
the material is numbers is to make any sense, their question must have
been different. They must have taken themselves to be explaining a
world of incorporeal perceptibles, not the world of corporeal percepti-
ble things that Plato was talking about, which he regarded as something
quite separate from numbers.
Nevertheless, Aristotle is evidently assuming from the start that the
Pythagorean project (including the appeal to numbers) was the same
project as the Ionian scientists project. He thinks that the question
was how to explain the perceptible world in terms of its constitutive
matter. Once the constitutive matter has been specified as non-separated
numbers, Aristotle concludes that the Pythagoreans had some funny idea
of the world. They were evidently explaining a rather incorporeal
world, with all its entities composed of numbers; but he does not
drop the idea that the Pythagoreans are to be assessed for their compe-
tence at reducing perceptible reality to explanatory components that are
immanent and not transcendent.45
rable to Platos notion of l]henir. It seems that at 987b11 Aristotle is assimilating

Plato and Pythagoreanisma tendency that was to have a long subsequent his-
tory by contrast with the careful attempt to draw distinctions between differ-
ent kinds of Platonism about numbers in the passages in Metaphysics N.
44 I have not questioned Aristotles claim that the Pythagoreans made things out
of numbers, because I am interested in trying to explain what his objection to
that thought is, and how it differs from our objections to the Pythagoreans use
of numbers in physics. It may well be that he was wrong to think that they
meant that numbers were a material cause, despite the fact that he surely had
access to the works of Philolaus and Archytas and had written extensively on
them. For a diagnosis of Aristotles mistake, and the evidence in Philolaus for
what he really meant, see Huffman (above, n. 9) 57 64.
45 The problem is surely only that Aristotle takes the reduction to be materialist in
its outlook. Without that assumption the project to reduce ontology to a system
of numbers is not self-evidently flawed. W. V. Quine entertains precisely this
project and investigates what it lacks, if anything, as a serious candidate in a
From Incomprehension to Admiration
I have suggested that there are striking similarities between the Pytha-
gorean concerns with number, ratio and harmony, and some of the ma-
terial that has been regarded as pioneering and profound in Anaximand-
er and Heraclitus. In asking why these moves should be considered
important and profound in the latter cases, I have suggested that they
meet with approval among modern scholars because the modern schol-
ars are assessing the Presocratic thinkers for their progress in a sequence
of developments in the direction of reductive materialist physics. Be-
cause they see the numerical patterns in the universe as immanent
p\hg of things, and do not presuppose a set of theoretical entities that
are numbers with properties of their own, the numerical fantasies of
these early Ionian thinkers are seen as an acceptableor even progres-
sive part of that generally materialist project.
It seems to me that this evaluation of the early Greek philosophers
displays an agenda that is built into our heritage of modern Presocratic
scholarship. The agenda is very evident in Barnes, but that is only be-
cause he is particularly blatant about expressing his prejudices in outspo-
ken terms. In practice he is following an existing tradition. One would
say that the tradition was Aristotelian in origin the rejection of fancy
metaphysical entities, the down-to-earth preference for specifying that
the world of particulars is what is real, the analysis of the Presocratics
as engaged in diagnosing the material cause, all these seem to be Aristo-
tles prejudices except that I think that it is a modern version of Ar-
istotelianism that has acquired a great deal of baggage from the Enlight-
enment, from logical positivism and from a more recent scientism that
equates truth with what can be proved by empirical methods. Aristotle
is certainly opposed to some of the things that he finds in Platonism,
such as the separation of Forms, but his opposition is not for quite
the same reasons as the reasons that modern Presocratic scholarship
would offer for why it doubts that Pythagorean numerology was a val-
uable contribution to Western Philosophys overall development.
number of works. See Quine, Ontological reduction and the world of num-
bers, in id., The Ways of Paradox and other essays (New York 1966), 199 207;
Ontological relativity, in id., Ontological Relativity and other Essays (New York
1969) 26 68; Propositional objects, in id., Ontological Relativity and other
Esays (New York 1969) 137 60. I am grateful to Nick Denyer for pointing
me to these references.
22 Catherine Rowett
In this second part of the paper I shall move beyond my initial

thought, that Pythagorean speculations are no worse than the comparable
bits of Heraclitus or Anaximander, if one is looking for empirically ver-
ifiable reasons in favour of a particular theory. My second thought is
more ambitious. I want to propose that there is something in the Pytha-
gorean enthusiasm for numbers that is far more significant philosophi-
cally, and has had far greater ramifications in the story of Western phi-
losophy than anything Anaximander or Heraclitus ever did, despite their
superior credibility in the contemporary western system of values. That
is, I am suggesting that we should not apologise for the Pythagoreans
tendency to idolise numbers, or certain particular numbers, nor should
we try to discard those bits and look for some cleaner bits of respectable
doctrine instead. Rather we should celebrate them.
A number of thinkers from Plutarch to Leibniz have anticipated my
point. We should start however with a well known passage attributed to
the fifth century Pythagorean Philolaus, which is quoted by Stobaeus.46
The point that is relevant to my topic is where Philolaus says that with-
out number it is not possible for anything to be thought or known:
ja p\mta ca lm t cicmysj|lema !qihlm 5womti oq cq oX|m te oqdm
oute mogh/lem oute cmysh/lem %meu to}ty.47
Philolaus B 4
The idea that having a number is a criterion of knowability is not one
that is often proposed or given prominence in epistemological discus-
sions, but it is worth comparing it to the point made by Parmenides
about the relation between being and knowability:
taqtm d( 1st moe?m te ja ovmejem 5sti m|gla.
oq cq %meu toO 1|mtor, 1m pevatisl]mom 1stim,
erq^seir t moe?m.48 Parmenides B8.34 6
It seems that Philolaus is giving to numbers a role very similar to the role
that is served by being and truth in Parmenides. There is no true think-
ing without being in Parmenides. There is no true thinking without
numbers in Philolaus. It is a very severe epistemology, in which nothing
46 Philolaus B 4 5; Stobaeus 1.21.7b-c.

47 And indeed all the things that are known have number. For without this it is
not possible for anything to be thought or known. Text from Huffman
(above, n. 9).
48 Its the same thing thinking and that whose thought it is. For you wont find
thinking without the reality, in which it is an utterance.
counts as knowing unless it is knowledge of numbers. There is only one

set of knowable objects, namely mathematicals, or at least numbered
items. So study of mathematics is not just one of the sciences, alongside
physics, but mathematics is the only science that relates effectively to
knowable objects.49
Secondly we may notice a quotation from Archytas (B 1) preserved
by Porphyry in his commentary on Ptolemys Harmonics, and the echo
of the same thought that is known from Platos Republic:
paqaje_shy d ja mOm t )qw}ta toO Puhacoqe_ou, ox l\kista ja cm^-
sia k]cetai eWmai t succq\llata. k]cei d( 1m t` Peq lahglatij/r eqhr
1maqw|lemor toO k|cou t\de Jak_r loi dojoOmti to peq t lah^lata
diacm~lem, ja oqhm %topom aqh_r aqtor oX\ 1mti peq 2j\stou vqom]m.
peq cq tr t_m fkym v}sior jak_r diacm|mter 5lekkom ja peq t_m jat
l]qor, oX\ 1mti, exeshai. peq_ te d tr t_m %stqym tawuttor ja 1pito-
km ja dus_ym paq]dyjam "l?m sav/ di\cmysim ja peq caletq_ar ja
!qihl_m ja oqw Fjista peq lysijr. taOta cq t lah^lata dojoOmti
eWlem !dekve\.50
Porphyry In Ptolem. Harm. 1.3, p. 56, Dring (quoting Archytas B 1)51
This passage confirms Platos claim, at Republic 530d, that the Pythagor-
eans called arithmetic, geometry, harmonics and astronomy sister sci-
ences, and it suggests that Archytas was the one who coined the phrase.
But my interest is not in that point, but in Archytass suggestion that one
49 See Huffman (above, n. 9) 173 7, responding to M. C Nussbaum, Eleatic con-

ventionalism and Philolaus on the conditions of thought, HSCP, 83 (1979) 63
108). I do not deny that Philolaus would probably allow that we can perceive
musical intervals and other relations without knowing the formulae, but I am
suggesting that when we know the formulae, the knowable things are the num-
bers, so that other things (ratios, harmonies etc) are knowable just in virtue of
being numerical, or having numbers (and their numbers are what we know
about them). This is to go somewhat beyond what is strictly justified by the text.
50 And now let us set alongside the words of Archytas the Pythagorean, to whom
the writings are most reliably attributed. He says in the work on mathematics,
right at the beginning, the following: The people who are versed in learned
subjects (mathematics?) seem to me to discern beautifully, and there is nothing
absurd in their thinking correctly about each of the things just what it is like;
for, since they discern beautifully with regard to the nature of the universe as
a whole, it is to be expected that they will observe beautifully about the partic-
ular things, just what they are like. They have handed down to us clear knowl-
edge concerning the speed of the stars and their risings and settings, and about
geometry and numbers, and not least about music. For these subjects seem to be
sister-subjects.
51 Text as in C. Huffman. Archytas of Tarentum: Pythagorean, Philosopher and Math-
ematician King, Cambridge 2005.
24 Catherine Rowett
would expect the experts in these sciences which are grouped together
because they work by theoretical manipulation of abstracted mathemat-
icals, not empirical data from physical bodies one would expect these
experts to be the ones who correctly discern the nature of things, and of
the universe as a whole. It is oqhm %topom, says Archytas, that these
people think correctly about things. But notice also the idea that they
do this jak_r they discern the workings of the universe beautifully.
And this surely links in to the idea that the workings of the universe
are themselves a fine object of attention.
That point is made more explicit by Plutarch in a passage in his
Quaestiones convivales:
psi lm owm to?r jakoul]moir lah^lasim, speq !stqab]si ja ke_oir
jat|ptqoir, 1lva_metai t/r t_m mogt_m !kghe_ar Uwmg ja eUdyka l\kista
d ceyletq_a jat tm Vik|kaom !qw ja lgtq|pokir owsa t_m %kkym
1pam\cei ja stq]vei tm di\moiam, oXom 1jjahaiqol]mgm ja !pokuol]mgm
!tq]la t/r aQsh^seyr. di ja Pk\tym aqtr 1l]lxato tor peq Eudonom
ja )qw}tam ja L]maiwlom eQr aqcamijr ja lgwamijr jatasjeur tm
toO steqeoO dipkasiaslm !p\ceim 1piweiqoOmtar, speq peiqyl]mour
d_wa k|cou d}o l]sar !m k|com, paqe_joi, kabe?m !p|kkushai cq
ovty ja diavhe_qeshai t ceyletq_ar !cahm awhir 1p t aQshgt pakim-
dqolo}sgr ja l veqol]mgr %my lgd( !mtikalbamol]mgr t_m !id_ym ja
!syl\tym eQj|mym, pqr aXspeq m b her !e he|r 1stim.
(Plat. Phaedr. 249c).52
Plutarch Quaestiones convivales 8.2.1, 718E (= DK 44 A 7a)
The thought developed by Plutarch in this passage is supposed to go
back to Philolaus in some sense,53 and indeed it is faintly reminiscent
of the passage from Sextus Empiricus which we noticed above,54
where Philolaus was said to have claimed that one gets an affinity
52 In all the so called (mathematical?) studies, the traces and images of the truth of
intelligible objects are reflected, as in even and polished mirrors; and most of all
geometry, according to Philolaus, being the source and mother-city of the
other studies, leads the mind up and converts it, like a mind purified and re-
leased effortlessly from perception. Hence Plato himself criticised the followers
of Eudoxus and Archytas and Menaechmus, who tried to divert the doubling of
the cube to instrumental and mechanical devices, as though they were trying to
obtain the two mean proportionals, however practicable, aside from rationality.
For this is to destroy and corrupt the good of geometry, when it is dragged back
to perceptible things and not carried up and not grasping eternal and bodiless
iconsthose things closeness to which makes god always be god.
53 The name Philolaus is not in the manuscript, but it is obtained by a plausible
emendation of a corrupt reading v_kaom in the manuscripts.
54 S.E. AM 7.92 (= DK 44 A 29).
with the harmony of the universe from assimilation to mathematical

knowledge. Here too, in Plutarchs passage, the thought attributed to
Philolausor built upon Philolauss foundations by Plutarchis that
handling numbers does something splendid for you.55 And for Plutarch,
the less empirical the science is, the better it is at this task. Plutarch sup-
ports this Platonist thought by reference to a legend according to which
Plato is supposed to have raised objections to the method for duplicating
the cube attributed (here) to Eudoxus, Archytas, and Menaichmus. The
details of the mathematics are not immediately relevant for now (except
perhaps to note that Archytas should be exempt from the criticism). 56
Plutarchs point is simply this: that we should not be attending to ma-
terial examples. We should be handling numbers. Thats how we get
to think of incorporeals, he thinks This is the Platonist thought:
that we need to get close to God by escaping from perceptible things.
Of course, Plutarch is feeding us Platonism served up on a bed of
Philolaus. Im not meaning to pretend that the importance or the signif-
icance of the Pythagoreans devotion to the elegance of numbers was
explicitly appreciated by them, at the time that they first developed
their fascination. On the contrary, I want to suggest that it is with hind-
sight that we can see that this was one of the most profound and lasting
legacies of Presocratic thoughtthe discovery of what we would call
incorporeals. It is a point made by Plutarch, and by Plato too of course,57
55 Huffman (above, n. 9) 193 194, takes the material from Philolaus to be very
brief, only the reference to geometry as the source and mother city, while all
the reflections on that are from Plutarchs Platonist context. On the other
hand it is not clear why Plutarch would be prompted to cite Philolaus at this
point if there were not some invitation to this line of thinking in the text to
which he is alluding. I think that Huffman is keen to exclude any hints of
proto-Platonism from Philolaus, and as a result he may be skimming the material
too hard.
56 There is something wrong with the story, though how exactly it has ended up
in this form in Plutarch is not entirely clear. In fact, it would appear that Archy-
tass solution to the problem of obtaining the two mean proportionals was more
theoretical and did not resort to practical methods as implied here. It makes no
sense to suggest that Archytas was one of the offenders against whom Plato
would have laid such a charge, therefore. The allusion does not seem to be
to any existing text of Plato, although the issue of the need for two mean pro-
portionals between cubic numbers figures in Platos Timaeus 31c-32b, and there
is a passage in the Republic 7.528a-d, which criticises stereometry for some fail-
ings that commentators have tried to link to the dispute mentioned by Plutarch.
See Huffman (above, n. 9) 344 401.
57 Plato Republic 527b.
26 Catherine Rowett
that learning mathematics and geometry helps us to turn our gaze up-
wards (meaning towards intellectual objects rather than corporeal or
perceptible ones), or to abstract the intelligible objects and intelligible
truths from material things. They are making a point about the useful-
ness of a particular kind of abstract discipline for intellectual exercise and
training. But in addition to that point, I would also want to add that
there is something important, even revolutionary, about the idea that
the most important reality might be the one that obeys the mathematical
rules, not the one that observably falls short, and approximates only
rather rarely to mathematical accuracy. In other words, Plato saw that
when mathematics talks about the p\hg that are numbers, it does not
treat them as p\hg of things (that is, of things that are more real than
their p\hg) but it treats the numbers as the most perfectly real things
(or at least the more perfectly real, as compared with bodily things).
This was an approach with which Aristotle was only partly out of
sympathy. It is true that he does not think that the Platonic separation
of mathematicals and forms from perceptible things is a good philosoph-
ical move. He prefers to think of the numbers as p\hg of bodies, and he
is pretty sceptical of the weird results when (as he supposes) the Pytha-
goreans imagine that things just are numbers or made out of numbers.
But the preference for the perfection of mathematics, and the location
of genuine truth and beauty in the intelligible world, is less alien to Ar-
istotles way of thinking than it is to the agenda with which modern
scholarship has approached the Pythagoreansthose prominent parts
of modern scholarship which have been responsible for relegating
them to the status of superstitious mystics.58
What do I mean? I mean that there is a seamless continuity between
Pythagorean awe at the perfect patterns in number, Parmenidess awe at
the eternity of Being, Platos awe at the Form of the Good, and Aris-
totles awe at the Unmoved Mover. All these are objects of love and ad-
miration, but their power is derived entirely from their beauty and per-
58 There are, of course, notable exceptions in modern scholarship on the Preso-

cratics. Most importantly perhaps, Carl Huffman, who has done much to
bring the contributions of neglected Pythagorean thinkers such as Philolaus
and Archytas to our attention in their own right, and show that they have seri-
ous philosophical and theoretical meat to offer. Others, including most prom-
inently Peter Kingsley, have sought to show why the religious and mystical side
of Pythagorean traditions needs to be taken seriously.
fection, not from any efficient or material causal efficacy.59 At the end of
the day, Aristotle too would locate the best and most perfect causal
power in the teleological cause, a cause discovered by abstract reasoning,
not by experimental science. The reason why Aristotle found the Pytha-
goreans puzzling was because he thought that they must be looking for
the material cause. He could not fit into his system a weird attempt to
explain bodies that have mass by appeal to entities that have no mass; but
his bemusement is created by the story that he has to tell about the de-
velopment of Presocratic philosophy. Presocratic phusikoi are (he thinks)
trying to answer the question Out of what, 1j t_mor, did the world
come? (or Out of what, 1j t_mor, is the world made?).60 Aristotle
glosses the stuff, out of which they suppose the world comes, as
their arche, and his account of his predecessors tends to be couched as
an analysis of their various attempts to cope with the logical complexity
of the notion of coming to be out of something, while supposing that the
something in question was logically playing the role of the material sub-
strate. So it makes sense that if he found the Pythagoreans saying that the
!qw^ is number (or saying what he took to be that claim), he would sup-
pose that something very weird was going on. But it was only weird be-
cause he was looking at their !qw^ as a material substrate. It is not at all
weird if you place it alongside the formal causes of the Platonists, or the
final cause that is so powerfully there in Aristotle. If, instead, you start by
wondering whether the Pythagoreans are talking about the explanatory
power of beauty, structure, form, and indeed teleology, in the universe,
the idea of appealing to patterns of numbers makes much more sense.61
If that was what they were doing, then their appeal to numbers as ex-
59 Some sources credit the Pythagoreans (Hippasus in particular) with the discov-
ery of irrational numbers, or particularly the incommensurability of the side and
diagonal in a square, and suggest that this was a challenge to their belief in the
mathematical perfection of the universe. (Porphyry Life of Pythagoras 246 7;
Clement Stromateis 5.58). But one might equally suppose that the fact that
the side and diagonal are commensurable when squared (effectively Pythagorass
theorem) would reveal a hidden rationality, a virtual rationality, in numbers that
were apparently irrational when treated as lines, and restore ones faith in the
idea of a mathematically coherent universe.
60 Aristot. Meta. 983b6 11; cf. Phys. 187a12 26.
61 Indeed there is evidence that Aristotle was partially aware of this alternative
construal, as for instance in his comments at Metaphysics N 1092b8, where
he admits that it is not clear in what sense numbers are explanatory of being,
and suggests (as the second alternative) that it is because harmony is (explained
as) a ratio of numbers and that this idea extends to explanations of other things.
28 Catherine Rowett
planatory principles of the universe is not just methodologically sound,

but also extremely perceptive.
So I suggest that Aristotles incomprehension might be created by
his own agenda, which informs his investigation and presentation of
the pre-Aristotelian history of ideas,in other words, by that story
that he tries to tell, of an investigation solely into material causes in
the Presocratic period but it does not reflect any ideological antipathy
on his part to the possibility that there might be other worthy causes to
investigate besides the material cause.62 In modern thought the lack of
appreciation for this pioneering work on numbers, which paved the
way for all the most influential thinking from Plato to the Renaissance,63
is due to something worse. It is due, I think, to a fundamental prejudice
in favour of reductionism, materialism, and mechanistic conceptions of
the physical world. It seems to me that, unlike Aristotle who thought
that the Presocratics were primitive in trying to explain things by refer-
ence to matter alone, contemporary scholars assume instead that they are
to be congratulated for exactly that. They think that progress was manifest-
ed in the Presocratics move towards the more and more mechanistic
theories of Anaxagoras and Democritus. They are embarrassed by Empe-
docles, and they often try to rescue him, by giving him one physical
poem in which mechanistic forces explain everything and there are
no appeals to immaterial values, so that they can conveniently ignore
the other bits which dont fit that ideal. And they find themselves
quite unable to stomach the Pythagoreans, when they discover that
even abstract maths is imbued with ethical and religious value. Math-
ematicians and philosophers just shouldnt be talking about numbers with
that kind of language, I hear them say. Its superstitious. It worries us.
The reason why we have come to think like that is because we have
been brought up with an agenda that is more ideologically blinkered
62 My diagnosis of the source of Aristotles incomprehension differs somewhat

from Huffman (above, n. 9) 57 64. This is largely because Huffman thinks
that Philolaus was investigating the corporeal world in the terms that Aristotle
envisages (so Aristotles mistake was in confusing the claim that things come
from limiters and unlimiteds with the claim that they come from numbers). I
am suggesting instead that Philolaus might indeednay surely didassign a
role to numbers in explanation, but not as a material explanation.
63 It would be good to say something about Plato, and particularly about those bits
of Plato that have been taken to be somewhat Pythagorean in inspiration (in-
cluding parts of the Phaedo), but that will have to wait for a more substantial
opportunity to treat it in its own right.
than Aristotles. A deep set ideological preference for mechanistic theory

prevents us from seeing that one mightone shouldwant to explain
what is beautiful and awe-inspiring about the world; and that the ex-
planation we give for beauty and wonder should not explain it away,
in such a way that there is no awe and no beauty to move us after all.
To encapsulate the point I want to make, and to remind ourselves
that the prejudice against the Pythagoreans is not universal, we should
look at a passage from Leibniz.
It is rightly said in the paper given to the Princess of Wales, and which Her
Royal Highness was kind enough to send to me, that next to vicious pas-
sions, the principles of the Materialists contribute much to support impiety.
But I do not think that there were grounds to add that the Mathematical Prin-
ciples of Philosophy are opposed to those of the Materialists. On the contrary,
they are the same except that the Materialists, following in the footsteps
of Democritus, Epicurus, and Hobbes, restrict themselves to mathematical
principles alone, and admit nothing but bodies, whereas the Christian
Mathematicians also admit immaterial substances. Thus it is not Mathemat-
ical Principles (in the ordinary sense of the term) but Metaphysical Principles
which ought to be opposed to those of the Materialists. Pythagoras, Plato,
and to some extent Aristotle, had some knowledge of these, but I claim to
have established them demonstratively, although in my Theodicy this was
done in a popular manner.64
There is a distinction to be made between the materialist way of doing
numbers, which uses numbers to give exact accounts of the behaviour
of bodies, and the metaphysical move, which admits immaterial substan-
ces. What the Pythagoreans give us, according to Leibniz, is the meta-
physical, which is, I would claim, the origin of real philosophy. It wasnt
Anaximander who started us moving towards real philosophy, nor De-
mocritus, nor even perhaps Heraclitus. It was Pythagoras, and he did it
when he told us to take our oaths by the tetraktys, and that justice is the
number 4 and kairos is the number 7.65
Dont get me wrong. I am not trying to say that metaphysical the-
ories are right, or that a metaphysical theory is better (as a theory) than a
materialist one. I am just saying that we would not have had a history of
western philosophy if there hadnt been Pythagoreans and Platonists:
that the metaphysical turn (initiated or preceded by the numerical
64 Gottfried Wilhelm Leibniz: Correspondence with Clarke: Leibnizs second paper,

in G. W. Leibniz, ed. C. I. Gerhardt Die Philosophischen Schriften von Gottfried
Wilhelm Leibniz (Berlin: Weidmann, 1890) vol. 7, p. 355. I am grateful to
Lloyd Strickland who kindly supplied this new translation for me.
65 Alexander In Meta. 38.10.
30 Catherine Rowett
turn) is what is distinctive, and that is what first sets the debate about the
nature of reality and the status of the perceptible world going. Parme-
nides does this, and the Pythagoreans do it, but the materialists
dont.66 And that is one reason why one might want to say that philos-
ophy started in southern Italy, not on the coast of Turkey, and that Par-
menides was, after all, an honorary Pythagorean.67
Bibliography
Algra, Keimpe. The beginnings of cosmology, in A. A. Long (ed.), The Cam-
bridge Companion to Early Greek Philosophy (Cambridge 1999) 45 65.
Barnes, Jonathan. The Presocratic Philosophers (2nd edn.; London: Routledge and
Kegan Paul, 1982).
Early Greek Philosophy. Harmondsworth 1987.
Betegh, Gbor. The Derveni Papyrus: Cosmology, Theology and Interpretation.
Cambridge 2004.
Burkert, Walter. Lore and Science in Ancient Pythagoreanism, trans. E.L. Minar Jr.
Cambridge, Mass. 1972.
Cornford, Francis MacDonald. Mysticism and science in the Pythagorean Tra-
dition, in Alexander P. D. Mourelatos (ed.), The Presocratics. (Garden City
1974) 135 60.
Couprie, Dirk L. Anaximanders discovery of space, in Anthony Preus (ed.),
Before Plato (Albany 2001) 23 48.
The Discovery of space: Anaximanders astronomy, in id., Robert Hahn,
and Gerard Naddaf (eds.), Anaximander in Context (Albany, 2003) 165
254.
Dicks, D. R. Early Greek Astronomy to Aristotle. London 1970.
Findlay, John N. Plato; The Written and Unwritten Doctrines. London 1974.
66 One feature of the prevalent view that presocratic philosophers were seeking
mechanistic and materialist explanations is the assumption that all or most of
the Presocratics were materialists in their metaphysics, and indeed that they
had no notion of the incorporeal at all. This is not the place to develop a
full defence of my claim that having a notion of the incorporeal is older and
more ordinary than the idea that there are no incorporeal entities. I see no rea-
son to suppose that it was hard for primitive thinkers to come up with an idea of
non-bodily powers and causes: assuming that there are such powers and causes
seems to me to be the norm in pre-scientific societies. I will develop this idea in
relation to the Presocratics in two further papers, as yet unpublished.
67 This paper was originally composed in 2005. The current version has not been
heavily revised to address more recent work, but it has benefited from discus-
sion with various audiences (in Samos and the B club in Cambridge in 2005,
and the pure mathematics seminar at UEA in 2006). Extensive written com-
ments from Carl Huffman on my first draft have been of immense value.
Furley, David. The Greek Cosmologists. Vol. 1. Cambridge 1987.
Hahn, Robert. Anaximander and the Architects. Albany 2001.
Proportions and numbers in Anaximander and early Greek Thought, in
Dirk L Couprie, Robert Hahn, and Gerard Naddaf, Anaximander in Con-
text. (Albany 2003) 73 163.
Heidel, William A. The Pythagoreans and Greek mathematics, in David Fur-
ley and R.E. Allen (eds.), Studies in Presocratic Philosophy (London 1970)
350 81.
Huffman, Carl. Philolaus of Croton. Cambridge 1993.
The Pythagorean tradition, in A. A Long (ed.), The Cambridge Companion
to Early Greek Philosophy (Cambridge 1999) 66 87.
Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King. Cam-
bridge 2005.
Kahn, Charles. Anaximander and the Origins of Greek Cosmology. New York
1960.
Kingsley, Peter. Ancient Philosophy, Mystery and Magic: Empedocles and Pythagor-
ean Tradition. Oxford 1995.
Leibniz, Gottfried Wilhelm. Correspondence with Clarke, in C. I. Gerhardt
(ed.), Die Philosophischen Schriften von Gottfried Wilhelm Leibniz (Berlin 1890)
7. 345 440.
Lloyd, Geoffrey E. R. Early Greek Science Thales to Aristotle. London 1970.
McKirahan, Richard. Philosophy before Socrates. Indianapolis 1994.
Naddaf, Grard, Anaximanders measurements revisited, in Anthony Preus
(ed.), Before Plato (Albany 2001) 5 23.
Nussbaum, Martha C. Eleatic conventionalism and Philolaus on the condi-
tions of thought, HSCP, 83 (1979) 63 108.
OBrien, Denis. Anaximanders measurements, CQ 17 (1967) 423 32.
Quine, Willard V. Ontological reduction and the world of numbers, in id.,
The Ways of Paradox and other essays (New York 1966), 199 207.
Ontological relativity, in id., Ontological Relativity and other Essays (New
York: Columbia University Press, 1969) 26 68.
Propositional objects, in id., Ontological Relativity and other Esays (New
York 1969) 137 60.
Robinson, Thomas M. Heraclitus: Fragments. Toronto 1987.
Vogel, Cornelia J. de. Pythagoras and Early Pythagoreanism: an Interpretation of Ne-
glected Rvidence on the Philosopher Pythagoras. Assen 1966.
Waterfield, Robin. The First Philosophers. Oxford 2000.
West, Martin L. Early Greek Philosophy and the Orient. Oxford 1971.
2. Pythagorean Communities
From Individuals to a Collective Portrait
Leonid Zhmud
In the middle of his doxographical discussion of Presocratic theories, Ar-

istotle makes an interesting psychological remark: It is what we are all
inclined to do, to direct our inquiry not by the matter itself, but by the
views of our opponents (De caelo 294b5). I think one can hardly find
a better motto for the Pythagorean studies of the last two centuries.
Most books on Pythagoras and early Pythagoreanism are highly polemi-
cal. This includes even such a paragon of objective research as Zeller,1 for
he, too, had his own target to attack, namely Rths History of Western
Philosophy,2 which accepted the entire ancient tradition on Pythagoras
as historically reliable. Zellers critical approach to the sources razed
Rths construction to the ground, so that very little remained of Pytha-
goras. Incidentally, what has remainedthe philosophical doctrine that
all is number, the astronomical theory of the spheres, and the concept
of the Central Firehas nothing in common with Pythagoras.
Admittedly, Zellers approach per se is sound and his distinction be-
tween the classical and the later sources is indeed crucial. The problem,
however, is that the classical authors did exactly what Aristotle said: they
were guided not by the matter itself, but by the views of their oppo-
nents. Aristotles own opponents were the Academics, and this fact
had a great impact on his treatment of the Pythagoreans. No other Pre-
socratic thinker was the object of such lively debate between philoso-
phers as Pythagoras. Starting with Xenophanes (3 7) and Heraclitus
(3 40, 81, 129), the entire fifth-century tradition concerning Pythagoras
is polemical. This is one of the reasons why this tradition had much
more to say about Pythagoras than about any other Presocratic thinker.
The first book about a Greek philosopher was Democritus Pythagoras
(68 A 33.1). At the same time, around 400 BC, a Sophist from Miletus,
1 E. Zeller, Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. 5 Vol. 1.

Leipzig 1895.
2 E. Rth, Geschichte unserer abendlndischen Philosophie. Mannheim 1846.
34 Leonid Zhmud
Anaximander the Younger, wrote a book on the Pythagorean symbola

(58 C 6). When the Academy created the monograph, the first genre
of philosophical historiography, which is devoted to an individual
thinker or a school, the first examples were Puhacqeia by Xenokrates
(fr. 2 Isnardi Parente) and On the Pythagoreans by Heraclides Ponticus
(fr. 22, 40 41 Wehrli). Speusippus wrote a book On the Pythagorean
Numbers (fr. 28 Tarn). In addition to the works about or against indi-
vidual Pythagoreans, Aristotle wrote two special monographs: About the
Pythagoreans (fr. 191 196 Rose), which contains a collection of the ma-
terials, and Against the Pythagoreans (fr. 198 205 Rose), which discusses
their philosophical and scientific views.
In the next generation, Aristoxenus works on Pythagoras and his
followers draw an idealized picture of the philosophers, scientists, and
politicians who lived according to their ethical principles.3 This is not
the picture that we find in Aristotle, or for that matter in Aristoxenus
biographies of Socrates and Plato, which are full of scandalous stories
(fr. 54a, 62, 67 Wehrli). Much less prejudiced, Dicaearchus also high-
lights Pythagoras, Socrates, and Plato as the heroes of his biographical
works and uses them to represent various forms of philosophical life.4
In Eudemus histories of mathematics and astronomy, the Pythagoreans
are quite dissimilar to the Pythagoreans of Dicaearchus, though not di-
rectly opposed to them.5
Thus, even when restricting ourselves to the classical sources, we still
get the same principal hypostases of Pythagoras found in modern scholar-
ship. Pythagoras is seen as an amalgam: a religious and moral teacher, a
politician, a philosopher and scientist, and the proportion of these quali-
ties he is assigned seems to be a matter of a personal choice. Burkert6
claims 90 % religion and 10 % politics with no science and philosophy,
while van der Waerden7 assigns 50 % religion and 50 % mathematics,
3 Peq Puhacqou ja t_m cmyqlym aqtoO (fr. 11 25), Peq toO PuhacoqijoO
bou (fr. 26 32), Puhacoqija !povseir (fr. 33 41), )qwta bor (fr. 47 50).
4 S. White, Principes sapientiae: Dicaearchus biography of philosophy, in W.
W. Fortenbaugh and E. Schtrumpf (eds.), Dicaearchus of Messana (New Bruns-
wick 2001) 195 236.
5 See L. Zhmud, The Origin of the History of Science in Classical Antiquity. Berlin
2006.
6 W. Burkert, Lore and Science in Ancient Pythagoreanism (Cambridge, Mass. 1972).
7 B. L. van der Waerden, Die Pythagoreer: Religise Bruderschaft und Schule der Wis-
senschaft. Zrich 1979.
2. Pythagorean Communities 35
whereas Riedweg8 views this as 98 % religion and 2 % philosophy, with

no scientific content. Now, I do not want to defend my own view of
these proportions,9 or if so, only in a very indirect way. Everyone who
knows Carl Huffmans works on Philolaus and Archytas10 would agree
that they are by far not as polemical as, for instance, Burkerts book, or
my own book. With respect to their subject matter, there are two obvious
reasons for this. First, Huffman does not need to prove that Philolaus was
a philosopher or that Archytas was a mathematician and a politician. Sec-
ond, there is no Pythagoras in Huffmans books, for Huffman does not
need him. I still have great difficulty in speaking about early Pythagorean-
ism without reference to Pythagoras,11 even though everything about him
appears to be highly disputable. In 1819, it was quite logical for A. Bckh
to start his discussion of Pythagorean philosophy and science with Philo-
laus, for he believed he could find there the undeniable written evidence
that either did not exist in the preceding period or was not so undenia-
ble.12 As we know, the case of Philolaus has proved to be more compli-
cated than Bckh thought; it needed the joint efforts of many scholars to
be settled. Yet it does not follow from this that the earlier period cannot
be reconstructed. Since it seems that Pythagoras himself is the main ob-
stacle to such a reconstruction, I am going to leave him alone for the time
being and turn to a much less problematic matter, namely the Pythagor-
eans.
Why are the schools followers in fact not as problematic as its
founder? Well, because they are different from their master. Pythagoras
pretended to possess supernatural qualities and was therefore the kind of
person to whom legends, even originally unrelated legends, tend to be
attached. Unlike Pythagoras, no historically known early Pythagorean is
connected with anything supernatural, mystical, or superstitious in the
reliable part of the tradition. The doctors Democedes and Alcmaeon,
8 C. Riedweg, Pythagoras. Leben, Lehre, Nachwirkung. Munich 2002.

9 L. Zhmud, Wissenschaft, Philosophie und Religion im frhen Pythagoreismus. Berlin
1997. See now an updated English version: Pythagoras and the Early Pythagoreans,
Oxford 2012.
10 C. A. Huffman, Philolaus of Croton: Pythagorean and Presocratic, Cambridge 1993;
idem. Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King,
Cambridge 2005.
11 See L. Zhmud, Some notes on Philolaus and the Pythagoreans, Hyperboreus 4
(1998) 243 270.
12 A. Bckh, Philolaos des Pythagoreers Lehren nebst den Bruchstcken seines Werkes
(Berlin 1819) 3 4.
36 Leonid Zhmud
the Olympic victors Milon and Iccus, the botanist Menestor, the philos-
ophers Hippon and Philolaus, and the mathematicians Hippasus and The-
odorus all appear in our sources to be closer to Anaxagoras than to Em-
pedocles. There is no evidence even for their belief in metempsychosis.
They are as normal as they can possibly be. It is their normality that has a
strong appeal to me, for if these people were Pythagoras students and fol-
lowers, we can learn something important from them about him and the
society he founded. If, furthermore, the Pythagoreans that we know of
were not the superstitious ritualists who adhered to rules forbidding
them to travel on main roads, use public baths, speak in the dark, step
over a yoke, sit on a bushel measure, stir the fire with a knife, etc.,13
then who were these ritualists? If they are nowhere to be found, we
must abandon the idea that such Pythagoreans ever existed.
To be sure, some of the Pythagoreans known to us did perform
miracles of faithfulness for their friends, like Damon and Phintias (Aris-
tox. fr. 31 Wehrli), but this is not a kind of miracle with which we
should be concerned. Nor should we worry about the rather abnormal
behavior of the Pythagorean athlete and general Milon: Aristotle calls
him pokvacor, according to the late sources, he devoured about
nine kilograms of meat a day and just as much bread and drank ten litres
of wine.14 Now, these stories are not just amusing anecdotes. Taken se-
riously, they reveal quite an important distinction. Milon, this Pythagor-
ean of the Pythagoreans, behaves in exactly the opposite way to what
could be expected of a true follower of Pythagoras. But how do we
know what should be expected of a true Pythagorean? In other
words: what sources do we use to create a composite image of a true
Pythagorean? Are they the same as our sources on the individual Pytha-
goreans? No, they are not. If we collect everything that is known about
the individual Pythagoreans and compare this with what is known about
anonymous Pythagoreans, Pythagoreans as a particular collective identi-
ty, we get very different pictures. Sometimes even one and the same au-
thor produces a different picture: Aristotles individual Pythagoreans
differ radically from his Pythagoreans in general, this time not in
their behavior, but in their doctrines.15 This is one of many reasons
why we ought to be very cautious about referring to the Pythagor-
13 See Burkert L&S (above, n. 6) 166 ff.

14 Arist. fr. 520 Rose (cf. NE 1106b3); Phylarch. F 3 FGrHist; Athen. 10.4.
15 Zhmud Pythagoras (above, n. 9) 433 ff.
eans as a collective identity, for this is the very area of the classical tra-
dition where we can expect to encounter the grossest distortions.
In the modern as well as in the ancient world, the stories told about a
social, ethnic, or cultural minority are often quite different from the sto-
ries told about the individuals who constitute these minorities. Though it
is not that the first are necessarily false and the second always true, the dis-
tinction between the two is quite important. If we proceed empirically,
our collective portrait of the Pythagoreans should look more or less
like a sum total of the traits common to all the individual Pythagoreans,
as well as those specific traits which are irreducible to a common denom-
inator. Certainly, oR Puhacqeioi is often just a faon de parler, behind
which the real figures are discernible, e. g., Archytas, who stands behind
the Pythagoreans in Platos Republic (530a531c), or Philolaus, whose as-
tronomical system Aristotle ascribes to some anonymous Pythagoreans
(De caelo 293a18 ff.). But equally often, the collective Pythagoreans do
not correspond to any known individuals, e. g., the Pythagoreans of
Anaximander the Younger, or those of Aristotles Metaphysics, or those
of Aristoxenus work Puhacoqija !povseir. In the last case as well as
in Theophrastus Metaphysics (11a27-b10), they are very much like the
Platonists. One can dispute these examples and adduce others, but this
hardly affects my general thesis: if actions or ideas allegedly peculiar to
the collective Pythagorean identity do not find independent confirmation
at an individual level, we will stay on safer ground by preferring individual
to collective evidence. Accordingly, still more suspicious are those testi-
monies about the Pythagoreans in general which plainly contradict the evi-
dence about individual Pythagoreans.
Some examples follow, beginning with the rules of conduct. The
traditional sources on Pythagorean vegetarianism are divided.16 Some
say that Pythagoreans did not eat meat, others that they abstained
from particular kinds of meat or particular parts of the animals. On an
individual level, strict vegetarianism is not attested, whereas consump-
tion of meat is. This means we should rather conclude that some Pytha-
goreans did eat meat, even if some others probably did not. Secondly, let
us look at their doctrines. Aristotle, who was followed by Theophrastus,
who was followed by the entire later tradition, persistently claims that
the core of Pythagorean philosophy is that all is number. We do
not find this thesis held by any of the Pythagorean thinkers, though
we find other Pythagorean ideas on number that differ both from the
16 Burkert L&S (above, n. 6) 180 ff.

38 Leonid Zhmud
Aristotelian version and from each other. I believe, therefore, that all is
number is an Aristotelian interpretation of various Pythagorean phi-
losophical and scientific ideas.17 Thirdly, we look at institutions. The
story that the early Pythagorean society was divided into mathematikoi
and akousmatikoi seems to be ineradicable from the scholarly literature,18
even though this story is found first in Porphyry (VP 37) and Iamblichus
(VP 80 89), and even though the word lahglatijr is first attested in
Platos Sophist (219c), whereas !jouslatijr is first found six centuries
later, in Clement of Alexandria (Strom. 5, 59). There is obviously some-
thing persuasive about this story that has made it so enduring. Burkert,
who once tried to show that it originated with Aristotle, now admits
that this is impossible to prove.19 Even if this story were a part of the
fourth-century tradition, we would not find in the sixth or the fifth cen-
tury any counterparts to the mathematikoi as they are described by Iam-
blichus and even less to the akousmatikoi. Should we believe that these
people existed but left no individual trace, whereas their collective por-
trait was kept secret to be disclosed only by Porphyry? I think it is better
to see this story for what it really is: as a construct of the Imperial age.20
Admittedly, the methodological individualism that I am professing is
not entirely unproblematic. In a sense, it is easier to follow its alternative,
definitional essentialism, i. e., to define and discuss specific Pythagorean
qualities or theories. If we speak of the Pythagoreans in general, there
is no need to bother about every individual Pythagorean: a deviant case
can always be treated as an exception. However, if one starts from the in-
dividual level, every single Pythagorean counts. In this case, the question
who is to be counted as a Pythagorean and according to which criteria?
becomes crucial. This is not an easy question, and modern research offers
widely differing answers. If one believes the late sources, the written Py-
thagorean tradition starts only with Philolaus, who lived a century after
Pythagoras. Accordingly, the Pythagoreans before Philolaus did not
write books, which means that those who did, e. g. Alcmaeon, Menestor,
and Hippon, were not real Pythagoreans.21 If one does not trust the late

18 See e. g. K. von Fritz, Mathematiker und Akusmatiker bei den alten Pythagoreern.
SBAW 11, Munich 1960.
19 Cf. Burkert L&S (above, n. 6) 192 ff.; id., Pythagoreische Retraktationen, in
W. Burkert et al. (eds.), Fragmentsammlungen philosophischer Texte der Antike
(Gttingen 1998) 314 5.
21 So Huffman Philolaus (above, n. 10) 15 16.
sources, but does trust Plato and Aristotle, the matter does not get any eas-
ier, for both of them avoided calling anyone a Pythagorean. Neither
Philolaus and his students in the Phaedo, nor Archytas in the Seventh Let-
ter, are called Pythagoreans. Was Platos teacher in mathematics, Theodo-
rus of Cyrene (43 A 2), a Pythagorean or a friend of Protagoras? Of
course, he could have been both, but Plato testifies only to the second.
Obviously he had his reasons to be reticent. Aristotles treatises are
quite densely populated with anonymous Pythagoreans and to a lesser de-
gree with individual Pythagoreans. He does mention by name Alcmaeon
(Meta. 986a27), Hippasus (Meta. 984a7), Hippon (Meta. 984a4; De an.
405b2), Philolaus (EE 1225a30), Eurytus (Meta. 1092b10), and Archytas
(Meta. 1043a21; Rhet. 1412a12; Pol. 1340b26), but he never tells us that
they were Pythagoreans.
If one looks not only to the ancient but to the modern authorities as
well, the situation appears less dramatic. But in Diels and in the other
modern collections of Pythagorean materials, we are faced with another
problem: there are too many Pythagoreans. Partly this is because the se-
lection criteria used are either not clear enough or not consistent. Diels
does not explain his criteria for considering someone a Pythagorean,
though he makes them quite visible, namely, by putting the evidence
from Aristoxenus catalogue of the Pythagoreans (Iamb. VP 267 = DK
58 A) at the beginning of part A. Except for one passing remark,22 I
have not found any explicit statements that he considered this catalogue
to derive from Aristoxenus and thus to be evidence of primary impor-
tance, although this is certainly what he believed. Further, Diels does
not always follow Aristoxenus. Thus, he places Cercops, Petron, Paron,
and Xuphus among the early Pythagoreans, though their names are lack-
ing in the catalogue. I think that he was wrong in all four cases.
According to Aristotle (fr. 75 Rose), the poet Cercops lived in Hes-
iods time, so he could not have been a Pythagorean.23 However, Epi-
genes, a grammarian of the Hellenistic age,24 in his book On the Writings
Attributed to Orpheus, calls Cercops a Pythagorean and ascribes two
poems to him: Zeqr kcor and EQr AVdou jatbasir (other authors at-
tributed these poems to Pythagoras). This evidence is not very reliable.
22 H. Diels, Antike Technik (Leipzig 1924) 23.

23 Burkert L&S (above, n. 6) 114, 130 n. 60; cf. DK 1.106.6 7.
24 Clem. Strom. 1.21.131. On Epigenes, see L. Cohn, Epigenes (16), RE 6
(1907) 64 65; cf. I. Linforth, The Arts of Orpheus (Berkeley 1941) 110 11,
114 ff.
40 Leonid Zhmud
Orphic poetry was always pseudonymous and there was no way to fig-
ure out the names of its real authors. Epigenes attributions must, there-
fore, have been guesswork, as was most of the other evidence of this
kind. In Cicero the reference to Cercops is linked to a quotation
from Aristotle, who claimed that there had never been a poet called Or-
pheus: Orpheum poetam docet Aristoteles numquam fuisse et hoc Orphicum
carmen Pythagorei ferunt cuiusdam fuisse Cercopis. 25 Only the first part of
this evidence derives from Aristotle (as is confirmed by a quotation
from Philoponus), whereas the second part goes back to Epigenes.26 Ar-
istotle would not call Hesiods contemporary a Pythagorean; more
importantly, he never calls anyone a Pythagorean.
We know about Petron (DK 16) only from a certain Hippys of
Rhegium, whose testimony is quoted by the Peripatetic Phanias of Ere-
sos; it is very likely that this is a forgery.27 Paron (DK 26) owes his ex-
istence to Aristotles mistake in taking the participle PAQYM to be a
proper name.28 Xuphus (DK 33) is mentioned only once in Aristotles
Physics (216b22). In the commentary to this passage, Simplicius calls
him a Pythagorean but it is impossible to verify his claim.
In Maddalenas and Timpanaro Cardinis works,29 Epicharmus, Ion
of Chios, Damon, Hippodamus, Polyclitus, Oenopides, and Hippo-
crates of Chios are considered to be Pythagoreans. This goes even fur-
ther than Diels and is absolutely too far. Their names are not found in
the catalogue and, moreover, no classical source considers any of them a
Pythagorean or even a pupil of the Pythagoreans. Even if Oenopides or
Hippocrates studied mathematics with the Pythagoreans, this fact alone
does not make them Pythagoreans.
Of the fourth century Pythagoreans on Diels list, another three
have to be removed. Timaeus of Locri (DK 49) owes his existence to
the Platonic dialogue and, later, to a Pseudo-Pythagorean text.30 Ocellus
25 Cic. De nat. deor. 1, 107 = Arist. fr. 7 Rose.

26 W. Kroll, Kerkops, RE 11 (1921) 314; Philop. In de an., 186.21 2 = Arist.
fr. 7 Rose.
27 See FGrHist 554 F 5 with comm.; J. Kerschensteiner, Kosmos: Quellenkritische
Untersuchungen zu den Vorsokratikern (Munich 1962) 209 10; L. Pearson, The
Greek Historians of the West (Atlanta 1987) 108 ff.
28 Burkert L&S (above, n. 6) 170; G. Martano, Il pitagorico Parone o il pitagor-
ico presente?, Elenchos 1 (1980) 215 224.
29 A. Maddalena, I Pitagorici. Bari 1954; M. Timpanaro Cardini, I Pitagorici: Testi-
monianze e frammenti. Fasc. IIII. Florence 1958 64.
30 H. Thesleff, The Pythagorean Texts of the Hellenistic Period (bo 1965) 202 ff.
of Lucania (DK 48) is mentioned in the catalogue, which means that

Aristoxenus regarded him as a historical person, but all the doctrines as-
cribed to him are Pseudo-Pythagorean.31 Lastly, the Pythagorean called
Lycon (DK 57) is in fact four different people.32 Since it seems unlikely
that Lycon of Tarent, as mentioned in the catalogue (57 A 1), will be
identified as being the same as the other three persons, what remains
of him is merely a name. But we are not interested in mere names,
for we have more then enough of them. We are looking for Pythagor-
eans with characteristic individual features that can be incorporated into
our collective portrait.
Why is it so important to look for the sources that explicitly call
someone a Pythagorean? Why not employ a doctrinal criterion, as is
employed in the case of the other schools? Indeed, a Hellenistic philo-
sopher can be regarded as a Platonist if he is known to belong to the
Academy or to profess specifically Academic doctrines. The problem
is that the Academy, the Lyceum, and the Stoa were institutionalized
schools with definite sets of doctrines, even if different at different
times. The Pythagorean school, in contrast, was founded neither as a
philosophical school, nor as an institutionalised school at all, but as a po-
litical society, a 2taiqea.33 Besides, Pythagoras teaching was never writ-
ten down and the school itself was dispersed both geographically and
chronologically, more than any other Presocratic school. This is why
we do not and should not expect to find anything resembling a Pytha-
gorean orthodoxy. As long as the Pythagorean school was alive, i. e., up
to the mid-fourth century BC, every Pythagorean philosopher devel-
oped his own views. Although there were certain similarities in think-
ing, it is difficult to identify one characteristic feature that was common
to all Pythagorean thinkers. Interestingly, where we do encounter a Py-
thagorean orthodoxy, e. g., in the pseudo-Pythagorean literature, it is
based on the Academic and Peripatetic interpretations of the Pythagor-
ean ideas, not on the authentic Pythagorean tradition. Therefore the
doctrinal criterion turns out to be of a limited value, though not invalid
as such. For if we find a common view in the medical theories of Alc-
maeon, Hippon, and Philolaus, this can confirm that they belonged to
31 Ibid., 124 ff.

32 See FGrHist 1110, with comm.
33 K. von Fritz, Pythagorean Politics in Southern Italy, New York 1940; E. Minar,
Early Pythagorean Politics in Practice and Theory (Baltimore 1942) 19 ff.
42 Leonid Zhmud
the same tradition, even if this view was not expressed by Hippasus and
Theodorus, who were not interested in medicine.
Admittedly, in Pythagorean science, i. e., in the four mathemata, the
situation looks different. There is such a common body of theories here,
and there are many more affinities among the views of different scien-
tists. Yet this consistency is not specifically Pythagorean; it is related to
the methods of the respective science. Hippocrates of Chios developed
Pythagorean geometry, Archytas solved the problem posed by Hippo-
crates, and Eudoxus studied geometry with Archytas, but neither Hip-
pocrates nor Eudoxus were Pythagoreans. What is specifically Pythagor-
ean in the exact sciences is the preoccupation with all four mathemata,
including arithmetic and harmonics. Indeed, before Pythagoras (or, if
you prefer, before Hippasus), theoretical arithmetic and mathematical
harmonics did not exist; and after Hippasus, the IoniansOenopides
and Hippocrates of Chios and to a certain degree also Democritusde-
veloped only geometry and astronomy, not the other two branches of
the quadrivium. This means, among other things, that Theodorus of Cy-
rene, who is mentioned in Aristoxenus catalogue and who taught all
four sciences (43 A 1, 4), was a Pythagorean and not just a friend of Pro-
tagoras. This means, furthermore, that Archytas predecessors, whose
knowledge of all four mathemata he praises at the beginning of his
work (47 B 1), were the Pythagoreans, and not Hippocrates of Chios
or the other Ionian mathematicians.34
This rather long digression was needed, I think, to make clear why
the question that I raised earlierabout how a Pythagorean can be de-
finedshould be settled on the basis of reliable classical sources. It is not
enough that a person calls himself a Pythagorean, like Lycon, the critic
of Aristotle, or the proto-Cynic Diodorus of Aspendus,35 for this is evi-
dence that they were not. But if someone was considered a Pythagorean
by his contemporaries or by the Pythagoreans themselves, this means
that he was judged by more complex and reliable criteria than we can
employ now. This means that he shared with the other Pythagoreans
not just one but many common features and at that it is these exact fea-
tures that define a Pythagorean, both from an internal and an external
point of view.
34 Pace Huffman Archytas (above, n. 10) 51 ff.

35 Lycon (57 A 4); Diodorus (Tim. FGrHist 566 F 16; Hermipp. fr. 24 Wehrli =
FGrHist 1026 F 26).
Now I come back to Aristoxenus catalogue of the Pythagoreans. The

first person to recognize that this catalogue, which is preserved in Iambli-
chus De vita Pythagorica, may be by Aristoxenus was Erwin Rohde.36
Diels, as I said, used this document for the Pythagorean chapters of his
Vorsokratiker, but did not go into detail concerning it. Later, Burkert
and Timpanaro Cardini briefly commented on the catalogue,37 suggesting
that it is based on genuine historical, probably documentary, evidence,
given that a list of 218 names organised according to 27 cities and nations
was not the kind of information that could be transmitted orally. A num-
ber of parallels between fragments of Aristoxenus and the catalogue make
his authorship quite certain. The evidence Aristoxenus relied on most
probably came from those last Pythagoreans with whom he was in con-
tact, namely Xenophilus, Phanton, Echecrates, Diocles, and Polymnastes,
the students of Philolaus and Eurytus (fr. 19 Wehrli). Another source of
information on the Pythagoreans was his father Spintharos, who belonged
to Archytas circle; he is twice mentioned in Aristoxenus biographical
works (fr. 30, 54a Wehrli).
What is meant by documentary evidence is not a formal member-
ship list of the Pythagorean society: such a list barely existed, if only be-
cause there was never a centralized Pythagorean community. Aristoxe-
nus list can be regarded rather as a reflection of the collective
Pythagorean memory concerning the prominent Pythagoreans of the
sixth, the fifth, and the early fourth centuries prominent not necessa-
rily in philosophy or science, but also in politics, or athletics, or medi-
cine. They were prominent members of Pythagorean communities, dis-
persed throughout the entire Greek world from Cyrene in Africa to
Cyzicus in Asia Minor. Some of them came from different cities to
study with a master, be it Pythagoras, Philolaus, or Archytas, but most
of them, I assume, remained in their own cities. Of the 218 Pythagor-
eans on Aristoxenus list, only about 60 are mentioned in any other
sources; the rest are no more than names. Moreover, we only have in-
formation that is of any use on about half of this small group of 60. The
largest group on Aristoxenus list (48 names) are all from his home city
Tarentum; the smallest (2 names) from Katana. There are only eight Py-
thagoreans in the catalogue who have no colleagues in the same city,
36 E. Rohde, Die Quellen des Iamblichus in seiner Biographie des Pythagoras

[1871], in id., Kleine Schriften. Vol. 2. (Tbingen 1901) 171.
37 Timpanaro Cardini (above, n. 29) 3.38 9; Burkert L&S (above, n. 6) 105
n. 40. See also Zhmud Pythagoras (above, n. 9) 109 ff.
44 Leonid Zhmud
and of these eight at least three are rather dubious figures. I would like
to point out here that we should approach Aristoxenus list critically,
like any other historical document, especially taking into account that
it comes from Iamblichus, who lived 600 years later. Though it does
not contain the names of the individuals who lived after Aristoxenus,
it is possible that some of the famous names on the list were added
later, after Aristoxenus. Being on the list does not guarantee that the
person in question really was a Pythagorean. In several cases, doubt re-
mains and on occasion there is evidence to warrant deleting a person
from the list. However, if there is no evidence to this effect, we can
consider the data of the catalogue to be a sufficient proof that the person
in question can be seen as a Pythagorean.
Though it is our most important document, the catalogue is by far
not the only one. Theophrastus mentions Hicetas of Syracuse (50 A 2),
who is not on the list, whereas his compatriot Ecphantus is wrongly
placed among the Pythagoreans from Croton (DK I, 446.11). Pythago-
ras contemporary, the famous Crotonian doctor Democedes, who mar-
ried a daughter of the Pythagorean athlete Milon (Hdt. 3, 127 137 =
19 A 1), is not on the list either. Obviously, in the course of catalogues
transmission, some names have been lost. For example, Aristoxenus fa-
ther Spintharos is absent, as well as Amyclas, though Aristoxenus men-
tions both him and his friend Cleinias (fr. 131 Wehrli), who is on the
list. Absent from the list are also Philolaus students Simmias and
Cebes (44 A 1a, B 15). The name of Parmenides teacher, the Pythago-
rean Ameinias, appears only in the Hellenistic biographer Sotion (D. L.
9.21 = 28 A 1). These significant names need to be added (as pluses)
to the catalogue.
On the other hand, the catalogue contains some names that clearly
should be removedminuses. For example, there are the ancient
lawgivers Zaleucus of Locri and Charondas of Catane, who seemed
to be associated with Pythagoras by the Pythagorean communities in
Locri and Rhegium as early as the fifth century. This means that Aris-
toxenus sources (fr. 17, 43 Wehrli) reflect a respectful but unreliable
historical tradition. Another famous duo are the miracle-workers, Aris-
teas and Abaris. Aristeas of Proconnesus, a shadowy figure from the late
seventh century BC, was the author of the Arimaspea, a poem describing
his journey in search of the Hyperboreans. During his lifetime, Aristeas
disappeared twice, and, according to Herodotus (4, 13 15), after the
second time, he reappeared 240 years later in Metapontum and told
the citizens to set up an altar to Apollo and dedicate a statue to himself.
In the catalogue, he is duly registered among the Pythagoreans from

Metapontum. Abaris, the mythical priest of Apollo and the expert on
the Hyperboreans, is listed in the catalogue as the only representative
of these legendary people. As Bolton showed, Aristeas and Abaris
were associated with Pythagoras in the fifth-century legendary tradition
and later in Heraclides Ponticus fantastic dialogues,38 so that Aristeas
miraculous powers, such as bilocation, were transferred to Pythagoras.
Thus, in this case too, the legendary and the historical tradition overlap.
Parmenides and Empedocles are also the sole representatives of their
cities. There seem to have been no Pythagorean communities in Elea
and Agrigentum, which means that in this case we can speak only of
the Pythagorean teachers of Parmenides and Empedocles. In the bio-
graphical tradition, both of them appear as students of the Pythagor-
eans.39 This could be the reason for including them in the catalogue, al-
though we do not know whether this attribution came before or after
Aristoxenus. The influence of Pythagorean ideas on Parmenides and
Empedocles is undeniable, but both these philosophers are too inde-
pendent and significant to be seen as completely integrated into Pytha-
gorean tradition. Rather, they should continue to be considered as Py-
thagorean sympathizers. The next and the last name to be removed is
Melissus, who is named together with five other Pythagoreans of
Samos. If there was a Pythagorean community on Samos, he could
have been a member even if, in terms of philosophy, he followed Par-
menides and Zeno. But to be certain we should strike him from the list,
for there is no need to be greedy.
After all these additions and subtractions, we can begin to make a
preliminary analysis. The first four generations of Pythagoreans (i. e.,
people born between 560 and 470 BC) with, at least to some extent,
discernible personalities can be placed in the following overlapping cat-
egories. First, the politicians, the largest category, containing the major-
ity of the names in the catalogue. Among the most prominent of them
was Milon, who won a battle against Crotons neighbour Sybaris around
510 BC.40 This victory made Croton the dominant city in Southern
Italy and, as the coins show, the neighbouring Pandosia, Temesa, and
38 J. Bolton, Aristeas of Proconnesus (Oxford 1962) 151 ff.

39 Empedocles (Alcidam. ap. D. L. 8.56 = 14 1 5; Theophr. 227 A FHSG = 31 A
7; Tim. FGrHist 566 F 14); Parmenides (Sotion ap. D. L. 9.21 = 28 A 1).
40 Diod. 12.9.2 10,1; Strabo 6.1.12 13 (both from Timaeus).
46 Leonid Zhmud
Kaulonia also became the dependent allies of Croton.41 The conquer-

ing of Sybaris caused a conflict within the ruling Crotonian aristocracy,
which led to Pythagoras flight to Metapontum. Aristoxenus describes
this conflict as a plot against Pythagoras, who had once refused the
rich aristocrat Cylon admittance to the Pythagoreans and thus made
him his enemy (fr. 18 Wehrli). Aristotle confirms the personal rivalry
between Cylon and Pythagoras and mentions another rival of Pythago-
ras, Onatas (fr. 75 Rose), who is listed among Crotonian Pythagoreans
(DK I, 446.13). As the arch-enemy of Pythagoras, Cylon is not on the
list, but if he really was the Crotonian exarch of the Sybarites, as Iam-
blichus says (VP 74), he must have been a Pythagorean.42 Anyway, we
know that political conflict also took place within the Pythagorean so-
ciety: Hippasus, for example, sided with Pythagoras enemies, while De-
mocedes remained Pythagoras supporter (Iamb. VP 255 f. = 18 A 5).
In the case of Hippasus and Democedes, we have two prominent
Pythagorean intellectuals who were politically active, as good citizens
were supposed to be. Milon was very successful both in politics and
in athletics, which leads us to our second category of Pythagoreans:
the athletes. From 532 to 488 BC, Croton achieved an extraordinary
series of victories in the Olympian Games.43 The catalogue gives us
the names of four Olympic victors: Milon, Astylos, Dicon, and Iccus
(DK I, 446.14. 20. 28, 447.14). There is no doubt that there were
many more. And there is no need to prove the importance of athletics
for the ethos of the ruling Italian aristocracy. We can safely assume that
many of the Pythagorean athletes were also politically active, as was the
case with Milon. Athletics was connected with the aristocratic way of
life, but it was also closely linked to medicine.
This brings us to our third category, the doctors. This group is
smaller than the politicians, but is a very important group for our under-
standing of the role of the natural sciences in ancient Pythagoreanism.
Iccus of Tarent won in the pentathlon in 476 BC and later became
an athletic coach and famous doctor (DK 25). He specialized in dietetics
and gymnastics and won Platos praise for his wisdom and temperate
41 Von Fritz Pyth. Politics (above, n. 33) 80 ff.; Minar (above, n. 33) 36 ff.
42 Minar (above, n. 33) 69 70; M. Giangiulio, Ricerche su Crotone arcaica (Pisa
1989) 311 n. 52; M. Bugno, Da Sibari a Thurii: La fine di un impero (Naples
1999) 41 2.
43 C. Mann, Athlet und Polis im archaischen und frhklassischen Griechenland (Gttin-
gen 2001) 164 ff.
way of life. Democedes (19 A 1) and Alcmaeon were two other pro-
minent representatives of the Crotonian school of medicine, which
stressed the importance of dietetics and gymnastics in maintaining
good health.44 The botanical book of Menestor was related to medicine
(DK 32), just as were the natural-philosophical writings of Alcmaeon
(24 A 1, B 4) and later, those of Hippon (38 A 11).
Our fourth category is natural philosophers, physikoi, according to
the Aristotelian terminology. Here we have Alcmaeon, Brontinus
(one of the addressees of Alcmaeons book; 24 B 1), Hippasus (18 A
1, 12), and Menestor. Among those born around 480 470 BC, there
are two other physikoi, Hippon and Philolaus.
The fifth category is mathematikoi, i. e., people concerned with any
of the four mathemata or with all of them. Strangely enough, we know
the names of only two mathematicians from the first 100 years of the
Pythagorean school: Hippasus (18 A 4, 14 15) and Theodorus (43 A
3 5). There is no doubt that there were many more, because by the
time when Oenopides and Hippocrates of Chios were active, i. e.,
about 450 430 BC, the bulk of the first four books of the future Ele-
ments of Euclid had already been written,45 which was an impossible feat
for only one or two people. Regrettably, we cannot identify the other
mathematikoi. For example, Parmenides teacher Ameinias could have
been a physikos, or a mathematician, or both. Alcmaeon was interested
in astronomy (24 A 4, 12), but certainly was not a mathematikos. Philo-
laus fits this term much better, especially as an astronomer and a special-
ist in harmonics (44 A 16 21, B 6).
So far we have identified fourteen prominent Pythagoreans belong-
ing to five overlapping categories. Now, let us imagine that we know
nothing about Pythagoras himself and all we have is a collective portrait
of his students and followers and the students of these followers. What
would an individual portrait of Pythagoras look like if we had to con-
struct it solely on the basis of the available collective portrait? I believe it
would be both natural and legitimate to assume that he had something
to do with activities that, already during his lifetime, became so distinc-
tive to the Pythagorean society. Surely, we should not expect a perfect
match, because for all his pokulaha (Heracl. 22 3 40) and pokutqopa
(Antisth. fr. 51 Caizzi), Pythagoras could not have been involved in all
44 See Zhmud Pythagoras (above, n. 9) 347 ff.

45 E. A. Neuenschwander, Die ersten vier Bcher der Elemente Euklids, AHES
9 (1973) 325 380.
48 Leonid Zhmud
these activities. But he would certainly have worked in some of these

fields and encouraged others.
Where, then, is the other side of Pythagoreanism: religion, magic,
mysticism, shamanism, ritual taboos, and so on? If Pythagoras was a
guru, as Riedweg suggests, where have all the other gurus gone? Did
I leave someone off the list? No, I did not. Then, could it be that Aris-
toxenus rationalistically-minded teachers struck all the gurus off the list?
I do not think so. In fact, they added two miracle-workers, even though
one of them, Aristeas, lived a century before Pythagoras, while the sec-
ond, Abaris, was legendary. The only other guru on the list would be
Empedocles, if we concede that he was a Pythagorean. There was
also Brontinus (DK 17), who was regarded by Epigenes as the author
of two Orphic poems, namely, Ppkor and Vusij (DK 15). In the
mid-fifth century, Ion of Chios asserted that Pythagoras was the author
of some Orphic poems (36 B 2), and his remark trigged all the further
attributions. It is misleading to present the early Pythagorean society as a
kind of workshop specialized in producing Orphic poems, as West
did.46 The more we know about Orphism, the more visible its profound
difference from Pythagoreanism becomes apparent. But even granted
that Brontinus was the author of some religious poems, this did not pre-
vent him from being a doctor, or a natural philosopher, or a politician,
as shown by the case of Empedocles.
Speaking about Empedocles in more appropriate ancient terms, he
was not a guru, but a divine man, he?or !mq,47 exactly like Pythagoras
before him. However, in contrast to Empedocles, of whose followers
we know nothing, Pythagoras founded a political society that outlived
him for at least fifty years and a school that existed till the mid-fourth
century. Both Pythagorean politicians and Pythagorean philosophers
and scientists took from him what they were interested in and what
they valued most. Incidentally, these interesting and valued things did
not seem to include magic, mysticism, and a variety of ritual taboos;
at least, taboos are not attested at all on the individual level and quite
poorly at the collective level. Herodotus says that Orphics and Pytha-
goreans did not bury people in woollen clothes (2.81). There were
some other attested taboos, for example, the prohibition of certain
kinds of meat and fish (Arist. fr. 194 Rose), or beans (cf. Emped. 31
46 M. L. West, The Orphic Poems (Oxford 1981) 7 ff.

47 For important qualifications of this notion, see D. S. Du Toit, Theios Anthropos,
Tbingen 1997.
B 136 141), although we do not really know how rigorously these

rules were observed. Anyway, there was nothing to fire the imagination
in these regulations and nothing that would make a religious sect of the
Pythagorean society.
What is then the basis for the widespread idea of a Pythagorean sect? 48
Putting aside Iamblichus akousmatikoi, it is the Pythagorean symbola, at-
tested first in Anaximander the Younger (58 C 6) and after him in Aris-
totle (fr. 194 196 Rose). If the symbola were strictly followed and if they
constituted that very Pythagorean way of life that was approvingly
mentioned by Plato (Res. 600a), then the Pythagoreans were indeed su-
perstitious ritualists and their society was a sect. There are plenty of rea-
sons, however, why this cannot be the case. It is anachronistic to speak of
a sect in archaic Greece. There were no sects at this time; there were
other religious and cultic communities, like hasoi or aqce_mer. Pytha-
gorean society was neither of them, it was a political 2taiqea. At no
time do we know of any specific Pythagorean cults or deities; their reli-
gion was the traditional polis religion. Pythagoras only known religious
innovation was metempsychosis, but even this was borrowed from Or-
phism along with several prescriptions that follow from this doctrine.
The so-called Pythagorean symbola contain almost a hundred prescriptions
that are hard to understand and much harder to follow.49 If we compare
them to the strictest charters of the real religious communities, like the
Jewish Essenes or the early Christian monastery of St. Pachomius, we im-
mediately see the difference between religious discipline and religious
folklore. Neither the Jewish nor Christian charters contain any nonsense:
every rule is clear and quite logical in the given environment; all of them
are enforced by various punishments for those who fail to follow them, so
that we can see a real, even though severe, life behind them. What kind
of life stands behind the symbola?
To take the acusmata seriously means an almost frightening constriction of
ones freedom in daily life. Whether a Pythagorean gets up or goes to
bed, puts his shoes on or cuts his nails, stirs the fire, puts on the pot, or
48 W. Burkert, Craft versus sect: the problem of Orphics and Pythagoreans, in

B. F. Meyer (ed.), Jewish and Christian Self-Definition, Vol. 3 (London 1983) 1
22.
49 On symbola, see: C. Hlk, De acusmatis sive symbolis Pythagoricis, Diss. Kiel 1894;
F. Boehm, De symbolis Pythagoreis (Diss. Berlin 1905); A. Delatte, tudes sur la
littrature pythagoricienne (Paris 1915) 271 2; Burkert L&S (above, n. 6) 166 ff.
50 Leonid Zhmud
eats, he always has a commandment to heed. He is always on trial and al-

ways in danger of doing something wrong.50
To appreciate this impressive but anachronistic picture properly, it is
crucial to understand that none of its elements correspond to the stories
told either about individual Pythagoreans or about Pythagoreans in gen-
eral. This picture emerges only if we put together all the sayings con-
tained in the symbola that were collected by Anaximander. But in inter-
preting sayings, we are dealing with folklore, not with reality. There is
no doubt that Anaximander dealt with folklore too, for he did not de-
scribe the way of life of any (named or anonymous) Pythagoreans. He
collected what he took to be the Pythagorean sayings and maxims
and interpreted them allegorically, i. e., according to one of the methods
of literary criticism available at that time.51 If he knew of a real Pytha-
gorean who did not break bread or step on nail parings, why did he in-
terpret these taboos allegorically? Aristotle believed that the taboos were
originally literal (fr. 195 Rose), which seems very plausible, but he too
had never heard of a Pythagorean really observing them. Except for a
few dietary prescriptions and burial customs, all the other taboos appear
only in the context of interpreting Pythagorean sayings. Nobody
would claim that by interpreting the sayings of Solon or any other of
the Seven Sages one could get closer to them as historical figures. It
is clear that making this distinction between folklore and historical real-
ity is a vital condition for the study of Pythagoreanism.52
Bibliography
Bckh, August. Philolaos des Pythagoreers Lehren nebst den Bruchstcken seines
Werkes. Berlin 1819.
Boehm, Fritz. De symbolis Pythagoreis. Diss. Berlin 1905.
Bolton, James. Aristeas of Proconnesus. Oxford 1962.
Bugno, Maurizio. Da Sibari a Thurii: La fine di un impero. Naples 1999.
Burkert, W. Craft versus sect: the problem of Orphics and Pythagoreans, in
B. F. Meyer (ed.), Jewish and Christian Self-Definition, Vol. 3 (London 1983)
1 22.
Lore and Science in Ancient Pythagoreanism, tr. E. L. Minar, Jr.. Cambridge,
Mass. 1972.
50 Burkert L&S (above, n. 6) 191.

51 He also interpreted the Homeric poems allegorically (Xen. Symp. 3.6)
52 See Zhmud, Pythagoras (above, n. 9) 192 ff.
Pythagoreische Retraktationen: Von den Grenzen einder mglichen Edi-
tion, in W. Burkert et al. (eds.), Fragmentsammlungen philosophischer Texte
der Antike / Le raccolte dei frammenti di filosofi antichi (Gttingen 1998)
303 319.
Cohn, L. Epigenes (16), RE 6 (1907) 64 5.
Delatte, Armand. tudes sur la littrature pythagoricienne. Paris 1915.
Diels, Hermann. Antike Technik. Leipzig 1924.
Du Toit, David S. Theios Anthropos: zur Verwendung von he?or %mhqypor und
sinnverwandten Ausdrcken in der Literatur der Kaiserzeit. Tbingen 1997.
von Fritz, Kurt. Mathematiker und Akusmatiker bei den alten Pythagoreern. Sitz.Ber-
lin.Akad.Wissen. 11. Munich 1960.
Pythagorean Politics in Southern Italy. New York 1940.
Giangiulio, Maurizio. Ricerche su Crotone arcaica. Pisa 1989.
Hlk, Cornelius. De acusmatis sive symbolis Pythagoricis. Diss. Kiel 1894.
Huffman, Carl A. Archytas of Tarentum: Pythagorean, Philosopher and Mathemati-
cian King. Cambridge 2005.
Philolaus of Croton: Pythagorean and Presocratic. Cambridge 1993.
Kerschensteiner, Julia. Kosmos: Quellenkritische Untersuchungen zu den Vorsokrati-
kern. Munich 1962.
Kroll, Wilhelm. Kerkops, RE 11 (1921) 314.
Linforth, Ivan. The Arts of Orpheus. Berkeley 1941.
Maddalena, Antonio. I Pitagorici. Bari 1954.
Mann, Christian. Athlet und Polis im archaischen und frhklassischen Griechenland.
Gttingen 2001.
Martano, Giuseppe. Il pitagorico Parone o il pitagorico presente? Elenchos 1
(1980) 215 224.
Minar, Edward. Early Pythagorean Politics in Practice and Theory. Baltimore 1942.
Neuenschwander, Erwin A. Die ersten vier Bcher der Elemente Euklids:
Untersuchungen ber den mathematischen Aufbau, die Zitierweise und
die Entstehungsgeschichte, Archive for History of Exact Sciences 9 (1973)
325 380.
Pearson, Lionel. The Greek Historians of the West. Atlanta 1987.
Riedweg, Christian. Pythagoras. Leben, Lehre, Nachwirkung. Mnchen 2002.
Engl. tr. by S. Rendall, Pythagoras: His Life, Teaching, and Influence. Ithaca
2005.
Rohde, Erwin. Die Quellen des Iamblichus in seiner Biographie des Pythago-
ras [1871], in id., Kleine Schriften. Vol. 2. Tbingen 1901.
Rth, Eduard M. Geschichte unserer abendlndischen Philosophie. Mannheim 1846.
Thesleff, Holger. The Pythagorean Texts of the Hellenistic Period. bo 1965.
Timpanaro Cardini, Maria. I Pitagorici: Testimonianze e frammenti. Fasc. IIII.
Florence 1958 64.
van der Waerden, Bartel L. Die Pythagoreer: Religise Bruderschaft und Schule der
Wissenschaft. Zrich 1979.
West, Martin L., The Orphic Poems. Oxford 1981.
White, Steven. Principes sapientiae: Dicaearchus biography of philosophy, in
W. W. Fortenbaugh and E. Schtrumpf (eds.), Dicaearchus of Messana (New
Brunswick 2001) 195 236.
52 Leonid Zhmud
Zeller, Eduard. Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. 5

Vol I. Leipzig 1895.
Zhmud, Leonid. The Origin of the History of Science in Classical Antiquity. Berlin
2006.
Some notes on Philolaus and the Pythagoreans, Hyperboreus 4 (1998)
243 270
Wissenschaft, Philosophie und Religion im frhen Pythagoreismus. Berlin 1997.
Pythagoras and the Early Pythagoreans. Oxford 2012.
3. Aristotle on the Pythagoreans
His Sources and his Accounts of Pythagorean Principles1
Richard McKirahan
Aristotles importance as a source on early Pythagorean thought is both

obvious and problematic. It is obvious because of his serious interest in
the thought of his predecessors (including the Pythagoreans), because he
provides far more information about the Pythagoreans than earlier au-
thors do, and because he lived before the contaminating effects of Neo-
pythagorean interpretations and reconstructions of the history of philos-
ophy.2 It is problematic because he interprets the Pythagoreans through
the lenses of his own philosophical concerns and his own theories, be-
cause his theories and his approach make him unreceptive and unsym-
pathetic to Pythagorean ways of thinking, and because he does not
name his sources. I am going to take up some issues in these problematic
areas. Specifically, I want to discuss his relation to his sources, whether it
is likely that he based his account of the Pythagoreans on only one
source: Philolaus in particular. I will present a case for a plurality of
sources and will explore some implications of this thesis for Aristotles
accounts of Pythagorean principles.
It hardly needs to be said that I have written this paper with Carl
Huffman in mind. He has presented a strong case that Philolaus is Aris-
totles principal source,3 and it is with due respect that I advance a dis-
1 This paper is a substantially rewritten and expanded version of my conference

presentation. I want to thank the other participants whose comments, objec-
tions, and conversation drew my attention to some points in the conference
presentation that needed further thought and elaboration. Except for the trans-
lations taken from the second edition (2011) of my Philosophy Before Socrates and
a few other small changes, the paper is identical with the version originally sub-
mitted for publication in 2005.
2 See W. Burkert, Lore and Science in Ancient Pythagoreanism (Cambridge,
Mass. 1972) 46: What Aristotle presents as the philosophy of the Pythagoreans
is truly pre-Socratic, unaffected by the achievements of Socratic-Platonic dia-
lectic, and not to be measured by their standards.
3 C. Huffman, Philolaus of Croton, Pythagorean and Presocratic (Cambridge 1993)
esp. 28 34. Huffman admits that Aristotle might also have made use of oral
54 Richard McKirahan
senting view. I have not, however, written this paper mainly as a rebut-
tal of his arguments. Instead, after an introductory section (Section 1), I
present some passages which prove beyond reasonable doubt that he
used more than one source in some of his discussions of Pythagorean
doctrines (Section 2). I then present some of Aristotles claims about Py-
thagorean principles, particularly those that have to do with numbers
(Section 3), and the fragments of Philolaus that bear on principles (Sec-
tion 4),4 with a preliminary discussion of the relations between the Ar-
istotelian material and Philolaus (Section 5) and some chronological
considerations that support the likelihood that Aristotle had access to
other sources than Philolaus book (Section 6). I then translate and com-
ment on four passages from the Metaphysics that provide Aristotelian
context for many of the doctrines he ascribes to the Pythagoreans and
give important clues about how he understood those doctrines, and
whether they can be fitted into a coherent ontology (Section 7).
Next I return to the claims of Section 4 and attempt to determine
which of them are authentic Pythagorean views and which are Aristo-
telian interpretations (Section 8). On the basis of the claims I find to be
authentic, I offer a brief sketch of a way Pythagorean thought on prin-
ciples may have evolved down to the time of Philolaus (Section 9). In
the final substantive part of the paper I turn to Aristotles treatment of
Pythagorean principles and attempt to account for his evident frustration
and lack of sympathy with their views. I suggest that Aristotles difficulty
in making sense of that aspect of Pythagorean thought is due to a fun-
damental difference between his way of approaching philosophical issues
and theirs (Section 10).
1. Aristotles Sources on the Pythagoreans
Aristotle mentions Pythagoras rarely (never in the works concerned

with metaphysics, epistemology, and nature), and refers to Philolaus
by name only once (EE 1225a33), in a context that has nothing to do
reports (32, 63), and even mentions that he might have made use of other
writing (63), but he does not give serious weight to these other possibilities,
especially since he (wrongly, see n. 19) rules out Archytas as a possible source.
For discussion of individual points in Huffmans treatment, see nn. 5, 13, 14,
19, 42, 43, 50, 52, 109, 117, 133, and pp. 98 9.
4 Philolaus B1-B6, quoted and translated in Section 4. Huffman considers these
fragments authentic. See n. 12.
Aristotle on the Pythagoreans 55
with metaphysical, physical, or mathematical speculation. When he

gives his accounts of Pythagorean thought, he speaks instead of the Py-
thagoreans or those called Pythagoreans.5 If we knew nothing more
about Philolaus, we would have absolutely no reason to suppose that
Aristotle based his accounts on Philolauss writings. On the other
hand, we are told that Philolaus wrote one or more books6 and that
Plato owned a copy of at least one of them,7 which if true may well
place a copy of the book within Aristotles reach. In addition, Philolaus
is the first Pythagorean we know of to have left a written account of his
philosophical views8 and is the only Pythagorean we know to have done
so prior to Archytas.9 Further, Aristotle wrote at least one work10 on the
Pythagoreans and another work (containing three books) on Archytas.11
These books would have been based at least partly on written Pythagor-
ean sources. Moreover, a number of genuine fragments of Philolaus
have survived that have a more or less close relation to Aristotles reports
of the philosophy of the Pythagoreans.12
5 The frequently made assertion that Aristotle commonly or often refers to the Py-
thagoreans as oR jako}lemoi Puhacoqe?oi (sometimes translated the so-called Py-
thagoreans) is false, e. g. Burkert (above, n. 2) 30. This expression occurs only six
times (Cael. 284b7, 293a20; Metaph. 985b23, 989b29; Mete. 342b30, 345a13) of
which four (those from De Caelo and Metaphysics) are found in contexts where
Aristotle refers to the Pythagoreans also as oR Puhacoqe?oi. By contrast, I count
thirty-nine occurrences of oR Puhacoqe?oi in the following works: Posterior Ana-
lytics, De Anima, De Caelo, Nicomachean Ethics, Magna Moralia, Metaphysics, Oeco-
nomicus, Physics, Problems and in the Fragments. Further, I see no reason to sup-
pose that Aristotle means oR jako}lemoi Puhacoqe?oi in the sense of the so-called
Pythagoreans. The passages cited by Huffman (Pol. 1290b40 and 1291a1, GA
734a19, De an. 410b28) equally support the view that the phrase means those
known as Pythagoreans, those I call Pythagoreans for my present purposes
or even those who call themselves Pythagoreans. The occurrences of jako}le-
moi in the Politics support this interpretation.
6 One: D.L. 8.85; three: Iamblichus VP 199. Huffman (above, n. 3) 12 14 rea-
sonably rejects the evidence for three books.
7 Timon F 54 (= DK 44 A8), D.L. 8.85.
8 D.L. 8.55 and other texts discussed by Huffman (above, n. 3) 15.
9 See discussion in Huffman (above, n. 3) 8.
10 D.L.s catalogue of Aristotle works (D.L. 5.25) names both Pqr tor Puhaco-
qe_our and Peq t_m Puhacoqe_ym. Burkert (above, n. 2) 29 n. 5 accepts both as
genuine.
11 D.L. 5.25.
12 Huffman (above, n. 3) 34 takes Fragments 1 7, 13, and 17, and Testimonia 9,
16 (part), 17 (part), 18 21, and 27 28 to be genuine.
On the strength of these facts, Huffman has proposed that since Ar-
istotles work on Archytas will have been based on Archytas writings,
Philolauss writings will be the only written source of Aristotles refer-
ences to those called Pythagoreans and the Pythagoreans in the
summaries of Pythagorean thought in his extant works.13
If this is true, and in particular if the Pythagoreans whose views Ar-
istotle records from time to time do not include Archytas, two questions
immediately arise. First, why does Aristotle not speak of Philolaus, but
of the Pythagoreans as a group? And second, how do we account for
discrepancies between Aristotles accounts and Philolauss fragments?
Regarding the first question, it is difficult to find a plausible explanation
for Aristotles failure to name Philolaus in the relevant passages.14 The
answer to the second question depends on the magnitude of the discrep-
ancies. If they are small, we may simply say that Aristotle may have
thought he was representing Philolaus accurately and he may not
have noticed the discrepancies even though we do. Further, Aristotles
lack of sympathy for Pythagorean views and his evident failure to make
sense of them may have resulted in distortions. But clearly, the greater
the discrepancies prove to be, the harder it is to explain them away. In
what follows, I will argue that some discrepancies are very great indeed.
On the other hand, if we suppose that Philolaus was not Aristotles
only source, there are several straightforward ways to explain Aristotles
preference for speaking of the Pythagoreans and his avoidance of men-
tioning any single Pythagorean, Philolaus in particular. First, his ac-
counts of Pythagorean ideas may not be based on Philolaus at all, but
on one or more other Pythagorean sources, written or oral, that are un-
known to us. In this regard we must not neglect lost works on the Py-
thagoreans, such as the book of Aristoxenus, whose sources we are not
in a position to name or evaluate. Second, his accounts may be based on
Philolaus and on other Pythagorean sources as well. If there was more
than one Pythagorean source at hand, Aristotle need not have expound-
ed them separately. It would have been natural for him to give a single
13 Huffman (above, n. 3) loc. cit. Philolauss book is the only likely candidate
(32); it is hard to say who Aristotles sources would have been if not Philolaus
(58 9). But Huffmans argument, even if sound, will apply only to Aristotles
lost works on the Pythagoreans, not to what he says about Pythagorean doc-
trines in his extant treatises. I can find no reason to exclude the possibility
that some of his remarks on the Pythagoreans, in particular his account of Eur-
ytus, are based on Archytas (see n. 19).
14 Even Huffman (above, n. 3) 34 seems to register unease on this point.
account stitched together from more than one source if he had thought
that the differences were not major, and especially if he thought it un-
important for his particular purposes to make fine distinctions among
the differing views.
I will proceed by giving some evidence that Aristotle used more
than one source (or a compendium of Pythagorean views, conceivably
one of his own works that was based on more than one source) for his
information on the Pythagoreans. In the subsequent treatment of Aris-
totelian passages on Pythagorean principles, I will point out some claims
that do not fit well with what we know of Philolauss views and which
may be easier to account for by supposing that they stem from different
sources. Later in the paper I will suggest (though it must remain only a
suggestion) that Aristotle did not even use Philolaus as his principal
source for Pythagorean doctrines, and that he may have regarded him
as somewhat out of the mainstream. If this is so, it is no longer myste-
rious that Aristotle did not identify Philolaus in particular as the source
of his information.
2. Some Pythagorean Doctrines

I begin with several testimonia that show that on some points the Py-
thagoreans had differing views.
De Caelo 300a14 17: T d( aqt sulba_mei ja to?r 1n !qihl_m sumtihe?si
tm oqqam|m 5mioi cq tm v}sim 1n !qihl_m sumistsim, speq t_m Puha-
coqe_ym tim]r.
The same consequence follows from putting the heaven together out of
numbers, for some people constitute all nature out of numbers, as some
of the Pythagoreans do. (tr. based on the Revised Oxford Translation) 15
Metaphysics 1080b30 3: lomadijor d tor !qihlor eWmai p\mter tih]asi,
pkm t_m Puhacoqe_ym, fsoi t 4m stoiwe?om ja !qw^m vasim eWmai t_m
emtym 1je?moi d( 5womtar l]cehor.
They all posit numbers to consist of abstract units except those of the Py-
thagoreans who declare the one16 to be the element and principle of the
15 J. Barnes (ed.), The Complete Works of Aristotle. The Revised Oxford Translation, 2
vols. (Princeton 1984).
16 I translate t 6m as the one (in scare-quotes). In most cases I take it as equiv-
alent to the number one.
things that are; those people suppose that numbers have magnitude. (tr.
based on the Revised Oxford Translation) 17
Meteorologica 342b29 33: t_m d( Ytakij_m timer jakoul]mym Puhacoqe_ym
6ma k]cousim aqtm eWmai t_m pkam^tym !st]qym, !kk di pokkoO te
wq|mou tm vamtas_am aqtoO eWmai ja tm rpeqbokm 1p lijq|m.
Some of the Italians called Pythagoreans say that it [sc. the comet] is one of
the planets, but that it appears at great intervals of time and only rises a little
above the horizon. (tr. based on the Revised Oxford Translation)
De Sensu 445a16 17: d d k]cous_ timer t_m Puhacoqe_ym, oqj 5stim euko-
com tq]veshai c\q vasim 5mia f`a ta?r asla?r.
What some of the Pythagoreans say is unreasonable. They declare that
some animals are nourished by odors alone. (my tr.)
I take it that Aristotle means one thing when he says that the Pythagor-
eans held a view and another when he says that some Pythagoreans
held a view, and what he means in the latter case is that other Pythagor-
eans held different views on the same subject, whether or not he tells us
what those other views are. In three instances he actually presents rival
Pythagorean views.
De Anima 404a16 19: 5oije d ja t paq t_m Puhacoqe_ym kec|lemom
tm aqtm 5weim di\moiam 5vasam c\q timer aqt_m xuwm eWmai t 1m t`
!]qi n}slata, oR d t taOta jimoOm.
The doctrine of the Pythagoreans seems to have the same idea; some of
them declared that the soul is the motes in the air, while others said that
soul is what moves them. (tr. based on the Revised Oxford Translation)
Meteorologica 345a13 18: t_m lm owm jakoul]mym Puhacoqe_ym vas_ timer
bdm eWmai ta}tgm oR lm t_m 1jpes|mtym timr !st]qym, jat tm keco-
l]mgm 1p Va]homtor vhoq\m, oR d tm Fkiom toOtom tm j}jkom v]qesha_
pot] vasim oXom owm diajejaOshai tm t|pom toOtom E ti toioOtom %kko
pepomh]mai p\hor rp t/r voqr aqt_m.
Of those called Pythagoreans some say that this [viz. the Milky Way] is the
path of one of the stars that fell from heaven during the alleged destruction
at the time of Phaethon, while others say that the sun once moved in this
circle; and so this region was scorched or met with some other affection of
this kind, because of the motion of these bodies. (tr. based on the Revised
Oxford Translation)
Metaphysics 986a15 17, a22 6: va_momtai d ja oxtoi tm !qihlm mol_-
fomter !qwm eWmai ja r vkgm to?r owsi ja r p\hg te ja 6neir, 6te-
qoi d t_m aqt_m to}tym tr !qwr d]ja k]cousim eWmai tr jat sustoi-
17 This translation differs considerably from the Revised Oxford Translation

(which, following Ross, fails to translate 1080b32). I take fsoi as identifying
a group of Pythagoreans rather than explicating what all the Pythagoreans did.
w_am kecol]mar, p]qar [ja] %peiqom, peqittm [ja] %qtiom, 4m [ja] pk/hor,
denim [ja] !qisteq|m, %qqem [ja] h/ku, AqeloOm [ja] jimo}lemom, eqh
[ja] jalp}kom, v_r [ja] sj|tor, !cahm [ja] jaj|m, tetq\cymom [ja]
2teq|lgjer.
Evidently, then, these thinkers consider that number is the principle both as
matter for the things that are and as their attributes and states. Others of
this same group say there are ten principles, which they arrange in parallel
columns limit and unlimited, odd and even, one and plurality, right and
left, male and female, resting and moving, straight and curved, light and
darkness, good and bad, square and oblong. (tr. based on Revised Oxford
Translation)
Did Aristotle derive this information from a single source? If he did, it
would almost have to be a compendium of their views rather than a sin-
gle original text such as that of Philolaus, which seems (like other pre-
socratic texts, as far as we can tell) to have set out its authors theory
without expounding other peoples contrasting views.
In one case Aristotle refers a particular doctrine to a Pythagorean
other than Philolaus.
Metaphysics 1092b10 12: ja r Euqutor 5tatte t_r !qihlr t_mor, oXom bd
lm !mhq~pou bd d Vppou, speq oR tor !qihlor %comter eQr t sw^-
lata tq_cymom ja tetq\cymom ovtyr !voloi_m ta?r x^voir tr loqvr
t_m vut_m.
This is how Eurytus decided what was the number of whate. g., this is
the number of man, that is the number of horse by imitating the figures
of living things with pebbles, like those who bring numbers into the forms
of triangle and square. (tr. based on the Revised Oxford Translation)
Since Eurytus is said to have been the student of Philolaus,18 it seems un-
likely that Aristotles source on this point would have been Philolauss
book. In fact, Theophrastus attributes this information to Archytas.19
I take it that these passages prove beyond doubt that Aristotle knew of
cases where different Pythagoreans held different views, and I take it as a
consequence that his information on at least these areas of Pythagorean
philosophy did not come from Philolaus or from any other single Pytha-
18 Iamblichus, VP 139, 148; references from Huffman (above, n. 3) 4.

19 Theophrastus, Met. 6a19 22. Theophrastus does not say that Aristotle used
Archytas as his source, but since Theophrastus uses the same examples (man,
horse) as Aristotle, it seems almost certain that Aristotle directly or indirectly
uses Archytas here, something not remarked on by Huffman, and an important
counterexample to his claim that Philolauss book was the only written source
that Aristotle used. See n. 13.
gorean source.20 This result establishes the possibility, if not the strong
likelihood, that Aristotle was indebted to a plurality of sources in other
cases where he reports disagreeing Pythagorean views, even where he
does not go out of his way to say that a particular view was held by
some Pythagoreans. In fact, we will shortly see that his accounts of Py-
thagorean views on the nature of first principles and on the relations be-
tween numbers and things contain a good deal of disagreement.
3. Aristotles Accounts of Pythagorean Principles:

A First Look
The following is a catalogue of the chief claims about principles and

about the relation between numbers and things that Aristotle attributes
to the Pythagoreans.
N1 The things that are, are composed of numbers.
De Caelo 300a14 17 (quoted in Greek above, p. 57): The same conse-
quence follows from putting the heaven together out of numbers, for
some people constitute all nature out of numbers, as some of the Pythagor-
eans do. (tr. based on the Revised Oxford Translation)
Metaphysics 990a21 2: tm !qihlm toOtom 1n ox sum]stgjem b j|slor.
This number out of which the cosmos is constituted. (tr. Revised Oxford
Translation)
Metaphysics 1080b17 19: 1j to}tou tr aQshgtr oqs_ar sumest\mai
vas_m tm cq fkom oqqamm jatasjeu\fousim 1n !qihl_m.
They say that sensible substances are constituted out of it [viz. number].
For they fabricate the entire heaven out of numbers. (tr. based on the Re-
vised Oxford Translation)
Metaphysics 1083b11 12 (C3.121): t d t s~lata 1n !qihl_m eWmai suc-
je_lema.
Bodies are composed of numbers. (tr. based on Revised Oxford Translation)
Metaphysics 1083b18 19 (C3.3): t coOm heyq^lata pqos\ptousi to?r
s~lasim r 1n 1je_mym emtym t_m !qihl_m.
20 By saying this, I do not mean to exclude the possibility that he used a single
work, such as a compendium of Pythagorean views, that was itself based on
more than one source.
21 This designation refers to the translation and discussion of the passages in ques-
tion below, in Section 7. The translation given in these cases is my own trans-
lation.
In fact they apply their propositions to bodies supposing that they consisted
of those numbers. (tr. based on Revised Oxford Translation)
Metaphysics 1090a22 3 (C4.4): 1po_gsam 1n !qilhl_m t emta.
They made the things that are, out of numbers. (my tr.)
Metaphysics 1090a32 (C4.6): poie?m 1n !qihl_m t vusij s~lata.
They make the natural bodies out of numbers. (tr. based on Revised Ox-
ford Translation)
N2 The things that are (or things) are numbers.
Metaphysics 986a21 (C1.13): !qihlor d], jah\peq eUqgtai, tm fkom oqqam|m.
The entire heaven, as has been said, is numbers. (tr. based on Revised Ox-
ford Translation)
Metaphysics 987b27 8: oR d( !qihlor eWma_ vasim aqt t pq\clata.
They say that the things themselves are numbers. (tr. Revised Oxford
Translation)
Metaphysics 1090a22 3 (C4.2): eWmai lm !qihlor 1po_gsam t emta.
They made the things that are, numbers. (my tr.)
N3 The things that are are number.22
Metaphysics 986a2 3 (C1.9): tm fkom oqqamm "qlom_am eWmai ja !qihl|m.
[They supposed] the entire heaven to be a musical scale and a number. (tr.
based on Revised Oxford Translation)
Metaphysics 1083b17 (C3.2): 1je?moi d tm !qihlm t emta k]cousim.
They say that the things that are, are number. (my translation)
N4 Things resemble numbers; the things that are, are by imitation of
numbers.
Metaphysics 985b27 31 (C1.3-C1.4): 1m d to}toir 1d|joum heyqe?m
bloi~lata pokk to?r owsi ja cicmol]moir, lkkom C 1m puq ja c0 ja
vdati, fti t lm toiomd t_m !qihl_m p\hor dijaios}mg t d toiomd
xuw^ te ja moOr 6teqom d jaiqr ja t_m %kkym r eQpe?m 6jastom blo_yr.
In these (viz. numbers) they thought they observed many likenesses to the
things that are and that come to be such and such an attribute of num-
bers being justice, another being soul and mind, another being the right
moment,23 and similarly for virtually all other things. (my tr.)
22 I suggest in n. 75 that there is no difference between N2 and N3.

23 I prefer this translation of jaiq|r to opportunity or due season.
Metaphysics 985b31 2 (C1.5): t_m "qlomi_m 1m !qihlo?r bq_mter t p\hg

ja tor k|cour.
They saw the attributes and the ratios of the harmonic intervals in numbers.
(tr. based on Revised Oxford Translation)
Metaphysics 985b32 3 (C1.6): $ lm %kka to?r !qihlo?r 1va_momto tm v}sim
!vyloi_shai psam.
All other things seemed to be made in the likeness of numbers in their en-
tire nature. (tr. based on Revised Oxford Translation)
Metaphysics 987b11 12: oR lm cq Puhac|qeioi lil^sei t emta vasm
eWmai t_m !qihl_m.
For the Pythagoreans say that the things that are, are by imitation of num-
bers. (tr. based on Revised Oxford Translation)
Metaphysics 1090a20 2 (C4.1): oR d Puhac|qeioi di t bqm pokk t_m
!qihl_m p\hg rp\qwomta to?r aQshgto?r s~lasim.
But the Pythagoreans, because they saw many attributes of numbers be-
longing to sensible bodies (tr. Revised Oxford Translation)
Metaphysics 1090a24 5 (C4.5): t p\hg t t_m !qihl_m 1m "qlom_ rp\q-
wei ja 1m t` oqqam` ja 1m pokko?r %kkoir.
The attributes of numbers are present in harmonic intervals and in the
heavens and in many other things. (tr. Revised Oxford Translation)
N5 The principles (!qwa_) of mathematical entities24 are the principles of
all the things that are.25
Metaphysics 985b25 6 (C1.1): tr to}tym !qwr t_m emtym !qwr ^hg-
sam eWmai p\mtym.
They believed that the principles of these things are the principles of all the
things that are. (my tr.)
N6 The elements (stoiwe?a) of numbers are the elements of all things.
Metaphysics 986a1 2 (C1.8): t t_m !qihl_m stoiwe?a t_m emtym stoiwe?a
p\mtym rp]kabom eWmai.
They supposed the elements of numbers to be the elements of all the things
that are. (tr. based on Revised Oxford Translation)
24 For this interpretation, see nn. 54, 55.

25 The sequel shows that Aristotle interprets these principles to be the principles of
number. See discussion of C1.8 below, p. 84.
N7 Number is a principle (!qw) both as matter and as attributes and

states.
Metaphysics 986a15 17 (C1.11): va_momtai d ja oxtoi tm !qihlm mol_-
fomter !qwm eWmai ja r vkgm to?r owsi ja r p\hg te ja 6neir.
Evidently, then, these people too consider that number is the principle
both as matter for the things that are and as their attributes and states. (tr.
N8 The principles are two: the limited and the unlimited ; the unlim-
ited itself and the one itself are the oqs_a of the things of which they
are predicated, and this is why the oqs_a of all things is number.
Metaphysics 987a13 19: (C2.1-C2.4): oR d Puhac|qeioi d}o lm tr !qwr
jat tm aqtm eQq^jasi tq|pom, tosoOtom d pqosep]hesam d ja Udi|m
1stim aqt_m, fti t pepeqasl]mom ja t %peiqom [ja t 4m] oqw 2t]qar
timr ^hgsam eWmai v}seir, oXom pOq C c/m E ti toioOtom 6teqom, !kk(
aqt t %peiqom ja aqt t 4m oqs_am eWmai to}tym m jatgcoqoOmtai,
di ja !qihlm eWmai tm oqs_am p\mtym.
The Pythagoreans similarly posited two principles but they added some-
thing peculiar to themselves, not that the limited and the unlimited are dis-
tinct natures like fire or earth or anything else of that kind, but that the un-
limited itself and the one itself are the oqs_a of the things of which they
are predicated. This is why they call number the oqs_a of all things. (tr.
N9 The oqs_a of something is the first thing it happens to belong to (the
number two is the essence of double).
Metaphysics 987a22 5 (C2.6): ja pq~t\ rp\qneiem b kewher fqor, toOt(
eWmai tm oqs_am toO pq\clator 1m|lifom, speq eU tir oUoito taqtm eWmai
dipk\siom ja tm du\da di|ti pq_tom rp\qwei to?r dus t dipk\siom.
They thought that the first subject of which a given term is predicable is the
oqs_a of the thing, as if one supposed that double and the number two
were the same because the number two is the first thing double applies
to. (tr. based on Revised Oxford Translation)
N10 The one is oqs_a; numbers are causes of the oqs_a of the other
things.
Metaphysics 987b22 5: t l]mtoi ce 4m oqs_am eWmai, ja l 6teq|m c] ti cm
k]ceshai 6m, paqapkgs_yr to?r Puhacoqe_oir 5kece, ja t tor !qihlor
aQt_our eWmai to?r %kkoir t/r oqs_ar sa}tyr 1je_moir.
He [Plato] agreed with the Pythagoreans in saying that the one is sub-
stance (oqs_a) and not a predicate of something else; and also in saying
that the numbers are the causes of the substance (oqs_a) of all other things.
(tr. based on Revised Oxford Translation)
N11 The elements (stoiwe?a) of number are the even (which is unlim-
ited) and the odd (which is limited); the one is from the even and
the odd, and number is from the one.
Metaphysics 986a17 21 (C1.12) toO d !qihloO stoiwe?a t| te %qtiom ja
t peqitt|m, to}tym d t lm pepeqasl]mom t d %peiqom, t d( 4m 1n
!lvot]qym eWmai to}tym (ja cq %qtiom eWmai ja peqitt|m), tm d( !qihlm
1j toO 2m|r.
The elements of number are the even and the odd, and of these the former
is unlimited, and the latter limited. And the one is from both of these
(for it is both even and odd) and number is from the one. (tr. based on
Revised Oxford Translation)
N12 When the one had been constituted, the nearest part of the un-
limited immediately began to be drawn in and limited by the limit.
Metaphysics 1091a15 18: r toO 2mr sustah]mtor, eUt( 1n 1pip]dym eUt( 1j
wqoir eUt( 1j sp]qlator eUt( 1n m !poqoOsim eQpe?m, eqhr t 5ccista toO
!pe_qou fti eVkjeto ja 1peqa_meto rp toO p]qator.
When the one had been constituted (whether from planes or from a sur-
face or from seed or from things which they are at a loss to say), the nearest
parts of the unlimited at once began to be drawn in and limited by the
limit. (tr. based on Revised Oxford Translation)
N13 The sustoiw_a (parallel columns)
Metaphysics 986a22 6 (quoted in Greek above, pp. 58 9): Others of this
same group say there are ten principles, which they arrange in parallel col-
umnslimit and unlimited, odd and even, one and plurality, right and
left, male and female, resting and moving, straight and curved, light and
darkness, good and bad, square and oblong. (tr. based on Revised Oxford
Translation)
N13 contrasts with the rest, and is marked as being the view of others
of this same group, that is, other Pythagoreans. The remaining texts are
concerned with numbers as principles of other things and with the prin-
ciples of numbers. They are an untidy collection. Their untidiness lends
itself to two ways of treating them, either as reflections of a single co-
herent ontology or as derived from different sources that represent dif-
ferent and incompatible views.
The first approach points towards a system of at least three tiers,26 as
follows.
26 Burkert describes a system along these lines at Burkert (above, n. 2) 32 ff. I pres-
ent a five-tier system below, p. 101.
Tier 1: the limited (or limit) and the unlimited, or the even and the odd
Tier 2: numbers
Tier 3: all things (that is, all things not included in Tiers 1 and 2)
No one of the claims N1-N12 sets out the entire system, but each of
them can be seen as revealing a part of it. The limited and the unlimited
(or the even and the odd) are the principles of numbers (some testimo-
nia favor the idea that the one is prior to and in some way a principle
of the other numbers27), which are in turn the principles of all things;
in a sense the limited and the unlimited are the principles of all things
too. The notion of principle involved is not spelled out or analyzed. It
may be metaphysical (in the sense of an essentially basic entity), genetic
(in the sense that the principles generate the things whose principles
they are), temporal (in the sense that the principles exist before the
things whose principles they are), material (in the sense that the things
are made up out of their principles), or a combination of these.
The overall picture that this approach produces is as follows. The
principles of numbers are the limited and the unlimited, or alternately
the even (which is unlimited) and the odd (which is limited). These
principles first generate (or compose or are the ingredients in the essence
of) the one. The one generates (or composes, etc.) the (other)
numbers. Numbers are the causes of the oqs_a of the remaining things,
which are generated by (or composed of, etc.) numbers, or are numbers,
or are by imitation of numbers. The variety of ways in which Aristotle
expresses the relation between numbers and things will reflect differen-
ces in his sources and also the difficulty he had in comprehending these
ideas and his fundamental lack of sympathy with them.
Where the first approach smoothes over disagreement among the
passages, the second approach emphasizes their discrepancies. Thus, it
is one thing to say (N2) that things are numbers, and quite another to
say (N1) that they are 1j numbers, whether that means that they are
constituted out of numbers as a statue is constituted out of bronze, or
that numbers generate them or something else. It is still another thing
to say (N6, cf. N5) that the principles or elements of numbers are the
principles or elements of all things, which is compatible with all things
being on the same level (whether ontological, causal, genetic or other-
wise) as numbers and not dependent on them at all. And yet another
thing to say (N4) that things are by imitation of numbers. We may
27 Metaph. 987a18 with discussion below, p. 100.

also doubt that N8s assertion that the oqs_a of all things is number and
N10s claim that numbers are causes of the oqs_a of the other things are
consistent. It is not clear how to unify N11s claim that the one is 1j
the even and the odd with N12s assertion that the Pythagoreans are at a
loss to say how the one was constituted. Or how the one is related
to the unlimited in N8: is the unlimited one of the principles of the
one, or is it identical with the one? Other inconsistencies are easy
to find in these claims as well. A straightforward way to account for
them would be to suppose that Aristotle is basing his accounts on several
sources which he considered to represent Pythagorean thought and that
they were broadly similar, yet importantly different.
These two approaches lead to two different assessments of Aristotles
treatment of his sources. On the first, the sources state their views im-
precisely and/or Aristotle does a poor job of understanding what they
say; on the second, the sources may well state their (contrasting)
views precisely and Aristotle may succeed in understanding their claims
or at least reporting them accurately (even though he finds it impossible
to make philosophical sense of them). The two approaches also lead to
two different assessments of early Pythagorean views on principles. The
first supposes that there is a Pythagorean orthodoxy behind Aristotles
various expressions (except for the outlier N13). A variation on this as-
sessment is that Aristotle is using only one source (perhaps Philolaus)
which is self-consistent. The second approach makes no such supposi-
tion. The different and inconsistent claims that Aristotle attributes to
the Pythagoreans are in fact more or less accurate expressions of different
Pythagorean beliefs.
This second approach is certainly worth considering and it is also
worthwhile to consider Philolaus in connection with it. If we do not
suppose from the start that Philolaus was Aristotles only source, or
that Aristotles source or sources, whoever or whatever they may
have been, need have told a coherent and consistent story, we are
free to look for variations, seeing Pythagoreanism as a movement rather
than as a monolithic theory (which would have been unique for its time;
we find no uniform set of doctrines among either the Milesians or the
Eleatics28) and even to evaluate Philolauss position within the move-
28 I name the two obvious candidates in early Greek philosophy for groups of
thinkers closely associated with one another. In fact, it is not infrequent even
now to read of the Milesian school and the Eleatic school of philosophy.
But the differences in the views of Thales, Anaximander, and Anaximenes are
ment. If his doctrines prove to be somewhat remote from the doctrines

that Aristotle describes, the explanation may be not that Aristotle was
inaccurate, but that he did not regard Philolaus as particularly represen-
tative of the movement. On the other hand, on due consideration it
may seem best to stick to some version of the traditional view, that Ar-
istotle meant in the main to attribute a single theory to the Pythagor-
eans, one that he found inconsistent and in ways obviously false. And
it may be best to decide that this single theory was that of Philolaus.
4. Philolauss Fragments on Principles
The main texts for Philolauss views on principles are the following.29
B1 " v}sir d( 1m t` j|sl\ "ql|whg 1n !pe_qym te ja peqaim|mtym, ja
fkor <b> j|slor ja t 1m aqt` p\mta.
Nature in the cosmos was joined from both unlimiteds and limiters, and the
entire cosmos and all the things in it.
B2 !m\cja t 1|mta eWlem p\mta C peqa_momta C %peiqa C peqa_momt\ te
ja %peiqa %peiqa d l|mom ouj !e. 1pe to_mum va_metai out( 1j peqaim|m-
tym p\mtym 1|mta out( 1n !pe_qym p\mtym, d/kom tqa fti 1j peqaim|mtym
te ja !pe_qym f te j|slor ja t 1m aqt` sumaql|whg. dgko? d ja t 1m
to?r 5qcoir. t lm cq aqt_m 1j peqaim|mtym peqa_momti, t d( 1j peqai-
m|mtym te ja !pe_qym peqa_momt_ te ja oq peqa_momti, t d( 1n !pe_qym
%peiqa vam]omtai.
It is necessary that the things that are be all either limiters or unlimiteds or
both limiters and unlimiteds; but not in all cases only unlimiteds. Now
since it is evident that they are neither from things that are all limiters
nor from things that are all unlimiteds, it is therefore clear that both the cos-
mos and the things in it were joined together from both limiters and un-
limiteds. The behavior of these things in turn makes it clear. For those
of them that are from limiters limit, those that are from both limiters
and unlimiteds both limit and do not limit, and those that are from unlim-
iteds will evidently be unlimited.
B3 !qwm oqd t cmyso}lemom 1sse?tai p\mtym !pe_qym 1|mtym.
well known: whatever the Milesians had in common, it was not that they ad-
hered to any specific doctrines. And the uniformity in the views of the Eleatics
is less than perfect. Notably, Aristotle is as impressed by the differences in the
views of Parmenides and Melissus as much as he is by their similarities (Aristotle
Phys. 184b16, 185b17 18; Metaph. 986b18 21).
29 I follow Huffmans text, but use my own translations, which differ from Huff-
mans in some respects.
There will not be anything that is going to know at all, if all things are un-
limited.
B4 ja p\mta ca lm t cicmysj|lema !qihlm 5womti oq cq bti_m
<oX|m> te oqdm oute mogh/lem oute cmysh/lem %meu to}ty.
And in fact all the things that are known have number. For it is not possible
for anything at all either to be comprehended or known without this.
B5 f ca lm !qihlr 5wei d}o lm Udia eUdg, peqissm ja %qtiom, tq_tom d
!p( !lvot]qym liwh]mtym !qtiop]qittom 2jat]qy d t_ eUdeor pokka loq-
va_, $r 6jastom aqt sgla_mei.
In fact, number has two proper kinds, even and odd, and a third kind,
even-odd, from both mixed together.30 Of each of the two kinds there
are many forms, of which each thing itself gives signs.
B6 peq d v}sior ja "qlom_ar de 5wei " lm 1st t_m pqacl\tym !_dior
5ssa ja aqt lm " v}sir he_am te ja oqj !mhqyp_mgm 1md]wetai cm_sim
pkm ca C fti oqw oX|m t( Gm oqhem t_m 1|mtym ja cicmysjol]mym rv(
"l_m cem/shai l rpaqwo}sar tr 1stoOr t_m pqacl\tym, 1n m sum]sta
b j|slor, ja t_m peqaim|mtym ja t_m !pe_qym. 1pe d ta !qwa rpqwom
oqw blo?ai oqd( bl|vukoi 5ssai, Edg !d}matom Gr ja aqta?r joslgh/mai, eQ
l "qlom_a 1pec]meto timi_m #m tq|p\ 1c]meto. t lm m blo?a ja
bl|vuka "qlom_ar oqdm 1ped]omto, t d !m|loia lgd bl|vuka lgd
Qsotaw/ !m\cja t toiaOta "qlom_ sucjejke?shai, eQ l]kkomti 1m j|sl\
jat]weshai.
Concerning nature and "qloma this is how it is: the being (1st~) of things,
which is eternal that is, in fact, their very nature admits knowledge that
is divine and not human, except that it was impossible for any of the things
that are and are known by us, to have come to be if there did not exist the
being (1st~) of the things from which the cosmos is constitutedboth the
limiters and the unlimiteds. But since the principles are not similar or of the
same kind, it would be completely impossible for them to be brought into
order if "qloma had not come upon them in whatever way it did. Now
things that are similar and of the same kind have no need of "qloma to
boot, but those that are dissimilar and not of the same kind or of the
same speed[] must be connected together in "qlomai if they are going
to be kept in an orderly arrangement.
B7 t pqtom "qlosh]m, t 4m 1m t` l]s\ tr sva_qar, 2st_a jake?tai.
The first thing that was joined, the one in the middle of the sphere, is called
the hearth.
B17 b j|slor eXr 1stim, Eqnato d c_cmeshai %wqi toO l]sou ja !p toO
l]sou eQr t %my di t_m aqt_m to?r j\ty, <ja> 5sti t t %my toO
l]sou rpemamt_yr je_lema to?r j\ty. to?r cq j\ty t jatyt\ty l]qor
30 I follow J. Barnes, The Presocratic Philosophers 2 (London 1982) 632 n. 31, in de-
leting this phrase.
1stm speq t !myt\ty ja t %kka sa}tyr. pqr cq t l]som jat

taqt\ 1stim 2j\teqa, fsa l letem^mejtai.
The cosmos is one. It began to come to be right up at the middle, and from
the middle <it came to be> in an upward direction in the same way as it
did in a downward direction, and the things above the middle are symmet-
rical with those below. For in the lower <regions> the lowest part is like
the highest part <in the upper regions>, and similarly for the rest. For both
<the higher and the lower> have the same relationship to the middle, un-
less they have been moved to another location.
5. Similarities and Differences between Philolaus and Aristotle
Philolauss account differs from Aristotles in a number of ways, of

which I will mention those that seem to me potentially significant.
First, Philolaus speaks of peqa_momta and %peiqa, limiting things and
unlimited things, or alternatively things that are limiting (or things that
limit) and things without limit, or limiters and unlimiteds (B1, B2, B3,
B6). These are the things out of which the cosmos is constituted or put
together (sum]sta) (B6). Philolauss account differs from the Aristoteli-
an material first in using plural active forms (limiters and unlimi-
teds) as opposed to the singular and frequently the passive in Aristotles
reports (the limited or limit, and the unlimited) (N8, N11, N12,
N13), which may point to a tendency to conceive of those entities in
more concrete terms than Aristotles account suggests: things that
limit as opposed to the more abstract expression limit; things that
are unlimited as opposed to the unlimited.31
Second, Philolaus holds that the cosmos is constituted or put togeth-
er out of the limiters and unlimiteds (B1, B6), as if they are physical
constituents, whereas Aristotle calls the one (seemingly identical with
limit) and the unlimited the oqs_a of the things of which they are pre-
dicated (N8).
Third, Philolaus places great weight on harmonia, while Aristotles
accounts make no mention of harmonia except in the sense of musical
harmonia, an attunement of a lyre according to fixed harmonic intervals,
corresponding roughly to our musical scale.32 In particular this holds for
the statement that they supposed the entire heaven to be a harmonia and
31 See Huffmans discussion (above, n. 3) 39 40.

32 Metaph. 985b31 (= C1.5), 986a3 (= C1.9), 1087b35, 1090a24 (= C4.5),
1093a14.
a number.33 In Philolaus it is a much broader notion. Limiteds and un-

limiteds are joined together (the verbs he uses are "ql|tteim, sumaql|t-
teim, which are related to the noun harmonia) by harmonia, and harmonia
must come upon (1pic_cmeshai) them (B6). Philolaus does use harmo-
nia in the sense of musical harmonia as well, prominently in fragment B6a
(not translated here), but the other fragments cited give it a wider sense
and a more central role in Philolauss metaphysics and cosmology. In
Philolauss view, the opposing natures of limiters and unlimiteds consti-
tute an obstacle to their uniting to form the cosmos and the things in the
cosmos. A third ingredient, harmonia, is needed to effect this union. On
the other hand, Aristotles account gives no hint that the way in which
things are constituted out of limit and the unlimited was problematic or
that harmonia has any role in this regard.
The fourth and most striking discrepancy between Philolauss account
and Aristotles concerns the role of numbers. In Aristotle numbers are
very prominent. They are generated from or otherwise based on the
even and odd, which are associated somehow with the unlimited and
the limited (N11), and things are then in turn dependent on numbers.
According to his most frequent formulations, they are constituted out of
numbers (N1) or they are numbers (N2). Indeed he gives the role of
numbers in the cosmos more prominence than any other feature of Py-
thagorean thought. By contrast, the fragments of Philolaus give the lead
roles to limiters and unlimiteds, not to numbers. It is not clear whether or
how numbers are derived from these basic entities. Numbers are impor-
tant, to be sure. Each thing gives signs of numbers (B5), and all the
things that can be known have number (B4). But what the cosmos
and all its inhabitants are constituted out of are limiters and unlimiteds,
not numbers (B1, B2). I will have more to say on this subject just below.
Now that these general points have been made, it is time to consider
the relations between Aristotles testimonia and the fragments of Philo-
laus in greater detail. To begin with, nothing in the extant fragments of
Philolaus has any direct connection with N1, N2, N3, N4, N7, N9,
N10, or N13. According to Philolaus the principles are limiters and un-
limiteds, not numbers (B1, contrast N7). Things are joined together out
of these principles, not from numbers (B1, B2; contrast N1). Further,
Philolaus does not say that all things are his principles (especially
B2, contrast N2, N3) or that they are by imitating the principles or
33 With Metaph. 986a3, cf. Cael 2.9.

that they resemble them (contrast N4), or that the being34 of a thing is
the first thing it happens to belong to (B6, contrast N9) or that the
one and the other numbers are the 1st~ of anything at all (B6, contrast
N10). Finally, the doctrine of the parallel columns (N13) is remote from
anything Philolaus says.35
Philolauss belief that all the things that are known have number
(B4) is importantly different from N1, N2, and N3, first because B4
concerns all the things that are known, not all things in general,
and second because B4 says that they have number, not that they
are or are constituted out of number or numbers. Nor does Philo-
laus talk of things imitating numbers or having likenesses to numbers
(N4). B4 gives the impression that having number is a necessary con-
dition for a thing to be known, but not a necessary and sufficient con-
dition for a thing to be, and for this reason Huffman sees the role of
number as epistemological rather than ontological.36 However, this
may be too strong; the one need not exclude the other. B4 makes it
plausible that Philolaus supposed that a thing that is known is known
because it has number and that it is known because of the number that
it has. Further, it is possible that he held that it is a things possession
of number (whatever that may mean and however the number it pos-
sesses may be related to the limiters and unlimiteds that constitute that
thing) that constitutes the thing as knowable. If this is correct, it leaves
open the possibility that number has an important ontological role
alongside its epistemological role.37 But even so, Aristotles reports on
the place of number in Pythagorean thought remain remote from
what we learn from B4.
Philolaus offers an ontology with at least two38 tiers: on the top tier
are the limiters and the unlimited, which are first principles, while the
bottom tier is constituted of what Philolaus describes as nature in the
34 Philolauss word is 1st~ rather than oqs_a. See discussion of N8 below,

pp. 72 6.
35 For this discrepancy, see Huffman (above, n. 3) 47 n. 1.
36 See Huffman (above, n. 3) 58, also 39, and his discussion of what having a
number means, on 64 73. My own opinion is that Philolauss ultimate unlim-
ited principles (whether there are one or many), which certainly do exist, do
not have number.
37 I cannot go further into this important topic in the present paper. I hope to ex-
plore it further in the future.
38 A case can be made for an indefinitely large number of tiers, but I do not have
the space to discuss this issue at present.
cosmos and the entire cosmos and all the things in it (B1). The es-
sential question here is where do numbers fit inare they on the top
tier, the bottom tier, somewhere in between, or are they not included
in the entities B1 refers to? Unfortunately the fragments provide insuf-
ficient help. The terminology of limiters and unlimiteds is so abstract
that it is not at all certain what it refers to,39 and it is quite unclear
whether the entire cosmos and all the things in it includes numbers
in addition to (what we regard as) more concrete objects.40
Further on this point, N12 is usually taken to describe the genera-
tion of numbers by the action of limit on the unlimited. If this is the
correct interpretation, it has no connection with Philolauss surviving
fragments (although this does not prove that he did not hold the
view). In any case, in N12 the one and the other generated entities
are not principles, so that numbers cannot belong to the top tier. How-
ever, there are some difficulties. In the first place, as N12 says, how the
one is constituted was not made clear. Next, the nature of the relation
between the one and the subsequently generated entities is obscure.
More radically, it is not even certain that N12 describes the generation
of numbers at all. Huffman has proposed an interpretation according to
which N12 has nothing to do with numbers; the one refers to the
central fire, and the passage is an account of the first stage of a cosmog-
ony.41 And most importantly, to assume that there is any connection be-
tween N12 and Philolaus would beg the question, since this is the very
subject under investigation.
39 Huffman (above, n. 3) 40 1.
40 An interpretation which I hope to develop in the near future is that for Philo-
laus the number of a thing is (in a suitably general sense) the measure of the
magnitudes of the entities (limiters and unlimiteds) that constitute an entity.
If this is correct, then for Philolaus things are not constituted out of numbers
or identical with numbers and do not resemble numbers, but it is possible
that he held that in some sense the being of a thing is or involves its number.
41 Huffman (above, n. 3) 203 4 with 211 14. However, in Burkerts view, to
the Pythagoreans number philosophy is cosmogony (above, n. 2, 36 and dis-
cussion 31 40). The generation of the number series is identical with the gen-
eration of the cosmos. (Cf. Aristotle, Metaph. 986a2 3: They supposed the
entire heaven to be a musical scale and a number.) Burkert is in agreement
with Ross: The Pythagoreans are trying to describe at one moment the forma-
tion of the number system and that of the physical universe, which is just what
on their principles they were bound to do. The number One is identified with
the central fire, as two was with the earth and seven with the sun. See W. D.
Ross, Aristotles Metaphysics (Oxford 1924) vol. 2, 484.
If Philolauss principles and derived entities do not include numbers,

it follows, barring some gross misunderstanding on the part of Aristo-
tle,42 that none of the Aristotelian texts that refer to numbers or their
principles is based on Philolaus. This will exclude, in addition to the
fragments already discussed (N1, N2, N3, N4, N7, N9, N10, and
N13), N5, N6, N11, N12, and some of the claims in N8. In other
words, if this is true, then virtually none of Aristotles remarks have
to do with Philolaus.43
42 Huffman argues that Aristotle did grossly misunderstand Philolaus. He is al-

most certain that N12 refers to Philolaus B7 (The first thing fitted together,
the one in the center of the sphere, is called the hearth), which, Huffman ar-
gues, is the beginning of Philolauss account of the generation of the cosmos
from limiters and unlimiteds. N12 could very easily be understood as a com-
mentary on this fragment. (Quotations from Huffman, above, n. 3, 62.) Huff-
man goes on to argue that Aristotles claims that the Pythagoreans believed N1
and N2 are misinterpretations of the cosmology. However, on the Ross-Bur-
kert view (see n. 41), it is possible to agree with Huffman that N12 refers di-
rectly to Philolaus B7, and yet see it as describing the first stage in the gener-
ation of the number series (as well as being the first stage in the generation of
the cosmos).
43 Huffman, on the other hand, finds that all the basic conceptual terms which
Aristotle assigns to the Pythagoreans are also found in Fragments 1 7 of Phi-
lolaus these concepts of limit, unlimited, harmony, number (divided into the
even, the odd, and the even-odd) and the generation of the cosmos around a
central fire; Huffman (above, n. 3) 28 9. Likewise, it seems pretty implau-
sible to me to suppose that Aristotle is basing his account of Pythagoreanism on
a book by a different Pythagorean, of whom we have no knowledge, and who
used all the same concepts as Philolaus, but who differed by explicitly equating
the even and odd with limiters and unlimiteds. Certainly, it is more reasonable
to explain Aristotles difference from what is found in the fragments of Philo-
laus as the result of his interpretation of what is going on there, especially since
we have ample evidence that Aristotle typically presents the views of others
under a heavy interpretation (ibid. 183). Whether this is a strong argument
for the thesis that Philolauss treatise is Aristotles source for the Pythagoreans
will depend on how closely Aristotles reports of Pythagorean views tally
with Philolauss fragments. If there is substantial disagreement, then Aristotles
source is unlikely to be Philolaus. Even a great deal of agreement does not
prove that Philolaus was Aristotles source in view of the real possibility that
what Aristotle and Philolaus have in common was passed down to Aristotle
through other Pythagorean sources. In fact, the surest evidence that Aristotle
used Philolaus would be cases where he agrees with Philolaus and disagrees
with other Pythagorean sourcesbut the absence of relevant early Pythagorean
sources other than Philolaus prevents us from identifying such cases.
On the other hand, if Philolaus includes mathematical entities, and

numbers in particular, in the classification the entire cosmos and all the
things in it, some of the Aristotelian material fits well with Philolauss
views on principles. N5 is a clear case. Limiters and unlimiteds will be
the principles of mathematical entities as well as of other entities, and so
they will be the principles of all the things that are (excluding the
principles themselves, of course). The same holds for N6 to the extent
that N6 is a restatement of N5.44 Some headway can be made towards a
Philolaic reading of N11 and N12, but problems still remain, principally
that the generation of entities described there does not obviously accord
with Philolauss ontology.
On the state of the evidence, it seems that Aristotles testimonia on
Pythagorean principles have little basis in Philolauss doctrines. Could
they, then, have come from other sources? Even though, as Huffman
shows, it is unclear what other sources than Philolauss book were at Ar-
istotles disposal, this problem is far less difficult than the problems that
the Philolaic interpretation brings, namely, why are so many of Aristo-
tles testimonia so remote from what we know of Philolaus and why
does Aristotle not attribute the relevant views to Philolaus by name?
6. Aristotles Other Sources
I will now present a consideration that points to an answer to this re-

maining question. As we know, some people and quite a few philoso-
phers lived to ripe old ages in ancient Greece. Xenophanes reached at
least 92, Democritus 90, Plato 80. The Macrobii, a work attributed to
Lucian, devotes three chapters (18 20) to enumerating long-lived phi-
losophers. One of them is Xenophilus, who was one of the Pythagor-
eans that Aristoxenus met45 and who is reported to have lived in Athens
to the age of 105 years. Interestingly, if Democritus was born in 460, he
was 76 when Aristotle was born in Stagira, not far from Abdera. It is
entirely possible from the chronological point of view that as a boy Ar-
istotle personally met Democritus, who we think of as a presocratic phi-
losopher! Could he then have met any Pythagoreans?
Huffman puts Philolauss birth between 470 and 440 and presents
evidence that he was still alive in 388, just four years before Aristotles
44 See p. 103.
45 D.L. 8.46.
birth.46 This would provide 388 as a terminus post quem for his death, but
it offers no evidence at all for how much later than 388 he died. In fact
the evidence for dating his birth as early as 470 is weak. It consists prin-
cipally of statements that Philolaus was the teacher of Eurytus and that
he and Eurytus were the teachers of the last of the Pythagoreans, who
include Echecrates, who is a (younger? older?) contemporary of Plato
(born c. 428).47 Huffman discusses the evidence, but makes some as-
sumptions which, although they are in general reasonable, need not
hold in a given case. For example, he assumes that teachers are older
than their students, and allows a ten or twenty year age difference.
But it is not always true that teachers are older than their students,
and it frequently happens that if they are older they are not much
older. Moreover, it is not always clear what teacher and student
mean in ancient contexts. The claim is typically of the form so-and-
so heard (Ejouse or di^jouse) so-and-so or so-and-so was a hearer
(!jqoat^r) of so-and-so. If Philolaus had written a philosophical book
and expounded it to others, might that not be enough to qualify them as
having heard him, no matter what their relative ages? And is there
any good reason to think that Philolaus could not have written such a
book by the time he was twenty or thirty? So if Philolaus was in fact
born as late as 440, he would have had plenty of time to have been
heard by the people who are said to have heard him. In addition,
he would have been only 52 in 388, and only 73 when Aristotle
came to Athens in 367. He might have lived another twenty years be-
yond thatand attended Platos funeral.48 On this accounting there is
46 Huffman (above, n. 3) 2 5.
47 See Plato, Phaedo 57a-b.
48 The following true story is indicative of how far back we can go in time in only
a few steps. My fourteen-year-old daughter has an English friend who once told
me that her grandfathers grandmothers gardener, who was a very old man, re-
ported that his grandfather said that the admiral was a very great man indeed.
Then she asked me who the admiral in question was. I ventured a guess: Nel-
son? No, she said. It was Drake. So in a very few steps ( just four jumps in
time back from my daughters friend) we are back in touch not with the hero of
Trafalgar, whose two hundredth anniversary is being celebrated this year
(2005), but with contemporaries of the man who destroyed the Spanish Armada
in 1588. We can imagine how this works out without any stretch of the imag-
ination. My daughter was born in 1991. My daughters friend was born in 1924.
Her father was born around 1885. So her grandfather could have been born
around 1835. His grandmother could have been born eighty years earlier in
1755. The gardener could have been 85 years older than her, so born in
no reason why Aristotle could not have met Philolaus himself. The case
is even easier to make for the possibility of his meeting later Pythagor-
eans such as Simmias, Cebes, and Echecrates, together with the other
members of the group known as the last generation of Pythagoreans, es-
pecially the long-lived Xenophilus, who for all we know was Aristotles
next door neighbor in Athens. In any case if Aristoxenus, whose birth is
placed in the period 375 360 (which makes him nine to twenty-four
years younger than Aristotle), met these last Pythagoreans, there can
be no objection to supposing that his teacher Aristotle could have
met them too. He surely had contact with Aristoxenus himself, and
in the context of the Academy and the Lyceum there may have been
no shortage of people who, if not Pythagoreans themselves, were famil-
iar with their views. Aristotle might even have used Aristoxenus book
on the Pythagoreans as his source.
Clearly, Aristotle found out about Pythagorean views from some-
where. Either he had access to written texts now unknown or he
may have found out from one or more people who claimed to know.
He knew about and reported some views that are unlikely to have
come from Philolauss book, for example the reports about Eurytuss
pebbles (which is in fact said to have been reported by Archytas49)
and the doctrine of the parallel columns. And it is far from clear that
Philolaus pursued the program of identifying numbers with such things
as justice, marriage, and the right moment (Metaph. 1078b22 3).50 If
the doctrine of the ten planets (including the counter-earth) moving
round the central fire belongs to Philolaus,51 then the view Aristotle at-
tributes to some Pythagoreans, that the comet is one of the plan-
1670. And his grandfather could have been born eighty years earlier in 1590.
When he was a boy he would surely have heard from his family and acquain-
tances about the greatness of the Admiral, who died in 1595. A similar concat-
enation of acquaintances could mean that Aristoxenus informant Xenophilus
and Philolaus as well could have associated with Pythagoreans who had
known Pythagoras personally.
49 See n. 19.
50 Huffman (above, n. 3) 286 8 floats the possibility, but his discussion is far from
proving the case.
51 Aristotle (attributes this cosmology to the Pythagoreans but Atius attributes
it to Philolaus in particular, Atius 2.7.7 (= Stobaeus, Ecl. 1.22.1d, DK 44 A16),
3.11.3 (= [Plutarch], Placita 895E; DK 44 A17), 3.13.2 (= [Plutarch], Placita
896 A, DK 44 A21).
ets52 cannot be his unless he is guilty of gross inconsistency. The same is

very likely to hold for a number of other cases, quoted above in Section
2, where Aristotle attributes views to some Pythagoreans, as if to
imply that there are other Pythagoreans who do not share them. In ad-
dition, Aristotle reports the view that the universe and all that is in it is
determined by the number 3, since beginning, middle, and end give the
number of the universe and the number they give is the triad
(Cael. 268a11), which seems to conflict with the importance the num-
ber 10 is given in the cosmology that includes the counter-earth. It may
be significant that at one point he refers to some of the Pythagoreans as
constituting all nature out of numbers and putting the heaven together
out of numbers (Cael. 300a14 17) (as if some Pythagoreans did not) and
at another he talks about those of the Pythagoreans who declare the
one to be the element and principle of the things that are; those people
suppose that numbers have magnitude. (Metaph. 1080b30) (again, as if
not all Pythagoreans shared that view about principles).
Could this information have come from Philolauss book? I think it
unlikely, since the surviving fragments do not give the impression that
Philolaus was concerned to set out the views held by all Pythagoreans
past and present, or any views other than his own. It seems to present
and argue for Philolauss own theories rather than provide a digest of
viewseven conflicting viewsheld by various Pythagoreans. In
these circumstances, the conclusion that Aristotle somehow or other
had access to a separate written or oral tradition becomes attractive,
maybe even inescapable.
7. Aristotles presentation of Pythagorean views
An examination of the passages in which these testimonia occur bears

out Huffmans remark that Aristotle tells a story about Pythagorean
views on principles.53 In some cases we can deconstruct the story and
52 Aristotle, Mete. 342b30. Huffman (above, n. 3) 240 recognizes the incompati-

bility and comments that the field of astronomy is one of the areas of Pytha-
gorean thought where we have clearest evidence of divisions among Pythagor-
eans (ibid., 239), but he does not acknowledge the problem this raises for his
thesis that Philolaus was Aristotles source.
53 It is clear that he is not just presenting a list of Pythagorean doctrines, but rath-
er telling a story that attempts to make sense of the Pythagorean outlook as a
see how Aristotle put it together to form the interpretation that he uses
in his own discussions. The following account of the four passages from
the Metaphysics that contain Aristotles discussions (as opposed to isolated
remarks) of Pythagorean views on principles and on the relation be-
tween numbers and things (C1: 985b23 986a21, C2: 987a13 27,
C3: 1083b8 13 & b17 19 and C4: 1090a20 5, a32 5) aims to
offer a precise translation and a commentary that explicates Aristotles
train of thought in his account of Pythagoreanism. The interpretations
given in the commentary will provide the basis of the subsequent at-
tempt to understand Aristotles remarks on Pythagorean principles and
on the relations between numbers and things (N1-N13) and to assess
their trustworthiness (Section 8).
C1 Metaph. 985b23 986a21

1m d to}toir ja pq to}tym oR jako}lemoi Puhac|qeioi
t_m lahgl\tym "x\lemoi pq_toi taOt\ te pqo^cacom, ja
1mtqav]mter 1m aqto?r [C1.1] tr to}tym !qwr t_m emtym !qwr 25
^hgsam eWmai p\mtym. 1pe d [C1.2] to}tym oR !qihlo v}sei
pq_toi, [C1.3] 1m d to}toir 1d|joum heyqe?m bloi~lata pokk
to?r owsi ja cicmol]moir, lkkom C 1m puq ja c0 ja
vdati, fti [C1.4] t lm toiomd t_m !qihl_m p\hor dijaios}mg
t d toiomd xuw^ te ja moOr 6teqom d jaiqr ja t_m %k- 30
kym r eQpe?m 6jastom blo_yr, 5ti d [C1.5] t_m "qlomi_m 1m !qih-
lo?r bq_mter t p\hg ja tor k|cour, 1pe d [C1.6] t lm %kka
to?r !qihlo?r 1va_momto tm v}sim !vyloi_shai psam, [C1.7] oR
d( !qihlo p\sgr t/r v}seyr pq_toi, [C1.8] t t_m !qihl_m stoi- 986a1
we?a t_m emtym stoiwe?a p\mtym rp]kabom eWmai, ja [C1.9] tm
fkom oqqamm "qlom_am eWmai ja !qihl|m ja [C1.10] fsa eWwom
blokoco}lema 5m te to?r !qihlo?r ja ta?r "qlom_air pqr
t toO oqqamoO p\hg ja l]qg ja pqr tm fkgm diaj|- 5
slgsim, taOta sum\comter 1v^qlottom. j#m eU t_ pou
di]keipe, pqoseck_womto toO sumeiqol]mgm psam aqto?r eWmai
tm pqaclate_am k]cy d( oXom, 1peid t]keiom B dejr
eWmai doje? ja psam peqieikgv]mai tm t_m !qihl_m v}sim,
ja t veq|lema jat tm oqqamm d]ja lm eWma_ vasim, 10
emtym d 1mm]a l|mom t_m vameq_m di toOto dej\tgm tm
!mt_whoma poioOsim. di~qistai d peq to}tym 1m 2t]qoir
Bl?m !jqib]steqom. !kk( ox d w\qim 1peqw|leha, toOt| 1stim
fpyr k\bylem ja paq to}tym t_mar eWmai tih]asi tr
!qwr ja p_r eQr tr eQqgl]mar 1lp_ptousim aQt_ar. va_- 15
momtai d ja oxtoi tm !qihlm [C1.11] mol_fomter !qwm eWmai ja
whole. This story is necessarily selective and leaves out important points in
order to preserve coherence; Huffman (above, n. 3) 179.
r vkgm to?r owsi ja r p\hg te ja 6neir, [C1.12] toO d !qihloO

stoiwe?a t| te %qtiom ja t peqitt|m, to}tym d t lm pe-
peqasl]mom t d %peiqom, t d( 4m 1n !lvot]qym eWmai to}-
tym (ja cq %qtiom eWmai ja peqitt|m), tm d( !qihlm 1j 20
toO 2m|r, [C1.13] !qihlor d], jah\peq eUqgtai, tm fkom oqqam|m.
Those called Pythagoreans took hold of mathematics; they were the first to
advance that study, and being brought up in it they believed that [C1.1] the
principles of these things54 are the principles of all the things that are. Since
[C1.2] numbers are naturally the first among mathematical entities,55 and
[C1.3] in these [viz. numbers] they thought they observed many likenesses
to the things that are and that come to bemore than in fire and earth and
water[C1.4] such and such an attribute of numbers being justice, another
being soul and mind, another being the right moment, and similarly for
virtually all other things; since, again, [C1.5] they saw the attributes and the
ratios of the harmonic intervals in numbers; since, then, [C1.6] all other
things seemed to be made in the likeness of numbers in their entire nature
and [C1.7] numbers seemed to be primary in all nature, they supposed
[C1.8] the elements of numbers to be the elements of all the things that
are, and [C1.9] the entire heaven to be a musical scale56 and a number.
And [C1.10] all the characteristics of numbers and harmonic intervals
that they found corresponding to the attributes and parts of the heaven
and to the entire ordering they collected and made them fit. If anything
was missing anywhere they eagerly filled in the gaps to make their entire
system coherent. For example, since they think the number ten is some-
thing perfect and encompasses the entire nature of numbers, they declare
that the bodies that move in the heaven are also ten, but since only nine
are visible they invent the counter-earth as the tenth. I have discussed
these matters more precisely elsewhere.
My purpose in pursuing this topic here is that we may understand for
these people too what they posit to be the principles and how they fall
under the causes I have mentioned. Evidently, then, these people too con-
sider that [C1.11] number is the principle both as matter for the things that
are and as their attributes and states, and they hold that [C1.12] the ele-
ments of number are the even and the odd, and of these the former is un-
limited, and the latter limited. The one is from (1j) both of these (for it
is both even and odd) and number is from (1j) the one; and [C1.13]
54 The principles of mathematics are definitions, axioms, and hypotheses; but as

the sequel shows (C1.8, C1.11-C1.12) Aristotle is here talking not about the
principles of mathematics but about the principles of mathematical entities.
See below, p. 83.
55 Since numbers have principles C1.8, they cannot be first principles, as Rosss
translation (Since of these principles numbers are by nature the first, in the
Revised Oxford Translation) has it. The antecedent of totym in b26 is to
be inferred (as with to}tym in b25) from t_m lahgl\tym (b24).
56 For this translation of "qlom_a, see below, n. 73.
the entire heaven, as has been said, is numbers. (tr. based on the Revised
Oxford Translation, with significant modifications)
C1 begins with the well known statement that the Pythagoreans were
the first to make serious progress in mathematics. The fact that that
modern scholarship has rejected57 this claim is a serious blow to the
credibility of Aristotles account of Pythagorean principles, since he
bases it on their alleged interest in mathematics. Still, one need not be
a serious mathematician in order to have a serious interest in numbers;
indeed a fascination with numbers and their power is widespread among
peoples who have not developed mathematics to any notable degree.58
Aristotle gives a tidy picture. The Pythagorean interest in mathe-
matics led them to conclude that the principles of mathematical entities
are the principles of all the things that are C1.1. They reasoned that
since numbers are the primary kind of mathematical entities C1.2 and
numbers are the first things in the whole of nature C1.7,59 the elements
of numbers are the elements of all the things that are C1.8, which is a
restatement of C1.1 (with elements apparently used as a synonym of
principles60). Aristotle mentions some of the connections the Pytha-
goreans discovered between numbers and things which could be used
in support of C1.7: first, likenesses between numbers and the things
that are C1.3, with examples C1.4, and second, attributes and ratios of
harmonic intervals which they discovered in numbers C1.5. Conse-
quently (d^)61 all other things appeared to be made in the likeness of
numbers C1.6, and numbers seemed to be primary in all nature C1.7.
The first thing to notice is that Aristotle presents two different kinds
of grounds for the view that the principles or elements of number are
the principles or elements of all things. First, C1.3-C1.4 that they
thought that they observed that some attribute of numbers was justice,
another attribute was soul and mind, and still another was the right
moment. Second, C1.5 that they saw the numerical properties of
harmonic intervals. How large was the difference between the second
57 Following Burkerts lead (above, n. 2) part VI.

58 For examples see Burkert (above, n. 2) 468 ff.
59 C1 makes no distinction between things that are and things that are in nature.
All entities are natural entities. Further the passage implies that for the Pytha-
goreans mathematical entities are natural objects. Numbers are the first things
in nature C1.7 and they are also the first mathematical entities C1.2.
60 See p. 85.
61 Or possibly indeed, in which case C1.6 is not presented as a consequence of
C1.4 and C1.5, but as further evidence for C1.7.
kind of grounds and the first can be judged from Alexanders explana-
tion in his comment on this passage.
They supposed that requital and equality were characteristic of justice and
they found these features in numbers, and so declared that justice was the
first number that is equal-times-equal.They said that the right moment
is the number 7, since things which are natural appear to have their decisive
moments of fulfillment in birth and growth by sevens. Humans, for exam-
ple. They are born in the seventh month and teethe in as many months,
and reach adolescence in the second span of seven years and get a beard
in the third They said that marriage is the number 5, because marriage
is the union of male and female, and according to them the odd is male and
the even is female, and this number is the first which has its origin from 2,
the first even number, and 3, the first odd. They declared mind and sub-
stance to be the one,62 since he spoke of the soul as the mind. They said
that because it is stable and similar in every way, and sovereign, the mind is
the unit and the one; and substance is the first thing. (Alexander in Met-
aph. 38,10 39,15)
The ways in which justice, the right moment, marriage, soul/mind
and substance are associated with numbers are very different from one
another, and can justifiably be called fanciful (cf. Aristotles they thought
that they observed [1d|joum heyqe?m]). They have virtually nothing to
do with mathematics. It appears that the Pythagoreans were determined
to make things fit their theory by hook or by crook. On the other hand,
the attributes and ratios of the harmonic intervals (2:1 for the octave,
3:2 for the fifth, 4:3 for the fourth) are derived systematically, have a
basis in nature (cf. they saw [bq_mter] and can be used to make certain
kinds of mathematical calculations.63 On the basis of this kind of evi-
dence, goes Aristotles story, the Pythagoreans concluded that all
other things appeared in their entire nature to be likened after numbers
C1.6. This broad generalization is the kind of move Aristotle attributes
to other presocratic philosophers as well.64
The second thing to notice is that Aristotle says neither that the Py-
thagoreans held that justice is a number but that such and such an at-
62 In this context, the one must mean the number one.

63 For example, the octave C-C* is the sum of the intervals C-G and G-C* (re-
spectively a fifth and a fourth), and the ratio of the octave is the product of the
ratios of the fifth and fourth (2:1 = 3:2 x 4:3).
64 Cf. Aristotles conjectures about Thaless grounds for claiming that water is the
material principle of things (Metaph. 983b22 27, quoted below, p. 114).
tribute of numbers is justice,65 nor that they claimed that the attributes
and ratios of harmonic intervals are numbers, but that they saw those at-
tributes and ratios in numbers, by which I think he must mean that they
saw that numbers have the same attributes and ratios as are found in the
harmonic intervals. Passage C1 makes better sense if we keep this in
mind. At this point he is recounting the grounds on which the Pytha-
goreans founded their strange view that the principles of numbers are
the principles of all things. The numerical basis of the harmonic intervals
was a real discovery, not a fancy. Also, the common points between,
say, justice and the number four, although perhaps arbitrarily chosen,
are comprehensible; and these discoveries are described in a way that
does not commit the Pythagoreans from the outset to any strange
views at all. The strangeness comes in first with C1.6 and more so
with the conclusions in C1.7 and C1.8.
C1.6 is not simply a generalization of C1.3. The earlier claim that
numbers have likenesses (bloi~lata) to things is symmetric. If A is like
B in some respect, then B is like A too, in the same respect. In particular
if some thing and some number have some attribute in common, so that
the number has a likeness to the thing,66 then by the same token the
thing has a likeness to the number. But being made in the likeness of
(!voloioOshai) is an asymmetric relation: if A is made in the likeness
of B, then B is in some way the original and A the copy.
C1.7 contains an important element of Aristotles account. Since for
Aristotle nature involves change whereas numbers are unchanging,
numbers are not part of nature;67 for the Pythagoreans they are. This
distinctively Aristotelian use of nature appears here as a key word
in his description of the Pythagorean view; that they included numbers
as part of nature will follow from the view that they considered numbers
65 However, he says elsewhere that they identified opinion, the right moment,
injustice, decision, and mixture with particular numbers (Metaph. 990a23 24).
The difference corresponds to the differences between Stages 1 and 2 in the de-
velopmental account of Pythagorean thought about principles which I give
below, p. 101.
66 Note that C1.3 says that in these [viz. numbers] they thought they observed
many likeness to the things that are and not in the things that are they ob-
served many likenesses to numbers.
67 For example, Metaph. M.3, 1078a2 9: mathematics deals neither with sensible
things qua sensible but with sensible things qua lines and planes (a para-
phrase).
to be the basis of natural phenomena as recounted in C1.4, C1.5, and

especially C1.6.
At this point we have the materials needed to reach the desired con-
clusion C1.1 as rephrased in C1.8, that the elements of numbers are the
elements of all the things that are. The reasoning is easy enough to fol-
low, although somewhat loose. We start with C1.2, that numbers are by
nature the first among mathematical entities. Then the additional claim
C1.3, that in numbers they supposed that they observed many likenesses
to the things that are and that come to be, is supported by C1.4 and
C1.5. C1.6, the general statement that all other things appear in their
entire nature to be likened after numbers, resembles C1.3, with the ad-
dition of the significant word nature, which here refers to the na-
ture of particular (sorts of) things.68 But it says more than C1.3. The
point is no longer that the things that are like numbers somehow or an-
other and in various ways, but that they are likened after numbers: num-
bers are the model, the original, the prototype. And they are likened
after numbers in their entire nature. The properties of a thing that are lik-
ened after numbers are not just some of the properties it might chance
to have; they are all the properties that constitute its nature.
C1.7 follows from C1.2 and C1.6. Since by C1.6 not only are
things like numbers, but also, as the prototypes of things, numbers
have priority; and further since (again by C1.6) this claim holds for
all other things, it follows that numbers are the first things in the
whole of nature. For in the first place they are prior to all things that
Aristotle would consider to be natural entities.69 And second, as I pointed
out above, he interprets the Pythagoreans as holding that numbers are
natural things. For the Pythagoreans, then, the category of natural things
will include not only what Aristotle places in that category but numbers
as well, together with other mathematical entities. And since by C1.2
numbers are by nature the first among mathematical entities, it follows
that they are the first among natural entities, or, as C1.7 puts it, the first
things in the whole of nature.
68 Nature in this sense may approximate to Aristotles technical notion of es-

sence. Cf. Aristotle, Metaph. 1014b35 1015a13. Aristotle may have meant
(anachronistically) to ascribe this idea to the Pythagoreans, but he may have
chosen the less technical word nature to convey a less precise idea, perhaps
amounting to the basic or most significant or notable properties of a thing.
69 I refer to what Aristotle describes as things that are by nature and things that
have a nature (Phys. 192b8, b32) in the sense of having an internal principle of
change and rest (ibid. 192b13 14).
Finally, C1.2 and C1.7 yield C1.8. Since by C1.2 numbers are the
first among mathematical entities, it follows that the principles of num-
bers70 are the principles of mathematical entities; since by C1.7 numbers
are the first things in the whole of nature, we reach the desired conclu-
sion C1.8 that the elements of numbers are the elements of all the things
that are, which is apparently intended as equivalent to the original thesis
C1.1 that the principles of mathematical entities are the principles of all
things.71
The immediately following claim C1.9 that the entire heaven is a
musical scale and a number, has no direct relation to the preceding dis-
cussion despite its immediate syntactical connection to C1.8.72 It is not
concerned with principles of things or numbers, and it is the only part of
passage C1 that identifies a physical entity with a number. In fact, it
identifies it as a number and a harmonia,73 which implies that it takes har-
monies to be numbers, an identification not found in C1.5. C1.9 intro-
duces C1.10, which gives an example of the way in which the Pytha-
goreans in question enthusiastically applied their mathematical ideas to
nature. The importance they gave the number ten led them to invent
a tenth celestial body since they could only find nine.
The final section of C1, C1.11-C1.13, is not a passage of continu-
ous exposition like C1.1-C1.10. It presents difficulties. The material
immediately preceding C1.11 suggests that Aristotle will state how
the Pythagorean doctrine on principles fits in with his own doctrine
of four causesbut that is the topic of only part of C1.11; C1.12 pro-
vides new information on their views on principles, and C1.13 repeats
material stated above whose connection with what is said in C1.11 and
C1.12 is not obvious.
70 There may be a shift here between two notions of principle: the principles of an
entity (number) and the principles of a science (mathematics). (See n. 54.) A
charitable reading would take C1.1 as speaking about the principles of mathe-
matical entities.
71 See above, p. 80.
72 C1.9 also introduces a thesis that will be important later on. See the discussion
of C1.13 below.
73 In C1.5 qlom_a is translated harmonic interval, referring, for example, to the
interval of the octave. Here, assuming that the reference is to the harmony of
the spheres, it refers to a system of harmonic intervals embodied in particular
musical notes, hence something close to a musical scale.
C1.11 states that the Pythagoreans evidently74 held that number75 is

the principle both as matter for the things that are and as their attributes
and states. This is a problematic passage, first because nothing in the pre-
ceding discussion indicates that numbers are the material causes of
things, and second because attributes and states of things are not in Ar-
istotles view principles or causes, and further, because the preceding
discussion has not said that the attributes and states76 of things are num-
bers. Regarding the first problem, even though C1.8 speaks of ele-
ments, which in Aristotle are frequently material elements,77 it is hard
to see how numbers can be thought of as material constituents of justice,
soul/mind, the right moment, or harmonic intervals, which are the
evidence on which C1.8 is based, particularly because C1.6, which is
a premise for C1.8, speaks only in terms of things being likened after
numbers and this is far from saying that they are composed of num-
bers. Worse, C1.8 does not say that numbers are the elements of things,
but that the elements of number are, so even if elements suggests the ma-
terial cause, the doctrine expressed in C1.8 is distinctly different from
that in C1.11.
The second problem is similar. The preceding material has spoken
of likenesses between numbers and things C1.3, which amounts to
numbers and things having the same attributes. It has declared that
things are made in the likeness of numbers C1.6 and has indicated
that numbers are prior to things C1.7. It has even identified the entire
heaven with a number C1.9. But it has not said straightforwardly that
the attributes of things are numbers or that numbers are the principles
of things as their attributes and states. I understand states to be an Ar-
74 Aristotles caution here (va_momtai mol_fomter) is a good indication that at least

in C1.11 we have Aristotles verdict rather than anything necessarily closely
based on Pythagorean sources.
75 Note the shift to number from numbers in the parallel passage C1.8. I do
not find any difference in meaning. In the passages I am investigating, the plural
is used more often than the singular, but I have not been able to discover that
Aristotle intends anything by using now one and now the other. The most im-
portant passages are C1.2,C1.4, C1.5, C1.6, C1.7, C1.8, C1.10, C1.13, C3.1,
C3.3, C4.1, C4.2, C4.4, C4.5, C4.6, plural) and C1.11, C1.12, C2.4, C3.1,
C3.2 (singular).
76 Even the mention of states (6neir) is unjustified. C1.4 and C1.5 mention attrib-
utes (p\hg), but not states.
77 This is the second meaning of stoiwe?om given in Metaph. D.3 (1014a31 35).
According to Diels, this is the most common meaning of the word in Aristotle
(H. Diels, Elementum [Leipzig 1899] 25).
istotelian term for relatively permanent attributes of things.78 Again,

these are different doctrines from what we already know, and they
are not justified by the earlier claims.
C1.12 seems to be the immediate continuation of C1.8 with its ref-
erence to the elements of number, as if C1.9 through C1.1179 had not
intervened. It identifies the even and the odd as the elements of number,80
describes those elements as unlimited and limited (respectively), and
seems to declare the even and the odd to be principles of the one81
and the one to be the principle of number. The wording of this last
claim presents difficulties:82 it declares the one to be from (1j) the
even and the odd, and number to be from the one. Aristotle stand-
ardly uses the preposition 1j to refer to the material cause, but it is difficult
to understand how the even and the odd can be material causes of the
one and how the one can be the material cause of number, as
C1.12 would be saying on this way of understanding 1j.
However, if we can make sense of C1.12 with this interpretation of
1j, some of the problems associated with C1.11 will be solved. To begin
with, it would be as hard for Aristotle, as it is for us to understand how
the even and the odd can be material causes of the one and how the
one can be the material cause of number. Since the preceding material
(other than the as yet unjustified claim that number is the principle as
matter for the things that are) would give Aristotle no reason to think
that material causes are under discussion, the most plausible explanation
is that Aristotle had a Pythagorean source for the views expressed in
C1.12 and that that source probably made use of the 1j locution. If
he did, he would be justified in mentioning that view here as relevant
to his concern to locate in earlier philosophers antecedents of his own
theory, even though that claim does not emerge clearly from the discus-
sion at C1.1-C1.10. And since it is virtually inconceivable that he
would have invented the ontology sketched in C1.12, I consider it cer-
tain that Aristotle had a Pythagorean source for the claims expressed
there. And they justify the surprising assertion of C1.11 that number
is the material cause of the things that are. We have a system in
78 For 6nir in this sense, see Cat. 8b7 8.

79 Or perhaps just C1.11.
80 For the singular, see n. 75 above.
81 I take this expression to mean the number one (which the Pythagoreans did not
regard as a number). See n. 16.
82 Rosss mistranslation of 1j as proceeds from is a reflection of these difficulties.
which the even and the odd (which are somehow related to the unlim-
ited and the limited83) are ultimate principles. They are the immediate
principles of the one;84 the one is the immediate principle of num-
ber. And in all these cases the notion of principle in play is that of the
material cause.
C1.13 repeats C1.9 almost verbatim, except that it omits mention
of the musical scale and says that the heaven is numbers rather than
number.85 Following the claim in C1.12 that number is from (1j)
the one, C1.13 completes the ontology. By asserting that the entire
heaven is numbers it links the sensible world with the preceding levels.
Although it does not state so explicitly, it seems to indicate that numbers
are the principles of the sensible world; we already have something close
to this claim in C1.7, that numbers seemed to be primary in all nature.
There remains the problem raised by the assertion in C1.11 that
number is a principle as the attributes and states of things that are. As
with the first problem, it is possible to ascribe it to a separate Pythagor-
ean source, but there is no need for this expedient, since it can be seen as
recapitulating points made earlier in the passage. Since numbers are pri-
mary in nature C1.7 Aristotle might well call them principles of other
entities. And he might well have interpreted the thesis that other entities
are made in the likeness of numbers in their entire nature C1.6 to in-
clude their being made in the likeness of number in their attributes
and states, that is to say, that numbers have the same attributes and states
that other entities do. Thus, the principles of things would be numbers
as having the attributes of those things. Further, some of the things
identified in C1.4 with attributes of numbers ( justice, the right mo-
ment) are themselves attributes rather than substances, which is another
83 See discussion of C2 below.

84 C1.12 includes an explanation of why the even and the odd are principles of
the number one: the number is both even and odd. The reason Aristotle
gives for this claim is that when one is added to an even number the result is
an odd number and when it is added to an odd number the result is an even
number (Aristotle Frag. 199, ap. Theo Smyrn. de math. ap. Plat. p. 22, 5 Hiller).
This explanation is unsatisfactory by itself, since it holds for all odd numbers.
However, one is the first number which has this property, and so, the principle
stated in C2.6 and exemplified by two and double implies that the number one
is identical with that property. Also, this explanation supports the interpretation
advanced below (p. 89) that (at least in some cases) the Pythagoreans held that if
one thing is a principle of another, then the former is predicated of the latter.
85 The difference between numbers in C1.13 and number in C1.9 seems en-
tirely negligible. See n. 75.
reason for saying that numbers are the principles of the attributes and
states of things. For these reasons I am inclined to see this claim of
C1.11 as an Aristotelian interpretation of Pythagorean views.
C2 Metaphysics 987a13 27
oR d Puhac|qeioi [C2.1] d}o
lm tr !qwr jat tm aqtm eQq^jasi tq|pom, tosoOtom
d pqosep]hesam d ja Udi|m 1stim aqt_m, fti [C2.2] t pepeqa- 15
sl]mom ja t %peiqom [ja t 4m] oqw 2t]qar timr ^hgsam
eWmai v}seir, oXom pOq C c/m E ti toioOtom 6teqom, !kk( [C2.3] aqt
t %peiqom ja aqt t 4m oqs_am eWmai to}tym m jatgco-
qoOmtai, [C2.4] di ja !qihlm eWmai tm oqs_am p\mtym. peq_ te
to}tym owm toOtom !pev^mamto tm tq|pom, ja [C2.5] peq toO t_ 1stim
20
Eqnamto lm k]ceim ja bq_feshai, k_am d( "pk_r 1pqacla-
te}hgsam. q_fomt| te cq 1pipoka_yr, [C2.6] ja pq~t\ rp\q-
neiem b kewher fqor, toOt( eWmai tm oqs_am toO pq\clator 1m|-
lifom, speq eU tir oUoito taqtm eWmai dipk\siom ja tm
du\da di|ti pq_tom rp\qwei to?r dus t dipk\siom. !kk( 25
[C2.7] oq taqtm Usyr 1st t eWmai dipkas_\ ja du\di [C2.8] eQ d l^,
pokk t 4m 5stai, [C2.9] d j!je_moir sum]baimem.
The Pythagoreans similarly [C2.1] posited two principles, but they added
something peculiar to themselves, that [C2.2] the limited and the unlimited
are not distinct natures like fire or earth or anything else of that kind,86 but
that [C2.3] the unlimited itself and the one itself are the substance
(oqs_a) of the things of which they are predicated. This is why [C2.4]
they call number the substance (oqs_a) of all things. On this subject,
then, they expressed themselves thus; and [C2.5] regarding the question
of what things are (t_ 1sti) they began to make statements and definitions,
but treated the matter too simply. For they would define superficially and
they thought that [C2.6] the first thing an indicated term applies to is the
essence (oqs_a) of the thing, as if one were to suppose that double and the
number two are the same because two is the first thing double applies to.
But surely [C2.7] to be double and to be two are not the same; [C2.8] oth-
erwise, one thing will be many[C2.9] a consequence which they actually
drew. (tr. based on the Revised Oxford Translation)
C2 occurs shortly after C1, still in the context of Aristotles discussion of
the principles and causes recognized by earlier philosophers. C2, which
concludes Metaphysics A5 and ends the treatment of the pre-Platonic
thinkers, is a brief appendix on two Pythagorean theses as yet unmen-
86 For this translation of v}sir, see LSJ s.v. vi. Contrast Rosss more interpretive
translation were not attributes of certain other things, e. g., of fire
tioned. Both theses are stated in the Aristotelian language of substance

(oqs_a) and essence (oqsa, t_ 1stim), predication (jatgcqoqe?shai),
and definition (bq_feshai). The first point, treated in C2.1-C2.4, is
closely related to C1.12. If read in connection with the earlier claim,
this section makes good sense. The two principles mentioned in C2.1
are the unlimited and the limited, which C1.12 associates with the
even and the odd. Either, then, the unlimited is identical with the
even and the unlimited with the odd, or they are different; if they are
the same, the unlimited/even and the limited/odd are the ultimate prin-
ciples; if they are different, the unlimited will be the principle of the
even and the limited the principle of the odd.
C2.2 makes the point that the Pythagoreans did not make fire or
some other thing of that sort a principle and declare that it was unlim-
ited or limited; rather they made the unlimited and the limited princi-
ples in their own right. C2.3 makes this point in Aristotelian idiom, refer-
ring to the unlimited itself (aqt t %peiqom), as opposed to an
unlimited [amount of] fire, or to fire (which happens to be unlimited).
Where C2.2 speaks of the limited and the unlimited, C2.3 speaks of
the unlimited itself and the one itself. The change is significant and I
will discuss it below (pp. 86 7). For the present, I focus on the general
point made here, that principles are the substance of the things of which
they are predicated. This claim is odd from Aristotles point of view,
since he holds that substances are not predicated of other things.87 Aris-
totle may be supposing that the Pythagoreans were committed to the
belief that if, say, the unlimited is a principle of something, that thing
must have the attribute of being unlimited. We do not at this point
have enough information at hand to determine whether this is a reason-
able interpretation of Pythagorean doctrine.88 But in any case, the idea
that the unlimited and the limited are the substance of the entities de-
rived from them fits well enough with the claim in C1.12 that they
are the material cause of those entities.89
The next thesis, that number is the substance of all things90 C2.4,
does not follow from C2.3 by itself, but it is a reasonable claim to
make in the light of the ontology implicit in C1. The important
87 Aristotle, Cat. 2a11 13, Metaph. 1029a18 27.

88 For another place where Aristotle may be attributing a similar way of thinking
to the Pythagoreans, see the discussion of C4 below, p. 94.
89 See discussion above, pp. 86 7.
90 For the singular, see n. 75.
thing to recognize is that we are several steps further down the ontolog-
ical hierarchy from C2.3, no longer at the level of ultimate principles
but at the derivation of the lowest level from numbers. Where C1
said that number is the principle as matter for the things that are
C1.11, C2 says that number is the substance of all things. The difference
between these two claims mirrors the difference between C1s and C2s
statements about the unlimited and the limited that were noted at the
end of the previous paragraph.91
The second topic begins at C2.5. This topic is advertised in C2.5 as
the Pythagorean contribution to the study of essence and definitions (t_
1stim, bq_feshai), but these terms do not recur in the passage, which (in
continuation of C2.4) is principally concerned with the relations be-
tween numbers and things, specifically about the way in which numbers
are the substance of things. The issue of numbers as principles drops out
of sight. The present concern is the identification of things with numbers.
The example Aristotle provides us in C2.6 treats the identification of
double with the number two. The grounds for this identification
are that the number two is the first subject to which the attribute dou-
ble applies. It applies to the other even numbers as well, but it is iden-
tified with the number two, and hence the number two is its substance.
The wording suggests that it is not an example that the Pythagoreans
themselves used, but that Aristotle thought it was typical of the reason-
ing that they did use. It is hard to see how this account would work for
any of the explanations of the identification of justice, soul/mind, sub-
stance, and the right moment with numbers that Alexander provides
in the passage quoted above (p. 81), since it is unclear in what sense the
associated numbers are the first things to which the entities in question
apply. Moreover, if Aristotle is right, it is hard to see how more than
one thing could be identified with the same number: how can several
things be first? On the other hand, these considerations by no
means indicate that Aristotle was mistaken in attributing this kind of rea-
soning to the Pythagoreans. In fact, the by hook or by crook ap-
proach the they used in amassing evidence for their views92 and the spe-
cificity of Aristotles statement here makes it more likely that he is
correct than mistaken.
Aristotle next goes on to make an objection to the practice of iden-
tifying things with numbers. The objection is simply that to be double
91 See also p. 107.

92 Above, p. 81.
and to be the number two are not the same C2.7. The expression to be
so-and-so (t eWmai followed by the dative) is one of Aristotles standard
technical ways of referring to essence. He means that the essence of
double is different from the essence of the number two, which entails
that the definition of two is different from the definition of double
which is indeed the case, even though the definition of double may
well contain the word two. And if they have different essences,
they are different. So when the Pythagoreans claim that the double is
the number two, they are saying that two different things are the
same, and therefore one C2.8. This objection does not depend on
the possibly non-authentic example of double and the number two,
or on the fact that two is the first number to which double applies,
even though it is stated in the terms of that example. The same result
will hold for the other identifications mentioned by Aristotle and ex-
pounded by Alexander (for example, that justice is the number four)
even if the kind of grounds for identifying double and the number
two that Aristotle describes in C2.6 does not apply in those cases. Aris-
totle can fairly object that to be justice and to be the number four are
not the same thing. This may be enough to justify the comment
C2.9 that the Pythagoreans are committed to this absurdity. But the
comment may also reflect the fact that in some cases several things are
associated with the same number, as Alexander says that soul/mind
and substance are identified with the one.
C3 Metaph. 1083b8 13; b17 19

b d t_m Puhacoqe_ym tq|por t0 lm 1k\ttour
5wei dusweqe_ar t_m pq|teqom eQqgl]mym, t0 d Qd_ar 2t]-
qar. t lm cq l wyqistm poie?m tm !qihlm !vai- 10
qe?tai pokk t_m !dum\tym [C3.1] t d t s~lata 1n !qih-
l_m eWmai sucje_lema, ja tm !qihlm toOtom eWmai lahg-
latij|m, !d}mat|m 1stim.
[C3.2] 1je?moi d tm !qihlm t emta k]cousim7 17
[C3.3] t coOm heyq^lata pqos\ptousi to?r s~lasim r 1n 1je_-
mym emtym t_m !qihl_m.
The Pythagoreans way [of treating this issue] in one way affords fewer dif-
ficulties than those discussed earlier, but in another way has others peculiar
to itself. For not making number separate removes many of the impossibil-
ities; but [C3.1] that bodies are composed of numbers, and that this is
mathematical number, is impossible. [C3.2] They say that the things
that are, are number. [C3.3] In fact [?at any rate] 93 they apply their prop-
ositions to bodies supposing that [?as if] they94 consisted of those numbers.
(tr. based on the Revised Oxford Translation)
C3 is an important passage for the relation between things and numbers.
In it Aristotle speaks of bodies, that is, physical objectsunlike the
things he mentioned in C1: justice, the right moment, and the
like. The absence of examples in C3 is sorely felt. It is clear from the
context that C3.1 represents a view Aristotle attributes to the Pythagor-
eans: that bodies are composed of (1j) numbers (N1). Also that C3.2
represents another such view: that the things that are, are number
(N2). The question is whether these are two representations of the
same view: did they hold both views or only one, and if only one,
which one. C3.3 has been taken to prove that the Pythagoreans at-
tempted to relate properties of number to properties of things, but
did not make the relation between numbers and things clear. Aristotle
sees the Pythagoreans applying mathematical speculations to bodies in a
way that suggests that they think that the bodies are composed of num-
bers, and on this basis he ascribes to them the thesis that all things are
numbers.95 The argument goes that C3.3 shows that Aristotle found
no clear statement that the Pythagoreans held N2. C3.3 gives Aristotles
reason for thinking that the Pythagoreans held N2, but in fact C3.3 re-
quires N1, not N2. N2 is simply Aristotles misunderstanding.
However, this interpretation is not decisive. It takes coOm as having a
limitative force (at least, at any rate) and r with the genitive abso-
lute as meaning as if (in effect, counterfactually). But coOm can also
have an emphatic force (in fact, indeed), and in several places in Ar-
istotle, in the Metaphysics in particular, it can well be translated in this
way. Also, r with the participle typically denotes the thought in
the mind of the subject of the principal verb without implicating
the speaker or writer96 and means in the belief that or supposing
93 The words in brackets here and in the following line are possible alternative
translations, which Huffman adopts. See discussion below.
94 I take s~lasim to be the antecedent of emtym. See H. W. Smyth, Greek Gram-
mar (Cambridge, Mass. 1956) 2072: The genitive of the participle [in the
genitive absolute] may stand without its noun or pronoun, and 2073a:
The genitive absolute may be used where the grammatical construction de-
mands the dative; cf. Thuc. 1.114.1 diabebgj|tor Edg Peqijk]our stqati
)hgma_ym Acc]khg aqt` fti
95 Huffman (above, n. 3) 61.
96 Smyth (above, n. 94) 2086.
that. If these alternative ways of taking the two particles are adopted,
C3.3 does not express Aristotles grounds for supposing that they held
N2. Instead it expresses the idea that in addition to holding N2, they
also did something else that stems from their belief (to which Aristotle
does not subscribe) that bodies are composed of numbers. In this con-
nection it is worthwhile to note that as in C3.2 Aristotle typically ex-
presses N2 as the things that are (t emta) are numbers, not bodies
are numbers, so the references to bodies being composed of numbers
in C3.1 and C3.3 may be another indication that Aristotle is not infer-
ring N2 from the consideration advanced in C3.3.97
C4 Metaphysics 1090a20 5; a32 5

oR d Puhac|qeioi di [C4.1] t bqm pokk t_m !qihl_m p\hg rp\q-
womta 21
to?r aQshgto?r s~lasim, [C4.2] eWmai lm !qihlor 1po_gsam t
emta, [C4.3] oq wyqistor d], !kk( [C4.4] 1n !qihl_m t emta di t_ d];
fti [C4.5] t p\hg t t_m !qihl_m 1m "qlom_ rp\qwei ja 1m
t` oqqam` ja 1m pokko?r %kkoir. 25
jat l]mtoi [C4.6] t poie?m 1n !qihl_m t vusij s~lata, [C4.7] 1j 32
l 1w|mtym b\qor lgd jouv|tgta 5womta jouv|tgta ja
b\qor, [C4.8] 1o_jasi peq %kkou oqqamoO k]ceim ja syl\tym !kk(
oq t_m aQshgt_m 35
But the Pythagoreans, because [C4.1] they saw many attributes of numbers
belonging to sensible bodies, [C4.2] made the things that are, numbers
(N2)[C4.3] not separate numbers, however, but [C4.4] [they made]
the things that are out of (1j) numbers. But why? Because [C4.5] the at-
tributes of numbers are present in harmonic intervals and in the heavens
and in many other things.
In that [C4.6] they make the natural bodies out of (1j) numbers,
[C4.7] things that have lightness and weight out of things that do not
have weight or lightness, [C4.8] they seem to be speaking of another heav-
en and other bodies, not of sensible ones. (tr. based on the Revised Oxford
Translation, with modifications)
Like C3 (but unlike C1), C4 contains both N1 and N2, and like C1
(but unlike C3), C4 gives an account of how the Pythagoreans reached
the theses it contains. Accordingly, it is a potentially valuable source of
information on their theory and the by now vexed question whether
they really held both N1 and N2, or only one (and if so, which one).
C4.1 says that they held N2 because they saw many attributes of num-
bers belonging to sensible bodies, and C4.5 that they held N1 because
97 See also the discussion of C4 below.

the attributes of numbers are present in harmonic intervals and in the

heavens and in many other things. I think that C4.1 and C4.5 are
meant to give the same grounds, and these are in fact the grounds
also mentioned in C1.3-C1.5. If I am correct, then, Aristotle is saying
that the Pythagoreans inferred both N1 and N2 from the same consid-
erations. This does not necessarily mean that N1 and N2 are synony-
mous or even equivalent, or that Aristotle takes them to be so. In fact
as we will see, there are reasons to suppose that Aristotle considers
them separate theses.
According to C4.1 and C4.2, then, the Pythagoreans concluded
that the things that are, are numbers because they saw that sensible bod-
ies possess many attributes of numbers. Aristotle may have credited them
with reasoning as follows. Numbers are the bearers of (certain?) attrib-
utes, so anything that has (any of ?) those attributes has them because it
(in some way) is those numbers. In effect, if a number has a given attrib-
ute, then anything that has that attribute is that number. But in what way
is it that number? C4 is part of a critical discussion of views on how
numbers exist. Plato held that numbers are Forms, and therefore self-
subsistent or separate entities (in the Aristotelian sense of existing in-
dependently and not being attributes of other entities) by participation
in which sensible bodies possess attributes. Aristotle objects to this
view on the grounds that Plato failed to prove that Forms exist, so
the explanatory role of Forms is unjustified (1090a2 20, cf. 987b8
10, b18 19). In C4 he says that the Pythagoreans did not hold that
numbers are separate C4.3, but believed that they somehow give
rise to the attributes of the things that are. Since numbers exist and
are not separate from things, and also are somehow causes of the attrib-
utes of things, they concluded that things are numbers (N2). Whatever
difficulties there may be for this theory, one objection that does not
apply is that it cannot account for how the attributes of numbers are
found in sensible bodies.98
But in what way are things numbers? Aristotle answers this question
by saying that the Pythagoreans held that (N2) things are composed of (1j)
numbers C4.4. Perhaps the reasoning he attributes to them is that if
numbers are not separate from bodies, they must be bodies of some
kind themselves, either identical with the bodies to which they give at-
98 Aristotle acknowledges this much at 1090a30 31. On this score the Pythagor-
eans fare better than Plato and Speusippus (the latter of whom he treats briefly
at 1090a25 30).
tributes or components of those bodies. Since one entity can have many
attributes, and since its different attributes may be due to different num-
bers, if a thing is simply identical with the numbers that cause its attrib-
utes, it may be identical with different numbers. But this would have the
impossible consequence that those different numbers are identical with
one anothera result that can be avoided by interpreting N2 as N1 and
allowing a thing to be composed of more than one number. C4 may
display traces of this progression of thought. First we have the thesis N2,
that the things that are, are numbers C4.2, which invites the question in
what way things are numbers. C4.3 states that the Pythagoreans did not
adopt one possible answer to this question, the Platonic answer that
numbers are separate entities, participation in which accounts for
the properties of things. C4.4 presents the answer Aristotle says they
did adopt. It is a surprising answer, since numbers may not seem to
be the right kinds of things to be components of sensible bodies.
C4.5, then, supports this answer by repeating a familiar Pythagorean
point (stated just above in C4.1), putting it in a slightly different way.
Where C4.1 said that attributes of numbers belong to (that is, are predi-
cated of: rp\qweim with dative) sensible bodies, C4.5 says that they are
present in (or belong in) (rp\qweim 1m with dative) the things that are, spe-
cifically, in harmonic intervals and in the heavens and in many other
things. It might be objected that the sensible bodies of C4.1 are not
identical with the things that are, on the grounds that some things
that are, are not sensible bodies. But shortly afterward C4.6, he repeats
C4.4 substituting natural bodies (which here is equivalent to sensible
bodies) for the things that are. Apparently, he does not here place any
weight on the difference between the things that are and sensible (or
natural) bodies, and he doubtless took the heavens and many other
things in C4.5 to include the latter class of entities.
This reading of C4 suggests that N2 was the basic Pythagorean for-
mulation and that N1 is a reformulation of N2 in a way that makes sense
of that claim when the entities in question are sensible bodies. (It is hard
to see how it would have been initially formulated with such things as
justice, soul/mind, and the right moment in mind.) Is it a Pythagor-
ean or an Aristotelian reformulation? Clearly enough, this is a question
which C4 by itself cannot answer; I shall return to it shortly.99
What conclusions can we draw from these passages about the Pytha-
gorean views on which C1 through C4 are based? Much of the discus-
99 Below, p. 98.
sion is framed from the point of view of Aristotles philosophical inter-

ests and doctrines, but it is based on what Aristotle took to be Pythagor-
ean doctrines. The main concerns of these passages are the Pythagorean
views on principles and ontology (C1 and C2), and the relations be-
tween numbers and things (C1, C3, and C4). I shall take up the latter
topic first.
As we saw above, Aristotle recognizes three main views: that things
are composed of numbers (N1), that things are numbers (N2) or num-
ber (N3),100 and that things resemble or are made in the likeness of or
are imitations of numbers (N4). C1 is the main source for something
close to N4; C3 and C4 for N1 and N2.101 C4 infers N1 and N2
from the same kind of evidence that C1 uses to infer that things are
made in the likeness of numbers (N4). It is clear that Aristotle believed
that the Pythagoreans based one or more of these three claims on the
kind of evidence brought in C1.4 and C1.5. They saw in numbers like-
nesses of different sorts of things and concluded what? That things
imitate numbers (N4), that things are composed of numbers (N1), or
that things are numbers (N2 and N3)? The examples in C1.4 and
C1.5 do not suggest the second of these claims (N1). They suggest
N4 most directly, but there is an important difference. C1 talks of
there being likenesses of things in numbers C1.3 and of things being
made in the likeness of numbers C1.6, but it does not say that things
are by imitating numbers (N4). This latter claim has important meta-
physical implications that are entirely absent from the wording of C1.
We can see how a Pythagorean with ontological interests might con-
clude N4 on the basis of the evidence in C1.4 and C1.5. Indeed, the
context in which N4 is reported strongly suggests that Aristotle had a
Pythagorean source for the claim.102 But N2 seems an equally good pos-
sibility for what a Pythagorean might have claimed on the basis of the
same evidence. C1, then, presents evidence, which I think it is safe to
take as genuinely Pythagorean, that things are like numbers, but leaves
it unclear whether N1 and N2 (which C1 does not contain) are genuine
Pythagorean doctrines.
Not that this is the whole story, since there is every reason to think
that Aristotle knew more about Pythagorean thought than he reports in
100 For the identity of N2 and N3, see n. 75.

101 C1.9 says that the whole heaven is a harmonia and a number, which points to-
wards N2.
102 See below, p. 102.
C1, and we have no good reason to reject N1 or N2 out of hand. In-

deed, C4 tells us that the Pythagoreans held both these theses and that
they held them on the basis of the very evidence supplied by C1,103
which means that the authenticity of N4 is no better supported by
the arguments Aristotle reports than the authenticity of N1 and N2.
Neither N1 or N2 is a rephrasing of N4, and it seems unlikely that Ar-
istotle would have invented either N1 or N2 solely on the basis of N4
and the evidence for it presented in C1. Moreover, since Aristotle takes
N1 and N2 to be different theses that are open to different kinds of ob-
jections,104 I think it unlikely that he would have corrupted N1 into N2
or vice versa. Aristotle is a hostile witness all right, but I do not think
that he would be so hostile as to invent problems; he found problems
enough as it was.
C3 and C4 contain both N1 and N2, but they put them in different
ways. Both passages state N2 in terms of the things that are (C3.2,
C4.2), but C3 states N1 in terms of bodies (C3.1, C3.3) and C4
states it both in terms of the things that are (C4.4) and in terms of
natural bodies (C4.6).105 I suggested above106 that N1 makes better
sense as an account of the nature of sensible bodies than it does for
the things that are in general, which for the Pythagoreans include
such non-sensible entities as justice, soul/mind, and the right mo-
ment. The evidence for the relation between numbers and things pre-
sented in C1.4, C1.5 and C4.5, shows that they were not concerned
103 C4 seems to repeat much of C1, but there are significant differences. C1 speaks
of things resembling numbers (N4), while C4 speaks of things being numbers
(N2) and being made from (1j) numbers (N1). Also, where C1 says that such
and such an attribute of numbers is justice, etc., C4 says that attributes of num-
bers belong to sensible bodies. Note that neither justice nor the other examples
given in C1 are sensible bodies, and further that C1 says that the Pythagoreans
saw the attributes and ratios of the harmonic intervals in numbers, while C4
says, vice versa, that the attributes of numbers are in them.
104 For an objection to N1, see Metaph. 1090a32 5 (= C4.6-C4.8) where Aristo-
tle complains that the Pythagoreans make the natural bodies out of numbers.
For N2, see Metaph. 1090a12 ff., where Aristotle objects that unlike the things
in the cosmos, numbers are eternal and unchangeable, and Metaph. 987a22 27
(= C2.6-C2.9) which points out that it follows for the Pythagoreans that if
things are identical with numbers, then one thing will be many.
105 Also compare C4.1 they saw many attributes of numbers belonging to sensible
bodies with C1.3 in these [viz. numbers] they thought they observed many
likenesses to the things that are and that come to be.
106 p. 95.
exclusively (or even mainly, perhaps) with sensible bodies. Those like
myself who think that a good deal of this evidence sounds early (I be-
lieve this to be the majority view) will be disinclined to think that the
original formulation of the relation between numbers and things was re-
stricted to sensible bodies, and hence will find that these considerations
favor the authenticity of N2.
What, then, of N1? If it is not Aristotles erroneous restatement of
N2, is it genuine too? Would the Pythagoreans have asserted two such
similar (and yet such different) claims as N1 and N2? I believe they assert-
ed both theses, and that N1 is in effect a Pythagorean reformulation of
N2 in a way that makes it more applicable to sensible bodiesa particular
kind of things that are. Aristotle takes them to hold that sensible bodies
are numbers in that they are constituted of them in the same way that a
chair is wood and nails and a man is flesh and blood. It is certain that the
1j locution was Pythagorean: Philolaus uses it in B1, B2, and B6, so Ar-
istotle could have taken this formulation from a Pythagorean source. This
way of expressing the relation between things and numbers may explain
Aristotles insistence that the Pythagoreans must have been thinking of
material causes.107 Possibly this is behind the move from principles in
C1.1 to elements in C1.8, since Aristotle frequently uses stoiwe?om
in connection with material elements.108
However, it would be unwise to leave the matter at this point, since
Huffman has rejected both N1 and N2 as due to an Aristotelian misin-
terpretation of Philolaus.109 To be sure, these theses are not found in
Philolauss extant fragments. In fact, N1, which declares that things
are 1j numbers, is incompatible with B1s claim that they are 1j unlim-
iteds and limiters. Huffman proposes that N1 and N2 (he does not clear-
ly distinguish between the claims) are Aristotles ways of summarizing
Pythagorean (i. e., Philolauss) doctrine.110 The doctrine in question is
Philolauss cosmology, including specifically the stage discussed in B7
107 See C1.11 and 986b6 8: They seem to range the elements under the head of
matter, for they say substance is constituted and fashioned out of these and that
these are present in it (adapted from the Revised Oxford Translation).
108 See n. 77.
109 See n. 42. It is an Aristotelian interpretation of Philolaus; Huffman (above,
n. 3)63. But this is too confident. Earlier in the same paragraph Huffman says
more cautiously (and, I believe, correctly), It seems most likely that he [Aristo-
tle] had Philolauss book and perhaps some other writing or oral reports. Huffman
gives no reason for eliminating or ignoring the second and third possibilities.
110 Huffman (above, n. 3)63.
(The first thing that was joined, the one in the middle of the sphere, is
called the hearth), which Huffman argues Aristotle mistakenly inter-
preted as the first stage in the generation of the number series, whereas
in fact it is an early stage in the generation of the cosmos. Huffmans ar-
gument is strong but I am not convinced that it overturns Burkerts
view that the Pythagoreans (or Philolaus) viewed the generation of
the cosmos as identical with the generation of the number series,111 so
that what B7 describes is both the first stage in the generation of the
number series and an early stage in the generation of the cosmos.
Huffmans other piece of evidence that Aristotle did not find N1 in
any Pythagorean text is also plausible but not decisive. It is Metaphysics
1083b17 19 (= C3.2-C3.3), which he paraphrases and translates as fol-
lows: Aristotle first states flatly that the Pythagoreans say that all things
are numbers, but then continues in the next sentence at least (coOm)
they apply heyq^lata to bodies as if they (the bodies) consisted of
those numbers. As Huffman understands the passage, the crucial
point is that the doctrine that all things are numbers is not presented
as something that the Pythagoreans explicitly stated, rather it is present-
ed as Aristotles own deduction from the way the Pythagoreans pro-
ceed.112 While this is a legitimate way to construe Aristotles words,
it is not the only way. On another way of understanding the particle
coOm, the sentence is translated They say that the things that are, are
number. In fact they apply their propositions to bodies supposing that
they consisted of those numbers. I defended this reading above.113 If
the sentence is taken this way, then far from showing that N1 and
N2 are Aristotles conjectures, it is actually evidence that his Pythagor-
ean sources, whatever they were, contained both N1 and N2.
As to the views on principles contained in C1 and C2 (especially
C1.12-C1.13 and C2.1-C2.4), all that remains to do is see whether
the doctrines of the two passages (treated above, pp. 89 90) can be
made into a coherent account, and if so, whether it is likely to be
based on a Pythagorean source.
I support an affirmative answer to the first of these questions.114 Ar-
istotle presents a Pythagorean ontology with several levels. C1.12 makes
the even and the odd (which are somehow related to the unlimited and
111 For references, see n. 41.

112 Huffman (above, n. 3)60 1.
113 See pp. 92 3.
114 I take up the second question below in Section 8.
the limited; either they are identical or the latter pair are the principles
of the former) ultimate principles. They are the immediate principles of
the one; the one is the immediate principle of number. C1.13
completes the ontology by giving a hint that numbers115 are the princi-
ples of the sensible world. The notion of principle involved here is ex-
pressed by the 1j locution; we are dealing with something that Aristotle
thinks approximates to his notion of the material cause.
C2.1 clears up the relation between the pairs even/odd and unlimit-
ed/limited. C2.1 states that there is one pair of ultimate principles, and
C2.2 makes it clear that this pair consists of the unlimited and the limited;
not anything else such as fire or earth (or, we may add, even) that is un-
limited, but C2.3 the unlimited itself. C2.4 says more clearly than C1.13
that number is the substance (again, presumably, in the sense of the ma-
terial cause) of all things, which must mean everything with the excep-
tion of the entities that are higher up in the ontology: the unlimited and
the limited, the even and the odd, the one, and number.
There remains the question deferred from above (p. 72): what is the
significance of the change from talk of the unlimited and the limited in
C2.2 to talk of the unlimited and the one in C2.3? It is clear that in
C2.2 Aristotle is dealing with the ultimate principles. In C2.3, I think,
he is dealing with the one as a principle of number and is alluding to
something not made explicit in C1.12, that the one is not the only
principle on which numbers depend; they depend on the unlimited as
well. The most useful text on this subject is Metaphysics 1091a15 18
(quoted above as evidence for N12), which talks of generation116 subse-
quent to the generation of the one. Most scholars think that this passage
refers to the generation of numbers. First the one was generated and
then in some obscure way, the unlimited was drawn in (presumably
by the action of the one on it) to form the other numbers.117 If this
is right, then the unlimited is (along with the limited) a principle of
the one and (together with the one) a principle of the other num-
bers.
115 I note the change from number in C1.12 to numbers in C1.13, but I do
not see that any difference is intended. See n. 75.
116 I use generation for want of a better term. I take it that what is being descri-
bed is not a process of generation in time, but of ontological dependence.
117 Even if Huffman is correct in thinking that N12 refers to the generation of the
cosmos (above, p. 72), it seems likely that some analogous procedure (possibly
even the same series of events) could have been used to account for the gen-
eration of numbers. See n. 41.
The resulting ontology has five tiers.

1. The unlimited itself and the limited itselfwhich are, respectively,
principles of:
2. The even and the oddwhich are jointly principles of:
3. The onewhich ( jointly with the unlimited) is a principle of:
4. Numbers (number)which in various ways are principles of:
5. Everything else (all things)
8. What did Aristotles Sources Tell him about Pythagorean

Principles?
I now return to the list of Aristotles reports on Pythagorean principles

(N1-N13). The problem is how to sort them outwhich are authentic
Pythagorean views and which are Aristotelian rephrasings and interpre-
tations? And of the authentic views, which ones are likely to represent
Philolauss theory? Whereas the list was composed of statements taken
out of context, in Section 7 I provided the contexts in which most of
them occur and discussed their meaning in context. Much of the fol-
lowing discussion is based on the results of that section.
I argued above that N1 and N2 are both authentic Pythagorean
views.118 Are they due to Philolaus? As for N1, it is true that Philolaus
uses the 1j locution in stating the relation between his principles and the
entities of which they are principles (B1, B2, B6). However, the claim
that things are 1j numbers corresponds to nothing in his preserved frag-
ments, and it actually conflicts with B1.119 Further I find it hard to be-
lieve that Aristotle would have corrupted B1 into N1. As for N2, there
is nothing in Philolauss fragments that remotely resembles the claim.
N1 and N2 must be regarded as not based on Philolaus.
As I said above, I take N3 to be equivalent to N2.120 Whether Aristo-
tles Pythagorean sources spoke of number as well as numbers I am not
willing to say. Nothing in Philolauss fragments corresponds to N3.
N4 assembles Aristotles statements that the Pythagoreans held that
(a) things have likenesses to numbers, (b) they have the same attributes
as numbers, (c) they are made in the likeness of numbers, and (d) they
118 See above, pp. 95 9.

119 See p. 98.
120 See n. 75.
are by imitation of numbers.121 At least some ingredients of N4 are very

likely to be authentic. I accept (a) and (b) as genuinely Pythagorean in
spirit and the examples which Aristotle gives in support of them as gen-
uinely Pythagorean, even though the wording of the generalizations in-
volved (In numbers they supposed that they observed many likenesses
and They saw the attributes ) is obviously Aristotles. (c) is a
stronger claim than (a) or (b) in that it makes numbers in some sense
the model or prototype of things.122 It would not be surprising if the
Pythagoreans actually held (c) on the basis of the observations that
grounded their claims (a) and (b).123
As for (d), Aristotle introduces it in the course of his comparison of
Pythagorean views with Platos. He says that where the former group
said that things are by imitation (l_lgsir) of numbers Plato said that
things are by participation (l]henir) in the Forms. The relation is the
same; Plato just changed the name. Plato does in fact use the word l]he-
nir and the related verb let]weim of the relation between sensibles and
Forms. And since Aristotle here is making a point not only about the
doctrines of Plato and the Pythagoreans but also about their terminolo-
gy, I suppose that Aristotle had a good reason to attribute the word
llgsir to the Pythagoreans.124 On the other hand, even though it
uses the Pythagorean term imitation, the statement that things
are or exist by imitation of numbers may not be authentic It has
a Platonic ring125 which is perhaps out of place in a fifth-century context
and may be due to Aristotles phrasing the point he is making in a way
that makes clear the parallel between the Platonic use of participation
and the Pythagorean use of imitation. Alternately, this very expression
may have occurred in a Pythagorean context of the generation contem-
porary with Plato. In any case, there is nothing in Philolaus that corre-
sponds with any of the four claims (a) through (d).
The temptation to see N5 as Aristotles attempt to express Pythagor-
ean ideas in the terminology of his own philosophy (!qw^ in the sense of
principle), and therefore unauthentic, would be irresistible if the word
did not occur in a similar context in Philolaus B6 (The \qwa_ are not
121 For (a), Metaph. 985b27 31 = C1.3-C1.4; for (b), ibid. 985b31 2 = C1.5;
for (c), ib. 985b32 3 = C1.6; for (d), ibid. 987b11 12.
122 See discussion of C1.6, p. 82.
123 See below, p. 116.
124 Evidently Aristotle knew the Pythagoreans had used the word l_lgsir Bur-
kert (above, n. 2) 44.
125 Cf. Plato, Soph. 259a7.
similar or of the same kind). Philolauss \qwa_ are the limiters and the
unlimiteds, which are undoubtedly principles in some sense. Indeed
since Philolaus may offer a two-tier ontology,126 with limiters and unlim-
iteds as the principles of everything else (presumably including num-
bers127), N5 may refer to Philolauss system and may even derive from
Philolaus.
N6 repeats N5 almost literally, with two exceptions. First, it substi-
tutes stoiwe?a (elements) for !qwa_ (principles), which I take to be
an Aristotelian paraphrase, intended to bring out Aristotles view of the
way in which the Pythagoreans made numbers their material cause.128
Second, it substitutes numbers for these, where the latter refers
not yet specifically to numbers but to mathematical entities. N5 and
N6 are each stated in only once, in fact, in the same context, passage
C1. I view N6 (whose sole source is C1.8) as simply an Aristotelian re-
statement of N5 (whose sole source is C1.1).129
N7, however, which is witnessed only by C1.11, is heavily laden
with distinctively Aristotelian terminology (matter, vkg; attribute,
p\hor; state, 6nir). The passage presents problems which I discussed
above,130 where I concluded that the claim that number is a principle
as matter is based on a Pythagorean source because it probably employed
the 1j locution, and the claim that it is a principle as attributes and states
is also an Aristotelian interpretation of Pythagorean views. If it is an ac-
curate (albeit Aristotelian) description of Pythagorean ideas, those ideas
are not found in Philolauss extant fragments.
Whereas N1 through N7 mainly have to do with the importance of
numbers, N8 through N11 are chiefly concerned with ultimate princi-
ples and the relations between principles and the things whose principles
they are.
N8 (= C2.1-C2.4) contains five claims. First, C2.1 it states that the
Pythagoreans posited two principles; second C2.2 it identifies the prin-
ciples as limit and unlimited. There is no difficulty in seeing Philolauss
limiters and unlimiteds at the root of this claim and we can therefore
126 See, however, n. 38.

127 On the traditional interpretation, B7 describes the generation of the number
one, and thus begins the generation of the number series. However, Huffman
has challenged this interpretation, arguing that B7 concerns the first stage in
Philolauss cosmology. See discussion above, pp. 85 8.
128 See discussion of this point above, p. 57.
129 See p. 80.
130 pp. 85 8.
take it at the very least as a close paraphrase of an authentic Pythagorean

view, even of Philolaus. It may also be the case that a different Pytha-
gorean source actually called the principles limit and unlimited.
Third, C2.3 refers to the unlimited itself and the one as prin-
ciples. On the interpretation proposed above (p. 100), they are not
first principles, but principles specifically of numbers. Alternatively, Ar-
istotle could be carelessly using the one as equivalent to the limited.
In any case this is un-Philolaic.
The fourth claim in N8, C2.3), that the unlimited and the one are
the oqs_a of the things of which they are predicated, sounds like Aristotle
interpreting Pythagorean doctrine in the vocabulary of his own system,
possibly under the influence of his occasional assimilation of Pythagorean
doctrines to Platonic views.131 It is a well chosen word, since it has a range
of meanings that conveniently includes essence and substance. I say
conveniently because N9 (C2.6) reports that the Pythagoreans regarded
the number two to be the oqs_a of double and does so in a context
where it is clear that oqs_a means essence (987a26 = C2.7), while ac-
cording to N1 and N7 numbers are taken to be the material cause or mat-
ter of things, and hence are in this way the substance of things.
The word oqs_a does not have a good pedigree in early Pythagorean
texts. (It is found twice in Philolaus B11, a fragment that is judged spu-
rious.) On the other hand, Philolaus uses the related word 1st (which
I translate being) twice in B6. The interpretation of B6 is uncertain.
As I interpret the fragment, 1st~ means being in an undifferentiated
sense, which has different applications in different contexts. In this I
think that it corresponds to Aristotles oqs_a, which (inter alia) some-
times refers to a things matter and sometimes to its essence,132 as I
noted above. In the context of B6 I believe that the word could be
taken in either of the two ways. Certainly if Aristotle knew the text,
it would be natural for him to paraphrase it in terms of oqs_a and to in-
terpret it in whatever sense of oqs_a seemed appropriate.
In B6, where Philolaus speaks of the 1st~ of things and the 1st~
of the things from which the cosmos is constituted, I think that he is
talking about the 1st~ of different groups of things. The things out
of which the cosmos is constituted, designates the principlesthe limit-
131 See Huffman (above, n. 3)207.

132 Burkerts discussion is instructive (above, n. 2) 255 7. He argues that 1st~ in
Philolaus is not to be understood as a material principle, but that in ps-Archytas
p. 19.19 Thesleff, s_a/1st~ is certainly the Aristotelian material principle.
ers and unlimiteds; in the phrase groups of things, things designates

things in general, which include not only the principles but also the
things that are and are known by usthe objects we know in the
world around us, which B1 tells us were joined from the principles.133
If I am correct, then the first part of B6 tells us that the 1st~ of things,
including both the principles and the derivative things that come to be in
the world around us, is eternal, and that we cannot know the 1st~ of any
of these things; the only thing about their 1st~ and nature that we can
know is that none of the things that we know (that is, none of the things
in the world around us) could have come to be if the 1st~ of the limiters
and unlimiteds out of which the world is constituted did not exist. This
claim makes sense if we take 1st~ to refer to material constituents. Phi-
lolaus is then saying that we cannot know the nature of the material con-
stituents of the principles. So even if we know that derivative entities are
composed of the principles, we cannot know what their (ultimate) mate-
rial constituents are either. This reading of B6 allows us to approach the
fourth claim of N8, that the Pythagoreans believed that the principles (viz.
the unlimited and the one) are the oqs_a of the things of which they
are predicated.
Two features of this statement need explaining: first, the un-Aristo-
telian claim that the principles are predicated of derivative entities and
second the potentially ambiguous claim that the principles are the
oqs_a of those entities. Regarding the first, B2 entitles us to think
that Philolaus would agree that the principles are predicated of the en-
tities derived from them, since it announces that the things that are
from limiters limit, those that are from both limiters and unlimiteds
both limit and do not limit, and those that are from unlimiteds will evi-
dently be unlimited. Regarding the second, Aristotles claims in N8
that make use of his own equivalent term, oqs_a, fall into line (to an ex-
tent) on the interpretation that in B6 1st~ refers to the material constit-
uents of things. For as I suggested above, Aristotle was sometimes in-
133 This interpretation is controversial. Huffman (above, n. 3) 130 thinks that it ap-
plies only to the principles, since he supposes that 1st~ means being in a
fused existential-cum-essential sense, and since B6 says that the 1st of
things is eternal. The things in the world around us are not eternal, so they
do not qualify for having an eternal existential being. But I do not find the argu-
ment for the fused sense compelling. The fact that the 1st~ is said to rp\qweim
does not tell any more in favor than against the view that 1st~ all by itself in-
cludes the notion of existence; if it included that notion, why bother to make
the assertion?
clined to view Pythagorean principles as the material causes of the de-

rivative entities (cf. N1, N7), particularly when the derivative entities
are physical bodies. And so the assertion in N8 that the unlimited and
the one, here identified as principles, are the oqs_a of the things
whose principles they are, is easy to understand if we take oqs_a to
refer to the matter of the derivative entities.
But the fit with B6 is not exact, since in N8 the principles them-
selves are said to be the oqs_a of the derivative things and B6 identifies
neither the principles nor their 1st~ with the derivative things or their
1st~. If, then, this part of N8 is based on a misreading of B6, it repre-
sents that fragment very inaccurately. Alternatively, if we suppose that it
is not an unsuccessful attempt to report the doctrine of that fragment, it
may well be influenced by some of Philolauss ideas even so. It is also
possible that Aristotles statement is based on some other Pythagorean
source that says clearly that principles are the being of derivative entities.
I pointed out in discussing C2134 that the fifth and final claim in N8,
this is why they call number the oqs_a of all things C2.4, does not
follow from C2.3, but makes sense in view of the entire ontology Ar-
istotle sketches if we understand it as applying at a lower level. What
needs explaining is the connecting phrase this is why. I think that Ar-
istotle means that number is the principle of the things below it in the
ontology in the same way as the unlimited and the one are principles
of the things just below them. Since the latter pair are principles of those
things in that they are their oqs_a, this is why number is the oqs_a of
all things (where all things refers to all things not already included
in the higher levels of the ontology135).
The claim that number is the oqs_a of all things does not fit with
what we know of Philolaus. While it might rely on some lost Pythagor-
ean source, it could also be another way of putting the idea of N1 or
N2: N1 if oqs_a is taken to mean substance with an emphasis on
its material aspect, and N2 if it is taken to mean essence.
I said something about N9 above.136 The only text that witnesses it
uses oqs_a in the sense of essence and seems to be an Aristotelian analysis
of Pythagorean claims of the form that some feature of the world (such
as double) is (in the sense of is essentially) some number (such as
two). Alexander gives some other examples: justice is the first number
134 pp. 89 90.

135 See p. 100.
136 p. 91.
that is equal-times-equal (i. e., four); the right moment is the number
seven; marriage is the number five, etc.137 There is nothing in Philo-
lauss fragments like these examples or the thesis that they illustrate.
An important point to bear in mind about N10 is that the single pas-
sage that attests it is not taken from Aristotles discussion of the Pytha-
goreans in Metaphysics A5, but is part of his treatment of Platos philos-
ophy in Metaphysics A6. It is part of a complex passage recently
explicated by Huffman,138 in which Aristotle speaks of Platos relations
to earlier philosophers, including the Pythagoreans. In particular, he says
that Plato used the word participation (l]henir) to express the relation
that the Pythagoreans expressed with the word imitation (l_lgsir)
(987b9 14, cf. N4), that like the Pythagoreans he had views on the na-
ture of mathematical entities, and that his views on this topic were like
those of the Pythagoreans in some ways and unlike them in others
(987b14 988a1). The subject of the sentence that witnesses N10 is
not the Pythagoreans, but Plato, understood from earlier in the
paragraph. The discussion is directed more at Plato than at the Pytha-
goreans and its vocabulary is more Aristotelian and Platonic than Pytha-
gorean. The passage, then, does not pretend to be closely based on any
Pythagorean source and can be expected to distort Pythagorean ideas in
order to make them comparable to Platonic theses.139
N10s first claim, that the one is oqs_a, can be interpreted in sev-
eral ways. (a) the one is oqs_a; (b) the one is an oqs_a; (c) the
one is the oqs_a of all other things (where of all other things is un-
derstood from 987b24 just below). The context shows that (c) is what is
meant; the previous sentence (987b18 21), which speaks of the Platon-
ic views that the Forms are causes of other things and that the elements
of the Forms are the elements of all things, also declares that Plato made
the great and the small principles as matter and he made the one prin-
ciple as oqs_a, which is best interpreted here as form in the Aristote-
lian sense, or essence. It is reasonable, then, to take N10 closely with
N8, which regards the one as a principle and declares the principles,
namely the unlimited and the one, to be the oqs_a of the things of
which they are predicated and number to be the oqs_a of all things.
137 Alexander, in Metaph. 38,10, quoted p. 81.

138 Huffman, Platos disagreements with the Pythagoreans: A response to Brisson
and Kahn, delivered at the Pacific Division meeting of the American Philo-
sophical Association, 2005.
139 For a likely instance of such a distortion, see above, p. 102.
The second claim of N10, that numbers are the causes of the oqs_a
of all other things shifts from talking about the one to talking about
numbers. (Cf. the change between the fourth and the fifth claims in
N8.). Where the fifth claim in N8 asserts that number is the oqs_a of
all things, N10 says that numbers140 are the causes of the oqs_a of the
other things.141 This is a different thesis from anything we have seen be-
fore, in particular from the claims that things are numbers (N2, N3) and
that numbers are the substance, matter, or essence of things (N1, N7,
N8, N9). While nothing precludes some Pythagoreans from having
made this rather careful claim, it sounds to me more like an attempt
on Aristotles part to make sense of such Pythagorean doctrines as the
ones reflected in N1 and N2, which were, from Aristotles point of
view, sheer and utter nonsense. It is hard to find anything in Philolaus
very close to the doctrines put forward in N10.
N11 and N12 must, I think, have authentic Pythagorean pedigrees
even though the passages that cite them they may contain some Aristo-
telian terminology. They are too specific and detailed for Aristotle to
have made them up or to have taken them to be the meanings of less
definite claims. The most striking thing about these two texts is how
different they are from one another. N11 seems to offer a kind of anal-
ysis of things into their constituents: number is composed of the one,
the one is composed of the even and the odd. It may even hint at an
argument: the elements of number are the even and the odd because
the one is composed of them and number is composed of the
one. On the other hand, N12 seems to describe the generation of
the number series by the operation of the one on the unlimited.
Even on Huffmans alternative interpretation of N12 as an account
of the first stage of a cosmogony, an account of the stages by which the
cosmos came to be over a period of time,142 the accounts in N11 and
N12 need not be incompatible, any more than a respectable cosmogony
today need be incompatible with current understanding of the physical
and chemical constituents of things in the universe. A cosmogony gives
a different kind of account of the universe than physics and chemistry
140 As elsewhere, I do not think that Aristotle intends any difference between
number and numbers; cf. n. 75.
141 Actually, the word translated causes in the standard English translations of the
Metaphysics is an adjective, not a noun: aQt_our (modifying !q_hlour), not aQt_a
or aUtiom. Numbers are responsible for the oqs_a of things. The idea is that if
you want to find the oqs_a of something, look to the relevant numbers.
142 See above, p. 72.
do, just as N11 gives a different kind of account from N12. The fact that
we have different kinds of accounts does not prove that they came from
different sources. In fact, both have links with Philolaus. N12 could well
stem from the cosmology of which B7 and B17 are the only surviving
fragments, as long as we understand the one in N12 to be the same as
the one in the middle of the sphere, which B7 calls the hearth (that is,
the central hearth of the cosmos), as Huffman proposes.143
N11 has links to Philolauss general account of limiters and limiteds
as principles, if we assume that he admits principles of numbers. How-
ever, the parenthetical remark that the even is unlimited and the odd is
limited may conceal a further step in this analytic account (allowing for
the slippage that is inevitable since Aristotle does not speak of limiters
and limiteds). In the terminology of B2, the odd is from limiters
and the even is from unlimiteds. The one is from the even
and the odd: from both mixed together, as B5 tells us: it is an
even-odd number. And the remaining numbers are from the
one. The resulting ontology is incompatible with the ontology pre-
sented in Philolaus B1,144 but could be based on a different Pythagorean
source.
N13 is presented as a belief held by others that fall in the range of
the Pythagorean umbrella. I have used it already as a piece of evidence
that Aristotle used more than one source.145 It is surely genuine: again, it
is inconceivable that Aristotle made it up. Also, it conflicts with Philo-
lauss doctrine that limiters and unlimiteds are two kinds of principles,
not one. To be sure in the parallel columns limit and unlimited come
first (counting them as one principle, not two), but the nine remaining
oppositions are not derived from that primary pair; there are ten prin-
ciples not two (or twenty).
143 Huffman (above, n. 3)202 ff.

144 See above, p. 71 2.
145 Above, p. 58 9.
9. A Developmental Sketch of Pythagorean Thought

on Numbers and Principles
It is clear that the claims attributed to the Pythagoreans in N1-N13 are

not all consistent.146 Some of the inconsistencies can be accounted for as
stemming from Aristotles difficulties in making sense of their views,
and others as being due to his attempts to state them in his own philo-
sophical language. But major inconsistencies remain. Accepting the
claims of N1-N13 as based on genuine Pythagorean doctrines requires
us to posit a plurality of sources, or at least a single source that is itself
based on a plurality of others. But rather than resting content with sim-
ply saying that the Pythagoreans were an inconsistent lot, I believe that
much of the inconsistency in Aristotles reports can be explained
through suitable reconstruction of the development over time of Pytha-
gorean thought on the subject of numbers and principles. In what fol-
lows I propose such an account, which is drawn entirely from the pas-
sages from Aristotle discussed earlier in this paper. It is important to
admit at the outset that Aristotle does not distinguish between earlier
and later Pythagorean views. What I offer is meant to be no more
than an outline of a possible history of their thought on these subjects.147
As I see it, Aristotles remarks reflect three stages, as follows.
Stage 1: Observation of resemblances between things and numbers.
Stage 2: Identification of numbers as the principles of things.
Stage 3: Identification of the principles of numbers, which are also the
principles of everything else.
The Pythagoreans (Pythagoras himself by many accounts) were (Stage 1)
struck by the numerical basis of the harmonic intervals and they found
(or supposed that they observed) numerical properties in other sensi-
ble phenomena as well.
They then148 (Stage 2) arrived at the view that numbers were basic,
that in some way everything was dependent on them, a view they may
146 I identified some inconsistencies above, pp. 65 6.

147 See Burkert (above, n. 2) 30. Burkert is skeptical of attempts like the present
one to discern a historical development of Pythagorean thought.
148 I do not insist that Stage 2 followed Stage 1 chronologically. It is possible that
the discovery that things are like numbers in certain ways was expressed in the
claims that all things are numbers, or are composed of numbers. Still, it is ob-
vious that there is an important conceptual and theoretical difference between
have expressed in a variety of ways, such as the things that are are com-
posed of numbers (N1) and the things that are, are number(s) (N2-
N3) or the things that are, are by imitation of numbers (N4). N7 and
the first part of N10 represent this stage, even if it is an Aristotelian re-
formulation as suggested above. Characteristic of Stage 2 is the identi-
fication of justice with the number four and the other identifications
mentioned in N4.
Finally (Stage 3), they turned their interest in principles to number
itself and developed views on what the principles of numbers are. Thus
N5 and N6 speak of the principles and elements of numbers and assert
that they (rather than numbers themselves) are the principles of all the
things that are, and N11 identifies the even and the odd as the elements
of numbers, and provides information about the derivatijon of numbers
from those principles. The derivation of numbers is also I think149 the
topic of N12. N10 is also relevant to this stage, though I think it also
reflects interests characteristic of Stage 2.
The interest in principles prevailed to the extent that numbers lost
their central role. The principles were identified as limit and the unlim-
ited (by Philolaus, as limiters and unlimiteds), and interest may have
shifted to accounting for things in terms of those principles rather
than doing so through numbers, which are themselves generated from
the same principles. The principles are said to be the oqs_a (or 1st~)
of the things that are dependent on them. This is the situation of N8
and N9, and perhaps the doctrine of parallel columns cited in N13. It
is also the situation of Philolaus.
All three stages are reflected in Aristotles testimony. He gives a
prominent position to the second stage, which struck him as importantly
different from the other early theories that he summarizes, and which
has connections with the topics of definition and essence, topics impor-
tant for Aristotle, especially in the Metaphysics. He does not claim that
there is a single Pythagorean orthodoxy, so he is not to be blamed if
he reports conflicting views in different places. Recall the passages
cited above in Section 2 that show him reporting that a particular
view was held by some of the Pythagoreans or by others. Even
so, and despite the differences he acknowledges, he evidently thought
that the Pythagoreans or those called Pythagoreans had enough
saying that one thing is like another and that one thing is another, even if the
former claim is the basis for the latter one.
149 For Huffmans disagreement, see p. 72.
in common to be given a common label and treated together. The de-

velopmental sketch I have presented justifies this practice, since there
are close links among the three stages and all of them are stages in a sin-
gle tradition of thought.
On this picture Philolaus belongs to the most advanced stage, as we
would expect from his chronological position, born over a century after
Pythagoras. A question that arises at this point is whether Philolauss
ideas or at least his more general and sophisticated approach were gen-
erally shared by the Pythagoreans of his generation and later, and here I
cannot say. Eurytus seems not have shared it; his trick of identifying
things with the number of pebbles needed to fill in their outlines
seems hardly worthy even of Stage 1. I wonder, then, whether Philolaus
was an outlier in the early Pythagorean tradition, advancing earlier
thought in a philosophically important way, but in a way that was not
pursued by his Pythagorean colleagues. I have not found evidence
that can decide this issue. But if it makes sense to talk about a history
of Pythagorean thinking about numbers and principles, this will be an
important topic to pursue.
10. Aristotles Frustrations with Pythagorean Thought
Aristotle says several times that the Pythagoreans (or some Pythagoreans)
made or constituted the heaven, or bodies, or natural bodies, or nature,
or the things that are out of numbers (N1). In his own system, this view
amounts to a claim that numbers are the material cause of things
which Aristotle finds obviously false. He also says that they said that
things (t emta, pq\clata) are numberswhich he finds equally
false. Armed with his distinctions between subject and predicate, be-
tween substance and attribute, between essence and accident, and in
particular in the light of his view that numbers are attributes, not sub-
stances, he cannot but regard these Pythagorean claims and most of
the other claims listed above (N1-N13) as arrant nonsense. No one
sharing these views of Aristotle would think of claiming that numbers
are the principles of things, let alone that things (including physical ob-
jects and other substances) are identical with numbers or made up of
numbers. On the other hand, the Pythagoreans put these ideas forward,
which suggests that they did not hold the Aristotelian views with which
they conflict.
In this last section of my paper I will propose an explanation why the

Pythagoreans put forward these views. I will not proceed by looking for a
view of reality that substitutes number for substance as the basic kind of
entity and perhaps turns other Aristotelian doctrines upside-down as well,
for I do not suppose that the fifth-century Pythagoreans had any devel-
oped ideas in these directions, any more than anyone else did before
Plato did and Aristotle after him. Instead I will take up some threads of
the Aristotelian evidence that I have not emphasized until now, and try
to make the case that the general approach of the early (Stage 1 and
Stage 2) Pythagoreans to these issues was quite different from Aristotles,
and resembled ways of thinking noticeable in Heraclitus, who may have
been contemporary with this period of Pythagorean development.
Much of the first part of passage C1 (C1.1-C1.10) describes Stage 1,
with the exception of the introductory material (C1.1 C1.2), C1.8 (in
which the interest in the elements of number is appropriate to Stage 3),
and C1.9 (which is characteristic of Stage 2). The remaining material in
C1 does not suggest either that the people who were enthusiastically
noticing that things resembled numbers in various ways had even
reached the point of declaring that things are made of numbers or are
identical with number, the theses characteristic of Stage 2. In fact the
Stage 1 Pythagoreans were onto the important idea that many important
features of things and their behavior and their relations can be described
numerically. According to the Aristotelian passage this valuable discov-
ery may have prompted a program of research: to identify the numerical
relations that things actually have.
Aristotle does not object to the discovery of the numerical attributes
and ratios of harmonic intervals. Recall that he says that they saw these
things, not that they just supposed that they observed them; this was a
genuine discovery, not a fancy.150
This is a long way from justifying the grand claims for number that
characterize Stage 2. Or so it appears to us, and so it did to Aristotle,
who directs his serious criticism at claims that reflect that stage. But it
may not have been so far for the early Pythagoreans. It seems to have
been characteristic of early presocratic thought to make sweeping pro-
nouncements on the basis of what we would consider grossly inadequate
evidence. Consider Aristotles statement about Thales basis for claiming
that water is the material principle.
150 See above, C1.5.

He may have got this idea from seeing that the nourishment of all things is
moist, and that even the hot itself comes to be from this and lives on this
and also because the seeds of all things have a moist nature and water is the
principle of the nature of moist things. (Metaph. 983b22 27)
Whether or not these were the grounds for Thales sweeping pro-
nouncement about the primacy of water, the fact that Aristotle thinks
they may have been says something about the kind of evidence he
thought the early natural philosophers used and the way they used it.
Closer to our time period is Heraclitus, who held that all things are one.
Heraclituss fragments associate several of his key concepts.151 For
example, he identifies the cosmos with fire and asserts that fire is
ever-living:
The cosmos, the same for all, none of the gods nor of humans has made,
but it was always and is and shall be: an ever-living fire being kindled in
measures and being extinguished in measures. (DK 22 B30)
He gives fire in its purest and most highly active form a controlling role
in the cosmos:
Thunderbolt steers all things. (DK 22 B64)
In the individual, the soul, which makes us alive and directs us, is fiery:
fire and soul are used interchangeably in DK 22 B36 and DK 22
B31:
It is death to souls to come to be water, death to water to come to be earth,
but from earth water comes to be and from water soul. (DK 22 B36)
The turnings of fire: first, sea; and of sea, half is earth and half fiery water-
spout. Earth is poured out as sea, and is measured according to the same
logos it was before it became earth. (DK 22 B31)
Further, fire as the all-controller seems to be identified with the logos,
according to which all things come to be:
For although all things come to be in accordance with this logos
(DK 22 B1)
The active nature of the universe is war and strife, which Heraclitus de-
clares is justice. Further strife and necessity are assigned the all-control-
ling role that also belongs to fire and the logos.
War is the father of all and king of all (DK 22 B53)
151 The following translations and discussion are taken from R. McKirahan, Philo-
sophy before Socrates 2 (Indianapolis 2011) 135 6.
It is necessary to know that war is common and justice is strife and that all
things happen in accordance with strife and necessity. (DK 22 B80)
Events in the cosmos, including transformations of one substance into
another and changes between opposites, are a necessary and universal
war and struggle which is needed to maintain the cosmos in a stable
condition. Whatever results in a particular case, it is part of the overall
process that rules the universe with justice. Heraclitus identifies the jus-
tice and therefore the strife and war in the universe with fire:
For fire will advance and judge and convict all things. (DK 22 B66)
Fire is also associated with God:
God is day and night, winter and summer, war and peace, satiety and hun-
ger, but changes the way fire, when mingled with perfumes, is named ac-
cording to the scent of each. (DK 22 B67)
Both fire and God take on different appearances in different situations,
but keep their own nature. In a sense, fire is the one behind the many,
the unity in all the diversity of the cosmos. Law too may form part of
the same cluster of concepts:
It is law, too, to obey the counsel of one. (DK 22 B33)
Heraclitus thus associates the logos, fire, soul, war, justice, God, and per-
haps law. In some sense they are the same, the ruling element in the uni-
verse, but precisely how they are the same is not clear. For example,
when I strike a match and create some fire, I am obviously not bringing
the eternal logos or God into being. It will not do to demand strict con-
ditions of identity in this context any more than it does when Heraclitus
says that day and night are one (DK 22 B57). In different settings these
concepts take on a variety of relations to one another, sometimes being
virtually identical and sometimes being almost separate. For example,
the burning match is not God, but is related to the cosmic fire (as a
part? by likeness? as an imperfect specimen? as a copy?) and in its small
area of active existence it performs functions which both symbolize and
are a part of the war and justice that rule the world. Heraclitus lacked
conceptual tools and analytical techniques for analyzing such assertions.
Broad claims of sameness or identity were easy to make and hard to chal-
lenge. This state of philosophy and the Greek language suited Heraclitus
purposes: his approach to the unity of the cosmos through the logos re-
quired associating ideas rather than analyzing or separating them. He
needed to bring things together before bringing them apart again.
I suggest that this same associative approach is behind the Pythagorean

move from Stage 1 to Stage 2. To move from saying that (some) things
resemble numbers and that numerical properties are important or even
the most important features of (some) things to saying that all things
are made in the likeness of numbers or that they are by imitation of numbers,
or that they are numbers or are composed of numbers may even have been an
inevitable step in the context of early presocratic thought. Once they had
come to believe that numbers and numerical relations are fundamental to
one or more kinds of entities and phenomena that they judged important
or revelatory, they made the generalization that they are fundamental to
all entities and phenomena, and they formulated this generalization in
ways that could apply across the board. Whereas things may resemble
numbers in very different ways, the tendency in Stage 2 was to erase
the differences by expressing the general point in a single formula, such
as all things are numbers or all things are 1j numbers, both of
which are deliciously ambiguous in how they can be interpreted. The
former can be taken as a straightforward identity statement (as if the oc-
tave is literally a ratio of two numbers), or as a statement of essence (as if
the essence or basic nature of the octave is the ratio 2:1), and perhaps in
other ways as well. The latter can be taken as a statement indicating the
composition or components of a thing (as if the octave was composed out
of a numerical ratio), or as a statement indicating the source from which a
thing is generated (as if the ratio 2:1 generated the octave) and perhaps in
other ways as well. Different ways of taking these generalizations will
work better for different things and phenomena; it is too much to
hope that any one way will precisely work for them all. But these fine
distinctions seem not to have been the concern of the Stage 2 Pythagor-
eans any more than they were for Heraclitus. Enough to have found a
formulation that can work for everything.
The generalizations that mark Stage 2 amount to a working hypoth-
esis. In effect they constitute a challenge to find the numbers for a wide
variety of phenomena, which would confirm the hypothesis. This seems
a reasonable way to account for the program of identifying such things
as justice, soul/mind and the right moment with particular numbers.
Alexanders apparently well informed account of the reasons for these
identifications (quoted above, p. 81) helps us by providing an account
of these connections. The reasons he gives for associating these things
with these numbers are disparate, and some of them appear accidental
and ad hocfeatures that agree with Aristotles picture of a group of
people determined to find numerical connections by hook or by
crook. It is hard to imagine what would have motivated the Pythagor-

eans to make these identifications unless they were already committed to
the view that numbers are somehow fundamental to things.
This approach to theory construction is the antithesis of Aristotles
approach. Where it tried to bring different kinds of things together, Ar-
istotle aimed to keep similar things apart; where it tried to synthesize Ar-
istotle preferred to analyze. For example, there are many ways in which
one thing or event can be an aQt_a (conventionally translated cause) of
another; Aristotle attempted to identify and distinguish among all those
ways. It turns out that we know relatively little about the relation be-
tween two things if we only know that the one is the cause of the
other. We are badly mistaken if we suppose that the one is the cause
of the other in one way when it is actually the cause of the other in a
different way. When he examines the views of earlier philosophers
that he finds relevant to the question of the different kinds of causes,
he is guilty of using Procrustean methods to fit what they said into
his own theory of the four causes. But he is trying to do something
positiveto translate the earlier theories into a more precise conceptual
framework. In effect he is trying to make sense out of their potentially
nonsensical claims. He believes that the senses of aQt_a he recognizes are
adequate to capture the meanings of earlier, vaguer claims, a belief that
may well fail in the case of the Pythagoreans. It seems a hopeless task to
locate their numerical accounts of phenomena under Aristotles head-
ings: material, formal, efficient, and final. We might suppose that the
ratio 2:1 is the formal cause of the octave, but in what Aristotelian
sense is the number 5 the cause of marriage? And then there is the Py-
thagorean 1j locution in N1, which is a common Aristotelian way of
referring to the material cause: but as Aristotle pertinently asks, how
can physical things be made out of numbers?152 And if it is difficult or
impossible to understand these accounts each as falling under one of
the four Aristotelian headings, it is much more so to see how anyone
could have thought that they all fall under the same one. This is the
kind of result that Aristotles philosophical approach can be expected
to have on products of earlier ways of thinking.
When Aristotle criticizes Pythagorean views he makes comments
along these lines.
152 See n. 104.

From [the Pythagoreans and from Alcmaeon] we can learn this much, that
the contraries are the principles of the things that are; and how many these
principles are and what they are, we can learn from [the Pythagoreans].
How to bring these principles together with the causes we have named
has not been clearly articulated by them. (Metaph. 986b2 6, tr. based on
the Revised Oxford Translation)
Down to the Italian school, then, and apart from it, philosophers have
treated these subjects [viz. the question of causes] rather obscurely.
(Metaph. 987a9 11, tr. based on the Revised Oxford Translation)
Regarding the question of what things are they began to make state-
ments and definitions, but treated the matter too simply. For they would
define superficially and they thought that the first thing an indicated
term applies to is the essence of the thing. (Metaph. 987a20 3, tr. from C2)
[Platos] divergence from the Pythagoreans in making the one and
the numbers separate from things, and his introduction of the Forms,
were due to his inquiries in the region of definitory formulae (for the ear-
lier thinkers had no tincture of dialectic). (Metaph. 987b29 32, tr. based on
the Revised Oxford Translation)
That bodies should be composed of numbers, and that this should be
mathematical number, is impossible. For it is not true to speak of indivisible
spatial magnitudes; and however much there might be magnitudes of this
sort, units at least have not magnitude; and how can a magnitude be com-
posed of indivisibles? But arithmetical number, at least, consists of abstract
units, while they say that the things that are, are number; in fact they apply
their propositions to bodies, supposing that they consisted of those num-
bers. (Metaph. 1083b11 19, tr. based on the Revised Oxford Translation,
cf. 1090a32)
It is absurd to construct an account of the generation of things that are
eternal, or rather it is an impossibility. There is no need no doubt whether
or not the Pythagoreans construct such an account, since they say clearly
that when the one had been constituted the nearest parts of the unlim-
ited at once began to be drawn in and limited by the limit. But since they
are constructing a cosmos and wish to speak the language of natural science,
it is fair to make some examination of their physical theories, but to let
them off from the present inquiry; for we are investigating the principles
at work in unchangeable things, so that it is numbers of this kind whose
generation we must study. (Metaph. 1091a12 22, tr. based on the Revised
Oxford Translation)
11. Conclusion
In this paper I have presented reasons to think that Aristotle did not base
his account of Pythagoreanism, and of Pythagorean principles in partic-
ular, entirely on Philolaus, and that he utilized more than one source.
The possibility that he had live informants is not excludedeither peo-
ple he would identify as Pythagoreans themselves, or people with per-

sonal knowledge of Pythagoreans. This view accounts for several other-
wise puzzling features of Aristotles accounts of Pythagorean thought,
including the almost complete absence of the name Philolaus, the refer-
ences to a plurality of conflicting views held by the Pythagoreans, and
some further inconsistencies in Pythagorean views as Aristotle reports
them. It also makes it plausible to look for different strata in the account,
which may reflect an actual historical development from the earliest
stages of Pythagoreanism down to the so-called last generation of the
Pythagoreans who heard Philolaus and Eurytus (D.L. 8.46).
I have sketched a possible history of Pythagorean thought on num-
bers and principles which I think is consistent with the Aristotelian evi-
dence I have presented. It is an incomplete sketch, but it provides a
rough basis for understanding some important features of Aristotles ac-
counts, in particular their emphasis on the Pythagorean views that things
are numbers or are composed of numbers. It also suggests the possibility
that Philolauss views on principles and the role of numbers on the one
hand were the acme of Pythagorean thinking in these areas, and on the
other hand, were not mainstream Pythagorean doctrines. Finally I have
attempted to account for Aristotles evident frustration with some of the
Pythagorean claims he reports and criticizes. Its not just that he judged
them falsein this they share the fate of most features of all pre-Aristo-
telian theories. Rather, they originated in a way of thinking which Ar-
istotle had rejected and against which he waged continuous warfare.153
Bibliography
Barnes, Jonathan (ed.). The Complete Works of Aristotle. The Revised Oxford
Translation, 2 vols. Princeton 1984.
153 It would have taken too much space in this already long paper to discuss
Zhmuds important article (L. Zhmud, All is number? Basic doctrine of Py-
thagoreanism reconsidered, Phronesis 24 [1989] 270 292), which argues that
Pythagorean number philosophy is the invention of Aristotle. Briefly, I agree
with Zhmud (290) that the discovery of the numerical basis of the harmonic
intervals was early (Zhmud attributes it to Pythagoras himself). Apparently un-
like Zhmud, I suppose that this discovery became the catalyst for the grand
claim that number is (somehow) fundamental to everything, which motivated
further Pythagorean interest in numbers that was manifested in mathematical
discoveries as well as in fruitless numerology, and which eventually led to ac-
counts of the principles of numbers themselves.
The Presocratic Philosophers 2. London 1982

Burkert, Walter. Lore and Science in Ancient Pythagoreanism. Cambridge,
Mass. 1972. Tr. E. L. Minar, Jr.
Diels, Hermann. Elementum. Leipzig 1899.
Huffman, Carl. Philolaus of Croton, Pythagorean and Presocratic. Cambridge 1993.
Platos disagreements with the Pythagoreans: A response to Brisson and
Kahn, a talk delivered at the Pacific Division meeting of the American
Philosophical Association, San Francisco, 2005.
McKirahan, Richard. Philosophy before Socrates2. Indianapolis 2011.
Ross, W. D. Aristotles Metaphysics. 2 vols. Oxford 1924.
Smyth, Herbert W. Greek Grammar. Cambridge, Mass. 1956.
Zhmud, Leonid. All is number? Basic doctrine of Pythagoreanism reconsid-
ered, Phronesis 24 (1989) 270 292.
4. Philolaus Critique of Heraclitus
Carl Huffman
There have not been many attempts to connect the philosophical sys-
tems of Heraclitus and Philolaus. At first sight, there appear to be
good reasons for this failure to see a connection. The two thinkers
come from opposite ends of the Greek world. Heraclitus was from
Ephesus in Asia Minor and Philolaus from Croton in southern Italy, al-
though he also visited in Thebes on the Greek mainland and may have
spent his last years in a different southern Italian city, Tarentum1. More-
over, Heraclitus probably wrote his book ca. 500 BC, while Philolaus
did not write until some seventy years later. Philolaus immediate pred-
ecessors were figures like Empedocles and Anaxagoras, whereas Heracli-
tus belonged to a still earlier generation and is thus less likely to have
elicited a direct response from Philolaus. Heraclitus is, moreover, fa-
mous for his paradoxical theses that apparent stability is, in fact, based
on change and that apparent opposites are in some sense one. Neither
this doctrine of the unity of the opposites nor the emphasis on the
role of strife and change in the cosmos are to be found in the fragments
of Philolaus. There have been, in addition, scholarly obstacles to con-
necting the two figures. The fragments of Philolaus had been thought
of questionable authenticity until fairly recently, so that there was noth-
ing to compare to Heraclitus. Heraclitus himself has been interpreted in
radically different ways. Scholars cannot agree whether Heraclitus is
talking about the cosmos as a whole or about human behavior, nor, if
he is talking about the cosmos, whether he is an exponent of a doctrine
of radical flux or the champion of a stable order. Thus, it is hard to
know which Heraclitus to compare to Philolaus.
Now that a core of the fragments of Philolaus can confidently be
regarded as authentic2, however, these fragments can help us to under-
stand both Philolaus and Heraclitus. There is not space here to argue in
1 C. A. Huffman, Philolaus of Croton: Pythagorean and Presocratic (Cambridge 1993)

6.
2 W. Burkert, Lore and Science in Ancient Pythagoreanism (Cambridge, Mass. 1972);
Huffman Philolaus (above, n. 1).
122 Carl Huffman
detail for the interpretation of Heraclitus on which I base the compar-

ison to Philolaus, if there is to be time to compare the two thinkers. My
paper can thus only claim to be a study of the relationship between Phi-
lolaus and one Heraclitus, although this is the interpretation of Heracli-
tus that I find most plausible and I will refer to scholarship supporting
this interpretation. I will argue that Philolaus and Heraclitus do share
two important themes. Both philosophers emphasize that reality does
not bear its meaning on its surface, i. e. that the phenomena give signs
of a deeper reality on which knowledge is based. Both philosophers
also emphasize the role of harmonia, fitting together, in explaining
the cosmos. They differ concerning what it is that the phenomena
give signs of and on the nature of the harmonia. In addition, although
nothing in the fragments of Philolaus suggests that he was responding
directly to Heraclitus or wrote with specific passages of Heraclitus
book in mind, I will also argue that the fragments of Philolaus offer a
strong implicit critique of Heraclitus metaphysical principles. Philolaus,
in part, accepts Heraclitus analysis of the substances and powers that
make up the cosmos, but his system can also be seen as a critique of
Heraclitus for his failure to give an adequate account of the structural
features of the world.
One of the most prominent themes in the fragments of Heraclitus is
the emphatic assertion that the truth about reality does not lie on its sur-
face; human beings must interpret the raw data of phenomena in order to
come to a deeper understanding of reality. The clearest statement of this
theme is found in B 123: v}sir jq}pteshai vike?, which, as Graham has
shown, should properly be translated not as a personification of nature but
impersonally: it is characteristic of nature to be hidden3. Our first en-
counter with the world gives us not the true nature of things but hints or
indications that point to that true nature. Heraclitus models his own dis-
course on this characteristic of the world. Thus, the description of the
Delphic oracle in Fr. 93 is usually taken as a description of both the phe-
nomenal world and also the words of Heraclitus: The lord to whom the
oracle at Delphi belongs neither tells nor conceals but gives a sign. An
oracle given by Apollo cannot be taken at face value and requires inter-
pretation. When Apollo tells Croesus that he will destroy a great empire,
Croesus comes to grief because he does not take time to interpret the ora-
cle. He does not enquire which empire is meant but simply assumes that
3 D. Graham, Does nature love to hide? Heraclitus B 123 DK, CP 98 (2003)

175 179.
4. Philolaus Critique of Heraclitus 123
the empire is that of his enemy, Cyrus4. Similarly, to understand either the
world or Heraclitus words, it is necessary to be able to interpret the
signs5. Thus, one must value what the eyes and ears reveal to us of the
world, as Heraclitus says in B 55, but also recognize that the eyes and
ears give bad evidence for those who have souls that do not understand
the language6. Human beings must understand the language of the phe-
nomenal world, if they are to interpret the signs that it gives about the
true nature of things. This emphasis on the hidden nature of reality
that is nonetheless close at hand, for those who can read the signs, explains
Heraclitus constant lament about the failure of most people to see the
truth, which is presented as common to all7, and his comparison of
most people to those who are asleep8.
Heraclitus uses several images to describe the reality to which the
phenomena point. In B 54, he says that the harmonie which is hidden
is better than the one that is evident. Once again the emphasis is on
going beyond that with which our senses initially present us, beyond
what can be seen, to what is hidden. This time, however, we are told
that the nature that lies hidden is an harmonia, i. e. a fitting together
or attunement of disparate things into a unity. The Greek word has sev-
eral different connotations, all of which may be relevant in interpreting
Heraclitus. A "qlom_a comes to refer to the fitting together of notes into
a musical scale, so that B 54 might be stating the paradox that the music
that is not evident, that we do not hear, is better than the music that we
do hear. In typical oracular fashion, Heraclitus challenges us to figure
out what is meant by music that is not evident. Kahn9 suggested that
we might think of the Pythagorean music of the spheres, which accord-
ing to the reports of the later tradition only the master could hear10. He
rightly rejects this approach, both because it is not certain that the con-
ception of the harmony of the spheres did go back to Pythagoras and
thus had been promulgated by the time of Heraclitus, and more impor-
tantly because it would be surprising to see Heraclitus embrace such a
distinctively Pythagorean doctrine, given his virulent criticism of Pytha-
goras elsewhere in the fragments. Pythagoras is called the prince of im-
4 Herodotus 1.91.
5 Kahn, H. 123 4.
6 B 107.
7 B 3.
8 B 1.
9 Kahn, H. 202 4.
10 Por. VP 30, Iamb. VP 66.
124 Carl Huffman
posters and is said to have practiced a polymathy that led not to under-
standing but rather to an evil trickery (Frs. 81, 40, 129). It is surprising,
then, that Kahn goes on to argue that by the hidden harmony Heraclitus
is referring to another prominent idea in the Pythagorean tradition,
which may go back to Pythagoras himself and is certainly present in Phi-
lolaus, i. e. the correspondence between the musical concords (what we
today call perfect intervals) such as the octave, fifth and fourth and the
whole number ratios 1:2, 2:3 and 3:4. According to Kahn, Heraclitus
non-evident attunement, the music which we cannot hear, is constitut-
ed by these whole number ratios. Kahns previous argument would
seem to work against his own suggestion here. Does it make sense to
suppose that in praising the hidden harmonie Heraclitus is adopting
as a central feature of his system a distinctive idea of that Pythagoras
whom he treats with such contempt elsewhere? Moreover, it is going
far beyond any evidence we have to suppose that Heraclitus was refer-
ring to the whole number ratios that are connected to the musical con-
cords; these ratios appear nowhere in the fragments of Heraclitus.
In order to make sense of Heraclitus in his own terms, it is necessary
to remember that in a more basic sense "qlom_a refers not just to musical
structure, but to any fitting-together of elements into a whole. Thus the
verb "ql|fy is used already in Homer to describe Odysseus fitting-to-
gether of pieces of wood to create a raft, and the "qlom_ai are the
joints11. It is thus reasonable to translate "qlom_a in some contexts as
structure. Kahn indeed suggests that hidden structure might
thus be taken as a general title for Heraclitus philosophical thought12,
and Hussey translates "qlom_g !vam^r as latent structure13. B 51 sug-
gests that it is this more general sense of harmonia that is most relevant for
Heraclitus, although the musical meaning is present as well: They do
not understand how in differing with itself it agrees, a backward turning
structure ("qlom_g) just like that of a bow or a lyre. The image of the
bow and the lyre in this fragment is complex, and I will focus on just a
few aspects of it here. The evident structure of the bow and lyre is the
combination of wood and string to form an apparently stable and static
object. Most people do not understand that this agreement of the parts
of the lyre or bow is based on a difference between them; the difference,
11 Od. 5. 247 8.
12 Kahn, H. 203.
13 E. Hussey, Heraclitus, in A. A. Long (ed.), The Cambridge Companion to Early
Greek Philosophy. (Cambridge 1999) 91.
the strife between the strings and the wood of the bow and the lyre, al-
though not evident, is nonetheless the true structure that underlies the
apparently static structure. It is not only the bow or the lyre that are held
together by such a structure, however. Heraclitus is arguing that there is
such a hidden fitting-together for everything in the world and indeed
for the world as a whole.
In our best source for B 51, Hippolytus, the harmonia of the bow and
lyre is described as backward turning (pak_mtqopor), while some
other sources report the common Homeric epithet for a bow, back-
ward stretched (pak_mtomor); scholars have been divided as to which
reading to accept. 14 Backward turning is clearly preferable as the lectio
difficilior. As Kahn15 has argued, however, this adjective is intended to
remind us of the tension on which the bow and lyre are based by an al-
lusion to the Homeric epithet, backward stretched. In the transmis-
sion of the text, this allusion was sometimes taken too seriously and
backward stretched was incorporated into the text instead of the
more unusual backward turning. At the same time, however, Kahn
argues that backward turning alludes to the turnings (tqopa_) or
changes of state of the cosmic fire, which are described in B 31 A:
The turnings of fire, first sea, and of sea the half is earth and the half
is lightning. B 51 thus indicates that the hidden harmonia of B 54 is
in part a reference to the constant transformation between opposites
such as the hot and the cold, which makes such opposites just phases
on a continuum of temperature. The hidden harmonia thus turns out
to be a reference to the famous Heraclitean unity of the opposites,
which asserts that apparent opposites such as hot and cold are, in fact,
one. What is real is not the transitory phases, hot and cold, but the uni-
fied continuum of temperature of which they are parts. Thus, in B 126,
Heraclitus asserts Cold things become hot, hot things become cold,
wet things dry up, dry things become damp.16
As Graham has suggested, it appears that Heraclitus is presenting a
critique of Ionian cosmology, which had mistakenly supposed that
powers and substances such as the hot or air were ultimate realities,
when they were just phases of the continual change, which is the
14 See, e. g., Kirk, H. 210 216, and G. Vlastos, On Heraclitus, AJP 76 (1955):
337 378.
15 Kahn, H. 199 200.
16 R. Dilcher, On the Wording of Heraclitus, Fragment 126, CQ 44.1 (1994)
276 277.
126 Carl Huffman
true reality.17 Graham puts it very well in saying, In effect each sub-
stance is not a permanent thing, a stable reality, but a stage in the se-
quence of transformations. What then is real? Only the process of trans-
formations itself.18 This then is the explanation of the role of fire in
Heraclitus. To quote Graham again, fire is the most changeable,
ephemeral, the least tangible and stable of things; thus it is the least sub-
stantial of all substances, the substance that is not a substance, the sub-
stance that is a process.19 I would suggest, then, that for Heraclitus
the opposites define and are united in continua in which powers,
such as different degrees of hot and cold, or substances, such as air
and water, are transitory phases. Thus the hidden harmonia is more likely
to be a reference to this famous Heraclitean unity of opposites than to
the whole number ratios that govern the musical concords. It is not
completely clear whether Heraclitus is arguing that there are a number
of different continua that unite opposites, e. g. a continuum of temper-
ature that unites hot and cold, a continuum of moisture that unites dry
and wet etc., or whether he is arguing that all oppositions are ultimately
united in one continuum. His assertions that all things, god, or the
wise are one suggest the latter interpretation.20 In B 67, god is descri-
bed as uniting a series of such continua: God, day night, winter summ-
er, war peace, abundance famine21.
At this point it is illuminating to turn to Philolaus in order to note
two similarities in approach with Heraclitus. First, Philolaus also empha-
sizes that human beings will not understand reality just by looking at the
phenomena presented to them by their senses. Things once again point
to a more basic reality. This theme is clearest in B 5. After identifying
two basic kinds of numbers, the even and the odd, Philolaus says, Of
each of the two kinds there are many forms, of which each thing itself
gives signs.22 The many forms of the even and the odd may just be the
series of even and odd numbers. Philolaus argues that each thing in the
world gives signs of or points to these numbers. He uses the same verb,
which Heraclitus used to describe the Delphic oracle as giving signs,
signs that are in need of interpretation. It is certainly not the case that
17 D. Graham, Heraclitus Criticism of Ionian Philosophy, OSAP 15 (1997) 1

50.
18 Graham (above, n. 17) 35.
19 Graham (above, n. 17) 37.
20 B 50, 67, 32.
21 Kahn, H. 278.
22 For the text of the fragment, see Huffman Philolaus (above, n. 1).
things appear on their surface to have anything to do with numbers but,

so Philolaus suggests, closer study shows that there are indications, signs,
that they are ultimately based on numbers. The clearest examples are the
whole number ratios that govern the musical concords to which Philo-
laus refers in B 6a. When an octave is heard, we do not immediately
perceive a number in it but, if we examine the phenomenon of the oc-
tave, we will discover that strings whose lengths are in the ratio 2:1 will
produce the interval of an octave, when plucked in succession. B 4
makes explicit what is implicit in B 5, that we only gain true under-
standing of things, when we come to understand the numerical struc-
ture to which they point and on which they are based: And indeed
all things that are known have number. For it is not possible that any-
thing whatsoever be understood or known without this. In order to
really understand anything, we must understand the numbers of
which it gives signs.
Philolaus, like Heraclitus, thinks that human beings first experience
things through their senses and he appeals to what the senses tell us
twice in a crucial argument in B 2. Philolaus regards it as manifest (va_-
metai) from our direct experience of the world that it is not composed just
of components that are unlimited or just of components that limit, but
that both sorts of components are necessary: it is manifest that they
[the cosmos and all the things in it] are neither from limiting things
alone nor from unlimited things alone. When we look at something
in the world, such as a forest, we see not just unlimited, unstructured
stuff, such as wood, nor just elements that impose limits and structure,
such as shapes. We see wood that has the structure of trees. I will return
to this important point in a moment but what is crucial here is to see that,
while Philolaus does appeal, here in B 2, to what is manifest to the senses
in order to establish certain general features about the cosmos, i. e. that it
is composed both of what is unlimited and what limits, B 4 and 5 show
that he, like Heraclitus, does not think that we gain a true understanding
of anything unless we go beyond what is manifest at first sight to grasp the
underlying structure, which governs what we perceive. Philolaus differs
from Heraclitus in making explicit that the underlying structure is numer-
ical. I will also return to this point in a minute.
A second striking point of similarity is that, as in Heraclitus, Philo-
laus too argues that understanding the nature of the cosmos and indeed
the nature of everything in the cosmos means coming to understand an
"qlom_a. The opening lines of Philolaus book stress that nature in the
world-order was fitted together ("ql|whg) both out of things which are
128 Carl Huffman
unlimited and out of things which are limiting, both the world-order as
whole and everything in it. This is an explicit assertion that nature, the
nature of the cosmos as a whole and of all things in it, is an attunement
of parts, is a structure. This intimate connection between nature and har-
monia is also seen in B 6 of Philolaus, which begins Concerning nature
and harmonia the situation is this This pairing of the two concepts
suggests that they are closely connected and that in order to understand
nature it is necessary to understand harmonia. Philolaus and Heraclitus
thus agree that things give signs of a harmonia. Again, however, the fur-
ther characterization of harmonia, which Philolaus will give in B 6a, as
having a numerical structure, the numerical structure of the diatonic
scale, shows that the ultimate nature of the harmonia is more explicitly
arithmetical in Philolaus than Heraclitus.
It is now time to turn from the broad similarities between Heraclitus
and Philolaus to the substantial differences that constitute Philolaus cri-
tique of Heraclitus. First, Philolaus appears to accept Heraclitus con-
ception of harmonia as a starting point but to move beyond it. For Her-
aclitus the hidden harmonie is constituted by his doctrine of the unity of
opposites, the surprising connection of opposites such as hot and cold,
up and down in one continuum. In part this connection is logical and
not empirical. Hot and cold, light and dark, up and down are only in-
telligible in terms of one another and thus cannot be separated logically
from one another. Heraclitus famous assertion of method that I
searched myself (B 101) may refer to his turning inward to examine
the logic of such opposites independent of experience. In the case of
the natural world, this logical connection becomes a physical connec-
tion through the doctrine of constant change, which makes hot and
cold, dry and wet, light and dark, just phases in larger continua. One
way of understanding Philolaus is to suppose that he has adopted Her-
aclitus point and accepts that the world is composed of an indefinite
number of such continua. Where do such continua appear in Philolaus?
I suggest that he made them one of the two basic types of components
of the cosmos and called them t %peiqa, the unlimiteds. In Heraclitus
unity of opposites doctrine, as I have sketched it above, the opposites are
one because they are parts of a constantly changing continuum. Such a
continuum is not made up of defined stable entities, but rather consists
of transitory phases. It is precisely because the parts of such a continuum
are not defined stable states that they can be said to be one. The contin-
uum of temperature cannot be labeled hot or cold or any other temper-
ature but is in itself without clear internal boundaries; this is the sense in
which it is apeiron. It is plausible to regard Philolaus as having reflected

on this aspect of Heraclitus thought and as having drawn the crucial
conclusion that what is most fundamental about opposites such as hot
and cold or substances such as fire and air is not their particular qualita-
tive character, but rather that they are parts of continua or are them-
selves continua. It is these continua without any internal boundaries,
i. e. t %peiqa, the unlimiteds, that are ultimate realities, not the transi-
tory phases of the continua.
Philolaus candidates for membership in t %peiqa are good candi-
dates for the kind of continua Heraclitus was talking about. In B 6a Phi-
lolaus treats sound as an %peiqom, and in Heraclitean terms sound can be
regarded as a continuum, which is the unification of the opposites high
and low pitch. A fragment of Aristotles book on the Pythagoreans,23
which appears to be describing Philolaus cosmogony, reports that
from the unlimited time and breath were brought in, as well as the
void, thus suggesting that time, breath, and void are themselves unlim-
iteds for Philolaus. Similarly, in B 7, since the central fire is described as
fitted together, and elsewhere when things are fitted together they are
said to be combinations of limiter and unlimited, it is probable that fire
is being treated as an unlimited, which is combined with the limiting
notion of center to form the central fire. All of these unlimiteds can
be regarded as Heraclitean continua. In the case of time and void, it
is the opposites before and after that are unified in a temporal and in
a spatial continuum, respectively. For Philolaus, at the level of both
the macrocosm and also the microcosm, the first step in generation is
for cooling breath to be drawn in by the heat of fire. The central fire
draws in cooling breath24 just as the newborn child does.25 Breath and
fire are parts of a Heraclitean continuum of elemental change which
unites the opposites hot and cold, but parts of continua are continua
themselves so that breath and fire are treated as themselves unlimiteds.
What Philolaus shares with Heraclitus in all these cases is the thesis
that what is most fundamental about such things as void, time, breath,
and fire, is not any specific qualitative characteristics they may have
but rather that they are continua. Philolaus, like Heraclitus, is asserting
the essential malleability of entities such as breath and fire, which other
Greek thinkers assumed to be fundamental realities.
23 B 201.
24 Aristotle Fr. 201.
25 Philolaus A 27.
130 Carl Huffman
Unlike Heraclitus, however, Philolaus shows no inclination to sup-

pose that these separate continua are somehow united in a single con-
tinuum so that all things are one. Instead, he seems to regard it as obvi-
ous that the world is composed of a variety of such continua, and in B 6
argues that further specification of a privileged set of continua is beyond
human competence: the being of things which is eternal and nature
in itself admit of divine and not human knowledge, except that it was
impossible for any of the things that are and are known by us to have
come to be, if both the limiting and the unlimited things did not pre-
exist. Our experience tells us that the world is constituted by limiting
things, to which I will turn in a minute, and by unlimiteds or continua.
There are a large number of continua (sound, time, space, fire, breath
etc.), but our experience does not allow us to say any more than that.
One striking thing about Philolaus adoption of Heraclitus point is
that what was for Heraclitus a startling and paradoxical assertion of
the unity of opposites, which had to be explained by a hidden harmonie,
is regarded as obvious. The unity of opposites in continua is now an un-
remarkable logical truth, and Philolaus does not invoke harmonia to ex-
plain how such opposites can be held together. Harmonia plays a differ-
ent role in Philolaus.
In order for the new role of harmonia in Philolaus philosophy to be-
come intelligible, it is first necessary to consider something new that Phi-
lolaus has seen and of which Heraclitus had only a partial glimpse: the
need for structural as well as material first principles in order to explain
the cosmos. Heraclitus harmonie of opposites is based on constant change
in a continuum, but he asserts in several places that this change is not ran-
dom or irregular and that it has measure. The clearest assertion of this
point is found in B 30: This cosmos, the same for all, neither any of
the gods nor of men made, but it always was and is and will be an
ever living fire, kindling in measures and being quenched in measures
(l]tqa). Thus, the constant change across a continuum that is the cosmos
proceeds according to measures. In another fragment Heraclitus asserts
that the sun will not go beyond his measures, but if he does the Furies,
the assistants of Justice, will find him out.26 It is controversial what the
measures of the sun are to which this fragment refers. The text of the frag-
ment in Plutarch uses the same word as in B 30, l]tqa. The quotation of
the fragment in the Derveni papyrus, however, suggests that the original
text may have been ouqour, which is an Ionic form of fqour, whose basic
26 B 94.
meaning is boundary or limit27. It has commonly been assumed that these

boundaries are all the regularities of the suns behavior. The Derveni
papyrus, however, quotes B 94 as forming a continuous text with B 3,
in which the sun is described as having the size of a human foot. Thus
the boundaries of the sun may be simply the boundaries of the apparent
size of the sun as equal to one human foot28. For my purposes here, how-
ever, exactly what the boundaries are is less important than the fact that
the sun has limits and measures. There is an interesting difference between
these measures of the sun and the measures of fire. In the case of the sun,
the measures appear to be imposed from outside by Justice and her min-
isters, the Furies. In the case of the cosmic fire, however, it sounds as if
the fire is self-regulating; the limits arise from the fire itself, which kindles
itself and quenches itself in measures. Long has suggested that it is this self-
regulating aspect of fire that, in part, led Heraclitus to give it its privileged
position in his system29.
Philolaus finds Heraclitus account of structure and measure in the
cosmos problematic. Where do these measures and limits come from?
Saying that they are imposed by the god is as good as giving no answer.
Nor is it enough to say that the cosmos is a fire behaving according to
measures. Why does the fire behave in this way? Philolaus argues that
we cannot explain the world just in terms of the unified continua,
which Heraclitus has identified as the true reality and which Philolaus in-
corporates into his own system as t %peiqa, unlimiteds. Principles of
measure and structure, limiters, must have equal status as first principles.
B 2 sets out the possibilities It is necessary that the things that are be
all either limiting, or unlimited, or both limiting and unlimited and not
in every case unlimited alone. By the things that are Philolaus
means the basic principles of the world. Logically these principles could
be 1) all limiting or 2) all unlimited or 3) both limiting and unlimited.
It is this third possibility for which Philolaus goes on to argue: both limit-
ers and unlimiteds are fundamental principles. Even in his statement of the
three possibilities, however, Philolaus makes clear that he is particularly
anxious to argue against the second possibility, i. e. that all things are un-
27 G. Betegh, The Derveni Papyrus: Cosmology, Theology and Interpretation. Cam-

bridge 2004.
28 D. Sider, Heraclitus in the Derveni Papyrus, in A. Laks and G. Most (eds.),
Studies in the Derveni Papyrus, (Oxford 1997) 140; see also A. Finkelberg, On
Cosmogony and Ecpyrosis in Heraclitus, AJP 119 (1998): 195 222.
29 Heraclitus, in E. Craig (ed.), Routledge Encyclopedia of Philosophy, London
1998.
132 Carl Huffman
limited, since in stating the third possibility, i. e. that things are both limit-
ing and unlimited, he adds pointedly, and not just unlimited. He ap-
pears to be correcting earlier thinkers, such as Heraclitus, who had sup-
posed that the world could be explained in terms of unlimiteds alone.
Philolaus argues that we must recognize that measures, limits and struc-
ture are features of reality, which cannot be eliminated, and that such
structure is not inherent in continua. The cosmos is not a fire alone,
but it has a central fire and other fires delimited within it.
The difference between Heraclitus and Philolaus regarding structure
can be further clarified by examining another Heraclitean image, the ky-
keon, a drink composed of barley, cheese and wine, which has to be stir-
red and still in motion when one drinks it in order to realize its full na-
ture. Mouraviev30 has raised important questions about the text of this
fragment, but appears to agree with most scholars that Heraclitus uses
the kykeon as an analogue to the cosmic order31; as constant motion is
necessary to keep the barley and cheese from settling out in the kykeon,
so constant motion is what maintains the celestial order of sun, moon
and stars, including the measures of the sun mentioned in B 94. Hera-
clitus thus appears to think that constant motion across a continuum is
all that is needed to explain the cosmos. The measures and order of the
cosmos somehow arise out of the constant change in the continuum,
analogously to the order that arises out of the ingredients of the kykeon
in motion32. It is this emergence of measures and limits out of the un-
limited continuum that Philolaus regards as unintelligible. In B 2, he
goes on to argue for the third logical possibility, that both limiters
and unlimiteds are first principles, by starting from our experience: it
is manifest that they [i.e. the cosmos and the things in it] are neither
from limiting things alone nor from unlimited things alone, i. e. in
the world around us, we easily identify continua like temperature,
water and earth, but also limits such as shapes. Heraclitus references
to measures and limits suggest that he would agree with Philolaus up
to this point. Philolaus goes on to argue, however, that it is clear
then that the world-order and the things in it were fitted together
from both limiting and unlimited things; the limiting and unlimited
features of the world we observe can only be explained on the assump-
30 S. N. Mouraviev, The Moving posset once again: Heraclitus Fr. B 125 in con-
text, CQ 46 (1996) 34 43.
31 Kahn, H. 194.
32 B 125.
tion that the world has as starting points, as !qwa_, both limiters and un-
limiteds. Philolaus continues, Things in their actions also make this
clear. For, some of them from limiting constituents limit, others from
both limiting and unlimited constituents both limit and do not limit,
others from unlimited constituents will be manifestly unlimited. Philo-
laus point is clearest in the last case, others from unlimited consti-
tuents will be manifestly unlimited. If the starting points are unlimited
what arises from them will be unlimited. From clay alone, all we get is
an undifferentiated quantity of clay, not a vase. It is not possible for limit
to arise from what is in its nature unlimited. Material in motion cannot
in itself explain the order that is observed in the cosmos. Heraclitus is
mistaken if he thinks that measures can arise from a burning fire.
Whereas Heraclitus tends towards monism, Philolaus always uses
plural nouns to describe his first principles, i. e. %peiqa, unlimiteds,
and peqa_momta, limiters. It would appear that he thought that the cos-
mos originated from an irreducible plurality of both limiters and unlim-
iteds. It is natural, then, to ask how many continua and how many struc-
tural principles there are. Philolaus, however, denies that this knowledge
is accessible to human beings. In B 6, when Philolaus focuses on the
most fundamental principles of reality, what he calls the the being of
things and nature in itself, rather than identifying a basic set of stuffs,
as, for example, Empedocles does, he says that such principles admit of
divine and not human knowledge. There is one large exception, i. e.
that it was impossible for any of the things that are and are known
by us to have come to be, if the being of the things from which the
world-order came together, both the limiting things and the unlimited
things, did not preexist. What human beings can know about ultimate
reality is precisely that the basic principles must include both continua,
unlimiteds, and also principles of structure, limiters, in order for the cos-
mos that we observe to have arisen. What humans cannot know is what
and how many unlimiteds and what and how many limiters are basic
principles. Philolaus does not see any way to determine that there are
just four elements, as Empedocles does. There may once again be a par-
tial similarity with Heraclitus on this point, since, on one interpretation,
Heraclitus also denies that it is possible to identify any one phase of the
changes in the material continuum that is the cosmos as more funda-
mental than any other. On this reading, he is not identifying fire as
the most basic of substances but rather uses fire as a symbol of the con-
134 Carl Huffman
tinuum in constant change, which he regards as more basic than any of

the phases of that continuum33. Heraclitus, however, appears to draw
the conclusion not that there are an indefinite number of continua all
equal to one another, as Philolaus does, but rather that there is just
one principle, the continuum, of which fire, water, earth and air are
phases; he refuses to posit structural first principles at all.
If we agree with Philolaus, however, that both sorts of basic princi-
ples must be presupposed, both limiters and unlimiteds, another prob-
lem arises. How is the relationship between limiters and unlimiteds to
be conceived? Heraclitus taught that opposites such as hot and cold
are connected by a harmonie, in that they are united as phases of constant
change across a continuum. I have argued that Philolaus accepted this
sort of unity. B 6 shows that Philolaus regards such a unity of opposites
as unremarkable and as not even requiring an appeal to a harmonia in
order to explain it. Hot and cold are not so different after all. They
are, in fact, alike as being parts of the same continuum. In B 6, Philolaus
argues first that things that are alike (blo?a) are not in need of any har-
monia. He is probably appealing here to the principle widely recognized
among early Greek philosophers that like naturally coheres with like34.
There is no need to invoke any further explanatory principle to explain
why water coheres with water or earth with earth. Philolaus takes the
point a step further, however. He argues that even things that are
bl|vuka, literally of the same tribe, naturally cohere. In Greek city-
states the tribe was the most general term of affiliation. At Athens you
were the member of a family first, then a clan, then a brotherhood or
phratry, and finally a tribe. Obviously a tribe could embrace people
with even diametrically opposed points of view. In the Hippocratic
treatises the term again indicates a similarity at the most general of levels;
it refers to copulating partners as being of the same species in order to
produce offspring35. Thus, Philolaus might well argue that opposites
that are part of a continuum, such as hot and cold, while not alike,
are part of the same tribe, are naturally related to one another and
not in need of any harmonia in order to fit them together.
The situation is different when we look at Philolaus basic principles,
limiters and unlimiteds. A limiting shape such as a circle and an unlimited
continuum such as water or earth are two quite different sorts of reality,
33 E.g., Graham (n. 17 above).

34 Democritus B 164.
35 Hipp. Nat. hom. 3.3.
yet Philolaus argues that these two radically different sorts of things must
work together in order to account for the cosmos. Thus, in B 6, Philolaus
asserts that since these beginnings [!qwa_, i. e., limiters and unlimiteds]
preexisted and were neither alike nor even related (of the same tribe,
bl|vukoi), it would have been impossible for them to be ordered, if a har-
mony had not come upon them. He supports this assertion by stating the
general principle that things that are unlike and not even related nor
even of the same order [reading Qsotac/], it is necessary that such things
be bonded together by a harmony, if they are going to be held in an
order. Thus, that master of paradox, Heraclitus, was dealing with some-
thing quite tame when he asserted that opposites such as hot and cold
were one. It is much more of a challenge to explain how center and
fire can be combined to form the unity known as the central fire. We
might well want to say that Philolaus is making some sort of category mis-
take in regarding center and fire as parts of the cosmos in anything like the
same sense, but Philolaus language in B 6 suggests that he is all too aware
of the seemingly unbridgeable gap between the sense in which fire is part
of our cosmos and the sense in which center is a part of our cosmos. He
emphasizes that limiters and unlimiteds are not just unlike (oqw blo?ai)
one another, they are not even of the same tribe (oqd( bl|vukoi). He
went on to use yet another adjective in order to indicate the vast differ-
ence between limiters and unlimiteds. He may well have used a neolo-
gism to highlight the difference, since the manuscripts provide the appa-
rently impossible reading not of equal speed (lgd Qsotaw/). Heidels
emendation lgd Qsotac/ (not of the same order) is perhaps the
most plausible. Philolaus would then, in fact, be recognizing that limiters
and unlimiteds do not belong even to the same category of being. Here is
a real gap, which we must bridge with some sort of attunement. Up to
this point I have avoided using the language of form and matter, and I
will continue to do so, in the belief that it is less misleading to understand
Philolaus insight in his own terms and those of his contemporaries and
predecessors. Nonetheless, it may be helpful to make the point by saying
that Philolaus is the first person to see that formal features are just as im-
portant as material features in explaining the world and is hence the first
person to grapple explicitly with the problem of how form and matter can
be related.
It is worth emphasizing one aspect of Philolaus answer to the prob-
lem of relating the limiting features of the world to the unlimited features.
As I have suggested above, since Philolaus can see no way in which limits
and measures can arise from what is unlimited, he argues that limiters no
136 Carl Huffman
less than unlimiteds must be assumed as first principles or starting points

for the cosmos (!qwa_). Given this argument, which suggests a strong par-
allelism between limiters and unlimiteds, it is not at all surprising that, as
the first thinker to take limiters seriously, Philolaus should suppose that
limiters are just as much components of the cosmos as unlimiteds. The
limiter sphere is just as much a part of the cosmos as an unlimited,
such as earth. This equality of treatment of limiters and unlimiteds is
shown in the strict parallelism between limiters and unlimiteds in all as-
sertions of his basic thesis about the cosmos; in every case limiters and un-
limiteds are joined by the simple conjunction and (ja) or the paired
conjunctions both and (ja ja, te ja_). This reading of Philolaus
treatment of limiters is confirmed by Aristotle. Aristotle only mentions
Philolaus by name once, but the philosophical system, which he assigns
to a group of thinkers whom he labels the so-called Pythagoreans
and which he dates to exactly the time when Philolaus was active, is so
similar to what we find in the fragments of Philolaus, that there is little
doubt that these fragments are one of Aristotles principal sources. One
of Aristotles central points about this system is that no separation between
numbers and things is recognized. He also complains that numbers are
treated as if they were physical bodies36. It seems to me that Aristotle is
slightly misreading Philolaus here, since Philolaus never puts numbers
on the same level as material objects; it is always limiters, not numbers,
which are paired with unlimiteds. Nonetheless, Aristotle is right to em-
phasize that, while Plato does separate forms from things, the Pythagor-
eans do not37. The fragments of Philolaus support such a conclusion in-
sofar as they treat the limiting formal aspects of reality on all fours with
the unlimited material elements and suggest no hierarchical relationship.
It is important to emphasize, however, that the evidence suggests that Pla-
tonic forms did not originate from Pythagoreanism but rather from Soc-
ratic dialectic, so that while Philolaus limiters are an attempt to account
for formal aspects of the cosmos, they have a quite different origin than
Platonic forms and hence are quite different in nature38. Aristotle regarded
it as unintelligible that the Pythagoreans should treat formal elements as if
they were material components of the cosmos, but as we have seen Phi-
36 Metaph. 1083b ff.

37 Metaph. 987b28.
38 C. A. Huffman, Philolaus (1998), in Edward N. Zalta (ed.), The Stanford En-
cyclopedia of Philosophy; <http://plato.stanford.edu/archives/fall2008/entries/
philolaus/>.
lolaus is not totally unaware of the difficulty. While he emphasizes in B 6

that the cosmos is not intelligible unless we assume limiters as first prin-
ciples alongside unlimiteds, he also highlights the difference between
these two sorts of principles by describing them as not even of the
same tribe. How then can these unlike but equally important principles
of the cosmos be combined? It is here that numbers enter Philolaus sys-
tem.
In B 6, Philolaus is emphatic that it is because limiters and unlimi-
teds are so unlike that they are in need of a harmonia, a principle to fit
them together. As we have seen, from Philolaus point of view, oppo-
sites like hot and cold, which Heraclitus thought in need of fitting to-
gether, are close relatives in comparison to a limiter like a sphere and an
unlimited like earth. B 6 does not say what this harmonia consists of, just
that one is required, if limiters and unlimiteds are ever to be combined
into a cosmos, into an ordered world: things that are unlike and not
even related nor of the same order [?], it is necessary that such things be
bonded together by harmony, if they are going to be held in an order.
In B 6a, which may be a direct continuation of B 6, Philolaus is much
more specific about what a harmonia is and begins by asserting that the
magnitude of harmonia is the fourth and the fifth. Here the harmonia is
being equated with the musical interval known as the octave, which is
composed of two intervals, the fourth and the fifth. The octave and
these two other intervals will then be identified with the ratios 2:1,
4:3, and 3:2 respectively. Thus, for Philolaus, a harmonia is not defined
by constant change across a continuum as in Heraclitus, but instead is a
number or more accurately a ratio. It looks as if the numbers of which
things in the world give signs are at least in part the numerical ratios that
govern the musical concords. Problems immediately arise. What sense
can be made of the idea that numbers are what fit together limiters
and unlimiteds? How can numbers make such unlike things as limiters
and unlimiteds, structural principles and continua, work together?
Meinwald has recently suggested that the idea of a ratio between a struc-
tural principle and a material principle is deeply problematic and that
this problem led Plato in his later metaphysics to introduce the great
and the small as a modification of Philolaus account of the unlimited.
Instead of supposing that there is just an unlimited continuum, the con-
tinuum is a dyad defined by the great and the small. The ratio then be-
comes more intelligible as a ratio between two unlimiteds such as hot
and cold, which express greater and lesser degrees of temperature. Ac-
138 Carl Huffman
cording to Meinwald, Plato then hoped that all forms might be defined
by such ratios of unlimiteds39.
Philolaus clear identification of harmonia with numerical ratios40
need not mean, however, as Meinwald suggests, that he envisioned
the ratios as ratios between limiters and unlimiteds. What he may
have been thinking can be most easily understood in terms of the proc-
ess of making something, e. g. a statue or a temple. Philolaus own ex-
ample seems to be the making of a musical scale, which at first sight ap-
pears unhelpfully abstract. Such an activity can be made concrete by
thinking of the tuning of an actual lyre or of experiments on a mono-
chord. I will use the monochord as an example, although there is no
good evidence that it was this of which Philolaus was thinking. The
string which is used to make the monochord is an example of something
that is unlimited, something that is a continuum. If a bridge is inserted
under the string somewhere, so that the whole string can no longer vi-
brate but only each of the parts, a limit has been imposed on the con-
tinuum. In the abstract the bridge could go anywhere along the contin-
uum, but once placed it defines a specific length of string. The two parts
of the string now have numbers, which are determined by the place-
ment of the bridge. If one part is twice as long as the other, then one
part has the number two in relation to the number one of the other.
Now bridges could be inserted at any point along the continuum but,
when placed in most positions, the intervals that arise from plucking
the two parts of the string will not be pleasing to the ear. If, however,
the bridges are placed so that the string lengths have the ratios 2:1, 3:2
and 4:3, the concordant intervals of the octave, fifth and fourth are pro-
duced. This example shows how it is that numbers do hold together
limiters and unlimiteds. In one sense, any time a limit is imposed on
an unlimited, any time a bridge is placed under a string, a number arises.
The string cut off by the bridge will have some length. If we are trying
to produce not just a random combination of limiter and unlimited, but
want to impose a limiter on an unlimited in a way that produces a pleas-
ing order, i. e. what the Greeks meant by a cosmos, then we will have to
impose a limit on the unlimited in a way that produces specific num-
bers. In the case of the octave, fifth and fourth these numbers are the
ratios 2:1, 3:2 and 4:3. The musical example may thus have suggested
to Philolaus that any time a limiter and an unlimited are combined, a
39 C. C. Meinwald, Platos Pythagoreanism, AP 22 (2002) 87 101.

40 Huffman Philolaus (above, n. 1) 60.
number results, and that thus it is a number that defines or fits together
the limiter and the unlimited, although the number is not the same as
either the limiter or the unlimited.
In studying Philolaus metaphysical scheme, it is easy to assimilate
numbers and limiters and to assume that limiters just are numbers. Ar-
istotle appears to make such an assimilation when he asserts that the Py-
thagoreans thought that things were made of numbers. In the surviving
fragments of Philolaus, however, numbers and limiters are clearly kept
separate. In statements of his central metaphysical principles, it is always
limiters and never numbers that are paired with unlimiteds. Thus B 1
asserts that nature in the world-order was fitted together out of things
which are unlimited and out of things which are limiting; it does not
assert that nature was fitted together out of things that are unlimited
and out of numbers. It is in B 6 that the connection between numbers
and limiters and unlimiteds becomes clear. The things from which the
world-order came together are again identified as both the limiting
things and the unlimited things with no mention of numbers. Philolaus
is thus pointing out that the world our senses reveal to us arises both
from continua such earth and also from limiters such as a spherical
shape. A sphere as a geometrical entity, however, need have no specific
size and hence is tied to no specific number. The propositions in Eu-
clids Elements that demonstrate the properties of spheres or circles are
never about spheres or circles with specific sizes but rather about a
sphere of whatever size. B 6 then goes on to point out, however,
that in order for a continuum and a limiter actually to be combined
to produce a physical entity, a harmonia or fitting-together must occur
and, when Philolaus goes on to discuss this fitting-together in B 6a, it
becomes clear that the continua and the limiter must be combined ac-
cording to a number: e. g. the sphere must be assigned a specific diam-
eter. Again, if the combination of limiter and unlimited is to produce a
cosmos, a pleasing order, the number that binds them will not be just any
number but one of a specific set of ratios such as those we find govern-
ing the concordant musical intervals. The crucial thing to recognize is
that the number is not identical with the limiter, but arises as limiter
and unlimited are combined in a specific entity.
When Philolaus asserts that it is a numerical fitting-together that
makes a limiter and an unlimited an ordered whole, he is not really ex-
plaining how it is that number serves to unite limiter and unlimited. In-
deed, in B 6, Philolaus admits his ignorance of how it is that this numer-
ical harmonia arises. He asserts of limiters and unlimiteds that it would
140 Carl Huffman
have been impossible for them to be ordered, if a harmonia had not come
upon them, but immediately adds in whatever way it came to be,
thus acknowledging that exactly how number combines limiter and un-
limited is beyond human knowledge. Despite this inability to explain
the precise mechanism of how number combines limiter and unlimited,
Philolaus is making an important and substantial point. He has observed
empirically that whenever a limiter and an unlimited are combined nu-
merical relationships emerge and that when the combination produces a
particularly pleasing order, such as that of a musical scale or indeed the
cosmos as a whole, the numerical relationships are of a specific sort, e. g.
whole number ratios. Philolaus project is thus ultimately to give an ac-
count of the world in terms of the numerical relationships to which
phenomena point. This project is taken over and significantly developed
by his successor Archytas, as I argue in my recently published edition of
the fragments of Archytas41.
It is now time to return to Heraclitus. My presentation of Heraclitus
and Philolaus critique of him takes a position in a long standing debate
about the essence of Heraclitus message. In my presentation the central
theme in Heraclitus is the paradoxical assertion of the unity of opposites,
which is in turn based on constant change across a continuum. This em-
phasis on the theme of constant change is in accord with the interpre-
tations of Guthrie, Popper, Vlastos, Stokes, and Barnes, among others.
The opposing side of the debate argues that Heraclitus emphasis is
not on change, but on the measure and order of the change in the
world. This view was advocated by Reinhardt and received its most
forceful presentation by Kirk42. In the epilogue to his book Kirk sum-
marizes Heraclitus achievement as follows:
The discovery of and emphasis on the arrangement of things, rather than
their gross material constitution is perhaps the most important one in
the history of archaic speculation. Heraclitus must be given full credit
for having developed this concept so as to produce the first reasonably co-
herent explanation of the world of experience.43
My argument suggests that Kirk is wrong on this point. Heraclitus great
achievement was precisely to give a sophisticated analysis of what Kirk
calls the gross material constitution of things. What earlier thinkers
41 Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King (Cam-

bridge 2005) 65 67.
42 For detailed references, see Graham (n. 17 above) 1 3.
43 Kirk, H. 430 434.
had supposed to be the basic material constituents of things water, air

etc. turn out to be just phases in a continuum that is united by the con-
stant change between phases. Philolaus gratefully accepts this analysis and
duly identifies the unlimiteds, which are just the sorts of continua that
Heraclitus identified, as one of his two basic kinds of principles.
There is not time here to develop all the arguments against Kirks
position. It is certainly true that Heraclitus recognizes that measure
and structure are important in the world, but Heraclitus recognition
of this fact is nothing new and is, in fact, an important point of contact
between Heraclitus and his predecessors. Anaximander thought that the
opposites that constituted the world were governed by an order, the
order that he identifies, in the sole surviving fragment of his book, as
justice or dike: the opposites give justice and retribution to each
other for their injustice according to the order of time44. Moreover,
there is evidence that, just as in Heraclitus cosmos the sun has to ob-
serve limits, Anaximanders cosmos was arranged according to clear
structural principles, which related the size of the earth to the distance
between the earth and the sun, stars and moon45. Heraclitus rhetoric is
not directed against people who fail to recognize structure or measure in
the cosmos, as we would expect it to be, if Kirks thesis were true. If
Heraclitus thought that his great achievement was the discovery of
and emphasis on the arrangement of things, we would expect to
find him saying things like people think that the world is just a chaotic
mixture and do not see the measures that govern it. Instead, when he
says in B 54 that the hidden attunement is better than the obvious
one, he seems to be saying that people do recognize structure in the
world, they just fail to understand its true basis. He instead repeatedly
emphasizes that his contemporaries and predecessors fail to recognize
that the world depends on strife and that the supposed opposites are
united by that strife: It is necessary to know that war is common
and justice is strife and that all things come to be in accordance with
strife and necessity46.
Someone might object that, when he calls on human beings to listen
to the logos, in B 1 and 50, Heraclitus is precisely calling on them to pay
attention to the structure in the world, which they have been neglect-
44 B 1.
45 C. H. Kahn, Anaximander and the Origins of Greek Cosmology (New York 1960)
85 98.
46 B 80.
142 Carl Huffman
ing. If logos is translated as structure in these fragments, however, it is

important to be careful about what is meant by structure in these con-
texts. The logos is a rational account of the universe, which most human
beings overlook, but it is illegitimate to assume from the word logos
alone either that Heraclitus is focusing on structure per se, as Philolaus
does in postulating limiters as first principles, or that Heraclitus is pro-
posing that the world is structured according to mathematical ratios
such as we find in Philolaus. The content of the rational account and
the type of structure which Heraclitus heralds must be developed
from the fragments themselves. In B 50, there is an explicit statement
of what the logos tells us: listening not to me but to the logos it is
wise to agree that all things are one. Thus, according to B 50, the
core message of the logos is the unity of all things, and in the rest of
the fragments of Heraclitus the unity that is emphasized repeatedly is
the unity of opposites. Thus it seems overwhelmingly likely that the
unity which the logos teaches is the unity of opposites, and hence the
constant change upon which that unity is based. Heraclitus is the proph-
et of a specific sort of structure, the unity of opposites, not of structure as
a fundamental principle.
It thus appears that the discovery of and emphasis on the arrange-
ment of things, rather than their gross material constitution, which
Kirk heralds as the most important one in the history of archaic spec-
ulation, should not be attributed to Heraclitus, but rather to Philolaus.
Heraclitus and his predecessors certainly saw the world as full of struc-
ture and gave bold characterizations of some of those structures, such as
Anaximanders cosmic justice and Heraclitus unity of opposites, but
their first principles remained material. Heraclitus gives the most sophis-
ticated analysis of the nature of these material constituents of the world,
but he nonetheless still tries to explain the world in terms of the constant
change within material continua. Philolaus is the first thinker to object
that the measures and structures that all early Greek thinkers, including
Heraclitus, observed in the world cannot arise from starting points or
first principles that consist only of such continua, only of unlimiteds.
Structural principles must be given the same fundamental status as ma-
terial principles in order to explain the structure which all observe in the
cosmos. Philolaus uses precisely the rhetoric that we would expect for
someone who was emphasizing that the structure of things is equal in
importance to their material constitution the rhetoric that is missing
in Heraclitus. In B 2, Philolaus objects pointedly that things are not
in every case unlimited alone and in his statements of his central thesis
the bothand structure underlines the message that structure, i. e.

limiters, is just as important as the continua, i. e., unlimiteds. Nature
in the world-order was fitted together both out of things which are un-
limited and out of things which are limiting47. Again, when he comes
to state what the harmonia is that holds the cosmos together in B 6a, Phi-
lolaus presents us with whole number ratios, not with the harmonie of a
bow or lyre, which turns out to be the unifying strife between oppo-
sites. In both Heraclitus and Philolaus, things give signs of a hidden na-
ture. In Heraclitus the hidden nature is a material process. It is only in
Philolaus that the hidden nature turns out to be a precisely defined
structure, a structure determined by numerical relationships.
Bibliography
Betegh, Gbor. The Derveni Papyrus: Cosmology, Theology and Interpretation.
Cambridge 2004.
Burkert, Walter. Lore and Science in Ancient Pythagoreanism, tr. E. Minar. Cam-
bridge, Mass. 1972. 1st German edn., 1962.
Dilcher, Roman. On the wording of Heraclitus, fragment 126, CQ 44
(1994) 276 277.
Finkelberg, Aryeh. On Cosmogony and ecpyrosis in Heraclitus, AJP 119
(1998) 195 222.
Graham, Daniel W. Heraclitus Criticism of Ionian Philosophy, OSAP 15
(1997) 1 50.
Does Nature love to hide? Heraclitus B123 DK, CP 98 (2003) 175 179.
Huffman, Carl A. Philolaus of Croton: Pythagorean and Presocratic. Cambridge
1993.
Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King. Cam-
bridge 2005.
Archytas, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy,
2008. http://plato.stanford.edu/archives/fall2008/entries/archytas/.
Philolaus, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy,
2008. http://plato.stanford.edu/archives/fall2008/entries/philolaus/.
Two Problems in Pythagoreanism, in D. Graham and P. Curd (eds.), The
Oxford Handbook of Presocratic Philosophy (Oxford 2008) 284 304.
Pythagoras, in E. N. Zalta (ed.), The Stanford Encyclopedia of Philosophy,
2009. http://plato.stanford.edu/archives/win2009/entries/pythagoras/.
Hussey, Edward. Heraclitus, in A. A. Long (ed.), The Cambridge Companion to
Early Greek Philosophy (Cambridge 1999) 88 112.
Long, Anthony A. Heraclitus and Stoicism, in id. Stoic Studies (Cambridge
1996) 35 57.
47 B 1.
144 Carl Huffman
Heraclitus, in E. Craig (ed.), Routledge Encyclopedia of Philosophy. London

1998. http://www.rep.routledge.com/article/A055.
Meinwald, Constance C. Platos Pythagoreanism, AP 22 (2002) 87 101.
Mouraviev, Serge N. The Moving posset once again: Heraclitus Fr. B 125 in
context, CQ 46 (1996) 34 43.
Pierris, Apostolos. Origin and nature of early Pythagorean cosmology, in K.
Boudouris (ed.), Pythagorean Philosophy (Athens 1992) 126 162.
Sider, David. Heraclitus in the Derveni Papyrus,in A. Laks and G. Most
(eds.), Studies in the Derveni Papyrus (Oxford 1997) 129 148.
Vlastos, Gregory. On Heraclitus, AJP 76 (1955) 337 378. Repr. in id. Stud-
ies in Greek Philosophy, ed. D. Graham. Vol. 1, The Presocratics (Princeton
1995) 127 50.
5. Heraclitus, the Rival of Pythagoras
Aryeh Finkelberg
As Walter Burkert observed,1 towards the end of the first century of the
Christian era a change in the Verstndnishorizont of the epoch brought
about a new reading of old texts; Heraclitus and Empedocles had
come to be regarded as the predecessors of the Platonic-Pythagorean
doctrine of the descent of the soul into the corporeal world. The au-
thentic evidence at our disposal does not validate the Platonists percep-
tion of Heraclitus as an eschatological thinker on a par with Pythagoras,
Empedocles, and Plato.2 Yet, considering the random character of our
collection of Heraclitus fragments, the lack of authentic evidence of
a full-fledged eschatological doctrine does not indicate decisively that
he had none. Theophrastus silence about Heraclitus eschatological
teaching carries little weight as he passed over Empedocles myth of
the fallen daimn as well. The Platonists list of the eschatologists was
fixed, with no traceable attempts to incorporate additional thinkers.
Therefore, Heraclitus inclusion in it cannot be explained as accidental;
so denying the testimony evidential value for Heraclitus will beg the
question and, considering that the other philosophers on the list were
really eschatologists, a special pleading too. Consequently, we should
admit that in our collection of Heraclitus fragments the eschatological
facet of his teaching is underrepresented, and we have no prospect of
forming any adequate idea of Heraclitus eschatology if we confine
our discussion to the authentic evidence.
1 W. Burkert, Plotin, Plutarch und die platonisierende Interpretation von Her-

aklit und Empedokles, in J. Mansfeld and L. M. de Rijk (eds.), Kephalaion. Fes-
tschrift de Vogel (Assen 1975) 137 146.
2 G. Teichmller, Neue Studien zur Geschichte der Begriffe I. Herakleitos (Gotha
1876) 74 75, and E. Rohde, Psyche (London 1925) 2.368) denied Heraclitus
the belief in the survival of the individual souls for systematic reasons, namely
because of its inconsistency with the theory of flux. For this and other reasons
this belief was denied to Heraclitus also by H. Cherniss, ACPPh, 297 n. 29; M.
Nussbaum, Xuw^ in Heraclitus, Phronesis 17 (1972) 153 69; M. Baltes, Die
Todesproblematik in der griechischen Philosophie, Gymnasium 95 (1988)
101 02; M. Conche, Hraclite (Paris 1986) 328; and others.
148 Aryeh Finkelberg
Macrobius, S. Scip. 14.19 (= DK 22 A 15), tells us that, according to

Heraclitus, our soul is a spark of the stellar essence (scintillam stellaris
essentiae). Macrobius source is uncertain, but Heraclitus association of
the souls with the stars is supported by Galen (4.786.10 Khn):
For he [sc. Heraclitus] said: Dry radiance (aqc^): soul is wisest [B 118],
asserting, for his part, that the dryness is the cause of intelligence (for the
word radiance indicates this), and one should regard as better the opinion
of those who, having discerned that the stars are both luminous and dry,
maintain that they have the highest intelligence.
To determine what Macrobius stellar essence is, we should turn to
Heraclitus account of the heavenly bodies which is intrinsically
bound up with his theory of exhalation. The theories at issue are report-
ed in Diogenes excerpt from Theophrastus (D.L. 9.7 11 = DK 22 A
1), the relevant portion of which runs as follows:
(9) Exhalations arise from earth and from sea; one kind is bright and pure
and the other is dark. Fire is increased by the bright exhalations, moisture, by
the others. He says that there are bowls in the heaven with their con-
cave side turned towards us, in which the bright exhalations are gathered and
produce flames, which are the stars. (10) The brightest and hottest is the
flame of the sun. For the other stars are farther from the earth and for that
reason give less light and heat, whereas the moon, though it is nearer the
earth, is borne through a region which is not pure. The sun, however, lies
in a lucid and unmixed region and that is why it gives more heat and
light
From the report we gather that the bright, dry, and pure exhalation is
contaminated by the dark, wet, and impure exhalations in the sublunary
region (an impure region through which the moon is borne), but
higher up it becomes pure and entirely bright (a lucid and unmixed re-
gion where the sun and the stars lie) and is set aflame in the bowls of
the stars.3 If the bright and dry exhalation fills the whole area between
3 Consequently, if, as Clement Strom. 5.14.104.3 (= B 31) tells us, the heavens,
and things they encompass, i. e., the heavenly bodies, are produced by sea;
i. e., the sea exhalation, the bright and dry exhalation rises from sea, while the
opposite, the dark one, rises from earth. Accordingly, in Diogenes c_meshai d(
!mahuli\seir !p te c/r ja hakttgr, $r lm kalpqr ja jahaqr, $r d sjo-
teim\r, the opposition $r l]m $r d] is to be construed as a chiastic reversal of the
main clause: Exhalations arise from earth and from sea, those (i. e., the latter)
bright and pure and those (i. e., the former) dark (so, for instance, in R. D.
Hicks Loeb translation and in T. M. Robinson, Heraclitus (Toronto 1987) 168.
5. Heraclitus, the Rival of Pythagoras 149
the earth and the uppermost heaven, it must comprise both atmospheric
air and heavenly fire.
According to Aristotle de An. 405a24 (=DK 22 A 15), Heraclitus
says that the principle is soul, if indeed soul is the exhalation of which
everything else is composed. Aristotles reasoning in equating fire, dry
exhalation, and soul is as follows: if soul is exhalation [and if exhalation is
the principle], then the principle is soul. Aristotle infers the sameness of
principle (fire) and soul from the synonymy of fire and exhalation, on
the one hand, and of soul and exhalation on the other. This inference
indicates that Heraclitus used the term soul synonymously with ex-
halation. Now, if Aristotle equates exhalation with fire, the exhalation
at issue must be bright and dry, and if this exhalation is the same as soul,
it forms soul on the cosmic scale, that is the divine world soul:
Heraclitus says that the soul of the world, which is an exhalation from moist
things in it, and the soul in living beings, which is from the exhalation that is
outside them and in them, are of the same kind (At. 4.3.12 = DK 22 A 15).
Commentators reject the report as crediting Heraclitus with a Stoic doc-
trine. This assessment, however, cannot be defended in view of Aristo-
tles testimony at de An. 405a24, and it is hardly consistent with B 36:
To souls it is death to become water, to water it is death to become earth,
but from earth comes water, and from water soul.
The shift from the plural souls to the singular soul in the otherwise
symmetrical statement must be deliberate and significant. The plural
souls are of course human ones perishing into water, whereas the sin-
gular soul into which water turns evidently refers to the bright and
dry exhalation that forms the soul of the world. This conclusion is re-
inforced by an Orphic imitation of the fragment (Clem.
Strom. 6.2.17.1 = 437 F Bernab = fr. 226 Kern): Water is death to
soul, an exchange for water, / and from water earth, and from earth
water again, / and from it soul, altering the entire sky.4 The traditional
4 5stim vdyq xuw0 [Sylburg, xuw cod.] h\mator d rd\tes<s>im !loib^,

[!loib^ secl. Bywater: wqd\tessi d ca?a Marcovich, acc. Bernab] j 1j d vda-
tor <p]ke> [suppl. West EGPhO 150] ca?a, t d 1j ca_ar p\kim vdyq j 1j toO
d xuw fkom aQh]qa !kk\ssousa [cod. : aQh]q !ma@ssousa Bywater]. The
verb !kk\ssy principally means make different, change, alter, hence also
change position, i. e., go to a place (LSJ, s.v. iii.2). M. L. West, The Orphic
Poems (Oxford 1983) 223, translates transferring to the universal aether, and A.
Bernab, Textos rficos (Madrid 2004) 61, which turns into the universal aeth-
er. These translations practically yield the same sense of altering the entire
view that Heraclitus souls come from water proves wrong: B 36 says
that while human souls perish into water, water itself is continually
transformed into soul-exhalation that forms the soul of the world and
not individual souls. Significantly, this reading avoids the inconsistency
which otherwise exists between B 36 and Macrobius report on the
heavenly origin of our soul.
Thus, Macrobius stellar essence is the bright and dry soul-exha-
lation gathered and set aflame in the bowls of the stars, which are the
heavenly bodies most distant from the earth. It follows that our soul
is a bit of the bright soul-exhalation, in effect a particle of the divine
soul, which comes from the farthest region of the heavens. And indeed,
if bright exhalation is the soul substance, it must be the substance of the
divine and the human soul alike, which, as Atius 4.3.12 (= DK 22 A
15) tells us, are of the same kind.
The soul of the world is divine by virtue of its physical nature, i. e.,
just because it is the dry fiery-airy body; consequently, if the human
soul is physically akin to the divine, it too must be divine. I turn now
to B 119, which has come down to us in three versions:
(i) Stobaeus Ecl. 4.40.23: Ghor !mhq~p\ da_lym.
(ii) Plutarch Quaest. Plat. 999D-E: Ghor !mhq~pou da_lym.
(iii) Alexander De fato 170.18 (= De an. 2.185.23): Ghor !mhq~pym da_-
lym.
Diels preferred Stobaeus !mhq~p\ and G. Bernardakis5 emended Plu-
tarchs !mhq~pou by the analogy of Stobaeus. However, as Cherniss6
pointed out, Plutarchs genitive is supported by Alexanders !mhq~pym;
a genitive form, probably Plutarchs singular, appears to be a superior
reading.7 Considering that elsewhere in Heraclitus Ghor means nature
(B 78: For the human nature has no insight but the divine does), the
preferred translation of the word in B 119 is nature, (non-moral) char-
acter.8 This, in turn, renders the traditional translation of da_lym as
sky (changing the whole aether, C. Osborne, Rethinking Early Greek Philo-
sophy (Ithaca 1987) 6: in being exhaled from water, soul either alters (in effect
forms) the entire heavens or joins them.
5 Plutarchi Moralia, Bibliotheca Teubneriana, 1888 96.
6 In the Loeb edition of Plutarchs Moralia, 13.1, 20 n 2.
7 Cf. J. Bollack and H. Wismann, Hraclite (Paris 1972) 328.
8 This is the sense in which Ghor occurs in Empedocles B 17.28 and probably in B
110.5. In Herodotus, the word never means moral character, disposition,
or the like; see J. E. Powell, A Lexicon to Herodotus, s.v. Ehea. In Plato,
Crit. 120D-121B the opposition t d !mhq~pimom Ghor:B toO heoO v}sir indi-
fate irrelevant, the required sense being god, divine being, demi-
god. This is precisely how Plutarch Quaest. Plat. 999D-E understands
Heraclitus statement:
Is it then his own nature (v}sir), as more discerning than productive, that
he [sc. Socrates] called god (he|m), just as Menander said For our mind
(moOr) is the god and Heraclitus said Mans nature (Ghor) is a da_lym.
Heraclitus idea of the human being, then, is basically the same as that of
Empedocles, who imagined man as a da_lym dressed in an alien robe
of flesh (B 126).
As we recall, Atius says that according to Heraclitus our soul is
made of two exhalations, the external and the internal. The connection
of our soul with external exhalation is explained in Sextus Adv.
math. 7.127 130:
[T]hat which encompasses us (t peqi]wom Blr) is rational and sane
(129) By having drawn in this divine reason (k|cor) through breathing
we, according to Heraclitus, become endowed with a mind9 In
sleep the mind that is in us is separated from its union with the encom-
passing (t peqi]wom) the only attachment preserved being that by breath-
ing as if by a root (130) But on awakening, having peeped through the
sensory passages as through a sort of windows once again and come in con-
tact with the encompassing, it becomes invested with the power of reason
once again [T]he portion of the encompassing hosted in our bodies
by the union (with the encompassing) is made like in kind with the
whole (of it).
H. Diels DG 210 11 argued that since Sextus says that we breathe in
the encompassing k|cor, and since what we actually breathe in is air,
the k|cor, and therefore the principle itself, must be air; this, as we
learn from Sextus Adv. math. 10.233, is Aenesidemus interpretation.
As I have suggested above, however, since in Heraclitus the atmospheric
air must be part of dry exhalation, the latter is the substance we breathe.
cates the synonymy of Ghor and v}sir, a sense quite frequent in Galen (e. g., at
11.9.1: t Ghor toO puqetoO). Note further Pindars affirmation that neither fox
nor lion will ever change its 1lvur Ghor (O. 11.19 20; cf. succemr Ghor in O.
13.13). On the early uses of Ghor see S. M. Darcus, Daimon as a force in shap-
ing ethos in Heraclitus, Phoenix, 27 (1974) 391 94 (although I cannot follow
her interpretation of Theognis, 963 70).
9 moeqo cim|leha. The translation endowed with a mind is suggested by the
correlation between moeqo cim|leha as a result of our having drawn in the en-
compassing k|cor and the subsequent account of the states of b 1m Bl?m moOr,
the mind that is in us; cf. R. Polito, The Sceptical Road (Leiden 2004) 153.
This makes the second premise of Diels reasoning unnecessary. Nor is it

for nothing that in his report Sextus nowhere mentions air. In this con-
nection I propose to consider the following passage of Aristotle (de
An. 411a7):
Some say that it [sc. soul] is intermingled in the whole universe This
presents some difficulties: why does the soul when it resides in air or fire
not form an animal, while it does so when it resides in mixture of the el-
ements, and that although it is held to be better when contained in the for-
mer? One might ask also, for which reason the soul in air should be better
and more immortal than that in living beings. [I]t is beyond paradox to
say that fire or air is a living being [T]hey say that it is through the re-
tention of a bit of the encompassing in the living beings that the living be-
ings become animate. But if the air breathed in is the same in kind, while
soul is made of different parts, etc.10
Aristotles last sentence makes it clear that the retention of a bit of the
encompassing is through breathing. Fire is not a convenient example of
a breathable substance yet Aristotle repeatedly mentions it along with
air.11 This indicates that Aristotle has in mind a particular thinker (or
thinkers) who believed that soul resided in fire, that it formed the
breathable encompassing, and that animals became animate by inhal-
ing this soul substance, which Aristotle, like Sextus, calls the encom-
passing. Aristotle knows of two thinkers, Heraclitus and Hippasus,
who posited fire as the principle. This leaves us with a choice: to
admit that Aristotles passage supports the core of Sextus account, or
to wonder why it should not be Hippasus, and him alone, whom Aris-
totle has in mind.12
An intelligent and rational substance which we breathe in and
which Aristotle and Sextus call the encompassing must be the same as
Atius exhalation which is outside us, i. e. the soul of the world. Sextus
statement by having drawn in this divine reason through breathing we,
according to Heraclitus, become endowed with a mind suggests that
the mind, a particle of the intelligent encompassing which is hosted
10 J.A. Smiths translation with minor changes.

11 From which H. Cherniss ACPPh 309 n. 72 infers that Aristotle is thinking of
Diogenes and Heraclitus.
12 Another lesson of Aristotles passage is that the Stoic conception of the human
soul as a particle (!p|spasla) of the divine traces back to the Presocratic
thinkers. The scholarly view of Atius report on the cosmic soul in Heraclitus
as a Stoic misattribution is therefore unwarranted.
in our bodies, enters us with our first breath. From Numenius (fr. 30
des Places) we learn what happens to souls on entering the body:
Heraclitus said that For souls it is pleasure or13 death to become wet [B
77] and that pleasure for them is falling into birth.
In falling into birth the soul becomes wet, evidently as a result of its
attachment to the body. The internal exhalation which, according to
Atius, is another constituent of our soul along with the dry exhalation
from outside, is thus the wet exhalation of the body. The soul is moeq\,
because it is a particle of the intelligent soul-exhalation and for this rea-
son it is best when it is most like it: A dry radiance: soul is wisest and
best (B 118). In becoming wet the soul grows different from the soul
substance, becoming, so to speak, less soul. As we learn from B 117,
wetness disrupts the souls activity: A man when he is drunk is led
by a beardless boy, stumbling, not perceiving whither he goes, having
his soul wet; and if the wetting impact of the body is unchecked,
the soul eventually perishes, turning into water (B 36). As the wetness
of the body damages the soul, Heraclitus figuratively portrays mortal
life as the souls death. Clement Strom. 3.3.21.1, 5.14.105.2 tells us
that Heraclitus calls birth death and that, according to him, the
path of the soul into the body is sleep and death..14 This is how
Philo and Sextus expound the meaning of B 62 and B 88, respectively;
I run their explanations parallel to Plato, Gorgias 492E-493A:
13 C Diels, l^ codd., see below, n 17.

14 Plut. de Is. 362A-B: And indeed, because the natural philosopher Heraclitus
says that Hades is the same as Dionysus in whose honour they rave and cele-
brate Bacchic rites [B 15], people reach this same opinion. For those who be-
lieve that the body is called Hades of the soul, as if the soul is deranged and
intoxicated in it, draw a close allegory. As O. Gruppe, Die griechischen Culte
und Mythen (Leipzig 1887) 1.649, observed, Plutarchs passage evinces Heracli-
tus belief that while in the body the soul is, as it were, inebriated and dead.
Philo, Leg. alleg. 1.108.1 5: Sext. Pyrrh. hyp. 3.230: Plato, Gorg. 492E-493A:
1. For he [sc. Heraclitus] says: But Heraclitus says that
We live their death, die their
life [B 62]
2. life and death are in us
both when we live and
when we die [B 88]:
3. I should not be surprised, if
Euripides was right when he said:
Who knows, is not to live to be
dead, and to be dead to live?
[fr. 638 Kannicht]. And we, in
truth, are perhaps dead.
For I myself already heard one of
the wise men saying that
4. since now, we are now
when we live, for when we live,
the soul is dead our souls are dead dead
and is buried in the body and buried in us, and the body is to us
as in a tomb; a tomb,
but if we be dead, while when we die
the soul would live the our souls return to
life of its own life and live
5. whereas that part of the soul in
which the desires reside happens
to be prone to seduction and
travels up and down.15
15 Philo, Leg. alleg. 1.108.2 5: Sext. Pyrrh. hyp. 3.230.2 6: Plato, Gorg. 492E-93A:
1. vgs cq [sc. Heraclitus]7 b d Jq\jkeit|r vgsim,
F_lem tm 1je_mym
h\matom, tehm^jalem d
tm 1je_mym b_om [B 62]
2. fti ja t f/m ja t
!pohame?m ja 1m t` f/m
Blr 1sti ja 1m t`
tehm\mai [B 88]
3. oq c\q toi haul\foil( #m eQ
Eqqip_dgr !kgh/ 1m to?sde
k]cei, k]cym t_r d( oWdem, eQ
t f/m l]m 1sti jathame?m, t
jathame?m d f/m; ja Ble?r
t` emti Usyr t]hmalem Edg
c\q tou 5cyce ja Ejousa
t_m sov_m
4. r mOm l]m, r mOm
fte f_lem, fte lm cq Ble?r f_lem,
tehmgju_ar t/r xuw/r tr xuwr Bl_m tehm\mai Ble?r t]hmalem
ja r #m 1m s^lati t` ja 1m Bl?m teh\vhai, ja t lm s_l\ 1stim
s~lati 1mtetulbeul]mgr, fte d Ble?r !pohm-sjo- Bl?m s/la,
eQ d !poh\moilem lem, tr xuwr !mabioOm
t/r xuw/r f~sgr ja f/m.
tm Udiom b_om, jtk
5. t/r d xuw/r toOto 1m
1pihul_ai eQs tucw\mei cm
oXom !mape_heshai ja
letap_pteim %my j\ty.
It is evident that Euripides verses, which Plato has Socrates cite, draw
on Heraclitus. In (4) verbal parallels in the three independent sources
suggest a paraphrase of Heraclitus, which authenticates Philos and Sex-
tus explanation of Heraclitus meaning in B 62 and B 88. Platos that
part of the soul in which the desires reside is of course his own, but his
travels up and down validates the report of Iamblichus (ap. Stob. i 49,
39.40) and Aeneas of Gaza (Theophr. 5.12) that Heraclitus posits the
travel of the soul up and down.16 As Proclus (in R. ii 20.23) puts it,
presumably elaborating Heraclitus, souls travel a circular path not
only led by the destiny as the saying goes, pastured by blows [B
11], but also leading themselves.17
So in a way, our birth is the death of our soul, and conversely our
death amounts to its life: Immortals are mortal, mortals immortal, liv-
ing their death, dying their life (B 62). Because of the double, literal
and figurative, meaning of life and death, and of the related wak-
ing and sleeping, our soul can be described in both ways: The same
thing is in us living and dead, and awake and sleeping, and young and
old (B 88). As the newborn soul of a mortal creature, it is young,
alive, and awake, but as a particle of the divine, it is old, dead, and
asleep.18 It emerges that the famous equation of body (s_la) with the
16 This report is found in the passages of Iamblichus and Aeneas which run parallel
to Plotinus 4.8.1. Since Aeneas exposition of Heraclitus, Empedocles, and Pla-
tos theories goes beyond that of Plotinus, he must have drawn, not on Plotinus,
as is usually assumed, but on a source common to him and Plotinus. Burkert
(above, n. 1) suggests that Plotinus (hence also Aeneas) source was a first-cen-
tury Alexandrian compilation, used also by Plutarch, Clement, and Hierocles.
17 As for the transmitted text of B 77 (cf. above, n 13), h\matom, rejected by
E. Zeller, in Zeller-Nestle, Philosophie der Griechen 6, 1.2.890 n. 2) and I. Bywa-
ter, Heracliti Ephesii Reliquiae (Oxford 1877), fr. lxxii), followed by Burnet,
Gigon, and others, is supported by Numenius quotation of B 62 next to B
77, Proclus as Heraclitus says, it is death for souls to become wet [B 77]
(in Plat. Res. 2.270.30), and, most importantly, the paraphrase (4) in the
above synoptic table where Heraclitus equates incarnate life with the souls
death. This equation makes it clear that the transmitted t]qxim l h\matom
must be wrong; one should read, with H. Diels, E (W. Kranz ja_ is paleo-
graphically less likely). That Porphyry, who borrows here from Numenius,
read E (or ja_) is indicated by his paraphrase of B 77 (in Tim. i, fr. 13.10 Soda-
no): that part of the soul, which is the seat of desire, inundated by the mois-
ture incident to generation is enfeebled and sinks into the flows of matter, and
this is another death of intelligent souls, to become wet [B 77], says Heraclitus.
18 Consequently, the explanation offered in the next, epexegetic clause (for these
having changed round are those, and those having changed round again are
souls tomb (s/la) is Heraclitus. But this is not surprising: the idea is a
corollary of his view of incarnation as the souls death.19
It is evident that if on entering the body the soul, a particle of the
dry exhalation, becomes mixed with the wet exhalation, it will quit
the body moist. This means that death does not deliver the soul from
the impact of bodily wetness.20 So one would expect to find in Heracli-
tus the theory that the disembodied soul is encumbered with bodily ves-
tiges, and consequently needs to be purified of the body while still in
the body. Such a theory would range with the ascetic and cathartic
ideas of the Archaic Age, typified in the Orphic and Pythagorean doc-
trines of pure life.
For souls it is pleasure or death to become wet (B 77). Souls derive
delight from the bodily life which costs them their own life: It is hard to
fight against desire, for whatever it wants, it offers to buy at the price of
the soul (B 85).21 Most men willingly pay this price, living the brutish
life of gratifying appetites: The many stuff themselves with food,
like cattle, while the best prefer one thing over anything else, ever-
these) is quite irrelevant, and U. von Wilamowitz proves right, pace H. Diels, in
rejecting this clause. Besides, even if life and death in Heraclitus are reciprocal,
they are such not in one and the same thing, hence not in us, but between
things. As to young and old, these are not reciprocal on any account.
19 The s_la-s/la doctrine is usually regarded as Pythagorean and/or Orphic. The
former attribution is merely by default, the latter is at odds with Platos words at
Cratyl. 400C, as Wilamowitz, Der Glaube der Hellenen 2.199 and many others
since then have pointed out. (For a recent defence of this attribution see A. Ber-
nab, Una etimologa Platnica: s_la-s/la, Philologus 139 (1995), esp. 218
23.) Platos wording is not the only difficulty in assigning the s_la-s/la doctrine
to the Orphics: recent decades have witnessed growing awareness that the sharp
dualism of soul and body was not part of the Orphic worldview: A. Boulanger,
Le salut selon lOrphisme, (Paris 1940) 73; U. Bianchi, LOrphisme a exist,
in id., Selected Essays on Gnosticism, Dualism and Mysteriosophy (Leiden 1978) 190;
L. J. Alderink, Creation and Salvation in Ancient Orphism (Chico 1981) 62 64, 83,
88 89; G. Casadio, Adversaria orphica, Orpheus 8 (1987) 392; R. Parker,
The Early Orphism, in A. Powell (ed.), The Greek World. (London 1995) 499.
20 Here is the answer to the difficulty raised by P. Tannery, Pour lhistoire de la sci-
ence hellne (Paris 1887) 190: if the difference between the da_lym before the
incarnation and the human soul consists in the impurity of the latter as a result
of its association with the body, why did Heraclitus attach importance to our
conduct in earthly life, if death restores the souls purity?
21 hul` l\weshai wakep|m f cq #m h]k,, xuw/r me?tai. Offer to buy, bid is a
usual sense of m]olai in the present and imperfect (LSJ, s.v. 1; J.E. Powell, Lex.
Herod., s.v. 2), a sense which suits the purport of the statement better than the
traditional translation buys: to fight hul|r is to decline its tempting offers.
lasting glory over mortal things (B 29). The everlasting glory alludes to
martial valour which Heraclitus praises in B 24 (Gods and men honour
those slain in battle) and evidently also in B 25 (Greater deaths win
greater destinies), which Plato (Cratyl. 398B) and Theodoretus (Gr.
aff. cur. viii 39) associate with B 24. Obviously, one who performs feats
of bravery and courage, laying down ones life, behaves in a way contrary
to gratification of the body.
At Strom. 4.4.14.4 Clement reports a theory of the ancients on the
difference between dying in battle and of disease:
And the ancients among the Greeks praise the death of those fallen in
war because he who dies in war passes away without fear of dying,
cut off from the body, without having suffered sickness in the soul nor
slackened, which men undergo in diseases. For they depart in a womanly
manner, longing to live on. For these reasons they let their soul go not
pure but carrying its appetites with itself, like weights of leadsave
those among them who have become renowned for goodness. There are
also those who die in war still with their appetites, and with them it
would make no difference whatsoever if they had died of disease.
Having set against these ideas his concept of purity as life in the knowledge
of God and in agreement with the Commandments, Clement proceeds:
Then Heraclitus says: Gods and men honor those slain in battle [B 24]
and Plato in the fifth book of the Republic [468E-69A] writes: Of those
who die in battle, whoever dies having distinguished himself shall we
not, first of all, say that he is of the Golden race?
The ancients among the Greeks whom Clement has in mind appear
to be Heraclitus and Plato. Since Plato never contrasts two kinds of
death or otherwise disapproves of death by disease, the theory must
be Heraclitus,22 the more so in that it involves the peculiar idea of
the disembodied soul burdened by bodily vestiges. This idea reappears
in Plutarch Rom. 28.8, who phrases it in distinctly Heraclitean physi-
cal terms, mentions Heraclitus by the name, and cites B 118:
For this [sc. soul] alone is from gods [Pindar, fr. 131b.2 3 Snell/
Maehler], for thence it comes thither it returns, not with the body, but
whenever it is most removed and separated from it and becomes complete-
ly pure and bare of flesh and holy. For this is a dry and best soul [B 118],
according to Heraclitus, darting out of the body like lightning out of a
22 As seen by G. Kirk, Heraclitus and Death in Battle (Fr. 24 D), AJP 70 (1949)
384 393.
cloud. But a soul mingled with the body23 and burdened with the body,
such as a heavy and misty exhalation, is hard to kindle and to carry up-
wards.
Clements and Plutarchs reports converge to produce a comprehensive
picture. Although a soul mingled with the body is still the dry exhalation
(for otherwise it would not be a soul at all, having perished into water),
it is literally fleshy because of an admixture of the wet exhalation of the
body and therefore heavy and misty, carrying its appetites with itself,
like weights of lead and is hard to kindle and to carry upwards. To
have ones soul pure one should make it bare of flesh, that is, dry.
This can be occasionally achieved through virtuous death in battle,
when one dies without fear of dying, cut off from the body, but oth-
erwise requires extraordinary goodness which consists in ones freeing
oneself from the attachment to bodily life. For when soul is removed
and separated of the body, it is darting out of the body like lightning
out of a cloud, altogether free of fear and desire, and therefore easy to
kindle and carry upwards.
Both Clement and Plutarch refer to the soul separated from the
body and free of appetites as jahaq\, pure. Heraclitus talk about
the purity and purification is confirmed by Iamblichus De myst. 5.15
(= B 69): There are two kinds of sacrifices: those by men wholly pu-
rified (!pojejahaql]mym pamt\pasim), as sometimes happens, though
rarely, to one man, as Heraclitus says, etc. Evidently, the completely
purified is one whose soul is as pure of the wet admixture as it was be-
fore it entered the body. If this is a rare accomplishment, the rest of the
souls must be impure in varying degree. As particles of the dry exhala-
tion, souls must become attached to different strata of the soul-exhala-
tion, and the purer they are, the higher they rise. Souls that are heavy
and misty and hard to kindle and to carry upwards must be situated
in the mixed, impure region, thus remaining close to the earth. By the
same token, a soul which is completely purified of wetness, darting
out of the body like lightning out of a cloud, reaches the pure and lucid
region of the heaven. Since this is the region of the sun and the stars,
arriving there the soul returns thither whence it came, hence recovers
its divinity. Incidentally, this sheds light on Heraclitus puzzling idea
of the heavenly bodies as bowls: these are the abode of fire-like
souls which otherwise would merge with the heavenly fire. If the size
23 s~lati pevuql]mg. v}qy = mix something dry with something wet, mostly
with a sense of mixing so as to spoil or defile (LSJ, s.v.).
of the sun is just as we see it (the width of the human foot; B 3),24 the
same must be true of the stars, which makes it plausible to suggest that
each star is a soul-da_lym.
The conclusion that the pure soul recovers its divinity finds support
in B 63 5mha d( 1|mti 1pam_stashai ja v}kajar c_meshai 1ceqt f~mtym
ja mejq_m. Before him [sc. the god] there they rise and become guard-
ians of those wakefully living and those dead. Hippolytus quotes the
fragment to prove Heraclitus belief in resurrection, and there is little
doubt indeed that Heraclitus speaks here of the souls of the dead
which awake and become tutelary deities. Such deities would nor-
mally be called da_lomer, and in fact Heraclitus, as has been repeatedly
observed,25 draws on Hesiod Op. 121 123: Since the earth covered
up this [sc. Golden] race, they have been holy da_lomer guardians
of mortal men.
It emerges that Heraclitus teaching involved an eschatological con-
ception similar to that of the Orphic-Bacchic gold leaves, burial tablets
dating from ca. 400 BC on, discovered throughout the Greek world
(West Sicily, Magna Graecia, Crete, Achaea, Elis, Thessaly, Macedonia).
The doctrine that appears from these inscriptions can be summarized as
follows: the soul is of divine origin; having kept pure during life on
earth, it breaks away from the grievous circle of reincarnations26 and,
facing in Hades the decision of Persephone, pleads for deification,
which is granted: You will be god instead of mortal.27 Similarly in
Empedocles B 112.4: I, an immortal god in your eyes, no longer mor-
tal. Given the divine origin of the soul, deificationin the gold leaves
and in Empedocles alikeis the souls homecoming.
Heraclitus doctrine shows the same pattern of belief the divine
origin of the soul, the circle of births, pure life as the agent of salvation,
and the final return and deification. To recall Sextus paraphrase of Her-
24 In the Derveni papyrus (col. 4.7 9) B 3 and B 94 are quoted as a continuous

piece. This suggests that the two sayings probably belonged together, in which
case B 3 is to be understood literally (cf. D. Sider, Heraclitus in the Derveni
papyrus, in A. Laks, and G. W. Most (eds.), Studies on the Derveni Papyrus (Ox-
ford 1997) 139 40).
25 The first time by J. Bernays, Gesammelte Abhandlungen (Berlin 1885) 1.324 25.
26 The gold plate from Thurii (from Timpone piccolo, 488.5 F Bernab): I flew
away from a deeply grievous and painful circle (j}jkou d( 1n]ptam baqupemh]or
!qcak]oio; cf. Empedocles B 115.8: !qcak]ar bi|toio letakk\ssomta jeke}-
hour, changing painful paths of life).
27 487.4 F; cf. 491.4 F Bernab.
aclitus words (Pyrrh. hyp. 3.230): When we live, our souls are dead
and buried in us, while when we die our souls return to life and
live; and since the body is the defiling tomb of the soul, when a
soul quits it and rises to life, the dead body is more fit to be thrown
out than dung (B 96). The inscription on the fifth-century Olbia
bone plate (463 T, Bernab) provides a concise parallel: Life, death,
life. Truth. Dio(nysus). Orphics.28
Given the state of our evidence, it seems wise not to attempt a de-
tailed comparison of Pythagoras eschatological ideas with those of Her-
aclitus. The typological coincidences between Heraclitus eschatology
and the Orphic doctrine as represented in the gold leaves prove thor-
ough and systematic; it therefore seems justified to infer a similar rela-
tion between Heraclitus and Pythagoras teachings. But although sim-
ilar typologically, these doctrines of course differ in the way they
moulded the shared eschatological beliefs. These differences may seem
secondary to us, but they were undoubtedly of paramount significance
for the proponents of these doctrines. Heraclitus deemed Pythagoras
the head of cheats (B 81) and threatened the initiates and bac-
chantsapparently the partisans of beliefs similar to those attested in
the gold leaveswith punishment for their impiety (B 14; cf. B 15).
Diogenes 9.6 (= DK A 1) tells us that Heraclitus book won such
great fame that devotees of his arose, called Heracliteans. There is no
suggestion in the report that the recognition was posthumous;29 in all like-
lihood, Heraclitus, like Pythagoras, was a renowned teacher of wisdom.
Heraclitus condemned Pythagoras as the head of cheats (B 81) and
called his wisdom artful knavery (B 129). This clearly distinguishes Py-
thagoras from other victims of Heraclitus abuse whom he censured as
silly, ignorant, and unworthy of their high reputation, but never as dis-
honest (A 22; B 40, B 42, B 56, B 57). As Pythagoras was famous for
28 OQVIJOI was read by A. Rusyaeva, Orfism i kult Dionisa v Olvii, VDI

(1978, no. 1) 87 88, and, on a re-examination of the inscription, by J. Vino-
gradov, Zur sachlichen und geschichtlichen Deutung der Orphiker-Plttchen
von Olbia, in P. Borgeaud (ed.), Orphisme et Orphe: en lhonneur de Jean Rud-
hardt (Geneva 1991) 81.
29 tosa}tgm d d|nam 5swe t s}ccqalla, r ja aRqetistr !p( aqtoO cem]shai
tor jkgh]mtar Jqajkeite_our. The sense founder of a philosophical school of
aQqetist^r, which LSJ suggests here, is otherwise unattested (the reference to
Vita Philonidis is a printing error corrected in the Revised Supplement) and evi-
dently is admitted to suit the presumption that Heraclitus won no recognition
in his life-time.
his doctrine of metempsychosiscf. Xenophanes B 7Heraclitus ani-

mus would suggest that eschatological ideas were at the heart of his
own teaching. More definite evidence seems to be furnished by Heracli-
tus literary style which compares well with the antithetic language of the
Orphic bone plates from Olbia (463 465 T Bernab) and the legomena of
the Eleusinian mysteries (of which, incidentally, he was hereditary priest,
as follows from Strabo, 14.1.3).30 If Heraclitus chose this language to
communicate his entire teaching, it seems hard to avoid the conclusion
that eschatology was its hub.
Bibliography
Alderink, Larry J. Creation and Salvation in Ancient Orphism. Chico, Ca. 1981.
Baltes, Matthias. Die Todesproblematik in der griechischen Philosophie.
Gymnasium 95 (1988) 97 128.
Bernab, Alberto. Una etimologa Platnica: s_la-s/la, Philologus 139
(1995) 204 37.
Textos rficos y filosofa presocrtica. Materiales para una comparacin. Madrid
2004.
Orphicorum et Orphicis similium testimonia et fragmenta. Fasculus 1, Poetae epici
Graeci. Testimonia et fragmenta. Munich/Leipzig 2004.
Bernays, Jacob. Gesammelte Abhandlungen. Berlin 1885 (repr. Hildesheim 1971).
Betegh, Gbor. The Derveni Papyrus. Cosmology, Theology and Interpretation.
Cambridge 2004.
Bianchi, Ugo. LOrphisme a exist, in id., Selected Essays on Gnosticism, Dual-
ism and Mysteriosophy (Leiden 1978) 187 95.
Bollack, Jean, and Wismann Heinz. Hraclite ou la Sparation. Paris 1972.
Boulanger, Andr. Le salut selon lorphisme. Memorial Lagrange (Paris 1940)
69 79.
Burkert, Walter. Plotin, Plutarch und die platonisierende Interpretation von
Heraklit und Empedokles, in J. Mansfeld and L. M. de Rijk (eds.), Kepha-
laion. Festschrift de Vogel (Assen 1975) 137 46.
Bywater, Ingram. Heracliti Ephesii Reliquiae. Oxford 1877 (repr. Heraclitus of
Ephesus, Chicago 1969).
30 The similarity of Heraclitus style to liturgical formulae of the mysteries was dis-
cussed by G. Thomson, From religion to philosophy, JHS 73 (1953) 79, 83,
who labelled Heraclitus language hieratic. The Derveni papyrus appears to
support Thomsons observation: in col. 4.6 Heraclitus seems to be referred to
as speaking like one who relates a sacred discourse: fspeq Ujeka Reqok|c\ (cf.
col. 7.7). Proposed by D. Sider (above, n. 24) 135, this restoration is adopted
in R. Janko, The Derveni papyrus: an interim text, ZPE 141 (2002) 8,
and G. Betegh, The Derveni Papyrus (Cambridge 2004) 10. Alternative restora-
tions are luhok|c\, vusiok|c\, and !stqok|c\.
Casadio, Giovanni. Aduersaria orphica et orientalia. A proposito di un libro

recente sull orfismo, Orpheus 8 (1987) 381 95.
Cherniss, Harold. Aristotles Criticism of Presocratic Philosophy. Baltimore 1935.
(ed.). Plutarchs Moralia. Vol. 13.1. Cambridge, Mass. 1976.
Conche, Marcel. Hraclite. Fragments. Text tabli, traduit, comment. Paris 1986.
Darcus, Shirley M. Daimon as a force in shaping ethos in Heraclitus. Phoenix,
27 (1974) 390 407.
Gruppe, Otto. Die griechischen Culte und Mythen in ihren Beziehungen zu den ori-
entalischen Religionen. Leipzig 1887 (repr. Hildesheim 1973).
Janko, Richard. The Derveni Papyrus: an interim text, ZPE 141 (2002) 1 62.
Kirk, Geoffrey S. Heraclitus and death in battle (fr. 24 D). AJP 70 (1949)
384 93.
Nussbaum, Martha C. Xuw^ in Heraclitus, Phronesis 17 (1972) 1 16, 153 170.
Osborne, Catherine. Rethinking Early Greek Philosophy. Ithaca, N. Y. 1987.
Parker, Robert. The Early Orphism, in A. Powell (ed.), The Greek World.
(London 1995) 483 510.
Polito, Roberto. The Sceptical Road. Aenesidemus Appropriation of Heraclitus. Lei-
den 2004.
Powell, John E. A Lexicon to Herodotus (2nd ed.). Cambridge 1938 (repr. Hil-
desheim 1977).
Rohde, Erwin. Psyche: The Cult of Souls and Belief in Immortality among the
Greeks. London 1925 (repr. Eugene, Ore., 2006).
Robinson, Thomas M. Heraclitus. Fragments. A text and translation with a commen-
tary. Toronto 1987.
Rusyaeva, Anna S. Orfism i kult Dionisa v Olvii, Vestnik Drevney Istorii,
(1978, no. 1) 87 104.
Sider, David. Heraclitus in the Derveni Papyrus, in A. Laks, and G. W. Most
(eds.), Studies on the Derveni Papyrus (Oxford 1997) 129 48.
Tannery, Paul. Pour lhistoire de la science hellne. Paris 1887.
Teichmller, Gustav. Neue Studien zur Geschichte der Begriffe. I. Herakleitos.
Gotha 1876 (repr. Hildesheim 1965).
Thomson, George. From religion to philosophy, JHS 73 (1953) 77 83.
Vinogradov, Jurij G. Zur sachlichen und geschichtlichen Deutung der Or-
phiker-Plttchen von Olbia, in P. Borgeaud (ed.), Orphisme et Orphe:
en lhonneur de Jean Rudhardt (Geneva 1991) 77 86.
The Orphic Poems. Oxford 1983.
Wilamowitz-Moellendorff, Ulrich von. Der Glaube der Hellenen. Berlin 1931.
Zeller, Eduard. Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. Edited
by W. Nestle (6th ed.), vol. 1.2, Leipzig 1920 (repr. Darmstadt 1963).
6. Early Natural Theology
The Purification of the Divine Nature
Herbert Granger
Aristotle, who provides us with one of our major sources of information

on the Presocratic philosophers, dubbed them physiologoi, natural phi-
losophers, because he took them to be devoted above all else to the
study of nature after the fashion of his own inquiry into the natural
world.1 These new intellectuals of the sixth and fifth century did spec-
ulate about the full range of natural phenomena. The earliest of them,
Thales, Anaximander, and Anaximenes of Miletus, took up a study of
astronomy, meteorology, seismology, zogony, anthropogony, and,
not least of all, cosmogony. Aristotles student Theophrastus and the
doxographical tradition he founded followed Aristotles lead in charac-
terizing the Presocratic philosophers as largely students of nature, and
this characterization has dominated their depiction throughout our his-
tory. There is, however, good reason for thinking that the Aristotelian
tradition distorts the enterprise of the Presocratic philosophers of the
sixth century. These earliest of philosophers are as much natural theo-
logians as they are natural philosophers.2 They undertook a radical
1 The references to Hecataeus are those of F. Jacoby, Die Fragmente der griechischen
Historiker, vol. 1 (Leiden 1957). The author is responsible for the translations,
although a heavy debt is owed to the translations of Heraclitus by C. H.
Kahn and to those of Xenophanes by J. H. Lesher.
2 In contrast with F. M. Cornfords enthusiasm for a connection between myth-
ology and Ionian natural philosophy, From Religion to Philosophy (London 1912),
J. Burnet, Early Greek Philosophy 4 (London 1930) 13-15, argues for the inde-
pendence of Milesian thought from theological speculation and for the sec-
ular character of the earliest Ionian philosophy. The Ionian natural philoso-
phers had purged their use of theos of its religious significance, which
always means first and foremost an object of worship, and these natural phi-
losophers used god only to indicate the eternity of their first principle, as
what is merely ageless and immortal. Burnet contends that this non-reli-
gious sense of theos already exists in Hesiod, who introduces gods no one
would worship, which are mere personifications of natural phenomena, or
even of human passions.
164 Herbert Granger
reform of the Greek conception of divinity, as well as a reformulation of

the Greek understanding of nature. The theology of the early Greek
philosophers goes hand-in-hand with their physics, since they had in-
herited a conception of divinity and nature that intertwines the two,
and thus the study of the one topic brings along the other. These
new intellectuals, however, may have undertaken their study of the nat-
ural world so that they may reach a better understanding of the divine
nature. Indeed, if they are taken to be theologians, rather than nascent
physicists, it is easier to make a case for them as the founders of phi-
losophy, as we have come to understand it.3 Those who seek the essence
of divinity have more in common with philosophers, who engage in a
W. Jaeger, The Theology of the Early Greek Philosophers (Oxford 1947), argues
against Burnet, without pressing the target of his criticism, for the serious theo-
logical enterprise of the early natural philosophers. G. Vlastos, Theology and
philosophy in early Greek thought, PQ 2 (1952) 97-123, in his criticism of
Jaeger, in effect favors Burnets thesis. The early natural philosophers, according
to Vlastos, are not philosophical theologians, since their divinity provides no
justification for cult practices of any sort. Pythagoras, however, is a genuine phil-
osophical theologian, since he established a cult for his followers, which is based
upon his religious beliefs. The natural philosophers in effect assimilate nature and
divinity by attributing the divine features of traditional religion to nature, and
their primary object is to understand nature, not to reform religion. Contrary
to Aristotles report, Vlastos argues that Anaximander never used the word theos
or theion for a description of the apeiron (n. 75). See Guthrie, HGP 1.89 n. 1 for
convincing arguments against Vlastos and in favor of Anaximanders use of theion.
J. Barnes, The Presocratic Philosophers (London 1979) 1.99, has no appreciation of
the theological dimension of the Milesians, who, he contends, pay merely lip-
service to divinity; whereas W. Burkert, Greek Religion (Cambridge, Mass. 1985)
310 311, portrays the philosophers of the sixth century as engaged in serious
theological activity. They did not act with impiety when they broke with tradi-
tional religion, but in the expectation that a better understanding of reality
would yield a better understanding of divinity.
3 K. Algra, The Beginnings of cosmology, in A. A. Long (ed.), The Cambridge
Companion to Early Greek Philosophy (Cambridge 1999) 45-65, illustrates the dif-
ficulty of classifying the natural philosophers. He has argued that the Milesians are
neither philosophical in the modern sense nor strictly scientific in either the
Baconian or Popperian rendition of science, and that they may best be appre-
ciated as protoscientists who stood upon the threshold of that part of ancient
philosophy that came to be known as physics. P. Curd, The Presocratics as
philosophers, in A. Laks and C. Louguet (eds.), Quest-ce que la philosophie prsoc-
ratique? (Villeneuve dAscq 2002) 115-138, does not do much better for the status
of the Milesians as philosophers when she defends what she regards in its fun-
damentals as the traditional interpretation of the Presocratic philosophers as sys-
tematic thinkers who engaged in rational inquiry.
6. Early Natural Theology 165
priori speculation, than those who undertake a physicalist enterprise that

intrinsically demands the backing of empirical verification. These theo-
logians may be counted as natural theologians, since they call upon no
Muse and depend on no divine revelation or sacred text. They depend
upon their own human capacities for their theological speculation,
while remaining under the influence of significant presumptions within
conventional religious belief. These natural theologians of the sixth cen-
tury divide in their thinking between those who postulate a transcen-
dent divinity beyond the range of full human comprehension, and
who, like Hesiod, place a mystery behind the cosmos, and those theo-
logians who postulate an immanent divinity accessible to human appre-
hension. In the fifth century immanent theism gains the upper hand,
and it is not until the fourth century that transcendent theism remerges
in a vigorous way in the theology of Plato, whose creator god, the
Demiurge, stands outside the cosmos it creates, just as the Forms of Pla-
tos metaphysics are beyond space and time. Aristotle follows Plato in
locating the primary deity of his theology outside nature.4 The thought
of Plato and Aristotle eventually joins that of the ancient Hebrews, and
together they issue in the divine transcendence that dominates religious
belief to this day in the West and in the Islamic East.
The early Presocratic thinkers were hardly alone in their revalua-
tion of theology. Their reform is the expression of the dissatisfaction
of just one group, among a number of dissatisfied groups in the sixth
century, with the religious tradition shaped by the Theogony of Hesiod
and the poems of Homer. The theocosmogony of the theologian Pher-
ecydes of Syros, who wrote just before the earliest authors in natural
philosophy, makes it evident that these earliest of physiologoi do not
stand alone in trying to replace the theogony of Hesiod with a new con-
ception of the relation between divinity and nature. Pherecydes lays
down a trinity of eternal deities, Zas, Chronos, and Chthoni, at the
foundation of reality (B 1). The god Zas takes on for the first time in
4 Plato and Aristotle place their major deity beyond the natural world, although
Plato applies the word theos generously and speaks of other deities within the
natural world, such as the heavenly bodies, and even the earth (Ti. 40b-c). Ar-
istotles views upon divinity are less certain across his career. In his lost De phi-
losophia he credits the stars with divinity (Cic. De nat. deor. 2.15.42, cf.
Metaph. 1074a38-b10), and in Cael. 1.9 he describes the sphere of the fixed
stars, which is in eternal circular motion, as the first and highest divinity
(279a32-b3). When he made this statement, Aristotle had not yet presumably
arrived at his idea of the transcendent unmoved mover.
166 Herbert Granger
surviving Greek literature the role of a creator of the cosmos. This god
weaves a robe upon which he embroiders the visible features of earth,
including the ocean that surrounds the earth, and he gives this robe
to his divine bride, Chthoni, as her wedding present (B 2). When
she dons the cartographical robe she assumes the name G, a common
name for the goddess Earth (B 1), and the familiar earth of our cosmos
takes form with the goddess as its divine substructure. The pseudony-
mous Orphic poets provide another important set of actors in this
broad reaction against Hesiod, and they, like Pherecydes, devise theo-
cosmogonies with their own demiurgic dimension.5 All these authors
in their reform give a prominent place to theogony or theocosmogony
in their expositions, probably because the traditional pattern Hesiod fol-
lows and strongly reaffirms locates the essence of the gods in their birth
and heritage and in their positions in the world-order. The discovery of
the Derveni papyrus scroll in 1962 makes it possible for classical scholars
to date the earliest Orphic poems within the sixth century and closer to
the time of Pherecydes.6 There is a broad discontent with the theolog-
5 For Orphism, see W. Burkert, Lore and Science in Ancient Pythagoreanism (Cam-
bridge, Mass. 1972); M. L. West, The Orphic Poems (Oxford 1983); W. K. C.
Guthrie, Orpheus and Greek Religion 2 (London 1952); id. The Greeks and their
Gods 2 (Boston 1954); I. M. Linforth, The Arts of Orpheus (Berkeley 1941).
An excellent summary is provided by R. Parker, Early Orphism, in A. Po-
well (ed.), The Greek World (London 1995) 483-510.
6 The Derveni papyrus is the remains of a commentary by an unknown com-
mentator upon an Orphic theogony, of which some nineteen verses of the
poem the commentator quotes. The date of the commentary is put at around
the last quarter of the fifth century. Parker (above, n. 5) 484 suggests that the
Derveni papyrus makes it plausible that the earliest Orphic poems may be from
the second half of the sixth century, whereas H. S. Schibli, Pherekydes of Syros
(Oxford 1990) 35-36, believes that the main doctrines of Orphism may now
be placed at least closer to the middle of the sixth century, at around the time
Pherecydes composed his book. The text of the Derveni papyrus remains un-
settled, however. The best text for the moment is T. Kouremenos, G. M.
Parssoglou, and K. Tsantsanoglou, The Derveni Papyrus (Florence 2006), for
a review of which, see R. Janko, BMCR 2006.10.29. There is a helpful collec-
tion of essays on the papyrus, edited by A. Laks and G. Most, Studies on the Der-
veni Papyrus (Oxford 1997), and R. Janko provides a significant supplement,
The Derveni Papyrus (Diagoras of Melos, Apopyrgizontes Logoi?): A new trans-
lation, CP 96 (2001) 1-32. For additional important work on the papyrus, see
West Orphic Poems (above, n. 5) 75-101 and G. Betegh, The Derveni Papyrus:
Cosmology, Theology, and Interpretation (Cambridge 2004).
ical tradition, and it draws together Pherecydes, the Orphic authors, the
Pythagoreans, and the sixth-century physiologoi. 7
In the sixth century the authority of the traditional poets was losing
its grip upon sophisticated imaginations for the reasons Hecataeus of
Miletus lays down in the proem of his Genealogiai, which is his inquiry
into the ancestry of the descendants of Heracles and Deucalion. Homer
and Hesiod had opened their great epics with an expression of their de-
pendence for their knowledge upon the Muses. Left to themselves hu-
mans know nothing and must rely upon rumor (Il. 2.486) for their
opinions about the great events in their history and about divinity. He-
cataeus challenges this poetic pessimism by opening his prose book with
a proud declaration of his epistemic autonomy. He takes credit for what
he writes, and he dismisses the stories of the Greeks because they are
many and absurd (F 1a). Hecataeus, whom Strabo identifies as a
student of Anaximander (14.1.7), finds the stories of the poets incredible
because they provide conflicting versions of the same events, but what is
more significant because the stories are in many cases just incredible in
themselves. Hecataeus finds Homers story of Heracles abduction of
Cerberus unworthy of belief (Il. 8.366 368), and he explains it away
as nothing more than Heracles capture of a snake that had come to
be called, the dog of Hades, because of its deadly bite (F 27, cf. F
25). Hecataeus does not lay down his opinions about the heroes dog-
matically, but he conjectures as it seems to me to be true, and propos-
es what he takes to be the probable truth of the matter.8
7 Some scholars think that Aristophanes depiction of Ionian science in the

Clouds provides evidence for the theological preoccupation of the natural phi-
losophers. C. H. Kahn, Anaximander and the Origins of Greek Cosmology (New
York 1960) 106-109, holds that the elaborate parallel Aristophanes develops
with the rites of initiation could not be merely an empty farce. Belief in
the old gods is no longer tenable because they have been pushed aside by
finer and truer divinities (Nu. 247-248, 367). R. Janko, The Physicist as hi-
erophant: Aristophanes, Socrates and the authorship of the Derveni Papyrus,
ZPE 118 (1997) 69-70, stresses the hierophantic qualities of the Ionian philos-
opher portrayed in the Clouds as evidence that some of the natural philosophers
presented themselves in the language of the hierophant and the magus. Janko
also maintains that P. Kingsley, Ancient Philosophy, Mystery and Magic: Empedocles
and the Pythagorean Tradition (Oxford 1995), esp. at 217-232, has forced scholars
to take seriously Empedocles as a hierophant who initiates his audience into sa-
cred mysteries.
8 Hecataeus is hardly a consistent rationalist, since he is willing to condone out-
rageous stories: e. g., a talking ram (F 17); the Ethiopian tribe of the Sji\poder,
168 Herbert Granger
The credibility of the poets was judged lacking over two independ-
ent issues. In the one case, the poetic tradition does not provide a suf-
ficiently convincing portrayal of divinity. The Presocratic philosophers
in their revaluations of divinity offer a reassessment that assigns it a
more stately position in the cosmos or beyond. In the other case, the
poetic tradition offers dismal prospects for humans after death. Homer
envisions for human souls an impoverished half-life shorn of most
human qualities, and Hesiod offers nothing better for the mass of hu-
manity. The pseudonymous Orphic poets devise theocosmogonies
that provide for the survival of human conscious personality within a
framework of metempsychosis, and under the tutelage of certain deities
the soul may achieve release from the cycle of birth and death and find
eternal bliss in its deification. Orphism and related forms of spiritualism
depend upon rites of initiation and make demands upon the individuals
behavior, which in the case of dietary restrictions come in direct conflict
with traditional cult sacrificial practices.9 The thinking of the historical
Pythagoras, whose career as a teacher of esoteric doctrines spanned the
second half of the sixth century, is more in keeping with Orphic beliefs,
from what may be gathered from the akousmata credited to him and
from the testimonial record.10 The Orphic poets, however, are infamous
who would lie upon their back and raise their feet above them so that they may
shade themselves from the sun (F 327).
9 Plato reports that the Orphics prohibited the consumption of any animal flesh,
and for sacrifices they made do with cakes of meal and grain soaked in honey
(Lg. 782c-d). The Pythagoreans were more liberal and allowed the consump-
tion of sacrificial beasts (Iamb. VP 82).
10 Our knowledge of early Orphism and the early Pythagoreans is slim, although
there is good reason to believe that they have some large-scale features in com-
mon. The followers of Orpheus, like those of Pythagoras, took a strong interest
in rituals of purification (Plato R. 364b-365a). The Orphics, like the Pythagor-
eans, held to strict dietary regulations, although they were more demanding
than the Pythagoreans, since they upheld a complete prohibition on the con-
sumption of animal flesh (Plato Lg. 782c-d, Eur. Hipp. 952-954, and Aristoph.
Ra. 1032); see Burkert L&S (above, n. 5) 180-183; they also held to some pro-
hibition on beans (Pausanias 1.37.4; cf. 8.15.3-4). But, what is especially signif-
icant, the followers of Orpheus probably advocated metempsychosis (Arist. de
An. 410b28-30, Plato Cra. 400b-c, cf. Phd. 62b, 70c, Ep. vii 335a, Lg. 870d-
e); see Burkert L&S (above, n. 5) 126, 133. The rhapsodic theogony indicates
that metempsychosis is a feature of later Orphism (F 338-9 Bernab); see also
Guthrie Orpheus (above, n. 8) 182-187; Parker (above, n. 5) 500. Burkert
(above, n. 5) 131-132 believes that the oldest sources do not indicate any dif-
ference between Orphic and early Pythagorean doctrines, although he believes
for their lurid renditions of divine behavior, which easily compete in

gruesomeness with the tales of Hesiod. Isocrates says of Orpheus that
he told stories about the gods such as no one would dare to tell
about their enemies eating of children and castrating of fathers and
fettering of mothers (Busiris 38 [A 14b]). Pherecydes also offers an im-
proved eschatology for humanity in his sponsorship of an eternal soul
that undergoes transmigration and reincarnation, and Porphyry tells us
that Pherecydes cosmos favors the souls comings and goings with
nooks and holes and caves and doors and gates through which souls
pass in their births and deaths (Antr. 31 [B 6]). Yet Pherecydes should
not be counted among the Orphic authors. His stories of the gods are
too mild for the Orphic imagination, and, in likeness with the Preso-
cratic philosophers reformulation of the divine nature, he refashions
the gods with respectable reputations.
The concern over divine dignity appears most vividly in the frag-
ments of Xenophanes of Colophon, who is a contemporary of Anaxi-
menes and Hecataeus, but who lived well into the fifth century. Despite
his belief in multiple gods, Xenophanes makes nothing of polytheism
and gives his attention to the One god among gods and men greatest
(B 23). Xenophanes develops his idea of the divine nature in accord
with moral criteria, and he complains that Homer and Hesiod credit
the gods with behavior that among men is a reproach and a disgrace:
thieving, adultery, and deceiving one another (B 11). The Greeks
had reached a level of moral discernment that did not permit the ethi-
cally astute, like Xenophanes, to find credible any longer the poets sto-
ries of the immoral behavior of the gods. Xenophanes, however, does
not reach beyond the traditional ideas of the Greeks for the fundamental
features of the divine nature. Rather, he improves upon the defects in
the Homeric and Hesiodic conception, which at its heart upholds the
goodness and power of the divine nature.11 The supreme god Zeus is
that the Orphics and Pythagoreans differed in their history and social circum-
stances. Guthrie, Greeks (above, n. 5) 311 n. 3 also does not think that Orphic
and Pythagorean religious beliefs may be separated, but he argues against the
suggestion that the followers of Orpheus were from the lower classes (326-332).
11 No full-blown arguments for the nature of divinity survive in Xenophanes re-
mains. Jaeger (above, n. 2) 49 says of Xenophanes opinions on divinity that
they are not really philosophical or based on logical proof, but arise
from an immediate sense of awe at the sublimity of the Divine. J. H. Leshers
view amounts to Jaegers: Xenophanes of Colophon: Fragments (Toronto 1992).
Xenophanes opinions upon the divine nature in B 23-26 are all dogma and
170 Herbert Granger
the best of the gods and the greatest in strength, v]qtat|r 1sti he_m
j\qtei te l]cistor (Hes. Th. 49). Even though Xenophanes writes in
verse, he calls upon no Muse for his authority. Instead, he offers a sharp-
er analysis of the nature of divinity, as he finds it in the poetic tradition,
and he formulates no proof of the greatest gods existence, which he
takes for granted, just as the Milesians do not undertake a proof of the
existence of their divine first principle.
When Xenophanes analyzes the traditional divine superiority in
goodness and in power, he finds the true nature of these features to
be often at odds with their poetic depictions. The tradition is riddled
with inconsistencies and inadequate renderings of divine properties.
Its depiction of the gods is incoherent when it maintains that the
gods are the deathless ones always existing, !ham\tym aQm
1|mtyv (Hes. Th. 21, Il. 1.290, 2.400, 3.296), and then insists upon
and recounts their birth. Aristotle reports that Xenophanes complains
that it is just as impious to speak of the birth as it is of the death of
the gods (A 12). Xenophanes theological task is to purify the idea of
the gods, and find pure words or purified words (jahaqo k|coi),
with which to honor them (B 1.14). The greatest god, as Homer
and Hesiod would agree, has exceptional power as an implication of
this greatness.12 Zeus is the most powerful of the gods (rpeqlem^r),
and many passages in Homer and Hesiod attest to his and the gods
great power (Il. 2.350, Th. 534). Zeus can make great Olympus
quake with the nod of his head, as Homer relates (Il. 1.528 530),
but Xenophanes improves upon this conception of divine power by
maintaining that the greatest god, without effort shakes all things by
the thought of his mind (B 25). Lesher takes this to be Xenophanes
outright rejection of the Homeric conception of the divine nature, as
what can accomplish great acts with little physical effort, and his replac-
ing it with a divinity that is causally efficacious without any physical ef-
all flat assertion (116), and Lesher is not sympathetic to the suggestion that
Xenophanes idea of divinity is based upon the Homeric tradition. Instead,
Xenophanes opinions arise from his new view of the nature of the divine
(113). For Xenophanes as a theologian, who very likely had metaphysical argu-
ments to back up his beliefs, see Barnes (above, n. 2) 84-94.
12 See Lesher (above, n. 11) 99 for references: e. g. Od. 5.4; Hes. Th. 548. Above
all, there is the passage at Il. 8.17-27, in which Zeus boasts of his being might-
iest of all the gods and of his ability to hang from a peak of Olympus all the
gods, together with the land and the ocean.
fort; Lesher (above, n. 11) 113. Instead, Xenophanes offers a more ad-
equate analysis of the divine power Homer deficiently articulates.
Xenophanes stresses that the attributes of divinity are those fitting
or seemly for it to possess (B 26), and these would include, besides
moral qualities, the divinitys absolute immobility or changelessness,
which may be a dimension of its absolute power, since the divinity
may accomplish what it wishes with its thought alone without any phys-
ical exertion. The Olympians are ageless, but in the traditional stories
they travel about, paying nocturnal visits, and staging epiphanies: in the
likeness of strangers from afar, assuming all sorts of shapes, they visit cit-
ies (Od. 17.485 486). Zeus, the most powerful of the gods and their
king, has his subordinates, Iris and Hermes, who bear his messages. The
centurion of the Christian Bible tells Jesus there is no call for his coming
to his unworthy home to heal his beloved servant, but say in a word,
and my servant shall be healed. For I also am a man set under authority,
having under me soldiers, and I say unto one, Go, and he goeth; and to
another, Come, and he cometh; and to my servant, Do this, and he
doeth it (Luke 7:7 8, King James Version). Kingly dignity, of the
sort proper to the mighty King of Persia, calls for the king to refrain
from running his own errands.
The divinitys immobility is one important reason Xenophanes pla-
ces it beyond nature, the domain of constant change and corruption.
The indignity of scurrying about after this or that objective is further
reason for prohibiting the divinity from busying itself with nature.
The traditional stories of the divinity that involve it intimately in the
day-to-day workings of nature foster popular depictions that are even
more degrading than those we find in Homer and Hesiod. Aristophanes
offers an instance of this indignity when he has his character Strepsiades
confess his belief that rain is just Zeus pissing through a sieve
(Nu. 373). When Xenophanes turns to nature he tries to show how
its workings on a daily basis are not those of divinity. A simple example
is his account of the rainbow, which is traditionally a messenger of Zeus
and an important omen of the divine will. Xenophanes acknowledges
the traditional religious status of the rainbow, when he identifies it as
she whom they call Iris, but he immediately purges the rainbow of
its religious significance by reducing it to mere cloud-stuff: a cloud
this too is by nature, j purple, red and greenish-yellow to behold (B
32). The rainbow is no longer an omen of the divine will, and two tes-
timonia bear witness to Xenophanes rejection of divination (A 52). Ac-
cordingly, his revaluation of divinity, which takes it beyond nature,
172 Herbert Granger
calls for a revaluation of nature as functioning in the ordinary course of

events without divine maintenance. Xenophanes, however, allows that
the greatest god shakes all things by his thought alone, and this would
suggest that the god has some causal role to play in nature.13 There is no
suggestion that the greatest god brings about the cosmos, and there is a
testimonium that affirms that Xenophanes believes in an eternal cosmos
(A 37). The eternal god may yet have some influence over the eternal
cosmos, just as Aristotles divine unmoved mover, who creates nothing
and who lies beyond space and time, is ultimately responsible for motion
within the cosmos.
The poetic tradition contains a tension in its conception of the gods
position within the surrounding natural world. The Olympians are
within nature, often capriciously disturbing the ordinary course of its
events that they have responsibility over, but they are also lodged far
from mortals upon Mount Olympus. Of Zeus, who dwells in halls
most high (Hes. Op. 8), Homer sings: He himself among the peaks
of the mountains sat triumphant, j looking upon the city of the Trojans
and the ships of the Achaeans (Il. 8.51 52). The aloofness of the
Olympians upon their calm mountain top contributes to their dignity
and anticipates imperfectly the idea of divine transcendence beyond
the hustle and bustle of nature. Yet the stress Xenophanes puts on divine
immobility might be directed just as much against the Milesians in their
new conception of the divine nature that calls for the eternal motion of
their first principle. Diogenes Laertius reports that Xenophanes is a critic
of Thales (9.18), as well as an admirer of him for his astronomical ac-
complishments (1.23), and Thales might be of the same opinion as his
Milesian successors about the mobility of divinity, especially if he iden-
tifies his first principle water with divinity (Hippol. Haer. 1.1.3).
Exceptional awareness is a traditional feature of divinity. Through
his oracle Apollo informs King Croesus of his unlimited knowledge
and its superhuman nature: I know the number of the sand and the
measures of the sea, j both the dumb I understand and the one not
speaking I hear (Hdt. 1.47.3). The eye of Zeus is all seeing and
all understanding (Hes. Op. 267), and well he knows all things
(Od. 20.75). Yet Zeus can be mistaken and fooled, as in the case of
Prometheus theft of fire, in which he hides fire away in the hollow
13 In B 34 god begot yellow honey, eQ l wkyqm 5vuse her lki, but Guthrie
HGP 1.376 takes these words to be a conventional phrase that means nothing
more than if honey had never existed.
fennel-stalk, where it escapes the detection of Zeus (Hes. Op. 52).

Thus far-seeing Zeus is hardly the perfect embodiment of omnis-
cience. Xenophanes holds of his greatest god, Whole he sees, whole
he thinks, and whole he hears (B 24),14 so that his deity is not restricted
in its epistemic enterprise by the limitations of anthropomorphic organs
for sensation and comprehension. In his efforts to purify the divine na-
ture of the defects of tradition, Xenophanes purges the gods of the
cruder anthropomorphisms of both Greeks and barbarians, and comes
through his own via negativa to a conception of the divine nature be-
yond the reach of most human predicates. His supreme god is not at
all in body like mortals or in thought (B 23). The god may be purely
cognitive powers, and thus it may provide the model for Anaxagoras
Nous, who has every understanding about everything and has the
greatest power, and who arranges the primeval motion of the primary
materials in an orderly fashion so that the cosmos comes to be (B 12).
Xenophanes has been slighted as a philosopher ever since the days of
Aristotle. Since Aristotles interest in his predecessors among the Preso-
cratic philosophers lay primarily in their natural philosophy, he concerns
himself little with the theology of Xenophanes. When Aristotle comes
to his rendition of the nature of causality, he dismisses Xenophanes and
Melissus, as being a little too rustic, r emter lijqm !cqoij|teqoi, to
claim his attention (Metaph. 986b26 27). Cherniss is even more per-
emptory in his dismissal when he describes Xenophanes as a poet
and rhapsode who found his way into the history of philosophy by
mistake.15 Yet Xenophanes, despite his composing in verse and his en-
gaging in performance-art, may be more representative of the thought
of the sixth-century philosophers than scholars have hitherto allowed.
His speculation is shaped primarily by theology, and this may be true
of most of the early philosophers. Accordingly, Xenophanes may pro-
vide a window upon the thought of those natural philosophers of
whom fewer traces of their opinions have survived. Most scholars ap-
preciate that Xenophanes is primarily occupied with theology rather
14 Diogenes Laertius 9.19 reports in his own rendition of B 24 that the embodied
god of Xenophanes also does not breathe, oqsam heoO svaiqoeid/, lgdm floiom
5wousam !mhqp\ fkom d bqm ja fkom !joeim, l lmtoi !mapme?m. Dio-
genes is also one of those who speak of this embodied god as spherical.
15 H. F. Cherniss, Characteristics and effects of Presocratic philosophy, JHI 12
(1951) 335. Barnes (above, n. 2) 84-94 is a notable exception, and he offers re-
constructions of what might have been Xenophanes arguments for divine
properties.
174 Herbert Granger
than natural philosophy, but they are unwilling to concede much in the
way of theological interests for the Milesians, Anaximander and Anax-
imenes. Barnes dismisses any theological-sounding language of the
Milesians as mere lip-service to the gods; Barnes (above, n. 2) 99.
Xenophanes, however, can be appreciated as having his antecedents
in the speculation of his predecessor Anaximander, who too should
be appreciated as placing divinity beyond nature, and who very likely
is the first to do so in a thorough way among the Greeks.16 Classical
scholars have not on the whole recognized the significant similarity in
the theological speculation of Anaximander and Xenophanes because
they have not given sufficient weight to the theological dimension in
Anaximanders thought. We have only one sure sentence-fragment of
Anaximanders book, and the rest of what we know of him comes large-
ly from Aristotle or sources dependent upon his student Theophrastus.
Nonetheless, a good case can be made for Anaximanders preoccupation
with theology and for his pious attitude towards the divine nature. And
the same may be said of his presumed student Anaximenes.
Anaximanders cosmogony has for its first principle the apeiron, the
boundless, and this word for the first principle may indicate its nature
as an unknown material substance, the indefinite, rather than its na-
ture as an infinite expanse of material. The apeiron may very well ex-
tend infinitely in space, just as it exists infinitely in time (A 15), and it
lies beyond and encloses the cosmos that arises from it. The testimonium
of Ps.-Plutarch expresses Anaximanders account of the coming-to-be
of the cosmos from the apeiron in vivid biological language:
what arose from the eternal and is germinative of hot and cold was sep-
arated off at the birth of this cosmos, and a kind of sphere of flame from this
grew around the air about the earth like bark around a tree (Strom. 2 [A 10]).
The cosmos comes to be when the apeiron, the eternal, throws off the
c|milom, a seed-like object that is germinative of hot and cold. Hot and
cold would exist in actuality only after they develop from the seed, and
thus neither the seed nor the apeiron from which the seed separates
off would actually be hot or cold. This conforms to those testimonia
that relate that the apeiron is neither water nor any other of the things
16 Pherecydes does not remove divinity altogether from the natural world, since
the goddess Chthoni serves as the divine substructure of the earth. For a com-
parison of Pherecydes and Anaximander, see H. Granger, The theologian
Pherecydes of Syros and the early days of natural philosophy, HSCP 103
(2007) 135-163.
called elements (A 9, Arist. Ph. 204b22 29), since in these earliest days
of cosmology fire, air, water, and earth had yet to be sorted out from the
properties they bear of hotness, coldness and the like.17 The verb sepa-
rate-off (!pojq_meshai), which describes the way the apeiron throws off
the seed, enhances the biological character of the passage, since sepa-
rate-off is used in medical works for the ejection of seed in conception.18
Hesiod too follows a biological and a genealogical model of explanation,
in which Earth gives birth to Heaven and the Sea, and Earth and Heaven
together give birth to the Titans. Anaximanders biological rendition of
cosmic generation is far less anthropomorphic, and more abstract than
Hesiods theogonic generation, and thus the biological language of Anax-
imander may be merely metaphorical.
The apeiron, in addition to engendering the cosmos, has an influence
on its state. It encompasses all things and all things pilots, as Aristotle
reports (Ph. 203b10 15 [A 15]), and in his testimonium Aristotle also
reports that the apeiron is divine. We find also in separate testimonia
the apeiron described as ageless (Hippol. Haer. 1.6.1 [A 11, B 2])
and deathless (Arist. Ph. 203b13 [A 15, B 3]), and Homer often
pairs these two adjectives (!c^qyr t( !h\mat|r te) to underscore the di-
vinity of deities or their instruments (e. g. Od. 5.218, Il. 2.447, 8.539,
12.323). Although these adjectives are never together in the same testi-
monium, Anaximander may have used them together, but even if he did
not, he may still have used them to mark the divinity of the apeiron and
to express in the poetic words of Homer his homage to its divinity.19 He
may use these adjectives above all to stress that the apeiron under a Ho-
meric description replaces the gods of Homer and Hesiod.20
The impersonality of Anaximanders divine apeiron is the earliest un-
disputed recorded reflection of the sixth-century movement in theolog-
ical speculation, which is especially vivid in the thought of Xenophanes,
away from divine anthropomorphism. The apeiron is not burdened with
17 For the opposites of hot, cold, and the like as quality-things, see F. M. Corn-
ford, Principium Sapientiae: A Study of the Origins of Greek Philosophical Thought
(Cambridge 1952) 160-163.
18 See H. C. Baldry, Embryological analogies in Pre-Socratic cosmogony, CQ
26 (1932) 27 with n. 8, and Kahn Anaximander (above, n. 7) 156.
19 For Anaximanders use of these adjectives, see KRS 117; Jaeger (above, n. 2)
202 n. 39.
20 Kahn Anaximander (above, n. 7) 43 suggests that when Diogenes Laertius says of
Anaximanders view, ja t lm lqg letabkkeim, t d pm !letbkgtom eWmai
(2.1), the Homeric !cqyr has become unchanging.
176 Herbert Granger
the limitations of human-like properties, and it allows no room for the

scandalous tales of the traditional polytheism, whose anthropomorphic
deities are heavily weighed down with human foibles: moral, emotional,
and even physical limitations. The divinity plays no role in the mainte-
nance of the weather or in bringing about other meteorological phe-
nomena, which, in addition to the weather, include any sort of phe-
nomena of the upper air, comets, meteors, lunar and solar eclipses.
Anaximander seems to pay particular attention to meteorological events.
Instead of divinities we find in his explanations the same mechanism of
separating-off that we found in his cosmogony:
Winds come to be when the finest vapors of the air are separated off
(!pojqimol]mym) and when by gathering together they are set in motion;
rain from the vapor that bursts forth from the earth beneath the sun, and
lightning whenever wind breaks out and cleaves the clouds (Hippol.
Haer. 1.6.7 [A 11]).
Zeus is no longer the wielder of lightning and the gatherer of rain clouds,
and the divinity for Anaximander, as in the theology of Xenophanes, is in
its transcendence beyond the ordinary workings of nature. Yet Anax-
imanders impersonal divinity has a personal dimension in that it pilots
all things, and its guidance may be what underlies the cosmic order.
The sole sentence-fragment from Anaximanders book maintains that re-
tributive justice rules over the interactions between opposing cosmic
forces: for they pay penalty and retribution to each other for their injus-
tice according to the assessment of time (B 1). Possibly, the divine apeiron
is the pilot that initiates or sustains the domination of this cosmic justice.
Hesiod and Homer would welcome a providential divinity, since they
depict Zeus as the enforcer of justice among humankind (e. g.
Od. 13.213 214; Op. 225 247). But in the case of Anaximander the
presumed providence would have nothing of the capriciousness of
the traditional gods and would be only the enforcement of a regular
and foreseeable exchange of position between the predominate and the
subordinate powers within the cosmos. Night and light must give way
to one another in an orderly pattern over the course of the day and of
the year.
There are indications of a high style of expression on Anaximanders
part, and he may adopt this style as the language appropriate for the ele-
vated topic of the divine nature and its tutelage over the cosmos it gives
rise to. The poetry of Hesiods hexameters provided him with his high
medium of piety, and Osborne contends that verse, rather than prose, is
the standard medium for the early philosophers.21 But our one sure frag-
ment from Anaximanders book is in prose, and Simplicius comment
upon the passage, or perhaps Theophrastus, that its language is rather
poetic (in Phys. 24.20 21), may signify nothing more than the com-
mentators opinion that Anaximander makes use of metaphors or images
drawn from the courtroom to describe the interaction between the cos-
mic opposites. Anaximander in all likelihood intends for his words to be
taken literally. Cosmic change is subject to the objectivity of disinterest-
ed justice, instead of the partiality of the fickle Olympians. Yet for all
that, Anaximanders prose may have possessed a high degree of artistry,
and he may borrow Homers poetic adjectives for divinity, !c^qyr and
!h\mator, for his apeiron. The authors of prose do not call upon the
Muses, who endow their minions with the power of versification.
Still these prose authors may endow their words with a certain majesty
for dealing with high topics. They may rise above the prosaic speech of
everyday life by making use of rhythmic expressions, assonance, striking
word-order, and vivid images, stylistic techniques that have their origin
in verse.22 Heraclitus brings this poetic prose to perfection. Dei-
chgrber argues convincingly for the pious character of the prose of
Anaxagoras when he writes of his principle Nous, and antiquity noted
his book for its pleasant and dignified composition (Bd]yr ja
lecakovq|myr Bqlgmeul]mom, D. L. 2.6). Deichgrber draws attention
to Anaxagoras use of solemn predication, in which there is a heap-
ing up of ponderous predicates upon the subject of Nous: unlimited,
self-controlled, alone, finest, purest (B 12).23 Heraclitus in his
pious prose does much the same when he describes his deity: The god
day-night, winter-summer, war-peace, satiety-hunger (B 67). These
grand, but conflicting, terms used to name the god also form a pair
of chiasmi in their stately progression. A chiasmus is a phrase constructed
on an abba pattern, in which the a-terms match in some fashion, as well
21 C. Osborne, Was verse the default form for Presocratic philosophy? in C.

Atherton (ed.), Form and Content in Didactic Poetry (Bari 1997) 23-35. For argu-
ments against Osbornes contention, see H. Granger, Poetry and prose: Xen-
ophanes of Colophon, TAPA 137 (2007) 403-433.
22 For more on the poetic features of the prose of the early philosophers, see J. D.
Denniston, Greek Prose Style (Oxford 1952) 2, who proposes that these authors
used their poetic prose as a medium consciously exalted to the level of an ex-
alted theme.
23 K. Deichgrber, Hymnische Elemente in der philosophischen Prosa der Vor-
sokratiker, Philologus 88 (1933) 347-361.
178 Herbert Granger
as the b-terms. In the first two pairs of terms the negative and positive
terms, winter and summer, are the inverse of the positive and neg-
ative terms, day and night, and the same arrangement holds for the
second two pairs of terms. The chiasmus has its provenance in poetry,
but it is also favored by the early authors of prose.24 When Aristotle de-
scribes Anaximanders apeiron as what encompasses all things and all
things pilots, this phrase, ja peqi]weim pamta ja p\mta jubeqmm,
is stylistically in the form of a chiasmus, in which the phrases two tran-
sitive verbs match, just as their plural objects correspond.25 This osten-
sible artistic feature might indicate that the passage records the words of
Anaximander, butwhat is also significantthe passage has a solemn
hymnal quality in its close repetition of all things; Jaeger (above,
n. 2) 30 31 with n. 43. Xenophanes makes conspicuous use of hymnal
language when he says of his supreme god, whole he sees, whole he
thinks, and whole he hears (B 24), in which the hymnal features consist
in the passages solemn, anaphoric repetition of whole and in the use
of the same basic grammar in all three phrases.26 The poetic artistry of
Anaximander may show itself after all in the words of B 1, 1n m d B
c]mes_r 1sti to?r owsi, ja tm vhoqm eQr taOta c_meshai jat t
wqe~m,27 which form a rhythmic balance between generation and de-
struction. At their core these words form an antithesis in the symmetry
of a chiasmus, in which the prepositional phrases from which things
and into these match as the a-terms, just as generation and de-
struction correspond antithetically as the b-terms.
A remark of Diogenes on the nature of Anaximanders book has
fostered the popular interpretation of his book as something less than
a finished, polished book, and as more of a handbook or an aide-mm-
oire. 28 Diogenes says of his book, t_m d !qesj|mtym aqt` pepo_gtai
24 For the prominent use of the chiasmus in early prose, see S. Lilja, On the Style of
the Earliest Greek Prose (Helsinki 1968) 133.
25 KRS 115 suggest that this phrase may form a single rhythmical unit.
26 In his examination of the hymnal qualities of the prose of Anaxagoras and Di-
ogenes of Apollonia, Deichgrber (above, n. 23) 360 cites B 24 of Xenophanes
as an example of hymnodic verse.
27 For what should count as the words of Anaximander in the context of Simpli-
cius quotation, see Kahn Anaximander (above, n. 7) 166-178.
28 An aide-mmoire could be a word-for-word rendition of what the performer re-
cites, but it might be merely notes that provide the performer with reminders as
he performs. Upon such an interpretation, Anaximanders book would be on a
par with lecture notes, which, as some scholars have thought, might just be
what his book is; i. e., nothing more than a basis for Anaximanders public lec-
jevakai~dg tm 5jhesim, which is usually translated in some such man-

ner as, He made the exposition of his resolutions in a summary fash-
ion (D. L. 2.2). The phrase, jevakai~dg tm 5jhesim, could also be
translated by the exposition in headings, and Anaximander may
have divided up his account into the topics that typically go into the
subsequent exhibitions of natural philosophy.29 Even if his book were
brief, its authors aspirations need not be any the less literary. Heraclitus
book was probably not long.
No certain fragments of Anaximenes, whom tradition treats, like
Hecataeus, as a student of Anaximander, have come down to us. Yet
tures: e. g. H. Frnkel, Early Greek Poetry and Philosophy (Oxford, 1975) 257
n. 9; H. Thesleff, Presocratic publicity, in S.-T. Teodorsson (ed.), Greek
and Latin Studies in Memory of Cajus Fabricius (Gteborg 1990) 112; C. H.
Kahn, Writing philosophy: prose and poetry from Thales to Aristotle, in
H. Yunis (ed.), Written Texts and the Rise of Literate Culture in Ancient Greece
(Cambridge 2003) 146. If Anaximanders book were lecture notes, we should
expect less polish to be put into its composition. See A. Laks, criture, prose,
et les dbuts de la philosophie grecque, Methodos 1 (2001) 131-151, for a de-
fense of the importance of literary prose for the early Greek philosophers. The-
sleff, Scientific and technical style in early Greek prose, Arctos 4 (1966) 90-
92, however, also surmises that the philosophers who wrote in prose prior to
the treatise of Anaxagoras composed in a gnomic style, which he takes to
have its culmination in the writings of Heraclitus. Thesleff allows that com-
position in pithy passages without an intrinsic connection between them could
take an artistic form. M. S. Schofield, An Essay on Anaxagoras (Cambridge 1980)
25, offers a more plausible view of the nature of Anaximanders book than The-
sleffs proposal of a gnomic style when he suggests that Anaxagoras was writing
in an Ionian tradition of philosophy whose standard mode of exposition was the
cosmogonical narrative. The events of cosmogony and its aftermath would
seem to call for a narration, in which passages are clearly linked in chronological
sequence, just as we find in Hesiods theogony, and the gnomic style would
seem thoroughly inappropriate for this. Kahn Anaximander (above, n. 7) 199
has surmised that Anaximander developed his book chronologically, as a
kind of epic poem, with a clear beginning, middle, and end, which begins
with the coming-to-be of the primal forces of the cosmos from the apeiron and
ends in the appearance of man. This narrative form for cosmogony and its after-
math is what we would expect from authors who have before them for their
precedent Hesiods chronological narration of theogony.
29 Laks (above, n. 28) takes this to be the correct interpretation of Diogenes com-
ment; see also Thesleff, Scientific and technical style (above, n. 28) 90 n. 5.
Kahn Writing (above, n. 28) 145-146 also suggests that Anaximanders book
set the pattern for the standard topics and the order of their display in the books
of the natural philosophers. See also W. A. Heidel, Hecataeus and Xeno-
phanes, AJP 64 (1943) 262.
180 Herbert Granger
we still have some indication of his writing also in a high style for pious
reasons. Anaximenes is reported to speak of his first principle, like
Anaximander, as boundless, as an apeiron, but Anaximenes apeiron is
very different from Anaximanders. It is not unknown, but is air and
thus very definite in character. As air it might be thought to be the
least perceptible of what may be observed in nature, when it is most
even, it is invisible (Hippol. Haer. 1.7.2 [A 7]), unlike earth, or fire,
or water, and thus Anaximenes might have hit upon air as something
of a definite, knowable nature, but the least discernable of what is avail-
able in observable nature, of what is less definite itself, so that it may
serve as the basis for all else. Anaximenes, however, still shares a great
deal with his alleged teacher. His apeiron is eternal, always in motion,
possibly infinite too, and it too is divine (At. 1.7.13 [A 10]). Air too
encompasses all things and perhaps pilots all things as well, as is indi-
cated by a testimonium of Atius that scholars debate as something close
to the very words of Anaximenes: Just as our soul, being air, holds us
together, so the whole cosmos breath and air encompass, oWom B xuw^ B
Blet]qa !q owsa sucjqate? Blr, ja fkom tm j|slom pmeOla ja !q
peqi]wei (B 2).30 Air gives rise to the cosmos, and the cosmos itself is air
and everything in it is nothing but air in condensed or rarefied forms (A
7).31 Anaximenes depersonalizes the natural world through his deper-
sonalization of the divinity by his making it into a natural, commonplace
stuff that goes into the constitution of every cosmic object.
Diogenes Laertius tells us that Anaximenes used simple and unre-
markable Ionic speech, j]wqgta_ te k]nei Y\di "pk0 ja !peq_tt\
(2.3), which is a judgment that may go back to Theophrastus (or to Apol-
lodorus). As Osborne stresses, this description of Anaximenes style does
not rule out his writing in a plain form of verse; Osborne (above, n. 21)
29. Diogenes words, however, would hardly lead one to suppose that
Anaximenes wrote in verse, especially since if he had done so the doxog-
raphers who were accustomed for many generations to the predominance
of prose for the philosophical medium would certainly have noted his ex-
ception, along with those of Xenophanes, Parmenides, and Empedocles.
30 Atius reports these words as if they were Anaximenes. For the dispute over
the authenticity of these words attributed to Anaximenes, see Guthrie HGP
1.131-132, who takes the passage to be largely authentic.
31 D. W. Graham offers reasons for thinking that this standard interpretation of the
monism of Anaximenes is mistaken: A Testimony of Anaximenes in Plato,
CQ 53 (2003) 327-337.
Thesleff, however, argues that the comment of Diogenes need not mean
that Anaximenes style was a dry and concentrated matter-of-factness,
but only that it did not possess the poetic fireworks of the prose of some-
one like Heraclitus; Thesleff (above, n. 28) 91 n. 6. Thesleff thinks that
possibly Anaximenes book possessed an elevation of style, and for
credible evidence he points to some of the phrasing in the testimonium
of Hippolytus, most likely an excerpt from Theophrastus, which could
make a plausible claim on being close to, if not exactly, the words of
Anaximenes.32 The pertinent part of the sentence reads: !]qa %peiqom
5vg tm !qwm eWmai, 1n ox t cim|lema ja t cecom|ta ja t 1s|lema
ja heor ja he?a c_meshai, t d koip 1j t_m to}tou !poc|mym
(Haer. 1.7.1 [A 7]). Although Thesleff makes nothing of it, other scholars
have noted the peculiarity of the word offspring (!p|comoi) for the
products of air, and have presumed that it is the word of Anaximenes,
which reflects the same biological model of explanation we found in
Anaximander and goes back to Hesiods genealogical explanations.33
The best candidates for the progeny of air, from which the rest
come to be, are the gods and divine things, which are most likely
the other elements (besides air) that go into the makeup of the cosmos,
fire, wind, cloud, water, earth, and stone, which arise most immediately
from the rarefaction and condensation of air. The list goes beyond the tra-
ditional four elements of fire, air, water, earth, but this narrowing of the
cosmic elements to these canonical four is probably due to Empedocles
who identifies them as the four roots of all things in his physics (B
6). Simplicius in his own paraphrase of what is probably the same excerpt
from Theophrastus also lists in the same order fire, wind, cloud, water,
earth, and stone as what emerge from air, and his wording makes it
clear that it is from these the other things come to be (A 5).
Thesleff finds notable the assonance and accumulation of the
opening of the sentence from Hippolytus, which is more art than
would be expected of the doxographers. But it is the peculiarity of
the accumulation that is especially striking, as the solemn invocation
of all that the first principle air is responsible for, namely, all things, in
the present, past, and future, as well as what counts as divinity, from
32 Kahn Anaximander (above, n. 7) 149-150 argues for Hippolytus testimonium as

a very close paraphrase of Anaximenes words. Barnes (above, n. 2) 1.97
finds that Hippolytus is garbled in his account and makes an absurdity of
Anaximenes views.
33 Kahn Anaximander (above, n. 7) 149-150, 156, and id. H. 137.
182 Herbert Granger
which come to be things now and things having been and things that
shall be and gods and things divine, with each component in the series
hammered forcefully by the reiterated particle ja_. The phrasing is not
that of a doxographer, but of Anaximenes who is praising his divine first
principle as the ground of all else that there is, and this phrasing can be
traced to a poetic ancestry. The phrasing of present, past, and future re-
calls closely the phrasing found in the poets for the extraordinary
breadth of the events prophets have access to. Homer uses similar phras-
ing to describe the great prophetic scope of the seer Calchas: dr dg t\
t( 1|mta t\ t( 1ss|lema pq| t( 1|mta, who knows things that are and
that shall be and that had been before (Il. 1.70). Hesiod uses exactly
the same wording as Homer when he describes the great scope of the
events the Muses have within their capacity to convey: eUqousai t\ t(
1|mta t\ t( 1ss|lema pq| t( 1|mta, relating things that are and things
that shall be and things that had been before (Th. 38, cf. 32). Anaxi-
menes may borrow this high-sounding language of poetry to express
his pious praise of the grandeur of the deity in its comprehensive re-
sponsibility for all that there is. In turn, he may be followed in his sol-
emn phrasing by subsequent natural philosophers for the same reasons.
Anaxagoras uses similar words to describe the great scope of the organ-
izing influence of Nous: ja bpo?a 5lekkev 5seshai ja bpo?a Gm ja fsa
mOm 1sti ja bpo?a 5stai, p\mta diej|slgse moOr, And whatever things
were to be both whatever were and as many as are now and whatever
shall be, all things mind ordered (B 12).34 Empedocles does the same
when he describes the great scope of what comes to be from his divine
elemental substances: 1j to}tyv cq p\mh( fsa t( Gm fsa t( 1sti ja
5stai, For from these all things as many as were and as many as are
and shall be (B 21.9).35 The language of Anaximenes may provide
the precedent for the pious phrasing of these two fifth-century natural
philosophers for their description of the great range of the productive
and organizing power of their first principles.
Anaximander proposes an eternal, impersonal divinity for the agent
that gives rise to our cosmos. This divinity remains largely mysterious,
and one of the few things Anaximander allows that we may know
about it is its state of eternal motion (A 9), which he may base upon
34 Simplicius quotes this line from the text of B 12 in four different ways. See
Schofield (above, n. 28) 145 n. 1 for the above version of the line.
35 Heraclitus uses a similar pattern when he speaks of the eternity of the cosmos as
fire ever living, Gm !e ja 5stim ja 5stai pOq !e_fyom (B 30).
its being alive. Our empirical knowledge is restricted to what emerges

from the divine apeiron and cannot include in its scope the apeiron itself.
The cosmos is the product of the opposites or the elemental substances,
of hot, cold, moist, dry, or of fire, air, water, earth (Arist. Ph. 204b26
28), which the apeiron throws off in its generation of the cosmos. What
we know through experience is restricted to the opposites or the ele-
mental substances and the sorts of things they give rise to through
their combination with one another in various ways. Although the apeir-
on throws off these opposites or elemental substances, it is not any one of
them, or even strictly a mixture of them. Instead, it may be best under-
stood as some sort of fusion of the elements, in which their contending
powers for heating and cooling, drying and moistening, are held in
abeyance, so that the apeiron may remain stable and not fly asunder, in
contrast with cosmic objects that in their volatility are subject to
decay and destruction.36 The apeiron is not within the range of what
we may know through perception, and, since our cosmos is made up
of these sorts of perceptible things, the apeiron transcends our perceptible
world. It remains, however, physical, in spite of its supercosmic status,
because it possesses the physical properties of spatiality and mobility.
Despite his strong disagreement with Hesiod, Anaximander has a
strong affinity with him when he lays down, like Hesiod, a mystery be-
hind the cosmos. In his Theogony Hesiod makes no effort to explain from
what the first deities of his theogony emerge (116 120), and this gap in
his explanatory scheme is glaring. Pherecydes may be thought to offer an
improvement upon Hesiods genealogical explanation when he introdu-
ces his eternal trinity at the foundation of reality; e. g., KRS 71. Hesiod,
however, may have a defense for his remaining silent upon the antece-
dents of Chaos, who comes to be first of all the gods. Hesiod might
plausibly argue that these antecedent conditions lie even beyond the pur-
view of the gods and the Muses. Chaos comes into existence, unlike any
of the other gods, without any epithet. The location of Chaos within
Hesiods conception of the world-order is far from certain, and to a
large extent scholars call upon the etymology of w\or for their supposition
that this first deity is a gap or chasm of some sort within the world-
order (Th. 700).37 At a late stage in his poem Hesiod holds that the de-
feated Titans have been banished to a place across dark Chaos, p]qgm
36 For the apeiron as a fusion of the opposites as quality-things, see Cornford

Principium (above, n. 17) 161-162.
37 E.g. Jaeger (above, n. 2) 13; Cornford (above, n. 17) 194 and n. 1.
184 Herbert Granger
W\eor foveqo?o (814), and Chaos without any mate gives birth to the first
divine offspring, Erebus and black Night, who are male and female
twin versions of darkness (123).38 The darkness of Chaos and its engen-
dering forms of darkness may be meant to indicate that Chaos comes
from circumstances about which nothing can be known.
The apeiron, from which our cosmos emerges, remains largely beyond
the reach of human comprehension. Anaximanders book probably did not
dwell much on the divine apeiron, beyond what the testimonial record re-
lates, since little can be known about it. Xenophanes may be allied in his
thought with Anaximander, since he proposes a transcendent deity who
lies beyond what humans may know: and of course the clear and certain
truth no man has seen nor will there be anyone who knows about the gods
and what I say about all things (B 34, tr. Lesher). Both men posit a mys-
terious divinity behind the workings of the visible cosmos, and the verse of
Hesiod may provide a precedent for their position when it depicts the
emergence of the world-order from unknown conditions.
The views of Anaximander and Xenophanes are in sharp contrast
with those of Anaximanders alleged student Anaximenes. Anaximenes
characterizes his own apeiron as divine, just as Anaximander does, but
Anaximenes makes his divine apeiron observable air. He lays down an im-
manent, knowable divinity, whereas Anaximander and Xenophanes pro-
pose a transcendent, unknowable divinity. Heraclitus of Ephesus, who is a
younger contemporary of Anaximenes and Xenophanes, is closer in his
theological speculation to Anaximenes. Although Heraclitus is no panthe-
ist, as Anaximenes might be plausibly construed, he does posit a divinity
that is common to everything in the cosmos, which is the whole of reality
for Heraclitus, but this divine commonality is comparable to the way the
laws of a city are common to its citizenry (B 114).
Most scholars have little problem in conceding the significance of
Heraclitus theological concerns.39 Any rendition of Heraclitus divinity
must put at its center his comments on the logos, of which B 1 provides
his opening words on the topic, the first sentence of which reads: Al-
though this Word is always men are uncomprehending, both before
hearing it and once they have heard. The nature of this Word has
provided a topic of controversy, and Barnes and West take logos to sig-
38 Although Erebus is a neuter noun, the deity plays a masculine role as the consort
of Night.
39 Barnes (above, n. 2) 99 willingly acknowledges Heraclitus devotion to divin-
ity.
nify nothing more than Heraclitus book.40 Heraclitus probably quib-

bles on logos as his book and as the Word by which all things come
to pass (B 1), which makes up the central content of his book. Long
ago Aristotle complains about the difficulty of reading Heraclitus,
which he traces to his disregard for proper punctuation, and Aristotle
appeals to the first sentence of B 1 for his illustration of syntactical am-
biguity (Rh. 1407b11 18). The lack of punctuation makes it impossible
to determine to which clause always belongs: the Word is always or
always men are uncomprehending. The always should be read with
both clauses, and therefore the sentence demands a double-reading.41
Ambiguity is a much-recognized feature of Heraclitus style, which
makes him hard to read, but he does not write obscurely for its own
sake. His obscure style is probably in imitation of the Pythian Apollo:
The Lord whose oracle is at Delphi neither declares nor hides, but
gives a sign (B 93). Ambiguity is the medium of prophecy, certainly
of Apolline communication, whether prophetic or not.42 Heraclitus ap-
propriates the ambiguity of the Pythian Apollo for his own medium be-
cause there is a significant parallel between the nature of things in the
form of the logos and the Apolline medium.
The logos, by which all things come to be, is the basis of human wis-
dom: Listening not to me, but to the logos, it is wise to agree that all
things are one (B 50). The unity of all things is something Anaximenes
would readily affirm, since in his monism all things are just variations on
air, just cooler or hotter forms of air, which amount to nothing more
than denser and rarer forms of air. Heraclitus does sound like a monist
when he says that out of all things one, and out of one all things (B
10), but the case for his monism is not strong, and it may be largely the
product of Aristotle and the doxography.43 B 30 provides the best tex-
40 M. L. West, Early Greek Philosophy and the Orient (Oxford 1971) 114-115;
Barnes (above, n. 2) 59.
41 Kahn, H. 92-95 argues convincingly for reading the always of the first sen-
tence of B1 with both clauses.
42 C. Sourvinou-Inwood, Delphic Oracle, OCD 3 445-446, suggests that the
ambiguity of Apolline oracles is not due to the Apolline prophets trying to
hedge their bets, but it is the acknowledgment of the ambiguity inherent
in the gods prophecy and of human fallibility in apprehending the future.
For the ambiguity of the Greek deities in communication and in nature, see
M. Detienne, Masters of Truth in Archaic Greece (New York 1996) 84-86.
43 Barnes (above, n. 2) 63 adopts Aristotles monistic interpretation of Heraclitus,
although he admits that the evidence is thin. Among others, D. W. Graham,
186 Herbert Granger
tual evidence for Heraclitus holding a monism like Anaximenes, in

which, instead of air, fire provides the stuff of all else: this cosmos
was always and is and shall be fire ever living. Fire plays a central
role in Heraclitus cosmos, but the mercantile imagery of B 90 bars it
from playing a monistic role: For fire all things exchange and fire for
all things, just as for gold goods and for goods gold. Fire stands to all
things just as gold in its role as a monetary unit stands to the merchandise
in the marketplace. A monistic comparison between fire and gold
should treat gold as the material that goes into the makeup of such ob-
jects as gold coins. Whatever the exact role of fire may be in the cosmos,
it is closely associated with the logos in that fire is connected with the
guiding-principle of the cosmos: Thunderbolt steers (oQaj_fei) all
things (B 64).
The unity Heraclitus has in mind is not the materialist monism of
Anaximenes, but the paradoxical unity of opposites. This hidden truth
of the unity of opposites contradicts common sense and the teachings
of the wise, but humans have supposed some things unjust and others
just (B 102). Within the essence of each thing a strife holds between
opposing powers, without which the cosmos and everything in it could
not retain their integrity, and this strife at the center of reality further
offends conventionality because it is beneficent and just (B 80). The op-
posing powers within the essence of each thing coperate by issuing in a
unified object through their striving with one another: They do not
comprehend how each thing quarrelling with itself agrees; it is a con-
nection turning back on itself, like that of the bow and the lyre (B
51; cf. B 54). There could be no bow or lyre without the tension
that arises from the powers of their wood and string for pulling in op-
posite directions. Thus the essence of things will display opposing prop-
erties just as Apolline speech will display multiple meanings. Sea water
provides an example of the unity of opposites through its exemplifica-
tion of the opposing properties of purity and impurity, for fish it is
drinkable and life-sustaining, for men it is undrinkable and deadly (B
61).
There are many other logoi, and Heraclitus knows of them: Of all
those whose logoi I have heard, no one reaches to the point so that he
may recognize that the wise from all is set apart (B 108). But there
is just the one logos that should be obeyed, and It is law too to obey
Heraclitus criticism of Ionian philosophy, OSAP 15 (1997) 1-50, argues

convincingly against Aristotles interpretation.
the counsel of one (B 33). The logos is common, vulgar, omnipresent,

and it shows itself, or hints at itself in everything. Its omnipresence is
after the fashion in which the law of a city is common over the
whole of the law-abiding, even the lawless, citizenry:
Those who speak with understanding must rely firmly on what is common
to all as a city must rely on law, and much more firmly. For all human laws
are nourished by one, the divine; for it prevails as it will and suffices for all
and is more than enough (B 114).
The logos is shared by all in the way a community practices or exercises a
custom or a m|lor. Even the lawless abide by the law in their evasion of
the law, which calls upon them to act in the way that allows them to
elude the law. Evasion, however, is impossible in the case of the logos:
All things fire coming suddenly upon shall judge and shall overtake,
p\mta t pOq 1pekhm jqime? ja jatak^xetai (B 66). It is perhaps not
too misleading to say that the divinity is present in the way the law of in-
ertia is present across the universe. The divinity as law-like in nature or as
itself a law is an original contribution to theology. The law-like nature of
the divinity brings out a new way of conceiving of the divine omnipre-
sence or divine immanence, in place of the monism of Anaximenes, and
it also brings out something of the power of the divinity over all things, as
the ruling-principle within the nature of things.
Heraclitus would seem to believe that the divinity in its law-like na-
ture just is the unification of opposites as this unity manifests itself in
each and every thing and as what is responsible for the order within
the cosmos and its eternal maintenance. In B 67 he brings together
the divinity, the unity-in-opposition, and very likely fire:
The god is day-night, winter-summer, war-peace, satiety-hunger. It alters, as
when mingled with perfumes, it is named according to the pleasure of each.
Heraclitus extols his god, and the grand names he gives it pass in chias-
mic procession from the cosmic to the human level. In his praise Her-
aclitus brings out the universal, common scope of god, in the broad
range of gods permutations, of the multiple ways the unity of opposites
manifests itself, through all the different perfumes. Plausibly, it is the
divine fire that embodies these contrasting perfumes through its al-
terations, presumably through its embodiment in law-like fashion in
everything that there is.
Xenophanes major contributions to theology lie in his introduction
of the immobility of divinity and in his replacement of the anthropo-
morphic model with, or its constriction to, a cognitive model. The di-
188 Herbert Granger
vinity is a purely cognitive agent, perceiving, thinking, and knowing,

perhaps a pure mind, which can bring about change through mere
thought alone, and which, therefore, provides the embodiment of the
greatest conception of pure power, in which the deed may coincide
with its thought. Heraclitus may be thought to follow Xenophanes in
his rejection of anthropomorphic cult: These images they pray to as
if one were chatting with houses (B 5). Heraclitus movement towards
the divine nature as law-like in character conforms with the movement
of the Milesian physiologoi and of Xenophanes away from a divinity cast
in personal terms. Heraclitus, however, escapes the scepticism that
threatens Xenophanes theology (B 34.1 2). In its ubiquity the deity
of Heraclitus is immanent and ready-to-hand, in everything, even in
the knower himself.44
Fire is psychic stuff (B 76), alive, fire ever living (B 30), and what
governs, steers or pilots all things (B 64). Is it providential in its steering all
things in such a way that it looks after the welfare of humanity? Does
thunderbolt steer everything towards the good? In typical Heraclitean
fashion the answer is both affirmative and negative: To god all things are
beautiful and good and just, but humans have supposed some things un-
just and others just (B 102). The good is perhaps not what can be fully
appreciated by mortals as good. The divine fire is cosmic intelligence if
we read the corrupted Greek of B 41 so that the cm~lg is the thought
by which all is piloted through all: 4m t sov|m 1p_stashai cm~lgm
bt], jubeqmtai p\mta di p\mtym.45 Heraclitus, and maybe Xeno-
phanes, may be taken to build upon Anaximanders idea of divine stew-
ardship. The retributive principle of justice that keeps in balance the op-
posing forces of the cosmos operates for the overall good of the cosmos
without any attention given to the fleeting goods of humanity.
Heraclitus is much more severe than Xenophanes in his criticism of
Greek religious practices. As any reader of Sophocles Antigone may con-
44 See Kahn, H. 267, who affirms in his comments on B 5 the immanence of

Heraclitus deity. Marcovich, H. 438, however, takes B 108 to be the expres-
sion of divine transcendence, the wise from all set apart. Marcovichs trans-
lation of jewyqislmom as different from would suggest something weaker
than divine transcendence.
45 Burkert Gk. Rel. (above, n. 2) 309 seems to read B 41 in this fashion. This ren-
dition of B 41 adopts the reading of Vlastos, On Heraclitus,AJP 76 (1955)
352-353 with n. 33, against that of Kirk, H. 386-390. Guthrie, HGP 1.428-
429 with n. 1 supports Vlastos rendition of B 41. See Marcovich, H. 449
for a thorough review of the text of B 41.
clude, Heraclitus outrages traditional reverence for the dead when he says
of corpses, in three words only, that they should be disposed of more
quickly than dung (B 96), where he picks the crudest word available
for dung (jopq_om). This is an insult and looks like a challenge to tradi-
tional piety, since it runs so counter to it: the funeral rites owed to the
dead are owed to the gods (Ant. 519, 1070 1076), and Creon and Anti-
gone, who in her own words dares a righteous crime (fsia pamouqc^-
sasa, 74), are contending over what is holy, eqac^r (521). The soul of
Patroclus appears in a dream to Achilles and urgently implores him to
bury his corpse quickly so that his agitated soul may pass the gates of
Hades and find repose (Il. 23.71). Heraclitus insult, however, offends
also the new beliefs of the Orphics, whose grave sites in southern Italy
and elsewhere indicate that the followers of Orpheus engaged in distinc-
tive burial practices.46 Xenophanes has nothing to match these harsh
words of Heraclitus. At most he reduces the goddess Iris to the stuff of
colored clouds. Yet what does Heraclitus criticism come to? Dung is
waste, offal, but also manure for new life. Is this insult less of an insult,
or not one at all? No matter what the event, it remains provoking.
Heraclitus might be thought to bring out the absurd features in tra-
ditional religious practice: They purify themselves in vain with blood
when defiled with it, as if a man who had stepped into mud were to
wash it off with mud. He would be thought mad if any man noticed
him acting thus (B 5).47 His pointing out the comparability of the ritual
practice of cleansing oneself of blood-guilt with a blood-sacrifice and of
the mad act of cleansing oneself of mud with mud looks like a reduction
to absurdity through an analogy. Blood, however, has opposing proper-
ties, polluting and purifying, which comes out neatly in Heraclitus
Greek by the double use of the single occasion of blood by the
two verbal forms flanking it: jaha_qomtai aVlati liaim|lemoi. Since
Heraclitus approves of the conjunction of opposing properties, one is
at a loss how to assess definitively the import of this passage. Yet he
does not approve of this bloody ritual, since he describes it as exercised
in vain (%kkyr).48
46 Parker (above, n. 5) 485, 496-498. Herodotus 2.81 reports that those who en-
gage in the rites of Orpheus cannot be buried in woolen garments.
47 This rendering of B5 is based on the proposal of Marcovich, H. 459.
48 C. Osborne, Heraclitus and the rites of established religion, in A. B. Lloyd
(ed.), What is a God? (London 1997) 40 n. 2, retains the %kkyr of the manu-
script, instead of the emendation of %kk\, which Kranz adopts, and she takes
190 Herbert Granger
Also B 5 might reconcile opposites in another way, since the puri-

fication ritual would appear to be madness, like cleansing oneself of mud
with mud, if it were not for its religious purpose. The first sentence of B
15 reconciles opposites more explicitly in this sort of way: what is
shameful if done without thought of Dionysus, the god of sexual
vigor and exuberance, is not shameful when done in his worship
when his devotees march in procession and chant the hymn to the
phallus. The word-play of B 15 is impressive, where Heraclitus estab-
lishes verbal associations between genitals, shame, most shame-
less, and Hades: aQdo?om, aQd~r, !maid]stata, and )_dgr. For Hera-
clitus these linguistic connections indicate a real connection between
the life-giving genitals and death, and when he identifies dried-up
Hades and juicy Dionysus he is indicating the unity life and death con-
stitute. The Orphic poems are filled with the assimilations of traditional
gods. But the truth they may bear is distorted, and the mysteries
(lust^qia) practiced among men they get themselves initiated into in
an impious way (B 14).49 These men are unaware that when they fol-
low the current practices in their initiations into the secret rites, their
pursuit of life in Dionysus in sex and wine is their pursuit of death in
Hades, since it is death for souls to become wet (B 77).
The unwholesome nature of the current practices of the mysteries is
that they lead their participants away from the truth of the logos and
bring about a dulling of the intellect. Falsehood dulls the mind. And
the sleepy-headedness of wine and of sex brings about a diminution
of the intellect and leads the initiate into sleep and intellectual death.
We should awake from the slumber of folly and become wakeful
watchers of the living and the dead (B 63). Scholars have often
taken these wakeful watchers to indicate Heraclitus belief in the res-
urrection of the dead, or at least of selected dead.50 The phrasing of the
fragment is reminiscent of the language of Hesiod when he describes
the task of the race of the golden men after their death, as immortal
this text to support her translation in a different way. The reversion to the
manuscript reading of %kkyr does not rule out the translation of in vain.
49 For the text for this rendering of Heraclitus B 14, see T. M. Robinson, Hera-
clitus: Fragments (Toronto 1987) 85-86 and Marcovich, H. 464. For the view
that Heraclitus has some regard for the mysteries, see Guthrie, HGP 1.476,
KRS 209-210, R. Seaford, Reciprocity and Ritual: Homer and Tragedy in the De-
veloping City-State (Oxford 1994) 226-228, 321-322.
50 Kahn, H. 254-256 is prominent among those who take B 63 to be evidence of
Heraclitus belief in the resurrection of the souls of some selected dead.
watchers of mortal men (Op. 253). Yet a better case can be made for
these wakeful watchers as the wise who have awakened from the stu-
por of their sleepy ignorance, since Heraclitus equates ignorance with
the subjectivity of sleep (B 89).51
Both Xenophanes and Heraclitus are possibly polytheists. This is
perhaps more certain in the case of Xenophanes (B 23.1), but Heraclitus
allows himself to speak of gods now and again (B 5, B 24). Neither
theologian abandons the tradition altogether. Xenophanes may be ap-
preciated as in effect affirming the basic presumptions of tradition
when he undertakes his purification of traditional religious language
under the direction of his more astute estimation of what is entailed
by the traditional belief in the goodness and strength of the greatest
deity (Hes. Th. 49). But Xenophanes does not make any apparent use
of the old names or any of the traditional epithets for the gods. His
greatest god remains unnamed, and Xenophanes remains more abstract
in his discourse on the divine nature, on gods and the greatest god. Her-
aclitus does not abandon the names or the epithets, but he uses them
unconventionally or even paradoxically. He follows, but deviates in im-
portant ways from, traditional religious thought. Below are a few nota-
ble examples of Heraclitus unconventional use of conventional reli-
gious language, stories, and conceptions.
(1) Heraclitus designates Apollo indirectly as the lord whose oracle is
in Delphi (B 93), which is a description and not a traditional
name or epithet for the god. Heraclitus mimics the obliquity of
the Pythian Apollo himself, who gives a sign, and scholars gen-
erally take his analysis of Apolline speech to be an indirect analysis
of his speech and the logos that makes up the heart of reality.52
(2) Artemis is the twin sister of Apollo, and Diogenes reports that
Heraclitus deposited his book at her great temple in Ephesus
(9.6). His dedication of his book to the goddess may amount to
its publication and to his making his thoughts available to the pub-
lic, rather than, as some scholarse. g., Kahn, H. 2have pro-
posed, his secreting them away from the vulgar. Heraclitus
51 For a detailed defense of this sort of interpretation of the wakeful watchers of

B63, see H. Granger, Deaths other kingdom: Heraclitus on the life of the
foolish and the wise, CP 95 (2000) 260-281.
52 E.g., Kahn, H. 123-124. J. Barnes, Aphorism and argument, in K. Robb
(ed.), Language and Thought in Early Greek Philosophy (La Salle 1983) 101, is
one of the few to question this common interpretation of B 93.
192 Herbert Granger
would not shun publicity, since he firmly believes that the truth of
the logos is common and open to anyone, and not a private posses-
sion of the privileged few. The twin children of Leto may have
furnished Heraclitus with the two sides of his display of the
truth: his medium of indirect expression and his public venue
for his publication.
(3) One, the wise alone, is not willing and is willing to be called by
the name of Zeus (B 32). The genitive Fgm|r may be used to al-
lude to the infinitive f/v to live,53 but at any rate this passage
brings sharply to the fore Heraclitus ambivalence towards tradi-
tion. He is willing to appropriate for his first principle a traditional
name of the supreme god of Homer and Hesiod, but he is not will-
ing to adopt it without qualification. There is no impiety in its use
and none in its neglect. The wise alone tolerates the use of Zeus
by mortals for its name, since the Zeus of tradition is the wisest and
the most important of the gods, but, like Xenophanes unnamed
deity, the wise in itself has no proper name.
(4) The only other surviving use of Zeus among Heraclitus frag-
ments is found in B 120 with an epithetical-like attribute: The
limits of dawn and evening is the Bear, and, opposite the Bear,
the watcher of luminous Zeus, AoOr ja 2sp]qar t]qlata B
%qjtor ja !mt_om t/r %qjtou owqor aQhq_ou Di|r. The star Arctu-
rus is the Bear-watcher or warder, b )qjtoOqor, and the Bear
Arcturus watches is the constellation Ursa Major, a mark for the
celestial pole-star. The time of rising of Arcturus is the middle of
September, and in Hesiod is a signal for fall and spring (Op. 566,
610); Kahn, H. 161 163. Zeus has a traditional association with
fire, as the wielder of lightning bolts, and as the hoarder of fire,
who will not share it with mankind. Hesiod recounts the theft
of Prometheus and describes the fire he steals as unwearying,
puqr l]mor !jal\toio (Th. 563), and this lack of fatigue might
be a feature of fire that Heraclitus found attractive in his efforts
to understand the limitless power of divinity.
(5) Heraclitus stresses the importance of fire as a guiding-principle in
Thunderbolt (jeqaum|r) steers all things (B 64), and the thunder-
53 This is a common interpretation of Heraclitus use of the genitive Fgm|r in B

32; e. g., Kahn, H. 268 271; Guthrie, HGP 1.463; Robinson (above, n. 49)
102.
bolt is famously the primary weapon of Zeus who delights in

thunder, teqpij]qaumor, and who thunders aloft, rxibqel]tgr.
(6) The Sun is often called on in oaths, as what sees and hears all
things: I]ki|r h(, dr p\mt( 1voqr ja p\mt( 1pajo}eir
(Il. 3.277). The cosmic fire of Heraclitus is like the traditional
Sun in its vigilance: All things fire coming suddenly upon shall
judge and shall overtake (B 66). But the fiery Sun sets and fails
to be ever vigilant, whereas the divine fire never sets so that Her-
aclitus may ask rhetorically: How would one ever escape the un-
setting? (B 16). Heraclitus recognizes something of the nature of
his ever-watchful divine fire in the traditional ideas of the fiery Sun
and of high-thundering Zeus.
(7) Fire is a kind of judge and enforcer in B 66, and Justice is spoken of
in similar language in B 28: Justice shall overtake the contrivers of
falsehoods and their witnesses, D_jg jatak^xetai xeud_m
t]jtovar ja l\qtuqar. The virgin Justice is a daughter of Zeus
(Hes. Op. 256), and Zeus is closely associated with retributive jus-
tice: Zeus who oversees all j men and punishes whoever trans-
gresses, Fer dr te ja %kkour j !mhq~pour 1voq ja te_mutai
fr tir "l\qt, (Od. 13.213 214; cf. Hes. Op. 225 247). Justice,
strife, war, fire, and even the logos seem to merge in Heraclitus
thought, although not in a way that makes for a clear formulation.
In Anaximanders cosmic speculation penalties and retributions are
handed out among the great cosmic powers through the assess-
ments of time. The measures of the sun are temporal and spatial
(B 94), and Heraclitus conviction about the stability or regularity
of nature disallows any story like that of Phaethon, the son of He-
lios, whom Zeus is compelled to blow out of the sky with thun-
derbolts when he imperils the earth with his uncontrolled ride
upon his fathers chariot (Ov. Met. 1.750 2.329). Yet the vigi-
lance of Zeus in his quick response to Phaethons threat is in keep-
ing with Heraclitus principles about the enforcement of order.
(8) In B 94 the Furies are the assistants of Justice, indeed her enforc-
ers, who assure that the Sun will keep to his measures. Scholars
have noted the possibility of word-play on the name of the Furies
and the word for strife, 9qim}er from 5qir,54 and Heraclitus may
take this etymological link between the words to support his belief
54 E. Hussey, Epistemology and meaning in Heraclitus, in M. S. Schofield and

M. C. Nussbaum (eds.), Language and Logos (Cambridge 1982) 54.
194 Herbert Granger
in the identity of Justice and Strife (B 80). Luminous Zeus

in B 120 corresponds to Justice in B 94. Zeus and Justice each
have trusted aides who assist them in supervising the orderly pat-
tern in the motions of those celestial bodies that are intimately in-
volved in important terrestrial events, such as day and night and the
seasons that divide the year.
(9) Dionysus is identical with Hades (B 15). The assimilation of deities
is also very likely a characteristic of early, as well as late, Orphic
poetry, and the equation of gods is a feature in allegorical analysis.55
Hesiod may too provide for Zeuss undergoing changes after a
fashion through his association with the different personifications
of Zeal, Victory, Might, and Force. These personifications always
sit at the side of Zeus, and are inseparable from him (Th. 388),
and there may be a parallel between the way these personifications
are intimately associated with Zeus and the way god for Hera-
clitus changes between day and night, winter and summer, war
and peace, satiety and hunger (B 67).56
(10) A much-used epithet of Zeus is father of men and gods, patq
!mdq_m te he_m te (Il. 1.544), and in B 53 the father of all is War,
who is also the king of all. Although Zeus is the king of the
gods, only on occasion does he go by the epithet he_m basiker
ja !mdq_m, king of gods and men (Hes. Th. 923). The god of
war is Ares, not Zeus. Heraclitus may find the unity of fire and
strife displayed in an assimilation of Zeus and Ares, and Heraclitus
may recognize these two traditional gods as the two sides of the
divine nature.
The wording of B 53 is outstandingly hymnal: War is father of all, and
king of all, and some he exhibits as gods, some as men, some he makes
slaves, some free, p|kelor p\mtyv lm pat^q 1sti, p\mtym d basi-
ke}r, ja tor lm heor 5deine tor d !mhq~pour, tor lm do}kour
1po_gse tor d 1keuh]qour. The chiasmic wording of B 53 has much
in common with the wording and content in the passage Hesiod
opens his Works and Days with, in which he praises Zeus for his majestic
55 Parker (above, n. 5) 495; Janko (above, n. 7) 73.

56 For this suggestion, see M. D. Northrup, Hesiodic personifications in Parme-
nides A 37, TAPA 110 (1980) 226 and n. 10, who draws a parallel between
Hesiods elaboration on the nature of Zeus through different personifications
and B67 of Heraclitus fragments.
power for lowering the famous and mighty and for raising the weak and
unknown:
fm te di bqoto %mdqer bl_r %vato_ te vato_ te,
Ngto_ t %qqgto_ te Dir lec\koio 6jgti.
N]a lm cq bqi\ei, N]a d bqi\omta wak]ptei,
Ne?a d !q_fgkom lim}hei ja %dgkom !]nei,
Ne?a d] t Qh}mei sjokim ja !c^moqa j\qvei
Fer rxibqel]tgr, dr rp]qtata d~lata ma_ei.
Through him mortal men alike are unnamed and named,
sung and unsung, by means of great Zeus.
For easily he strengthens, and easily the strengthened he crushes,
Easily the conspicuous he diminishes and the obscure he increases,
Easily he straightens the crooked and the arrogant he withers,
Zeus who thunders aloft, who in the highest halls dwells (Op. 3 8).
What is Heraclitus trying to achieve with the unconventional use of
conventional religious ideas and language?
Xenophanes in B 1 judges the traditional divine stories to be impi-
ous, and he puts his recommendations in the pious language of men
must hymn the god and make do with reverent stories and purified
words, eqv^loir l}hoir ja jahaqo?si k|coir. He calls upon his audi-
ence to make a libation and engage in prayer, and at the conclusion of B
1 he stresses that it is good always to be mindful of the gods. Nothing
of this sort of invocation of piety, which follows so closely traditional
language and practices, appears in Heraclitus. Yet in contrast with Xen-
ophanes, Heraclitus does not purge his speech of the traditional divine
names and epithets. Pythagoras too uses conventional religious termi-
nology and ideas for novel purposes, as the akousmata credited to him
would indicate,57 and perhaps he uses this conventional language to
some degree to reassure his auditors that he is still working within the
tradition, or not too far from it, and to soften the novelty of his opin-
ions. After all, the Pythagoreans make concessions to traditional practi-
57 For instance, the islands of the blest the Pythagoreans identified with the
Sun and the Moon (Iamb. VP 82), and Tartarus they also designated as a
place of punishment (Arist. Apo. 94b33). The sea is the tear of Cronos; the
Great and the Little Bear are the hands of Rhea; the Pleiades the lyre of
the Muses; the planets the dogs of Persephone; daemons can be trapped
in bronze (Porph. VP 41). The oracle of Delphi is identified with the tet-
ractys (fourness), which, in turn, is identified with the harmony the Sirens
sing within (Iamb. VP 82). Iamblichus records one akousma that identifies Py-
thagoras with the Hyperborean Apollo (140).
196 Herbert Granger
ces. Unlike the followers of Orpheus, they accommodate in their diet-

ary restrictions the traditional animal-sacrifices of the Greeks.58 Pythago-
ras or the early Pythagoreans may hide behind traditional religious lan-
guage and practice. Heraclitus, however, gives the traditional language
and stories a distorting twist that shocks his audience and leaves it puz-
zled, when he speaks of that which alone is wise, as what is not willing
and is willing to be called by the name of Zeus (B 32). Some readers
merely dismiss the puzzles, others are stirred to thought, enough so that
they may awaken from their unreflective slumber. Heraclitus provokes
his audience into piety, of which a wakeful state of awareness provides a
necessary condition for a pious state of mind. Genuine piety is contem-
plation of divinity, or, as Xenophanes would say, it is always good to
be mindful of the gods.
Heraclitus does not, however, twist the language of religion merely
to provoke. Traditional religious language and the practices and stories it
informs remain intrinsically related to the truth of the logos. Heraclitus,
however, is no allegorist, like Theagenes of Rhegium, who tries to re-
deem the old mythology by constructing artificial interpretations of the
tales of the gods so that he may tease somehow from them the truths of
reality that Homer expresses in a veiled fashion; see the testimony col-
lected at D-K 8. Instead, Heraclitus recommends that the astute may de-
tect through careful analysis of the traditional religious language and its
stories the truth that they bear in an obscure way, just as the discrimi-
nating may be aided in their appreciation of the unity of life and
death, when they recognize that the word for the death-dealing bow
through a mere shift in accent, from bi|r to b_or, can be appreciated
as the word for life (B 48). Words and the practices they shape cannot
fail to be a key to reality, even when they are misused by Homer and
Hesiod, or by the priests of the Eleusinian Mysteries, or even by the fol-
lowers of Orpheus. Thus, despite his criticism of tradition, Heraclitus
58 The Pythagoreans had elaborate dietary restrictions, which included, in addi-

tion to the celebrated prohibition on beans, forbidding the consumption of
the mallow and certain fish. The Pythagoreans did, however, allow the con-
sumption of sacrificial beasts, and this kept them from coming in conflict
with the traditional sacrificial rituals: eQr lma t_m f]ym oqj eQsqwetai
!mhqpou xuw, oXr hlir 1st tuh/mai di toOto t_m huslym wq 1sheim
lmom, oXr #m t 1sheim jahj,, %kkou d lgdemr f]ou (Iamb. VP 85). The Or-
phics were more demanding, and they prohibited, according to Plato, the con-
sumption of any animal flesh. For sacrifices they used cakes of meal and grain
soaked in honey (Lg. 782c-d).
cannot abandon the traditional way of talking about the divine nature
for an altogether new way, although only through twisting the tradi-
tional religious stories and their language around in unexpected ways
can he find the proper way to reveal the divine truth. Homer is
wrong to have Achilles pray for the disappearance of strife between
gods and men (A 22, Il. 18.107), since Homer fails to appreciate that
the father of all, if properly understood, is War (B 53) and that
Justice is Strife (B 80). In accord with the reformed idea of divinity
initiated by the Milesians and taken up by Xenophanes, Heraclitus refa-
shions the old religious formulas so that they may reveal the true nature
of divinity and may furnish the proper litany for its praise.59
The transcendent or supernatural divinity of Anaximander and
Xenophanes aids in the depersonalization of the nature Hesiod popu-
lates with highly anthropomorphic gods and goddesses. As Hesiod
points out, even the best of seers can never know the mind of Zeus
(fr. 303 Merkelbach-West),60 and when natural events are under the
control of whimsical deities, science would not be possible. Even
though the immanent theism of Anaximenes and Heraclitus brings to-
gether nature and divinity, these theologians too are in a position to
provide an improved appraisal of nature, since, like Anaximander and
Xenophanes, they rid the divine nature of its cruder traditional anthro-
pomorphisms. They tame nature through their domestication of the di-
vinity that makes up the heart of nature. The depersonalization of the
divine nature in the rejection of traditional anthropomorphisms rids
the divinity of its human-like will and its predilection for falling into
capriciousness. Even though in immanent theism divinity and nature
are identified or intimately bound up together, as in the case of Hesiods
theology, divinitys depersonalization allows for predictable regularity in
nature, and this ushers in the stability of the cosmos which was the dom-
inating desideratum of Hesiod as well. Hesiod endeavors in his Theogony
to reassure his audience of the permanence of the current world-order
under the just rule of Zeus, who, after his violent ascension to sover-
eignty, has shown himself capable of overcoming all unsettling obstacles
59 Burkert Gk. Rel. (above, n. 2) 309-310 takes Heraclitus to rebuild the bridge
to tradition with his ceremonial prose or prose art after the Milesians, with
their matter-of-fact prose, had dismantled it.
60 Hes. Fr. 303 M-W ap. Clem. Alex., Strom. 5.14.129: lmtir d( oqd( eXr 1stim 1pi-
whomym !mhqpym, j fstir #m eQdeg Fgmr mom aQciwoio. It is uncertain to
what work this fragment belongs.
198 Herbert Granger
that may come his way. The natural philosophers who are among the
new thinkers of the sixth century provide, however, an assurance,
which is more convincing, of cosmic stability in their reassessment of
the divine nature.61
Bibliography
Algra, K. The Beginnings of Cosmology. In A. A. Long (ed.), The Cambridge
Companion to Early Greek Philosophy (Cambridge 1999) 45 65.
Baldry, H. C. Embryological analogies in Pre-Socratic Cosmogony, CQ 26
(1932) 27 34.
Barnes, J. The Presocratic Philosophers. Vol. 1. Thales to Zeno. London 1979.
Aphorism and argument. In K. Robb (ed.), Language and Thought in Early
Greek Philosophy. (La Salle 1983) 91 109.
Betegh, G. The Derveni Papyrus: Cosmology, Theology, and Interpretation. Cam-
bridge 2004.
Burkert, W. Lore and Science in Ancient Pythagoreanism. Trans. E. L. Minar, Jr.
Cambridge, Mass. 1972 [1962].
Greek Religion. Trans. J. Raffan. Cambridge, Mass. 1985 [1977].
Burnet, J. Early Greek Philosophy. 4th edn. London 1930.
Cherniss, H. F. Characteristics and effects of Presocratic philosophy, JHI 12
(1951) 319 345.
Cornford, F. M. From Religion to Philosophy. London 1912.
Principium Sapientiae: A Study of the Origins of Greek Philosophical Thought.
Cambridge 1952.
Curd, P. The Presocratics as philosophers. In A. Laks and C. Louguet (eds.),
Quest-ce que la Philosophie Prsocratiques? Cahiers de philosophie, 20 (Ville-
neuve dAscq 2002) 115 138.
Deichgrber, K. Hymnische Elemente in der philosophischen Prosa der Vor-
sokratiker, Philologus 88 (1933) 347 361. Repr. in id. Ausgewhlte Kleine
Schriften, ed. H. Gartner, E. Heitsch, and U. Schindel (Hildesheim 1984)
67-81.
Denniston, J. D. Greek Prose Style. Oxford 1952.
Detienne, M. Masters of Truth in Archaic Greece. Trans. J. Lloyd. New York 1996
[1967].
61 I owe a debt of gratitude to Apostolos Pierris for his generous invitation to par-
ticipate in the Fifth Symposium of Ancient Philosophy and for his thoughtful
organization of this conference, which brought together an international col-
lection of scholars, of the most congenial natures, from whose considerable
learning I benefited greatly, and for which I am deeply grateful. I also wish
to express my gratitude to David Sider and Dirk Obbink for their collecting
and editing the papers of the conference and for their bringing them to publi-
cation.
Frnkel, H. Early Greek Poetry and Philosophy. Trans. M. Hadas and J. Willis.
Oxford 1975 [1951].
Graham, D. W. Heraclitus criticism of Ionian philosophy, OSAP 15 (1997)
1 50.
A Testimony of Anaximenes in Plato, CQ 53 (2003) 327 337.
Granger, H. Deaths other kingdom: Heraclitus on the life of the foolish and
the wise, CP 95 (2000) 260 281.
The Theologian Pherecydes of Syros and the early days of natural philos-
ophy, HSCP 103 (2007) 135 163.
Poetry and prose: Xenophanes of Colophon, TAPA 137: 403 433.
Guthrie, W. K. C. Orpheus and Greek Religion. 2nd edn. London 1952.
The Greeks and their Gods. 2nd edn. Boston 1954.
Heidel, W. A. Hecataeus and Xenophanes, AJP 64 (1943) 257-277.
Hussey, E. Epistemology and meaning in Heraclitus, in M. S. Schofield and
M. C. Nussbaum (eds.), Language and Logos (Cambridge 1982) 33-59.
Jaeger, W. The Theology of the Early Greek Philosophers. Trans. E. S. Robinson.
Oxford 1947.
Janko, R. The Physicist as hierophant: Aristophanes, Socrates and the author-
ship of the Derveni Papyrus, ZPE 118 (1997) 61-94.
The Derveni Papyrus (Diagoras of Melos, Apopyrgizontes Logoi?): A new
translation, CP 96 (2001) 1-32.
Review of The Derveni Papyrus by T. Kouremenos, G. M. Parssoglou,
and K. Tsantsanoglou, BMCR 2006.10.29.
Kahn, C. H. Anaximander and the Origins of Greek Cosmology. New York 1960.
Writing Philosophy: Prose and Poetry from Thales to Plato, in H. Yunis
(ed.), Written Texts and the Rise of Literate Culture in Ancient Greece (Cam-
bridge 2003) 139-161.
Kingsley, P. Ancient Philosophy, Mystery and Magic: Empedocles and the Pythagorean
Tradition. Oxford 1995.
Kouremenos, T., G. M. Parssoglou, and K. Tsantsanoglou. The Derveni Papy-
rus. Edited with Introduction and Commentary. Studi e testi per il Corpus dei pa-
piri filosofici greci e latini, vol. 13. Florence 2006.
Laks, A., and G. Most. Studies on the Derveni Papyrus. Oxford 1997.
criture, prose, et les dbuts de la philosophie grecque, Methodos 1 (2001)
131-151. http://methodos.revues.org./document139.html.
Lesher, J. H. Xenophanes of Colophon: A Text and Translation with a Commentary.
Toronto 1992.
Lilja, S. On the Style of the Earliest Greek Prose. Helsinki 1968.
Linforth, I. M. The Arts of Orpheus. Berkeley 1941.
Northrup, M. D. Hesiodic personifications in Parmenides A 37, TAPA 110
(1980) 223-232.
Osborne, C. Was verse the default form for Presocratic philosophy? In C.
Atherton (ed.), Form and Content in Didactic Poetry (Bari 1997) 23-35.
Heraclitus and the rites of established religion, In A. B. Lloyd (ed.), What
is a God? Studies in the Nature of Greek Divinity (London 1997) 35-41.
Parker, R. Early Orphism. In A. Powell (ed.), The Greek World (London
1995) 483 510.
Robinson, T. M. Heraclitus: Fragments. Toronto 1987.
200 Herbert Granger
Schibli, H. S. Pherekydes of Syros. Oxford 1990.

Schofield, M. S. An Essay on Anaxagoras. Cambridge 1980.
Seaford, R. Reciprocity and Ritual: Homer and Tragedy in the Developing City-State.
Oxford 1994.
Sourvinou-Inwood, C. Delphic Oracle, OCD 3 (1996) 445 446.
Thesleff, H. Scientific and technical style in early greek prose, Arctus 4 (1966)
89-113.
Presocratic publicity, In S.-T. Teodorsson (ed.), Greek and Latin Studies in
Memory of Cajus Fabricius (Gteborg 1990) 110-121.
Vlastos, G. Theology and philosophy in early Greek thought, PQ 2 (1952)
97-123. Repr. in id. Studies in Greek Philosophy, ed. D. W. Graham.
Vol. 1. The Presocratics (Princeton 1995) 3-31.
On Heraclitus, AJP 76 (1955) 337-368. Repr. in SGP 1.127-150.
West, M. L. Early Greek Philosophy and the Orient. Oxford 1971.
The Orphic Poems. Oxford 1983.
7. Heraclitus on Measure and the Explicit Emergence
of Rationality
Anthony A. Long
Heraclitus made remarkable contributions to the idea and the ideal of ra-
tionality. In particular, he prefigured many of the distinctive ways by
which Plato and subsequent Greek philosophers conceptualized this car-
dinal notion. These are the two propositions I intend to substantiate in
my paper. No one, I presume, needs to be persuaded that Heraclitus
logos involves rationality in some sense or senses of that word.1 The in-
terest of the inquiry turns entirely on what Heraclitus himself was seek-
ing to express with the term logos and on why, in recounting his own
logos, he uses such words as l]tqom, m|lor, "qlom_g, j|slor, d_jg,
cm~lg, syvqos}mg, num|r, and blokoce?m; 2 for I assume that Heraclitus,
rather than invoking any pre-existing concept of rationality connoted by
the word logos, found himself largely in the process of discovering that
1 I intend this generalization to encompass both the minimalist interpretations of

logos, favored by M. L. West, Early Greek Philosophy and the Orient (Oxford
1971) 124 129, and J. Barnes, The Presocratic Philosophers 2 (London 1982) 59,
(for whom logos is simply Heraclitus account, without metaphysical and theo-
logical connotations) and interpretations represented by Kirk, H. 65 71, 188
189, which take the terms referents to include divine law. I side with Kirk
see Long, Heraclitus and Stoicism, in id., Stoic Studies (Cambridge 1996) 47
51in thinking that Heraclitus adumbrated the Stoic conception of a universal
causal principle, which is not to say that he anticipated their notion of a benev-
olently designing and ruling deity. Actually, even Barnes (ad loc.) allows that
Heraclitus B 1 makes it clear that his account must include or embody some-
thing like a general law of nature. From the copious literature, I select the
following as particularly helpful programmatic statements concerning the Her-
aclitean logos: Kahn, H. 102, rationality as a phenomenal property manifested
in intelligent behavior, and Hussey, Heraclitus in Long (ed.), The Cambridge
Companion to Early Greek Philosophy (Cambridge 1999) 93, who says of the au-
thority that the logos enjoys. It can be none other than the impersonal kind of
authority that is intrinsic to reason or rationality. See also Dilcher, H. ch. 2.
2 To which one could add t\nir if DK 22 A 5 line 15 (fr. xliiiB Kahn), derived
from Theophrastus, was a word Heraclitus, following Anaximander, applied to
the temporal determinacy of cosmic changes. See below, p. 206 for Platos use
of the same word at Gorgias 503e ff.
202 Anthony A. Long
very thing. This is not to say that he created it from nothing. He had
available to him, as we see from the inherent meanings of logos, such
concepts as meaningful discourse, account, ratio, and reckoning; and
he also had available, as we see from the other Greek terms I have listed,
such further concepts as measure, proportion, balance, law, structure,
order, arrangement, judgment, plan, moderation, commonality, and co-
herence. His logos is or has or bespeaks all of these notions, and accord-
ingly it comprises much that we moderns associate with rationality. My
thesis, in essence, is that Heraclitus discovered an idea and ideal of ra-
tionality by incorporating such notions in his logos, something that no
one, to the best of our knowledge, had done before.
If there is a single term that captures the essence of rationality, as
Heraclitus conceived of it, we should opt for the word l]tqom, measure.
Heraclitus focus on measure, and on the conceptual relations and im-
plications of measure, whether for cosmology or mind or ethics,
was his most salient contribution to providing an explicit formulation
of rationality. I emphasize explicit, to avoid giving the impression that
I take Greeks before Heraclitus to have implicitly lacked rationality or
rational practices. The metrical form and structure of Homeric epic,
Hesiods catalogues of divine genealogy, social practices for settling dis-
putes by means of witnesses and the citation of evidence, the political
convention of making speeches pro and contra, counting and measuring
all of these, like human language as such, are implicit manifestations of
rationality. Long before Heraclitus, too, the Greeks had an intuitive un-
derstanding that behavior is normative (syvqome?m) precisely to the ex-
tent that it is sensible or moderate or orderly (jat j|slom or jat lo?qam).
What Heraclitus discovered, so I propose, was not how to make all of
this explicit, but how to articulate rationality in terms of measured or
proportional processes both in (what we would call) non-animate nature
(including, specifically, the unity of opposites) and also in analogous dis-
course, mental disposition, and day-to-day conduct.
Notice, too, that what I am attributing to Heraclitus is an idea, not a
complete account, much less an analysis, of rationality. Rationality com-
prises much more than we find in Heraclitus, including, for instance,
formal proof, rules of logic, or the best means to achieve a given end.
It also involves procedures, such as reductionism and economy of ex-
planation, which he uses but which are probably not presupposed by
his explicit interest in measure and proportion. Heraclitus conceptual-
7. Heraclitus on Measure and the Explicit Emergence of Rationality 203
ization of rationality, nevertheless, had a profound influence on Greek

philosophy.3 It also articulated and helped to propagate the pre-philo-
sophical notion that order and moderation are essential to the proper
functioning of things, whether human or divine.
In order to make the case for Heraclitus seminal contribution to the
idea of rationality, I adopt a three dimensional methodology. Of prime
importance, obviously, are the conceptual connections Heraclitus invites
us to make between the terms l]tqom, m|lor, and so forth, that figure in
his most suggestive statements. However, I shall not start from his own
text because any progress we can make in understanding Heraclitus
idea of rationality requires us to situate that idea within his cultural con-
text. Which brings me to the other two dimensions of my methodology.
When we do the history of philosophy, we generally proceed by ask-
ing how a given thinker positions himself in relation to his predecessors,
following the forward arrow of time. Heraclitus himself, moreover, re-
quires us to do this because he alludes so pointedly and critically to the
older authorities Homer, Hesiod, Pythagoras, Hecataeus, and Xeno-
phanes4 He must suppose that these authorities have failed to grasp and
live up to his idea(l) of rationality, which does not imply, of course,
that he has failed to learn from them or could have thought as he does
without them. Hence a further dimension of my inquiry will involve
us in asking what we can learn about this idea by viewing it against the
background of earlier Greek thought. Important though this background
is, what we can learn from it is severely limited by the huge gaps in our
knowledge. What, for instance, was it about Pythagoras that irked Her-
aclitus so much? If Pythagoras had the interests in musical ratios attributed
to him by the later tradition, Heraclitus should, it would seem, have
counted him an important ally in his own pursuit of rationality and cos-
mic harmony. Thus Kahn, H. 204 suggests that. Heraclitus conception of
cosmic order in terms of k|cor and "qlom_g should be seen as a general-
ization of the Pythagorean notion of musical ratios. Maybe so. But I find
this suggestion hard to reconcile with Heraclitus polemic against Pytha-
goras as the prince of swindlers (B 81).
3 Plato, Aristotle, and Stoics frequently signify the normative sense of rationality
with the expression aqhr k|cor, Translators typically render this phrase by
right reason or correct rule, but they would do better, in my opinion,
to opt for correct ratio or right proportion, in line with what I take to
be Heraclitus seminal conception of rationality.
4 Xenophanes B 17, 40, 42, 57, 67.
204 Anthony A. Long
It is tempting, on the other hand, to think that Anaximander escap-

ed Heraclitus censure precisely because the latter sympathized with
some of his Milesian predecessors ideas. The two thinkers shared an in-
terest in measurement, balance, and cosmic justice.5 All of these notions
are fundamental components of Heraclitean rationality. A complete
papyrus of Anaximanders book might show us even greater lexical
and conceptual affinity to Heraclitus than we are able to state from
our meager documentation. But one thing I very much doubt that it
would reveal is Anaximanders interest in human psychology, ethics,
and politics.6 We have every reason to suppose that Anaximander, like
his fellow Milesians, was primarily a cosmologist, albeit one whose in-
terests were broad enough to include the origins of life and also geog-
raphy. Heraclitus had the much more complex ambition of integrating
cosmology with (what we would call) ethics, psychology, epistemology
and politics.7 It is that integration, so I think, that makes him such a
seminal figure for understanding the distinctive evolution of rationality
in subsequent Greek philosophy.
This observation brings me to the third dimension of my method-
ology. Rather than working from Heraclitus own words or from his re-
lation to his predecessors, I want initially to study his idea of rationality
through what I take to be its afterlife. An obvious way to do that would
take us forward to the Stoics, who drew so much inspiration from re-
flection on Heraclitus. As I have argued elsewhere, they interpreted
Heraclitus with great skill and insight.8 In the Stoic conception of a cau-
sally coherent world, pervaded and guided by the divine and fiery logos
(of which we human beings are integral parts), the resonance of Hera-
clitus is obvious, as it also is in the Stoic project of a life in agreement
with nature. In this paper, rather than looking to the Stoics, I propose
5 Should we detect an implicit criticism of Anaximander in B 80 where Heracli-

tus identifies justice with strife? Hussey (above, n. 1) 110 n. 5, takes Heraclitus
to be making an implicit correction of Anaximander here, but he also (89)
thinks it may be significant that Heraclitus does not attack any of the Milesians
by name.
6 For a very different view of Anaximanders primary interests, see G. Naddaf,
The Greek Concept of Nature (Albany 2005) 92 112, whose highly political in-
terpretation of Anaximander is hardly grounded in any firm evidence.
7 Hence, I take it, the doxographical claim (D.L. 9.5) that his discourse On nature
treated three subjectsthe universe, politics, and theology.
8 Long H. and Stoicicm (above, n. 1), which is a revised version of Heraclitus
and Stoicism, Philosophia 5 6 (1975 1976) 133 156.
to approach Heraclitus from basic constituents of rationality as concep-

tualized by Plato. It is Plato, I will argue, who retrospectively provides
us with the best conceptual threads, so to speak, for tracing and stitching
together the disparate components of Heraclitean rationality.
Plato and Heraclitus on syvqos}mg and measure
When scholars pair Plato and Heraclitus, they generally do so to contrast

them. For Plato, writes Kahn, H. 4), Heraclitus is the theorist of uni-
versal flux. This is quite right as a report on Platos statements about
Heraclitus in the Cratylus and Theaetetus. According to those dialogues,
Heraclitean flux excludes knowledge. I do not doubt that Plato interpret-
ed Heraclitus accordingly; as Aristotle says (Metaph. 1, 987a32 4), and
that it was Platos interpretation of Heraclitean flux as the condition of
phenomena that strongly motivated his theory of changeless Forms. In ad-
dition, I believe, like many scholars, that the reading of Heraclitus as es-
sentially a flux theorist who excluded all knowledge is quite mistaken.9
Why, then, do I turn to Plato, to elucidate Heraclitus idea of rationality?
I do so not because I take Plato to have read Heraclitus directly as a
theorist of rationality, in the way that I shall advocate. My strategy, rath-
er, is to look to Plato for an idea of rationality that seems too strikingly
similar to what I find in Heraclitus to be accidental. Has Heraclitus,
then, influenced Plato after all? I do not know. Perhaps so, or perhaps
it is rather the case that Heraclitus has influenced the intellectual climate
in which Plato found himself. What matters, so far as I am concerned, is
not Platos actual interpretation of Heraclitus or the direct influence of
Heraclitus on Plato, but the congruity of two great minds operating in
more or less the same cultural milieu.10
According to B 112 Heraclitus identified the greatest excellence
(!qet^) and wisdom with syvqome?m.11 Keeping that statement in
mind, let us turn to Plato Gorgias 507e6 508a8. The context is Socrates
9 See Long, Finding oneself in Greek philosophy, Tijdschrift voor Filosofie 54

(1992) 255 79, for a detailed defense of this view.
10 See R. Wardy, The unity of opposites in Platos Symposium, OSAP 23 (2002)
1 61, who argues that Plato makes consistently positive use of Heraclitean per-
spectives in the Symposium.
11 Cited by Stobaeus 3.1.178. Doubts about the authenticity of this fragment are
well answered by Kahn, H. 120.
206 Anthony A. Long
refutation of Callicles undiscriminating hedonism. Here is a summary

of his immediately preceding argument (503d-507c):
1. Every craftsman imposes organization (t\nir) and order (j|slor) on
the materials with which he works.
2. As with artifacts, so too with bodies and souls.
3. The name for organization and order in the soul is m|lor, which
condition is equivalent to justice and syvqos}mg.
4. In fact the excellence of anything is a function of t\nir and j|slor.
5. Therefore it is the s~vqym soul that is excellent.
Having proved that psychic excellence consists in order and syvqos}mg,
Socrates continues:
The wise say, Callicles, that heaven and earth and gods and humans, are
bound together by community and friendship and order and moderation
and justice. That is why they call this universe order, and not disorder
and immoderation I think you are not applying your mind to these
things, but are failing to see that geometrical equality [or proportion]
has great power among both gods and humans. You think you should prac-
tice getting more (pleonexia) because you neglect geometry. (507e-508a)
vas d oR sovo_, Jakk_jkeir, ja oqqamm ja c/m ja heor ja !mhq~pour
tm joimym_am sum]weim ja vik_am ja josli|tgta ja syvqos}mgm ja
dijai|tgta, ja t fkom toOto di taOta j|slom jakoOsim, 2ta?qe, oqj
!josl_am oqd !jokas_am. s d] loi doje?r oq pqos]weim tm moOm to}toir
!kk k]kgh]m se fti B Qs|tgr B ceyletqij ja 1m heo?r ja 1m !mhq~poir
l]ca d}matai, s d pkeomen_am oUei de?m !sje?m ceyletq_ar cq !leke?r.
In this remarkable passage Plato invokes cosmic order as a universal binding
force, glossing that force as community, friendship, order, syvqos}mg, and
justice, and finally, as geometrical equality or proportion. He also exploits a
pun, in the manner of Heraclitus, on j|slor and j|slior. Echoes of Her-
aclitus are unmistakable, not only in the shared terminology, including the
shared reproach concerning lack of understanding (Heraclitus B 1 kamh\mei,
Plato k]kghem), but most importantly in the macrocosm/microcosm analo-
gy. I am thinking particularly of B 114:
Those who speak with intelligence must rely on what is common to all
things [num|r, implying the commonality of m|lor] as a city relies on its
law, and still more resolutely; for all human laws are nourished by one
law, the divine one. For it has all the power it wants, and suffices and
more than suffices for all.
nm m|\ k]comtar Qswuq_feshai wq t` num` p\mtym, fjyspeq m|l\ p|kir,
ja pok Qswuqot]qyr. tq]vomtai cq p\mter oR !mhq~peioi m|loi rp 2mr
toO he_ou jqate? cq tosoOtom bj|som 1h]kei ja 1naqje? psi ja peqic_-

metai.
Just like Platos Socrates in the Gorgias, Heraclitus grounds his ethics in a
cosmological model, treating the world order itself as just or law-gov-
erned. He does not use the phrase geometrical equality in the extant
fragments, but proportionality is central to his conception of the balance
that governs natural processes. He insists (B 94) that justice governs the
measured behavior of the sun. And, just as Socrates stresses the power of
geometrical equality, so Heraclitus emphasizes the universal control ex-
ercised by divine law. We shall hardly go wrong if we connect his
grounds for the excellence of syvqome?m with the thought, which Soc-
rates advances here, that appropriate human norms are validated by their
correspondence to the cosmic order.
Commentators on the Gorgias often look to Pythagoreanism as Pla-
tos inspiration in this passage. I do not exclude that as part of the back-
ground. Plato is often indebted to multiple predecessors. In this context,
with its allusion to cosmic friendship and community, Empedocles nat-
urally comes to mind. But I am encouraged that Terence Irwin also re-
fers to Heraclitus B 114 in his commentary on the Gorgias passage (1979,
226), where he observes that Platos interest may be no more in Pytha-
goreanism than in Presocratic theory in general. Heraclitus is the first
Presocratic who quite certainly used the word j|slor in reference to the
universe (B 30). There is good reason to credit him with being the
prime mover of many of the ideas that Plato presents here.12
What may we infer from the Gorgias concerning Platos idea of ra-
tionality?
Socrates in the dialogue has argued that injustice presupposes igno-
rance about ones own good because injustice involves a disorderly and
uncontrolled self; a self that is to flourish, on this view, needs to be regu-
lated and duly proportioned. By appealing to geometrical equality and
connecting it with syvqos}mg, Plato implies (1) that syvqos}mg consists
in a measured, balanced, and moderated mentality; (2) that that mentality
is the essential condition of happiness; and (3) that it is equivalent to wis-
dom and the application of intelligence. Thus the Gorgias adumbrates the
great teaching program of the Republic whose would-be guardians need
12 With Grg. 492e10 493a3 (Platos citation of Euripides, Who knows if living
is being dead, and being dead is living? and the s_la/s/la identification), cf.
Sextus Empiricus, PH 3.230, who attributes something very similar to Heracli-
tus, as noted by Dodds in his edition ad loc.
208 Anthony A. Long
mathematical training in order to facilitate the rule of reason over their

lives and those of their fellow citizens.
In this mathematical agenda Plato went far beyond anything proposed
by Heraclitus. Yet, we may wonder what it was that authorized Plato to
suppose that mathematics has an intrinsic connection with morality.13
Outside the Greek context this connection is likely to appear quite bi-
zarre. What did these two domains (so different from one another from
our modern perspective) have in common, in Platos eyes? The answer
suggested in numerous Platonic contexts is proportion or measure.
In the myth of divine judgment with which Plato concludes the
Gorgias, Socrates offers a graphic description of a tyrants soul (524e-
525a):
[The divine judge] saw nothing healthy in the soul, but that it was filled
with scars from false oaths and injustice .. and everything crooked from
lying and deception, and nothing straight, from being reared without
truth. And he saw that, from power and luxury and excess and failure to
control actions, this soul was brimming over with disproportion and ugliness.
jate?dem oqdm rcir cm t/r xuw/r, !kk oqk_m lestm rp 1pioqji_m
ja !dij_ar ja p\mta sjoki rp xe}dour ja !kafome_ar ja oqdm
eqh di t %meu !kghe_ar tehq\vhai ja rp 1nous_ar ja tquv/r ja
vbqeyr ja !jqat_ar t_m pq\neym !sulletq_ar te ja aQswq|tgtor c]lou-
sam tm xuwm eWdem.
We are so accustomed to Platos pairing of truth and beauty that it is
tempting to accept this association without reflecting on its conceptual
assumptions. Yet once one thinks about it, the question of its meaning
clamors for elucidation. Truth in the sense of mere fact is often ugly.
What makes truth beautiful, according to the passage I have just quoted,
is to be sought from the following cluster of ideas health, honesty, jus-
tice, moderation, straightness, control, and proportion. The truth of
something, according to this model, is normative. Just as lying falls out-
side the norms of proper discourse, so truth also characterizes the way
something should actually be if it is functioning optimally where the
condition of functioning optimally is balance, straightness, and proper
proportion. Hence Platos liking for musical and mathematical illustra-
tions of truth. A harmonious sound is true to the relevant ratios of
the tetrachord. A well constructed square is true to the principle that
its four sides are equal and subtend four right angles.
13 See M. Burnyeat, Plato on why mathematics is good for the soul, PBA 103
(2000) 1 81, for brilliant suggestions in response to this question.
The beauty of truth, then, is a highly theoretical notion, grounded

on ideas of symmetry and balance. This notion, as we see in the Gorgias,
enables Plato to represent a tyrannical soul as devoid of truth precisely
because such a soul is unbalanced, given to excess, and therefore ugly.
Such a soul fails to impose a proper ratio on itself and its actions. Like
irrationals in mathematics, which defy whole number ratio relation-
ships, a disorderly soul is irrational, and irrational by virtue of its lack
of proportion.
Accordingly, when Plato talks about the beauty of truth he is not
saying that the content of any true statement (or fact) must be beautiful.
The beauty of truth picks out the relation that anything with a purchase
on truth (whether a soul or an action or a statement) has to the norms of
straightness and proper proportion. Hence we can see why Plato could
regard truth telling itself as beautiful, in as much as it corresponds and
coheres with what is the case.
There is much more one could say about Platos conceptualization
of rationality in terms of measure and proportion. This cluster of ideas is
especially prominent in the Philebus, where it is measure that heads the
classification of goods (66a). But, I think I have said enough here to pro-
vide a Platonic perspective or retrospective on Heraclitus. Let us return
with some detail to Heraclitus B 112:
Sophronein is the greatest virtue and wisdom: to speak and do things that are
true, understanding them in accordance with their nature.
syvqome?m !qet lec_stg ja sov_g, !kgh]a k]ceim ja poie?m jat v}sim
1pa@omtar.
Like Kahn, H. ad loc., I punctuate the fragment after sov_g, thus making
the entire text a comment on good sense or practical wisdom, which is
how we may gloss syvqome?m. This punctuation, rather than the placing
of a comma after lec_stg, suits the sense and rhythm of the passage, with
its division into three cola. The last colon, understanding [sc. true] things
in accordance with their nature, echoes Heraclitus programmatic state-
ments in B 1, where he presents himself as the spokesman of natures
logos, setting forth an account of the way things are.14 His audience, as
he regretfully says there, continually fails to heed his account, because
they retreat into their private worlds. Here, in B 112, he asks us to
take the message of B 1 to constitute the essence of syvqome?m, glossing
that virtue as speaking and doing things that are true.
14 I fail to see why Kirk, H. 390, characterizes B 112 as a banal paraphrase.

210 Anthony A. Long
What does Heraclitus mean by doing things that are true? Kahn,
H. 122 puzzles over this curious and seemingly unparalleled expression.
He suggests that Heraclitus wants to say that the man whose thinking is
sound will not hide the truth but signify it in his actions as in his words.
This suggestion does not go far enough, in my opinion. Taking our lead
from Plato, we shall hardly go wrong if we endow Heraclitus !kgh]a
with connotations of straightness, balance, and due proportion. Thus
his truths, as the context of B 112 requires, are norms of properly
measured action as much as they are norms of thought and speech.
Plato regularly couples balanced (l]tqior) with such commenda-
tory words as s~vqym, jak|r, jahaq|r, %qistor, and b]baior.15 In the
Republic (3, 412a5) the most musical and harmonious person is the
one who best combines physical and musical education and applies
them to his soul in the most measured way (letqi~tata). The mark
of an unmusical and unshapely soul is !letq_a (6, 486d5). These con-
texts are in line with the passage I already discussed from the Gorgias,
but the Republic proposes still more precise connections between ration-
ality and balance or proportion.
Reviewing the newly formulated community, Socrates invites his in-
terlocutors to agree that it could be called s~vqym and master of itself
if its better constituent rules over its worse one (4, 431b-c). He continues:
You all would also find especially among its children and women and slaves
a multitude of all kinds of desires and pleasures and pains but simple
and moderate ones, which are guided by logismos in association with intel-
lect and correct opinionthese you would encounter only among a few
people, ones with the best natural constitutionand the best education.
ja lm ja t\r ce pokkr ja pamtodapr 1pihul_ar ja Bdom\r te ja
k}par 1m pais l\kista %m tir evqoi ja cumain ja oQj]tair tr d] ce
"pkr te ja letq_ar, aT d let moO te ja d|ngr aqh/r kocisl` %comtai,
1m ak_coir te 1pite}n, ja to?r b]ktista lm vOsim, b]ktista d paideuhe?sim.
Moderate or balanced emotions are the outcomes of reasons rule, or, to
be more precise, the rule of calculation. We may translate kocisl|r by
reason or rationality, but, in doing so, we shall miss Platos own insights
if we overlook the mathematical connotations of the word. kocisl|r is
the disposition or faculty in virtue of which a soul has the capacity to
rule itselfthat is to say, impose order and balance on its emotions
15 E.g., R. 339b8 syvq|myr te ja letq_yr; 5, 466b6 l]tqior ja b]baior ja %qi-

stor; Phd. 86c2 jak_r ja letq_yr; 108c3 jahaq_r te ja letq_yr.
and desires through subjecting them to appropriate calculation or meas-

urement or proportion.
The same cluster of ideas recurs like a Leitmotif throughout the Re-
public. Truth, Socrates proposes, is akin to proportion (1lletq_a, 6,
486d7). Therefore, he infers, persons who are naturally suited to grasp
the truths of reality (i. e. potential philosophers) require an 5lletqor
mentality. In the dialogues final book Socrates offers measuring, count-
ing and weighing as the activities that correct optical and other illusions
generated by the senses (10, 602d). These, he says, are the function of
the souls calculative (logistikon) part, which is its best part precisely be-
cause it puts its trust in measure and calculation.
The message of these passages can, of course, be amplified by refer-
ence to other parts of the Platonic corpus. In the Protagoras (356d-357d)
Socrates proposes that it is the measuring craft (letqgtij tewm^)
which will prevent people from making mistakes in determining what
is good for themselves, where goodness is construed as achieving a pre-
ponderance of pleasure over pain.16 Similarly in the Philebus (55e) all
true crafts require measure and such measuring methods as weighing
and counting. The final Platonic passage I offer as retrospective to Her-
aclitus comes from the Laws (4.715e7 716b5), where the Athenian
Stranger, imagining himself addressing the colonists of the newly found-
ed community, preaches to them as follows:
God, so the ancient account tells us, holds the beginning, the end, and the
middle of all beings and proceeds without deviation in his natural revolu-
tion. He is always accompanied by justice, who punishes those who forsake
the divine law. One who intends to flourish sticks to justice and is compli-
ant and well ordered, but the person puffed up by arrogance who thinks
he needs no ruler or guide is abandoned by God and utterly ruins
himself, his household and his city.
b lm d he|r, speq ja b pakair k|cor, !qw^m te ja tekeutm ja l]sa
t_m emtym "p\mtym 5wym, eqhe_ peqa_mei jat v}sim peqipoqeu|lemor t`
d !e sum]petai d_jg t_m !pokeipol]mym toO he_ou m|lou tilyq|r, Hr b lm
eqdailom^seim l]kkym 1w|lemor sum]petai tapeimr ja jejoslgl]mor, b d]
tir 1naqher rp lecakauw_ar , r oute %qwomtor oute timr Bcel|mor
de|lemor jatake_petai 5qglor heoO 2aut|m te ja oWjom ja p|kim
%qdgm !m\statom 1po_gsem.
16 Note also Prt. 326b on the educational need for eqquhl_a and eqaqlost_a, and
Phlb. throughout, especially 25e, 26a, and 65a.
212 Anthony A. Long
The moral of the story is that human beings should align themselves
with God. How So? 17
There is only one way so to act, and only one account thereof, viz. the ancient
logos that like is friend to like, provided that they conform to measure; things
that do not conform to measure, on the other hand, are friend neither to
themselves nor to those that do conform. For us it is God who must
pre-eminently be the measure of all things, much more so than any
human being, as they say [implicitly correcting Protagoras celebrated dic-
tum]. To become befriended to one of this sort, it is necessary to become as
like to it as possible, and that means according to this logos that whoever of
us is sphrn is Gods friend, because such a one is like God.
T_r owm d pqnir v_kg ja !j|kouhor he`; l_a, ja 6ma k|com 5wousa
!qwa?om, fti t` lm blo_\ t floiom emti letq_\ v_kom #m eUg, t d %letqa
oute !kk^koir oute to?r 1ll]tqoir. b d her Bl?m p\mtym wqgl\tym l]tqom
#m eUg l\kista, ja pok lkkom E po} tir, r vasim, %mhqypor tm owm
t` toio}t\ pqosvik/ cemgs|lemom, eQr d}malim fti l\kista ja aqtm
toioOtom !macja?om c_cmeshai, ja jat toOtom d tm k|com b lm
s~vqym Bl_m he` v_kor, floior c\q. (716cl-716d2)
This passage is replete with intertextuality. It corrects Protagoras and it also
strongly recalls Hesiods Zeus whose rule over mortals in the Works and
Days (256 85) is assisted by his daughter Justice and her inexorable atten-
tion to human conduct. But Plato is not simply parroting Hesiod. His di-
vinity in this passage (as in book 10 of the Laws, the Timaeus, and the Phil-
ebus) is characterized in language that resonates with scientific overtones.
Plato seems to identify divinity with the worlds natural revolutions; and
the idea of divinity as the principle of cosmic order is reinforced by the
focus on measure. With all this leading on to characterization of the person
who is s~vqym, we are clearly within a conceptual context that can recall
Heraclitus and his prescription to follow a divine law.18
17 Ti. 90c-d is similar in thought and expression; see Long Eudaimonism, divin-
ity and rationality in Greek ethics, PBACAP 19 (2003) 134 137.
18 Plato echoes Heraclitus linguistically as well as conceptually: note especially the
repeated instances of logos and also jat v}sim and he?or m|lor. England [1921,
ad loc.] on the authority of the scholiast and Eusebius, identifies Platos ancient
logos with an Orphic saying: Zeus is beginning, Zeus middle, and from Zeus
are all things accomplished, Zeus foundation of earth and starry heaven. Plato
clearly had a hallowed saying in mind, but we should not assume that he cited it
as being specifically Orphic. The sentiment that Zeus encompasses all things oc-
curs in Greek tragedy (cf. Aes. Ag. 160 ff., and Sophocles Tr. 1278) and Hera-
clitus gives his own expression to it in B 67. As with Gorgias 503d ff., I do not
claim that Plato had Heraclitus in mind here, but I see no reason to exclude that
as a possibility.
The measures of Heraclitus

Plato, as we have seen, repeatedly spells out intrinsic connections be-
tween measure or proportion, calculation, balance, law, intelligence
and syvqos}mg. Heraclitus largely requires us to make those connec-
tions for ourselves. I now pursue more of the relevant connections,
with Platos cosmological passage from the Laws serving as a guideline.
We may start from unequivocal certainties. As envisioned by Hera-
clitus, the world is:
1. An everlasting process of the balanced, measured, and proportional
changes of fire. (B 30)
2. The world, in its diurnal, seasonal, and annual cycle, is god. (B 67)
3. All things are steered by (divine) intellect. (B 41)
4. Cosmic order is an expression of divine justice. (B 94)
In these striking statements we see Heraclitus giving his own expression
to the Platonic idea that cosmic order is a process and system of measure
constituted by divinity. No less Platonically, he associates cosmic order
and divinity with the principle of justice. What this signifies in both
philosophers is a notion of natural law, meaning that norms of con-
duct and retribution for misconduct are not simply human institutions,
but mandated by the structure of reality. Hence Heraclitus, just like
Plato, finds it appropriate to apply propositions about the worlds order-
ly structure to the human domain, as he does explicitly in
5. Human law is nurtured by divine law. (B 114)
or, as he does by implication in passages that emphasize the good-
ness of order, regularity and control, such as
6. One should fight for law as one does for a city wall. (B 44)
7. Of prime importance is the control of vbqir (which is the opposite
of syvqos}mg). (B 43)
8. Anger causes psychological damage. (B 85)
9. Justice seizing liars and false witnesses is inevitable. (B 28)
10. syvqome?m is the prime excellence and wisdom. (B 112)
I propose that we call all ten of these Heraclitean propositions meas-
ures, trading on the multiple senses of the English word measure to sig-
nify the following range of meanings: (i) determinate quantity; (ii) pro-
portion or ratio; (iii) moderation; (iv) limit; (v) rule or standard; and (vi)
political or judicial decision.
Heraclitus has different words to signify these different senses of
measure but we shall not go wrong if we privilege his recourse to
214 Anthony A. Long
the term l]tqom. He applies this word to the worlds constantly balanced
changes (B 30), to the suns due and regular behavior (B 94), and, in its
verbal form, to the equivalence in quantity of the change from sea water
to earth and back again (B 31). These are instances of measure in the
sense of (i) determinate quantity; (ii) proportion or ratio; and (iv)
limit. For measure in the sense of (iii) moderation we have syvqos}mg;
for (v) rule or standard m|lor; and for (vi) judicial decision d_jg.
Heraclitus own words and their striking affinity to Platonic usage
establish the cardinal importance of measure to the connections he
intuited between three fundamental domains: 1, the way the world is
structured as a physical system; 2, the worlds divine governance and
governor; and 3, the norms of human conduct. Should we, then, use
measure as our translation of the Heraclitean logos? At least one scholar
has done so, but that translation, tempting though it is, does not capture
Heraclitus creative use of the term in his most programmatic passages.19
He did not discover the importance of measure as a criterion for acting
appropriately. That notion was already present in his culture. His great
innovation was to take measure as the key to understanding structure,
balance, and good order. Perhaps the best English word to translate
his logos is rationale. As such, it incorporates the measures of language
and thought, if these are deployed objectively in the interests of truth.
With this in mind, I now review Heraclitus most important state-
ment (B 1):
Of this rationale that is so always people are not in touch, both before hear-
ing it and after they have first heard it.
toO d k|cou toOd( 1|mtor !e !n}metoi c_momtai %mhqypoi ja pq|shem C
!joOsai ja !jo}samter t pq_tom.
I translate !n}metoi by not in touch in order to register what I take to
be Heraclitus punning allusion to num|r, common, His audience, as
he says in B 2, lives in a private world, unheeding of the common
19 Kirk, H. 39 finds measure, the most common meaning, judged purely by stat-
istical criteria in the extant fragments. Referring to K. Freeman, Ancilla to the
Pre-socratic Philosophers (London 1956) 116, he writes [She] has well stressed
that the concept of measure is implicit in the Logos of Heraclitus, but rightly
observes that it makes but little sense to translate it thus in fr. 1, as Freeman
does. Kirk himself favors formula for B 1, 2, and 50. Other translations he
surveys include word (Burnet) meaning of Heraclitus teaching (Snell),
argument (Verdenius), and account(ing) (Minar). See also Rowett at n.
19 in this volume.
logos. By calling it common, Heraclitus is saying that the logos is both

publicly available and also objective; it is not his personal account,
and his audience are to listen not to him but to it (B 50)20. The objec-
tivity of the logos is registered in its everlasting validity.
Heraclitus continues:
For although all things come to pass in accordance with this rationale,
human beings resemble those without experience even though they expe-
rience such words and deeds as I expound, when I distinguish each thing
according to its nature and explain how it is.
cimol]mym cq p\mtym jat tm k|com t|mde !pe_qoisim 1o_jasi, peiq~le-
moi ja 1p]ym ja 5qcym toio}tym, bjo_ym 1c digceOlai jat v}sim diai-
q]ym 6jastom ja vq\fym fjyr 5wei.
The rationale that Heraclitus states equips him both to account for all
happenings and to provide a taxonomy of particular things, and so ex-
plain their natures. With his insights into measure, proportion and bal-
ance as the key to understanding the world Heraclitus arrived at an idea
of rationality that includes the following attributes objectivity, imper-
sonality, law-like authority, coherence, generality, intelligibility, and ac-
countability. Also, by emphasizing the commonality of the logos and be-
rating people for failing to heed it, Heraclitus implies that rationality is
essential to an authentic and excellent human lifea life which one lives
awake and fully cognizant of natures rationale.
Did Heraclitus connect the measures that constitute the worlds ra-
tionale with his ideas about the soul? Two fragments bear on this ques-
tion:
You would not discover the boundaries of the soul by going, traversing
every route; so deep is the logos that it has.
xuw/r pe_qata Qm oqj #m 1ne}qoio, psam 1pipoqeu|lemor bd|m ovty
bahm k|com 5wei. (B 45)
Soul has a self-increasing logos.
xuw/r 1sti k|cor 2autm aunym. (B 115)
If, contrary to the first of these fragments, the soul were limited in its
range, would it be capable in principle of understanding the structure
of the everlasting Heraclitean world? What we seem to learn here
20 The single ms. of Hippolytus (the source of B 50) was unsure whether d|c-
lator or k|cor was correct in the quotation. So Kirk, H. 66, who gives com-
pelling reasons for accepting the latter reading.
216 Anthony A. Long
and in the next fragment is that the soul is potentially boundless, which I
take to mean unlimited in its capacity to measure, i. e., understand na-
tures rationale. The more the soul expands, as it were, the more it is
capable of putting itself in touch with the common logos. 21
Antecedent and contemporary ramifications
No more than Parmenides, Empedocles or even Plato, did Heraclitus ex-

press his revolutionary philosophy by coining new terminology. The idea
of rationality that I have attributed to him draws on language which is as
old as Homer and Hesiod, one or both of whom use the words k|cor,
l]tqom, m|lor, "qlom_g, j|slor, d_jg, s~vqym, and num|r. Heraclitus orig-
inality is conceptual; it consists in his meanings and the contexts in which
he uses these words. In the next part of this paper I give a short linguistic
survey as background for viewing his innovations.
k|cor
In its epic usage, k|cor, far from connoting rationality, pertains to dis-
course that is deceptive rather than true, as in the formula aRl}kioi
k|coi. Calypso tries to soothe Odysseus with such words (Od. 1.56),
a usage that looks forward to the enchanting powers of logos advanced
by Gorgias in the Encomium of Helen. Pindar, by contrast, tends to associ-
ate k|cor with such words as !kgh^r and oq xeud^r.22 His usage suggests
that, though k|cor has begun to acquire normative connotations, its truth
or correctness needs to be specified as such, as in Herodotus 2.17.1 where
aqhr k|cor refers to the correct reckoning concerning the flooding of
the Nile.23
The tally of k|cor in Presocratic philosophers other than Heraclitus
is surprisingly meager. Setting aside instances where the term means
simply what is said or the familiar antithesis of k|cor/5qcom, the
21 Cf. Kahn, H. 130: A logos so profound and limitless can scarcely be distinct
from the universal logos, according to which all things come to pass.
22 E.g. Pi. O. 1.28, 4.21, 7.68, P. 1.68.
23 Guthrie HGP 1.420 424, gives numerous examples of the usage of k|cor, but
most of them, in the nature of the evidence, are from authors later than Her-
aclitus.
only instances worth reporting are Parmenides B 7.5 jq?mai k|c\ and
Leucippus B 2 oqdm wq/la l\tgm c_metai, !kk p\mta 1j k|cou te
ja rp( !m\cjgr (nothing occurs at random but all things from logos
and by necessity).24 If this latter citation is authentic, it would represent
the closest parallel to Heraclitus conception of the worlds rationale or
determinate measure.
There seems to be no fifth-century parallel for Parmenides instru-
mental dative.25 In its context, the command judge by k|cor is in-
tended to detach his audience from reliance on their senses. What
they are to judge is the goddesss contentious elenchos. Taking elenchos
to mean proof, it is tempting to interpret the command as the re-
quirement to assess the goddesss argument by discussing it. Parmenides
gives an extraordinary demonstration of rationality in action, so to
speak. He offers nothing comparable to Geraclitus on the conceptual
connections between k|cor, l]tqom, syvqos}mg, and so forth. The
later Presocratics, according to our record, had still less to say.
l]tqom
The world of Homer and Hesiod conforms to measure. We may think, for
instance, of the five scenes represented on the Shield of Achilles (Il. 18.
478 ff.), especially the first scene depicting earth, sea, and heavens, or the
equidistance from earth of sky and Tartarus (Theog. 719).The expanse of
the sea is actually conveyed by the word l]tqa (e. g., Op. 648). As the
world at large is a measured or balanced structure, so in the human domain
the concept of measure or due proportion serves archaic thought as its prin-
cipal ethical norm. Thus Hesiod tells his farmers to preserve measures
(l]tqa vuk\sseshai), glossing his injunction by saying that jaiq|r (propor-
tion) is best in everything (Op. 694)26
Here we have a classic statement of the ethics of syvqos}mg, which
Homer likes to express in such phrases as jat j|slom or jat lo?qam.
He characterizes the recalcitrant Thersites as someone unmeasured in his
24 I discount instances of logos from ethical fragments attributed to Democritus,

both on grounds of lateness and also from doubts about their authenticity.
25 Peter Kingsley may be right to question the reading k|c\. He suggests to me in
conversation k|cou, taking it with 5kecwom.
26 According to a fragment of the Melampodia (Strabo 14.1.27 = fr. 278 M-W),
Hesiod used a word for true, 1t^tulor, in connection with the numerical cal-
culation of a l]tqom (!qihlr 1t^tulor l]tqou).
218 Anthony A. Long
speech (!letqoep^r), minded to utter disorderly words (5pea %josla)

and act oq jat j|slom (Il. 2.212 14). We would hardly call Thersites
irrational for his speaking out of line. But the semantic connections be-
tween j|slor and l]tqom, which become so prominent in Heraclitus
and Plato, facilitated the application of these words to mark intelligence.
For an instructive instance, I cite Pindar, I. 6.71 2, where the poet
praises Lampon for pursuing measures in his judgment, and also keep-
ing to measures [sc. in his actions?], with a tongue that does not exceed
good sense: l]tqa lm cm~l di~jym, l]tqa d ja jat]wymck_ssa d(
oqj 5ny vqem_m.27
Clearly Heraclitus audience was familiar with the general idea of
measured and moderate mentality, captured in the maxim lgdm
%cam. What makes his thought so challenging is the relationship he sug-
gests between this traditional norm and a world whose structure explic-
itly conforms to measure(s). Does any earlier text have a bearing on
these Heraclitean propositions?
The answer is to be found in an elegiac couplet of Solon (fr. 16
West). The lines are cited by two Christian authors, who interpret
them as a very wise statement about God: It is very difficult to
know the obscure measure of intelligence, which alone holds the limits
of all things (cmylos}mgr d( !vamr wakep~tat|m 1sti mo/sai j l]tqom,
d d p\mtym pe_qata loOmom 5wei).
Given the absence of any context for these lines, we have to inter-
pret them from other statements by Solon. In his longest surviving
poem (fr. 1), after praying for justly gained prosperity, Solon reflects
on the justice of Zeus. The god sees the end of all things (l. 17),
but, sure though he is in his punishments, wrongdoers may be punished
through their innocent children. Thus human beings are unable to dis-
cern the fairness or logic of Zeus punishments. Mortals have their own
particular mentalities and ambitions, over-confident that they will suc-
ceed. The future is unpredictable, for we are in no position to know in
advance the fortunes gods will dispense to us. In a further isolated line
(fr. 17), reiterating the obscurity of the cmylos}mg that encompasses
everything, Solon states the pessimistic message of his long poem as fol-
lows: The mind of the immortals is completely obscure to human be-
ings (p\mt, d( !ham\tym !vamr m|or !mhq~poisim). The Christian
source of this line says that here Solon follows Hesiod, probably with
27 Note also Solon fr. 13. 52 West on the craftsman who knows the l]tqom of
lovely sov_a.
reference to Op. 483 4: The mind of Zeus is different at different

times and hard for mortal men to know.
There can be little doubt that Solons cmylos}mg alludes to some
kind of divine intelligence or plan, vaguer though this is in its reference
than his Christian sources took it to be. How, then, should we interpret
his fr. 16? Vlastos took cmylos}mg to apply specifically to human jus-
tice.28 He supported this proposal by reference to Theognis 1171 2,
The best thing gods give mortal men is cm~lg. cm~lg holds the limits
of everything. Theognis, in a clearly ethical context, appears to be al-
luding to Solon, but rather than agreeing with his predecessor, as Vlastos
suggested, Theognis has corrected Solons epistemological and theolog-
ical pessimism by making cm~lg a divine gift to human beings. I take
Solon to be saying that the mind of the controlling divinityor the
mind of the gods in generalis virtually inscrutable for mortals.
Does he express nothing more than that quite traditional thought?
The intriguing words are l]tqom and p\mtym pe_qata. Solon postulates
a bounded universe that conforms to the measure of cosmic intelligence.
With these thoughts Solon seems to anticipate Heraclitus so closely that
one is tempted to regard the latter as directly alluding to him, especially
if we compare Heraclitus B 41 on the wisdom of knowing the cm~lg
that steers everything through everything with Soloms cosmic clylo-
s}mg that holds the limits of everything.
But the allusion, if I am right, is also highly critical. Although Her-
aclitus and Solon agree in taking the world to be intelligently governed
according to measure, they disagree on the implications of that fact for
human beings. For Solon human beings have no access to any logos that
makes the divine measure accountable. For Heraclitus, by contrast, the
obscurity of nature (B 123) and the superiority of its non-evident
structure (B 54, connected with Solon fr. 16 by Vlastos loc. cit.) are
challenges to the human intellects powers of discovery. Thanks to
the commonality of logos, human beings have the capacity to engage
in science and thus close some of the traditional gap between the mortal
and the immortal.
We may now bring in Xenophanes for three reasons. First, like
Solon, he envisioned a divinity that controls all things by thought (Xen-
oph. B 25). Secondly, like Solon again, Xenophanes divinity has a mind
that mortals are in no position to fathom (B 23). Thirdly, Xenophanes
withholds knowledge from the statements he makes about all things,
28 G. Vlastos, Solonian justice, CP 41 (1946) 67.

220 Anthony A. Long
consigning human cognition to mere opinion or conjecture (B34 5).

Xenophanes, of course, is one of the older authorities whom Heraclitus
explicitly castigated (B 40). I now want to argue, with reference to that
fact and what we have just observed in Solon, that this antecedent ma-
terial has much to offer us in drawing conclusions about Heraclitus idea
of rationality.
Giving a measured account/rationale of all things
As early as Homer and Hesiod the Greeks showed their awareness that
the physical world is an orderly system. They did not then call it a
j|slor, but they envisioned it as a limited and tripartite structure of
heaven, earth and seas, and underworld, with each of these three do-
mains under its own divine manager.29 They also envisioned the
world and its inhabitants as a regulated system, with its regulation some-
times assigned to Zeus and sometimes to Moira. Without sharply distin-
guishing these regulative powers from one another, in referring to Zeus,
they emphasized foresight and intelligence, while their references to
Moira implied necessity and an embryonic sense of causal connected-
ness.
This conception of the world, though imprecise in its details, was
rationalistic in presupposing that events do not happen for no reason.
Myth and religion provided explanations, albeit personalist ones, such
as the anger of Zeus or the resentment of Poseidon, or invoked vaguely
impersonal agencies, such as fate. The conception was also rationalistic
in the prudential sense reflected in the value of syvqos}mg, which sig-
nified the wisdom of self-restraint, especially with regard to ones appro-
priate attitude to the controlling divinities.
What this archaic outlook completely lacked, so far as one can see,
was the idea that human intelligence might be able to enter the orderly
cosmos it already intuited and discover laws of nature for itself, gaining
some authentic access, as it were, to the operations of the divine men-
tality. Hence the epistemological pessimism we find in Hesiod, Solon,
and Xenophanes. All three authors agreed that, while the supreme di-
vinity itself is far seeing and controlling, human beings can achieve
29 Cf. Il. 15.187 189, the division of the world into three domains for Zeus, Pos-
eidon, and Hades; and Hesiod, Theog. 720, which makes heaven as high above
earth as the underworld is below earth.
no likeness to god in these respects or gain a reliable measure of in-

tellectual control over their own lives.
Heraclitus, by contrast, thought that human beings not only could
but should acquire insight into nature, as he calls it; for, without that in-
sight, they could not live wakeful, authentic and intelligent lives.30
Thus, probably with strong stimulus from Milesian cosmology, he set
ancient philosophy on an epistemological course quite opposite to ar-
chaic pessimism. He intuited the unifying power of structure, measure
and proportion in the worlds physical processes; took these to be in-
stantiated in the operation of divine intelligence; and, in his greatest
and most far-reaching innovation, posited human capacity to think
and speak commensurately i. e. in accordance with nature, and there-
fore rationally.
As I have written elsewhere: Heraclitus relation to the world is, in
a sense, an attempt to think the thoughts of god. This would be a very
unsafe thing to do if you were dealing with Olympian Zeus.31 Heracli-
tus Zeus, by contrast, is only equivocally the Olympian deity (B 32).
Non-equivocally, his Zeus is the deep structure of the world and the
worlds predictable and regular processes, as manifested in the suns
cycle and the alternation of day and night (B 67). As I have also pro-
posed elsewhere, we may characterize the general project of early
Greek philosophy as accounting for all thingsaccounting, not in
the sense of enumerating facts, but giving a systematic explanation
why the world is the way it appears to be.32 Heraclitus did not invent
the concept of the world as a bounded totality. Yet, he was the first
thinker, so far as we can see, to engage in second-order reflection on
the world, by which I mean reflection on what it means to account
for that totality, the universe as such.
In this paper I have investigated the conceptual constituents of his
idea of rationality, focusing on measure, proportion, and structure, I
could have explored his applications of that idea, by studying his tech-
niques for rousing his audience from their epistemological slumbers,
30 The difference between the gods and humanity, traditionally almost unbridge-
able, is for Heraclitus inessential, Hussey (above, n. 1) 103. As Hussey observes
(ibid. 104), in reference to such fragments as B 78 79, which contrast the di-
vine and the human: this is a matter of character not of nature . That human
nature is perfectly capable of achieving real understanding is shown by B 113
and B 116.
31 Long, H. & Stoicism (above, n. 1) 273.
32 Long, Finding oneself (above, n. 9) 10 13.
222 Anthony A. Long
such as his challenges to common sense distinctions between day and

night, up and down, mortal and immortal, or his polemics against rec-
ognized authorities. These techniques, no less than the material I have
discussed, were a crucial part of his contribution to rational inquiry,
but they are less directly relevant to the idea of rationality in the way
it was pursued by Plato and later philosophers.
Given how little of Heraclitus we possess in sheer number of words,
the challenge he presents to interpretation is quite remarkable. As is the
case with Parmenides, or even more so, with Heraclitus there is always
something else for a commentator to add. I have said nothing in this
paper about his cryptic and aphoristic style. For anyone who finds
these traditional marks of Heraclitus obscurity an impediment to my
main argument, I respond as follows.
Obscurity just for its own sake is the enemy of reason, and clarity is a
philosophers principal virtue. True enough. But there are no grounds, in
my opinion, for thinking that Heraclitus is ever obscure purely for the
sake of mystification. He often writes with such limpidity that his state-
ments are memorable precisely because of that. When he is quite ob-
scureas for instance in B 62, Immortal mortals, mortal immortals
I take it that the obscurity is philosophically motivated. He takes
on the role of the Delphic oracle in order to challenge his audience to
come up with their own interpretations of his remarks, and so achieve
not only critical distance from their unreflective preconceptions but
also open themselves to rethinking the conventionally exclusive distinc-
tion between mortal and immortal things.
In this latter respect, Heraclitus was the true precursor of Platos
Socrates. Both thinkers require the persons they engage with to follow
their respective k|cor wherever it leads. Again like Socrates, Heraclitus
revels in paradoxes, meaning controversions of standard opinions. To
his great credit, he does not claim to know much in particular detail
concerning cosmology, as some Presocratics will do in ways that,
though highly imaginative, are little more than hand-waving by later
scientific standards. More importantly, his riddling statements, in their
balanced structure, rhythms, and measures, are a representation of the
rationality he finds at work in the cosmos and, mutatis mutandis, in him-
self as the spokesman of its logos. 33
33 This paper began life as my contribution to the discussions of Heraclitus which

took place in the summer of 2005 as part of the Symposium Philosophiae Antiquae
Quintum, organized by Apostolos Pierris. I am most grateful for his invitation to
Bibliography
Barnes, Jonathan. The Presocratic Philosophers. 2 London and New York 1982.
Burnyeat, Myles F. Plato on why mathematics is good for the soul,
Proc.Brit.Acad. 103 (2000) 1 81.
Freeman, Kathleen. Ancilla to the Pre-socratic Philosophers London 1956.
Hussey, Edward. Heraclitus, in A. A. Long (ed.), The Cambridge Companion to
Early Greek Philosophy (Cambridge 1999) 88 112.
Irwin. Terence. Plato. Gorgias. Oxford 1979.
Long, Anthony. A. Finding oneself in Greek philosophy, Tijdschrift voor Fi-
losofie 54 (1992) 255 79.
Heraclitus and Stoicism, in id., Stoic Studies (Cambridge 1996; repr. Ber-
keley 2001) 35 57.
Eudaimonism, divinity and rationality in Greek ethics, Proc. Boston Area
Colloquium Anc. Philos. 19 (2003 [2004]) 123 43.
Naddaf, Grard. The Greek Concept of Nature. Albany 2005.
Vlastos, Gregory. Solonian justice, CP 41 (1946) 65 83. Repr. in id., Studies
in Greek Philosophy (Princeton 1995) 32 56.
Wardy, Robert. The unity of opposites in Platos Symposium, OSAP 23
(2002) 1 61.
participate, and also to Dorothea Frede for giving me the further opportunity to
present my views at the Hamburg conference on Leib und Seele in der antiken
Philosophie. I thank the conference participants for their responses and also An-
drea Nightingale and Chiara Robbiano, who sent me written comments on the
paper. The version that appears here has previously been published in D. Frede
& B. Reis (ed.), Body and Soul in Ancient Philosophy (Berlin 2009) 87 110.
8. On the physical aspect of Heraclitus psychology1
With New Appendices
Gbor Betegh
The most significant and possibly most far-reaching novelty of Heracli-

tus conception of the soul is a considerable extension of the xuw^s
role. As commentators have emphasized, in Heraclitus aphoristic utter-
ances the xuw^ emerges for the first time as an integrated center of
motor, cognitive and emotive functions.2 Heraclitus thus has a major
role in the process through which the word xuw^ acquires the broad
but fairly unified sense it has in the texts of the Classical and later peri-
ods. This novel characterization of the soul, moreover, is an integral part
of what appears to be the general thrust of Heraclitus entire philosoph-
ical project: a shift from an impersonal, objectivist description of the
physical world, as practiced by his Milesian predecessors, to an approach
in which questions about the nature of things in the world on the one
hand, and reflections on the human being who is striving to understand
the nature of things, and seeks to communicate his findings to other
1 This paper has grown out of ideas I first presented at the Symposium Philoso-
phiae Antiquae Quintum organized by Apostolos Pierris and then at the Jubilee
meeting of the Southern Association of Ancient Philosophy in Oxford. A ver-
sion was published in Phronesis 52 (2007) 3 32, which is reprinted here slightly
revised. I added a series of appendices in which I present the discussions to
which I referred to as forthcoming in the Phronesis paper. For the research
contained here I received support from the ERC_HU BETEGH09 research
grant. At the Symposium, Aryah Finkelberg read a paper (in this volume)
that overlaps with my discussion at some points, although his main interest is
in Heraclitean eschatology. I am grateful for the audiences in Ephesos and Ox-
ford. Comments by David Charles, Christopher Gill, Katalin Farkas, Aryah Fin-
kelberg, Charles Kahn, Tony Long, Stephen Menn, Malcolm Schofield, and
David Sedley were particularly helpful. I owe much to discussions with
Roman Dilcher. I would like to express my gratitude to David Sider for all
his help with the present version.
2 See especially M. Nussbaum, Xuw^ in Heraclitus, Phronesis 17 (1972) 1 16,
153 170; and M. Schofield, M. Heraclitus theory of the soul and its antece-
dents, in S. Everson (ed.), Companions to Ancient Thought. Vol. 2: Psychology
(Cambridge 1991) 13 34.
226 Gbor Betegh
human beings, on the other hand, are intimately linked. In such a com-
plex, epistemologically oriented and reflexive approach the center of
cognition will naturally acquire a place of prominence; and Heraclitus
takes xuw^ to be the center of cognition. The joint effect of these de-
velopments resulted in a new, and in a sense strikingly modern, concep-
tion of the self that has sometimes been compared to Thomas Nagels
objective self.3
If this is the most momentous feature of Heraclitus philosophical
psychology, why should we bother too much with the seemingly less
exciting physical aspect of his conception of xuw^? The answer, I main-
tain, is that Heraclitus seems to hold at the same time that self-reflection
and introspection, although important, are not sufficient to gain knowl-
edge about the soul. The soul is integrated into the physical world, and
reflections about the physical world can lead us to some crucial insights
about the nature and working of the soul. Understanding the soul is the
key to understanding human nature and the world, and the interrelation
of the two; but, on the other hand, understanding the cosmic order and
the major physical processes is vital to gaining knowledge about the
soul.4 Heraclitus thus appears to agree with the implication of Socrates
rhetorical question in the Phaedrus: Do you think it is possible to un-
derstand the nature of the soul in a worthwhile manner without under-
3 See A. A. Long, Finding oneself in Greek philosophy, Tijdschrift voor Filosofie

54 (1992) 257 279; and C. Gill, La psychologie prsocratique: quelques ques-
tions interprtatives, in P.-M. Morel and J. F. Pradeau (eds.), Les Anciens sav-
ants (Cahiers Philosophiques de Strasbourg, 12). (Strasbourg 2001) 169 190.
Because of all these considerations, it would probably be more appropriate to
translate Heraclitus xuw^ as mind. I continue to render it as soul for rea-
sons of convention.
4 The fact that these two aspects of Heraclitus philosophy cannot be severed has
rightly been emphasized by Dilcher H. 53 54, with footnote 1 containing a
criticism of Kirks attempt to insulate the cosmic fragments. Dilcher seems
to start out by treating the cosmic aspect as secondary to the psychic (cf. e. g.
p. 17), yet he eventually states that [i]t is vital for the argument that man
and cosmos illuminate each others structure by this correspondence. Therefore,
this double focus should not be imbalanced by giving priority to either side.
() Both sides are equally essential (93). I am in full agreement with these for-
mulations, even though I disagree with specific parts of Dilchers interpretation
of Heraclitus psychology, on which see below. See also the almost Heraclitean
formulations in Long (above, n. 3) 271: In disclosing truths about nature, Her-
aclitus is disclosing himself, or rather, in disclosing himself he is disclosing truths
about nature; he is discovering the knower in the knowable and the knowable
in the knower.
8. On the physical aspect of Heraclitus psychology 227
standing the nature of the whole? (Phdr. 270c1 2).5 The soul is part of
the whole, and cannot be properly understood without having a grasp
on the way it is integrated in the physical world.
In what follows, I shall approach Heraclitus views on the xuw^ from
the side of its physical and metaphysical aspects. I shall concentrate first on
the metaphysical framework which allows the souls integration into the
physical world and creates the conceptual space for its physical identifica-
tion. Second, I shall turn to the physical characterization of the soul and
to the way it is integrated into physical processes. In this part, I shall try
to tackle the vexed question whether the soul for Heraclitus is fire or air,
say something about the philosophical advantages of the option taken by
Heraclitus, and briefly discuss Sextus report of Heraclitus psychology.
Finally, I should add a methodological point. It has sometimes been
urged that because of Heraclitus mode of expression, the intentional ob-
scurity of his pronouncements and his apparent refusal to expound his
views in continuous, argumentative prose, the interpreter ought not try
to impute a clear-cut doctrine or theory to him. I would certainly
agree that the interpreter has to show the utmost sensitivity to the formal,
literary and linguistic features of the Heraclitean fragments. Yet I would
still not renounce the idea that Heraclitus by his pronouncements ex-
pressed a set of fairly worked out and specific views about the topics
he dealt with. The celebrated B 93 states that the lord whose oracle is
the one in Delphi neither declares nor conceals, but gives signals (b
%man ox t lamte?|m 1sti t 1m Dekvo?r oute k]cei oute jq}ptei !kk
sgla_mei) and this statement has customarily been taken to describe Her-
aclitus own mode of expression as well. But the implication of the
apophthegm is neither that Apollo does not have a clear and specific
view on the matter at hand, nor that the recipient should renounce dis-
covering what the lord of the oracle thinks about the mattereven if one
cannot hope to arrive at a definitive answer and dissipate all ambiguity.
B 36 is the central fragment for locating the souls place in physical proc-
esses:
5 Xuw/r owm v}sim !n_yr k|cou jatamo/sai oUei dumatm eWmai %meu t/r toO fkou
v}seyr;
228 Gbor Betegh
xuw0sim h\mator vdyq cem]shai,

vdati d h\mator c/m cem]shai,
1j c/r d vdyq c_metai,
1n vdator d xuw^.
For souls it is to die to become water,
for water it is to die to become earth,
from earth water is being born,
from water soul.
The first remarkable point about this fragment is that soul appears here
together with two main elemental masses, water and earth, in such a
way that soul seems to be treated on a par with the two physical stuffs.
This observation may appear obvious, but will receive some refinement
later. The statement links the three terms, soul, water, and earth, by
specifying the way they transform into each other so that the series of
changes creates a cycle. That the focus is on the changes, and not merely
on the states, becomes especially evident when we realize that the word
h\mator refers not to the state of being dead, but to the event of dying.
Apart from philological considerations,6 this rendering is strongly rec-
ommended in this case by the fact that Heraclitus explains h\mator by
the verb cem]shai. I try to indicate this fact by translating h\mator by
the infinitive to die, instead of the more usual noun death. The
translation is being born in the latter part of the fragment tries to cap-
ture the continuous aspect of the present indicative c_metai and the con-
trast with the aorist cem]shai.
The use of the terms to die and is being born to describe the
transformations between soul and water and water and earth is doubly
provocative. First, it is challenging to apply these terms to stuffs like
water and earth that are usually considered lifeless. Second, it is possibly
even more intriguing to speak about the dying and birth of the
soul. Surely, the xuw^ has strong connections to life in the traditional,
Homeric, conceptionso much so that it is customarily called in the
literature the organ of life. Yet in the traditional conception it is
6 E. Hussey, Heraclitus on living and dying, The Monist 74 (1991) 518 519
gives a very helpful concise review of the Homeric and Hesiodic occurrences
of the word h\mator. By this overview Hussey seeks to show, to my mind suc-
cessfully, that h\mator usually refers not to a state but to a process or event, and
should hence be translated by dying rather than death. Hussey does not dis-
cuss B 36, which he thinks originally consisted only of the four words xuw0sim
h\mator vdyq cem]shai (529, n. 18), which he translates as It is death to souls
to become moist.
not the xuw^ that does the living, but the organism is thought to live as
long as the xuw^ is present and active in it. Correspondingly, it is the
organism, not the xuw^, which dies when the xuw^ departs from it.7
Similarly, it is highly unusual to speak about the birth of the xuw^ as
such, distinct from the organism it is part ofjust as we would not nor-
mally speak about the birth of either the mind or the brain as such.8
The interpreter can respond to these challenges by depleting the sig-
nificance of the words dying and being born to arrive at the innoc-
uous meanings cease to exist and come into being. The more in-
teresting alternative, however, is to accept that Heraclitus attaches
(some form of) life to stuffs that are not usually conceived as living.
When this latter option is taken, it becomes significant that the formu-
lation of the cycle is not symmetrical. Describing the way down, from
xuw^ to water, from water to earth, Heraclitus speaks about dying
into the next phase of the process, whereas on the way up, from
earth to water, from water to xuw^, he does not mention dying, only
becoming or birth (c_metai). When, on the way down, souls transform
into water, the souls die and this is how water comes into being (or is
born). When, on the way up, water transforms into soul, soul comes
into being and is born. By this transformation water ceases to bebut
I doubt that Heraclitus would describe the same transformation as the
dying of water; water dies only on the way down, when earth is
born from it, but not on the way up, when it is transformed into
soul. Dying characterizes the way down.9
Incidentally, if it is true that only things that live can die, xuw^ in
Heraclitus ceases to be the bearer of life. For water is distinct from
xuw^, yet it has to have some form of life if it is to die when earth
comes into being from it.
7 This has been excellently brought out by Nussbaum (above, n. 2). For the same
point in Plato, see F. Karfk, Die Beseelung des Kosmos: Untersuchungen zur Kos-
mologie, Seelenlehre und Theologie in Platon Phaidon und Timaios. Leipzig and Mu-
nich, 2004. For further elaboration on this point, see Appendix 1.
8 See Appendix 2.
9 B 36 has many variants and paraphrases, fourteen listed in Marcovich H. ad loc.
(his fr. 66). Among these, there are only three that speak about dying on the
way up as well: Plut. de E 392c; Marc. Ant. 4.46.1, and Max. Tyr. 41.4, all
three listed by Diels under B 76. I agree with those commentators like Marco-
vich H. ad loc., Kahn H. ad loc., and Dilcher H. 67 n. 1, who take these to be
(Stoicizing) paraphrases; contra J.-F. Pradeau, Hraclite: Fragments (Paris 2002)
284 285, 296.
230 Gbor Betegh
The wording of the fragment shows a further notable point. The

first and last words of the sentence refer to soul. This is clearly an indi-
cation of the cyclical nature of the transformations captured in the frag-
ment. There is a difference, however. In its first occurrence, the word is
in the plural, whereas in the second it is in the singular. In the following
analysis, I shall assume that this shift from plural to singular is signifi-
cant;10 I base it on the conviction that Heraclitus uses language with ut-
most care and his formulations show a remarkably high level of con-
sciousness. The shift from plural to singular shows, first of all, that
when we get to the end of the sentence we have not as yet made the
full circle of the cycle of transformations. To close the circle, we still
need to get to the plural xuwa_ from the singular xuw^. As we shall
see, much hangs on the way we account for the difference between
the plural and the singular form, and how we conceive of the process
that leads from the singular xuw^ to the plural, individuated xuwa_.
We may ask, for example, what type of process is the one that delivers
plural xuwa_ from the singular xuw^? Does it already involve a measure
of dying as the processes on the way down? Or is it primarily an event
of birth as is characteristic for the processes of the way up? Or is it a
completely different type of process?
Moreover, this difference between singular and plural calls attention
to a semantic feature that may have important metaphysical bearings.
The terms water and earth refer in the fragment to elemental mass-
es, and correspondingly function as mass terms (or non-count nouns as
linguists prefer to call them).11 They do not refer to individuated things
with definite borders but to stuffs. The plural xuwa_ is all the more sig-
nificant because it shows that Heraclitus here uses the term xuw^ as a
count nounjust as we would expect, because we are accustomed to
think about souls as individuated entities. This means that we should
qualify our preliminary contention that Heraclitus treats soul on a par
with water and earth: not quite, because the word souls refers to in-
dividuated things, whereas water and earth refer to stuffs. We can
accordingly pinpoint a difference between the first two phases of the
series of transformations. The first phase refers to the process through
which souls conceived as individuated things lose their identities and
10 See also Finkelberg in this volume.

11 Admittedly, count noun and non-count noun or mass term are our cat-
egories. Yet these linguistic categories clearly apply to Greek as well (cf. K-G
2.15 16 on the use of Stoffnamen, and the way they can be used in the plural).
cease to exist as they dissolve in an unindividuated stuff, water. The sec-

ond phase in contrast refers to a process in which some part of water is
transformed into another stuff, earth.
But what about the singular xuw^, with which the fragment ends? I
would maintain that at the end of the sentence xuw^ is used as a mass
term, much like water and earth, and not as a singular count noun stand-
ing for a class of things, souls. If it were not so, Heraclitus could have
used the plural here as well. I assume, therefore, that what is described
in the last colon of the sentence is not the way individuated souls are
born from water but rather the way some part of water is transformed
into another stuff, soul. The death of individual souls in the first
colon of the fragment may or may not correspond to the death of indi-
vidual human beings, depending on whether Heraclitus believed in the
post mortem survival of individual soulsa question on which I would
suspend judgement.12 The birth of soul in the last part of the fragment,
in contrast, does not refer to the birth of individual human beings but to
the way soul stuff is generated.13 The missing link in the cycle, the shift
from the singular xuw^ to the plural xuwa_, is the way we get individ-
uated souls from soul stuff.
The idea of using soul or mind as mass terms, and the corre-
sponding idea of conceiving soul or mind as stuffs, is quite alien for
us. Yet it was not so in archaic and later philosophical texts. As I cannot
12 For the most recent defence of the survival of individual souls, see Finkelberg in
this volume.
13 The shift from the plural to the singular must pose a problem also for interpret-
ers like Marcovich H. 362 and Dilcher H. 67 69, who take the fragment to
refer exclusively to processes within human beings. If the fragment describes
the microcosmic cycle within the living human being, it is unclear to me
why it speaks about the death of souls in the plural and then switches to soul
in the singular. Dilcher thinks that the only alternative to the microcosmic in-
terpretation is eschatological, one that ascribes to Heraclitus the view that the
soul first leaves the body, lingers around for a while and then dies into
water. It may well be that the first phase of transformation, from souls to
water, happens within the human being, but that does not oblige us to say
that all the transformations described in the fragment occur within the micro-
cosmos: the fragment speaks about these transformations in general terms. For
a critique of Marcovichs physiological interpretation, see Schofield (above, n.
2) 15 21. Kirk H. 341followed by Marcovich H. 361states that even if
the main thrust of the fragment would suggest that xuw^ stands for cosmic
fire, parallel to the other two cosmic masses, the plural xuwa_ rules this out. I
agree that to be the case at the beginning of the fragment, but nothing hinders
that things change with the second, singular, occurrence of the word.
232 Gbor Betegh
provide a detailed survey of the relevant texts within the limits of the
present paper, let me only mention some notable examples. Anaxagoras
clearly uses the word moOr as a mass term when he says in B 12 ad fin.
mind is all alike, both the larger and the smaller [of it] (moOr d pr
floi|r 1sti ja b le_fym ja b 1k\ttym).14 Similarly, when Diogenes
of Apollonia maintains in B 4 that human beings and the other animals
live by means of the air as they breathe. And this is for them both soul
and intelligence (%mhqypoi cq ja t %kka f_ia !mapm]omta f~ei t`
!]qi. ja toOto aqto?r ja xuw^ 1sti ja m|gsir) he treats air as a stuff
that enters the body with breathing, but which, on the other hand,
functions as soul and intelligence in us. It is more striking to see that
Socrates, in both Xenophon and Plato, entertains a similar conception
of soul and mind. In the Memorabilia, Socrates tries to corner the irreli-
gious Aristodemus by the following salvo of questions:
Do you think you have some intelligence (vq|milom),15 but there is nothing
intelligent anywhere else at all? And this, when you know that you have in
your body only a tiny measure of earth which is so huge, and a little bit of
the large extension of waters, and that your body was constructed by re-
ceiving only a minute portion of all the other things which, surely, exist
in great quantities? But as to mind (moOr), which is thus the only thing
that does not exist anywhere else, you think you managed by some
happy chance to snatch it, and you believe that those immensely great
and infinitely many masses are in such a well-ordered state due to some-
thing lacking intelligence? (1.4.8)
Just as our bodies contain some measure of the stuffs earth and water, we
also have a measure of mind in us. Much like Anaxagoras, Xenophons
Socrates treats mind as a mass term that we can have a measure of.
Platos Socrates develops a strikingly similar argument in the Philebus
to the effect that just as fire, earth and other material components of
14 So also S. Menn, Plato on God as Nous (Carbondale 1995) 26: We may begin
by recalling the undeniable fact that Anaxagoras uses nous as a mass term.
Menn argues that expressions like moOm 5weim are not Greek for have a mind
but rather for something like have reason or be reasonable, so that nous
for Anaxagoras and other authors is reason (a mass term) and not mind (a
count noun). My point is rather that Anaxagorasjust as some of his predeces-
sors, contemporaries, and successorscould use words like moOr and xuw^ as
both count-nouns and mass terms. I shall continue to translate moOr as mind
even if the English word cannot capture this double usage.
15 My translation follows the text in M. Bandini and L.-A. Dorion, Xnophon:
Mmorables. Vol. 1, Paris 2000, who, following Muretus, based on Bessarions
Latin version, delete the words 1q~ta coOm ja !povqimoOlai.
us come from, and are nourished by, the respective cosmic masses, so
our soul comes from its cosmic counterpart.16 Finally, we can think of
the Timaeus as well, where the imagery of soul stuff, its mixing, mould-
ing, and portioning by the Demiurge is particularly conspicuous.
The gist of what I have said about Anaxagoras, Diogenes of Apol-
lonia, and Socrates argument in the Memorabilia may be present already
in Anaximenes. The only relevant text is B 2: oWom B xuw^, vgs_m, B Ble-
t]qa !q owsa sucjqate? Blr, ja fkom tm j|slom pmeOla ja !q
peqi]wei (Just as our soul, which is air, keeps us together, so do the
breath and air encompass the whole world); yet the wording, authen-
ticity and interpretation of this fragment is too contested to draw any
firm conclusions.17
From the point of view of terminology, these passages are largely in-
congruent as they speak about xuw^, moOr, vq|milom, and m|gsir. What
connects them is the underlying metaphysical assumption according to
which that which is the bearer of mental functions in us is a stuff that oc-
curs also elsewhere in the world in smaller and larger quantities. Human
beings show mental functions, and live, in so far as they have a share in
that stuff. In this respect xuw^ and moOr are like other elemental constit-
uents in us, and can be used as mass terms just like fire or earth. xuw^
and moOr at the same time function as count nouns in the same texts in
referring to individuated portions of xuw^ and moOr stuff.
A possible corollary of this approach is that the stuff in question does
not need to be in a human, or animal, body to show mental functions.
More exactly, if a theorist wants to hold that this stuff shows mental func-
tions only when it is in a human (or animal) body, he needs to provide
specific reasons why this should be so.18 A further corollary is that the cos-
16 Phlb. 29a-30d, esp. 30a about the soul. The relationship between the arguments
in the Memorabilia and the Philebus, and the possibility that they go back to the
historical Socrates, are interesting questions that cannot be discussed here. See
D. Frede, Platon. 3.2. Philebos (Gttingen 1997) 215 n. 183 and D. Sedley, Cre-
ationism and Its Critics in Antiquity (Berkeley 2007) 82, 217 19.
17 The most detailed discussion of this fragment is in K. Alt, Zum Satz des Anax-
imenes ber die Seele: Untersuchung von Aetios Peq )qw_m, Hermes 101
(1973) 129 164, who concludes that it has mistakenly been attributed to Anax-
imenes. For a much more optimistic view, see, e. g., A. Laks, Soul, sensation,
and thought, in A. A. Long (ed.), The Cambridge Companion to Early Greek Phi-
losophy (Cambridge 1999), esp. 252. Schofield (above, n. 2) 23 24 provides a
balanced assessment of the fragment and compares it to Heraclitus views on the
soul. See also below, Appendix 3.
18 Aristotle makes this point explicit in De An. 1.5 411a8 23.
234 Gbor Betegh
mic mass of soul (or mind) can be assigned cosmological roles because it
can also be described as the greatest individuated portion of soul, or mind,
stuff. Indeed, from this aspect the cosmic mass of this stuff can function as
a cosmic divinity. The ambiguity between an elemental mass and a cor-
responding cosmic god is a familiar phenomenon that we can observe for
example in the cases of Oceanus and Gaia, or water and earth. The char-
acteristics of this cosmic god will be those that we assign to soul, or mind,
in a human being: control, motor and cognitive functions and so forth.
It is time to come back to Heraclitus. By analyzing B 36, I hope to
have been able to make a case for the point that the fragment starts out
by speaking about individual souls, but ends in speaking about soul con-
ceived as a stuff. A very brief historical survey has shown that such a
conception is fully in line with some representative theories of the ar-
chaic and classical periods. This constitutes the general metaphysical
framework for the physical interpretation of the Heraclitean soul.
B 36 indicates that Heraclitus agrees with those who think that soul
can transform into other stuffs, but apparently stands apart in describing
that process in terms of dying. The central questions that have
emerged from the preceding discussion are the following: (Q1) What
is xuw^ in physical terms? and (Q2) How can we conceive the shift from
soul to souls?
II
A number of fragments show that Heraclitus, in conformity with what

the metaphysical framework allows, ascribes mental functions and char-
acteristics not only to human beings, but to natural cosmic phenomena
as well. If there is a stuff that intrinsically shows mental functions, it will
show mental functions also when it is not in a human body. In B 118
Heraclitus states that A gleam of light: dry soul, the wisest and the
best aqc ngq xuw sovyt\tg ja !q_stg.19 A natural phenomen-
onthat can be a beam of sunlight or a thunderboltis identified as
soul, and described in superlative epistemological and ethical terms.
When in B 64 Heraclitus asserts that The thunderbolt steers these
things, all of them t\de p\mta oQaj_fei jeqaum|r), he ascribes to the
19 The philological arguments marshaled by J. Bollack and H. Wismann, Hraclite

ou la sparation (Paris 1972) 325 327, who show, I think conclusively, that the
subject of the sentence is aqc^ and the predicate is xuw^, qualified by the ad-
jective ngq^.
thunderbolt the governing, coordinating function that other fragments

(e. g. B 117) assign to the human soul. If we accept that the rhetorical
question in B 16 How will one escape the notice of that which
never sets? (t l dOm|m pote p_r %m tir k\hoi;) refers to cosmic
fire,20 we can see that Heraclitus also ascribes (some form of) cognition
to the cosmic mass of fire.
What is common in these fragments is that psychic functions are at-
tributed to fire or fiery things in natural phenomena. But does this entail
that xuw^ is simply to be identified with fire? The standard interpreta-
tion holds that this is so, whereas the alternative interpretation, champ-
ioned in the most developed form by Charles Kahn, maintains in con-
trast that xuw^ is, or is primarily, air.21 Note that this interpretative
choice has important bearings on my overall thesis about the metaphys-
ical framework. For if it is true that the human soul is air, while the
bearer of psychic functions on the cosmic side is fire, then it is a serious
threat to my whole story about explaining mental functions in human
beings by saying that we have a share of the stuff that has mental proper-
ties in general and also on the cosmic scale.
The question of the material identification of the soul, whether it is
fire or air, is intimately bound up with the question whether or not
Heraclitus theory of elemental masses included air as a distinct stage
in the cycle of transformations. We have seen that B 36 mentions
only three stages in the cycle: soul, water, and earth. Fragments such
as B 30 and 90, together with the entire doxographical tradition starting
with Aristotle, make it certain that fire had a place of prominence in
Heraclitus teaching, and thus clearly secure the place of fire.22 To this
we can also add B 31a: puqr tqopa pq_tom h\kassa, hak\ssgr d
t lm Flisu c/, t d Flisu pqgst^q (turnings of fire: first sea, of
sea half earth, the other half prstr).
This fragment, just like B 36, speaks about the transformations link-
ing cosmic masses and phenomena. The transformations in this case are
20 Cf., e. g., Kahn H. 274 275.

21 See Kahn H. 238 240. Kahns view has been accepted also by T. Robinson,
Heraclitus: Fragments (Toronto 1987) 104 105.
22 It is an interesting but rarely remarked fact that not all ancient sources agreed
that fire is Heraclitus principle. Cf. Cicero ND 3.35: sed omnia vestri, Balbe, sol-
ent ad igneam vim referre, Heraclitum ut opinor sequentes, quem ipsum non omnes in-
terpretantur uno modo. Aenesidemus, for example, took air to be Heraclitus prin-
ciple. Cf. S.E. M. 10.233, with discussion in R. Polito, The Sceptical Road:
Aenesidemus Appropriation of Heraclitus (Leiden 2004) 154 161.
236 Gbor Betegh
not described as dying and being born, but referred to as the turnings
(tqopa_) of fire. Moreover, accepting that pqgst^q, notoriously diffi-
cult to translate, refers to some fiery atmospheric phenomenon,23 we
have yet again a cycle in which the first and last terms are closely related
but are not quite identical. Earth is common to B 31a and 36, whereas
seawater in B 31a refers to the cosmologically most prominent part of
water mentioned in B 36.
All these fragments seem to deliver a coherent image according to
which the third element beside water and earth is fire. The word !^q
only occurs in the different variants listed under B 76, of which I
quote here only the version transmitted by Maximus Tyrius:24 f0 pOq
tm c/r h\matom ja !q f0 tm puqr h\matom, vdyq f0 tm !]qor
h\matom, c/ tm vdator (Fire lives the dying of earth, and air lives
the dying of fire; water lives the dying of air, and earth that of water).
This text, however, has customarilyand I think with good rea-
sonsbeen described as a Stoicizing paraphrase of B 36.25 Shall we con-
clude, then, that air is missing from Heraclitus system altogether?
Against such a conclusion, some powerful arguments have been raised.
To begin with, it is unclear how one could claim that water immediately
turns into fire in B 36.26 Moreover, I think that Charles Kahn is simply
right in insisting that Heraclitus cannot have offered a theory of the
natural world in which the atmosphere was omitted.27 And this is es-
pecially so, I would add, for someone writing after Anaximenes, who
23 For arguments for this view, see e. g. Kirk H. 330 331.

24 The word !^q is missing from Marcovichs index verborum. Other versions of
B 76 are in Plut. De E 392 C and Marc. Ant. 4.46.
25 So also Kahn H. 153. Note that different versions of B 76 are much less care-
fully formulated than B 36. They do not indicate the cycle by ending the frag-
ment with (a version of) the first word as B 36 does and do not indicate the
chain of transformations by always repeating the outcome of the previous proc-
ess which becomes the starting point of the next phase of transformation. In-
stead, Marcus Aurelius adds the words ja 5lpakim at the end of the fragment.
Remarkably, Plutarchs version has only three terms, fire, air and water, as he
omits earth. Note also, that if the version of Maximus Tyrius, or that of Marcus
Aurelius, is genuine, it invalidates the observations I have made above about
dying being characteristic of the way down. This is however not true of Plu-
tarchs version, which preserves this aspect of B 36.
26 One might refer to instances when a liquid, such as the oil in a lantern, is burn-
ing, and this is how fire is being born from something liquid. But then one
should just as well refer to cases when something solid, earth-like, such as wood
or coal, is burning.
27 Kahn H. 140.
made air his principle, and Xenophanes, who was pre-eminently inter-
ested in various atmospheric phenomena and clouds. Simply ignoring
what he does not agree with is, I think, wholly uncharacteristic of Her-
aclitus. It remains true, on the other hand, that the relevant fragments
consistently list only three stages in the cycle, and I do not think that
adherents of the four element interpretation of Heraclitus have been
able to address this question in a satisfactory way. The ideal solution
should be able to account for the fact that Heraclitus prefers to speak
about three terms in the processes of transformation without, however,
excluding atmospheric air. But how could that be done if the places of
fire, water, and earth have already been secured?
When we come to the more specific question whether the human
soul is air or fire, prima facie it seems more probable that it is to be iden-
tified with fire. First of all, we have just seen that fiery phenomena, such
as a gleam of light and the thunderbolt, have psychic functions. There is
also the force of analogy with other ancient theories. In all other early,
and not so early, theories, when there is a primary form of matter, the
xuw^, or the bearer of the most important psychic faculties, is identified
with that form of matter.28 This interpretation can also align some an-
cient doxographical reports on its side.29
Yet, once again, the alternative view can refer to some ancient in-
terpretations, including that of Philo of Alexandria.30 Just as important,
it can also present some powerful internal arguments supporting the
view that soul is air and not fire. First of all, there are fragments in
which we hear about wet souls, such as the soul of the drunken man
in B 117: !mq bj|tam lehush0, %cetai rp paidr !m^bou svakk|le-
mor, oqj 1pa@ym fjg ba_mei, rcqm tm xuwm 5wym (A man when he is
28 Some further arguments of less force have also been adduced. For example, G.
Vlastos, On Heraclitus, AJP 76 (1955) 364 365, referred to the observable
fact that the living body is warm whereas the corpse is cold. This, however, is
not based on any of the surviving texts of Heraclitus. KRS 204, on the other
hand, state that while Anaximenes built on the Homeric view of the xuw^
being some kind of breath, Heraclitus abandoned this idea in favour of another
popular conception of the soul, that it was made of fiery aither. I was unable to
find any clear evidence for such an assumedly popular conception prior to Her-
aclitus.
29 Theodoret. Gr. aff. cur. 5.18 Paqlem_dgr d ja ^ppasor ja Jq\jkeitor
puq~dg ta}tgm (sc. tm xuw^m) jejk^jasim.
30 Philo, De Aet. Mundi 21 quotes the first part of B 36 and then states that: xuwm
cq oQ|lemor eWmai t pmeOla tm lm !]qor tekeutm c]mesim vdator jtk. Aene-
sidemus also thought that the soul is air for Heraclitus. On this point, see below.
238 Gbor Betegh
drunk is led by a beardless boy, staggering, not knowing where he steps,

having his soul wet.).
The controlling, motor functions of the drunken mans soul are se-
verely impaired because his soul is wet. The question naturally arises
how could Heraclitus maintain both that the xuw^ is fire and that it
can get wet? Answering this objection by claiming that in such cases
part of the soul dies does not quite convince.31 For B 117 speaks not
about someone who has a reduced-size fiery xuw^, because part of his
soul has died, but someone who has a wet xuw^. Replace xuw^ with
its assumed physical description, fire, and we are back to the absurd
idea of wet fire. Consider also B 77: xuw0si t]qxim l h\matom
rcq0si cem]shai (for souls it is pleasure, not dying, to become moist).
This fragment, preserved by Porphyry quoting Numenius, has often
been considered spurious or corrupt. Marcovich, for example, treated it
only as a reaction to B 36. Many others have retained it, but offered var-
ious emendations. So some have bracketed l h\matom as an interpolat-
ed gloss.32 Another option, offered by Diels and followed by numerous
scholars, is to emend the transmitted l^ to E, with the result that becom-
ing wet is pleasure or death for souls. The motivation for emending this
otherwise well transmitted and intelligible text is clear: editors and in-
terpreters have tried to harmonize B 77 with B 36 on the assumption
that becoming wet is death for souls. But as Jaap Mansfeld has pointed
out in his defense of the transmitted wording of B 77, rcq0si cem]shai,
becoming wet, is not the same as vdyq cem]shai, becoming
water.33 And the same point applies also to the drunken mans soul
in B 117. When the soul becomes moist or wet, it approaches water
without becoming water as yet; it approaches dying without dying as
yet. Thus B 77 in its transmitted form is perfectly compatible with
both B 36 and 117but only if soul is something that can become
wet, which fire surely cannot do.
The discussion in this section adds two further questions to the list
started at the end of the previous section: (Q3) How can we find a place for
air in a system that apparently counts with only three major cosmic masses? and
31 See Schofield (above, n. 2) 29: When the soul becomes moist or wet its fire is
put outit dies (fr. 36) either entirely or in part, as in drunkenness, fr. 117, or
in sleep, fr. 26. On the interpretation of xuw^ as fire Heraclitus does not need to
be seen as committed to the absurd concept of wet fire.
32 So, e. g., O. Gigon, Untersuchungen zu Heraklit (Leipzig 1935) 106.
33 J. Mansfeld, Heraclitus on the psychology and physiology of sleep and on riv-
ers, Mnemosyne 20 (1967) 9.
(Q4) How could we account for the fact that some manifestations of xuw^ for
Heraclitus are fiery whereas others are wet?
III
When we contrast the terms of the modern discussion about the phys-
ical identification of the Heraclitean soul with the earliest and most de-
tailed ancient testimonies, one point becomes striking. The earliest ac-
counts do not identify the Heraclitean xuw^ with either air or fire, but
say that soul, for Heraclitus, is exhalation (!mahul_asir).34 And they do
so even when one would really expect them to identify soul as one of
the elements. This is a conspicuous feature of Aristotles account of
the physical characterization of the Heraclitean soul in De Anima 1.2:
Diogenes, like some others, [held the soul to be] air because he believed it
to be the most finely grained and the principle; this is the reason why the
soul knows and moves, in so far as it is primary, and the other things are
from this, it knows; in so far as it is the most finely grained, it originates
movement. Heraclitus says too that the principle is soul, since it is the ex-
halation from which the others are composed. (De An. 1.2 405a21 6)
Aristotle is entirely happy in this context to say that for Diogenes the soul
is air, just as for Democritus spherical atoms, for Hippo water, for Empe-
docles the organ of cognition is a mixture of the four elements. He even
says that for Heraclitus, just as for Diogenes, xuw^ is the principlewhat
he does not say is that it is fire, even though elsewhere (e. g. in Meta. 1.2
984a7 = Her. A 5 DK) he identifies Heraclitus principle as fire. On the
other hand, he does not say that soul for Heraclitus is air, even though it
would have been quite an obvious move considering that he is just point-
ing out the strong parallels between Heraclitus and Diogenes on the ques-
tion of the material identification of xuw^. Aristotle could jolly well iden-
tify Heraclitus soul with either air or fire in this context, but he did not
do so. Instead, he said that soul for Heraclitus is exhalation.35
34 Arist. De An. 1.2 405a24; Aet. 4.3.12.; Cleanth. apud Ar. Did. fr. 39.2 Diels =
Her. B 12.
35 This feature of Aristotles testimony has clearly caused some uneasiness for Kirk.
After having quoted Aristotle, he continues by saying that This occurs in a con-
text in which Aristotle is anxious to find a common term for soul and !qw^; by
!mahul_asir here he means a kind of fire, but has deliberately chosen a vague
term for it (Kirk H. 275). But why would Aristotle prefer to use a vague
term in this context, where fire or air would clearly better suit his purposes?
240 Gbor Betegh
At this point, I need to discuss very briefly what Aristotle may have
meant when he stated that soul for Heraclitus is exhalation. Exhalation is
a central concept in ancient meteorology in general, but is particularly
important in Aristotles theory. Indeed, Aristotle explained most atmos-
pheric and meteorological phenomena by taking exhalations to be the
material cause of these phenomena.36 Aristotle famously distinguished
between two kinds of exhalation, dry and moist, both starting from
the earth.37 Dry exhalation arises from the earth heated by the sun
and reaches up to the fiery sphere in the sky, situated between the
sphere of air and that of aither. Moist exhalation, in contrast, arises
from the water in and on the earth and, being heavier, rises only up
to the sphere of air (Mete. 1.3). On this view, then, exhalations extend
from the level immediately above the earth, through the layer of air, up
to the uppermost fiery region which is contiguous with the divine aith-
er. Moreover, different fiery phenomena in the sky, such as shooting
stars, lightning, torches (dako_), and goats (aWcer), are explained
as ignition of dry exhalations.
Clearly, not all of Aristotles own theory of exhalations can be ascri-
bed to Heraclitus. To begin with, Aristotles insistence that exhalations
arise from earth is in obvious contrast with what we have seen in B 36: if
Heraclitus xuw^ is indeed exhalation, this exhalation must arise from
water and not from earth. B 31a, quoted above, also implies that trans-
formations to and from earth must go through a liquid state of matter,
water or sea, and that the fiery atmospheric phenomenon of pqgst^q
again comes from water/sea and not earth.
A further complication comes from Diogenes Laertius doxographical
report in 9.9 11, going back to Theophrastus, according to which Her-
aclitus had not only a bright exhalation arising from the sea and nourish-
ing the sun, but also a dark exhalation coming from the earth so that this
dark exhalation produces the darkness of night and winter. But as Kirk
(1954) 270 6 has, to my mind, convincingly shown, the ascription of
an earthly exhalation to Heraclitus is the result of some confusion with
the Aristotelian theory.38 The main objection to Diogenes report, I
36 For the most recent comprehensive discussion, see L. Taub, Ancient Meteorology
(London 2003) 88 92.
37 There is a discussion in the literature whether both exhalations are hot, or the
moist exhalation is cold, but this point does not need to concern us here.
38 We encounter a further level of confusion in Atius 2.17.4 = Her. A 11 DK,
who says that it is the exhalation coming from the earth that nourishes the
think, is the same as above: the fragments that speak about the transfor-
mations between elemental masses show that earth only transforms into
water and then water transforms into further fiery atmospheric phenom-
ena and xuw^. Diogenes testimony, on the other hand, contains an addi-
tional piece of information, not mentioned by Aristotle. Diogenes says
that the celestial bodies are bowls (sj\vai) in which some part of the ex-
halation coming from the sea is burning.39 This doctrine corresponds to
the idea that exhalations extend from the mist lingering on the surface
of waters to the fiery phenomena in the upper sky.
Retaining the core of Aristotles testimony, but stripping off what is
incompatible with what we find in Heraclitus original fragments, we
get that xuw^ for Heraclitus is the exhalation that rises from water ex-
tends to the sky and the fire of the celestial bodies.40 The first part of this
conclusion is reinforced by the apparently independent information
provided by Cleanthes, preserved by Arius Didymus, that, according
to Heraclitus, souls, too, come out of moisture by exhalations ja
xuwa d !p t_m rcq_m !mahuli_mtai, Ar. Did. fr. 39.2 Diels =
Her. B 12), whereas the connection between xuw^ and stellar matter,
is supported by the testimonies of Macrobius and Galen.41
stars, even though in 2.20.16 he says that the sun is nourished by the exhalation
arising from the sea, and in 2.28.6 he says that both the sun and the stars are
nourished by the bright exhalation coming from the sea. The idea that the
stars are nourished by dark exhalations is also contradicted by D.L. 9.10. One
may wonder whether this confusion comes from that part of the doxography
that connects the dark exhalations with night.
39 Kirk H. 269 279 and Kahn H. 291 293 come to opposite conclusions about
the different elements of the Theophrastean doxography. For Kirk, the doctrine
of celestial bowls is a secure piece of information, because this is such an un-
usual idea, so different from the sort of thing which would be invented by The-
ophrastus (270), whereas in Kahn view one must suspend judgment on the
authenticity of the doctrine of sj\vai. Kirk, on the other hand, tries to refute
the double exhalations, whereas Kahn says that this is what seems most authen-
tic in Theophrastus report (H. 293). As should have become clear, on the as-
sessment of this part of the doxography in D.L., I tend to side with Kirk.
40 In a way, the line of interpretation I am taking is closest to the one proposed by
Robert B. English in a paper published in 1913: Heraclitus and the soul,
TAPA 44 (1913) 163 184. Englishs paper, as far as I am aware, is completely
ignored in more recent discussions.
41 Macr. S. Scip. 14.9 = Her. A15: (animam) scintillam stellaris essentiae. Galen 4.786
Khn, not in DK. For the possible eschatological implications of this doctrine,
see Finkelberg in this volume.
242 Gbor Betegh
If we accept that Aristotle, our earliest doxographical source on Her-

aclitus psychology, was basically correct in saying that xuw^ is exhalation
for Heraclitus, we get clear and coherent answers to three of the questions
I have formulated above. Let us start with (Q1): What is xuw^ in physical
terms? The xuw^ is all states of matter covered by exhalations from the
lowest level of atmospheric air to the uppermost layer of celestial fire.
This general answer in turn effortlessly offers solutions to the two puzzles
raised at the end of section 2: (Q3) How can we find a place for air in a
system that apparently counts with only three major cosmic masses? If we
accept that xuw^, one of the three terms of the cyclical transformation de-
scribed in B 36, covers the whole spectrum of matter covered by exha-
lation, then this term will include celestial fire, all kinds of fiery phenom-
ena and atmospheric air. On the interpretation that follows Aristotles
testimony, it is correct both to say that Heraclitus distinguished three
main cosmic masses and to accept that he did not ignore atmospheric
air. Moreover, because the phenomenon of exhalation produces a gradual
shift in the different physical properties of matter (from wet to dry, from
cold to hot, from dark to bright), the entire cosmic region from the low-
est part of atmospheric air to heavenly fire forms a continuum. That there
is no such clear borderline between air and fire as there is between air and
water or water and earth can receive some empirical justification. In the
language of B 36, it means that the continuum extending over the range
of exhalation is not punctuated by death and birth, whereas in the
language of B 31a, we can say that the gradual shift from fire to atmos-
pheric air does not involve a turning. Of course, a more refined analysis
can distinguish a variety of different states of matter and phenomena with-
in this spectrum: mist is not the same as the thunderbolt, a cloud is not
pqgst^q or aqc^, and so forth.
Note also that it is not specific to xuw^, understood as exhalation,
that it designates one basic mass of matter within which different parts
can be distinguished. The cosmic mass of water comprises brine, the
fresh waters of springs, rivers and lakes, while the cosmic mass of
earth can be analyzed into sand, soil, rock and so forth. This point
makes it easier to understand the apparent inconsistency between the
terms of B 36 and 31a. Earth is common to both, but in the case of
the remaining two terms B 36 speaks about the more general, compre-
hensive cosmic masses (xuw^ and water), whereas B 31a mentions only
the cosmologically, or cosmogonically, more important parts of these
masses (fire and sea).
The above answer to (Q3) immediately provides a solution to (Q4)

as well: How could we account for the fact that some manifestations of soul for
Heraclitus are fiery, whereas others are wet? We have just seen that within
the range of matter covered by exhalation, there are misty, foggy, wet
parts, but drier, brighter and fierier parts as well. Indeed, it is the
basic idea behind the concept of exhalation that it connects qualitatively
different states of matter located in the atmosphere and the heavenly re-
gion. Note, again, that showing contrary properties is not particular to
xuw^. The mass of earth also covers a wide range of states that can man-
ifest opposite physical properties from the dry, soft, bright and barren
sand through the moist, dense, dark and fertile soil to dry and hard
rocks. Different forms of water also show contrary absolute and relation-
al properties; just to mention one, salty seawater is lethal for human be-
ings, but the fresh water of springs is healthy.
Incidentally, these considerations should remind us of the fact that
the terms of the transformations described in B 36 are not the elements
as later philosophers in the wake of Empedocles conceived of them.
Earth and water in B 36 are not elementary forms of matter with a
fixed set of properties,42 but large masses that comprise also contrary
characteristicsand this applies to xuw^ as well.
The view that xuw^ encompasses a wide continuum of physical
states offers considerable theoretical advantages for Heraclitus philoso-
phy. First of all, it is an unmistakable feature of Heraclitus thought,
that he, more than any other Presocratic, emphasizes the differences
in the epistemological and moral states of people. On the one hand,
his diatribe against the lamentable stupidity and wretched baseness of
the many is a central element of his whole discourse. Fragments to
this effect abound and there is no need to quote them here. Yet, on
the other hand, Heraclitus clearly implies that this is not only what is
available for humans, for people would have the possibility to improve
themselves, and there are some, including of course Heraclitus himself,
who have actually attained a much higher state in becoming wise and
morally excellent. It remains true however that the perfection of the di-
vine is beyond the limitations of the human condition.
Now, as we have seen, the soul is the center of cognition and ethical
behavior. Moreover, we have also seen that the moral and intellectual
properties of the soul are correlative to its physical properties, in partic-
ular whether it is wet or dry. Maintaining that the xuw^ can manifest
42 So also KRS 204, n. 1.

244 Gbor Betegh
opposite intellectual and moral properties and maintaining at the same

time that these intellectual and moral properties are correlative to the
relevant physical properties, entail that the soul stuff must be such that
it can manifest opposite physical properties. The soul can be foolish
and wise, virtuous and wretched, because it can be both dry and wet.
So it is not the case that part of the soul dies when one debases oneself
by being stupid, drinking too much, or being wicked, as the identifica-
tion of xuw^ with fire would require. The soul in becoming wet clearly
approaches death, being on the way to water, but has not as yet died.
Nor is it the case that the soul becomes something else, a different
stuff, not-soul, when it reaches an excellent, divine state, as the identi-
fication of xuw^ with air would require.43 The most fiery, and hence
best state of the soul is reserved for the cosmic fire, and for the fire burn-
ing in the heavenly bowls. But these forms of fiery xuw^ show the same
type of mental, intellectual, ethical properties as the more airy human
xuw^ doesonly on a much higher level. This connection was clearly
seen by Galen when he writes:
For those who hold that the soul is the form of the body will be capable to
maintain that it is not dryness, but the equality of mixture, that makes it
more intelligent; in this they differ from those who hold that as much as
the mixture becomes dryer, so much will also the soul become more intel-
ligent. But should we not agree with Heraclitus followers that dryness is
the cause of intelligence? For also he himself said: A gleam of light
(aqc^): dry soul, the wisest,44 again esteeming that dryness is the cause
of intelligence. Indeed, one must hold this superior doctrine, considering
the fact that the stars, which are luminous (aqcoeide?r) and dry, possess
the highest form of intelligence. (Galen, Quod Animi Mores Corporis Temper-
atura Sequantur 4.786 Khn; following Mllers Teubner text, I omit the
words t cq t/r aqc/r emola toOt 1mde_jmutai marked.)
43 So I cannot agree with Kahn H. 251 when he says that What he [sc. Heracli-
tus] probably meant, therefore, but what would be difficult to say (since psych
means life) is that the passage of the psych into celestial fire might be both the
death of psych and at the same time its attainment of the highest form of life.
44 Galen omits the last word of the fragment (ja !q_stg) presumably because he
concentrates on the cognitive capacities stemming from dryness.
IV
There is only one question from our list that has not as yet received an
answer: (Q2) How can we conceive the shift from soul to souls? The gist of
the answer, I would maintain, may come from a doxographical report
by Sextus. The first part of the most elaborate ancient discussion of Her-
aclitus psychology in M. 7.127 30 runs as follows:
(127) But what this [Heraclitus common and divine reason] is must be ex-
plained concisely. It is a favourite tenet of our Physicist that that which sur-
rounds us is rational and intelligent. (128) [Sextus lists poets who assumedly
held the same doctrine before Heraclitus] (129) According to Heraclitus it
is by drawing in this divine reason (k|com) in respiration that we become
endowed with mind (moeqo cim|leha), and in sleep we become forgetful,
but in waking we regain our senses. For in sleep the passages of perception
are shut, and hence the mind (moOr) in us is separated from its natural unity
with the surrounding medium; the only thing preserved is the connection
through breathing, which is like a root. So when separated, our mind loses
its former power of memory. (130) But when we awake it goes out again
through the passages of perception as through so many windows, and by
contact with the surrounding medium it regains its rational power. Just
as coals that are brought near the fire undergo a change and are made in-
candescent, but die out when they are separated from it, just so does the
portion of the surrounding medium which resides as a stranger in our bod-
ies become nearly irrational (%kocor) as a result of separation, but by the
natural union through the multitude of passages it attains a condition
which is like in kind to the whole. (trans. based on Polito and Kahn)
It has usually been maintained that this testimony contains practically
nothing that we could attribute to Heraclitus. Kahn H. 293 6 discusses
the text in an Appendix and concludes that the heavily Stoicizing inter-
pretation preserved by Sextus is without any authority for the modern
interpretation of Heraclitus (p. 269). In the most recent extensive
treatment of this passage, Roberto Polito,45 on the other hand, lists
powerful arguments, some of which go back to Diels, against the idea
that the Heraclitus doxography in Sextus is of Stoic origin. Admitting
that the idea of a cosmic Reason (logos) is indeed Stoic, he points out
that there is no reason why a Stoic would interpret this cosmic Reason
to be air or attribute such an idea to Heraclitus. Moreover, it is not a
Stoic idea that we partake in the rational principle by breathing in the
atmospheric air that surrounds us. Finally, the view that in sleep we
are separated from Reason and our minds are inactive is un-Stoic,
45 Polito (above, n. 22) 149 172.

246 Gbor Betegh
and indeed there is evidence that the Stoics had a different interpretation
of the state of sleep in Heraclitus.46 Polito then argues forcefully that the
source of Sextus is Aenesidemus interpretation of Heraclitus. Polito
seeks to explain the admittedly Stoic elements in the text (such as divine
Reason) by suggesting that Aenesidemus himself was working not on
Heraclitus book, but on an authoritative, systematic, word-by-word
commentary on Heraclitus, for which he thinks Cleanthes exegesis of
Heraclitus is the most plausible candidate.47
Now what is it that we can still attribute to Heraclitus from this ac-
count? In particular, how shall we assess the role assigned to breathing?
Is there anything Heraclitean in it? Although Polito provides a detailed
analysis of the Hellenistic medical pedigree of the breathed-in-soul doc-
trine, and the reasons why Aenesidemus espoused it, he tentatively
leaves open the possibility that it was present in Heraclitus as well.
He, however, points out that it is somewhat problematic to square
this with the idea that soul for Heraclitus is fire or fiery.48 But this is
not any longer a problem when we accept that everything from the
lowest part of atmosphere to the heavenly fire counts as soul for Hera-
clitus. And we can know from Diogenes of Apollonia B 4 and Aristotle
that there were pre-Hellenistic versions of the idea of a breathed-in-
soul.49 The doctrine of breathed-insoul, together with the idea that dif-
ferent layers of exhalation show different levels of intellectual powers in
connection with the differences in the relevant physical properties, can
explain the description of sleep in Sextus as well. Because in sleep we are
only in connection with the lowest layers of air through breathing, our
souls are only nourished by the least intelligent part of external soul stuff.
It is sufficient to keep us alive, and to maintain our souls in a stand-by
position, as it were. Note also that when the text says that through
breathing we become moeqo_, it cannot mean that we become intelligent
in the strong sense of the word. As the rest of the passage makes it clear,
to become intelligent it is not enough to keep breathing. I thus agree
with Polito (above, n. 22) 153 that the expression must mean something
like breathing supplies us with psychic matter or endowed with mind;
in other words breathing is a necessary but not a sufficient condition for
46 Calcidius, Comm. in Plat. Tim. ch. 251, p. 260 W.

47 Polito (above, n. 22) 167.
48 Ibid. 146 148.
49 Arist. De An. 1.5 411a17 21. The doctrine is attributed specifically to the Or-
phic writings in 410b27 30.
becoming intelligent. This may, by the way, apply also to Cleanthes

testimony preserved by Arius Didymus fr. 39.2 Diels = Her. B 12, ac-
cording to which aR xuwa !mahuli~lemai moeqa !e c_momtai. The
meaning cannot be that the souls continuously (!e_) become intelligent.
It is the prerogative only of some people to become truly intelligent, and
breathing will clearly not suffice to attain that state. For a more intelli-
gent functioning of the soul we need to be in contact through the senses
with what surrounds us and what is permeated and physically trans-
formed by the sun in daylight. It is only through perception, and pri-
marily through sight, that we become in contact with the bright and
the fiery.50
The same passage in Sextus also tells us that the soul or mind in us is
separated from its natural unity with the surrounding medium and
that it is a portion of the surrounding medium which resides as a
stranger in our bodies. This, I think, provides us an answer to the re-
maining question (Q2) How can we conceive the shift from soul to souls? The
move from the singular soul to the plural souls occurs when part of the
continuous but layered soul surrounding us and extending to the sky
gets drawn in and trapped in our bodies through breathing. Moreover,
the passage in Sextus also implies that this process can be conceived as a
partial dying of the soul. This interpretation, as Aryeh Finkelberg shows,
is confirmed by a passage in PH 3.230 where Sextus explicitly attributes
to Heraclitus the view that our souls are dead and buried in us. Philo
who, as we have seen above, is one of those ancient interpreters who
thought that the Heraclitean soul is airy, attributes a very similar view
to Heraclitus. Philo in his interpretation of Heraclitus B 62 says that,
according to Heraclitus when we live the soul is dead and buried in
the body as a tomb (Leg. Alleg. 1.108.2 5).51 According to this inter-
pretation, then, the shift from the common, universal, singular soul to
the particular, individual, private souls is already part of the way
down and is characterized by dying. It seems to me that even if the
wording and some details of Sextus testimony in M. 7.127 30 show
the traces of Stoic and Aenesidemean interpretation, the basic idea
and the general thrust are entirely congenial to Heraclitus and are per-
50 The exceedingly enigmatic B 26 may support such an interpretation. Note also

the use of sb]mmumtai, central to B 26, in Sextus text.
51 Finkelberg in this volume argues that the famous sma-sma doctrine in Platos
Gorgias 492d-493a is also Heraclitean.
248 Gbor Betegh
fectly compatible with what I have tried to reconstruct from other sour-
ces.
Let me wrap up now with some methodological and terminological

considerations on the question whether or not, and if so with what qual-
ifications, we can use technical terms coming from contemporary phi-
losophy of mind in our interpretation of the Presocratics in general and
of Heraclitus in particular.
Although everyone seems to agree that the Presocratics thought that
the bearer (or bearers) of mental functions are bodies of some sort that
have physical properties, it has sometimes been questioned whether it is
appropriate to describe the Presocratics, and Heraclitus in particular, as
materialists or physicalists.52 Some commentators have maintained that it
is not quite legitimate to apply these terms because the view that the
bearers of mental functions are bodies showing physical properties re-
mained uncontested all through the period. On this view, the dualist,
i. e. Platonist, alternative has to be already on the table so that physical-
ism may emerge as an option consciously chosen and vindicated against
a rival theoretical position.53 I would certainly agree that the formulation
of the dualist position had an important role in the theoretical elabora-
tion of the physicalist position, but I do not think that the presence of
dualism as a theoretically elaborated alternative is necessary for charac-
terizing a view as physicalist. If that were the case, one could just as
well claim that it is illegitimate to label ancient ethical theories eudai-
monist, because the eudaimonist assumptions remained uncontested
and there was no theoretically elaborated alternative on offer. More-
over, the appearance of the Platonic position in itself shows that dualism
was an available conceptual possibility, and when Plato formulates his
52 I use the somewhat awkward formulation bearer (or bearers) of mental func-
tions because different Presocratics use different words (such as xuw^, moOr,
m|gsir, and vq^m) to describe this entity, and may distribute different psychic
and cognitive functions among more than one of these. Materialism and physi-
calism is sometimes used interchangeably. For some authors, materialism is the
broader term because it applies also to some pre-theoretical views (see below).
Physicalism, by contrast, is broader in the sense that it accommodates forces etc.
which are physical but not necessarily material.
53 Cf. J. Barnes, The Presocratic Philosophers (London 1982) 475; and Gill (above,
n. 3) 170.
dualism, he formulates it in opposition to the philosophical views of his

predecessors and contemporaries. All in all, I would stick to the idea
that, quite simply, a conception or theory can legitimately be called
physicalist if it satisfies our definition of physicalism. So the question
is whether or not the views we can assign to the Presocratics, such as
that the bearer of mental functions is a physical entity, satisfy our fav-
oured definition of physicalism.
Another, connected, worry has been that only a sufficiently explicit,
theoretically elaborated position can appropriately be described as physi-
calist, while the pronouncements of the Presocratics had not as yet
reached that level of explicitness and theoretical elaboration. As to the
first component of this objection, I would agree with Terence E. Horgan,
who in his entry on physicalism in Blackwells Companion to the Philos-
ophy of Mind has emphasized the importance of distinguishing between
(i) certain vague, partially inchoate, pre-theoretic ideas and beliefs
about the matter at hand; and (ii) certain more precise, more explicit,
doctrines or theses that are taken to articulate or explicate those pre-the-
oretic ideas and beliefs.54 In making this distinction Horgan does not
question that (i) can be called physicalist or materialist.55 So even if the
views of the Presocratics, and especially what is available to us from
them, are certainly lacking in explicitness and theoretical elaboration, I
do not think that anyone can deny that they satisfy at least (i).
Things get of course immensely more complicated once we try to
be more specific, because, first, there is no non-contested general defi-
nition of physicalism and, second, on any account, the term covers a be-
wilderingly great variety of different theories and conceptions. It is thus
desirable to specify, first, what we mean by physicalism in the case of
the Presocratics in general and, second, more important, how to char-
acterize the individual Presocratic texts in this respect.
Let me now try to specify what I take to be the most important dif-
ference between Presocratic and modern theories. The question that
dominates modern discussions in philosophy of mind is whether or
not, and if so how, mental functions and phenomena are related to ana-
tomically recognized parts and physical constituents of our bodies. The
ancient texts I have referred to approach this question from a different
54 T. Horgan, Physicalism, in S. Guttenplan (ed.), A Companion to Philosophy of

Mind, (London 1994) 471.
55 With an avowedly arbitrary terminological decision, Horgan then reserves ma-
terialism for (i) and uses physicalism for (ii).
250 Gbor Betegh
angle. They start from the assumption that there is a stuff in the world
that is responsible for mental functions and life, and that we show men-
tal functions and live because this stuff, along with other stuffs, is present
in us.56 The ancient authors then formulate alternative hypotheses about
what this stuff is and how it works in human beings and in the world at
large. Some authors, e. g., Diogenes of Apollonia and probably Anaxi-
menes, identify this stuff with an elemental mass, whereas other authors
such as Anaxagoras and the Socrates of the Memorabilia and Philebus,
treat mind and soul as stuffs apart. A connected point of contention is
whether or not the stuff responsible for vital and mental functions par-
ticipates in the transformations among different stuffs. Diogenes of
Apollonia and Anaximenes think that it does, whereas Anaxagoras, Soc-
rates and Platos Timaeus hold that it is insulated from transformations
among other stuffs. The central question of the ancients is not how a
physical stuff can under certain conditions show also mental properties,
but rather they take it for granted that there is such a stuff, material or
otherwise, and that it inheres in us as well.
Heraclitus, more specifically, has repeatedly been called an identity
theorist.57 As must have been clear from what I have said above, I also
agree that, for Heraclitus, ones moral and intellectual condition is iden-
tical with the physical state of ones soul. But due caution, and some
qualification, are in order here as well. Modern type-identity theorists
are generally interested in the relationship between the mental proper-
ties of intentional states and sensations, such as having a certain thought
or having a certain sensation, and the corresponding bodily states. Her-
aclitus, in contrast, seems to be much more concerned with more or less
standing conditions or dispositions such as being wise, being morally ex-
cellent and so forth, which can be expressed on the dry-wet axis. He
certainly treated drunkenness in this context, but I wonder if he went
any further and analyzed particular mental states, such as having a par-
ticular thought, desire or sensation in terms of the physical properties
of the soul. It may be significant in this respect that Theophrastus has
56 As Tim Crane has reminded me, the closest modern analogy may be Henri
Bergsons vitalism.
57 So, e. g., G. S. Kirk, Heraclitus and death in battle, AJP 70 (1949) 392, fol-
lowed by Kahn H. 249, quoted also by Schofield (above, n. 2) 15.
practically nothing to say about Heraclitus in his extensive discussion of

Presocratic theories of perception.58
Appendix 1
On the life and death of the soul and the elements
I have argued above that the formulation of B 36 is out of the ordinary

not only because it apparently ascribes life and death to various material
stuffs, but also because it attributes life and death to the xuw^. Usually, it
is the human being (or other kinds of organisms) who lives and dies, and
death means, according to the standard archaic conception, that the soul
departs from the body. In the rare occurrences where the xuw^ is said to
be destroyed, the word quite clearly stands for the whole person in a
synecdoche, or more probably to the life of the person, as e. g. in Ae-
schylus Agamemnon. (Aesch. Ag. 1465 6 speaking about Helen: r
l_a pokk_m j !mdq_m xuwr Dama_m ak]sas(). But even in such cases,
xuw^ is said to be destroyed and not to die. The intended meaning,
even if taken literally, is then that the life force of many people got de-
stroyed and hence many people died. Apart from Heraclitus, the earliest
text in which the xuw^ is said to die is Sophocles Antigone 559 60: s
lm f0r, B d 1l xuw p\kai j t]hmgjem seems to stand for the human
being, as has been well explained by Nussbaum.59 To what Nussbaum
and others said about dying, we might add the other side. cem]shai, be-
coming or being born, if possible, is even less easily applicable to the
xuw^. To my knowledge, no early text speaks about the birth of the soul
and, indeed, it is difficult to imagine in what context this idea could
occur. And again, the relationship between xuw^ and life is just as prob-
lematic as the ones between xuw^ and death and xuw^ and birth: it is
the human being, and not the xuw^, who lives by virtue of xuw^.
The only pre-fifth century text in which the xuw^ might be said to
live is Pi. P. 3.61 2 l^, v_ka xuw\, b_om !h\matom j speOde. But
here again, it is unclear whether the xuw^ strives after a deathless life
58 The point about Theophrastus has been made by Laks (above, n. 17) 254. The-
ophrastus only claim, i. e., that Heraclitus, together with Anaxagoras, thinks
perception to be by contraries, seems to be a possibly unwarranted inference
from the role of opposites in Heraclitus.
59 Nussbaum (above, n.2) 153 4.
252 Gbor Betegh
for itself or, which seems the more probable option, for the human
being to whom it belongs.
Pindars verse brings us in turn to the adjective !h\mator applied to
xuw^. The earliest text in which the soul is said to be !h\mator is the
famous passage in Herodotus where he describes the Egyptian and Py-
thagorean doctrine of reincarnation (2.123). When applied to the xuw^,
it is not immediately obvious what the opposite of !h\mator is, exactly
because of the problematic relationship between xuw^ and death as dis-
cussed above. It seems that in this early context the opposite of an
!h\mator xuw^ is not merely, and not even primarily a xuw^ which
dies. Rather, it is opposed to a xuw^ which, as in Homeric eschatology,
continues to exist after the death of the individual, but goes to Hades
and stays there indefinitely, and therefore ceases to be the soul of a living
being. A xuw^ might then be qualified !h\mator in so far as it continues
to be present in, and bring life to, (different) living organisms.
This body of evidence, both positive and negative, strongly suggests
that in speaking about the xuw^, death and life are not simple opposites,
and are not simply equivalent with the existence or non-existence of the
xuw^. Similarly, ascribing death or destruction to the xuw^, on the one
hand, and claiming it to be !h\mator, on the other, are not simple op-
posites and do not create a complete disjunction. This is so precisely be-
cause in the traditional, Homeric, conception the xuw^ continues to
exist, without however bringing life to a living organism.
B 36 appears thus to redefine the scope of application of the terms
death and birth, by applying them to the xuw^ and the elements. If
so, B 36 can provide a coherent interpretative context for B 62:
!h\matoi hmgto_, hmgto !h\matoi,
f_mter tm 1je_mym h\matom,
tm d 1je_mym b_om tehme_ter.
Immortal mortals, mortal immortals,
living the death of these,
and dying the life of these.
This fragment is often praised for its perfect composition, and, at the
same time is often called one of the most enigmatic of Heraclitus say-
ings.60 Without intending to dissipate all its magnificent obscurity, let
60 Cf. e. g. Marcovich H. ad loc. 240: The meaning of the saying is obscure.

Schofield (above, n. 2) 32: Clearly it is a riddle which can be solved only
by the reader who has thought his or her way through the whole Heraclitean
logos already. B 62 has received a thorough re-examination by B. Bossi, Acer-
me say that the fragment can be mapped seamlessly on the subjects of

B 36. In view of Heraclitus general physical theory, the elements can
be called immortal, in so far as there is no moment in the history of
the world in which any of these completely and definitively61 ceases
to exist.62 On the other hand, B 36 explains in which sense they are sub-
ject to death, and also in which sense one comes to life from the death of
the other. If so, the fragment, at least at one level of interpretation, is not
giving a redefinition of gods and humans and the relationship between
the two groups, but rather redraws the possible sphere of application of
the terms immortal and mortal, in close connection to the redefini-
tion of the sphere of application of the terms coming to life and
dying that we have seen in B 36.
An immediate additional confirmation for this line of interpretation
comes from the last part of B 30:
pOq !e_fyom, "pt|lemom l]tqa ja !posbemm}lemom l]tqa.
Ever-living fire, kindling in measures and quenching in measures.
The adjective !e_fyom not only ascribes life to fire, but functions as a
synonym for !h\mator. The last two clauses, by contrast, can be
taken as functional synonyms of coming to life and dying. On the
basis of these considerations, I think Maximus Tyrius is basically correct
in connecting B 62 with the road up and down of B 60, and in merging
it with the elemental cycle of B 36, to come up with the following free
paraphrase63 :
letabokm bqr syl\tym ja cem]seyr, !kkacm bd_m %my ja j\ty, jat
tm Jq\jkeitom ja awhir aw f_mtar lm tm 1je_mym b_om, !pohm^sjomtar
d tm 1je_mym fy^m. f0 pOq tm c/r h\matom, ja !q f0 tm puqr h\ma-
ca del significado del Fragmento B 62 (D.-K.) de Herclito, in E. Hlsz (ed.),

Nuevos Ensayos sobre Herclito (Mexico 2009) 285 314.
61 Note that the view thus remains compatible with attributing the theory of ek-
pyrosis to Heraclitus. The elements may go out of existence for a while, but
come back again in the next cosmic cycle.
62 Bossi (above, n. 60) lists this reading as one of the four main interpretative op-
tions, but does not discuss it. See also M. A. Santamara lvarez, La transmi-
gracin del alma de Ferecides de Siros a Pndaro, in A. Bernab et al. (eds.),
Reencarnacin: la transmigracin de las almas entre Oriente y Occidente idente (Madrid
2011) 246 8.
63 Thus I agree with those, like Kahn H. ad loc., who treats Maximus text as not
being verbatim. On why I think that the part of Maximus speaking about the
death of fire is mistaken, see below.
254 Gbor Betegh
tom vdyq f0 tm !]qor h\matom, c/ tm vdator. diadowm bqr b_ou ja

letabokm syl\tym, jaimouqc_am toO fkou.
What you see is the transformation of bodies and becomings, the exchange
of the roads up and down, according to Heraclitus; and those that again and
again are living the life of these, and die the life of these. Fire lives the death
of earth, and air lives the death of fire; water lives the death of air, and earth
that of water. What you see is the taking over of life and the transformation
of bodies, the renewal of the whole (41.4.i-k).
Understanding B 47 as speaking about the terms of the cosmic processes
is a distinctly uncomplicated interpretation as compared to the ones that
understand it as speaking about gods and humans. Yet it is not simplistic,
because, as I have argued, it involves a radical reinterpretation of the
possible references of immortal and mortal, life and death, closely
connected to what we have seen in B 36 and 20. In this sense, Heraclitus
appears to be an important antecedent for the way Empedocles speaks of
the elements as immortal mortals for example in B 26:
in this respect they come to be and have no constant life, but insofar as
they never cease from constantly interchanging, in this respect they are al-
ways unchanged in a cycle. (Trans. Inwood)
And in the same context, we could also mention Aristotles GC 2.10 in
which the elements come-to-be and pass-away to approximate an eter-
nal life through their continuous cyclical transformations.
I should like to stress that the fact that B 47 can receive a satisfactory
and coherent interpretation when connected to B 36 and 20 does in no
way mean that it cannot have further levels of signification. We are, after
all, speaking about a Heraclitean fragment.
Appendix 2
Anaximenes and Heraclitus on the soul
There is one important fragment from the period before Heraclitus that
appears immediately relevant to the physical aspect of Heraclitus theory
of xuw^. This is Anaximenes B 2:
oWom B xuw^ B Blet]qa !q owsa sucjqate? Blr, ja fkom tm j|slom
pmeOla ja !q peqi]wei.
Just as the xuw^, which being our air, keeps us together, so do the breath
and air encompass the whole world.
The authenticity and interpretation of this fragment is hotly debated.

The text, preserved by Atius in his inventory of !qwa_, has no parallel,
and no other source tells us anything on Anaximenes views on the soul
apart from a remark by Philoponus. In commenting on Aristotles con-
tention in De Anima 1.2 405a21 2 that Diogenes and some others (ja
2teqo_ timer) took air to be the !qw^ and the soul, Anaximenes, says Phil-
oponus, is a good candidate for being one of the unnamed others. Ad-
mittedly, this is not very compelling. Indeed, numerous interpreters
have doubted whether Anaximenes said anything at all about the
soul.64 Moreover, almost all agree that the wording cannot be original,
mainly on account of the use of the terms sucjqate? and j|slom. Be-
cause of these reasons, many have claimed that the text preserved by A-
tius has nothing to do with Anaximenes. Earlier, Reinhardt and others
had maintained that the view expressed in B 2 is sheer Stoicism.65 Karin
Alt, who has offered the most detailed historical analysis of the fragment,
has successfully disproved this thesis. Ultimately, however, she also
thinks that the attribution to Anaximenes is unfounded and suggests,
not very convincingly to my mind, that it is merely the result of confus-
ing Anaximenes with Diogenes of Apollonia.66 All in all, I still find Vlas-
tos formulation the most judicious: though much of the wording of
this fragment is doubtful, there is no good reason to doubt that it para-
phrases an analogy by Anaximenes himself; Vlastos (1993) 147.67
The basic purport of the fragment is relatively clear and uncontest-
ed. It states (1) that the xuw^ in us is a stuff that exists also in the world at
large, and (2) that this stuff is responsible for both psychic and cosmic
functions. This much is already sufficient to establish an important con-
nection between Anaximenes and Heraclitus; all the more so as Anax-
imenes seems to be the earliest formulation of this idea. It is fair to as-
sume, moreover, and as far as I am aware no one denies it, that
Heraclitus knew Anaximenes doctrines. If so, it is a reasonable conjec-
64 G. Whrle, Anaximenes aus Milet (Stuttgart 1993) 15: Ob sich also Anaximenes
tatschlich ber die Seele in irgendeiner Weise geuert hat, lt sich kaum mit
Sicherheit sagen.
65 K. Reinhardt, Kosmos und Sympathie (Munich 1926) 213, followed by, among
others, J. Kerschensteiner, Kosmos (Munich 1962) 83.
66 Alt (above, n. 17) 157 64.
67 Vlastos in the same paper also emphasizes the importance of Anaximenes B 2
for the understanding of Heraclitus, although I derive different conclusions
from this evidence.
256 Gbor Betegh
ture that Heraclitus own views are to some extent dependent on, and
contain a critical reaction to, Anaximenes theory.
Let us see then in what way Heraclitus reacted to Anaximenes in
formulating his own views on the soul, starting with the general char-
acteristics of the two doctrines and their place in the respective physical
theories they form part of. First a point of contact: in accordance with
the general tendencies of Ionian physics, both Anaximenes and Heracli-
tus allowed that different forms of matter turn into each other. But then
comes the contrast in the description of the transformation: condensa-
tion and rarefaction for Anaximenes, cooling and warming, moistening
and parching for Heraclitus (cf. B 126).
The second point of comparison concerns the role of the soul in
both human beings and the cosmos at large. Here things become tricky,
especially regarding Anaximenes. This is so because there are good rea-
sons to think that the word sucjqate? that describes the function of the
air in human beings, is not Anaximenean.68 If it still carries over some-
thing of the sense of the original, it seems that the function ascribed to
air in both roles is something like holding together.69 With respect to
Anaximenes, anything further is a matter of conjecture. Thus, one may
think that the air also governs or steers both the cosmos and the human
being.70 Yet it is an open question whether sucjqate?, or whatever
stood in its place in Anaximenes original, had anything to do with gov-
erning or steering. It seems to me just as possible that, closer to the Ho-
meric meaning of the word xuw^, what the air does is keeping the cos-
mos together and keeping it functioning. The fact that pmeOla is
emphasized on the cosmic side may support this interpretation in so
68 As many have pointed out this verb is suspicious because it is only attested
much later, first in Plutarch, and even then rather rarely. It is unclear what
may have stood in its place in Anaximenes text. But it is just as unclear to
me why one found this word appropriate to inflict on Anaximenes and thus
create a less than obvious analogy with peqi]wei.
69 The best attested meaning of sucjqat]y is to keep something together by au-
thority or force (as a general keeps together his troops in Plut. Phoc. 12), so
when one tries to find a common denominator between sucjqat]y and
peqi]wei, the best bet I think is still that the air holds together and keeps in a
functional state both the world and the human beings. The cosmos would
fall apart and stop functioning without the encompassing air, just as the
human being disintegrates and stops functioning when the soul leaves it. Hold-
ing the living being together is a distinct function of the soul also in Aristotles
discussion towards the very end of De Anima 1 at 411b7 9.
70 See the different possibilities lined out in KRS 160.
far as blast or wind is the active manifestation of air. As long as the air is
there and remains active, the human being and the cosmos remain func-
tional and do not decompose.
Some interpreters went further and claimed that Anaximenes air
must also be intelligent. Interestingly, this claim is based not on the
function of air in human beings, but on the assumption that the air
also governs Anaximenes cosmos.71 This, however, seems to me a rath-
er bold interpretative leap, for the text does not even say in so many
words that the air governs the cosmosit only says that it encompasses
it. This stronger interpretation, I suggest, reads back to Anaximenes a
move that was made by Heraclitus in reaction to Anaximenes. What
makes this move possible is precisely the novelty of Heraclitus concep-
tion of xuw^, which consists in making the xuw^ the center of both
motor and cognitive functions.72
So Heraclitus critical assessment of Anaximenes position on the
soul thus far would be something like this. Anaximenes was correct in
conceiving the soul as a portion of some stuff, which also has cosmic
functions and which can turn into other stuffs. But he is not quite
right in describing the role of this stuff in both the cosmos and the in-
dividual human being. For the main function of the xuw^ in both con-
texts is cognition and intelligence, and closely related to it, control. It
does not merely hold things together, but determines the behavior of
both the cosmos and individual human beings.
Now what about the physical identification of the soul? On the in-
terpretation I suggest, Heraclitus critical attitude towards Anaximenes
turns out to be more complex than either simple refusal or agreement.
He can agree with Anaximenes in saying that the soul, in the large ma-
jority of people, is atmospheric air. He can nevertheless point out that
Anaximenes is hubristically overoptimistic when he considers that that
form of matter which in most human beings functions as soul is the
most important, nay, divine form of matter that fullfils the central cos-
mic functions as well. It is sheer hubris to identify the souls in these peo-
ple with the highest state of soul stuff. These souls, though parcels of the
same cosmic mass, are qualitatively different from the highest form of
soulthey are wet, approaching water, whereas soul in its ideal state
is dry and fiery.
71 So e. g. Laks (above, n. 17) 252: and air was probably also responsible for what
we call thought, since it governs Anaximenes universe.
72 On this, see above pp. 225 6.
258 Gbor Betegh
Appendix 3
Heraclitus and the Orphic writings on the soul
The connection between Heraclitus and Orphism is a dauntingly com-

plex issue, with a long history in the literature, which I cannot discuss
here in any detail. Let me, however, point out that the interpretation
of Heraclitus I have offered above creates a connection between two
otherwise unconnected Orphic fragments, and in turn receives some
cumulative support from them.
Clement, our primary source for B 36, reports the fragment thus:
Orpheus composed the following verses:
Water is dying for the soul, for waters, earth.
From water earth comes to be, from earth water again,
from that soul that turns into the whole of aither.
Heraclitus, who put together his account from these verses, writes some-
thing like this:
For souls it is to die to become water,
for water it is to die to become earth,
from earth water is being born,
from water soul.73
So Clement starts by quoting the verses he attributes to Orpheus (OF
226 = 437 F Bern.) that show an uncanny resemblance to B 36.74
Then he claims that Heraclitus in composing B 36 was inspired by,
or simply plagiarized, Orpheus. In all probability, Clement here reverses
the direction of influence, and the author of the Orphic verses was
drawing on Heraclitus.75
73 Clem. Strom. 6.2.17 iqv]yr d poi^samtor

5stim vdyq xuw0, h\mator d rd\tes<s>im !loib^,
1j d vdator <lm> ca?a, t d( 1j ca_ar p\kim vdyq
1j toO d xuw fkom aQh]qa !kk\ssousa
Jq\jkeitor 1j to}tym sumist\lemor tor k|cour d] pyr cq\vei xuw0sim
h\mator vdyq cem]shai, vdati d h\mator c/m cem]shai, 1j c/r d vdyq c_me-
tai, 1n vdator d xuw^.
74 OF 226 = 437 Bern. has recently been re-examined by M. Kahle, OF 437 and
the transformation of the soul, in M. Herrero et al. (eds.), Tracing Orpheus
(Berlin 2011) 153 158.
75 See, in this sense, A. Bernab, A. Textos rficos y filosofa presocrtica: materials para
una comparacon (Madrid 2004) 61 2. I am however not entirely convinced by
Bernabs argument that the Orphic text shows Stoic influence. We dont have
From the point of view of the reconstruction I have offered above,

it is noteworthy that the Orphic formulation does not retain the distinc-
tion between the singulars and the plural of the Heraclitean original. On
the other hand, it makes explicit that in the step described in the final
verse, xuw^ joins the cosmological element that constitutes the heavenly
region. What, however, makes the Orphic recycling of Heraclitus B 36
particularly significant for my purposes is its possible connection with
another set of Orphic fragments about the soul. One of the earliest evi-
dence about Orphic conceptions of the soul comes from Aristotles De
An. 1.5 410b27 411a1 (OF 27 = 421 F Bern.76):
This is what applies also to the account contained in the so-called Orphic
poems. For it claims that the soul, carried by the winds, enters from the
universe as we breathe. This is not possible to happen in the case of plants,
nor of certain animals, given that not all of them breathe.
The view expressed here contains the gist of the doctrine Sextus attrib-
utes to Heraclitus, i. e. that the soul enters the body by breathing in the
atmospheric air. Indeed, Iamblichus interpretation of Aristotles refer-
ence comes even closer to my proposed reading of Heraclitus:
5oij] ce lm aqtr b iqver wyqr rpokalb\meim eWmai ja l_am tm xuw^m,
!v Hr pokkr lm eWmai diaiq]seir, pokkr d ja !l]sour 1pipmo_ar jah^-
jeim 1p tr leqistr xuwr !p t/r fkgr xuw/r.
At least it seems that Orpheus himself supposes that the soul is separate and
unique from which there are many portions, and many and immediate in-
halations reach these individual portions from the universal soul (Iambli-
chus de An. ap. Stob. 1.49.32.104 108)
The fragment preserved by Vettius Valens (317.19 Pingree = OF 228b
= 422 F Bern.), moreover, shows that the Aristotelian evidence is not
isolated: !]qa d( 6kjomter xuwm he_am dqep|leha, When we draw in
the air, we pluck for ourselves the divine soul. From the point of
view of our present concern, it is particularly significant that the Aristo-
the standard Stoic list of the four elements (as in some versions of B 36), and,
despite the evidence of Arius Didymus fr. 39 = SVF 2.821, the mention of the
aQh^q is not conspicuously Stoic either.
76 For a recent discussion of this text, see C. Megino, Aristteles y el Liceo ante
el orfismo, in A. Bernab & F. Casadess (eds.), Orfeo y la tradicion rfica (Ma-
drid 2008) 1281 1306; and id. Presence in Stoicism of an Orphic doctrine on
the soul quoted by Aristotle (De Anima 410b 27 = OF 421), in M. Herrero de
Juregui et al. (eds.), Tracing Orpheus (Berlin 2011) 139 146.
260 Gbor Betegh
telian evidence ensures that the author of the Orphic writing in question
adopted this view independently of any Stoic influence.
There is of course no guarantee that the Orphic rephrasing of B 36
and the Orphic doctrine reported by Aristotle ever formed parts of a co-
herent and explicit Orphic theory of the soul.77 Yet I find it remark-
able that they correspond perfectly to the respective parts of the Hera-
clitean theory as I have tried to reconstruct it. Naturally, one can come
up with different scenarios to explain this body of evidence. For in-
stance, one could speculate that first some Orphics espoused the doc-
trine of the breathed-in soul from whatever, non-Heraclitean, source,
then the Stoics mistakenly attributed the same view to Heraclitus, and
then some later Orphics accepted this Stoicizing (mis-)interpretation
of Heraclitus, and this in turn gave them the motivation to adapt Her-
aclitus B 36. Yet, I find it just as conceivable that we dont need to pre-
suppose such zigzags, and hypothesize that some relatively early Orphics
were impressed directly by the Heraclitean theory of the soul, both that
the xuw^ can be identified with the element that extends from the at-
mospheric air to the heavenly fire and that we receive shares of it by
breathing it in.
Bibliography
Alt, K. Zum Satz des Anaximenes ber die Seele: Untersuchung von Aetios
Peq )qw_m, Hermes 101 (1973) 129 164.
Bandini, M., and L.-A. Dorion, M. Xnophon Mmorables. Vol. 1. Introduction
gnrale, Livre I. Collection Bud. Paris 2000.
Barnes, J. The Presocratic Philosophers, 2nd ed. London and New York 1982.
Bernab, A. Textos rficos y filosofa presocrtica: materiales para una comparacon,
Madrid 2004.
Bollack, J., and H. Wismann. Hraclite ou la sparation. Paris 1972.
Bossi, B. Acerca del significado del Fragmento B 62 (D.-K.) de Herclito, in
E. Hlsz (ed.), Nuevos Ensayos sobre Herclito. Actas del II Symposium Heracli-
teum Mxico 2006 (Mexico 2009) 285 314.
English, R. B. Heraclitus and the soul, TAPA 44 (1913) 163 184.
Frede, D. Platon. Vol.3, part 2. Philebos. Gttingen 1997.
Gigon, O. Untersuchungen zu Heraklit. Leipzig 1935.
Gill, C. La psychologie prsocratique: quelques questions interprtatives, in
P.-M. Morel and J. F. Pradeau (eds.), Les Anciens savants (Cahiers Philoso-
phiques de Strasbourg, 12). Strasbourg (2001) 169 190.
77 Interestingly, recent treatments of the relevant Orphic fragments in Megino

(2011, above, n. 77) and Kahle (above, n. 75) do not make the connection.
Horgan, T. Physicalism, in S. Guttenplan (ed.), A Companion to Philosophy of
Mind ( London 1994) 471 479.
Hussey, E. Heraclitus on living and dying, The Monist 74 (1991) 517 530.
Kahle, M., OF 437 and the transformation of the soul, in M. Herrero de Ju-
regui et al. (eds.), Tracing Orpheus. Studies of Orphic Fragments (Berlin 2011)
153 158.
Karfk, F. Die Beseelung des Kosmos: Untersuchungen zur Kosmologie, Seelenlehre
und Theologie in Platon Phaidon und Timaios. Leipzig and Munich, 2004.
Kerschensteiner, J. Kosmos: Quellenkritische Untersuchungen zu den Vorsokratikern.
Munich 1962.
Kirk, G. S. Heraclitus and death in battle, AJP 70 (1949) 384 393.
Laks, A. Soul, sensation, and thought, in A. A. Long (ed.), The Cambridge
Companion to Early Greek Philosophy (Cambridge 1999) 250 270.
Long, A. A. Finding oneself in Greek philosophy, Tijdschrift voor Filosofie 54
(1992) 257 279.
Mansfeld, J. Heraclitus on the psychology and physiology of sleep and on riv-
ers, Mnemosyne 20 (1967) 1 29.
Megino, C., Aristteles y el Liceo ante el orfismo, in A. Bernab & F. Ca-
sadess (eds.), Orfeo y la tradicion rfica. Vol. 2 (Madrid 2008) 1281 1306.
Presence in Stoicism of an Orphic doctrine on the soul quoted by Aristo-
tle (De Anima 410b 27 = OF 421), in M. Herrero de Juregui et al. (eds.),
Tracing Orpheus: Studies of Orphic Fragments (Berlin 2011) 139 146.
Menn, S. Plato on God as Nous. Carbondale 1995.
Nussbaum, M. Xuw^ in Heraclitus, Phronesis 17 (1972) 1 16 and 153 170.
Polito, R. The Sceptical Road: Aenesidemus Appropriation of Heraclitus. Leiden
2004.
Pradeau, J.-F. Hraclite: Fragments. Paris 2002.
Reinhardt, Karl. Kosmos und Sympathie. Munich 1926.
Robinson, Thomas. Heraclitus: Fragments. Toronto 1987.
Santamara lvarez, M.A. La transmigracin del alma de Ferecides de Siros a
Pndaro, in A. Bernab et al. (eds.), Reencarnacin: la transmigracin de las
almas entre Oriente y Occidente (Madrid 2011).
Schofield, Malcolm. Heraclitus theory of the soul and its antecedents, in S.
Everson (ed.), Companions to Ancient Thought. Vol. 2: Psychology (Cambridge
1991) 13 34.
Sedley, David. Creationism and Its Critics in Antiquity. Berkeley 2007.
Taub, Liba. Ancient Meteorology. London and New York 2003.
Vlastos, Gregory. On Heraclitus, AJP 76 (1955) 337 368; reprinted in Vlas-
tos, Studies in Greek Philosophy. Vol. 1. The Presocratics (Princeton 1993)
127 151.
Whrle, Georg. Anaximenes aus Milet. Stuttgart 1993.
9. How Not to Conceive Heraclitean Harmony
Roman Dilcher
Harmony is presumably one of the terms that is most often connected

with Heraclitus philosophy, as it were, almost like a title that indicates
the character of his thought as a whole. But the difficulty with the Her-
aclitean concept of harmony is that it is hard to tell what its difficulties
are. From a semantic point of view, "qlom_a denotes something that is
joined together, a connection or a fitting. In this way, harmony can be,
and usually has been, taken to express the structural side of what Hera-
clitus calls more concretely j|slor: the order of the world that is estab-
lished by fire and its measures. Thus, harmony may be seen as a more
concrete explanation of how everything happens according to logos
(jat k|com, B 1). On the other hand, the concept of harmony leads
to what is often believed to be the central thesis of Heraclitus philoso-
phythe so-called unity of opposites. The main text for the concept of
"qlom_a is B 51: They do not understand how going apart with itself it
accords: a back-stretched fitting, such as that of bow and lyre, oq
numisim fjyr diaveq|lemom 2yut` blokoc]ei pak_mtomor "qlom_g
fjyspeq t|nou ja k}qgr.1
Now, although the unity of opposites is an almost ubiquitous for-
mula in contemporary interpretations, there are some fundamental
problems with it which I shall address briefly and a little dogmatically:2
1. In Heraclitus we often find the term for unity (6m), but never the
Greek word for opposites, 1m\mtia. This is a simple fact, and a fact
that at the very least needs to be addressed by anyone believing in
a unity of opposites. B 51 provides the term that comes closest
1 For a discussion of the textual problems see Kirk, H. 203 21; Kahn, H. 195.
2 For a more thorough discussion I may refer to my own book (Dilcher, H.), ch.
V. A slightly different account why Heraclitus does not use the term opposites is
given by T. Buchheim, Die Vorsokratiker (Munich 1994) 78 ff. From the per-
spective of the history of interpretations, it was Karl Reinhardt who succeeded
in shifting attention away from the flux doctrine to the problem of opposites as
being the central issue in Heraclitus: Parmenides und die Geschichte der griechischen
Philosophie (Bonn 1916) 220.
264 Roman Dilcher
to speaking of opposites: diaveq|lemom. In contradistinction to the

talk of opposites, diaveq|lemom is an expression in singular form,
that which is going apart, which is differing (from itself). To
this we may without hesitation apply a predicate of unity: it is
one, but whatever this means it will still be uninformative since
the subject is conceived of as a single phenomenon in the first
place, and not as two opposing forms. However, what is in fact pre-
dicated of it in B 51 is 2yut` blokoc]ei, it accords with itself.
Whichever way this is explicated, it certainly will not yield anything
resembling the common interpretations of opposites; i. e., that on
the surface there are, and are believed to be, two opposite phenom-
ena (such as day and night), in which some sort of (hidden) unity
can subsequently be discovered, such that from disparate appearan-
ces the philosophers synthetic view will arrive at an underlying
unity. In short, the formula of unity of or in opposites is right
from the beginning in glaring contradiction with the main relevant
text of Heraclitus.
2. The collection of Heraclitean fragments contains endless instances of
what can formally be classified as opposites. These alleged opposites
are very diverse, and accordingly the logic of their being one would
perforce be almost equally diverse. Heraclitus nowhere tells us that
these various instances should all be taken to be examples of one
doctrine of which they are illustrations and which should be extract-
ed from them. If one nevertheless insists on doing so, the onus of
defending such a thesis will be on the interpreter, and this onus is
a heavy one indeed. For either one will reach the conclusion that
Heraclitus has, to say the least, overstated his case, as many examples
do not match the thesis of unity once this is formulated in a logically
sound way. Or else one will have a very vague notion of unity that
can encompass all the diverse instances precisely by virtue of being
vague and indeterminate. In addition to vagueness, this way of in-
terpretation will produce emptiness. In illustrating this one vague
theory of unity, the various fragments will tend to be degraded
and become nothing more than illustrations. If their main point is
just to demonstrate unity again and again, they will lose their
power to say something of their own.
3. The specific interpretation of the formula unity in opposites that has
in some way or other become predominant (Kirk here being the
starting point) leads to further problems. In order to avoid the
charge of contradiction (from Aristotle to J. Barnes), the unity of
9. How Not to Conceive Heraclitean Harmony 265
the opposites is usually said to amount not to strict identity (e. g.,
day is night), but rather to some sort of underlying unity. According
to this interpretation, day and night are one in respect of something
else, for instance of the continuum of time they both occupy. Hon-
orable though this solution may be, it has the flaw of turning Hera-
clitus surreptitiously into a premature Aristotelian. The term un-
derlying which is so frequently und unhesitatingly used in
scholarship has its origin in Aristotelian physics and metaphysics,
which becomes plain once it is rendered in Greek: rpoje_lemom
(from which come Latin sub-stratum and sub-stantia).3
In Physics Alpha, Aristotle develops for the very first time this notion of
rpoje_lemom by starting with the traditional (i. e., Presocratic) ways of
accounting for movement through opposites and showing that some
third principle underlying movement from one opposite to the other
is needed. Interpreting Heraclitus along such Aristotelian lines will
not only spoil Aristotles innovative point but will also deprive Heracli-
tean unity of its most characteristic trait which in B 51 is expressed as a
back-stretched fitting. No underlying substratum will be able to possess
that tension that bow and lyre are meant to display.
As long as the formula of unity of opposites cannot be matched with
Heraclitus own conceptualization (of unity as well as of those phenom-
ena we usually class as opposites), such as can be gathered from B 51, it
will be a mere phantasma of scholarship. It may after all not be incidental
that Heraclitus does not directly introduce his notion of "qlom_g. What
he says is first of all that they do not understand it. So harmony seems
to be a special instance of the general human tendency to incomprehen-
sion. In addition, B 54 tells us that an unapparent fitting is stronger than
an apparent one ("qlom_g !vamr vameq/r jqe_ttym). While it is not
clear what bearing this distinction has on understanding the back-
3 B 67 has sometimes been adduced as implying a Heraclitean notion of the un-

derlying. If Frnkels supplement oil is accepted then this can indeed be in-
terpreted as in some way underlying the various spices. Yet this is still a far
cry from both the Aristotelian conception and from the sort of idea that is
used to explain the alleged unity of opposites. In B 67, god is presented
not as a unity underlying a specific pair of opposites but as being all
those that are named. The structure in question here is also special in that
the differences among the various phenomenal instances are to be explained
not as being opposites but as being different tastes or rather names of the
same thing (cf. Dilcher, H. 124 5). Neither tastes nor names can properly
be said to be opposed to one another.
266 Roman Dilcher
stretched fitting of B 51 as manifest in bow and lyre, it appears that in

order to understand Heraclitean harmony we also have to pay attention
to the distinction between the apparent and the unapparent. So in order
to understand the harmony of B 51 the first question that needs to be
addressed is why men tend to fail to comprehend it. The Heraclitean
concept of harmony turns into the problem how harmony is to be prop-
erly conceived.
There is one Heraclitean fragment that deals explicitly with the
problems of cognition of what is apparent. It is one of the longest extant
fragments of Heraclitus, containing no less than 37 words. And yet, al-
though it belongs to the few texts with a more discursive structure, and
though there are no problems of textual transmission,4 it has seldom re-
ceived proper attention. I am speaking of B 56, the legend of Homer
and the lice riddle. It runs as follows: 1ngp\tgmtai oR %mhqypoi pqr
tm cm_sim t_m vameq_m paqapkgs_yr jl^q\, dr 1c]meto t_m :kk^mym
sov~teqor p\mtym. 1je?m|m te cq pa?der vhe?qar jatajte_momter 1ng-
p\tgsam eQp|mter fsa eUdolem ja 1k\bolem, taOta !poke_polem, fsa
d oute eUdolem out( 1k\bolem, taOta v]qolem.
The translation provides no problems: Men are mistaken with re-
spect to the cognition of apparent things, in a way similar to Homer,
who was wiser than all the Greeks. For he was deceived by boys killing
lice who said: What we have seen and caught, we leave behind; but
what we have not seen and not caught, we carry with us. Heraclitus
here refers to the legends that were told about Homers life, several
compilations of which are transmitted in the Vitae Homeri. 5 The story
goes that Homer, after having spent some time on Samos, went to
the island of Ios. There he was approached on the beach by some
boys fishing, whom he asked whether they had caught anything. Instead
of admitting that they were going home empty-handed, they answered
by presenting the riddle. Unable to make sense of their words, Homer is
reported to have passed away.
Now, Heraclitus would hardly have included this little story just for
purposes of entertainment. Indeed, presenting a riddle appears quite in
line with the style of an author notoriously called the Obscure. While
many fragmentsand especially those that have central philosophical
relevanceare in themselves enigmatic, usually consisting of condensed
words with multiple meanings, the legend with the lice riddle appears
4 Cf. Marcovich, H. fr. 21 (pp. 81 ff).

5 Homeri Opera 5, ed. T. Allen, Oxford 1912, p. 184 ff.
rather plain. This may partly explain why this fragment has been so
thoroughly neglected. Still, if Heraclitus is presenting what he has to
say about cognition in the form of a riddle, or more precisely through
a story containing a riddle, then we have here the same enigmatic style
as a literary genre that in other fragments is displayed on the semantic
level. If Heraclitus thought as a whole presents itself as difficult to un-
derstand like an enigma or even like an oracle, then surely B 56 is sup-
posed to provide some clue for finding a way to understanding.
This story is given as an explanation for a general human cognitive
failure. Given the selective state of transmission, it could well have hap-
pened that only the first sentence would have been transmitted: People
are mistaken in respect to knowledge of apparent things. In this case,
our fragment would have been a perfect example of a general saying
(such as one often finds for instance in Stobaeus). The range and specific
application of such a saying, bereft of its explanatory context, is neces-
sarily open, and it would have been left to the interpreter to find a suit-
able context for it. Our fragment would then appear as just another re-
petition of those attacks against peoples misunderstanding with which
Heraclitus discourse began. While in many instances there is simply
no other choice than resorting to guesswork, B 56 offers the more or
less unique opportunity to determine the specific point of the saying
by referring to Heraclitus own explanation. The Homeric legend is
presented explicitly and pointedly to highlight the specific way in which
people are mistaken.
B 56 works by comparison. People are said to 1ngp\tgmtai, to be
mistaken, or more literally to be deceived or to deceive themselves.
This shortcoming is to be understood not as just as, speq or similar,
but as nearing, being about equal to, paqapkgs_yr, that which hap-
pened to Homer. The legend thus provides no full parallel but just an
approximation. On both sides of the comparison the verb 1napatm is
used, so that what can be understood about Homers situation is to be
transferred to that of men in general. The point of comparison is Hom-
ers inability to solve the riddle which, in Heraclitus view, comes close
to the general human lack of understanding. Although !patm is usually
translated as deceiving (this meaning here even strengthened by the
prefix 1n-), this will not adequately describe Homers state of mind.
He is in fact not deceived by the boys into believing something that
is untrue, but is rather put into puzzlement. Instead of a straightforward
answer such as: No, we had no luck with fishing, we have just spent
our time getting rid of lice, the boys utter their enigmatic words. In
268 Roman Dilcher
giving this answer, they count on Homers preconceptions about what

to expect from a fishing trip, and thereby succeed in outwitting him.
Delusion may be a more adequate description. While Homer was ac-
tively put into this state of mind by the boys (1ngp\tgsam in aorist
form), people are, as it were, constantly deluded by themselves: 1ng-
p\tgmtai in the middle voice of perfect aspect. Thus, by understanding
the mischief done to Homer, we will be able to gain some insight into
the prevailing epistemic condition in which Heraclitus thinks we all are
constantly entangled.
Homers delusion may be comparable to the general delusion, but it
is by no means obvious to what the latter relates. In the legend, the
question is whether and what the boys have caught on their fishing
trip, and the delusion is caused by the somewhat peculiar answer of
which Homer failed to make sense. That which is to be illustrated by
this legend Heraclitus says to be cm_sir t_m vameq_m, the cognition
of apparent things. Now unto itself this almost amounts to a paradox.
For whatever it may be that is said to be apparent, by its very being ap-
parent it must be supposed to be accessible to knowledge. Being appa-
rent is opposed to being unapparent or veiled; matters that are public,
that is, publicly known or at least knowable, as against what is secret.6
Given Heraclitus frequent hints that he is concerned with what is in
some way hidden and elusive we would rather expect the claim that
people are mistaken precisely in what is not so apparent.
At first glance, it may be tempting to suggest a simple solution. The
whole phrase not recognizing what is apparent can be taken to desig-
nate not a certain range of objects of knowledge but rather to emphasize
the failure of men. So it would boil down to nothing more than an un-
derlining of their incapacity for knowledge despite things being apparent.
In line with this interpretation is the proposal to take into account
Homers blindness.7 So just as Homer was physically not able to see,
so the upshot would be, people are metaphorically blind with respect
to the logos. Yet, although Homers puzzlement is indeed reminiscent
of the incomprehension people are said to experience in B 1, this sol-
ution will not do. Not only is Homers blindness not mentioned in
Heraclitus version, but the implication of this reading would be that
the only thing that can be learned from the fragment about the right ap-
proach to cognition is that it is the opposite of being blind. It would just
6 Cp. the various instances in LSJ s.v.

7 Cf. Kahn, H. 111 12.
amount to something like opening ones eyes and taking a straight look
at what allegedly is apparent before us. So what is meant by apparent
would consequently not be apparent in the vulgar sense, but only to the
philosophers eyes. Heraclitean insight would thus have to be reached
by some sort of direct intuition, but we would not gain any indication
as to how to achieve it. On this reading, the fragment would be unin-
formative on the crucial point.
Furthermore, and more importantly, this reading would positively
spoil the point of the riddle. Within the situation of the legend, it fol-
lows that Homer would have been able to understand the riddle if only
he still had had sight. This way he would have been able to observe the
doings of the fisher-boys and just see they had been catching lice. How-
ever, if Homer, or anyone else, had visual access to what was going on
there would be no need for a riddle in the first place. If the solution is
available visually, then a riddle has no point. A riddle calls for sagacity,
not for better observation or gaining more external information. The
fact that Heraclitus here calls Homer the wisest of the Greeks serves
to underline that what it takes to solve a riddle is above all wisdom. So
in highlighting the basic human failure this fragment gives an indirect
clue as to how to be wise.
Riddles are there to be solved by thinking about them, by finding the
right answer through listening to the words and making sense of them.
It is for this reason that the form of a riddle is congenial to the character
of Heraclitus thought. When confronted with a riddle one is prompted
to reconsider the implicit preconceptions one has about what is at issue.
Often the solution can be found if the words are taken in an unusual
way whereas if one clings to the normal expectations it remains enig-
matic.8 So, not being able to solve a riddlethe situation of
Homerindeed comes close to what is said of men in B 1: they hear
the logos but they remain uncomprehending. What they hear from
Heraclitus is certainly not just noises, but words they understand and
sentences they can follow. What they hear even has a meaning to
them but they still cannot make sense of it. Or rather, precisely because
they can make some sense of the sentences they read they conclude
that it seems to be nonsense. It all appears absurd to them just as the
boys riddle seemed to Homer. This allows for the conclusion that B
56 is actually meant to provide, by means of a comparison to a tricky,
8 Cf. the extensive material of ancient riddles collected by K. Ohlert, Rtsel und
Gesellschaftsspiele der alten Griechen, Berlin 1886.
270 Roman Dilcher
but comprehensible situation, an analysis of the problem of coming to

terms with Heraclitus logos.
As far as Homer is concerned, it is clear that what he had expected
was the kind of thing that is usually caught by fishermen. So, what he
could not move his mind away from is the usual way of thinking, and
his failure results from clinging to what appears self-evident and obvious,
or in the words of B 18, he was not able to expect the unexpected
and therefore could not find what is hard to track. In other words,
he was deluded pqr tm cm_sim t_m vameq_m, he was mistaken in basing
his attempts to gain knowledge on what appears to be apparent. So the term
apparent can suitably be understood from the delusory situation itself.
The implicit claim is not that he failed to recognize what is apparent, but
rather that what he ought to have done is to question his expectations
and to direct himself towards what is not apparent. Within the frame-
work of the legend, this realm of the hidden is materialized by the lice.
So, as also urged by B 54, we are advised to turn to what is unap-
parent. But this advice is by itself again uninformative as it gives no
guideline how to proceed. Homer was stuck although he realized that
the answer apparently was not the obvious result of hunting. The
only guideline for Homer was to find the clue in some way in the
words of the riddle, and as Heraclitus thinks we are in no other situation
we finally have to pay attention to the wording itself.9
In presenting his version Heraclitus apparently assumes the legend to
be known. The background story is condensed to the minimum while
by contrast the enigmatic words the boys utter are quoted in full. Com-
pared to the various versions of the Vitae Homeri, Heraclitus even has
expanded the saying. There we have (with slight variations) the follow-
ing line of hexameter:
fss( 6kolem kip|lesh(, fsa d( oqw 6kolem veq|lesha.
9 Even those interpreters who have dealt with B 56 usually shy away from ana-
lyzing the riddle. Cf., e. g., Uvo Hlscher, Anfngliches Fragen (Gttingen 1968)
138, Auf die Lsung des Rtsels kommt nichts an, es geht lediglich um die
Form als Rtsel, und die besagt: auch die Dinge stellen ein paradoxes Geheim-
nis dar, das gleichwohl am Tage liegt; sie sind selber ein Rtsel, das zu lsen ist
man mu nur die Chiffre lesen knnen, das heit, man mu das Sichtbare als
Zeichen verstehen lernen, als das Sich-Anzeigen des Unsichtbaren. Yet how
could we learn to read the chiffre without attending to the riddle? Without
a k|cor for solving the riddle the solution would amount to sheer intuitional
guesswork.
Heraclitus version, fsa eUdolem ja 1k\bolem, taOta !poke_polem, fsa

d oute eUdolem out 1k\bolem, taOta v]qolem, differs in two significant
points: first, he uses the present tense !poke_polem instead of the aorist
form. He envisages not a unique action in the past, something that the
boys did before meeting Homer; instead, this action is to be understood
as ongoing, on an equal level with and simultaneous to the other part of
their activity, v]qolem. Leaving behind and carrying along are presented
as interrelated aspects of the same activity, the first relating to what has
been caught, the latter to what has not been caught.
Second, while the traditional verse just has 6kolem for designating the
action of catching, Heraclitus instead speaks of eUdolem and 1k\bolem, that
is a double action that is evidently chosen to suit better to be transferred to
the problems of cognition. This choice of words is remarkable by itself.
Without any tradition of epistemological vocabulary at hand, Heraclitus
here uses in conjunction those two terms that will later play a fundamen-
tal role in the history of metaphysics and epistemology: seeing, as the
sense which gives the best and most differentiated access to the world,
stands for sensory perception in general, or even metaphorically for intel-
lectual intuition. On the other hand, taking denotes the intellectual side
of cognition. In Greek, aRqe?m (the word in the versions of the Vita Hom-
eri) is not often used in an epistemological context, whereas kalb\meim
(the word chosen by Heraclitus) is here quite usual; it means to appre-
hend, to take something as, to understand.10 In Latin the equivalent is
con-cipere, whence conceptus, and in modern languages various expressions
of grasping (in German Begreifen and Begriff); all of them make use of
the metaphor of laying hands on something and seizing it so that it will
be understood and possessed by the mind.
Now, what is even more remarkable is that by employing the riddle
for explaining cognition, or more precisely the common failures of cog-
nition, Heraclitus apparently sets out to undermine this view: What we
have seen and taken, we leave behind; but what we have not seen and
not taken, we carry with us. In both the literal and the metaphorical
sense, this comes as a paradox; it fundamentally clashes with what is be-
lieved to be the natural order of grasping something. What is to be
caught or conceived is usually sought after in order to acquire it. It is
something that cannot be left behind as it is not had in the first
place, and that is the very reason for going after it. The intention of
catching something is to possess it and to carry it home. Transferred
10 LSJ s.v. A I 9.
272 Roman Dilcher
to cognition: grasping something is done by conceiving, that is, by

forming a concept of it through which the matter in question is
had by the mind.
Here we arrive at the core of the delusion. In B 56 we do not learn
only that Heraclitus insight comes like a riddle that is to be solved by
questioning our usual beliefs about the world. The message of the frag-
ment is not confined to recommending in an undetermined way to ex-
pect the unexpected or to turn to the unapparent. Rather, we are of-
fered a positive analysis of what is wrong with the usual conception
of cognition. What Homer could not make sense of is the reversal of
the business of catching; in the present case, it does not aim at acquiring
and possessing, but at getting rid of something.
Homer may be imagined as pondering what on earth the boys might
have caught, if it apparently was not fish. He might have been prepared to
look for other possible and impossible objects of their hunt. Still, this
would only have resulted in ignorance. His puzzlement is not just the re-
sult of being unable to find a solution, a suitable object fitting the descrip-
tion (such as might have happened if the answer to the riddle were too
far-fetched). His delusion arises from realizing that there can be no
such object in principle given, that is, his unquestioned assumptions
concerning what catching is all about. Homer is positively in a predica-
ment; the harder he looks for a solution, the more impossible it appears
to be.
It is interesting to note how Herrmann Frnkel accounts for the un-
usual action: What men see with their eyes and grasp with their hands,
they throw away without understanding it and making it their own, but
the true realities they carry like lice in their clothing which escape
both their eyes and their hands.11 While Frnkels remark about
where the true realities lie surely hints at the right direction, he still
seems not to realize the consequences of following this hint. If what
is sought here is something that lies hidden, like lice, near or on oneself,
then grasping it will of necessity not be done in the usual way. By re-
garding the action of throwing away as a failure, Frnkel himself mistak-
enly confuses the peculiar business of the boys with the human error
that is illuminated here. Yet, what the boys are doing is clearly not
meant to illustrate the human failure itself, but on the contrary to be
that which causes the delusion. It is Homers state of mind that is the
11 H. Frnkel, Early Greek Poetry and Philosophy (Oxford 1975) 372 3. Frnkel ap-
pears to be the only interpreter who has tried to analyze the riddle itself.
point of reference for understanding that of people. Frnkels interpre-

tation is illuminating as it also betrays the reason for this confusion: he
explicitly equates understanding with making it their own. Obvious-
ly, he could not make sense of throwing away as being an essential part
of the troubling phenomenon, and so he believed it was part of the at-
titude that Heraclitus wanted to criticize. This is a striking example of
the power of the conventional paradigm that understanding is essentially
about making something ones own and possessing it.
If the scene at the beach is to serve as the model for understanding in
which way people fall short in cognition, then it will not do to point
vaguely at some sort of hidden reality. What is at stake is not just a pe-
culiar and possibly remote object of cognition that is to be found out by
unusual means. It is the very nature of understanding itself that is in
question here, its direction and its manner. What is highlighted by
the legend is not a simple error or failure; it is (in modern parlance) a
category mistake.
Two conclusions can be drawn from B 56. The first comes as no
surprise. The main reason for Homers not solving the riddle proves
to be his conviction that what one is after is something external to
the person doing the hunt, something that is to be caught by approach-
ing and finding it. Transferred to the problems of cognition, this means
that in Heraclitean philosophy we are not dealing with anything resem-
bling knowledge of external objects in the world, be it the harmony of
the cosmos or a universal cosmic law or a unity in all phenomenal op-
posites. Whatever it is, it must not be conceived of as anything objec-
tive, lying there like an object to be discovered, to be seen and grasped
by the philosophers mind. True knowledge concerns something to
which ones relation is not external, and it involves the person who is
trying to understand in a special way.
This claim is just the epistemological consequence of the main
theme of the Heraclitean discourse. In B 1 it is stated that the subject
matter will be the words and deeds of men of which they are peculiarly
unaware.12 Accordingly, the way in which Heraclitus thinks he is distin-
guished from them is precisely that he has made the effort to go in
search of himself (B 101). Knowledge of oneself is apparently in some
way the main concern of Heraclitus philosophy, and it is this sort of
knowledge that we are supposed to seek. In B 56, this is symbolized
12 For a detailed interpretation of B 1, see Dilcher, H. ch. I.

274 Roman Dilcher
by catching lice on oneself, as against hunting for fish or any other ob-
jects in the world.
The second conclusion from B 56 is more perplexing. The riddle also
shows how the nature of what is to be known has implications for the
right way to deal with it; in the imagery of the legend, lice are not caught
in the same way as fish. Now, the appeal to go in search of oneself could
be understood as retreating from the outward world and concentrating
on the state of the soul or some inner self. While we do find in Hera-
clitus much attention to the soul and its states (for the first time in philos-
ophy), Heraclitus logos apparently does not exclude consideration of the
cosmos and the so-called physical world. B 56 may help to get the point
right. The riddle indicates that for Heraclitus it is not any form of intro-
spection that is called for, but rather a different mode of understanding
of oneself as well as of the world one lives in. The boys action involves
not only a reversal of direction by turning to oneself, instead of reaching
out into the world. While it is conceivable that lice (or anything similarly
close to oneself) are to be caught simply by laying hands on them and
keeping them, the crucial point that causes the delusion is that the lice
are caught in order to get rid of them. This may appear natural in respect
to lice, but it is not natural any more if this action is to serve as the clue for
understanding what understanding oneself is about.
While self-knowledge is in some way or other a standard theme in
the philosophical tradition, it is this reversal of the action of gaining such
knowledge which requires explanation. As the riddle so pointedly and
effectively puts it, the knowledge in question is nothing that after
being taken can be had and as it were be carried along (v]qolem).
Self-knowledge not only cannot be gained in the same way as knowledge
of external objects, it is, according to Heraclitus, not even something
that can be gained in the strict sense at all, in the sense of making it
ones own and possessing it; rather, it is to be left behind (!poke_polem).
So whatever the procedure is like, in the end nothing is left in our
hands, and the effort will be in vain; at least so it will necessarily appear
to standard beliefs.
When turning to Heraclitus claims to knowledge, our first (and to
many interpreters also the lasting) impression appears to be that he be-
lieved himself to have discovered the one and only truth which all man-
kind fails to recognize. This universal claim seems to be presupposed by
his many attacks against men in general and almost all intellectual au-
thorities in particular (B 56 appears to be just another instance of this
on first reading). Yet, on closer inspection it is rather difficult to sub-
stantiate this impression by any explicit claims to such universal know-

ledge. In contrast with Parmenides, Heraclitus never uses the word !k^-
heia to mark his own insight (B 116 being a general remark on wisdom).
Even more striking is the fact that Heraclitus often speaks of cicm~sjeim,
but always in a negative way: people are regularly criticized for not re-
cognizing something.13 Including B 56, there are no fewer than eight
instances of this kind. The evidence is even clearer with the word
cm~lg. In B 41, to understand cm~lg is attributed to the one Wise,
and in B 78 we have the statement that human ethos has no cm_lai,
while divine ethos has. Finally, there is the absence of cicm~sjeim
where you would expect it most: given the central importance of
self-knowledge, Heraclitus may be supposed to claim to have achieved
at least this. In fact, however, he only makes the assertion: I went in
search of myself, 1difgs\lgm 1leyut|m (B 101). Heraclitus may have
considered it a major step to do this, and it apparently was a point he
felt was important enough to pronounce, but still it is a modest claim.
This is all the more striking when the saying is understood as in some
way reacting to the Delphic maxim cm_hi saut|m. In this context be-
longs also B 45, which forcefully states the inscrutability of the soul:
The limits of soul you will not find out, even if you travel any road.
So deep is its reasoning. So it should not be assumed that the Heracli-
tean logos either proceeds from or aims at a standpoint of insights that
can be taken as knowledge in any usual sense.
Heraclitus apparently did not find it suitable to present his insight in
the epistemic terms of truth and knowledge that seem so natural (if not
unavoidable), and which were to be the core of philosophy and science
later on, and this despite or rather because of his vivid interest in the
various modes of knowledge. Instead of gaining for oneself a result, like
a big fish from which one can make a living, Heraclitus presents an ac-
tivity of letting go of what has been seized. As this is the opposite of
what we expect knowledge to be, it may be tempting to conclude
that Heraclitus has the negative aim of a skeptic who takes pride in get-
13 This important discovery was already made by B. Snell, Die Ausdrcke fr den
Begriff des Wissens in der vorplatonischen Philosophie (Berlin 1924) 66 7. See
also his pertinent remark in Die Sprache Heraklits, in his Gesammelte Schriften
(Gttingen 1966) 136 n. 3: Erkennen kann fr Heraklit hchstens eine gtt-
liche cm~lg, menschliches cicm~sjeim bleibt immer im Unzulnglichen steck-
enaber darin zeigt sich, wie schwach in ihm der Impuls zum Erkennen war
und der Glaube, damit an ein wertvolles Ziel zu gelangen. Eine Resignation
gegenber dem ihm wesentlichen Wissen liegt Heraklit durchaus fern.
276 Roman Dilcher
ting rid of all opinions, as he thinks no truth is in human reach. This

conclusion is equally unsatisfactory, however, as, irrespective of claims
to truth, it in no way matches the presentation of ideas which we
find in the Heraclitean fragments. Heraclitus logos may be evasive but
it is not negative in the sense of aiming at emptiness.
Maybe the legend with the riddle has been overstretched in the at-
tempt to extract some sort of Heraclitean epistemology. Or maybe rath-
er, it has not yet been read carefully enough. For what Heraclitus says
the legend will illuminate is not the right way of reasoning, but the de-
lusion of men. It is only on the background of their preconceptions
about knowledge that a peculiar epistemic activity which does not
end in getting hold of anything will appear negative, disappointing,
and vain. The sceptical idea of deconstructing claims of knowledge is
still based upon the concept of knowledge as acquiring a correct view,
only that this is considered to be impossible. What Heraclitus presents,
by contrast, is the way in which the pre-philosophical view cannot
make sense of the reversed activity in question. What is called for is
not quite giving up all aspirations to knowledge, but rather abandoning
the standard beliefs about what true knowledge is like.
Where this will lead is still unclear. B 56 is unique in explaining the
epistemic illusion of mankind, but it is not unique in addressing the
epistemic idea which is to be gained here. Already in B 2 men are criti-
cized for making something their own, for creating a private under-
standing whereas the logos is something num|m. Thus it is logosto be
translated as reasoningwhich stands for the philosophical way of
thinking that avoids the shortcomings of the conventional understand-
ing of knowledge. The word num|m here is hard to translate. Usually it
has been fitted into a more general picture according to which men
are like sleepers caught in a private world of their own (with the
Stoic formulation in B 89). Heraclitus appeal would then be that one
arises from slumber and open ones eyes in order to see the common
world in which we all live. On this construal, however, the main
way to gain insight would again be modeled after the paradigm of seeing
something external; it would amount to some sort of immediate intu-
ition of what is obvious. The problem addressed in B 2, however, is
that of adequately comprehending the logos, and for this reason the
model to be referred to for gaining the right way of thinking is not see-
ing but hearing in its full sense:14 what is heard is primarily not noises
but words, and not just words, but words meaningfully strung togeth-
erin Greek, k|cor. It is therefore misleading to construe Heraclitean
insight by way of a model of sensory perception; that is, by any sort
of immediate awareness or direct intuition. As both B 1 and 56 tell
us, what is to be achieved is understanding and making sense of
words deliberately strung together, and the suitable Greek word for
this is numi]mai, to bring together.15 People remain uncomprehending
(!n}metoi, in B 1) as they fail to bring together the words of the Hera-
clitean logos, just as Homer failed to make sense of the boys riddle. Also
the word num|m in B 2 ought to be interpreted in the same direction.
And a failure to bring it together (oq numisim) is what happens to people
also in respect to the back-stretched harmony of B 51. In the light of B
56, it appears that these are different manifestations of the same mistaken
attitude. Therefore, not comprehending the back-stretched fitting(B
51) can be read as another comment on what it specifically is that
makes the Heraclitean logos difficult to understand; for if they fail to
bring together those artfully und deliberately composed words of Hera-
clitus this failure concerns the adjustment of those words, their connec-
tion and specific structure, in Greek their "qlom_a. Thus, the problem of
conceiving Heraclitean harmony is in fact coupled with the basic human
problem of coming to grips with the logos, and an explanation for this
predicament is offered by B 56.
Now, the main factor for not understanding could be determined as
the tendency to make what was seen and taken ones own so as to
possess it. There is another fragment that may shed further light on
what this may amount to, B 28: doj]omta cq b dojil~tator cim~sjei,
14 B 101a Eyes are better witnesses than ears has often been taken as the basis for
a Heraclitean epistemology. Given the lack of context, however, it is difficult to
determine the scope of this saying. Rather than valuing one sense over another,
this fragment is presumably concerned with criticizing the reliance on hearsay
(cp. K. Robb, The Witness in Heraclitus and in early Greek law, The Monist
74 (1991) 664 ff). As B 107 demonstrates, sense perception as such is for Her-
aclitus something that needs interpretation. So we are in any way led back to
the problems of properly understanding a language.
15 Cf. Snell Ausdrcke (above, n. 13) 48: diese Beziehung zum menschlichen
Geist ist durch sumi]mai deswegen gegeben, weil es von der Wahrnehmung
durch das Ohr ausgeht. Dessen Objekt ist menschliche vernnftige Rede, wh-
rend das Auge auf die Gegenstnde des Raumes gerichtet ist, deren Ordnung es
erkennt. According to Snell 41, the basic meaning in Homer is to be rendered
as etwas durch das Gehr in sich aufnehmen und ihm geistig folgen.
278 Roman Dilcher
vuk\ssei. Whatever the correct wording may be,16 the gist of the frag-
ment is again criticism of knowledge, here expressed by the paradoxical
phrase doj]omta cim~sjei: he, whoever this dojil~tator is supposed to
be, recognizes what is only doj]omta. As everyone familiar with Parme-
nides knows, the translation of doj-words is notoriously difficult, as it
can range from d|na in the plain sense as view, opinion to more neg-
ative senses such as seeming, mere belief. The latter meaning occurs
for instance in B 17, there in opposition to cicm~sjeim: they do not
learn and know, but they form beliefs of their own, oqd lah|mter
cim~sjousim, 2yuto?si d doj]ousi. Again, we encounter the idea of
having something for oneself (2yuto?si), which here can be understood
as leading to the fabrication of doxa. It is likely that doj]omta in B 28
should be understood on the same lines. Now, the interesting informa-
tion in B 28 is that cicj~sjeim is here coupled with vuk\sseim, which is
the exact opposite of !poke_peim. It is therefore advisable to take vuk\s-
seim here as specifying this epistemic activity. Normally one guards and
protects something from getting lost, one cares for it so that it may re-
main intact. Not to let go but to guard something one has taken in
cognition will accordingly mean that it is transformed into a doxa, a
firm view one entertains about the matter in question. Such a fixed be-
lief is what one can carry along and thereby become a d|jilor among
menjust as, for instance, Pythagoras who, Heraclitus says, contrived
even a wisdom of his own (1poi^sato 2autoO sov_gm, B 129).
A recurrent epistemological theme emerges in Heraclitus criticism
of human incomprehension. The contrast between letting go and
carrying along is vital to understanding the Heraclitean logos. The
mistaken, though common and in fact natural attitude leading to incom-
prehension results in forming beliefs and con-cepts by which what
has been seen and taken can be carried along. If this way of under-
standing as making something ones own ends in forming fixed doc-
trines, then another fragment that has rarely been interpreted calls for
consideration. B 70 has something to say about human doxai: pa_dym
!h}qlata t !mhq~pima don\slata, childrens toys are the human
opinions. Again as in B 56, we hear of playing children.17 Playing is
16 The grammatical forms of the first and the last word as given above are the most
common accepted emendations; cf. Marcovich fr. 20 (p. 78 ff).
17 As B 52 (aQm pa?r 1sti pa_fym jtk) shows, Heraclitus ought to be taken seri-
ously in what he says about the role of playing children in human life. For in-
terpretation, I may again refer to Dilcher, H. 153 ff.
what the boys do with Homer, and what they play with is precisely
Homers opinions. Likewise Heraclitus plays with his readers. So B 70
might suitably be understood as expressing what human beliefs are
there for: to be played with. This might be an indication of a possible
way to deal with that epistemic predicament.
To sum up: the initial question was how to conceive Heraclitean
harmony, and after realizing the problems inherent in formulating a co-
herent concept or doctrine of harmony (or unity) we now reach the re-
sult that attempting to gain and fix any such doctrine is precisely the
mistake humans tend to make. The problem with the Heraclitean con-
cept of harmony turns out to be that harmony is not to be conceived.
Even more, any concept or doctrine in its normal sense, so at least it
seems, will fail to make sense of Heraclitus logos. This epistemic idea
that can be gained from B 56 may still appear puzzling, but it is no
more and in no other way puzzling than the character of Heraclitus
logos itself. In fact, it proves to be a very precise description of the use
of concepts or themes that can recurrently and on different levels
be observed in Heraclitus. What is once taken up as the subject of dis-
course or as an explanatory notion will then be abandoned again, or it
will reappear with different and new connections.18 So B 56 in fact
points at an essential feature of the logos by issuing a warning. If we pro-
ceed on conventional epistemological assumptions, we shall wreak
havoc on the logos and fail to conform with its specific structure, and
that is, with its harmony.
How a more positive way of coming to terms with Heraclitus logos
might look is still out of reach. Yet in B 56 there is an indication. Her-
aclitus does not put us in the situation of Homer not knowing how to
solve the riddle. Instead, he tells us the solution right from the begin-
ning: the boys, he says, were killing lice, vhe?qar jatajte_momter. By
knowing this solution, the reader is able to understand Homers delu-
sion. The words that illuminate for us the situation at the beach may
also illuminate for us, in a way, the meaning of this whole business.
After all, the boys killing of lice is the reason, in the framework of
the legend, for the thoroughgoing negativity that shatters all attempts
to get straightforward answers from Heraclitusafter reading B 56 no
less than before. In the pictorial language of the legend, this is the result
of presenting a trifling activity like killing lice instead of bringing us the
expected catch of fish that would give us a mouth-filling meal. Likewise
18 Some descriptions can be found in Dilcher, H. ch. 6.

280 Roman Dilcher
our hunger for anything definite is frustrated; no doctrine to fill our in-
tellectual appetite is forthcoming. Even so, killing lice is useful, in that
one thereby gets rid of something annoying. What one gets rid of is in
Greek called vhe?qer, and this word may well be understood from the
root vhe_qeim, to destroy, ruin, eliminate. So somewhat secretly we
here have two words for the same negative activity. The boys are bring-
ing death to the lice which in fact are themselves called destroyers. The
peculiar activity could thus be rendered as killing the killers. What this
may mean for the nature of true understanding must remain open,
but it seems to have something to do with keeping oneself alive.
So, B 56 in the end sheds some important light on Heraclitus
thought as a whole. In view of the fact that it is the only fragment
that actually explains what is wrong in the belief and practice of the nor-
mal pre-philosophical way of thinking, it is a guideline as to how to deal
with Heraclitus main insight. Considering the epistemic predicament
that mankind is in according to Heraclitus, one may well say that with-
out heeding the claims of B 56, the logos will forever remain incompre-
hensible ( just as he says in B 1). In this way, this fragment holds the same
fundamental function as does in Platos philosophy the simile of the
cave: showing the way out of human shortcomings by engaging in phi-
losophy. Surely, Heraclitus is more reticent about what the wise will ex-
perience. Instead of a colorful imagery (as in Plato), we are confronted
with a deceptive but inconspicuous riddle. Yet, by paying close atten-
tion to what it says, we can not only rid ourselves of some delusions,
but what we are in fact also doing is, to some extent, listening and con-
forming to the logos. Only for those who expect big fish will, whatever
they see and take, be and remain a lousy tale.
Bibliography
Buchheim, Thomas. Die Vorsokratiker, Munich 1994.
Frnkel, Hermann. Early Greek Poetry and Philosophy. Oxford 1975.
Hlscher, Uvo. Anfngliches Fragen. Gttingen 1968.
Ohlert, Konrad. Rtsel und Gesellschaftsspiele der alten Griechen. Berlin 1886.
Reinhardt, Karl. Parmenides und die Geschichte der griechischen Philosophie. Bonn 1916.
Robb, Kevin. The Witness in Heraclitus and in early Greek law, The Monist
74 (1991) 638 676.
Snell, Bruno. Die Ausdrcke fr den Begriff des Wissens in der vorplatonischen Phi-
losophie. Berlin 1924.
Gesammelte Schriften. Gttingen 1966.
10. Heraclitus on Logos
Language, Rationality and the Real
Enrique Hlsz
Almost1 all interpreters of Heraclitus have always had a ready answer to

the question: What was Heraclitus?2 The mainstream reply assumes as
self-evident that he was a philosopher. In what follows I, too, will take
for granted that he was a philosopher, but I shall start by pointing out
that universal attribution of a philosophical character to Heraclitus
thought does not really imply monolithic unanimity, since the term
philosophy can be (and usually is) understood quite equivocally.
And I would like to insist that Heraclitus (and, for that matter, all
other Presocratics) probably did not use the words philosophy, phi-
losopher, or philosophize, and so could not think of himself or of his
own activity as a thinker in these terms.3 Thus the concept of vikosov_a,
1 Agustn Garca Calvos views about the non-philosophical character of Heracli-

tuss thought are among the few exceptions that I know of. See A. Garca
Calvo, Razn comn. Edicin crtica, ordenacin, traduccin y comentario de los restos
del libro de Heraclito [sic] (Madrid 1985). Although Giorgio Colli, in La sapienza
greca, III. Eraclito (Milan 1980), sometimes seems to deny him implicitly philo-
sophical status because he usually calls him a sage (un sapiente), one can find
also in his valuable work statements like this: La filosofia di Eraclito, per il fatto
che esprime la verit assoluta, pu dirsi senz altro il k|cor (172).
2 The lucid treatment of M. L. West happily puts it in just these words. See M. L.
West, Early Greek Philosophy and the Orient (Oxford 1971). West certainly seems
to credit Heraclitus with a philosophy, and I believe that he is right when he
affirms that physical cosmology was not central in Heraclitus thought. As
to what sort of philosophy, then, Heraclitus was, West interprets it (closely fol-
lowing K. Reinhardt, and perhaps not so far from Burnet) as having a religious
end-purpose. My drift is that Heraclitus philosophy lies rather in his metaphy-
sics, understood as a rational methodic and systematic concern with knowledge
and the real. See E. Hlsz Piccone, Lgos. Herclito y los orgenes de la filosofa
(Mexico City 2011).
3 The verb vikosov]y and words of this family are first attested in fifth-century
authors (notably Herodotus) but are wholly absent from all genuine Presocratic
fragments. The sole exception which could be pointed out is a fragment of
Heraclitus (B 35), but in all likelihood, as Marcovich shows in his H., the ref-
282 Enrique Hlsz
as a regulatory hermeneutical idea, is most of the time presupposed, and

the validity of its attribution to Heraclitus is hardly ever fully acknowl-
edged and argued for. In point of fact, and even if there are good
grounds for doing so, when we call the Presocratics philosophers,
we are projecting retrospectively a language that was shaped at a later
time, mainly by Plato and Aristotle, from which everybody else took
it. Although there is some continuity in the shaping of the idea of phi-
losophy from Plato to Aristotle, there are also striking differences be-
tween their respective conceptions, including their historical approaches
to the Presocratics, particularly to Heraclitus. In Aristotle, he appears as a
self-contradictory thinker and writer, and as one of the vusijo_, his pe-
culiar monistic thesis about the !qw^ being fire. The image of Heraclitus
in Plato is more nuanced, less physicalistic, centered as it is on flux and
the unity of opposites. I therefore want to emphasize the relevance of
the question of the nature and the origins of philosophy for any inter-
pretation that chooses to see Heraclitus precisely as a philosopher.
In this paper I shall try briefly to articulate a general view about who
Heraclitus the philosopher is. My thesis is that he is both a metaphysi-
cian (an ontologist and epistemologist) and a moral philosopher. I shall
take k|cor as the center of my sketch, and against semantically reductive
interpretations I will argue that the full meaning of logosa word best
left untranslated in the textsinvolves the notions of language, knowl-
edge, reality and action, a complex unity pervaded by a common ration-
ality. I will focus first on the metaphysical dimension of logos, and sug-
gest that this is the ground on which Heraclitus builds his idea of mans
nature (which is centered in knowledge, itself essentially connected to
both being and action, the ontological and the ethical dimensions). At-
tribution to Heraclitus of a theory on logos depends partly on whether
one acknowledges its ontological status, and partly on whether one ac-
cepts that there is semantic unity and philosophical consistency through-
out all documented instances of Heraclitean usage.
erence there to vikos|vour %mdqar is Clements, not Heraclitus. Cf. P. Hadot,

Quest-ce que la philosophie antique? (Paris 1995). All references to fragments of
Heraclitus are made according to DK, ch. 22, although the Greek text assumed
is mostly that of Marcovichs edition. Translations into English are my own,
unless otherwise noted.
10. Heraclitus on Logos 283
Telling the Logos: Heraclitus as Metaphysician
It is remarkable that Heraclitus himself explicitly describes his own atti-

tude and procedure in terms which amply justify the ancient opinion
that the thinking he displays is philosophy indeed, even if he would
not actually word it like that. In the opening of his book,4 after dealing
with mens lack of understanding about a logos which is always the same
(literally, always this) and by which all things happen, he offers the
earliest recorded description of what came to be called philosophy,
when he goes on to characterize his own procedure as di^cgsir, a nar-
rative explanation5 dealing with words and deeds, according to na-
ture (jat v}sim6). Yet more precisely, he presents his own rational ex-
planation as diaiqe?m, a dividing of each thing and a showing it forth
(verbally7) as it truly is (vq\feim fjyr 5wei). What we can call his phi-
losophy and his method, he presents as the unfolding or telling of the
true and universal logos, closely linked to knowledge (explanation, anal-
4 For the Greek text, see below, note 12. That B 1 is the !qw^ of the book relies
on the authority of Aristotle and Sextus Empiricus, and is nowadays widely ac-
cepted. Not implausibly, many have proposed that it could have been preceded
by a few words like Heraclitus of Ephesos says as follows. It remains possible
that other preserved fragment or fragments preceded B 1 (B 50 is a recurring
candidate); cf. L. Tarn, The first fragment of Heraclitus, Illinois Classical
Studies, XI, 2; S. Mouravievs recent reconstruction (Heraclitea, IV.A, 2011)
conjectures a series of nine fragments before it. Some think that B 1 and at
least B 114 and 2 (in that order) belong together here and form a neat proem.
5 The Greek noun di^cgsir, meaning something like narrative exposition and/
or explanation, corresponds to the verb actually used, digceOlai (set out in
detail, describe; LSJ s.v.).
6 Heraclitus use of v}sir is arguably the oldest preserved within the philosophical
tradition (since he probably wrote earlier than Parmenides, and the term is not
documented in the scarce authentic fragments of Anaximander and Anaxi-
menes). V}sir is one of Heraclitus main philosophical watchwords endowed
with considerable linguistic density, to use Kahns words. Note in the passage
quoted from B 1 the apo koinou construction of the adverbial phrase jat v}sim,
affecting all three surrounding verbs. For the three other instances of the word
in the fragments, see B 112, 123, and 106the last probably not a full verbatim
quotation. Put in direct style, v}sir Bl]qgr "p\sgr l_a 1sti would be a mini-
malistic and reasonable version of the saying. See Mouraviev, H. III. 3. i-iii for
his reconstruction of B 106.
7 This seems to be the best sense for vq\fym in context, and one that is consistent
with usage at Heraclitus time. Cf. LSJ s. v. vq\fy, 2, show forth, tell, de-
clare.
284 Enrique Hlsz
ysis and manifestation) of genuine being as v}sir.8 In fact, the formula

peq v}seyr Rstoq_a, an investigation on the nature of things, first
documented in Plato,9 which became (through Aristotle) the basis of
the standard generic characterization of presocratic philosophy, is
likely based on Heraclitus actual language. Platos view differs from Ar-
istotles significantly in that Heraclitus is treated as a metaphysician rath-
er than as a natural philosopher10.
Heraclitus stance contrasts with the oblivious disposition of other
men, who are likened to sleepers (thus projecting back the idea of
awareness as a fundamental trait in Heraclitus own philosophical atti-
tude). Because of this self-awareness, a certain self-reflexivity goes
well with k|cor from the very start. In any case, k|cor is undeniably
present as an explicit topic at the starting-point or !qw^ of Heraclitus
book, so it appears to be a true principle at least in a literary, and espe-
8 Pace H.-G. Gadamers opinion that v}sir carries no philosophical import yet in
Heraclitus; cf. H.-G. Gadamer, Der Anfang der Philosophie (Stuttgart 1996) 41
42 = 34 35 of the Enlish translation. The opposite view of the idea of v}sir as
an early basic metaphysical concept was, of course, defended by M. Heidegger
in his Einfhrung in die Metaphysik and essays Aletheia (Heraklit, Fragment 16)
and Logos (Heraklit, fr. 50 DK) (all in the Gesamtausgabe, Frankfurt, various
dates). Although Gadamer says he agrees with Kirk (in KRS), his claim misrep-
resents the latters own statements, which do not really back that v}sir has no
philosophical weight whatsoever in Heraclitus. In the better and wider treat-
ment of B 123, Kirk actually states that the broad general sense of v}sir,
from which all specialized senses are derived, is essence or nature, the way
a thing is made, [] the way it normally behaves, and that the most common
early sense of v}sir is being, though the idea of growth is not excluded and
may be emphasized on particular occasions. See Kirk, H. 227 231,
esp. 229. This is a far cry from v}sir not being philosophically significant in
Heraclitus, even if it does not mean Nature, as it allegedly does in Aristotle.
9 Phaedo 96a7. It is crucial for getting the right sense in context not to translate as
nature tout court, but as the nature of things.
10 For Platos reception of Heraclitus, see T. Irwin, Platos Heracliteanism, PQ
27/106 (1977) 1 13; M. Adomenas (2002), The fluctuating fortunes of Her-
aclitus in Plato in A. Laks-C. Louguet (eds.), Quest-ce que la philosophie prsoc-
ratique (Villeneuve-dAscq 2002) 419 447; See also E. Hlsz Piccone, Scra-
tes y el orculo de Delfos (una nota sobre Platn, Apologa de Scrates 20c4
23c1), Theora 14 15 ( June 2003) 71 89, and Anmnesis en el Menn plat-
nico, Apuntes filosficos 22 (2003) 61 79. See id., n. 2 above, as well as id., La
imagen de Herclito en el Cratilo y el Teeteto de Platn, in E. Hlsz (ed.), Nue-
vos ensayos sobre Herclito (Mexico City 2010) 361 389; and id. Platos Ionian
Muses: Sophist 242 d-e, in B. Bossi and T. M. Robinson (eds.), Platos Sophist
Revisited (Berlin 2012).
cially in a narrative, sense. It can be seen as an !qw^ in a fuller philo-

sophical sense, too (although not as either a material or an efficient
cause).
The Proem in Heraclitus book could have run something like this11:
[B 1] Although the logos is this always men become uncomprehending both
before hearing and once they have heard it. For about all things that happen
in accordance to this logos, they are like the unexperienced experiencing
words and deeds such as those I set out in detail according to nature distin-
guishing each thing and showing it as it truly is. But other men neglect all
things they do awake just as all things they are oblivious of while asleep.
[B 114] Those who speak with sense must strengthen themselves with that
which is common to all, just as the city with the law, and more strongly
still. For all human laws are nurtured by a single one, the divine [law].
For it commands all as much as it wants to, and it suffices for all and is
still left over.
[B 2] That is why one must follow what is common, but although the logos
is common, most people live as if they had a private wisdom of their own.12
Much that concerns the purpose and nature of Heraclitus philosophy
hangs on just what the word logos means here. That the meaning is a
rather complex one is the view most widely held among modern schol-
11 I do not claim exhaustivity here, as there could be a few other fragments com-
ing between B1 and B 114 (B 19 and 34 come to mind). As to the relationship
between B 114 and 2, I follow Marcovich and others in treating it as a single,
continuous whole. This would require more argument that I can give here, but
it is prima facie likely, on grounds of style (i. e. narrative, diegematic form) and
of content affinity (i. e., the num|m), that the texts go together. It is worth no-
ticing that Kahns insightful ordering does not follow this line, nor does Mour-
avievs reconstruction.
12 B 1: toO d k|cou toOd 1|mtor aQe_ !n}metoi c_momtai %mhqypoi ja pq|shem C
!joOsai ja !jo}samter t pq_tom7 cimol]mym cq p\mtym jat tm k|com
t|mde !pe_qoisim 1o_jasi, peiq~lemoi ja 1p]ym ja 5qcym toio}tym, bjo_ym
1c digceOlai jat v}sim diaiq]ym 6jastom ja vq\fym fjyr 5wei. tor d
%kkour !mhq~pour kamh\mei bj|sa 1ceqh]mter poioOsim, fjyspeq bj|sa evdom-
ter 1pikamh\momtai.
B 114: nm m|\ k]comtar Qswuq_feshai wq t` num` p\mtym, fjyspeq m|l\
p|kir, ja pok Qswuqot]qyr. tq]vomtai cq p\mter oR !mhq~peioi m|loi rp
2mr toO he_ou7 jqate? cq tosoOtom bj|som 1h]kei ja 1naqje? psi ja peqic_-
metai.
B 2: di de? 6peshai t` <num`>, toO k|cou d 1|mtor numoO f~ousim oR pok-
ko r Qd_am 5womter vq|mgsim.
286 Enrique Hlsz
ars,13 though there are some exceptions. A line of interpretation going

back to John Burnet keeps regularly reappearing. According to this
view, we would be facing an expression that does not entail any partic-
ularly interesting philosophical overtones, the meaning of which would
be fairly simple, even trivial and standard. A common element in differ-
ent representatives of this view is the stress laid on the exclusiveness of
the generic meaning of language, a move that usually turns out in
translations of logos as Word, discourse, account, and the like.14
Oddly enough, this approach does not make that great a difference in
the long run, for Heraclitus logos is not, on any interpretation, just
any discourse, and soon becomes by sheer weight of its own contents
and meaningful cross-references to other concepts and images, a (if
not the) fundamental epistemological standard.
Now, without denying that the meaning of logos includes Heracli-
tus own account or discourse, reducing it to his Word (with a capital
w, as Burnet would have had it) just will not do. It is not necessarily mis-
leading to insist that one must translate logos as discourse or ac-
count, since it is expressly stated (here and elsewhere) that logos is
something heard (and not understood) by most men. But to claim
that this is all logos means is, again, something else. It seems much
more likely that the texts themselves support quite the opposite claim,
and that a purely linguistic reading15 of logos will not suffice to satisfy
obvious contextual requirements.
The very first sentence is syntactically ambivalent, in a way that per-
mits or even demands that both possible constructions are kept.16 If for-
ever (aQe_) is construed with the logos being this, or with the existing
13 Starting with Diels, adherents to this general view include W. Jaeger, R. Mon-
dolfo, M. Heidegger, O. Gigon, H. Frnkel, M. Marcovich, C. Eggers, and C.
H. Kahn, all in the bibliography below.
14 Cf., besides Burnet, M. L. West, C. Diano, T. M. Robinson, J. Barnes, and M.
Conche, all in the bibliography below.
15 Such as Wests (see above, n. 2, pp. 124 ff.) insistence that logos refers to Her-
aclitus discourse and nothing else (my italics).
16 As one can see, for example, in Kahns translation: Although this account
holds forever, men ever fail to comprehend (97). That the syntax of the first sen-
tence can work both ways (i. e. k|cou toOd 1|mtor aQe_ or aQe_ !n}metoi c_momtai
%mhqypoi ) is made clear by Aristotles famous complaint in the Art of rhetoric
(iii, 5, 1407b 11 ff.). See D. Sider, Word order and sense in Heraclitus: Frag-
ment One and the river fragment, in K. Boudouris (ed.), Ionian Philosophy
(Athens 1989) 363-368.
(or real) logos,17 even if we take this to mean merely that the discourse is
always true (taking the expression in the weaker veritative sense, rath-
er than in the fuller and more natural predicative sense, or in the stronger
existential one), this does not by itself exclude an intended deeper
meaning18. In fact, the language that Heraclitus uses rather suggests the
eternal existence of its objective contents: if he is saying, at the very
least, that his own account holds true throughout, this could very well
be due to his implicitly attributing irrestricted ontological permanence
to its proper object19 (there could even be an intentional, yet subtle,
play between the logos being (this) forever (fde b k|cor 1m aQe_) and
all things becoming according to it). Burnet may have been right that
logos should not be translated as reason (as a cognitive faculty), but
his interpretation of it as Word misses the objective sense, which
could be better conveyed by the French expression, raison dtre. Logos
as objective rationality pertains to the formal aspect of the real, the way
things are and happen, the structure in all change. Seen in this light,
and without being named, logos is mirrored in the eternal kosmos,
which is an ever-living fire (B 30), kindled and quenched in measures
(l]tqa). More obliquely perhaps, logos is implied in the river statement (B
12), the quintessence of which is identification and opposition of the
same rivers (potalo?si to?sim aqto?sim) and ever different waters
(6teqa ja 6teqa vdata). The common thought here is permanence in
change, sameness in difference, unity in plurality. All these formulae are
but variations that approach the same unitary and dynamic objective ra-
tionality.
If by logos Heraclitus meant only his own discourse, how is it, then,
that men are expected to understand it before having listened to it? 20
How could he say, not only that it is always the same or that it exists
17 Rather than with !n}metoi c_momtai %mhqypoi , although this needs not to be
excluded.
18 Especially since Heraclitus careful expression mimics one Homeric formula
that refers to the blissful existence of the gods, heo aQm 1|mter , Il. 1.290,
494; 21.518; 24.99; Od. 1.263, 378; 2.143; 3.147; 4.583; 5.7; 8.306, etc.
19 As he does in B30, where logos is likely to be implied in the measures, which
surely are as eternal as the world-order, the kosmos itself. As with v}sir, Hera-
clitus usage of j|slor is the first one attested in the history of philosophical lan-
guage.
20 So rightly argued by Kahn, H. 98.
288 Enrique Hlsz
forever21 and that all things happen according to it, if he did not
mean to point to some kind of objective cosmic principle, a single uni-
versal law, as suggested in B 114? And would not it be a bit unlikely
that, by calling the logos common (num|m, B 2), Heraclitus intended to
express only the universality and the formal unity of his own account,
but not the actual complex unity of its subject-matter? Questions like
these are perhaps among the reasons why Diels corrected himself trans-
lating Weltgesetz, Law of the world (appending this latter meaning
to the much simpler original Wort).22 In a different line, but likely
moved by similar reasons, Marcovich translated it as Truth, while
Kirk proposed the formula of things. Kahn, H. 22, interprets it as
being at once the discourse of Heraclitus, the nature of language itself,
the structure of the psukhe and the universal principle in accordance
with which all things come to pass.
Logos must include the generic notion of language, and specifically
Heraclitus own discoursive account as part of its surface meaning, but it
also must refer, as its proper object and the solid ground for true knowl-
edge, to the nature of things themselves, to being or reality (as the
text also suggest: cf. the parallel jat tm k|com, jat v}sim in B 1). It is
no mere coincidence that he insists that v}sir tends to be hidden,23
just as logos is said here to be, from the wits of most men. Ignorance
of the rationale of things themselves, omission of the ontologically
basic facts of logos and phusis lie behind his famous criticisms of venerable
poets as Homer (B 42, 56), Hesiod (B 40, 57, 106) and Archilochus (B
42), and representatives of Ionic Rstoq_g as Xenophanes (B 40), Pytha-
goras (B 40, 129) and Hecataeus (B 40). Hiddenness dialectically refers
precisely to what is manifest (vameq|m), and is best interpreted in an
epistemological context (rather than a purely ontological one).
21 As it is plain that I take aQe_ with the participle 1|mtor, my reading tends to be a
metaphysical one. If the minimalist view is prefered, the phrase would mean
simply that the logos as discourse is true [Burnet], or that it holds through-
out [West], with no ontological overtones.
22 Cf. Diels, Die Fragmente der Vorsokratiker 1,2 (Berlin 1903, 1906) 1, ch. 12: Fr
dies Wort. Alles geschieht nach diesem Wort. Weltgesetz was added in
Diels second edition of Herakleitos von Ephesos (Berlin 1909).
23 DK 22 B 123: v}sir jq}pteshai vike?. I owe to Daniel Graham the point that
vike? does not mean here loves, but (in view of the infinitive) something like
it is usual for. See his Does nature love to hide? Heraclitus B 123 DK, CP
98 (2003) 175 179. Compare with B 87, n. 29 below.
For Heraclitus, logos refers to the rational aspect of reality, to objec-

tivity (in an ontological as well as in an epistemological sense), and it is
thus fair to call it a principle of being. This strong objective sense is re-
inforced by semantic closeness of logos and metron, as shown by the ex-
pression letq]etai eQr tm aqtm k|com, is measured [sc. either earth
or sea] in the same proportion, in B 31b (a text in which the surface
meaning seems to have pushed the notion of language to a remote back-
ground). Metron is a key notion in the cosmic order, which is ever-liv-
ing fire (B 30). In fact, and even leaving aside the enlightening analogy
with the single divine law in B 114,24 a high ontological status of logos is
also backed by B 2, which closes the Proem by calling the ever-real logos
common or shared (surely by all things).25 Logos is opposed to most
mens fancy, ironically refered to as a wisdom of sorts, which is pri-
vate or peculiar to each (Qd_g vq|mgsir), an oxymoron which imme-
diately recalls the analogical image of men as oblivious sleepers.26 So the
logos in the Proem presents the reader with a characterization of human
life in epistemic terms, as Kahn, H. 100 aptly puts it. The background
seems more metaphysical (ontological) than cosmological (in a narrow
physicalistic sense), and the foreground could be perceived as epistemo-
logical and ethical at the same time.
By giving expression to the notion of the rationality of the real,27
Heraclitus points at once to a vital unity of being and speech, word
24 An analogy clearly refered to and greatly appreciated by Cleanthes (perhaps so

much that he not only took it over but considerably enriched it) in his Hymn to
Zeus. Although analogy of logos and nomos is particularly visible, it is by no
means the only significant point of affinity. It is worth recalling line 21 (sh
6ma c_cmeshai p\mtym k|com aQm 1|mta), which clearly takes on Heraclitus B
1, and supports the reading proposed here. Logos was fused with Fire and
Zeus in the Stoic tradition from Cleanthes on. On these questions, see A. A.
Long, Heraclitus and Stoicism, in id. (ed), Stoic Studies (Berkeley 1996)
35 57.
25 Note, besides the formula t` num` p\mtym in B 114, the echoes of k|cou toOd
1|mtor aQe_ (B1), and toO k|cou d 1|mtor numoO (B 2) in B 80s tm pkelom
1mta numm.
26 B 89 connects the sleepers with private worlds of their own. With Marcovich, I
fear that Plutarch, the source here, could be freely paraphrasing B 1 and 30.
27 I submit that the Stoic view on Heraclitus logos is, on the essentials, correct. I
believe that A. A. Long has got it right when he writes that [i]t is thoroughly
misleading to label Heraclitus logos metaphysical and that of the Stoics moral.
In both systems logos is a principle of being and a principle of morality. [] In
calling their own active power in the universe logos the Stoics were expressing
290 Enrique Hlsz
and content, and includes the ontological structure or nature of the

contents of language itself within the semantic range of logos. Heraclitus
choice of words suggests that remedying mens need for knowledge
should begin with a fuller understanding of their own language. To
say it again in Kahns words: logos means not simply language but ra-
tional discussion, calculation, and choice: rationality as expressed in
speech, in thought and in action (H. 102). Logos is simultaneously a
metaphysical and an anthropological principle. It is not just an accident
that, in what looks like a rare praise, Bias of Priene is credited with a
greater or better logosusually translated as esteemthan the
rest.28 Nor is it trivial to state that for a stupid man it is usual to be
amazed at every logos,29 or, again, that logoi heard by Heraclitus are
cut off from the wise.30 Heraclitean explicit usage (in all nine fragments
containing the term) reveals an awareness of the tension between speech
and wisdom, since not every logos is in fact wise. At the same time, in-
telligent speech is endowed with the power of making phusis manifest.
Heraclitean utterance is closely shaped after the oracular Delphic model:
Apollo, who represents the highest standard, says not, neither hides,
but gives a sign.31 The twin themes of speaking and hearing give
logos a concrete texture, and develop into fuller ideas (involving much
more than mere perception), rich in epistemic and moral nuances:
They [most men?] know not how to hear, nor to speak (B 19),
and like the deaf, though present, they are absent (B 34). Universal
the closest affinity with Heraclitus (above, n. 24) 51. See also Long in this vol-
ume.
28 B 39: 1m Pqim, Bar 1cmeto b Teutley, ox pkym kcor C t_m %kkym. It is not
at all clear who these others specifically are: the citizens of Priene, Greeks
generally, perhaps the other six sages? Approval of at least one of Bias moral
gnomai is explicit in B 104, so it seems that, in spite of the usual meaning of
the Ionic expression pkym kcor (= higher esteem), logos must include
what Bias himself said, and he must have suceeded in showing things as they
truly are.
29 B 87: bkn %mhqypor 1p pamt kc\ 1pto/shai vike?. This deals with the re-
ceptive aspect of the process of acquiring knowledge, implicitly contrasting
hearing with the more creative abilities of speaking and acting.
30 B 108: bjsym kcour Ejousa, oqder !vijme?tai 1r toOto ste cimsjeim fti
sovm 1sti pmtym jewyqislmom. I take pmtym as masculine and referring
back to bjsym. The targets would seem to naturally include some great figures
named in other fragments, such as Homer, Hesiod, Archilochus, Hecateus,
Xenophanes, and Pythagoras.
31 B 93: b %man ox lamte?m 1sti t 1m Dekvo?r oute kcei oute jqptei !kk sgla-
mei.
presence of logos as a voice men constantly hear is implicitly paired with

its absence from their conscious minds: that with which they deal con-
tinuously, this they differ from, and things they encounter every day
seem alien to them.32
The universal ontological dimension of logos becomes clearer still
when we look at B 50:
Having listened, not to me, but to the logos, it is wise to say in agreement
(blokoce?m): all things are one33.
The first point of interest here is the distinction between Heraclitus
own words and the logos men should heed. Logos is not reducible to
speech, not even to that which is spoken of, but suggests a voice and
a continous presence, perceived but unrecognized by most men (B
72, 17). We have in B 50 a characterization of the attitude of the intel-
ligent speakers (legontes) of B 114, which is opposite to that of the sleep-
ers. Not surprisingly, here we find that (true) wisdom lies in men speak-
ing together with one another and saying the same thing (blokoce?m) that
logos utters once and again, the oneness of all things (4m pmta). A couple
of other fragments (B 41, 32) contain the expression 4m t sov|m, which
joins immediately unity and wisdom34.
Heraclitus logos is not only a complex rational concept, but a pow-
erful, yet simple image, almost a metaphor. So, instead of choosing be-
tween language and the well-ordered nature of the real as a single
basic meaning, necessarily excluding one of the two, I propose we
32 B 72: lkista digjem_r blikoOsi (kc\ t` t fka dioijoOmti), tot\ dia-

vqomtai, ja oXr jah Blqam 1cjuqoOsi, taOta aqto?r nma vametai. Probably
not a verbatim quotation in its entirety (it is likely that the words in the paren-
thesis, as well as the final formula, are at best paraphrases).
33 B 50: oqj 1loO !kk toO kcou !josamtar blokoce?m sovm 1stim 4m pmta
eWmai.
34 B 41: 4m t sovm7 1pstashai cmlgm btg jubeqm/sai pmta di pmtym.
B 32: 6m, t sovm loOmom, kceshai oqj 1hkei ja 1hkei Fgmr emola. Both
texts seem to allude to logos, though in a very elliptical manner: in the first
case, through the universality and immanence of the gnome, and through the
passive form legesthai, which is the object of wants not and wants, in the sec-
ond. The theme of unity is itself as intricate as that of logos. No fewer than ten
fragments refer explicitly to t 6m: besides those already refered to, cf. B 10
(closely echoing B 50), and B 29, 33, 49, and 121, all of which exploit the po-
litical side of the idea of unity. Unity is typically viewed as synthesis in Hera-
clitus, and is expressed with the term harmonie, used in B 51 and 54. See E.
Hlsz Piccone, La unidad de la filosofa de Herclito Tpicos 28 (2005)
13 46.
292 Enrique Hlsz
ought to keep both, and interpret Heraclitus logos as the language of the
real35: the language of objective rationality and the rationality of lan-
guage. Logos appears as a voice36 coming from the real itself, the voice
of meaningful and intelligent language, which coincides with the struc-
tural objective single form in things themselves and is mirrored in Her-
aclitus own carefully articulated statements.
Unity of all things (4m pmta), as B 50 puts it, is a proper character-
ization of the content of the same logos which is called in the Proem for-
ever real and common or shared by all things, and which is itself a unity
of opposites. Heraclitean logos is, first and foremost, the universal ration-
ale of reality: it is a permanent, formal principle, a sort of law that pertains
to reality itself, conceived of and presented as a rational language. The
explicit notion of a community of all things through logos (and thus
of all men, and of all they say and do) refers to the togetherness or in-
tegrity of being, in itself and as presented in true language. Communion
of intelligent language with the intelligible nature of things is indeed the
suggested basis for true knowledge on the level of human action, and
the reason why logos is a prescriptive, and not merely a descriptive prin-
ciple: being and action are connected through logos as the rational lan-
guage of the real.
Heraclitus criticism of mens epistemic negligence is based on his
attributing them a universal and natural capacity to understand37,
which is nevertheless seldom achieved: even the senses can be deceptive,
but only because the data they provide can be incorrectly interpreted by
their soulswhich in turn must be seen as seat of the faculty of under-
standing: Eyes and ears are bad witnesses, he writes elsewhere, for
men who have barbarian souls.38 A barbarian soul is, of course, one
that does not speak the language of the real, and whose defective dis-
course or account is, because of its being irrational, thus separated
from the wise (B 108). One important point Heraclitus is making
here is that, for good or ill, the human soul is the natural seat for the
35 Cf. T. M. Robinson, Heraclitus. Fragments (Toronto 1991) and id., Heraclitus

and logosagain, in E. Hlsz (ed.), Nuevos ensayos sobre Herclito (Mexico City
2009) 93 102.
36 Cf. C. Eggers Lan, La doctrina heracltea del logos, Nova Tellus 5 (1987) 11 ff.
37 With which he explicitly credits all men in B 113, echoed also in B 116. See
below nn. 51 52.
38 B 107: jajo lqtuqer !mhqpoisim avhaklo ja ta baqbqour xuwr 1wm-
tym.
faculty of apprehending, understanding and expressing the rationality

and true nature of things.
Heraclitus on Logos and Self: Some Ethical Fragments
An essential link between the human39 soul and logos is focused in B 45:
He who travels every path would not find the boundaries of soul, [as he
goes]: so deep a logos does it have.40
Immediate connection between the soul and logos is apparent, a point
that strenghthens that rationality (and not just language) comes into
play here. Many interpreters have been baffled by the curious paradox
about the souls limits and its suggested virtual unlimitedness or immen-
sity: if Heraclitus is saying that the soul is limited, then why is it that its
boundaries could not be reached? And if he is saying that soul is, in
some sense, infinite, how are we to understand the positive reference
to its peqata? Puzzlement is probably induced, not just by Heraclitus
unmistakable paradoxical style, but also by the readers (understandable)
eagerness to find a sort of definition of the souls nature. But Heraclitus
is perhaps not concerned here (is he ever?) with solving the definitional
aporiai that might trouble his readers or hearers, but with making a point
about the presence of logos (again best interpreted as a rational principle),
precisely within the soul of man. He is succeeding to set down, in a short
string of words, the limits of the souls cognitive power. The souls na-
ture is not exhausted here, but is displayed in a set of other fragments. In
particular, B 3641 implies the souls mortality, and a primarily (though
39 Among the possible hermeneutic scenarios, I favor an anthropological one,

which seems the most likely. This line of interpretation is compatible with
more cosmological and physicalistic approaches of Heraclitus views on the
soul.
40 B 45: xuw/r peqata [Qm] oqj #m 1neqoi, b psam 1pipoqeulemor bdm7 ovty
bahm kcom 5wei. I assume the alternative reading to Diogenes Laertius text,
proposed by Gbor Betegh in The limits of the soul: Heraclitus B 45 DK.
Its text and interpretation, in E. Hlsz (ed.), Nuevos ensayos sobre Herclito
(Mexico City 2009), 391 414. For yet another reading see J. Mansfeld, Her-
aclitus Fr. 22 B 45 D. K. A Conjecture, Elenchos 31 (2010) 117 122, who,
relying on B 18, proposes oqj 1neuq^sei instead of oqj #m 1neqoio.
41 B 36: xuw0sim hmator vdyq cemshai, vdati d hmator c/m cemshai7 1j c/r d
vdyq cmetai, 2n vdator d xuw^. (For souls it is death to become water, for
294 Enrique Hlsz
perhaps not exclusively) cosmological context seems likely there. Al-

though I will not go into the matter, the allegedly Heraclitean doctrine
of the soul as fire does not seem to be backed by the fragments them-
selves (and it would be hardly useful anyway for gaining new insight
into the meaning of B 45). The connection with selfhood seems
more promising.
The background image seems to be an alternative description of the
personal experience Heraclitus alludes to in B 101: I went in search for
myself42. If identity of soul and self is assumed,43 this would then be a
journey one takes within ones soul. The whole territory to be traversed
must be soul itself, and the imaging of the road and the unsuccessful
search seems to unfold at first in a spatial, horizontal dimension. As
the traveller surely must be the owner of a soul, the drift is that the jour-
ney within oneself is a cognitive reflexive experience, not to be ascribed
to just any soul,44 but to the intelligent one. Now, the point of this hor-
izontal approach is that the limits are not to be found. Is there a moral
lesson in here? If some prescriptive intention is present, it cannot be
not to search, but quite the contrary. Comparison with other themati-
cally akin texts on searching, as for instance B 18 (He who does not
expect, will not find the unexpected, for it is hard to find and to under-
stand45) and B 22 (Those who search for gold dig much earth and find
little),46 suggest that Heraclitus point is the difficulty, not the radical
impossibility, of the souls cognition. The ethical impact of the cogni-
water it is death to become earth. But water is born from earth, and from water,
soul.)
42 B 101: 1difgslgm 1leyutm. This was sometimes interpreted in antiquity (e. g.,
Diogenes Laertius 10.5) as meaning that Heraclitus was self-taught. But more
could be going on here: the traditional idea of self-knowledge is essentially
linked to sophrosune, which is precisely the virtue praised in B 112. The sentence
also mimics the delphic cm_hi saut|m. To know oneself is, of course, not just
psychological self-awareness, but requires being able to understand what one is
and is not, and so to know ones proper place. See Julia Annas exploration of
Heraclitean self-knowledge in Alcibiades I, Lovers, and Charmides: Self-knowl-
edge in Early Plato, in D. OMeara (ed.), Platonic Investigations (Washington
1985) 111 138.
43 A point denied by Marcovich, H. 57.: [] neither d_fgshai means the same as
cim~sjeim, nor 1leyutm as xuw^.
44 The restrictive character of the interpretation was suggested to me by Conches
text. Contrast with B 116, which I take to refer to self-knowledge as a universal
faculty, not a universal accomplishment.
45 B 18: 1m l 5kpgtai !mkpistom oqj 1neuqsei, !meneqemgtom 1m ja %poqom.
46 B 22: wqusm cq oR difllemoi c/m pokkm aqssousi ja erqsjousim akcom.
tive search gains in clarity if the soul is to be thought of at once as the

object being searched for, as well as the active searching subject.
The final point changes the readers perspective once more: limits
are out of reach, and at the same time, are determined precisely as
such, because of the deep logos that the soul has. Depth of logos sug-
gests a vertical47 dimension: and even if limits are, for every person,
impossible to reach, logos must be at the beginning and at the end (B
103) 48 of every path one can take, whatever its direction (B 60: %my
jty, up and down),49 just as it must be both within and without.50
So the upshot of all this is that xuw^, in spite of its limitedness, can in
principle come to know of all things by virtue of the logos within (for
logos is, as already we have seen, universal or common to all things
soul itself included). This line of interpretation is consistent with what
we find in other relevant fragments, which together put xuw^ in an in-
tensely anthropological scenario. It will be noticed that the picture I am
presenting amounts to a basically optimistic view, both ethically and
epistemologically, to which someone could object that it seems to flatly
contradict Heraclitus repeated condemnation of mens epistemic negli-
gence. But this incoherence is merely apparent, and not real. For it is
not all men, not always, and not fatally, who are so vigorously critized
by Heraclitus (but merely most men). When he explicitly refers to
all men, what he actually says is thinking is common to all51 and
all men partake in knowing themselves and being of a sound mind52.
Now, from the viewpoint of this interpretive sketch of B 45, B 115
makes perfect sense:
47 I take the suggestion of focusing on the image of B 45 through a double view-

point, consisting in a horizontal axis and a vertical one, mutually complemen-
tary, from Marcovichs commentary ad loc in H.
48 B 103: numm cq !qw ja pqar 1p jjkou.
49 B 60: bdr %my jty la ja ut.
50 For other suggestive interpretations of B 45, see A. Lebedev, Psyches peirata (Il
termine psyche nei frammenti cosmologici di Eraclito 66 67 M (17 26), M.
Ghidini Tortorelli, Limite e circolarit in Eraclito: frr. 45 e 103 D.-K. (27
46), and P. Rosati, Eraclito, in cammino, Intorno al fr. 60 D.-K. (47 62),
all in M. Capasso, F. De Martino, P. Rosati (eds.), Studi di filosofia preplatonica
(Naples 1985).
51 B 113: numm 1sti psi t vqome?m.
52 B 116: !mhqpoisi psi ltesti cimsjeim 2autor ja syvqome?m.
296 Enrique Hlsz
To the soul belongs a logos which increases itself.53

The limits of soul are out of reach, for they are unstable and changing.
And how could it be otherwise, since the logos on which they depend is
said to grow by itself ? The metaphor of growth (affecting both logos and
soul) must stand for the progress of an intelligent souls self-knowledge
but, at the same time, also for the souls ontological structure, its distinc-
tive nature, or v}sir. Marcovichs philosophical objection to the authen-
ticity of B 115that the notion of a changing measure is implausible in
Heraclitusis not convincing: although it is true that Heraclitus theory
implies that measure is something constant, it is a bit too much to reduce
it to a numerical ratio.54 And, besides that, would it not be an argu-
ment in favour of authenticity that the bold idea of a permanent stan-
dard of change that constantly changes by itself (2autm aunym) is the su-
preme paradox? It is tempting to connect the logos that increases itself of
B 115 with the sentence in B 84a, which lacks an explicit grammatical
subject: it rests by changing.55
Thus, both B 45 and 115 would point to the specific presence of
logos in the anthropological dimension: all things change according to
logos, but natural (non-human) change, although analogous to human
becoming, would not, after all, be identical with it. To derive again
from the originary examples, Heraclitus says that GGkior, the Sunitself
an important paradigm for the rationality of change, by its being fiery
and always new (B 656)will not overstep his measures.57
53 B 115: xuw/r 1sti kcor 2autm aunym. I cannot make the case for the authen-
ticity of B 115 fully here. It may be summarily laid down, however, that the
onus probandi falls rather on the side of those who deny it is genuine (the source
here is Stobaeus, who also provides the whole series B 108-B 119). Explicit at-
tribution to Socrates (by some unattentive scribe) is not an unsurmountable ob-
jection.
54 I set this apart from his other two objections, which are of a different nature. Cf.
Marcovich, H. 569 ff.
55 B 84a: letabkkom !mapaetai, the first part of the only DK fragment transmit-
ted by Plotinus.
56 I must say that I assume a different reading of B 6, one of the solar fragments.
I propose that what Heraclitus wrote includes the expression always new (b
Fkior oq l|mom jah\peq Jq\jkeit|r vgsim, m]or 1v Bl]q, 1st_m, !kk !e m]or). If
one could follow A. Garca Calvos suggestion (above, n. 1), perhaps the addi-
tion of [ja utr] would be a fitting end. The result would be something like
this: the Sun (so Heraclitus says) is not only new every day, but new always
[and the same]. On my view, Heraclitus could be criticizing a theory (if theory
it could be called) attributed independently to Xenophanes, a known target in
Human conduct, on the other hand, is nothing like this: by his very na-
ture, man is constantly overstepping his measures: ~bqir is a recurring
fact everywhere in human life, either in a political context, as wanton
violence,58 or in the individuals everyday existence (as in the drunkard
with a wet soul in B 11759). Human unreasonableness has a place in the
cosmic order of Heraclitus.
But again, so does the possibility of true excellence, even if this will
be restricted to a few, the true %qistoi.60 In my view, Heraclitus is an eth-
ical optimist at heart, his harsh criticism of his fellow citizens61 and his
well-known unflattering views on men in general notwithstanding.
For, all things considered, those who speak with intelligence, the owners
of non-barbarian souls with ever-changing limits are there too, side by
side the unintelligent hearers. And perhaps even they can change, and
learn to think well, see, hear, speak and act. B 112 wraps it up neatly:
Being of a sound mind is the greatest merit and wisdom: to say what is true
and to act according to nature, as those who pay heed.62
Supreme virtue and wisdom are equated with sound thinking, which
consists of the unity of speech and action. As in B 1, the adverbial phrase
jat vsim is in apo koinou construction, and can be taken as modifying
the three surrounding verbs (kceim, poie?m and the participle 1paomtar),
strongly evoking the language of the proem. Heraclitus moral ideal is
thus quite philosophical, and it could be fairly characterized as an ethical
intellectualistic theory: moral virtue is deeply rooted in knowledge and
the nature of things themselves, including men. One important feature
of this ideal is that it seems feasible, reasonable and earthly, as is befitting
another authentic fragment (B 40). See my Heraclitus on the sun, in R. Pat-

terson, V. Karasmanis, and A. Hermann (eds.), Presocratics and Plato. A Festschrift
in honor of Charles H. Kahn (Las Vegas 2012) 31 52.
57 B 94: GGkior cq oqw rpeqbsetai ltqa.
58 B 43, tr. after Kahn.
59 B 117: !mq bjtam lehush0, %cetai rp paidr !mbou svakklemor, oqj
1paym fjg bamei, rcqm tm xuwm 5wym.
60 That is, not the so-called best in B 29, but the best in terms of their wis-
dom, not their birth, the bearers of syvqome?m, being of a sound mind, itself
identified as supreme merit (!qet^) and wisdom (sovg) in B 112.
61 B 121, and implicitly, many others of which the clearest are B 29, 49 and 104.
For a different reading of oR d pokko in the second clause of B 29, see Sider in
this volume.
62 B 112: syvqome?m !qet lecstg ja sovg7 !kgha kceim ja poie?m jat vsim
1paomtar.
298 Enrique Hlsz
to the author of the famous saying: Mans character is his fate (B

119)63. So, in symmetric contrast with the wet soul of the drunkard,
temporarily devoid of the most elementary functions (such as being
able to walk or just to stand up and having a sense of direction), Her-
aclitus points us to the virtues of just being sensible. In Heraclitus uni-
verse theres also room, somewhere, for a dry soul (one endowed with a
divine ethos, B 7864). This is the best, as it is the wisest. But Heraclitus
does not stop there. He also calls this, the best and wise soul, a beam of
light.65
This last and very visual image might be somehow connected with
the auditive image of logos through B 16:
How could someone be unaware of that which never sets? 66
Since this question illustrates so well the nature of logos itself, I suggest
that it complements Heraclitus implicit imperative (I mean something
like %joue k|cou, listen to logos67). Do nothe seems here to tell
usturn a blind eye to that which is always present. His intended
point is here that we picture something like a hypersun that never
sets, a source of light that never fails. And what, better than logos,
could claim to be the ultimate foundation for intelligibility and enlight-
enment? The philosophical unity of Heraclitus logos presents, in addi-
tion to its diverse meanings, which we already touched upon earlier,
an ethical (and perhaps also a political) dimension. And, inasmuch as
logos includes the idea of measure and the rational and unifying relation-
63 B 119: Ghor !mhqp\ dalym. I read into this Heraclitus recognition of moral
autonomy and resposibility of the individual.
64 B 78: Ghor cq !mhqpeiom lm oqj 5wei cmlar, he?om d 5wei. This I take as a
good example of diairesis kata phusin, which focuses on internal opposition with-
in (human) ethos.
65 B 118: aqc ngq xuw sovyttg ja !qstg The basic meaning of aqc^ is
light of the sun (LSJ s. v.). It could perhaps mean here just a bright
light. I follow Kahn in keeping Stobaeus text. In all fairness, it should be
said that there are alternative readings to that proposed here. Another possibility
would be A dry gleam of light is the wisest and best soul, but then what
might a dry gleam of light mean?
66 B 16: t l dOmm pote p_r %m tir khoi. The sentence goes well with B 17: oq
cq vqom]ousi toiaOta pokko_, bj|soi 1cjuqeOsim, oqd lah|mter cim~sjousim,
2yuto?si d doj]ousi. (For many men does not think straight about such
things as they come across, nor do they know after having learned them, but
fancy themselves they do).
67 This implicit imperative Ive modelled after Hesiods formula %joue d_jgr,
heed justice (Erga 213, cf. 275 1p\joue d_jgr ).
ship between things, it operates also as a sort of poetic principle. All this
has to do with the internal physiology of being, thought, language and
action in Heraclitus complex metaphysics. Not only is logos a funda-
mental metaphysical concept, but it is at the centre of Heraclitus
view of man as an ethicaland a politicalbeing.
Bibliography
Adomenas, Mantas.The fluctuating fortunes of Heraclitus in Plato, in A. Laks
and C. Louguet (eds.), Quest-ce que la philosophie prsocratique? / What is Pre-
socratic philosophy (Villeneuve-dAscq 2002) 419 447.
Annas, Julia. Self-knowledge in early Plato, in D. OMeara (ed.), Platonic In-
vestigations (Washington 1985) 111 138.
Barnes, Jonathan. The Presocratic Philosophers. London 1979.
Betegh, Gbor. The limits of the soul: Heraclitus B 45 DK. Its text and inter-
pretation, in E. Hlsz (ed.), Nuevos ensayos sobre Herclito (Mexico City
2009), 391 414.
Burnet, John. Early Greek Philosophy. 4 London 1930.
Capasso, Mario, F. de Martino, F., and P. Rosati (eds.). Studi di filosofia prepla-
tonica. Naples 1985.
Colli, Giorgio. La sapienza greca, III. Eraclito. Milan 1980.
Conche, Marcel. Hraclite. Fragments. Paris 1986.
Diano, Carlo, and G. Serra. Eraclito. I frammenti e le testimonianze. Venice 1980.
Diels, Hermann. Die Fragmente der Vorsokratiker. 1 4 Berlin 1903 22.
Herakleitos von Ephesos. 2 Berlin 1909.
Eggers Lan, Conrado. Los filsofos presocrticos. Madrid 1978.
La doctrina heracltea del logos, Nova Tellus 5 (1987) 9 18.
Frnkel, Hermann. A thought pattern in Heraclitus, in A. P. D. Mourelatos
(ed.), The Pre-Socratics. A Collection of Critical Essays (New York 1974)
214 228.
Dichtung und Philosophie des frhen Griechentums. Munich 1962.
Gadamer, Hans-Georg. Der Anfang der Philosophie. Stuttgart 1996 [The Beginning
of Philosophy, English tr. by R. Coltman. New York 1998].
Garca Calvo, Agustn. Razn comn. Edicin crtica, ordenacin, traduccin y comen-
tario de los restos del libro de Heraclito [sic]. Madrid 1985.
Ghidini Tortorelli, Marisa. Limite e circolarit in Eraclito: frr. 45 e 103 D.-
K., in Capasso, de Martino, and Rosati 27 46.
Gigon, Olof. Des Ursprung der griechischen Philosophie. Basel, 1947 [Los orgenes de
la filosofa griega. De Hesodo a Parmnides (Spanish tr. by M. Carrin). Ma-
drid 1980].
Graham, Daniel W. Does nature love to hide? Heraclitus B 123 DK, CP 98
(2003) 175 179.
Hadot, Pierre. Quest-ce que la philosophie antique? Paris 1995.
Heidegger, Martin. Einfhrung in die Metaphysik, in Gesamtausgabe, Band 40.
Frankfurt 1935.
300 Enrique Hlsz
Altheia (Heraklit, Fragment 16) (Eng. tr. as Aletheia (Heraclitus, Frag-

ment B 16), ch. iv of Early Greek Thinking. The Dawn of Western Philoso-
phy. Tr. D. F. Krell and F. A. Capuzzi. San Francisco 1984.)
Logos (Heraklit fr. 50 DK), in Gesamtausgabe, I, Abteilung, Bd. 7: Vortrge
und Aufstze (Frankfurt 1954) 211 233. (English translation as Logos. Her-
aclitus, Fragment B1, chap. ii of Early Greek Thinking. The Dawn of Western
Philosophy. Tr. David Farrel Krell & Frank A. Capuzzi, San Francisco, 1984.
Hlsz Piccone, E. Scrates y el orculo de Delfos (una nota sobre Platn, Apol-
oga de Scrates 20c4 23c1), Theora 14 15 ( June 2003) 71 89
Anmnesis en el Menn platnico, Apuntes filosficos 22 (2003) 61 79.
La unidad de la filosofa de Herclito, Tpicos 28 (2005) 13 46.
La imagen de Herclito en el Cratilo y el Teeteto de Platn, in E. Hlsz
(ed.), Nuevos ensayos sobre Herclito (Mexico City 2010) 361 389.
Lgos. Herclito y los orgenes de la filosofa. Mexico City 2011.
Heraclitus on the sun, in R. Patterson, V. Karasmanis, and A. Hermann
(eds.), Presocratics and Plato. A Festschrift at Delphi in honor of Charles Kahn
(Las Vegas, 2012) 31 52.
Platos Ionian muses: Sophist 242d-e, in B. Bossi and T. M. Robinson
(eds.), The Sophist Revisited (Berlin 2012) 98 106.
Irwin, Terence. Platos Heracliteanism,PQ 27 (1977) 1 13.
Jaeger, Werner. The Theology of the Early Greek Philosophers, The Gifford Lectures
1936. Oxford 1936. (Spanish tr. as La teologa de los primeros filsofos. Mexico
City 1980. Tr. J. Gaos).
Paidea, Die Formung des griechischen Menschen, Vol. 1. Berlin 1933; Spanish tr.
as Paideia: los ideales de la cultura griega, libro I. Mexico City 1973. Tr. W.
Roces-J. Xirau.)
Lebedev, Andrei. Psyches peirata (Il termine psyche nei frammenti cosmologici
di Eraclito 66 67 M, in Capasso, de Martino, and Rosati 17 26.
Long, Anthony A. Heraclitus and Stoicism, in id. (ed), Stoic Studies (Berkeley
1996) 35 57.
Mansfeld, Jaap. Heraclitus Fr. 22 B 45 D. K. A conjecture, Elenchos 31
(2010) 117 122.
Mondolfo, Rodolfo. Herclito. Textos y problemas de su interpretacin.4 Mexico
City 1966.
Mouraviev, Serge N. Heraclitea, III. 3. i-iii., Sankt Augustin, Academia, 2006.
Heraclitea, IV. A, Refectio. Liber ut ad nobis restitutus, Sankt Augustin, Aca-
demia, 2011.
Robinson, Thomas M. Esiste una dottrina del logos in Eraclito?, in L. Rossetti
(ed.), Atti del Symposium Heracliteum 1981 (Chieti 1981) 1.65 72.
Heraclitus. Fragments. A text and translation with a commentary. Toronto 1991.
Heraclitus and Plato on the language of the real, The Monist 74 (1991)
481 490.
Heraclitus and logosagain, in E. Hlsz (ed.), Nuevos ensayos sobre Her-
clito (Mexico City 2009) 93 102.
Rosati, Pierpaolo. Eraclito, in cammino. Intorno al fr. 60 D.-K. in Capasso,
de Martino, and Rosati 47 62.
Sider, David. Word order and sense in Heraclitus: Fragment One and the river
fragment, in K. Boudouris (ed.), Ionian Philosophy (Athens 1989) 363-368.
Tarn, Leonardo, The first fragment of Heraclitus, Illinois Classical Studies,
XI, 1986, 1 15
11. Once More unto the Stream
Daniel W. Graham
There is a debate that has been going on for the better part of a century
now, as to what Heraclitus utterances about the river are, and what they
mean. The traditional view accepts Diels fragments 12, 49a, and 91 as
different statements about the river, that in general support the view of
Plato, Aristotle, and the ancient tradition that for Heraclitus all things
are in flux like a river, and nothing remains constant. But in 1916 in
a book on Parmenides, Karl Reinhardt argued that a close look at B
12 showed that Heraclitus point was quite different: instead of stressing
the transience of things, he was arguing the constancy of change.1 With
variations, his view has been followed by a number of leading scholars,
including Geoffrey Kirk and Miroslav Marcovich, both of whom accept
only B 12 as a genuine statement of Heraclitus.2 Other scholars, how-
ever, have continued to accept one or both of the other alleged frag-
ments and to defend a radical flux reading.3
As ever, the debate comes down to this: can one, on Heraclitus
view, step twice into the same river? The traditional interpretation
1 K. Reinhardt, Parmenides und die Geschichte der Griechischen Philosophie (Bonn

1916) 206 207 et passim.
2 G. S. Kirk, Natural change in Heraclitus, Mind 60 (1951) 35 42; Kirk, H.
367 380; Marcovich, H. 194 214. Cf. also A. Rivier, Un emploi archaque de
lanalogie chez Hraclite et Thucydide (Lausanne1952) 10. Kirk also accepts B 91
(b), but not the problematic B 91(a), which will be discussed below.
3 G. Vlastos, On Heraclitus, AJP 76 (1955) 337 78; R. Mondolfo, I fram-
menti de fiume e il flusso universale in Eraclito, Revista Critica di Storia della
Filosofia 15 (1960) 3 13; K. Popper, Kirk on Heraclitus, and on fire as the
cause of balance. Mind 72 (1963) 386 92; Guthrie HGP 1, ch. 7; M. C.
Stokes, One and Many in Presocratic Philosophy (Washington, D.C. 1971),
ch. 4; J. Bollack and H. Wismann, Hraclite: ou la sparation, Paris 1972; C. J.
Emlyn-Jones, Heraclitus and the unity of opposites, Phronesis 21 (1976)
89 114; J. Barnes, The Presocratic Philosophers (London 1982), ch. 4; M. Con-
che, Hraclite: fragments (Paris 1986) 459. T. M. Robinson, Heraclitus, Toronto
1987, accepts B 49a but is skeptical about B 91(a). For a good statement of the
revisionary view, see A. Petit, Hraclite: La captation de la marge, tudes Phi-
losophiques (1988) 207 19.
304 Daniel W. Graham
says no, the revisionary interpretations says yes. Recently, Leonardo

Tarn has criticized the revisionary view of Reinhardt, Kirk, and Mar-
covich, and put forward a new reading of the evidence, broadly exon-
erating the traditional interpretation.4 To my knowledge, this is the only
strong defense of the traditional reading to be published in the last twen-
ty years. I have elsewhere defended the revisionary interpretation or
something like it,5 and I here wish to provide a reply to Tarns article,
which, though admirably clear and forceful, seems to me to create more
problems than it solves. In my argument I shall (I) present Tarns argu-
ment, (II) raise challenges for it, (III) interpret B 12, and (IV) show how
on this interpretation Heraclitus has a coherent and defensible theory of
flux. In addressing this problem I am conscious of stepping into a stream
of ever-changing waters, but I believe there are some landmarks that
will get us safe across.
Here are some of the more important statements and reports about the
river:6
T1 [B 12] potalo?si to?sim aqto?sim 1lba_mousim 6teqa ja 6teqa vdata
1piqqe?.
On those stepping into rivers staying the same other and other waters flow.
(Eusebius, Praeparatio Evangelica 15.20; Arius Didymus fr. 39.2)
T2 [A 6] Heraclitus, I believe, says that all things pass and nothing stays, and
comparing existing things to the flow of a river, he says you could not step
twice into the same river [dr 1r tm aqtm potalm oqj #m 1lba_gr] (Plato
Cratylus 402a)
T3 According to Homer, Heraclitus, and all their tribe, all things are in
motion like streams. (Plato Theaetetus 160d)
T4 [Cratylus] finally decided that he could not express any thought in
speech but only wiggled his finger; and he criticized Heraclitus for saying
4 L. Tarn, Heraclitus: The River-fragments and their implications, Elenchos 20

(1999) 9 52. Repr. in id. Collected Papers (1962 1999) (Leiden 2001) 126
167.
5 Cf. D. W. Graham, Heraclitus criticism of Ionian philosophy, OSAP 15
(1997) 1 50, where, although I argue for a third position, my position is closer
to the revisionary view; see Graham, Explaining the Cosmos: The Ionian Tradition
of Scientific Philosophy (Princeton 2006), ch. 5.
6 My own translations throughout, except as otherwise noted.
11. Once More unto the Stream 305
that it is not possible to step twice into the same river. For he thought it was
not possible to step in even once! (Aristotle Metaphysics 1010a12 15)
T5 [B 49a] potalo?r to?r aqto?r 1lba_mol]m te ja oqj 1lba_molem, eWl]m te
ja oqj eWlem.
Into the same rivers we step and do not step, we are and are not. (Heraclitus
Homericus Homeric Questions 24)
T6 [B 91] (a) It is not possible to step twice into the same river [potal`
cq oqj 5stim 1lb/mai dr t` aqt`] according to Heraclitus, or to come
into contact twice with a mortal being in the same state. But by its swiftness
and speed of change (b) it scatters things and in turn gathers themor rather,
not in turn or later, but at the same time it comes together and separates
approaches and departs. (Plutarch On the E at Delphi 392b)
T7 potalo?r cq dr to?r aqto?r oqj #m 1lba_gr, r vgsim Jq\jkeitor,
6teqa cq 1piqqe? vdata.
Into the same rivers you could not step twice, as Heraclitus says, for other
waters flow on. (Plutarch Natural Questions 912a)
According to Tarn, B 49a (T5) is not a real fragment of Heraclitus, but
only a doublet of B 12 (T1), which restates in more everyday language
what B 12 is saying more profoundly but obscurely (42 3). Tarn also
rejects B 91(b) as alien to Heraclitus message, but accepts B91(a) (T6)
as a genuine fragment, which lies behind statements such as T2, 4, and
7, and has a very different message from B 12 by virtue of adding the
word twice.7
Tarn points out that if we take B 91(a) seriously, Heraclitus is not
saying without restriction that all things pass (A 6) or all things
flow,8 because we can step into the river once at least; hence Heraclitus
defends a kind of measured flux rather than a complete flux (14).9 Fur-
thermore, Plato is stretching his view for his own dialectical purposes
(14 16). Something like B 91(a) has the authority of Plato, Aristotle,
and Cratylus behind it. But with its use of twice and its reference
to the river in the singular rather than the plural, it is highly unlikely
that B 91(a) could be derived from B 12 (16 19).
7 Others, such as Kahn, H. 168 169 also accept B 91(a), and see n. 3 above.
Some of my arguments will apply to all who defend the alleged fragment.
8 A. Lebedev, The Cosmos as a stadium: Agonistic metaphors in Heraclitus,
Phronesis 30 (1985) 143 145, warns against conflating imagery of running
and flowing; but there seems to be some conflation here already.
9 As Plato Tht. 152d-e seems to suggest; in its most extreme interpretation, 182d-
183a.
As to B 12, Tarn convincingly shows that in its context, deriving

from Cleanthes, it has been misapplied as a metaphor for the birth of
souls (21 8).10 He criticizes Reinhardt and Kirks readings of B 12, ac-
cording to which it belongs to a cosmological context, as being unsub-
stantiated (30). Most interpreters of B 12, Tarn argues, have failed to
explain why potamos is in the plural; any adequate interpretation should
explain this (33). The participle embainousin is in the dative as expressing
the judgment of those entering the river, and the waters are different
from their point of view:
though in a sense two (or more) people are in the same river, the river
water in which each is, is different. The cosmos is common and one
to all men who are really awake but most people, like sleepers, treat
their own environments each as a separate and private cosmos . [I]n
the continuous flow of the different and different waters that bathe
them, men find confirmation of their belief that each one lives in his
own private universe. (35 6)
On this interpretation, B 12 is concerned with the common misappre-
hension that it is possible for men to live in their own private universes
(52), while B 91(a) is about the universal change of individual things,
but it is not so radical as the view Plato ascribes to him (50 1).
10 Arius Didymus commenting on Cleanthes preserves B 12 (from Eusebius Prae-

paratio Evangelica 15.20.2 3 = Arius Did. fr. 39.2 3 Diels), followed by the
sentence, And souls are nourished by moist exhalations, which some editors
have read as a continuation and specification of B 12 (see e. g., L. Gemelli Mar-
ciano, Die Vorsokratiker, (Dsseldorf 2007) 1:318. Serge Mouraviev defends this
reading in his Hraclite: les fragments du fleuve: B12: le contexte: T 261,
Rev. Philos. Anc. 26 (2008) 4 39, and H. II.A.1, T 261. He provides a valuable
reading of the context, but I am not persuaded that it establishes the second sen-
tence for a number of reasons, including the word !mahuli_mtai, whose use is
not attested before Aristotle, and clashes with the non-technical language of
Heraclitus (on the absence of technical terminology, see L. Gianvittorio, Il dis-
corso di Eraclito: un modello semantico e cosmologico nel passaggio dall oralit alla scrit-
tura (Hildesheim 2010) 59 68. On the context of B 12, see Dilcher, H. 177
183. See now M. Colvin, Heraclitus and material flux in Stoic psychology,
OSAP 28 (2005) 257 72, who argues persuasively that one of the Stoic com-
mentators pairs Heraclitus passage with a quotation or paraphrase from Zeno of
Citium to establish his Stoicizing interpretation.
II
Tarns account of the river fragments is a novel and interesting one. It

identifies two different river fragments, which belong to different con-
texts and have different messages, and it develops a complete new inter-
pretation of B 12 while giving a traditional interpretation of B91(a).
The most innovative component of Tarns interpretation is his
reading of B 12. The essence of his interpretation is the claim (quoted
above) that in the continuous flow of the different and different waters
that bathe them, men find confirmation of their belief that each one lives
in his own private universe (36). Now Heraclitus does indeed claim
that the many live their lives like sleep-walkers, each in a private world:
Other men are unaware of what they do when they are awake just as they
are forgetful of what they do when they are asleep. (B 1, in part)
Although this Logos is common, the many live as if they had a private
understanding. (B 2, in part)
What is strange about Tarns reading is the curious juxtaposition by
which what Heraclitus says of the many, he takes to be what they say
about themselves. Heraclitus nowhere says the many view themselves
that way, and it is difficult to imagine that this could be his point.
For surely if the many had reached this Socratic level of self-awareness,
they would already be open to Heraclitus message. Indeed, Heraclitus
almost expressly contradicts Tarns interpretation:
Many do not understand such things as they encounter, nor do they learn
by their experience, but they think they do. (B 17) 11
Heraclitus point must not be that the many believe they live in private
worlds and are ignorant of the world around themlike Cartesian sub-
jectsany more than sleep-walkers believe they are sleep-walking. It is
that they live like dreamers when they claim to be fully awake. The
many would maintain that they are fully aware of the world about
them, even while they make unreliable and subjective judgments
about it. Heraclitus never seems to take the beliefs of the many seriously
enough to examine for what reasons they believe what they believe; he
simply points out that they are unaware of the objective realities around
them. It seems highly unlikely that he is engaged in some kind of meta-
critique of their beliefs and their grounds for those beliefs. Furthermore,
it seems anachronistic to expect that people living around 500 BC
11 Cf. B 34.
would concerned about issues of existential subjectivity, not to mention

that Heraclitus would think they were.
Furthermore, Tarn faults other interpreters of B 12 for not taking
seriously the question of why river is in the plural. Yet when he gives
his own interpretation, he says the following:
The primary plural is embainousin, while the plural rivers may be due ei-
ther to Heraclitus desire to emphasize the universality of the river-state-
ment in B12, or to euphonic reasons and word-play, or to both . (35)
After all the fuss about the plural forms, this seems a rather feeble reason
for discussing rivers in the plural. Almost any other interpreter could
avail himself of the same reasons, and many have.12 There is no impor-
tant sense in which Tarns interpretation makes any special use of the
plural form, nor does he have more to say about euphony and word-
play, which are central and indeed indispensable to other interpreta-
tions. What we are left with is just a generalizing force, which any in-
terpreter who sees the fragment as having a general application, could
equally appeal to.
There are, however, serious problems for the traditional interpreta-
tion of B 91 too. For this text seems say that all things are changing so
much that one cannot encounter the same physical object (e. g. river)
twice. The implication that everything is constantly changing would
seem to entail, as Plato was quick to point out, that we cannot even
use words with constant meaning, or if we can, we cannot describe
the world without contradicting ourselves.13 Tarn acknowledges this
challenge before he fully explains his own position:
If carried to an extremeas by Cratylus and othershis doctrine would
amount to a denial of the law of contradiction. But in all probability he
[Heraclitus] himself did not carry his doctrine of the universality of change
to that extreme. (11)
Presumably the extreme Tarn refers to is the claim that all things are
changing all of the time.14 But if change is universal, how could Hera-
clitus not mean that? Even if he were not aware of the implications of
his theory, he would be guilty of propounding a theory that transgresses
the law of non-contradiction. Now admittedly, there are others who
12 E.g., Kahn H. 167.

13 See previous note, as well as Pl. Timaeus 49c-e.
14 Barnes (above, n. 3) 65 69 worries about this issue and determines that for
Heraclitus everything is always flowing in some respects (69, Barnes italics).
would agree that Heraclitus ignorantly falls into the trap of defending a
theory that entails contradictions.15 But however popular, and ancient,
this interpretation may be, it cannot provide a charitable reading of
Heraclitus, inasmuch as self-contradiction is the ultimate philosophical
failure.
Tarn makes one qualification to the traditional interpretation of B
91(a). What kinds of things, he asks, are always changing according to
Heraclitus? Heraclitus must only be thinking of individual things, he an-
swers, not abstract qualities such as Plato sometimes has in mind (49
50). But given that restriction, change for Heraclitus was as universal
as he could imagine, and the river provided a kind of paradigm example
of constant change (50 1). Yet even on this more restricted interpreta-
tion, a contradiction arises within Heraclitus philosophy. For whereas B
91 says that a person cannot step twice into the same river, B 12 says that
a person who steps into the same rivers encounters ever different waters;
but the latter statement surely presupposes or entails that the rivers are
the same. Thus, on Tarns view, B 91 and 12 turn out to be incompat-
ible statements. Curiously, Tarn recognizes this problem early in his
presentation:
[Heraclitus] probably did not realize that, if both the object and the subject
are constantly changing, it would not make sense to say upon those who
step into the same rivers different and different waters flow; one could do
so only if one assumes a stable element not only in the rivers but also in
those who step into them. (11)
Here one awaits some sort of explication that will show how Heraclitus
is not abysmally stupid; but in vain, as Tarn simply moves on to another
topic. Since, by Tarns own account, Heraclitus theory of change has a
universal application, it must apply even to the benighted waders slog-
ging through the river. But according to B 12 the rivers are the same,
and the waders encounter them as such. Not only does B 91 by itself
entail contradiction, but B 91 is incompatible with B 12.
Finally, Tarns defense of B 91(a) (= T6) as an independent text
seems problematic. In considering Plutarchs rendering in T7, he says
Plutarch has conflated B 91 and 12 (45), and tries to explain the con-
flation from the context. But his only real evidence that there is a con-
flation is the difference in language between B 91(a) and T7. Now here
I would like to suggest that we look at T7 without assuming that B 91(a)
15 E.g., Barnes (above, n. 3) 79 81.

is a fragmentwithout, that is, begging the question. T7 is just B 12

with B 91(a) substituting for the first half of the statement. Notice
how close it is to B 12:
T1 = B 12 T7 T6 = B 91(a)
potalo?si to?r potalo?r cq dr to?r aqto?r oqj potal` cq oqj
aqto?sim #m 1lba_gr, 5stim
1lba_mousim (r vgsim Jq\jkeitor), 6teqa cq 1lb/mai dr
6teqa ja 6teqa 1piqqe? vdata. t` aqt`.
vdata 1piqqe?.
In every case the word for river(s) comes first (not the natural word
order in Greek any more than English). In T7, if we leave out the in-
ferential particle gar and the dis of the interpretation (cf. Plato in T2) 16
we have the same three first words as in B 12 (with Attic rather than
Ionic endings),17 then the same fourth word, though in a finite form
rather than a participle; then the same fifth word and (skipping Heracli-
tus duplication and Plutarchs inferential particle) the same last two
words. The most natural supposition here is that in T7 Plutarch is sub-
stituting what he regards as an equivalent expression for the original T1.
If that is so, then he regards B 91(a) as just a stand-in for the first half of
B 12, not an independent and different statement.18 Since Plutarch
knows Heraclitus writings well, and is in a position, as we are not, to
judge whether there are two river fragments or only one, this substitu-
tion offers powerful confirmation for the view that B 91(a) is just a para-
phrase of B 12.19 We could, of course, wonder whether Plutarch had
misread B 91(a) or overlooked its context (whatever that might have
been). T6 shows parallels also, though here river is rendered in the sin-
gular. But now Plutarchs testimony works against Tarns reading and
undermines his claim that there are two independent statements with
different contexts. For lacking direct access to Heraclitus original
manuscript, Plutarch is the best witness we have for the relation of
the two alleged fragments. He gives a close paraphrase in T7 and a
loose paraphrase in T6; but T7 just is T1 with some inferential particles
16 K. Reinhardt, Heraklits Lehre vom Feuer, Hermes 77 (1942) 60 n. 24, argues

out that the dis goes back to Cratylus, not Heraclitus.
17 See C. D. Buck, The Greek Dialects (Chicago 1955) 88.
18 Cf. Reinhardt (above, n. 1) 207n. 1.
19 Cf. Reinhardt (above, n. 1) 18 n. 24.
(of the kind Heraclitus eschews in order to produce meaningful ambi-

guities) and some interpretation.
There is, however, one significant difference between Plutarchs para-
phrase in T7 and T1, and that is the presence of oqj. Yet we cannot rea-
sonably claim that this word has dropped out of T1, where it would make
no sense. It surely comes from a prior interpretation as manifested already
in Platos T2, and has become so ingrained that it has reversed the mean-
ing of the real statement of Heraclitus, namely that in T1 = B 12.
All in all, there are serious problems for Tarns interpretation of the
river fragments. B 12 (as interpreted by Tarn) says something at odds
with Heraclitus view of human understanding; the alleged fragment
B 91(a) by itself entails a contradiction; B 12 and 91(a) are incompatible;
and Plutarch treats B 91(a) as interchangeable with B 12.
III
The simplest hypothesis concerning the river fragments is, I believe, that
there is only one, B 12, of which B 91(a), no less than B 49a, is a para-
phrase. But there is one thing wanting for a convincing case, namely
that we be able to show for B 91(a), as Tarn has for B49a, how an in-
telligent reader could misconstrue it so badly as to turn it into a state-
ment that you cannot step twice into the same river. Against the as-
sumption that there are plural river fragments, I hope I have shown
that the alleged fragments other than B 12 are too much like B 12, as
well as too contradictory of it.
Can we, then, reconstruct a plausible account of how B 12 might
have come to be read as B 91(a) and A 6? Marcovich rightly argues
that Cratylus and those who read B 12 as he does must take embainousin
as iterative in force: every time someone steps into the river, different
waters flow.20 Cratylus must also have focused on hetera kai hetera to
the exclusion of toisin autoisin. Then he can claim that each initial
step into the river is a step into different waters. This would be an in-
teresting interpretation, but a hasty one. For since Heraclitus expressly
states that the rivers are the same, it really does not matter for the iden-
tity of the rivers that the waters are different. Nor, in fact, does it matter
whether we think of the entry into the rivers as repeated (iterative) or
ongoing (progressive); in either case the rivers are the same. Tarn ob-
20 Marcovich, H. 206.
jects to Marcovichs reading: why single out merely the action of en-
tering the river and disregard the role of the observer once he is inside
the river? This would unnecessarily restrict the universal validity of the
river-statement in B 12 (33). But this reply will not do. First, it does
not really matter what stage of the river-crossing Heraclitus is talking
about: the truth that the waters are always changing and the rivers stay-
ing the same still applies. Second, this interpretation does not really re-
strict the universality of the statement. Finally, to complain that the role
of the observer is being ignored begs the question, since it is only on
Tarns interpretation that the waders observations make any difference.
Tarn thinks that since Plato and Aristotle most likely had access to
Heraclitus text, they could not be so mistaken in their reading of Hera-
clitus as to miss his point. Yet recent studies have shown how conven-
tional ancient interpretations could be. Not only did the doxographical
tradition perpetuate demonstrably incorrect readings of the Presocratics,
but there was a kind of proto-doxographical tradition that began with
the sophists, and which tended to get recycled and repeated by Plato,
Aristotle, and other early readers of the Presocratics.21 It is quite possible
that Cratylus reading, which likely influenced Platos, came from an
earlier interpreter, namely Hippias of Elis in his Sunagog. 22 And Aristo-
tles reading continues that of Cratylus and Plato without so much as a
notice that they could be wrong (T4). If the interpretation given above
of Plutarch is correct, we see in Plutarch how a scholar with access to
Heraclitus text could read it without noticing the discrepancy between
something like A6 and the actual words of Heraclitus.
It appears from the verbal similarities among the three alleged frag-
ments, from the messages they convey and the settings in which they are
reproduced, that we can account for all other statements about the river
on the basis of B 12 alone. As for Tarns reading of B 12, that seems
unpromising at best. Tarn makes a good deal of the context of the
two alleged fragments he recognizes, and he does a good job of rejecting
the context in which B 12 was quoted. But when all is said and done,
that simply means that in fact we have no context for B 12. We must, as
21 B. Snell, Die Nachrichten ber die Lehren des Thales und die Anfnge der
griechischen Philosophie- und Literaturgeschichte, Philologus 96 (1944)
170 82; J. Mansfeld, Studies in the Historiography of Greek Philosophy (Assen
1990) 126 146 et passim.
22 Mansfeld (above, n. 21) 84 96; A. Patzer, Der Sophist Hippias als Philosophiehis-
toriker (Freiburg 1986) 33 42, 49 55.
it were, infer the context from the meaning of the fragment and not vice
versa. It is, indeed, possible, as one prominent view of Heraclitus text
would have it, that there is no extended argument in the book in any
case, but rather a collection of epigrams and riddles.23 If that is so,
even having the original book would not necessarily provide an illumi-
nating context for the meaning of the fragment. In any case, we do not
have the original sequence in which B 12 appeared, and hence the
whole issue of the context is an exercise in futility.
Fortunately, the fragment is complex in structure and rich in impli-
cations, and this very fact speaks both for its being a genuine quotation,
and for its being in a certain sense an autonomous text that can be ap-
preciated in the absence of a discursive context. In this interpretation I
follow especially Kahn, but also Kirk and Marcovich, and leave for an-
other occasion questions of the connection between this fragment and
Heraclitus larger message.
The major contrast in the fragment is between what is the same and
what is other. It is given by the words of the text that the rivers are the
same, the waters are other. The changing waters of the river flow onto
the people fording them. Accordingly, ever different waters from the
same rivers flow onto people who ford them. This is the obvious reading,
if anything in Heraclitus can be said to be obvious. But there is another
reading. The term toisin autoisin the same is sandwiched between pota-
moisi and embainousin, and it agrees not only with the former (masculine
dative plural), but also with the latter. In fact, as Kahn points out, Hera-
clitus has gone to a good deal of trouble to create this (syntactical) ambi-
guity, since any one of a number of changes could avoid it (changing the
word order, making one term singular, different word choice, etc.).24 And
indeed Heraclitus seems to favor putting an ambiguous term between two
possible referents.25 Let us, then, construe toisin autoisin with embainousin:
other waters flow onto the same people stepping into rivers. Again there
23 This was Diels view, embodied in his scheme of numbering the fragments al-
phabetically by source. Kahn, H., looks for a more coherent exposition, also
defended by Barnes (above, n. 3), but the view is opposed by H. Granger, Ar-
gumentation and Heraclitus book, OSAP 26 (2004) 1 17.
24 Rivier (above, n. 2) first called attention to the ambiguity, and argued that it
should be eliminated by deleting 1lba_mousim, which he argued was inserted
by a later editor.
25 E.g., B 1, 119, 51; see Graham, Heraclitus and Parmenides, in V. Caston and
D. W. Graham (eds.), Presocratic Philosophy: Essays in Honour of Alexander Mour-
elatos (Aldershot 2002) 35 37.
is a contrast between the same and other. But this time what is said to
remain the same are the persons fording the river. Curiously, we get pre-
cisely the opposite sense from that envisaged by Cratylus (and Plato and
Aristotle and Plutarch ), stressing the fact that the same action cannot be
repeated twice because of radical flux. To the contrary, the people who
encounter the river stay the same (on one reading); and the rivers they
encounter (on another reading) also stay the same. There is a flux here
(other and other waters), but the consequences of the flux are not chaotic
non-identity, but, surprisingly, some sort of continuity and hence identity
through time.
At least one more feature needs to be noted in the interpretation,
again following Kahn: by putting the first four words of the statement
in the dative plural, Heraclitus paints with sound. He creates the sounds
of a bubbling river, which, in the second clause seems to speed up and
hurry off in a series of alliterations. Heraclitus is a poet as well as a phi-
losopher, even if he does not confine himself to the conventional
rhythms of Greek poetry.
IV
We see that B 12 is wonderfully complex and suggestive, a text worthy

of Heraclitus and one that could not arise by chance or from a careless
rendition of another text. Paraphrases lose complexity and information,
as have B 49a and 91(a). But the process of interpretation has left us
without a context (which the passage may have never really had in
any case) and without any trust in the paraphrases that ancient interpret-
ers derived from it.26 Let us turn briefly to those who have championed
B 12 as the dominant or only river fragment and see how they take it.
According to the originator of the interpretation I am following,
Not a single fragment expresses the thought that all things are in flux, that
in general only transition and change, never continuity and persistence are
found which will show us where the panta rhei is truly at home. The basic
idea of Heraclitus is much rather the most precise opposite imaginable to
the flux doctrine: persistence in change, constancy in alteration, tauton
in metapiptein, metron in metaballein, unity in division, eternity in destruc-
tion. (Reinhardt [above, n. 1] 206 207)
26 Some have wished to find the context in Arius Didymus. On this, see above, n.
10.
Kirk thinks we should be careful not simply to assimilate B 12 to the

coincidence of opposites because it is not as concrete as other examples,
it is not explicitly framed in this way, and because there is a danger that
the same-other contrast will collapse into incoherence. Rather,
What these river-fragments are intended to show, I believe, is the regularity,
the order, the metron or measure, which Heraclitus believed to underl[ie]
and to control natural change in all its forms . The river-fragments, then,
seem to exemplify not the constancy of change but the regularity of natural
change in one particular manifestation. (Kirk H. 36 37) 27
Marcovich, on the other hand, sees no reason to follow Kirk in his em-
phasis on the regularity of change, but instead defends the discussion as
an instance of the coincidence of opposites, which Kirk rejects.28
While these three interpretations mark an advance on those based
on B 91, they all seem to miss the point to some degree. In the first
place, the image Heraclitus presents us with does indeed depend on a
flux: other and other waters flow. This is a kind of given in the
scene. Of course, a good deal depends on just what we make of the
flux portrayed in the image, but it needs to play some decisive role.
As the first and third commentators recognize, there has to be some
sort of connection between the twoas is often the case in Heraclitus
discussion of opposites. But then it is the precise relationship that is per-
haps most important. And here, it appears, Reinhardt wants to subordi-
nate flux to constancy; this seems problematic. Second, contra Kirk, B 12
is not about measure, though it may presuppose the notion: it says noth-
ing about the amount of water involved. Further, to downplay the
image as applying only to some kinds of change, as Kirk does, is to
miss the power of the image. It is true, no doubt, that as it stands B
12 expresses just a single concrete image. But the image is powerfully
suggestive, and there is no sign that he marks any boundaries to its po-
tential applications. Finally, contra Marcovich, there is something that
makes this case fundamentally different from most other cases of the co-
incidence of opposites: in at least many cases, opposites are connected
because they (that is, opposite things) change into one another (B 88,
126).29 But in B 12 we are not talking about hot and cold or waking
and sleeping, but a river and its contents. The opposition occurs at a
27 Kirkss italics. Compare Kirk, H. 366.

28 Marcovich, H. 212 213.
29 For discussion of this, see Graham Heraclitus criticism (above, n. 5) 38 44,
Explaining the Cosmos (above, n. 5) 122 124.
higher categorial level altogether, between, in Aristotelian terms, a sub-

stance and its matter. A substance and its matter are not opposites; the
opposition arises between the sameness of form and the difference of
matter: paradoxical, but not contradictory. If we want to classify this
case as a coincidence of opposites, we should still recognize it as a differ-
ent sort of animal than most other cases. The river example somehow
represents an exceptional and unusually profound case of contrast.
As to the continuing value of the image, we may call a modern phi-
losopher to the witness stand, namely David Hume:
as the nature of a river consists in the motion and change of parts; tho in
less than four and twenty hours these be totally alterd; this hinders not the
river from continuing the same during several ages.30
The river is interesting to Hume for precisely the reason it was interest-
ing to Heraclitus: it is an example of an entity that remains the same
even though its components change; or better yet, precisely because its
components change. Indeed, a figure much closer in time to Heraclitus
seems to echo the image. Aristotle observes:
[S]hall we say that while the race of inhabitants remains the same, the city is
also the same, although the citizens are always dying and being born, as we
call rivers and fountains the same, although the water is always flowing away
and more coming? (Aristotle Politics 1276a34 39, revised Oxford trans.)
Here rivers remains the same although their contents change, thus po-
tentially serving as a reliable model for the sameness of city states. In this
putative analogy, the transtemporal identity of rivers is assumed as a
given.
And this brings us back to the question of context. Is there a con-
text? What is it? Should there be a context? A similar question is some-
times raised concerning the road fragment: The way up and down is
one and the same (B 60). Is this a statement roads, or about the ele-
ments, or about mortality, or about the cosmos, or something else?
Both statements are, we might say, almost commonplaces, about every-
day experiences. What makes them powerful is just the vividness of the
concrete example. The road fragment is about roads and routes. The
river fragment is about rivers and people. But both of them have impli-
cations for life, nature, cosmology. There is a unity that roads have:
30 Treatise I.4.6, p. 258 Selby-Bigge. Here Hume is concerned with conventional

assessments of identity more than the strict identity which he denies to sensible
objects.
whether we travel uphill or downhill, into town or out of town, there is

only one road. If we go out of town on the road, we come back on the
same road. This seems to be true of processes of elemental change, and
of cosmic alteration. There is a kind of principle involved that manifests
itself in many contexts, a kind of logos. As Kahn has persuasively argued,
Heraclitus builds his treatise on the twin principles of linguistic density
and resonance. In linguistic density a multiplicity of ideas are expressed
by a single word or phrase. Resonance denotes a relationship by
which a single verbal theme or image is echoed from one text to another
in such a way that the meaning of each is enriched when they are under-
stood together.31 Further, These two principles are formally comple-
mentary.32 In effect the resonance between texts allows a passage to
transcend its immediate context, whatever it was, and to connect
with other passages with similar themes and motifs. If this is so, the
question of context becomes less urgent and less decisiveas well as vir-
tually unanswerable in the present state of our knowledge. We must ex-
tract the meaning from the text itself, in all its richness, and then in sec-
ond place try to link it to passages with similar themes. The result is a
network of associations rather than a linear argument.33
With the river image a commonplace about rivers comes to the fore.
When we cross a river (the Greeks, we must remember, had very few
bridges, even in Roman times), we pass through constantly changing
waters. The river is the same, and we are the same. Indeed, without
the flowing waters, there would be no river; the river is constituted
by the changing waters. As we wade into the river, we confront ever
changing waters, yet we are the same. By parity of reasoning, if we
did not confront an every-changing experience, we would not have ex-
perience. In a certain sense, our own reaction to our experience consti-
tutes us as stable subjects.
The scholar who sees this relationship most clearly is Charles Kahn:
What is emphasized is that the structure and hence the identity of a given
river remains fixed, despite or even because its substance is constantly
changing. And if the parallel mentioned above is pressed, something similar
is indicated about the structure and identity of individual human beings.
(Kahn, H. 168)
31 Kahn, H. 89. See also a recent study using Kahns assumptions, Gianvittorio
(above, n. 10).
32 Kahn, H. ibid.
33 Which Gianvittorio (above, n. 10), ch. 2, explores as reti di significazione.
In this logos a flux is essential: without it there would be no rivers, and

no waders. But paradoxically, the changing surroundings make possible
a unified reality and a unified experience. A river is an enduring entity
that gets its very being from its ephemeral material; a self is an enduring
entity that gets its continuity from its changing encounters. There is a
flux doctrine here, but not the flux doctrine that worried Cratylus,
Plato, and Aristotle. Here the flux does not undermine any possibility
of experience, but rather, mirabile dictu, makes experience possible. In
fact, the message here is not so different from Aristotles own doctrine
concerning metabolism and biological existence, that higher-level struc-
tures or forms endure through exchanges of matter.34 Indeed, the sub-
lunary world maintains itself through the generation and destruction of
its material basis:
Nature always strives after the better. Now being is better than not-
being: but not all things can possess being, since they are too removed
from the originative source. God therefore adopted the remaining alter-
native and fufilled the perfection of the universe by making coming-to-
be uninterrupted: for the greatest possible coherence would thus be secured
to existence . (Aristotle On Coming To Be and Passing Away 336b28 33,
trans. Joachim)
In the same way, it appears that for Heraclitus the whole cosmos is ever-
lasting precisely because parts of it are constantly undergoing material
changes (B30). Thus we seem to be justified in seeing the flux doctrine
as having application to the cosmos as well as to rivers and people. In
Heraclitus world, relatively stable structures supervene on constantly
changing matter.
B 12 is the river fragment, Heraclitus one and only statement of the
flux. The sage of Ephesus believes in the flux of matter, but not a flux
that erodes all landmarks and makes knowledge and communication im-
possible. Rather, his flux brings stability to rivers, life to animals, coher-
ent experience to men, and everlasting existence to the cosmos itself. Or
rather, there are regularities that govern flux, that order the exchanges of
opposite characteristics. While the changes are continuous, the regular-
ities balance them out and ensure a higher-level order in the world. Like
Adam Smiths Invisible Hand that maintains the balance of market
forces in an economy, Heraclitus scheme maintains a balance of mate-
rial forces. The mysterious entity he calls the Logos can be seen as just
34 A theme stressed by M. Furth, Substance, Form and Psyche: An Aristotelian Meta-

physics. Cambridge 1988.
the expression of systematic order in nature. No less real than the con-
stant exchange of hot and cold, wet and dry, earth, water, and fire is the
constant replacement of what is lost by a new supply, drawn from what
is lost of the contrary stuff. The world needs constant flux and inter-
change at the level of material, on which a hidden law operates to main-
tain balance. The cosmos is an ever-living fire precisely because it lives
in its own transformation of opposites; it stays the same as it changes.35
There is one way that Heraclitus theory is profoundly different
from that of all other early philosophers of nature. Whereas they attempt
to explain change in terms of constant realities and their modifications,
Heraclitus attempts to explain constancy in terms of ongoing change.
Whereas in some metaphysical sense they are all philosophers of stasis,
for whom (especially from Parmenides on) stability is fundamental
and change problematic, Heraclitus is a philosopher of process, for
whom change is fundamental and stability problematic. But if he is a
process philosopher, he is not a nihilist or an epistemological or onto-
logical anarchist. His world is orderly through and through, the forces
balanced, the elements proportionate, the cosmos stable. Like a river.36
Bibliography
Barnes, Jonathan. The Presocratic Philosophers. Revised ed. London 1982 [1979].
Bollack, Jean, and Heinz Wismann. Hraclite: ou la sparation. Paris 1972.
Buck, Carl Darling. The Greek Dialects. Chicago 1955.
Colvin, Matthew. Heraclitus and material flux in Stoic psychology, OSAP
28 (2005) 257 72.
Conche, Marcel. Hraclite: fragments. Paris 1986.
Emlyn-Jones, C. J. Heraclitus and the unity of opposites, Phronesis 21 (1976)
89 114.
Furth, Montgomery. Substance, Form and Psyche: An Aristotelian Metaphysics.
Cambridge 1988.
Gemelli Marciano, Laura. Die Vorsokratiker. 3 vols. Dsseldorf 2007 09.
Gianvittorio, Laura. Il discorso di Eraclito: un modello semantico e cosmologico nel pas-
saggio dall oralit alla scrittura. Spudasmata 134. Hildesheim 2010.
Graham, Daniel W. Heraclitus criticism of Ionian philosophy, OSAP 15
(1997) 1 50.
35 B 30, 31, 90.

36 My thanks to my fellow symposiasts for helpful comments. On process philo-
sophy, see now my Heraclitus as a process philosopher, Philosophy Study 2
(2011) 1 8.
Heraclitus and Parmenides, in V. Caston and D. W. Graham (eds.), Pre-

socratic Philosophy: Essays in Honour of Alexander Mourelatos (Aldershot 2002)
27 44.
Explaining the Cosmos: The Ionian Tradition of Scientific Philosophy. Princeton
2006.
Heraclitus as a process philosophers, Philosophy Study 2 (2011) 1 8.
Granger, Herbert. Argumentation and Heraclitus book, OSAP 26 (2004) 1
17.
Kirk, Geoffrey S. Natural change in Heraclitus, Mind 60 (1951) 35 42.
Lebedev, Andrei. The Cosmos as a stadium: Agonistic metaphors in Heracli-
tus, Phronesis 30 (1985) 131 50.
Mansfeld, Jaap. Studies in the Historiography of Greek Philosophy. Assen 1990.
Mondolfo, Rodolfo. I frammenti de fiume e il flusso universale in Eraclito,
Revista Critica di Storia della Filosofia 15 (1960) 3 13.
Mouraviev, Serge. Heraclitea II.A.1: Hraclite dphse: la tradition antique et me-
dieval. Sankt Augustin 1999.
Hraclite: les fragments du fleuve: B 12: le contexte: T 261, Revue de Phi-
losophie Ancienne 26 (2008) 4 39.
Patzer, Andreas. Der Sophist Hippias als Philosophiehistoriker. Freiburg 1986.
Petit, Alain. Hraclite: La captation de la marge, tudes Philosophiques (1988)
207 19.
Popper, Karl R. Kirk on Heraclitus, and on fire as the cause of balance, Mind
72 (1963) 386 92.
Reinhardt, Karl. Parmenides und die Geschichte der griechischen Philosophie. Bonn
1916.
Heraklits Lehre vom Feuer, Hermes 77 (1942) 1 27.
Rivier, Andr. Un emploi archaque de lanalogie chez Hraclite et Thucydide. Lau-
sanne 1952.
Robinson, Thomas M. Heraclitus. Toronto 1987.
Snell, Bruno. Die Nachrichten ber die Lehren des Thales und die Anfnge
der griechischen Philosophie- und Literaturgeschichte, Philologus 96
(1944) 170 82.
Stokes, Michael C. One and Many in Presocratic Philosophy. Washington, D.C.
1971.
Tarn, Leonardo. Heraclitus: The River fragments and their implications,
Elenchos 20 (1999) 9 52.
Vlastos, Gregory. On Heraclitus, AJP 76 (1955) 337 78. Repr. in G. Vlas-
tos, Studies in Greek Philosophy, vol. 1 (Princeton 1995) 127 150.
12. Heraclitus Ethics
David Sider
Investigations into Heraclitus ethics are scarce on the ground. It is not

that, apart from some that have been been overlooked because they are
universally considered spurious, his several comments that touch upon
human behavior have been ignoredthere are too few fragments for
any to go uncommented for longbut that they seem isolated and
not obviously calling out to be made part of a systematic theory of eth-
ical behavior. Furthermore, it has to be said that some of these scarce
attempts to study Heraclitus ethics are disappointing.1 Nor does the
subject of ethics as such receive much attention in more comprehensive
treatments. It is true that ethics is discussed elsewhere in the huge Hera-
clitus bibliography, but rarely head-on as a subject to be investigated and
analyzed in its own right.2 Two noteworthy contributions to the liter-
1 Thus, A. N. Zoumpos, Das ethische Urteil bei Heraklit, Platon 11 (1959)

420 423, raises the Aristotelian distinction between a-priori and a-posteriori
reasoning, only to find that this does not apply to Heraclitus, which does not
get us very far: Kurz gesagt nimmt Herakleitos das ethische Urteil als Produkt
des Gttlichen Logos an, das naturgemss also in Voraus in den Menschen vor-
hande ist. (One could compare Anaxagoras Nous, which also shows up in an-
imals.) Zoumpos relies too heavily on Plato for his understanding of Heraclitus,
especially those passages in the Cratylus and Theaetetus that characterize him pri-
marily as one relying on a theory of flux to explain his epistemological con-
cerns. Zoumpos thus concludes that der Mensch kein sichers Urteil bilden
kann, since there can be no Seiende, existence, in Heraclitus. There is, how-
ever, logos, styled as Weltgesetz by Zoumpos, which can offer guidance to
men, should they care to follow it. C. J. Broniak, Heraclitus, Parmenides,
and Plato on living the good life, Dialogue (Milwaukee) 30 (1987) 28 37, de-
votes only a little over one page to Heraclitus (28 29), with very little attempt
to work the fragments themselves into a coherent system. C. Eggers Lan, Ethi-
cal-religious meaning of fr. 30 D.-K. in L. Rossetti (ed.), Atti del symposium
heracliteum 1981 (Rome 1983) 1.291 299, does not follow through on the
promise of his title.
2 Some few examples chosen from the best and most useful books on Heraclitus:
Charles Kahn, The Art and Thought of Heraclitus (Cambridge 1979) does not
even include a listing for ethics in his index; Miroslav Marcovich, Heraclitus:
Greek Text with a Short Commentary (Merida 1967) does index ethics, but
322 David Sider
ature are a brief contribution by Robert Bolton, which acknowledges

that it is only a preliminary study, and the more substantial investigation
by J. W. Evans, which, because it was only an unpublished doctoral dis-
sertation, has not received the attention it deserves.3 It was with this near
vacuum on the subject in mind that I planned to address the matter at
the conference on Pythagoras and Heraclitus, little knowing that Enri-
que Hlsz (who calls Heraclitus a moral philosopher) and Tony Long
would be addressing this very question directly and that others, most
notably Aryeh Finkelberg and Gabor Betegh, would be shedding signif-
icant light on the matter. Still, our various approaches complement each
other, and each may have something to contribute to a new overall ap-
preciation. In particular, Tony Long has properly noted how misleading
it can be to read Plato back into Heraclitus. Yet my approach to Her-
aclitus ethics based on the fragments themselves will lead to a theory
that is consistent with much that is in fact found in Plato.
he assumes without argument that Heraclitus ethics were very much those of
the ruling class of his native Ephesos, which entails that Heraclitus never devel-
oped an ethical system of his own. Cf. too M. Adomenas, Heraclitus on reli-
gion, Phronesis 44 (1999) 109, who refers to Heraclitus antipopulist ethics.
This will be disputed below.
3 R. Bolton, Nature and human good in Heraclitus, in K. I. Boudouris (ed.),
Ionian Philosophy (Athens 1989) 49 57; J.W. Evans, Heraclitus and Parmenides as
Moral Philosophers, diss. Yale 1970. Bolton approaches Heraclitus ethics via the
notion of natural law as it would have incorporated Heraclitus idea of physis.
Placing Heraclitus broadly in an early form of the nomos/physis controversy,
Bolton relies most heavily on B 114 nm m|\ k]comtar Qswuq_feshai wq t`
num` p\mtym . tq]vomtai cq p\mter oR !mhq~peioi m|loi rp 2mr, toO
he_ou. From this, along with some other supporting texts, Bolton reasonably
concludes not only that is there a natural (because divine) law, but also that
men can access it and trust in it. Note also B 116 !mhq~poisi psi l]testi
cim~sjeim 2yutor ja syvqome?m, where self-knowledge is coupled with so-
phronein (54). Bolton then offers some criticisms that could be brought to
bear, but his article remains a good place to start. Evans more comprehensive
treatment argues that Heraclitus ethical views are inseparably entwined with
his cosmology and metaphysics (39).
M. Fattal, Paroles et actes chez Hraclite: Sur les fondements thoriques de laction
morale (Paris 2012), appeared as this article was on the point of submission to
the publishers.
12. Heraclitus Ethics 323
Before beginning, however, it should be noted that, despite the modern

neglect, several voices in antiquity were in agreement that the book of
Heraclitus did indeed devote a significant portion of its contents to ethics
or politics, a subject not always easy to distinguish from ethics in the frag-
ments, nor, I suspect, in Heraclitus own thought. Thus, Diogenes Laer-
tius 9.15 records that many have written commentaries on Heraclitus,
including the grammatikos Diodotos, who denied that his syngramma was
about nature, but rather that it was about politics, and that the material
on nature was placed there (only) as a paradigm (ou vgsi peq v}seyr
eWmai t s}ccqalla !kk peq pokite_ar, t d peq v}seyr 1m paqade_c-
lator eUdei je?shai). Again from Diogenes (9.12) we learn that this same
Diodotus described Heraclitus book as an !jqibr oQ\jisla pqr st\hlgm
b_ou, accurate steerage toward lifes harbor.4 Diodotus Grammaticus is
unknown, but whoever he was, he probably was not a professional phi-
losopher, most of whom in the ancient world seemed ready if not eager
to ignore the ethical views of Socrates predecessors.5 Still others, accord-
ing to D.L. ibid., call Heraclitus book a cm~lom( Ah_m, tq|pou j|slom
6ma t_m nulp\mtym, a guide of conduct, .6
Diogenes Laertius also records the interesting statement (9.5) that t
d veq|lemom aqtoO bibk_om 1st lm !p toO sum]womtor Peq v}seyr,
di-qgtai d( eQr tqe?r k|cour, eUr te tm peq toO pamtr ja pokitijm
ja heokocij|m. Heraclitus book, although peq v}seyr on the
whole, has been divided into three logoi, one on t pm, which must
include at least cosmological matters; a second on politics, and a third
on divine matters. A straightforward reading of this passagefocusing
4 The Greek is a iambic trimeter, with its caesura preceding the fourth rather than
the third princeps. Since grammatikoi were teachers, Diodotus may have com-
posed verses to serve as mnemonic aids for his students. For a recent discovery
of a teachers verses directed to his students, see R. Cribiore, D. Ratzen, & P.
Davoli, A teachers dipinto from Trimithis (Dakhleh Oasis), Journal of Roman
Archaeology 21 (2008): 170 91.
5 Note how Aristotle in his Ethics, although coming to grips with Plato in EN
1.6, does not provide in this work the same sort of overview of presocratic
or sophistic thought on the subject that he does in his Physics and Metaphysics.
6 The reading of the codd., printed here, is not easy to construe, and has been
frequently altered:
ja cm~lom( Ah_m, <ja jakm> j|slom tq|pym 2m|r te sulp\mtym Diels;
ja cm~lom( Ah_m, <toO h( fkou> j|slou tq|pim 2m|r te sulp\mtym te
Hicks.
324 David Sider
on the present 1st_would allow for somebody simply adding chapter

headings, without any further alteration, but this would hardly merit
mention by Diogenes. The perfect has been divided would never
have been used if this were Heraclitus own division. By concentrating
on the odd phrase !p toO sum]womtor, however, which seems to mean
that the book was originally continuously on, or continually returning
to, the subject of nature, we get the more interesting statement (one
that Diogenes would have found deserving of mention) that Heraclitus
book suffered rearrangement into the three rubrics/sections he names. If
so, there must have been a higher percentage of the book concerning, at
least on a first reading of Heraclitus notoriously ambiguous prose,
human nature (politics and ethics, as later understood).7 And as Dei-
chgrber noted, these topics are three of the six into which Cleanthes
divided philosophy, which suggests that the rearrangement and division
into three parts were due to a Stoic.8
These important secondary testimonia aside, Heraclitus was clearly
concerned with both individual behavior (B 43 vbqim wq sbemm}mai
lkkom C puqjaz^m), as well as that of his city (B 44 l\weshai wq tm
d/lom rpq toO m|lou). It remains, however, for us to place these gno-
mic thoughts into a more complex system of ethical behavior.
Whatever the precise form of Heraclitus original book, it is clear
that those who could read it acknowledged that a significant part con-
cerned ethicseven if, as Diodotus said, on the surface it seemed to be
speaking about nature. That is, as I believe but cannot argue for at
length here, Heraclitus moved rapidly from topic to topic, so
that, as one possible example, his two statements about the hidden
(but not unrecoverable) meaning of both Apollo and of nature (B 93,
123) could have come close together in his first edition but have
7 The main problem lies in the phrase !p toO sum]womtor. The related adjective
sumew]r is regularly applied to continous words, phrases, etc. (LSJ s.v. I 2).
The verb, however, does not seem to be used in this sense, but perhaps its
basic sense of embrace, comprise, hold together can suggest a translation of
the phrase as from that which holds the work together as a whole, i. e.,
that physis is a constant topic throughout the work.
8 K. Deichgrber, Bemerkungen zu Diogenes Bericht ber Heraklit, Philologus
93 (1938 39) 19. The remaining parts are dialectic, rhetoric, and ethics (D.L.
7.41), the last of which, as I have suggested, may have been too hard to distin-
guish from politics in Heraclitus writings. On the other hand, R. Hirzel, Un-
tersuchungen zu Ciceros philosophischer Schriften (Leipzig 1882) 2.70 178, argued
that Cleanthes six-part division was an expansion of that which he found in
Heraclitus. See further, Dilcher, H. 188 189.
been segregated later into the theological and the physical parts, respec-
tively.
II
Let us begin our attempt to provide an overall view of Heraclitus eth-

ical thought by considering B 29 aRqeOmtai cq 4m !mt "p\mtym oR %qi-
stoi, jk]or !]maom hmgt_m oR d pokko jej|qgmtai fjyspeq jt^mea,
which is regularly taken as an expression of Heraclitus approval of his
own aristocratic classas seems obvious from the standard translation
of the first clause (in whatever language): The best men [a term the no-
bility applies to itself; see Nietzsches Genealogy of Morals] choose one
thing in place of [= over] all things: everlasting fame from mortal
things.9 It is also further thought that Heraclitus is here alluding to Si-
monides famous lyrics on the fallen at Thermopylae (531 PMG):
t_m 1m Heqlop}kair ham|mtym 1

1mt\viom d toioOtom out( eqqr
ouh( b pamdal\tyq !lauq~sei wq|mor. 5
Keym_dar, 7
Sp\qtar basike}r, !qetr l]cam kekoipr
j|slom !]ma|m te jk]or
To ignore Heraclitus use of the word for the moment, Simonides, pre-
sumably very soon after 480 B.C., is the first to use the adjective !]maor
in this extended sense. Up until the beginning of the fifth century,
!]maor is always applied to water (Homer, Hesiod).10 For Simonides
9 This is my translation. Some examples from others: I migliori, in cambio di

tutto, ci, una cosa scelgono: la gloria eterna presso i mortali (A. Capizzi, Era-
clito e la sua leggenda [Rome 1979] 75; Denn eins gibt es, was die Besten allem
anderen vorziehen: den ewigen Ruhm den vergnglichen Dingen (DK); Ils
prennent une chose contre toutes, les meilleurs, gloire toujours jaillissante des
mortels ( J. Bollack & H. Wissmann, Hraclite ou la sparation [Paris 1972]
128. It will be clear from my translation that I take !mt "p\mtym as a neuter
phrase (contra Bollack-Wissmann), and hmgt_m as neuter (but, contra Diels, not
with an understood !mt_; see further, below).
10 The simple !] for !e_ was used by Pisander F 12 Bernab and perhaps by Pindar
P. 9.8 as well, if Hermanns conjecture is accepted. Ignored here are two ex-
amples from Pindar: P. 1.6 !em\mou puq|r (fire is fluid) and the metaphorical
O. 14.12 !. til\m, in the belief that Simonides poem would have been recited
in Ephesus soon after composition, whereas Pindars epinicia (for patrons in
326 David Sider
and Pindar, everlasting is an easy metaphorical extension of ever-

flowing, but how would Heraclitus, whether on his own or by bor-
rowing from a predecessor, have made use of the word in this extended
sense?Heraclitus, who so famously made his readers question the sta-
bility or sameness of rivers?11 For him, ever-flowing fame all by itself
would be a suspect phrase all by itself; add the word hmgt_m and it seems
to lose all positive force, especially when B 12 potalo?si to?sim aqto?sim
1lba_mousim 6teqa ja 6teqa vdata 1piqqe? is read with to?sim aqto?sim
taken !p joimoO with rivers and the steppers-in, so that the stability of
men too is subject to doubt.12 To be more specific, Heraclitus himself
would have no objections to things in flux (everything is), since he
has the ability to understand such things; but, as so many of his frag-
ments show, he is not prepared to grant this ability to others.
To return to B 29, there may indeed be men who chose one thing
over all, jk]or !]maom hmgt_m, but are these men really the best? Once
this is put in question, the plural aristoi also begins to look suspicious. It
is not one class of men whom Heraclitus would put forth as a model, but
rather one man, whom in fact he calls the best and all by himself worthy
of many: B 49 eXr 1lo l}qioi, 1m %qistor .13 Similar is B 33 m|lor ja
bouk0 pe_heshai 2m|r.
The relative chronology between Heraclitus and Simonides use of
jk]or !]maom is unknown. The ancient testimony for the death of the
former is rather confused, but 484 would seem to be the absolute earliest
date; a later date remains quite possible.14 If soand this is what I be-
Orchomenus and Etna, respectively) would have had a more limited fame soon
after composition.
11 See Graham in this volume.
12 See D. Sider, Word order and sense in Heraclitus: Fragment One and the river
13 As Marcovich ad loc. notes, Heraclitus original words are open to doubt, since
this reasonable reconstruction from the several citators forms, less one initial syl-
lable (e. g., 5sh( Marc.), an iambic tetrameter. Scythinus versification may have
supplanted the original, which itself may well have had a metrical clausula. In
this case, taking note of the fact that several witnesses insert the phrase !mt pok-
k_m, we may entertain the notion not only that this phrase was original, but also
that B 29, with 6m, !mt_, and %qistoi, originally appeared very close to, and in
obvious contrast to, B 49.
14 Serge Mouraviev, Heraclitea III.1: Recensio: Memoria (Sankt Augustin 2003)
124 126, has argued for placing Heraclitus death as late as 460, but this has
been criticized by Aryeh Finkelberg in his review of this volume, SCI 25
(2006) 151. For the inconsistent testimony, see further Jaap Mansfeld, The
lieve to be the caseit seems to more likely that Heraclitus was re-
sponding to Simonides (and the favorable reaction his poem no doubt
received in Ephesus) than the reverse.15 Read this way, Ephesus
best are in fact not the best, and Heraclitus would be playing on
words in a way that becomes more familiar later in the fifth century.16
On the surface, where The Best are to see nothing but a compliment,
hmgt_m has been taken as masculine with an understood pq|r, by
men; or neuter with a somewhat more easily understood !mt_, instead
of mortal things. I am happy with either of these (or both) as the sur-
face meaning, but, keeping the underlying pejorative meaning in mind,
we can also understand it as a simple causative genitive: The [soi-disant]
Best choose an ever-changing reputation (kleos) for mortal things.
The entire fragment has always been read as if the two clauses were
in complete contrast, although there is in fact no evidence for a l]m in
the first clause. Thus, instead of a contrast between the upper and lower
classes, as is usually understood, the second clause, following the first as
explained above, can now be rendered and the majority [sc. of them,
the aristoi] glut themselves like cattle. In other words, Heraclitean eth-
ics loves to hide. oR %qistoi are not in fact %qistoi, and some of them are
no better than oR pokko_, the people they generally despise. Thus, al-
though Heraclitus may not be a friend of oR pokko_, neither is he to
be taken as a staunch defender of the upper classes. More on this later.
Along the same lines, but somewhat more obvious is B 28 doj]omta
b dojil~tator cim~sjei vuk\sseim,17 the person whose doxa is greatest
knows how to hold on to his beliefs, which also admits of both com-
plimentary and pejorative interpretations, depending on whether doxa
means ones own judgement (cf. Parm. B 1.30, which also combines
d|nar and dojoOmta, along with doj_lyr; 8.51) or the judgement in
which one is held by others, i. e., fame (cf. Sol. 13.4, 34 and CEG
396 [vi c. BC] dr d ( Qm %mhqpoir d|nam 5weim !cahm, Grant that
chronology of Anaxagoras Athenian period and the date of his trial, Mnemo-
syne 32 (1979) 68 69.
15 Another close echo between the two is the phrase vbqim sbemm}mai (Her. B 43 &
Simonides 3 PMG), which could go in either direction, if in fact each is not
independently elaborating on the Homeric phrases s. w|kom and s. l]mor.
16 Cf., e. g., Euripides Ba. 395 t sovm oq sov_a, which in thought returns us to
Heraclitus B 40 poukulah_a m|om oq did\sjei, as was noticed by Sandys ad Eur.
Ba. 395. Note also Thuc. 3.82.4 7 on how words changed meaning.
17 My text remains as close to the mss. as possible; for cicm~sjy + inf., cf. Soph.
Ant. 1089 cm` tq]veim, learn how to.
328 David Sider
he have a good reputation among men). It is beginning to look as

though Heraclitus did not direct his ethical statements at those men
who were most in need of them. With this, compare the ethical belief
attributed to Heraclitus in an anonymous collection: Gnomol.
Vatic. 315 (B 135) sumtolyt\tgm bdm 5kecem eQr eqdon_am t cem]shai
!cah|m, becoming good is the shortest path to eqdon_a. As the high
D-K number (above 126) indicates, this was considered spurious by
Diels, as it was also by Marcovich, and it is true that few would want
to credit Heraclitus with the banal, and perhaps perverse, thought that
becoming good is the shortest path to glorybut what if this too
were originally a statement that, although superficially appealing to fool-
ish seekers after fame, also contained the message that becoming good
(however Heraclitus may have expressed it in his own words) was the
most direct route to eu-doxia, virtue, much as in Pindar Nem. 3.40
succeme? d] tir eqdon_ l]ca bq_hei, he is greatly powerful in his in-
stinctive good thoughts, although Heraclitus doubtless would intend
something more intellectual than Pindar.
Since paths usually have destinations or goals, it may be that another
usually neglected testimonium is relevant: Clem. Strom. 2.130 (A 21)
v\mai toO b_ou t]kor eWmai Jq\jkeit|m tm eqaq]stgsim. See
also Theodoret, who said that Heraclitus chose eqaq]stgsir instead of
Bdom^.18 In what context anyone would prefer eqaq]stgsir, satisfac-
tion, to pleasure is not easy to reconstruct. Satisfaction, furthermore,
seems an unlikely lifes goal for the amusingly irascible and arrogant
Heraclitus.19 Perhaps this late word conceals an original pun on, or a
misunderstanding of, eq + !qet^; or, somewhat more likely, on eq +
!qiste}y. Compare the Homeric line aQ]m !qiste}eim ja rpe_qowom
%kkym (Il. 6.240). If so, this testimony nicely complements B 135, so
that thought, virtue, becoming good, and being best (in the
proper way) all are key ideas in Heraclitus ethics. Of these, doxa is es-
pecially important in that it points towards the intellectual basis of his
ethics, as we shall see further below. For now, lets briefly adduce Alber-
tus Magnus de veget. 6.401 (B 4) Heraclitus dixit quod Si felicitas esset in
18 Graec.Aff. 11.7 (p. 185 Marc.) ja Jq\jkeitor d b 9v]sior tm lm pqosgco-

q_am let]bake, tm d di\moiam jatak]koipem !mt cq t/r Bdom/r eqaq]stgsim
t]heijem.
19 Heraclitus was lecak|vqym ja rpeq|ptgr (D.L. 9.1); see further Moura-
viev (above, n. 13) 25 6. Plato, to whom we shall be comparing Heraclitus
later, avoids these charges by having his best man be someone other than him-
self.
delectationibus corporis, boves felices diceremeus, cum inveniant eqobom ad com-

edendum, which if nothing else strongly suggests that for Heraclitus
happiness lies in the delights of the soul,20 but perhaps we do not
need Albertus for this unstartling thought.
Stob. 3.1.178 (B 112, here without punctuation) syvqome?m !qet
lec_stg ja sov_g !kgh]a k]ceim ja poie?m jat v}sim 1pa@omtar allows
for several possible construals,21 but, however one parses it, the frag-
ment, although consistent with the ones we have been examining,
clearly goes further in associating human virtue with, superlatively,
mental activity, which, furthermore, is now associated not only with
understanding nature but also with its expression in words, presumably
for the purpose of convincing others of this truth; i. e., in order to better
them. So, then, one construal: The greatest virtue and wisdom is to be
sophronto speak and act truly, perceiving (things) in accord with na-
ture.22 But there is no way to pin this sentence down. Just as Heraclitus
in the river fragment (see above) elevates the puzzle from word to the
underlying idea (from what word is modified by same to what
does it mean to say something is the same?), so too here one puzzles
over sophrosyne, arete, etc., first on the level of grammar but then on a
more philosophical one.
III
Thus far, we have been slowly building up a picture of Heraclitus ethics

from individual fragments and testimony that directly touch upon the
subject. Let us now try a wider approach, based on the idea that Hera-
clitus logos represents (inter alia) an account or system that unifies a dis-
parate array of objects or data: a unity that comprises and makes sense of
20 Again, as in the case of satisfaction, one wonders what Heraclitean word or

root felicitas actually represents. A good guess would be eqdailom-, which would
allow for some Heraclitean punning of the sort we have been examining.
21 See, e. g., Marcovich, H. 96 (who regards these words as a reminiscence of B
114). On the accuracy of Stobaeus quotations from Heraclitus, see Dilcher, H.
21 n. 30, who points out that B 116 is transmitted in a more archaic form by
Sextus than by Stobaeus. In the case of B 112, however, the difficulties mod-
erns have in parsing argues for an accurate transcription by Stobaeus.
22 !kgh]a poie?m is only slightly odd (contra Marcovich) and in any case can be for-
given as a slight zeugma. One may also take syvqos}mg as the predicate of
!qet^ and sov_a, both modified by lec_stg; I dont think Heraclitus would
mind. Or, understand sov_a as the subject of k]ceim. Or . (one could go on).
330 David Sider
what even may be regarded as opposites. Since this logos (see in partic-
ular B 1) seems to be all inclusive, it should, if Heraclitus is to be con-
sistent, serve to frame an ethical theory as well. It comes as no surprise,
therefore, that here too Heraclitus is fond of associating, or rather equat-
ing, what ordinary people think of as opposites.
Should the good man be just, as might seeem obvious in a pre-
Sophistic age? Heraclitus, at least at first glance, seems to be saying
something otherwise: Origen C.Cels. 6.42 (B 80) eQd]<mai> d wq
tm p|kelom 1|mta num|m, ja d_jgm 5qim, ja cim|lema p\mta jat( 5qim
ja wqe~m, it is necessary to know that war is universal and justice is
strife; and that everything comes about in accord with strife and neces-
sity.23 Men may want to throw up their hands at trying to deal with
such an universe, but, as Pangloss might put it, t` lm he` jak
p\mta ja !cah ja d_jaia, %mhqypoi d $ lm %dija rpeik^vasim $
d d_jaia (Porphyry Qu.Hom. ad Il. 4.4 = B 102). Gods, that is, recog-
nize that what men may choose to call strife and justice are in fact sub-
sumed under the latter, which comprises the former as well, much as (in
Greek and other languages) day (the 24-hour cycle) comprises day (the
hours of sunlight) and night.24
If so, mankind too may divide into two camps, comparable to day/
night and justice/strife. The best man and the fool (bk\n, B 87)25 have,
though, more than humanity in common; they are also united by their
city wall as well as by the more abstract civil nomos. D.L. 9.1 (B 44)
l\weshai wq tm d/lom rpq toO m|lou fjyspeq te_weor should prob-
ably not to be taken as an aristocratic statement (as Marcovich does
without argument), but as a more comprehensive one that does not dis-
tinguish social classes, all of whom are protected the one law and the
one wall. The wall, though, solidly visible and fixed in place, has a
23 The necessity of knowing what comes about through necessity makes for
an interesting rhetorical and epistemological kyklos. On this fragment, see fur-
ther Evans (above, n. 3) 41.
24 And in biological terminology, man (the species) comprises man (the gender)
and woman, which allows for an article such as T. McKeown and R. G. Re-
cord, Observations on foetal growth in multiple pregnancy in man, Journal of
Endocrinology 8 (1952) 386 401; cf. the occasional B %mhqypor. It is interesting,
but ultimately not very helpful, to compare Heraclitus statements about
wholes, parts, and opposites with those of Aristotle about universals, partic-
ulars, and negations; cf., e. g., Int. 20a16 31.
25 What Heraclitus thinks about foolish men is nicely surveyed by Dilcher, H. 18
26.
known history and a manifest structure. Where should its equivalent,

the nomos, come from? Not from hoi aristoi, but from ho aristos, as we
saw earlier in the discussion of B 33 and 49, which demonstrate the eth-
ical facet of the one and the many. That is, just as one should listen to
(= comprehend) the logos, which makes sense of the universe, one
should (a fortiori, if the logos is to be all inclusive) look to/atttend to/
comprehend the unity that guides human behavior. The city wall is
its symbolic physical manifestation; a nomos devised by the best man
keeps the demos within less visible but more pervasive limits.26 It is of
course only loosely that one speaks of the (one) nomos, when in fact
what is meant are all the laws of the city and the customs of its citizens.
And a curse on those of his fellow Ephesians, especially hoi aristoi, who
do not recognize the value of ho aristos, saying Bl]ym lgd eXr am-stor
5sty, Let not even one of us be the most beneficial (to the city) (Strabo
14.25 = B 121).27
Still, the (one) nomos should not be lost sight of. Its unifying force is
more explicit in B 114 (Stob. 3.1.179) nm m|\ k]comtar Qswuq_feshai wq
t` num` p\mtym, fjyspeq m|l\ p|kir, ja pok Qswuqot]qyr. tq]vomtai
cq p\mter oR !mhq~peioi m|loi rp 2mr toO he_ou, which makes m|lor
equivalent to t joim|m. Its word play also nicely links the struggle for t
joim|m with an intellectual endeavor (nm m|\ ~ t` num`). Note too how
t joim|m in this political context does not distinguish among classes. Fur-
ther, the simile formed by fjyspeq makes the most important activity of
intelligent people (i. e., those capable of nm m|\ k]comtar) essentially a
political struggle, even if it is for their own ultimate good. (For the last
sentence in this context translates readily into tqev|leha cq p\mter oR
%mhqypoi jtk) Complementary to this thought is D.L. 9.73 (B 47) l
eQj/ peq t_m lec_stym sulbakk~leha. A similar pairing of the political
with the material is D.L. 9.2 (B 43) vbqim wq sbemm}mai lkkom C puq-
jaz^m, where the hybris to be quenched is more likely to be that of an-
26 Thus it is the m|lor which constitutes a city so that it is more than just an as-
sembly of people. It shapes the behaviour of the individuals and so makes a true
unity (Dilcher, H. 49 f.). P. Wheelwright, Heraclitus (Princeton 1959) 88, of-
fers a semireligious idea of Heraclitus city wall which nicely complements the
one offered here: It is a kind of magical encirclement, representing and guar-
anteeing some kind of supernatural protection. See also Evans (above, n. 3)
16 f., who puts B 44 into the larger context of Heraclitus statements about
nomos.
27 Heraclitus somewhat maliciously has the Ephesians indulge in rhetorical over-
kill, as if more than one person could be the the most (of anything).
332 David Sider
other person or group rather than ones own, although I would not want
absolutely to exclude the latter.28
The thrust of B 114, therefore, is for intelligent men to direct their
speech towards the best, in large part because divinely guided, political
activity. As is shown by B 2, that which is common is the logos, aware-
ness of which directs men to live better lives: Sext. Emp. adv.math. 7.133
di de? 6peshai t` num`, toO k|cou d( 1|mtor numoO f~ousim oR pokko
r Qd_am 5womter vq|mgsim. Taken together, then, B 44, 114, and 2 make
a strong case for living an ethical life.29 That is, the fragments we have
looked at are designed to direct mens actions, the 5qca mentioned in B
1, of which some men are ignorant.30
Once again, it seems that, although Heraclitus was born into the
aristocracy, he clearly was not prepared to defend each and every mem-
ber of that class, certainly not those who banished Hermodorus (B 121),
who had been a moloh]tgr for the Ephesians.31 It is even more notewor-
thy that Heraclitus did not despise the lower classes. He seems, in fact, to
have argued that everybody has the same potential for intellectual en-
lightenment: Stob. 3.5.6 (B 116) !mhq~poisi psi l]testi cim~sjeim
2yutor ja syvqome?m and Stob. 3.1.179 (B 113) num|m 1sti psi t
vqom]eim. These are, for Heraclitus time, remarkable statements. What
Heraclitus demanded of himself (cf. Plut. adv. Col. 1118c = B 101 1di-
fgs\lgm 1leyut|m, I examined myself),32 and what he regarded as the
greatest virtue, sophrosyne (see above, on B 112), are open to everybody.
28 An internal hybris would fit well with B 85 hul` l\weshai wakep|m. Arguing
for external hybris is, e. g., C. Diano and G. Serra, Eraclito: I frammenti e le tes-
timonianze (Milan 1980) 186 f., and Marcovich, H. 532 (against Kirk and Ver-
denius), who, however, following his usual line, adducing Theognis, argues that
the hybris is that of the demos alone.
29 Marcovich, in fact, combines B 114 and 2 to produce his fragment 23. It should
also be noted that in these and other fragments Heraclitus does not hesitate to
tell men what they ought to do, which is the hallmark of the ethicist as opposed
to a dispassionate observer of human behavior: wq^ in B 35 (of doubtful authen-
ticity, however), 43, 44, 80, 114; de? in B 2; and %niom in B 121.
30 B 1 (in part) !pe_qoisem 1o_jasi peiq~lemoi ja 1p]ym ja 5qcym toio}tym
bjo_ym digceOlai.
31 For what little testimony there is on Hermodorus, see Mouraviev (above, n. 13)
16 f. An untrustworthy account (Pomponius ap. Digest. 1.2.2.4, Plin. NH
34.21) has Hermodorus advising the first decemvirate in the mid-fifth century,
which would suggest that he was a contemporary of Heraclitus.
32 Which in fact Plutarch and others immediately link to the Delphic cm_hi sau-
t|m. Marcovichs I asked myself and his rejection of the usual meaning are
This helps to explain his anger at fools: they had the chance but re-
jected it. Hence too his attitude towards children, which also is relevant
to his ethical theory. Note Hippol. 9.9.3 4 (B 52) aQm pa?r 1sti pa_fym,
pesse}ym paidr B basikg_g, human life is a child playing childish
games; to the child belongs the power of a king.33 It is not that an
adult should act like a child; quite the reversecf. Celsius ap. Orig.
C.Cels. 6.12 (B 79) !mq m^pior Ejouse pqr da_lomor fjyspeq pa?r
pqr !mdq|r, but in each child is the hope and the potential to be
the one, the best, whose bouk^ will be followed by the city. Children,
however, should not learn from or be influenced by their childish parents:
Marc. Aur. 4.46 (B 74) oq de? pa?dar toje~mym, children should not be
of their parents; that is, they should think and act for themselves, since
parents corrupt their children.34 Note the assumption of childhood inno-
cence in Strabo 14.25 (B 121) %niom 9ves_oir Bbgdm !p\cnashai psi
ja to?r !m^boir tm p|kim jatakipe?m, Hang the Ephesians and leave
the city to the boysboys, by the way, who are capable of leading
their tipsy elders when they have lost their way: Stob. 3.5.7 (B 117)
!mq bj|tam lehush0, %cetai rp paidr !m^bou svakk|lemor, oqj
1pa@ym fjg ba_mei, surely not a sentence meant to be limited to its literal
meaning.
Heraclitus ethics, then, is inextricably linked to his epistemology and
politics. His urging everyone to exercize ones own logos in order to recog-
nize the external logos of the cosmos entails a ethical and political scheme in
which one is persuaded by the one best person, who can only be the one
who exercises this capacity best. An intellectualist theory of ethics, to be
sure, and, further, one that should remind us of that found later in Plato,
especially in the Republic, where also is expressed the extreme view that
it might be advisable to rid the city of adults and start from scratch with
the young, who are to be guided by the best, i. e., most philosophical, citi-
zen. It would be a gross methodological mistake to argue that because the
unacceptable. Heraclitus is merely emphasizing the process through which one

gets to know oneself.
33 <r> pa?r (many editors) is unnecessary.
34 It has to be acknowledged that Marcus Aurelius cites this passage cites B 74
along with others in order to illustrate his immediate point, that being (t
emta) displays not a simple progression but an amazing range of relationships.
His source of Heraclitean fragments, however, is almost certainly simply a list
of excerpts bereft of context. See further my article The Fate of Heraclitus
book in later antiquity, in E. Hlsz (ed.), Nuevos ensayos sobre Herclito: Actas
del Symposium Heracliteum Secundum (Mexico City 2010) 443 458.
334 David Sider
few disjointed fragments of Heraclitus ethical theories are consisent with

the far more extensive and complex theories of Plato, the latter (or Socra-
tes) must have derived his ethics from the formerbut the similarities be-
tween the two are striking, and one may conclude this investigation won-
dering whether there was more in the lost parts of Heraclitus ethics that
would strike us as Platonic.35
Bibliography
Adomenas, Mantas. Heraclitus on religion, Phronesis 44 (1999) 87 113.

Bollack, Jean, & H. Wissmann, Hraclite ou la sparation. Paris 1972.
Bolton, Robert. Nature and human good in Heraclitus, in K. I. Boudouris
(ed.), Ionian Philosophy (Athens 1989) 49 57.
Broniak, Christopher J. Heraclitus, Parmenides, and Plato on living the good
life, Dialogue (Milwaukee) 30 (1987) 28 37.
Capizzi, Antonio. Eraclito e la sua leggenda. Rome 1979.
Cribiore, Raffaella, D. Ratzen, & P. Davoli, A teachers dipinto from Trimi-
this (Dakhleh Oasis), Journal of Roman Archaeology 21 (2008): 170 191.
Deichgrber, Karl. Bemerkungen zu Diogenes Bericht ber Heraklit, Philo-
logus 93 (1938 39) 12 30.
Diano, Carlo, and G. Serra, Eraclito: I frammenti e le testimonianze. Milan 1980.
Eggers Lan, Conrado. Ethical-religious meaning of fr. 30 D.-K. in L. Ros-
setti (ed.), Atti del symposium heracliteum 1981 (Rome 1983) 1.291 299.
Evans, James W. Heraclitus and Parmenides as Moral Philosophers, diss. Yale 1970.
Fattal, Michel. Paroles et actes chez Hraclite: Sur les fondements thoriques de laction
morale. Paris 2012
Hirzel, Rudolf. Untersuchungen zu Ciceros philosophischer Schriften. Leipzig 1882.
Mansfeld, Jaap. The chronology of Anaxagoras Athenian period and the date
of his trial, Mnemosyne 32 (1979) 39 69.
Sider, David. Word order and sense in Heraclitus: Fragment One and the river
The Fate of Heraclitus book in later antiquity, in E. Hlsz (ed.), Nuevos
ensayos sobre Herclito: Actas del Symposium Heracliteum Secundum (Mexico
City 2010) 443 458.
Wheelwright, Philip. Heraclitus. Princeton 1959.
Zoumpos, Anastasios N. Das ethische Urteil bei Heraklit, Platon 11 (1959)
420 423.
35 The rather large literature on Platos use and reminiscence of Heraclitus, none
of which will be cited here, concentrates almost entirely on epistemological and
cosmological matters.
Contributors
Gbor Betegh is Professor of Philosophy at the Central European Uni-
versity, Budapest. He has published primarily on ancient natural philo-
sophy, metaphysics, and theology. He is the author of The Derveni Papy-
rus: Cosmology, Theology and Interpretation (2004).
Roman Dilcher teaches in the philosophy department at the University

of Heidelberg. His philosophic interests include ancient philosophy
(Presocratics, and Aristotle), metaphysics, practical philosophy, phe-
nomenology, hermeneutics, and aesthetics. He has published Studies in
Heraclitus (1995).
Aryeh Finkelberg (PhD 1985, the Hebrew University of Jerusalem) is

Senior Lecturer at the Philosophy Department of Tel Aviv University.
He is the author of a series of articles on the Presocratics and Plato and is
currently finishing a book on Heraclitus.
Daniel W. Graham is Professor of Philosophy at Brigham Young Uni-

versity. He is the president of the International Association for Preso-
cratic Studies. His books include Aristotles Two Systems (1987); Explain-
ing the Cosmos: the Ionian Tradition of Scientific Philosophy (2006); (as co-
editor with Patricia Curd) The Oxford Handbook of Presocratic Philosophy
(2008); The Texts of Early Greek Philosophy (2 vols., 2010); and Science
Before Socrates: Parmenides, Anaxagoras, and the New Astronomy (2013).
Herbert Granger is Professor Emeritus of Philosophy at Wayne State

University. He has published a book on Aristotle (1996) and a number
of articles on Aristotle and on the Presocratic Philosophers.
Carl Huffman is Research Professor at DePauw University. He has

written two books, Philolaus of Croton: Pythagorean and Presocratic and
Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King,
both published by Cambridge University Press, and has edited a third
book, Aristoxenus of Tarentum: Discussion. He has held fellowships
from the John Simon Guggenheim Foundation and the National En-
dowment for the Humanities. He was a visitor at the Institute for Ad-
336 Contributors
vanced Studies in Princeton during the tenure of a fellowship from the

American Council of Learned Societies.
Enrique Hlsz Piccone (Mexico City, 1954) works as a teacher and

researcher in Ancient Philosophy at the National Autonomous Univer-
sity of Mexico (UNAM) since 1978. He authored Logos: Herclito y los
orgenes de la Filosofa (Mexico City, 2011). He was the organizer of the
Second Symposium Heracliteum in 2006, and edited the Proceedings
under the title Nuevos Ensayos sobre Herclito (Mexico City, 2009). On
behalf of the International Association for Presocratic Philosophy
(IAPS), he organized and hosted its III Biennial Conference in Mrida,
Yucatn, 2012). He has published several articles in specialized journals
and collective books, most recently Heraclitus on the sun (2012),
Platos Ionian Muses. Sophist 242 d-e (2013), and Heraclitus on
vsir (2013). His current area of research is centered on Platos recep-
tion of Heraclitus.
Anthony A. Long is Emeritus Professor of Classics and Irving Stone

Professor of Literature, at the University of California, Berkeley. He
is the author and editor of many books on ancient philosophy, including
most recently The Cambridge Companion to Early Greek Philosophy (1999),
Epictetus: A Stoic and Socratic Guide to Life (2002), and From Epicurus to
Epictetus. Studies in Hellenistic and Roman Philosophy (2006). A set of lec-
tures Long delivered in 2012 at Renmin University, Beijing, is to be
published under the title Greek Models of Mind and Self. He is also co-
author, with Margaret Graver, of a forthcoming translation of Senecas
Moral Letters to Lucilius.
Richard McKirahan is the Edwin Clarence Norton Professor of Classics

and Professor of Philosophy at Pomona College in Claremont, California,
USA. He has published a book on Presocratic philosophy (Philosophy Be-
fore Socrates, Hacket 1994, second ed. 2011) and several articles on Preso-
cratic thought. He edited the second edition of A. H. Coxons book on
Parmenides, which received the Philosophy Book of the Year award in
2009 from ForeWord Reviews. He is President of the Society for An-
cient Greek Philosophy, has been an Overseas Visiting Scholar at St.
Johns College, Cambridge and a Fulbright Senior Fellow in Greece,
and has received fellowships from the National Endowment for the Hu-
manities.
Contributors 337
Dirk Obbink, a fellow and tutor at Christ Church, is the Lecturer in

Papyrology and Greek Literature in the Faculty of Classics at Oxford
University. His interests include Greek philosophy, poetry, and religion,
on which he has published widely, including Philodemus On Piety:
Part 1, Critical Text with Commentary (1996); and Anubio: Carmen Astro-
logicum Elegiacum (2006).
Catherine Rowett (formerly Catherine Osborne) is Professor of Philo-

sophy at the University of East Anglia, Norwich U.K. Her relevant
publications in the field of Presocratic Philosophy include Rethinking
Early Greek Philosophy, Presocratic Philosophy: a Very Short Introduction,
and Dumb Beasts and Dead Philosophers, as well as two volumes covering
Philoponus commentary on Aristotles discussion of the Presocratics in
Book 1 of the Physics (in the Ancient Commentators on Aristotle ser-
ies). She has also written widely on Plato, Aristotle, and late antique and
Early Christian thought. She is best known for her views on methodo-
logy in research on Presocratic philosophy (against the use of isolated
fragments, and in favor of proper use of the ancient and late antique in-
terpreters) and for her revisionary interpretation of Empedocles, built
on the one poem hypothesis and an integrated hylozoic reading of
the elements and other beings in Empedocles physical theory (a reading
defended in Rethinking Early Greek Philosophy and a range of later
articles and reviews).
David Sider teaches at New York University and writes on Greek po-
etry and philosophy, especially when they overlap, such as in Empedo-
cles, Parmenides, Plato, Philodemus, and in didactic poetry in general.
He has edited The Fragments of Anaxagoras (second edition, 2005), The
Epigrams of Philodemos (1997), and Theophrastus On Weather Signs (2007).
Leonid Zhmud was born in 1956 in Lvov (Ukraine). He studied at the

Department of Ancient History, Leningrad University. After graduation
in 1982, he taught history at a school for three years. From 1987 he
has been working at the Institute for the History of Science and Technol-
ogy (Russian Academy of the Sciences) in St. Petersburg, currently as
Leading Academic Researcher. In 1990 1992, he was the Alexander
von Humboldt Research Fellow at Constance University (Germany),
in 1995 1996, he was a Junior fellow at the Center for Hellenic Studies
(Washington, DC) and in 1998 1999, he was a member of the Institute
for Advanced Study (Princeton, N.J.). In 2000 2001, he was a fellow of
338 Contributors
the Wellcome Trust for the History of Medicine (London), in 2002

2003, a fellow of Wissenschaftskolleg zu Berlin, and in 2006 2007, a fel-
low of the Netherlands Institute for Advanced Study, Wassenaar. His re-
cent books are The Origin of the History of Science in Classical Antiquity.
(Berlin 2006) and Pythagoras and the Early Pythagoreans (Oxford 2012).
General Index
Abaris 44 45, 48 Aristotle 3 n.1, 7 n.9, 18 21, 27

Academics 33 29, 33 34, 36 40, 50, ch. 3 pas-
Aeneas of Gaza 155 sim, 129, 136, 139, 149, 152, 163,
Aenesidemus 151, 235 n.22, 237 165, 173, 185, 235, 239 242,
n.30, 246 247 265, 282, 284, 303
Akousmata, akousmatikoi 38, 49, 168, Aristoxenus 34, 37, 39, 41 46, 48,
195 56, 74, 76
Alcmaeon 35, 38 39, 41 Artemis 191
Allegory 50, 153 n.14, 194, 196 see Atomists 11, 16
also Style
Ambiguity see Style Barnes, J. 4 7, 13, 16, 21, 174
Ameinias, teacher of Parmenides 44, Beauty 26 27, 29, 208 209
47, 118 Birth 153, 155, 159, 168 170, 175,
Amyclas 44 184, 228 231, 242, 251 252,
Anaxagoras 14, 15 n.26, 28, 36, 121, 306
173, 177 179, 182, 232 233, Bckh, A. 35
250 251, Body 153 160, 173, 231 n.13, 232
Anaximander 6 11, 17, 21 21, 29, 234, 237, 244, 247, 251, 259
34, 167, 174 184, 187 188, 193, Bolton, J. 45
197, 204 Bolton, R. 322
Anaximander the Younger 37, 49 Bow 124 125, 143, 186, 196, 263
50, 141 142, 163 164 266
Anaximenes 163, 169, 174, 179 Breath 129, 151 153, 180, 233, 237
187, 197, 233, 236, 250, 254 257 n.28, 246 247, 254, 259 260
Anthropomorphism 173, 175 176, Brontinus 47 48
187 188, 197 Burkert, W. 3 n.1, 34 35, 38, 43, 72
%peiqom 6 8, 164 n.2, 173 180, n.41, 99, 147
183 184 Burnet, J. 163 n.2, 286 287
!po joimoO 283 n.6, 297
Apollo 44 45, 122, 172, 185, 191, Cebes of Thebes 44, 76
195 n.57, 227, 290, 324 Cerberus 167
Ares 194 Cercops 39 40
Arius Didymus 241, 247, 259 n.75, Chaos 183 184
306 n.10, 314 n.26 Charondas 44
!qw^ 27, 102 103, 133, 136, 255 Chiasmus 177 178
Archytas 4, 23 25, 35, 37, 39, 42 Children, boys 129, 153, 155, 210,
43, 54 56, 59, 76, 140 218, 238, 266 273, 277 280,
Arcturus 192 333
Aristeas 44 45, 48 Cleanthes 241, 246 247, 289 n.24,
Aristophanes 167 n.7, 171 306, 324
340 General Index
Clement of Alexandria 38 Echecrates 43, 75 76

Cleinias 44 Ecphantus 44
Continua 126, 128 135, 137, 139, Empedocles 15, 28, 36, 45, 48, 121,
141 143 133, 147, 150 151, 159, 180
Cornford, F. M. 3, 182, 207, 216, 239, 243, 254, 337
Cosmos 9, 13 14, 17 18, 60, 67 Epicharmus 40
70, 72 74, 97 n.104, 99, 104, Eschatology 147, 160 161, 169,
108 109, 114 115, 118, 121 252
122, 127 128, 130 133, 135 Epigenes 39 40, 48
143, 165 166, 168 169, 172 Ethics 28, 202, 204, 207, 217, 219,
176, 179 184, 186 188, 197, 243 244, 248, 282, 289, 293
207, 220, 222, 226 n.4, 256 257, 299, ch. 12 passim
273 274, 287 n.19, 306, 318 Eudaimonism 248
319 Eudemus 34
Cratylus 304 305, 308, 311 312, Eurytus 39, 43, 56 n.13, 59, 75 76,
314 112, 119
Cube, duplication of 24 n.52, 25 Exhalation 149 153, 156 158,
Cylon 46 239 243, 246, 306
Existence 105 n.33, 318, 321 n.1
Daimn 17 n.36, 147, 150 151, Fire 8, 33, 63, 72, 73 n.43, 76, 89,
155 158 114 115, 125 126, 129 135,
Damon 36, 40 148 149, 152, 158, 182 n.35,
Death 114, 149, 153 158, 160, 186 188, 192 194, 213, 227,
168 170, 190, 196, 228, 231, 231 n.13, 235 239, 241 242,
238, 242, 244, 251 254, 293 n.41 245 246, 253 254
Demiurgy 165 166 Flux 121, 147 n.2, 205, 263 n.2, 282,
Democedes 35, 44, 46 47 326, ch. 11 passim
Democritus 6 n.6, 15, 16 n.33, 28 Form and matter 135
29, 33, 42, 74, 217, n.24, 239
Derveni Papyrus 17 nn.35 36, Gadamer, H.-G. 284 n.8
130 131, 159 n.24, 161 n.30, 166 Graham, D. 125 126
Diatonic scale 128
Dichaearchus 34 Harmony 3 n.1, 6, 10 14, 16 17,
Diocles 43 21, 25, 69 70, 84, 96 n.101, 122
Diodorus of Aspendus 42 126, 128, 130, 134 135, 137
Diodotus grammaticus 323 324 143, 203, ch. 9 passim,
Diogenes of Apollonia 152 n.11, 178 Hecataeus of Miletus 167, 169, 203,
n.26, 232 233, 239, 241, 246, 288
250, 255 Heidegger, M. 284 n.8
Diogenes Laertius 160, 172, 173 Hermodorus of Ephesus 332
n.14, 175 n.20, 178 181, 191, Hesiod 159, 163 n.2, 165 171,
240 241, 293 n.40, 323 324 175 176, 179 n.28, 181 184,
Dionysus 153 n.14, 190, 194 190, 192, 194, 196 97, 202, 212,
Divinity 12, 158 159, ch. 6 passim, 217, 220, 288
212 213, 219 220, 234 Hicetas 44
Division 73 n.43, 118, 283, 314 Hippias 312
Dyad 7, 137 Hippodamus 40
General Index 341
Hippasus 27 n.59, 26, 39, 42, 46 Lycon 41 42

47, 152 Lyre 69, 124 125, 138
Hippocrates of Chios 40, 42, 47
Hippon 36, 38 39, 41, 47, 239 Macrobius 148, 150, 241
Hippys of Rhegium 40 Marcovich, M. 188 n.44, 231 n.13,
Huffman, C. 35, 53, 54 56, 72 75, 328, 252 n.60, 281 n.3, 285 n.11,
77 78, 98 99, 107 109 288, 294 n.43, 296, 303 304,
Hume, D. 316 311 313, 315, 316 n.13, 330, 332
Hybris 297, 331 332 nn.29, 32
Mathemata, mathematikoi 38, 42, 47
Iamblichus 38, 43 44, 46, 49, 155, Measure, metron 201, ch. 7 passim,
158, 195 n.57, 259 253, 287, 289, 296 298, 315
Iccus 36, 46 L]henir 20 n.43
Identity 115 116, 220, 250, 265, Meinwald, C. 137 138
294, 311, 314, 316 317 Melissus 45, 67 n.28, 173
Infinite, infinity see %peiqom Milon 36, 44 46
Intermediaries, intermediates 18 19 Menestor 36, 38, 47
Ion of Chius 40, 48 Metempsychosis 36, 49, 161, 168
Monism 133, 180 n.31, 185 187
Jesus 171 Music (of the spheres) 4, 8 9, 11,
Justice 17 n.36, 29, 61, 76, 79 82, 61, 79, 84, 87, 123, 140
85, 87, 90 92, 95, 97, 106, 111,
114 116, 130 131, 141 142, Narrative 283 285
176 177, 188, 193 194, 197, Nature 10 11, 17, 57, 60, 67 68,
204, 206 208, 211 213, 218 71, 77, 79 84, 87 88, 105, 114,
219, 330 122 123, 127 128, 133, 139,
143, 150 151, ch. 6 passim, 201
Kahn, C. 8 n.13, 123 125, 179 n.1, 202, 204, 209, 215 216,
n.28, 201 n.1, 203, 205, 210, 235, 219 221, 225 227, 282 285,
241 n.39, 244 n.33, 245, 283 n.6, 288, 290 293, 296 298, 318
288 290, 313 314, 317 319, 323 324, 329
Kirk, G. 140, 142, 201 n.1, 214 n.19, Neopythagoreanism 3
231 n.13, 239 n.35, 240, 241 n.39, Number ch. 1 passim, 33, 37 38,
284 n.8, 288, 303, 313, 315 ch. 3 passim, 126 128, 136 140
Kykeon 132 Number atomism 3
Number ratio 124, 126 127, 140,
Law 115, 184, 186 188, 202, 206 143, 209
207, 213, 220, 273, 285, 288, 289
n.24, 292, 319, 323 n.3, 330 331 Ocellus 40
Law, divine 201 n.1, 207, 211 213, Oenopides 40, 42, 47
289, 323 n.3 Olbia, bone plates 160 161
Leibniz 22, 29 Onatas 46
Leucippus 16 n.33, 217 One, The 7 8, 57, 63 67, 69, 71
Life see Death 73, 77, 79, 81, 88 89, 91, 100,
Limited see Unlimited 104 109, 115, 118
Linguistic density 283 n.6, 317 see Opposites 13, 115, 121, 125 126,
also Style 128 130, 134 135, 140 143,
Lucian 74 177, 183, 186 187, 190, 202,
342 General Index
253, 263 265, 273, 282, 292, Races of men 157, 159, 190 191
315 316, 319, 330 see also Unity Ratio 9 10, 12 16, 21, 62, 80 81,
of opposites 116 117, 124, 126 127, 137
Orpheus, Orphics 40, 159 160, 168 140, 202 203, 208 209, 213
n.10, 166, 168 169, 189, 196, 214, 296
258 259 Riddles 252 n.60, 266 267, 269
72
Rivers 287, ch. 11 passim, 326, 329
Pak_mtqopor, pak_mtomor 125
Parmenides 3 n.1, 15, 22, 26, 30, 45,
67, 180, 216 217, 222, 275, 278, Separation 21, 26, 136, 245
319 Simmias 44, 76
Paron 39 40 Simonides 325 327
Petron 39 40 Sleep 123, 151, 153, 155, 190 191,
238, 245 246, 276, 284 285,
Phaethon 58, 193
289 n.26, 291, 306 307
Phanton 43
Solon 218 220
Pherecydes of Syrus 165 167, 169,
Sophocles 188 189, 251
174 n.16, 183 Soul 58, 61, 79 81, 85, 90 92, 95,
Philo of Alexandria 237 97, 114 116, 123, ch. 5 passim,
Philolaos 3 4, 6 7, 13, 15 n.31, 20 168 169, 80, 189 190, 206,
n.44, 22 25, 35 39, 43 44, 47, 208 211, 215 216, ch. 8 passim,
53 54, 56 57, 59, 66 67, 69 274 275, 292 298, 306, 329
77, 98 99, 101 112, 118 119, Speusippus 18 n.39, 34, 94 n.98
ch.4 passim Spintharus 44
Philoponus 255 Symbola 34, 49 50
Philosophy, nature of 281 282 Stoics, Stoicism 149, 152 n.12, 203
Phintias 36 n.3, 204, 229 n.9, 236, 245 247,
Pindar 151 n.8, 216, 218, 252, 325 258 260, 276, 289 nn.24, 27, 306
n.10, 326, 328 n.10, 324
V}sir see Nature Style, prose 115, 161, 167 n.7, 176
Plato 19 20, 23, 24 n.52, 25 26, 182, 185, 190 191, 195, 222,
39, 46 47, 55, 63, 94, 102, 107, 227, 230, 266 267, 285 n.11,
118, 136 138, 155, 157, 165, 168 293, 313 314, 317
n.9, 196 n.58, 201, 205 214, 218, S_la-s/la 155 156, 160, 207
222, 232, 248, 280, 282, 284, 303, n.12, 247 see also Body
305, 308 309, 312, 318, 322, 328 Syvqos}mg 294, n.42, 329, 332 see
n.19, 333 334 also Measure
Platonism 20 n.43, 21, 25
Plotinus 155 n.16 Tetraktys 3 n.1, 29
Plutarch 22, 24 25, 150 151, 157 Taboos 48 50
158, 309 312, 332 n.32 Tarn, L. 304 312
Polyclitus 40 Theagenes of Rhegium 196
Polymnastes 43 Theodoret 157, 328
Polytheism 169, 176, 191 Theodorus of Cyrene 36, 39, 42, 47
Pythagoreans chs. 1 3 passim, 123 Theogony, rhapsodic 168 n.10
124, 136, 167, 168 nn.9 10, 195 Timaeus of Locri 40
196 Tqopa_ 125
General Index 343
Unity of opposites 126, 128, 130, World order see Cosmos

134, 140, 142, 186 187, 205
n.10, 263 265, 273, 282, 292 Xenophon 232
Unlimited 6 7, 13, 59, 63 74, 79, Xenophanes 14 16, 33, 74, 161,
86 90, 99 101, 103 109, 111, 169 179, 184, 187 189, 191
118, 127 143, 192, 195 197, 219 220, 237,
288, 290 n.30, 296 n.56
Vegetarianism 37, 48 Xenophilus 43, 74, 76
Water (not rivers, q.v.) 17, 79, 81 Xuphus 39 40
n.64, 113 114, 132, 134, 141,
149 150, 158, 172, 174 175, Zas 165
181, 186, 214, 228 232, 234 Zaleucus 44
244, 254, 258, 293 n.41 Zeller, E. 33
Index Locorum Potiorum
Aeschylus Meta. 983b6 27

Ag. 1465 6 251 985b23 86a21 78
985b25 62
Alexander Aphrod. 985b27 61
In Meta. 38.10 29, 81 985b31 61
986a1 62
Anaxagoras 986a2 72
B 12 177, 182 986a15 58, 63
986a17 64
Anaximander 986a21 61, 64
A 1.1 175 986b2 118
10 174 986b9 98 n.107
11 176 987a9 118
15 175 987a13 63, 88
B1 141, 176, 178 987a20 118
2 180 987a22 63
987b11 61
Anaximenes 987b22 63
A7 181 986b26 173
B2 233, 254 987b27 19 n.41
987b29 118
Archytas 990a21 60
A 7a 24
1010a12 305
B1 23
1083b8 91
1080b17 60
Aristophanes
Nu. 247 8 167 n.7 1080b30 57, 77
367 167 n.7 1083b11 60, 118
373 171 1083b18 60, 61
374 1090a20 18, 61, 93
1090a22 61
Aristotle 1090a24 61
An. 404a16 58 1090a32 61
405a21 239, 255 1091a12 118
410b27 259 1091a15 63
411a7 152 1092b8 27 n.61
Cael. 268a11 77 1092b10 59
279a32 165 Mete. 324b29 58
300a14 57, 60, 77 345 a13 58
GC 336b28 318 Phys. 203b6 15 7 8
346 Index Locorum Potiorum
Pol. 1267a34 316 1.6 160

Fr. 7 40 1.7 11 148
1.5 323
Aristoxenus 1.12 323
Fr. 77 Mller 11 n.17 1.15 323
5 239
Calcidius 6 304, 305
In Pl.Tim. ch.25 246 n.46 11 240 n.38
15 148, 149, 150, 241
Cicero 21 328
Nat.Deor. 1.107 40 B1 114, 123, 214, 283
3.35 235 n.22 nn.1,6, 285 286,
307, 332 n.30
Clement Alex. 2 285, 307, 332
Strom. 1.21.131 39 n.24 3 17 n.36, 123, 131, 158
3.3.21.1 153 4 328 329
4.14.4 157 5 189
5.14.105.2 153 11 155
6.2.17 258 12 241, 247, 287, ch. 11
passim, 326
P.Derveni 14 160
4.7 8 158 n.24 15 153 n.14, 160
16 193, 298
Carmina Epigraphica Graeca 17 278, 307
396 Hansen 327 18 270, 294
22 294
Diogenes of Apollonia 24 156, 157
B4 232 25 156
26 247 n.50, 254
Empedocles 28 193, 213, 277, 327
B 17.28 150 n.8 29 325, 326
21.9 182 30 114, 130, 185 186,
110.5 150 n.8 207, 213, 253, 287,
112.4 159 289
115.8 159 n.26 31 17, 114, 125, 148 n.3,
126 151 235, 242, 289
32 192, 196, 221, 291
Euripides 33 115, 326
Ba. 395 327 n.16 36 114, 149 150, 227
228, 229 n.9, 242
Galen 243, 252 253, 258,
4.786 K 148, 241 n.41, 244 293 294
39 290
Hecataeus 40 220, 327 n.16
F 27 FGrHist 167 41 188, 213, 219, 275,
291
Heraclitus 43 213, 297, 324, 331
A 1.1 328 n.19 44 213, 324, 330
Index Locorum Potiorum 347
45 13, 215, 293 102 188, 330

47 254, 331 103 295
48 196 107 123, 292
49 326 108 186, 289, 292
49a 303, 305, 314 112 205, 209, 213, 297,
50 142, 185, 215, 291 329
292 113 295, 332
51 124, 186, 263 264 114 187, 206 207, 213,
52 333 285, 289, 291, 322
53 114, 194 n.3, 331 332
54 123, 141, 219, 265, 115 215, 295 296
270 116 295, 322 n.3, 332
55 123 117 153, 237, 297, 333
56 ch. 9 passim 118 148, 153, 157, 298
60 253, 316 119 150, 298
62 153 155, 222, 247,
120 13, 192
252 253
121 331, 332, 333
63 190
64 114, 186, 192, 234 123 122, 219, 288
66 115, 193 129 278
67 12, 17, 115, 126, 177, 135 328
187, 194, 213, 221,
265 n.3 Heraclitus Homericus
69 158 24 305
70 278
72 291 Herodotus
74 333 1.37.3 172
76 236 2.17.1 216
77 153, 155 n.15, 156,
190, 238 Hesiod
78 150, 298 Op. 3 8 195
79 333
80 115, 194, 289 n.25, 121-3 159
330 Fr. 278 M-W 217 n.26
81 160, 203 303 197 n.60
84a 296
85 156, 213 Hippolytus
87 290, 330 Ref. 1.6 6, 8 nn.10 11
88 153 155
89 289 n.26 Isocrates
91 303, 305, 306, 308 Bus. 38 169
311
93 185, 191, 227, 289 Leucippus
94 13, 131, 132, 158 B2 217
n.24, 213, 297
96 159, 189 Maximus of Tyre
101 128, 273, 275, 294 41.4 253 254
101a 277 n.14
Orpheus Tht. 160d 304

T 463 Bernab 159 Tim. 31c-32b 25
Parmenides Plutarch
B 1.30 327 E ap. Delph. 392b 305
7.5 217 Is. 362a-b 153 n.14
8.34 6 22 Qu.Nat. 912a 305
Qu.Conv. 718e 24
Pherecydes v.Rom. 28.8 157
B1 165
2 166 Porphyry
In Ptol.Harm. 1.3 23
Philo
Aet.Mund. 21 237 n.30 Proclus
In Resp. 2.20.23 155
Philolaus
A 29 14 15 Sextus Empiricus
B1 67 AM 7.126 15 n.29
2 67, 127, 132 7.127 30 151, 245
3 67 Pyrrh. 3.230 159
4 22, 68
5 68, 126 Simonides
6 68, 128, 130, 133 531 PMG 325
137, 139
6a 127, 129, 139 Solon
7 68, 73, 129 1 West 218
17 68 69 16 218, 219
17 218
Pindar
P. 3.61 2 251 Sophocles
N. 3.40 328 Ant. 559 60 251
I. 6.71 2 218 1089 327 n.17
Fr. 131b S-M 157
Stobaeus
Plato 1.49.32.104 259
Gorg. 507e-508a 206
525a 208 Theodoret
Lg. 715e-16b 211 Gr.Aff.Cur. 5.18 237 n.29
716c-d2 212 11.7 328 n.18
Phdo. 86c 210 n.15
108c 210 n.15 Theognis
Phdr. 249c 24 1171 2 219
270c 226 227
Resp. 339b 210 n.15 Thucydides
412a 210 1.114.1 92
431b-c 210
466b 210 n.15 Vettius Valens
528a-d 25 317.19 Pingree 259
Xenocrates 24 173, 178

Fr. 9 Heinze 10 n.16, 11 n.17 25 170, 219
26 171
Xenophanes 32 171
A 1.19 173 34 184, 219
12 170 35 219
B1 195
5 188 Xenophon
11 169 Mem. 1.4.8 232
23 169, 173, 219
Contributors
Gbor Betegh is Professor of Philosophy at the Central European Uni-
versity, Budapest. He has published primarily on ancient natural philo-
sophy, metaphysics, and theology. He is the author of The Derveni Papy-
rus: Cosmology, Theology and Interpretation (2004).
Roman Dilcher teaches in the philosophy department at the University

of Heidelberg. His philosophic interests include ancient philosophy
(Presocratics, and Aristotle), metaphysics, practical philosophy, phe-
nomenology, hermeneutics, and aesthetics. He has published Studies in
Heraclitus (1995).
Aryeh Finkelberg (PhD 1985, the Hebrew University of Jerusalem) is

Senior Lecturer at the Philosophy Department of Tel Aviv University.
He is the author of a series of articles on the Presocratics and Plato and is
currently finishing a book on Heraclitus.
Daniel W. Graham is Professor of Philosophy at Brigham Young Uni-

versity. He is the president of the International Association for Preso-
cratic Studies. His books include Aristotles Two Systems (1987); Explain-
ing the Cosmos: the Ionian Tradition of Scientific Philosophy (2006); (as co-
editor with Patricia Curd) The Oxford Handbook of Presocratic Philosophy
(2008); The Texts of Early Greek Philosophy (2 vols., 2010); and Science
Before Socrates: Parmenides, Anaxagoras, and the New Astronomy (2013).
Herbert Granger is Professor Emeritus of Philosophy at Wayne State

University. He has published a book on Aristotle (1996) and a number
of articles on Aristotle and on the Presocratic Philosophers.
Carl Huffman is Research Professor at DePauw University. He has

written two books, Philolaus of Croton: Pythagorean and Presocratic and
Archytas of Tarentum: Pythagorean, Philosopher and Mathematician King,
both published by Cambridge University Press, and has edited a third
book, Aristoxenus of Tarentum: Discussion. He has held fellowships
from the John Simon Guggenheim Foundation and the National En-
dowment for the Humanities. He was a visitor at the Institute for Ad-
336 Contributors
vanced Studies in Princeton during the tenure of a fellowship from the

American Council of Learned Societies.
Enrique Hlsz Piccone (Mexico City, 1954) works as a teacher and

researcher in Ancient Philosophy at the National Autonomous Univer-
sity of Mexico (UNAM) since 1978. He authored Logos: Herclito y los
orgenes de la Filosofa (Mexico City, 2011). He was the organizer of the
Second Symposium Heracliteum in 2006, and edited the Proceedings
under the title Nuevos Ensayos sobre Herclito (Mexico City, 2009). On
behalf of the International Association for Presocratic Philosophy
(IAPS), he organized and hosted its III Biennial Conference in Mrida,
Yucatn, 2012). He has published several articles in specialized journals
and collective books, most recently Heraclitus on the sun (2012),
Platos Ionian Muses. Sophist 242 d-e (2013), and Heraclitus on
vsir (2013). His current area of research is centered on Platos recep-
tion of Heraclitus.
Anthony A. Long is Emeritus Professor of Classics and Irving Stone

Professor of Literature, at the University of California, Berkeley. He
is the author and editor of many books on ancient philosophy, including
most recently The Cambridge Companion to Early Greek Philosophy (1999),
Epictetus: A Stoic and Socratic Guide to Life (2002), and From Epicurus to
Epictetus. Studies in Hellenistic and Roman Philosophy (2006). A set of lec-
tures Long delivered in 2012 at Renmin University, Beijing, is to be
published under the title Greek Models of Mind and Self. He is also co-
author, with Margaret Graver, of a forthcoming translation of Senecas
Moral Letters to Lucilius.
Richard McKirahan is the Edwin Clarence Norton Professor of Classics

and Professor of Philosophy at Pomona College in Claremont, California,
USA. He has published a book on Presocratic philosophy (Philosophy Be-
fore Socrates, Hacket 1994, second ed. 2011) and several articles on Preso-
cratic thought. He edited the second edition of A. H. Coxons book on
Parmenides, which received the Philosophy Book of the Year award in
2009 from ForeWord Reviews. He is President of the Society for An-
cient Greek Philosophy, has been an Overseas Visiting Scholar at St.
Johns College, Cambridge and a Fulbright Senior Fellow in Greece,
and has received fellowships from the National Endowment for the Hu-
manities.
Contributors 337
Dirk Obbink, a fellow and tutor at Christ Church, is the Lecturer in

Papyrology and Greek Literature in the Faculty of Classics at Oxford
University. His interests include Greek philosophy, poetry, and religion,
on which he has published widely, including Philodemus On Piety:
Part 1, Critical Text with Commentary (1996); and Anubio: Carmen Astro-
logicum Elegiacum (2006).
Catherine Rowett (formerly Catherine Osborne) is Professor of Philo-

sophy at the University of East Anglia, Norwich U.K. Her relevant
publications in the field of Presocratic Philosophy include Rethinking
Early Greek Philosophy, Presocratic Philosophy: a Very Short Introduction,
and Dumb Beasts and Dead Philosophers, as well as two volumes covering
Philoponus commentary on Aristotles discussion of the Presocratics in
Book 1 of the Physics (in the Ancient Commentators on Aristotle ser-
ies). She has also written widely on Plato, Aristotle, and late antique and
Early Christian thought. She is best known for her views on methodo-
logy in research on Presocratic philosophy (against the use of isolated
fragments, and in favor of proper use of the ancient and late antique in-
terpreters) and for her revisionary interpretation of Empedocles, built
on the one poem hypothesis and an integrated hylozoic reading of
the elements and other beings in Empedocles physical theory (a reading
defended in Rethinking Early Greek Philosophy and a range of later
articles and reviews).
David Sider teaches at New York University and writes on Greek po-
etry and philosophy, especially when they overlap, such as in Empedo-
cles, Parmenides, Plato, Philodemus, and in didactic poetry in general.
He has edited The Fragments of Anaxagoras (second edition, 2005), The
Epigrams of Philodemos (1997), and Theophrastus On Weather Signs (2007).
Leonid Zhmud was born in 1956 in Lvov (Ukraine). He studied at the

Department of Ancient History, Leningrad University. After graduation
in 1982, he taught history at a school for three years. From 1987 he
has been working at the Institute for the History of Science and Technol-
ogy (Russian Academy of the Sciences) in St. Petersburg, currently as
Leading Academic Researcher. In 1990 1992, he was the Alexander
von Humboldt Research Fellow at Constance University (Germany),
in 1995 1996, he was a Junior fellow at the Center for Hellenic Studies
(Washington, DC) and in 1998 1999, he was a member of the Institute
for Advanced Study (Princeton, N.J.). In 2000 2001, he was a fellow of
338 Contributors
the Wellcome Trust for the History of Medicine (London), in 2002

2003, a fellow of Wissenschaftskolleg zu Berlin, and in 2006 2007, a fel-
low of the Netherlands Institute for Advanced Study, Wassenaar. His re-
cent books are The Origin of the History of Science in Classical Antiquity.
(Berlin 2006) and Pythagoras and the Early Pythagoreans (Oxford 2012).
General Index
Abaris 44 45, 48 Aristotle 3 n.1, 7 n.9, 18 21, 27

Academics 33 29, 33 34, 36 40, 50, ch. 3 pas-
Aeneas of Gaza 155 sim, 129, 136, 139, 149, 152, 163,
Aenesidemus 151, 235 n.22, 237 165, 173, 185, 235, 239 242,
n.30, 246 247 265, 282, 284, 303
Akousmata, akousmatikoi 38, 49, 168, Aristoxenus 34, 37, 39, 41 46, 48,
195 56, 74, 76
Alcmaeon 35, 38 39, 41 Artemis 191
Allegory 50, 153 n.14, 194, 196 see Atomists 11, 16
also Style
Ambiguity see Style Barnes, J. 4 7, 13, 16, 21, 174
Ameinias, teacher of Parmenides 44, Beauty 26 27, 29, 208 209
47, 118 Birth 153, 155, 159, 168 170, 175,
Amyclas 44 184, 228 231, 242, 251 252,
Anaxagoras 14, 15 n.26, 28, 36, 121, 306
173, 177 179, 182, 232 233, Bckh, A. 35
250 251, Body 153 160, 173, 231 n.13, 232
Anaximander 6 11, 17, 21 21, 29, 234, 237, 244, 247, 251, 259
34, 167, 174 184, 187 188, 193, Bolton, J. 45
197, 204 Bolton, R. 322
Anaximander the Younger 37, 49 Bow 124 125, 143, 186, 196, 263
50, 141 142, 163 164 266
Anaximenes 163, 169, 174, 179 Breath 129, 151 153, 180, 233, 237
187, 197, 233, 236, 250, 254 257 n.28, 246 247, 254, 259 260
Anthropomorphism 173, 175 176, Brontinus 47 48
187 188, 197 Burkert, W. 3 n.1, 34 35, 38, 43, 72
%peiqom 6 8, 164 n.2, 173 180, n.41, 99, 147
183 184 Burnet, J. 163 n.2, 286 287
!po joimoO 283 n.6, 297
Apollo 44 45, 122, 172, 185, 191, Cebes of Thebes 44, 76
195 n.57, 227, 290, 324 Cerberus 167
Ares 194 Cercops 39 40
Arius Didymus 241, 247, 259 n.75, Chaos 183 184
306 n.10, 314 n.26 Charondas 44
!qw^ 27, 102 103, 133, 136, 255 Chiasmus 177 178
Archytas 4, 23 25, 35, 37, 39, 42 Children, boys 129, 153, 155, 210,
43, 54 56, 59, 76, 140 218, 238, 266 273, 277 280,
Arcturus 192 333
Aristeas 44 45, 48 Cleanthes 241, 246 247, 289 n.24,
Aristophanes 167 n.7, 171 306, 324
340 General Index
Clement of Alexandria 38 Echecrates 43, 75 76

Cleinias 44 Ecphantus 44
Continua 126, 128 135, 137, 139, Empedocles 15, 28, 36, 45, 48, 121,
141 143 133, 147, 150 151, 159, 180
Cornford, F. M. 3, 182, 207, 216, 239, 243, 254, 337
Cosmos 9, 13 14, 17 18, 60, 67 Epicharmus 40
70, 72 74, 97 n.104, 99, 104, Eschatology 147, 160 161, 169,
108 109, 114 115, 118, 121 252
122, 127 128, 130 133, 135 Epigenes 39 40, 48
143, 165 166, 168 169, 172 Ethics 28, 202, 204, 207, 217, 219,
176, 179 184, 186 188, 197, 243 244, 248, 282, 289, 293
207, 220, 222, 226 n.4, 256 257, 299, ch. 12 passim
273 274, 287 n.19, 306, 318 Eudaimonism 248
319 Eudemus 34
Cratylus 304 305, 308, 311 312, Eurytus 39, 43, 56 n.13, 59, 75 76,
314 112, 119
Cube, duplication of 24 n.52, 25 Exhalation 149 153, 156 158,
Cylon 46 239 243, 246, 306
Existence 105 n.33, 318, 321 n.1
Daimn 17 n.36, 147, 150 151, Fire 8, 33, 63, 72, 73 n.43, 76, 89,
155 158 114 115, 125 126, 129 135,
Damon 36, 40 148 149, 152, 158, 182 n.35,
Death 114, 149, 153 158, 160, 186 188, 192 194, 213, 227,
168 170, 190, 196, 228, 231, 231 n.13, 235 239, 241 242,
238, 242, 244, 251 254, 293 n.41 245 246, 253 254
Demiurgy 165 166 Flux 121, 147 n.2, 205, 263 n.2, 282,
Democedes 35, 44, 46 47 326, ch. 11 passim
Democritus 6 n.6, 15, 16 n.33, 28 Form and matter 135
29, 33, 42, 74, 217, n.24, 239
Derveni Papyrus 17 nn.35 36, Gadamer, H.-G. 284 n.8
130 131, 159 n.24, 161 n.30, 166 Graham, D. 125 126
Diatonic scale 128
Dichaearchus 34 Harmony 3 n.1, 6, 10 14, 16 17,
Diocles 43 21, 25, 69 70, 84, 96 n.101, 122
Diodorus of Aspendus 42 126, 128, 130, 134 135, 137
Diodotus grammaticus 323 324 143, 203, ch. 9 passim,
Diogenes of Apollonia 152 n.11, 178 Hecataeus of Miletus 167, 169, 203,
n.26, 232 233, 239, 241, 246, 288
250, 255 Heidegger, M. 284 n.8
Diogenes Laertius 160, 172, 173 Hermodorus of Ephesus 332
n.14, 175 n.20, 178 181, 191, Hesiod 159, 163 n.2, 165 171,
240 241, 293 n.40, 323 324 175 176, 179 n.28, 181 184,
Dionysus 153 n.14, 190, 194 190, 192, 194, 196 97, 202, 212,
Divinity 12, 158 159, ch. 6 passim, 217, 220, 288
212 213, 219 220, 234 Hicetas 44
Division 73 n.43, 118, 283, 314 Hippias 312
Dyad 7, 137 Hippodamus 40
General Index 341
Hippasus 27 n.59, 26, 39, 42, 46 Lycon 41 42

47, 152 Lyre 69, 124 125, 138
Hippocrates of Chios 40, 42, 47
Hippon 36, 38 39, 41, 47, 239 Macrobius 148, 150, 241
Hippys of Rhegium 40 Marcovich, M. 188 n.44, 231 n.13,
Huffman, C. 35, 53, 54 56, 72 75, 328, 252 n.60, 281 n.3, 285 n.11,
77 78, 98 99, 107 109 288, 294 n.43, 296, 303 304,
Hume, D. 316 311 313, 315, 316 n.13, 330, 332
Hybris 297, 331 332 nn.29, 32
Mathemata, mathematikoi 38, 42, 47
Iamblichus 38, 43 44, 46, 49, 155, Measure, metron 201, ch. 7 passim,
158, 195 n.57, 259 253, 287, 289, 296 298, 315
Iccus 36, 46 L]henir 20 n.43
Identity 115 116, 220, 250, 265, Meinwald, C. 137 138
294, 311, 314, 316 317 Melissus 45, 67 n.28, 173
Infinite, infinity see %peiqom Milon 36, 44 46
Intermediaries, intermediates 18 19 Menestor 36, 38, 47
Ion of Chius 40, 48 Metempsychosis 36, 49, 161, 168
Monism 133, 180 n.31, 185 187
Jesus 171 Music (of the spheres) 4, 8 9, 11,
Justice 17 n.36, 29, 61, 76, 79 82, 61, 79, 84, 87, 123, 140
85, 87, 90 92, 95, 97, 106, 111,
114 116, 130 131, 141 142, Narrative 283 285
176 177, 188, 193 194, 197, Nature 10 11, 17, 57, 60, 67 68,
204, 206 208, 211 213, 218 71, 77, 79 84, 87 88, 105, 114,
219, 330 122 123, 127 128, 133, 139,
143, 150 151, ch. 6 passim, 201
Kahn, C. 8 n.13, 123 125, 179 n.1, 202, 204, 209, 215 216,
n.28, 201 n.1, 203, 205, 210, 235, 219 221, 225 227, 282 285,
241 n.39, 244 n.33, 245, 283 n.6, 288, 290 293, 296 298, 318
288 290, 313 314, 317 319, 323 324, 329
Kirk, G. 140, 142, 201 n.1, 214 n.19, Neopythagoreanism 3
231 n.13, 239 n.35, 240, 241 n.39, Number ch. 1 passim, 33, 37 38,
284 n.8, 288, 303, 313, 315 ch. 3 passim, 126 128, 136 140
Kykeon 132 Number atomism 3
Number ratio 124, 126 127, 140,
Law 115, 184, 186 188, 202, 206 143, 209
207, 213, 220, 273, 285, 288, 289
n.24, 292, 319, 323 n.3, 330 331 Ocellus 40
Law, divine 201 n.1, 207, 211 213, Oenopides 40, 42, 47
289, 323 n.3 Olbia, bone plates 160 161
Leibniz 22, 29 Onatas 46
Leucippus 16 n.33, 217 One, The 7 8, 57, 63 67, 69, 71
Life see Death 73, 77, 79, 81, 88 89, 91, 100,
Limited see Unlimited 104 109, 115, 118
Linguistic density 283 n.6, 317 see Opposites 13, 115, 121, 125 126,
also Style 128 130, 134 135, 140 143,
Lucian 74 177, 183, 186 187, 190, 202,
342 General Index
253, 263 265, 273, 282, 292, Races of men 157, 159, 190 191
315 316, 319, 330 see also Unity Ratio 9 10, 12 16, 21, 62, 80 81,
of opposites 116 117, 124, 126 127, 137
Orpheus, Orphics 40, 159 160, 168 140, 202 203, 208 209, 213
n.10, 166, 168 169, 189, 196, 214, 296
258 259 Riddles 252 n.60, 266 267, 269
72
Rivers 287, ch. 11 passim, 326, 329
Pak_mtqopor, pak_mtomor 125
Parmenides 3 n.1, 15, 22, 26, 30, 45,
67, 180, 216 217, 222, 275, 278, Separation 21, 26, 136, 245
319 Simmias 44, 76
Paron 39 40 Simonides 325 327
Petron 39 40 Sleep 123, 151, 153, 155, 190 191,
238, 245 246, 276, 284 285,
Phaethon 58, 193
289 n.26, 291, 306 307
Phanton 43
Solon 218 220
Pherecydes of Syrus 165 167, 169,
Sophocles 188 189, 251
174 n.16, 183 Soul 58, 61, 79 81, 85, 90 92, 95,
Philo of Alexandria 237 97, 114 116, 123, ch. 5 passim,
Philolaos 3 4, 6 7, 13, 15 n.31, 20 168 169, 80, 189 190, 206,
n.44, 22 25, 35 39, 43 44, 47, 208 211, 215 216, ch. 8 passim,
53 54, 56 57, 59, 66 67, 69 274 275, 292 298, 306, 329
77, 98 99, 101 112, 118 119, Speusippus 18 n.39, 34, 94 n.98
ch.4 passim Spintharus 44
Philoponus 255 Symbola 34, 49 50
Philosophy, nature of 281 282 Stoics, Stoicism 149, 152 n.12, 203
Phintias 36 n.3, 204, 229 n.9, 236, 245 247,
Pindar 151 n.8, 216, 218, 252, 325 258 260, 276, 289 nn.24, 27, 306
n.10, 326, 328 n.10, 324
V}sir see Nature Style, prose 115, 161, 167 n.7, 176
Plato 19 20, 23, 24 n.52, 25 26, 182, 185, 190 191, 195, 222,
39, 46 47, 55, 63, 94, 102, 107, 227, 230, 266 267, 285 n.11,
118, 136 138, 155, 157, 165, 168 293, 313 314, 317
n.9, 196 n.58, 201, 205 214, 218, S_la-s/la 155 156, 160, 207
222, 232, 248, 280, 282, 284, 303, n.12, 247 see also Body
305, 308 309, 312, 318, 322, 328 Syvqos}mg 294, n.42, 329, 332 see
n.19, 333 334 also Measure
Platonism 20 n.43, 21, 25
Plotinus 155 n.16 Tetraktys 3 n.1, 29
Plutarch 22, 24 25, 150 151, 157 Taboos 48 50
158, 309 312, 332 n.32 Tarn, L. 304 312
Polyclitus 40 Theagenes of Rhegium 196
Polymnastes 43 Theodoret 157, 328
Polytheism 169, 176, 191 Theodorus of Cyrene 36, 39, 42, 47
Pythagoreans chs. 1 3 passim, 123 Theogony, rhapsodic 168 n.10
124, 136, 167, 168 nn.9 10, 195 Timaeus of Locri 40
196 Tqopa_ 125
General Index 343
Unity of opposites 126, 128, 130, World order see Cosmos

134, 140, 142, 186 187, 205
n.10, 263 265, 273, 282, 292 Xenophon 232
Unlimited 6 7, 13, 59, 63 74, 79, Xenophanes 14 16, 33, 74, 161,
86 90, 99 101, 103 109, 111, 169 179, 184, 187 189, 191
118, 127 143, 192, 195 197, 219 220, 237,
288, 290 n.30, 296 n.56
Vegetarianism 37, 48 Xenophilus 43, 74, 76
Water (not rivers, q.v.) 17, 79, 81 Xuphus 39 40
n.64, 113 114, 132, 134, 141,
149 150, 158, 172, 174 175, Zas 165
181, 186, 214, 228 232, 234 Zaleucus 44
244, 254, 258, 293 n.41 Zeller, E. 33
Index Locorum Potiorum
Aeschylus Meta. 983b6 27

Ag. 1465 6 251 985b23 86a21 78
985b25 62
Alexander Aphrod. 985b27 61
In Meta. 38.10 29, 81 985b31 61
986a1 62
Anaxagoras 986a2 72
B 12 177, 182 986a15 58, 63
986a17 64
Anaximander 986a21 61, 64
A 1.1 175 986b2 118
10 174 986b9 98 n.107
11 176 987a9 118
15 175 987a13 63, 88
B1 141, 176, 178 987a20 118
2 180 987a22 63
987b11 61
Anaximenes 987b22 63
A7 181 986b26 173
B2 233, 254 987b27 19 n.41
987b29 118
Archytas 990a21 60
A 7a 24
1010a12 305
B1 23
1083b8 91
1080b17 60
Aristophanes
Nu. 247 8 167 n.7 1080b30 57, 77
367 167 n.7 1083b11 60, 118
373 171 1083b18 60, 61
374 1090a20 18, 61, 93
1090a22 61
Aristotle 1090a24 61
An. 404a16 58 1090a32 61
405a21 239, 255 1091a12 118
410b27 259 1091a15 63
411a7 152 1092b8 27 n.61
Cael. 268a11 77 1092b10 59
279a32 165 Mete. 324b29 58
300a14 57, 60, 77 345 a13 58
GC 336b28 318 Phys. 203b6 15 7 8
Pol. 1267a34 316 1.6 160

Fr. 7 40 1.7 11 148
1.5 323
Aristoxenus 1.12 323
Fr. 77 Mller 11 n.17 1.15 323
5 239
Calcidius 6 304, 305
In Pl.Tim. ch.25 246 n.46 11 240 n.38
15 148, 149, 150, 241
Cicero 21 328
Nat.Deor. 1.107 40 B1 114, 123, 214, 283
3.35 235 n.22 nn.1,6, 285 286,
307, 332 n.30
Clement Alex. 2 285, 307, 332
Strom. 1.21.131 39 n.24 3 17 n.36, 123, 131, 158
3.3.21.1 153 4 328 329
4.14.4 157 5 189
5.14.105.2 153 11 155
6.2.17 258 12 241, 247, 287, ch. 11
passim, 326
P.Derveni 14 160
4.7 8 158 n.24 15 153 n.14, 160
16 193, 298
Carmina Epigraphica Graeca 17 278, 307
396 Hansen 327 18 270, 294
22 294
Diogenes of Apollonia 24 156, 157
B4 232 25 156
26 247 n.50, 254
Empedocles 28 193, 213, 277, 327
B 17.28 150 n.8 29 325, 326
21.9 182 30 114, 130, 185 186,
110.5 150 n.8 207, 213, 253, 287,
112.4 159 289
115.8 159 n.26 31 17, 114, 125, 148 n.3,
126 151 235, 242, 289
32 192, 196, 221, 291
Euripides 33 115, 326
Ba. 395 327 n.16 36 114, 149 150, 227
228, 229 n.9, 242
Galen 243, 252 253, 258,
4.786 K 148, 241 n.41, 244 293 294
39 290
Hecataeus 40 220, 327 n.16
F 27 FGrHist 167 41 188, 213, 219, 275,
291
Heraclitus 43 213, 297, 324, 331
A 1.1 328 n.19 44 213, 324, 330
45 13, 215, 293 102 188, 330

47 254, 331 103 295
48 196 107 123, 292
49 326 108 186, 289, 292
49a 303, 305, 314 112 205, 209, 213, 297,
50 142, 185, 215, 291 329
292 113 295, 332
51 124, 186, 263 264 114 187, 206 207, 213,
52 333 285, 289, 291, 322
53 114, 194 n.3, 331 332
54 123, 141, 219, 265, 115 215, 295 296
270 116 295, 322 n.3, 332
55 123 117 153, 237, 297, 333
56 ch. 9 passim 118 148, 153, 157, 298
60 253, 316 119 150, 298
62 153 155, 222, 247,
120 13, 192
252 253
121 331, 332, 333
63 190
64 114, 186, 192, 234 123 122, 219, 288
66 115, 193 129 278
67 12, 17, 115, 126, 177, 135 328
187, 194, 213, 221,
265 n.3 Heraclitus Homericus
69 158 24 305
70 278
72 291 Herodotus
74 333 1.37.3 172
76 236 2.17.1 216
77 153, 155 n.15, 156,
190, 238 Hesiod
78 150, 298 Op. 3 8 195
79 333
80 115, 194, 289 n.25, 121-3 159
330 Fr. 278 M-W 217 n.26
81 160, 203 303 197 n.60
84a 296
85 156, 213 Hippolytus
87 290, 330 Ref. 1.6 6, 8 nn.10 11
88 153 155
89 289 n.26 Isocrates
91 303, 305, 306, 308 Bus. 38 169
311
93 185, 191, 227, 289 Leucippus
94 13, 131, 132, 158 B2 217
n.24, 213, 297
96 159, 189 Maximus of Tyre
101 128, 273, 275, 294 41.4 253 254
101a 277 n.14
Orpheus Tht. 160d 304

T 463 Bernab 159 Tim. 31c-32b 25
Parmenides Plutarch
B 1.30 327 E ap. Delph. 392b 305
7.5 217 Is. 362a-b 153 n.14
8.34 6 22 Qu.Nat. 912a 305
Qu.Conv. 718e 24
Pherecydes v.Rom. 28.8 157
B1 165
2 166 Porphyry
In Ptol.Harm. 1.3 23
Philo
Aet.Mund. 21 237 n.30 Proclus
In Resp. 2.20.23 155
Philolaus
A 29 14 15 Sextus Empiricus
B1 67 AM 7.126 15 n.29
2 67, 127, 132 7.127 30 151, 245
3 67 Pyrrh. 3.230 159
4 22, 68
5 68, 126 Simonides
6 68, 128, 130, 133 531 PMG 325
137, 139
6a 127, 129, 139 Solon
7 68, 73, 129 1 West 218
17 68 69 16 218, 219
17 218
Pindar
P. 3.61 2 251 Sophocles
N. 3.40 328 Ant. 559 60 251
I. 6.71 2 218 1089 327 n.17
Fr. 131b S-M 157
Stobaeus
Plato 1.49.32.104 259
Gorg. 507e-508a 206
525a 208 Theodoret
Lg. 715e-16b 211 Gr.Aff.Cur. 5.18 237 n.29
716c-d2 212 11.7 328 n.18
Phdo. 86c 210 n.15
108c 210 n.15 Theognis
Phdr. 249c 24 1171 2 219
270c 226 227
Resp. 339b 210 n.15 Thucydides
412a 210 1.114.1 92
431b-c 210
466b 210 n.15 Vettius Valens
528a-d 25 317.19 Pingree 259
Xenocrates 24 173, 178

Fr. 9 Heinze 10 n.16, 11 n.17 25 170, 219
26 171
Xenophanes 32 171
A 1.19 173 34 184, 219
12 170 35 219
B1 195
5 188 Xenophon
11 169 Mem. 1.4.8 232
23 169, 173, 219

(Sozomena 14) David Sider, Dirk Obbink (Eds.) - Doctrine and Doxography - Studies On Heraclitus and Pythagoras-Walter de Gruyter (20

Uploaded by

Copyright:

Available Formats

(Sozomena 14) David Sider, Dirk Obbink (Eds.) - Doctrine and Doxography - Studies On Heraclitus and Pythagoras-Walter de Gruyter (20

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

(Sozomena 14) David Sider, Dirk Obbink (Eds.) - Doctrine and Doxography - Studies On Heraclitus and Pythagoras-Walter de Gruyter (20

Uploaded by

Copyright:

Available Formats

Doctrine and Doxography

David Sider and Dirk Obbink

Library of Congress Cataloging-in-Publication Data

Bibliographic information published by the Deutsche Nationalbibliothek

2013 Walter de Gruyter GmbH, Berlin/Boston

10. Enrique Hlsz

11. Daniel W. Graham

12. David Sider

We do not know a great deal about Pythagoreanism in the early days

1 As regards Pythagoras himself, this is somewhat controversial. Burkert denies

But what kind of a passion for numbers do we find in the Pythagor-

2 J. Barnes, The Presocratic Philosophers (London 1982) 100 101.

And thirdly he invites us instead to condemn the whole Pythagorean

Pythagoras and the early Presocratic tradition.

One not particularly original route to defending the Pythagoreans

7 The terminology is widespread, particularly in material from Philolaus. See Phi-

And this is what is divine, for it is deathless and indestructible, as Anax-

10 Hippol. Ref 1.6.

thought, in D. L. Couprie, Robert Hahn, and Gerard Naddaf, Anaximander in

grounded, fantasy about the relative sizes of the heavenly rings, he is

any particular occurrence.20 These high-minded practices have an unfore-

posites) is also seen as a key concept in Heraclitus thought, as it is in

22 The impression that the Pythagoreans have a systematic technical terminology

rpost\seyr !qihl|r di ja b jqitr t_m p\mtym k|cor oqj !l]towor m

in Heraclitus too.32 But there is no reason to think that the addition of

32 E.g., B 30, 31, 79, 82, 83.

Pythagorean mathematics, by contrast, tries to make numbers do the

37 See Furley (above, n. 14) 52 3 for a brief discussion of this thought.

ematical phenomena, he supposes, made the Pythagoreans go for the

So the Pythagoreans have just the one world, but it is apparently

rable to Platos notion of l]henir. It seems that at 987b11 Aristotle is assimilating

From Incomprehension to Admiration

In this second part of the paper I shall move beyond my initial

46 Philolaus B 4 5; Stobaeus 1.21.7b-c.

counts as knowing unless it is knowledge of numbers. There is only one

49 See Huffman (above, n. 9) 173 7, responding to M. C Nussbaum, Eleatic con-

with the harmony of the universe from assimilation to mathematical

58 There are, of course, notable exceptions in modern scholarship on the Preso-

planatory principles of the universe is not just methodologically sound,

62 My diagnosis of the source of Aristotles incomprehension differs somewhat

than Aristotles. A deep set ideological preference for mechanistic theory

64 Gottfried Wilhelm Leibniz: Correspondence with Clarke: Leibnizs second paper,

In the middle of his doxographical discussion of Presocratic theories, Ar-

1 E. Zeller, Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung. 5 Vol. 1.

Anaximander the Younger, wrote a book on the Pythagorean symbola

whereas Riedweg8 views this as 98 % religion and 2 % philosophy, with

8 C. Riedweg, Pythagoras. Leben, Lehre, Nachwirkung. Munich 2002.

13 See Burkert L&S (above, n. 6) 166 ff.

16 Burkert L&S (above, n. 6) 180 ff.

17 Zhmud Pythagoras (above, n. 9) 394 ff.

22 H. Diels, Antike Technik (Leipzig 1924) 23.

25 Cic. De nat. deor. 1, 107 = Arist. fr. 7 Rose.

of Lucania (DK 48) is mentioned in the catalogue, which means that

31 Ibid., 124 ff.

34 Pace Huffman Archytas (above, n. 10) 51 ff.

Now I come back to Aristoxenus catalogue of the Pythagoreans. The

36 E. Rohde, Die Quellen des Iamblichus in seiner Biographie des Pythagoras

In the catalogue, he is duly registered among the Pythagoreans from

38 J. Bolton, Aristeas of Proconnesus (Oxford 1962) 151 ff.