my recognizing and acknowledging the virtues of the construction of QMR

itself. So, finally, I am keen to make clear that I learned a great deal from this
book, and that I found much in it to agree with and more to admire. The
Nature of Contingency deserves to be read by researchers in the many fields
that it covers and will – I expect – be widely appreciated for the imaginative
1364impressive
and Book Reviews
work that it is.

References
Divers, John and Jade, Fletcher 2020, ‘(Once again) Lewis on the Analysis
of Modality’, in Synthese 197.

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

Everett, Hugh 1973, ‘The Theory of the Universal Wave Function’, in B.,
DeWitt and N., Graham, The Many-Words Interpretation of Quanum
Mechanics (Princeton NJ: Princeton University Press)..
Kripke, Saul 1959, ‘A Completeness Theorem in Modal Logic’, in Journal of
Symbolic Logic 24 1963, ‘Semantical Considerations on Modal Logic’ in.
Acta Philosophica Fennica, 16 1980, Naming and Necessity (Oxford:
Blackwell).
Lewis, David 1968, ‘Counterpart Theory and Quantified Modal Logic’, in
Journal of Philosophy 65.
—— 1986, On The Plurality of Worlds (Oxford: Blackwell).
—— 1992, ‘Review of D.M. Armstrong, A Combinatorial Theory of
Possibility’, in Australasian Journal of Philosophy, 70.
Plantinga, Alvin 1987, ‘Two Concepts of Modality. Modal Realism and
Modal Reductionism’ in Philosophical Perspectives..
Quine, Willard V.O. 1981, ‘Response to Kripke’, in his Theories and Things
(Cambridge MA: Harvard University Press).
Shalkowski, Scott 1994, ‘The Ontological Ground of the Alethic Modality’,
8 in Book Review Review, 103.
Philosophical
Stalknaker, Robert 2012, Mere Possibilities (Princeton NJ: University of
Princeton Press).
van Inwagen, Peter 2006, ‘In Defence of Magical Ersatzism’ in Philosophical
Quarterly 56.
Mind, Vol. 00 . 0 . 2021  Mind Association 2021
Williamson, Timothy 2013, Modal Logic as Metaphysics (Oxford: Oxford
University Press).

Book Review
University of Leeds, England
doi: 10.1093/mind/fzab004
john divers

The Primacy of Metaphysics, by Christopher Peacocke. Oxford: Oxford

University Press, 2019. Pp. xiii þ 218.
Peacocke introduces The Primacy of Metaphysics with the apt observation
that ‘[t]here can be few issues as fundamental as the relation between the
metaphysics of some domain and our ways of thinking about it’ (p. 1). Is
there, for example, any relation of explanatory priority between them?
Peacocke’s treatment of this topic is, characteristically, challenging, deep,
and subtle. My critique here will try to engage the book in its own terms.
The book proposes that ways of thinking about a domain cannot be ex-
planatorily prior to its metaphysics: such ‘meaning-first’ positions are
excluded in principle. Peacocke concludes that the space of philosophical
possibilities is exhausted by the union of ‘metaphysics-first’ cases, in which
the metaphysics is explanatorily prior, and ‘no-priority’ cases, in which nei-
ther metaphysics nor accounts of ways-of-thinking-about are explanatorily
Mind, Vol. 131 . 524 . October 2022 © Mind Association 2022
prior. Here I offer some reasons for hesitation.

1. Preliminaries to primacy
Let me preview two broad considerations that might keep one from adopting
metaphysics of some domain and our ways of thinking about it’ (p. 1). Is
there, for example, any relation of explanatory priority between them?
Peacocke’s treatment of this topic is, characteristically, challenging, deep,
and subtle. My critique here will try to engage the book in its own terms.
The book proposes that ways of thinking about a domain cannot be ex-
Book Reviews
planatorily prior to its metaphysics: such ‘meaning-first’ positions1365
are
excluded in principle. Peacocke concludes that the space of philosophical
possibilities is exhausted by the union of ‘metaphysics-first’ cases, in which
the metaphysics is explanatorily prior, and ‘no-priority’ cases, in which nei-
ther metaphysics nor accounts of ways-of-thinking-about are explanatorily
prior. Here I offer some reasons for hesitation.

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

1. Preliminaries to primacy
Let me preview two broad considerations that might keep one from adopting
Peacocke’s position. I will then return to these considerations, afterward, for
additional discussion.
First broad consideration: Peacocke’s claims recommend attention to the
very idea of explanatory priority. Sometimes, asking whether one sort of thing
is explanatorily prior to another is like asking whether the moon is to the left
of the sun. Not only is there no good answer, neither is it quite just a ‘no-
direction’ case; one element of my hesitation about Peacocke’s view is un-
certainty about whether the relation of explanatory priority itself holds only
relative to some other parameter.
Three related qualms around this first consideration:
2 (i) Book Review
Should explanatory priority be viewed as asymmetric? Might there
fail to be no priority even while it is not the case for either mean-
ings or metaphysics that they are ultimately asymmetrically prior to
the other , explanatorily (because there are multiple relations of
priority, including a kind of priority for metaphysical accounts)?
(ii) How do relations between the metaphysics of a domain and our
explanations of or accounts of intentionality connect
Mind, Vol. 00 . 0 . 2021
to relations
 Mind Association
between metaphysical matters and intentional entities themselves?
doi:10.1093/mind/fzab028
Couldn’t our understanding of the metaphysics of a domain be in a
way derivative from ways of thinking about it (so that there’s a
priority of intentional items with respect to our account of the
metaphysics) even while the elements of the domain themselves
are not understood as in any way derivative from anything inten-
tional (so that there’s no ‘corresponding’ relation of priority among
all the relevant entities themselves)? Involvement in explanation is
not itself the explanation of an involvement.
(iii) Might there be non-explanatory varieties of involvement in an ex-
planation? Is what Peacocke calls ‘the Primary Thesis’ (to be con-
sidered in detail below) validated or rather refuted by an
arrangement on which the metaphysics of a domain and the con-
tents and meanings about it are each necessarily mentioned in any
account of the other but without that mention itself serving an
explanatory role? Perhaps there’s no giving an elucidation of the
metaphysics of a domain, or of our ways of thinking about it,
without making presuppositions about the other.
Second broad consideration : Peacocke offers two general arguments against
meaning-first
Mind, views—the
Vol. 131 . 524 . October‘argument
2022 from understanding’ and Association
© Mind the ‘argument
2022
from rationality’. The argument from understanding endeavours to convert a
point about intentionality into a conclusion seen as incompatible with any
meaning-first view. The key thesis is that ‘[w]hich relations a thinker can
stand in to an entity depends on the correct metaphysics of that entity’ (pp.
11-12, emphasis added). Later, the argument from rationality begins (p. 25)
 sidered in detail below) validated or rather refuted by an
arrangement on which the metaphysics of a domain and the con-
tents and meanings about it are each necessarily mentioned in any
account of the other but without that mention itself serving an
1366 explanatory
Book Reviewsrole? Perhaps there’s no giving an elucidation of the
metaphysics of a domain, or of our ways of thinking about it,
without making presuppositions about the other.
Second broad consideration : Peacocke offers two general arguments against
meaning-first views—the ‘argument from understanding’ and the ‘argument
from rationality’. The argument from understanding endeavours to convert a

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

point about intentionality into a conclusion seen as incompatible with any
meaning-first view. The key thesis is that ‘[w]hich relations a thinker can
stand in to an entity depends on the correct metaphysics of that entity’ (pp.
11-12, emphasis added). Later, the argument from rationality begins (p. 25)
with a thesis about the dependence of the rationality of judgment on what it
takes for that judgment to be true. That thesis is then used to extract a
position on which an account of grasping a content must have built into
it some sensitivity to the metaphysics of the items thought about. Notice that
in both cases, the crux involves a relation of dependence.
These two arguments are supposed to show us how there can be only
metaphysics-first or ‘no-priority’ validations of the Primary Thesis. They
would in that sense take us beyond that thesis itself.
[Primary Thesis]
The metaphysics of a domain is involved in the philosophical explanation of the
nature of the meanings of sentences about that domain; andBook Review of 3a
the metaphysics
domain is involved in the philosophical explanation of the nature of intentional
contents (ways of representing) concerning that domain. (p. 4)
This exploits a notion of involvement that is in principle compatible with a
derivative status
Mind, Vol. 00 for whatever’s involved in the explanation
. 0 . 2021 of what,
 Mind so that
Association F
2021
might thus be involved in an account of G while being subordinate explana-
torily. (Perhaps it is rather the role of G in an account of F that is explana-
torily significant.)
The argument from understanding and the argument from rationality
must show us that in this case that abstract possibility cannot ultimately
be countenanced. To do that, they should not be susceptible to an interpret-
ation on which their plausibility is owed precisely to corresponding flexibility
in the dependence deployed in both arguments. The second broad element of
my hesitation about the primacy of metaphysics is my worry that the relevant
debt cannot be discharged.

2. The Primary Thesis

Let’s begin by attending to the crucial thesis that structures much of the
book. Very soon after offering it, after a useful chart in which metaphysics-
first and no-priority cases alone are the sub-nodes of the Primary Thesis
(above), Peacocke claims explicitly that it ‘stands opposed to what we can
call meaning-first views’ (p. 7 ). If this is to preview the conclusion of the
considerations to follow, then there is no immediate problem—ultimately, if
the arguments succeed, then in light of those additional substantive philo-
sophical considerations there can be no meaning-first validation of the
Primary Thesis. And Peacocke does himself admit before proceeding to the
Mind, Vol. 131 . 524 . October 2022
arguments that ‘[w]e have here then, in summary, three © Mind Association 2022
large-scale positions
in philosophy. . . the metaphysics-first view, the no-priority view, and the
meaning first view’ (p. 8).
But if he sees the Primary Thesis as already itself establishing relations of
explanatory priority, that should give us pause. Peacocke soon says, for ex-
Let’s begin by attending to the crucial thesis that structures much of the
book. Very soon after offering it, after a useful chart in which metaphysics-
first and no-priority cases alone are the sub-nodes of the Primary Thesis
(above), Peacocke claims explicitly that it ‘stands opposed to what we can
call meaning-first views’ (p. 7 ). If this is to preview Book Reviews of1367
the conclusion the
considerations to follow, then there is no immediate problem—ultimately, if
the arguments succeed, then in light of those additional substantive philo-
sophical considerations there can be no meaning-first validation of the
Primary Thesis. And Peacocke does himself admit before proceeding to the
arguments that ‘[w]e have here then, in summary, three large-scale positions

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

in philosophy. . . the metaphysics-first view, the no-priority view, and the
meaning first view’ (p. 8).
But if he sees the Primary Thesis as already itself establishing relations of
explanatory priority, that should give us pause. Peacocke soon says, for ex-
ample, that although ‘for the most part Dummett does not write in terms of
explanatory priority, he does endorse a series of claims that imply the falsity
of the Primary Thesis’ (p. 9). Now, even if Dummett endorsed, in whatever
terms, a series of claims about explanatory priority, it’s hard to see how that
would engage the Primary Thesis directly. Again, that thesis appears to be
compatible with any relation of explanatory priority, so long as it yields—as
perhaps it must—an involvement of the metaphysics of the domain in the
account of the meanings of corresponding sentences.
To be clear, I’m not sure that Dummett’s claims are about explanatory
priority. Ironically, they can appear to be simply affirmations of the Primary
Thesis—according to which conditions on the metaphysics of a domain are
(exhaustively)
4 implicit in any theory of meaning. But my point here is that
Book Review
even if they were about priority, that would not entail Dummett’s opposition
to the Primary Thesis, only perhaps to the Primacy of Metaphysics. Again,
the Primary Thesis makes a claim about involvement; the Primacy of
Metaphysics demands priority.
Mind, Vol. 00 . 0 . 2021  Mind Association 2021

3. Explanatory priority
Peacocke views the relation of explanatory priority as creating a ‘fundamental
division between a stronger and a weaker way in which the Primary Thesis
can hold for a given domain’ (p. 7). But he will argue that the thesis is
opposed to ‘meaning-first views, which hold that theories of meaning and
intentional content. . .are always explanatorily prior to the metaphysics of the
domain’ (p. 7 ). Given that the Primary Thesis is stated in terms of ‘involve-
ment in the philosophical explanation of’, which is not asymmetric, the
claimed opposition requires development, which Peacocke does indeed ex-
plicitly seek to provide. As noted above, his agenda incorporates two general
arguments against meaning-first views: I will address them both below. But
the framework for those arguments seems already to exclude the relativity of
the relevant variety of explanatory priority. Otherwise, it might be that while
there’s a respect of asymmetric explanatory priority for metaphysics, there is
another respect relative to which content takes precedence in giving an ac-
count of the metaphysics of a domain.
I have in mind, at least initially, just the difference between (a) explanation
with respect to metaphysics—which exploits the metaphysics of a domain in
order to show how the elements of that domain are apt for what might be
understood as representation in accord with how we seem to represent
Mind, Vol. 131 (b)
them—and . 524explanation
. October 2022 © Mind Association
with respect to representation—which uses 2022
and
relies on our ways of representing the elements of a domain in order to make
sense of its metaphysics.
Sometimes explanation is a kind of grounding of one thing in another,
even if our understanding of the ground itself relies (in another way) on the
 arguments against meaning-first views: I will address them both below. But
the framework for those arguments seems already to exclude the relativity of
the relevant variety of explanatory priority. Otherwise, it might be that while
there’s a respect of asymmetric explanatory priority for metaphysics, there is
another respect relative to which content takes precedence in giving an ac-
1368
count of Book Reviews
the metaphysics of a domain.
I have in mind, at least initially, just the difference between (a) explanation
with respect to metaphysics—which exploits the metaphysics of a domain in
order to show how the elements of that domain are apt for what might be
understood as representation in accord with how we seem to represent
them—and (b) explanation with respect to representation—which uses and

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

relies on our ways of representing the elements of a domain in order to make
sense of its metaphysics.
Sometimes explanation is a kind of grounding of one thing in another,
even if our understanding of the ground itself relies (in another way) on the
thing grounded. But explanation can also involve making sense of something
in terms of contents we presuppose as referring to that thing: a domain our
representations of which cannot be brought together into a coherent account is a
domain we cannot understand.
So although there may be one kind of explanatory dependence of our
account of the metaphysics of a domain on the contents and meanings
that refer to items in that domain, there may also be—compatibly and con-
temporaneously—a different relation, also explanatory, that gives priority to
the metaphysics, showing how the account of the contents and meanings—
the very things we first relied on, as one kind of basis, in explaining that
metaphysics—makes sense only in terms of that same metaphysics.
Peacocke might see his Chapter 1, §6 (‘Discovery, Explanation, and a
Theory of Conditions’) as responsive to the present concern. Key there is
Book Review 5
Reichenbach’s distinction between the context of discovery and the context
of justification, which Peacocke seeks to redeploy as a distinction between the
order of discovery and the order of philosophical explanation. Peacocke
applies that
Mind, Vol. 00 .distinction
0 . 2021 to several representative examples, inAssociation
 Mind each case2021
in
defence of ‘the metaphysics-involving position stated in the Primary Thesis’
(p. 26). Again, we need not reject involvement: but I doubt the sufficiency of
the Primary Thesis for the Primacy of Metaphysics.
In some cases and contexts, discovery is itself simply the enactment of a
particular sort of understanding and the concomitant achievement of one
sort of explanation. The order of discovery should not be contrasted unre-
strictedly with the order of philosophical explanation. Sometimes discovery
proceeds in the order it does because there is one variety of explanation
available (and not another), an explanation in which our ways of represent-
ing a domain play an essential role. The order of discovery and the order
explanation do not seem to me to contrast sufficiently deeply to support the
distinction Peacocke favours.
Peacocke uses Lewis’s Modal Realism to illustrate and support his distinc-
tion and his claim of primacy. If we resist Modal Realism on the basis of
doubts about how we could understand any associated modal discourse, that
shows a priority of the theory of understanding, according to Peacocke, only
in the order of discovery: ‘It does not by itself show anything at all about the
relative standing of the metaphysics and the theory of understanding in
respect of explanatory priority’ (pp. 27-28). There is certainly a variety of
explanatory priority with respect to which the basis of our resistance to
Modal Realism shows little (about the relative standing of metaphysics and
Mind, Vol.or
content 131meaning).
. 524 . October 2022 is a way of viewing the
But there © Mind Association
relation 2022
between
explanation and understanding that would rather limit the significance of
the distinction: if, in particular, any lack of understanding suffices for some
absence of explanation, and any absence of explanation suffices for some lack
of understanding, then our basing resistance to Modal Realism in a lack of
 Peacocke uses Lewis’s Modal Realism to illustrate and support his distinc-
tion and his claim of primacy. If we resist Modal Realism on the basis of
doubts about how we could understand any associated modal discourse, that
shows a priority of the theory of understanding, according to Peacocke, only
in the order of discovery: ‘It does not by itself show anything at all about1369
Book Reviews the
relative standing of the metaphysics and the theory of understanding in
respect of explanatory priority’ (pp. 27-28). There is certainly a variety of
explanatory priority with respect to which the basis of our resistance to
Modal Realism shows little (about the relative standing of metaphysics and
content or meaning). But there is a way of viewing the relation between

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

explanation and understanding that would rather limit the significance of
the distinction: if, in particular, any lack of understanding suffices for some
absence of explanation, and any absence of explanation suffices for some lack
of understanding, then our basing resistance to Modal Realism in a lack of
understanding is an implicit prioritization of one kind of explanation. Our
metaphysical theory is revealed as answerable to how we understand the
domain.
Peacocke urges that once ‘we have [a] positive non-Lewisian metaphysics,
then what is involved in understanding’ (p. 28) can be described in terms of a
thinker’s relations to what that account delivers. But if failure to provide an
adequate description in those terms is treated as a prior condition of ad-
equacy on any such metaphysical account, there seems still to be a residual
priority for an account of our understanding relative to the metaphysics.
Peacocke attributes to Dummett a view according to which, as we might
put it, the theory of meaning associated with a given domain need not derive
from prior metaphysical presuppositions and yet suffices to settle and resolve
metaphysical
6 questions and controversies ‘without residue’. Once the theory
Book Review
of meaning is constructed, a picture of the metaphysics is forced upon us.
Moreover, Dummett thinks there is no content to any ‘non-metaphorical’
metaphysical thesis that does not ‘amount to’ a thesis in the theory of mean-
ing. That
Mind, Vol. 00is .how high the seas of language rise.
0 . 2021  Mind Association 2021
But it is not compulsory to interpret Dummett, as Peacocke does, as
holding that ‘the metaphysics of a domain is legitimate insofar, and only
insofar, as it can be construed as making claims founded in the correct mean-
ing theory for a language about that domain’ (p. 10, emphasis added). While
it’s clear enough that he would oppose a reverse relation of ‘founding’,
Dummett’s main point seems to be that any metaphysical thesis that cannot
be correlated with a thesis in the theory of meaning is, well, meaningless.
While Peacocke’s conclusion is that ‘[m]etaphysics, on Dummett’s view,
simply could not have the status it would need to have for the Primary
Thesis to be correct’ (p. 10), Dummett can instead be viewed as holding to
a kind of no-priority view, on which language and metaphysics are rather
explanatorily entwined—inextricably and without remainder. Metaphysics
could not have the status it would need to have for Primacy; but the
Primary Thesis is not blocked.
There are accordingly several related approaches on which there might be
no priority: and one of them leaves room for a kind of ‘meaning-first’ pos-
ition compatible with the Primary Thesis (while only on another approach is
that thesis blocked). Entanglement no-priority positions are ones in which,
both, meanings are involved in metaphysical accounts and metaphysics is
involved in accounts of meanings—where those involvements are either (i)
unrelativized,
Mind, (ii) .relative
Vol. 131 . 524 Octoberto2022
the same parameter, or (iii)© relative to different
Mind Association 2022
parameters, potentially in ways that don’t yield an ultimate, more abstract,
priority of either relative to the other. It is the possibility of relativization to
different parameters that allows for an asymmetric priority for meanings in
an account of the metaphysics of a domain (if in a no-ultimate-priority-
a kind of no-priority view, on which language and metaphysics are rather
explanatorily entwined—inextricably and without remainder. Metaphysics
could not have the status it would need to have for Primacy; but the
Primary Thesis is not blocked.
1370
There Book Reviews several related approaches on which there might be
are accordingly
no priority: and one of them leaves room for a kind of ‘meaning-first’ pos-
ition compatible with the Primary Thesis (while only on another approach is
that thesis blocked). Entanglement no-priority positions are ones in which,
both, meanings are involved in metaphysical accounts and metaphysics is
involved in accounts of meanings—where those involvements are either (i)

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

unrelativized, (ii) relative to the same parameter, or (iii) relative to different
parameters, potentially in ways that don’t yield an ultimate, more abstract,
priority of either relative to the other. It is the possibility of relativization to
different parameters that allows for an asymmetric priority for meanings in
an account of the metaphysics of a domain (if in a no-ultimate-priority-
because-equipoised-asymmetric-prioritizations sort of way), even while pre-
serving the Primary Thesis.
There are also Independence no-priority cases, in which neither meaning
nor metaphysics is even involved in the philosophical explanation of the
other—not even relative to different parameters. Neither is ultimately, or
even in a relativized way, prior in the order of explanation, but unlike in the
kind of case Peacocke calls ‘no priority’, the Primary Thesis does not hold.

4. Arguments from understanding and rationality

Peacocke thinks support for his line follows, given the right background
framework, from the following key premise:
[Relations]
Which relations a thinker can stand in to an entity depends Book
on theReview 7
correct meta-
physics of that entity. (pp. 11-12)

Plausibly, relations into which an entity can enter depend on its nature:
Mind, Vol. 00 . 0 . 2021  Mind Association 2021
Peacocke infers that the metaphysics of a domain ‘constrains’ the theory of
concepts of entities in that domain and constrains the theory of meaning for
language about that domain. Given a suitably flexible understanding of ‘con-
strains’, that seems right too. If the relevant notion of constrains is just built
up from the same notion of dependence that figures in [Relations ], a notion
that’s correlated with the notion of involvement that made the Primary Thesis
plausible in the first place, then the conclusion may follow. But if we do not
go beyond that notion of constraint, then it is harder to see how even
meaning-first views need be excluded.
If one cannot account for something without mentioning something else,
then the latter is in that way involved in that account. But that does not yet
give the latter a philosophically significant asymmetric explanatory priority.
It may also be the case that you cannot account for the latter without
mentioning the former.
Contrast [Relations] with the following, more committal and less plaus-
ible, claim:
[Relations*]
Whenever something stands in a relation to o, that is so asymmetrically in virtue of
what o 131
Mind, Vol. is (independently)
. 524 . Octobermetaphysically
2022 like as a prior matter.
© Mind Association 2022
And now consider a foil: the metaphysics of a domain is ‘involved in’ the
philosophical explanation of the nature of the properties instantiated by the
elements (objects, properties, relations) of that domain. That would be a
plausible metaphysics-involving thesis. In giving a theory of the properties
then the latter is in that way involved in that account. But that does not yet
give the latter a philosophically significant asymmetric explanatory priority.
It may also be the case that you cannot account for the latter without
mentioning the former.
Contrast [Relations] with the following, more committal and less plaus-
ible, claim: Book Reviews 1371

[Relations*]
Whenever something stands in a relation to o, that is so asymmetrically in virtue of
what o is (independently) metaphysically like as a prior matter.

And now consider a foil: the metaphysics of a domain is ‘involved in’ the

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

philosophical explanation of the nature of the properties instantiated by the
elements (objects, properties, relations) of that domain. That would be a
plausible metaphysics-involving thesis. In giving a theory of the properties
of a domain, we are at the same time giving a theory of the nature of its
elements—of how they are of various kinds. And [Relations] holds: which
properties an entity can instantiate depends on the correct metaphysics of
that entity.
But [Relations*] is less plausible: when a property is instantiated by an
entity, is that ‘in virtue of’ what the entity is (independently) metaphysically
like, as a prior matter? How could which properties an entity instantiates be
in general determined, with priority, by what it’s like metaphysically? The
object’s metaphysical nature appears to be explained precisely in terms of the
properties it instantiates. At the same time, the opposite order of priority
seems no more immediately attractive.
This possibility challenges Peacocke’s argument from understanding. I
think even meaning-first positions can sustain the thesis that the intentional
relations in which a thinker can stand to an entity depend on the correct
metaphysics of that entity.
If P entails Q, then, (even) if it is not the case that Q entails P, there are
already
8 two contraposed
Book Review dependence relations, with something like reversed
polarity, in place: P has a kind of asymmetric priority with respect to Q
because it is logically stronger. It suffices for Q’s truth; whereas Q does not
suffice for P’s truth. Given P, Q’s situation is settled; but Q’s truth doesn’t
have the same effect on P. So P is in a way prior. We
Mind, Vol. 00 . 0 . 2021
could say that P is
 Mind Association 2021
‘involved in an account of’ Q here to represent Q’s being made true by P—
emphasizing P’s sufficiency for Q.
At the same time, P in another sense asymmetrically depends on Q: P can
be true only if Q is true. P can’t manage to be true without Q’s truth
obtaining. And the reverse of this dependence relation is not in place for
Q: Q can be true without P’s being true so it doesn’t in the corresponding
way depend on P. That gives Q a different kind of priority: P depends on it in
a way that it does not depend on P. We would now say that Q is involved in
an account of P, emphasizing Q’s being a partial contributor to the obtaining
of the stronger condition P. Affirming a dependence relation between P and
Q is potentially ambiguous and compatible with very different logical rela-
tions between them.
Later in the chapter, Peacocke offers ‘a second general domain-
independent argument in support of’, as he puts it (p. 25), ‘the Primary
Thesis’. To repeat a point from above: what’s wanted is not support for a
Primary Thesis about involvement so much as an argument that would take
us from the Primary Thesis to the exclusion of meaning-first witnesses of it
(that is, to the Primacy of Metaphysics).
Mind, Vol. 131 . 524
Peacocke’s . October
second 2022is about rational sensitivity.
argument © MindItAssociation
observes 2022
that
anyone that grasps the constitutive concepts ‘can make a rational assessment,
in various circumstances, whether to judge that Fa’ (p. 25) . But, Peacocke
continues,
[Rationality]
an account of P, emphasizing Q’s being a partial contributor to the obtaining
of the stronger condition P. Affirming a dependence relation between P and
Q is potentially ambiguous and compatible with very different logical rela-
tions between them.
Later in the chapter, Peacocke offers ‘a second general domain-
1372 Bookargument
independent Reviews in support of’, as he puts it (p. 25), ‘the Primary
Thesis’. To repeat a point from above: what’s wanted is not support for a
Primary Thesis about involvement so much as an argument that would take
us from the Primary Thesis to the exclusion of meaning-first witnesses of it
(that is, to the Primacy of Metaphysics).
Peacocke’s second argument is about rational sensitivity. It observes that

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

anyone that grasps the constitutive concepts ‘can make a rational assessment,
in various circumstances, whether to judge that Fa’ (p. 25) . But, Peacocke
continues,
[Rationality]
Whether it is rational to judge Fa in given circumstances depends on what it takes
for it to be true that Fa (p. 25).

Since ‘what it takes for Fa to be true depends on the metaphysics of [a and

F,] there must already be a sensitivity to the metaphysics of the object and the
property built into the condition for grasping the singular concept a and the
predicative concept F’ (p. 25). Peacocke is helpfully forthcoming about some
limitations, one dialectical and another concerning its scope, of this argu-
ment. (I accept and will not pause over those here.) But he does think the
argument is sound.
Focus, however, on the claimed dependence of the rationality of a judg-
ment on what it takes for that judgment to be true. There’s a good sense in
which the rationality of a judgment is independent of what it takes for it to
be true, at least if what it takes to be true is understood as going beyond
relations among the concepts that constitute the content of the judgment
itself—as going beyond the level of sense, to the level of reference.
A good initial example exploits demonstratives and other Book indexical
Review con- 9
cepts and meanings. Castor and Pollux might each think he is in Greece. But
the rationality of the judgment made by either twin does not depend on
which of them is having the (one and the same) thought. In at least one
important sense,
Mind, Vol. 00 the rationality of the judgment does not
. 0 . 2021 depend
 Mind on what
Association it
2021
takes for the judgment to be true: what it takes for Castor’s thought to be
true is different from what it takes for Pollux’s. But the rationality of either’s
thought is determined by the same phenomena.
You see a barn and think, that’s a fine home for livestock. I look at an
indistinguishable fake barn, a papier-maché replica, and think the same
thought: that’s a fine home for livestock. On at least one way of making sense
of the data in this sort of example (there are, of course, other ways), what it
takes for those thoughts to be true varies. Indeed yours is true and mine is
not. Our thoughts involve the same content; but they refer to different
things.
Peacocke’s second claim, that ‘what it takes for Fa to be true depends on
the metaphysics of a and F’, holds—what it takes for any thing to be a fine
home for livestock depends on whether it provides shelter and the like. But
while the content of the judgment will, in different circumstances, determine
different things as the referent of ‘that’, the variations at that level—the level
of reference—do not affect the rationality of the judgment.
Demonstratives and pure indexicals are not isolable special cases: ration-
Mind, Vol.
ality 131general
is in . 524 . October 2022 independently of the ©level
determined Mind Association 2022
of reference.
Consider a descriptive thought: the inventor of the zip was ingenious. In
some contexts, what it takes for that thought to be true is for Julius to be
ingenious. In other contexts, someone else invents the zip, and it’s their
ingenuity that matters. The rationality of the thought will be independent
of the data in this sort of example (there are, of course, other ways), what it
takes for those thoughts to be true varies. Indeed yours is true and mine is
not. Our thoughts involve the same content; but they refer to different
things.
Peacocke’s second claim, that ‘what it takes for Fa to be true depends on
the metaphysics of a and F’, holds—what it takes for Book Reviews
any thing to be a1373
fine
home for livestock depends on whether it provides shelter and the like. But
while the content of the judgment will, in different circumstances, determine
different things as the referent of ‘that’, the variations at that level—the level
of reference—do not affect the rationality of the judgment.
Demonstratives and pure indexicals are not isolable special cases: ration-

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

ality is in general determined independently of the level of reference.
Consider a descriptive thought: the inventor of the zip was ingenious. In
some contexts, what it takes for that thought to be true is for Julius to be
ingenious. In other contexts, someone else invents the zip, and it’s their
ingenuity that matters. The rationality of the thought will be independent
of those variations.
More generally, if thoughts are constituted by properties (serving as con-
cepts) and the level of reference is determined through the instantiation of
those constitutive properties, and the pattern of property instantiations can
vary, then there may be variation in what it takes for a thought to be true,
even while there is no variation in what it takes for the judgment to be
rational.
Admittedly, in every case, trivially, the rationality of a judgment does in a
way ‘depend on what it takes for the thought to be true’: in the case above of
the inventor of the zip, for example, it depends, in each context, on the
inventor of the zip—whoever they are in that context—being ingenious. In
the demonstrative case, the rationality of the thought depends on, to put it
generally, that—whatever that is—being a fine home for livestock. And in the
indexical case, for both Castor and Pollux, it depends on what it takes for
him (respectively!) to be in Greece.
But while this way of interpreting the claim of dependence enables
[Rationality] , it limits the force of the argument. We have to abstract
from the
10 veryReview
Book level of reference that is meant to be implicated. We should
not say that ‘what is referred to’ in any demonstrative thought is always the
same thing: that. When we allow that, in a sense, for each of Castor and
Pollux what it takes for their thought I am in Greece to be true is the same, we
have
Mind, abstracted
Vol. 00 . 0 . from
2021 the level of reference: those thoughts refer
 Mind to different
Association 2021
twins. No more is the same time always referred to in any thought containing
an element for, as it were, now.
In the same way, the what-it-takes-for-it-to-be-true of a thought consti-
tuted by the inventor of the zip is not the same across contexts in which Julius,
or, say, his older sister Julie, invented the device. In one context what it takes
for that same thought to be true is for Julius to be ingenious; in the other it’s
his sister’s ingenuity that matters. This isn’t to fail to distinguish dependence
on truth condition from dependence on truth. The variation in truth condi-
tion is indeed not necessarily accompanied by variation in truth (value): Julie
would reliably have been as ingenious as Julius. It’s to distinguish the role of
truth condition as something instantiated at the level of reference from its
role as thinkable. If what it takes for Fa to be true depends on the metaphysics
of a and F, then whether it’s rational to judge Fa does not depend on what it
takes for it to be true.

Mind, Vol. 131 . 524 . October 2022 © Mind Association 2022

5. Last orders
I want then in closing to discuss a possible alternative, which may in fact be
simply an alternative take on Peacocke’s own project, on which that project’s
conclusion is better called the ‘Inescapability’ or even the ‘Insuperability’ of
or, say, his older sister Julie, invented the device. In one context what it takes
for that same thought to be true is for Julius to be ingenious; in the other it’s
his sister’s ingenuity that matters. This isn’t to fail to distinguish dependence
on truth condition from dependence on truth. The variation in truth condi-
tion is indeed not necessarily accompanied by variation in truth (value): Julie
1374 Book Reviews
would reliably have been as ingenious as Julius. It’s to distinguish the role of
truth condition as something instantiated at the level of reference from its
role as thinkable. If what it takes for Fa to be true depends on the metaphysics
of a and F, then whether it’s rational to judge Fa does not depend on what it
takes for it to be true.

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

5. Last orders
I want then in closing to discuss a possible alternative, which may in fact be
simply an alternative take on Peacocke’s own project, on which that project’s
conclusion is better called the ‘Inescapability’ or even the ‘Insuperability’ of
Metaphysics , with no implicit claim to asymmetric priority.
Some evidence that Peacocke’s intentions are accordingly modest is avail-
able in Chapter 6: ‘Ontology and Intelligibility’ . There Peacocke confronts
cases in which there has been a radical misconception of reality. These cases
‘do not merely involve specifications that could not be fulfilled. Rather. . .the
relevant misconceptions fail even to specify a genuine condition whose pos-
sibility or impossibility could be set up for assessment’ (p. 171). Focusing
especially on the case of absolute space, though viewing his analysis as gen-
eralizing, Peacocke offers what he calls the ‘excess dimension’ diagnosis:
An ontology is illegitimate if it [invokes entities] of such a kind that there is
actually no saying—there is no account possible—of what it would be to be
thinking of one rather than another of these entities. . .. (p. 183)

The diagnosis here establishes a basic condition of adequacy on metaphysical

theories: that they make possible an account of what it would be to refer to the
items in the corresponding domain. So here intentionality has a certain pride
of philosophical place: there’s no pursuing a metaphysical account in its own
terms and just hoping for the best as far as any (accordingly subordinated)
theory of content and meaning might later go. According to the excess di-
mension diagnosis a ‘defect in [a] proposed metaphysics’Book can Review
be precisely 11its
failure to provide for ‘a good account of concepts of [the] alleged domain’
(p. 202).
Peacocke urges that ‘when we have an explanation from the nature of the
distinction
Mind, Vol. 00 between
. 0 . 2021 good and bad metaphysics for Mind various proposed
Association 2021
domains, of why there cannot be good fundamental reference-conditions
for certain proposed concepts, we have thereby an example of the philosoph-
ical explanatory priority of metaphysics over the theory of intentional con-
tent’ (p. 202). But when the distinction between good and bad metaphysics
used in the explanation is itself made out in terms of the possibility of coherent
reference conditions, as in the excess dimension diagnosis, the situation is less
clear. The goodness of our metaphysics now appears to be subordinate to the
availability of an account of the corresponding intentional relations. In the
end, this unclarity sustains some of my hesitation over Peacocke’s Primacy.
I hope to have both granted concessions and raised concerns about steps
Peacocke appears to take in the book. While we need offer no resistance to
the Primary Thesis itself, it is harder to see how to go beyond it, to any
asymmetric explanatory priority for metaphysics, any priority that is not
matched
Mind, Vol. by
131 another—perhaps relative to a different parameter—on
. 524 . October 2022 which
© Mind Association 2022
metaphysical accounts are rather posterior. What is by contrast not hard to
see, from the beginning of Peacocke’s book to its end, is that—to echo
something he says about Dummett—even should we hesitate over his pos-
ition, one of the many ways in which we are deeply indebted to Peacocke is
for his crystallization and treatment of the compelling issues in this book.*
for certain proposed concepts, we have thereby an example of the philosoph-
ical explanatory priority of metaphysics over the theory of intentional con-
tent’ (p. 202). But when the distinction between good and bad metaphysics
used in the explanation is itself made out in terms of the possibility of coherent
reference conditions, as in the excess dimension diagnosis, the situation is less
clear. The goodness of our metaphysics now appears toBook beBook Reviews
Review to1375
subordinate 11
the
availability of an account of the corresponding intentional relations. In the
end, this unclarity sustains some of my hesitation over Peacocke’s Primacy.
Peacocke
I hope to urges that ‘when
have both granted weconcessions
have an explanation
and raisedfrom the nature
concerns aboutofsteps
the
distinction between good and bad metaphysics for various
Peacocke appears to take in the book. While we need offer no resistance to proposed
domains,
the Primary of Thesis
why there
itself,cannot be good
it is harder to fundamental
see how to go reference-conditions
beyond it, to any
for certain proposed concepts, we have thereby an example of the that
philosoph-

Downloaded from https://academic.oup.com/mind/article/131/524/1364/6374529 by University of Cambridge user on 02 February 2023

asymmetric explanatory priority for metaphysics, any priority is not
ical explanatory priority of metaphysics over the theory
matched by another—perhaps relative to a different parameter—on which of intentional con-
tent’ (p. 202). But when the distinction between good and bad
metaphysical accounts are rather posterior. What is by contrast not hard to metaphysics
used from
see, in thethe
explanation
beginning is itself made out book
of Peacocke’s in terms
to of
itsthe possibility
end, of coherent
is that—to echo
something he says about Dummett—even should we hesitate over hisispos-
reference conditions, as in the excess dimension diagnosis, the situation less
ition, one of the many ways in which we are deeply indebted to Peacockethe
clear. The goodness of our metaphysics now appears to be subordinate to is
availability
for of an account
his crystallization and of the corresponding
treatment intentional
of the compelling issuesrelations. In the
in this book.*
end, this unclarity sustains some of my hesitation over Peacocke’s Primacy.
I hope to have both granted concessions and raised concerns about steps
Universityappears
Peacocke of Texastoattake in the book. While we need offer no resistance
Austin david sosa to
the Primary Thesis itself, it is harder to see how to go beyond it, to any
[email protected]
doi: 10.1093/mind/fzab028
asymmetric explanatory priority for metaphysics, any priority that is not
matched by another—perhaps relative to a different parameter—on which
metaphysical
Vagueness andaccounts
Thought,are byrather
Andrewposterior. What is Oxford
Bacon. Oxford: by contrast not hard
University to
Press,
see, from the beginning
2018. Pp. xviii + 340. of Peacocke’s book to its end, is that—to echo
something he says about Dummett—even should we hesitate over his pos-
ition, one of the many ways in which we are deeply indebted to Peacocke is
for his crystallization and treatment of the compelling issues in this book.*
1. Presentation
It’s difficult nowadays to write an interesting new book on vagueness, but
Andrew
University Bacon has succeeded.
of Texas david sosa
at Austin He hasn’t done so by putting forth revolutionary
[email protected]
views about the Sorites paradox or the nature of borderline cases, but rather by
doi: 10.1093/mind/fzab028
asking deep questions about the role of vagueness in thought and action and
providing inspiring answers on the background of a carefully crafted theory of
propositions. Bravo!
In essence, Bacon’s main theses are to the effect that, while vague proposi-
tions often play a crucial role in the acquisition of knowledge, the role they play
in *thought
Thanks toand action isPeacocke
Christopher by andforlarge derivative
valuable comments, on with
the which
role played by precise
only restrictions of
propositions.
space precludedFor lack of
my fuller space, inThanks
engagement. this review, I’ll mostly
also to Sebastian focus
Rödl, onparticipants
and to these theses
in a
2019 conference on Peacocke’s book at the University of Leipzig, for the philosophical context
and set aside the few parts of the book dealing with more traditional problems
in which this review was born.
in the vagueness debate such as the logical problems represented by the Sorites
paradox and by borderline cases, concerning which Bacon’s views are broadly
epistemicistic
Mind, Vol. 00 . (I
0 .discuss
2021 those views in depth in Zardini 2022b,  Mindwhich thus com-
Association 2021
plements this review). Substantial parts of the book which I’ll also leave out
from this review are devoted to criticisms of some prominent approaches and
assumptions in the vagueness debate. While it can be expected that all those

*
Thanks to Christopher Peacocke for valuable comments, with which only restrictions of
space precluded my fuller engagement. Thanks also to Sebastian Rödl, and to participants in a
2019 conference on Peacocke’s book at the University of Leipzig, for the philosophical context
in which this review was born.

Mind, Vol.
Mind, Vol.131
00 . . 0524
. 2021
. October 2022 © Mind
Mind Association
Association 2021
2022