Tagmemics

WITHDRAWN
University of
illinois LiDrary
at Urbana-Cha
mpaign
UNIVERSITY OF
ILLINOIS LIBRARY
CENTRAL CIRCULATION BOOKSTACKS
The person charging this material is re-
sponsible for its return to the library from
which it was borrowed on or before the
Latest Date stamped below.
Theft, mutilation, and underlining of books are reasons
for disciplinary action and may result in dismissal from
the University.
, 333-8400
TO RENEW CALL TELEPHONE CENTER
UNIVERSITY OF ILLINO IS LIBRAR Y AT URBANA-CHAMPAIGN
gec 17 92
JAN 0 3 1994
JUN 41 2004
FEB 24 HUY
SEP *6 2004
write new due date below

When renewing by phone, L162
previous due date .
Digitized by the Internet Archive
_In 2022 with funding from
Kahle/Austin Foundation
https://archive.org/details/linguisticconcep0000pike
Linguistic Concepts
KENNETH L. PIKE
Linguistic
Concepts
An Introduction to
Tagmemics
University of Nebraska Press Lincoln and London

i)
Portions of Part I appeared originally in
Reading about Language, by Charlton Laird
and Robert M. Gorrell,
copyright 1971 by Harcourt Brace Jovanovich, Inc.
Copyright 1982 by the

University of Nebraska Press
Manufactured in the United States of America
The paper in this book meets the guidelines for

permanence and durability of the
Committee on Production Guidelines for
Book Longevity of the Council on
Library Resources.
Library of Congress Cataloging in Publication Data
Pike, Kenneth Lee, 1912-

Linguistic concepts.
Bibliography: p.
Includes index.
1. Tagmemics. I. Title
P160.P48 415 81-19814
ISBN 0-8032-3664-6 AACR2
Contents
List of Figures ix
Preface xi
PartI The Observer and Things 1
Introduction 3
Chapter 1: Theory 5
1.1. Value of a Theory 5
1.2. Weaknesses of a Theory 6
1.3. Selection of a Theory 7
Chapter 2: Survey of Tagmemic Theory 10
2.1. Beliefs Underlying Statements 10
2.2. Complementarity of Perspective 11
2.3. Units as Known 13
2.4. Hierarchy of Interlocking Hierarchies 13
2.5. Context as Relevant 15
Chapter 3: Particles 19
. 3.1. Things and Nonthings as “Things” 19
3.2. Particles in Language 21
3.3. Particles in Linear and Spatial Order 23
Chapter 4: Wave 24
4.1. The Merging of Sounds in Sequence 24
4.2, Nucleus and Margin of Wave 26
4.3. Types of Wave 27
Chapter 5: Field 30
_5.1. Unit Determined by Relationship to Context 30
5.2. Sounds in Dimensional Patterns 31
5.3. Clauses in Dimensional Patterns 33
5.4. A Poem as a Field Structure 36
5.5. Verbal Cues 38
vi Contents
Part II The Unit 39
Introduction 41
Chapter 6: Contrast and Identification 42
6.1. Recognition via Contrast and Similarity, within Field 42
6.2. Matching for Contrast 43
6.3. Contrast in Contextual Frames 45
6.4. Contrast Seen via Matrices 47
6.5. Contrast in Verbal Meaning and Behavior 49
Chapter 7: Variation 52
7.1. Referential Identity in Spite of Change 52
7.2. Changes in Pronunciation 54
7.3. Changes in Grammar 55
7.4. Indeterminate Segmentation 57
Chapter 8: Distribution 60
8.1. Distribution of Units as Relevant to Their Identity 60
8.2. As a Member of a Substitution Class 62
8.3. As Part of a Structural Sequence 62
8.4. As a Point in a System 65 .
Part III Hierarchy 67
Introduction 69
Chapter 9: Grammatical Hierarchy 70
9.1. A Psychological Constraint on Complexity 70
9.2. Behavioral and Hence Semantic Relevance of Levels
of Grammatical Structure 72
9.3. Pairs of Etic or Emic Hierarchical Levels
Sharing Behavioral Impact 73
9.4. The Unit-in-Context (the Tagmeme) 75
9.5. Diagrams and Formulas for Unit-in-Context 79
Chapter 10: The Phonological Hierarchy 84
10.1. Noncoterminous Borders of Phonological and
Grammatical Hierarchies 84
10.2. Relevance Criteria for the Pairing of Levels 85
10.3. Constraints of Focal Attention 86
10.4. Source of the Terms Emic and Etic 87
10.5. Simultaneous Features—and Where Do You Stop? 87
10.6. Wave Characteristics of Phonological Units 89
10.7. Field Characteristics of Phonology Seen via
Intonation 91
Contents vii
Chapter 11: The Referential Hierarchy 97

11.1. Levels in the Referential Hierarchy 97
11.2. The Referential Tagmeme 100
11.3. Some Grammatical and Phonological Alternatives in
the Representation of Referential Talking Space 103
11.4. Reality Relative to a Frame of Reference, versus
Grammatical Form 105
Part IV Context 107
Introduction 109
Chapter 12: Form-and-Meaning 111
12.1. Meaning as Essentially Tied to a Physical
Component in a Unit 111
12.2. Hierarchically Extended Lexicon as the Substance
Manifesting, Simultaneously, the Forms of the
Three Hierarchies 112
12.3. Humor via Focus on Intersecting Form and Meaning in
Hierarchies 115
12.4. Form and Meaning Interlocking in Poetry 115
Chapter 13: Sharing as Prerequisite to Change 118
13.1. Change Involving the Sharing of Particles 119
13.2. Change Involving the Sharing of Wave Components 119
13.3. Change Involving the Sharing of Field Components 122
Chapter 14: Universe of Discourse 126
14.1. Lack of Coherence in Speaker-Hearer Interaction 127
14.2. Failure of Coherence of Speaker with Reality as
Seen by Others 129
14.3. Behavioral Universals in Language Learning 131
14.4. Why Translation Is Possible 131
14.5. Poetry Can Concentrate the Signaling of a
Universe of Discourse 134
14.6. Conclusion: Unit-in-Context within Context 135
References 137
Index 141
Figures
ikeale Theory like a window 6

eee Choice of theory 8
2.1. Interlocking tagmemic concepts 11
2.2. Merging of perspectives 12
2.3. Units change 14
2.4. Meaning of a word 16
2.5. Sharing and change 17
3.1 . Things, processes, and relations 20
Be: Static view of sound 21
eos A limerick 22
4.1. Smearing of sounds 25
4.2 . Merging of words 25
4.3. Extreme smearing of sounds 26
4,4, A play as a standing wave 28
Dake Units and context 31
5.2. A phonological field 32
5.3. The highest point of the tongue 34
5.4. The root of the tongue 35
5.5. Dimensional matrix of Spanish clauses 36
5.6. Large units and field structure 37
ae 6A; Contrastive components 43
6.2. Frame technique 45
6.3. Tones and variants 46
6.4. Tone contrast in frames 48
6.5. Unit-times-component matrix 49
6.6. Matrices and meanings of lexicon 50
6.7. Words and contrastive senses 50
ygale Identity and variability 53
7.2. Conditioned variants of field structure 56
7.3. Segmentation 58
Outs Context and distribution 61
x Figures
8.2. Constraints on occurrence of vowels and consonants 63

8.3. Contrastive distribution of sentences 64
9.1. Recognition of chunks 71
9.2. The unit-in-context 75
9.3. A tree diagram 78
9.4. A generalized four-cell tagmemic notation 82
LOS Noncoterminous units of grammar and phonology 85
10.2. Field structure of English intonation 93
10.3. Poem as read by James Squires 94
10.4. Poem as read by Austin Warren 95
10.5. Voice quality and meaning of poem 96
10.6. Phonological unit-in-context 96
11.1. Referential event tagmeme 101
11.2. Referential person tagmeme 102
AGES} Truth of statement and frame of reference 104
11.4. Frame of reference 106
12.1. Hierarchies of reference, grammar, and phonology 113
12.2. Hierarchies and jokes 114
12.3. Form and meaning in a poem 116
13.1. The pun 119
13.2. Change in meaning-viewed-as-wave 120
13.3. Sharing of a referential component 121
13.4. The taxonomic tree 123
13.5. Sharing of a taxonomic feature 124
14.1. Failure of reply to meet expectations 128
14.2. Slippage of elicited universe of discourse 128
14.3. Mixture of universes of discourse 129
14.4. Inconsistency with reality 129
14.5. Political coherence 130
14.6. Context and definitions of words 133
14.7. Poetry and signals 134
Preface
In this volume person (and relation between persons) is given

theoretical priority above formalism, above pure mathematics, above ide-
alized abstractions.
A person, as observer, has choice. This choice extends to adoption of a
theory or a temporary or permanent combination of theories (see Chap-
ter 1). A theory may be adopted because it is exciting to the observer, is
useful to reach some material or cultural goal, or helpful to understand
oneself, others, or one’s environment.
In the theory surveyed here (see Chapter 2, for summary), discussion be-
gins (Part I) with observer variableness. One can choose to see the world as
made up largely of discrete bits, particles (Chapter 3). This has the advan-
tage of obviousness, since well-known “things” serve as the model for ex-
tension to less easily observed units, and the relation of one unit joined to
another to make a larger thing is also close to experience. The danger is
that the analyst will get carried away with its usefulness and assume that
the world is nothing but discreteness—a partitioning—of nonoverlapping
elements. But fusion, merging, gradients, change, growth, education, and
indeterminacies are also a part of experience and are difficult to capture if
only a particle theory is available.
In order to meet this need, the theory adds the perspective of wave
(Chapter 4). The merging of units and the overlapping of borders are to be
expected, and not to be treated as a theoretical disturbance. But there is no
suggestion here that some items are particles, while others are waves;
rather, any particle can be viewed (for certain temporary purposes) as a
wave, and any wave can be viewed (for other temporary purposes) as a
particle.
Similarly, all events or situations or items or components or persons can
for certain purposes of the observer be viewed for a time as points in a set
of relationships (that is, as making up a field, Chapter 5). The relationships
between these points or things or events or persons can be focused on as
comprising one view of reality. But here also the emphasis must be re-
xii Preface
peated that when life and the universe are viewed as nothing but relation-
ships, without adequate provision for viewing the world of man as made
up of particles and waves, insurmountable difficulties will eventually be
encountered. And again, the solution is not to treat some elements as re-
lations and others as units, but to be able to use all three approaches as
alternative ways to view every situation as static, dynamic, or relational ac-
cording to the current needs and interests of the observer.
In Part II, emphasis shifts to the units themselves. Yet even here the hope-
less attempt to eliminate the observer in favor of scientific detachment or of
objectivity is dismal. Units must be experienced or observed or deduced or
imagined. In language, combinations of features are units, to human view,
only when there is purpose, meaning, relevance, or significance of some
kind to give coherence to a set of sensations or memories or involvements or
imaginations. Unless one item is perceived as different from another (tac-
itly, by no means necessarily explicitly), it cannot be treated as a separate
unit (Chapter 6); and only when such differences are present are the fea-
tures involved serviceable as identificational features for units in the ab-
sence of such contrasting units.
Similarly, every repeatable or identifiable unit of human experience has
a range of variation within it, or in the human’s perception or experience
or imagination of it (Chapter 7). The stance varies, and the sensations with
it, even when the flowing river remains the same or the table is still a table.
The persistence of units seen via their identificational-contrastive features,
along with variableness in these units, forces us to a theoretical position for
all of rational behavior (not just linguistics). Such a persistent, perceptual
unit is termed an emic one (drawn from the linguistic term phonemic)—an
entity seen as “same” from the perspective of the internal logic of the con-
taining system, as if it were unchanging even when the outside analyst
easily perceives that change. (Meanwhile, the term etic, from the linguistic
term phonetic, labels the point of view of the outsider as he tries to pene-
trate a system alien to him; and it also labels some component of an emic
unit, or some variant of it, or some preliminary guess at the presence of
internal emic units, as seen either by the alien observer or as seen by the
internal observer when somehow he becomes explicitly aware of such vari-
ants through teaching or techniques provided by outsiders. )
In addition, a physical component is posited for every emic unit of be-
havior. This adds a further constraint to all possible views of the universe
as seen through this theoretical perspective as a whole. No item of human
behavior can be completely abstracted from all physical settings, or from
all physical components; no purely abstractional system can serve in it,
whether of ideas or of postulated systemic relational elements. Each must
Preface xiii
have its physical component. That is, the theory is not completely mecha-
nistic, since it has purpose, meaning, significance, and human relevance
tied to it; and it is not completely abstract, since it has a physical compo-
nent for every emic unit. (For thought itself, the physical component would
be composed of some kind of physical activity in the brain.)
Yet along with identificational-contrastive features and variants with
physical components, there is a third requirement for the existence of emic
units. They must be appropriate to some place in a system; they must be
appropriately distributed (Chapter 8). Random occurrence does not make
a system. And man without pattern is not man.
The distributional component, from a different perspective, results—
when items are seen as lined up in a structure—in the presence of a rela-
tional field. So we already find that there is no partitioning here between
distribution and field; they are related concepts used for slightly different
purposes under related but different conditions. So also, variants can be
seen in relation to merging or growing—and variation overlaps on the dy-
namic perspective of wave.
Why, then, do we find value in such a multiple perspective? Why not
Just one, kept neatly partitioned into its parts? Life won’t allow such isola-
tionism of fact from fact, man from man, view from view, or man from fact
and view.
Reductionism is inadequate. We do not have access to the ultimate mini-
mum units; our successors may find even smaller bits of matter or of en-
ergy, just as our contemporaries have gone deeper than did those Greeks
who talked earlier about atoms. And there are thresholds where the whole
cannot be equated to parts in mere combination; hierarchy is needed, with
threshold phenomena (see Part III). Then any discipline can enter wher-
ever it chooses, by an arbitrary—human—choice for pleasure or profit,
and build upwards to expanding complexity, and downwards to descend-
ing complexity (not downward to ultimate simplicity, since the parts are,
once more, viewable only by man in system, in structure, in emic relations
to causation as deduced by man in man’s system—and this is not simple).
Biology may choose to enter via a living cell, going up, for example, to man
and down to photosynthesis—or something else. In linguistics, I personally
choose to enter not at the level of the sentence, nor even by use of a feature
of a sound (for example, voicing), but at the level of social interaction of
person with | person. This leads directly to dialogue, personal response, and
definition of sentence in relation to dialogue (see Chapter 9). It leads also to
pronunciation elements such as the relatively isolatable syllable (see Chap-
ter 10). And it opens the door to the systematic study of encyclopedic par-
ticulars (Chapter 11)—of particular men, particular events, particular ta-
xiv Preface
bles, particular ideas as seen by man and related to him. Again, it circles
back on particles as particular things, or on a background underlying a
(tacit) field.
But is such complexity logically justifiable? Is only logical simplicity to
be acceptable to the academic community? Here I quote the logician Su-
sanne K. Langer: “We are no longer limited to propositions that are simple,
obvious, and generally entertained. If we chance upon a fairly complex and
even surprising proposition, from which very many simple ones would fol-
low, we are perfectly justified in taking the former as a postulate, and de-
riving the others from it” (1953:185). And: “One’s aim in formulating an
algebra is always to reach as soon as possible the greatest number of im-
portant propositions. Which propositions are ‘important’ depends upon
the use one makes of the algebra” (308).
Tagmemic theory—the name for my theory of unit-in-context as pre-
sented here—is more complex than some theories in demanding that con-
text be considered at every step (Part IV): that is, in all perception and ex-
perience and knowledge. Form must not be treated apart from meaning
(Chapter 12). Humans perceive forms in relation to situational function,
whether that function is perceived in relation to nonhuman cause (a thun-
derstorm, perhaps), or human cause (a war). In each case, meaning of
some kind is present—or else the emic form cannot be perceived. (An un-
known object may be perceived emically as an unknown “thing”—with
borders perceived in relation to prior experience.) Similarly, change
(Chapter 13) occurs only in relation to some merging of unit with unit, or
in relation to a containing field. So, once more, the perspectives of particle
and of wave merge with field, with context, with form-meaning com-
posites—and with a universe of discourse (Chapter 14). The last of those
approaches field structures from the viewpoint of background system,
rather than (usually) concentrating on the relationship of those units ex-
plicitly under analytical attention.
Again, an objecting query can be raised: Why so many different ways of
approaching related things? And, again, we draw on Langer for our re-
ply—where she tells us that unless we have alternative ways of specifying
data relations, we do not have a useful algebra: “The possibility of making
deductions from given facts depends so often on the form in which these
facts are given, that a large supply of interchangeable forms is the first re-
quirement for an interesting system. The establishment of more such forms
is, therefore, our first ambition in developing the algebra, and the theorems
we are most anxious to prove are such as yield more laws of manipulation”
(1953212).
The approach here is designed to serve in a very wide range of circum-
Preface xv
stances. The principles have been applied, for example, to football games
(Pike 1967a: chap. 4), party games (1967a: sec. 1.2), church services
(1967a: chap. 3), a breakfast scene (1967a: sec. 5.2), and—briefly—to so-
ciety (1967a: chap. 17) as well as to language. Human emic experience is
the target, not merely linguistics.
Tagmemic theory developed by accident. I did not set out to develop such
a theory. I was involved in training linguistic research workers for the study
of preliterate languages. As teachers, we did not know to what part of the
world our students might eventually go or to what kind of language. We
had to teach them by general principles, in part, to be ready for anything.
As we became more and more general—and were able to teach more and
more in the limited time span, since one principle could be adapted to
many different situations by minor adjustment—I felt like a man climbing
a mountain. At first he could see a small way; as he climbed higher, he
could see farther; at the top, he could suddenly see in all directions—not
just farther in one direction. This suggests the manner in which I gradually
discovered, to my utter surprise (and delight) that the principles selected
for linguistics were equally applicable to anthropology—and eventually to
other phases of human activity.
From 1935 to 1948 I had been largely involved in studying pronuncia-
tion—phonetics, phonemics, intonation of American English, tone lan-
guages. In 1948 I asked whether or not there might be a “phoneme of
grammar”—and if so, whether the concept would be as helpful as the pho-
neme to those of us investigating preliterate structures. With this question
in mind, I eventually worked out contrast, variation, and distribution for
the tagmeme (for a while called “grameme,” in the first publications of this
view in 1954 [Pike 1967a]). Eventually, the analysis of the unit was elabo-
rated into the tagmeme as unit-in-context, with the four cells, as described
below in section 9.4.
In 1959 I wanted to reach a more general audience, and used the meta-
phor of particle, wave, and field to try to do so. It turned out that this view
of perspective was very useful as an approach to the observer and his role,
to supplement the approach through units. In 1960, further development of
approach through field began when I asked whether or not there might be
in grammar an analogue of the phonetic chart, but with the chart treated
seriously as a legitimate emic unit of the observer; this led to numerous
publications helpful in the study of clauses, of affixes, and eventually of dis-
course itself
Following the question about a “phoneme of grammar” in 1948, the the-
ory had developed far enough by 1950 that my wife Evelyn began to make
up artificial problems to help teach the approach (much as an arithmetic
xvi Preface
book might have problems about two plus two, without mentioning partic-
ular apples). A few of the exercise problems from that year are still pre-
served in Pike 1967a:212-17.
In 1966, I drafted the present book on concepts, and had half of it mim-
eographed for classroom use by 1968. But I ran into a major problem:
Meaning, although treated in 1967a: 598-640, was not as clearly handled
as I needed, if I were to feel comfortable in finishing the book; so I put it to
one side for over a decade. In 1971, however, Evelyn and I began a ped-
agogical text to implement the teaching of the semantic and discourse char-
acteristics of the theory (with exercises to include the kind of work we had
done before, but to go beyond it). Here Evelyn had a semantic break-
through, with “purpose” being added explicitly to the treatment of the
events of the referential hierarchy (replacing our older lexical hierarchy,
but retaining lexical items as the substance manifesting the new hier-
archy). In addition, she pushed the notation for the four-celled tagmeme
consistently from the highest levels of discourse down to the morpheme—
which I had not myself done. These matters were then included in Pike and
Pike 1977, and tested in the classroom for several years both at the Univer-
sity of Michigan and the University of Oklahoma.
With this behind us, I then returned to the present book, seeking once
more to try to write for an audience which might be interested in the gen-
eral principles involved. I wish them enjoyment as they try to read it. They
may have to struggle with some of the data and claims, of course, since the
views presented may seem to them to be buried in some irrelevant way. But
this may often be the price of pleasure. Suppose, for example, that one
wants to know the analysis of
WOWOLFOL
It may take a bit of study to see that it hides a wolf in sheep’s clothing.
=.
The Observer and Things
Introduction to Part I
When man studies “things,” he injects part of himself into their

definition. What is a chair, if there is no man to sit on it? A flute, with
no player? A concert, with no listeners? A saw, with no carpenter? The
relevance or intended use of a thing is part of its nature as experienced by
us—a component added to it by its designer or user or deduced by an
observer.
One approach to studying language emphasizes that man as a user of
language affects the nature of the units of his language. His reactions to
language become part of the data for the study of language, because de-
scribed expectations of his reactions are part of the definitions of the struc-
ture of language.
The list and kind of things men find will vary radically if they adopt dif-
ferent theories as tools with which to search for these units. The theory is
part of the observer; a different theory makes a different observer; a dif-
ferent observer sees different things, or sees the same things as structured
differently; and the structure of the observer must, in some sense or to some
degree, be part of the data of an adequate theory of language. A particular
language, of a particular culture, in relation to a particular person with his
particular history constitutes an implicit theory for that person.
Tagmemic theory is, in this respect, a theory of theories which tells how
the observer universally affects the data and becomes part of the data. No
wonder, therefore, that tagmemic theory cannot stop with confining its in-
terest to mere language, but must view language in the broader context of
the study of ordinary lay nonverbal behavior as well as in the context of the
behavior of that special language observer, the linguist.
Chapter 1 discusses my view of the nature of theory itself. Chapter 2 gives
some basic concepts accepted by tagmemic theory as a foundation for its
development. Then Chapter 3 discusses the first of three kinds of perspec-
tive which the observer can tacitly or explicitly adopt: the view of the uni-
verse as made up of particles. Chapter 4 covers comparable ground, show-
4 The Observer and Things
ing that the same things or situations can be viewed as ifthey were waves.
Chapter 5 changes the perspective to one of field—that is, the same data
once more, but this time with units viewed as points or relations in a pat-
tern as part of an even larger system.
Theory
“Why study theory? Why not just be practical?” So speak those

who fail to realize that the line between theory and practice is blurred—
that in many situations only an approach to theory will allow practical re-
sults to be obtained in reaching one’s goals. Today’s practicality is often no
more than the accepted form of yesterday’s theory.
1.1. Value of a Theory
A theory is like a window.

The intellect, in order to get outside itself and to interpret the sense data
impinging on the body, needs in advance some kind of idea of the way in
which the data may turn out to be organized. Then it can search for
pattern. —
A thegry in this sense is directional (see fig. 1.1). By looking out of a
south window we get one view, but out of a north window a different view.
Both lead to partial insight into one’s surroundings, but in different direc-
tions. Sometimes, however, the same view may be seen through two dif-
ferent windows. Similarly, different theories may each contribute insight
into the nature of patterns of language. If we look at the same data through
different theories, we may see different aspects of a pattern.
A theory must be simpler than reality if it is to be helpful. It attempts to
strip away from attention those items which are not important to the ob-
server at the moment. In this way it helps obtain answers to particular
Figure 1.1. A theory, like a window, can look at only part of the data,
and in one direction at a time.
questions on a narrow front by simplifying the task of investigation. Only if

a theory is simpler than that reality which it is in part reflecting is it useful.
In physical situations, a model may be physical—a construction, to
scale, which can be destroyed or distorted. The model is less expensive to
destroy than the real thing would be. Conceptually, on the other hand, a
mathematical model may allow one to manipulate in more detail and more
extensively some of the characteristics of reality without interference by
other characteristics not built into the model. The relation of two plus three
does not need apples—or bridges—to clutter the addition process.
It follows, therefore, that a model allows more exhaustive and systema-
tic tests for relationships between certain variables than can be handled or
isolated with ease in reality itself. By exhausting possibilities of certain re-
lationships one can test certain implications of those relationships without
the expense involved in trying to test them when other factors interfere.
1.2. Weaknesses of a Theory
Any theory may have a weakness at the point of its greatest

strength. Since a theory looks in a particular direction, as in the window
illustration, it may tell us nothing about data or characteristics of reality
which must be seen from some other vantage point. A scientific theory is
good only if it leaves out wisely those materials which are relevant to other
questions but not to those immediately being answered. But the necessary
valuable simplicity of a particular theory can destroy its usefulness if it
Theory 7
happens to leave out data which are in fact at that moment important. Un-
fortunately, the observer cannot always be sure what is important; a mis-
take here will hurt him.
All theories eventually are doomed to be outmoded. A useful theory in-
vestigates a point of interest to a particular observer at a particular time. If
the theory is successful, that particular problem will have been solved by it.
Interest then switches to other problems, requiring further theory. But these
other problems are of the deepest kind only if they involve data deliberately
or unintentionally left out by the otherwise successful theory. The next stage
of investigation requires a model which is more inclusive. When that one,
in turn, has made its contribution, it needs to be replaced by another with
wider, or different, perspectives which cover other parts of the physical or
conceptual universe.
A good theory is like a smooth, clean window. A bad theory may allow
us to see, but with excessive distortion, blurring, or a filtering out of some
useful information relevant to the task at hand. A window which has wavy
glass may distort reality. Nevertheless, if one wishes to look out of a room,
one had better have a glass window which is wavy than have no window at
all. A dirty window allows one to see something, even though what one
sees may be blurred. To have a poor theory is better than having no theory
at all.
1.3. Selection of a Theory
In trying to choose or to build a theory we should seek an orga-

nized, systematic arrangement of general principles which will help us to
understand something about our physical or conceptual world. We wish
for insight into the nature of the setting of our lives, our behavior, and the
things with which we must cope.
Since we want a theory to help us, a theory may be viewed as a concep-
tual tool. A good theory is a useful one. Usefulness, in turn, is relevant to
some purpose, to some goal. This implies that theories may be good or bad,
relative to the sociological setting in which they are found (see fig. 1.2). A
dentist’s drill and a steam shovel both are useful for excavation, but not at
the same spot. Einstein’s theory of relativity may not comprise the most
profitable mathematics for designing a culvert—but most of the world’s
traffic at some time rolls over culverts, whether on rails or at an airstrip.
The freight which a linguistic theory may be called on to carry may in-
clude the refinement of techniques for teaching foreign languages, the
Figure 1.2. Choice of a theory is in part pragmatic. One may select a

theory as one selects an airplane. How far must the plane go? With what
load? With what available runways? How expensive is it to obtain? To
use? (For this analogy, see Frank 1957 :357.)
preparation of an alphabet for a preliterate culture, the teaching of fresh-

man composition, and the provision of frames of reference to help us un-
derstand the relation of language to culture, of language to psychology or
philosophy, or of language to life and action.
A theory is most likely to be useful if its results, or predictions, can be
easily tested. In general, this can be done by two radically different meth-
ods: the inductive one works from the data toward the theory, and the de-
ductive one works from the theory toward the data. Both are useful and
effective. In the inductive approach, some available data may be carefully
classified as elements appropriate to parts of a guessed-at larger pattern.
To test the accuracy of the guess, more data are taken; the analyst then tries
to fit these into the pigeonholes of the pattern. If all goes well, one knows he
is right—until he is proved wrong.
Theory 9
In a deductive, formal approach, one guesses, to reach an idea of a

larger pattern, without worrying much how that pattern was arrived at.
Then, to test the hunch, a formal theory is built and used. Such a theory can
be thought of as a kind of intellectual machine with three parts: an initial
set of axiomatic sentences containing primitive terms, not defined by the
theory; secondly, an interpretation, preferably through mathematical for-
mulas which look from these initial statements toward the observed data;
and, third, specific predictions about data to be found (compare Carnap
1955 : 207, 210.) This time one knows one is right—until proved wrong—if
a satisfactory number of the data not in hand at the start, but now checked,
are just those which the machine predicted should have been there.
A part of the world which we wish to investigate may be likened to a
mysterious old castle. One boy may enter it by scaling the high walls; an-
other, by smashing through a sagging door; a third, by crawling through
an old escape tunnel. Their first-stage excited reports to each other seem to
describe different, unconnected buildings. Later, after they have met some-
where inside the castle, it becomes clear that each can find all the areas re-
ported by the others. Each avenue of entrance, however, has its own pecu-
liar directness. To the dungeon? Use the tunnel!
While working with data, all scholars from time to time utilize parts of
various underlying theories, not just one. They differ in the proportion of
energy devoted to each. In publication, on the other hand, a scholar often,
but not always, presents all of his conclusions from just one viewpoint.
Logical consistency may appear to him to be desirable in the presentation
of results, even if it is impossible during the stages in which the data are
being found and analyzed.
Survey of Tagmemic Theory
2.1. Beliefs Underlying Statements
No statement can be made seriously unless preceding it there is

in the speaker’s thoughts an underlying set of beliefs which he holds firmly,
but cannot prove. Ordinary statements and theoretical statements share
this restriction. In some sense, man cannot begin with known facts; he has
to begin with some kind of commitment, such as the commitment to believe
in the existence of facts or the possibility of obtaining knowledge at all.
Often some of the beliefs underlying scientific statements are so far re-
moved from the immediate problem at hand that the speaker does not try
to trace them back that far. For example, the scientist normally begins with
the belief—which he does not question, and which would not be shaken by
an opposing view—that there is a world to be investigated. Such beliefs
may appear to be common sense. Nevertheless, when a new basic theory is
being presented, many of one’s older beliefs may be challenged. Occasion-
ally some belief long held will be modified in the face of new kinds of evi-
dence—with far-reaching results in action based upon these beliefs.
When differing scientific views clash, one needs to try to discover what
kind of scientific reality, or truth, their authors believe in and start with.
This responsibility rests upon us in linguistics; and our particular commit-
ments need to be made explicit for the reader so that he can compare them
with implicit or explicit commitments of other scholars. We attempt, there-
fore, to give here a collection of some of the concepts involved in tagmemic
theory. No one basic concept of the theory can be explained (for under-
standing) or exploited (for usefulness) without reference to each of the oth-
ers. A theory is, in this sense, a single interlocking whole, just as a circle is
not a circle if one part is left out.
Survey of Tagmemic Theory 11
Some Conceptual Tools

Perspectives as complementary observer standpoints
Elements seen as particles
Elements seen as waves
Elements seen as fields
Units as structured
With contrastive-identificational components (features)
With variant manifestations
With distribution appropriate to class, sequence, and system
Hierarchies of units as parts of wholes

Phonological
Grammatical
Referential
Contexts as relevant
To form-meaning composites
To change
To universe of discourse
Figure 2.1. A set of interlocking tragmemic concepts which can be thought

of as a set of tools for analyzing language and for describing human
behavior. Various others (for example, norm, role, nucleus, lexicon, emic)
are related to these.
The concepts presented here are useful for understanding human behav-
ior, including the nature of language, and for finding the structure of that
behavior. A number of the most crucial concepts are listed in figure 2.1.
Discussion of these in later chapters will make up the principal content of
this book. Although many of the problems represented by this list may be
traced back to the horizons of history, the synthesis of them given here is
new (Pike 1967a); it is given the name tagmemics, related to the Greek
word from which we get the English word tactics; it helps to suggest the
relevant structured arrangement of behavioral units relative to an in-
sider’s (emic) view of a behavioral system.
2.2. Complementarity of Perspective
Two photographs may appear identical to the casual observer.

If, however, these have been taken from vantage points just a few inches
apart, and seen separately by the eyes through a stereoscopic arrangement,
the brain synthesizes these views into a three-dimensional experience. Sim-

ilarly, there are three views of linguistics which cover approximately the
same material and which in some respects are similar, but which are dif-
ferent enough to allow a far richer experience if the linguist uses all three
than if he uses only one. The three-dimensional experience of the stereo-
scopic view is in part added by the observer, and in some sense is not di-
rectly in the flat photographs themselves. Similarly, in language no brute
facts are available—all of them are filtered through the receiving mecha-
nisms of the observer. His experience of the factness around him is affected
by his perspectives.
Within tagmemic theory there is an assertion that at least three perspec-
tives are utilized by Homo sapiens. On the one hand, he often acts as if he
Figure 2.2. Merging of perspectives in a simple model can show how three
different designs can point to the same units. The units can be shown as
if they were isolated particles (A), or as peaks of flowing waves (B), or as
points at the intersection of lines (C). All can be superimposed (D), giving
the same set of units. The composite enriches our experience of life.
were cutting up sequences into chunks—into segments or particles. At

such times he sees life as made up of one “thing” after another. On the
other hand, he often senses things as somehow flowing together as ripples
on the tide, merging into one another in the form of a hierarchy of little
waves of experience on still bigger waves. These two perspectives, in turn,
are supplemented by a third—the concept of field in which intersecting
properties of experience cluster into bundles of simultaneous characteris-
tics which together make up the patterns of his experience (see fig. 2.2).
2.3. Units as Known
The human being experiences the world around him as made

up of units. For an element to be treated as a unit it must be considered as
different from (in contrast with) other units. Only those characteristics
which make one unit different from another can be used to identify that
unit. The child knows this, as when he asks, “What is the difference be-
tween an elephant and a mailbox?” And adds, “If you don’t know, I won’t
send you to mail my letters!”
In order to know a unit adequately for purposes of identification, we
must know the way it changes from time to time without ceasing to be it-
self. How much different can something be and yet not be something else?
What are its variants? (See fig. 2.3.)
Third, a unit is well known only if one knows where it can appropriately
be found (its distribution must be known). Unless we know where it is
appropriate for a unit to come, we cannot use it in the right places, and we
will appear awkward, unintelligible, or foreign. A noun and a verb, for
example, occur at different places in a sentence.
2.4. Hierarchy of Interlocking Hierarchies
Man experiences the world as if it were made up of wheels

within wheels. Little items are in larger ones, larger ones within still larger
ones, until the world becomes one thing after another—larger and larger
—or smaller and smaller: “There is a frog on the log in the hole in the
bottom of the sea.”
In language we find three kinds of hierarchy which are partly indepen-
dent but which at the same time interlock with each other. In the gram-
matical hierarchy, meaningful lexical chunks make up parts of words en-
tering into larger chunks of structure, which in turn enter into still larger
Susie: 1 year old Susie: 5 years old Susie: 13 years old
Figure 2.3. Units change at different times or places. Talking to Susie

appropriately depends upon recognizing both the continuity and change
in her personality.
chunks. At the bottom of this hierarchy are affixes—that is, prefixes like
re- of return, or suffixes like -s of returns, or roots like the turn of returns.
These make up words, like returns, as a whole; or phrases like will be
returning; or clauses like when the tiger will be returning to his kill; or
sentences like Tomorrow is the time when the tiger will be returning to his
kill; or paragraphs discussing a topic announced by such a sentence, de-
veloped by further sentences such as So we better be prepared. Monologues
carry long or short speeches of an individual, enclosed in exchanges be-
tween speaker and hearer, such as:
First speaker: The tiger will return.

Second speaker: So What?
And conversations can carry on further, with various speakers.

On the other hand, the way which a man pronounces may also be ar-
ranged in the form of wheels within wheels—smaller items within larger
and larger and larger ones. Cat is a syllable in which three sounds (spelled
c, a, t) occur. In the long word prestidigitator, however, there may be
several syllables, with just one accented syllable—which we may call
stressed—so that the whole word forms a single larger pronunciation unit,
that is, one stress group. Stress groups become important in the meter of
many poems; the line sometimes forms a still larger pronunciation unit;
and the entire poem, when read aloud, may be pronounced in such a way
that a person can tell that the reader is going to stop—evidence that it too is
a pronunciation unit of some kind, in the total phonological hierarchy.
In addition to the grammatical and phonological hierarchies, we assume
(at least in our tagmemic theory, not necessarily in this form in other theo-
ries) that there is a referential hierarchy. The referential hierarchy includes
the talk-concepts which people have about things and about events or
about features or situations relating such things and events, which they
observe—or imagine—and talk about (or think about). It is, then, the ob-
served (or imagined) and talked-about reality which is concerning us here.
We are simply silent about the “thing-in-itself,” which may exist apart from
any human observer; things and situations or events enter into our analysis
only when some perspective is involved. (There is a theistic perspective—
of which my own epistemology would be a sample—which would involve
an Observer whose hidden-to-us views are not part of our scientific equip-
ment; but such a perspective is relevant to various discussions concerning
the nature of reality.)
Identity of a talk-concept referential unit is specified for a particular time
and situation by paraphrase, that is, by the ability to say the same thing in
other ways which the hearer and speaker can agree on as being the same
concept for their joint temporary purposes. For example, Joey’s mother can
say to a close neighbor and be understood (but not necessarily successfully
by a stranger) Joey just came home, or My boy just arrived, or It is my boy
who just arrived, or My boy just arrived. Such different ways of saying the
same thing include a relevant close similarity of basic content, but often
differ sharply in terms of focus affected by the lexical choice, by grammati-
cal form, or by phonological emphasis. Each of these differences is mean-
ingful, in our view, but attributed to a different hierarchy. And it is impor-
tant to point out that we do not put into the referential hierarchy the thing
or the concept abstracted from the speech or the observer. Observer;
speech, and experience of thing, situation, or events are kept tied into a
package in the units of the referential hierarchy.
2.5. Context as Relevant
Language itself can be viewed as action—as a kind of behavior.

When people talk to other people, they may wish to influence them to act
differently, to believe differently, or to interrelate with them in some so-
cial way.
If language did not affect behavior, it could have no meaning. The bits of
Figure 2.4. The meaning of a word comes from an experience of its be-
havioral and lexical contexts. The making of a formal dictionary defi-
nition comes later, with the lexicographer exploiting some of those
contexts.
forms which we have discussed—sounds, words, sentences—are mean-

ingful not in and of themselves, not just because they exist, but because
when used in social environments they do in fact ultimately affect behavior.
The meanings grow out of this social relationship. The command Jump!
implies “jump” only because it has been used in situations where this part
of meaning is somehow reacted to vigorously (see fig. 2.4).
In treating language as behavior, therefore, we first make the point that
language elements are combinations of form and meaning. We try very
hard to avoid studying form by itself or meaning by itself. We deal with
them both together. We can never discuss either of them unless, lurking
somewhere in the background, is the other. Even when a person tries to
talk about the isolated forms of words, he knows that they are meaning-
ful—or he knows that somebody knows that they are meaningful—or he is
not handling language. Similarly, if a person tries to make a classification

of isolated possible meanings, he is likely to end up without helpful results
unless somehow these meanings arise from words which are tied into some
language system or systems of cultural behavior (including scientific be-
havior and its classificatory devices).
Unless one item shares some contextual feature with another, it can have
no impact on it—no action on it at an “unconnected” distance. For one
man to be able to shoot another, they must belong to the same universe and
both be subject to the same physical laws. To talk to one another, men must
share some kind of language. Change involves sharing.
It is often convenient to speak of a bridge over which change passes—
with the bridge being the item or context shared in such a way that it si-
multaneously belongs to (or contains) the two elements or the two events.
Across a river a bridge is shared by both sides (see fig. 2.5). Across a time
change, events share the joining moment. The first two sounds of kitten
smear into one another, by sharing mouth positions part of the time. Dur-
ing the /k/, for example, the middle part of the tongue rises high in the
mouth getting ready for the vowel; during the start of the vowel, the back of
the tongue is gradually leaving the top of the mouth, where it has closed the
mouth for the /k/. The consonant affects the vowel, and the vowel affects
the consonant—change works in both directions.
(Engineers building the bridge)
ae ‘’
A: Army engineers, bridginga river, call to their colleagues:
Why are you shooting at us?
(Enemies)
B: The friends reply:
An enemy might come across that bridge.
Figure 2.5. Sharing occurs for change to occur. So change may work in
both directions.
When one has two words in a row, or several words scattered through-
out a paragraph, the mind has to connect them in some way as belonging
to the same area of shared discourse—a universe of discourse. Thus traf-
fic, bridge, automobile, transport, and the like come under the general
topic of traffic. Similarly, spring, summer, fall, and autumn are seasons.
Bicycle, tire, seat, and chain may all be expected to occur in certain kinds
of discussions concerning bicycle repair. These universes of discourse are
determined, not by the words themselves, but by the relationship of the
words to the larger culture within which they are used. A linguistic uni-
verse of discourse, therefore, links language and society.
Various linguistic concepts can be illustrated by using this pun: Did you
know that Robinson Crusoe started the forty-hour week? He got his work
done by Friday. As a single unit—a joke—it can be viewed as a particle
made up of various smaller particles (including two sentences, for exam-
ple). It contrasts with other jokes. Referentially, it may be paraphrased in a
variety of ways: We know that Robinson Crusoe started the forty-hour
week because he got his work done by Friday. Change from the universe of
discourse of our social structure to that of Crusoe’s personal relations oc-
curs over the identity of pronunciation which serves as the bridge which
links the day Friday to the person Friday.
The characteristics of a linguistic theory need to be, insofar as possible,
those which are required by human nature. Tagmemic theory insists that
the efficiency of the brain in using language, and of the child in learning all
kinds of nonverbal patterns of activity, is due to the fact that many of the
deepest principles of the structure of human nature are equally responsible
for activity of nonverbal and verbal types; the child does not have to learn
or utilize a wholly different set of components for his different activities.
A general theory of language is useful in proportion to its exploiting of
these shared features. Tagmemics grew up somewhat like the view which
grows in climbing a large mountain. At first one sees but a short distance in
one direction. As he climbs higher, more of the countryside is in view.
When he reaches the top, he can see in all directions. So also with tag-
memic theory. It started with a struggle with a few local linguistic prob-
lems. As more languages were tackled and as general principles were
needed to analyze languages totally unknown to the linguistic community,
it happened that the practical principles enunciated turned out to cover hu-
man behavior in general, that is, it became a theory of language in relation
to a unified theory of the structure of human behavior (see the title of Pike
1967a).
Particles
3.1. Things and Nonthings as “Things”
Human nature experiences the world—sometimes—as made

up of particles, that is, as “things.” Houses, trees, and people seem to be
obvious physical objects. But often an experience or concept which ob-
viously is not an object is also talked about as if it were one (see fig: 3,1);
and is perhaps even counted or measured. Such a concept may be illus-
trated by the term movement, which implies that something is going on;
but movements can be counted, like houses (I saw two houses; I saw two
movements).
An abstraction like beauty or force can be put in a sentence such as
Beauty is desirable (as Food is desirable); Some force moved us or The
bomb moved us). A class of items can be treated as if it were a unit: the
American contingent. Or a sample of a class can be somehow treated as
representing, the whole class: He is the average American. A symphony
may be heard as a unit, even though it may take an hour to play.
Our mental computers seem to be able to operate only if the whole world
is processed into bits, each of which can when appropriate be treated as a
particle, punched onto our mental IBM cards.
The normal relaxed attitude of the human being in most of his actions
treats life as if jit were made of particles. Perhaps this is one reason why
many people think that a language is just a collection of words one after the
other. (Translation, to them, appears merely to require the translator to
look up a word in a bilingual dictionary and replace it by the proper
equivalent.)
(a melody)
house in decay
VO/e,
(flowing letters)
Bare
es Ty
(beauty)
ee Oe ess ee
wes
process
(force)
Figure 3.1. Things, processes, and relations can all be viewed.as “parti-
cles.” This provides a static view, in which a melody can be experienced
as a unit, even though it takes time; a house, though in the process of
falling down, is still a thing; letters of the alphabet are often treated as
separate items, even when inseparable in cursive script.
Particles 21
3.2. Particles in Language
In the word kit there is a sequence of three units of sound. The

first is the pronunciation of /k/. This is the same unit as found at the begin-
ning of cat (in spite of the change in spelling) or character.
A speech sound is made by the moving parts of the mouth, throat, and
lungs. If one takes an X ray moving picture of these motions, a single frame
of the movie taken at the central part of the sound gives a static picture of it.
In Figure 3.2 /t/ is shown this way.
Words too are particles. They often include within them small particles.
The central (nuclear) part of a word is the stem, which carries the princi-
pal meaning of the word. Smaller parts (affixes) affect the way in which the
principal meaning is to be interpreted. In Candoshi of inland Peru (Cox
1957), for example, the word tayanchshatana is made up of parts: ta-va-
nch-sha-t-a-na. This means ‘be-recent-complete-next-individually-I-em-
Figure 3.2. A static view of a sound represents it as one frame from an

X-ray moving picture.
phatically,’ or ‘I was recently and completely just then individually em-

phatically there,’ or ‘I’ve stayed there then.’
Some language particles are larger than words. If we say the ugly dog bit
the mail carrier, the whole is a single clause construction, a unit, a parti-
cle. The particle the ugly dog is a smaller unit, a phrase which serves as the
subject of that clause. Other phrases, however, could have taken its place.
We could have said A small terrier bit the mail carrier or Something bit
the mail carrier, and so on. The whole range of possible actors in this rela-
tionship to the verb phrase (bit) and to the object (mail carrier) makes up
a class of replaceable forms. This class plus its functional role in relation-
ship to the rest of the clause can be treated as a tagmeme particle called
“subject-as-actor”; the mail carrier would be an instance of (or represent)
a different functional unit, an object of the clause, as bit would be an in-
stance of the transitive predicate tagmeme. (The class sequence noun-verb-
noun [N V N] by itself is not enough to represent the structure of the clause,
since it does not specify the actor-action-undergoer role relationships of its
parts.)
Noun classes may themselves be viewed as particles, however, of a kind
different from constructions, relational parts of a construction, or separate
words.
Often we have expectations of a form which is under way, long before
the form is fully visible. This may be true, for example, of a limerick (see
fig. 3.3) even if we hear only the first line. In There was a young lady from
Clyde, the stresses on was, lady, and Clyde start a phonological pattern
which may alert us to something coming. The choice of the words there
and young contributes to our beginning expectation. The grammatical
structure of was after there adds a further clue, as does the choice of
lady—or some analogous noun. By the time we add the second line Who
ate some green apples and died, we are probably quite certain, now, that
we are headed into a limerick. The rhyming of Clyde with died, the rhythm,
and the lugubrious flavor, all encourage this guess.
There was a young lady from Clyde

Who ate some green apples and died.
The apples ferménted
Inside the laménted
And made cider inside her insides.
(/sa'dar unsa'‘dar unsa‘dz/)
Figure 3.3. A limerick is not a random sequence of sentences, but a single
internally complex coherent unit, a “high-level” particle. (Source unknown.)
Particles 23
The expectation of limerick form leads us to hope for some kind of diffi-
culty which is solved only in the last line by a surprise verbal twist. In this
limerick, the pleasure is heightened by the pun of cider with (in)side’er. No
such expectation of form, and delight in finding it, could occur unless the
limerick as a whole were some kind of particle, a unity.
Once the hearer (or creator) of the unit grasps the fact that a particular
kind of unit has begun, the unit’s general pattern itself takes over and de-
mands its formal completion. One is driven by the form once the form is
begun and recognized. Language is a pattern of such particles on many
interlocking hierarchical levels.
3.3. Particles in Linear and Spatial Order
The analysis of some phase of life as made up of a string of

particles may be called a linear kind of order. The mind is able to under-
stand life, in part, by thinking of it as one thing after another. These linear
distributions have a great grasp on us.
Learning a different order of alphabet, for example, is not easy. Learning
to count also takes considerable effort. In many languages of the world
counting has never been developed. In some languages of New Guinea, for
example, people (until recently, at least—I myself met some in 1949)
count: one thing, two things, many things, and then must stop; one man
translated his counting system into Pidgin English for me as one fellow, two
fellow, plenty fellow. In eastern Peru I met, years ago, the man who had
invented the number eight for the language Piro.
Linear order may be replaced (or accompanied) by items arranged in
two-dimensional or three-dimensional space. Books may be scattered on a
table; flowers may be arranged in a garden; rooms may be divided into
upstairs and downstairs floor plans. Sounds can be charted in rows and
columns which show characteristics which they share. (The sound /p/
shares with /b/ and /m/ the fact that the lips are closed when it is made. )
Such spatially ordered arrays represent particles as being within a field.
The linear order of particles and their ordering in a field are both impor-
tant to linguistics.
Wave
4.1. The Merging of Sounds in Sequence
When one tries to read a paragraph as if it were made up of

completely separate particles of sound, an astonishing result is heard. The
material is very jerky and practically unintelligible. It is worth trying in
order to hear the effect (see fig. 4.4). Natural speech never sounds like that.
Why not, if speech is made up of particles of sound? Surely the analysis of
speech into separate chunks is in some manner false—a model useful for
some purposes, awkward for others.
The same material will sound very different if, instead of putting spaces
between the sounds, one reads the lines very, very slowly indeed—as if a
phonograph record were slowed down to half speed—and lets one’s
mouth movements glide slowly from one sound into another. This also is
well worth trying for its effect in comparison to what one would expect if
sounds were separate particles. They drift into one another—have no clear
starting and stopping points.
Instruments show that the sounds do in fact slur into one another. One
sound is not finished before the next one is begun. Speech is more like
sloppy cursive handwriting in this respect than it is like printing. The dif-
ference between bee and boo, for example, can be seen by looking in the
mirror—lips are rounded for the /b/ of boo while they are getting ready for
the following vowel.
Words may be so run together that they become completely or partially
simultaneous (see fig. 4.2). Some years ago my teenage son used to pro-
Wave 25
A. O-1-d-M-o-th-e-r-H-u-bb-a-r-d-w-e-n-t-t-o-th-e-c-u-pb-
oar-d
Figure 4.1. Sounds smear together in natural speech. If each sound is

artificially separated by a pause (A), speech becomes jerky and unin-
telligible. If one reads B equally slowly (like a phonograph at half speed),
one can see in a mirror the continuously changing mouth positions.
A. I don’t want to.

B. [a-1-0-4-n-5]
Figure 4.2. Words can merge. The word not (A) has already weakened to
n’t. A whining child may merge this much farther—almost into a se-
quence of nasalized vowels (B). (Pronounce the schwa [a] as the vowel in
cup. The tilde [~] means the air comes out the nose at the same time, to
make nasalized vowels.)
nounce the phrase I don’t know (along with a shrug of the shoulders) using
just the sound m, with a hum which rises, then falls considerably, and rises
a little bit. If we hear something like that in a foreign language, it may
sound incredible (see fig. 4.3). In Chinantec a similar smearing of m over a
couple of syllables has led to some verbs where the entire conjugation is
given with the mouth closed. (Most of us do a bit of this kind of thing in
mom for English ‘no’, and mhm for ‘yes’.)
A. (English) m | ‘I dén’t know’
B. (Chinantec) m‘~ ‘T ask for’

ie ‘He asks for’
mom ‘You [singular] ask for’
m]- mpm ‘We [I and he] ask for’

m|: me ‘You [plural] ask for’
m|-[> ‘We [I and you] ask for’
Figure 4.3. Smearing of sounds can be carried to such lengths that only a
hum, with pitch, is left. In English the sentence 1 don’t know may become
m (A); a line drawn over the letters shows the English emic scale offour
relative pitch levels: just above the letter, high; just below, mid; extra high
and low may also occur. (One may whistle the pitches to get a crude idea
of the pronunciation.) In Chinatec (B) of Mexico, there are also four sig-
nificant pitches, but as regular parts of ordinary words. (Perhaps an old
stem vowel has smeared into the m, and then the m has smeared over
onto the suffix. The dot represents extra length; the [o] is a glottal stop,
as in the middle of a sharply pronounced Oh oh!) (Data from Robbins
1961.)
4.2. Nucleus and Margin of Wave
The clearest or steadiest part of a sound during slow pronun-

ciation can be called the nucleus of a wave. The borders, or transitions,
from one sound to another can be called the margins. In wave types which
interest us we often look for their nuclei and margins. If our attention is
almost exclusively on the nucleus, however, we may be thinking in terms of
particles. When our attention is on the whole flowing movement from mar-
gin to nucleus to margin, we are thinking in terms of the unit as a wave.
Human nature requires that people focus attention in some direction—
on some element or concept or view. The person who tries to watch con-
struction out of the corner of his eye while he drives down the road is likely
to end in the ditch. Language reflects this capacity for directing focus.
Stressed pronunciation can be used to emphasize or highlight parts of a
sentence, as in I said THEY must go! So can grammar, as in It is indeed the
CAT that swallowed the RAT where the phrase it is indeed forces a focus on
CAT. If the two kinds of attention crisscross, we may call the phonological
Wave 27
one emphasis (the RAT) and the other focus (It is indeed the cat). When
whole paragraphs are involved, the focused nuclear beginning sentence
may be called a topic.
4.3. Types of Wave
Even when no special attention is called for, the dynamics of

behavior organizes speech and action into wave units with nuclei and
margins.
We may call a grammar wave one in which there is a central, crucial,
nuclear part of a construction which may or may not be preceded or fol-
lowed by less important elements which in some way modify it. When such
a wave is made up of just one word, the nucleus may be called a stem, and
the margin a prefix or a suffix as in inactive or in singing. A two-word
phrase, smaller than a clause, may also have its head and modifier, as in
my (margin) alligator (nucleus). When two clauses are in a sentence, one
may be independent, nuclear, and the other dependent, as in The doctor
treated the boy (nucleus) after he broke his leg on a skate board (margin).
Similarly, a question-answer relation gives a higher level of nucleus-
margin: Will you come? Perhaps.
A phonological wave can be as short as a single sound or as large as a
lecture. Each may have nucleus and margins. In figure 4.1B, the drawled
pronunciation allows one to hear the slow (premargin) approach to the
center (nucleus) of the separate sounds, and the slow postmargin, a depar-
ture from that central part. In figure 4.3, the entire sentence I don’t want to
is a single, larger phonological wave. The I don’t is premargin, to is post-
margin, and want is the stressed syllable making up the nucleus of the
stress-group wave. The smearing of the English words, in such a situation,
is often the result of pronouncing the premarginal section very fast. The
nucleus is usually pronounced more slowly, and is less likely to be lost or
changed.
A referential wave shows highlighting of reported events, or of argu-
ment, or exposition. Thus a story can be viewed as a wave: a premargin
may set the stage; the plot warms up}; the climax (the nucleus) occurs; the
hero gets out of trouble and lives happily ever after (postmargin). For a
very special kind of story wave, see figure 4.4.
The use of waves in analysis does not eliminate need for the use of parti-
cles. Even a wave, when it is viewed as a unit, may itself be treated as a
particle, just as the particle, when viewed in its physical smearing aspects,
can be viewed as a wave.
“The Christmas Dinner”
Figure 4.4. A play as a standing wave may focus attention on the un-
changing role-structure of a family with a grandmother as premargin,
parents as nucleus, a child as postmargin. These roles and their inter-
relations can be kept constant while different generations grow into and
out of these roles. In Wilder (1963), a scene at a Christmas dinner is kept
constant, while some ninety years pass as a wave: baby in basket turns
into mother sitting at the table; mother gradually drifts into the slot of
grandmother; old ones die and drift off stage.
A pebble dropped into a pond sends out a wave in all directions from the
center. So also, a word may comprise a lexical wave with a central mean-
ing normal to the most frequent or nonspecial set of contexts and a set of
marginal meanings which occur when the central meaning is modifed by
other words in the context. For example, to run is basically (in my intui-
tion) referring to the action of a man (or creature) in his feet moving
rapidly toward or away from some place; but a run in a stocking, or to
run the business, or to run into trouble, or He ran out of words are margi-
Wave 29
nal. In such instances, we may also sometimes speak of the nuclear mean-
ing as normal, and the marginal one as off-norm. Slang, or idioms, may
therefore be treated as off-norm, relative to the usage of the community as a
whole, or frequency of usage, or some criterion of appropriateness as seen
by the community.
An item prominent in a general setting, or in a painting, can be viewed
for some purposes as an instance of a perceptual wave, with focus on the
item or figure as nucleus and on the setting as marginal.
A wave view can even be used to characterize the sweep of scientific his-
tory. The nuclei of a science over time would be periods of relative stability
in its presuppositions, methodology, and objectives. Kuhn (1962) would
call such a nucleus the “continuation of a research tradition,” or “normal
science” (11), or a “paradigm” (23) of a theory (ix). The prenucleus of
such a wave for him could be called “the genesis” or emergence (ix) of a
new theory as it gains its status (23). The postnucleus of one of these waves
would be seen when researchers notice, among other things, “the insuffi-
ciency of methodological directives, by themselves, to dictate a unique sub-
stantive conclusion to many sorts of scientific questions” (3). Kuhn’s own
focus, however, might be said to aim at a discussion not of nucleus,
postmargin, or premargin, but at the juncture in between the waves, the
“transition” (xi) moment itself, or the “extraordinary episodes” which “are
the tradition-shattering complements to the tradition-bound activity of nor-
mal science” (6).
Field
We have discussed the way in which elements can be viewed by

themselves as particles, or as waves smearing into some kind of continuum
whose prominent parts make up nuclei. Now we turn to sets of relation-
ships which occur when units are linked to one another by their presence
in some larger system. A total set of relationships and of units in these rela-
tionships we call a field.
5.1. Unit Determined by Relationship to Context
No item by itself has significance. A unit becomes relevant only

in relation to a context. Outside such a relationship the item will be neces-
sarily uninterpretable by the observer. A simple circle, for example, may be
interpreted by an observer as a numeral, if it is in the context of the number
1,000. On the other hand the circle may be interpreted as a letter of the
alphabet in the context pop—or as eyes, nose, or teeth, or buttons, or a
wheel, in a picture (see fig. 5.1). When, therefore, one asks what is the
circle in one of these circumstances, it seems clear that its “factness” is rel-
ative to the observer.
The observer interprets data relative to context—within a field. In this
sense a datum of sensation “becomes” a “fact” of human significance when
seen relative to its distribution in an environment—relative to a field as
structured in the perception or imagination of some person.
Field 31
A. 1,000
Figure 5.1. Units determined by their relationships to context. A circle is

perceived as a talked-about fact only in relation to a field and to an
observer of that field.
5.2. Sounds in Dimensional Patterns
Symbols representing phonological units of a system can be

placed in rows, such that each sound represented has at least one feature
(or component, property, characteristic) in common; units in a column
also share a feature (see fig. 5.2A). If the symbols for the sounds are re-
moved but the presence of features is shown by arrows (see fig. 5.2B), a
circle around the place formerly occupied by a consonant symbol contains
a point of intersection of two of the arrows. Either the unit can be treated as
merely a “point” at the intersection of features of a field, or the presence of
the units canbe treated as setting up the field by the relations between
them. This kind of display allows us to show that several (intersecting)
characteristics are simultaneously present. In this simple diagram, only
A. With tongue With mouth

With lips tip closing closed at the
closed the mouth soft palate
Stopped air stream

without vocal cord p t k
vibration
With vocal cord

d
vibration
Air escaping from

the nose i
B.
Figure 5.2. A phonological field seen as a pattern of sounds with features

which intersect (A) and as a set of points at the intersection offeatures
(B).
two dimensions are suggested as relevant—but in principle a field of char-

acteristics can be specified.
A vowel system, like a consonantal one, may be viewed as a field. If we
mark the highest point of the tongue on the “pictures” of /i, a, u/ we can
use these points as a crude basis for a three-by-three diagram in which /i/
and /u/ are in the top row, in front and back columns, with /a/ in the lower
row, middle column. The implied field structure can be used to describe
Field 33
other vowel sounds as well (see fig. 5.3). X-ray pictures by various scholars
have shown this scheme to be not completely accurate, but the general
scheme continues to be useful on a practical phonetic basis, as an approx-
imation of the way the sounds can be heard in “perceptual” space.
In baby talk the setting of the grid may as a whole be moved higher in
the mouth, with the result that a smaller part of the mouth cavity is used.
When all of speech is thus transformed, it sounds as if it were coming from
the small mouth cavity of a child. If, on the other hand, the whole vowel
triangle is pushed farther front, the general effect may be more like
Spanish.
Further dimensions than high versus low and front versus back can also
enter into the articulatory field. Work (for example, Stewart 1967, Pike
1967b) on African languages has emphasized that the movement of the
root of the tongue in the throat opening may serve as a further more or less
independent dimension. When the root of the tongue is pushed front, leav-
ing the top of the tongue more or less unchanged (see fig. 5.4A), a rather
hollow quality may result from the greater openness of the throat cavity. A
whole set of sounds in many languages of West Africa may be affected by
such a movement. The closest American English equivalent might be the
difference between the vowel quality of the word feet versus that of fit (/i/
versus /t/); of boot versus put (/u/ versus /u/); of cape versus kept (/e/
versus /€/). Figure 5.4B shows how the vowel chart can be shown with a
third dimension, tongue root position, to suggest this. A simpler chart can
be given by subdividing rows to show this difference (fig. 5.4C).
5.3. Clauses in Dimensional Patterns
In tagmemic theory we have attempted to develop the use of

charts. to show the field structure of clauses (Pike 1962, Pike and Pike
1977 :140—43, 236-37) just as charts have been used for centuries for pho-
nology; each row (or column) indicates a feature shared by that row (or
column). One subset of the clauses of Spanish can be shown in figure 5.5A3
their abstracted features occur in rows and columns. A feature of indepen-
dence contrasts with dependence: Les hacen traversuras ‘They do tricks to
them’; versus porque hacen la presentacion a base de estudio de casos ‘be-
cause they base their presentation on the study of cases’. A feature of in-
transitivity (Without object allowed) contrasts with transitivity (with object
required), and with an equational feature: Hablé con el dueno del café ‘I
spoke to the owner of the cafe’ (intransitive, without direct object); Les
hacen travesuras ‘They do tricks to them’ (transitive, with ‘tricks’ as direct
Front Central Back
Figure 5.3. The highest point of the tongue can be used to set up points
on a grid. Vowels can then be described, crudely, in reference to such a
scheme. In A, the grid is seen superimposed on the mouth. In B, it is
abstracted and labeled. The plan is pedagogically useful. Physical details
must be modified and refined by presentation of X-ray studies.
Field 35
Front Center Back
(of main part of tongue)
; Root front
H
abe Root back
Root front
Root back
Root front
Root back
Figure 5.4. The root of the tongue, pushed front, opens the throat and
sometimes gives a hollow quality to a sound (dotted line in A). Backing
the tongue in the throat may choke the vowels or consonants. This char-
acteristic can be shown by a third dimension of a field display (B), or by
subdividing rows of a two-dimensional display (C).
A. Transitive Intransitive Equative
Independent Independent Independent Independent

Transitive Intransitive Equative
Dependent Dependent Dependent Dependent

Transitive Intransitive Equative
B. Primary (Basic) Secondary (Derived)
Transitive Intransitive Equative Passive Impersonal Descriptive
Independent
Dependent
Figure 5.5. Dimensional matrix of Spanish clauses. The units may them-
selves be presented, with certain features important to their contrasts,
labeled A; or an extra dimension (here, primary versus secondary—that
is, basic versus derived) may be presented as some kind of subdivision of
a two-dimensional display (B, after Brend 1968).
object); Eso fué el regreso de Ocotlan para acd “That was the return from
Ocotlan to here’ (equative).
Various added dimensions of contrast can be handled either in three-
dimensional display or as a subdivided two-dimensional one (fig. 5.5B).
5.4. A Poem as a Field Structure
It is intriguing, also, to study much larger units than the sen-

tence in reference to field structure. Suppose, for example, that we take
Robert Browning’s poem “Memorabilia” (see 1895, and fig. 5.6). One
could choose to interpret the poem as representing a kind of field of mis-
cellaneous particles within the memory. Within this field two happenings
are described: The first deals with Browning’s meeting of a man who met
Shelley. The shock of this caused Browning to pale, since even such an indi-
rect contact with the great man Shelley was staggering. The second event
deals with the finding of a feather. In some sense each event seems to be
perceived as abstracted from experience, as if without relationship to time
before or after. Therefore they seem strange, because unchanging, and so
somehow eternally new, like some carved figure frozen on a Grecian urn.
Art and emotion share these characteristics. I well remember, myself,
when a friend, after taking his three-month-old baby to the hospital for
Field 37
A. Memorabilia
Ah, did you once see Shelley plain,

And did he stop and speak to you,
And did you speak to him again?
How strange it seems and new!
But you were living before that,
And also you are living after;
And the memory I started at—
My starting moves your laughter!
I crossed a moor, with a name of its own
And a certain use in the world no doubt,
Yet a hand’s-breadth of it shines alone
*Mid the blank miles round about:
For there I picked up on the heather,

And there I put inside my breast
A moulted feather, an eagle-feather!
Well, I forget the rest.
Robert Browning
B. Field-Memory
Particle Wave
Event I: Meeting a man I II

who met Shelley
Event II: Finding a feather IV Ill
Figure 5.6. Large units may show field structure. The Roman numerals
in the cells (B) represent the stanzas of the poem (A) by Browning. As
particles (1, IV) the events seem isolated and unchanging, therefore
strange and new. As waves (II, III) the events are known to have been tied
to a before and an after, with movement and integration into the real
world as a whole. (From Pike 1965:291.)
emergency brain surgery, said, “As the elevator door closed behind the cot
with my child on it, time stood still.” The picture congealed.
Under appropriate circumstances, the same event can be viewed as par-
ticle, or as wave, or as field—or as a combination of the three, with one
more prominent than the other in the perception of speaker or of hearer. A
person’s emotional state or his immediate or remote experience affects his
conscious or unconscious choice of priorities to be given to one of the three.
And his language often reveals this choice.
5.5. Verbal Cues
Verbal cues that an observer has chosen to adopt (partially at

least) a field view may often be seen through his use of words or phrases
such as pattern, structure, arrangement, blueprint, plan, network, system,
relationship. Some words can be grouped together in sets of threes, sug-
gesting (for me) particle, wave, and field respectively:
static, dynamic, relational

item, process, relation
point, line, space
list, rule, pattern
identifying, making, organizing
abstraction, fusion, intersection
repetition, change, integration
But there is an analytical awkwardness about observer reality here: at

any particular moment there can be an overlap or an indeterminacy
among his perspectives. Whereas an observer is sometimes likely (in my
judgment) to think of a particular unit as a particle, he may at other times
find himself thinking of the same unit as changing in a wavelike fashion,
and at still other times as a point in a contextual field within which it gets
its meaning.
An observer may shift his attention from an item to its background,
thereby treating the background as a whole as itself aunit. One may note,
from an airplane at night, the wave character of moving or changing lights
in a city below, then focus on the general pattern of arrangement of the
lights, and then change once more to admire the city as a whole. Focus has
shifted from wave to field to particle perspective. Or one can combine
them, by saying, “Note the beautiful changing pattern of lights of that city
as a whole.”
Ht
The Unit
a -
a 7 »mall? “ a . _ a
nab ie
Introduction to Part II
In Part I, we talked about ways in which we can look at things

or experience: The world can be viewed as if it were made up of particles,
of waves, or of field. Perspective has been the focus of our attention, with
analytical focus itself coming under focused attention. In Part II we go over
much of the same ground, but with the observer in the background and the
observed data themselves in the foreground.
Chapter 6 tells us how we can recognize a unit when we find it. This is
important. If we could not recognize units—things, people, actions—how
could we live as social human beings? Chapter 7 discusses the fact that a
unit may be found or perceived in a variety of forms. Chapter 8 shows that
a unit to be significant must be appropriate to certain times, places, or
situations.
Contrast and Identification
6.1. Recognition via Contrast

and Similarity, within Field
Recognition of a unit involves a curious difficulty: We do not

know what something is unless we know something about what it is not.
Recognition is, in part, negative. In part, it deals with the contrast between
units, rather than with the identification of isolated units. One does not
know what a horse is if he cannot tell it from a cow. One does not know
Susie James if one cannot tell her from Susie Pennyfeather (see figure 6.1).
The same components which help us to see that one unit differs from
another also help us to recognize the unit when it is no longer close to the
one which we separated it from. If, for example, we treat seal and zeal as
two different words, with the /s/ and /z/ being two different sounds as
components of the two words, then we can use either /s/ or /z/ to recognize
words where they do not pair up in contrasts—as zebra is clearly recog-
nizable even when no *sebra can be found (with the asterisk before the
word meaning that no such instance is known). Features which are con-
trastive in relation to some contexts are identificational in others.
When we go around trying to find units, we could get useless answers if
we were to ask only “What is X not like?” If we want to know, for example,
what a door is, we hardly feel enlightened if we are told only that it is nei-
ther a star nor a pebble nor a skunk. A component of a contrast seems to be
interesting only if we feel that somehow the two units being contrasted are
similar. Random contrasts, without reference to similarity, interest us only
7
Contrast and Identification 43
Long-nosed Joey
“Are you Joey, or the other boy? Oh, I see. You’re not Joey —
you don’t have a big nose.”
Figure 6.1. Contrastive components which have served to differentiate

pairs of units can serve also to identify them when apart from one
another.
occasionally (as, for example, when we want to talk about randomness it-
self—comparing cabbages, kings, and pebbles).
But what do we mean by similar? Units will appear to be similar only if
they share some component within the same universe of discourse. Joey
and Billy of Seattle can be compared easily, since they share maleness,
youth, and the same community. Concepts of contrast and similarity, then,
circle back.on the concepts of field and observer interest.
6.2. Matching for Contrast
Many of the most interesting problems of contrast and dif-

ference occur when the items to be compared are very similar indeed. Oc-
casionally, for example, my wife has wanted me to buy her a spool of
thread. Until she has given me close instructions, it has turned out cata-
strophically; thread color and cloth color did not match. She taught me to
44 The Unit
take along a slip of cloth, but comparing spool with cloth still would not
serve. She instructed me to unwrap a bit of the thread, put it on the cloth,
and hold it in natural light, moving it around a little bit; then, and only
then, I began to get the kind of match she needed. Colors can be told apart
better when they are matched side by side in proximity—even, one might
say, in the same tiny, temporary “universe of discourse” of color and light
and setting.
Similar considerations apply to sound. We hear differences best when
we hear them in a near identical context. Good procedure for phonological
analysis encourages us to listen to the respective sounds in contexts that are
as much alike as possible. Often this means in word pairs which are other-
wise identical (as seal and zeal, which are called minimal pairs because
the difference is restricted to the sounds (/s/ and /z/) under attention). In
Mazatec, for example, there are words which differ by the presence or ab-
sence of a difficult-to-hear /h/ before or after the consonant, as in tho ‘gun,’
-hto ‘rotten,’ or its absence in to ‘fish’ (I omit the tones).
The beginner’s ear may play tricks on him and refuse to listen at all, and
“tell” him that the words sound the same. Under such circumstances, it is
very helpful if the meaning is different. The human being learns most easily
to pick up contrasts between elements which are highly significant to him.
If, for example, two similar words are such that one is obscene and the
other is not, it is astonishing how much faster he learns to hear the relevant
contrast of sound than he does if he has no social problem growing out of
error.
Here again there is an observer component in linguistics. Seeing or hear-
ing or learning is facilitated when the observer has a stake in the outcome.
We are born to talk with meanings; we are not built to talk only nonsense.
Even our analytical equipment works best when these conditions are taken
into account. Language is not merely a set of unrelated sounds, clauses,
rules, and meanings; it is a total coherent system of these integrating with
each other, and with behavior, context, universe of discourse, and observer
perspective. It is a form-meaning composite (see Chapter 12). Units can
contrast in their respective forms, in their meanings, or in both.
In studying an unknown language, the linguist must crack its code. He
must find and learn to recognize and use the units of speech which allow
the native speaker to understand and be understood, to act and to react
appropriately; he must learn to recognize and utilize units from an in-
sider’s view.
When such units are sounds of a certain language system, they are called
phonemes; each of them is phonemic in relation to that particular lan-
guage. Generalizing from phonemics, I coined the term emic in 1954 (Pike
1967a: chap. 2) to represent any such unit of language, from any level of
any hierarchy (see below, chapters 9-11), or from any nonverbal system.
(Variants of emic units will be discussed in Chapter 7.)
6.3. Contrast in Contextual Frames
If we think of the different possibilities of pronunciation of the

third syllable in figure 6.2 as different gates through which the speaker can
walk, then we can think of the process of speaking as the process of picking
one’s way by some path through one gate after another. In figure 6.2, the
path goes through mid (tone 2) in the first syllable, then through high (1),
through high again in the third syllable, neutral (3), low (4), mid (2). But
in the third syllable the alternate potential choices which were not taken
are shown as dotted, leaving the actual chosen path as a solid line. When
the full set of choices is in fact available, the analyst can compare the con-
trastive levels of the tones of these items of the substitution set with each
other as well as with that of the high frame tone of the second syllable (for
this technology, see Pike 1948).
The choices of the substitution list are independent of one another (any
one of them can be chosen at that place); they are relevantly different from
one another (they contrast with each other at that place and may—or may
Mid:
Neutral:,
Low: '
Figure 6.2. A frame technique chooses a sentence or phrase in which

words can bé*successively replaced in one slot. The number of alternate
contrastive pitch paths in the substitution slot often represents the po-
tential of the tone system at that point. A high, unchanging frame tone is
especially useful. (From Pike 1952, fig. 6.)
46 The Unit
not—make a difference of meaning when used); they are consistently dif-

ferent in this context (there is no haphazard—free—variation between
them at this point such that one overlaps in pronunciation with another).
That is, they are emically different.
The lack of variability would seem obvious in a situation involving gates
through fences. But in the use of pitches of the voice, that is not true. So we
Figure 6.3. Tones may have variants conditioned by mood of the speaker.
Let us suppose that we have two tones, represented as weights sus-
pended between springs (A). When a person speaks emphatically, the
tones may be farther apart (B); a quiet or phlegmatic mood (C) may
move the tones closer together. In addition they may jiggle a bit, in free
variation around a norm. (From Pike 1952, fig. 1.)
use a supplementary illustration (fig. 6.3) of variation of tone (anticipating

more theoretical discussion of variation in Chapter 7). Under excitement,
the whole frame-cum-tone may be raised; or its tones may be stretched fur-
ther apart (fig. 6.3B); or in hushed speech around a camp fire they may be
substantially lowered (fig. 6.3C). Yet in each such instance, they may con-
tinue to be consistently different from one another.
(Other problems, not under attention here, may give further headaches
to the analyst: different consonants or vowels or a different number or kind
of syllables of the words of the list may also cause changes; and lexical re-
placements in the frame itself may cause sharp differences of replace-
ment—morphophonemic changes—of one emic tone by another in words
on either side of it. An unrecognized replacement of a tone of the frame
itself, furthermore, may cause errors concerning the perceived pitch of the
words of the list. As a partial defense against unrecognized but disruptive
changes, one is well advised [especially in languages of Latin America] to
seek for a frame in which some one of its tones is always as high as or
higher than any word of the substitution list. In this circumstance, any
word of the list which is at the same level as that high frame tone can itself
be considered emically high.)
In figure 6.4A is given a list of utterances from the Mazatec of Huautla,
Mexico, in which ‘causative’ is si’; in 6.4B ‘to place, or to repeat, or to
place repeatedly’ is vse’; in 6.4C ‘continuative’ is ti’. In each of these groups
the substitution list contains words with the respective tones from 1 to 4. By
putting all of the sets together, one can see that there is a random relation of
the consonants starting the words of the substitution list, so that those con-
sonants do not destroy the tone contrast (by implying that the pitch dif-
ference is conditioned by the consonant). The pitch contrasts are therefore
independently, consistently different—that is, emically contrastive.
Consonants—and vowels—may be similarly contrasted with one an-
other in frames. Usually, however, there is less variability in their pronun-
ciation growing out of moods or other contexts.
6.4. Contrast Seen via Matrices
We have already seen that sounds can be lined up in charts, or

matrices (figs. 5.2A, 5.4C). These show the relation more than with just
one sound, or than with one set of sounds in a particular frame. Rather,
they line up the contrasts of a subsystem of sounds—which may itself be
treated as an emic unit. A three-dimensional matrix of sounds has been
48 The Unit
seen in figure 5.4B. These charts may be labeled as component-times-com-

ponent displays (that is, with components labeling the rows and the col-
umns). The component-times-component presentation was used again in
figure 5.5, but for a display of some Spanish clauses rather than for sounds.
Other kinds of matrices may show advantageously other features of a
system; among these is a display of units as columns, components as rows,
and an indication of a component’s presence (by a plus sign) or absence
(by a minus sign) in the cells. In figure 6.5 a still different type of matrix is
shown for a display of clause types as rows, the naming of classes of tag-
memes as columns, but the naming of the particular tagmeme of that class
in the appropriate cell. For Zapotec, of Mexico (Pickett 1960: 35), this helps
A1. si’ hnti’ ‘[he] makes [something] dirty’

si*S$i? ‘[he] makes it dry’
si’Ski® ‘[he] gives medicine to’
si'Ski* ‘[he] counts’
A2. si‘he’* ‘{he] asks for’

si*te? ‘{he] makes [something] dance’
si‘he* ‘[he] fattens up [something]’
si'kao* ‘{he] touches’
si‘oya* ‘[he] makes string’
A3. si'Sa? ‘[he] works’

si‘jnta’ ‘(The] borrows’
si‘¢éa* ‘[he] looses’
si‘kao* ‘{he] touches’
B. vee*Cchi* ‘Che] pays’

voe'Si? ‘{he] dries [something]
vee’ thi? ‘[he] spins [something]’
voeSki* ‘[he] counts’
C. ti‘ncha‘ntze*_—_ ‘‘[he] is talking persuasively’

ti‘ncha‘*nthai’ ‘[he] is talking in defense of?
ti‘ncha‘to® ‘[he] is speaking as he passes by’
ti‘n¢éha*kao* ‘{he] is talking with’
Figure 6.4. Tone contrast in frames may help in the analysis of the num-
ber of emic tone levels. A high frame which is high and unchanging in
some contexts, as in the Mazatec of Huautla, Mexico, allows for recogni-
tion of contrast (data from Eunice V. Pike 1958:95—165).
Zapotec
ee
:5 E
as!
S90) S 3 ~ 5 ~ +
= ¢
ne} E
5x =
erin
ves
4) eaneae
One
p
a fs
bee eo
Ab eee See. etaer
Intransitive + IntrDeclPred + DepS +IndS — —
declarative
Transitive + TranDeclPred + DepS +IndS +0 —
declarative
Personal referent + PRefDeclPred - DepSey-IndSs as 0) = PRef
declarative
Figure 6.5. A unit-times-component matrix for clauses of Zapotec (Mex-

ico). Each clause differs from each of the others by a difference in the
predicate. There is also the obligatory presence or absence of object, and
optional presence or obligatory absence of personal referent.
to show the contrastive structure of the clause types, in terms of labeled

elements in the sequence in which they occur in their respective clauses.
Notice, in this figure, that there is an obligatory predicate for each, but that
each predicate is different from the others, and that there is the obligatory
occurrence of an object in the second and third rows, but with an optional
referent in the third only.
6.5. Contrast in Verbal Meaning and in Behavior
We have illustrated contrast between sounds and contrast be-

tween clauses. Contrasts between meanings of lexical units can also be
seen through matrix displays. For words-in-cells (comparable to the
sounds-in-cells in fig. 5.2), note figure 6.6, where the components male
versus female, and royal persons (reigning versus nonreigning) versus ani-
mal are abstracted from the data in the cells and used to label the rows and
columns of the matrix.
Sometimes words are pronounced alike, but have different meanings.
The contrast between their meanings sometimes is clear because of the dif-
ferent contexts in which the words with these particular meanings are al-
lowed to occur. The word big’ has several lexical meanings which are not
50 The Unit
Male Female
Royal Reigning king queen

Nonreigning prince princess
Animal tiger tigress

Figure 6.6. Matrices can display meanings of lexicon, with units ar-
ranged and labeled according to shared features.
How big is X?,

or X is very big,
big X X is big or X is bigger.
big, a big, rock The rock is big,. How big, is the rock?
big, my big, sister *My sister is big,. *How big, is my sister?
big, a big, fool *The fool is big,. How big, a fool is he?
Figure 6.7. Words may have contrastive senses. The word big can refer to
physical size (1), to elder relation (2), or to the degree to which a charac-
teristic is present (3). Inappropriate usages are marked with asterisks.
(The matrix is adapted from Lamb 1964: 74.)
easy for the beginner to differentiate. If we try to use the word big in vari-
ous different contexts, however, they fit some of the contexts correctly, but
not others. One may say It’s a big rock, or The rock is big, or Is the rock
very big?, or How big is the rock?, or The rock is bigger. But in the sense
big of my big’ sister, one cannot say My sister is bigger or How big is my
sister? In figure 6.7 these differences are shown in a matrix.
It is not only in language, however, that contrast is relevant to meaning,
purpose, or culture. A person, in English, may signal Good-bye! by waving
the hand downward several times. (Spanish speakers sometimes misin-
terpret the English gesture as meaning come.) To signal the contrastive
meaning come in English, one is likely to turn the palm of the hand up-
ward, and move the index finger upward and inward.
This pair of English gestures is part of a whole set of gestures (culturally
conditioned manners of eating, sitting, or even smiling) which make up
another universe of discourse with its contrastive features, varying accord-
ing to language or geographical area. Some years ago in Australia, for ex-
ample, I amused myself by trying to tell, by looking at advertisements in the
locally published journals, whether or not the photographs were taken of
Australian girls or of American girls. Often I could be quite sure. In those

instances, the smiles were radically different; American girls smiled with
the corners of the mouth a bit open, whereas the Australian girls frequently
smiled with the corners of the lips closed. When I called this to the atten-
tion of the Australians, one of them replied, “Yes, those American pictures
look like toothpaste advertisements!”
Variation
Normal social behavior requires that we be able to recognize

identities in spite of change. Unless we can do so, there can be no human
society as we know it (see fig. 7.1, and compare fig. 2.7).
7.1. Referential Identity in Spite of Change
The problem of identity within change is not new. Heraclitus

said: “Into the same river we both step and do not step. We both are and
are not” (fragment 81, in Patrick 1889). Or one may fail to find identity at
all: “Into the same river you could not step twice, for other [and still other]
waters are flowing” (fragment 41). If the river is made up of molecules of
water, then—since the molecules have changed—you can never bathe in
the same river again; the molecules will have flowed by you.
In this chapter we are especially interested in those features of a unit
which may change without causing the loss of recognizability of the unit.
The appropriate observer readily recognizes that certain changes do not
affect deeper identities.
The pattern as a whole can change subtly and gradually over a period of
time. The child grows into a man (compare fig. 2.7). From stage to stage of
growth and change the parents have no difficulty recognizing him. But for
me a picture of a child a few months old may in some instances not be at all
distinguishable from that of his brother when he was the same age.
The outside analyst—an observer alien to a native system—may have a
great deal of difficulty in determining precisely those features which are
Variation 53
Yesterday
&)
He: I can’t take you to the party.

She: Why not?
He: You're not the same girl I promised
— you’re wearing
your hair in a bun!
Figure 7.1. Refusal to acknowledge identity within variability would stop

all normal behavior.
diagnostic for recognizing minimum structurally and behaviorally relevant

segments of sound (phonemes) which to the native speaker—the inside ob-
server—are reacted to as in some way unchanged entities. Nor can the out-
side analyst expect that the native actor can tell him what the specific cues
are which allow him to recognize the unit.
How well J remember once when my own boy, at about six years of age,
was lost for a time. Ditches were being dug all around our dormitory to put
in telephone lines. Dusk came, police were called, a search was made.
“What does he look like?” “Well, he’s tow-headed, has a dimple in his chin,
and was last seen wearing a pair of overalls.” Since he had no outstanding
physical characteristics, it was impossible for us, his own parents, to give a
description which was very helpful. Changeable bits (the overalls) were
included along with more permanent components (his light-colored hair—
which, however, has darkened in the years since then). The overall pat-
tern, including many unanalyzed bits, carried the total signal of identity—
for his parents—in some kind of known field patterns beyond words.
54 The Unit
Even persistent obvious cues have a range of change. There is no set of

unchanging, precisely same, distinctive components in any language, or
across languages. The analyst may act—falsely—on the assumption that
once enough cues have been identified to differentiate sounds, these alone
are the cues actually used by the native speaker. But the speaker has—and
uses—more cues than he needs in an idealized context. Some variable ones
are redundant from the viewpoint of minimum descriptive elegance, but
important to communication—under noisy conditions, for example. The
analyst may desire a succinct, traditional logical definition which aims to
limit the data in a definition to a minimum set of criteria which will place a
unit in a taxonomic classification. This kind of definition refers (a) to ele-
ments which differentiate a unit from similar units, and (b) to elements
which place it in a taxonomic hierarchy: for example, a chair differs from a
table, but both belong to the class furniture. Tagmemic definition—or de-
scription, if one prefers the term—differs from classical definition by insist-
ing on the inclusion of further defining characteristics: (c) the range of
variability of an item, as treated in this chapter, plus (d) the appropriate
range of occurrence of a unit as it is distributed in system (or field) rather
than merely its distribution in class or sequence (see chapters 5, 8, and 14).
7.2. Changes in Pronunciation
In listening to his own English, the beginner, furthermore, may

find this chapter difficult to understand or to follow, precisely because he
doesn’t notice the changes in his own native speech. From this point of
view one can say that he is reacting to the structure of his own language as
if it were a kind of model or theory in which (most of the time) he leaves
out wisely any major attention to minor variants (see discussion of models
in section 1.2). If he were constantly focusing attention on details irrelevant
to signaling contrastively the normal meaning or purpose of behavior, he
would never be able to concentrate on the crucial intent of its messages. Yet
at the same time—and in a paradoxical relation to the variants—he must
be able at any moment to pick up cues from minor characteristics of units if
the major components get lost or covered (as under noise). In addition, the
variants can simultaneously be crucial signals of style, or attitude, which
are important from a social viewpoint. If, for example, Joe says to Sue,
“That’s a beautiful hat you’ve got on!,” but says it with a contemptuous
growl, she will understand the deprecative meaning as overriding the (op-
posite) lexical one.
Variation can affect not only the pronunciation of the individual pho-
Variation 55
nemes, but can lead to addition, replacement, or deletion of some of the

sounds in the words—that is, to morphophonemic changes. The command
Come here! may become C’m’ere! in some rapid speech. Many Ameri-
can speakers whom I have checked pronounce Ed had edited it as
/ededédededut/ (with /¢/ the vowel of dead, /./ the vowel of bit, provided
they pronounce the sentence fast, without pause, and with stress only on
the third syllable). Yet because of the morphophonemic changes, the sen-
tence pronounced in isolation is often unintelligible to these same speakers.
Context (universe of discourse, field) is relevant to the set of mind within
which understanding takes place. In part, one hears what one expects to
hear. In another kind of morphophonemics words may share sounds, as
/Z/ may be shared by as you in rapid pronunciation.
7.3. Changes in Grammar
Units of grammar may have free or conditioned variants just as

can phonemes and morphemes. Free variation in the manifestation of the
subject of a clause tagmeme can be seen in the replacement of one word by
another in the same slot. In The tiger swallowed the boy and The snake
swallowed the boy, both the tiger and the snake fill the subject slot. Since
the grammatical relationship in these instances is the same, the change is a
kind of free variation.
Conditioned variation can be seen in the predicate: Compare The tiger
sees the canary with The tigers see the canary. The change from The tiger
to the tigers forces the change from sees to see; the free change from a sin-
gular to a plural subject forces a conditioned change from singular to plu-
ral predicate. That is, the presence of a freely chosen plural variant of sub-
ject tagmeme conditions the presence of the plural variant of the predicate.
Variation can also affect the placement of parts of grammatical construc-
tions (for example, in clauses). The word yesterday can either precede or
follow I came from Detroit. Viewed more or less in isolation, the place-
ment of the time tagmeme is free to occur either early or late in the clause.
In a paragraph context, there are instances in which having it early gives it
prominence, or focus. That is, the placement choice is locally free, but
focus- or style- or discourse-conditioned. Compare the sentence I came
from Chicago today but yesterday I came from Detroit with I came from
Chicago today but I came from Detroit yesterday. Compare also He gave
me the book with He gave the book to me (but not, comfortably, *He gave
the great big book I used to own in Ann Arbor to me, since we prefer to
have the long element at the end).
56 The Unit
Constructions sometimes have abbreviated variants also: She can sing

better than I can sing, or She can sing better than I can, or She can sing
better than I. There are also restrictions in the utilization of particular lexi-
cal items with particular grammatical ones. Here the lexical and gram-
matical systems affect one another. We can say You know the way, or Thou
knowest the way (in elevated poetry or prayer) but not the mixture *You
knowest the way. The appropriate occurrence of units involves the whole
pattern of a language—not merely its isolated bits.
Sometimes it is convenient to view a subsection of a field structure as
itself a single unit which can have variants. In figure 7.2A a matrix is given
showing certain of the subject suffixes of some Fore (New Guinea) verbs
(from Pike 1963:8-10). No clear set of simple particles can be found
which signal the meanings for singular and plural, although the composite
forms with first and second person are clearly different from each other.
Thus /w/ is singular with first person and plural with second person,
whereas /n/ is plural with first person, but singular with second person. If,
now, to the words containing forms in figure 7.2A a further suffix /n/
meaning ‘emphasis’ has been added, every form of figure 7.2A changes
radically into the form seen in 7.2B. Yet the same pattern of crisscross is
retained. For some purposes it is useful to say that the entire pattern (or
field) has been conditioned by the /n/, rather than set up separate rules for
each separate item or part. (At some other times, when the separate parts
need to be in view, an alternate description may be preferred. In tagmemic
theory, a field view of conditioning does not exclude a particle view of the
same data for a different purpose.)
It would be convenient to have some one term which refers to variants of
all kinds of units (both phonological and grammatical). A solution is to call
a variant of any emic unit an allo-unit (from a Greek work meaning
‘other’). Thus we can have allophones of phonemes; allomorphs of mor-
phemes, allomatrices, and so on.
A. First Second B. First Second
Singular (u)w (aa)n Singular (0) # (aa)mpe
Plural (u)n (aa) w Plural (o)mpe (aa)#
Figure 7.2. Conditioned variants of a field structure can be seen in Fore

(New Guinea) subject suffixes. A crisscross occurring between /w/ and /n/
in A is paralleled by crisscrossing zero (#) and /mpe/ in B. The change
from A as a whole to B occurs before an emphatic suffix /n/.
Variation 57
7.4. Indeterminate Segmentation
We have implied that the separation of segments in a sequence

is at times difficult since a small element which lies between two larger
units can be partially or completely shared by each of them when they are
partly fused (see fig. 4.2 and as you in section 7.2). In such circumstances
a person can know that two units are present in a word even though he
may not know where the two units respectively end and begin. Sometimes,
that is, segmentation is indeterminate—as can be seen in the following
illustration.
Myra Barnard and Janette Forster, members of the Summer Institute of
Linguistics, were working as a research team in a remote aboriginal vil-
lage, where the planes of the institute shipped them supplies. It became
awkward to keep the accounts (as in dividing up charges for a half kilo of
carrots to each member of the pair) every time a joint shipment was sent.
The finance office, therefore, encouraged these girls (and other pairs of
field linguists) to adopt fused names. The Barnard-Forster combine was
called Barnster (see fig. 7.3A).
At their staging center, furthermore, each pair shared a mailbox; and
pairs often got social invitations addressed to the name combine. One day I
was invited to lunch with Barnster. Mary Jane Gardner also happened to be
living in their house, sharing part of their center expenses but not their
tribal expenses. To the three of them there was now applied a triply fused
name: Garnster. I asked them to tell me which part of the name belonged
to which girl. With Barnster it appeared easy. Barn- represented Bar-
nard; -ster represented Forster. This would represent a simple particle
view, with obvious unambiguous breaks between the parts; this kind of
segmentation allows the parts to be easily identified and entered into a dic-
tionary or morpheme listings.
But with Garnster, many difficulties arose. Gardner claimed the Garn-
as all her own—providing a particle segmentation, to be sure, but leaving
poor Barnard as a zero feature (#), unmanifested (as in 7.3C1). Barnard
protested—so Gardner allowed her to claim simultaneously the -n- (C2);
this shared -n- would represent the border fusion in wave view. But Bar-
nard still revolted, demanding instead that -arn- be all hers, with Gardner
being left only with G- —from a particle view, but with a different segmen-
tation (B, C3), perhaps a compromise there would have been C4 or even
C5. Thus far, each of these (apart from zero and fusion) preserves the con-
tiguity of the parts. In C6, however, Forster is robbed of her -er by the
grasping of Gardner who adds it discontinuously to her earlier-in-the-word
58 The Unit
A. Barnster = Barnard + Forster
The garn
GARNSTER caren rel
t)
No! Only
the g is
yours.
C. Garnster = Gardner + Barnard + Forster
= Gar------+---n
= Garn-er-+ ----n---- +----st--
+ -a-n----+--r-ster
Figure 7.3. Segmentation may be indeterminate even when the presence of

units is certain. In a remote village a pair of girls shared transportation
services (A), and a joint fused name for the bill. Elsewhere a third girl (B)
joined the combine. Alternate possible segmentations of the fused name
(C) point up the fact that variation brought on by fusion may lead to
arbitrariness of segmentation. (From Pike 1964:85.)
Variation 59
material. In reverse, C7 transfers the -r- of the first syllable analytically

(not in pronunciation) over to Forster.
A particle view demands segmentation, even if it be arbitrary, acting as
ifit were possible to make a clean-cut segmentation even when in fact it is
not. A wave view asserts the presence of units, without needing specific
identification of border points (or arbitrary pretense of such points) for its
kind of description of the fused unit-as-a-whole. When variation arises
from fusion of segments, treatment by the two perspectives differs radically:
choice between the two depends on the varying needs of the analyst. It is
the observer perspective or judgment, not the data, which is variable here.
Distribution
8.1. Distribution of Units as Relevant

to Their Identity
Some place must be the point of origin for the coordinates

which allow one to identify oneself in a place in the larger world. Some-
thing must tell us where we are, beyond reference to our immediate en-
vironment. Reference to successively larger patterns of occurrence, to a
larger universe of discourse, is necessary if one is to know the significance
of a person, a thing, or a word. Knowing that I’m behind my nose does not
tell me how to find the way to town. I need to be oriented in reference to a
larger context, to a universe—or to a discourse. The I of a discourse (in
relation to you) suggests that the speaker knows something about his social
relations. But unless one goes beyond oneself and his neighbor to a hier-
archically ordered outside world, he is in some sense lost within himself.
Even the schoolboy senses this, and can get excited when it dawns on
him that he fits a slot in a larger structured world. How well I remember
sitting at a desk in a one-room red brick schoolhouse, when I was in the
third grade. I wrote out my address as being “Woodstock, Connecticut,
USA, World.” I received a second stage of pleasure many years later when I
had called to my attention that Joyce (1968:15) said of one of his charac-
ters, “He turned to the flyleaf of the geography and read what he had writ-
ten there: himself, his name, and where he was” (see our fig. 8.1).
The same problem of identity and belonging affects adults who are just
becoming aware of a larger world beyond their province. Some years ago I
was talking with teachers in a very remote nook of the world, where the
Distribution 61
Stephen Dedalus
Class of Elements
Clongowes Wood College
Sallins
County Kildare
Treland
Europe
The World
The Universe
Figure 8.1. The school boy relishes an address (Joyce’s Portrait of the
Artist as a Young Man) which places him in a context. Seeing himself
thus distributed into the world, his self-awareness is given sharper
focus.
students had traveled no more than a few miles in their entire lives. Yet the
teachers had to prepare them for life in a world of threatening war and of
United Nations debate. Only a few months were available beyond the fifth
grade to get the students ready to teach others. What could then be said to
help them understand their place in the world? Perhaps:
See how the brook winds down through this little valley? Now you
go over the mountain. There is another valley—(Yes, we know
that—one of our uncles wandered over there once and was killed.)
Well, ifvou go into another valley, and another one, and another
one, you come to some water with salt in it. There you may see a
house that floats and you can push with a motor. (Oh, like the one
we hear in the airplane going overhead.) After the house swims that
way as far as you have walked, then as much again, as many times
as there are days in a moon, you come to another place where there
are people and valleys. That is just the beginning. . . .
62 The Unit
And so on. In teaching the isolated, they had to teach about distribution.
Knowledge of identity includes knowledge of relevant distribution.
8.2. As a Member of a Substitution Class
A unit, whether a person or word, is in part characterized by its

membership in a class of replaceable units which may appropriately occur
in the same place(s) in a particular kind of structure. The set comprises a
distribution class of units. For example, kitten, dog, train, bicycle are
members of a (sub)class of nouns which can appropriately substitute one
for another as head of a noun phrase; compare the phrase a large kitten
with a large dog, a large train, a large bicycle. These phrases, as wholes,
are themselves members of a higher-level class of such phrase units which
may occur appropriately in the object slot of an English transitive clause
such as Joe saw a large kitten. (For tones in a substitution list in a frame,
see fig. 6.4.)
8.3. As Part of a Structural Sequence
A word or the class of words of which it is a member is in part,

therefore, characterized or identified by the structured sequences (gram-
matical constructions, syntactic units) of which it is a part. That is, a unit
is in part defined by the constructions in which it occurs, in addition to the
class of items of which it is an appropriate substitute member within such
structured sequences. The nouns just listed are in part recognized by the
fact that they can occur in a noun phrase, preceded in that phrase by an
article (the kitten), or an adjective (the lovely kitten), or followed by a
modifying phrase (the kitten I used to own).
Similarly, a person, like a word in a grammatical construction, has a
place in a particular social structure which in part identifies him by pro-
viding for him certain social roles—say as a son of an elderly father and
also simultaneously as father of his own young children, or as the head of
the house. In turn, he himself comprises part of that larger kinship struc-
ture, just as the noun phrase does not exist without words.
In pronunciation, also, the place where sounds may occur in a sequence
of sounds is important in specifying one of the crucial characteristics of that
sound. For /h/, in English, it is not enough to say that the air comes out of
the mouth without appreciable friction at any one point, without the vocal
cords vibrating and with the passage to the nasal cavity closed. It is also
Distribution 63
A. English: /glunpst strimz/ from [He] glimpst streams [in the val-
ley below]; but not */mpststrop/as a single word.
B. Seri (Mexico): /kaafsx/ ‘to go fast’, and /ptkamn/ ‘lobster’, but not:
*/ptkaaf$x/; /kiaai/ ‘to cost’, but not: */ kiaaoeoap/.
Figure 8.2. Various constraints limit the appropriateness of the occur-

rence of a consonant or vowel in types of sequences. English has
/mpststr/ (A), crossing the word barrier, but the sequence does not occur
at the beginning of English words. In Seri (Mexico, based on Moser and
Moser 1965) one can have large consonant clusters (or large vowel clus-
ters) (B), but not both together in a single word. (If we assign a value of
two to each vowel, and one to each consonant, no syllable adds up to
more than ten or so units of value.)
important to specify that the /h/ may come at the beginning of words but
not at the end of them. For the -ng of hang the reverse is true. One can be
startled to see how queer it sounds to pronounce hang backwards (as ng-a-
h). Often it is as difficult to pronounce familiar sounds in strange environ-
ments as it is to pronounce strange sounds in any environment. Spanish
speakers have a difficult time with words like school; they do not have a
consonant cluster of /s/ plus /k/ at the beginning of words. They may pro-
nounce the word with a distortion, putting a vowel before the /s/, that is,
eschool /eskul/.
Each language has its own built-in limits to the number of vowels or con-
sonants it can have together. English can have a monster like /mpststr/—
but only crossing a border between words, and not at their beginning (see
figure 8.2).
Whenever,in phonology or in grammar, two languages differ in the dis-
tribution of their sounds or tagmemes, the speaker of the one may find it
awkward to acquire the structure of the other. Problems of language learn-
ing can be expected where differences of distribution of otherwise similar
units are found. The beginner needs drills to teach him not only new
sounds but old sounds in new combinations. This is another way of saying
that he doesn’t really know a sound until he knows its distribution and
can control it in all slots appropriate to it in a particular language.
Various kinds of elements control the distribution of behavioral units.
Linguistic controls, for example, may involve lexical, phonological, gram-
matical, or referential elements—including style, or coherence within a so-
cial universe of discourse. The cluster shm /Sm/, for example, occurs in
English only in special lexical items like schmoo.
64 The Unit
Grammatical control of the structure of consonant clusters can be seen in

that certain of them occur only across morpheme boundaries in which par-
ticular lexical items are involved. Note /ks-6-s/, of sixths, in which the nu-
meral morpheme followed by the ordinal suffix -th and the plural -s lead
to the large cluster.
Certain exclamations like Tsk!, on the other hand, may be restricted to
occasions which are socially defined. (This particular sound which is only
vaguely implied by this spelling of it, may be called a click. It is made by
closing off both the front and back parts of the mouth with the tongue,
lowering the mid part of the tongue a bit to make a partial vacuum, and
releasing the tongue tip to allow the air to rush in with a hiss or a pop. It
may be used informally to express lugubrious disapproval. )
Differences of distribution of grammatical units are as crucial to their
definition as are distributional constraints to the specification of emic units
of sound. If we have the sentences Why did you come? and Joe is coming
today, how can we know that they represent different syntactic structures?
One difference is that this question has an interrogative slot for why. An-
other is the meaning difference—the fact that one inquires, whereas the
other states. But a third, important, difference is the class of responses
which can appropriately come after each of them (see fig. 8.3). After Why
did you come?, one does not expect to hear So I see; but that might be ap-
propriate after Joe came today. Such contrastive distribution ofgrammati-
cal constructions in larger constructions—here in dialogue—can serve as
one of the criteria for identifving them as emically different. Until one
knows the appropriate distribution of constructions, one cannot talk a lan-
guage well; one does not know where to put words, clauses, sentences—or
paragraphs or conversations.
Utterance Response
A. Question: Why did you come? Answer: Oh—just because.

Or: He told me to.
Not: * Yes.
B. Statement: John is coming Reply: Oh?
today. Or: I believe you.
Not: *Oh—just because.
Figure 8.3. The contrasting distribution of sentences in larger conversa-

tional or narrative structures may help differentiate and define them as
emically contrastive.
Distribution 65
8.4. As a Point ina System
One can in part define a unit by the place it fills in a matrix of

units (an n-dimensional system of units). For sounds in such a display of a
system, see figures 5.2—6; for tone, figure 6.3. For affixes in a matrix, see
figure 7.2; for word meanings, figure 6.6; for clauses, figures 5.5, 6.53 for
stanzas of a poem, figure 5.6. The labeling of the rows and columns of such
a display (fig. 5.2A) tells us something of the characteristics of the units
listed there.
But their place in a matrix can also be used as a further kind of feature
which can help identify and contrast units. If in our preliminary analysis of
a strange language we have minimal pairs to contrast sharply all but one
pair of sounds in a phonetic chart, lesser evidence in connection with this
systemic data may justify the conclusion that the remaining pair is emically
contrastive also. In English, one finds pie / buy, tie / die, choke / joke,
fie / vie, seal / zeal—but not ship / *zhip. One does, however, have mis-
sion vision, where the first consonant of the word pair also differs, along
with /§/ and /Z/; and rush rouge, where the vowels differ, along with the
final consonants. The strong contrastive pattern of the system as a whole
supports the conclusion that the /8/ and /Z/ of mission and vision are dif-
ferent phonemes.
Grammatical patterns may also need analogy in a matrix to help deter-
mine some systemic contrasts. Human nature is such that for verbal or non-
verbal systems to be operable at all, they must have some symmetry about
them. Totally random structures are not serviceable in human interaction;
a field structure lies at its heart. This does not affirm that no exceptions can
occur, nor does it affirm that no irregularities will be found. Rather, it af-
firms that some degree of systematization can be found in every function-
ing structure. It follows that human nature cannot be described accurately,
purely in terms of isolated units of sound, purely in terms of classes of such
sounds without reference to system, or in terms of a miscellaneous list of
rules for combining them. Nor can mere lists or taxonomies of events, ges-
tures, behavior elements, abstract semantic components, or abstract formal
rules of action describe human activity in a way satisfactory to us. Some-
where, to some degree, the distributional relation of a unit to a background
system must enter the definition of every unit either as an implicit under-
lying assumption or as an explicit part of the descriptive mechanism.
7 oe.
A Tr i
wy aes > oe
; wi sf
: yorred) ar
ce, MY
MV ,!, i a _ od \ fa
7 taj bing rig ar &, A 2
sol :? pial! Te Peet if i; :
? 94," AR a tld 7?
¢ r oe a eee | PA ri ral pre wot
= a” 4
Lo AeA‘wed: AS a i
a” bm,
tad YH
ma 7.
il
5 A
’
aan)
n ’ , pied? Rae ay re —s, vey
iis xP, fae irs : ; : i 7

: [ia
Ht
Hierarchy
‘
-
fi
']
s
4
.
in
i
= i.
J
St
-_ H
7 o*
*
- + Be,
;
bd ‘
~
’
y
}
A
~,
1
i “ :
: 7
A
ft)
" .
~ ‘ ys ‘
B iy)
Introduction to Part II
In Part I we emphasized certain characteristics which the ob-

server brings to his study of data. The raw material experienced is in some
sense inaccessible to us. The observer enters into the warp and woof of the
pattern which he seems to observe and report. His report reflects an input
from the combination of his perspectives of particle, wave, and field.
In Part II the observer was less in focus, as units observed in the data
became the target of study. Even there, however, the observer intruded with
judgments whether two different items were emically the same (separate
instances of a particular behavioral structural unit), or whether they were
rather noncontrastive etic variants of an emic unit.
Now, in Part III, the emphasis shifts again. Here it centers upon three
facets of language behavior and structure which interlock with one an-
other, but are also somewhat (but by no means completely) independent
one from another. These structural facets may themselves be called hier-
archies. In the next three chapters special attention is directed, one chapter
each, to the linguistic hierarchies of grammar, phonology, and reference.
The linguistic terms and concepts have analogies in nonverbal behavior.
Grammatical Hierarchy
9.1. A Psychological Constraint on Complexity
The relevant structural inclusion of wheels within wheels, of

successive inclusion of parts within a whole, with the whole in its turn be-
coming one part of a still larger whole, is an important component of our
approach. We arrived at and have maintained this emphasis because of
our own experience with linguistic data. But it is very helpful to find that in
the discipline of psychology there is strong support for a hierarchical view.
Miller in a now classical article in 1956 summarized and discussed the
evidence. The crucial point: There is a “channel capacity” (86) affecting
“the span of immediate memory” which “is limited by the number of
items” which he calls “chunks of information” (92) or “units or chunks”
(93), such as “steps” from letter to word to phrase (93). But—and this is
vital to understanding both his approach and ours—“The span of memory
seems to be almost independent of the number of bits per chunk” in the
areas studied (93). And: “Our language is tremendously useful for re-
packaging material into a few chunks rich in information” (95); and the
limit of chunks per larger chunk seems to be “the magical number seven,
plus or minus two” (note his title). But it needs emphasis that this number
is that of the chunks—the sections, words, phrases, clauses—not the
“bits,” that is, not the contrastive features of the chunks.
Figure 9.1 gives an illustration of the fact that we can recognize at once,
without counting, the number of dots which are present when there are
only a very few, but that we cannot do the same when there are many. If the
Introduction 71
Figure 9.1. Recognitions of chunks is limited by their number (appro.i-

mately seven) rather than by the number of simultaneous features lead-
ing to their internal complexity. In A one can easily see four dots, with-
out counting. In B one cannot tell at a glance what the number of dots is,
whereas in C one can immediately recognize that there are four groups of
dots (cf. Miller 1956).
72 Hierarchy
dots are grouped, however, we can once more recognize the number of the
groups, without counting them, if these too are few (see Miller 1956:90).
Similarly, we can easily hear three words in three big cats even though it is
impossible for us to note quickly where and how many phonological fea-
tures are present (voicing, various relevant tongue positions, friction, stop-
page of the air stream, and others).
9.2. Behavioral and Hence Semantic Relevance of

Levels of Grammatical Structure
But a “level” of a hierarchy in our view of grammar is much

more than the mere degree of inclusiveness in successively larger chunks.
Each major type of level has some kind offunction in behavior. Its struc-
tural arrangement, its form, has simultaneously a general kind of impact
on the hearer (which we can call the meaning of that level). We are well
acquainted with the fact that s-k-u-n-k, is a spelling form (reflecting its
pronunciation form), but that this is not all; it carries the meaning of a
native American animal which puts out an incredibly repulsive odor. The
term derives from an Algonquian word (from which the name Chicago
also is said to have been derived).
Note, for example, a difference between That skunk smells bad and
that bad-smelling skunk; the first formally asserts that a state is true; the
second merely labels the characteristic to refer to it without putting it in the
form of deliberate affirmation. The first might start conversation (once
Hellos are out of the way), but the second would not normally do so. (We
say “normally,” here, without repeating it in every other relevant instance,
since imaginative students or poets can almost always find some context—
an off-norm situation—where an unexpected form might actually show
up. Norms are important to include in a theory at some point; in tag-
memics they enter in relation to a wave view.) And the meaningful affirma-
tion characteristic, with various modifications according to context (norm
or off-norm), we can call proposition, for our purposes; the meaning of the
label level we will call term.
But we drew upon the conceptof conversation in order to reach this con-
clusion. That is, the relative independence, or relative isolatability, of prop-
ositions is definable only in relation to a still higher level, that of conversa-
tion. And this conversation level includes a meaning function of social
interchange. It is this factor which then becomes our entrance point into
many of our linguistic definitions. Language is a part of behavior. It is not
Grammatical Hierarchy 73
isolated from nonverbal action, but is integrated with it, and takes its mean-
ing from social settings, intents, and reactions.
In between the levels of social interaction and of proposition we may
wish to posit another: theme development. We all know that a lecture is
(normally expected to be) more than a sentence; but few of us would care
to call it a conversation. Rather, it often takes a topic, in the form of a prop-
osition, and then makes comments about it. A paragraph can do that. So
can a lecture on a larger scale, as a single continuous “speech” (except in
off-norm situations when, for example, the presumed continuous speech is
interrupted by heckling comments from a hostile audience or by Amens
from some religious ones).
9.3. Pairs of Etic or Emic Hierarchical Levels

Sharing Behavioral Impact
And now we can generalize a bit more. It is useful to have a

preliminary, generalized (etic) view of pairs of types of levels around the
world before checking to see whether any one language has them all (it
may not), all of them behaviorally relevant (emic) to it, or whether it adds
more or lacks some. At the top, the social interaction pair can include both
a single dialogue exchange between two people or a conversation between
various people. Comparably, the theme-development pair—a bit lower in
the hierarchy—can include a paragraph or a monologue (with numerous
kinds of disjunctively defined variants not important to us here). Proposi-
tions—still lower—can be in the form of clause or of sentence. Terms can
be in the form of word or of phrase.
Morphemes and morpheme clusters in words comprise a still further
pair of levels—the lowest one in our treatment of the grammatical struc-
ture. A root such as berry would be the simple member of the pair; black-
berry wouldbe a morpheme complex (here, a compound stem) in black-
berries; the -s ‘plural’ would be another simple morpheme, an affix. In
each pair type, that is, there can be smaller (minimum) or longer (ex-
panded) kinds of specific levels within the general kind of level.
A caution: If one wishes them to reflect behavior-as-it-is-lived, defini-
tions often cannot usefully be made as simple or as clean-cut as one might
think that they ought to be. Some definitions need to be disjunct—that is,
almost a sum of several different definitions leading to a kind of emically
defined behavioral impact. Others can be conjunct—that is, a single rule.
In tennis, for example, a game is won not by a single, definable kind of
74 Hierarchy
physical action, but by a person. Any tennis player—even a very young

one—knows that there are several different ways (by disjunct definition) to
win a game: by serving four winning points in a row without the opponent
getting any; or by eventually getting at least four, but also at least two more
than the opponent, without having allowed him to get more than two
ahead after the first three; or by winning by default if for some reason the
officials rule that the game is validly won whereas the opponent does not in
fact play—or finish playing—the scheduled game. The attempt to describe
winning is not simple.
Similarly, the definition of sentence has given great difficulty, in past gen-
erations, to persons looking for a single-faceted (conjunct) definition in
which one descriptive claim fits all the varied instances. But in our ap-
proach the sentence is disjunctively defined. As a start, it first is given a
subdefinition as a minimal kind of proposition which may start a conversa-
tion after the Hellos—although this is also one definition of an independent
clause such as Look at that skunk over there! (which is simultaneously
sentence and independent clause in that environment). But as a second
subdefinition a sentence may have two or more clauses in it, one of which,
for example, is marginal and the other nuclear (note the wave view enter-
ing the description), as in If you want to be careful, watch out for that
skunk over there! For some important purposes, a sentence may have a
third subdefinition, which reflects a total response no longer than these
others mentioned, but also shorter ones: A word is a response sentence in
[Who is coming today? Response:] Lancelot; but not in [What did you see
today? Response:] A big’ smelly skunk! When Lancelot stood alone, in the
first of these two responses, it was simultaneously a word and a sentence;
when it is included in a longer response, it is still a word but not a sen-
tence. That is, the definitions of sentence, clause, phrase, word, and mor-
pheme interlock—overlap—in relation to this occurrence at different times
and places in discourse (for details see Pike and Pike 1977). In addition, we
have to call on the concept of norm lest we get forced into treating gl- in
glance, glitter, glow, gloom as a morpheme comparable to skunk.
A further caution: Sometimes we mention paragraph, word, and the like
as if, for example, a word comprises such a level. But when we speak a bit
more carefully, we say that an etic level is made up of a set of such ele-
ments—for example, all the words of a language together make up the etic
level of word.
A third caution: If the researcher into the structure of a specific language
decides that there is no emic contrast—say, between etic word and etic
phrase—but just the emic level of term, then the set of all etic words and
phrases would comprise just one emic level in that language.
9.4. The Unit-in-Context (the Tagmeme)

The four features of a tagmeme are in part independently vari-
able; they are also mutually dependent on each other, with interlocking
components and definitions. In general, these four components help an-
swer questions well known to everyone: (1) Where (in what slot) in the
immediate structural setting does some concrete part occur? (2) What set
(class) of units can be parts of that larger containing unit, in that same
place in the unit? (3) Why is any member of that set of units relevant
(meaningful) to the sets of units in other slots at that same immediately-in-
view structure? (4) How does the generalized frame of reference, on this or
other levels of the hierarchy, reflect or exercise control (give cohesion) to tie
the parts together? We shall now discuss these again, in more detail (see
also the four-cell diagram in fig. 9.2).
(1) Slot (2) Class(es)
Where (the position) What (the items)
Specific place of part in whole General set of items substitutable

appropriately in the slot
Wave characteristic, with Particle characteristic
nuclear or marginal relation
Syntagmatic relations Paradigmatic relations
(3) Role (4) Cohesion
Why (the relevance) How tied to other units
Specific function of the set to General background materials

other sets in the including from any level of the hierarchy
whole which are controlled by or
controlling the item in view
Behavioral meaning Field characteristic, systemic

structure
Pragmatic relations Framework relations

Figure 9.2. The unit-in-context (the tagmeme) has interlocking features,
each quasi-independent, but each dependent on all of the others. Hence
the unit in view here is not the isolated lexical item, but a set of elements
and features of elements mutually relevant—hence the term unit-in-
context. (See notation throughout Pike and Pike 1977, and display 12.3.)
76 Hierarchy
(41) Structural slots: Hierarchical grammatical structures are charac-

terized by parts making up wholes, which in turn make up other larger
entities. But for a general entity to be recognizable as emically the same,
with etic variation involving substituted parts, those parts must somehow
be seen as occurring within constant structural positions.
These slots frequently differ as to nuclear versus marginal relations to the
structure as a whole. These relations can occur at all levels of the hier-
archy: Within the word skunks there is a nuclear root (with role of item)
and a marginal suffix -s. Within the sentence If I can, I shall certainly give
that skunk a wide berth, the marginal if clause (with role of condition)
has a nuclear independent clause following it (with role of statement).
Within the paragraph It is good to avoid skunks. They are smelly crea-
tures. Why run the risk of having to bury one’s then-forever-unwashable
clothing?, the It is sentence is the nucleus, and the They are sentence is
marginal, followed by further explanation in the form of a rhetorical-
question sentence. Within the dialogue exchange
Abe: Why can’t you use your clothing after meeting a skunk?
Bill: Because the smell won’t wash out!
there is first a nucleus, followed by dialogue margin.

(2) Substitution class: At any point in a grammatical structure we can
expect to be able to substitute one of various lexical items for another with-
out affecting the proportion (the general, not lexically specific, relation-
ship) between the various kinds of chunks of the unit. Thus, instead of The
skunk lent its X to Y, one might hear The man lent his presence to the
occasion, or A big boy whom I used to know lent his top to a friend, or The
skunk gave flavor to that incident. Here The skunk, The man, and A big
boy whom I used to know are all members of the substitution class noun
phrase, and are appropriately occurring (distributed) within that subject
position. (Note that each would change the message; that is a different kind
of change—a referential one—which will be taken up in Chapter 11. Once
more, also, “normal” is in view: some items like -s ‘plural’ are not appro-
priately replaced by anything else under some conditions; in such instances
a class may be limited to a class of one—or to one plus zero.)
(3) Role: But the same class may have a sharply different role from one
sentence to another. If I say I smelled the skunk, the role of the skunk in the
act is quite different from that in The skunk smelled me. Who is doing
what? In the one case it is Iwho am doing the smelling; in the other it is the
skunk. But in the passive form, I am smelled by skunk, I am not doing the
smelling, even though I am the subject of the clause. It is this kind of fact
which forces a long-known difference between form and function in gram-
matical relations. Specifically, we have insisted from the very first publica-
tion on tagmemic theory (Pike 1967a; 1st ed., 1954: sec. 7. 6) that there was
not just one subject tagmeme (called grameme there, but tagmeme in the
second edition, 1967a), but rather a set of various subject tagmemes, dif-
fering by role meaning. Thus there was subject-as-actor distinct from the
tagmeme subject-as-goal (but with “goal,” a Bloomfieldian term, referred
to as “undergoer” in Pike and Pike 1977, or “patient” in the work of case
grammarians such as Fillmore 1968).
There is, furthermore, etic variability in the meaning of such roles:
While retaining the same kind of grammatical formal proportion, the sub-
stitution of particular lexical items may change the message relation (the
referential relation) between the parts. For example, in I killed the stinking
skunk to get it out from under the house, I am there a voluntary actor; but
in I smelled that intolerable, unavoidable odor, then I am an involuntary
actor. (This attention to etic versus emic differences of role meaning dif-
ferentiates the Pike and Pike approach sharply from that of some other ap-
proaches to case. As a further difference, our approach insists on studying
such role relations at every level of the grammatical hierarchy, not just on
the level of clause.)
(4) Framework and control (cohesion): Something has to tie words to-
gether in some kind of coherent whole, or talk and the perception of the
world revealed by that talk would be a random madhouse. Talk structures
must have their parts in internal systemic agreement, often marked as such
by particular bits. For example, Skunks are mild, beautiful, and pleasant
unless frightened or disturbed differs from That skunk is lifting his tail—
beware, that’s a signal that it is getting ready to be offensive! The back-
ground frame of reference includes number (one versus more), and this is
reflected in Skunks are versus That skunk is, and the words is and are are
selected (controlled by) the number difference of skunk versus skunks. In
addition, it is’selected because it is controlled by the fact that animals are
usually (apart from personification in folk tales or other circumstances)
treated as nonhuman and referred to by the neuter pronoun. And the pres-
ent tense is used both in are and is, controlled by the grammatical frame of
reference which is that of a “telling style” which is talking about more or
less permanent characteristics which can be represented by that tense
(with the speaker talking now, about the situation now). From language to
language, however, the choice of the particular kinds of items to be for-
mally marked “for coherence varies (for example, in sight versus out of
sight, older versus younger, reported by the speaker as known from his
experience versus reported but with the speaker refusing to take responsi-
bility for its accuracy).
The contrast between slot, class, role, and cohesion can also be related to
yxajuo9]
15a [paves
é
(Z)
uoyesiaau0D
(G-V)
()] [e105 punos8yoeq Surfjo.nuos
[euosiad [saouasayau
tt)
gnapuNn
(¢)
(z) ydy.se.eg
pue enSojouow
(d-a)
(»)] asuay,pue [euosiad asuasajas [petionuos
gugentl
(Z)(¥)
jo
pue
sump
aouayuas
jonuoD
aaneso.ajuy
78 Hierarchy
uossad
(F)][pouauajyas
(Z) aayisueyy, xajduio5 aouajues
(qd) (7) aanisuentg
yoeads (7) aagtsueyuy
(F)] asuay [(pate.quos asnejopue
(9)
pue
aoua)ues
(d)
asne[o
~
(F)]
(*)] asuay,pue Japuas
asuay,
e0ua}uas
[pato.nu02
[pettonuos
(z) yuapuadag
esnejs(qq)
(Z)
(F)] esuar, [Pattenuos
(eg)
(F)]
esnejo
asuay,
aagisuesy,
[Petfo-gu0s
(z) uoyounfuog
(tqq) (Z) aaisseg sayisues]
9sSNne]D
(7) [puns 1004 (arugg)
(+)] asuay [Petlonuos
(Z) unou-junog
aseryd (z) jeuontsodaig
(¥)] sjoquoy s2quinu[@qa) esesyd (arqa)
(7) aAISseg SADISUB.Y (r)] [ruasy
quan asesyd (9) aue[D :pajsassns (a) oy. Aused
(v) yeym pauaddey
ise] (F)]} JayseWasual
WIGS , J, NOA JUBM
0} ayouq
dn ye
“Tysru caonig pattoquos[("4) say -8ns
aq
uoysas
yey)
198Woy
‘[nyares
feme
Ayred e Suraey aiam a (eg) dyunys

(qa)
& ajquuioy
JOpo (Iq)sem A{uappns —_—pay|auis (arqg)
Aq Apogdreaa
(144)
uaym
other terms known to linguists: (1) syntagmatic relation to a sequence of

parts in a larger whole, (2) paradigmatic class, of appropriate replaceable
parts in a comparable place, (3) pragmatic relevance of behavioral func-
tion of that kind of chunk, (4) agreement of items.
9.5. Diagrams and Formulas for Unit-in-Context
For effective work, however, one must be able to apply his gen-
eral ideas to specific detail. Yet when the detail is enormous in its variety, as
in language, the mechanism for representing it is found to be complex. A
theory or notation which is too simple may omit crucial behavioral details.
A mathematics without a notation would be less useful than it is with.
The change from roman numerals to arabic numerals makes the practice
of multiplication different. Practice, as well as theory, changes the world.
What kind of notation can help to move us from a view of language as
made up of words as isolates (with their inner structure making up the
whole of grammar), or sentences as isolates (a view which thinks that there
is no grammar beyond the sentence), to a view that all units are perceived
as relevant only in relation to an including context or set of contexts? Or
that even language as a whole cannot be isolated “as a separate science” (a
dream of some of my teachers) without dependence on items outside it in
culture as a whole?
In order to illustrate one kind of tagmemic notation, we give a brief sam-
ple in figure 9.3 as a part of the grammatical hierarchical structure of a
Figure 9.3. A tree diagram, tagmemically labeled, can show successive

hierarchical inclusion of grammatical constructions as units (as immedi-
ate constituents) within a particular larger construction. (1) A slot label
on a branch shows its inclusion in a higher unit, for ecample, as nuclear
or marginal to its wave features; (2) beneath the branch a function label
shows its role in relation to the including whole; (3) at the point ofjunc-
tion (the node) a construction label gives the name of the part which
happens at the moment to be filling the including slot; and (4) along with
that construction name there is sometimes given in brackets a suggestion
as to controlling or controlled features in relation to other items at the
same or at different levels of the including hierarchy or of the system.
Because of space limits here, only a sample of the relevant labels are
given, and structures inside the verb phrase and noun phrase are omit-
ted as well as details of most of the clauses and sentences.
80 Hierarchy
single conversation outlined in a way somewhat like a (very old) classical

sentence diagram (or tree structure), but with the parts labeled by the four-
cell tagmemic features. (For more detailed illustration, and for the technol-
ogy by which one arrives at such tagmemic analyses, see Pike and Pike
1977.)
The tree diagram of figure 9.3 splits the sequence of the conversation into
its successive immediate constituents. It first suggests, by the dotted line to
the upper right, that the conversation is embedded in a situation larger
than the conversation itself. At the bottom of the tree, the actual conversa-
tion is given, with pairs labeled for reference as A to D. At the top, just at
the lower end of the first line slanting to the left, is the number 2, which
refers to Cell 2, the class; the name of the class at this point is conversation,
covering the indicated span from A to D; and immediately below that, in
brackets, is a suggestion of its relation to background material or linguistic
social system, from the cohesion Cell (4). This sequence then breaks down
into two tagmemes, leading to right and left branching lines. The left
branch is marked above the line for Cell 1, as nucleus, and below that same
line for Cell 3, as having the role of initiating the conversation; the inter-
rogative sentence comprises the filler of Cell 2 at the lower end of that
branch. The right-hand branch is marked as marginal, and response,
manifested by paragraph, in which tense and personal reference are con-
trolled for cohesion.
The paragraph B—D in turn breaks down into a nucleus and two mar-
gins which are respectively topic, contingent suggestion, and result, man-
ifested by a transitive complex sentence (B), a bitransitive speech clause
(C) which is here simultaneously a sentence, and an intransitive clause (D)
which is here simultaneously a sentence. The complex sentence B is then
broken into clauses Ba and Bb; and the passive transitive clause root (Bbii-
iv) is broken down into subject, predicate, and adjunct. For lack of space,
further breakdown is not given. (But for illustration of bitransitive with
tagmemes of subject, predicate, adjunct as undergoer [direct object] and
adjunct as scope [indirect object], see Pike and Pike 1977.)
There is much detail in such a figure. Nevertheless, one major gap is
present there. It gives only the structure of one particular conversation, or
one particular clause or phrase type at a time. It does not show how these
are instances of broader recurrent patterns, each with a range of variability
involving more or fewer optional tagmemes. In part to fill this gap, we give
one sample of a more general formula in figure 9.4. Such a formula repre-
sents a large number of potential instances built on that pattern. In this
sense it simultaneously represents (or generates) them all.
An English independent declarative bitransitive clause is chosen for rep-

resentation in that figure, but limited to its root form (omitting modifiers of
time phrases and others), and represented by an abbreviated name. Four
tagmemes are represented, with three of them obligatory (subject, predi-
cate, adjunct as undergoer [object]), and with one of them as optional (ad-
junct as scope [indirect object]). The characterizing formula follows the
equals sign. Then each tagmeme is given, first with an indication as to its
normally required presence (shown by plus—though the tagmeme may be
omitted in some special discourse contexts where the context is clear), or as
to its optionality (shown by a plus-minus sign). Then comes an abbrevia-
tion for the characteristic of the slot (subject, predicate, adjunct), followed
after a hyphen by the role feature (actor, statement, undergoer, scope). A
colon is then given, following which the class of fillers of the slot is indi-
cated (for example, noun phrase, declarative verb phrase, a noun phrase
beginning with to or with some comparable particle). In this particular
abbreviation no cohesion feature is given. (For discussion of this kind of
notation, see Pike and Pike 1977:12-—18, 36—37, 39—40, 44-46, 412-54.)
In the passive clause root Bbii—iv of figure 9.3, the formula would have to
be changed, to make the subject have the role of undergoer instead of actor.
The data are so complex that different scholars have given primary at-
tention to different selections from the data, with different resulting types of
formulas. Bloomfield, for example, used the roles (our term, not his) of the
tagmemes of a clause root to label the clause. For example, his terms “ac-
tor,” “action,” “goal” (or “undergoer”) (1933 :267) would represent our ac-
tor, statement, and undergoer in a transitive clause root. For Jespersen
(1937), on the other hand, the labels were more like those of our slots, as
subject-predicate-object. Fries (1952:76-—86) chose, for notation, word
classes similar to noun and verb, but labeled them with the numbers one
and two. A notation based on the cohesion cell could also be invented, I
suppose, but I do not recall having seen one; perhaps one could experi-
ment, for the same clause root, with the concept of relation to number, for
example, and label the subject-predicate-object as controlling unit, con-
trolled unit, and noncontrolled unit.
But under some circumstances, it is a decided loss if one loses easy refer-
ence to different phases of the unit-in-context—circumstances, for exam-
ple, where the normal relation of role to class is overridden. Note the
question directed by a starchy lady at a sniffling youngster: Little boy, do
you have a handkerchief? To which he replied: Yes, ma’am, but Pm not
allowed to lend it to anybody. Here the lady’s impatient intent (role) was a
command; but its relation to class was off-norm since it was spoken with
“y uy juapuadapuy sayerepaqisueyig
sa Isne][DJOOY
(T) joalqns (7) sadoigunou joo1 (T) espe aayesepa
quan q aseryd
jo
(2)
joa{qns unouo.d suites
AF
junog unou aseryd
Ste
(€) JIo}oV (F) jDa[qns sjo.1yUOs Jaquinu
jo (¢) JUSTa}eIS (¥) yoveds aanisueyiq
‘oseryd
ayeotpaid ymsy aouasaid Ayfenjnur
yeurMoN pe]jo.yu0s
Aq Surjosjuos
pue pay[o.jQuos
Aq
as.INOISIp ar} aouasaid
jo -se-joun[pe
adoos
) joamput,,(,Joaf{qo
pue
-se-joun({pe
“10}9e-se-joa{qns
1a081apun (.J9alqo,,)
saquinyy payjonuos
Aq
1aquinu
Jo au} ‘yoafqns
(T) ounfpy Joarrput,,)
| (Z) 07 yim unou aseryd (1) unfpy
40) | (Z) adUa}Uag
JO asNe]D
JO SNOLIEA
(,Joa{qo uoyes07Taseryd (.d9a(qO,, Surpnyoul—sa
Juoule}e}s
dA}
UOT}eD0T
pi0oMm ‘sateSO.11a}UrAeIePap
‘aayesadumt
10 Jayjo
juno unou aseryd
unouolg
puy] “ayo SuIpnpour
+ +
[enso[ouour
(¢) adoag
10) Joasput,,
| (F) Japuas)
pue Jaquinu (¢) Ja0S1apuy)
(¥) Japuas
pue Jaquinu
(,joalqo pay[o.ju09
Aq asnoastpSuIas 10) «}99lqo,, pe[jo.ju0
Aq as.moosip
10 ,Je038,,
JO
(,quayed,,
-_ JHIDLEPedepul
= dN‘?V-S+ dAP2a:e}S-d+dNO:N-PV+ IN9O}}3SPV+
interrogative form (class). The boy, however, interpreted it as if it were a

request. (See Pike and Pike 1977:49-50 for a systematic display of fur-
ther illustrations.) The voice quality which the woman used was probably
brusque—and should have signaled its off-norm intent, but the boy missed
the signal.
Figure 9.4. A generalized four-cell tagmemic notation. In A the descriptive

terms are written out for easy reading by the person unacquainted with
the abbreviated symbols. In B a partial summary is given, to show the
way the symbols are more likely to be abbreviated when less of the data
needs to be in view. This formula represents a clause root such as Bill
gave the skunk to the garbage man or Bill gave him the skunk, with
further variation of the order in which the adjunct-as-scope may come in
the clause.
The Phonological Hierarchy
We have seen a number of characteristics of units of the pho-

nological hierarchy, anticipated in other contexts. Sounds may smear to-
gether (figs. 4.1-3). They enter a field structure (figs. 3.6, 5.2—4). They
may carry relevant tone contrasts (figs. 6.4—5). They may be displayed in
component-times-unit charts (fig. 5.2). Their place in a word affects the
difficulty in pronunciation, and their number and arrangement affect sylla-
ble structures (fig. 8.3). Now we focus on the phonological hierarchy per
se, instead of using its characteristics incidentally to illustrate other ele-
ments or principles.
10.1. Noncoterminous Borders of Phonological and

Grammatical Hierarchies
In the last chapter we discussed chunks of material arranged

into grammatical units. But at the same time, the same substance—for
example, the same sentence—is arranged differently, into phonological
chunks. And these segments, like the grammatical ones, are hierarchically
ordered—but with their borders sometimes not coinciding, and therefore
forcing an analysis of the system of language into distinct hierarchies. Thus
skunks was one syllable but two morphemes (root plus suffix), as well as
being one word; but horrible was one morpheme (and one word) but three
syllables. Similarly, to be careful is just one stress group (it has just one
accent, shown by the acute mark over the vowel), but it contains three
words, with four syllables and four morphemes (see fig. 10.1).
The Phonological Hierarchy 85
I?’m—Grammatically: Word border occurs between I and am.

Phonologically: No syllable border occurs within I’m.
Figure 10.1. Noncoterminous units of grammar and phonology are the

Justification for treating them as distinct hierarchies.
10.2. Relevance Criteria for the Pairing of Levels
In the last chapter we showed that there were behavioral rea-

sons for organizing our description of the successive inclusion of gram-
matical chunks into pairs of kinds of chunks; there was a social relevance
to such a structuring scheme. So, for pronunciation, there are sharp dif-
ferences in the function of different levels of chunks in the hierarchy. Just as
in grammar the clause and sentence have a kind of independence (relative
to the other levels) in being able to initiate a conversation (after a Hello/),
so also there is a kind of phonological independence which is more likely to
be related to the syllable and to the stress group (a pair like clause and
sentence) than it is to an isolated sound (like k-) or to a cluster of sounds
(like sk- or -nk in skunk) less than the syllable. It is relatively simple, in
most languages, to teach even a preliterate to pronounce a phrase in terms
of syllables, but very difficult to have him separate the sounds one at a time.
(Note that for the latter I do not mean the names of the letters as given
when a word is spelled, with /w/, for example, named to sound like
“double-you.”)
Even for professionals in phonetics it takes some concentration to be able
to isolate the sounds one after another. So this independence characteristic
is important. Given that fact, it is less surprising to learn that writing in
terms of syllables has been invented various times in history, but that al-
most every alphabet—perhaps every one—ever invented has had some
kind of remote indebtedness to an accident in history arising in the Middle
East which abstracted some signs for an alphabet (at first consonants only)
from the hieroglyphs of Egypt. These at first were merely word pictures, of
which some twenty-two were later extended in usage to represent any of
the syllables whose initial consonant was the same as one of that small sub-
set of the many hieroglyphs. (See Gelb 1963 for a more precise statement of
this history. For the pairing of phonological levels, see Tench 1976. And see
sec. 10.6 below for consonant clusters.)
86 Hierarchy
10.3. Constraints of Focal Attention
But this last section reminds us of figure 9.1, where we saw that
the mind inevitably grasps things, features, or subunits by clustering them
into units, when the nature of the mind makes it impossible, say, to count
or to identify separately all of the included items or features at a glance. So
also, here in phonology, a hierarchy of chunks is necessary for our under-
standing of human behavior.
The minimum (shortest) such phonological chunk (segment) which
meets the requirements laid down in section 6.3 for being independently,
consistently different from other chunks is the phoneme. The articulatory
movement toward the nucleus of such a wave of sound (sec. 4.2), or the
movement away from that nucleus, can be important and contrastive. This
can be seen, for example, if one merely hums the sounds m, n, and the
sound -ng' of song, without a vowel; they sound much more alike than they
seem to us if we speak aloud the words sum, sun, sung—where the glide
off from the vowels helps to enhance our perception of the contrast. Sim-
ilarly, the (usually) longer vowel in cub may help us to tell the -b from the
-p of cup, where the vowel may be shorter. Yet neither the glide away from
the -u of sung nor the glide toward the -b of cub attains, by itself, the status
of a phonemic chunk.
It must be emphasized again, however, that such a relevant sound is an
emic chunk, not an etic one: that there are phonetic variants of each such
phoneme chunk, depending upon the environment in which the phoneme
occurs (as, for example, the /n/ differs slightly after or before -u- in sun or
numb, or as it differs slightly before the different vowels of numb, nymph,
neck, noose, nap, know). Machine recording of such sounds shows up dif-
ferences very clearly. But the normal (nonphonetician) speaker of English
does not notice the specific details as such—he cannot talk about them in
academic terms or identify verbally the differences—even though he may
hear as a queer foreign accent the way foreigners depart from our norms in
using them. In the same way we can recognize a friend as the same person
when he smiles or frowns, laughs or cries, but we may be unable to de-
scribe him (that is, put into words the detail about his face as a chunk) so
that he can be picked out ina crowd (unless, that is, he has some very
strange characteristic such as a nose twisted horribly).
It comes as a shock, therefore—or as incomprehensible and therefore
almost incredible at first—to learn that such minute differences can be the
basis of word differentiation in some language other than our own. For ex-
ample, the two p sounds of English paper are so different, as caused by
their position in relation to the stressed first syllable, that the same phonetic
difference (the aspiration—the puff of breath—after the first but not after
the second) can signal a difference of meanings (such as /t*i/ ‘circle,’ versus
/ti/ “it burns’) in the Jalapa Mazatec Indian language of Mexico (data from
Judy Schram),
10.4. Source of the Terms Emic and Etic
And it was specifically my academic inheritance of this ver-

balized knowledge from former scholars, plus my own application of it in
the analysis of languages hitherto unwritten, which forced me to see that
comparable situations existed in all human behavior. I extended the terms
phonemic and phonetic to behavior as a whole by dropping the phon- (im-
plying pronunciation), and using only their endings emic and etic. The
tagmeme, as discussed in sections 9.4—5, is one of these units, where the
concept (but not the term) was developed in response to my question: Is
there a “phoneme” of grammar which will have analogous characteristics
of contrast, variation, and relevant distribution (as presented in Chap-
ters 6—8)?
10.5. Simultaneous Features—

and Where Do You Stop?
I may be unable to describe my friend’s face in such a way that

others can quickly recognize it, but that is no reason to decide that the
physiologist cannot study—and talk about—the function of the eye and its
parts, or that the biologist cannot talk about the cells of the eye, or the
chemist the molecular structure of a few of its bits, or the physicist the
atoms in some of it or the way in which light acts in relation to a lens. A
person talks about what he is interested in and about things which he has
knowledge of inherited from former scholars or has discovered for himself.
And wherever he stops, there is always the opportunity for someone else
to probe in more detail into the hierarchy of the universe as a whole. In
phonetics one can wonder how the atoms affect the muscles which drive
the tongue to articulate a /t/. Or one can wish one could know the size of
the units which control the eardrum, or nerve fibers, or brain components
so that one can hear twenty thousand vibrations per second. Yet one must
stop arbitrarily, somewhere, knowing that there is no end. My conclusion:
One cannot ever work adequately with a philosophical reductionism
which demands that one begin with the smallest bits of the universe, but
88 Hierarchy
must rather adopt the view that each person (and each discipline) jumps in
wherever he pleases and works up and down the hierarchy of reality until
he finds answers to questions tantalizing him, or until he gets tired, or runs
out of ability-in-the-face-of-the-current-state-of-the-discipline. People will
differ in these judgments and choices.
In this chapter, my attention has thus far ended in its “downward” view
by concentration on the minimum chunk, the phoneme. But in section 5.2,
and in figures 5.2—4, I was discussing simultaneous components (fea-
tures) of phonemes in relation to their articulatory fields. There was a
sharp discontinuity of phoneme to feature because of the possibility of
simultaneity of features in a single chunk: The /b/ of maybe is made up,
in part, of the simultaneous presence of the vibration of the vocal cords
(called voicing) plus the closure of the lips and of the passage from the
back of the throat into the nose. This in part comprises a consonant which
has cessation, for the moment, of air coming out the mouth and nose (that
is, it is a stop).
Many scholars prefer to give much greater prominence to such simul-
taneities than I do, however. They may treat them with little (if any) discus-
sion of the kinds of problems I have been interested in here. That is, they
may try to list a set of features occurring in many languages, but treat them
implicitly as if in any one language they were unchanging. This view is
inadequate. The features themselves must be treated as emic, and as eti-
cally variable. For example, the English phoneme /b/ by no means always
shows up as having clear vibration of the vocal cords throughout its entire
pronunciation. Other characteristics, such as the lengthening of a vowel
before a word-final /b/, may help to signal its contrast with a word-final
/p/ (as in cub versus cup). These factors of sequence conditioning, with
relevant bits carried by adjacent sounds, are not easily accommodated
within a set of features theoretically treated as if they were wholly invar-
iant. (A person who wishes an introduction to the feature approach might
consult Hyman 1975; in a foreword to it, Victoria A. Fromkin says, “Chap-
ter 2 deals with the basic building blocks of phonology—distinctive fea-
tures.” For further extensive historical introduction to varieties of phono-
logical theory and practice, see Fischer-Jgrgensen 1975.)
Going up the hierarchy, like going downwards, also has its difficulties.
The handling of tone and stress are related to phonological units larger
than the phoneme—that is, to distribution within the syllable, or within a
phrase larger than that. But phonology goes still further. A lecture as a
whole, for example, may have phonological characteristics as a unit, in
that one may sometimes detect when the speaker is running down towards
the end, by the characteristics of his voice quality.
10.6. Wave Characteristics of Phonological Units
The way we print books makes it easy for us to think of sounds

as particles, since the letters in a word are separated by spaces. In long-
hand writing, the smear between them is more visible orthographically. In
pronunciation, the smear and overlap is, however, substantial (as we saw
in secs. 4.1—2, with nucleus versus margin illustrated and discussed). Now
we add further detail about phonological hierarchy.
Phone (a particular variant of a phoneme, from some particular kind of
environment): for example, /n/; with a premargin on-glide towards it (as
in the glide toward -n from the -u of sun); followed by its nuclear peak
(when the closure of the mouth passage is completed by the tongue tip);
followed by the postmargin off-glide (as the tongue releases).
Syllable (or an etic variation of an emic syllable): with a premargin on-
glide to it comprising one consonant (in all languages possible, and in
some languages obligatory), or more than one consonant (like str- in
strict); with a nuclear element comprising at least a vowel or a syllabic
consonant (the -i in strict or the -m- in the exclamation hm/), and some-
times two or more vowels (as in the /a'/ of strike, according to the analysis
of English adopted here—but as /ay/ in the same word, according to some
alternative analyses), or a vowel plus a consonant closely joined with it
such as h or a glottal stop (the closure of the vocal cords) in some lan-
guages; and with a postmargin consonant or group of consonants, in some
languages (as -mpst of glimpst).
Stress group as a wave: see section 4.3, with want as the nucleus of the
stress-group wave of I don’t want to, pronounced very fast, smearing the
words as in figure 4.3.
Phonological paragraph: the nuclear theme material may sometimes
be pronounced slowly, distinctly, loudly, and with relatively high pitch,
whereas a marginal element, such as an aside (often written in paren-
theses), may be more rapid, softer, lower in pitch, and perhaps less dis-
tinct—or even set off by pauses.
Phonological discourse: may comprise a monologue marked somewhat
like the paragraph just mentioned, but with the characteristics spread over
a large unit. In addition, it may build up to a discourse climax, and at the
end taper off phonologically. (If the speaker gives these finalizing signals,
but speaks tog long and doesn’t end there, his audience may comment
wryly that he missed several good stopping places.)
Utterance and response: voice quality and speed are to some degree
likely to be set by the initiating speaker (who is acting in a nuclear pho-
nological role). For example, a question showing urgency and hurry may
90 Hierarchy
elicit a more rapid, tense reply (the margin) than would have been received
from a quiet question asked in front of a fireplace.
The speaker of English must be prepared for many surprises, however.
Instead of something which he might think of as syllables, he will find in
some languages that units of length may be important, emically. For exam-
ple, in the Mixtec of San Miguel el Grande, Mexico, every isolated pro-
nounceable unit must have two length elements. The unit can be made up
of two etic syllables (like English syllables), as in kata ‘to sing’; or it may be
composed of just one long etic vowel, after a consonant, as in ka: ‘will
climb’ (with a long vowel shown by a raised dot), which must be analyzed
as /kaa/ with two emic vowels (analogous to words like kata ‘will sing,’
but with no consonant between the emic vowels). And the tones contrast:
kaa ‘is climbing’, kaa ‘metal’, kata ‘is singing’-—the acute as high tone,
grave as low, and zero marking as mid tone (compare Pike 1948:79/n).
Etics and emics, that is to say, must enter the analysis of every language
at every level of its phonological hierarchy (as well as its grammatical and
referential ones). This fact poses a major theoretical and practical chal-
lenge to all researchers in language analysis. The answer to the simple
question What is a syllable? is still difficult—after decades of discussion
here and there. How many syllables are in C’mere! (Come here!)? Or in
*Kyu (Thank you, in some British speech)? Or in ’S cool today! (It’s cool
today!—compare school today)?
To experience this kind of problem for oneself, one may ask a native
speaker of Japanese how many syllables, thinking as a Japanese, he hears
in the English word skates. They have sometimes told me five—seldom
fewer than three. They may, in the first case, break it up into something like
sa-ke-ee-ta-sa; that is, the consonant cluster sk may appear to them as a
broken down syllable, with a voiceless vowel inserted after the s- and be-
fore the -k-. Few other simple experiments are so accessible to the phonetic
layman, and as directly convincing of the need for a systematic analysis of
language, in place of the normal lay assumption that one merely goes to it
to find its sounds.
In Chinatec, referred to earlier in figure 4.3, there can be two kinds of
syllables, one with a sharp, quick, let-go decrescendo, and one with a held
controlled length; compare (in the pronunciation of some of us) the sharp,
short pronunciation of the final syllable of English celery or effigy versus the
long final syllable (perhaps with some kind of secondary or tertiary stress)
of chickadee (a bird, one morpheme), or refugee (grammatically com-
plex—refuge plus suffix). There continue to be difficulties in the analysis of
syllables, even in English. In some dialects a word like more or mower or
fear appears more like two syllables than it does in other dialects of En-
glish. Such difficulties, however, should not lead to the discarding of the
concept of syllable—which we need very much—but rather to further at-
tention on etic-emic relations, and to the wave character of syllables which
allows for indeterminacies in segmentation at the borders.
In stress groups, likewise, we can have surprises as we study our own or
other languages. In English, in an angry, protesting That isn’t what I
said!, the syllable may be very short and loud, with the -n’t short and
weaker, with much lower pitch; but in a pleading pronunciation of the
same sentence, one may hear the opposite—a long, drawn-out is-, with a
slow drop throughout the rest of the sentence. The Culina language of Peru
has a special emphasis type used sometimes under surprise, reflecting a
dangerous situation—as in spying a tiger near one (see Pike 1957); the
final syllable of the sentence has a sharp, quick rise followed by a sharp,
quick fall and decrescendo. In Aguaruna of Peru (Pike 1957) a further pat-
tern comes in the chanting of a shaman. The end of, say, each sentence
ends in a quaver (a chanting lilt of voice pitch up and down on the syllable).
10.7. Field Characteristics of

Phonology Seen via Intonation
We have seen that to view sounds only as isolated chunks is not

adequate; they must be sometimes seen as waves overlapping and smear-
ing into each other (figs. 4.1-3), and they must be seen as units in a con-
trastive system or field (figs. 5.2—4), so that features of tone can be seen as
contrasts (fig. 6.4) which lead to choices which can be compared to a per-
son choosing to go through one of several gates (fig. 6.2). In this section we
wish to illustrate that even in English a pitch choice is possible also, with a
similar set of four gates, or relative levels of intonation contrast, but con-
tributing to structures not at the level of syllable nucleus (illustrated for
Mazatec in fig. 6.4), but at the level of the stress group. They act not as
parts of lexical units such as words or morphemes, but rather as contribut-
ing to pitch contours spreading over words, phrases, clauses, or higher
units, and contributing to the signaling of speaker attitude (or attitude
which he attempts to elicit from the hearer).
In general, a rising pitch contour ending an English phrase implies to
the hearer that the speaker has not given finality to what he is saying; either
he is going to add another phrase (as before a comma), or he is hoping that
the hearer will reply (to a question of certain types), or he may be pausing
before going further. (Note that I did not say that this is the only way these
things can be signaled, or that these items always signal them; an intersec-
92 Hierarchy
tion of various items can cause these normal expectancies to be overridden,

in ways beyond the scope of this brief section. For fuller discussion, see
Pike 1945.) If, on the other hand, the speaker makes the pitch fall after he
has stressed a syllable, he may be calling attention to the importance of the
word or phrase which is stressed. If, further, the stressed syllable is fol-
lowed by unstressed syllables on that same level, there is often an implica-
tion that important things are to come or have been left unsaid (either of
lovely expectancies or of impending doom). For example, if I say, slowly,
His brain is addled, with a rise on brain, the hearer waits to see what I
have to say about that brain. If, rather, I say His brain is addled, with pitch
dropping immediately on brain, I am singling out the brain rather than
some other part (that is, I could add but his muscles surprisingly enough
continue coordinating so that he could still win at the Olympics). If, on the
contrary, I were to say with a completely level mid pitch Be careful of that
boy’s brain, I might be implying lest you damage it permanently in letting
him do that run on the bobsled without wearing a helmet.
See figure 10.2 (from Pike 1964b: 108), where contours at the lower left
of the diagonal are all falling, while those on the diagonal itself are level
and those to the upper right are rising. That is, each of these major subdivi-
sions has numerous further varieties—with, for example, those including
an extra-high pitch involving emphasis or surprise or politeness in some
way (for detail. and illustrations, see Pike 1945.) The chart as a whole
shows a set of English intonation contours as comprising a field structure
involving stress and pitch. In that figure we have used E for extra-high, H
for high, M for mid, and L for low, with a degree sign for the stressed sylla-
ble followed by a gradual stepping or gliding down or up (or, in off-norm
circumstances of special intensity, with a rapid change).
But these are just a part of the possibilities open to us. We can also have a
change point in the middle of a contour, such that we first drop from high
to low, and then, without any further stressed syllable intervening, rise (for
example, in a contour °H-M-L, as in [as for the] groove, where the fall on
groove singles it out for sharp attention, and the rise on the end of that
same syllable leaves the hearer knowing that something is to follow). In
addition, there can be unstressed syllables preceding the stressed one,
without pause, which help set the stage for the word which has been put
into focus by the stress. These pitches carry meanings also, but less strik-
ingly so, and are not shown in figure 10.2 (but see Pike 1945 for discussion
of them).
Poets, I feel, should be allowed to say what they think and feel. It should
be clear, here, that many scholars do not agree. Part of the meaning of what
a man says—often the most important part—is in the intonation. Com-
Figure 10.2. Field structure composed of English contours of stressed syl-

lables followed by unstressed ones, in various pitch patterns. The degree
sign indicates a stressed syllable; E is extra-high, H is high, M is mid, L
is low. To the upper right are rising contours, expressing incompleteness.
To the lower left are falling contours, focusing attention on the stressed
word or phrase. On the diagonal are level ones, indicating strong im-
plication of something important left unsaid as of that moment (from
Pike 1964
:108).
pare, for example, the phrase That brain of his is magnificent, spoken
with extra-high pitch on the stressed syllable of magnificent, then stepping
down to low somewhat slowly (impressively), and with appropriate voice
quality; it may imply admiration. But the same sentence spoken with the
stressed syllable on mid pitch, stepping down to low at the end of the word,
and with a snarling voice quality, may imply that the speaker does not be-
lieve it and that the man referred to is somewhat of an idiot. Surely such
differences are important to a poet. And here we argue that, since many of
these meanings are carried by intonation, the right of the poet to express
the feelings of his mind carries with it the right to mark his poems for pitch
and voice quality—and the added obligation on teachers of English and
drama to instruct their students in the mechanisms of utilizing these factors
in the writing of poetry, or in writing contrastive ways of speaking a partic-
ular bit of'a.play on stage. Yet when I have suggested this to some teachers,
they have resisted the idea, preferring to leave the interpretation rights to
the reader and denying them to the poet (whose brain is often assumed to
be incapable of this kind of creativity—which in fact may be true of some
of this generation and may continue thus until students are taught to be
literate in intonation, by being given an intonation alphabet, along with
training in its.se).
I have tried writing a number of poems marked for pitch in this way (see
for example Pike 1967a:61,70) and have made their oral reading available
through educational videocassettes (Pike 1977). In figures 10.3—4, how-
ever, I give a poem of Emily Dickinson’s marked in ways to indicate how it
94 Hierarchy
The ‘brjain within its ‘groo-ve

Runs ‘evenly and 'true:;/
Tod ratstheywater tbat,

When ‘flood:s have ‘slit the ‘hills,
And ‘scooped a ‘turnpike for them ’se-lv-es,
And ‘blotted ‘out: the ‘mi-Il-s!/
Figure 10.3. Poems may be marked for pitch, as I have done for this one
by Emily Dickinson. High pitch is represented by a line just above the
letter, mid pitch just below, and low pitch substantially below the letter.
A single slash line (as after true) suggests that the pitch stays level and
steady, rather than falling and fading away. Stress is shown by a verti-
cal stroke before the accented syllable. A raised dot indicates extra
length. This reading was by a poet (James Squires) who read it for me
very gently and slowly.
was read for me in quite different ways by James Squires, a poet, and by
Austin Warren, a literary critic (for these and a third marking, see Pike
1967a:529-—30). The reading by the poet was done very quietly with rela-
tively little variation in pitch, and with breaks between the high-level pho-
nological units shown by vowel lengthening (indicated by a raised dot after
a consonant or vowel). The reading by the literary critic was much more
dramatic, with more pauses, since he felt—he told me—that every word is
important.
But, as we have implied, voice quality, not just pitch as such, enters into
the meaning of the whole. To see how this is suggested crudely, see figure
10.5, where I indicate to the left of one of my own poems a suggestion as to
the style of reading of that line.
In order to understand the signals which a speaker sends to a listener,
then, the hearer must, in fact, react tacitly to a vast total field structure,
largely unknown to him, of intersecting dimensions of acoustic patterns re-
flecting the movement of the articulatory organs in the formation not only
of consonants and vowels but also of pitch, voice quality, rhythm, and
pause in relation to hierarchical levels.
At any one point, at any one level, for any one system, some of these mat-
ters can be captured in a notation of phonological unit-in-context—the
phonological tagmeme—applied to pronunciation. We close this chapter
by noting, in figure 10.6, just one illustration of this type (from Pike and
Pike 1977: 364), where in a poem the slot under focus of the analysis might
be the end of a line, where rhyme ties that line to others.
The |’brai-n within its "groove: /
Runs’ / "elvenly and/‘tru e;//
But? / / eta] splin|ter ‘swerve,

en —/
(f) “Twere '‘easjier for you
To put the|'walter back

(s)
When|'floods have|'sl-it|the ‘hijll-s, /
And/|"scooped a ‘turnpike for themsyelves,
(s)
And|"blotted ‘out: the ‘mi-ll-s! / /
sate
Tle,
Figure 10.4. A poem can be read with heightened pitch contrasts, as in

this reading by a literary critic (Austin Warren), one of whose specialities
was the study of Dickinson. The line levels again indicate general pitch
contrasts (not vibrations per second, nor a fixed ratio between them); the
lines between’a stressed syllable and an end point or change point are
written as level, for convenience, but in fact may drift gradually up or
down toward the change point. The raised /e/ indicates an added vowel
sound, lightly given, of somewhat the quality of cup; the /h/ is audible
breathiness (when indrawn, an arrow points to the left); the double
slash line is a pause which gives a signal of something being finished
(even though, after true, the sentence is not grammatically finished); the
pitch drifts downward, and the intensity fades. An extra-high pitch is
needed for this reading, shown by an extra-high line. The f means “fast”;
the s means “slow.”
96 Hierarchy
Voice Quality The Poem Itself:

Chanting style: "The Power is MINE!”
Detached, low pitch: The tyrant chants.
With breathiness: ' Yet he pants;
With short pause Breath/ comes/ short.

groups, and in-drawn
breath after each:
Tense vocal cords, God squeezes.
with lengthened second
word:
And pretended sneeze Sovereign sneezes... .

or two during the
pronunciation:
With nonverbal And po0f!!!
breathy exclamation,
rising to a crescendo:
Figure 10.5. Voice quality affects the meaning of a poem. Here in this
poem of Pike’s (1967c:76) the tension of the vocal cords, breathiness,
direction of audible breath intake, crescendo, speed change, and pauses
are added to pitch (not shown) for suggesting the effect desired by its
author.
Slot Class
End of a poetic line, ina Words ending in a certain set
rhyming poem of -VC syllables
Role Cohesion
With the function (in elegant Controlling some agreements
poetry) of calling attention to of form. For example, in Dick-
semantic relations which inson, (h)ills with (mills, and
might not otherwise come to (tr)ue with (vou
attention
Figure 10.6. A phonological unit-in-context is a tagmeme where pronun-

ciation is involved. The four features (already treated extensively for
grammar) show up here on the level of the line-as-a-unit, in a set of lines
and words tied by rhyme.
The Referential Hierarchy
We have already mentioned, either directly in relation to the

referential hierarchy, or indirectly in relation to meanings or behavior ac-
tivities which relate to that hierarchy, a number of items: In figure 2.4 the
meaning of jump is shown as related to the social and physical back-
ground. In figure 4.4 the Christmas dinner scene is shown as a continuing
wave nucleus over time—which, here, we would treat as a referential con-
stant. The specific identity of Joey as long-nosed, is referential (fig. 6.1),
as is the identity of the “same” river in spite of molecular replacements
(sec. 7.1), and the girl with changed hairdo (fig. 7.1). So too is the contrast-
ive feature array of semantics for king and queen (fig. 6.6), or big sister
(fig. 6.7). These feature arrays are a kind of representation of a field struc-
ture within the referential relations, as the Christmas dinner represented
a perceived referential wave structure, and as the identity of Joey or of
the river captures a component of referential structure in relation to a par-
ticle view. , _
11.1. Levels in the Referential Hierarchy
In the grammatical hierarchy we made a sharp break between

a proposition of some kind (in the form of clause or sentence) and the men-
tion of a word (for example, a noun labeling a thing, or person, or item);
the proposition was considered relatively more independent, and the term
relatively more dependent (see secs. 9.2—3). In the phonological hierarchy
there was more independence attributed to the syllable than to the pho-
98 Hierarchy
neme, in an analogous way (sec. 10.2). Now, for the referential hierarchy,
we treat as more independent the telling about an event than the mention
of a name of something. Thus the named person Socrates is treated on a
lower hierarchical level than the action affirmed in the sentence Socrates
drank himself to death; the event as described is treated as higher in rank
than the items involved in the event and mentioned in connection with it.
At the lower of these two levels one has, then, the members of the cast
of a play or of an event. And these members are—in the normal, basic sit-
uation—particular members mentioned in a particular play or a partic-
ular event. The particular members of the cast may be—if well enough
known—entered into an encyclopedia. There one could find various Henrys
of history—each of them a different person; or various Johns—for exam-
ple, as different popes. Each would be a different referential entity, al-
though the name would be the same lexically, and hence identical in each
occurrence in so far as its membership in a grammatical class (personal
noun) is concerned.
But there is a further, crucial difference between grammatical and refer-
ential units at this point. The lexical item Socrates must remain as a noun,
the same noun, in all its occurrences. But as simultaneously representing—
for the moment—a referential item, it has great flexibilityofparaphrase
without loss of its referential identity, even though it is no longer a noun.
Thus one can say You know, the man I was talking to you about that Plato
wrote about, or I refer to Plato’s teacher who had to drink the hemlock, or
even You know who I mean, or the great Socrates, or he. Notice, however,
that these are accepted by the hearer (or reader) only ifthe specific context
warrants it and allows the participants (speaker and hearer) in the discus-
sion to be satisfied that the same person is being talked about. It is their
(emic) judgment of the identity, not that of someone who is an outsider to
the conversation, which provides the appropriate criteria for the referential
identification involved.
Note further that the fact that a unit of the referential hierarchy can be
paraphrased in numerous different grammatical forms is crucial evidence
that referential and grammatical hierarchies are not the same, since they
often are nonisomorphic (for phonology versus grammar, see fig 10.1).
Their forms are often different (even though necessarily homophonous in
some of their manifestations), since every manifestation (mention) of a ref-
erential unit must be in some grammatical form; the difference is that the
emically same referential item may be in many different grammatical
forms.
On the other hand, there is a parallel: As grammar can have the paired
levels of word and phrase, both of which are terms, so the referential level
The Referential Hierarchy 99
of identity (Jones 1977) can have a particular member of a cast belonging

to a larger group in the cast: Socrates was a member of a philosophical
group, and as a child he belonged to some kind of biological and social
family group with father and mother. At the higher level too there are par-
allels. As the grammatical level of proposition can have a clause (which if
independent and unmodified can be simultaneously a particular kind of
simple sentence, by disjunct definition) paired with complex sentences of
various types, so an event may be part of a larger event, which in turn is
part of a still larger event. In addition, a more complex situation frequently
is relevant on the higher level: Two “separate” events may be occurring at
the same time, which later turn out to merge into a single event. (Boy [with
one history] meets girl [with a different history]; they fall in love, marry,
and live happily ever after.) Such a merging event complex also serves as a
unit of the referential hierarchy analogous to the complex sentence (or even
higher unit) of grammar.
Background expectancies can be forced to merge in the hearer’s mind by
the speaker’s use of a pun. Thus Socrates drank himself to death merges
the expectancies raised by the mention of the dialogues of Plato with the
expectancies caused by the mention of excessive use of liquor. They can
also merge, on a much more serious level, when an author acts as a mem-
ber of a cast in the performance of a play which he has himself written—a
kind of literary incarnation in his own creation. (For an elegant referential
analysis of a Carib, Guatemala, folktale in which a related situation is ana-
lyzed, along with extensive hierarchical treatment of the housing arrange-
ment, social involvement, and tagmeme formulas of a referential type, see
Howland 1981.)
For a lower level of the referential hierarchy than that of the identity (or
person or prop or item), I tried using contrastive features of a unit (that is,
elements of its physical or semantic componential analysis) as seen by the
analyst or as seen and commented on by a native actor serving as a lay
analyst of.some item on his own culture. There is no break in the theory
from the discussion of the philosopher or linguist to the discussion of the
man in the street or the preliterate speaker of a language. (See figs. 5.2-6
and 6.6—7 for charts whose rows and columns are labeled for features of
phonology, of grammar, of semantics, or of poetry. An extensive summary
of the bibliography of semantic features is found in Nida 1975.) That is,
once an item is abstracted from the normal contexts of its being used and is
focused on by an analyst for discovering or describing its inner properties,
the named concept, in its various paraphrases, thus abstracted becomes an
element of the referential hierarchy at the level of identity (cast, person,
item), whereas its contrastive features in turn become elements of a level
100 Hierarchy
below that one. Recently, however, Evelyn Pike has felt that we should treat
contrastive features as components of units at every level of each hierarchy.
Then, for the level below that of identities in the referential hierarchy, we
might place relationships. This might include, among others, the relation-
ship between an attribute and an entity, or a geographical relation be-
tween two entities, or the direction of action of an entity toward or from a
place or thing.
11.2. The Referential Tagmeme
Just as slot, class, role, and cohesion were useful in the analysis
of grammar (fig. 9.3) and of phonology (fig. 10.6), so the scheme is useful
here. Purpose, for example, finds its place in the role cell, in the descrip-
tion of an event; and so likewise does the relevance in the description of a
member of the cast or of a prop. On the event level it is relevant, for exam-
ple, that a joke is intended when one says Socrates drank himself to death.
On the level of cast, likewise, it is relevant to know who is the hero—or the
villain—in a film (and unless one does know, one cannot understand the
outcome of many stories or films). That is, the role cell may be filled explic-
itly by the analyst, or tacitly by the naive reader; pure formal analysis with-
out reference to meaning or purpose or relevance cannot provide an analy-
sis which can satisfy those who wish to know sensibly “what happened.”
Such roles can be complicated. An event may have a purpose leading to a
larger purpose; or motives may be mixed (dual). And a person’s role may
appear different from the point of view of different observers, each of
whom has his own interpretations of (or guesses about) the purposes of
others. Thus, in the referential realm, we may sometimes discuss the ac-
tions of different observers separately, since their purposes are different
(see, for several different levels of observer, Pike 1981).
An event as described may enter a marginal slot in a series of events
leading toward the crucial, nuclear event under attention. The item filling
that slot of the description is one member of the paraphrase set identifiable
by the speaker as (emically) the same from his point of view, and as accept-
able as a replacement for it in this place in his account, or in other places of
that same account (see Fig. 11.1).
But for the narration of an event to be intelligible to a hearer, there must
be a degree of cohesion of actual expectancies of the hearer with the expec-
tancies expected of him by the speaker, and a degree of coherence—if the
hearer thinks that facts are being recounted—with the view of reality as
believed in by the hearer. If the speaker is perceived by the hearer as having
no such coherence (and no joking intent, for example), the hearer may
WHERE the subevent occurred, as nu- WHAT happened, emically defined as

clear or marginal to the including being members of a paraphrase set ac-
event: here, nuclear, as a subevent of ceptable to the narrator:
the larger event of The meeting at the
prison (69): when he drank the poison [67];
Then raising the cup to his lips,
in the prison with Socrates on the quite readily and cheerfully he
day when he drank the poison [67] drank of the poison [159].
the crowning point at the end of a di-

alogue which included a discussion of
the possibility of avoiding it.
WHY the actor performed the event— HOW the event coheres with the un-
or the cause deduced by the narrator: derlying belief system of the narrator,
or of the hearer as expected or re-
I have spoken many words in the
ported of him by the narrator (or with
endeavor to show you that when I
truth as seen by the “outside” analyst
have drunk the poison I shall leave
in relation to some other frame of
you and go to the joys of the blessed.
reference):
[156]
Why should he [the real philoso-
In this discussion, furthermore, his
pher] repine at that which he had
purpose was to put truth above life
been always pursuing and desiring
and above his own reputation; and his
[i.e., death and dying] [76], with
wish was to push others to truth:
death in which the soul is released
And I would ask you [he said] to be from the body and the body is re-
thinking of the truth and not of So- leased from the soul. [77]
crates; agree with me, if I seem to
you to be speaking the truth; or if
not, withstand me might and main,
that I may not deceive you as well as
myselfinmy enthusiasm, and like
the bee, leave my sting in you before
I die. [120]
Figure 11.1. A referential event tagmeme in which the subevent occurs as

merely one tagmeme in a series which represents the larger event: the
dialogue offriends with Socrates, when in the jail they tried to persuade
him not to drink the hemlock. The larger particular-event setting gives
the slot. The larger background setting gives cohesion to—and in part
controls—the whole, and makes sense out of it. The specific purpose of
the specific actor under attention (Socrates) provides the data for the role
cell. A set of paraphrases are taken from Jowett’s translation of the Di-
alogues of Plato, or from the introduction to it by Kaplan (1952); page
numbers refer to that edition.
102 Hierarchy
WHERE the member of the cast fits a WHAT the cast member is named, as a
larger unit of the cast: member of a paraphase set appropri-
ate to this narrative:
Socrates as the nuclear member of
the discussion group, in the prison, } Socrates, a friend, he, him, his, I,
along with Phaedo, Apollodorus, me, my, you, yourself; [Kaplan: ]
Simmias, Cebes, Crito, and an ethical leader of the young and un-
attendant, flagging conscience of the rulers and
citizens of Athens [ix], the main fig-
ure in most of the dialogues [xi], the
seventy-year-old Socrates. [2]
THE FUNCTION OR ROLE of the cast HOW the cast member is related to an
member as he sees himself, or is seen underlying belief system which affects
by others, or by the narrator: his actions:
of all men of his time whom I have therefore let the cup be brought
known, he was the wisest and justest . . . for I do not think that I should
and best. [160] gain anything by drinking the poi-
son a little later; I should only be ri-
diculous in my own eyes for sparing
and saving a life which is already
forfeit’ [158]; ‘I ought to be grieved
at death, if I were not persuaded in
the first place that I am going to
other gods who are wise and
good ... I donot grieve... forI
have good hope that there is yet
something remaining for the dead,
and as has been said of old, some
far better thing for the good than for
the evil. [75]
Figure 11.2. A referential person tagmeme, the principal speaker in the

dialogues of Plato; characterized in the Phaedo as the nuclear member of
a small discussion group in the prison at the time of the death scene,
with the role of a wise man who has a belief about afterlife which affects
and controls his sayings at the time of his drinking the poison. The par-
aphrase set which includes Socrates includes also the pronouns used by
himself and others, as well as indirect reference to him as a friend and
the characterizations of him in notes by Kaplan. Other tagmemic charac-
terizations could have been made of the prison itself, the poison, and the
setting as a whole. (Page numbers refer to Kaplan 1952.)
judge the speaker to be insane or lying. Here, then, truth and falsehood,
relative to a given frame of reference, find their way into the description.
Each referential level will have its own tagmemes. Like an event, so also
a member of the cast may be represented in relation to slot, class, role, and
cohesion (see fig. 11.2 for Socrates treated this way, in relation to the di-
alogue Phaedo).
11.3. Some Grammatical and Phonological

Alternatives in the Representation
of Referential Talking Space
We assume here that a person lives in referential space—in in-

terlocking tacit or implicit frames of reference of time sequence, spatial
array, physical relationships, logical coherence, social structures, psycho-
logical involvement, belief systems about reality, intersecting events, and
others. No person can bring into words all at one moment the content of
each of these in all its detail. Selection must be made if one is to talk at all;
the balance must be left unstated at that moment, even though it remains as
a strong force controlling much of the talk.
Furthermore, within any one language or culture the ordering of the
talking presentation of any one of these frame types will be considered here
the normal (or basic) order when it is the first learned in childhood or
most frequently used later. But off-norm forms of presentation may occur
when the speaker has some reason (tacit or explicit) for selecting a less
basic item for highlighting, since at that moment, to him, it seems impor-
tant beyond its normal place in the hierarchical or sequential or matrix
structure of the frame as a whole. When this happens, the special focus
desired can be obtained by special grammatical arrangements, or they may
be forced, in part, by the place of the item in the larger discourse or by the
emotional reaction of the speaker.
In a classical syllogism, for example, major and minor premises are fol-
lowed by a conclusion: All men are mortal; Socrates was a man; there-
fore, Socrates was mortal. A belief system underlies the premises of such a
syllogism. If this belief seems to someone to be challenged, and if he is out-
raged by it, the syllogism may be reordered, as in the following interaction
between two speakers:
First speaker: Socrates lives on!

Second speaker: Socrates? He died! He was a man, wasn’t he? (And
don’t all men die?)
104 Hierarchy
s O 515 29324055 CM7, fa), ) 0) lsh a ake ;
9+4=13
Figure 11.3. The truth of a statement is not to be confused with the cor-
rectness of its grammatical form; it must rather be judged relative to the
frame of reference within which it is stated. Here, the addition of 9 + 4 in
relation to the model of a line gives 13 (A); on a clock face of 12 hours,
it gives 1 (1:00) (B); on a clock face with 0 and 2 hours, 2 + 2 gives 1 (C).
(Metaphor, likewise, may carry truth judged relative to the joint expec-
tancies of speaker-hearer acting within the same frame of reference of
expectancies, intent, and literary genre.)
The grammatical structure is radically altered in the utterance of the sec-

ond speaker by the order of its elements (with conclusion first), by the
change from statement to question, by the optional occurrence of the first
premise (as shown by the parentheses), by the change of focus (with atten-
tion to the man and to the fact of his death), by the emotional overtones (of
expostulation), and by the deliberate or pretended misunderstanding of the
first by the second speaker (where the second takes the first literally rather
than figuratively).
The underlying syllogism is unchanged; but the change of speaker-to-
audience relations, implicit in the second telling, is a referential change
superimposed upon the syllogism itself. (Here we have referential levels of
observer relevance; see Pike 1981). Here, as in many other instances, a
change in pronunciation (for example, the exclamatory form) may help
alert the listener to know that an off-norm form of grammar or of reference
or of their interrelation is occurring.
11.4. Reality Relative to a Frame of

Reference, versus Grammatical Form
It is important not to confuse correctness of formation of a sen-

tence grammatically with truth relative to a particular referential frame-
work. A statement can be factually correct, or true, even though badly
stated. And a lie can be elegantly framed by a con man to cheat a victim. In
terms of the theory presented here, it is grammatically correct to say Plato
drank the poison—fully as correct grammatically as to say Socrates drank
the poison—but false to our referential belief about history. The primary
student, then, must not be accused of bad grammar when his word ar-
rangements are impeccable but his sums are faulty (see fig. 11.3, where
nine plus four equals thirteen or one, relative to a particular frame of
reference).
It is important, therefore, in listening, to seek to understand the intent
and general frame of discourse of the author. Jones (1977:v), for example,
discusses theme in expository discourse “in terms of nuclearity in the refer-
ential hierarchy”; and she treats theme as “referential prominence” but
focus as “grammatical prominence” (pp. vi—vii). She also deals with “gram-
matical devices for highlighting theme” (pp. 169-223).
Understanding may disappear for listeners when respective frames of
understanding of the universe differ for speaker and listener, but it can be
the source of shared humor when both speaker and listener know that de-
liberate incongruity of frames of reference is involved (see fig. 11.4).
106 Hierarchy
A. The moon is more important than the sun—the moon gives light by
night when it’s needed—the sun during the day when it’s light
anyhow.
B. A drunk, running around a tree three times: I’m lost in the impen-
etrable forest.
C. Psychiatrist: Tell me sir, why are you standing on one foot?

Patient: If I lifted the other one I’d fall down.
D. A man, to the barber, when he had been nicked by the razor: Give me
a glass of water, please.
Barber: Why?
Man: I want to see if my neck leaks.
Figure 11.4. Unless the frame of reference is shared and coherent, there
may be misunderstanding, error, or humor. In A there is implied lack of
knowledge on the part of the purported speaker, understood by the lis-
tener as being humorous error. In B the misunderstanding of the physi-
cal situation by the speaker is attributed by the hearer to the influence of
liquor on the speaker. In C the speaker is assumed to be out of touch
with reality for mental reasons. In D we are supposed to imagine the
barber listening to the wry humor directed at him, involving a deliber-
ate—and caustic—implication of known (shared) implicit impossibility.
In E there is further subtlety—the listener pretends to understand, pre-
tends to reinforce sympathetically the speaker’s claim, but in fact jabs
back with an implicit vicious attack.
Context
Introduction to Part IV
A field view of a discipline as an intersection of various param-

eters of analyst viewpoint on the world outside of him has the great advan-
tage of flexibility. One can come at the data from various theoretical or
empirical directions, and therefore mine a richness and multiple interlock-
ing cohesion of principles which is lacking if one attempts to follow a single
rigid logical sequence of presentation (or a single presumed real taxonomy
or rule structure of objective reality) apart from observer choice (or the
arbitrariness of temporary analyst-observer interests changing according
to the needs of the moment or of the era). On the other hand, for a person
who wants a tightly organized nonoverlapping approach such a view is
disconcerting. Nothing in it seems stable, everything seems everything else,
and observer freedom seems to be the death of scientific detachment and
precision. Even the fact that over the eras the views of scientific truth
change radically (and the certainty of then-certain facts) does not deter him
from obtaining intellectual peace by a commitment to a scientific truth
which will not last—but which for the moment is popular (or, stated more
elegantly, is the current dominant paradigm—compare Kuhn 1962).
One can avoid many problems of the moment and pacify the snarls of a
dominant scientific paradigm by agreeing to accept its overpowering pos-
tulates. (“Do you want to know a sure way to conciliate a tiger? Let him
swallow you.”) But if those postulates cancel the possibility of including the
observer himself in the situation and among the data observed, and hence
of observer flexibility in choice of viewpoint and in the possibility of adopt-
ing different viewpoints for different purposes of the moment, one has
allowed his scientific freedom to be swallowed up and science itself to
become nothing but an unattainable abstract ideal in a dreamed-of un-
reachable reductionism, or else in an abstracting idealism which splits me
from you, us from things, and things from human knowledge.
The next three chapters cover some of the same data as seen earlier. This
is inevitable, in tagmemic theory, since each of its perspectives covers all
the same data—as we began to show in the chapters on particle, wave,
110 Context
and field. Here, then, I add emphases rather than new material. In Chap-
ter 12 there is emphasized the necessity of keeping form and meaning to-
gether—in contrast with some other schools of thought. In Chapter 13 I
discuss change again, but with emphasis on the fact that change does not
occur in the absence of a framework (or of bumping, or merging, or shar-
ing, or some other way of bridging the gap between the items involved). In
Chapter 14, higher-level frames are under attention—the universes of dis-
course within which rational talk can occur.
Form-and-Meaning
In this chapter I am rejecting all views of language or life which

claim that abstract forms comprise basic reality. Here, as over against cen-
turies of work by some scholars, form is a term used to refer to concretely
structured elements which include a physical component, not idealized
disembodied realities, essences, ideas—or even rules or relations. In addi-
tion to form, a meaning, relevance, value, significance, deduced cause, re-
sult of deduced cause, or some other observer-related component is always
demanded in our affirmation of the existence of any unit of rational behav-
ior, of the existence of the observability of any concrete object or event, or of
any object or event as deduced by man or imagined by man.
12.1 Meaning as Essentially Tied

to a Physical Component in a Unit
Each unit (see Part II) of each level of each hierarchy—gram-

matical, phonological, and referential—has a physical component. In tag-
memic theory there is no semantic hierarchy abstracted away from all
physical manifestation; the referential hierarchy is not a semantic one per
se, but is a different and simultaneous structuring of the same stuff, which
is also structured by the grammatical and phonological hierarchies (see
also the combination of form and meaning as discussed in sec. 2.5). Sup-
pose, for example, that we read aloud onto tape the sentence Socrates died
after he'd drunk the poison. If we were to scissor and snip away each
112 Context
sound (or syllable or stress group) one after another, what would be left?
Nothing. If we take a second copy of the tape and snip away the grammati-
cal dependent and independent clauses (or the conjunctions, subjects,
be left? N othing. If, from a third copy,
predicates, and objects), what would.
we snip away the stated causing subevent (the drinking of the poison,
which happened—but was not stated—first), then the stated effect sub-
event (the dying, which happened second but which was stated first), what
would be left? Nothing. There are, then, in this view three separate struc-
turings of the same substance, but structurings which may or may not have
the same beginnings and endings or sequential (or other) relationships of
all or of some of their parts (for nonisomorphic grammar and phonology,
see also fig. 10.1).
12.2. Hierarchically Extended Lexicon as the

Substance Manifesting, Simultaneously,
the Forms of the Three Hierarchies
The substance shared by grammar, phonology, and reference

we call lexicon. But the term carries more meaning, in this context, than
that of mere dictionary listing; rather, it includes the particular utterance
or part of an utterance under attention at the moment, provided that it is a
unit containing at least one morpheme. It may be a single morpheme, as
had, or a variant (allomorph) of that morpheme, as ’d; or a phrase such as
had drunk; or an entire particular sentence (not the abstract pattern of a
sentence); or an entire poem or discourse in its particularity. That is, by
hierarchical extension of the lexicon we mean the particular language-
manifested elements at any level of the hierarchy, whether dictionary en-
tries (lexicon proper), or phrases, sentences, discourses, or conversations.
Meanings in language as used by the man in the street are treated as
features relevant to behavioral impact, including the elicitation of under-
standing. In grammar, these may include the meanings of role (function) of
a tagmeme (as in secs. 9.4—5, figs. 9.3-4). In phonology, it can come from
the impact of voice quality (see fig. 10.5). In reference, it includes the
meaning of lexicon (fig. 2.4), contrastive features of lexicon (figs. 6.6—7),
or meanings of particular words, phrases, sentences, poems, discourses, or
conversations. Each of the units manifesting each level of each hierarchy
carries some kind of behavioral impact. Here, each of these is treated as a
meaning of some kind, as part of a concrete unit manifesting the form of
one of the hierarchies.
In figure 12.1 I try to suggest this relationship (building on suggestions
Form-and-Meaning 113
language
utterances, in- * 8
e cluding lexicon o %.
Ro
Ye
proper and hierarchi- % %
cally extended lexicon
Class(es)
Cohesion
Phonological Tagmeme
Figure 12.1. Hierarchies of reference, grammar, and phonology are not

abstracted from shared substance, but retain the physical component
contributed by the particular speech event itself. An emic utterance is a
behaviorally meaningful simultaneous combination of three structures
which share their momentary lexical manifestation including lexicon, ex-
tended hierarchically to embrace the reading of a poem or the presenta-
tion of a lecture.
from Evelyn Pike—see Pike and Pike 1977:365) by a triangle labeled as

this lexical shared substance of language utterances, but with tagmemic
four-celled rectangles sprouting from each of the sides, representing the re-
spective hierarchies. The dots on the triangle sides suggest that the lexical
item is part of all three of the hierarchies at the same time. It is a concrete
form-meaning behavioral composite we are dealing with, not a set of ab-
stractions away from substance. (A comparable point could be made by
using the figure of a cube; its three dimensions would represent the three
hierarchies, intersecting in specific utterances, which represent the three
dimensions simultaneously. )
114 Context
Reference
12
Phonology
A.Mary: I didn’t say yes to John the first time he proposed.

Susan: Yes, I know — you weren’t there, my dear.
B. (Mother puts diaper on baby, without powder) Child: Mommy, you
forgot to salt him.
C. What is it that the more it dries the wetter it gets? A towel.
D. Why is the electric chair period furniture? Because it ends a sentence.
E. Why do Swiss cows wear bells? Because their horns don’t work.
F. She must be in her middle flirties.
G. What is a cross between an abalone and a crocodile? A crock
of bologney.
H. Why are three leaves from a stone? Because an elephant sighs.
Figure 12.2. Prominence given to the hierarchies differs from one joke to
another. Any one of the three may be exploited for the unexpectedness of
its relationships; or any combination may be used. In A, no pun—no
phonology—is involved, but the implication—not expected by Mary—
is that John had earlier proposed to someone else. In B, the child con-
fuses talcum powder with salt, referentially, and expresses this in a
referential-grammatical idiom. In C, the term dries is grammatically
first intransitive, then transitive (that is, it dries something). In D, there
is a triple pun (period as a unit of time or punctuation, sentence as a
grammatical unit or a judge’s verdict, ends as in ending a life or a writ-
ten utterance). In E the attention shifts from pasture needs to the pseu-
doneeds of warning by blowing horns, by way of the phonology (the pun).
In F the phonology mixes flirt with thirties, making up a new lexical
item, flirties. In G there is a switch from two words (abalone and croco-
dile) to a noun with a phrase modifier (a crock of baloney) by way of
partial phonetic similarity, fused. In H there is no natural or focused
coherence of grammar, phonology, or reference between question and an-
swer, so that the noncoherence is itself intended to be puzzling, and
hence “funny.” Some people would call this a shaggy dog story.
12.3. Humor via Focus on Intersecting

Form and Meaning in Hierarchies
The relationships, however, are much more complex than can

be seen in figure 1. If we ignore for the moment the details of the four-celled
elements and treat their relation only in relation to the momentary promi-
nence of one or more of the three hierarchies, we can also show this inter-
section by means of a Venn diagram, as in figure 12.2. Humor is there clas-
sified in relation to the kinds of hierarchical sharing which are especially
in focus for a particular joke. (All three hierarchies are present in each
joke, but one or more are selected for special attention.)
Query: Why is getting up at four o’clock in

the morning like a pig’s tail?
Answer: It’s [twfli].
Here, if twirlly and too early are deliberately pronounced alike, the single
phonological answer to both implies two different simultaneous lexico-
grammatical-referential answers, referring respectively to the two parts of
the question.
12.4. Form and Meaning Interlocking in Poetry
Poetry, like humor, exploits the relations between the simul-

taneous hierarchies. It uses phonology to force to attention (of the compos-
ing writer, then the reader) certain normally unexpected connections which
may carry pleasure, shock, or depth of understanding. In figure 12.3 I give
a poem of mine in which referential connections are suggested by rhyme
and in which grammatical form is affected by the pretended discourse
structure (that of a telegram, cutting out various minor words). As for
grammar, for example, note the lack of a subject before the verb need (even
though it is not an imperative clause), justified by the telegraphic form; and
note lack of subject and verb auxiliary I am before feeling queer. See also
the balance—an essential property of poetry—of hoping and waiting, as a
kind of grammar rhyme. As for phonological elements, there is not only
rhyme balancing the second and third lines of both stanzas (here—queer,
quick—sick) but a tie to the referential implication (that is, implication:
Since yow’re not here, Ifeel queer; and Unless you come quick, P’ll get sick).
Unless the rhyme forces attention to referential meaning to lock words into
a universe of discourse and cohesion in some way, rhyming poetry can be-
come mere jingle. (Note, also, that the capital letters help suggest the tele-
116 Context
TELEGRAM
CAN YOU COME

NEED YOU HERE
FEELING QUEER
HOPING.
GRAB A TRAIN
MAKE IT QUICK
’LL GET SICK
WAITING.
Figure 12.3. A poem interlocks form and meaning in an intricate complex
vastly more variable than can be handled by a fixed set of rules or cross-
hierarchical mappings to predict them. The poem is by Pike.
graphic format.) Referentially, in addition, the first lines of the two stanzas
balance, with parallelism of come (using a query as a semicommand) with
grab a train (as direct indication of command to implement the implicit
command-request). Similarly, the suggestion of illness (need you . . . feel-
ing queer) is more fully developed in the second stanza (I'll get sick waiting).
But in poetry as in jokes, in seriousness as in humor, I do not treat an
abstract semantic component divorced from the concrete body of expres-
sion. Meaning is tied to form, so that it exists only where there is form of
some kind; it is never postulated either apart from form or apart from an
observer who emically structures the concrete or pretended material which
he sees, imagines he sees, deduces the presence of, or creates in nightmare
or in science fiction. The interrelationships of grammatical arrangement,
phonological manifestation, and referential emically significant event, per-
son, or prop are tied to the presence of an observer as part of the total situa-
tion. The linguist is just one observer among others, including the native
speaker.
In handling data such as those implied in the appreciation of a small
total poem (not just its separate sentences), its concrete, physically man-
ifested words, sentences, rhymes, and background experiences are rele-
vant, along with its meaning, implications, author intent (where known),
and reader interpretation (where known). This leads to differences be-
tween tagmemic theory, with its emic generalizations and notation, and
those theories which offer a quasi-mathematical abstraction of rules or
symbolic systems away from such concreteness and observer relevance.
The difference is clearly seen, perhaps, by contrasting this preference of
mine with a statement by Chomsky (1962: 129), in which symbols only ex-
plicitly comprise his philosophical and procedural focus:
Motivated now by the goal of constructing a grammar, instead of a

rule of procedure for constructing an inventory of elements, we no
longer have any reason to consider the symbols ‘NP’, ‘sentence’, a
‘VP’, etc., that appear in these rules to be names of certain classes,
sequences, or sequences of classes of concrete elements. They are
simply elements in a system of representation which has been con-
structed so as to enable us to characterize effectively the set of En-
glish sentences in a linguistically meaningful way. [Italics added]
Compare also Chomsky and Morris Halle, in a preface to Chomsky 1966:

“the purely formal aspects of language, envisaged as a mathematical ob-
ject” [italics added].
My own view seems to be more related to an approach by a physicist
than to one by a mathematician or logician. Purely formal aspects of a
poem—or of any other manifestation of language—do not satisfy my inter-
ests, excitements, or applied needs, no matter how helpful they have been
and continue to be in deepening the understanding of the discipline in
many phases of language. I seek, therefore, for a balance of attention be-
tween form (including physical and systemic relations and contexts) and
meaning (including behavioral impact from the three hierarchies of gram-
mar, phonology, and reference). The four-cell tagmemic notation is a com-
ponent of our attempt to tackle these concerns.
I would have comparable reactions to any view of language, therefore,
which might treat it as exclusively made up of relations of relations of rela-
tions (with no given theoretical stopping point) or to a view of language
which assumes that one can first treat all of meaning, and then and then
only appropriately study the relation of that meaning to manifesting form.
Tagmemic theory affirms that language contains and is composed of emic
form-meaning composites.
Sharing as Prerequisite
to Change
We live in a universe of physically flowing, merging, smearing,

overlapping perceived entities. We live in a universe of waves (Chapter 4).
No view which deals exclusively with perceived sharp-cut units, things,
classes can handle Heraclitus’s river (see sec. 7.1), the merging of pho-
nemes (figure 4.2), or the intersection of phonology, grammar, and ref-
erence (Chapter 12). Logical partitioning has a necessary place in per-
ception, when the observer “sees” particles (Chapter 3) as if they were
completely separate. But it must not be allowed to become the logical tiger
which swallows us (Introduction to Part IV).
13.1 Change Involving the Sharing of Particles
When two syllables are viewed as a pair of waves in sequence,

rapid speech may make them merge at their borders. This can lead to tem-
porary or permanent change in the phonemic content of morphemes man-
ifested by syllables in such a way that a new sound—not present in either,
under different or slower conditions—is shared. Thus in the phrase as you
like (/zez yu 1a‘k/), both /z/ and /y/ may disappear and be replaced by /z/
(as in azure). The two morphemes, and the two syllables, both share the
one newly present sound. Here the meaning is unchanged, and no unex-
pected shift of attention occurs. Speed-induced fusion such as this, if re-
tained, may lead to a permanent change of morphemic phonological shape.
In a pun, on the other hand, the fact of sharing is a deliberate part of the
communication—which by design forces a change from thinking of one
imagined or expected situation to thinking of another. Suppose, for exam-
ple, that we ask:
Sharing as Prerequisite to Change 119
Catt’ll = cattle
‘Why is a mouse like a pile of hay?’

‘The cat'll eat it.’ or ‘The cattle eat it.’
Figure 13.1. In a pun, two different sentences can share the same form of
pronunciation. Out of context, the utterance manifesting either one is
ambiguous with the other. In a pun, the ambiguity is designed to force a
change of attention from one area to another (expected to unexpected) or
to force a vibration of attention between two possible understandings.
Question: Why did the widow with two

sons name her ranch “Focus”?
Answer: That’s where the sons raise meat.
Or: That’s where the sun’s rays meet.
The alternative answers may be pronounced exactly alike. That is, in the
reply, the single utterance (physical item) could be the manifestation of ei-
ther of two different sentences (form-meaning composites); the two share
the same phonological particles. (For another illustration, see fig. 13.1.)
The genre relies on this fact.
13.2. Change Involving the

Sharing of Wave Components
From a different perspective, the preceding illustrations can be

viewed as the sharing of waves—since a sound can be seen as a wave, as
can a syllable. So I turn to a much different kind of wave—a wave of mean-
ing of a lexical item, which was mentioned earlier in connection with run
in section 4.3. There, the lexical meaning of a word was viewed as change-
able, with a central (normal) meaning, and one or more marginal mean-
120 Context
ings. Here, however, we are interested in its meaning variants as involving

the influence of context.
The central meaning will usually be considered the one which is learned
earliest in life, is used most frequently, is most physical in its reference, and
is used analytically as the most convenient basis for descriptive order or
rule derivation. When the criteria clash, different analysts may reach dif-
ferent conclusions—which will in general be mechanically convertible the
one to the other, communicable across theories or descriptions, and each
useful in its applied context. Such judgments are usually—and usefully—
made on the basis of the intuition of the analyst or of the native speaker,
subject to changing judgments as more instances are brought to attention.
The centrality versus marginality of meaning—the wave character—of
run can be suggested by figure 13.2. In its marginal usages, the word run
is found to share (or be modified by) some component of meaning implied
in the context containing it; simultaneously, the word may affect the mean-
ing of the contextual words. When an object is added to a clause, run may
be transformed metaphorically from (a person himself) running to a place
into (a person) forcing something which has no feet to move; the added
grammatical transitivity there affects the meaning of the verb, as in to run
_- Central meaning, learned early in life,

ae more frequently used, often physical
in reference, less modified by
context.
run—on feet, _ Marginal meaning, changed by the

to or from a presence of an object, forcing a tran-
place sitive interpretation, grammatically,
of the verb.
run a business
- Marginal meaning, often metaphoric,
run a temperature “7 frequently heavily dependent upon
specific conditioning context.
run to seed ~
Raat Still further marginality, rare in
the language as a whole, though
perhaps—as in slang—frequent for
a special style.
Figure 13.2. A change in meaning-viewed-as-wave is forced by its shared

context, whether that be physical, mental, social, or linguistic.
Oh,
so that’s
why they
call them
pigs.
A. Mother’s admonition to B. Child’s comment, upon seeing

city child. pigs, later, in the country.
Figure 13.3. The sharing of a referential component of eating style can

lead to a metaphor, a socially marginal meaning of a wave of meaning.
a business. The change can go further, to instances where the running and
the item changing it are both metaphorical, as in to run to seed, where
seed has little to do with the person affected, but suggests that life’s produc-
tive period has ended.
The background (physical, mental, social, linguistic) within which a
word (or other unit) has been learned may force a change from the socially
accepted central meaning to a personal marginal one. The choice of native-
speaker norm is often dependent upon the history of the speaker’s own ex-
perience, or solidified into a dead metaphor by the past experience and
usage of the community, when accompanied by loss of the original en-
vironment stimulus for the old change. In figure 13.3 the context is the city
training of the child versus the country training (considered to be the so-
ciety norm, in this instance); the original marginal meaning has now be-
come the central one for the off-norm child. The word pig as meaning an
animal has come to mean, for the city child, a sloppily eating child. The
city child’s experience of hearing the term has been related to contexts in
which its eating style has been objected to by the parent. The child then
uses its particularized interpretation of norm to interpret the meaning of
the word pig when it senses the physical characteristics shared by pig and
122 Context
self. (Based on a lost reference somewhere in Bloomfield’s work—but re-

membered because of shared amusement at the shift in sharing. )
13.3. Change Involving the

Sharing of Field Components
In Chapter 5 we showed field structures in terms of two-dimen-

sional diagrams, with contrastive features as rows and columns for sounds
(figs. 5.2-4), grammatical clauses (fig. 5.5), and for a referential poetic
structure (fig. 5.6). Later we used comparable charts for showing contrast-
ive semantic features of words (figs. 6.6—7). If one restricts oneself to just
one entrance into such a matrix, one can transform the chart into a tree
structure. This has the advantage that one particular set of relationships is
selected and clearly diagramed and gives the basis for a useful taronomy
of part embedded in whole, with species differentiated from genus. Such
taxonomies are used again and again in scientific work of many kinds. In
addition, they can be used to describe the talked-about world of a group of
speakers whose native taxonomy is very different from our own. In figure
13.4 a brief suggestion is given of the way such an approach can be used to
partition a universe of experience.
A difficulty with such diagrams, however, may appear when certain
characteristics are seen to be shared by two branches of the tree after they
have been separated; if the branches are twigs far separated, the feature
may be awkward to describe. The implication is that such a simple tree
approach is inadequate for representing some of the complexities of be-
havior as lived. This can lead to surprise at unexpected relations, when
expectation has been built on a tree instead of matrix relations, to a disillu-
sioned rejection of an elsewhere useful taxonomy, or to the desirability of
supplementary rephrasing of a problem in dimensional matrix terms. In
figure 13.5 I give one instance where the expectancy of the nonoccurrence
of a characteristic, on the basis of experienced differentiation in the pro-
jected taxonomy, is fatal when in fact that characteristic is shared by the
two branches.
An n-dimensional matrix gives more flexibility than does a tree diagram
for the observer to choose alternative sets of viewing priorities (alternative
entrance points or directions into the matrix) at different particular mo-
ments, for different needs and interests. Instead of talking about plant ver-
sus animal, and cat versus dog or mouse, one may wish to discuss tempo-
rarily various divisions of creatures according to hair types, for example.
B.“Is it a plant?” “No, it’s an animal, it has spontaneous

movement, and is a sentient being.”
“Is it a cat?” “No, it growls, it doesn’t meow.”
“Is it a collie?” “No, it is a small pet with curly hair.”
Living organism
SSS
Plant Animal
Cat
Figure 13.4. A taxonomic tree illustrates one way of partitioning a con-

ceptual universe into genus and species. It has the advantage of leading
to clear-cut logical definition, but the disadvantage of imposing (for the
moment at least) one viewpoint, one set of priority criteria; and it may
conceal implicitly some characteristics—important for other purposes—
shared by twigs far removed from each other in the diagram.
124 Context
A mouse in a hole in the

wall says: ‘There is a cat out-
side. I’m not stupid, P’ll wait for
it to go away before I leave
this hole’
The mouse suddenly pricks

up its ears, and hears: ‘Growl,
meow! Growl growl! Meow!’
and says, ‘Aha, a dog has come
and is fighting the cat. Pll
escape while the cat is busy.’
Out it darts.
The cat grabs the mouse, holds

it up, and says in triumph:
“How truly my father taught me
that if you want to live well you
must become bilingual.’
Figure 13.5. The cat’s sharing of a taxonomic feature—a capacity for

growling which is normally attributed only to dogs—catches the mouse
by surprise. In trusting to his taronomy, here too simple, the mouse
loses his life to the “bilingual” cat.
The mind of man has an extraordinary capacity for entering into its n-
dimensional memory from a vast variety of starting points to search for
data in it. I enjoy working with matrix (field) theory which in a small mea-
sure reflects a bit of this important human capacity.
Jokes may exploit negative expectancies of taxonomy or matrix shape or
content, since man’s mind can enjoy unexpected implications of shared
characteristics. A cat, for example, cannot bark. If, however, one pretends
that a specialist “bilingual” cat can do so, pleasure can be invoked, as in
figure 13.5. Nonoccurrence of a feature leads to contrast (Chapter 6), which
is crucial to the identification and understanding of the nature of units.
A theory based on just one taxonomy, just one partitioning of charac-
teristics, treating the universe as entailing “nothing but” some one (clearly
important) set of features, is in danger of epistemological death. Formal-
ism, for example, is a useful way of describing certain differences between
behavior patterns. But if meaning, relevance, purpose, and observer per-
spective are not directly provided for in the viewing model, mental blind-
ness to the more complex nature of humanity may be the price paid for the
temporary elegance of the taxonomy or mathematical rules.
Universe of Discourse
A number of approaches, with overlapping components, relate

to context. In Chapter 12, both form and meaning were simultaneously rel-
evant. In Chapter 13 it was change which was in view, with no change
occurring unless something contextual was shared. Now we speak of con-
text in a much broader way, in relation to universe of discourse. This refers
to the general or temporary or somewhat permanent frame of reference,
either tacit or explicit, within which social interchange is taking place; it
can include topic, style, genre, discipline, or general speaker or hearer ex-
pectancies. It can reflect the cohesion component of any tagmeme of the
referential hierarchy, in relation to truth or falsehood, specific history, or
specific or particular encyclopedic background; or it can include general
relation to the situation in respect to space, time, society, or personal psy-
chology; or it can include phonological cohesion factors of voice quality
controlled by style or emotional situation; or it can include grammatical
cohesion factors which control or are controlled by literary form. Such co-
hesion features of the hierarchies are touched upon in sections 11.4, 10.7,
and 9.4—the latter with special reference to the tagmeme as itself unit-in-
context.
In addition, distribution as an essential feature of an emic unit is treated
extensively in Chapter 8. But the term “distribution” was there much more
restricted than is universe of discourse. Distribution was referring to the
paradigmatic occurrence of a set of units in a particular structure (or struc-
tures) occurring as in some place (or specific set of places) in such an ex-
plicit structure. Universe of discourse as a term refers rather to the broader
background, often tacit, and often having relevance to a very wide variety
of matters without explicit relation to their particular hierarchical structur-
ing in particular units of grammar or phonology or reference, but having a
pervasive influence on them all, without necessarily being tied to one par-
ticular level (or closed set of levels) or manifestations or constituents of
those structures immediately at hand. So we are at a very high level of co-
herence of the human being in his joking or lying, his fears and his hates,
his biography or his social immersion, his explicit assumptions or his un-
stated but controlling beliefs. Beyond the pure mathematician in his ab-
stractions lies person in commitment to the sheer joy of the pattern chase.
We too can have fun together, searching for evidence that the man in the
street—not just the professional of some kind—is tacitly aware of, and ex-
ploits for pleasure, such “universes” accessible to him.
14.1 Lack of Coherence in

Speaker-Hearer Interaction
When a speaker addresses his audience, he may expect it to ex-

pect something of what he himself expects. If the hearer does not do so,
then communication may fail completely or go awry. By Ciardi (1959 : 847)
this expectation is called the sympathetic contract in relation to poetry. For
him, the relation between the one who writes the poem and the one who
receives it must include a shared belief or value system, permanent or tem-
porarily in the imagination, if the artistic event is to occur. Thus, he says,
“the badness of bad poetry can always and only be located in the quality of
the sympathetic contract.” But the same expectancies affect prose writing
and ordinary conversation.
If the speaker asks a question or makes a comment, awaiting a reply, he
does so against the background of such expectations. If the hearer ignores
this fact in replying, conversational chaos may engulf the original speaker.
Figure 14.1 shows how the reply can answer the actual surface question
but totally fail to meet the about-to-be-buried expectation.
In the implied universe of discourse here, Whippletree was speaking in a
way which should have clued in Mr. Axelgrease that Whippletree wanted
a remedy for his sick cow—not a mere answer to the surface question.
Whippletree assumed that Axelgrease would assume that Whippletree
wanted helpful medical advice—and acted accordingly, using the sup-
posedly prescribed turps as a hoped-for remedy. When his cow died, he
was puzzled—he had expected better advice from his friend. So he in-
128 Context
quired again. The shock. Turps was not a remedy for Axelgrease, either.
(Note, further, that in my reporting this old joke, I made up names to help
signal the fact that I was back in a farming community: the whippletree, as
a swinging bar tying traces to a cart, was normal to the universe of dis-
course of that community of sixty years ago before machines took over.)
In figure 14.2, on the other hand, the universe of discourse behind each
question is noncoherent with that of the answer.
In figure 14.3 the speakers imply that there is a relation between two
universes of discourse, by asking the hearer to guess some kind of unex-
pected relationship—and thus cuing a probable joke.
Four weeks (and sixty years) ago: Mr. Whippletree asks Mr. Axelgrease:
Alex, what did you give yr’old cow when she was sick?
Mr. Axelgrease: Turps.
Two weeks later: Alex, what d’you say you gave yr’old cow when she was
sick?
Mr. Acxelgrease: Turps.
Two weeks later still: Alex, what d’you say you gave yr’old cow when she
was sick?
Mr. Acxelgrease: Turps.
Mr. Whippletree: Well I gave turps to mine, and she died!
Mr. Axelgrease: So’d mine.
Figure 14.1. Failure of reply to meet expectations of speaker can give

difficulty.
A. Medical professor: What happens when a body is immersed in water?

Student: The phone rings.
B. Patient: Doctor, what should I do if I get run down?
Doctor, harassed: Take down the license number.
C. Teacher: Name me four members of the cat family.
Student: Father cat, mother cat, brother cat, sister cat.
Figure 14.2. Slippage of elicited universe of discourse can be deliberate,

even in serious situations, for wry humor; or it may be through failure
to understand. In A, a medical student is presumably trying to enjoy a
joke at the professor’s expense. In B, the doctor seems to have been
pressed beyond endurance, and may have to start over for serious con-
sultation. In C, a very young student may have been serious, but has
failed to learn the implicit taxonomic guidelines of the underlying class.
A. Query: Ifyou pour hot water down a rabbit hole, what do you get?
Answer: Hot cross bunnies.
B. Query: Why is a moth flying around a lighted candle like a gate
blowing in the wind?
Answer: If it kéeps-6n-it-singes-its wings.
Or: If it kéeps-6n-its-hinges-it-swings.
Figure 14.3. The speaker may imply a mixture of universes of discourse,

as in much humor. In A, the clause What do you get, in appropriate
contexts, invites the hearer to guess at some kind of association. A hot
cross bun is a biscuit marked with frosting in the shape of a cross, and
eaten by some groups at Easter time; the word bun ties the association
to the first syllable of bunny, as a pet name for rabbit, while hot reflects
the temperature of the water, and cross the annoyance of the wetted rab-
bit. In B, the answers can be homophonous, if pronounced rapidly and
smoothly, as suggested by the hyphens and stress marks.
A. One duck hunter to another, in the pouring rain: Can you imagine my
wife wanting me to go shopping in wet weather like this!
B. One farmer to another: I entered my mule in the Kentucky Derby. I
didn’t expect him to win, but I thought that the association might do
him some good.
Figure 14.4. Inconsistency with reality, as seen in us by others, can be

annoying or ludicrous. In A, a man sees rainy weather as a deterrent to
shopping but not to hunting; the wife is expected to disapprove the dis-
tinction. In B, the farmer suggests that a mule may receive a good influ-
ence, by associating with the right company; we are expected to doubt it.
14.2. Failure of Coherence of Speaker

with Reality as Seen by Others
As we listen to a person’s statements, they may appear to us

to be inconsistent. In figure 14.4, the first illustration portrays a man as
inconsistent, in that he reacts differently to comparable circumstances; in
the second, a man humorously suggests the improbable relevance of an
association.
In figure 14.5, fiction exploits expectancies ofa politico-social universe of
130 Context
(Yawn) I am a loyal
subject of the King, A tory! A spy!
God bless him. Hustle him!
Away with him!
Figure 14.5. Political coherence can be lost, with disastrous social conse-
quences. Here the fictional character Rip Van Winkle does not know that
times have changed, since he has just wakened from years of blissful,
loyal sleep.
discourse, by showing how Rip Van Winkle got into trouble when he tried
to be politically polite and loyal, but was interpreted—having been asleep
for many years and not knowing that there had been a change of polit-
ical climate in the meantime—as being hostile to unrecognized-as-new
government.
14.3. Behavioral Universals in Language Learning
In one sense, all languages are equivalent: any one of them is

thoroughly adequate for its native speakers to discuss anything which in-
terested their grandfathers. And any of these can be expanded to include
adequate reference to any new topic or item, provided it is given time
enough and provided that not too many new items are thrust at it from the
outside at any one period. (When an industrialized culture, however, first
contacts a nonindustrialized one, the many new concepts and accompany-
ing correlative vocabulary items may give it a verbal traffic jam. The capac-
ity to develop or to borrow new terms is a universal of human experience,
in all languages.
Another factor which is universal is that all cultures must talk about
many items which are related to their survival. In all cultures people must
eat to live. All cultures must recruit new members. All must have role dif-
ferentiation, communication, shared cognitive orientations and goals, and
regulation of disruptive behavior, as well as means of socialization—that
is, the induction of an individual into the roles and subsystems of the so-
ciety (see Aberle et al. 1950). It is reliance on these factors which makes it
practical for a person to assume, when he visits a culture which has a lan-
guage which has never been written (of which there are many in fact still
in existence), to assume that it will be a normal or ordinary language (not
the language of birds or beasts), and to assume that he can find shared
bridges of behavior over which he can pass to learn it. And it is these facts
which have allowed me for many years to be willing to try in an hour, in
public, to start learning a language by gesture only (that is, in a mono-
lingual demonstration, without using English or any other language shared
by the other party to the demonstration). Many colleagues of the Summer
Institute of Linguistics have also given such demonstrations repeatedly.
(For one on TV videotape, see Pike 1977, Program 5.)
14.4, Why Translation Is Possible
It is only the presence of such behavioral universals which

makes it other than a wild dream to assume that translation from any
one language to any other should be expected as being possible. Transla-
tion may for our purposes here be considered a special instance of cross-
cultural paraphrase. From this viewpoint many things which have been
mentioned earlier in this volume become helpful in understanding why
132 Context
translation is possible at all between cultures which seem so different on

the surface.
Every culture has its own “theories”—its cultural windows as ways of
looking at the world as a particular setofuniverses of discourse; and just as
in our culture people can talk together (somewhat) even if they differ as to
assumptions, so they can learn to talk across language differences. In each
there can be misunderstandings; in each, different presuppositions can
block, temporarily or permanently, appreciation for a viewpoint (Chap-
ter 1). In each, there will be differences of observer perspective (static, dy-
namic, and relational as basic ones, chapters 3—5). In each there will be
systems of contrast (Chapter 6) and areas of variation (Chapter 7), so that
emic units must be identified for each before clear understanding can be
achieved. For each there will be a particular set of units of grammatical
and phonological types (chapters 9-10) which must be mastered, for
transfer of concepts into the relevant form, in order to have viable and ex-
pressible form-meaning composites (Chapter 12).
But the nonlinguist may still wonder how it is possible to use words in a
translation when the words in the second language never mean exactly
what they mean in the first. It is important to see that here, also, this trans-
lation problem is not different in principle from that facing the English
speaker who enters a new discipline—say that of linguistics. One can get
formal, logical definitions of new technical terms in relation to including
genus and differentiating species (see fig. 13.4); or one can learn the terms
through a set of illustrations, or from a book-length discussion.
In general, however, the way we normally learn the meaning of new
words is by hearing them in contexts. For the moment, it may be helpful to
think of a new word as a partial vacuum into which flows meaning from
the environment. As it is heard in more and more environments, the word
picks up more and more marginal meanings. Notice that a large dictionary
gives not only definitions of words but samples excerpted from various
writings. In fact it is the study of the contexts of the excerpts which leads
the lexicographer to make the definition, not the definition which leads to
the meanings.
This implies that a context has very high power to determine or to
change a meaning—vastly greater than would be suspected until we have
studied the matter (see fig. 14.6). That contexts can form or change mean-
ings is a very legitimate and valuable (in fact crucial) characteristic of
language.
It allows for the adaptations of ordinary vocabulary for scientific pur-
poses or for the discussion of new culture habits. Translation, furthermore,
I have some good property: hold in possession

It has three parts: consists of
I have a letter to write: obligation
I have enemies: stand in relationship to
I had bad news: received
He had the gall to refuse: characterized by
I had a cold: experienced
Thad a fight: performed
I had an opinion: entertained in mind
I had the children stay: caused to
We'll have no more of that: allow
He had only a little French: was competent in
We had him, then: in position of disadvantage
He had been had by his partner: put at disadvantage
I had my rights: able to exercise them
She had a baby: bore
They had dinner: partook of
They can be had for a price: bribed
Figure 14.6. Dictionaries use context to reach definitions of words. From

such citation forms, the lexicographer can deduce the usage offorms and
their elicited impact on behavior or understanding (data adapted from
Webster’s 1963). This is possible only because contexts in part shape
meanings.
is a special instance of the modification of terms by context; terms never

quite match across the two languages, but the new contexts of the trans-
lated document may bring sufficient change to the starting meanings of the
words used to allow them to communicate with a degree of accuracy suffi-
cient for the purposes of the facts or behavior discussed. On the other hand,
the word sincere normally means ‘free from dissimulation, not pretending
that which is false’; but if one says Always be sincere, followed immediately
by the addition of whether you mean it or not (as on a joking postcard), the
added context forces sincere to mean the opposite (that is, as presenting a
false front of apparent integrity even when there is none). It is almost im-
possible to resist the implication of that change in such a context. Such con-
textual power can appear to an observer as damaging or constructive, or as
diabolical or heavenly, depending on his estimate of the legitimacy and lo-
cal usefulness of the change elicited.
Items being translated, furthermore, presumably have contexts unfamil-
134 Context
iar to the new readers, or else there would have been no reason to translate
them. And the contexts new to the readers must themselves elicit sufficient
change in the meanings of the words used to allow for the transfer of mean-
ing across the initial culture gap.
14.5. Poetry Can Concentrate the

Signaling of a Universe of Discourse
Concentration of features of contrasting universes of discourse

are found in humor, as we have seen in illustrations already given in this
chapter. Poetry can do the same, but even more subtly. In figure 14.7 a tiny
poem of E. E. Cummings combines the use of orthography, line spacing,
and brief lexical content to suggest a whole social-political setting. My pho-
netic interpretation of the data has been heavily influenced by the work of
Axelrod (1944:89).
Graphically, in the last line there, by spelling pause as “paws”, the au-
thor suggests a relation to the first line, where applause is spelled as “ap-
plaws.” Since applause is made by hands, and since “paws” are the hands
of animals, the heavy nature of the applause is forced on the reader. The
separate lines suggest separate stress groups—two stresses for fellow, in
lines two and three, and two stresses for citizens in lines four and five. The
“isn’ts” suggests an unvoiced ending of citizens—which in turn suggests a
foreigner speaking English. This, further, implies that the stress groups are
foreign also, lacking sufficient nonstress on the last syllables of fellow and
citizens. Granted these interpretations, the occasion emerges: there is a pa-
triotic celebration with an appropriate address. The speaker addresses the
audience as one with them. But he is a naturalized citizen, still retaining a
applaws)
fell
ow
sit
isn’ts”
(a paw s
e.e, cummings
Figure 14.7. Poetry may concentrate signals identifving a universe of dis-

course. (Copyright 1944 by E. E. Cummings; renewed 1972 by Nancy T.
Andrews. Reprinted from Complete Poems 1913-1962 by permission of
Harcourt Brace Jovanovich, Inc.).
foreign accent; and sometimes a new member of the national family can
feel more deeply patriotic on such occasions than a person born into it.
Thus an extraordinary hint is given in brief compass of an extensive back-
ground, an interlocking set of elements signaling a universe of discourse.
14.6. Conclusion: Unit-in-Context within Context
No unit relevant to human beings exists without its having a re-

lation to a system of interlocking types of context. The units exist in a vast
matrix of n-dimensional intersecting relations within which the specific
unit is distributed and which comprise our universe with our cognitive
frames of reference.
This contextual factor gives the tagmeme its alternate technical name—
unit-in-context. Each tagmeme (see sections 9.4—5) involves four kinds of
contextual distribution of a unit, any one of which can be under temporary
focus. At any one moment it is manifested by one element of a set of ele-
ments (in Cell 2, Class) such that any one member of the set is substitutable
for that member of that set in that context without changing any structural
characteristic of the kind of pattern of which it is a part. Yet that set of
elements is a part of an immediately larger pattern, which contains it and
which it helps make up; and the part comes in a particular functional
place (Cell 1, Slot) in relation to that larger pattern. But that function has a
relevance, significance, purpose (Cell 3, Role) relative to the action or ob-
servation or judgment of an observer of that unit. And the unit is embedded
in a larger set of general kinds of backgrounds (Cell 4, Cohesion), a refer-
ential set of beliefs, situations, precedents, biographies, histories, logic,
sanity or lunacy, science or fiction, seriousness or humor, observation or
imagination, and others; and these help hold units together in a coherent
observer viewpoint, or in a set of happenings or things or causal relations
and controls. Similarly, there may be both grammatical patterns of agree-
ment or contrast, and phonological ones.
A unit-in-context may be observed at any level of any one of the hier-
archies. They comprise part of the warp and woof of every grammatical,
generalized linear pattern of verbal or nonverbal patterns of behavior.
No hierarchy exists except in relation to the others. No simple necessary
single mapping relation is present between the hierarchies of tagmemes.
No simple set of rules is adequate to capture all the possibilities or to pre-
dict at one particular moment which choice will be made by the observers
involved. If one wishes his theory to handle this human flexibility, he must
allow his theory to have some indeterminacy of choice at the moment under
136 Context
observation. The theory must be able to describe any one variation of inter-
action, after it has occurred, without being forced to predict in advance
which one must inevitably occur.
This points up a crucial assumption of the theory: Personal interaction is
given priority over identification of things or abstractions. It is persons
who perceive, imagine, and sense relations. It is persons who talk about
these concepts. But persons have limits. Therefore they must have hier-
archy, and hierarchy intersecting with hierarchy. Only thus can they signal
the needed complexity of their experience, and do so in a way that will not
stifle them.
References
Note: For extensive documentation and discussion of tagmemic the-

ory in historical perspective, see Pike 1967a; for bibliography on grammatical
methodological matters in tagmemics, see Pike and Pike 1977. For this current
volume, where the general background principles (rather than their sources) are
being presented, it seems inappropriate to duplicate the prior discussion. The
references given here, therefore, are largely those from which illustrations are
taken—and these, in turn, are usually drawn from data with which I have had
direct or indirect contact. A few—but only a few—further references are to mate-
rials which I refer to or quote from directly, as supplementary to my own experi-
ence or to the earlier documentation just referred to.
Aberle, D. F., A. K. Cohen, A. K. Davis, M. J. Levy, Jr., and F. X. Sutton. 1950. The
functional prerequisites of a society. Ethics 60.100—111.
Aschmann, Herman P. 1946. Totonaco phonemes. International journal of Ameri-
can linguistics 12.34—43.
Axelrod, Joseph. 1944. Cummings and phonetics. Poetry 65.11.88-94.
Bloomfield, Leonard. 1933. Language. New York: Holt, Rinehart and Winston.
Brend, Ruth M. 1968. A tagmemic analysis of Mexican Spanish clauses. The
Hague: Mouton.
Browning, Robert. 1895. The complete poetic and dramatic works of Robert
Browning. Cambridge, Mass.: Houghton Mifflin.
Carnap, Rudolf. [1939] 1955. Foundations of logic and mathematics. Foundations
of the unity of science: toward an international encyclopedia of unified sci-
ence, ed. Otto Neurath, Rudolf Carnap, and Charles Morris, 1:139—214. Chi-
cago: University of Chicago Press.
Chomsky, Noam. 1957. Syntactic structures. The Hague: Mouton.
. 1962. A transformational approach to syntax. Third Texas conference on
problems of linguistic analysis in English, May 9-12, 1958, 124-69. Austin:
University of Texas.
. 1966. Cartesian linguistics: a chapter in the history of rationalist thought.
New York: Harper and Row.
Ciardi, John. 1959. How does a poem mean? An introduction to literature, pt. 3.
Boston: Houghton Mifflin Co.
138 References
Fillmore, CharlesJ.1968. The case for case. Universals in linguistic theory, ed. by
Emmon Bach and Robert Harms, 1—80. New York: Holt, Rinehart and
Winston.
Fischer-Jorgensen, Eli. 1975. Trends in phonological theory: a historical introduc-
tion. Copenhagen: Akademisk Forlag.
Frank, Philip. 1957. Philosophy of science: the link between science and philoso-
phy. Englewood Cliffs, NJ.: Prentice-Hall.
Fries, Charles C. 1952. The structure of english. New York: Harcourt Brace.
Gelb, I. J. [1952] 1963. A study of writing. 2d ed. Chicago: University of Chicago
Press.
Howland, Lilian G. 1981. Communicational integration of reality and fiction.
Language and communication 1.89—148.
Hyman, Larry. 1975. Phonology: theory and analysis. New York: Holt, Rinehart
and Winston.
Jespersen, Otto. 1937. Analytic syntax. Copenhagen: Munksgaard.
Jones, Linda K. 1977. Theme in english expository discourse. Edward Sapir
monograph series in language, culture, and cognition 2. Lake Bluff, Ill.: Jupi-
ter Press.
Joyce, James. [1916] 1968. A portrait of the artist as a young man. New York:
Viking Press.
Kaplan,J. D., ed. 1952. Dialogues of Plato. Trans. by Benjamin Jowett. New York:
Pocket Books.
Kuhn, Thomas S. 1962. The structure of scientific revolutions. Chicago: University
of Chicago Press.
Lamb, Sydney. 1964. The sememic approach to structural semantics. Trans-
cultural studies in cognition, ed. by A. Kimball Romney and Roy G. D’An-
drade, 57—58. American anthropologist, vol. 66, no. 3, part 2, special
publication.
Langer, Susanne K. [1937] 1953. An introduction to symbolic logic. 2d ed. New
York: Dover.
Miller, George A. 1956. The magical number seven, plus or minus two: some
limits on our capacity for processing information. Psychological review
63.81-—97.
Moser, Edward, and Mary B. Moser. 1965. Consonant vowel balance in Seri
(Hokan) syllables. Linguistics 16.50-—67.
Nida, Eugene A. 1975. Componential analysis of meaning: an introduction to se-
mantic structures. The Hague: Mouton.
Patrick, G. T. W., Trans. [1888] 1889. The fragments of the work of Heraclitus of
Ephesus on nature: translation from the Greek text of Bywater. Baltimore:
Murray. .
Pickett, Velma. 1960. Hierarchical structure of Isthmus Zapotec. Language disser-
tation no. 56. Baltimore: Linguistic Society of America.
Pike, Eunice V. 1970. Review Aspects of phonological theory, by Paul M. Postal.
Lingua 25.30—46.
References 139
Pike, Kenneth L. 1945. The intonation of American English. University of Michi-

gan publications in linguistics no. 1. Ann Arbor: University of Michigan
Press.
- 1948. Tone languages: A technique for determining the number and type
of pitch contrasts in a language, with studies in tonemic substitution and
fusion. University of Michigan publications in linguistics no. 4. Ann Arbor:
University of Michigan Press.
- 1952. Operational phonemics in reference to linguistic relativity. Journal
of the acoustic society of America 24.618-25.
- 1957. Abdominal pulse types in some Peruvian languages. Language
33.30—35.
- 1962. Dimensions of grammatical constructions. Language 38.221—44.
. 1963. Theoretical implications of matrix permutation in Fore (New
Guinea). Anthropological linguistics 5.8.1—23.
» 1964a. Name fusions as high-level particles in matrix theory. Linguistics
6.83—91.
. 1964b. On the grammar of intonation. Proceedings of the fifth interna-
tional congress of phonetic sciences, 105-19.
. 1965. Language—where science and poetry meet. College English
26.283—92.
. [1954, 1955, 1960] 1967a. Language in relation to a unified theory of the
structure of human behavior. 2d ed. The Hague: Mouton.
. 1967b. Tongue-root position in practical phonetics. Phonetica
17.129-—40.
. 1967c. Stir, change, create: poems and essays. Grand Rapids, Mich.:
Eerdmans.
. 1977. Pike on language. (Series: 3/4-inch videocassettes [NTSC standard]
and 16mm kinescopes.) Ann Arbor: University of Michigan Television Cen-
ter. Program no. 1: Voices at work [phonetics]; Program no. 2: The music of
speech [pitch and poetry]; Program no. 3: Waves of change [the how and
why of change in language]; Program no. 4: The way we know—the value of
theory in linguistics study; Program no. 5: Into the unknown [learning an
unknown language by gesture—a monolingual demonstration].
. 1981. Tagmemics, discourse, and verbal art. Michigan studies in the
humanities.
Pike, Kenneth L., and Evelyn G. Pike. 1977. Grammatical analysis. Summer Insti-
tute of Linguistics publications in linguistics 53. [2d ed., with section refer-
ences—but not page numbers—unchanged, 1982.]
Potter, Ralph K., George A. Kopp, and Harriet C. Green. 1947. Visible speech.
New York:-Van Nostrand.
Robbins, Frank E. 1961. Quiotepec syllable patterning. International journal of
American linguistics 27.237—50.
Stahlke, Herbert, and Ruth M. Brend. 1967. The use of matrices in the prepara-
tion of language textbooks. Language learning 17.37—44.
140 References
Stewart,J.M. 1967. Tongue-position in Akan vowel harmony. Phonetica

16.185—204.
Suharno, Ignatious, and Kenneth L. Pike. 1976. From Baudi to Indonesian. Irian
Jaya: Cenderawashi University and Summer Institute of Linguistics.
Tench, Paul. 1976. Double ranks in a phonological hierarchy. Journal of linguis-
tics 12.1—20.
Webster’s seventh new collegiate dictionary. 1963. Springfield, Mass. G. & L.
Merriam Co.
Wilder, Thornton N. [1931] 1963. The long Christmas dinner. New York: Harper
& Row.
Index
Abstract, 15, 19, 111, 127; and person, 52—54; sharing in, 17, 118-25. See
x4; and physical, xiii; and system, 65 also Wave
Actor, 22, 77; role of, 78, 82 Chinantec, 90; tone in, 26
Adjunct, 78, 82 Choice, 45—46; of theory, 135-36
Agent, 78 Circle and field, 31
Aguaruna, 91 Circumstance as role, 78
Algonquian, 72 Class: distribution in, 62; of nouns, 82;
Alphabet, 85, 89 in paragraph, 80; proportion with,
Analysis: alternative, 120; componen- 76; in slots, 782—79; as substitution
tial, 99; formal, 100 items, 62, 75. See also Tagmeme
Arithmetic and truth, 104 Clause, 78—80, 82; in Zapotec, 49
Attention. See Focus Clock, 104
Author and intent, 116. See also Ob- Cognition. See Emic; Universals
server Coherence. See Cohesion
Cohesion, 75, 127—30; and agreement,
Baby talk, 33 79; and background, 75, 121, 135;
Barnard, Myra, 57—58 expectancy in, 100, 127—28; in hu-
Behavior, 49—54; contrast in, 49—51; mor, 130; number in, 82; in refer-
controls on, 63; identities in, 52—54; ence, 135; and systemic control, 77.
language learning in, 131; levels in, See also Universe of discourse
72—74; nonverbal, 18, 50—51 Colors, matched, 43—44
Big, meanings of, 50 Complexity: via channel capacity, 70;
Bitransitive clause, 78 constraint on, 71—73, 86—87
Browning, Robert: “Memorabilia,” Component. See Feature
36-37 Concepts, 11, 99
Conjunction as class, 78
Carib, 99 Consonant clusters, 63
Cat, 119, 124 Context, 109-35; definition of, 130; and
Cell. See Tagmeme; Unit-in-context field, 30-31; form and meaning in,
Change, 13, 17, 52-56; in grammar, 111-17; humor in, 115; and mean-
55—56; in meaning, 132—33; mor- ing, 132—33; relevant, 15-17; and
phophonemic, 47; in pronunciation, unit, 30-31. See also Cohesion; Slot;
54—55; referential identity in, System; Universe of discourse
142 Index
Continuum. See Wave taneous, 31, 88; of sound in field, 32;

Contrast, 42—51,64. See alsoEmic; Unit taxonomic, shared, 124. See also
Conversation, 72, 78. See also Prag- Contrast; Emic; Unit
matics; Speaker Field, 30—38; and clause, 33—36; and
Counting, 23 change, 122—25; and contrast,
Criteria. See Contrast; Emic; Etic; 42-43; intonation in, 91-95; and
Feature morphology, 56; pattern in, 31—33;
Culina, 91 phonological, 32; and space, 23
Culture, 131-32 Focus, 26; by change of grammar, 105;
Cummings, E. E.: “applaws),” 134—35 constraint on, 86—87; emotional,
105; in form and meaning, 115; and,
Data: raw, 69; and theory, 3, 8 message, 54. See also Observer; Per-
Declarative clause root, 82 spective
Definition, 73—74 Fore, 56
Dickinson, Emily, 93-96 Form, 115-17; and humor, 115; and
Dictionary, 132. See also Encyclopedia; interpretation, 93; of intonation, 92;
Lexicon and meaning, 16, 44, 115-17; and
Dimension. See Matrix poetry, 115-17; and sympathetic
Discourse, 82, 89. See also Universe of contract, 127; voice quality in, 96.
discourse See also Meaning
Distribution, 60—65; appropriate, xiii, Formalism, xi, 116—17, 125. See also
13, 52; proportional, 76 Logic; Mathematics
Drama, 28 Formula, 80
Forster, Jannette 57—58
Emic, 44—45, 47; insider’s view of, 11, Frame: for reference, 106; tone tech-
44—45; and phonology, 90; source of nique by, 45, 48; and universe of dis-
term, 87; unit, xii—xiii. See also Con- course, 126. See also Observer
trast; Culture; Etic; Identity; Ob- Fusion, 118; in Chinantec, 26; in mor-
server; Phoneme; Thing pheme sequence, 58. See also Wave
Encyclopedia, xiii, 98, 126. See also
Dictionary Gardner, Mary Jane 57—58
Entity. See Identity Generation, 81
Epistemology, 125 Gestures, 50—51
Etic, xii, 73—74, 87. See also Emic Goal. See Undergoer
Event, 98, 100, 101. See also Reference Grammar: borders of, 84—85; con-
Expectation, 3, 100, 127—28. See also straints on, 64; emic units in, 87;
Role and etic level, 74; and form, 105-6;
formal proportion in, 77; hierarchy
Fact and observer, 30 in, 70—83; and lexicon, 112—14;
Feature: and chunk constraint, 70, 72; and phonology, 85, 98; and refer-
componential, 99; contrastive, 43, ence, 103—5; and semantic rele-
99; diagnostic, 53; etic, 88; identifi- vance, 72—73; in social structure, 62;
cational, 43; minimum descriptive, of syllogism, 103; talking order in,
54; nonoccurrence of, 125; simul- 103; variant in, 55—56. See also
Index 143
Clause, Discourse, Form; Morpheme; Manifestation, 55. See also Feature;

Tagmeme Form; Substance and hierarchy;
Variant
Have, 133 Margin of wave, 26—27
Heraclitus, 52, 118 Mathematics: and formalism, 125;
Hierarchy, 69-106; grammatical, multiple forms in, xiv; notation in,
70—83; and humor, 114; and inter- 79; and person, xi; and poem, 117;
locking elements, 13—15, 84—85, 98, and simplicity, xiv
112-14, 135; levels of, 72, 73-74; Matrix, 47—49, 122—24; clauses in, 36;
and, lexicon, 112-14; phonological, variant of, 56
84—96; referential, 15, 97-106; and Mazatec, 44, 47-48, 87
reductionism, xiii, 87—88; social, 99 Meaning: and behavior, 75, 112; cen-
History as wave, 29 tral, 28, 119-20; componential anal-
Humor: and hierarchy, 114-15; and ysis of, 99; and context, 132—33; as
noncoherence, 17, 105—6, 124-25, contrastive and homophonous, 50;
128-29; and pun, 118-19 and form, xiv, 16, 116-17; and hu-
mor, 115 (see also Humor); of
Identity, 52—54, 99. See also Contrast; idiom, 29; as impact, 72, 112; and
Emic; Thing levels of grammar, 72—73; and lexi-
Idiom, 29. See also Meaning cal contrast, 16, 50; in phonology,
Indirect Object, 80, 82. See also Role 112; and physical component,
Intent and metaphor, 104 111-12; in poem, 116; and purpose,
Interrogative, 78 xvi; and understanding, 112; and
Intersection. See Field; Matrix voice quality, 96; as wave, 119. See
Intonation, 91—95 also Form; Reference; Role
Intransitive clause, 78 Memory, 70, 124
Item. See Emic; Identity; Thing Metaphor and truth, 104
Minimal pairs, 44
Japanese syllable, 90 Mixtec, 90
Joke. See Humor Model. See Theory
Monolingual demonstration, 131
Kuhn, Thomas's. 29, 109 Monologue, 73
Morpheme, 73, 118. See also Grammar
Langer, Susanne K. xiv Morphophonemic change, 47,
Language as theory, 3 54-56
Language learning, 63, 131
Length, 90, 94 Name, 98. See also Lexicon; Meaning;
Levels, 77-100; pairs of, 85. See also Reference
Hierarchy Native speaker. See Observer
Lexicon: and behavioral contexts, 16, Norm in wave, 72
50; dictionary definition of, 133; ex- Nucleus: and person in group, 102;
tended, 113; and hierarchy, 112—14; and stem of word, 21; of wave,
as wave, 28 26-27. See also Wave
Logic, xiv. See also Formalism Number, 56, 78, 82
144 Index
Object. See Undergoer Phonology, 84—96; distribution of, 63;

Observer, 3—37, 111; alien, 52—53; field of, 32; fusion in, 118; hierarchy
and distribution, 61; etic view by, of, 15, 84—96; lexicon, 112—14;
xii; and fact, 30; and focus, 26; free- merging sounds in, 24—26; morpho-
dom of, 109; and hierarchy, 136; phonemic change in, 54; and stress,
limits to, 136; as native speaker, 14-15, 88-89; tongue in, 34. See
52—53; and perspective, xi, xiii, 12, also Morphophonemic change; Pho-
38, 41; and person, xi; and phonol- neme; Tone
ogy, 86—87; and reality, 129-30; ref- Pike, Evelyn: exercises by, xv; and lex-
erential tagmeme for, 102. See also icon as shared substance, 113; and
Emic; Focus; Identity; Perspective; purpose added, xvi; and referential
Theory hierarchy, 100
Occurrence. See Distribution Pitch. See Phonology; Tone
Order, of talking, 103. See also Plato, 98, 101, 102, 105
Grammar Poetry, 36-38, 93—96; elevated style in,
56; form and meaning in, 115-17;
Paradigmatic, 75, 79. See also Class intonation in, 93-95; sympathetic
Paragraph, 73, 89. See also Grammar contract in, 127; universe of dis-
Paraphrase, 15; and concept, 100; ref- course in, 134—35
erential identity of, 98; translation Practice. See Theory
as, 131 Pragmatics, 75, 79. See also Behavior;
Particle, 19—20, 23; and change, Observer; Reference; Speaker
119-20; class as, 75; partitioning of, Prayer, 56
xi. See also Unit~ Predicate, 78, 80
Partitioning, 118. See also Particle Prefix, 27
Passive, 78 Prepositional phrase, 78. See also
Patient, 77. See also Role Grammar
Pattern, 31—36; characteristics of, 135; Process as particle, 20
clause in, 33—36; and field, 31-33; Proposition, 72, 73
in formula, 81; overall, 53. See also Psychology, 126. See also Reference
Field Pun. See Humor
Perspective, 3—4; choice of, 38, 593 Purpose, 100. See also Reference; Role
complementarity in, 11-13; focus in,
41; theistic, 15; verbal cues to, 38. Reality: hierarchy in, 87—88; non-
See also Field; Observer; Particle; coherence with, 129; and observer,
Wave 129-30; and theory, 5—6. See also
Phaedo (Plato), 102 Cohesion; Reference
Philosophy, 116-17. See also Formal- Recognition, 42—43. See also Contrast;
ism; Logic; Mathematics; Observer; Emic; Identity; Observer; Reference;
Plato; Reductionism; Theism; The- Thing
ory; Truth Reductionism, 87—88, 109; and thresh-
Phone, 89. See also Etic; Phonology olds, xiii
Phoneme, 44, 86; feature of, 88; variant Reference, 97-106; components shared
of, 89. See also Emic; Phonology in, 121; and lexicon, 112-14; para-
Phonemic, xii phrase in, 15, 98, 100; poetic, 112;
Index 145
purpose in, 15; and reality, 105-6, xiii; syllogism change by, 105; utter-
129; and thing-in-itself, 15; types of, ances by, 64. See also Observer
135. See also Meaning Species and genus, 122. See also
Relation, 20, 100, 128. See also Field; Taxonomy
Matrix Statement, 78
Repetition. See Particle Stem, 77
Result, 78, 80 Stress, 14-15, 88, 92; in poem, 94; as
Role, 77—82, 100; social, 62; universals wave nucleus, 89. See also Hier-
of, 131. See also Grammar; Mean- archy; Phonology
ing; Tagmeme Structure, simultaneous, 111. See also
Run, 28—29, 120 Hierarchy; Tagmeme
Style, 54. See also Grammar; Phonology
Schram, Judy, 87 Subject, 22, 78, 82
Science: in context, 79; and truth, 109; Substance and hierarchy, 112-14
vocabulary of, 132—33; as wave, 29 Substitution. See Paradigmatic;
Scope, 80, 82. See also Role Reference
Segmentation, 57—59. See also Fusion; Syllable, 89, 90, 118
Wave Syllogism, 103-5
Semantics. See Meaning; Reference Sympathetic contract, 127
Sentence: definition of, 74; topic, Syntagmatic relation, 79. See also Wave
26—27; types of, 82 System, 47—48, 65
Sequence. See Distribution; Wave
Seri, 63 Tagmeme, 75—83; class in, 75; in
Sharing, 17; of meanings, 50; and pho- clause, 81; cohesion in, 75; concepts
nology, 18; of referential compo- in, 11; and context, 135; features of,
nents, 121. See also Change; Fusion; 75; history of, xv, 18, 87; notation
Wave for, 78—83; phonological, 95; refer-
Slang, 29 ential, 100—103; slot and role in, 75;
Slot, 75, 76, 78, 80, 82. See also Distri- of subject, 22; theory of, 10-17; as
bution; Tagmeme unit-in-context, xiv. See also Gram-
Smile, 50, 51 mar; Hierarchy; Phonology; Refer-
Social interaction, xiii, 72—73. See also ence; Unit-in-Context
Conversation; Observer; Pragmatics; Taxonomy, 54, 122—24
Speaker
Society, 99 “Telegram” (Pike), 116
Socrates, 98-105, 111 Tense, 78
Sound, 31—33; and distribution, fiermi., 72573
63-64; smeared, 25—26; static view Theism, 15
of, 21. See also Phoneme; Phonology; Theory, 3, 5—18; choice of, 8, 135-36;
Poetry « context in, 15—17; in culture, 132;
Space, 23, 103 hierarchy in, 13-15; language as,
Spanish, 33, 63 54; tagmemic, 10—18, 78-79,
Speaker, 72, 75, 78—793 attitude of, 91; 82-83; usefulness of, xi, 5—7. See
and conversation, 78—79; expecta- also Emic; Observer; Perspective;
tions of, 127—29; and interaction, Tagmeme
146 Index
Thing, 3—4; as particle, 19-20; as Unit-in-context, 75—79; in context, 135;

thing-in-itself, 15. See also Emic; diagram and formula for, 79-83. See
Identity; Observer also Tagmeme
Time as role, 78 _ Universals, 131
Tone: in Chinantec, 26; frame tech- Universe of discourse, 126—36. in hu-
nique in, 45, 48; in larger units, 88; mor, 129; for identification, 60;
in Mixtec, 90; and variants, 46 shared, 18. See also Cohesion; Dis-
Tongue, 34—35 course; Pattern
Topic, 80; as role, 78; in sentence,
26-27 Variant, 46, 52—59. See also Change;
Translation, 131-34 Etic; Wave
Transitive clause, 78 Voice quality, 54, 88, 91, 93; cohesion
Tree, 78—80, 122-23 by, 126; and meaning, 96, 112
Truth, 103—5, 126
Undergoer, 77—78, 80, 82. See also Wave, xi, 24—30; and change of mean-
Role ing, 120; and norm, 72; phonologi-
Unit, 11, 39-65, 111-12; as allo, 56; cal, 24-27, 89-91; shared, 120-22;
and context, 31; and contrast, xii; and slot, 75; and stress group, 89
discontinuous, 57; and distribution, Word, 13; contrastive meaning in, 50;
60—61; features of 13; in field, meaning of, 16; merging, 25; stem
30-31, 37, 56; in matrix, 49; non- of, 21. See also Lexicon, Meaning,
coterminous, 85; and observer, 31; of Reference
phonological wave, 89-91; recogni- Writing, 85, 89.
tion of, 71; socially defined, 64; and
Tagmeme, 75—79; variant of, 13. See
also Particle Zapotec, 48—49
ISBN 0-8032-3664-6

Tagmemics

Uploaded by

Copyright:

Available Formats

Tagmemics

Uploaded by

Document Information

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

Tagmemics

Uploaded by

Copyright:

Available Formats

What are some of the concepts discussed in the book?

What are some of the concepts discussed in the book?

What is the structure of the book?

What is the structure of the book?

WITHDRAWN

write new due date below

University of Nebraska Press Lincoln and London

Copyright 1982 by the

Manufactured in the United States of America

The paper in this book meets the guidelines for

Library of Congress Cataloging in Publication Data

Pike, Kenneth Lee, 1912-

PartI The Observer and Things 1

Part II The Unit 39

Part III Hierarchy 67

Chapter 11: The Referential Hierarchy 97

Part IV Context 107

ikeale Theory like a window 6

8.2. Constraints on occurrence of vowels and consonants 63

In this volume person (and relation between persons) is given

When man studies “things,” he injects part of himself into their

“Why study theory? Why not just be practical?” So speak those

1.1. Value of a Theory

A theory is like a window.

questions on a narrow front by simplifying the task of investigation. Only if

1.2. Weaknesses of a Theory

Any theory may have a weakness at the point of its greatest

1.3. Selection of a Theory

In trying to choose or to build a theory we should seek an orga-

Figure 1.2. Choice of a theory is in part pragmatic. One may select a

preparation of an alphabet for a preliterate culture, the teaching of fresh-

In a deductive, formal approach, one guesses, to reach an idea of a

2.1. Beliefs Underlying Statements

No statement can be made seriously unless preceding it there is

Some Conceptual Tools

Hierarchies of units as parts of wholes

Figure 2.1. A set of interlocking tragmemic concepts which can be thought

2.2. Complementarity of Perspective

Two photographs may appear identical to the casual observer.

the brain synthesizes these views into a three-dimensional experience. Sim-

were cutting up sequences into chunks—into segments or particles. At

2.3. Units as Known

The human being experiences the world around him as made

2.4. Hierarchy of Interlocking Hierarchies

Man experiences the world as if it were made up of wheels

Susie: 1 year old Susie: 5 years old Susie: 13 years old

Figure 2.3. Units change at different times or places. Talking to Susie

First speaker: The tiger will return.

And conversations can carry on further, with various speakers.

2.5. Context as Relevant

Language itself can be viewed as action—as a kind of behavior.

forms which we have discussed—sounds, words, sentences—are mean-

not handling language. Similarly, if a person tries to make a classification

(Engineers building the bridge)

3.1. Things and Nonthings as “Things”

Human nature experiences the world—sometimes—as made

3.2. Particles in Language

In the word kit there is a sequence of three units of sound. The

Figure 3.2. A static view of a sound represents it as one frame from an

phatically,’ or ‘I was recently and completely just then individually em-

B. veeCchi ‘Che] pays’