The Instruction of Imagination Language PDF

T h e I n s t r u c ti o n o f I m a g i n ati o n
9780190256623-Dor.indb 1 03/06/15 6:17 PM

F o u n d ati o n s O f H u m a n I n t e r a c ti o n
General Editor
N.J. Enfield, Max Planck Institute for Psycholinguistics, Radboud University, Nijmegen,
and the University of Sydney
This series promotes new interdisciplinary research on the elements of human sociality, in par-
ticular as they relate to the activity and experience of communicative interaction and human
relationships. Books in this series explore the foundations of human interaction from a wide
range of perspectives, using multiple theoretical and methodological tools. A premise of the series
is that a proper understanding of human sociality is only possible if we take a truly interdisciplin-
ary approach.
Series Editorial Board

Michael Tomasello (Max Planck Institute Leipzig)
Dan Sperber (Jean Nicod Institute)
Elizabeth Couper-Kuhlen (University of Helsinki)
Paul Kockelman (University of Texas, Austin)
Sotaro Kita (University of Warwick)
Tanya Stivers (University of California, Los Angeles)
Jack Sidnell (University of Toronto)
Recently Published in the Series

Agent, Person, Subject, Self
Paul Kockelman
Exploring the Interactional Instinct
Edited by Anna Dina L. Joaquin and John H. Schumann
Relationship Thinking
N.J. Enfield
Talking About Troubles in Conversation
Gail Jefferson
Edited by Paul Drew, John Heritage, Gene Lerner, and Anita Pomerantz
The Instruction of Imagination
Daniel Dor
9780190256623-Dor.indb 2 03/06/15 6:17 PM

The Instruction
of Imagination
Language as a Social
Communication Technology
Daniel Dor
9780190256623-Dor.indb 3 03/06/15 6:17 PM

1
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide.
Oxford New York

Auckland Cape Town Dar es Salaam Hong Kong Karachi
Kuala Lumpur Madrid Melbourne Mexico City Nairobi
New Delhi Shanghai Taipei Toronto
With offices in
Argentina Austria Brazil Chile Czech Republic France Greece
Guatemala Hungary Italy Japan Poland Portugal Singapore
South Korea Switzerland Thailand Turkey Ukraine Vietnam
Oxford is a registered trade mark of Oxford University Press

in the UK and certain other countries.
Published in the United States of America by

Oxford University Press
198 Madison Avenue, New York, NY 10016
© Daniel Dor 2015
All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted, in any form or by any means, without the prior
permission in writing of Oxford University Press, or as expressly permitted by law,
by license, or under terms agreed with the appropriate reproduction rights organization.
Inquiries concerning reproduction outside the scope of the above should be sent to the Rights
Department, Oxford University Press, at the address above.
You must not circulate this work in any other form

and you must impose this same condition on any acquirer.
Library of Congress Cataloging-in-Publication Data

The Instruction of Imagination : language as a social communication technology / Daniel Dor.
p. cm. — (Foundations of human interaction)
Includes bibliographical references and index.
ISBN 978–0–19–025662–3 (hardback)
1. Communication—Technological innovations. 2. Human-computer interaction.
3. Language and languages—Usage. 4. Technological innovations—Social aspects.
5. Sociolinguistics—Technological innovations. I. Title.
P96.T42D67 2015
401—dc23
2014050097
9 8 7 6 5 4 3 2 1
Printed in the United States of America on acid-free paper
9780190256623-Dor.indb 4 03/06/15 6:17 PM

In Memory of Yehuda Elkana
9780190256623-Dor.indb 5 03/06/15 6:17 PM

Uttering a word is like striking a note on the keyboard of
the imagination.
—Ludwig Wittgenstein, Philosophical Investigations
9780190256623-Dor.indb 6 03/06/15 6:17 PM

Content s
Series Editor’s preface ix

Acknowledgments xi
1. Introduction 1
2. The Functional Specificity of Language 16
3. How the Technology Works 34
4. Sign and Meaning 60
5. The Spirals of Relativity 86
6. Production and Comprehension 103
7. The Social Autonomy of Syntax 123
8. The Universality of Diversity 147
9. Acquisition as a Collective Enterprise 164
10. The Evolution of Language and Its Speakers 184
11. Conclusion: Reassembling the Puzzle 216
Notes 221
References 225
Author Index 247
Subject Index 253
9780190256623-Dor.indb 7 03/06/15 6:17 PM

9780190256623-Dor.indb 8 03/06/15 6:17 PM
S e r i e s e d it o r ’ s p r e fa c e
There would be no hope of understanding the foundations of human interac-

tion without a proper understanding of language. But many have suggested that
the reverse is not true. Is social interaction relevant for explaining what is at the
core of the human capacity for language? Some have said that the social func-
tions of language are merely convenient side effects of what is nothing more than
a private tool for thought. Daniel Dor’s account of language—what it is, why
it exists, how it is learnt and used, how it evolved—shows why that view has
little merit. He casts the fundamental questions of language in a compelling new
light, adding weight to a perspective that is slowly but surely taking hold in the
cognitive science of language. The required shift in thinking is not only to see
that language is socially constructed in a non-trivial sense but to see that its es-
sential function is to affect and direct others, for social ends. As Dor puts it, the
construction of language is “not an individual project.” This applies at numerous
scales: in the process of acquiring language in a person’s lifetime, in the historical
evolution of languages in our communities, and in the original emergence of the
language capacity in our species. The key precondition for language, as we know
it, is the existence of a truly collective mode of thinking and acting. While few
would deny that language is at the foundation of interaction, Dor’s work now
makes it harder than ever to deny the reverse: that social interaction is at the
foundation of language.
N. J. E.
Sydney, March 2015
9780190256623-Dor.indb 9 03/06/15 6:17 PM

9780190256623-Dor.indb 10 03/06/15 6:17 PM
A c k n o wl e d g m e n t s
If it weren’t for good conversation, there would be no point to intellectual life,

and it would definitely make no sense to spend too much time thinking about
language. Good conversation is where language functions at its best—not just
in spite of, but also because of the fact that it is so often an uphill battle against
misunderstanding. Over the very many years in which I have been trying to put
this theory on its feet, I was lucky enough to enjoy such conversations with Mi-
chael Arbib, Mira Ariel, Shoshana Blum-Kulka, Jerome Bourdon, Mati Breski,
Eve and Herb Clark, Mary Dalrymple, Dan Dediu, Jean-Louis Dessalles, Eli
Dresner, John Du Bois, Miran Epstein, Nick Evans, Daniel Everett, Tecum-
seh Fitch, Miriam Fried, Ad Foolen, Ofer Gal, David Gil, Simona Ginsburg,
Snait Gissis, Lila Gleitman, Adele Goldberg, Nathalie Gontier, Uri Hadar,
Oren Harman, Gal Hertz, Ray Jackendoff, Sverker Johansson, Adam Kendon,
Brian Klug, Chris Knight, Ivan Krastev, Ehud Lamm, Steven Levinson, Erez
Levon, Jerome Lewis, John Lucy, Anat Matar, Irit Meir, Bracha Nir, Irene Pep-
perberg, Stanley Peters, Amit Pinchevski, Shalini Randeria, Malka Rappaport-
Hovav, Haim Rechnitzer, Jacqueline Rose, Wendy Sandler, Michal Schechter,
Yael Sharvit, Chris Sinha, Dan Slobin, Dan Sperber, Yosef Schwartz, Luc Steels,
Deirdre Wilson, Emily Wyman, Noam Yuran, Yi Zheng, Jordan Zlatev, and
Ghil’ad Zuckerman. I am grateful to all.
I would also like to thank my students at Tel Aviv University, who have been
struggling for years now with a theory in the making, and the participants in
talks I gave about the theory, at different stages of its development, in Tel Aviv,
London, Budapest, Amsterdam, Nijmegen, Lisbon, X’ian, and Berlin.
Special thanks are due to three of my teachers, each of whom had a deep
impact on me and my work. Ruth Berman, firmly suspicious of vacuous theo-
rizing, deeply knowledgeable of language as it is in reality, has been a constant
source of inspiration and support for more than twenty years now. Paul Kipar-
sky, the best theoretical linguist I have ever met, taught me everything I know on
how to approach linguistic analysis. Peter Sells, the best teacher of linguistics I
have ever had, recognized the potential of the theory at a very early stage, helped
9780190256623-Dor.indb 11 03/06/15 6:17 PM

xii • Acknowledgments
and encouraged me, and gave invaluable strategic advice at a time when the pros-
pects of publishing did not look very promising. Thank you, Ruth, Paul, and
Peter: it has been an honor.
I owe a great debt of gratitude to Nick Enfield, the editor of the series Foun-
dations in Human Interactions. Nick read an early version of the manuscript,
managed to look beyond its obvious faults and see what I was trying to do—and
then directed me in the process of turning it into something worth publish-
ing. Thank you, Nick, it really wouldn’t have happened without you. And many
thanks to all the good people at Oxford University Press, the two anonymous
reviewers, Hallie Stebbins, Jamie Chu, Bactavatchalane Gogulanathan, Susie
Hara and Carolyn Napolitano. It has been a pleasure working with you.
More than twenty years ago, evolutionary biologist Eva Jablonka and I started
a conversation on language and its evolution, and the conversation has simply
continued without interruption ever since. We developed a dynamic model of
the co-evolution of language and its speakers, and wrote extensively on this and
related topics. After so many years of collaboration, none of us can really remem-
ber who brought which idea to the table. I am forever grateful to her for her
penetrating wisdom and creativity, her total honesty, and most importantly, for
being such a wonderful friend.
What I owe to Lia Nirgad, my very own Lia’le, lies safely beyond the power of
words—unless they are used for poetry. My wife, my love, my best friend in the
world, my partner in activism, my closest reader, the artist of language that I live
with—thank you for just being yourself. Gabi, Talia, and Daniella, thank you
for believing all these years, and for being exactly who you are. I am very lucky
indeed. Galia and Ramon, thank you for always having been there.
Finally, I dedicate this book to the memory of Yehuda Elkana, an intellec-
tual adventurer and entrepreneur, a man of wisdom, passion, and endless energy,
the only man I ever met who could see the entire map of science and its history,
the only chef who has ever moonlighted as a university rector—my mentor, my
friend, in many ways my father. I cannot even begin to describe what it meant
for me to have Yehuda in my life. I only know I miss him every day. May his soul
rest in peace.
9780190256623-Dor.indb 12 03/06/15 6:17 PM

T h e I n s t r u c ti o n o f I m a g i n ati o n
9780190256623-Dor.indb 13 03/06/15 6:17 PM

9780190256623-Dor.indb 14 03/06/15 6:17 PM
1 I n t r o d u c ti o n
This book attempts to tackle the fundamental question of the linguis-

tic sciences: what is language? It is a very old question, and the fact
that the answers keep changing testifies to its complexity and impor-
tance, to the enormous intellectual effort invested in it through the
ages—and, perhaps paradoxically, to the fact that the human condi-
tion is always in transition; it forces us to constantly re-examine and
adjust our conceptions of ourselves, and at every point along the way,
whenever we manage to take a look at ourselves from a new angle,
we find that we also see language in a new light. There is not much
choice here: the way we understand ourselves is inextricably and per-
manently entangled with the way we understand language.
The understanding of language that I wish to offer in these pages
begins with a return to a long tradition of thought—a tradition that
was expelled from the linguistic sciences by the cognitive revolution,
fifty years ago, and is now being locally re-considered in certain quar-
ters of the field: the human condition is deeply social, and language is
a social entity. It is a property of the community, of the social network,
the product of a collective process of invention and development. It
resides between speakers, not in them, at a level of organization and
complexity that transcends the individual mind—and cannot be re-
duced to it. The place to look at for the essence of language is not the
mind-brain. It is social life.
This, however, is only the beginning, and as such, not very in-
formative. The question has to be: what type of social entity is lan-
guage? I will argue that it is essentially a communication technology.
What this means, as a first approximation, is that language has to
be ontologically classified together with the other communication
technologies humans have invented, such as the book, fax, tele-
phone, computer games, and Facebook—not together with social
institutions (such as government or the family), or cognitive capaci-
ties (such as vision or rationality). Like other technologies, language
has to be constructed before it can be used, and it constantly de-
velops and changes as a result of usage. It spreads and propagates
9780190256623-Dor.indb 1 03/06/15 6:17 PM

2 • t he ins t r uc t ion of imaginat ion
like other technologies, and like them it has its experts, role models, innova-
tors, guards, rebels, users, and abusers—active and passive to different degrees.
Most importantly, its modus operandi is best analyzed in technological terms.
Language is the first communication technology we ever invented. It has rev-
olutionized human life and actually changed us as a biological species. It is
still, today, the most powerful technology we use. To understand it, and its
dialectic entanglement with the human condition, we have to adopt a social-
technological approach to its study.
At the center of the social-technological approach to be developed in this
book, stands the understanding that language cannot be a general-purpose com-
munication technology. There is simply no such thing: had there been, we would
never need to invent anything else. Every communication technology is func-
tionally specific. Every technology employs a specific functional strategy, and the
specificity of the strategy determines the technology’s functional envelope: what
it can do with high levels of efficiency, where its efficiency declines, where it col-
lapses, and what it cannot do to begin with. To get to the bottom of language, to
turn the acknowledgment of its social essence into a scientific theory, we have to
ask: What is the functional specificity of language as a socially constructed com-
munication technology?
The functional specificity of language, I will argue, lies in the very particu-
lar functional strategy it employs. It is dedicated to the systematic instruction of
imagination: we use it to communicate directly with our interlocutors’ imagina-
tions. All the other systems of intentional communication, technological or not,
used by humans and other species (with the possible exception of bee dances),
work with what I will call the experiential strategy. They provide materials for
the interlocutors to experience with their senses and thus allow for the actual
sharing of experience. The experiential strategy is thus inherently limited: an ex-
perience can only be shared if it can be experienced. Language is the only system
that goes beyond the sharing of experience. It allows speakers to intentionally
and systematically instruct their interlocutors in the process of imagining the
intended experience—instead of directly experiencing it. The speaker provides
the receiver with a code, a plan, a skeletal list of the basic co-ordinates of the
experience—which the receiver is then expected to use as a scaffold for experi-
ential imagination. Following the code, the interlocutor raises past experiences
from memory, and then reconstructs and recombines them to produce novel,
imagined experiences. Language is thus the only system that allows for commu-
nication that actually bridges the experiential gaps between speakers. In doing
that, it opens a venue for human sociality that would otherwise remain closed.
This is the secret of its success, and it is also the way it actively participates in the
construction of the human individual as a social being.
9780190256623-Dor.indb 2 03/06/15 6:17 PM

Introduction • 3
Nothing in the social-technological approach to be developed here denies

that language raises crucial questions about our cognitions. On the contrary:
language is a technology that requires much of its users, a whole array of cogni-
tive (and emotional) capacities, and these deserve our full attention. What the
approach insists on is that in order to understand the capacities (in acquisition,
in actual communication)—and in order to understand how they evolved—
we must first of all position the technology itself at the right ontological level,
where it belongs. We have learned an enormous amount in the last fifty years
about language-related cognition, and we now know enough to finally abandon
the computer metaphor—the idea that language, like other cognitive capacities,
should be investigated as if it were a piece of software in the computer that we
call the mind-brain—and replace it with the updated metaphor of the Web: lan-
guage is indeed a piece of software—a communication software. It resides on the
Net. Human individuals are end-users: they download copies of the software into
their mind-brains, and these allow them to communicate with one another (to
the extent that the copies are similar enough). What we understand today about
the processes and capacities involved in the downloading and use of the soft-
ware by the end-users should teach us exactly this: that the essence of the soft-
ware itself does not lie there. The essence of the software lies in the fact that it
resides between the end-users; that it facilitates something that is not within
the individual capacity of any of them. As Humboldt (1936 [1999], p. 42) puts
it, “the existence of languages proves that there are also mental creations which
in no way whatever pass out from a single individual to the remainder, but can
only emanate from the simultaneous self-activity of all.” Language cannot be
explained in terms of the cognitive dynamics taking place within the individ-
ual mind-brain, for the exact same reason that social networks on the Internet
cannot be explained on the basis of whatever is happening in the personal com-
puters of the networks' users. As I will try to show, the fact that the cognitively
oriented theories of language currently under discussion in the literature are so
baroquely complex testifies to exactly this: they are attempts to achieve an im-
possible goal. Language is a much simpler, much more reasonable entity, but to
see that, we have to see it for what it is.
My hypothesis, then, is that everything in need of theoretical explanation
in linguistics and around it follows, directly or indirectly, from the essence
of language as a socially constructed, imagination-instructing communication
technology: the architecture of language, the way it is socially constructed and
the way it works in actual communication; the properties of words and their
meanings; the inextricable relationship between language and truth; the dia-
lectical relationship between language and the ways we experience the world;
the patterns of its grammatical structures; its dual life as a universally diverse
9780190256623-Dor.indb 3 03/06/15 6:17 PM

phenomenon; the inherent difficulty and fragility of linguistic communica-

tion, its successes and failures; the dynamics of language change; the intrica-
cies of language-based identity; the political power of language; the nature of
language acquisition; and the evolutionary dynamics that brought language
about in the first place: pre-linguistic human collectivities gradually distanced
themselves from ape sociality, and developed a multi-faceted and highly so-
phisticated culture, still without words—a culture that brought the strategy
of experience-sharing to the limits of its functional envelope. This was, at the
very same time, the foundation upon which language was built—and the cause
for its emergence. As human existence came to depend more and more on the
overall quality of the collective, the need for a way of communicating that
could go beyond the limits of the envelope gradually turned into a necessity—
as always, the mother of invention. As the explorations into the new realm of
communication stabilized to form the first working prototypes, human evo-
lution entered a new era. Language continued to evolve at the cultural level,
and it dragged human societies and human individuals—their behaviors and
identities, cognitions and emotions, physiologies and genetic makeups—into
a fascinating web of co-evolutionary spirals. First we invented language. Then
language changed us.
It goes without saying that I have no intention of actually explaining every-
thing in need of explanation about language—in this book or in general. What I
will try to do is present a general theory of language explicit and detailed enough
to provide a new conceptual foundation for the scientific inquiry into everything
in and around language. Philosopher of science Carl Hempel’s classical defini-
tion of scientific theory captures what I aim to achieve with great precision and
elegance. It deserves to be quoted in its entirety:
A scientific theory might therefore be likened to a complex spatial net-

work. Its terms are represented by the knots, while the threads connect-
ing the latter correspond, in part, to the definitions and, in part, to the
fundamental and derivative hypotheses included in the theory. The whole
system floats, as it were, above the plane of observation and is anchored to
it by the rules of interpretation. These might be viewed as strings which
are not part of the network but link certain points of the latter with spe-
cific places in the plane of observation. By virtue of these interpretive
connections, the network can function as a scientific theory: from cer-
tain observational data, we may ascend, via an interpretive string, to some
point in the theoretical network, thence proceed, via definitions and hy-
potheses, to other points, from which another interpretive string permits
a descent to the plane of observation. (Hempel 1952, p. 36).
9780190256623-Dor.indb 4 03/06/15 6:17 PM

Introduction • 5
Hempel’s characterization of scientific theory as a spatial network of terms, defi-

nitions, and hypotheses, floating above the plane of observation as an interpreta-
tive map, makes it immediately clear that the issue is not just the explanation of
this set of data or the other, or the formulation of hypotheses and predictions, or
the development of tools for empirical investigation. What is at stake is the es-
tablishment of a complex web of relationships (of many different types) between
ideas, arguments, observations, and interpretations that eventually produces a
general sense of clarity, coherence, and unity and allows for that freedom of in-
terpretative movement, up and down, between the plane of theory and the plane
of observation, so beautifully described above. The challenge of theory is of the
architectural type.
What I will do, then, is begin with the careful construction of the theoretical
network—with its terms, definitions, and hypotheses—and then show how it
allows for a major re-thinking of the plane of observation, on three interrelated
levels. First, I will select a representative set of problem-clusters from the differ-
ent sub-domains of the linguistic sciences and show how the theory offers to
deal with them—not in order to provide them with full accounts, but in order to
demonstrate their susceptibility to the type of treatment offered by the theory.
The goal here will be to show that the theory offers a fresh outlook on the various
problem-clusters; suggests new answers to foundational questions; resolves per-
sistent difficulties; allows for serious re-arrangements of the data; invests major
findings and discoveries in the different fields with new meaning; releases the in-
vestigation from deadlocked controversies and unwarranted stipulations; simpli-
fies analyses; opens new venues for empirical research; and so on. As the theory
attempts to re-position language in the social domain, the most important chal-
lenge here will be to demonstrate that it does not lose sight of the enormous
achievements of linguistics as a cognitive science. There is no reason to make that
mistake again.
The technical discussions of the various problem-clusters will thus be as ex-
plicit and detailed as the task of demonstration requires, not more than that.
They should be read as attempts to show, in a rather informal manner, how the
theory allows for new types of principled explanations—and what more serious
work on the empirical issues should look like. Most of the examples, moreover,
are going to be in English, but nothing is implied by this choice—definitely not
that there is something universal about the conventions of English. It is exactly
because language is universally a conventional system, that any demonstration
will have to make use of the particular conventions of this language or the other.
The fact that you and I are at the moment using English for communication
suggests that a demonstration in English would probably be the most reasonable
choice.
9780190256623-Dor.indb 5 03/06/15 6:17 PM

At the second level, I will try to demonstrate that the particular analyses of
the various problem-clusters converge to produce a new and unified outlook on
the general picture of human language: that the theory presents enough evidence
to support the claim that all the different problem-clusters, currently subjected
in the literature to highly specialized and incompatible explanatory apparatuses,
are in fact susceptible to the same type of treatment; that the theory re-arranges
the problem-clusters with respect to each other in fruitful ways, and raises new
explanatory bridges between them; that the entire move shows the way toward
the reassembly of the puzzle of human language as a unified phenomenon.
At the third level, I will be interested in the impact of the theory on the rela-
tionships between theoretical linguistics and all the other disciplines around it
which have a vested interest in the linguistic plane of observation—all the way
from psychology and biology to sociology, communication studies, and critical
studies. I will try to show that the theory creates new harmonies between theo-
retical linguistics and its neighbors; moves them closer together and helps ratio-
nalize their relationships; opens new venues for mutual influence, inspiration,
and co-operation; and thus in effect helps re-position language in its rightful
place within the overall picture of human life. Here, again, the most important
challenge will be to demonstrate how the theory re-connects linguistics with the
socially oriented traditions (including the postmodern ones)—without severing
the ties it has worked hard to establish with its cognitively oriented neighbors.
1.1 Setting the Theoretical Stage: Where Are We Today?

Fifty years ago, a proposal for a general theory of language was put forward by
Noam Chomsky. Not everybody agreed, of course, but no one offered a com-
prehensive alternative. Chomsky’s theoretical strategy was clear, bold, and very
exciting, and it remained virtually untouched through the entire developmental
history of his theory, from the logical structure of linguistic theory (1955) and syn-
tactic structures (1957), through the aspects model (1965), government and bind-
ing (1981) and barriers (1986), all the way to the current version of the theory,
the minimalist program (1995). At the center of his spatial network, Chomsky
positioned a set of definitional statements (often characterized, quite mislead-
ingly, as working hypotheses):
(a) Language is a generative system: it allows for the production of an infinite

number of sentences from a finite arsenal of primitives. The constitutive
principles of the generative system are abstract and formal. They are the es-
sence of language.
9780190256623-Dor.indb 6 03/06/15 6:17 PM

Introduction • 7
(b) The fact that we acquire language and use it means that we know the gen-
erative system. We have linguistic competence. The essential question about
language is thus a question about individual cognition: how does the speaker
know the generative system? How does competence come about?
(c) Knowledge does not emerge from experience: the constitutive principles of
language cannot be learned from the input offered to the language-acquiring
child by the surrounding community of speakers. The input is “meager and
degenerate”; the principles are too abstract. The child must come to the word
already equipped with the essential knowledge of language. Competence
must be innate.
(d) Children are capable of acquiring whichever language they find around
them, which means that the constitutive principles are the same in all the
languages of the world. The observed differences between languages are sur-
face phenomena. Where it counts, all languages are identical (or almost iden-
tical, as in the program of principles-and-parameters, Chomsky and Lasnik
1993): they are all founded on a Universal Grammar (UG), coded in our
genes.
(e) The scientific goal of Linguistics is to uncover the principles of UG, and thus
to figure out the essence of the innate and autonomous language organ (an
organ of the mind, irreducible to the physical properties of the brain). Lin-
guistics is a branch of psychology (itself a branch of biology). “The tasks of
the psychologist,” Chomsky writes, “divide into several sub-tasks. The first
is to discover the innate schema that characterizes the class of potential lan-
guages—that defines the ‘essence’ of human language. This sub-task falls to
that branch of human psychology known as linguistics; it is the problem
of traditional universal grammar, of contemporary linguistic theory. The
second sub-task is the detailed study of the actual character of the stimula-
tion and the organism-environment interaction that sets the innate cogni-
tive mechanism into operation . . . A third task is that of determining just
what it means for a hypothesis about the generative grammar of a language
to be ‘consistent’ with the data of sense.” (Chomsky 1998)
From this heavy set of definitional statements emerges a complex and highly par-
ticular strategic attitude toward the task of interpreting the plane of observation.
To begin with, empirical work should concentrate on the observed patterns of
generativity, especially as they manifest themselves in the syntactic structures
of languages. Syntax is where the essence of human language lies. The patterns
of syntactic generativity should be figured out on the basis of native speakers’
grammaticality judgments, not on the basis of their actual usage of language: lin-
guistic performance is full of mistakes, hesitations, repetitions, and repairs (and
9780190256623-Dor.indb 7 03/06/15 6:17 PM

this is why it cannot provide the child with sufficient input); competence is only
reflected in judgment. Through the analysis of speakers’ judgments, the linguist
discovers the principles that govern them—those innate principles that make
up the human capacity for language. These principles are abstract and formal,
which means that the analysis cannot, and therefore should not, attempt to
relate them to the meanings of the sentences, or to the communicative intents
of their speakers, or to the context of conversation. In this challenge of distilla-
tion, everything that does not directly relate to the foundations of generativity
is moved aside. Other issues may be interesting, even important, but they are
not essential: performance, meaning, communication, context, and also general
cognition, and social learning, and language change (the linguist’s task is that
of synchronic, not diachronic analysis), and society, culture, semiotics, rhetoric,
literature—virtually everything that the non-linguist might consider relevant
for linguistic research. Generativity deserves to be investigated in isolation.
This was a radical strategy indeed, and it brought about a huge revolution in
our understanding of language. There was, however, something deeply paradox-
ical in the way it did that: the theory itself, as it continued to be developed by
Chomsky and his colleagues of the generative camp, developed into an exceed-
ingly esoteric discourse, baroquely complex and deeply abstract, often interested
more in the internal relationships between its terms, definitions, and hypotheses
than in the strings of interpretation that were supposed to anchor the whole the-
oretical machinery to the plane of observation. With time, even the judgments
lost much of their importance. In a way, this was inevitable. The third task of
the psychologist-linguist, as Chomsky defined it, was “that of determining just
what it means for a hypothesis about the generative grammar of a language to be
‘consistent’ with the data of sense.” If the child could not trust the input, there
was no a priori reason to assume that the linguist should. Empirical science does
not grow very well on rationalist ground.
This, however, was only one side of the coin, actually the less interesting.
Chomsky’s general perspective sent a shock wave through the linguistic com-
munity and beyond it, across the scientific world. His picture of language in-
vested everything around it with new meaning and new energy. The issue was no
longer the investigation of languages. This was passé. It was not even that hidden
software of language inside our minds, not as such. The issue was the essence of
being human. More than anybody else in the second half of the twentieth cen-
tury, Chomsky was the one who asked the formative question: as a present-day
Plato, he formulated the question as one about the computational foundations
of the human mind. It was no accident that computers, computer science, and
artificial intelligence emerged at the very same time: to program machines that
do what we do, we need to discover the programs that we carry within.
9780190256623-Dor.indb 8 03/06/15 6:17 PM

Introduction • 9
Chomsky’s question was a major engine behind the stellar appearance of

cognitive science in the world of science, and language came to be permanently
positioned at the very center of the drama. A huge number of new research pro-
grams on language and human cognition appeared, and virtually all of them
were strategically designed to deal with different facets of Chomsky’s ques-
tion. Many corners of the plane of observation were visited for the first time,
and many were re-examined in new ways. New specialties developed, and sub-
disciplines went from first programmatic papers to yearly international confer-
ences. From the very beginning, the booming field acquired a certain energy,
highly argumentative and polemic: researchers were either coming back from
their travels on the observational plane with the ultimate proof that Chomsky
was perfectly right, or with the strongest demonstration possible that he was
utterly wrong. In this sense, Chomsky’s program directed the development of
his rivals as much as it directed that of his supporters. They played against him
on his field.
The last fifty years, then, have seen enormous developments in language re-
search, and it seems fair to state that most of the work has been dedicated to the
attempt to bring back into the puzzle those pieces that Chomsky deliberately
left out. Meaning and communicative interaction were brought back with the
development of functional and cognitive linguistics (Fauconnier 1994, Fillmore
1968, Lakoff 1987, Langacker 1987), and then frame semantics (Fillmore 1982,
Fillmore and Baker 2009) and the theory of constructionism (Croft 2001, Gold-
berg 1995, 2006, Östman and Fried 2005); with the development of pragmatic
theory (Austin 1962, Searle 1969, Grice 1975, Levinson 2000, Sperber and
Wilson 1986); with the emergence of a variety of semantic theories (Jackendoff
1983, 1990, Montague 1973, Rosch 1975, Wierzbicka 1996); and with the ad-
vances made in the understanding of lexical semantics and its relation to syntax
(Jackendoff 1983, 1990, Cruse 1986, Dowty 1979, Kiparsky 1997, Levin 1993,
Levin and Rappaport Hovav 1995).1 Structure and meaning are now understood
to be connected in bewilderingly complex webs of interrelations.
Society and culture came back into the puzzle with the sociolinguistic quest
for socially determined variation (Eckert 2000, 2005, Labov 1966, Trudgill
2011); the anthropological linguistic search for the relationships between lan-
guage, culture and mind (Duranti 1997, Gumperz and Levinson, 1996); and
the growing understanding of the intersubjective nature of language (Meltzoff
and Brooks 2007, Tomasello 2008, Tomasello et al. 2005, Zlatev et al. 2008).
All these have contributed to the current positioning of language somewhere
between human cognition and human society: for many in the field, the im-
plicit working assumption is that language is grounded both in cognition and
in society.
9780190256623-Dor.indb 9 03/06/15 6:17 PM

All of the above, together with what we now know from conversational anal-
ysis (Enfield and Stivers 2007, Sacks 1995, Shegloff 2007) and interactional lin-
guistics (Ochs et al. 1996, Couper-Kuhlen and Selting 2001), has challenged
the distinction between the original notion of competence and the attested pat-
terns of linguistic performance. All this, together with the results accumulated
in linguistic typology (Comrie 1989, Croft 2002)—and the simple fact that
more and more languages came to be deeply researched—has changed our views
on the universality of language. What we see today is a web of similarities and
differences between languages—many restricted, implicational universals; not
too many absolute ones: nothing, definitely, that looks like a set of properties
worthy of the name Universal Grammar (Evans and Levinson 2009, Levinson
and Evans 2010).
The enormous advances we have made in the understanding of language ac-
quisition (Bates et al. 1995, Berman 2004, Bloom 1970, Bruner 1983, Clark
2009, Elman et al. 1996, Slobin 1997, Tomasello 2003), actual linguistic pro-
cessing (Levelt 1989, Harley 2008), and brain activity related to language
(Ahlsén 2006) now position language somewhere between its original auton-
omy and the realm of general human cognition—between innate constraints
(much softer than originally suggested) on language acquisition and usage and
the general human capacity for learning, especially for social learning. All this
has renewed our interest in the dynamics of language change (Aitchison 2001,
Croft 2000), especially the process of grammaticalization (Deutscher 2005,
Lehmann 2002, Traugott and Heine 1991, Traugott and Dasher 2002), and has
positioned language, again, somewhere in between the synchronic and the dia-
chronic. A rich and lively discourse, unimaginable fifty years ago, now attempts
to tackle the most difficult question of all (which Chomsky himself refused to
deal with for four decades): the question of the evolution of language (Arbib
2003, Bickerton 1990, Botha 2003, Botha and Knight 2009, Christiansen and
Kirby 2003a, b, Corballis 2003, Deacon 1997, Donald 1991, Dor and Jablonka
2000, 2010, Dor, Knight, and Lewis 2014, Dunbar 1998, Fitch 2010, Hurford
2007, 2011, Larson, Deprez and Yamakido 2010, Lieberman 2007, Pinker and
Bloom 1990, Richerson and Boyd 2005, Steels 2001, 2014, Tomasello 1999,
2008, and many more).
We have moved quite a long way from Chomsky’s original picture, and we
know much more than we ever did. But the accomplishments did not come with-
out a price: language, the entire thing, has disappeared on the way. Today, we do
not have a general theory of language. We have many pieces of the puzzle, but the
puzzle itself does not assemble. The plane of observation has been parcelized, and
different areas came to be governed by different rules. The science of language
has developed into an extremely fragmented field, in which different explanatory
9780190256623-Dor.indb 10 03/06/15 6:17 PM

Introduction • 11
apparatuses, incompatible with each other, serve the theoretical needs of highly
specialized sub-domains. As a consequence, when we look at our fragmented
pieces of language, situated as they are away from Chomsky’s original picture, we
still look at them, as if by default, from the point of view of that very same picture.
We have found new answers, but they are still answers to Chomsky’s question.
Our partial theories of language are still about human cognition; they are still
answers to the question of knowledge. We have very different grammars, but we
still think about them in the same way, as computational characterizations of
levels of cognitive representation. We struggle to find a place for diversity within
a framework that is still universalistic. We look at language as a social entity
from the point of view of social psychology.
At the most foundational level, then, what I intend to suggest in this book
amounts to the claim that the pieces of the puzzle, the partial answers, would
only fall into place if we agree to look at them as partial answers to a very dif-
ferent question, if we manage to re-interpret them, as Hempel puts it, under the
light of a new network of terms, definitions, and hypotheses. To find language
again, to assemble the pieces, we need a new organizing principle, a new theo-
retical characterization of the essence of language. I will suggest that we need to
look at language as a socially constructed communication technology, not as a
cognitive capacity.
1.2 Language as a Communication Technology

The general idea that language is a socially constructed tool of communication is
far from new. It has featured prominently in the linguistic discourse of the first
half of the twentieth century (Saussure 1916, 1966, Sapir 1921, Meillet 1921,
Gardiner 1932, and others); it has informed much of the semiotic literature on
language (Itkonen 2003, 2008, Zlatev 2009); and it has been debated in the phi-
losophy of language (Wittgenstein 1953, Lewis 1969, Davidson 1984, Dummett
1996). Most of the other disciplines interested in language—communication
studies (Peters 2000), sociology, anthropology, literature, critical studies, and so
on—have always worked with a general and often implicit view of language as an
institution of society.
In the wake of the Chomskian revolution in linguistics, the idea of language
as a tool was explicitly marginalized. Reddy’s (1979) toolmakers’ paradigm was a
notable exception: its influence on my theory is enormous, and I will get back to
it in chapter 3. In the last three decades, however, the idea has gradually come to
be accepted, implicitly or explicitly, in many quarters of the linguistic sciences:
if the capacities involved in language acquisition and usage are general human
9780190256623-Dor.indb 11 03/06/15 6:17 PM

capacities, rather than innate pieces of linguistic knowledge, then language

itself, the object of acquisition, must be out there—available for the child in the
social-cultural environment. “This perspective on human communication and
language,” writes Tomasello (2008, p. 10–11), “thus basically turns the Chom-
skian proposal on its head, as the most fundamental aspects of human commu-
nication are seen as biological adaptations for cooperation and social interaction
in general, whereas the more purely linguistic, including grammatical, dimen-
sions of language are culturally constructed and passed along by individual lin-
guistic communities.” For Croft (2000, p. 26), “a language is the population of
utterances in a speech community,” whereas “a grammar is the cognitive struc-
ture in a speaker’s mind that contains her knowledge of her language.” Based on
his lifelong inquiry into the lives of the Pirahã in Brazil, Everett (2012, p. 6) says
that “languages are tools. Tools to solve the twin problems of communication
and social cohesion. Tools shaped by the distinctive pressures of their cultural
niches—pressures that include cultural values and history and which account in
many cases for the similarities and differences between languages.”
These formulations and others are very important developments. They do
not, however, go beyond the level of declaration. As Botha (1992, p. 237) puts
it, “to say that something—for example an entity, structure, practice or whole
realm of reality—is ‘social’ is not to characterize it ontologically in a fundamen-
tal way.” The authors position language in the social domain and then move on
to more pressing issues: Tomasello to the social psychology of communication,
Croft to language change, Everett to the relationship between language and cul-
ture. Under some readings, this theoretical dynamic has also characterized the
writings of de Saussure and Sapir: according to Koerner (1982), de Saussure was
interested in the development of a general theory of signs—a semiological, or
semiotic theory—and he “appears to have claimed that language is a social fact
just because of its semiological character” (p. 57). Sapir’s work often seems to
deal more with the psychological significance of language than with its social
nature. The same is true of the literature in sociolinguistics and anthropological
linguistics: the social-cultural essence of language is stated informally and then
taken for granted, and the discourse concentrates on the web of relationships
between language and other social-cultural facts—identity, class, power, values,
gender, and so on. This is what Coulmas (1998, p. 3) calls “the theoretical deficit
of sociolinguistics.” In the philosophy of language, the debates revolve around
the question of meaning and its determination vis-à-vis the questions of conven-
tion, usage, and truth. With the exception of Wittgenstein, there is relatively
little that attempts to investigate the ontological foundations of the social reality
of language. There is a regular pattern here: the declarative endorsement of the
social-cultural nature of language has allowed researchers, past and present, to
9780190256623-Dor.indb 12 03/06/15 6:17 PM

Introduction • 13
delve deeply into different aspects of the context within which language oper-
ates, with extremely significant results that will play a central role in the con-
struction of my theory. Language itself, however, the social-cultural entity as
such, has somehow been neglected on the way: it is still severely under-theorized.
What are we looking for, then? Well, basically what we need is a much better
understanding of what language does—a theoretical characterization of its func-
tion as a socially constructed tool of communication. Crucially, we should not
confuse the question of what language does with the question of what we, as
communicators, do with it. These are two separate levels of functional analysis.
To see the difference, consider the following example. Suppose I wish to inform
you that I’m arriving on Tuesday, and ask you to wait for me at the train station
at 7:15 p.m. And assume that I can do this in three different ways: send you a fax,
an email, or a regular letter by mail. So, two questions: (a) What would I do with
each of these technologies? And (b) What would each of the technologies do if
I chose to use it? As far as the first question is concerned, it seems I’ll be doing
the same two things in all three cases: telling you something and asking you to
do something. In Jakobson’s (1960) terms, I will be performing two functions:
the referential and the conative. In Searle’s (1969) terms, I will be performing
two speech acts: the assertive and the directive. If we wish to break the speech
acts down to their locutionary, illocutionary, and perlocutionary components
(Austin 1962), we’ll find that nothing there hinges on my choice of technology
either. As far as the second question is concerned, however, each of the technolo-
gies would give us a radically different answer. The fax would look at my writ-
ten note as visual information, convert it into a bitmap, and transmit it directly
through the phone system to your fax. My email software would read my note
as text, convert it into an electronic signal, and send it to my email server, from
where it would be re-directed through the Internet to your server, and then to
your computer. The mail system would require an envelope and a stamp, and it
would physically carry my written note, through a series of processing centers, all
the way to your door. The essence of the three technologies does not lie in what
I do with them, but in the unique and specific functional strategies that they
employ as technologies. This, then, is the first point: to understand language as a
technology, we have to understand what it does, not what we do with it. We have
to figure out its functional strategy as such.
Positioning the question of functional specificity at the center of our the-
oretical quest, we may finally break away from the deadlocked debate over the
functionality of language as it developed on Chomsky’s field, under Chomsky’s
question, in the last fifty years. The debate revolved around a much more fun-
damental question: Is language functional or not in the first place? The genera-
tivists claimed that it simply was not; the functionalists insisted that it was a
9780190256623-Dor.indb 13 03/06/15 6:17 PM

(cognitively based) general-purpose system of communication (for a thorough

discussion of the debate, see Newmeyer 1998). If we agree to think of language
as a technology, we have to immediately reject both positions. On the one hand,
the essence of a functioning technology simply cannot lie in the non-functional
specificity of its architecture: non-functionality may explain useless technology,
but it is not an explanatory option with such an enormously successful technol-
ogy as language. On the other hand, no technology is ever general-purpose. Just
like any other technology, the essence of language must lie in the fact that its
function is specific. This is where we should start.
As I will show throughout the book, this new perspective does much more
than simply reject the two rivaling positions. It actually allows us to accept many
of the positively formulated observations of both camps, while rejecting their
negatively formulated explanations. No, language is not non-functional or non-
specific. As a technology, it cannot be any of those. Yes, language is functional
and specific. As a technology, it must it be both. From our new perspective, we
will find again and again that the two camps looked at two sides of the same
coin, concentrating on the observations of their own side, and working hard
to discredit the idea that there is another. Released from the debate, we will be
able to re-evaluate the relevant observations and assign them a new meaning.
Having rejected Chomsky’s question, we will even find, from time to time, that
we should adopt some of his answers.
1.3 The Structure of the Argument

The next two chapters provide a first systematic look at the spatial network of the
theory—its terms, definitions, and hypotheses. In chapter 2, I define the specific
functional strategy of language, and show how it is different from the strategies
employed by all the other communication systems, technological or not, that are
used by us and other biological species. In chapter 3, I present a technical descrip-
tion of language the way I see it—its constitutive parts, their social construction,
the way they fit, how they function together to allow for the production and in-
terpretation of complex sets of instructions for imagination, and the parameters
that determine the overall quality of the instruction.
Following this definitional discussion, chapters 4 and 5 descend to the
plane of observation and deal with a set of issues having to do with meaning.
In chapter 4, I show how the theory re-frames some of the foundational ques-
tions of word meaning: how do words mean? How do their meanings connect
them to the world of experience and to other words? Why do they manifest the
very particular behavioral patterns that they do? In chapter 5, I present a major
9780190256623-Dor.indb 14 03/06/15 6:17 PM

Introduction • 15
re-interpretation of the question of linguistic relativity, as a question about the

dialectic influence of a technology on its users—a re-interpretation that posi-
tions the question in its rightful place at the very center of linguistic theory, as-
signs the accumulated results in the field with new significance, and opens the
way toward new ways of research.
Chapter 6 takes us back up to the spatial network of the theory, with a more
detailed definitional discussion of the processes involved in the production and
comprehension of linguistic utterances. It then descends back to the plane of ob-
servation, and shows how the definitional description captures some of the most
important recent findings in psycholinguistics and how the stormy relationship
between psycholinguistics and general linguistics can be rationalized. In chap-
ter 7, I use all this to deconstruct the question of syntax, demonstrate the very
different way the theory handles syntactic complexity, and claim that Chomsky
was actually right in his insistence on the autonomy of syntax from general, indi-
vidual cognition—but for the wrong reason: syntactic complexity is not a matter
of individual cognition; it is socially constructed, prescriptive, and specifically
suited for the instruction of imagination.
Chapters 8 and 9 are all about variability. In chapter 8, I show how the theory
captures the attested patterns of linguistic diversity around the world, and how
it re-conceptualizes the universality of language as a foundationally social fact—
not a cognitive one. This carries important implications for research. In chap-
ter 9, I claim that language-related cognitive capacities are also variably spread
within communities of speakers. The universal mind of the human speaker is a
myth, and so is the universal mind of the language-acquiring child. This move
allows for a new hypothesis about the collective essence of the process of lan-
guage acquisition, which also sheds new light on ongoing construction of sign
languages around the world, in Nicaragua and elsewhere.
Chapter 10 connects all these issues together in a new hypothetical expla-
nation of the evolution of language as a collectively constructed communica-
tion technology. Based on my work with evolutionary biologist Eva Jablonka,
and on a wealth of insights from current evolutionary biology, the evolution of
the human species in general, the evolution of technology, and the evolution of
language, I claim that the specific function of the instruction of imagination
was collectively invented, on the basis of everything that was achieved by pre-
linguistic societies, before individual speakers were specifically adapted to it.
Our variable, innate dispositions toward the participation in the collective ac-
tivities of language emerged for a technology that was already there in the social
domain. In the conclusion, I show how the theory, with its derivative explana-
tions of the different areas on the plane of observation, begins to allow for the
systematic reassembly of the puzzle of language.
9780190256623-Dor.indb 15 03/06/15 6:17 PM

2 T h e F u n c ti o n a l S p e c i f i c it y
of L anguage
The question, then, is this: What does language do as a socially

constructed communication technology? What is its functional speci-
ficity? I will make my way toward the proposed answer in three stages.
The third stage will culminate with the definitional characterization of
language as a technology designed by cultural evolution for the specific
function of the instruction of imagination. This will be based on a cer-
tain perspective on animal communication in general, to be developed
in the second stage, a perspective that immediately exposes the consti-
tutive difference between language and all the other communication
systems used by animals, including humans. To develop this perspec-
tive, I will have to begin with a definitional discussion of a concept that
will play a central role in the entire book: the concept of experience.
2.1 Experience, Experiencing, and the Experiential Gap

The notion of experience has a complicated history. It has emerged
from the late nineteenth century to the middle of the twentieth, in a
series of seminal works in philosophy (Dewey 1938, Heidegger 1927,
Merleau-Ponty 1945 [2005], Wittgenstein 1953) and psychology
(James 1890, 1904, Vygotsky 1978, 1986). The notion was developed
as a counter-reaction to the dominant view of the human mind, based
on Cartesian philosophy, as a solipsistic machine, thinking about the
world from the isolated viewpoint of the detached observer. It thus
systematically highlighted the pragmatic, situated, embodied, and in-
tersubjective essence of our lives as mental creatures. When modern
cognitive science came on stage, with its Cartesian conception of
cognition as a system of formal computational operations on abstract
symbol-like entities, the notion was predictably expelled from the dis-
course: we have already seen Chomsky’s rejection of the very idea of
experience as an explanatory concept. Then, in the last twenty years,
when cognitive scientists began to look for new ways to contextualize
9780190256623-Dor.indb 16 03/06/15 6:17 PM

The Functional Specificity of Language • 17
cognition, the concept finally came to its rightful place—at the center of a new
scientific perspective on cognition, variously known as situated cognition, em-
bodied cognition, and distributed cognition (Robbins and Aydede 2009). The
perspective is committed to three central ideas: “First, cognition depends not
just on the brain but also on the body (the embodiment thesis). Second, cogni-
tive activity routinely exploits structure in the natural and social environment
(the embedding thesis). Third, the boundaries of cognition extend the boundar-
ies of individual organisms (the extension thesis)” (p. 3). As far as language is
concerned, this perspective inspired a new focus on its intersubjective founda-
tions: the human capacity for cooperation, collective thought, and cultural pro-
duction; the facts of experience sharing, social learning, mimesis and imitation,
mind-reading, joint attention, and empathy (Clark 1996, Tomasello 1999, 2008,
2009, Donald 1991, Zlatev 2008, and many others).
In light of all this, the definitional conception of experience that I intend
to position at the center of my theory may strike some readers as strangely
reactionary. I will suggest that, in order to understand language, we have to
take a step back and contemplate the extent to which human experience, with
all its situatedness, is still a private phenomenon. There are many reasons for
this move, and they will show themselves throughout the discussion. At the
moment, let me just stress that I do not deny that our mental lives are heavily
intersubjective. I just insist that we need to put intersubjectivity on the side (if
only for a short while), as a definitional move, because it masks the functional
essence of language. This is mainly because the technology of language is not
just immersed in an intersubjective world—it also actively participates in the
construction of intersubjectivity. In this sense, the insistence on the private
nature of experience would probably sound much less problematic, actually
quite trivial, if we imagined a world in which we would no longer be allowed
to use the technology—to speak, sign, or write to each other. No stories, no
questions, no opinions, no explanations, no advice. We would still be able to
do much together, to communicate on a lot of levels, to be with each other and
feel each other, but the pain of epistemic solitude (that we have no problem
identifying in ourselves even when language is around) would obviously be
felt much more severely. This is what we need to touch in order to understand
language.
Let me begin, then, where I adopt the perspective of the experiential tradi-
tion in its entirety. First, the notion of experience, and experiencing, should be
understood as referring holistically to everything that happens to us as mental
creatures: feeling, thinking, understanding, seeing, hearing, imagining, wish-
ing, and so on—and also, importantly, everything that happens in our ner-
vous system when we do, move, touch, react, try, succeed, and fail. Theoretical
9780190256623-Dor.indb 17 03/06/15 6:17 PM

distinctions can be made between these different aspects of experience, but the
essence of our experiential lives lies exactly in the fact that they are all intercon-
nected in a holistic system. Like all other animals with a nervous system, we
live in experience.1 Oakeshott (1933, p. 10) makes this point in the following
charismatic paragraph:
Reflection on the character of experience finds in most men a mind filled

with prejudice and confusion; and not uncommonly these prejudices and
this confusion will be found to spring from distinctions elevated into
differences. One such distinction is that which divides experience into
the part which may properly be called thought, that which (because it is
more elementary or immediate) falls short of the condition of thought,
and that which passes beyond the condition of thought. Thought, we
have been told, is a particular mode of experience, which must be dis-
tinguished at once from mere consciousness, from sensation, from per-
ception, from volition, from intuition, and from feeling. And we shall
perhaps find ourselves to have inherited also beliefs about the order and
relative validity of such modes of experience.
Now, it need not be denied that, for some purposes, such an analysis
of experience may be relevant and useful. These names certainly stand for
what can be distinguished. Nevertheless, it is difficult to understand how,
if it be pressed to its conclusion, the issue can be anything but one of error.
For in the end, a consciousness which is mere consciousness (and not a
thinking consciousness) turns out to be a mere contradiction; sensation,
when it is isolated, turns out a meaningless abstraction; and intuition
achieves independence only to discover that the price of it is nonentity.
And the view I propose to maintain is that experience is a single whole,
within which modifications may be distinguished, but which admits of
no final or absolute division.
Second, the concept of experience carries a commitment to a dynamic and

emergent view of the process of experiencing. “According to my view,” writes
James (1904, p. x), “experience as a whole is a process in time, whereby innumer-
able particular terms lapse and are superseded by others that follow upon them
by transitions which, whether disjunctive or conjunctive in content, are them-
selves experiences, and must in general be accounted at least as real as the terms
which they relate.” The process of experiencing is a “quasi-chaos” (p. y). Knowl-
edge emerges in a non-linear fashion. As our experiences accumulate in our
embodied minds (leaving their traces in our nervous system), we detect similar-
ities and analogies between them, and construct generalizations—experiential
9780190256623-Dor.indb 18 03/06/15 6:17 PM

generalizations, always analogue, holistic, fuzzy, and context-dependent—

which then color, shape, and sometimes determine the way we further experi-
ence. This is how we learn. This does not deny the possibility that our nervous
systems might be innately biased, in different ways, toward certain ways of ex-
periencing and accumulating experiences (Rupert 2009, Elman et al. 1996).
What it does deny, and very strongly so, is the idea that our general cognition
can be described—let alone explained—in terms of the manipulation of ab-
stract symbols. It thus denies the existence of a Fodorian language of thought
(Fodor 1975) and the existence of conceptual structure as a level of cognitive
representation (Jackendoff 1990). 2
This is absolutely crucial, because language, as opposed to experience, is a
system of symbol manipulation. The assumption that general cognition is a
system of a similar type has thus created an illusion of sameness between the
way we know (and think and feel) and the way we speak—between the cat-
egorizations of our experiences and the categorizations of our languages. What
that implied, in turn, was that the mapping between language and general
cognition—the mapping that we perform when we actually use language—is a
relatively transparent business. The understanding of the holistic and emergent
nature of experience makes it clear that the challenge is of a different order of
complexity. To capture the functional specificity of language, we have to get a
good grasp of this complexity.
This is why I also have to insist on the private nature of experience. Not in
order to deny the fact of intersubjectivity, but in order to highlight the magni-
tude of the challenge of language. In a world of complete intersubjectivity, we
would all be experiencing the same things in the same ways. In such a world, we
would not need language, definitely not a language as complex as the languages
we use. Intersubjectivity is an important component of our experiential lives, but
we still experience it from our private perspective. We experience on our own,
within ourselves, even when we experience together. We are separated from each
other by experiential gaps: the processes of experiencing that we go through are
inaccessible to the others. The gaps are variously deep, depending on a long series
of parameters, including the level of intimacy in our relationships, but they are
always there.
Two contextualizing comments seem to be required at this point. The first has
to do with the evolution of human experiencing. I agree with Tomasello (1999,
2008, 2009), Donald (1991), Zlatev (2008), and many of the authors in Dor,
Knight, and Lewis (2014) that early humans must have gone through an entire
stage of evolutionary development, before language, that resulted in the emer-
gence of a new type of human culture: co-operative, mimetic, and intersubjec-
tive. In chapter 10, I will present a detailed hypothesis on how language emerged
9780190256623-Dor.indb 19 03/06/15 6:17 PM

from within this culture. What I do not agree with is the implicit assumption
that the evolution of human experiencing into the intersubjective dimension
somehow erased whatever was there before. The unparalleled human capacity
for intersubjectivity evolved from an experiential foundation much closer to so-
lipsism, and this foundation is still there, at the very core of our experiential lives:
intersubjectivity, like everything else, is always experienced privately. Failing to
see this leads to another illusion of sameness between the experiences of different
individuals, which, again, masks the challenge that language has to overcome.
Language is indeed made possible by intersubjectivity, both ontogenetically
and philogenetically, but what it does goes beyond intersubjectivity. To under-
stand what language does, we have to begin with human individuals as private
experiencers.
Second, nothing in the conception of human experiencing suggested here
denies the Vygotskian conception of private experiencing as socially constructed
(Vygotsky 1978). The very idea of the zone of proximal development is based on
the acknowledgment that the process of socialization involves privately experi-
encing individuals. The zone is “the distance between the actual developmental
level as determined by independent problem solving and the level of potential
development as determined through problem solving under adult guidance, or
in collaboration with more capable peers” (p. 86). As a Marxist, Vygotsky may
have held to the belief that his educational process may eventually bring children
to the end of the developmental voyage, to a place where socialization done right
actually allows their levels of performance to meet their potentials. As a develop-
mental psychologist he probably knew that the gap is never really closed. Where
I do cross paths with Vygotsky, and this is quite paradoxical, is on the role of
language in the developmental process. Vygotsky seems to take all social expe-
riencing to involve language, so he only distinguishes private experiencing from
social experiencing. For me, there are three processes rather than two: (a) private
experiencing (independent problem solving), (b) social experiencing without
language (solving problems under guidance and in collaboration, without the
usage of language), and (c) social experiencing with language (solving problems
together while talking about it.) In my terms, then, Vygotsky was interested in
the distinction between (a) on the one hand, and (b) and (c) on the other, whereas
I’m interested in the distinction between (a) and (b) on the one hand, and (c) on
the other. My own view on the influence of language on development—language
as such as opposed to the entire world of social experiencing—is more modest
and more nuanced than Vygotsky’s. We shall get back to this in chapter 5.
The insistence on the private nature of experience, then, highlights the fact
that every human individual, each of us, experiences the world in different ways.
Every individual looks at the world from his or her own egocentric perspective;
9780190256623-Dor.indb 20 03/06/15 6:17 PM

everyone carries a different baggage of memories, different private histories of in-

teraction with different worlds. Every individual comes to rely on different strat-
egies for understanding the world; each is by nature (and by enculturation) more
deeply attuned to certain aspects of the world than to others. Different indi-
viduals have different attention spans; different perceptual capacities; a different
talent for detail; a different eye for distinction; a different capacity and interest
in generalization; different conceptions of the order of things, of what belongs
with what, what is similar to what, what is the cause or effect of what. They have
different ideas of what is beautiful, frightening, useful, important, or interesting.
They are different in gender, age, social status, physical strength, emotional de-
velopment, curiosity, patience, and anxiety level. Every human individual lives in
a different experiential world. This, again, does not deny that there are similari-
ties. If there were not any, we could not speak at all. It just highlights the fact that
speakers always come into a conversation from the two sides of an experiential
gap. The gap is where the conversation takes place.
This definitional characterization of experience—holistic, emergent, non-
symbolic, private, variable—leaves very little that is shared (in both senses of the
word) by all humans. It concentrates on the differences that we have between us.
In this foundational sense, it marks a radical departure from current views. For
Descartes and Kant, and for the great majority of cognitive scientists (Elizabeth
Bates and Steven Levinson are notable exceptions), the individual is the site of
the universal. We are all separate but similar. In the experiential, intersubjective
tradition, we are similar because we are not separate. As far as the social sciences
are concerned, we are similar to other members of our culture (or sub-culture),
because we are enculturated together. My claim is that we have to begin our jour-
ney with the acknowledgment that the individual is the site of variability. Each
human individual lives in a private, experiential world that is different from that
of the others, and is inaccessible to them. This is a foundational fact about our
experiential nature, and it is the foundational obstacle to communication that
language, as a social invention, set out to circumvent. Human cognition partici-
pates in the story of language not just as part of the origin, but also—much more
importantly from the point of view of language—as the original problem that
had to be solved by a social technology.
2.2 The Definition of Intentional Communication

Humans are obviously not the only species that communicates, and language
is not the only system that we, as humans, use for communication. To figure
out the functional specificity of language, we have to find a way to distinguish
9780190256623-Dor.indb 21 03/06/15 6:17 PM

it from all the others. This comparative strategy has a long history (see Hauser
1996 and Fitch 2010 for extensive reviews). The most cited comparison these
days is Hockett’s (1960), to which we shall return later. Here, I will suggest a
different approach to the question, based on the notion of experience as defined
above. The approach will be presented in three stages. First, a distinction will
be proposed between intentional and non-intentional communication systems.
Language will belong in the first set, together with an entire series of non-
linguistic systems. Second, all the intentional communication systems apart
from language will be characterized as experiential, and will be shown to be
either presentational or re-presentational. Third, language will be characterized
as the only intentional system that works with a non-experiential strategy, which
I will call instructive. The functional specificity of language lies in this strategy.
Language instructs the imagination.
I will define intentional communication in a deliberately maximalist way: in
intentional communication, every act of communication is driven in real time
by an experiential intention, on the side of the communicator, to modify the ex-
perience of another individual (or more) in a certain way. What this means, in
terms of our above discussion of experience, is that intentional communication
is by definition an attempt to overcome the obstacle of the experiential gap: to
get something across to the other side. The definition is maximalist in four ways.
First, it is not interested in the fact that different acts of intentional communi-
cation modify different aspects of the other’s experience: some target the other’s
behavior, some the other’s emotions, some the other’s beliefs. Second, it is not in-
terested in the amount of control the communicator has over the communica-
tive act: the experiential urge to communicate may be triggered automatically
by fixed environmental cues, or it may be volitional (to different extents). Third,
the definition ignores the distinction between what is innately given and what is
learned in the act of intentional communication. Fourth, it puts aside the ques-
tion of the availability of the experiential intent to the communicator’s reflective
consciousness: to intend to do something and to know that you intend to do it
are very different things, and the former is not dependent on the latter.
This definition of intentional communication leaves out quite a lot that is
clearly non-intentional. Evolution has provided many species with exquisite means
of communication that do not require any intentional activity. As Fitch remarks,
“the coloring of an orange and black caterpillar informs potential predators of its
nasty taste, but the caterpillar presumably has no psychological representation
of this ‘meaning’” (p. 190). The definition also marks as non-intentional all the
different ways in which a non-communicative behavior of an individual provides
the others with important information: a sick individual may behave in ways that
betray fragility and weakness, and thus make it clear to predators that this one is
9780190256623-Dor.indb 22 03/06/15 6:17 PM

going to be easy, but this is not intentional communication. Many communication

systems used by animals in the wild seem to reside in the gray area between the in-
tentional and the non-intentional. Vervet monkeys, for example, do not produce
their alarm calls when no conspecifics are around (Cheney and Seyfarth 1985), but
does that necessarily mean that when they do, they produce them with an inten-
tion to alarm the others? The jury is still out, and we may actually never know—
but nothing in our argument here hinges on the answer. What does matter is the
fact that we have a large set of communication systems that are clearly intentional:
they are all driven in real time by an intent (conscious or not) to modify the expe-
rience of the others (whether their behavior, emotions, or beliefs.)
2.3 Experiential Communication: The Sharing of Experience

All systems of intentional communication, apart from language, are what I will
call experiential systems. They allow communicators to convey their experien-
tial intents in ways that are in and of themselves experiential. They allow for the
sharing of experience. There are two general types of experiential systems: presen-
tational and re-presentational. The great majority of systems used by humans and
other species are presentational. A frowning expression, a smile, a hug, a kick,
a threatening posture, demonstrations of physical strength, different forms of
mimesis (and teaching through mimesis), manual gestures, grunts and screams,
music and dance—all these employ the strategy of presentation. In all of them,
the communicating individual behaves in a way that is, at the very same time, de-
termined by his or her experiential intent and made available for experiencing by
the receiver. Presentational communication takes place within experience, in the
here-and-now that is experientially shared by both sides to the communication
event. Its very essence lies in the fact that the internal experience of the commu-
nicator, whatever it is that gives rise to the entire event, is not hidden away, but
is made present, revealed, mirrored in perceptible behavior, which immediately
and directly becomes part of the experience of the receiver. There are, to be sure,
many important differences between the systems: some imply different levels of
dialogical coordination (sometimes turned into ritualized convention); others do
not. Some require a certain amount of learning (on both sides); some seem to be
totally instinctive. The point that unites them, however, is that they all take place
within the immediate context that is experientially shared by the communicator
and receiver at the moment of communication. They do not attempt to bridge
the experiential gap between the two sides. They become functional only where
the gap itself is reduced, if only momentarily, by the actual fact of being together.
We, humans, share many of our presentational systems with other spe-
cies, especially the apes, but we have also extended the functional envelope of
9780190256623-Dor.indb 23 03/06/15 6:17 PM

presentational communication—in two ways. First, we developed mimetic and

demonstrational capacities, with eye contact, pointing, and gestural sophisti-
cation, which allow us to achieve communicative goals unimaginable for the
apes—precisely those capacities that contribute to the unique level of intersub-
jectivity that we have even without language (Donald 1991, Tomasello 2008,
Zlatev 2008).
Second, we have invented an entire series of communication systems that
work with a different strategy, still experiential but perfectly revolutionary—
the re-presentational strategy. In re-presentational systems—drawings, paint-
ings, maps, musical recordings, photographs, and movies—the communi-
cator’s experiential intent is recorded, frozen in time, delivered from the
here-and-now of the communicator’s experience into the here-and-now of the
receiver, where it is “melted back,” so to speak, brought back to life for the re-
ceiver to perceive. Much of the original complexity of the experiential intent
is lost in the process. No system of re-presentational communication is ever
capable of bringing the entire original experience into the present. But sys-
tems of re-presentation are nevertheless capable of restoring different types of
experiences to different degrees. They follow the holistic and analogue logic of
presentational communication, transferring (dragging, if you will) the experi-
ential intent of the communicator, across time and space, into the experiential
world of the receiver. This is why the products of re-presentational commu-
nication acquire their significance from their similarity with their presenta-
tional counterparts, the corresponding directly presented experiences that
they are supposed to stand for. In semiotic terms, they are iconic: they provide
their receivers with echoes, or silhouettes, of what they would perceive had the
original experience been directly presented to them. Just like presentational
systems, re-presentational systems do not attempt to bridge the experiential
gap between the sender and the receiver: they use various technical means to
allow for the sharing of experience across time and space, without actually
bridging the gap.
2.4 Linguistic Communication: The Instruction of Imagination

The functional specificity of language lies in the fact that it works with a radi-
cally different strategy. I will call it the strategy of instructive communication. In
instructive communication, the very attempt to share an experience, whether di-
rectly or through mediation, is abandoned in principle: the communicator does
not try to make some of his or her experience perceptibly present to the receiver.
Rather than doing that, the communicator provides the receiver with a code, a
plan, a skeletal list of the basic coordinates of the experience—which the receiver
9780190256623-Dor.indb 24 03/06/15 6:17 PM

is then expected to use as a scaffold for experiential imagination. Following the

code, the receiver raises past experiences from memory, and then reconstructs
and recombines them to produce a novel, imagined experience.
As I will suggest below, the technology of language consists of two compo-
nents, the symbolic landscape, a collectively coded model of the world, and the
communication protocol—a set of normative rules for the regulation of instructive
communication. The two components allow speakers to channel coded instruc-
tions for imagination, through the symbolic landscape, into their interlocutors’
minds. The instructions, then, do not stand for the original experience. They
instruct the receiver in the process of the mental creation, in imagination, of a
totally independent experience—an experience that is supposed to reflect the
communicator’s experience not because it is perceptually based on it, but because
it is (approximately) of the same type. In experiential communication (presenta-
tional and re-presentational), the sender communicates: “this is my experience.”
In instructive communication, the sender communicates: “my experience is of
this type—try to imagine.”
The essence of this strategy, and its implementation, lies in the fact that
it requires a huge amount of collective effort to make it work, prior to actual
communication—an effort of experiential mutual identification for language.
In instructive communication, the listener is not invited to share an experience
with the speaker, but to create an independent experience, on the basis of the
skeletal formulation of the received code, within his or her own experiential
world—in isolation from the experiential world of the speaker. In the creative
activity of imagination, the listener may in principle imagine in a wide variety
of ways, all of which would always follow the analogue complexities of his or
her own experiential world, never that of the speaker. The code should thus be
able to instruct the listener in a process in which he or she has to create not only
a more or less focused image (an object of imagination, not necessarily a visual
representation)—but also a focused image that more or less corresponds to the
original experience of the speaker: an image of the same type. This is a very am-
bitious goal. The strategy of instructive communication does this through the
coordinated investment of enormous social energies in the never-ending process
of careful mapping and marking of those points in experience, and those ways
of speaking, which the different speakers within the community may, more or
less reliably, count on in the process: “when I use this word, imagine a thing of
this type (not that)”; “when I use this word together with this one, imagine this
type of experiential relationship (not that)”; “when I arrange the words in my
sentence this way, imagine you look at the whole experience from this type of
perspective (not that)”; and so on. It is precisely in this sense that language is a
technology: each of its components has to be built before it can be used.
9780190256623-Dor.indb 25 03/06/15 6:17 PM

2.5 The Definition of Imagination

As the recently growing literature on imagination readily acknowledges, the gen-
eral concept of imagination escapes all attempts at a comprehensive definition.
It is involved in almost every human mental state and mental activity, and it
takes on a different shape and form in each of them. Stevenson (2003), for ex-
ample, formulates twelve conceptions of imagination, from “the ability to think
of something not presently perceived, but spatio‐temporally real,” through “the
ability to think of things that one conceives of as fictional” and “the non‐rational
operations of the mind, that is, those explicable in terms of causes rather than
reasons,” to “the ability to sensuously appreciate works of art or objects of natural
beauty without classifying them under concepts or thinking of them as useful”
(p. 238). All these are obviously important aspects of our complex entanglement
with imagination. For David Hume, and then for Emanuel Kant, it provided a
major key to the understanding of all the products of our mental processes—
knowledge and belief, artistic judgment and creativity, moral judgments and
fantasy. For the Romanticists, it was the defining capacity of the creative genius.
In the philosophical discourse of the last century, it was variously discussed in
relation to (and sometimes in contrast to) perception, belief, desire, pretense,
memory, and visual imagery.
There is, however, a certain common thread that runs through many of the
conceptualizations of imagination—a distinction between basic imagination on
the one hand and creative imagination on the other. As Mithen (2007) argues, we
may share the capacity for basic imagination with other animals (although we may
never be sure): “when a chimpanzee manufactures a stick for extracting termites,
it may be engaged in foresight—imagining the process of poking the stick into the
hole, extracting it, and then licking off the termites—and it may also be contem-
plating alternative choices, with regard to either what size of stick to use or whether
to eat termites or fruit” (p. 4). What other animals seem to lack is creative imagina-
tion: “the ability to combine different types of knowledge and ways of thinking to
creative novel ideas and insights” (p. 4). In Scruton’s (2009) words, creative imag-
ination “always involves the summoning or creating of mental contents which are
not otherwise given (as they are given, e.g., in perception and judgment)” (p. 348).
As is always the case with such distinctions, there is no exact demarcation line be-
tween basic and creative imagination. They lie on a continuum: the prototypical
tokens of the two, on the two poles of the continuum, are clearly distinguished by
the relative freedom of the imagined content from the totality of experience that
is already there, in the imagining mind, at the moment of imagining (including
everything that is being experienced at the time of imagining). The chimpanzee’s
imagination is (supposedly) locked within the confines of its experiential world;
9780190256623-Dor.indb 26 03/06/15 6:17 PM

the poetic genius is (supposedly) capable of breaking away into a world totally un-
conditioned by experience. As Addis et al. (2008) show, in their analysis of age-
related changes in the capacity to imagine future events, the issue is one of flex-
ibility. They define the cognitive capacity that allows for imagination as “a system
that can flexibly recombine details from past events into novel scenarios” (p. 33).
The chimpanzee and the poet do the same thing, but the poet is radically more
flexible. What they do share is the foundational capacity, developed to very differ-
ent extents, to re-combine experiences into new experiences that are not themselves
directly given by experience. This will be my working definition of imagination.
As opposed to all the other systems of communication, then, the technology
of language allows speakers to communicate directly with their interlocutors’
imaginations, to instruct them through processes of experiential re-combination
that result in imagined experiences—experiences that are not given to them by
their own experience. As I will suggest later on, this is how we value our lin-
guistic exchanges. Other things being equal, the more your words force me to
imagine—the more communicatively valuable they would be for me (provided
that I manage to meet the challenge). Listening to sports on the radio, we need as
much play-by-play description as we can get (radio sportscasters are often virtu-
oso speakers); on television, where we visually experience the game by ourselves,
it would be a horrible nuisance. This is so not just for assertives, but for all speech
acts: I can use language to order you to do something you have already experi-
enced doing, but the communicative value of language really shows itself when I
order you to do something you have never done before, when you have to imagine
what you need to do before you can do it.
The capacities involved in language use and in imagination are thus intimately
related. The emergence and further evolution of the technology was entangled,
from the very beginning and throughout, with the evolution of the uniquely
flexible human imaginative capacity (Ginsburg and Jablonka 2014, Jablonka,
Dor, and Ginsburg 2012, Dor and Jablonka 2014). Every advancement in the
technology required more imagination; every new imagined experience required
further technological development. The more dependent human society became
on the technology of language, the more it came to acquire imagined properties
in and of itself. There is a straight line here from the very origins of language all
the way to the imagined communities of modern nation-states (Anderson 1983).
2.6 Only Language Bridges the Gap (and Always Partially)

This, then, is the centerpiece of the theory to be developed here: the strategy
of the instruction of imagination is the key to the understanding of language.
9780190256623-Dor.indb 27 03/06/15 6:17 PM

It is the key because it achieves something that no other strategy is capable of

(with the possible exception of the dancing rituals of the bees): it allows for com-
munication that actually bridges the experiential gaps between interlocutors. To
see what this means, let us take a step back and look at experiential commu-
nication again. There, everything that is communicated must be presented (or
re-presented) for perception within the experiential here-and-now of the com-
munication event. The communicative act is one of showing (in the general sense
of exposing for the senses to grasp, not necessarily in the visual sense). This im-
plies that whatever cannot be shown cannot be communicated. The functional
envelope of experiential communication is inherently limited. It is also flexible,
of course, stretchable and extendable—the human species has taken the art of
showing to unprecedented places, with ritual, theatre, music, and dance (and re-
cently the movie)—but the inherent limit is always there: something can only be
communicated if it can be shown.
There is, then, at every given moment, for every human individual, an entire
set of un-sharable experiences, experiences that lie outside the functional enve-
lope of experiential communication, beyond the gap: things that the individual
may want to communicate—but is incapable of showing. Generally speaking,
such experiences are of two types. The first is simply everything that the in-
dividual has ever experienced in the past (and by implication, everything that
was experienced in other places) that the individual cannot show. Other ani-
mals simply cannot communicate anything about their past experiences. In his
comparison of language with animal communication, Hockett (1960) refers to
this as the capacity for displacement: “Man is apparently almost unique in being
able to talk about things that are remote in space or time (or both) from where
the talking goes on. This feature—‘displacement’—seems to be definitely lack-
ing in the vocal signaling of man’s closest relatives, though it does occur in bee-
dancing.” Bertrand Russell puts it succinctly in his famous saying—“no matter
how eloquently a dog may bark, he cannot tell you that his parents were honest
though poor.” We, humans, have developed the re-presentational tools that
allow us to communicate better and better echoes of our past experiences, but
there is always much that is left incommunicable—incommunicable not because
of technical limitations, but in principle. A hundred years ago it was unimagina-
ble that I could go to a party with a video camera, and show you the entire event,
in high quality, the following day. As far as re-presentational communication is
concerned, we really live in revolutionary times. But even today, even with the
video, I cannot show you that this woman in red, right there by the door, is the
cousin of my ex-wife. For that, I still need words. The point, then, is not displace-
ment as such, as Hockett would have it (also see Bickerton 2009 and Mufwene
2013). The issue is displacement of that which cannot be shown.
9780190256623-Dor.indb 28 03/06/15 6:17 PM

All this, however, is only half of the story. There is also a very wide variety
of inner experiences that are very difficult to show—and are thus experientially
incommunicable. To be sure, experiential communication is exquisitely suita-
ble for the communication of emotions, especially social emotions—in humans
and other animals. Beyond the emotional level, however, there is much that re-
mains un-showable. The fact that I am worried, for example, may show itself
on my face, but nothing in my perceptible behavior can tell you why—unless
what I am worried about is present for perception in the here-and-now of the
communication event. If I am worried about something that happens elsewhere,
or something that is about to happen in the future, the object of my concern
will remain a secret as long as we do not use language. All of us live through a
never-ending experiential dynamic in which we perceive, organize, and evaluate
our perceptions, compare what we perceive with everything we have experienced
before, try to understand what the new bits of knowledge mean, think about the
implications of what we understand, try to figure out how to react, what to do,
and then plan the action, and so on and so forth—these are the inner workings
of our minds, and they remain safely out of the reach of experiential communi-
cation (again, unless they are directly concerned with salient components of the
perceptible here-and-now of the communication event). The essence of language,
then, lies in the fact that it allows for the communication of exactly these types
of experiences. The instructive strategy makes them communicable.
What this implies, in terms of functional specificity, is a foundational divi-
sion of labor between language, on the one hand, and the entire set of experiential
systems on the other. Language works best where experiential communication is
hopeless, and is quite useless where experiential communication is most effec-
tive and efficient. We have already mentioned emotional communication, and
the way it is best served by experiential means. The essence of emotion lies in its
holistic and analogue complexity, and this essence is communicated with exquis-
ite proficiency by the look on our face, the tone and color of our voice, and the
way we move our body. Language, however, in order to instruct the imagination
of the listener, dissects the analogue continuum of emotional experience, giving
names to certain points on the continuum that have come to be mutually identi-
fied by the members of the community: happiness, sadness, anger, jealousy, trust
(Aitchison, 2001). These obviously help, but as Deacon (1997) says, “a rich and
complex language is still no substitute for a shocked expression, a muffled laugh,
or a flood of silent tears, when it comes to communicating some of the more im-
portant messages of human social relationships.”
The same is true of another experiential domain, that of practical instruc-
tion (Dor 2000). Consider the challenge of teaching someone how to tie a knot
with a rope. There are at least three ways to do this. You may take a piece of
9780190256623-Dor.indb 29 03/06/15 6:17 PM

rope and directly demonstrate (a form of presentational communication); or you

may make a drawing that visually depicts the process (a form of re-presentational
communication); or you may try to communicate the process through language,
and tell the person you are trying to teach what to do. (Note that some other
systems of communication will not do at all: you cannot teach someone how
to tie a knot through music or facial expression.) It is easy to see, then, that the
three strategies—physical demonstration, drawing, and telling—are not equally
suitable for the task. Physical demonstration is the most appropriate: the task
lies at the very center of its expressive envelope. What has to be communicated
is a certain practice, a certain ordered set of manual moves. To learn how to tie a
knot is to become experienced in a manual practice, and the best way to do that
is by directly experiencing someone else doing it in front of you, then trying to
do it yourself, then looking at the other person again, and so on, until you know.
The experience (knowing how to tie a knot) is communicated directly through
experience (watching someone else doing it and trying to do it yourself).
The same task becomes more cumbersome when we try to achieve it through
the mediation of drawing—a means of re-presentational communication. The
practice is still communicated visually, but the experience is no longer directly
communicated as a unitary experience: it is broken down into a set of consecu-
tive stages, each of which is re-presented as a static posture, frozen in time, of the
hands vis-à-vis the rope. This is what drawing is all about: it is a system capable
of re-presenting visual moments, abstracted away from the temporal flow of ex-
perience, frozen into static form. Moreover, there is no experience of depth in
the drawing: the three-dimensional experience of the manual practice is reduced
into a two-dimensional re-presentation. We can no longer look at the process
from different angles, first from here, then from there. And the physical experi-
ence of touch is gone. The drawing does not communicate the feel of the rope,
its elasticity, its material reality. The entire dialogical dynamics of the physical
demonstration is lost: we cannot ask the drawing to come closer, do it again,
more slowly this time, and see if we are doing it right. The strategy of drawing
makes it less suitable for the task at hand. The task becomes more difficult, but
it is still possible.
Now, let us turn to language. Consider the following set of instructions (from
a do-it-yourself manual quoted in Aitchison 1996) for tying a figure-of-eight:
(a) Pass the end of the rope over the standing part of the rope.
(b) Take the end under the standing part away from the loop.
(c) Bring the end of the rope back up over itself toward the loop.
(d) Pass the end down through the loop.
(e) Pull tight.
9780190256623-Dor.indb 30 03/06/15 6:17 PM

The figure-of-eight is probably the most basic knot, but the most impressive
thing about this set of instructions is their sheer complexity. They are extremely
difficult to follow, and they were probably quite difficult to formulate. This is
so, because the translation of the experience of the manual practice into a mes-
sage that would allow the listeners to imagine what they have to do requires the
dissection of the practice into a series of stereotyped events that are already mu-
tually identified by speakers and listeners. The act of tying the knot, which is a
single, continuous move in the world of experience, has to be broken down into
a set of consecutive steps that already have a name: passing and taking, bring-
ing and pulling. These are types of moves that the symbolic landscape already
acknowledges. In experience, moreover, the rope is made to move around itself,
but in language, exactly because of the instructive strategy, the thing that moves
has to be designated differently than the thing it moves about. The rope, then,
which is clearly a single entity in the world of experience, has to be broken down
into separate parts: the end of the rope, the standing part, the loop. These are,
again, imaginary entities that are there, in the set of instructions, because they
can be named. To learn how to tie the knot by means of language, we have to be
able to identify these parts by their name, understand where they are supposed
to be on the rope that we experience in our hands. And, of course, we have to try
and figure out what it would mean, in the world of experience, to bring some-
thing back up over itself toward something else. An analogue trajectory of motion
which is trivial to demonstrate, and very easy to draw, becomes a nightmare
when we try to approximate it by combining four types of basic spatial relations
which language already acknowledges. All this is extremely complex, but as long
as we are dealing with the most basic knot, it is still possible. With more com-
plex knots, language simply breaks down. In more complicated practical tasks—
teaching someone how to drive a car, play the violin, practice t’ai chi—language
is virtually useless. It sometimes plays an auxiliary role, but it is hopeless on its
own. And consider all the types of experiential intents, emotional and other-
wise, which are communicated extremely efficiently through music, dance, and
pantomime. And all those situations where, quite literally, a picture is worth a
thousand words. Language, then, is not a general-purpose system of communica-
tion. It is specifically designed for the communication, through the instruction
of imagination, of experiences that cannot be experientially communicated.
This division of labor between language and experiential communication is
no accident. It is actually the most important fact about human communication,
the strongest reflection, as I will propose toward the end of the book, of the evo-
lutionary process that we went through as a communicating species: experiential
communication (in the presentational version) was there before. We share most
of our presentational tools with the apes. For a very long time, we developed
9780190256623-Dor.indb 31 03/06/15 6:17 PM

these tools further and further (most importantly with pointing and mimesis),
pushing the envelope of presentational communication step by step, and turning
more and more of the components of the here-and-now of the communication
event into communicable materials. The advances in presentational communi-
cation allowed human societies to make unprecedented leaps in material cul-
ture with the control of fire, the invention of cooking, and the sophistication of
tools, and to revolutionize social life with new divisions of labor, an entirely new
level of co-operative behavior, alloparenting, and collective inventiveness. Then,
we found ourselves in an unprecedented situation: on the one hand, we reached
the limit of presentational communication; on the other hand, material culture
and social structure continued to develop in complexity, and maintaining them
demanded deeper and deeper levels of information exchange—of exactly the
types that were very hard to share experientially. Language, a new technology
dedicated to the instruction of imagination, was the invention that solved the
problem.
It did not all happen overnight. Language began to develop the moment it
was born, and it still does. But when the Rubicon was crossed, when interlocu-
tors began to systematically construct bridges (always fragile and tentative) over
the experiential gaps between them, a new era began in human life. The instruc-
tion of imagination changed everything. Individuals were still living in the here-
and-now of private experiencing (we still do), but additional elements began to
penetrate their worlds from the outside, things they did not experience but were
told about. If there ever was a Copernican moment in the history of humankind,
this was it: almost all animals experience; the apes know how to follow the ex-
periences of the others in order to learn about the world; pre-linguistic humans
learned how to direct the experiences of others, and let others direct theirs—but
with language, humans could finally begin to experience for the others, and let
the others experience for them. Individuals began to learn to imagine events that
happened to others, and learn how to take them into account in their own deci-
sions. Memories from the past gradually turned into new objects of communi-
cation. Active remembering-for-speaking—keeping an experience fresh in your
mind until you tell the others—became a necessary capacity. Humans gradually
learned all the different things they could do with the new technology: delve
deeper and deeper into the others’ minds, ask questions and receive answers,
compare judgments about the world and argue about them, discuss the details of
future plans, tell others what to do, tell others the truth or lie to them. In their
proto-arguments, they gradually came to realize the extent to which they looked
at the world in different ways. In their proto-dialogues, they made their first steps
toward mutually identified descriptions of the world. We are still very much in
the process.
9780190256623-Dor.indb 32 03/06/15 6:17 PM

There is a complex story here, and we will get back to it in chapter 10. At
the moment, the important thing is this: we began our quest with the question
of the functional specificity of language, and we have now identified it as the
unique communicative strategy of the instruction of imagination. Let us, then,
look at the way language is built to implement this strategy.
9780190256623-Dor.indb 33 03/06/15 6:17 PM

3 How the Technology Works
How does the technology of language bridge the experiential gaps

between its speakers? How does it allow them to instruct each oth-
er’s imaginations? In this chapter, I will take the first steps toward a
technical model of the technology. My objective here is strictly defi-
nitional, to put the parts of the technology on the table, call them by
their names, connect them together, and see how they are constructed
and how they work.
The chapter opens with a general claim about the essence of
the social construction of language: the entire technology is col-
lectively constructed through an iterative social process that I will
call experiential mutual-identification for language. The basic idea is
quite simple: as speakers whose experiential worlds are different, we
have to work together to create a model of the world that we can
tentatively agree on and thus use as a channel for instructive com-
munication. We also have to agree on sets of norms for the use of
the technology, to make sure that the listener interprets the instruc-
tions in a way that is similar enough to that intended by the speaker.
I will define the process, discuss some of its properties, and begin
to show how it determines the particular architecture of language.
From the mutual-identification of meaning, emerges the symbolic
landscape, the mutually identified model of the world that allows for
the instruction of imagination. The landscape comprises of all the
signifieds of the language’s signs, which are connected in three dif-
ferent ways to other entities: each signified (i) points at a mutually
identified cluster of experiences; (ii) is connected to other signifieds
on the landscape by a web of semantic relations; and (iii) is marked by
a signifier. From the mutual-identification of communicative behav-
ior, emerges the communication protocol, a set of mutually identified,
formal norms for the production and interpretation of instructions
for imagination. I will then present an informal characterization of
linguistic communication as a process of conversion, where speakers
translate their experiential intents into formal instructions, which
their listeners use to imagine the intended meaning. Finally, I will
9780190256623-Dor.indb 34 03/06/15 6:17 PM

How the Technology Works • 35
define a general notion of instructive success, and show how the theory, as op-
posed to all other general theories of language, accepts the fragility of language
as one of its constitutive properties. Taken together, these first definitional
moves, it is to be hoped, will provide us with a clear enough picture of the tech-
nology and some of its constitutive properties. In the following chapters, the
picture will get more technical, and begin to find its theoretical contextualiza-
tions and empirical justifications.
3.1 Language is Constructed by Mutual-Identification

To see what is involved in the effort of mutual-identification for language, we have
to take a step back, and look at the effort of experiential mutual-identification as
such—and its most crucial component, the activity of pointing.
The discovery of the crucial significance of pointing in human life is one of
the most important achievements of the cognitive sciences in the last two de-
cades. Today, four interrelated facts seem to have been established almost beyond
doubt. First, apes in the wild, who use a rich array of gestures in their presenta-
tional communication, very rarely point (Liszkowski et al. 2006). Second, apes
in captivity learn to point, but only in order to ask humans to hand them pieces
of food that lie out of their reach. In one case, reported in Call and Tomasello
(1994), apes have been documented to use pointing in order to direct the at-
tention of a human to a hidden tool, that he had to use in order to retrieve the
food they were interested in. Third, human infants begin to intentionally point
and understand pointing around their first birthday, and in doing so they reflect
their innate propensity and motivation for co-operation and intersubjectivity
(Tomasello et al. 2007). Fourth, pointing plays the most central role in language
acquisition: “children’s initial skills with a conventional language emerge on the
heels of their initial skills with pointing—typically by only a couple of months”
(Tomasello et al. 2007, p. 718). What is it about pointing, then, that positions it
so clearly at the border between us (with our language) and the apes? Tomasello
and his colleagues explain things this way:
The first thing to note is that, by itself, pointing is nothing. If you and
I are walking down the street talking about the weather, and I stop and
point for you in the direction of a bicycle leaning against a tree, without
any other context, you will be totally mystified as to what I could possibly
be intending to communicate. The reason why you will be mystified in
this situation is that you do not know either what I am directing your
attention to (what I am referring to) or why I am directing you to it (what
is my motive). With regard to the what question, Wittgenstein (1953)
9780190256623-Dor.indb 35 03/06/15 6:17 PM

demonstrated decades ago that pointing always underdetermines the

intended referent without some “form of life” or shared context within
which the pointing occurs. Am I pointing at the whole bicycle? Or the
special kind of polyvinyl seat? Or the color? Or the metal material it is
made of? The possibilities are limitless, and demonstrate, perhaps sur-
prisingly, that the pointing gesture can actually indicate radically dif-
ferent perspectives on one and the same perceptual situation. Pointing
simply directs someone’s attention to a location in the perceptual envi-
ronment, but to correctly identify the intended referent requires that the
communicator and the recipient know together that the indicated loca-
tion is in some way relevant to some larger context they share (see Sperber
& Wilson, 1986, on the key role of relevance assumptions). We will call
this larger context, following Clark (1996), common ground or, some-
times (when we wish to emphasize the shared perceptual context), the
joint attentional frame . . . What we mean by “know together” is simply
that two individuals both know that they both know. (Tomasello et al.
2007, p. 706).
Pointing, then, only becomes meaningful when based on an already established,

intersubjective common ground. This is also what makes language possible. The
apes lack the capacities required, and the human species evolved to have them.
Infants thus develop the capacities, using them for pointing and then for lan-
guage acquisition.
This understanding of pointing, I would like to suggest, puts the cart before
the horse. Pointing does much more than direct the attention of the other to a
component of the already established common ground. It constructs common
ground between individuals whose experiential worlds are different. It serves us
when common ground is required, not when it is already there. Let us take an-
other look at Tomasello’s bicycle example. In real life, people do not walk around
pointing at random bicycles for the sake of a thought-experiment. They do so if
the bicycle is huge, or extremely beautiful, or totally destroyed, or decorated in a
strange way, and so on. They point because they understand that their interlocu-
tor has not seen what they have, that there is no common ground. The pointing
indicates the acknowledgment of the experiential gap, and the will to close it.
And, indeed, in cases of this type, where A points and the thing pointed at is
visually salient, B looks, and then sees and acknowledges (with a “wow,” or some-
thing of the sort), and then the two of them look each other in the eyes and smile.
Then, and only then, can they start talking about the bicycle. It is true that in
some cases, the thing pointed at would not be very salient, and materials that are
already part of the common ground between the interlocutors would participate
9780190256623-Dor.indb 36 03/06/15 6:17 PM

in the determination of visual attraction. This, however, is exactly the point. For
these materials to already belong in the common ground, A and B should have
already gone through an earlier round of pointing and acknowledgment. If we
walk together in Cochin, India, for example, you may not understand my excited
pointing if you do not know that I am looking for the Hebrew scriptures on the
walls that lead to the old synagogue. In order to close the gap, we have to look at
the guidebook, point at the scriptures, and agree: this is what we look for. If we
do not have a book, I can tell you what I am after, but then I have to find a way
to show you what Hebrew looks like, and so on and so forth: it’s pointing all the
way down.
Pointing by itself, out of context, is not nothing. It actually captures the entire
story. It is one of the most important achievements of the human species, because
it allow us to systematically mutually identify elements of our experiences—not
just our experiences of the outside world, but also our experiences of our own
behavior, our relationships, and so on—and thus work together to reduce the
experiential gap between us. Indeed, apes in the wild are capable of following
the gaze of others: they have a theory of mind (Call and Tomasello 2008). This,
however, is still an individualistic capacity: the one whose gaze is followed does
not intentionally participate in the event. It is only with pointing, and only in
humans, that the circle is closed. The fight against the solitude of private experi-
encing becomes a collective, communicative, dialogical effort.
It is important, then, to appreciate the complexity involved in every event of
mutual-identification. It begins with A’s detection of a new and important ele-
ment in the environment; and then, immediately, A’s detection of the fact that B
has not yet detected the element; A’s will to make the other see; A’s pointing; B’s
detection of the pointing and the turning of B’s gaze at the pointed direction; B’s
detection of the thing pointed at; B’s acknowledgment; and then a quick meet-
ing of the eyes for the sealing of the mutuality of the experiential identification.
We humans are extremely good at this; we have evolved for that, both cognitively
and morphologically. As Kobayashi and Kohshima (2001) show, for example,
the properties of the human eye—such as the unique, large, and exposed white
sclera—are adaptations for enhanced signaling by eye gaze. Pointing is exactly
what allows us, as privately experiencing individuals whose experiential worlds
are different, to construct intersubjective common ground within the here-and-
now of being together. This is why the discovery that apes in captivity learn to
point, but only for humans and only at food (Call and Tomasello 1994), is so
fascinating. It shows that the difference between us and them is not a simple
matter of the individual capacity for learning: when apes are introduced to the
technique of mutual-identification by the humans around them, the apes can
then learn how to do it. And yet, they only use the technique to get food. They
9780190256623-Dor.indb 37 03/06/15 6:17 PM

do not show any interest in the dialogical activity of mutual-identification as

such. As every parent knows, human infants consider it the most exciting thing
in the world.
Crucially, the complexity of mutual-identification implies a certain funda-
mental fragility. The danger of failure is always very close: B may not notice the
pointing; look in the wrong direction; look in the right direction and fail to
see what A pointed at; notice something else where A pointed, or more than
one thing (this is the essence of Wittgenstein’s observation quoted above, and
of Quine’s (1960) Gavagai thought experiment); or notice the exact same thing
that A pointed at and still understand it differently. This last option will play a
central role in the entire discussion below. A and B may actually leave the event
of mutual-identification with a clear feeling that mutual-identification was
achieved, whereas in effect each of them thinks they mutually identified some-
thing else. In spite of their efforts, the experiential gap is always still there.
To see how mutual-identification allows for the construction of language,
then, we need to add another layer of complexity to the story. Assume that when-
ever they mutually identify a feature of their experiences, A and B continue imme-
diately to mutually identify a sound pattern, or a manual gesture, that will from
now on mark in their minds the bit of experience they have just mutually identi-
fied. If they manage to do this, they would in effect construct a linguistic sign, a
discrete instructor of imagination (or, in a different scenario, a linguistic norm, a
discrete convention for instructive communication). If they continue to do this
for a very long time, working harder and harder to raise the levels of instructive
success between them, they will inevitably go through a very particular process,
that will eventually create a new language—with its symbolic landscape and pro-
tocol. I will call this the process of experiential mutual-identification for language.
To be sure, the process does not rely throughout on the actual physical ac-
tivity of pointing. Tomasello et al. (1996), for example, show that children are
also capable of acquiring word meanings in non-ostensive contexts, by watching
adults engage with the world and talking about it. From an early stage onward,
language itself participates in the process not just as the outcome, but also as a
means: we mutually identify for language by speaking about it. This is the essence
of dialogue. As Clark and Brennan (1991) show, a major part of any conversation
is dedicated to the grounding of a shared lexicon between the participants. This,
among other things, is how the communal lexicons of specialized communities—
professions, sports fans, and neighbors—emerge and stabilize (Clark 1998).
What is common to all these is not the finger-pointing itself, but the essence:
the systematic engagement of the speakers in the effort of mutually identifying
experiential commonalities and their formal signification—for the sake of in-
structive communication. It is a process that never ends.1
9780190256623-Dor.indb 38 03/06/15 6:17 PM

3.2 The Landscape: The Mutual-Identification of Meaning

To get a sense of the particular nature of the process of mutual-identification for
language, let us try the following thought-experiment. Assume a community of
four human members, A, B, C, and D, and assume that they have absolutely no
language in common: how they found themselves together we need not know,
but now they have to communicate on a regular basis. Let us, then, allow them
to begin to construct a new language from scratch.
Suppose that A begins by pointing at a chair and uttering tʃɛər. The point-
ing means: “let this experience be common.” It acknowledges the experiential
gap, and calls upon the others to gather around, concentrate their attention on
that particular object, and make a mental note of the fact that the experience of
looking at the object, of identifying the object as a focal point of attention, was
something that was, at that moment, shared by all the others. In uttering tʃɛər,
A suggests: “in the future, to get the others to imagine an experience of this type,
let us say tʃɛər.” Assume, then, that B, C, and D do indeed look in the right di-
rection, and manage to meet Gavagai’s challenge: they associate the sound pat-
tern with the chair itself, not one of its legs, or the upholstery, or the color. Now
they have agreement on what I call the sign’s experiential-anchor, the specific
chair A was pointing at, which is now marked by all four as an experience shared
with the others.
This, however, is only the beginning. The crucial question now is how the
four of them would generalize the meaning of their first sign: what, for each of
them, would be other experiences of the same type, experiences for which the sig-
nifier tʃɛər would be adequately used? (Remember that they cannot talk about
it.) The foundational fact of individual experiential variability forces us to the
conclusion that they would zoom in on different ideas. For the sake of simplic-
ity, let us try the following scenario: what A had in mind in the first place was
the chair as something used for (the physical experience of) sitting, so for A,
everything used for sitting could from now on be called tʃɛər. B, however, has
never seen a chair before. Having no real idea of what the object is (what it is
supposed to be), B would register tʃɛər as designating things that look like this.
For B, the sign would be adequately used to designate objects whose visual ex-
periences are similar enough to the original experience-anchor. C, on the other
hand, would be mostly impressed by the fact that the object pointed at was a reg-
ular, four-legged dinner-table chair, not an armchair, an office chair, a recliner,
or a beanbag. C, evidently experienced in the world of furniture, would decide
that tʃɛər designates dinner-table chairs. Finally, assume that D, exhausted after
many hours of walking, would happily see the designated object as A saw it:
something to sit on.
9780190256623-Dor.indb 39 03/06/15 6:17 PM

At this very early stage, then, our four individuals already have everything
they need in order to find themselves systematically misunderstanding each other
whenever any of them attempts to communicate about office chairs, armchairs,
cushions and beanbags, recliners and stools—and everything which might, from
this angle or the other, look like the original chair. Assume, for example, that
D sees a stool somewhere, and then goes back to tell the others, with a combi-
nation of hand waving and the signifier tʃɛər, that there is a chair—something
used for sitting—for them to pick up. They go along with him and see the stool.
For A, all this would be natural. B and C, however, would probably be confused.
Both would think (not for the same reasons): “what is this? This is not a chair.” C
might think: “had I known that this was a stool, I wouldn’t have taken the trou-
ble to come here in the first place.”
The first sign of the new language, then, already raises the question that will
follow our four individuals all the way to their full-fledged language: when does
something in the experience of any of the members of the community count as
different enough vis-à-vis what has already been established, to deserve its own
place in the collective model of the world? The question is that of the threshold
of demarcation, and it is a social question through and through. There is nothing
objectively necessary in the distinction between a chair and a stool, nothing that
exceeds the necessity of dedicating different signs to chairs with four and three
legs, expensive chairs and cheap chairs, wooden and metal chairs, comfortable
and uncomfortable chairs, new and old chairs, and so on—ad infinitum. Indeed,
the great majority of experiential distinctions which any of our four individu-
als would notice in his or her lifetime between types of chairs (“this one looks
funny”; “that one is heavier than the one we carried yesterday”) would probably
never find its way into their gradually emerging language. For the distinction be-
tween the chair and the stool to enter their language, one of the four individuals
would have to make a point, to insist, and the others would have to agree. Assume,
then, that C points at the stool and says stuːl. Doing this, C indicates that the
experiential extension of the sign chair should somehow be limited, delineated:
stools should be excluded. They should be mutually identified as such, and re-
ceive their own sign. So, says C, “I insist that this new experience which we are
now sharing is different from the one we have named tʃɛər. In the future, let us
refer to experiences of this type by using the signifier stuːl.” It is important to see
that C does not have to do all this—he or she may very well decide not to—and
that the others do not have to agree: as we shall see later, the social dynamics that
develop within the community determine much of what eventually emerges. If C
does intervene, and the others do agree, they will now have two signs.
This, then, is the crucial point: following their second round of negotiation,
our four individuals would not have just isolated and mutually identified two
9780190256623-Dor.indb 40 03/06/15 6:17 PM

types of experiences from everything they have experienced—they would have

also isolated and mutually identified one distinction from the huge set of distinc-
tions that come and go in their experiential lives: stools and chairs are different.
The two signs will no longer just be associated in their minds with their experi-
ences. They will be connected to each other, in a way that reflects the short history
of their linguistic negotiation. This will be the first semantic connection on the
semantic landscape of the emerging language. A first step toward the construc-
tion of an entire model of the world of experience, whose signifieds and their
semantic relations reflect not the actual properties of the world, nor the expe-
riential world of this individual or the other, but the entire history of mutual-
identification in the community. This is why it is so important to see that this
first distinction, between stools and chairs, would not be as natural for the others
as it would be for C, who initiated the second round. Each of the three would
have to ask, again: in what way are chairs and stools different? D would think:
“we used the sign tʃɛər for everything that can be used for sitting; this thing that
we just named stuːl is probably used for something else.” B would think: “this
one I know: we used to sit on these when we were children; that thing that we
called tʃɛər I still don’t understand.” A would think: “I don’t see why we need
an extra sign, but if C insists. . . .” New rounds of mutual-identification would
follow, with new signs and new semantic connections, in new contexts and with
new initiators, new bridges over the experiential gaps and new failed attempts
to cross it. If and when our four individuals ever reach a relatively stable under-
standing of the difference between chairs and stools, the two signs will already
be held together, in a fine-grained semantic configuration, by everything in their
neighborhood on the semantic landscape.
The web of semantic connections on the symbolic landscape emerges as a nec-
essary consequence of the fact that each and every round of experiential mutual-
identification for language involves not just the pointing and acknowledgment of
a new experiential cluster—but also the re-configuration of everything that has
already been established that is related in this way or the other to the new sign.
Every round of mutual-identification (besides the very first) says two things, not
just one: (i) this type of experience is now mutually identified and marked by this
signifier; (ii) this type of experience is now demarcated from all the experiences
we have already mutually identified, and is related to them in such and such
ways.2 When our four community members will begin to go beyond the mutual-
identification of simple physical entities, and get into the adventure of mutually
identifying more complex experiences—such as parts of things, properties, and
eventualities, their symbolic landscape will become populated with semantic
connections at a much higher rate. If different animal species, for example, are to
be distinguished on the symbolic landscape, the signs referring to them as species
9780190256623-Dor.indb 41 03/06/15 6:17 PM

must abstract away from everything else that distinguishes between them in ex-
perience, including, for example, their sex. If male and female animals need to
be distinguished on the symbolic landscape, the signs referring to them as male
or female must abstract away from everything else, including species member-
ship. Once these two demarcations are established, however, they are forced into
a standardized relationship that stereotypically reconnects what was originally
disconnected and isolated.
All this goes much deeper when we consider signs referring to eventualities,
such as “break,” “dance,” “eat,” “know,” “tall,” and so on—what we usually think
of as verbal and adjectival meanings. To receive their own signs, eventualities do
not just have to be abstracted away from their experiential complexities—they
have to be abstracted away from the very entities which embody them. Yeats fa-
mously asks the question in Among School Children: “How can we know the
dancer from the dance?” There is nothing in the event of dancing that is not,
at the very same time, a fact about the dancer. Dancing only happens as long as
the dancer dances, and the dancer is only a dancer as long as the dancing event
takes place. Watching a dance (visually experiencing it) is looking at the dancer.
To assign the eventuality itself a sign, the semantic landscape has to disembody
the dance, and leave the identity of the dancer unspecified. The sign referring to
the dancer, however, must be semantically connected to the sign referring to the
eventuality, because a dancer is to be identified as different from other people
on the sole basis of the fact that he or she participates in a dancing event. What
this means is that eventualities come to be associated on the symbolic landscape
with their stereotypical participants. This is what argument roles (Fillmore 1968,
Fillmore 1982, Jackendoff 1972, and so on) are all about: as I will suggest in
chapter 7, they are generalizations constructed by mutual-identification. And
eventualities are connected to other eventualities by stereotypical relations such
as cause, reason, and goal; they take place in time, and so on. All these categories,
and others, reflect the emergent consensus among the community members with
respect to their mutually identified world: “in the model of world through which
we exchange instructions for imagination, things are either male or female, sin-
gular or plural, animate or inanimate; they participate in events of certain well-
defined types, which take place in time; they have causes and goals, and so on.”
This understanding of semantics, as an emergent, collectively coded realm
of meaning, functionally built by mutual-identification for the instruction of
imagination, marks a radical departure from the way semantics is conceptualized
in the literature. As we shall see, it determines much of what happens in actual
linguistic communication, and it helps resolve a host of traditional difficulties
in the study of semantics. It also allows semantics to explain much more about
language than is generally assumed: its explanatory repercussions reach as far as
9780190256623-Dor.indb 42 03/06/15 6:17 PM

the formal complexities of syntax, the dialectics of linguistic relativity, the politi-
cal power of language, and much more. At the moment, however, let me high-
light the single most important fact about the mutual-identification of meaning:
the symbolic landscape only correlates very partially with the pictures of the
world that emerge in the private experiences of each of the individual commu-
nity members. This is so for a wide variety of reasons. To begin with, our experi-
ential worlds are analogue: the categorical distinctions that emerge in our minds
throughout a lifetime of experiencing are continuous, fuzzy, dynamic, transient,
heavily detailed, constantly variable, and context-dependent. The world of the
symbolic landscape, on the other hand, is discrete: it is a set of isolated values and
their distinct relations. This is a necessary result of the very essence of mutual-
identification. On the landscape, for example, the property fragile is semantically
connected with a host of fragile objects, with the event break, with other proper-
ties such as frail and delicate, with the antonyms sturdy and strong, and so on. In
the analogue world of experience, however, things are never that simple. Objects
are never either fragile or sturdy, either frail or strong. Eventualities in which
things break or remain intact do not arrange themselves neatly on the two sides
of a clear line of demarcation. To really understand the susceptibility of physical
objects to manipulations such as breaking, fracturing, folding, and so on, we
have to experience working with them. To teach this, we have no choice but to
use presentational communication. But if we wish to make others imagine an
object with certain properties, the analogue richness of our experiences will not
do. The much simpler model of the symbolic landscape provides us with discrete
points of reference, fixed by mutual-identification, with which we can instruct
their imaginations.
The fact that most of us are not as lucky as our community of four, that we
have to make use of languages handed down to us by past generations, deepens
the discrepancies between our experiential worlds and the symbolic landscape of
our languages. In our experience, for example, not all things have gender. Chairs
and stools, as we experience them, are not male or female. In the landscapes of
many languages, however, they are. It may well be the case that the demarca-
tion of inanimate objects into male and female reflected very deep experiences
of those people, who in the distant past made the collective decision to assign
gender to all the objects, regardless of animacy, but for us this is no longer the
case. Nevertheless, if we speak any of these languages, and if we wish to talk
about inanimate objects, we are obliged to adopt the logic of the symbolic land-
scape, and treat them, for the purposes of communication, as if they were male
or female. As Slobin (1996) puts it, we have to accept the categories for the pur-
pose of thinking for speaking. This is one way in which the historical depth of
the collective project reflects itself on the symbolic landscape. The full-fledged
9780190256623-Dor.indb 43 03/06/15 6:17 PM

4 4 • t he ins t r uc t ion of imaginat ion
landscapes that we use in our languages are by their very nature conservative,
which means that they are also quite often obsolete: this is the major reason why
we consistently work together to upgrade them.
Finally, and most importantly, the logic of mutual-identification implies that
the symbolic landscape reflects the accumulated compromise between the experi-
ential worldviews of its different speakers: the landscape only includes the primi-
tives, categorizations, and relations that came to be pointed at, negotiated, and
collectively accepted. This understanding positions different speakers of the lan-
guage in different positions vis-à-vis the landscape, in terms of the experiential
clusters they associate with the signifieds; the semantic areas they feel experien-
tially comfortable with and those they only partially connect to their experiential
worldviews; and the areas of their experiential worlds that for them are not sup-
ported by the landscape, and are thus destined to remain mute, unless they manage
to produce innovative usages of language and get them mutually identified.
Every speaker of language, every one of us, walks around with two mis-
matched worldviews—private-experiential and social-linguistic—and the mis-
matches themselves are variable. Not only is there an external experiential gap
between us as individuals; there is also an internal gap within ourselves, between
ourselves as experiential creatures and ourselves as language users. This founda-
tional fact will follow us for the rest of the book: it explains many of the particu-
lar properties of word meanings, much of the structural complexity of language,
and the inherent fragility of the process of linguistic communication. It is where
the drama of linguistic relativity takes place: the landscape and experience are
caught in a never-ending cycle of bi-directional influence, both at the collec-
tive and individual level. The fact that different individuals live in variable mis-
matches explains much of the variability in individual communicative behavior,
capacity, motivation, and success. And it explains the connection between lan-
guage and social power: communicative success rises as the space between the
two worldviews shrinks, and the process of mutual-identification is thus also a
struggle for symbolic control within the community.
3.3 The Linguistic Sign as a Discrete Instructor

The linguistic sign functions as a discrete instructor of imagination by maintain-
ing a triadic relationship between the signified and the signifier, the neighbor-
hood, and the experiential cluster (see Figure 3.1).
At the heart of the sign lies the signified, an isolated point on the sym-
bolic landscape—the trace of the foundational act of mutual-identification. It
is collectively marked by the signifier, the mutually identified form of the sign;
it is held in place by the web of semantic connections to other signifieds in its
9780190256623-Dor.indb 44 03/06/15 6:17 PM

Experience The symbolic Linguistic

landscape sound
Analogue cluster Digital signified Signifier

around semantically
experiential-anchor connected to other
signifieds
figure 3.1
neighborhood; and it points at the experiential cluster, a fuzzy set of private ex-
periences, clustered around the experiential-anchor, the actual experience that
was mutually identified by the interlocutors. The experiential cluster itself lies
outside the sign, in the realm of private experience. It is pointed at by language.
The sign “chair,” for example, constructs in the minds of English speakers an
association between (i) the signifier tʃɛər; (ii) the signified chair, semantically
interconnected on the symbolic landscape with the signifieds stool, armchair, up-
holstery, legs, table, furniture, comfortable, sit, and many others; and (iii) a private
cluster of chair-related experiences—chairs of different sizes and shapes, their
look, smell and feel, various sitting experiences and so on—centered around the
experience-anchor of some very familiar type of chair, which, at some point in
the past, found its way into the mutual-identification event (or consecutive set of
events) in which they learned what the sign meant.3
It is easy to see that this triadic structure, like that of a three-part rocket, is
specifically designed for the instruction of imagination. The intent to talk about
a chair emerges in the speaker’s mind at the private-experiential level: the object
of the will to communicate is something in the cluster of chair-related experi-
ences; the experience activates the signified that is attached to the cluster, thus
translating the inaccessible, private intent into a term that is already mutually
identified; the signified activates the mutually identified signifier; which acti-
vates actual pronunciation. The listener hears and analyzes the stream of speech,
and isolates the signifier; the signifier activates the signified; which then raises
9780190256623-Dor.indb 45 03/06/15 6:17 PM

the experiential cluster in the listener’s imagination – from the private chair-
experience of the speaker, through the collective mutual-identification of “chair-
hood,” to the chair-imagination of the listener.
And it is also easy to see the fragility of the sign as a discrete instructor:
for different speakers, the signified chair points at different clusters of private
experiences. This is also the case with the other signifieds in the neighborhood,
which means that the webs of relations around the signified chair may also be
different to this extent or the other. Mutual-identification is always a matter of
degree, and the potential for miscommunication always remains. The listener
may always imagine something different from what the speaker intended. This
is why we spend so much of our speaking time trying to decide what we mean
by the words we use (Clark 1998). As a matter of fact, the single component
of the sign that a community of speakers may mutually identify with almost
perfect confidence is the signifier: what it sounds or looks like. This is a fact of
enormous importance: it helps explain the amazing, inborn capacity that we
have for the detection of even the slightest deviations from the norms of accent
(Mehler et al. 1988).
3.4 The Protocol: Mutually Identifying Communicative Norms

The communication protocol does for the speakers exactly what the symbolic
landscape does for them with respect to their experiences of the world—only
this time with respect to their experiences of communication. Different speak-
ers experience events of communication—instances of communicative success
and communicative failure, confidence and confusion, understanding and mis-
interpretation, trust and suspicion, sincerity and deception—in a wide variety
of ways. They occupy different positions in these communication events; they
look at them from their own particular perspectives (as speakers, as listeners).
They come into them from different histories and different social positions, with
different habits and sensitivities, different capacities (for expression, understand-
ing, trust); different attention spans, sets of beliefs, expectations, and interests.
They share a moment of communication, but this moment is nevertheless an at-
tempt to construct a very narrow, very shaky bridge over the constant experien-
tial gap. This is a very precarious endeavor: it can only work to the extent that the
participants in the collective effort manage to coordinate their efforts. In order
to be able to do that, they have to be able to mutually identify, at every given
moment, what the other person is trying to do. The communication protocol,
then, emerges through the entire history of a language’s development, from the
constant social negotiation of normative ways of speaking that the members of
the community mutually identified as co-operative usage of language. When the
9780190256623-Dor.indb 46 03/06/15 6:17 PM

members of a community mutually identify a norm, what they say to each other,
across the experiential gap, is: “here we do the same thing.” (And again, just like
before, they never really do.)
Consider speech-acts, for example. Our experiential communicative intents
very rarely, if ever, emerge in our minds as digitally demarcated intents to either
say or ask something, to order, or promise, or predict, or deny, and so on. They are
multi-layered, variable, vague, dynamic, analogue. We wish to express something,
and what we wish to express is as complex as the experience within which the
communicative intent emerged. Coupled with the foundational fact of the ex-
periential gap, this analogue complexity constitutes a major obstacle to commu-
nication: we very often find it difficult to understand what the person speaking
to us is trying to do (“is this a promise or a threat”), and our experiential histories
often lead us to the wrong conclusion. Speech-acts, then, are socially negotiated,
stereotypical communicative behaviors, highlighted and isolated from the ex-
periential continuum of communication, which, when practiced according to
a set of mutually identified conventions, allow for the successful mediation of
the speaker’s intention across the experiential gap. When conventionalizing a
speech-act, what the members of the community agree on is this: “from now
on, when we behave this way—when, in these particular contexts, we use this
intonation, this word order, this gesture—we mean to ask a question (or make a
promise, or tell a story).”
We have learned from Austin (1962) that speech-acts are ways of doing things
with words. But where do they come from? What is the source of the distinctions
between different speech acts—between assertions and questions, promises and
requests? Austin himself, and later John Searle (1969), maintained that speech-
acts are natural phenomena, distinguishable from each other (and dependent
upon each other) on the basis of their logical properties as forms of communi-
cation. The formal properties of speech acts (as analyzed, for example, by Searle
and Vanderveken 1985), are facts about the very essence of language as a logical
system of communication: when a speech-act is performed under certain, well-
formedness conditions ( felicity conditions), the speech act is successful. When
the felicity conditions are not met, the speech-act fails.
This conception was famously attacked by Derrida (1977), who has claimed
that meanings are always polysemic, contexts are never totally determined, in-
tentions are never transparent—and speech-acts are hence always different,
opaque, indeterminate.4 The option of failure is not something that lies outside
the domain of language, something that can be systematically avoided by follow-
ing a certain formal logic. It is always there, within language, as “the very force
and law of its emergence” (p. 17), because the reality of language is always an
“endless alternation of essence and accident” (p. 16), never the stable, idealized
9780190256623-Dor.indb 47 03/06/15 6:17 PM

system envisaged by Austin and Searle. The essence of speech-acts lies precisely
in the fact that they cannot be exhaustively defined.
I think that Derrida is right, but so are Austin and Searle. Speech-acts do
have specific formal properties, but they have them not because they reflect
an inherent logic of communication, but because they are mutually identified
norms for communication—conventionalized norms that are necessary because
(not in spite) of the fact that the experiential world within which linguistic com-
munication has to take place looks just like Derrida describes it. What Derrida
is talking about is not language itself, but the experiential gap that is “the very
force and law of (the) emergence” of language, and he is perfectly right: the ex-
periential gap is never really bridged. Austin and Searle, on the other hand, are
talking about language and its conventional norms, and because of that—they
are not idealizing. They are describing the idealized conception that lies at the
very basis of language as a conventional system—the idea that perfect commu-
nication across the experiential gap would be possible if only all members of the
community managed to perfectly mutually identify everything in the worlds
and their communications.
The fact that the conventions of the protocol emerge as compromises be-
tween speakers whose experiences of communication are different highlights
their normative essence. They do not just emerge, as Lewis (1969) would have
it, as solutions in co-ordination games, and they are not just expectations that
we have concerning the future linguistic behavior of the others and of ourselves.
They are statements of what speakers should do in their linguistic behavior, and
they behave exactly like norms in other social domains (Bicchieri 2006, Gil-
bert 1989). They impose collective demands on individual speakers to behave
in ways that very often contrast with their own communicative inclinations.
This is why speakers have judgments to share with linguists: when a speaker
follows all the norms of the language to the letter, the end product (the actual
fragment of speech) is judged by the other members of the community as gram-
matical (or otherwise well-formed). When making a grammaticality judgment,
the members of a linguistic community are not actually interested in the frag-
ment of speech itself. They ask themselves whether the speaker has followed the
linguistic norms established by their community. They do not ask: “is this sen-
tence grammatical?” They ask: “does the speaker obey our rule?” This is why, as
the entire sociolinguistic literature shows, grammaticality judgments are always
identity judgments. Eliciting grammaticality judgments from native speakers is
always a political act, and grammaticality judgments are always methodologically
complex, variable, and vague. Speakers are defensive about grammaticality judg-
ments: as far as they are concerned, it is the very normative foundation of their
community that is being tested.
9780190256623-Dor.indb 48 03/06/15 6:17 PM

(We linguists, for our part, have become so used to the distinction between
descriptivism and prescriptivism that we no longer see that the object of linguis-
tics as a descriptive science is itself a system of prescription. We often say to our
native speakers: “forget what they taught you at school about the way the lan-
guage should be spoken; just tell us how this sentence sounds to you.” We think
we are going beyond conventional norms, touching something much deeper. Ac-
tually, we are telling our native speakers: “forget about the norms you have only
heard about; pay attention only to those you have already internalized.”)
The protocol, then, includes everything that pragmatics, sociolinguistics,
anthropological linguistics, discourse analysis, and conversational analysis have
taught us about the actual process of linguistic communication. At its core, as
I will suggest in chapter 6, the communication protocol includes a practical
guideline for the actual production and comprehension of utterances—a set of
ordered procedures which lead the speaker, step by step, through the process
of experience-to-speech translation, and allows the listener to follow the same
route backward, from speech to experiential interpretation. The set of proce-
dures is related to the actual processes of production and comprehension in the
exact same way laws are related to the dynamics of social life: it specifies the
normative ways in which the process should take place. To the extent that it is
internalized, and to the extent that it is enforced, it actually regulates and con-
strains the process in real time. The set of procedures does not, however, do more
than regulate and constrain. The activity of communication is not brought to
life by the conventions which regulate it. It originates from, and is motivated by,
the very will to communicate. The protocol governs linguistic communication
in the way laws govern social life: not all laws are equally internalized; not all
laws are equally enforced; not everybody obeys the law all the time. This is the
reason for another similarity between linguistic communication and its conven-
tions and social life and the law: communication is always regulated by conven-
tions, but it is also, at the very same time, a constant attempt to break away from
them (Pinchevski 2005). This understanding should help us come to terms with
a host of phenomena that reside at the psychological boundaries of linguistic
communication—the refusal to communicate, the sense of betrayal of self that
accompanies the linguistic socialization of a private intent, the notion of the
secret, and Freudian slips.
3.5 Instructive Communication as a Process of Conversion

Let us, then, use all this to take a first look at instructive communication as
a process of conversion. In conversion processes, inputs of one type are trans-
formed into outputs of another. The telephone, for example, converts sound
9780190256623-Dor.indb 49 03/06/15 6:17 PM

waves into electrical signals in the microphone, and the signals are converted
back into sound waves in the earphone. The important point for us is that the
telephone performs its function with a single conversion on each side of the com-
munication event: sound to electricity on the speaker’s side, electricity to sound
on the listener’s.
In this sense, experiential communication of the presentational type is sim-
ilar to the telephone. It performs a single conversion on each side: from private
experience to perceptible behavior, and then from perceived behavior to private
experience. Systems of re-presentational communication also require a single
conversion, but of a very different type. Consider painting, for example: it con-
verts one type of visual percept, the view as it was privately experienced by the
painter, into another type of visual percept, the painting, that is then presented
for perception to the viewers. The essence of re-presentation lies exactly in this
fact, that it remains in the domain of perception, that it converts perceived ma-
terial into perceived material.
What about language, then? Well, according to the accepted view, the
uniqueness of language lies in the fact that it performs two conversions instead of
one. The communicative intent is not converted directly into perceptible behav-
ior. There is another level in between, a level of formal structure. When we pro-
duce an utterance, we begin by converting our intended meaning into a formal
configuration of words, morphemes, and other types of constructions. It is this
configuration, not the original intent, that we then convert into perceptible be-
havior when we actually speak. The same happens on the listener’s side. Perceived
behavior (the stream of speech) is converted into a formal structure: the words
and their formal relations are extracted from the stream of speech, isolated, rec-
ognized, and registered. Then, the formal configuration is converted again, to
produce a meaningful interpretation.
This picture, I would like to suggest, requires a serious revision. The secret
of language, the ingenious trick that allows for instructive communication, is
the fact that language performs three conversions, not just two. And the heart
of the trick is the fact that the first conversion is not from meaning to form. It is
from meaning to meaning: from communicative intent to instructions for im-
agination, from experiential meaning to the symbolic landscape—from private
meaning to collective meaning.
The production process begins, as in all the other types of intentional com-
munication, with an experiential intent in the mind of the speaker: something
that is, for that particular speaker, at that particular time, the mental object of
the intention to communicate as defined above. The intent is experiential and
private, and is thus, crucially, prior to language. It is the original material that the
technology receives from the outside, the material it has to process.5 In the first
9780190256623-Dor.indb 50 03/06/15 6:17 PM

conversion, then, the normative rules of the communication protocol guide the
speaker, step by step, in the conversion of the experiential intent into an ordered
set of instructions for imagination. I will use the term message to refer to this
ordered set. In the first conversion, then, the intent is converted into a message.
The building blocks of the message are the inherent meanings (the signifieds) of
the words and constructions, chosen by the speaker on the basis of the intent.
The first conversion, then, transfers meaning from the private realm to the realm
of the social. By doing this, it makes it communicable: it turns something that is
privately experienced into instructions that the others can understand.
It is the message, then, the ordered set of instructions for imagination, that
provides the input for the next conversion, which is indeed the conversion to
formal structure. The message is converted into a formal configuration consist-
ing of the perceptible structures of the words and constructions (their signifiers),
and the phonological, morphological, and syntactic relationships between them
(we shall see later how these relationships are re-defined within the theory). I will
call this configuration the utterance. Finally, in the third stage, the utterance is
converted into perceptible behavior—actual speech. Sound waves (or visible mo-
tions) are produced for the interlocutor (listener, viewer) to experience.6
The process is governed throughout by the normative rules of the protocol,
which also govern the process as it takes place, in the opposite direction, on
the side of the interlocutor. As in all other types of intentional communication,
the comprehension process begins with the perception of behavior: we listen to
the stream of speech, or view a stream of visible motion, and take them in. The
perceived material is then converted into a mental representation of the utter-
ance. The stream is phonetically, phonologically, morphologically, and syntacti-
cally analyzed, and the signifiers and their structural relations are identified.
The utterance is then converted into the message, the ordered set of instructions
for imagination constructed by the speaker. Finally, the message activates the
interlocutor’s imagination, instructing it to retrieve from memory certain types
of experiences, and arrange them together in a particular way, in order to create
a private imaginary experience. Not the speaker’s experience, but an experience
of the same type. This imaginary experience, I will call the interpretation. The
entire process, then, is depicted in the flowchart in Figure 3.2.
We will get back to this process, in more technical detail, in chapter 6. What
is important at the moment is the fact that the flowchart redefines the relation-
ship between the cognitive and the social. What happens inside the speaker’s
mind is the socialization of the private intent. What takes place inside the listen-
er’s mind is the privatization of the social message. Language mediates between
private experience and the social world. This is what it does as a social technol-
ogy. This is how it bridges the gap.
9780190256623-Dor.indb 51 03/06/15 6:17 PM

Speakers’s
9780190256623-Dor.indb 52
Speakers’s Speakers’s
intent message utterance
The Process of
(experiential, (semantic, (structural, Linguistic
private) social) social) Communication
intent
translated
into message message
translated
(socialized,
into
simplified, sound
interpretation
stereotyped, waves
changed)
(complexified,
analogized,
contextualized,
imagined)
listener’s listener’s listener’s
social utterance message interpretation
(structural, (semantic, (experiential,

mental social) social) private)
physical
figure 3.2
03/06/15 6:17 PM
3.6 The Fragility of Instructive Communication

It goes without saying that not all instances of linguistic communication end
successfully. Some result in complete failure, and some achieve their goals only
partially. The understanding of language as an imagination-instructing com-
munication technology requires a new definitional perspective on the success-
failure continuum and its determinant parameters. The fact that we are dealing
with a technology forces us to distinguish between two questions: (a) instructive
success: Has the technology performed its task successfully? and (b) perlocution-
ary success: Has the speaker been successful in achieving his or her communica-
tive goal? We have already touched on this distinction in our discussion of the
letter, email, and fax, in chapter 1. Whether or not you would actually come to
pick me up from the train station (whether or not, in Austin and Searle’s terms,
my perlocutionary act results in success) is not the same question as whether or
not your fax machine received my note in good form. By the same token, when
we talk, the question of whether or not I managed to persuade you, make you
answer my question, or convince you to commit to doing something for me is
not the same question as whether or not we used the technology of language suc-
cessfully in our interaction. Let us, then, begin with the following definition of
instructive success:
An instance of imagination-instruction is successful to the extent that

the listener constructs an imagined experience that is similar to the
speaker’s intent.
There’s quite a lot that needs to be said about this definition. First, it measures
the entire process as described in the last section, from the speaker’s intent to
the listener’s interpretation. As such, it reflects the foundational positioning of
language in the social domain, between interlocutors: success in the usage of
the technology of language is a collective matter. The speaker and listener make
different contributions to the effort, but the success of the interaction depends
on both (along with many other parameters). It is not enough for the speaker
to do everything right. If there is no one there to listen, no imagined experi-
ence on the other side of an experiential gap, the activation of the system on
the speaker’s side results in complete failure. (Talking to oneself, of course, is
a different matter: the same individual, speaking in two voices, works on both
sides of the technology). Second, the definition is formulated in relative, rather
than absolute, terms: an instance of linguistic communication is instructively
successful to the extent that. The more similar the imagined experience is to the
intent, the more successful is the instruction. Third, because the experiential gap
9780190256623-Dor.indb 53 03/06/15 6:17 PM

is always still there, linguistic communication at its very best is always only an
approximation.
The quality of instruction, then, depends on a wide range of intercon-
nected parameters. For the sake of presentation, I will discuss them as if they
could be comfortably split into four classes: physical, cognitive, linguistic, and
experiential.
To begin with, in line with the mathematical theories of communication
following the Shannon-Weaver model (Weaver and Shannon 1963), an event of
instruction is successful to the extent that the act of transmission, from speaker
to listener, is performed under suitable physical conditions. Other things being
equal, for example, the level of noise on the auditory (or visual) channel is in-
versely correlated with instructive success.
Second, instructive success is correlated, positively this time, with the total
sum of language-related capacities the two sides bring into the event. Other
things being equal, difficulties in the production of speech sounds on the speak-
er’s side, or hearing problems on the listener’s side, reduce the levels of instruc-
tive success. The same is true of the relevant levels of cognition (and emotion)
recruited for the task of language, from phonetic deciphering, through lexical
retrieval and grammatical organization, to the efficiency and flexibility of ex-
periential re-combination in imagination, memory capacity, general reasoning,
general communicative capacities, and so on. In terms of the software-hardware
distinction, all this may be thought of as relating to the quality of the cognitive
hardware in both minds involved in the exchange.
All this is quite straightforward. Things become more subtle in the next two
classes: linguistic and experiential. The linguistic parameters have to do with the
proficiencies of the interlocutors with the copies of the linguistic software they
use in the exchange. Proficiency is the overall ability of the speaker to speak a
certain language. It is not the same as the capacity to speak in general: being an
expert speaker is not the same as being an expert speaker of English. Proficiency
includes competence as one of its components: the speakers’ knowledge of the
normative rules and regulations of the protocol and the symbolic landscape. In
terms of communicative success, then, there are two issues here:
(i) the quality of the two copies: the levels of proficiency the speaker and lis-
tener have with the particular language variation they are using. Other
things being equal, higher levels of proficiency on both sides would provide
for higher levels of instructive success.
(ii) the similarity between the two copies: the extent to which the two partici-
pants are competent speakers of the same language variation. Other things
being equal, the extent to which the speaker and listener have the same
9780190256623-Dor.indb 54 03/06/15 6:17 PM

normative expectations from each other, both in terms of meaning and

communicative behavior, will determine the level of their communicative
success.
Last, and most important, two experiential parameters play crucial roles in de-
termining the quality of instruction:
(i) Instructive success is inversely correlated with the width of the overall gap
between speaker and listener. Other things being equal, the closer the expe-
riential worlds of the interlocutors are to each other, the more successfully
they will communicate.
(ii) Instructive success is inversely correlated with the width of the internal gap,
within each speaker, between the realm of experience and the realm of lan-
guage. Other things being equal, higher levels of compatibility between the
interlocutors’ experiential worlds and the language variations that they use
will provide for higher levels of instructive success.
To be sure, because we are interested in success or failure of particular communi-

cative instances, the linguistic and experiential parameters should be relativized
to the contents and context of each and every instance. In a conversation about
Beethoven and his influence on the Beatles, communicative success depends
much more on the interlocutors’ music-related experiential worlds, and their
command of the relevant jargon, than it does on the similarities between their
experiences with skiing, or their command of its specialized jargon.
Let me state the same idea in a slightly different way. Suppose we start with
a certain experiential intent, in the mind of A, and ask: how easy would it be for
A and B to bring about a situation where B imagines an experiential interpreta-
tion that is, for their communicative purposes, similar enough to A’s intent? How
quickly would A say: “I think you have understood what I was trying to say.”
Each of the above parameters contributes in its own way to the answer: lower
levels of noise, higher levels of language-related capacities, proficiency with the
language variations and compatibility between the variations, narrower experi-
ential gaps between the interlocutors and within them—all these would make
the task easier. Higher levels of noise, lower levels of capacity, proficiency and
compatibility, and wider gaps would require more effort. From a certain point on,
the task would become impossible.
There is no way, then, around this fundamental fact: an enormous lot has to
work right in order for interlocutors to succeed in the challenge of instruction,
and very little is required for them to fail. This is nothing but a realistic appraisal
of the essential fragility of the technology of language—one of the darker sides
9780190256623-Dor.indb 55 03/06/15 6:17 PM

of the ingenious invention. We spend much of our speaking time hesitating,

looking for words, trying out formulations, and then realizing this is not what
we wanted to say. We try to clarify and ask for clarifications, ask for and provide
missing background information, repair mistakes by ourselves and elicit repairs
from others. We constantly interrupt each other, negotiate the meanings of our
words, assess the width of the gap between us, get entangled in webs of mis-
understanding, and look for creative ways to finally get across. As the growing
literature in conversational analysis on the phenomenon of repair is beginning
to show, conversations (like other social interactions) are not just “characterized
by the routine occurrence of troubles, ‘hitches,’ misunderstandings, ‘errors,’ and
other infelicities” (Hayashi et al. 2013, p. 1). They actually show systematic pat-
terns of co-operation in repair, governed by a normative infrastructure for the
organization of the conversation. The organized set of practices of repair is used
by interlocutors to try to make sure that “troubles may be addressed at or near
the point where they occur, and may be potentially resolved more or less imme-
diately before there is intolerable divergence of the participants’ intersubjective
understandings of what is going on in the interaction” (Enfield and Sidnel 2014).
This level of organization is in fact so fundamental for language that some of
its components seem to be universal to an extent that the search of universals
in the structures of languages could only dream of. As Enfield et al. (2013), for
example, show in their large-scale comparative investigation of repair strategies,
all the languages of the sample include very close variations on the same primary
interjection, used in all the languages for the signaling of misunderstanding—a
monosyllable with an open non-back vowel [a, æ, ə, ʌ], often nasalized, usually
with rising intonation and sometimes an [h-] onset: huh?
That conversations are infested with hurdles is indisputable. What this means
is a totally different matter. The hurdles have indeed attracted the attention of
psychologists and psycholinguists, anthropologists and sociolinguists, discourse
analysts and conversation analysts, but the simple fact is that no general linguistic
theory (with the possible exception of Sapir) has ever taken them to constitute an
essential fact about language itself. General linguistics has always been interested
in language at its very best. The presupposition has always been that the nature of
language shows itself most clearly when everything works right. Language as such
is always ready to produce optimal results. The objects of inquiry in general lin-
guistics are the fully grammatical sentence, the totally understood utterance, the
perfectly combined constituents, the most adequate interpretation. Everything
that goes wrong is delegated to psycholinguistics, to be explained as a matter
of cognitive processing—an altogether different thing. To be sure, processing
mistakes have been used in the literature as linguistic evidence (e.g., Fromkin
1973), but the logic of inquiry here still follows the above presupposition: we
9780190256623-Dor.indb 56 03/06/15 6:17 PM

can learn about language as an optimal system from instances where other causal
elements prevent the speaker from achieving perfect execution. In this sense,
general linguistics has never really detached itself from the prescriptive tradi-
tion within which it was born: the prescriptive discourse has simply camouflaged
itself in the colors of “legitimate scientific idealizations.”
The theory of language as a technology for the instruction of imagination,
then, is the first to accept the fragility of language into its theoretical core. In
this, it owes much to Reddy (1979). It is one of the most beautiful pieces ever
published in linguistics: the unforgettable opening sentence says “I should like
to respond to Professor Schön’s chapter by replaying his theme several octaves
lower.” Professor Schön is Donald Schön, a philosopher who spent most of his
career thinking about the cognitive and communicative processes that we go
through, as individuals and collectivities, when we try to solve a problem. He
showed that much of what happens to us is determined by the way we set the
problem, or frame it, for ourselves. Disagreements about the solution very often
emerge not from factual differences, but from different framings. What we need
to do, then, instead of working hard to categorize the problem itself, is to exam-
ine the categories with which we work. Real progress in the resolution of dis-
agreements can only be achieved by frame restructuring on both sides.
Reddy agrees, with enthusiasm, but then suggests that all this lies very close
to the surface. It only reflects the fact that the very same thing is happening at a
much deeper level (hence the octave lowering): what Schön talks about is actu-
ally a problem of communication. It underscores the fact that the communication
of thoughts between people is a hugely difficult enterprise. We speak to each
other all time, but most of our conversations end with failure: we do not under-
stand each other. So, in order to solve the problems that reality faces us with, we
have to turn our attention to the way we talk about them, to the problems that
we face when we communicate. But then, Reddy says, Schön’s principle appears
again. Instead of trying to approach the communication problem directly, we
have to concentrate on the way we frame the problem for ourselves: the way we
talk about our failures in conversation.
Reddy thus set out to investigate the way speakers (of English) describe
“what’s wrong” with their conversations and “what needs fixing.” He collected
a corpus of hundreds of expressions in regular language, catalogued and classi-
fied them, and showed that they all reflect a single, overarching conception of
communication, that he dubbed the conduit metaphor. According to this concep-
tion, “language transfers thoughts and feelings” from the mind of one individual
to the mind of the other. The speaker “inserts” his or her meaning into words,
and combinations of words, just like material things are put into “containers”
or “conveyers.” The listener simply unpacks the containers and “extracts” from
9780190256623-Dor.indb 57 03/06/15 6:17 PM

them the thoughts and feelings of the speaker in their entirety. What this con-
ception implies is that language, the conduit itself, is a perfect tool of communi-
cation. When everything is done right, thoughts and feelings are actually carried
along and transferred from mind to mind. Communicative success requires no
explanation: it “appears to be automatic.” Failure, on the other hand, does re-
quire explanation, and in Reddy’s corpus it is always explained as a personal fail-
ure of one of the sides, especially of the speaker. Here are some of his examples:
(1) You still haven’t given me any idea of what you mean.
(2) Try to pack more thoughts into fewer words.
(3) Your words are hollow—you don’t mean them.
(4) That remark is completely impenetrable.
(5) Your thoughts here don’t quite make it across.
(6) The thought is there, although I grant that it’s sunk pretty deep in
paradoxical language.
These assertions, then, frame the communication problem according to the con-
duit metaphor, and implicitly or explicitly suggest the solution: the speaker has
to improve.
The only problem, says Reddy, is that the conduit metaphor is wrong: “Actu-
ally, no one receives anyone else’s thoughts directly in their minds when they are
using language. Mary’s feelings . . . can be perceived directly only by Mary; they
do not really ‘come through to us’ when she talks. Nor can anyone literally ‘give
you an idea’ since these are locked within the skull and life process of each of us”
(p. 287). If we use the wrong metaphor to frame our failures to communicate,
we may very well be looking for the solution in the wrong place. We need frame
restructuring.
Reddy thus moves on to suggest an alternative view of linguistic commu-
nication, that he calls, somewhat misleadingly, the toolmakers’ paradigm. He
constructs a thought-experiment, in which a group of toolmakers try to discuss
different methods for making tools such as rakes, rock-picks, and hoes. The tool-
makers are positioned vis-à-vis each other in a very particular way. They all live
in a wagon-wheel shaped compound, each in his own pie-shaped sector. The en-
vironments in the sectors are different: some sectors are more woody, for exam-
ple, others more rocky and bare. Crucially, the toolmakers have no access to the
others’ sectors. They can’t see or hear what is happening on the other side of the
spoke. They cannot even exchange samples of the tools they try to build. They live
in “radical subjectivity” (p. 292). At the hub of the wheel, however, there is “some
machinery which can deliver small sheets of paper from one environment to an-
other.” All the toolmakers can do is exchange “crude sets of instructions—odd
9780190256623-Dor.indb 58 03/06/15 6:17 PM

looking blueprints scratched on special sheets of paper that appear from a slot
in the hub and can be deposited in another slot—and nothing more” (p. 292).
This, argues Reddy, is a much more realistic model of what is actually happening
to us when we use language to communicate. What it implies, as opposed to the
conduit model, is that it is not failure that needs to be explained—but success.
“Partial communication, or divergence of readings . . . are not aberrations. They
are tendencies inherent in the system” (p. 295).
This is exactly right. Processing problems obviously make things worse, as
do many other components of reality, but the most important cause of the ubiq-
uity of failure is the essential nature of language itself. Language functions as
a bridge over the experiential gaps between speakers, constructed by mutual-
identification from the two sides of the gap. Because of this, the bridge is inher-
ently very shaky. Attempts to communicate fall off the bridge not just because of
the wind, and not just because of carelessness and so on, but because the bridge
itself breaks down very easily. Walking across the bridge of language is a pre-
carious endeavor even under the best conditions. The human species has been
working hard to overcome the bridge’s technical flaws, to make it more stable, for
probably a half a million years by now: the frustrating levels of communicative
success it allows for are still, quite simply, the best it has to offer.
9780190256623-Dor.indb 59 03/06/15 6:17 PM

4 Si g n a n d M e a n i n g
In the first three chapters, we have looked at enough of the defini-

tional foundations of the theory for us to be able to make our first
descent onto the plane of observation. This chapter will be devoted
to words and their meanings. I will discuss one central line of devel-
opment in the empirical study of lexical meaning—from the rise and
fall of the definitional-componential approach, through the emergence
of prototype theory, to the current investigation of polysemy. I will be
mostly interested in the minimal meaning-bearing entities that we
informally call words, but the principled account of their meanings
to be developed here applies just as well to all the constructions of lan-
guage whose meanings cannot be compositionally computed from the
meanings of their parts (Goldberg 2006, Tomasello 2003, Verhagen
2009) and are thus to be thought of as genuine linguistic signs—
mutually identified as such by the community of speakers.
The literature on lexical meaning is vast and the issues are nu-
merous and complex (see Geeraerts 2010 for a history of the disci-
pline). In the most generalized sense, the developmental dynamics of
the field over the last fifty years can be characterized in two dimen-
sions. First, research has gradually moved from the attempt to under-
stand the meanings of words in isolation to the attempt to explain
them within their contexts—both grammatical and communicative.
Second, the discipline has seen a gradual move away from a cogni-
tively based perspective on word meaning that we may think of as au-
tonomous, toward a cognitively based perspective that is general. The
autonomous perspective can be described as the convergence of three
ideas: one observation concerning the lexicon and two overarching
hypotheses concerning language as a whole. The first is quite intui-
tive (and actually much older): it maintains that there is, in fact must
be, a distinction between the meaning of a word as such, and all the
meanings that we associate with the word in our view of the world, a
distinction between word knowledge and world knowledge, between
the lexicon and the encyclopedia. The intuition here is clear: not eve-
rything that we think and feel about cats is included in the meaning
9780190256623-Dor.indb 60 03/06/15 6:17 PM

Sign and Meaning • 61
of the word cat. The second idea is highly theoretical, and fifty years ago it was
also brand new: it maintained that the foundational properties of language, as
an innate cognitive capacity, are qualitatively different from those of general
cognition. This was the Chomskian paradigm: language is a cognitive entity sui
generis, autonomous from general cognition and not reducible to it. The lexicon-
encyclopedia distinction thus seemed to fit the theory perfectly: the lexicon is
where autonomous linguistic meaning resides. The third idea had to do with the
truth-oriented perspective on the semantics-pragmatics distinction: the meaning
of a sentence, as opposed to the meaning intended by the speaker, was supposed
to equate the set of conditions in the world under which the sentence would be
true. The autonomous lexical meanings of the words, together with their gram-
matical relationships, were thus supposed to provide everything required for
truth-evaluation—which means, among other things, that they had to be con-
text-independent. The three ideas, taken together, painted a clear picture: lexical
meaning is semantic, autonomous from world knowledge, stable across contexts,
and truth-evaluative; general cognition provides additional context for the prag-
matic interpretation of the speaker’s intended meaning.
With time, more and more arguments emerged against the autonomous per-
spective, and most of them were targeted against the two overarching hypoth-
eses, attempting to show that knowledge of language is not autonomous from
general cognition, and that truth-evaluation is either irrelevant or performed at
the level of the encyclopedia. From this effort emerged different versions of what
Geeraerts (2010) calls a maximalist approach to word meaning: “a maximalist
approach to semantic description abandons the idea (which is strongly present in
structuralist and generativist theorizing) of achieving some form of autonomous
semantics, and goes for a type of meaning description that radically embraces
the idea that there are close and inseparable ties between ‘word knowledge’ and
‘world knowledge’” (p. 572). Most researchers today (Levinson 1997 calls them
lumpers) subscribe to the idea that the distinction between the lexicon and the
encyclopedia is theoretically unnecessary—that words as such do not have mean-
ings: they provide access to clusters of structured information in general cogni-
tion. Some others (Levinson’s splitters) insist that the distinction is necessary,
but also characterize it in very different ways. Importantly, the splitters agree
with the lumpers on the formulation of the question itself as a question about the
individual mind/brain: to what extent, and in which way, do the lexicon and the
encyclopedia constitute two separate levels of cognitive representation?
The programmatic account to be developed in the following pages, then,
begins by moving the entire question from the cognitive to the social domain.
To understand the intricacies of sign meanings, we have to position them be-
tween individuals, as discrete instructors of imagination, mutually identified
9780190256623-Dor.indb 61 03/06/15 6:17 PM

by speakers whose experiential worlds are different. The context within which
word meanings should be understood is social. This move, along with the
definitional characterization of the sign presented in chapter 3, section 3.3,
changes the picture in five constitutive ways. First, it re-defines the notions of
the lexicon and encyclopedia: the lexicon is the symbolic landscape, the stor-
age of everything that was collectively mutually identified; the encyclopedia
includes everything that emerges in private experience. What this means, and
this is the second point, is that the distinction between the lexicon and the
encyclopedia not only exists—it is a necessary functional property of language
as a social technology. There is no language without it. In this sense, perhaps
surprisingly, my theory sides fairly and squarely with the generative position,
re-framed in one crucial sense: the lexicon-encyclopedia distinction does re-
flect the autonomy of language from general cognition—but not because it is
a cognitive capacity of a different kind, but because it is not a cognitive capac-
ity at all. It reflects the autonomy of language, as a socially constructed tool,
from the private experiences of its speakers. Third, moving the question to the
social domain pluralizes what has always been discussed in the singular: we
are no longer talking about the relationships between a lexicon and an ency-
clopedia; there are as many encyclopedias as there are speakers, with gaps of
varying widths between them, and as many copies of the lexicon as there are
speakers, also variable to different degrees. Fourth, the lexicon does not simply
reflect conceptual structures in general experience. It is normative: it super-
imposes on private experience a collectively constructed order, an order that
emerges through iterative mutual-identification for the purposes of instructive
communication. The superimposition determines the way we speak, not the
way we experience, but as we shall see in the next chapter, it is also entangled
in a bi-directional spiral of influence with our private experiences: this is the
essence of linguistic relativity. Fifth, as detailed in chapter 3, the process of
iterative mutual-identification necessarily produces signs whose signifieds are
connected in two very different ways to other meanings: the signified points at
an experiential cluster around an experiential anchor, and is semantically con-
nected to other signifieds in its neighborhood on the symbolic landscape. The
lexicon is both arranged within itself and pointing at the encyclopedia. The
two meaning relations, moreover, correlate only partially. They emerge from
two types of processes. Speakers’ judgments and behaviors having to do with
word meanings are thus determined together by the two types of meaning rela-
tions. What this means, and this a pattern we will frequently meet throughout
this book, is that the arguments brought about by both sides to the debate
actually highlight different facets of a reality that is more complex than previ-
ously imagined, but also much easier to explain.
9780190256623-Dor.indb 62 03/06/15 6:17 PM

In the next section, 4.1, I will present a skeletal overview of the empirical
issues, from the attempt to provide words with componential definitions in terms
of necessary and sufficient conditions, through the discovery of prototype effects
in lexical meanings and the emergence of prototype theory, to the investigation
of polysemy and its interpretation in the context of communication. Along the
way, we will collect an intriguing list of clues to the essence of word meanings, as
well as a set of unresolved problems. In sections 4.2–4.4, I will aim to show how
the clues converge to support the conception of word meanings as instructions
of imagination, and how, within this conception, the problems may eventually
be resolved.
4.1 Componential Definitions, Prototypes, and Polysemies

The beginning of the modern quest for the essence of lexical meaning is usually
identified with the Katzian version of the classical theory (Katz and Fodor 1963,
Katz and Postal 1964, Katz 1972). The classical theory assumes a clear demar-
cation between encyclopedic and lexical meaning, and asserts that autonomous
lexical meaning is represented by componential definitions of the type that we
find in dictionaries. The definitions are formulated in terms of final sets of nec-
essary and sufficient conditions. All the components of the definition (formally
represented by Katz as semantic markers and distinguishers) are denotative con-
ditions: for an entity in the world to be worthy of the sign, it has to meet all of
them at once. The sign chair, for example, is accompanied by a definition such
as “a physical object, which has four legs and a back, and is used for sitting.” The
definition allows us to identify the referents of the sign—actual chairs in the
worlds—by the compositional identification of its semantic components.
Today, it is clear that this conception cannot be right—for at least three rea-
sons. First, it turns out that definitions are very rarely exhaustive: there are always
going to be non-chairs wrongly included in the definition, and genuine chairs
left out. The point is not that exhaustive definitions are hard to find; in the great
majority of cases they are impossible. The definition of the chair we looked at
before, for example, is obviously too tentative—it leaves out chairs with three
legs, without legs, without a back, and so on—but the fact of the matter is that
no amount of tinkering would make it exhaustive. Fodor (1981) demonstrates
this beautifully with the definition of the verb paint. He begins with “X covers
Y with paint,” but this doesn’t work: we couldn’t describe an event in which a
paint factory exploded and covered spectators with paint with the sentence: “the
factory painted the spectators.” X has to be an agent. But “X is an agent and X
covers Y with paint” is not good either: if X kicks over a bucket of paint which
9780190256623-Dor.indb 63 03/06/15 6:17 PM

then covers his shoes, that does not mean that X painted the shoes. We need X
to cover Y with paint intentionally, but this is not enough: Michelangelo inten-
tionally covered the ceiling of the Sistine Chapel with paint, but he did not paint
the ceiling. He painted a picture on the ceiling. So, how about “X is an agent,
and X intentionally covers Y with paint, and X’s primary intention in doing this
is to cover Y with paint”? This is already quite complex, but still lacking: every
time Michelangelo dipped his brush in paint, he did that with the primary in-
tention of covering it with paint, but he definitely didn’t paint the brush. The ex-
ercise seems to have no end. It should be noted that certain types of words, such
as terms for mathematical concepts and kin relations, do allow for exhaustive
definitions—we will get back to this later—but the consensus in the literature is
that this is the exception rather than the rule. Much of the research thus contin-
ues to look for necessary components of word meanings, but the other criterion,
that of sufficiency, has been abandoned.
The second problem with the definitional approach is this: definitions are
circular. In the above example, all we wanted was to define the verb paint, and
now we have to find exhaustive definitions for such signs as agent, intention,
and cover. These will require their own detailed definitions, which are going
to be at least as complex as the one for paint: the definition of intention will
probably prove the hardest. And even if we managed to do all this, we would
only find ourselves facing a new layer of required definitions—and so on, ad in-
finitum. One possible way out of this quandary is to try and find a small set
of basic signs that could be used to compositionally define all the others, and
treat them as the undefinable foundations of the entire lexicon. This is the basic
idea behind Wierzbicka’s natural semantic meta-language (NSM) perspective
(Wierzbicka 1996, 1997, Goddard and Wierzbicka 1994). Wierzbicka famously
gathered a list of around sixty semantic primes, presumably undefinable mean-
ings such I, YOU, SOMEONE, BODY, THIS, THE SAME, GOOD, BAD,
THINK, WANT, TRUE, BELOW, INSIDE, VERY, PART OF, and so on,
with which she claimed all word meanings in all languages can be paraphrased.
The primes themselves are innately given, and are also inter-translatable between
all languages: their basic meanings are universal. All the other words are defin-
able, learnable, and as culture-specific as needed.
The idea inspired a wave of cross-linguistic inquiry into word meanings, with
fine-grained methodologies and insightful results (e.g., the papers in Enfield and
Wierzbicka 2002), and I agree with Goddard (2002) that “if (this perspective)
is valid, it would have very important consequences for linguistics, and so it de-
serves to be taken seriously.” I fail to see, however, as do other commentators
(Barker 2003, Riemer 2006, Jackendoff 2007, and see the “bad arguments” in
Goddard 1998), how the perspective lives up to its explanatory promise: in what
9780190256623-Dor.indb 64 03/06/15 6:17 PM

sense are such primes as GOOD, BAD, and TRUE simpler than the others?
How can they possibly be universal? Have we not been debating them since
the beginning of time? Why should there be such a correlation between what
is simple and what is universal? How can the primes be innate? Do we not al-
ready know (Elman et al. 1996) that representational innateness of this type is
a biological impossibility? Most importantly, how does the construction of the
often-cumbersome definitional paraphrases à la Wierzbicka explain anything?1
It definitely does not provide the necessary level of exhaustiveness required by
the definitional approach it attempts to salvage. As Riemer (2006) notes, NSM
seems to focus not on the question “what is happening when I understand the
meaning of a word?,” but on “how can I explain the meanings of words (to
others)?”. Theoretically speaking, this is a non-starter.
The problem of circularity, then, was a major impetus behind the move of
the burden of explanation from the lexicon to the encyclopedia, from the word
itself to the piece of general cognition associated with it: what seem to be the
properties of our words actually reflect the properties of our experiential worlds,
conceptualized in different ways. This is the essence of the maximalist approach,
and it cuts through the Gordian knot of the circularity problem: definitions of
words made of other words are no longer required.
The third problem with the definitional approach opened the way for a new
perspective: word meanings very often manifest prototypicality effects (Rosch
1975, 1978, Mervis and Rosch 1981, Taylor 1995, Geeraerts 1997). In the clas-
sical theory, all the members of the set denoted by the word are of equal status.
Experimental evidence, however, shows that some members of the set are often
judged by speakers as more or less representative of the set. Thus, the robin is the
best example of the category “bird,” followed by the sparrow, the blue jay, and the
bluebird. Chickens, ostriches, and peacocks are considered the worst examples
(Rosch 1975). Chairs and tables are better examples of the category furniture
than mirror and shelves. Rather than conform to a final set of necessary and suf-
ficient conditions, the members of these sets show a general family resemblance.
The most celebrated example is of course Wittgenstein’s (1953) discussion of the
concept “game”:
Consider for example the proceedings that we call “games.” I mean

board-games, card-games, ball-games, Olympic games, and so on. What is
common to them all?—Don’t say: “There must be something common, or
they would not be called games”—but look and see whether there is any-
thing common to all.—For if you look at them you will not see something
that is common to all, but similarities, relationships, and a whole series of
them at that . . . I can think of no better expression to characterize these
9780190256623-Dor.indb 65 03/06/15 6:17 PM

similarities than “family resemblances”; for the various resemblances be-

tween members of a family: build, features, colour of eyes, gait, tempera-
ment, etc. etc. overlap and criss-cross in the same way. And I shall say:
“games” form a family. (p. 31–32)
Related to this is the fact that for many words, the sets have fuzzy boundaries,
blurred at the edges. Labov (1973), for example, has demonstrated this with the
words cup, bowl, and vase. Speakers were shown pictures of dishes of different
sizes and shapes, and found it hard to classify them neatly into the three sets. An
impressive amount of variability was found between individual speakers, and it
turned out that additional information—about the usage of the dish—played
a significant role in the classification: filled with coffee, the dish is more easily
categorized as a cup; filled with mashed potatoes—as a bowl.
Rosch’s Prototype theory set out to account for all this. According to this
theory, in its original form, the components of the concept are still there, but
they lose their status as necessary and sufficient conditions. At the center of the
set denoted by the word, we find the most prototypical exemplars of the cate-
gory, and they do meet all the conditions. But around them appear less proto-
typical exemplars, and they only share various subsets of the conditions—hence
the emergent sense of family resemblance. The least prototypical entities only
share a few properties with the others (the ostrich, for example, does not fly),
and they might share properties with members of other sets, which explains the
fuzziness at the boarders. Today, the psychological reality of prototypes seems to
be widely accepted. In the cognitive-functional camp, it is taken to be a central
fact about cognition at large: our conceptual world is arranged prototypically,
not just within the single concept, but also (as we shall see later), in the relation-
ships between the different senses of polysemous words. Taylor (1995), following
Geeraerts (1990), argues that prototypicality-based conceptualization is actually
more efficient than the rigid type of conceptualization based on necessary and
sufficient conditions, which he calls Aristotelian: “with only Aristotelian cat-
egories at our disposal, new data would often demand, for their categorization,
the creation of new categories, or a re-definition of existing categories. On the
other hand, new entities and new experiences can be readily associated, perhaps
as peripheral members, to a prototype category, without necessarily causing any
fundamental restructuring of the category system” (p. 53). In Jackendoff’s (1983,
1990) generative model of the lexicon, prototypicality is incorporated into a
formal system that in many ways reflects the logic of the classical theory, with
what Jackendoff calls preferential (P-) features: the feature (CAN FLY) in the
definition of BIRD, for example, is a P-feature, which allows it to be cancelled in
the case of the OSTRICH.
9780190256623-Dor.indb 66 03/06/15 6:17 PM

There are, however, good reasons to suspect that prototype theory by itself
cannot explain the observable facts. Three problems are especially important.
First, whereas the classical theory turned out to be too strong, prototype theory
seems to be too weak: the boundaries between concepts are too fuzzy. If all there
is to a sign is a general sense of family resemblance, and different members of the
set designated by the sign may be more or less prototypical, then nothing seems
to prevent entities, which clearly do not belong in certain sets, from becoming
non-prototypical members of these very sets. Consider camels and mice. If the
set designated by camel is identifiable on the basis of family resemblances, what
is it that prevents a mouse—evidently a mammal with four legs—from joining
the family as a very distant relative?2 The second problem is more technical: it
has to do with the way words, or concepts, combine to produce compositional
meanings (Osherson and Smith 1981, Fodor and Lepore 1996, Connolly et al.
2007). A combination such as pet fish only inherits some of the prototypical
properties of its components, and the prototypical pet fish (say, the guppy) is
neither the prototypical pet nor the prototypical fish. The third problem with
prototype theory is this: it seems that speakers sometimes make prototypical-
ity judgments that are surprisingly unrelated to what they actually think of as
the meaning of the word in question. This is most evident in those rare types
of words that do have exhaustive definitions: terms of kin relations and math-
ematics. Speakers, for instance, consistently describe prototypical grandmoth-
ers as elderly, old-fashioned women who make chicken soup and spoil children
with candy, but also consistently agree that the word grandmother refers to “a
female parent of a parent.” Someone who looks and behaves like a prototypical
grandmother may not be a grandmother at all, and someone may be a genuine
grandmother without meeting any of the prototypicality conditions mentioned
above. As Margolis and Laurence (2003) put it: “Mrs. Doubtfire (the Robin Wil-
liams character) may look like a grandmother, but Tina Turner really is a grand-
mother” (p. 196). Armstrong et al. (1983), to take another example, use Rosch’s
experimental methodology to show that speakers have prototype intuitions even
with respect to such well-defined concepts as odd and even numbers. The num-
bers 4 and 106 are equally even, and speakers know that, but they still judge 4 as
prototypical and 106 as non-prototypical.
What to do with all this? The major line of reasoning in the current litera-
ture is this: to move further in our understanding of word meaning, we probably
have to stop looking at words in isolation—and try to figure them out on the
basis of their behavioral patterns in context. Two types of contexts are especially
significant. The first, which will be discussed in chapter 7, is the syntactic con-
text: the syntactic configurations within which a word may appear in a sentence.
The second is the interpretative context: everything that happens to the word’s
9780190256623-Dor.indb 67 03/06/15 6:17 PM

meaning in the process of its interpretation. Much of the discourse today at-
tempts to figure out the contextual influence on the interpretation of polyse-
mous words, and much of this discourse stems directly from prototype theory.
Generally speaking, the patterns of lexical interpretation in actual com-
munication reveal the impressive extent to which words meanings are flexible,
ambiguous, variable, and context-sensitive. Polysemy is one central facet of this
reality. It is usually defined as “the association of two or more related senses with
a single linguistic form” (Taylor 1995), and the definition already exposes the
great difficulty of the topic: how do we decide how many senses a word has, and
whether or not they are related to each other? How do we distinguish between
ambiguity that derives from polysemy from ambiguity derived from homonymy?
Homonyms are linguistic forms that have two or more unrelated senses. The pro-
totypical examples seem to be easy enough. Bat, bank, and sole are homonyms:
each is associated with two unrelated meanings. Run, head, and movement are
polysemous: the different meanings in each case (running as jogging and as par-
ticipating in elections; the head as a body part and as manager; movement as
physical and political) are semantically related to each other. Where to put the
line between these prototypes is not an easy question. The same problem asserts
itself on the other side of polysemy: how do we distinguish polysemy (and hom-
onymy) from vagueness, the fact that word meanings are unspecified with respect
to this property or the other? Is the word friend vague with respect to gender, or
do we have to assume that it polysemous between male and female friend? And
there is more: how do we represent polysemy in our theory of word meaning—do
we represent instances of polysemy as a single word with more than one mean-
ing, or as separate words with the same form and related meanings? And how is
the essence of polysemy to be explained?
Whether clear distinctions between homonymy, polysemy, and vagueness
are possible, or for that matter necessary, is a matter of considerable debate. As
Tuggy (1993) and Geeraerts (1993) show, the different tests proposed for the
distinctions produce insufficient results. Three observations with respect to
polysemy, however, seem to be generally accepted. First, polysemy is a funda-
mental fact about human language (as opposed, for example, to formal com-
puter languages). Second, polysemy seems to present a problem only when we
look at the word in isolation; interpreting the word in context very often dis-
ambiguates it. If someone runs to the store, and someone else runs for presi-
dency, the expressions “to the store” and “for presidency” determine which of
the meanings of the polysemous run is to be used in interpretation (this is also
true of homonyms). Third, whichever way we think about it, polysemy does
not seem to bother speakers to the extent that it challenges linguists. This is
what Taylor (2003), following Ravin and Leacock (2000), calls the paradox
9780190256623-Dor.indb 68 03/06/15 6:17 PM

of polysemy: “The paradox is that, whereas polysemy raises all kinds of theo-
retical and methodological issues for semanticists, and practical issues for
lexicographers and for workers in natural language processing and automatic
translation, speakers of a language rarely experience polysemy to be a prob-
lem at all” (p. 647). At first sight, the fact that polysemy is disambiguated by
context seems to explain the paradox, but it actually accentuates it in at least
one way. As Taylor argues, “a sentence containing n words each of which is
m-times polysemous will in principle have n×m potential readings. It is com-
monly thought that context will serve to disambiguate the senses of a poly-
semous word. But if polysemy is ubiquitous, the disambiguating context will
itself most likely also be many-ways ambiguous. It is not surprising, therefore,
that disambiguation is a major issue in natural language processing. . . . What
is surprising is that for human language users, disambiguation, most of the
time, is not an issue at all” (p. 647–648).
How, then, should we account for polysemy? The literature seems to offer
four general alternatives, which unfortunately do not always concentrate on
the same types of observations. The first approach is couched within the gen-
eral perspective of cognitive linguistics (Brugman 1988, Lakoff 1987, Fillmore
1982, Geeraerts 1993, Taylor 1995): it takes the lexical entry itself to be noth-
ing more than a point of access into a rich set of cognitive models (or frames),
which enumerates the different senses one by one and specifies the conceptual
connections between them. The cognitive models are arranged in radial net-
works that are built around the idea of prototypicality—used this time to ex-
plain the family resemblances not between different words, but between the
different senses of the same polysemic word. At the center of the network re-
sides the prototypical meaning of the word, with its constitutive properties,
from which the other meanings are radiated. Lakoff (1987) demonstrates this,
for example, with the word mother (p. 74–76): the prototypical meaning of the
concept accessed by the word has such properties as giving birth, contributing
genetic material, nurturing, being married to the father, and being the closest
female relative. These properties may then be extended to produce the non-
prototypical usages of the word: “necessity is the mother of invention” is a met-
aphorical extension of the birth-giving property; the verb to mother in such a
sentence as “he wants his girlfriend to mother him” includes a metaphorical ex-
tension of the nurturing property; and the notions of mother and daughter used
in the description of formal tree diagrams extend the closest relative property.
Polysemy of this type, then, reflects the principled nature of human cognition,
which categorizes the world prototypically and metaphorically (and also meto-
nymically and so on). One major problem with analyses of this type is that they
end up splitting words into long sets of very fine-grained senses, and require
9780190256623-Dor.indb 69 03/06/15 6:17 PM

additional updates whenever a new usage of the word is found in discourse.

Very often they seem to turn vagueness into what Cuyckens and Zawada (2001)
call rampant polysemy.
The second type of approach, developed mainly in Pustejovsky’s (1995) gen-
erative lexicon, attempts to solve the problem by squeezing all the different senses
of the polysemic word into a highly complex lexical entry, which is built in such a
way as to allow the context of interpretation to select the relevant sense. Pustejo-
vsky’s formal and extremely complex model is designed for computational goals,
and is thus very unlikely to reflect what is happening inside human heads: every
lexical entry includes specifications of argument structure, event structure, lex-
ical inheritance structure, and four types of qualia, and computational results
are generated with a set of formal procedures such as type-coercion. As opposed
to the Lakoff-style cognitive analyses, Pustejovsky is interested in what he calls
regular polysemy, and many of his examples demonstrate the depth of the par-
adox we have discussed above: as speakers, we do not feel that breakfast is poly-
semous between a food reading and an event reading, or that book has an object
reading and an information reading. For Pustejovsky, books and breakfasts are
dot-objects: their lexical entries are overspecified for both meanings, and complex
computational procedures make sure that the right interpretation is selected in
the relevant context.
The third approach attempts to explain things in terms of a division of labor
between lexical and conceptual structures. In the formal camp, this strategy
is mostly represented by the theory of two-level semantics (Bierwisch 1983, Bi-
erwisch and Lang 1989) and in the cognitive camp by Evan’s (2009) theory of
lexical concepts and cognitive models (LCCM). In both conceptions, but in very
different ways, every lexical item has a unitary, abstract, schematic, and under-
specified lexical meaning that is different in essence from the rich conceptual
structures with which it is associated. The range of polysemous interpretations
of the lexical item is produced through the dynamic process of interpretation,
by different rules and principles. The most cited example from Bierwisch has to
do with the word university, which (counter-intuitively again) is supposed to be
polysemous—referring either to the institution or to the building. The lexical
meaning of the word only contains one property—the purpose of the university:
λx PURPOSE (x,w) & PROCESSES OF HIGHER EDUCATION (w). (The
λ-abstraction means: the property of x such that.) The conceptual structure asso-
ciated with the word, however, does include such concepts as building, campus,
and so on, and a special conceptual shift makes sure that these are selected in
the relevant contexts, such as “the university is not far from here.” Evans (2009)
proposes an analysis similar in spirit, but very different in theoretical style, in
which the word France, for example, is enigmatically characterized as having the
9780190256623-Dor.indb 70 03/06/15 6:17 PM

schematic lexical meaning [france], which provides access to an ordered set

of cognitive models—the nation-state model, the geographical landscape model,
the holiday destination model, and so on. The models, and a series of specific
principles of interpretation, make sure that the word is interpreted appropriately
in sentences such as “France is a country of outstanding natural beauty” and
“France is one of the leading nations in the European Union.”
The fourth and final approach attempts to explain polysemy as a pragmatic
phenomenon, and is most elaborately developed by Falkum (2011), working
within Sperber and Wilson’s (1986) relevance theory: meanings encoded by lan-
guage only provide the listener with “a minimal input, or clue” to the mean-
ing intended by the speaker in the particular occasion of communication. The
listener infers the speaker’s intended meaning on the basis of the principle of
relevance: other things being equal, a hypothesis concerning the speaker’s inten-
tion will be selected if it provides the listener with the maximal ratio between
contextual effects (the extent to which the hypothetical interpretation produces
a “worthwhile difference” in the listener’s representation of the world) and the
amount of cognitive effort required to produce the effects. Earlier work in lexi-
cal pragmatics (see Wilson and Carston 2007) had used this conception to ac-
count for lexical narrowing (the use of drink without an object to refer to alcohol
drinking) and loosening (the use of boiling in an expression like “it’s boiling out-
side”), and Falkum extends this type of analysis to polysemy. The different senses
of polysemous words emerge from creative usages by speakers, on specific occa-
sions, which require equally creative efforts of relevance-based inference on the
side of listeners. In these efforts, the encoded meanings are narrowed, loosened,
and so on—to produce the most relevant interpretations. The different meanings
of polysemous words are thus occasion-specific, which means that the context
does not disambiguate them—it actually creates them. To the extent that these
new interpretations are conventionalized, new words are created, which are syn-
chronically independent from their ancestors, but remain pragmatically related
to them at the diachronic level.3
Let us stop here, then, and take stock. What we have at our disposal at the
moment is the following set of nine open questions:
1. The definitional approach and its failure have taught us two things: approxi-
mations to definitions are useful and easy to produce, but they are very rarely
exhaustive. Why is that?
2. Prototype theory captures something right about word meaning, but it is too
weak at the boundaries. How can we correct for that?
3. How do we account for the fact that words such as even number and grand-
mother, which are exhaustively definable, manifest prototypicality effects?
9780190256623-Dor.indb 71 03/06/15 6:17 PM

4. Why is prototypicality not inherited in such combinations as petfish?

5. What is polysemy, as manifested in run, book, university, mother, and the
other examples, and how is it disambiguated by context? How can we account
for polysemy without rampant sense enumeration on the one hand and unre-
alistic formal complexity on the other?
6. How can we explain the paradox of polysemy, the fact that it does not seem to
bother speakers as much as it bothers researchers? (This is probably the same
as asking: how are humans different from computers?)
7. What do all these issues imply with respect to the separation between the lex-
icon and the encyclopedia? What is represented where?
8. What exactly does pragmatic inference contribute to all these issues?
9. How do we account for these issues without invoking innateness?
In the following, I will try to show how the theory developed here allows for
a new type of unified answer to these questions. On our way there, however,
I will have to highlight the single most important blind spot shared by all the
accounts we have looked at: lexical variability—not variability between the lex-
ical inventories of different languages, but variability within the same linguistic
community.
4.2 The Blind Spot: Where Is Social Variability?

As speakers of the same language, we do not have the exact same meanings
for the same words. Some of the authors mentioned above acknowledge this
fact, but none of them considers it a fundamental issue in need of explana-
tion. This is obviously true of the formal perspectives, but also of the usage-
based models. Geeraerts’ (2010) acknowledges this explicitly in the following
paragraph:
Prototype-theoretical studies generally tend to ignore the question

whether and to what extent the prototype structure of the category
is shared by all the speakers of the language. This is remarkable, be-
cause the prototypical model of category structure might be plausibly
interpreted as involving social variation over individuals rather than
just psychological variation over contexts of use. For instance, if a lexi-
cal category consists of a core reading A and peripheral senses A1, A2,
and A3, the subconcepts might be variously distributed over subgroups
of the linguistic community. A natural assumption could be that at
least the core reading A would be shared by all, and only the peripheral
9780190256623-Dor.indb 72 03/06/15 6:17 PM

senses A1, A2, and A3 would be distributed in different combinations

over different individuals or groups. According to such a model, the
central application of the prototypically structured concept, the proto-
type, implicitly receives a social function rather than a purely psycho-
logical one. The shared knowledge of the speakers of the language then
consists of a common central reading plus rules of semantic extension
that could lead to either A1, A2, or A3. But whether a particular indi-
vidual ever puts rule 1 into practice to derive A1 from A (and further,
whether he or she subsequently stores A1 in his/her long-term seman-
tic memory) would depend on the specific circumstances of the indi-
vidual’s linguistic history, i.e. on whether he or she ever experienced
a context in which A1 was relevant. In most actual studies of proto-
typicality in linguistics, however, the social interpretation is largely
ignored. (p. 255)
As we shall see, Geeraerts’ suggestion that the core reading may be shared by
all speakers while the peripheral ones may be variably spread, is only one way in
which variability within the community may assert itself. What is more impor-
tant at the moment is the almost total absence of research projects that attempt
to expose the actual levels of variability between speakers with respect to lexical
meanings. It seems that whenever such an effort was attempted, serious levels of
variability were found. We have already seen this in Labov’s dish experiments,
and the same is true of McCloskey and Glucksberg (1978): they asked their sub-
jects to make membership decisions on 540 candidate exemplar-category name
pairs, such as apple and fruit, and found high degrees of agreement for proto-
typical category members and for non-members, but serious disagreements on
intermediate-typicality items. This was done in two sessions, and the subjects
were also individually inconsistent with respect to the intermediate items from
one session to the next.
It is fascinating to discover, then, that the single most important experiment
on variability of this type was performed more than a hundred years ago: its
story has recently been told by Levelt (2013). The German experimentalist Karl
Marbe, working together with Albert Thumb, performed a word association
task—one of the first of its kind—with eight subjects, measuring the amount of
time it took them to react and produce their answers. What he discovered, and
published in a 1901 monograph, came to be known as Marbe’s law: when the
subjects produced similar associations, their reaction times were faster. Response
agreement and response time were positively correlated.
Let us, then, try a thought-experiment of our own. Consider the dish draw-
ings used by Labov in his 1973 paper (Figure 4.1).
9780190256623-Dor.indb 73 03/06/15 6:17 PM

1 2 3 4
5 10 13
6
11 14
7
12 15
8
16 18
17
9 19
figure 4.1
Instead of asking what each of these should be called, let us begin with the
fundamental fact of experiential variability. The dishes can be split into sets in
very many different ways, each of which may or may not be adopted by different
individuals. They are different in size, shape, width, height, functionality, and
so on; some are more beautiful or elegant than others, more regular or strange;
they are easier or more difficult to make, to use, to store, to wash—and so on and
so forth. Different individuals, in different contexts, would also make variable
judgments on the differences themselves: for some people, for some purposes, all
the dishes would simply be of the same type; for others, in other circumstances,
each would make a category of its own.
Assume, then, that we let a community of speakers mutually identify the
dishes, pointing at one, acknowledging the pointing and giving it a name, then
pointing at another and naming it, and so on, iteratively, until the entire domain
is linguistically categorized. Let us, moreover, make sure that the community
does not just approach the process as an intellectual exercise: the goal is to reach
9780190256623-Dor.indb 74 03/06/15 6:17 PM

the optimal level of instructive communication, where speakers will be able to

effectively talk about the dishes without showing them to each other. What can
we say in advance about the end result? Four points are crucial. First, we have
to acknowledge that we cannot predict what the semantic field will look like,
because it will depend on the entirety of the social dynamics: different speak-
ers will bring different experiential perceptions and communicative interests
into the process of mutual-identification, and also different capacities, sensi-
tivities, and social standings. Some categorization proposals will be accepted,
others will be rejected. Because of this, if we ask different communities to go
through the same process, we should expect to find patterns of similarities and
differences between their end results, of the type that we actually find between
languages. To be sure, the idea that different communities would end up con-
structing different semantic fields around the set of dishes does not mean that
certain categorizations will not appear and re-appear in many of them, maybe
even all of them. What it does mean is that in order to explain the dynamics,
we will no longer need to find a universal subset of categories, and postulate
their innateness, as Wierzbicka, Jackendoff, and many others do. If certain cat-
egorizations turn out to be universal, if it turns out that they manifest no vari-
ability between individuals, it stands to reason that their mutual-identification
would be easier to achieve, but that does not imply that they would necessarily
be the most important categories to mutually identify for instructive commu-
nication. They would thus participate in the same type of game as the more
variable ones, and their mutual-identification would abide by the same social
logic as the others.
Second, we can predict that the end result in each community will be com-
patible, to different degrees, with the experiential categorizations of their dif-
ferent speakers: the higher the level of compatibility, the easier it will be for the
speaker to use the semantic field for instructive communication. Because of this,
different subgroups within each community may settle on different categoriza-
tions, designed for different communicative interests, and some individuals may
find it easier or more difficult to move from one scheme to the other.
Third, we can predict that in the course of the negotiation process, speakers
would also necessarily mutually identify some of the properties of the dishes, and
give them their own names: width and height, handles and legs, shapes and sizes.
The emergent field on the symbolic landscape would include not just names for
the different types of dishes, but also contrasts and similarities, parts and func-
tions, and so on. Thus, each speaker would eventually represent the field on two
levels, only partially correlated with each other—an experiential level and a sym-
bolic level: one analogical, prototypical, rich in detail, variable and experiential,
consisting of the actual dish experiences; the other discrete, schematic, based on
9780190256623-Dor.indb 75 03/06/15 6:17 PM

social negotiation, consisting of the signifieds and the web of semantic relations
between them, designed for instructive communication.
Fourth, and here we begin to get into the technical issues, different end re-
sults in different communities will produce different prototype structures: if
dishes 1, 2, and 3 end up receiving the same name, dish 2 will probably be judged
as the most prototypical; if dishes 2, 3, and 4 receive the same name, 3 will be the
prototypical one, and 2 will be demoted to a less prototypical status. The sense of
family resemblance between the dishes will stay the same, but the prototypical-
ity judgments will change. This is so crucial that I would like to ask the reader
to spend some time playing with this: concentrate on Figure 4.1, try different
categorizations—and actually see how the cups change from regular to strange,
and vice versa, depending on their linguistic categorizations. The semantic field
determines, or at least heavily influences, the patterns of prototypicality.
This observation may be interpreted in two ways. First, minimally, proto-
typicality as such is indeed a foundational property of human cognition, as
Geeraerts and Taylor claim, but the actual patterns of prototypicality around
words are based on the socially constructed patterns of language. Second, maxi-
mally, prototypicality itself is a fact about the interaction between language as
a social technology and private experience, which as such does not assign pro-
totype status to its analogue categorizations. Experimental work with pigeons
(Jitsumori et al. 2011) and monkeys (Smith et al. 2008) shows that the animals
can be trained to show prototypicality effects in their categorizations of abstract
stimuli, but the experiments do not support the hypothesis that the effects actu-
ally reflect a prototype-based cognition. If genuine prototypicality does indeed
turn out to emerge only in the process of mutual-identification for language, this
would make perfect functional sense: instead of naming each and every dish by
itself, anchors could be pointed at and named as the central representatives of
their categories, and the rest could be referred to as non-prototypical members.
4.3 Explaining the Phenomena

Let us, then, get back to our list of open questions, beginning with the issue of
definitions. The error of the classical theory turns out to have only been this,
that it mistook the web of semantic relations surrounding a signified for its def-
initional components. The essence of the web is not definitional, and it is not
meant to be. It emerges, gradually and tentatively, through the process of iter-
ative experiential mutual-identification for language, as members of communi-
ties reach agreements about distinctions and relationships between elements in
their experiences that are, for them, worthy of communication. It is the product
of a haphazard, unplanned, evolutionary process, in which signs and relations
9780190256623-Dor.indb 76 03/06/15 6:18 PM

are added to the symbolic landscape, in order to meet the growing and chang-
ing communicative needs of certain communities of speakers, in certain specific
social settings. There is no reason to assume that such a process should result in
logically well-formed definitions. The signifieds chair and sit, for example, are
semantically associated on the landscape not because they were meant to define
each other, but, first, because the members of the relevant communities agreed
that chair-experiences and sitting-experiences were to be recognized as things to
talk about, and, second, because for all of them the activity of sitting was mutu-
ally identified as something that distinguishes chairs from other types of entities.
As associations of this type multiply in the iterative process of mutual-
identification, however, the set of conventionalized semantic relations between
a signified and its neighbors gradually comes to approximate a definition—and
this is why such approximations are eventually useful and easy. Consider Fodor’s
paint: we know what painting is from our experiences of painting, not from
anywhere else, but the scenarios that Fodor constructs to show that definitions
are impossible actually help us distinguish, for the sake of instructive commu-
nication, between experiences of painting and other experiences in which paint
is involved. The components of his imperfect definitions are there not to define
the essence of painting: they are the mutually identified semantic properties that
tell us how, on the symbolic landscape, the activity of painting is different from
other types of activities. Thus, when we wish to describe Michelangelo dipping
his brush into the paint, we will probably not use the verb paint, because we have
mutually identified painting as an activity involving the primary intention to
cover a surface with paint. This we did not need to discuss explicitly: the experi-
ential anchors we pointed at in the process of the mutual-identification of paint
all involved such a primary intention.
The point, then, is not just that the set of semantic connections does not
go beyond the level of approximation: the logic of the instructive strategy im-
plies it does not have to. The function it is supposed to fulfill on the symbolic
landscape is much more modest: not to tell speakers what things are, or what
the world is like, but only to direct speakers toward a certain tentative set of
distinctions, to tell them which things should be thought of as different from
which (for the sake of conversation), and which things are to be thought of as
connected to which (again, for the sake of communication). What these differ-
ent things actually are is something the speakers have to figure out for them-
selves, on the basis of their experiences and the experience of the specific event
of communication in which the signs are used. The amount of information
required for this communicative function is much smaller than that required
for the definitional function. There is no reason to assume that it should ever
amount to a definition—unless we are talking about mathematics, which is
9780190256623-Dor.indb 77 03/06/15 6:18 PM

the human project of defining certain essences on the basis of pure logic. In
this sense, the idea of exhaustiveness itself can now be seen in its true colors,
as a prescriptive attempt to clean the emergent mutually identified structures
of language to such an extent that language itself could tell us what things are,
independently of our private experiences. The entire formal tradition, in lex-
ical semantics and elsewhere, should be understood as a continuation of this
prescriptive tradition.
The meaning of a sign cannot be reduced to the semantic relations it main-
tains with others on the symbolic landscape, but neither can it be reduced to the
prototypical complexities of the experiential cluster it points at. The process of
mutual-identification inevitably involves two dynamics, not one: the identifica-
tion of distinctions and other types of connections between sets, and the identi-
fication of similarities within sets. The two dynamics take place at two different
levels of meaning: the symbolic landscape distinguishes and connects between
sets; experience identifies similarities within them. Once the distinctions are
semantically made, then, prototypicality theory may tell us which entities are
closer to the anchor, and thus more exemplary members of the cluster—but it
cannot contribute to the construction of demarcation lines between clusters.
This is why it cannot ensure that mice will be excluded from the set of camels.
The fact that signifieds, then, whose very essence lies in being discretely
different from each other, point at analogue experiential clusters around an-
chors, implies that speakers are instructed by the symbolic landscape to focus
on certain types of distinctions between experiences that are otherwise not that
different from each other. Grandmothers, for example, are indeed similar in all
kinds of ways to other human beings: being old, making soup, handing candy to
children—all these are properties which are spread all over the experiential con-
tinuum. The speakers who at some point decided to add the sign grandmother
to their language were attempting to abstract away from this experiential con-
tinuum, and highlight a single property at the expense of others—because they
were specifically interested in talking about mothers-of-parents as such. This was
a necessary move, not in spite of, but because of the fact that there are family
resemblances between grandmothers and other types of people. Many of the
grandmothers who served as experience-anchors in naming events in which
the word grandmother was stabilized were indeed busy spoiling children with
candy, which means that for many speakers old ladies with candy came to be
perceived as prototypical members of the experiential cluster associated with the
sign grandmother. But what they came to know about the properties of pro-
totypical grandmothers had very little to do with what they had to know about
the distinction between grandmothers and non-grandmothers in order to use
the sign appropriately. Grandmotherhood is represented on the two levels—the
9780190256623-Dor.indb 78 03/06/15 6:18 PM

experiential and the symbolic—in two different ways: one for knowledge, the
other for instructive communication.
To understand why pet fish doesn’t inherit the prototypical properties of pet
and fish, and then to explain polysemy, we need to delve deeper into the process
of mutual-identification: assume that speaker A wishes to talk about a certain
experience that has not been mutually identified and named by the community.
What innovative strategies can A adopt? There are actually no more than three:
A may choose to treat the experience as a non-prototypical token of an experi-
ence type that has already been mutually identified and use the name given to
it to refer to the new experience; A may attempt to launch a process of mutual-
identification of the experience by pointing at it and proposing to give it a new
name; or A may try to refer to the new experience with a creative usage of exist-
ing signs, expecting the listener, B, to pragmatically infer the intended meaning
from relevance considerations, as Wilson, Carston, and Falkum suggest. These
three strategies are not just technical variations. They actually mark the three
epistemic positions the speaker may adopt in considering the relationship be-
tween the new experience and the symbolic landscape: proposing a new name
for the new experience implies that the experience is categorically different from
everything that has been mutually identified before; using a regular word, with
its regular meaning, implies that the experience is a new variant of a familiar
category; proposing a creative use of existing words—similar enough to be rec-
ognized and different enough to be recognized as different—implies that the
experience is distinct but related to what has already been mutually identified by
the community.
The innovative speaker who adopts the third strategy, then, makes the first
move in such processes—but then the others have to react. In many cases, the
listeners would manage to pragmatically infer the speaker’s meaning, and the
conversation would move on. In others, they would fail to understand, repair
strategies would be employed—the speaker, for example, would replace the ex-
pression with another one—and the conversation would continue. In some cases,
however, the interaction would turn into a genuine mutual-identification event,
where the innovative usage would be accepted as pointing at the new experience.
Nothing in this process implies that the prototypical properties of the origi-
nal words should be inherited by the innovative sign. Like all other signs, pet
fish points at its own experiential cluster: fish of the types that we find in aquari-
ums. The prototypicality judgments that emerge there have to do with the dif-
ferent types of fish, their sizes and shapes, their colors and so on. The semantic
relations that remain on the symbolic landscape, between pet fish and its parents,
pet and fish, register the memory of the mutual-identification event, and indicate
the relatedness of pet fish to other pets on the one hand, and other fish on the
9780190256623-Dor.indb 79 03/06/15 6:18 PM

other. Importantly, the semantic relatedness does not seem to play a real role in
the actual use of the word for communication: speakers and listeners translate
the experiential cluster of aquarium fish into the word pet fish, and vice versa.
If we now turn to polysemy, we can see that the process naturally distin-
guishes between genuine cases of polysemy and cases where it is theoretically
unnecessary. Consider the genuine examples of run, movement, and head. Each
of them carries a clear memory: the innovative speaker pointing at the person
managing the department and suggesting to call him or her the head of the de-
partment; the others understanding the dialectic relationship—seeing that the
new experience resides outside the experiential cluster including physical heads
of different types and shapes, but also that the new experience shares some simi-
larities with the physical head—and agreeing to coin a new word, head 2 , with an
identical signifier, that bears a semantic relationship to the original one, head1.
Crucially, these are two separate signs, not the same sign with two meanings,
because each of them points at a different experiential cluster. Note, moreover,
that the semantic connections between the two polysemous words should not
necessarily be registered by all speakers in the same way. Consider movement1
and movement2 . For all speakers of English, the two signs point at two very dif-
ferent experiential clusters, but speakers may have variable conceptions of their
semantic relations: for some, the relationship would have deep implications for
the understanding of politics; some would vaguely recognize that the relation-
ship is there; others would be surprised when shown the relationship (just as
some speakers would be surprised to discover the relationship between breakfast
and the breaking of fast).
All this is necessary with head1,2 and movement1,2 , but not with the ex-
amples that served us before to accentuate the paradox of polysemy. As speak-
ers we feel there is no polysemy in book and in breakfast, and this is exactly
because, until we try to teach the meanings of these words to computers, there
really is none. Book is not polysemous between an information and an object
reading, and breakfast is not polysemous between a food and an event read-
ing. We were introduced to these two words in mutual-identification events in
which certain experiences where pointed at and named: breakfast experiences
and book experiences. Our individual breakfast experiences were probably
quite different, but in their mutual-identification we isolated and highlighted
a certain experiential complexity—not just the food types, not just the first
meal of the day, not just the event with its often ritualistic manners, but all of
these together. It was this experiential complexity that the sign came to dis-
tinguish from other experiences which were, in this way or the other, similar
to it: grabbing something from the fridge (“why don’t you sit down and have
a proper breakfast?”); eating at other times during the day; eating other types
9780190256623-Dor.indb 80 03/06/15 6:18 PM

of food, and so on. The same is true of the book. The mutually identified ex-
perience was not that of a physical object or the carrier of information, but of
a physical object containing information. That was the entire point: the book
was mutually identified as such to distinguish it from other physical objects,
containing other types of things (or nothing at all). At the level of the sym-
bolic landscape, the two signifieds came to be surrounded by signifieds point-
ing at related experiences, at properties of the experiences and so on—eat,
have, come down for, dinner, food, morning; read, write, buy, carry, interesting,
library. This is all we need.
Pustejovsky has to treat book and breakfast as inherently polysemous—
marked as the dotted-types physobj•info and event•meal—because he works,
along with the entire formal literature, with the conception of semantic com-
positionality: the semantics of a complex expression is built from the semantics
of its components. His analysis should thus prevent a result in which “carry the
book” eventually means “carry the information” and “read the book” means
“read the physical object.” All this, however, simply does not arise within the
perspective presented here: this is not how we, as human speakers, use lan-
guage. As we saw in chapter 3, the acts of imagination that combine the el-
ements in these two expressions take mutually identified experiences, not
semantic components, as their objects. The expression “carry the book” in-
structs the listener’s imagination to bring book experiences and carrying expe-
riences from memory and imagine their intersection; the expression “read the
book” instructs the listener to imagine the intersection between a book expe-
rience and a reading experience. The intersections come out differently, as they
should, because read and carry are different—not because of hidden polysemy
in the lexical item book.
The case of university, dealt with by Bierwisch, seems to highlight another
facet of this reality. Just as before, the polysemy is unwarranted: for most of us,
the experiential cluster pointed at by the word involves both the function as an
institution for higher education, and the physical site, the campus or building,
as well as other experiential properties. It should be remembered, however, that
different speakers may potentially associate very different experiential clusters
with the word—all the way from the green campus where you mother teaches,
to that mysterious institution that the rich people in the big city go to for some
unclear reason. Someone may say something like “the bus station is in front of
the university” without knowing what the purpose of the building is, whereas
someone else may say something like “if you don’t study you won’t be able to go
to university” without ever seeing a campus or getting into a university building.
Such variable representations will produce problems of communication between
different speakers, and their social distribution, among teenagers for example,
9780190256623-Dor.indb 81 03/06/15 6:18 PM

may tell us quite a lot about their future education (or lack of it). As far as each
individual speaker is concerned, the polysemy postulated by Bierwisch is unwar-
ranted, but polysemy is still a fact at the social level—and as such it is an impor-
tant factor.
Putting all this together, we may now take a closer look at Lakoff’s analysis
of the word mother, in which the properties of the prototypical mother—giving
birth, nurturing, being married to the father, contributing genetic material, being
the closest female relative—are extended into the different non-prototypical
usages of the word on the radial network. Lakoff’s analysis makes things look or-
derly and symmetrical, but the elegance unfortunately hides a much more com-
plex and dynamic emergent reality. Two observations are significant. To begin
with, the set of properties of the prototype itself does not just lie there peacefully:
each of the properties is a site of social variability, actually social struggle—in
many ways today more than ever. When surrogacy allows women to become
mothers without giving birth; when the rights of biological and non-biological
mothers are contested in courtrooms all around the world; when single mothers
(together with same-sex parents and many others) change the very concept of
family—the word mother, even in its most prototypical usage, is not just moving
in front of our eyes, but is also represented differently by individuals who occupy
different positions in the drama.4
The second observation, more important for our current purposes, has to do
with the radial extensions from the prototypical mother. For Lakoff, the nurtur-
ing property is metaphorically extended in “he wants his girlfriend to mother
him”; the birth-giving property is extended in “necessity is the mother of inven-
tion”; and the family relation is extended in the depiction of an immediately
dominant node on a formal tree as a mother node, as in “the immediately dom-
inant node is the mother node.” This elegant symmetry misses out on the fact
that the three extensions are of very different types. In the first case, the speaker
actually does nothing to extend the prototypical meaning of the word mother
itself. He or she only says: “he wants his girlfriend to belong in the experience
cluster pointed at by the word mother as we know it.” To understand the utter-
ance, listeners have to imagine the girlfriend as the mother of the guy, and prag-
matically infer—based on the world knowledge that the girlfriend cannot give
birth or genetic material to her boyfriend, and that such interpretations would
anyway make little sense—that the speaker intended to mean that the guy wants
his girlfriend to nurture him like a son. There is no polysemy here, and no con-
ventionalization of a change in the lexicon. In the second case, the innovative
speaker means: “the relationship between necessity and invention, which is ob-
viously distinct from the biological relationship between mother and offspring,
is nevertheless metaphorically related: it is almost as if necessity gives birth to
9780190256623-Dor.indb 82 03/06/15 6:18 PM

invention.” Here, again, listeners need no more than pragmatic inference to in-
terpret the expression. If the metaphorical relationship is mutually identified,
and the expression is conventionalized, what is added to the lexicon is not a new
word, but the entire construction, the proverb “necessity is the mother of inven-
tion.” Again, no polysemy is required.
The third case, however, is already qualitatively different: the innovative
speaker launches a genuine mutual-identification event, in which an immedi-
ately dominant node on a tree is pointed at and named mother. What is thus
added to the lexicon is a new word, mother2 , pointing at a distinct experiential
cluster, which bears a semantic relationship to mother1—a semantic relationship
that may or may not remain represented in the minds of future speakers. Here,
for the first time, polysemy is genuine.
All this takes us a long way toward the resolution of the paradox of
polysemy—the fact that polysemy constitutes a major problem for semanticists,
lexicographers, and computational linguists, whereas “speakers of a language
rarely experience polysemy to be a problem at all” (Taylor 2003, p. 647). Much
of the research attempts to gather the variable meanings, associated with the sign
by different speakers within the community, into a single representation: they
work with the assumption of the idealized community, and thus, paradoxically,
mistake the social variability for variability inside the individual mind. Speakers,
for their part, are only interested in language as a technology for instructive com-
munication. When they speak, they proceed in their production process from
the experiential cluster to the sign: they do not encounter the fact that language
includes both head1 and head2 , because they begin with either the experience of
a physical head or the experience of a manager, and go directly to the right sign.
As listeners, they do face the challenge of ambiguity, but the challenge is exactly
the one involved in the disambiguation of homonyms such as bank1 and bank2 .
The signifier directs the listeners to two different signifieds, which point at two
different experiential clusters, one of which needs to be selected. It is selected, as
anywhere else, by pragmatic inference, within the context created by the entire
discourse and everything around it. The semantic relationship between head1
and head2 does not participate in the disambiguation process. It does assert itself,
however, when speakers are subjected to psycholinguistic experimenting. In Kle-
pousniotou (2002), for example, subjects heard a sentence priming the different
meanings of an ambiguous word—homonymous or polysemous—without ac-
tually including the words themselves. At the same time, they were presented
with a string of letters on a screen, and asked to decide whether the string was a
real word of English. The reaction time for polysemous words was significantly
shorter than for homonyms. The semantic connection seems to produce greater
priming effect.
9780190256623-Dor.indb 83 03/06/15 6:18 PM

4.4 Conclusion
The account proposed in this chapter is obviously programmatic, a first attempt
to re-interpret the empirical issues on the basis of the new theory. There is also
much more that needs to be said about signs and their meanings. In the remain-
der of the book, I will touch on certain psycholinguistic issues in word process-
ing (chapter 6); the relationships between word meaning and syntactic behavior
(chapter 7); inter-linguistic variability in word meaning (chapter 8); the acquisi-
tion of word meaning (chapter 9); and the evolutionary emergence of signs as dis-
crete instructors of imagination (chapter 10). Nevertheless, I believe this chapter
already takes a significant first step toward the understanding of the phenomena.
Every sign, simple or complex, is represented by the individual speaker as a
triplet: a signified on the symbolic landscape that is semantically connected to
other signifieds, marked by a signifier, and pointing at a single experiential clus-
ter around an experiential anchor. The sign thus connects two levels of meaning
representation—the private-experiential and the social-semantic. The pointing
function from the signified to the experiential cluster is normative: it parcelizes
the world of experience into signifiable clusters, based on the process of mutual-
identification. Different speakers within the same community may represent the
same signs differently, in terms of the experiential clusters they point at, their
prototypical and less-prototypical properties, and their semantic relations. This
is a constant cause of miscommunication.
The web of semantic connections between signifieds describes the ways in
which the clusters are different from each other and related to each other. It
emerges from the iterative process of mutual-identification, and as such it allows
for approximate definitions—but very rarely for exhaustive ones. Prototype
theory handles the complexities within the clusters quite well: the anchor is the
most prototypical token, and the others are judged as more or less prototypi-
cal based on their similarities with the anchor. Prototypicality, however, is not
supposed to handle the parcelation, which is semantic and normative. Whether
prototype effects appear without language is an open question. Words such as
even number and grandmother demonstrate the fact that the two types of judg-
ments refer to different levels of representation: they are exhaustive at the seman-
tic level, prototypical at the experiential level.
The construction of a complex sign to point at a new experiential cluster cre-
ates a new prototypical pattern around the cluster and does not inherit the pro-
totypical properties of the original clusters pointed at by the combined signs.
Every sign points at a single experiential cluster: signs are never polysemous
within themselves. Sense enumeration is unnecessary, much of the polysemy
postulated in the literature is unwarranted, and so are the formal apparatuses.
9780190256623-Dor.indb 84 03/06/15 6:18 PM

Language, however, includes sets of signs, two or more, which have the same sig-
nifier but nevertheless point at distinct experiential clusters: these may either be
homonyms or polysems. If they are polysems, they are most probably the result
of an innovative mutual-identification process in the history of the language, in
which the original name was borrowed for the purposes of pointing at the new
cluster. This past process may or may not be registered in the semantic relations
represented by individual speakers. Speakers do not experience polysemy as a
problem: when talking, they always go from the experience to the signified point-
ing at it; as listeners, they decide which experience type to bring from memory,
within the context, on the basis of relevance considerations. The semantic con-
nections registering the relations between polysemous words do not participate
in the communication process. As far as the process is concerned, there may not
be a difference between polysems and homonyms.
The lexicon and the encyclopedia are thus different in essence: one is social-
semantic, the other private-experiential. They are, however, related to each other
dialectically. Signs are regularly used in a variety of innovative ways for the ex-
pression of ad hoc meanings, with more of the burden of interpretation falling
on the listeners’ capacity for pragmatic inference. Such inferences, however, do
not directly affect a change in the lexicon. The lexicon changes only when a new
sign is mutually identified, pointing at a different experiential cluster. Nothing
here requires innateness of the concrete type, a priori components or properties
of word meaning. The observable phenomena are determined by the functional
logic of language as a technology for the instruction of imagination. The actual
use of the technology, including the selection and interpretation of words, re-
quires a wide array of cognitive capacities—some of which are probably partially
genetically based (see chapters 9–10).
9780190256623-Dor.indb 85 03/06/15 6:18 PM

5 T h e S p i r a l s o f R e l ati v it y
Up to this point, we have seen how pieces of private experience find

their way, through mutual-identification, into the linguistic fabric of
socially constructed meaning—and how the interaction and friction
between the normative meanings of language and the private world of
experiential meanings help explain the functional logic of language,
and some properties of its signs. But is it not also the case that lan-
guage, with its normative meanings, influences the way we experi-
ence privately? This is the question of linguistic relativity. Over the
past twenty years, a vibrant field of research has emerged around this
question, and in this chapter I would like to suggest that the theory
of language presented here allows for a major re-interpretation of the
empirical results accumulated in the field—and thus opens the way
toward a new understanding of the dialectic relationship between lan-
guage and experience.
The modern empirical investigation of linguistic relativity devel-
oped from within a much older philosophical and ideological tradi-
tion (for extensive reviews, see Gumperz and Levinson 1996; Lucy
1992a,b, Leavitt 2006, and Everett 2013). In section 5.1, I will de-
scribe this development in some detail because of its direct relevance
for the state of the discourse today. What separates the modern field
from its traditional background is mainly the move from a deter-
ministic view, according to which we are forced to look at the world
through the lens of our language, to a softer conception of influence,
which allows for the question, for the first time, to be investigated
empirically. The accumulated results in the field very strongly indicate
that language does influence the way we experience when we speak,
along the lines of Slobin’s (1996, 2003) conception of “thinking-for-
speaking.” There are also very good experimental demonstrations of
the influence of language on the way we experience in general, but
here there is still no consensus.
In section 5.2, then, I will begin to show how the theory re-
interprets the empirical results and assigns them with new significance.
9780190256623-Dor.indb 86 03/06/15 6:18 PM

The Spirals of Relativity • 87
Slobin’s “thinking-for-speaking,” re-interpreted as “experiencing-for-instructing,”

will be shown to derive as a necessary implication from the technological essence
of language: as with any other technology, we must learn to experience in specific
ways in order to make it work. In section 5.3, I will suggest that the extent to
which language-specific modes of experiencing influence the process of experi-
encing in general depends on a long list of parameters: every individual stands
at the center of a vector space of influences, experiential and linguistic, and the
actual contribution of language depends on the susceptibility of the individual
to these influences, and on the direction, temporal order, and strength of the dif-
ferent vectors. The tensions between the experiential and linguistic influences
may be resolved in a number of principled ways, which seem to be reflected in
the different empirical results. The tensions, moreover, may not just influence
experience—they also change language. The two opposite processes—the influ-
ence of language on private experiencing, and the experience-based invention of
new linguistic forms—connect language and experience in a never-ending cycle
of bi-directional influence, always spiraling over the foundational constant of the
experiential gap.
The emergent picture, then, is dialectic, cyclic, and highly variable. As I will
suggest in section 5.4, however, the theory predicts also that in certain specific
configurations, the influence of language on private experiencing should be very
strong—almost deterministic: cases where experiencing is only possible while
speaking; and cases where private experiences are built on the sole basis of in-
structed imagination. These predictions seem to be supported by observation,
and they open new venues for empirical research. All this puts a new complexion
on the original philosophical question, and helps re-connect the scientific study
of language with the critical discourse on the relationship between language, ide-
ology, and propaganda.
5.1 From Determinism to Influence

The philosophical and ideological discourse on the relationship between lan-
guage and cognition emerged from the very fruitful collision of four of the most
dramatic developments of modernity, which culminated simultaneously in the
nineteenth century. First, the rise of the nation-state as a political model, with
its romantic ideology of nationalism—together with the rise of the print indus-
try (the first serious example of modern capitalism) that helped turn the native
languages of Europe into national languages—all these brought along a vested
interest in a view of language as both an exact reflection of the national spirit,
the Volksgeist, and a major determining factor in its construction. The interest
9780190256623-Dor.indb 87 03/06/15 6:18 PM

was not simply academic. True to its nationalistic origins, the perspective was
formulated in a way that somehow revealed the superiority of the languages we
speak (we Europeans, very often we Germans) over the languages of the others,
and thus, by implication, the superiority of our spirit over theirs. Our language
allows for sophisticated thinking, for a sound view on the world; theirs does not.
“A people which did not distinguish its feelings much and did not distinguish
them sharply,” wrote Johann Gottfried Herder (1772), “a people which did not
have enough heart to express itself and to steal expressions mightily, will also be
less at a loss because of nuances in feeling, or will make do with slothful semi-
expression” (p. 114).
All this was accompanied by the second development, the fact that Euro-
peans, in the course of the project of colonialism, discovered more and more
languages around the world that were ostensibly very different from the lan-
guages known to them at the time. Travelers, adventurers, and priests began to
describe and analyze these languages, and suggest ideas as to the relationships
between them and the cultures within which they emerged. A huge field of in-
terest quickly emerged which eventually, in the twentieth century, matured into
the twin disciplines of linguistic anthropology and anthropological linguistics.
Here, too, the discourse began with the notion of European superiority as a clear
presupposition. It turned into a serious science with Franz Boas’s insistence that
there are no primitive languages, that all languages should be investigated on
their own grounds.
The third development, which has received less attention than it deserves in
the historiography of the field, was secularization: the question of linguistic rela-
tivity in its modern form could only begin to emerge with the weakening of the
conviction that both human language and human thought, whichever way one
thinks about them, are the divine creation of God. Herder, who is often quoted
for his conception of language and national spirit, actually opens his treatise, as
late as 1772, with the sentence: “I do not want to pursue the hypothesis of the
divine origin of language any further on a metaphysical basis, for its ground-
lessness is clear psychologically from the fact that in order to understand the
language of the gods on Olympus the human being must already have reason
and consequently must already have language.” He then dedicates a good part of
the first half of the text to the quarrel with the divine origin perspective on other
bases. The question of relativity, and the idea of the unbreakable tie between lan-
guage and national spirit, were also Herder’s way to try and understand what it
means to be human in a world where humans, not God, are responsible for who
they are.
Finally, there was Kant’s philosophy of mind. Following two centuries
of struggle between Rationalists and Empiricists on the nature of human
9780190256623-Dor.indb 88 03/06/15 6:18 PM

knowledge, Kant (1768) found a way out. Rationalists claimed that true know-
ledge could only emerge in pure reasoning, based on a priori intuitions, as far
away as possible from sense experience; the Empiricists insisted that sense expe-
rience is the source of all knowledge. (This, too, was a struggle about religion: as
long as mind and language are God-given, rationalism must be correct.) What
Kant suggested, put very crudely, was the following. First, we have no access to
objective reality, to things in themselves. Second, what we perceive of the world,
our sense experiences, are already structured by the a priori categories and pure
intuitions of our own rational mind. Both sides to the debate thus turned out to
be right: knowledge is achieved empirically—on the basis of a priori reasoning.
We look at the world through the lens of reason.
Kant’s bold move was decidedly universalistic—the categories and intu-
itions are shared by all rational minds—but it immediately opened the door
for a relativistic re-formulation: what if we look at the world through the cat-
egorical lenses of our different languages? This was Wilhelm von Humboldt’s
view: every language has a “genius” of its own, an inner form that determines
everything in it, and because “language is the formative organ of thought”
(Humboldt 1999, p. 54), the genius of every particular language determines
the way its speakers think. The Humboldtian view was taken up, with adjust-
ments, by generations of future scholars, and was developed into a full-fledged
Neo-Kantian perspective by Ernst Cassirer (1923 [1955], 1944). Cassirer’s
theory of language, virtually unnoticed in the linguistic discourse of the last
half-century, has been a constant inspiration for my own work. As I show in
Dor (1999), Cassirer thinks about language—as well as the other symbolic
forms of myth, art, and science—as a system whose formal structures embody
a categorical worldview that is prior to experience in the Kantian sense, and
thus functions as a “means of ‘objectifying’ sensory impressions” (Cassirer
1955, p. 158).
All this background, then, is important for four reasons. First, it high-
lights the fact that the idea of linguistic relativity emerged from and within a
philosophical and ideological discourse—not as a hypothesis but as a convic-
tion. This was also true of Benjamin Lee Whorf: his famous Hopi examples
are not used to determine whether “we dissect nature along lines laid down
by our native language” (1956, p. 212), but to demonstrate what is necessarily
the case: “users of markedly different grammars are pointed by their gram-
mars toward different types of observations and different evaluations of ex-
ternally similar acts of observation, and hence are not equivalent as observers
but must arrive at somewhat different views of the world” (p. 221). For him,
linguistic relativity was a principle. Edward Sapir had a more nuanced view
of the matter, but for him, too, the influence of language on human thought
9780190256623-Dor.indb 89 03/06/15 6:18 PM

could not be doubted. What came to be known as the Sapir-Whorf Hypoth-

esis was not a hypothesis. Second, from the very beginning, two central and
interrelated issues remained implicit and vague: (a) are we talking about lan-
guage and thought as separate entities, so that the first determines the other,
or about language as the organ of thought, so that the two are inseparable?;
(b) what exactly is the nature of the causal relationship between language and
thought? How does it work? Is the relationship really unidirectional? Does
thought not influence language as well? Third, the background makes it clear
that the conviction of linguistic relativity implies a great attraction for what-
ever is variable between languages, while at the same time suggesting a new
dialectic outlook on the relationship between universality and diversity in
human nature—including language, thought, and culture: what is universal
about human nature, what makes it so different from other species, is the fact
that it is so variable. All languages determine their speakers’ worldviews, and
each does so in its own unique way. There is a straight line leading from this
understanding to Evans and Levinson’s (2009) refutation of the myth of lin-
guistic universals. Fourth, the background makes it clear why the principle
of relativity emerged as a deterministic conception: if we look at the world
through the lens of language, there is no way around the conclusion that lan-
guage determines what we see.
Whorf ’s principle of linguistic determinism was accepted with enthusi-
asm by the social sciences of the twentieth century, and it played an impor-
tant role in the eventual emergence of the linguistic (or cultural) turn in the
social sciences (cf. Geertz 1973)—with its view of culture as the social process
of meaning-construction and meaning-communication. More than anything
else, the principle gave social scientists a reason to believe that they could de-
cipher the logic of social thinking, in different societies, through the careful
interpretation of their discourse, linguistic and other. In the linguistic sciences
of the Chomskian era, on the other hand, the idea was rejected with contempt.
Pinker (1994) famously called it a conventional absurdity, “a statement that
goes against all common sense but that everyone believes because they dimly
recall having heard it somewhere and because it is so pregnant with impli-
cation” (p. 57). The rejection was rooted in the universalistic conception of
human cognition, in the Fodorian conception of the language of thought and
Chomsky’s conception of language as an autonomous formal system, and in
the conviction that language and thought, so conceived, are separate modules,
each with its own essence. Pinker, however, also made two valid points in his
critique. First, Whorf ’s empirical work was clumsy at best, and his examples
provided no real evidence for his claim. Second, and this is an argument usu-
ally ignored by Pinker’s rivals, our own experiences as speakers or writers
9780190256623-Dor.indb 90 03/06/15 6:18 PM

should teach us that language and thought are far from being so tightly cor-
related with each other:
We have all had the experience of uttering or writing a sentence, then

stopping and realizing that it wasn’t exactly what we meant to say. To
have that feeling, there has to be a “what we meant to say” that is different
from what we said. Sometimes it is not easy to find any words that prop-
erly convey a thought. When we hear or read, we usually remember the
gist, not the exact words, so there has to be such a thing as a gist that is not
the same as a bunch of words. (p. 57–58)
For a long time, there was no communication between the opposing camps. Re-
search on relativity with respect to color names began in the 1960s (Berlin and
Kay 1969), but the breakthrough had to wait for the 1990s. In the pioneering
works of Lucy (1992a,b), Choi and Bowerman (1991), Gumperz and Levinson
(1996), Levinson (1996) and Slobin (1996), and then Boroditsky (2003) and
many others, Whorf ’s principle was translated into a scientific question. Five
conceptions were involved in the translation: (a) the idea that language might be
the organ of thought was abandoned, and the issue was formulated in terms of a
relationship between two levels of cognitive representation; (b) Whorf ’s radical
view was replaced with the hypothesis that the structures of language influence
patterns of thought, rather than determine them; (c) the question was theoreti-
cally separated from two others, that is, whether human thought in general is in-
fluenced by the very existence of language as a semiotic system, and whether the
actual functioning of language in discourse influences thinking; (d) a research
strategy was developed that attempted to first look for correlations between
structural patterns in different languages and the habitual patterns of thought
manifested by their speakers, and then to investigate the development of these
correlations in children, in order to expose the causal relationship; (e) fascinating
experimental techniques were developed to capture patterns of thought inde-
pendently, in non-linguistic tasks, in order to avoid the circularity of Whorf ’s
argument (the very circularity that still haunts some of the social sciences when
they rely on speaking for social understanding.)
The accumulated results now show that speakers of different languages are
indeed sometimes inclined to think in ways that are correlated with their lan-
guages. Lucy (1992b) and Lucy and Gaskins (2001, 2003), for example, show
a correlation between the grammars of English and Yucatec Maya and the pat-
terns of their speakers’ thoughts in two related domains. First, the singular-
plural distinction is obligatory for almost all types of nouns in English, and
optional only for animate nouns in Yucatec Maya. Second, inanimate nouns
9780190256623-Dor.indb 91 03/06/15 6:18 PM

in Yucatec Maya refer to their material essence (like the English mass nouns
water and sugar), and in order to count them, the numerals must be accompa-
nied by numeral classifiers of different types. In picture recognition and se-
lection tasks, speakers of Yucatec Maya were less sensitive to the number of
inanimate entities in the pictures. When presented with object triads, two of
which were identical in material and two of which were identical in shape or
function (a plastic comb with a handle, a plastic comb without a handle, and a
wooden comb), Maya speakers saw greater similarity between the two plastic
combs; English speakers preferred to go with the shape. Levinson (1996) and
his colleagues asked speakers of Dutch and Tzeltal (another Mayan language)
to pick up physical objects and re-position them, in a certain way, somewhere
else. Dutch regularly refers to the positions of objects with respect to each
other in relative terms: to the right of, to the left of, in front of, behind, and so
on. Tzeltal, quite remarkably, works only with absolute terms (like our north
and south): uphill, downhill, and so on.1 Levinson and his colleagues found
that speakers of the two languages preferred to re-position the objects in space
in ways that correlated with the languages: Maya speakers re-positioned them
in the same absolute direction; English speakers placed them in the same po-
sition relative to themselves. What, then, is the source of such correlations?
The claim is that children use the linguistic categories they acquire in order
to carve the world they are getting accustomed to. Lucy and Gaskin (2001),
for example, show that Yucatec Maya children under the age of seven actually
prefer shape to substance, and only change their preference after they master
their language’s complex nominal system.
Quite obviously, not everybody is convinced (cf. Li and Gleitman 2002, Glei-
tman and Papafragou 2005, Bloom and Keil 2001). The critics, mostly coming
from the generative tradition, are especially emphatic in their objection to the
causal hypothesis: even if the correlations are there, they may be the result of
many possible causal scenarios. It may be the case, for example, that the cause and
effect are reversed: Maya children take seven long years to become experientially
acquainted with the importance of materials and substances in their physical
culture, and only manage to acquire the nominal system when they can already
understand it. The correlation, in other words, “might be due to cultural fac-
tors independent of language” (Bloom and Keil 2001, p. 356). It may also be the
result of the complex interaction between causal factors, including the physical
environment in which the experiments are performed. Li and Gleitman (2002)
replicated Levinson’s experiment, this time with speakers of the same languages,
but in different physical environments, and claimed to show that the preference
for absolute directions is strengthened in open spaces, where stable landmark cues
help situate the speaker in absolute space. In their rebuttal, Levinson et al. (2002)
9780190256623-Dor.indb 92 03/06/15 6:18 PM

showed significant weaknesses in Li and Gleitman’s experiments and conclusions,

but I think the general point is still valid: the empirical correlations are enough
to suggest a significant relationship between patterns of language and patterns of
thought, but the overall argument for direct influence of language on thought is
not strong enough.
For many researchers today, a reasonable compromise seems to revolve around
Slobin’s (1996, 2003) suggestion, inspired by Boas, to take a step back and dis-
tinguish between two possible types of influence: that of language on thinking
in general, and that of language on thinking-for-speaking. We may never know
to what extent language influences general thought, says Slobin, but we may be
quite confident that, at the very least, the obligatory grammatical components
of language force speakers to attend to particular aspects of experience, in order
to be able to fit what they wish to say into an utterance. Slobin (1996) reports
on a part of a larger project (Berman and Slobin 1987), in which speakers of five
languages were presented with a picture-story, without words, which they were
then asked to tell in their own words. The stories told by the speakers highlighted
properties of the picture-story, in correlation with the grammatical patterns of
the languages. In one of the pictures, a boy lies on his back after evidently falling
from a tree, and a swarm of bees is chasing a dog across the picture. Speakers of
English and Spanish, languages that mark progressive aspect, described the two
events as aspectually different: “the boy fell out . . . and the bees were flying after
the dog.” Speakers of German and Hebrew, languages that do not mark progres-
sive aspect, treated the events as aspectually similar: “the boy fell and the dog ran
away (from the bees).” Slobin’s point, then, is that the speakers of the different
languages would not necessarily see the scenes differently as long as they were
only looking, but once asked to describe them they had to obey their grammars
and thus pay attention to those properties of the experience highlighted by them.
Papafragou et al. (2008) add support to this claim. Tracking the eye-movements
of English and Greek speakers, who were asked to look at motion events and
either prepare a verbal description or just memorize them, Papafragou and her
colleagues found that speakers’ eyes rapidly focused on the event components
typically encoded in their native language, but only when asked to describe the
events.2
For Papafragou and others, all this implies that the general relativity hypoth-
esis is probably wrong: language does not really influence our cognition at large. I
think this is too hasty. The influence is there, but it is much more dialectic, varia-
ble, and complex than the philosophical heritage of the question allows us to see.
The empirical results of the type obtained by Lucy, Levinson, and many others
are extremely revealing, but they need to be re-interpreted within a different the-
oretical conception of language and its relationship to individual cognition.
9780190256623-Dor.indb 93 03/06/15 6:18 PM

5.2 Experiencing-for-Instructing
To re-think the relativity question, we must finally move away from the
philosophical-ideological conception of language as an interpretative scheme.
Language is a communication technology, not an epistemic organ. It does not
stand between us and the world: we do not experience through language. We
experience, we use the technology of language for instructive communica-
tion, and the two processes are entangled with each other in a variety of ways.
At the center of the entanglement lies the fact that language carries its own se-
mantic model of the world—the symbolic landscape, constructed through the
social process of experiential mutual-identification for language. Every speaker,
then, lives in two worldviews at the same time, one for knowing and one for
speaking—the first privately experienced, the other socially constructed. The
question is about the causal relationships between these two worldviews.
The first dividend of this move is the fact that it positions the question of in-
fluence between two levels of meaning, instead of its usual positioning between
a level of structure and a level of meaning. It thus resolves an implicit difficulty
in the traditional formulation: even if it turns out that the influence is there, the
two representational levels are still very different in essence; why should it be the
case that the formal structures of language influence our worldview? A causal
relationship between two levels of meaning makes much more common sense.
The second dividend is that it widens the scope of the question. The target of
influence is not just thought, but experiencing in general: how do the meanings
that we use for linguistic communication influence our private experiences of
the word? Third, it corrects a misperception implicit in the formulation of many
of the research projects, which equate cognitive variability with linguistic influ-
ence and cognitive universality with innateness. This leaves out everything in
the individual’s experiential life that serves as non-linguistic input for learning.
Children and adults do not either know a priori or learn through language. They
learn by doing, touching, playing, watching, and listening. Most importantly,
they learn about social life not only through language, but also through active
participation (always privately perceived) in non-linguistic social activities (this
is the domain neglected by Vygotsky).
Fourth, the proposed perspective relativizes the entire question: different in-
dividuals experience the world differently and use language differently, which
means that we should expect variable levels and types of influence within the
community. In this sense, the most significant fact about the entire set of ex-
perimental results accumulated in the field is that they are all statistically
significant—not absolute. Lucy and Gaskins (2003), for example, report triad
experiments in which English speakers preferred material to shape 23% of the
9780190256623-Dor.indb 94 03/06/15 6:18 PM

time, while Yucatec Maya chose shape 39% of the time. In Levinson (1996),
95% of the Dutch speakers went with relative directions, and 75% of the Tzeltal
speakers chose the absolute ones. Not all subjects are influenced in the same way.
Most importantly, the new perspective turns Slobin’s thinking-for-
speaking—re-formulated as experiencing-for-instructing—from a compromise
position into the most important key to the entire question. Every technology
ever invented forces its users to experience in specific ways in order to make the
technology work properly. When we drive, we have to direct our visual and au-
ditory attention to aspects of the physical environment that the person sitting
next to us may safely ignore. When we play a musical instrument, we have to de-
velop that mysterious capacity of experiencing with our ears and fingers together,
which we seldom need anywhere else. Experiencing-for-instructing is exactly the
same phenomenon. Speakers must be able to pay attention to those components
and properties of their experiences that are required by their language’s norms,
those elements that were highlighted and signified by their communities as ex-
periential commonalities, for use in instructive communication. Learning to
experience-for-instructing is simply learning to use the technology of language.
Putting things this way, one thing becomes immediately clear: every tech-
nology that we know exposes a unique pattern of variability between individu-
als, and language is no different. The ability to describe a picture in words, for
example, is far from being equally shared by all speakers. Other things being
equal, then, we should expect significant differences in the general capacity to
experience-for-instructing—all the way from great difficulty to masterly achieve-
ment. On the one end of the continuum, we find individuals who find it very
difficult, or impossible, to describe a simple picture in words. On the other end,
we sometimes read a paragraph from a great author and realize: we have had this
experience but we could never describe it so well. We thus acknowledge that the
author is capable of experiencing-for-instructing in ways we that we cannot. All
across the continuum, we experience Pinker’s frustration, trying to find ways to
translate our experiences into words with varying levels of success. Moreover, be-
cause individuals experience differently, they should be expected to experience-
for-instructing differently in terms of contents too. Speakers, for example, can
only use a linguistic construction to the extent that they manage to experience-
for-instructing it, which means that we should expect them to avoid it if they do
not. Different individuals thus probably avoid different components of what the
technology offers, and feel more confident with others.
With all this, we may finally re-formulate the general relativity question in
the following way: to what extent, and under which conditions, does the proc-
ess of experiencing-for-speaking influence the way different individuals expe-
rience in general? Shen and Gil’s (to appear) experiments on hybrid perception
9780190256623-Dor.indb 95 03/06/15 6:18 PM

make it easy to demonstrate the new question. They presented speakers of dif-
ferent languages with drawings of imaginary hybrids they had never seen before.
The hybrids were made, half and half, from images of humans, animals, plants,
and inanimate objects (like a hammer and a gun). These four types of entity—
human, animal, plant, and inanimate object—constitute the four levels of what
is known as animacy hierarchy, which turns out to play an important role in
many languages. In one experimental setting, the subjects were asked to describe
the hybrids. In another setting, the subjects were asked to make non-linguistic
judgments about them—allocate the drawings, for example, to different sets. A
statistically significant majority of the subjects took the animacy hierarchy into
account in their descriptions. They described one of the hybrids, for example,
as “a person with an upper body of a fish,” not as “a fish with human legs.” This,
however, was not how they behaved when asked to perform non-linguistic tasks.
There, the hierarchy did not seem to play a significant role. The remarkable fact
about these experiments, then, is that the subjects were inexperienced with re-
spect to these objects. They had never seen them before. What would happen,
then, if the subjects kept on experiencing hybrids, then got into the process of
their mutual-identification and naming, and then began to spend time talking
about them? Would they eventually zoom in on the animacy hierarchy in their
non-linguistic categorization as well? This, as opposed to the original question,
is no longer a question that awaits a simple yes-or-no answer. The extent of influ-
ence depends on a very long list of variables.
5.3 The Vector Space of Influence

Let us think about the entire experiential world of any individual, at any
moment, as a vector space of experiential and linguistic influences, pushing and
pulling in all directions. As a first approximation, let me suggest that the accu-
mulated impact of these influences depends on two types of variables—the ex-
periential profile of the individual standing at the center of the vector space, and
the strength, direction, and temporal order of the vectors themselves.
The first type of variable has to do with the susceptibility of the individual to
the influence of language vis-à-vis other types of influences. Individuals have dif-
ferent experiential styles, and we may expect them to be different in their ability
and will to allow the socially constructed view of the world reflected by language
to interfere with their private experiences, and in their ability and will to expe-
rience the world away from language. Other things being equal, we may expect
this variability to correlate with the amount of time and energy the individual
spends with language, both as a child and as an adult. The more he or she looks at
9780190256623-Dor.indb 96 03/06/15 6:18 PM

the world in order to talk about it, and the more he or she learns about the world
by listening to other people talking about it, the more influence the worldview
of language would have over the construction of his or her experiential world.
This is exactly like saying that professional drivers are more inclined than the
rest of us to look at the road through the lens of driving even when they are not
actually sitting at the wheel. There is a possible venue for empirical research here,
with a clear prediction: crossing the results of influence-experiments with data
concerning individual linguistic habits and proficiencies, we should expect a pos-
itive correlation. Children who spend a lot of time reading, for example, should
be more susceptible to the influence of language.
Individuals may also differ in terms of their trust in the carriers of the lin-
guistic worldview, and in the general truthfulness of language. There is an entire
issue here that is totally ignored in the literature. On the one side of the con-
tinuum, when speakers are convinced that they speak a sacred tongue, reflecting
godly truth, they would probably be very deeply affected by their language. (In
this sense, again, the original deterministic view may carry with it some reli-
gious baggage.) On the other side, when speakers realize that the language they
have to use is an artificial construction built and enforced for the purposes of
thought-policing, they very often learn to experience-for-instructing (they have
no choice), but in their own world of experience, they nevertheless keep them-
selves away from the worldview reflected by the language. This was the attitude
adopted by many speakers toward the newspeak dictates of the great totalitar-
ian regimes of the twentieth century. Even the extent to which we reflexively
understand the effects of language on our worldview changes something in the
effect: we may live by metaphors, but their effect is probably reduced after we
read Metaphors We Live By (Lakoff and Johnson 1980).
The second type of variables determining the overall effect of language on
experience has to do with the vectors of influence themselves: (a) the linguis-
tic vector and the experiential ones may be different in strength: in the extreme
scenarios, the individual may be exposed to a certain realm of meaning only
through language or only through experience. In between, some of the vectors
may be stronger or weaker than the others; (b) the vectors may appear in different
temporal orders: if the experience is already there, language may strengthen or
weaken it; if language arrives first, it may effectively direct the individual toward
or away from the experience, but this is far from trivial: a linguistic construction
cannot direct the individual anywhere if he or she cannot assign an experiential
interpretation of some sort to the construction. Pieces of language that the in-
dividual cannot understand have no effect. Because of this, individuals must be
already close enough to the relevant way of experiencing for language to direct
them there; (c) the vectors may arrange themselves with respect to each other in
9780190256623-Dor.indb 97 03/06/15 6:18 PM

different configurations: the linguistic vector may direct the individual toward a
way of experiencing that he or she is attracted to anyway by experience; different
experiential influences may direct the individual in different directions, one of
which is correlated with the categorization of language; and the linguistic cate-
gorization may fully or partially conflict with the experience of the individual.
Let us look at these configurations one at a time. The first is the most straight-
forward: the individual has to learn to experience-for-instructing in a way that
he or she finds experientially natural, and the two vectors converge to produce
a strong effect. Language serves two functions in this configuration: first, as
just another vector, it strengthens the inclination toward the relevant way of
experiencing. Second, and much more importantly, it provides the individual
with the sense of social affirmation that comes with the experience of mutual-
identification. This is the best-of-all-worlds scenario. Language and experience
emerge in the individual’s mind genuinely correlated, and are thus relatively
transparently translatable into each other. With respect to meanings involved in
such situation, we may expect the speaker to be able to use language for instruc-
tive communication easily and efficiently. Consider, for example, the system of
absolute directions in Tzeltal. The notions of uphill and downhill have real and
stable correlates in the physical experience of the speakers. This is what the ter-
rain they live in feels like. It seems reasonable to assume that for many of them,
when they were children, the linguistic terms strengthened an experiential in-
clination that also emerged on its own. Importantly, however, this was not nec-
essarily true of all of them. There were probably those who had a harder time
reading the terrain by themselves, and these probably benefited more from lan-
guage: it actually directed them toward a way of experiencing that they would
find it harder to zoom in on their own (and there were probably also those who
found it difficult to find their way around, with language or without it). Even in
such situations, then, language probably played slightly different roles for differ-
ent speakers, with the end result of a statistically significant pattern of language-
experience correlation in the entire group. What both Li and Gleitman (2002)
and Levinson et al. (2002) show together, then, is that these patterns of influence
are always sensitive to nuanced changes in the arrangement of the vector space.
In the second configuration, different experiential influences direct the indi-
vidual in different directions, and language helps (or forces) him or her to choose
between them. Consider, for example, Choi and Bowerman’s (1991) comparison
of English- and Korean-speaking children (see also Bowerman and Choi 2003).
Whereas English distinguishes between being on-something and being in-
something, Korean distinguishes between maintaining a tight-fit and a loose-fit
with something. For example, in English, a CD is in its pocket like an apple is in
a bowl, and a cover is on a pot like a cup is on the table. In Korean, the disk and
9780190256623-Dor.indb 98 03/06/15 6:18 PM

the cover are marked as being in tight-fit to the pocket and pot, whereas the apple
and the cup are marked as being loosely placed with respect to the bowl and the
table. Choi and Bowerman show that Korean- and English-speaking children
do indeed pay more attention to the categorization of their own language. What
seems to be happening is this: both ways of categorization are independently
supported, to this extent or the other, by experience. Children, as we all know,
are fascinated by both. Quite obviously, the fact that language highlights one of
them at the expense of the other sends children a very clear message: one way of
looking at the world is more important, more relevant than the other, because
the adults mention it all the time. The children must learn to experience the adult
categorization for instructing, which means that most of them would probably
choose to focus on the worldview suggested by their language: it is there, it is
valued, it is mutually identified and thus communicable. And yet, even here,
it is important to see that the effects should be expected to be variable. Some
English-speaking children would still be most impressed by the aesthetics of
tight and loose fit, and some Korean-speaking children would still be most im-
pressed by the distinction between being on-something and being in-something.
The third configuration is more dramatic: the linguistic vector and the rele-
vant experiential vectors (for a certain individual) point in opposite directions.
Boroditsky’s (2003) work on grammatical gender is the only research project I
am aware of which may be interpreted as an attempt to tackle such opposition.
The main fact here is that many languages do not just assign grammatical gender
to animate entities, but also to inanimate ones. On the symbolic landscapes of
these languages, entities such as chairs, knives, and apples have gender just like
humans, dogs, and cats. If the children acquiring these languages have any expe-
riential understanding of the difference between males and females, they simply
must know that chairs and knives do not really have gender. When they wish to
talk about these objects, however, their language forces them to relate to them as
if they did. So, does this have an effect on the way they actually perceive the ob-
jects? Boroditsky shows that it sometimes does: speakers of Spanish and German
were asked to mention three adjectives that came to their mind for describing
different objects, and the sets of adjectives, taken together, amounted to descrip-
tions of characteristically masculine or feminine entities, in correlation with the
grammatical gender assigned to these objects in both languages. The word for
key, for example, is masculine in German and feminine in Spanish—and most of
the descriptions of keys provided by speakers of the languages conformed with
this distinction. In this case, language seems to have at least a partial effect on
experience when experience itself contradicts it. This is a very important finding:
children are forced by language to memorize the gender affiliation of entities
which are not experienced as gendered, and they probably find this easier to do,
9780190256623-Dor.indb 99 03/06/15 6:18 PM

less arbitrary to grasp, by imagining that the entities do have gender. It is crucial
to remember, however, that the effect of all this is much more limited than it
seems at first sight: speakers of German and Spanish do not end up believing that
keys have gender. The tension between both worldviews remains.
In the case of inanimate gender, the tension does not seem to have serious
communicative, experiential, or practical implications. It is just there. In many
other cases, the tension positions itself at the very center of our mental and emo-
tional lives. The discrepancy between the way things are experienced, externally
and internally, and the way they are described (especially the way they are habitu-
ally described) is a constant factor in human life. It has attracted the attention of
post-modern thinkers, determined to deconstruct the linguistic worldview (see
Lyotard’s 1983 [1988] notion of the differend). It also plays a crucial role in the
work of clinical psychologists.
Conflicts between linguistic and experiential influences may thus be par-
tially resolved by an experiential shift, and they may remain unresolved, thus
constituting a constant obstacle both for communication and for the under-
standing of the world. There is, however, another way: linguistic innovation.
Language and experience may send opposing messages to the individual, and he
or she may actually try to use his or her experience to change language. This is
the origin of all instances of mutual-identification for language, and it is also the
source of many grammatical innovations. Modern Hebrew, for example, plural-
izes the word woman with the masculine marker im and the word father with the
feminine marker ot. Children very often refuse to use these plural forms. They
re-arrange the morphology, and say aba’ im and ishot. This indicates that children
are not just the most natural candidates for deep influence—because they are
inexperienced and trusting, flexible and eager to speak—they are also capable of
resisting the influence of language where it conflicts with their own experiences.
What does all this amount to? To begin with, the influence of language
on its speakers resembles the influence of any other technology on its users:
it depends on the extent to which individuals use the technology, depend on
it, feel comfortable with it, and trust it, and it depends on the specific proper-
ties of the technology and their functionality for the individual. Other things
being equal, those who feel more comfortable with the worldview of language
will find it easier to communicate. Second, in order for a linguistic category
to participate in the shaping of the individual’s worldview, the individual has
to be able to identify it with something at the relevant experiential level. So,
language may strengthen an experiential tendency, and it may help the indi-
vidual choose between different experiential tendencies, but when the relevant
category does not connect to anything in the individual’s experiential world
there is no effect. No speaker ever uses everything that language has to offer.
9780190256623-Dor.indb 100 03/06/15 6:18 PM

Third, even where a linguistic category does influence the experiential world-
views of some individuals, there is no reason to assume that it influences all of
them in the same way. Different speakers assign it with different experiential
interpretations.
This, however, is only one side of the coin. Language and experience do not
maintain a unidirectional, linear relationship between them. The relationship
is bi-directional, dialectic, highly dynamic, and cyclical. At any given moment
in the life of a linguistic community, different individuals and groups of in-
dividuals find themselves occupying different positions in the cycle: there are
always those who find that language, as it is, lacks something they need in
order to express their intents—or in order to understand their interlocutors.
They invent new ways of speaking: a word, a phrase, a metaphor, a politeness
marker, a pattern of argument. The inventors are not necessarily aware of their
special status: they are driven by the need to communicate. Other speakers
may adopt the inventions, because they echo their own experiences, or because
they wish to sound as if they do, or because they reveal something they did
not see before, or simply because they have to. Social power plays a crucial role
in all this. Yet others may reject the inventions or ignore them, because they
go against their own experiences, or because they wish to distance themselves
from such experiences, or from the inventors, or simply because they do not
understand them. In the process of negotiation that ensues, some of the inven-
tions may be gradually accepted into some version of the language and begin to
influence more and more individuals—always to variable degrees and in vari-
able ways.
5.4 Back to Determinism

Perhaps surprisingly, the re-interpretation of linguistic relativity as the dialectic
relationship between experiencing-for-instructing and experiencing at large im-
plies a new outlook on the original question in its nineteenth-century formula-
tion: the question of the influence of language on thought. It should probably be
answered in the affirmative. The issue was heavily biased by the fact that it had
mainly been raised by philosophers, for whom the notion of thinking was identi-
cal with philosophical thinking—that mode of experiencing that is only made
possible by such specifically linguistic activities as argumentation and persua-
sion, negotiation and explanation, affirmation and refutation. These activities
cannot be performed without language, and in a very real sense they are negotia-
tions about language, collective attempts to rationalize the symbolic landscape.
Philosophers live in a world of experiencing-for-philosophical-speaking, which
implies that the meanings of language should have an especially deep influence
9780190256623-Dor.indb 101 03/06/15 6:18 PM

on their thinking. This, however, is a particular case of Bourdieu’s (2000) scho-

lastic bias: most humans are not philosophers.
Generalizing this point, we may predict that the influence of language on
experiencing should be at its peak exactly when experiencing is not possible in
the absence of language. This is the essence of the Vygotskian conception. A dra-
matic demonstration of this has recently been made in the domain of numeric
cognition. Based on his father’s work on Pirahã (Everett 2012), Everett (2013)
argues as follows: Pirahã lacks a numerical system, and only contains hói for
“small size or amount,” hoí for “somewhat larger size or amount,” and baágiso for
“cause to come together” or “many.” Speakers of the language were asked to per-
form tasks designed to test their recognition of numeral correspondences, such
as to produce an exact mimic of a series of log-stomping or rowing movements
and to match a uniformly spaced line of spools of thread with a line made of an
equal number of rubber balloons. Everett reports a clear result: “The speakers of
this anumeric language struggle with the recognition of exact correspondences
between numerosities over three” (p. 96). He then interprets the result in the
regular way: the anumericity of the language causes the cognitive difficulty. This
misses out on what seems to be the main issue. All the other domains of experi-
ence investigated in the literature—space and time, gender and color—are open
to independent experiencing away from language. Experiencing higher numbers
of things is only possible while counting. The Pirahã do not count. They manage
wonderfully in their environment without it. There is no reason for them to
either invent or adopt a numerical system—or develop numerical cognition.
Language, then, determines experiencing in those cases where experiencing is
not possible without it.
There is, however, another dimension to all this. According to the theory,
language should also reach the peak of its influence on private experience where
it manages to replace direct experiencing—and build an entire experiential world
built purely on imagination. This is an essential key to the understanding of ide-
ology and propaganda: they make their impact not by directly configuring lived
experience, but through the construction of an entire imaginary world—which
then reflects back on experience. The imagination of Heavenly entities; the
myths of past heroes; the imagined communities of nation-states; the imaginary
worlds of commercials; the imaginary body images of the fashion industry—all
these are variations on the very same theme. This conception carries implications
for the political critique of modern ideology (cf. Eagleton 1991), some of which
I discuss in my own work on the construction of political hegemony (Dor 2004,
2005b). I will leave their further elaboration for future research.
9780190256623-Dor.indb 102 03/06/15 6:18 PM

6 P r o d u c ti o n a n d C o m p r e h e n s i o n
In chapter 3, we looked at linguistic communication as a conversion

process: following the conventions of the protocol, speakers translate
experiential intents into semantic instructions for imagination, then
translate these into formal utterances and articulate them; listen-
ers process the strings of speech they receive into formal utterances,
translate these into semantic instructions for imagination, which ac-
tivate their imaginations to produce an experiential interpretation. In
this chapter, I would like to delve deeper into this process and high-
light some of its constitutive properties.
The processes of production and comprehension that interlocu-
tors go through in the course of communication are obviously cogni-
tive, but the theory of language as a social communication technology
implies a crucial distinction between two issues that are perceived
as one and the same as long as language and its usage are thought of
as merely cognitive: (a) the ways in which speakers and listeners ac-
tually process language in real time and (b) the ways in which their
language, as a social entity, prescribes its own processing. As a first ap-
proximation, the distinction may be understood by comparison with
other technologies. When we buy a fax, a TV, or a computer, we also
receive a booklet of instructions, an operational manual. The manual
prescribes the way the technology should and should not be used.
Characteristically, the ways we eventually use the technology are
quite different from the prescriptions of the booklet—on two levels.
First, we very often do things that the booklet forbids and neglect to
do things it requires. Second, as we accumulate experience with the
technology, by ourselves or with others, we adopt our own procedures
and habits. The communication protocol, then, should be thought of
as the equivalent of these booklets: the set of procedures prescribed
for production and comprehension by the community, through the
never-ending process of the mutual-identification of communication
experiences. In actual speech, experienced speakers develop habits of
communication that do not necessarily follow the protocol: the pro-
cesses of production and comprehension, for example, are prescribed
9780190256623-Dor.indb 103 03/06/15 6:18 PM

as serial, but speakers and listeners make shortcuts, plan ahead, use parallel proc-
essing and so on.
The comparison, however, should not be stretched too far. The communi-
cation protocol does more than prescribe the ways of communication. Differ-
ent individuals adopt different ways of speaking and listening, but to be able to
communicate across the experiential gaps between them they have to produce
similar-enough outputs—to converge on the same outside attractor. The pro-
tocol socially defines the challenge that speakers and listeners have to meet for
their communication to succeed. This view implies a new conception of the rela-
tionship between competence and performance. As we have seen in chapter 3, the
speaker’s competence (part of the speaker’s proficiency) is his or her internalized
knowledge of the mutually identified norms of the language, at the levels of the
protocol and symbolic landscape. The speaker’s performance is the way he or
she actually uses language in the course of communication. As opposed to the
usage-based view of language (Croft 1991, Givón 1995, Bybee 2006, Goldberg
2006), according to which the distinction is unwarranted because knowledge
of language is based on the patterns of performance, the theory of language as a
communication technology re-instates the distinction along the lines originally
proposed by de Saussure’s (1916) notions of langue and parole: performance is
parole, and competence is knowledge of langue. Obviously, I have no quarrel
with the idea that acquiring competence is also a matter of learning from usage,
but the contents of competence, the normative rules and regulations of lan-
guage, do not simply emerge from usage: they emerge from the ongoing social
activity of mutual-identification for language. Competence, then, is not a matter
of innate knowledge. It reflects the depth of socialization of the individual as a
speaker of his or her language. As such, it shows itself most clearly in speakers’
grammaticality judgments, because there they concentrate on the normativity
of the sentence, not on the attempt to produce or interpret it. What this means
is that linguistics should not choose whether to concentrate on grammaticality
judgments (as the generative tradition does) or only look at patterns of perfor-
mance (in line with usage-based tradition), but take as its object of inquiry the
gaps between the two: where do speakers’ judgments match their performance
and where do they not? What are the patterns of deviation? How should they
be explained?
All this re-arranges the relationship between linguistic theory and psycho-
linguistic research. Psycholinguistics investigates everything about language-
related cognition, including actual processing, the capacities and fluencies that
make it possible, and the way they emerge in development. Linguistic theory,
no longer a cognitive science, investigates the technology itself—and as a major
part of this effort it attempts to expose the socially constructed attractor around
9780190256623-Dor.indb 104 03/06/15 6:18 PM

Production and Comprehension • 105
which speakers and listeners try to converge. This re-arrangement, as I will claim
below, rationalizes the relationship between the two disciplines.
In the first part of this chapter, then, I will present a more detailed hypothet-
ical description of the processes as prescribed by the protocol—for one simple
sentence in English. As always, the technical discussion should be read as a dem-
onstration, not as a full analysis. I will have very little to say about the last stages
of production and the first stages of comprehension, from form to sound and
vice versa. I will concentrate on the translation between the two levels of mean-
ing that allows for the unique functionality of language: from the experiential
intent to the message, the ordered set of instructions of imagination, and from
the message to the imagined interpretation. On the speaker’s side, I will sug-
gest that language does not simply let speakers translate their intentions into
messages at will. The process is serially prescribed by the normative procedures
of the protocol. This is so because message construction is the single most chal-
lenging stage in language production: the speaker has to collect discrete signs
from the symbolic landscape and arrange them in a way that will both (a) reflect
his or her analogue and richly detailed experiential intention and (b) allow the
listener to accurately construct an imagined experience, on the basis of the mes-
sage, that would reflect the speaker’s original experiential intent. This is the key
to the function of language, but it is also the key to its fragility. The procedures
of the protocol are there to make sure the speaker and listener do their parts in
mutually identified ways, and thus maximize the chance of communicative suc-
cess. On the listener’s side, I will suggest a re-framing of the process of interpre-
tation: the semantic reconstruction of the message is only a temporary stage on
the way to comprehension, which takes place in two pragmatic stages: first, the
imagination of the speaker’s intent on the basis of message’s instructions; then,
the inference of implicature.
In the second part of the chapter, I will try to demonstrate how the picture
is dialectically related to some of the major patterns of actual usage as they are
analyzed in the psycholinguistic literature. I will concentrate on one side of the
process, that of production, which has received much less attention in the liter-
ature than comprehension. In the next chapter, I will show how the prescriptive
conventions of the protocol, as they are described here, converge to produce the
observable syntactic patterns of language.
6.1 Message Construction: From Essentials to Details

Message construction, I would like to suggest, is prescribed by the protocol as
a serial process, in which the protocol iteratively presents speakers with forms
9780190256623-Dor.indb 105 03/06/15 6:18 PM

that include open slots to fill in with signs from the symbolic landscape. Once
the first form is filled in and submitted, the lexical choices made in it determine
the blank slots in the next form, and so on and so forth, until the speaker decides
that enough material has been collected from the symbolic landscape. The itera-
tive logic of the forms leads the speaker in the process of experience-to-message
construction—along a path that goes from essentials to details.
The resemblance of this iterative process to the way we fill in official forms
on the Internet is not accidental. Language is formal in exactly this sense, that
it provides the speaker with a closed set of pre-determined forms of expression,
and forces him or her to translate his or her intent into these forms. This under-
standing of formalism follows Sapir’s (1921) view, which deserves to be quoted
in its entirety:
Language is in many respects . . . unreasonable and stubborn about its

classifications. . . . It must have its perfectly exclusive pigeon-holes and
will tolerate no flying vagrants. Any concept that asks for expression must
submit to the classificatory rules of the game, just as there are statistical
surveys in which even the most convinced atheist must perforce be la-
beled Catholic, Protestant or Jew or get no hearing. It is almost as though
at some period in the past the unconscious mind of the race had made a
hasty inventory of experience, committed itself to a premature classifica-
tion that allowed of no revision, and saddled the inheritors of its language
with a science that they no longer quite believed in nor had the strength
to overthrow. Dogma, rigidly prescribed by tradition, stiffens into for-
malism. (p. 93)
In the first stage of message construction, which I will call decide, the speaker
is required to make the most radical act of translation: to abstract away from the
analogue complexities of his or her experiential intent, and translate it, reduce
it, into a stereotyped kernel of meaning—the message kernel—that the language
(and thus the listener) will be able to identify. The message kernel is a convention:
different languages (different communities of speakers) present speakers, in dif-
ferent contexts, with different varieties of message-kernels. For the purposes of
our demonstration, let us concentrate on one foundational variety that seems to
be very widespread. In this variety, the message kernel is a meaning unit of the
type—SPEECH-ACT (topic-entity) (eventuality).
This message-kernel should be thought of as a conventionalized demand, for-
mulated by the community and addressed to the speaker, to specify, from the
very beginning, what (or whom) he or she is going to talk about; what happened
(or happens, or will happen); and what he or she intends to say about it—tell a
9780190256623-Dor.indb 106 03/06/15 6:18 PM

story, ask a question, make a prediction, and so on. It is a social requirement,

which came to be conventionalized through a long and difficult process in which
listeners had to stop the people speaking to them and insist, time and again:
“Before you go on, could you please try to arrange the story in your mind and
begin with the essentials?”
The message kernel requires the speaker to decide, from the very outset,
whether he or she wants to Assert(John)(dance); Ask-whether(Bill)(tall);
Promise(child)(eat), and so on. These meaning-triplets are as of yet untranslata-
ble into English sentences: they lack an indication of tense, for example. Each of
them is a first step toward the construction of a message, a first step toward the
socialization of the speaker’s intent. The speaker is only allowed to select a speech-
act out of the pre-established set of speech-acts offered by the conventions of the
language, and is only allowed to select an entity and an eventuality, as the objects
of the speech-act, out of the pre-established set of entities and eventualities on the
symbolic landscape. The entire set of possible message-kernels, everything that the
language allows for, is already specified prior to the act of communication.
The challenge of decide is probably the toughest challenge we face as speak-
ers. Imagine you come back home from work and find that a very precious object
of yours (an expensive vase, for instance) has been broken to pieces. You have
very good reasons to suspect a certain person, but you are not quite sure. He is
still there, but he is not saying anything. You are sad, angry, confused. Now, im-
agine you decide to approach him; what are you going to say? Are you going to
ASSERT(vase)(break)? BLAME(you)(break)? ASK-whether(you)(break)? ASK-
why(you)(break)? ASSERT(I)(angry)? THREATEN(I)(hit)? DEMAND(you)
(compensate)? On the one hand, all of these are probably going through your
mind at the same time: they are all there because the will to communicate that
emerges in your mind is as complex as the experience you are going through.
On the other hand, each of the above options, once highlighted and isolated,
carries with it an entire array of discursive implications: assume you ASK-
why(you)(break), and then discover that it was not him after all; or assume that
you ASSERT(vase)(break), and the person reacts with what seems to you like a
fake expression of surprise. What are you going to do then? To be sure, language
allows you to formulate a series of utterances, each of which is based on one of
the above options. Language, however, also obligates you to formulate them one
after the other. The question, then, still remains: which is the one you are going
to formulate now?
In the second stage of message construction, which I will call specify, the
speaker is presented with a closed set of extensions—obligatory and optional—
to the eventuality chosen for the message-kernel. He or she is required to fill in
the obligatory extensions, and is allowed to fill in the optional ones. This stage
9780190256623-Dor.indb 107 03/06/15 6:18 PM

may be thought of as the second step in the collective attempt to enforce a certain
order on the speakers of the language: “Now that we know what you are talking
about, can you please tell us, in a way that we will understand, what happened?”
The end result of this stage is what I will call the basic message.
A speaker of English who selects a message kernel such as ASSERT(boy)
(break), for example, is presented at the second stage with the following set of
obligatory and voluntary extensions to the eventuality break (in technical terms,
this means that the protocol should have access to the signified break on the sym-
bolic landscape; this will become important in the next chapter).
Assume, then, that our speaker fills in the blank form (Figure 6.1), as shown
in Figure 6.2.
* patient _____
agent _____
instrument _____
goal _____
Extensions of event-type (break): reason _____
manner _____
* tense _____
time _____
place _____
figure 6.1
* patient vase
agent boy (topic)
instrument
goal
Extensions of event-type (break): reason
manner
* tense past
time
place
figure 6.2
9780190256623-Dor.indb 108 03/06/15 6:18 PM

In the third stage, complete, the speaker is required to further develop the
basic message into what I will call the complete message. Each of the content selec-
tions made by the speaker in the former stage now becomes the base for another
round of extensions, some of which are obligatory, requiring the speaker’s speci-
fication. The extensions are filled in, and, if necessary, prompt another round,
and so on, until all obligatory spaces, and all the spaces chosen by the speaker, are
specified according to convention. This is the third step in the collective effort:
“Now that we know what you are talking about, and what happened, you may
add some detail of the type we can understand.” The process may, for example,
look like Figure 6.3.
Having specified the vase as the patient, the speaker is presented with a set of
extensions to the patient-vase, a set of conventionalized questions, some of which
he or she must, some of which he or she may answer. Is the patient-vase a single
entity or a plurality of entities (a vase or vases)? Does it have a property that needs
to be mentioned? Did it (or does it, or will it) participate in another eventuality
that has to be reported? Then, the choice of singularity leads to another obliga-
tory extension: is this singular entity already mutually known by the speaker and
the listener (is it definite)?
The end-product of these three stages, the complete message (Figure 6.4), is
thus a formal representation of a selected set of interconnected signs from the
symbolic landscape, hierarchically arranged by order of selection. This semantic
hierarchy, as will be shown in the next chapter, determines most of what we reg-
ularly think of as hierarchical syntax.
Before we move on to the construction of the utterance, let me make two
additional comments. First, the obligatory status of some of the extensions
we have looked at is normative. Obligatory extensions are there, different sets
decide specify complete

sing/plural sing def/indef inite indef
patient vase
property
assert
other event
break
boy sing/plural sing def/indefinite def
agent boy property tall
other event
time frame past
figure 6.3
9780190256623-Dor.indb 109 03/06/15 6:18 PM

complete message
assert
speech act indefinite
definiteness
*
singular
* singularity
vase
* patient
break
eventuality
past
* tense
boy boy
topic agent
tall
property
singular
* singularity
definite
definiteness
*
figure 6.4
in different languages, because the communities within which the languages

evolved decided to obligate their speakers to include certain meaning compo-
nents in their messages, whether or not these components featured prominently
in their communicative intents. In our example, the speaker who is working
his or her way up from the message kernel ASSERT(boy)(break) toward the
complete message may not be particularly interested in specifying whether
the boy broke a few vases or a single one. The community of English speakers,
however, has made a collective decision, through a certain period of its cul-
tural evolution, always to obligate speakers to specify whether a spoken object
is singular or plural. The members of the community decided that of all the
categorical distinctions available on the symbolic landscape, this was a distinc-
tion so important, so crucial, that it had to be a necessary component of every
possible message. Our speaker has no choice, if he or she wants to maintain the
position of a normative speaker in the community (and thus be regarded as a
reliable interlocutor), but to comply with the obligation, and provide the listen-
ers with more information about the broken object (or objects) than he or she
might consider essential to the message. When a community turns a categori-
cal distinction from the symbolic landscape into an obligatory component of
every possible message, the members of the community, in their role as listen-
ers, impose a minimal specificity requirement on themselves as speakers. When
speakers are presented with such requirements, they have to teach themselves
9780190256623-Dor.indb 110 03/06/15 6:18 PM

to pay attention to certain aspects of their experiences that they might not be
interested in as such, in order to be able to provide their listeners with the re-
quired information.
The obligatory extensions, then, do not necessarily reflect the core categories
of our experiences. They reflect those categories of our experiences that are most
crucial for instructive communication to succeed. The listener has to construct
an imagined experience in his or her mind, based on the instructions sent by the
speaker. To be able to do this, the listener has to anchor the imagined experience
in some certainties: it happened in the past, it was an event of breaking, there
was just one vase. Such anchoring is simply not required in experiencing. This is
exactly why obligatory extensions provide us with the best demonstrations of the
real-time influence of language on the way we experience-for-instructing.
The second comment is this: some of the operations in the third process of
message construction, from the basis to the complete message, allow for recursion:
the slot other event, for example, lets the speaker begin another round that will
eventually result in an embedded sentence. For scholars in the generative tradition,
recursion has always been the pinnacle of language, the single human universal—
according to Hauser, Chomsky, and Fitch (2002)—that separates us and our lan-
guage from all the other species and their communication systems: it allows for
the generation of infinitely complex sentences from a finite set of building blocks.
One problem with this conception is that recursion is not universally shared by all
languages (Evans and Levinson 2009): I will get back to this in chapter 8. At the
moment, let me stress another crucial point: the idea that recursion generates in-
finity is valid as a logical statement, but this logical statement is only applicable to
mathematics. As far as language is concerned, it is nonsensical: nothing in human
life is infinite. Recursion has been shown to play an important role not just in lan-
guage, but also, for example, in toolmaking (Greenfield 1991), but no one would
suggest that the capacity to re-apply manual operations to their own outputs—
and then to do it again and again—allows for the making of infinitely complex
tools. The recursive operations in message construction (in those languages that
have them) do allow for the production of very complex messages, much more
complex than the messages produced in non-recursive languages, but this is all.
6.2 Utterance Construction and Execution

The construction of the utterance on the basis of the complete message is also pre-
scribed as a serial process with three stages. The first stage, signify, comprises two
processes. The signifieds of the complete message, hierarchically ordered as they
are on the basis of their order of selection, are replaced by their respective signi-
fiers. It is useful to think about this process, metaphorically of course, as we think
9780190256623-Dor.indb 111 03/06/15 6:18 PM

assert assertive intonation

speech act speech act
indefinite
definiteness a
singular definite-singular
singularity
vase vase
patient patient
break broke
eventuality eventuality+
past tense
tense
boy boy
topic-agent topic-agent
tall
property tall
property
singular
singularity the
indefinite-singular
definite
definiteness
figure 6.5
about the copying of a DNA sequence into RNA. In most cases, replacement takes
place one-to-one: most signifieds on the symbolic landscape are, by their very con-
ventional nature, uniquely attached to single signifiers. Sometimes, replacement
would take place many-to-one, as for example in the case of irregular morphology.
Second, some of the protocol-based relations on the semantic hierarchy are trans-
lated into their respective function words: prepositions, agreement markers, and
so on. This is the essence of the content-function word distinction: content words
come from the symbolic landscape (this is why they are open class); function words
are determined by the forms of the protocol (this is why they are closed class). Let
me, then, call the product of this stage—the basic utterance (see Figure 6.5).
In the next stage, linearize, the signifiers comprising the basic utterance
are arranged in linear order—following the linearization conventions of the
language. The output of this state will be called the linear utterance. This is the
second step toward the construction of the full utterance, which must be ready to
be transmitted as a stream of speech. This is why the process of linearization takes
place only after replace, and applies to signifiers, not to signifieds: it can only
apply to structural entities with a temporal dimension. For our purposes, let us
assume the following partial set of linearization conventions in English:
(i) Linearization takes place iteratively, bottom-up on the selection hierarchy;

(ii) quantificational extensions (such as singular-plural), and property exten-
sions, should appear to the left of their bases; the remaining extensions
should appear to the right of their bases;
9780190256623-Dor.indb 112 03/06/15 6:18 PM

(iii) When obligatory and non-obligatory extensions appear on the same side of

the base, obligatory extensions should always be on the left;
(iv) The topic-entity of the basic message should always be on the left of
eventuality.
Each of these conventions wears its functional utility for instructive commu-
nication on its sleeve. The first convention implies a process in which every
base is always located on the utterance with its extensions adjacent to it. This
is what strict word order (which is a fact about English) is all about: it copies
the semantic relations between the signifieds into the linear order relation-
ships between the signifiers. This, again, is a prescriptive requirement: “Always
mention the base with its extensions together, so that we understand that the
details are about that thing.” The other three conventions organize linear order
on the basis of the message’s properties. The prescriptive convention in this
case is this: “situate the different participants in your message in mutually
identifiable positions along the utterance, so that we will be able to recognize
which is which.”
In our rudimentary example, then, linearization would proceed in the
manner shown in Figure 6.6.
In the next stage, connect, the by-now ordered signifiers of the linear
utterance are structurally connected to each other, and a set of phonologi-
cal procedures are applied to them. Two sets of conventions are involved here:
bottom-up process of linearization
3 2 1
the vase
by convention (ii)
broke the vase

by convention (ii)
the tall boy broke the vase

by convention (iv)
the tall boy

by convention
(ii) and (iii)
figure 6.6
9780190256623-Dor.indb 113 03/06/15 6:18 PM

morphological conventions determine the extent to which two signifiers or

more, which are adjacent to each other, are to be phonologically fused into a
single word or phrase; phonological conventions determine what such fusion
should sound like.
In the last stage, speak, the full utterance goes through a set of phonetic pro-
cesses (deriving from the phonetic conventions of the language), and is actually
realized as a stream of speech.
6.3 The Procedures of Comprehension

Listeners follow the conventions of the protocol as they go through the pro-
cess of comprehension, all the way from the analysis of the stream of sound
to the imagination of the experiential interpretation. The conventions they
follow are the same: the fact that the conventions are mutually identified by
speakers and listeners is what makes communication possible in the first place.
The conventions are followed, however, in different ways, because the chal-
lenges that the listeners face on their way are of different types. As they move
along the process of comprehension, listeners go through four stages. Let me
call them identify, replace, reconstruct, and imagine. Then, some-
times, they also have to make pragmatic inferences on the basis of relevance
considerations.
In the first stage, identify, listeners apply a set of sound analysis proce-
dures to the incoming stream of speech. Using the conventions of phonetics,
and then phonology and morphology, the listeners attempt to break down the
stream of speech into a set of identified signifiers—still arranged in linear order
but already disconnected from each other. The output, then, is the linear utter-
ance. It is already at this early stage that a major difference is revealed between
linguistic production and linguistic comprehension: the level of proficiency re-
quired for the task of comprehension is always lower than that required for the
parallel task of production. Breaking down the stream of speech into identifi-
able signifiers is always easier than fusing a set of signifiers into an identifiable
stream. This is so, among other things, because there is an inherent element of
opportunism in the process of comprehension that cannot be a part of produc-
tion. When the speaker is missing a word, communication is halted (speech is
broken); when the listener fails to identify a signifier or two at the first stage of
comprehension, the entire process that still lies ahead provides him or her with
ample opportunity for compensation (an opportunity that may or may not be
used, of course; compensation, in and of itself, is a serious challenge). There is
a constant element of guessing in the process of comprehension that eventu-
ally allows listeners to produce reasonable interpretations, at different levels of
9780190256623-Dor.indb 114 03/06/15 6:18 PM

complexity, of utterances which they themselves could not have produced. It is

always there, from the first stages of language acquisition all the way to the abil-
ity of adult readers to understand (and identify with) works of literature they
could not have written themselves.
In the second stage of comprehension, replace, the by-now disconnected
and identified signifiers are copied onto their respective signifieds. This is the
stage in which lexical ambiguity asserts itself, for the first time, as a constant
danger. Ambiguity (lexical and other) is one of the deepest reflections of the fact
that speakers and listeners are not simply engaged in parallel-but-inverse tasks
of coding and decoding. Ambiguity is always only a problem of comprehension:
speakers are not confronted with homonyms, for example, because they always
proceed from an already selected signified to the signifier that is associated with
it. The problem of synonyms, on the other hand, is only a problem of production:
listeners are never confronted with the need to choose between two signifieds of
very close meanings, associated with two different signifiers.
In the third stage of comprehension, reconstruct, the listener faces the
challenge that lies at the very heart of what we usually think of as semantic
analysis: the challenge of the re-organization of the copied signifieds and the
reconstruction of the speaker’s complete message in its original, hierarchical
form. This is where the entire set of message construction conventions comes
into play, together with the conventions of linearization (in languages such as
English in which linear order is, by convention, associated with meaning). This
is the stage in which structural ambiguity becomes a constant concern: listen-
ers face the danger of re-associating the signifieds in ways that are licensed by
the symbolic landscape, but that do not correspond to the speaker’s original
message.
It is crucial to see, then, that the end result of reconstruct—a mental
representation, in the mind of the listener, of the speaker’s complete message—
is only a temporary stage on the way toward the understanding of the speaker’s
intent. It is only in the last stage, imagine, that the listener must try to meet
the challenge—the pragmatic challenge—of figuring out what the speaker
tried to convey. The listener has to reconnect the signifieds of the reconstructed
message, and their semantic relationships, to their experiential clusters; re-
analogize what was originally digitalized by the speaker; re-contextualize what
was de-contextualized; and bring back to the private domain of experience
what was originally translated into the language of social agreement. In this
stage, everything that is conventional in language receives its experiential in-
terpretation, which means that it also includes the inference of what Levinson’s
(2000) calls generalized conversational implicatures, such as the inference of
not all from some.
9780190256623-Dor.indb 115 03/06/15 6:18 PM

Linguistic comprehension is thus made possible by the conventions of lan-

guage, but it is also, at the very same time, totally dependent on the ability of the
listener to go beyond the conventions, to figure out what it was that the speaker
had to reduce, digitize, and stereotype, in order to be able to communicate in the
first place. A communication event proves to have been successful to the extent
that the listener has managed to gain an experiential insight that is as rich, as
complex, as inexpressible, as the intent that gave rise to the entire event—to the
extent that the listener has managed to transcend language as a social system and
re-enter the private domain of experiential cognition. This is why we usually only
remember the gist of what we heard, not the exact wording (Sacks 1967).
We sometimes say to our children: “You only understand what you have
heard, or read, once you manage to say it in your own words.” This is exactly the
point. When we repeat what we have heard in the same words, we never let go
of language: we never go beyond the realm of social convention. We only begin
to understand what we have heard when we manage to embed it within our own
experiential grasp of the world around us, when we go beyond the words of the
message and let our new understanding interact with our experiences in inex-
pressible ways. Then, if we wish to restate what we now understand, we have to
launch a new process in which we are the ones who face the challenge of reduc-
ing our communicative intent into the conventionalized forms by our language.
The very fact that this new communicative intent emerged within our experi-
ence, not that of the original speaker or writer, implies that we would probably
find ourselves (subconsciously) making different decisions along the process of
production—selecting a different message-kernel, choosing different signifieds,
building a different message.
This is precisely why the entire process of linguistic communication is so
fragile. Understanding depends on our ability to go beyond the conventions of
language and re-enter the private domain of experience, but the very fact of the
experiential gap implies that whatever we do, we always re-enter our own expe-
riential world, never that of the speaker. Every understanding is also, at the very
same time, a misunderstanding.
In is only after imagine, when we have the full experiential interpretation,
that we sometimes face the challenge of conversational implicature. The listener
suspects that the speaker could not have meant to say this, and tries to recon-
cile the experiential interpretation with everything else in his or her experiential
world. Sperber and Wilson (1986) capture this explicitly in their insistence that
relevance-based inference is a non-linguistic affair. There is a general point here
that needs to be stressed: pragmatic theories of comprehension emerged in the
twentieth century in an environment in which semantics was still thought of as
the only necessary tool for the analysis of meaning. The days were the great days
9780190256623-Dor.indb 116 03/06/15 6:18 PM

of logic and computation, and the theories tried to show, quite modestly, that in
some particular cases, semantic analysis was not enough, and that in some other,
more extreme cases, it actually produced the wrong results. Pragmatic analysis
was needed only in those instances where it seemed that the speaker’s intent was
not fully reflected in semantic interpretation of the utterance. The notion of im-
plicature, as formulated by Grice (1975), was constructed not as a general con-
cept of interpretation, but as a supplementary concept—something that needed
to be introduced into the analysis of meaning only in cases where semantics was
not enough. Sperber and Wilson (1986) later contended that, rather than fol-
lowing Grice’s four maxims, the speaker actually follows a single principle: the
principle of relevance. Both conceptions, then, implicitly accepted the reification
inherent in language itself—and thus, in consequence, in the theory of seman-
tic interpretation—that the symbolic landscape of a language, together with its
communication protocol, maps onto the experiential worlds of the speakers in
a perfect manner, and thus, in effect, obliterates the experiential gaps between
them: as long as the speaker follows the conventions of the language to the letter,
the semantic content of the speaker’s message is all the listener needs in order to
fully understand what the speaker attempted to say. It is only when the speaker
violates the conventions—intentionally or by mistake—that an additional layer
of interpretation is required.
All this, however, is nothing but an idealization. Comprehension cannot end
at the level of semantic interpretation, because semantic interpretation cannot
take the listener beyond the meaning conventions of the language. The stage of
imagine is already pragmatic. Pragmatic interpretation does not wait for those
special cases when the speaker ostensibly violates the conventions of the lan-
guage. It is always there, as the essence of the entire process of comprehension.
6.4 The Psycholinguistics of Production

The psycholinguistic literature on the production and comprehension of lan-
guage is immense (for overviews, see Harley 2008, Spivey et al. 2012, Levelt
2013). In this section, I will try to do no more than hint at some first connec-
tions between the major discoveries made in the field of production and the
model presented above. I will not discuss the processes of comprehension, and
will settle for the following remark: one of the major issues in psycholinguistics
is whether production and comprehension should be thought of as distinct pro-
cesses, employing different computations and based on different capacities, or as
two facets of the same integrated reality. Most of the literature clearly goes for
the first option, but Pickering and Garrod (2013) have recently proposed a de-
tailed model that integrates both processes. The authors show that speakers also
9780190256623-Dor.indb 117 03/06/15 6:18 PM

partially perform as listeners to their own words, and listeners covertly imitate
the actions of speakers, thus in fact partially performing as speakers. This is a
very important idea, and it is theoretically strengthened by my perspective. Both
sides of the communication process work with the same mutually identified
prescriptions—otherwise they would not be able to understand each other—
but they implement the prescriptions in different ways. Speakers, for example,
translate their private experiential intents into messages; listeners imagine on
the basis of the messages. The two processes are regulated by the same prescrip-
tions, but they nevertheless imply radically different challenges. This is reflected
in the constant struggle between speakers and listeners in mutual-identification
events: speakers have to make sure that their language allows them to express
themselves. Listeners have to make sure that they understand.
The most impressive fact about the current relationship between linguistics
and psycholinguistics is how slim it is. Following a short period of co-operation
in the two first decades of the Chomskian era, and with some counter-examples
later on, the two disciplines have gone their separate ways. The dominant models
of language and its processing in both of them look like models of two very dif-
ferent systems. On the surface, this is a reasonable reflection of the competence-
performance distinction—models of performance should not necessarily look
like those of competence—but the divide is actually much deeper. Psycholin-
guistics very often find linguistic theorizing too abstract and too removed from
psychological reality; linguists consider psycholinguistics too shallow, too close
to the hardware of communication to allow for insights about the software.
In terms of this divide, the theory of language as a technology for the in-
struction of imagination sides fairly and squarely with the psycholinguistic po-
sition. In a very real sense, it may be read as an attempt to construct a theoretical
understanding of language that would pass through the bottleneck of current
psycholinguistic research.
I have already said something about the type of relationship I see between
linguistic theory and psycholinguistics: the theory describes and explains the
social attractor, the set of mutually identified norms and regulations of the lan-
guage, and psycholinguistics deals with everything that happens to speakers and
listeners around it. Putting things this way, we may now distinguish between
two types of psycholinguistic phenomena: those determined by the technologi-
cal properties of language as the external attractor, and those that emerge in
the actual effort to meet the challenge. Consider the question of seriality, for
example. According to Garrett’s (1975, 1988, 1992) model, language produc-
tion is a serial process: in every stage a different type of computation takes place,
with only one thing happening at each stage, and the computations do not in-
teract between stages. A lot of speech-error data supports this, but there are also
9780190256623-Dor.indb 118 03/06/15 6:18 PM

counter-examples. Garrett’s conception implies that the system can only compute
one message at a time, but speech errors show that we sometimes try to produce
two in parallel. In one of the most beautiful examples (Harley 2008, p. 404), the
speaker seems to have attempted to say both “I’m making some tea” and “I’m put-
ting the kettle on”—and ended up with “I’m making the kettle on.” This seems to
imply parallel processing. Other observations indicate that the computations at
the different levels interact with each other, and information flows in all direc-
tions, even in reverse—from later stages to earlier ones (Dell 1986, Harley 1984).
This is called feedback, and it explains why word substitutions such as “darn
bore” for “barn door” emerge more often as real words rather than non-words,
and why word substitutions are often phonologically similar to their targets. As
things stand in the literature, there is a debate here: production is either serial
or parallel and interactive. It cannot be both. In terms of my proposal, however,
both sides of the debate actually highlight two sides of the same coin. Language
production is prescribed as a serial process—and performed interactively. On
the one hand, the prescribed seriality reflects the logic of the technology: it is
designed for the staged production of a single utterance at a time, and nothing
in the process requires interaction between the stages. In actual processing, how-
ever, production is handled the way everything else is handled in the brain, with
massive parallel processing. Speakers, moreover, encounter a wide array of prob-
lems to overcome, including, as we have seen, the challenge of choosing a single
message kernel out of the many that suggest themselves. In the above example
with the kettle, then, the speaker evidently could not decide—and two produc-
tion processes were launched at the same time. This was obviously made possible
by the speaker’s cognition, and it is worth asking how and why, but the output
was an error—something that the protocol did not permit. Garrett’s model thus
seems to capture the foundational fact that language, as a technology, requires
its speakers to produce outputs serially constructed in specific ways. The parallel
cognitive processes, the feedback, the planning ahead—all these are required for
the speaker to be able to produce such outputs. The speech errors presented by
both sides of the debate are all indications of failure to meet this challenge.
Consider feedback again: why should there be a mechanism that looks back
like that? As Levelt (1999, p. 225) puts it, “its functionality can hardly be to
induce speech errors.” As Postma (2000) shows in his review of speech errors
(and their self-repair), the different approaches to the question converge on the
notion of self-monitoring: throughout the entire process of production, stage by
stage, we monitor our outputs to make sure they are appropriate—on many dif-
ferent grounds. Errors that are more similar phonologically to the target, or those
that look like real words, may thus be able to escape under the radar more easily
than others. Self-monitoring, however, is involved to some extent in everything
9780190256623-Dor.indb 119 03/06/15 6:18 PM

we do, not just in language. It is not required as such by the protocol. Whereas
seriality is specifically required by the linguistic technology, feedback is a general
cognitive mechanism, specifically adapted to the requirements of language the
way it is adapted in other way to the requirements of other challenges.
This division of labor allows for the following hypothesis: we should find
much less individual variability in psycholinguistic phenomena that emerge
from the foundational properties of the linguistic technology and much more
individual variability in phenomena having to do with the cognitive dynamics
taking place in the communication process. We all try to meet the same exter-
nal challenge, but we do it in different ways. The most robust discoveries in the
field should thus reflect the inherent properties of the technology. This, as far
as I can tell, is indeed the case. I will discuss three points. First, the protocol
prescribes that speakers handle words in two stages—first the signified, then the
signifier. This is the most important discovery made in the investigation of word
production (Levelt 1989). Among other things, researchers have shown that (a)
in a series of production tasks, speakers tend to re-use words in later tasks that
they used in earlier ones, but this priming effect only takes place when the words
are the same semantically, not phonologically: homophones do not produce the
effect (Wheeldon and Monsell 1992); (b) in picture-word interference experi-
ments, word production is sensitive to semantic interference only in the first
150–200ms, from which point on semantic interference is replaced by phono-
logical facilitation (Schriefers, Meyer, and Levelt 1990); (c) the different brain
regions responsible for semantic and phonological production are activated in
sequence (Indefrey and Levelt 2004); (d) the tip-of-the-tongue phenomenon,
where we feel we know the meaning of a word but cannot retrieve its form, is
best explained as a difficulty in moving from the first stage to the second (Brown
and McNeill 1966, Caramazza 1997).
Second, the protocol prescribes that the grammatical construction of the
utterance should take place in two stages—the first hierarchical and semanti-
cally based, the second devoted to the positioning of signifiers in linear order.
This is the established view in the field. In Garrett’s model, the functional level
is where words are selected and assigned to their roles as subjects, objects, and
so on in their syntactic frames; the positional level is where linear ordering takes
place. This is supported mainly by speech-error observations. Thus, for example,
semantically based word substitution can take place across long distances in the
utterance, which means they happen before linearization; sound substitutions,
however, are already constrained to very short linear distances. Content words
and function words, moreover, never substitute for each other, and function
words are often positioned in their right place even if the content words they
are attached to are substituted. In Fromkin’s (1973) famous example, where “a
9780190256623-Dor.indb 120 03/06/15 6:18 PM

weekend for maniacs” is produced instead of the intended “a maniac for week-
ends,” the nouns are exchanged but the plural morpheme—"s”—stays where it
should be. According to Garrett, this means that content and function words are
brought into the process at two different stages: content words are selected at the
functional level; function words only appear at the positioning level. This, too, is
predicted by the theory: content words are indeed selected in the process of mes-
sage construction; function words are determined by the protocol’s forms, and
are brought in only in the signify stage —after the completion of the message.
The third, more general point has to do with the fact that regular speech
production (in every language) is infested with errors, pauses, hesitations, and re-
pairs. All these seem to appear where the theory predicts, because they expose the
inherently fragile points in the workings of the technology. As Beattie and But-
terworth (1979) show, for example, the distribution of hesitations is determined
mainly by problems in semantic planning. Hesitations mostly appear when we
experience difficulty in translating our experiential intents—holistic, analogue,
and private—into the digital forms dictated to us by the protocol. Importantly,
when we experience such difficulties, we often have no problems translating our
experiential intents into presentational communication: we produce mimetic
gestures, such as those of using the telephone, when we can’t find the word (But-
terworth and Beattie 1978). Clark and Fox Tree (2002), based on Clark (1996),
show that English speakers use the fillers uh and um to tell their interlocutors
that they are taking a pause in order to make a translation decision—uh for short
pauses, um for longer ones. The authors show that “uh and um are conventional
English words, and speakers plan for, formulate, and produce them just as they
would any word” (p. 73). This is exactly what the theory would predict: the fillers
are prescribed by the protocol.
General linguistic theory has always been interested in language at its best.
Broken speech has been delegated to psycholinguistics. The theory developed
here implies that broken speech is where the essence of language as a technology
is most clearly revealed. The entire production process, including much of the
feedback, is almost always subconscious. The first time we get consciously ac-
quainted with our own message is when we hear ourselves speak. We construct a
message, translate it into an utterance, start speaking it, listen to the first words,
realize that this was not what we meant to say, stop, hesitate, try again, and so
on and so forth. We enter a conscious feedback loop, with ourselves and our lis-
teners, in which we struggle to approximate our intent. The essence of language
emerges in broken speech as the real-time struggle, within the speaker, between
the private will to communicate and the social conventions of language.
There is, quite obviously, much that is different between my theory and the
production models discussed in the field. The most important difference has to
9780190256623-Dor.indb 121 03/06/15 6:18 PM

do with syntax. In Garrett’s model, as well as in Bock’s (1986) and others, the
construction of the utterance’s hierarchical syntactic structure is a major part of
the process. In Levelt’s (1989) two-staged model of word production, the entity
involved in the first stage, the lemma, also specifies the syntactic properties of
the word, its grammatical category and function, and such syntactic features as
grammatical gender. My description of production, however, does not include a
level of hierarchical syntax. The explanatory burden that falls on the shoulders
of hierarchical syntax—both in production and comprehension and in general
linguistic theory—is split in my theory between three components: (a) the se-
mantics of signifieds; (b) the prescriptive forms of the protocol; and (c) the con-
ventions of linear order. This is a major theoretical move, and the next chapter is
devoted to its explication and argumentation.
9780190256623-Dor.indb 122 03/06/15 6:18 PM

7 T h e S o c i a l A u t o n o m y o f S y n ta x
Having looked deeper into the processes of production and compre-

hension as prescribed by the protocol, we are now in a position to take
a fresh look at the question of syntax. Over the past fifty years, the
study of syntactic complexity has taken center stage in linguistics:
within the generative paradigm, it has been considered the most im-
portant window through which we can peak into the formal, hierar-
chical, and recursive capacities that separate the human mind from
all other minds in nature, and language from all other communi-
cation systems, including the hypothetical protolanguage spoken by
our own ancestors. Chomsky has revised his syntactic theories five
or six times throughout this period, and the literature now includes
dozens of theories, most of which are notational variants on his own,
some of which—especially Culicover and Jackendoff’s (2005) simpler
syntax, Joan Bresnan’s (2001) lexical functional grammar and Pollard
and Sag’s (1994) head-driven phrase structure grammar —differ more
substantially. Much of the effort in the functional-cognitive camp has
been dedicated to the construction of meaning-based grammars (Lan-
gacker 1987, Givón 1995, Bybee 2006, Evans and Green 2006). In the
last twenty years, some of the authors on both sides managed to find
common ground around the compromise position of construction-
ism (Fillmore 1982, Goldberg 2006, Goldberg and Jackendoff 2004,
Croft 2001). All sides to the debate follow Chomsky’s foundational
conception of syntax as the most important reflection of the universal
properties of the human mind.
The theory of language as a communication technology for the in-
struction of imagination does not just demote syntax from this privi-
leged position—it actually denies the existence of hierarchical syntax
as a level of representation. Syntactic complexities as observational
patterns do not reflect an inner syntactic reality: they emerge from
the conversion of subsets of the protocol’s prescriptions. Tree-like syn-
tactic patterns, in other words, are observational properties of linguis-
tic outputs, not explanatory properties of the linguistic system. The
hierarchical dimension emerges from the semantic process of message
9780190256623-Dor.indb 123 03/06/15 6:18 PM

construction; the linear dimension emerges from the process of linearization. In

this chapter, I will try to show how this works.
I will frame the discussion around what is probably the most important
debate in linguistics in the last half century: the debate over the autonomy of
syntactic patterns. Chomsky’s characterization of language as a non-functional
entity was not just another component of his overall, philosophically inspired
conception of language as an innately given, universal cognitive capacity. At its
core, it was founded on a set of basic empirical observations (mostly in English
and typologically similar languages) that seemed to indicate that the syntactic
patterns manifested by language cannot be correlated with patterns of meaning
and communication. The argument based on these observations was simple: the
products of a system whose essence lies in its function have to ostensibly reflect
this function in their structures. If the essence of language lies in the function
of the communication of meaning, then the observable patterns of structure it
manifests should correlate with patterns of meaning. But they do not. They seem
to be autonomous.
Most of the debate that emerged around the claim of autonomy thus con-
centrated on the empirical question of the precise degree of autonomy of the rel-
evant structural patterns from considerations of meaning and communication.
The functionalists kept on demonstrating that there were deep resemblances be-
tween the relevant structural patterns and general patterns of meaning and com-
munication. Generative grammarians, for their part, kept insisting that this was
not enough: the correlations have to be exact. In a strange way, most of the debate
never got beyond what Chomsky himself already said, back in 1957, in his Syn-
tactic Structures. Toward the end of the book, having presented some counter-
examples to certain functionalist proposals, he wrote:
These counter-examples should not, however, blind us to the fact that

there are striking correspondences between the structures and elements
that are discovered in formal, grammatical analysis and specific semantic
functions. None of the [functional] assertions is wholly false; some are
very nearly true. It seems clear, then, that undeniable, though only im-
perfect correspondences hold between formal and semantic features in
language. (p. 101)
For Chomsky, the fact that the correspondences were so inexact suggested that
“meaning will be relatively useless as a basis for grammatical description.” The
correspondences should be “studied in some more general theory of language
that will include a theory of linguistic form and a theory of the use of language as
subparts.” We should proceed in the investigation of structure as a purely formal
9780190256623-Dor.indb 124 03/06/15 6:18 PM

The Social Autonomy of Syntax • 125
entity, and only then, “having determined the syntactic structure of the lan-
guage, we can study the way in which this syntactic structure is put to use in the
actual functioning of language” (p. 102). For Chomsky’s opponents, the fact that
the correspondences were so striking implied that meaning should be the basis
for grammatical description, which means that, ultimately, the formal proper-
ties of language should (and could) be explained in functional terms. The debate
thus became, quite rapidly, a battle over fine-grained judgments. The question
came to be: how striking should a correspondence be in order for it to be con-
sidered exact? In most cases, the struggle has not led any of the sides to inspiring
results. In some cases, where the struggle was eventually replaced with the search
for theoretical common ground, important results were achieved. We shall get
back to them shortly.
In this chapter, then, I will claim that the entire struggle was fundamentally
misguided. Chomsky’s demonstrations of structural autonomy do not imply
non-functionality, and they do not need to be explained away in order to salvage
the functional essence of language. It is precisely in Chomsky’s demonstrations
that the essence of language, as a social system of instructive communication, is
most clearly revealed.
Two different levels of meaning, rather than one, are involved in our relation-
ship with language—the level of private-experiential meaning, and the level of
normative-semantic meaning. The functional essence of language lies in the fact
that it allows for the approximate translation of meanings of the first type into
meanings of the second, and vice versa. Chomsky’s demonstrations deal with the
relationship between the patterns of regularity manifested by utterances and the
patterns of private-experiential meaning. His observations are thus correct. The
formal properties of utterances are indeed autonomous from experiential mean-
ings: they have to be. They are social—experience is cognitive and private; they
are discrete—experience is analogue. Utterances, however, are built from mes-
sages: their formal properties reflect the social and discrete properties of their
messages. The observable syntactic patterns of language, then, are intimately cor-
related with normative-semantic meaning. Both sides to the traditional debate
are thus both right and wrong. First, the syntactic complexities of language are
indeed autonomous from general, experiential meaning. This, however, does not
imply non-functionality; it is actually the key to the functional specificity of lan-
guage. Second, syntactic patterns are indeed autonomous, but not as cognitive
representations; what they indicate is the social autonomy of language as a spe-
cialized technology.
Autonomy claims come in two versions: paradigmatic and syntagmatic. Par-
adigmatic claims have to do with the defining properties of the structural build-
ing blocks of language: such entities as parts of speech (noun, verb, adjective, and
9780190256623-Dor.indb 125 03/06/15 6:18 PM

their likes) and grammatical functions (subject, object, and so on). The question
here is what determines whether a certain linguistic expression, in a certain sen-
tence, is a noun or a verb, a subject or an object. Syntagmatic generalizations have
to do with the distributional patterns manifested by words and other expressions
across sentences. Here the question is what determines whether a certain linguis-
tic expression may appear in a certain configuration together with another in a
sentence. Let us begin with the paradigmatic issue of the autonomy of parts of
speech.1
7.1 Parts of Speech as Formal Slots

The functionalist view on parts of speech has traditionally been that they have
to be defined notionally, in terms of their meanings (Lyons 1977, Langacker
1987, Anderson 1997, 2006): a noun is a word that names a person, a place, or
an entity; a verb is a word that names an event or a process; an adjective is a word
that names a quality, and so on. This is indeed what we feel intuitively as speak-
ers. The generative view—which in fact has a much longer history—has been
that notional definitions of this type do not account for the observable syntactic
facts. The most ready examples are words that by the notional definition should
be verbs or adjectives, but are in fact nouns: words such as fire, crime, party, and
war name events and processes, but they are not verbs; words such as beauty
and health name properties, but they are not adjectives. The verb dance and the
noun dance clearly name the same type of eventuality, but they play very differ-
ent roles in the syntax. Generativists have thus insisted that terms such as verb
and noun should be defined on the purely formal basis of distribution (Baker
2003): verbs are those words that appear in verbal positions in the sentence;
nouns appear in nominal positions; and so on. This means that the paradigmatic
question at hand can only be resolved syntagmatically. Everything is a matter of
distribution.
As these arguments stand, the generativist position must be closer to the
truth: it predicts part of speech membership much more efficiently. The posi-
tion, however, suffers from three problems. First, it does not go beyond the level
of description: no explanation is provided for the distributional facts; second,
it fails to account for our notional intuitions; third, it cannot capture the huge
typological diversity in the parts of speech systems in the languages around the
world. In the last two decades, with the exception of Baker (2003), the gener-
ative discourse seems to have lost interest in the question. In the functionalist
camp, discourse has turned its attention to the facts of diversity, and the most
developed theories are Hengeveld’s (1992) and Croft’s (2001). Hengeveld at-
tempts to maintain something of the distributional spirit of the generative view
9780190256623-Dor.indb 126 03/06/15 6:18 PM

within a generally functionalist perspective: words can appear in one of four

syntactic slots, which are functionally defined as propositional functions: the
head of a predicate phrase, the head of a referential phrase, a modifier within
a predicate phrase, and a modifier within a referential phrase. A word that
can appear in one of these slots, “without further morphosyntactic measures,”
would respectively be a verb, noun, adjective, or adverb. Croft postulates three
functions instead of four—predication, reference, and modification—but also
incorporates a notional level into his theory. Words belong in three semantic
classes: objects, actions, and properties. As opposed to the classical notional
conception, however, Croft thinks about this semantic level in prototypical
terms: prototypical nouns, for example, are non-relational, stative, permanent,
and non-gradable. Both chair and war are nouns in the sense that they are used
for reference (and here Croft and Hengeveld say the same thing), but Croft also
captures the fact that war is less prototypical as a noun than chair: it is non-
stative and transitory. The two authors thus explain the typological facts in
slightly different ways.
Two issues are involved, then: the inherent meaning of the word and the posi-
tion it is assigned to in the sentence. This division of labor, and its consequences,
are naturally incorporated into my theory. First, the positioning is determined
by the formal slots of the protocol. A noun is a word that can be selected to fill in
the slot entity; a verb can be selected for eventuality, and so on. Second, whether
a word is fit to fill in these slots is indeed a matter of its meaning, but here the
theory offers a new interpretation of the observable facts. The question is: why
should the semantics of parts of speech have prototypical structure? The answer
I would like to suggest is this: all sides to the debate, traditional and current,
have taken the concept of naming to be a systematic function from experiences
(physical entities, places, events, states, and qualities) to signs. Under this shared
conception, the traditional notional perspective would be right only if the prop-
erties of signs could be exhaustively determined on the basis of the experiential
clusters associated with them. Indeed, this is not the case. There is nothing in
the experience-clusters associated with such words as war or beauty that implies
that they should be nouns. For the generativists and for Hengeveld, this implies
that the notional side of things should be left out of the picture. Croft solves part
of the problem by showing that these words are non-prototypical nouns, but he
leaves the main question intact: why are war and beauty nouns in the first place?
How is it that they are used for reference?
The answer, then, emerges from the fact that naming is not a descriptive
function from the world of experience to the world of signs. It is prescriptive:
from signs to the world of experience, from the realm of social convention to
the realm of private experiencing. The symbolic landscape does not mirror the
9780190256623-Dor.indb 127 03/06/15 6:18 PM

complexities of experience. It constructs its own simplified, stereotyped world-

view, superimposing on the analogue world of experience a set of digital dis-
tinctions and generalizations that the speakers of the language found necessary
for the purposes of communication. What this means is that the traditional
notional definition of parts of speech, which is wrong as is, is actually perfectly
right in reverse order. Not: “If something in the world of experience is a person,
a place, or an entity, then the word referring to it is a noun,” but: “If a word re-
ferring to something in the world of experience is a noun, then the thing in ex-
perience that the word refers to is to be thought of, by prescription, as a person,
place, or entity.” Not: “If a thing in the world of experience is an eventuality,
then the word referring to in language is a verb,” but: “If a word referring to a
thing in the world of experience is a verb, then that thing in experience which
the word refers to is to be thought of, for the sake of communication, as an
eventuality.”
The fact that the digital categorizations of the symbolic landscape are super-
imposed, by social convention, on the world of experience, thus means that these
categorizations may find themselves superimposed on experiences which might
not, at first sight, suggest themselves as candidates for the relevant category. This
is the very essence of prescription. Such words as war, dance, and beauty are nouns
in English not because there is something inherently nominal about them. They
are nouns because the speakers of the language decided, at some point in the
cultural evolution of the language, that it would be useful to be able to refer to
the entire sets of war-experiences, dancing-experiences and beauty-experiences,
as if they were entities. The capacity of referring to wars, dancing, and beauty as
entities is functional in the most profound sense of the word: it enhances the ex-
pressive capacities offered to the speakers by their language. Being able to refer to
wars, dancing, and beauty as entities, the speakers may now qualify them (“un-
usual beauty”), count them (“three wars”), use them as topic-entities (“the dance
went on for hours”), and so on and so forth.
As far as language is concerned, then, war and beauty are not non-
prototypical. They are used for reference, and that is it. The two nouns, how-
ever, are non-prototypical as far as our experiences are concerned. The fact that
the correspondences between parts of speech and experiential meanings are
prototypical—both striking and imprecise—is exactly what we should expect
to find. The syntactic distinctions between parts of speech are perfectly corre-
lated with the discrete, prescriptive meanings of the protocol and the symbolic
landscape. These prescriptive meanings, however, are correlated only impre-
cisely with the analogue categorizations we make in our experiential words.
Consequently, our parts of speech are only imprecisely correlated with our
experiences.
9780190256623-Dor.indb 128 03/06/15 6:18 PM

7.2 Grammatical Functions as Regulative Prescriptions

The very same line of argumentation applies to grammatical functions. Consider
the subject. Functionalists claim that the syntactic notion of subject can and
should be defined in terms of its communicative function—as the topic of the
sentence, what the sentence is about, and as its most prominent argument (Li and
Thompson 1976, Langacker 1999). Generativists contend that this is sometimes
true, but not always, and it is definitely not true when the subject seems to have
no semantic significance (when it is expletive). Thus, in the English sentence “It is
raining,” the subject it is not what the sentence is about, simply because it is very
unclear what it stands for in the first place. For generativists, this has been taken
as evidence for the contention that subjecthood is a formal notion: the subject
is the external-argument position on the syntactic tree. English sentences with
expletive subjects thus provide us with the first formal property of the subject
position in English: it must always be lexically filled. No reference is made in this
conception to topichood or to any other meaning-related concept. Subject posi-
tion and its properties are autonomously syntactic.
As the debate stands, and this is already familiar, the generative perspective is
closer to target. The subject cannot be defined as a function from the private ex-
perience of aboutness to a certain structural configuration, for the simple reason
that it appears in situations where the experience is not there. Once again, to
resolve the problem we have to reverse the order of things and look at the sub-
ject not as a natural function from experience to language, but as a prescrip-
tive function—from language to experience. The contention that the subject is
what the sentence is about should not be read as a description of something that
is simply there, but as a conventionalized, regulatory principle, which a com-
munity of speakers decides to impose on itself precisely because the speakers of
the community do not always naturally specify for their listeners what it is that
they are talking about. A prescriptive principle, through which the members of
the community (acting as listeners) tell themselves (as speakers): “From now on,
whenever you tell me something, make sure that you remember to tell me what
(or who) you are talking about. It will help me understand.”
For this prescriptive principle to be efficiently implemented, two separate
issues have to be seen to: first, a special convention needs to be added to the set
of conventions of message construction of the language, that will oblige speakers,
whenever they set out to construct a new message, to select a single entity from
the symbolic landscape and designate it as the topic. This is exactly what the spec-
ification of the topic-entity in the message kernel is there for. It forces the speaker to
decide, at the very outset of the process of experience-to-speech translation, what
he or she wants to talk about. Second, a special convention (or set of conventions)
9780190256623-Dor.indb 129 03/06/15 6:18 PM

needs to be added to the conventions responsible for the construction of the ut-
terance, one that will oblige speakers to mark the topic-entity in a specific way,
mutually identified with the listeners, so as to allow them to distinguish between
the topic-entity and the rest of the entities in the sentence. In English, the con-
vention chosen for this task is a linearization convention: the topic-entity has
a designated default position in the linear order of the utterance. The fact that
all this is a prescriptive effort, that specifying the topic is not a natural tendency
shared by all speakers, crucially implies that its application cannot be a function
of the speakers’ experiential intents. The conventions oblige speakers to choose
and mark a topic-entity, whether or not there is indeed a certain entity that they
themselves experience as the thing they intend to talk about. This is why the sub-
ject is required (in English) even in those cases where it is expletive. This very
fact, that the notion of subject is indeed autonomous from experiential meaning,
is thus the best indication of the specific type of prescriptive function it fulfills.
Here is another way to say the same thing: much of the debate around sub-
jecthood was also, implicitly, about the competence-performance distinction. It
is well known that speakers of English often produce sentences without subjects,
as in the answer “Went out to get something to eat” to the question “Where
is John?” In situations of this type, it is clear that the subject is not function-
ally necessary: the question already determines what the answer is going to be
about. For functionalists, this meant that the subject was not really obligatory
in English, and that its distribution was indeed a function from the experience
of communication to the actual usage of language. For generativists, the impor-
tant point was not what was happening in the course of performance, but what
types of grammaticality judgments speakers produced when asked to consult
their linguistic intuitions. Here, speakers of English very clearly indicated that
they knew that the subject was obligatory. They did not always behave accord-
ing to their knowledge, but they knew nevertheless. The two competing camps
thus highlighted (and discredited) different types of empirical observations in
the course of their argumentation, with the debate quite often deteriorating into
an argument about presuppositions.
The realization that the subject is neither a natural function of communica-
tion nor a purely formal entity, but a regulative mechanism, a set of prescriptive
conventions designed to organize the experience of communication, immedi-
ately reveals that the empirical observations used by both sides to the debate were
not only valid—but also significant. In most cases, speakers clearly indicate in
their grammaticality judgments that they have indeed internalized the relevant
conventions of their language: they know the law. When explicitly asked about
their knowledge of the law, they demonstrate it. They thus indicate that the sub-
ject is normatively obligatory in English. They do not, however, always behave
9780190256623-Dor.indb 130 03/06/15 6:18 PM

according to the law: social norms are not always followed. Specifically, in cer-
tain communicative situations, speakers may feel that the normative behavior
is superfluous, because whatever the norm was supposed to regulate in the ex-
perience of communication is already there. This is exactly the situation in the
above example: aboutness has already been determined by the question, and the
responder may feel free to break the law and produce a subjectless sentence with-
out the risk of communicative failure. This is exactly what we do whenever we
encounter regulative laws: we usually refrain from crossing the street on a red
light, but when we do—we (hopefully) do it in situations which we judge as rela-
tively safe. And note: we may sometimes feel that it is safe to cross the street on a
red light, but we nevertheless stay in place—because there is a child walking with
us. We are more normative in the company of children. This is an important fact
about language acquisition.
7.3 Constructionism and Selection

Let us move on to autonomy claims of the syntagmatic type. The question here
is how the distributional patterns of words (and other signifiers) in sentences
are to be accounted for. Two issues are involved. The first is the problem of selec-
tion. The relevant observations show that there are systematic restrictions on the
co-occurrence of word-types in sentences: different types of words seem to be
allowed, and disallowed, to appear with other types of words, in different types
of constructions. The question is: how are these restrictions to be explained? The
second problem is a more complicated variation on the same topic—the prob-
lem of long-distance dependencies: some of the systematic restrictions on the
co-appearance of words (and other signifiers) in syntactic constructions seem
to involve complex considerations of structural distance. How are these to be
accounted for?
It was around the question of selection that the struggle between functional-
ists and generativists eventually evolved into an attempt, shared by scholars from
both schools of thought, to reach theoretical common ground. In the last three
decades, the attempt has developed into a full-fledged, alternative perspective on
the question of selection called constructionism, which is at the same time (and
for the first time) functionally oriented and empirically robust. At the core of
the new perspective lies the understanding that the syntagmatic observations
of selection, for a long time used as evidence for autonomy, are actually corre-
lated with much finer-grained notions of meaning and communication than
the more traditional functionalists and cognitivists recognized. These finer-
grained notions of meaning, moreover, are not general functional principles of
9780190256623-Dor.indb 131 03/06/15 6:18 PM

communication, but the specific meanings of the words and constructions in-
volved in the observable distributional patterns. The following examples, from
Fillmore (1970), are by now classic material. The verbs hit and break are some-
times allowed to appear in the same syntactic configurations—but are also often
disallowed to do just that:
(1) a. John broke the stick

b. John hit the stick
(2) a. The stick broke
b. *The stick hit
(3) a. *I broke John on the leg (meaning: I broke John’s leg)
b. I hit John on the leg (meaning: I hit John’s leg)
What Fillmore noticed was that these behavioral patterns (and many others)
are not just manifested by these two particular verbs, hit and break. Verbs such
as bend, fold, shatter, and crack behave exactly like break; verbs like slap, strike,
bump, and stroke behave exactly like hit. The behavioral patterns are thus associ-
ated not with verbs, but with verb classes, and the classes are distinguished from
each other on the basis of a set of fine -grained semantic categorizations. Break,
bend, and fold are change-of-state verbs: they indicate, as part of their meanings,
that their object came to change its state as a result of the relevant eventuality.
Hit, slap, and stroke are surface-contact verbs: they indicate that their object came
to be touched on its surface as a result of the eventuality denoted by the verb. In
English, all change-of-state verbs appear in syntactic configurations (1) and (2),
but not in (3). All surface-contact verbs appear in (1) and (3), but not in (2). The
syntactic restrictions in the above examples are thus correlated with the seman-
tics of the classes of verbs involved, and the objects they take play different se-
mantic (or thematic) roles under the different classes. The stick is a patient under
break (1a) and a theme under hit (1b).
How is this to be explained? Different constructionist theories suggest dif-
ferent technical variations on the explanatory principle that Goldberg (2006)
calls the Semantic Coherence Principle. Informally speaking, all parts of syntax,
simple and complex, are constructions—systematic pairings of form and mean-
ing. Words and other types of constructions (like the possessed body-part con-
struction in (3), X on X’s body part) may appear together in the same sentences
(which are themselves complex constructions) if the meanings of the different
words and constructions are semantically compatible with the meanings of the
sentential constructions. The meanings of the building blocks of a sentence have
to be compatible with the meaning of the sentence itself. Very informally, con-
structionism takes sentences to be systematic mappings of thematic roles onto
9780190256623-Dor.indb 132 03/06/15 6:18 PM

grammatical functions. Sentences construct perspectives on eventualities. The

construction in (1) maps an agent (the instigator of the eventuality) onto the
subject, and either a patient or a theme onto the direct object. It thus provides
a way of describing transitive eventualities (eventualities in which an entity
does something to another entity)—from the point of view of the entity that
instigated the eventuality. The construction in (2) maps a patient onto the sub-
ject: it provides a way of describing eventualities in which an entity undergoes
a certain change (not necessarily as a result of something else doing anything
to it)—from the point of view of that entity. The construction in (3) involves
the possessed body-part construction, whose intrinsic semantics indicates that
it can only be used to refer to themes (the details of the semantics are not impor-
tant at the moment). The different constructions may thus appear together in
the relevant sentences only if their semantics are compatible: this is why break,
with its patient, may appear in constructions (1) and (2), and hit, with its theme,
may appear in (2) and (3).
All this makes good functional sense. It is still formulated, however, in a way
that severely reduces its theoretical significance. Two issues are involved. The first
has to do with the semantic categorizations that turn out to determine the rele-
vant syntactic patterns. What exactly are those semantic categorizations? Where
do they belong in the general picture? The constructionist view, formulated most
clearly by Jackendoff (1983, 1990), is that the relevant semantic categorizations
emerge on the level of general cognition known as conceptual structure. Lan-
guage, according to this perspective, provides structural tools for the commu-
nication of meanings, which are in and of themselves independent of language.
Everything said here so far, however, indicates that the relevant categoriza-
tions are linguistic, and thus social: the fact that they highlight specific facets
of experiences, abstracting away from the totality of their complexities; the fact
that they isolate eventualities from the very entities that embody them; the fact
that they assign stereotyped roles to these entities; the foundational fact that the
classes are digitally, as opposed to analogically, demarcated. Three additional ar-
guments are especially important. First, as many constructionists have explicitly
shown, the sentences which turn out to be ungrammatical because of the semantic
incompatibility between their parts—sentences like “*the stick hit,” or “*I broke
him on the leg”—are nevertheless perfectly comprehensible in terms of general
meaning: we can tell what they are supposed to mean, but we also feel that this
is not the way to say it in English. If the problem with these sentences is, at the
very same time, language-specific and semantic, then the semantics should be
language-specific. There seems to be no way around this conclusion.
Second, one of the most impressive facts about the categorizations is that there
are so few of them. This fact, too, has been discussed by many, most importantly
9780190256623-Dor.indb 133 03/06/15 6:18 PM

by Steven Levinson (1997) and William Frawley (1992). The number of catego-
rizations that seem to determine syntactic distributions in languages is surpris-
ingly small. There is a huge number of semantic categorizations one could think
of that play an important role in the way we conceive of the world but do not
seem to be syntactically relevant in any language. Consider, for example, the dis-
tinctions between friend and foe, important and trivial, expensive and cheap, en-
dangered and safe—and so on. None of these seem to play a role in the syntax of
languages. There does not seem to be anything in experience that would explain,
on the basis of the way we understand the world, why certain categories turn
out to determine syntactic patterns and others do not. The categorizations that
turn out to be grammatically relevant are not necessarily the most important for
experiencing, or the easiest to experience. They are the categories most required
for the successful instruction of imagination.
The third and most important indication that the categorizations are lin-
guistic is the fact that they capture classes of verbs, not the verbs themselves.
The differences in meaning between the members of the same class (the differ-
ences between break, fold, and bend, for example) seem to be invisible as far as the
grammatical facts are concerned. The forms provided by language in the second
stage of message construction are based on the signified chosen to fill-in the
eventuality slot in the message kernel. The signifieds break, fold, and bend
point at different experiential clusters, and are thus separate from each other
on the symbolic landscape. The protocol, however, is not interested in the full
meaning of the signified. It is only there to make sure that the speaker meets the
minimal specificity requirements established by the community, and provides
the listener with the right material to be able to reconstruct the intended experi-
ence. Break, fold, and bend do have different inherent meanings, but they share
the same requirement for additional material: specify the entity undergoing the
change-of-state.
There is a radical hypothesis lurking here that I should explicate: I do not see
that there is any other cognitive or communicative situation where such require-
ment is even an issue. Experiencing the eventuality of breaking is at the same time
experiencing the entity being broken, and the same is true of the experiential
communication (by demonstration) of such events. The entire structure of the
requirement—if you specify x, make sure you specify y—seems to be uniquely
linguistic. Could it be, then, that the semantic categorizations themselves emerge
only for language? Could it be that we do not need to see the similarity between
break, fold, and bend when we are not experiencing-for-instructing? Could it be
that other animals do not need it at all? If this is indeed the case, it may provide
us with a major key to the technologically driven co-evolution of human cogni-
tion and language: an emergent requirement for instructive communication that
9780190256623-Dor.indb 134 03/06/15 6:18 PM

forces the emergence of an entirely new view of the world. This way or the other,
the entire set of arguments presented above strongly indicates that construction-
ist theories do not merely describe sets of systematic relationships between lin-
guistic form and general meaning: they are extended investigations into the very
essence of language as a social system. Syntactic analysis is a social science par
excellence.
All of which leads us to the next issue: constructionist theories are formu-
lated in terms of cognitive representation. The semantic coherence principle, for
example, in its different versions, is thought of as a fact (or a hypothesis) about
the logic of the arrangement of linguistic representations in the mind. The prin-
ciple, in other words, is taken to be a component of a grammar: a functionally
oriented, semantically based grammar—but a grammar nonetheless. Once we
think of the principle, however, as a prescriptive convention (as the demand to
connect elements to each other only if they are semantically compatible), we can
easily see that it captures the very essence of the process of linguistic production
as it is governed by the conventions: it is exactly what happens when we produce
grammatical sentences. The principle does not describe a state of correspond-
ence between meaning and form. It describes the process in which form (a linear
arrangement of signifiers) is produced on the basis of meaning (a hierarchical ar-
rangement of signifieds).
Let us take another look at the examples in (1)–(3). Instead of devising a
grammar that would distinguish between them, let us try to reconstruct the pro-
cesses of their production. The message kernels for the sentences involving the
verb break should be:
(1a’) ASSERT(John)(break)
(2a’) ASSERT(stick)(break)
(3a’) ASSERT(I)(break)
In all these cases, the speaker is presented by language with the set of extensions
of break. We have already seen the illustration shown in Figure 7.1 in the previ-
ous chapter.
In (1a’), then, the speaker assigns the signified stick as the obligatory patient,
the signified John (which is already the topic-entity) as the agent, and the sig-
nified past as the tense. In (2a’), the speaker assigns the signified stick (which is
already the topic-entity) as the obligatory patient, and past as tense. The agent is
not obligatory here, and for a very good reason: the event signified by break is of
the type change-of-state, and the conventions obligate the speaker to specify the
identity of the entity that changed its state, the patient, and nothing else. In (3a’),
the speaker cannot specify the complex signified possessed body-part (John on
9780190256623-Dor.indb 135 03/06/15 6:18 PM

* patient _____
agent _____
instrument _____
goal _____
Extensions of event-type (break): reason _____
manner _____
* tense _____
time _____
place _____
figure 7.1
John’s leg) as patient, because the complex signified is a variation on the semantic
notion of theme—it is not semantically connected to change-of-state signifieds
on the symbolic landscape in the first place. (3a’) is thus an ungrammatical sen-
tence because the message it signifies is not made possible by the conventions of
message construction. The basic messages produced from (1a’) and (2a’) are then
developed into their complete messages (we have already seen how this works),
copied into their respective utterances and executed. The end results are the full
sentences (1a) and (2a).
The message kernels for the sentences involving the verb hit should be:
(1b’) ASSERT(John)(hit)
(2b’) ASSERT(stick)(hit)
(3b’) ASSERT(I)(hit)
The set of extensions presented to the speaker, in these cases, would be the one
associated with hit (Figure 7.2).
Two elements are different here: first, the set of extensions specifies an oblig-
atory theme, rather than a patient. Second, the set of extensions obligates the
speaker to specify either an agent or an instrument—and for a very good reason:
hit is a surface-contact verb, which means that it necessarily involves the con-
tact between two entities. One is the theme, and the conventions obligate the
speaker to specify what type of entity the second one is. This, again, is a prescrip-
tive demand: “Whenever you tell us about a hitting event, make sure that you
tell us what hit what.” In (1b’), then, the speaker assigns the signified stick as the
obligatory patient, the signified John (which is already the topic-entity) as the
9780190256623-Dor.indb 136 03/06/15 6:18 PM

* theme _____
agent _____
one of these
( obligatory )* instrument _____
goal _____
Extensions of event-type (hit): reason _____
manner _____
* tense _____
time _____
place _____
figure 7.2
obligatory agent, and the signified past as the tense. In (2b’), the speaker assigns
the signified stick (which is already the topic-entity) as the obligatory patient,
past as tense, and nothing else. The speaker thus fails to specify an obligatory
extension, and the sentence cannot be produced. In (3b’), the speaker specifies
the theme with a complex signified whose semantics is thematic. The process
launched with (1b’) and (3b’) is then continued, to produce the sentences in (1b)
and (3b).
Each grammatical sentence, then, is the product of a specific convergence of
prescriptive conventions. The sentence in (1a), “John broke the stick,” is what it
is because (i) the signified break requires, by convention, agents and patients; (ii)
John is a suitable agent; (iii) the stick is a suitable patient for break (very specif-
ically so, unbreakable entities would not do); (iv) the conventions of language
obligate the speaker to specify a topic-entity; (v) John is a suitable topic-entity; (vi)
past is a suitable tense; (vii) each of these signifieds is by convention marked by a
specific signifier; and (viii) the linearization conventions of the language specify
how these signifiers should be arranged in linear order. Complex utterances are
the products of complex convergences of conventions. Let me, then, try and dem-
onstrate how this perspective may help us resolve the fourth and last problem
associated with the autonomy claim: the problem of long-distance dependencies.
7.4 Island Constraints and the Convergence of Conventions

The problem of long-distance dependencies, also known as island-constraints,
has taken center stage in generative grammar (and as a result also elsewhere),
virtually since its inception (Chomsky 1973, Ross 1967). Island-constraints are
9780190256623-Dor.indb 137 03/06/15 6:18 PM

usually classified as either strong or weak (Szabolcsi 2006). Weak islands are
highly variable and easily violable: as Hofmeister and Sag (2010) show, on the
basis of experimental data, their patterns of acceptability are correlated with
different determinants of processing effort. Strong islands are exceptionally
robust, which is why they have always provided linguists with the toughest
challenge. Here, then, I will only discuss one type of strong island. Consider
the contrasts between the grammatical and ungrammatical questions in (4)
and (5). There seem to be serious constraints on the formulation of questions
of this type:
(4) a. Who did the girl hit? (Answer: The girl hit the boy who delivered the
pizza.)
b. *What did the girl hit the boy who delivered? (Answer: The girl hit
the boy who delivered the pizza.)
(5) a. Whom did your obsession with Madonna annoy? (Answer: Your
obsession with Madonna annoyed your father.)
b. *Whom did your obsession with annoy your father? (Answer: your
obsession with Madonna annoyed your father.)
Generative grammarians have always taken observations of this type to be the

best demonstrations of the autonomy of syntax. The constraints involved in these
observations were thus described and explained as purely formal limitations on
the possible arrangements of syntactic elements with respect to each other. Very
informally, the issue was formulated in terms of constraints on transformations:
certain structural positions in sentences disallow the movement of words from
within them to the outside. They are islands. In (4b), the noun-phrase the pizza
is structurally embedded within another, larger noun-phrase, the boy who deliv-
ered the pizza, which is an island. The same is true of Madonna in (5b), which is
embedded within the noun-phrase your obsession with Madonna. Over the years,
generative grammarians have developed increasingly complex formal theories
designed to explain all this, which were unfortunately only capable of capturing
(some of) the phenomena with formal stipulations that did not seem to be inde-
pendently motivated in any real sense.
Functionalists, cognitivists, and pragmaticists have attempted to find more
substantial ways to explain the relevant phenomena, proposing two types of
theories. Deane (1991) attempts to explain the phenomena, much like those
of weak islands, in terms of the cognitive task involved in language processing.
He suggests that restrictions on transformations are correlated with the cogni-
tive notion of attention: when listening to question sentences, and other such
complex structures, listeners have to simultaneously attend to two parts of the
9780190256623-Dor.indb 138 03/06/15 6:18 PM

syntactic structure, and figure out which part of the sentence the question word
stands for. This strains their limited working memory. The constraints on trans-
formations thus ensure that sentences are only generated if they are not too
complicated to process, if their structures do not overstrain the attention of the
listener. The second type of theory, most clearly articulated by Erteschik-Shir
and Lappin (1979) (see also Erteschik-Shir 2007), argues that constraints on
transformations are correlated with the pragmatic notion of dominance: a syn-
tactic element in a sentence is dominant if, by uttering the sentence, the speaker
intends to direct the attention of the listener to the meaning of the relevant syn-
tactic element. Syntactic elements can be moved away from their original syntac-
tic positions only if they are potentially dominant.
These two theories, then, attempt to explain the relevant syntactic facts on
the basis of the communicative experiences of the speaker and the listener, and
because of that, they suffer from the same problems encountered by the func-
tionalist theories discussed earlier: they attempt to explain very clear, discrete
observations on the basis of analogue and variable notions such as attention
and intent, which is why they are only capable of providing first approximations
toward the phenomena, and not of capturing them at the required level of spec-
ificity. Explaining the syntactic facts on the basis of the cognitive capacities and
intentions of speakers and listeners implies that speakers and listeners with dif-
ferent capacities and intentions should have different judgments with respect
to the relevant facts. The same grammatical complexity, for example, should
be judged as grammatical by listeners with a better capacity for attention, and
ungrammatical for listeners who find the task more demanding. This is clearly
not the case. Most importantly, the two theories implicitly accept the generative
formulation of the entire problem, as the grammatical problem of transforma-
tion constraints. They consequently attempt to construct an experientially based
grammar for the relevant syntactic phenomena, and this, as we have already seen,
is exactly what cannot be done.
Both theories, however, do capture something significant. They are good
descriptions of the experiential difficulties involved in communicative events
in which the relevant sentences are produced. Dean’s theory describes the dif-
ficulty involved in the process of comprehension: the listener identifies the
question-word, understands that the speaker is asking something, and must
now figure out what the speaker’s question is about. Erterschik-Shir’s theory,
for its part, describes the major challenge of production: the speaker has to
select and mark the particular semantic element as the thing-to-be-asked.
What we should thus be looking for, in order to explain the phenomena, is
a set of prescriptive conventions, a set of rules for the regulation of commu-
nicative events of this type—those conventions that were socially developed,
9780190256623-Dor.indb 139 03/06/15 6:18 PM

and mutually identified by speakers and listeners, in order to overcome these

difficulties.
As a first step, let us take a deeper look into the essence of the speech-act of
asking. The speaker who asks actually does two things at the same time: he or she
performs the assertive speech-act, telling the listener about a certain eventuality,
and also the interrogative speech-act, asking the listener to indicate the identity
of a certain semantic element of the eventuality, which the speaker marks as un-
specified. Thinking about all this as a prescriptive convention, in line with the
general characterization of speech-acts suggested in chapter 3, section 3.4, the
combined speech-act may be formalized as the following requirement, made by
the community and addressed to the speaker: “If you wish to ask us something,
tell us who and what we are talking about—which type of eventuality took place
and who was involved—and make sure that you specify for us a single element of
this eventuality, and mark it as the thing you are asking about.”
There is a certain restriction involved in this requirement, which lies at the
heart of the matter. What the restriction says to the speaker, in effect, is this:
“Make sure that you ask your listener about the event you are telling about; don’t
tell your listener about one event and then ask about another.” This, to be sure,
is a very reasonable demand. Listening to someone talking about one event and
then asking about another is confusing enough when the talking and the asking
take place one after the other. It is indeed a challenge (in terms of attention) to
keep track of the speaker’s intentions in this type of situation. It is even more
difficult, however, to do this when the two speech-acts take place at the same
time—and this is exactly the essence of the combined speech act of asking. The
conventional restriction, then, is there to minimize this confusion. It is pre-
scribed for all speakers, at all times, whether or not they (or their listeners) actu-
ally find themselves straining their capacities or losing sight of their intentions.
This is already familiar: the restriction bears a striking resemblance to a certain
essence in the experience of communication, but it is nevertheless, by its very
functional definition, autonomous from this experiential essence. It is not there
as a reflection of this essence. It is there to regulate it.
As a first approximation, then, all this translates into the following charac-
terization of the message kernel for asking:
ASK(unspecified element of eventuality)(topic)(eventuality)
This message kernel allows the speaker, for example, to formulate such kernels as:
ASK(reason)(John)(break)
ASK(patient)(John)(break)
9780190256623-Dor.indb 140 03/06/15 6:18 PM

ASK(agent)(unspecified)(break) (a question about the agent, which is the

topic)
ASK(time)(John)(break)
And so on.
As strange as this may seem, this very simple characterization already ex-
plains the facts in (4) and (5), and a host of similar observations. Let us take
another look at our examples. The message kernels for the grammatical sentences
in (4a) and (5a) are the following (the experiencer in (5a’) is the thematic role of
the cognitive entity that goes through an internal change-of-state as a result of
the eventuality):
(4a’) ASK(theme)(girl)(hit)
(5a’) ASK(experiencer)(obsession)(annoy)
These are good kernels. They may be developed into their respective basic mes-
sages, and then into their complete messages. The signifieds of the complete mes-
sages may then be replaced by their respective signifiers (including the signifier
who for the unspecified theme and the unspecified experiencer), and these signi-
fiers may then be linearly ordered—with an additional linearization convention
that specifies that the question-signifier should be positioned as the first word of
the sentence. Thus (4a) and (5a) are judged as grammatical because they can be
produced in accordance with a certain convergent set of conventions, some of
which are specifically about the speech-act of asking.
The message kernels for the ungrammatical (4b) and (5b), however, directly vi-
olate the entire prescriptive principle of the speech-act. The speaker is supposed to
tell the listeners about an eventuality and ask about one of the semantic elements
involved in that same eventuality, but the semantic elements that are supposed to
be the things asked about in (4b) and (5b) are not involved in the relevant even-
tualities at all. In (4b), the speaker tells of an eventuality of hitting, but the pizza
is not involved in it. The pizza is involved in another eventuality, that of deliver-
ing, which is part of the additional information that the speaker was allowed to
specify about the boy, the specified theme of the hitting event—in the answer sen-
tence. As we have seen, language allows the speaker, whenever he or she specifies
an entity in the process of message construction, to further specify whether the
entity also participated in another eventuality, and in the answer sentence to (4b),
the speaker decided to specify deliver. None of this, however, can happen in the
parallel question sentence, because the major prescription governing the produc-
tion of questions is the specification that the thing asked about should play a the-
matic role in the asserted eventuality. The pizza is not involved in the first place.
9780190256623-Dor.indb 141 03/06/15 6:18 PM

The same is true of (5b). The speaker tells about the eventuality of annoy-
ing, but Madonna is not involved in this eventuality at all. What is involved in
the eventuality of annoying is the obsession. Madonna is the object of the ob-
session, not of the annoyance, and she gets into the process of production only
in the answer-sentence, where she is specified as additional information about
the obsession. In the question-sentence, she is not part of the process at all. Ma-
donna, and the pizza, in other words, are just like all other things in the world
not involved in the eventuality chosen by the speaker in the message-kernel: the
conventions of language prescribe that they cannot be asked about. Madonna
and the pizza can only be asked about if they themselves participate in the even-
tuality asserted by the speaker. The questions in (6) and (7), for example, are
producible and thus grammatical, because the asserted eventualities are those of
the delivering and the obsessing:
(6) What did the boy whom the girl hit deliver?
(7) Who did you obsess with in a way that annoyed your father?
Here, however, questions about the girl and the father are not producible, be-
cause they are involved in the eventualities of hitting and annoying:
(8) *Who did the boy who hit deliver the pizza?
(9) *Who did you obsess with Madonna in a way that annoyed?
What all this means is that the original formulation of the question as a question
about transformations is fundamentally wrong. Each of the question-sentences
we have looked at is indeed related to a corresponding answer-sentence, but the
relationship is not structural—it is semantic. The pairs of sentences are, quite
simply, questions and answers about the same eventualities. They are related to
each other because they are founded on related semantics. They are produced
independently, however, and their production processes are governed by differ-
ent sets of conventions. The set of conventions for the production of questions is
more restrictive than the parallel set of conventions for the production of asser-
tive sentences, because it includes the additional restriction on the semantics of
the unspecified element, the thing asked-about. The conventions of the language
allow the speaker to tell about one event and then tell about another in the same
message (“the girl hit the boy who delivered the pizza”), but they do not allow
the speaker to tell about one event and then ask about another (*“what did the
girl hit the boy who delivered”). Because of that, the set of questions produci-
ble by language about a certain eventuality is always smaller than the parallel
set of assertions about the same eventuality. This is a semantic fact: it is a fact
9780190256623-Dor.indb 142 03/06/15 6:18 PM

about the linguistic conventionalization of communication. If we mistake it for

a structural fact, and think about the question-sentences as generated, by trans-
formation, from their assertive counterparts, we find ourselves having to explain
why some semantic elements that may appear, in specified form, in the answer-
sentences, cannot re-appear, transformed into question-words, in the question-
sentences. This seems to be the wrong way of thinking about the entire problem.
Let us take this one step further. The observations we are looking at have
come to be called long-distance dependencies, because it seems that some ques-
tion-sentences (and other types of sentences) that are more complicated than
those in the ungrammatical examples above, are nevertheless perfectly gram-
matical. Consider, for example, the example in (10):
(10) What did Eve say that Mary thought that Bill told her that Dean was
going to buy in the supermarket?
At first sight, this sentence seems to go against everything we have said so far.
The asserted event in the sentence is that of saying, but the thing asked-about,
the thing bought in the supermarket, is involved in the event of buying. To un-
derstand what is happening here, let us take another look at the line of events in
the sentence—saying, thinking, telling, and buying. All but the last denote what
we may call epistemic events: events in which a cognitive entity makes a judgment
about an event (Dor 1996). Other epistemic events are remembering, informing,
knowing, doubting, explaining, assuming, believing, and so on. What all of these
eventualities have in common is a certain stereotyped semantic relationship be-
tween a cognitive entity and an event, the factuality of which the cognitive entity
asserts, or doubts, or affirms, and so on. As it turns out, then, sentences like (10)
may be lengthened at will, but only as long as all the eventualities involved, apart
from the very last, are epistemic events. A sentence like (10) with other types of
eventualities would have to look like (11):
(11) *What did Eve say that Mary broke the table that Bill told her that Dean
was going to buy in the supermarket?
Or like (12):
(12) *What did Eve say that Mary met Bill who walked with Dean who was
going to buy in the supermarket?
These just don’t work. All the events apart from the last one have to be epistemic.
They specify the speaker’s assertion: “Based on what Mary thought that Bill told
9780190256623-Dor.indb 143 03/06/15 6:18 PM

her, I ask you what it was that he was going to buy in the supermarket.” The mes-
sage kernel for asking, that I presented before as ASK(unspecified element of even-
tuality) (topic)(eventuality), should thus be slightly more complex. It should allow
the speaker to choose an unspecified element of any eventuality—as long as the
speaker makes sure that it is the asserted eventuality. The message kernel is thus:
ASK(unspecified element of asserted-eventuality)(topic)(eventuality). It allows for
the production of sentences like (4a) and (5a), and also for the production of (10),
but it nevertheless disallows the production of all the ungrammatical sentences
we have looked at. The fact that the protocol allows for recursive usage of epi-
stemic signifieds allows speakers to produce utterances not just about what they
themselves think or believe, doubt, or conjecture—but also about what (they
think) other people do. Speakers may assert more things, tell their listeners more
about the epistemic sources of what they say, and, probably most importantly,
say things about what other people think about the world, without necessarily
committing themselves to the truth of their thoughts. A speaker may say: “Eve
said that Mary heard Bill saying that Dean was buying a loaf of bread in the su-
permarket, but I don’t believe that.” A speaker may ask “What did Eve say that
Dean bought in the supermarket?” without committing to anything apart from
the epistemic chain that leads from Bill all the way to him or herself. The message
kernel in its corrected formulation, then, is a complex compromise between the
interests associated with speakers and listeners, which crystallized, over a certain
period of time in history, as a set of conventions for the formulation of questions.
All this is obviously a demonstration, not an attempt at a full analysis. There are
also many other types of islands, requiring other types of explanation. What I
do hope to have achieved here is to show how the theory allows for a new prin-
cipled perspective on the question of long-distance dependencies, a perspective
that may now be further developed and tested.
7.5 Conclusion
The generative argument about syntactic complexity has always been accom-
panied by a sort of meta-observation, intended to highlight the importance of
the relevant empirical facts for the overall theory of language. Native speakers,
so it was claimed, are genuinely surprised when presented with ungrammatical
sentences such as (4b) and (5b). They testify that they have never come across
anything similar, but they nevertheless immediately know that the sentences are
ungrammatical. What this was supposed to imply was that the speakers could
not have learned about the ungrammaticality of these sentences from their past
experiences, which means that they must have grasped the ungrammaticality on
the basis of their innate knowledge of language. The entire line of argumentation
9780190256623-Dor.indb 144 03/06/15 6:18 PM

is seriously flawed, but the meta-observation is in some strange way absolutely

right: we really do not come across ungrammatical sentences like (4b) and (5b).
This is a significant fact not because it indicates that language is innate, but be-
cause we do come across so many other types of ungrammatical sentences. Many
types of ungrammaticality are familiar to us, but (4b) and (5b) are not. They
really are surprising.
What all this testifies to is this: in our regular lives as speakers, not as lin-
guists, all the sentences we ever encounter as candidates for grammaticality
judgment are produced by speakers who intend to say something. We are thus
familiar with two types of ungrammaticality. Speakers may either intend to
follow the conventions to the letter, but fail to do so, or they may deliberately
break the conventions in order to communicate something (just like in Grice’s
maxim violations). Characteristically, when speakers fail to follow the conven-
tions (because they only have partial command of the language, or because they
are exhausted, and so on), they manifest a certain reduced capacity with respect
to the entire cluster of conventions they have to follow, or at least with respect
to a major subset. What we generally find in situations of this type are instances
of ungrammaticality spread all over the sentence. We rarely, if ever, come across
situations in which speakers manage to produce complex sentences, following
all the conventions of the language to the letter—except one. A speaker who pro-
duces a sentence like (4b) or (5b) by mistake must have perfect command of the
entire cluster of conventions involved in the production of such sentences—and
a total misunderstanding of the conventional nature of the speech-act of asking.
This is a highly unlikely state of affairs.
The other type of ungrammaticality we are familiar with is the one produced
deliberately by the speaker. It is, however, very difficult to think of a commu-
nicative intent that would be served by sentences like (4b) and (5b)—unless, of
course, we think of them as instances of poetry. Poetry is one of the strategies
adopted by innovative speakers in their struggle to push the expressive envelope
of language further: it is, by its very definition, a deliberate attempt to violate the
social conventions of the language, to go beyond them. It is actually quite easy
to imagine (4b) or (5b) appearing as lines in an avant-garde poem, and it is also
easy to see that, as part of a poem of this type, they would be supposed to surprise,
shock, disturb the conventional order of language and thus reveal its constrain-
ing nature. This is exactly what happens to speakers who are asked by linguists
to provide a judgment about a sentence of this type: the sense of shock is actually
more acute, because the sentences are not presented as poetry—but as potential
instances of regular speech. Innateness has nothing to do with this.
The syntactic complexities of language do not reflect a level of syntactic
knowledge given to us by our genes, and the hierarchical syntactic tree—which
9780190256623-Dor.indb 145 03/06/15 6:18 PM

may still be used as a convenient tool for description—is essentially an epiphe-

nomenon. It conflates a hierarchical semantic dimension with a linear struc-
tural dimension. What the two dimensions have in common—together with all
other parts of language—is the fact that both are governed by sets of mutually
identified prescriptive norms. The norms of linearization, moreover, are corre-
lated (to variable degrees in different languages) with the hierarchical semantics
of the message: linear strings are built from complete messages, and the other
way around. Observations of syntactic complexity thus reflect the prescriptive
meanings of the symbolic landscape—not the experiential meanings of private
cognition. This is why they are autonomous, just as Chomsky claimed. They are
autonomous, however, not because they are cognitively unique, but because they
are collective. Syntactic analysis, the crown jewel of cognitive science, is a funda-
mental branch of the social sciences.
9780190256623-Dor.indb 146 03/06/15 6:18 PM

8 T h e U n i v e r s a l it y o f Di v e r s it y
The last fifty years of linguistic research have seen a gradual shift of
perspective on the all-important question of the relationship be-
tween universality and diversity in language. The Chomskian para-
digm promised to uncover the common structural foundations of all
languages (the veritable UG), but the hope has gradually diminished
in most linguistic circles, and much more attention has been given
to the enormous variability exhibited by the world’s languages. The
far-reaching implications of the shift were explicated in Evans and
Levinson’s (2009) bold attack on “the myth of language universals,”
probably the most important linguistic paper of the new millennium’s
first decade. Reviewing the vast literature in typology and descriptive
linguistics, Evans and Levinson show “how few and unprofound the
universal characteristics of language are” (p. 429) at the levels of pho-
netics, phonology, morphology, syntax, and semantics. They decon-
struct the generative arguments concerning the universality of such
notions as constituency and recursion, arguments that were based on
the examination of a handful of languages, very often only English.
Instead of the reductive, philosophically based picture of UG, they
present a sophisticated, empirically based and multi-dimensional evo-
lutionary conception of the emergence of linguistic patterns, as “stable
engineering solutions,” from the interaction between many different
motivational causes—functional, communicative, cognitive, and
systemic. The debate that followed the paper, in the Behavioral and
Brain Sciences publication, in Lingua, and elsewhere, testified to the
crucial significance of the issue, and showed how entrenched the
belief in universality is among scholars of language, even those who
have already rejected everything else that Chomsky ever said. The dis-
missive tone of some of the commentators reminded many readers of
earlier times, when the myth was at its hegemonic peak. Fortunately,
we are no longer there.
I accept Evans and Levinson’s argument in its entirety. I also be-
lieve, however, that some of the objection—and some of the tone—
emerges from the feeling that the argument takes us back to square
9780190256623-Dor.indb 147 03/06/15 6:18 PM

one, leaving us with no answer to the most fundamental question of all. The idea
of language universals may be a myth, but the universality of language is not. It
is a fact. Without universals, how can we conceptualize the universality? The
issue can be demonstrated quite dramatically with the opening citation ironi-
cally chosen by Evans and Levinson for their paper, a quote from Pinker (1994)
who says that “according to Chomsky, a visiting Martian scientist would surely
conclude that aside from their mutually unintelligible vocabularies, Earthlings
speak a single language” (p. 232). Here is Chomsky himself in a paragraph
quoted by Sandler and Lillo-Martin (2006, p. 5):
As human beings, we are naturally interested in the differences among

humans; we take the similarities among humans for granted. So, we’re
interested in the way humans look different from one another and [the
way that] their faces are so different, and so on. But from the point of
view of, say, some Martian, we would all look alike. Just as from our
point of view all frogs look alike. Now from the point of view of the
frog, they look, I’m sure, very much different from one another, be-
cause they’re interested in the differences among frogs. We just notice
the overwhelming respects in which they’re similar. If we can make the
leap of the imagination that enables us to look at ourselves the way we
look at other organisms, we will quickly discover that we’re remarkably
alike.
On the one hand, then, the Martian argument exposes everything that is
wrong in Chomsky’s rationalist perspective, based as it is on a “leap of imagina-
tion”: the tension between his insistence that we humans, driven by our natural
interests, should forever be suspicious of the way we perceive the world, and
his calm conviction that we can simply tell how things really are by imagining
how non-existent extra-terrestrial organisms, devoid of any natural interests of
their own, would objectively see things from very far away; the idea that look-
ing at frogs somehow turns us into objective perceivers; the idea that we are
all most impressed by the similarities between frogs (quite obviously, anybody
who knows a bit more about frogs than Chomsky seems to is most impressed
by the differences). All these imply so much disdain for everything empirical
in the human sciences, such conviction that what can never be refuted must be
true, that it leaves no room for diversity in the picture of human language. On
the other hand, the Martian argument captures something very deep that does
not actually need the Martian help: we, as existent, terrestrial human organ-
isms, feel very deeply that when we speak our different languages we are actu-
ally doing the same thing. We have no problem identifying people as speaking
9780190256623-Dor.indb 148 03/06/15 6:18 PM

The Universality of Diversity • 149
to each other, as opposed to doing other things, even when we can understand
nothing of the conversation. There is a universal essence there that we intu-
itively recognize: it does tell us, at some basic level, that we are “remarkably
alike” after all.
Evans and Levinson persuasively put to rest Chomsky’s suggestion that
the universal essence lies in the set of structural universals shared by all lan-
guages, but they leave the question open. Their move, as Goldberg (2009)
writes in a positive commentary, is first and foremost an attempt to break
away from essentialism: “‘Universal Grammar’ (UG) provides an essential-
ist answer to the question, What is human language? . . . (a) UG is assumed
to offer an underlying invariant essence of language, although surface reali-
ties are recognized to differ; (b) UG is assumed to have clear boundaries:
all human languages, and no other communication systems, are assumed
to share UG; (c) UG is viewed as unchanging over time; and (d) UG is as-
sumed to be innately determined. . . . It is time we let go of the essentialist
mindset and embraced motivation as linguistic explanation” (p. 456). This
anti-essentialist motivation is shared by many in today’s literature: there is
not much choice when the hegemonic perspective in the field is such a her-
metically closed, irrefutable, essentialist theory as Chomsky’s. The emergent
result of the anti-essentialist effort, the evolutionary, motivational view, is a
great achievement—superior to the generative perspective on all fronts. At
the same time, however, the approach pays the price of its own commitment:
it does not have a good answer to the essentialist question, what is human
language? Evans and Levinson, like Goldberg and others (cf. Sinha, 2014)
conceptualize language as a bio-cultural hybrid, “a biological system tuned
to a specific linguistic system, itself a cultural historical product” (p. 446).
The biological system “is designed to deal with the following shared Hocket-
tian design features of spoken languages: the use of the auditory-vocal chan-
nel with its specialized (neuro)anatomy, fast transmission with output-input
asymmetries, . . . multiple levels of structure (phonological, morphosyntactic,
semantic) bridging sound to meaning, linearity combined with nonlinear
structure (constituency and dependency), and massive reliance on inference.”
This is a very sophisticated and very realistic description of language, but it
does not answer the question of essence.
In this chapter, then, I would like to show how the theory of language as
a socially constructed technology for the instruction of imagination helps
take Evans and Levinson’s argument another step forward. I will suggest a re-
formulation of the entire question, beginning with my essentialist answer to the
question of what human language is, and then deriving the required levels of di-
versity from that essence.
9780190256623-Dor.indb 149 03/06/15 6:18 PM

8.1 Universality Is a Social Fact

The first point I would like to make may strike some readers as counter-intuitive,
but I am convinced that it should serve as the foundation for the entire discus-
sion: we have been looking for the universality of language in the wrong place.
The universality of language is not a cognitive fact. It is social. Language is a uni-
versal fact about humanity not in the sense that all human minds have language,
but in the sense that all human societies have it.
This is not a call for the replacement of one meta-theoretical presupposition
with another. It is, quite simply, the replacement of an invalid empirical obser-
vation with a valid one. Not all human minds have language, but all societies
do. There are many human minds without language: the minds of all human
children before they acquire language; autistic minds (not all of them); apha-
sic minds; the minds of people with a wide range of brain injuries and affective
pathologies; and most importantly—the minds of human individuals who have
not been exposed to the social activity of language at the right age. We may have
good stories to tell about each of these, but the crucial fact is that we need these
stories. As far as human societies are concerned, no stories are required. There
is not a single human society that has no language. All human societies use dif-
ferent variations of the same technology, locally designed by cultural evolution
for the universal function of the instruction of imagination. This is an absolute
universal. We should start here.
From this fundamental assertion emerges a distinction between two levels
of description: essentialist and emergent. At the essentialist level, where all lan-
guages of all societies are treated as one and the same, the essence of language
implies a long and substantial set of constitutive and absolute universals, true of
all languages. At the emergent level, everything else about language is allowed to
be as diverse and multi-causal as society itself. Most importantly, the extent of
diversity at the emergent level is predicted from the set of absolute universals at
the essentialist level.
8.2 Essential Universals and Emergent Properties

Essential universals are definitional in the most concrete sense of the word:
we identify a certain communication technology as a language if and only if
manifests the entire set of constitutive properties. All languages are socially
constructed communication technologies, designed through the collective, iter-
ative process of experiential mutual-identification for the specific function of the
instruction of imagination. All languages comprise a large symbolic landscape
with triple-natured signs, and a protocol with prescribed procedures (strong or
9780190256623-Dor.indb 150 03/06/15 6:18 PM

weak, stable or flexible) for the production and interpretation of utterances. All
languages translate experiential intents into formal messages and translate mes-
sages into experiential interpretations, thus mediating between the private and
the social. All languages require their speakers to experience for instructing,
and all of them are entangled in a bi-directional spiral of influence with their
speakers’ private experiences. In all languages, utterances reflect (not necessarily
syntactically) the hierarchical nature of the complete message. In all languages,
grammatical patterns are determined by clusters of prescriptive conventions. All
languages necessitate a long process of acquisition and practice. All this is very
general, but also, at the same time, very particular. All the languages of the world
are technologies of the very same type. Everything that is absolutely universal
about human languages is a constitutive functional property of the technology
as such.
Note that this move already solves the most fundamental problem that Evans
and Levinson (2009) and Levinson and Evans (2010) face in their discussion.
Many of their commentators wonder why they do not consider some of Hock-
ett’s design features as universal properties of language. The authors rightly ex-
plain that “if questions of what is universal in language are to have any empirical
content, they cannot at the same time be definitional of what language is, or
their presence is merely tautological—a bit like ‘discovering’ that it is universal
for even numbers to be divisible by two” (p. 13). Within my perspective, how-
ever, the tautology disappears. Those design features that are universal—duality
of patterning, semanticity, productivity, and discreteness—are no longer def-
initional: they emerge from the definitional characterization of language as a
technology.
At the definitional level of essential universals, then, all languages are the
same. Beyond this level, however, everything may be as diverse as required. Em-
pirical research in linguistics should thus always begin with the default assump-
tion of absolute diversity. Unless otherwise demonstrated, all languages are by
definition different from each other. Every particular language (every dialect,
every variation) is a unique system of instructive communication. Every language
embodies, in each and every one of its conventions, a totally unique social his-
tory of negotiation and struggle, in a unique community, with uniquely different
individuals, at different times and places, under uniquely different social condi-
tions. Every component of every language—every word, every category, every
prescriptive norm—emerged and came to be established as part of the effort to
allow for the communication of different types of experiences across different
types of experiential gaps. Every language reflects a different set of communi-
cative interests and compromises, a different set of realities and perceptions, a
different history of innovation. Every language is an entity sui generis—a totally
9780190256623-Dor.indb 151 03/06/15 6:18 PM

different variation on the very same topic. There is just one universal constraint
on diversity: all the variable properties of all languages should be analyzable
as “stable engineering solutions” to the foundational problems of instructive
communication.
It is important to see that the patterns of similarity between languages found
in the typological literature—even those properties that would eventually turn
out to be true of all languages—all these find their proper place in this picture.
As opposed to a theory based on universality, which by its very definition bans
whatever is variable, a theory of variability happily accepts all existing patterns of
variability. Not everything can be equally different from everything else: if that were
the case, the linguistic system would again be governed by a uniform principle.
Every system based on variability always includes something that is universal.
Among all the observable patterns of difference between languages, there may
also be some in which the pattern of difference is partial, or even total similar-
ity. Such patterns, however, are not necessarily foundational of the system. They
are not common denominators. They reflect emergent commonalities between
languages, points where different languages found similar solutions to similar
problems—points of convergence between the evolutionary histories of the lan-
guages. There is much to be learned from these convergences about language and
its evolution, but in order for them to be understood, they should be analyzed
against the background of diversity.
8.3 The Human Potential for Linguistic Diversity

Nothing is lost if we adopt this perspective, but the potential dividends are enor-
mous. As long as we look at the languages of the world from the universalist point
of view, we have no choice but to define our task in the most counter-empirical
way possible—as the search for the human potential for language. The question
is not why existing languages are the way they are, but what a possible human lan-
guage looks like. Chomsky’s innateness argument is sometimes misunderstood
as the claim that some of the potential for knowledge is already actualized in
the minds of children prior to their encounter with real languages—that chil-
dren already know some of the things that the empiricists say they should learn
from their linguistic environment. This, however, is not what Chomsky says.
His theory is not a theory of learning. What he does is reverse the relationship
between potentiality and actualization: those formal components of linguistic
knowledge that Chomsky claims are innately given to us as human beings are
not a priori actualizations—they are the potential for language that all humans
share. They are the foundational capacity. They define what a possible human
9780190256623-Dor.indb 152 03/06/15 6:18 PM

language may look like. All those languages out there, in the real world—they are
the actualizations of this potential. Each and every one of them is a variation of
human language. Languages emerge from knowledge, not the other way around.
Thus, for example, the conception of Principles and Parameters (Chomsky and
Lasnik 1993) does not contend that children already have a parameterized grasp
of what different languages such as English, Japanese, and Italian look like. The
theory contends that English, Japanese, and Italian are three of the possible ac-
tualizations of the parameterized capacity that makes these children what they
are as human beings. Children do not acquire language. They activate it. They
actualize their own potential for language. Language grows in them.
The goal of Chomsky’s theory, then, is the uncovering of the human poten-
tial for language, and because of that, linguistic theory is obliged to ignore the
facts of linguistic variability—not out of negligence, but as a fundamental meth-
odological principle. This is why Chomsky’s insistence that universal general-
izations may be deduced from a handful of languages, sometimes a single one,
makes perfect sense within the theory: whatever is essential about the human
potential for language must manifest itself in the essential properties of each and
every language that actually exists. There is no point in going through hundreds
of languages: we may as well concentrate on one or two (or a dozen).
This is an astonishing argument, the best Platonic argument of the twentieth
century. The next logical step, however, is even better: there is no point in trying
to verify linguistic hypotheses in large typological sub-sets of the world’s lan-
guages, because even if we could check all the thousands of existing languages,
and all the languages that there ever were, how would we know they make up a
representative sample of all possible languages? Here is Newmeyer (2005):
(T)he (tens of thousands of?) languages that have ever existed may just
be a small percentage of possible human languages, among which are
certainly those which chance, rather than principle, has decreed should
never have a native speaker. (. . .) What reason might we have to conclude
that the distribution of features that any possible available sample mani-
fests represents anything more than some accidental product of the vicis-
situdes of human history? That is, what reason might we have to conclude
that any implicational relationships we might find among these features
represent some linguistically significant fact to be explained? (p. 306)
Two or three languages are enough—but then again, not even all the languages
that have ever existed are enough. There must be a problem here. An empirical
science of language cannot take as its goal the analysis of what a possible lan-
guage might look like. There are no languages that chance decreed should never
9780190256623-Dor.indb 153 03/06/15 6:18 PM

have a native speaker. Linguistics has no other choice but to concentrate on what
there is. Between the handful of languages that are relatively well researched,
and the infinity of possible languages, there are around six thousand languages
to be analyzed—and many of them are not going to stay around for long (Crys-
tal 2002, Dorian 1989, Evans 2011b, Nettle and Romaine 2000, Skutnabb-
Kangas 2000).
It is important to see, then, that the functionalist perspective on language—
different as it is on all other fronts from the generative one—has never really dis-
tanced itself from the universalist search for the human potential for language.
The debate between both sides has centered on the nature of the potential: is it
formal or functional? internal to language or external? The assumption of the
human potential was the foundational presupposition upon which the debate
was based. Where Chomsky imagines a formal UG underlying all languages,
functionalism imagines an underlying universal set of meanings and commu-
nicative functions. There are culture-dependent meanings and functions, of
course, but they reside at the periphery. The core is universal. It includes such
functional universals as iconicity and dominance, semantic hierarchies, infor-
mation flow, topic, focus, and so on. These functional universals play the role of
external motivations: they determine the way linguistic structures develop. For
most functionalists today, this is mainly a diachronic process: the relevant uni-
versals influence the historical process of grammaticalization (Deutscher 2005,
Lehmann 2002, Traugott and Heine 1991, Hopper and Traugott 2003). The
universal functional determinants, however, do not all influence the structures
of languages in the same way. They produce potentially conflicting results, and
because of that, they compete (Bu Bois 1985). In different languages, certain de-
terminants have a greater influence than others, and the end result is a web of
partial similarities between large subsets of the world’s languages.
Here is a concrete example. Functionalism uses the notion of iconicity to
refer to the idea that structures are the way they are because they resemble their
meanings. Givón (1985) explains the logic of the notion in the following way:
“All other things being equal, a coded experience is easier to store, retrieve and
communicate if the code is maximally isomorphic to the experience” (p. 189).
Consider, then, the following iconicity principles: (i) old information appears
first in the utterance; new information appears last; and (ii) ideas that are con-
ceptually close to each other are placed together, or close to each other, in the
utterance. It is easy to see that the two principles would arrange the elements
of the utterance (as a first approximation, the words) in two different ways. The
first principle would linearize the words by novelty, the second by proximity. The
two principles, then, are in competition (and there are additional principles in
competition with both of them). In different subsets of the world’s languages,
9780190256623-Dor.indb 154 03/06/15 6:18 PM

different principles are more powerful than others, and the result is the variety
of word orders we see.
All this makes much more sense than Chomsky’s Universal Grammar. It is a
theory that respects variation, that attempts to explain diversity—not explain it
away. There is also no doubt that the relevant functional categories—including the
iconicity of novelty and the iconicity of proximity—are (and have been) important
in the lives of languages. There is, however, no reason to assume that they have
ever been universal. Functionalism attempts to explain structural phenomena on
the basis of their function as carriers of general meaning. In the terms proposed
here, the goal is to understand the way languages are shaped by experiences—the
experiences to be expressed through language, and the experiences of linguistic
communication. These experiences, however, are variable by their very definition.
The attested variability between languages is not a reflection of a deeper level of
experiential universality. Behind the structural variability that is observed on the
surface there is nothing but a deeper level of experiential variability. The ways lan-
guages eventually carry meanings—the meanings themselves, the regulation of
information flow, the semantic hierarchies in the symbolic landscapes—all these
are social compromises, emergent patterns of prescriptive order that are there in
the first place only because the very foundation of language lies in the variability
of experience.
The two types of iconicity we have discussed above have influenced the struc-
tures of languages not because all human beings share a natural tendency to
begin by talking about what is already known and then proceed to the news, and
not because we all share a natural tendency to arrange our sentences by concep-
tual proximity—but precisely because we do not. Consider children, especially
when they are excited: they begin with the news; they jump back and forth be-
tween the different facets of the experience they are trying to express. They are
not interested in iconicity. They are trying to communicate. The two types of
iconicity are prescriptive principles. They are demands made by the community
and addressed to the speakers: “Could you please tell me about the first thing
first, and then the other?”; “Can you please refer to the properties of the refer-
ent you talk about together with the referent itself, so we can tell they belong
together?” The principles are there (in those languages that have them) because
some innovative individuals realized, at certain points in the languages’ evolu-
tion, that communication would become more effective if speakers could be
made to arrange the elements of their utterances in ways that reflect something
of the order of their experiences. These innovative individuals were probably
more sensitive to iconicity than others, and they were probably more attuned to
the problems of unorderly communication: they may have experienced more dif-
ficulty than the others in understanding what was being said, and they may have
9780190256623-Dor.indb 155 03/06/15 6:18 PM

found themselves looking for ways to express something about the internal order
of their own experiences, because this mattered to them.
One way or the other, their innovations launched different social processes
in different languages, in which they had to struggle, among other things, with
other prescriptive innovations that were trying to pull these languages in other
directions. There is always competition, and in this sense the functionalist per-
spective is right, but the competition is always language-internal. The complex
web of partial similarities between the iconicity-related patterns in different lan-
guages does not result from a competition between universals; it results from
a competition between variables. Most importantly, there is no good reason to
assume that the very notion of iconicity was developed everywhere in the same
way. The distinction between new and old information, for example, has proba-
bly taken different shapes and forms in different communities throughout his-
tory. It is not obvious. We are not born with it. It encapsulates an entire socially
developed philosophy. Meaning and communication are just as variable as the
structures of languages. There would be no reason for the degree of structural
variability empirically observed had it not reflected a parallel level of variability
at the foundational level of the social negotiation and struggle over meanings
and their communication.
What all this means is that the question of the human potential for language
should be replaced with the question of the human potential for linguistic in-
novation: a question about diversity, not about the universal. The potential does
not manifest itself in those solutions found and implemented by the majority
of linguistic communities around the world—these were probably the easiest
solutions (and it should be interesting to ask why). The human potential for lin-
guistic creativity manifests itself first and foremost in the range of differences
between all those languages that reside, from the point of view of the universal-
ist paradigm, at the peripheries—the range of differences between exceptional
languages.
8.4 Could Recursion Be an Exceptional Phenomenon?

It is also important to see that this perspective carries implications for the
very distinction between regular and exceptional languages. Consider syntac-
tic recursion, for example. The idea that certain syntactic computations, such
as sentence embedding and the nesting of the possessive markers, can apply
freely to their own outputs—has featured prominently in the universalistic
discourse on language. For Hauser, Chomsky, and Fitch (2002), to which we
shall come back in chapter 10, recursion is the essential property of the Faculty
9780190256623-Dor.indb 156 03/06/15 6:18 PM

of Language in the Narrow sense (FLN), the one thing that separates human
language from all the other systems of communication in the biological world
(see also the volume of Biolinguistics 2011 on recursion). As Evans and Levin-
son (2009) show, however, syntactic recursion is actually severely restricted in
some languages, and probably missing altogether from some others. The Ab-
original language Kayardild, for example, allows for no more than one level of
embedding (Evans 1995). More dramatically, the Amazonian language Pirahã
shows no signs of recursion at all:
There are many words in English that tend to be found in sentences with
recursion. English words like “think,” “believe,” “say,” and “want” all re-
quire subordinate clauses. We see this in examples like “John said that he
was coming,” where “he was coming” is a subordinate clause. An English
speaker cannot say “I think you” without the following clause, because
that is just the way the verb “think” is programmed in English. Pirahã
lacks all words of this type. No verbs in Pirahã require a subordinate
clause. This is one way that Pirahã culture can eliminate, by lexical con-
vention, recursion from its grammar. Rather than saying “John said that
he was coming,” the Pirahãs would say “John spoke. He is coming.” That
is ambiguous, making what is in English a single sentence in effect a small
story. Or rather than say, “John thinks he is coming,” they would handle
this the same way, “John spoke. I am coming.” (Everett 2012, p. 278)
Syntactic recursion, then, is not universal. But what does that mean? How dif-
ferent should English, Kayardild, and Pirahã be from each other in order for
the first to allow for free syntactic recursion, the second to restrict it to one em-
bedding, and the third to disallow it altogether? For Evans and Levinson, the
key lies in the fact that the speakers of Kayardild and Pirahã have no problem
expressing complex ideas in non-syntactic ways: “Although recursion may not
be found in the syntax of (these) languages, it is always found in the concep-
tual structure, that is, the semantics or pragmatics—in the sense that it is always
possible in any language to express complex propositions” (p. 444). I think this
is only partially true. These languages do allow for complex propositions, but
they contain no machinery for the cyclic application of computations to their
outputs. All languages produce one utterance at a time, and what is conveyed in
English with a single, recursive utterance, is conveyed in Everett’s examples with
two simpler ones. The semantics is spread differently in English and Pirahã—
note how the two variations of the second utterance in Pirahã, “he is coming”
and “I am coming,” influence the interpretation of the verb speak—but there is
no recursion. To the extent that we think about recursion as a universal cognitive
9780190256623-Dor.indb 157 03/06/15 6:18 PM

capacity foundational for language, we are still in trouble: why is it not free in
all languages?
Within the perspective developed here, the question is no longer one of cog-
nition. The cognitive capacity for recursion may be (variably) universal.1 It may
be involved in many non-communicative human activities, and it may or may
not appear in the cognitive workings of other animals (Pepperberg 1992). The
question is technological: whether, and to what extent, the protocols of the lan-
guages of the world include pieces of machinery that allow for recursion in the
construction of the complete message. Eventually, this is a question about the
social-cultural process of mutual-identification: whether, and to what extent,
communities of speakers were systematically engaged in the effort of maxi-
mizing the amount of meaning squeezed into a single utterance. When we put
things this way, one undeniable fact all of a sudden takes center stage: “Recursive
structures appear to be far less frequent in spoken language than in written lan-
guage” (Sakel and Stapert 2010, p. 11). Spontaneous speech in English very rarely
includes more than a single embedding: spoken English is much more similar to
Kayardild than it is to written English, and Pirahã is just a more extreme version
of the same reality. In speaking, the output fades rapidly: this is Hockett’s third
design feature. There is an inherent interest to keep it as simple as possible. The
revolutionary invention of writing keeps the output in front of our eyes for as
long as we need. It is possible to squeeze much more into each and every utter-
ance. Languages with a written tradition thus include mechanisms for the recur-
sive embedding of additional messages within the same utterance, a feature that
is still used mostly in writing. As far as spoken language is concerned, Kayardild
and Pirahã may not be the exception: they may actually be good representatives
of the rule. Recursion is not a universal feature of human language, syntactic or
otherwise. It may be a very late development in the technological evolution of
some of the languages of the world.
8.5 Sign Languages: Data in Search of Theory

The sign languages used by deaf communities around the world present the uni-
versalist position with one of its toughest challenges: they do not have sound
systems. As Evans and Levinson put it, “at a stroke . . . they invalidate such gen-
eralizations as ‘all natural languages have oral vowels’” (p. 433). Instead of the
sounds of spoken languages, sign languages use complex combinations of hand
shapes and locations, arm locations and movements, and facial and bodily ex-
pressions. This is not an easy matter. Sign languages are so different from spoken
languages that they had to wait until the last third of the twentieth century to
9780190256623-Dor.indb 158 03/06/15 6:18 PM

be recognized by the linguistic community as full-fledged languages. Together

with some of their commentators, Levinson and Evans suggest that language in
general should be conceptualized as a modality general system, whose outputs
can be expressed in different modalities. In the terms suggested by Sandler and
Lillo-Martin (2006), sign languages teach us that we need to separate the code of
language from its mode: as far as the code is concerned, the two language types
are the same; the mode, however, allows for more than one option, and is thus
variable. Sandler and Lillo-Martin, working with the generative paradigm, also
attempt to explain the mode in universalist terms: the building blocks of the
utterance—hand shapes, positions, movements, and directions—are analyzed in
the terms of the phonology and phonetics of spoken languages. They quote one
suggestion according to which hand movements and hand positions should be
thought of as the vowels and consonants of sign language.
There are three problems with this approach. First, it is unclear what the uni-
versal code, language minus modality, is supposed to include. Is it that part of
language that deals with meaning? Does it include syntax, for example? Second,
and more importantly, the different options opened up by different modali-
ties are also evidenced in the code. In sign languages, for example, whenever an
entity is mentioned for the first time in the conversation, the speaker moves the
hand signing its name to a certain point in the space between both interlocutors,
where the entity is “located.” From that moment on, references to this entity
will be made by pointing at the location. As the conversation goes on, the space
comes to be filled with all the different entities mentioned, and utterances refer-
ring to them are constructed with the locations as nouns. This allows for simul-
taneous signing: signifiers can be produced at the same time by both hands, the
face, and the body, and entire sentences may be communicated in one stroke.
Consider, for example, a sentence indicating that two of the referents already
mentioned in the conversation did the same thing, something like the English
sentence “they fell asleep.” In many sign languages, the signer of the equivalent
sentence would move both hands to the locations in space of the two referents,
and make the sign for “fall asleep” with both hands, one for each. The point,
then, is that this is a totally different semantic construction of the event. In the
spoken sentence, both referents are construed as a plural entity, which is reported
to have fallen asleep. In the signed sentence, falling asleep is attributed separately,
but at the same time, to each of the referents. The two types of language reduce
the same experience into two different types of skeletal descriptions, and these
are systemically related to the possibilities and constraints implicit in the differ-
ent modalities. The modes are different, but so are the codes.
The third problem has to do with the universalistic foundation of the entire
discussion. There is nothing wrong with the idea that the different modalities of
9780190256623-Dor.indb 159 03/06/15 6:18 PM

spoken and signed languages have deep effects on the architectural properties
of both language types, and there is nothing surprising in the observation that
there are partial similarities between them: as typological research has shown,
all languages show some similarities with some others. The problem is that ob-
servations of similarity are presented as immediate indications of universality,
whereas the attested facts of variability are never acknowledged as such; they are
only accepted to the extent that they can be universalized. This is disturbing:
less than fifty years ago, sign languages were still thought of as forms of panto-
mime. The idea that full-fledged languages may take shape in a modality other
than sound was theoretically unimaginable. Early typological reports thus in-
cluded such totally obvious absolute universals as: “All languages have vowels.”
The devastating understanding that there are actually many languages without
vowels should be allowed to raise the suspicion that languages are much more
variable than we thought they were, that there is, in other words, something
fundamentally wrong with the presupposition. There is nothing to be gained,
apart from the salvation of the presupposition itself, from the attempt to show
that sign languages do have vowels after all, the only difference being that they
are pronounced by the hands. The superimposition of the categories of spoken
languages on sign languages misses out on the very essence of sign languages. It
does not let them speak in their own terms. This understanding was forcefully
formulated by Slobin (2008):
As researchers in the language sciences, we have a choice to begin with

one or another of two basic strategies, which I refer to as “theories in
search of data” and “data in search of theories.” I maintain that the first
approach, which begins with an established theory based on spoken lan-
guages, keeps us from seeing deeper into the nature of signed languages
and ultimately into the nature of human languages in general. The second
approach—data in search of theories—is the one that is leading us to new
insights. It is an approach that can have a deep impact on linguistics—
now that we’ve passed beyond the stage of demonstrating that signed lan-
guages are “real languages” (p. 117–118).
Looked at from the social-technological viewpoint, the problem simply disap-

pears. Every language, spoken or signed, is a system with its own architectural
logic, and the logic asserts itself throughout the entire language. The unique-
ness of sign languages cannot be limited to the mode. Just like spoken languages,
sign languages are technologies for the instruction of imagination, invented and
developed by deaf communities to allow their members to communicate across
the experiential gaps between them. At the definitional-functional level, they
9780190256623-Dor.indb 160 03/06/15 6:18 PM

are exactly the same entity. At the essential level of description, they share every-
thing with spoken languages: the symbolic landscapes with triple-leveled signs,
the prescriptive protocols, the translation from the private-experiential to the
social-semantic and so on. They are socially constructed in exactly the same way,
through the ongoing process of experiential mutual-identification, and they take
the same amount of time to learn and master. Beyond that, however, everything
about them can be as different as shown by the data.
8.6 Toward a Science of Linguistics

In many sectors of the linguistic community, the enthusiasm with the univer-
salist program, back in the sixties, gradually came to be replaced by a deep sense
of disappointment. Many linguists today, making the most important empirical
discoveries about languages near and far, say that they work “without a theory.”
Chomsky presented his program as a rise up the ladder of scientific abstraction—
from “mere description” to “explanatory theory”—which means that for him,
the retreat to the empirical facts implied a step back, a step down, to a lesser
science. This could not be otherwise: a theory whose entire logic is based on the
conviction that the observable facts—variable, chaotic, complex—are nothing
but obstacles on our way to the understanding of the essence that lies hidden
behind them—universal, orderly, simple—cannot but develop this type of con-
descending attitude toward the actual description of what there is.
The universalist program, however, was never a move up the ladder of science.
In a very real sense, it was a move down, from a descriptive science (with all of
its problems) back into the realm of pre-science—where the process of investi-
gation centers around the attempt to impose the logical implications of philo-
sophical principles on observable phenomena. This can be demonstrated quite
nicely by a comparison with Ptolemy’s astronomy, probably the most ingenious
of all pre-sciences. Ptolemy attempted to capture the exact movements of the
stars and planets around the earth. The observations showed very clearly that the
planets performed from time to time what was called “retrograde movement”:
they slowed down, stopped, moved backward, and then returned to their regular
orbit. (The movement, of course, results from the fact that the earth moves too,
together with the planets, around the sun—but for Ptolemy, the earth was still at
the center of the universe, totally motionless.) The challenge that Ptolemy faced
was that of settling the apparent contradiction between the observations of ret-
rograde movement and Plato’s principled assertion that all the movements of the
heavenly bodies are always uniform and perfectly circular. Ptolemy’s solution was
an extremely sophisticated and refined version of an idea that informed much of
9780190256623-Dor.indb 161 03/06/15 6:18 PM

Greek astronomy before him: the planets are carried and moved by a complex set
of spheres rotating around the earth. Every planet has a major sphere, the defer-
ent, that rotates around the earth. The planet, however, is not directly attached
to the deferent. There are smaller spheres that are attached to the deferent, called
epicycles, and then there are more epicycles attached to those (sometimes up to
six per planet), and the planet itself is attached to the last one. The fact that there
are as many epicycles as there are, and that they all move at once, means that
the movements of the planets would seem to us to be imperfect. But in the skies,
nothing really moves back, slows down, or stops. Each and every sphere, deferent
or epicycle, is perfectly circular, and the movement of the entire system is always
totally uniform. The contradiction is resolved. The observations are explained:
they have been made to conform with the philosophical principle. Chomsky’s
program works in exactly the same way. It is, after all, inspired by the same Greek
philosopher.
For linguistics to become a real science, an explanatory science, it must aban-
don the presupposition of universality and come to terms with the fundamental
fact of diversity—diversity all the way down. There is series of challenges, meth-
odological and conceptual, involved in the re-thinking of linguistics as a science
of diversity. The most important and the most immediate one is the challenge of
reflexivity. Linguistics must learn how to look at languages in their own terms, to
avoid the universalist temptation of imposing the familiar categories of heavily
researched languages on the structures and meanings of the other languages of
the world (Levinson and Evans 2009).
Language is a complex system, determined by a multiplicity of causal fac-
tors. Some of these factors are naturally more crucial to the system as a whole.
Others are secondary, contingent. Linguistic analysis is thus bound to lead the
researcher, working on a particular language, toward a tentative distinction be-
tween core and peripheral phenomena. There is no way to avoid this distinction,
and there is no need to: it is exactly what we need in order to understand the inter-
nal order within the multiplicity of causal factors. The assumption of universal-
ity, however, imposes on the distinction between core and periphery a parameter
that is totally foreign to the essence of language: if the universal human capacity
for language is the essential causal factor determining language, every language,
then the question of core and periphery becomes an inter-linguistic question: core
and periphery in a particular language are determined by the extent to which
that language shares its different architectural properties with other languages.
Architectural properties that the language shares with the others are thought of
as core phenomena; they are determined, as in any other language, by the univer-
sal capacity. Unique architectural properties, developed by the language “on its
own,” thus require additional explanation: in the generative literature they are
9780190256623-Dor.indb 162 03/06/15 6:18 PM

usually demoted to the status of contingent, periphery phenomena, or explained

away as surface phenomena that hide a deeper universal reality.
The universalist assumption compels the linguist to look at particular lan-
guages in terms of their similarities with others. But languages cannot be un-
derstood unless they are analyzed on their own, in their own terms, one at a
time. The distinction between core and periphery cannot be inter-linguistic. It
must be language-internal. There is a certain logic to every language: an emer-
gent complexity of a very unique type, a certain implicit order in the way signs
are arranged on the symbolic landscape; a certain understanding of the essence
of convention; a different etiquette of communication; a different regime of pre-
scription; a particular level of tolerance and capacity for change; a very particular
division of labor between semantics and pragmatics; a unique pattern of social
stratification; a unique set of relationships with other languages; and, of course,
different signifiers connected to different signifieds and different experiential
clusters; different conventions of message construction; different word orders;
different morphologies, and so on. What turns out to be a core phenomenon
in one language may be peripheral in another, and totally absent from a third.
A phenomenon that appears in many languages may be a core phenomenon in
some and peripheral in others. Whether or not a certain phenomenon in a cer-
tain language also appears in other languages is perfectly immaterial to the ex-
planation of that phenomenon in that language. Languages cannot be variable
only at their peripheries. They must also be variable to their very core.
When variability comes to be equated with contingency, linguistic theory
loses its very foundation. Variability is not a contingent fact about language: it
is its very essence. Everything that is essential in every particular language (to-
gether with everything that is contingent about it) emerged and came to be es-
sential (or contingent) through a very particular social history, located in space
and time, in which particular communicative problems forced members of a par-
ticular society to find new and different ways to communicate new and differ-
ent experiences. The social histories of languages are different, and the languages
themselves must be at least as different as their histories—as different as the so-
cieties within which they emerged.
9780190256623-Dor.indb 163 03/06/15 6:18 PM

9 A c q u i s iti o n a s a C o l l e c ti v e
Enterprise
As long as language is thought of as an individual cognitive capacity,

the challenge of language acquisition—the task of deciphering such a
system of enormous complexity—falls on the narrow shoulders of the
individual child. The question is thus: how does the child manage to
meet this formidable challenge? For Chomsky, the answer lies in what
the child know a priori: the universals postulated within the genera-
tive framework are the innately given foundations of language that
the child brings to his or her encounters with the particular language
spoken in the environment. Chomsky’s rivals in this field of acquisi-
tion, working within different versions of the language development
paradigm, attempt to understand how linguistic knowledge actually
develops in the child’s mind—in most cases without the postulation
of pre-given linguistic universals (Arnon et al. 2014, Bates et al. 1995,
Berman 2004, Bloom 1970, Bruner 1983, Clark 2009, Elman et al.
1996, Slobin 1997, Tomasello 2003). The great majority of their work
concentrates on the interactions between child and caregivers, par-
ents, and others: the child learns language from them.
The theory of language as a socially constructed communication
technology paves the way toward a new hypothesis: language acqui-
sition may not be an individual project in the first place. Everything
about language suggests that it is actually a collective enterprise: chil-
dren construct and develop their language together with others—their
parents and other adults, but also, much more importantly than gen-
erally assumed, with their peers. If this is correct, the overall challenge
is spread between the participants, and the task facing each individual
child is dramatically reduced. What this implies, quite paradoxically,
is that Chomsky’s foundational suspicion was right all along: the in-
dividual child does not decipher the entire complexity of language on
the basis of external input. This is done by the group.
In this chapter, then, I would like to make the first steps toward
a collective-acquisition hypothesis. I will begin with two prelimi-
nary moves, both intended to release the individual child, each and
every member of a language-acquiring group, from the shackles of
9780190256623-Dor.indb 164 03/06/15 6:18 PM

Acquisition as a Collective Enterprise • 165
universalism. First, while most of the ongoing research on language acquisition

concentrates on the dynamics, capacities, and learning strategies that seem to
be shared by most or all children, I will follow the minority views of Bates et
al. (1988, 1995) and Levinson (2012) in suggesting that we need to acknowl-
edge the enormous variability between language-acquiring children, on all these
fronts. Children, just like the speaking adults they will become, are very differ-
ent from each other. Second, I will suggest that the attested variability does not
just reflect the variable patterns of the children’s exposure to language, but also a
level of innately given cognitive variability. The deciphering of language requires
the combination of a huge array of capacities, and these capacities are variably
spread across the population of children. No two children are alike. Why this
must be the case will be explained in the next chapter: it is a necessary result of
the evolutionary emergence of language and its speakers.
Then, partially adopting Harris’s (1998) insistence on the crucial importance
of peer-group dynamics in the process of language acquisition, I will suggest that
we may need to turn our investigative gaze from the interactions between indi-
vidual children and their parents and other adults—which are already well re-
searched and obviously very important (Arnon et al. 2014, Tomasello 2003)—to
the interactions between children and their peers—at the park, kindergarten,
and school: children bring bits and pieces of the adult language into the processes
of experiential mutual-identification they go through in their peer groups—and
thus collectively construct their own language with their friends. This means
that the differences between children may actually play a crucial constructive
role in the acquisition process: different children play different roles in the col-
lective process.
At the moment, there is only some preliminary evidence to support this idea
(Köyman et al. 2014, Ervin-Tripp 2014, Stephens and Matthews 2014) There is,
however, something else: the fact that groups of children manage to invent lan-
guage by themselves, for themselves, even in the absence of a target language in
the adult environment. This is how Nicaraguan Sign Language emerged, and a
similar process has also been observed in other sign languages. In the last section
of the chapter, I will try to show how the collective-acquisition hypothesis allows
for a new explanatory analysis of these processes.
9.1 Releasing the Individual from Universalist Assumptions

For most non-linguists, the idea that adult speakers are extremely different
from each other in their linguistic skills is self-evident. Speakers are different
in the ways they manage to express themselves, translate their experiences into
9780190256623-Dor.indb 165 03/06/15 6:18 PM

instructions for imagination, analyze the structures of complex sentences, imag-

ine on the basis of semantics, follow discourse, persuade others, explain, spot a
contradiction in an argument, tell (and understand) a joke, solve a crossword
puzzle, get away with a lie. Some individuals write novels, litigate in court, or
move crowds of people to action by their words.1 Others can read (some of) the
novels and enjoy them (identify with them), or follow the discussions in court.
Many others yet find these tasks very demanding, quite often impossible. Knowl-
edge of language is not equally spread among all its speakers. Different versions
of the technology are implemented in different ways, to different extents, in
the minds of different speakers. There is, to be sure, a certain basic level of skill
shared by most speakers, which raises a fundamental question: how did human-
ity get to this point? This question will be discussed in the next chapter.
It is fascinating, then, to see how the fact of individual variability—such a
central element in our experience as speakers—was made to disappear, as if by
magic, in the cognitive science of linguistics. A combination of three moves was
involved. The first is what Levinson (2012) calls the original sin of the cognitive
sciences: it simply states that linguistic knowledge (like all other human cogni-
tion) should be universally and equally shared—because human beings are all the
same. This is the Martian argument. There are variations between individuals
(and we, as humans, might feel that we are more interested in them), but they
are, as always, negligible details. The essence is what makes us all human, the fact
that we have a mind (a mind, in the singular) of a certain type. The second move
was the introduction of the distinction between performance and competence:
what individuals manage to do with their languages has nothing to do with what
really counts—the abstract, static knowledge of the structural foundations of
language. Many non-linguistic factors may influence the way people work with
language: great speakers and writers may be more gifted in terms of their gen-
eral intelligence (they may have quicker parsers), their aesthetic sensitivities,
their personal confidence, their understanding of the world, their imagination,
or their voice. As far as their knowledge of the foundations of language is con-
cerned, however, they are just the same as everyone else. Behind the observable
facts of variable performance lies the hidden layer of universal competence.
The third move—the most important for our present purposes—was the in-
troduction of the distinction between native language acquisition and second lan-
guage learning. There are, to be sure, major differences between the process of
language acquisition at an early age and the process of second language learning
later in life, but the distinction was introduced to say much more than just that.
It stated that language acquirers and language learners do different things: ac-
quisition is a unique process, only possible within a certain critical age frame. It
takes place naturally, and results in perfect, native command of the language. It
9780190256623-Dor.indb 166 03/06/15 6:18 PM

develops in similar ways in different children—and it reveals something deep and

essential about the human capacity for language. Learning, on the other hand, is
a shallow phenomenon. It takes place later, when acquisition is no longer possi-
ble. It is a haphazard (and often artificial) process, in which different individuals
manifest all possible levels of variation, and it never leads to native proficiency.
The essence of linguistic knowledge, the object of the cognitive science of linguis-
tics, reveals itself in the universal patterns of acquisition. Everything else is detail.
Individual variability between children, then, has been a major topic of in-
vestigation in the domain of second language learning (Dörnyei, 2005, 2006,
Ehrman et al. 2003), but has generally been ignored in the literature on native
language acquisition. The strongest dissenting voice was Elizabeth Bates’s (Bates
et al. 1988, 1995). As opposed to the great majority of studies, which were (and
are) based on fewer than 25 children, Bates et al. (1988) analyze the distribu-
tion of acquisition patterns in a sample of more than 1,800. As opposed to the
other studies, they refuse to concentrate on the mean rate of development, and
highlight the range of difference. Thus, they show that even the famous “vocab-
ulary burst” that all children supposedly go through around the age of two is
not really there. The mean for reported expressive vocabulary at this point is 312
words—but the range “goes from a low of 89 to a high of 534.” Bates and her
colleagues discuss a wide array of differences between individual children: in the
rate of vocabulary growth, the combination of words and the emergence of sen-
tence complexity, the understanding of speech-acts, the gaps between what the
children can comprehend and what they can produce, and the developmental re-
lations they exhibit between the growth of their lexicons and that of their gram-
mars. Moreover, they show that children have different acquisition styles and
strategies, that manifest themselves as early as the stage of babbling—with some
infants showing more interest in words and others in intonation patterns—and
continue to diverge with time, in patterns of under-generalization and over-
generalization, in the acquisition of closed-class words, the usage of nouns versus
pronouns, the consistency of rule application, and so on and so forth. Bates and
her colleagues conclude:
The Average Child is a fiction, a descriptive convenience like the Average

Man or the Average Woman. Theories of language development can no
longer rely on this mythical being. Any theory worth the name will have
to account for the variations that are reliably observed in early language
learning. (p. 26)
This change of perspective carries extremely important implications for our view
of normal child development, implications that I will not pursue here: as long
9780190256623-Dor.indb 167 03/06/15 6:18 PM

as the average is thought of as the normal, every deviation is a potential indica-

tion of a problem in need of special treatment. As far as my theory of language is
concerned, the facts of deep variability are exactly what we should expect to find:
like all other technologies, language exposes every level of difference between its
users that carries implications for its successful acquisition and usage. The range
of differences develops and stabilizes as children go through their variable expe-
riential histories of engagement with the technology.
Let us, then, consider another human skill, like chess playing, that has
never been claimed to be universal: how would we conceptualize the rela-
tionship between the learning of the game from the environment and the set
of innate capacities that the chess-learning child brings to the game? Quite
obviously, we would not be tempted to postulate innate knowledge of chess.
We would probably formulate the question—and then the answers—in terms
of the dynamic interaction between two complex levels of variability. On the
one hand, we would need detailed descriptions of the experiential histories
of different players—from those who have only heard about the game all the
way to the great masters: the age in which they began to learn the game; their
teachers, their teachers’ knowledge of chess, their teachers teaching capacities;
the amount of time they dedicated to the game every day; the skills of their
playmates; and so on and so forth. On the other hand, we would have to rec-
ognize something that, in the case of chess, we would not hesitate to simply
call talent—an enormous set of variable innate capacities, all of which would
be unevenly distributed among the different players: everything involved in
learning, remembering, understanding, analyzing, planning, and inventing
chess moves. The game of chess, as the task out there, requires a certain com-
plex set of skills, which means that it exposes a corresponding set of differences
in the innately given capacities that different individuals bring with them to
the game. We would thus find that (a) different players, who have undergone
relatively similar experiencing processes, have nevertheless developed differ-
ent levels (and types) of skill; and (b) different players, with relatively similar
talents for chess, have nevertheless developed into different players because of
differences in their exposure to the game. We do have innate dispositions for
the game of chess, but they participate in the story as emergent variables, not
as foundational constants.
This, I would like to suggest, is exactly how we should conceptualize the
acquisition of language: different children come to the world with variable
innate dispositions toward language—just as they do toward anything else.
They encounter language in different ways too, and what they end up with
is a complex cumulative function of both talent and exposure. In the next
chapter, I will get back to the innate side of things, define it more explicitly,
9780190256623-Dor.indb 168 03/06/15 6:18 PM

and show how it must have emerged. At the moment, let us turn our atten-
tion back to the process of exposure to language. Here, the theory of lan-
guage as a communication technology suggests that the process may actually
be collective.
9.2 Language Acquisition as a Collective Effort

There is a certain fundamental fact about language acquisition, which is as im-
portant as it is ignored in the literature. In his preface to Harris (1998), Steven
Pinker calls it “a strange factoid” in the “True-but-Inconvenient file” of language
development: “children always end up with the language and accent of their
peers, not of their parents . . . no one in psycholinguistics had ever called atten-
tion to this fact, let alone explained it, but here was a theory that did” (Harris,
p. xxiii). Together with other observations, the fact that children zoom in on the
language variety of their peers is incorporated by Harris into a strong argument
against the nurture assumption: the idea that children’s personalities are deter-
mined by their parents. Children, argues Harris, develop into members of their
social communities mainly through interactions with their peers. This general
argument has attracted much attention, positive and negative, and it obviously
raises serious questions having to do with parental responsibility—questions I
will not dwell on here. As far as language is concerned, Harris highlights two
types of issues. First, whenever children acquire one language at home and an-
other in kindergarten or at school, they eventually reach native command of the
second, not the first. This is most evident in immigrant families. Second, in most
cases, the very idea that parents are supposed to teach language to their children
seems to be is a relatively new, Western one:2
The language lessons we give our infants and toddlers are a peculiarity
of our culture. In parts of the world where people still live in traditional
ways, no lessons are given and parents generally do very little conversing
with their babies and toddlers—they consider learning the language the
child’s job, not the parents’. According to psycholinguist Steven Pinker,
mothers in many societies “do not speak to their prelinguistic children at
all, except for occasional demands and rebukes. This is not unreasonable.
After all, young children plainly can’t understand a word you say. So why
waste your breath in soliloquies?” Compared to American toddlers, the
two-year-olds in these societies appear retarded in their language devel-
opment, but the end result is the same: all the children eventually become
competent speakers of their language. (p. 65)
9780190256623-Dor.indb 169 03/06/15 6:18 PM

Children in traditional societies, then, acquire their language “in the play group”:
Their conversational partners are other children. Older children simplify

their speech a bit when they’re talking to younger ones, but they do not
provide the kind of language instruction that parents give their toddlers
in our society — the question-asking, the patient rephrasing of the learn-
er’s poorly phrased statement, the smile or pat when something is said
exceptionally well. So the children in traditional societies learn language
at a slower rate. But learn it they do. They all become competent users of
the language that is spoken in their community. And they all become
socialized. (p. 150)
How, then, is the language spoken at home, and the entire culture it carries with
it, involved in the process? Harris suggests that children “bring that culture with
them to the peer group, but they do it carefully, tentatively. They are alert to
signs that there might be something wrong with it—that it might not be the
culture of the Out-Theres” (p. 197). For Harris, then, the main issue is one of
identity: the child always chooses to belong in the peer group. And how does
all this reflect on the classical question of acquisition? If the challenge is that of
learning from peers, how does the child manage to meet it? Here, Harris resorts
to the type of generativist innateness postulated by Pinker: language “is one
of the things that we inherit from our ancestors but that does not vary among
normal members of our species, like lungs and eyes and the ability to walk erect.
Every human baby born with a normal brain is equipped with the ability and
desire to learn a language. The environment merely determines which language
will be learned” (p. 28).
This marriage between the peer-group hypothesis and universal innateness
(from the author who rightly insists in Harris (2006) that all children are dif-
ferent from each other), together with the emphasis on immigrant children,
weakens Harris’s linguistic argument to a considerable extent. The environment
does not merely determine which language will be learned, and the issue is not
just of one of identity. The point is much more fundamental: it stems from the
functional essence of language as a communication technology, and from the
essentially variable positions that we, as speakers and acquirers, occupy vis-à-vis
the technology.
The first point I would like to stress is this: there seems to be something
wrong in the very idea that children are engaged in an attempt to acquire the
language of the adults around them (the target language). The adult language as
such, the entire system, is not their object. It is not just that the language is too
complicated; the point is that it simply runs way over their heads. The adults
9780190256623-Dor.indb 170 03/06/15 6:18 PM

Acquisition as a Collective Enterprise • 17 1
use the technology, in complicated ways, to instruct each other’s imaginations

about experiences the children do not recognize. What children are trying to
do, at every point in the course of their linguistic development, is improve their
ability to communicate their own experiences—their childish experiences. Moti-
vated by their own communicative, epistemic, and social needs, they only engage
in the mutual-identification of experiences that are relevant to them. This they
do with whoever is willing to participate, but the patterns of acquisition in the
non-Western world strongly indicate that the most natural partners for such
mutual-identification are other children. This is not just a matter of language.
Experiential mutual-identification may be the most important thing children do
with their peers. They are obsessed with it: they become friends whenever they
discover they experience something in the same way; they can become enemies
in a second when they realize they see something differently. This, among other
things, is why childhood may be the most blessed period for some, and a never-
ending nightmare for others.
Taking all this as our starting point, we may conceptualize the entire process
as a spiraling, collective dynamic. Children constantly search the adult language
for conventional components that can be used for the communication of experi-
ences that are relevant to them. Rather than trying to figure out the language
as such, they are merely trying to pluck materials that they can use. Crucially,
different children—with their different intents, capacities, learning strategies,
and types of exposure at home—pluck different components of the adult lan-
guage. Different children are attracted to different words; some are more fas-
cinated by the words’ sounds, others by their meanings; some are attracted to
logic; others relish metaphors; and so on and so forth. Then, children introduce
their findings into their discourse with their peers—not just in conversation
but also, maybe more importantly, in pretend-games and other imagination-
based collective activities (Wyman 2014, Whitehead 2014). Every individ-
ual child is thus presented by his or her friends with new components of the
language—already discovered by someone else, already isolated from the en-
tirety of the adult language. Processes of mutual-identification are launched, in
which the children take another step in the identification of the similarities and
differences between their experiences. What is then accepted into the children’s
language, and what is rejected, is determined by the social, communicative, and
epistemic needs of the children, needs that emerge from their experiences, their
social relationships, the power dynamics they are involved in, and the general
level of capacity and interest manifested by the majority of children. Sugges-
tions by precocious children, for example, are very often rejected by their peers,
and precocious children do indeed sometimes find it easier to talk to adults. As
the children grow and develop, additional experiences enter their worlds, new
9780190256623-Dor.indb 171 03/06/15 6:18 PM

capacities emerge that allow for more plucking, and then more sophisticated
levels of mutual-identification are achieved, and so on and so forth. Slowly but
surely, the bits and pieces constantly brought from the outside accumulate to
construct what will eventually become an adult-like language.
What this collective model implies, then, is a dramatic reduction in the mag-
nitude of the challenge every individual child has to face. Whenever they bring
an isolated component of language into the group, children already do most of
the work for the others. The component now appears within the simple and rel-
evant discourse of the group, surrounded only by components that have already
been mutually identified. It is much easier to learn. Different children, more-
over, do what comes naturally to them: where they encounter problems they are
implicitly helped by their friends. Some children play more significant roles in
the construction of the children’s language, others stay more passive and adopt
whatever has been constructed by the others. The entire set of innately given
capacities that a child must recruit should be sufficient not for the acquisition of
adult language—but for the participation, in some role or other, in this collective
project.
In most non-Western cultures, the involvement of adults in this process
seems to be of three types. First, and most important, is the fact that the adults
communicate with their children, mainly on practical issues, and thus engage
in mutual-identification processes with them regarding these issues (Brown
2014). Second, the adults also speak with other adults, around the children, and
thus allow the children to engage in emulation learning (Boesch and Tomasello
1998): watch, listen, and try to pluck. Third, adults are engaged in emotional and
practical relationships with their children. What the adults do not seem to do,
in most cases, is attempt to enter the children’s own experiential worlds, identify
with the ways they experience, participate in their pretend-games, communicate
with them at their level—and teach them language. All this is exactly what par-
ents do in the West, and this is why, in their child-directed speech, they try to
sound like children. Motherese is the imitation of child speech, and what it indi-
cates is a deep shift in the entire cultural conception of childhood, parenthood,
and child-parent interaction.
As the literature on the history of childhood shows very clearly (Ariés 1962,
Pollock 1983, Clarke 2004), the shift began to appear in Europe of the seven-
teenth century—with the rise of the new ideology of the middle-class family
as a nuclear unit, focused on the upbringing and proper education of its chil-
dren. Whether this means that in medieval Europe the idea of childhood “did
not exist,” as Ariés famously claimed, is probably still an open question: Pollock
(1983) presents the best case against it. What seems to be established beyond
doubt is that the shift was characterized by a growing fascination with the
9780190256623-Dor.indb 172 03/06/15 6:18 PM

experiential lives of children—as naïve creatures of nature on the one hand, as

sinful creatures in need of discipline on the other—and in both cases what the
fascination implied was that children required education. Infants began to spend
more of their time in the home, away from other infants, and children began
to spend more of their time in educational institutions, thus spending less time
communicating with their friends. Within this context, the growing involve-
ment of parents in the acquisition process, together with the emergence of child-
directed speech, seems to be a natural and highly required development. It also
became one of the most crucial class markers in the emergent capitalist society
of Europe: working-class children and their parents did not participate. I am
quite convinced there is much to be learned, even today, from a systematic class-
sensitive investigation of language acquisition processes in Western societies.
At the moment, as I already indicated, there is only preliminary evidence
for the collective-acquisition hypothesis. What we need is not hard to imagine,
but probably much harder to implement than parent-oriented studies: long-term
studies of the process of language construction in groups of children—in differ-
ent societies—that would be able to follow the children both at home and with
their friends, and track down the micro-dynamics in which language compo-
nents are plucked by individual children, brought into the group, accepted, imi-
tated and learned by the others, and incorporated into the children’s language.
What we do have, however, is the undeniable fact that children are capable of
doing this even when there is no target language around them.
9.3 The Message from Nicaragua and Al-Sayyid

The basic facts of the Nicaragua story are by now widely known: in 1977, a first
group of about twenty-five deaf children was brought to Managua to attend a
special educational program for the deaf, the first of its kind in Nicaragua. After
the Sandinista Revolution in 1979, the number of children grew to about a hun-
dred. In 1980, a vocational school for deaf adolescents was opened, and in 1983,
there were already over four hundred students in both programs. The instructors
in the programs attempted to teach the children to lip-read and fingerspell, but
the attempt failed. Then, the teachers realized that the children were communi-
cating with each other using signs that the teachers could not follow. Linguist
Judy Kegl was invited to try and figure out what the children were doing. What
she discovered quickly captured news headlines, way beyond the circles of lin-
guistics. The New York Times called it a linguistic big bang: the children invented
a language by themselves—and the language they invented clearly manifested a
level of lexical, morphological, and syntactic complexity they could not possibly
9780190256623-Dor.indb 173 03/06/15 6:18 PM

have picked up from their environment. Judy Kegl, Ann Senghas, and other re-
searchers showed that Nicaraguan Sign Language (NSL) distinguishes between
different parts of speech; that its verbs may be marked for agreement with their
objects, and inflected for person, location, number, and manner; that the lan-
guage allows for multiple-verb constructions (known as “verb sandwiches”); and
so on and so forth (Kegl 1994, Kegl et al. 1999, Senghas and Coppola 2001,
Senghas et al. 2004, 2005). In one of the clearest demonstrations, Senghas et al.
(2004) concentrated on the expression of complex motion events such as “rolling
down a hill” or “climbing up a wall.” They compared the way signers described
these events with the gestures produced by hearing Spanish speakers while de-
scribing the same events. When asked, for example, to describe an event (from an
animated cartoon) in which a cat rolls rapidly down a steep street in a wobbling
manner, the Spanish speakers accompanied their speech with holistic gestures
that resembled the actual trajectory of the cat’s movement. The signers, on the
other hand, produced two separate signs, one for the manner and one for the tra-
jectory, and then assembled them into a sequence. They thus described the event
in the discrete and combinatorial fashion that distinguishes genuine language
from mimics and gestures.
In the case of Al-Sayyid Bedouin Sign Language, the circumstances of emer-
gence were different: the language seems to have been created over the last sev-
enty years within an already stable community. The fundamental linguistic fact,
however, is the same: Al-Sayyid Bedouin Sign Language includes structural com-
ponents that could not have been learned or adapted from external input. The
language manifests (among other patterns) regularities of word order: a robust
subject-object-verb order in sentences and a head-before-modifier order within
phrases. Such patterns are nowhere to be found in the spoken languages of the
area (Sandler et al. 2005, Aronoff et al. 2008, Meir et al. 2010).
How is all this to be explained? For Kegl et al. (1999), the emergence of lan-
guage with no input was the ultimate proof of Chomsky’s linguistic bioprogram:
“the language-emergence process in Nicaragua constitutes one of the few cases
in which the human bioprogram for language or innate human-language capaci-
ties, by virtue of no coexisting language input, have been forced to take a singu-
lar role in shaping the emergent language.” The authors explain the processes in
three parts: first, deaf children already used small sets of home signs for commu-
nication with their family members, before they came to Managua—when they
were still isolated from each other. When the first cohort of children came to-
gether and the children began to interact, a pidgin language arose on the basis of
the home signs they brought with them—a rudimentary communication system
with no grammatical complexity. Then, when the second cohort of younger chil-
dren arrived, the pidgin that they found was enough to trigger their innately
9780190256623-Dor.indb 174 03/06/15 6:18 PM

given knowledge of language: they restructured and regularized the impover-

ished pidgin and created the full-fledged, grammatical language that they now
speak. The fact that the older children of the first cohort kept on using their
pidgin, and never learned the grammatical structures created by the younger
children, was proof that they were already beyond the critical age for language
acquisition: the innate knowledge could no longer be triggered.
In later work, Senghas et al. (2005) developed a much more sophisticated ac-
count of the process, highlighting the sociocultural history of the creation of the
language (including the role played by inter-generational relationships within the
emerging community), and showing that the language did not emerge abruptly,
in one cohort of children, but gradually increased in complexity from one cohort
to the next. The bioprogram, however, remained a major explanatory component
in the account:
NSL could emerge only when a cohort of adolescents and adults provided
the social and linguistic environment from which it grew, and ensured
the perpetuation of its signs and conventions. The grammatical elements
to be perpetuated, however, depended on a complementary role that only
children are equipped to play. Their capacity to acquire grammatical sys-
tematicity (even where it is absent in the environment) is essential for the
initial appearance of linguistic structure. (p. 303)
Sandler and her colleagues are very careful to keep their work on Al-Sayyid at
the descriptive level. In Sandler et al. (2005), however, they make the following
statement: “The appearance of this conventionalization at such an early stage in
the emergence of a language is rare empirical verification of the unique proclivity
of the human mind for structuring a communication system along grammat-
ical lines.” (Sandler et al., p. 2665) These formulations are remarkable for two
reasons. First, they are totally uninformative. They teach us nothing new about
the languages and their emergence. The descriptions of the two languages are
extremely informative, and so are the descriptions of their social backgrounds,
but the fact of emergence itself, the single fact that makes these languages so cru-
cially important, is eventually only explained away: what seems at the empirical
level to be new was actually there all along, just waiting to be triggered. Second,
the authors systematically refrain from using the word invention in their analy-
ses. They write about the human capacity “to acquire grammatical systematicity
(even where it is absent in the environment).” The question of invention is trans-
lated into a question about acquisition—and is immediately lost in translation.
All this could not be otherwise: the emergence of these languages lies beyond
the explanatory power of linguistics as a cognitive science. It is exactly where
9780190256623-Dor.indb 175 03/06/15 6:18 PM

the cognitive program collapses. The languages did not emerge from within the
minds of their users, children or adult. They emerged between them. They are
both collective projects. Let us, then, get back to the beginning, and see how the
pieces of evidence (including those left in the background by researchers) may be
re-interpreted to produce another explanation of the process.
We know much more about the beginning of the NSL story than we know
about Al-Sayyid, where the language probably originated more than seventy
years ago. Two major observations concerning the pre-history of NSL require
our immediate attention. First, when the children began to use their language
to tell their own story to outsiders (hearing people, researchers, and journalists),
they testified that before they came to the school, when they were still at home,
they understood much more of the world around them than they could com-
municate. A twenty-year-old signer called Aleman told New York Times corre-
spondent Lawrence Osborne: “I can remember my childhood, but I can also
remember not having any way to communicate.” Aleman told Osborne about
an incident in which he was hit by a fire truck when he was little, “sprinkling his
account with small scenes and characters.” Explaining why he only began to sign
at the age of fifteen, he said: “I couldn’t learn the language earlier because I grew
up in the forest. It was during the war, too, and since my father was a contra, we
were always hiding, being hunted down by Sandinistas. So I remember guns,
fear, hiding. When I came to Bluefields I was amazed. I was like . . . (pause) . . .
Babar in the big city going in the elevator for the first time.” Aleman did not
know the Babar story before he came to school, and he may have only learned the
names of the rivaling forces after he began to sign, but the rest should be enough
to demonstrate that he had spent his childhood experiencing his world, work-
ing hard, like any other child, to put some order and meaning into the events
around him.
It should also be enough to demonstrate that he had to do all this on his
own, and that he was longing for communication. Not just for communication
about the immediate experiences of the here-and-now. He wanted to commu-
nicate about the world. There were so many things he wanted to say and ask.
This longing is the deepest demonstration of the fact that all of us, modern
humans, are born with crave for linguistic communication. Native speakers of
regular languages cannot remember life before language, but Aleman does: it
was a nightmare.
The second observation concerning the pre-history of the language has to do
with the small sets of home-signs they used for communication with their par-
ents. It is crucial to see that these sets are languages, not just “rudiments of lan-
guage,” as Senghas puts it. The children and their parents began to invent their
own languages: they mutually identified certain elements of their experiences,
9780190256623-Dor.indb 176 03/06/15 6:18 PM

and marked them with mutually identified signs that they used for instruc-
tive communication—every family in its own particular way. As Senghas et al.
(2004) reports, “the home sign systems developed by Nicaraguans appear to
have varied widely from one deaf person to another in form and complexity”
(p. 1780). It is true that the home-signs had much more general meanings than
the signs of full-fledged languages, but this could not be otherwise: when the vast
world of private experience is only criss-crossed by a handful of linguistic catego-
rizations, every category can be made to cover a lot of ground. Semantic specific-
ity gradually emerges with the number of signs. It is also true that the home-signs
were much more iconic than the signs of full-fledged spoken languages, but this
is also hardly surprising. Sign languages are generally more iconic than spoken
languages: they take place in the visual domain. The question, then, is this: why
did the children and their parents stop developing their languages?
I see no reason to assume that they could not in principle continue working
together on their language projects. It is easy to imagine a couple of parents to a
deaf child, who live in isolation and refuse to send the child away to a boarding-
school, who have already heard about sign languages and decide to dedicate their
time and creative energy to the project of the invention of a full-fledged language
with (and for) their child. The parents would have to spend entire days with the
child, not just see him or her after work. They would have to be extremely sen-
sitive to the experiential view of the child who lives without sound, and totally
insensitive to the fact that the language would not allow the child to commu-
nicate with anyone but them. All this, quite fortunately, does not make a very
reasonable scenario, but it is definitely not impossible. The home-sign projects
were probably stopped because the conditions of this unreasonable scenario were
never met. The parents did not know about the possibility of sign language, they
did not have the time or the patience, they found it more and more difficult to
understand their children—on the other side of the experiential gap—without
the help of spoken language, and most importantly, they probably gave up hope
of ever communicating with the children. They saw them as inherently limited,
as “eternal children,” incapable of learning. They consequently established the
use of a small number of signs to allow for co-existence around the house, and
settled for that. The process of mutual-identification requires full participation
of all sides. A single isolated child, hearing or deaf, cannot do it alone.
All this changed dramatically when the first children arrived at school. Sud-
denly, some of the people on the other side of the experiential gap were just like
them: deaf children, hungry for friendship and communication. The children
began to experience together (in the beginning, only for a few hours a day). Using
the home-signs they had brought with them, they tried to reach at least the level
of mutual-identification already established with their parents. They probably
9780190256623-Dor.indb 177 03/06/15 6:18 PM

discovered quite quickly that their home-signs were different: signs that seemed
similar turned out to refer to different experiences; signs that looked different
turned out to refer to the same thing. Some signs were more difficult to produce
than others, or more difficult to understand. They began to mutually identify.
Senghas reports that “signers occasionally disagreed over the appropriate use and
meanings of particular signs”:
Interviews with older deaf people today suggest that initially there was
often confusion about what signs referred to which referent. As in other
such cases, as the lexicon became more and more conventionalized with
this emerging linguistic community, certain linguistic forms were left
behind, for reasons varying from efficiency and ease of production to the
charismatic nature of a particular signer. (p. 291)
The children were not just negotiating their signs in the technical sense. They
were negotiating their experiential worlds, raising first bridges across the gaps be-
tween them. They were learning about the world together, and they were learn-
ing to identify (the same things) with each other. They probably found it easier
to identify with some of their peers than others: not everybody became friends
with everybody else. They began to develop that particular sense of belonging
that is based on mutual-identification. This was the real miracle: the children
were rescued from their epistemic solitude. For the first time, they could begin to
socialize their private experiences.
Thus, the children gradually mutually identified more and more components
of their shared experiences, memories, and hopes. Some of them, for example,
began to mutually identify manners and trajectories of motion and mark them
with iconic signs. They could thus already describe the cat’s motion in the car-
toon, not just follow it with a gesture. Some of their friends, however, still ges-
tured. Different variations of the collective worldview began to emerge, and with
them, endless conversations, arguments, and power struggles. New problems of
miscommunication began to appear, and the children made their first explora-
tory attempts to resolve them. As Senghas et al. (2005) report, “signed conver-
sations from this period were characterized by frequent redundant phrases for
clarification of reference.” The children also began to experiment with spatial
modulations and sometimes produced a sign together with a movement toward
or away from a particular location, to mark co-reference. A few years later, the
children became adolescents, and then, young adults. They grew up together.
They made conscious decisions “to form enduring and informal relationships,
including the establishment and control of the Managua-based Deaf Associa-
tion. These enduring relationships include friendships, participation in the Deaf
9780190256623-Dor.indb 178 03/06/15 6:18 PM

association, marriage and domestic partnerships, sometimes despite opposing

pressure from individuals, families or institutions to do otherwise” (p. 294).
When the children of the second cohort arrived in school, there was already a
genuine linguistic community waiting for them. As far as I can see, the children of
the second cohort did three things with the language they found. To begin with,
they acquired what was already there, as part of their linguistic environment, and
spread the inventions of the first cohort among them. Most of them used the signs
for manner and direction of motion in their descriptions of the cartoon. Second,
they continued developing the symbolic landscape of the language, inventing more
and more signs—some of which were dedicated to more sophisticated meanings.
Having received (and further developed) the tools for the mutual-identification of
the major entities and eventualities around them, they could gradually turn to the
exploration of more fine-grained aspects of their experiences. The children devel-
oped a classifier system, with signs designating different forms and shapes of phys-
ical entities; they invented signs to mark person, number, and location, and so on.
As the complexity of the symbolic landscape gradually increased, a web of seman-
tic relations between the signs began to emerge, and these came to be signified too.
Third, and most importantly, the children of the second cohort began to develop
a communication protocol, with mutually identified, prescriptive conventions
for language usage. They restricted the usage of spatial modulations to one side
of the body, thus disambiguating the signification of co-reference; they began to
mark the signs designating the different participants in the events they described
in order to indicate more clearly who did what to whom; they collectively decided
that some of the fine-grained meanings mentioned above were minimal specificity
requirements, and so on. There was a functional reason for all this: the growing
complexity of the language, the growing speed of production, and the growing
number of users—all these must have gradually begun to reduce the general levels
of communicative success. Each of the prescriptions emerged, in the form of explor-
atory behavior (very often a gesture), as an attempt to resolve a specific problem of
understanding—and was then conventionalized and stabilized through a long
process of negotiation between speakers and interlocutors. Some of the conven-
tions resembled those manifested by other languages around the world; some were
different and unique. With the new conventions came new problems, and then
more conventions. The end result of the process was what we now know as NSL.
The history of the Al-Sayyid Bedouin Sign Language is different in many
ways: it has developed within an already established community, and many of the
hearing individuals around the deaf have a good command of the language. As
Kisch (2008) shows in her sociological analysis, the emergence of the language
was entangled—probably from the very start—with social issues of kinship,
gender, and identity. The growing body of analyses of the language, however,
9780190256623-Dor.indb 179 03/06/15 6:18 PM

indicates that the very same processes were involved in its emergence. Kisch
demonstrates, for example, how some of the signs used today still carry on their
sleeve much of their experiential history: the sign for adult woman or mother is
based on the iconic representation of a burqa (a traditional form of veiling) no
longer practiced, and the sign for butter derives from the iconic representation of
the churning of milk, by now rarely seen. Sandler et al. (2005) show an interest-
ing correlation between the emergence of morphological and phonological com-
plexities in the language and the process of lexical conventionalization—exactly
as predicted by my theory. And again: some of the properties of the language are
well known from other languages, and some are unique. Based on the analysis
of other sign languages, Sandler and her colleagues expected to find verb agree-
ment and classifier systems in the language, but instead found a very strict world
order—not a “typical” sign language property.
Back to Nicaragua: Both Kegl and Senghas work with the conviction that
the essence of language lies in its grammatical complexity. Because of that, they
describe a process in which a general, gesture-like system of signs was devel-
oped (abruptly or gradually) into a bona fide language. The children of the first
cohort, who negotiated their home signs in the beginning, and the children of
the second cohort, who developed the grammatical devices of the language later
on, are conceptualized as actors of different types: the older children created
the community; the younger children created the language. The achievement
of the first cohort was social and communicative. It could be done by adults
too. The achievement of the second cohort was linguistic. It could only be done
by very young children, still in their innately given critical period for language
acquisition. The children of the first cohort are thus often described in almost
tragic terms, much like the People of Israel who wandered for forty years in
the desert and never made it to the Promised Land. By the time the language
emerged, they were already too old. Until today, all they have is their impover-
ished sign-system.
There is actually nothing wrong in this line of reasoning, apart from the fact
that it is based on a false conviction. The essence of language does not lie in its
grammatical complexity. Such complexity, as we have already seen, is not even a
universal fact about languages. Why, then, didn’t the children of the first cohort
adopt the grammatical inventions of the younger signers? Were they really too
old—individually, biologically—for the task? I do not think so. There is no doubt
that something of the ease of language acquisition fades with age: throughout
the evolution of language, individuals were also selected for their ability to join
the activity of language as early as possible. But the reduction in the ease of
acquisition is not abrupt. Acquisition does not all of a sudden become impos-
sible. It gradually becomes more difficult. Different children, moreover, bring
9780190256623-Dor.indb 180 03/06/15 6:18 PM

variable capacities into the collective process—and for some of them, the criti-
cal age stretches for a much longer time than for some others. Hence, if biology
were really the issue, we should have expected some of the members of the first
cohort to adopt some of the inventions of the second. Spatial modulation, for
example, has already been partially used in the first cohort, and the younger chil-
dren restricted it to one side of the body. I see no reason to assume that adopt-
ing this convention was beyond the capacity of all the older children, even as
adolescents—provided that they were motivated to do it. Even in second lan-
guage learning, the more gifted and motivated adolescents and adults manage
to achieve a very good command of the conventions of the language and their
usage. They never reach native levels of proficiency, but some of them actually
come remarkably close. The fact that no one in the first cohort adopted any of
the new conventions suggests that the older children ignored the innovations of
the younger ones—not individually, but as a group.
The members of the first cohort spent the first crucial years in school as an
extremely isolated community. They could only communicate within them-
selves, and they only mutually identified their experiential worlds with each
other. They came to know each other very intimately, and they must have devel-
oped an exceptionally strong sense of group identity. There is very good reason
to believe that the language they developed served them quite well. They did
not need the new conventions: they understood each other without them. As
we all know, intimately close interlocutors often use radically impoverished
language: there is so much that has already been experienced together, so much
that has already been mutually identified. Intimate interlocutors may use a
word or two to convey something to each other that would require very hard
work, and very complex sentences, if any of them attempted to communicate
the same thing to strangers. As more and more children arrived at the school,
the members of the first cohort probably found it significantly more difficult
to communicate with them. In a very real sense, for them, this was no longer
a problem of communication within the group. It was a problem of commu-
nication with the members of another group—a problem of communication
between a very tightly knit group of adolescents, with their unique experiences
and concerns, and a group of much younger children, just beginning to con-
struct their own collective understandings: there were many more of them,
much wider experiential gaps that had to be bridged, many more problems of
miscommunication that had to be resolved. The members of the first cohort,
however, probably never participated in this process, at least not systematically
(even if they encouraged the children to innovate, and even if they met them
on a regular basis in public events and celebrations). In this sense, despite the
special circumstances, they were not very different from groups of adolescents
9780190256623-Dor.indb 181 03/06/15 6:18 PM

all over the world. Children very often adopt pieces of language from adoles-
cents, but such adoption in the opposite direction is very rare.
From a certain point, then, the dynamics of mutual-identification, negoti-
ation, communication, and identity-construction involved in the development
of the language no longer included everybody all the time. The community had
reached a certain level of social complexity, and the dynamics of language devel-
opment began to be constrained by the emerging patterns of social stratification.
Different dynamics probably began to take place in different corners of the com-
munity. As we know it today, the NSL has (at least) two age-related dialects, not
because some of the children were too old to keep up with the pace of innovation,
but because they retired—as a group—from the social activity of innovation.
It is important to see, however, that throughout the entire process, all the
Nicaraguan children were busy doing the same thing: motivated by neces-
sity, and to the best of their ability, they participated in the collective activity
of mutual-identification as it was taking place at the moment. Their language-
craving cognition did not just come into the story when the children of the
second cohort began to develop the protocol. It was there from the very be-
ginning, expressed in behavior to the extent that there was collective activity to
participate in. Throughout the process, the collective effort was centered on the
attempt to improve the technology—again, as it was at that moment. As the
language developed, the challenges kept changing. The linguistic challenges of
the first cohort were no less daunting than those of the second—rather more so,
actually. The children of the second cohort joined a linguistic community, and
further developed the language they received. The children of the first cohort re-
leased themselves from the solitude of their early years by their own hands (quite
literally), and invented a language almost from scratch. The revolution was def-
initely theirs.
9.4 Acquisition, Invention, Evolution

Where does all this leave us? On the one hand, what children are capable of
doing by themselves is much more than just acquiring an existing language: they
can construct it for themselves. On the other hand, they can only do it as a group,
not as individuals. They are equipped with variable innate capacities for the par-
ticipation in the collective process of language construction, not for the individ-
ual task of deciphering the entirety of language. There are three scenarios, then:
in the Western scenario, children are capable of participating in the collective
process; they are surrounded by adult language; and the adults around them ac-
tively participate in the endeavor. Children construct their language with their
9780190256623-Dor.indb 182 03/06/15 6:18 PM

peers and with their parents, and when they have to choose, they go with their
peers. In the prototypical non-Western scenario, the variable innate capacities
and the adult language are there, but adults do not actively teach language to
them: the children pluck materials from the adult language and construct their
own language together. In the third scenario, the variable innate capacities are
there, but there is no adult language around. Here, children go through the very
same process without, or with very little, external input. The collective process
is made easier by the existence of adult language, but it is not dependent on it.
In a very deep sense, then, what children do all over the world, in these dif-
ferent scenarios, is actually very similar to what our ancestors did when they first
invented the new technology—with one crucial difference. As I will try to show
next, our ancestors managed to invent and construct language not just without
parental participation, not just without external input—but also without the
specific innate capacities required by language the way we know it today.
9780190256623-Dor.indb 183 03/06/15 6:18 PM

10 T h e E v o l u ti o n o f L a n g u a g e
and Its Speakers
The question of the emergence and evolution of language is recog-

nized today as one of the central questions in the whole of science—
certainly the most crucial issue we have to contend with if we wish to
understand how the human species came to be what it is. In the last
three decades, the question has inspired an unprecedented wave of re-
search, in which scholars from a wide array of disciplines—linguistics
and philosophy, the different branches of psychology, anthropology
and sociology, paleontology and archeology, evolutionary biology
and genetics, primatology and ethology, neuroscience and computer
science—have been collaborating to find and interpret clues to
what actually happened. The fact that we have no direct evidence—
language does not leave material traces behind it—forces us to adopt
a detective’s mindset, searching for pieces of circumstantial evidence
that we then try to piece together into theoretically plausible hypoth-
eses. Disciplinary boundaries lose their significance: every piece of
evidence counts.
The question of the evolution of language, however, is not just im-
portant for its own sake. It should also be properly understood as the
most crucial bottleneck that any theory of language should be able to
squeeze through. With all the advances in the linguistic sciences, evo-
lutionary biology is still light years ahead of us in maturity, sophistica-
tion, insight, and methodology: we have a much better understanding
of the nature of evolution than we have of the nature of language. For
every theoretical model of language we should thus ask: how is this
evolvable?
In the final account, then, the deepest paradox of Chomsky’s pro-
gram is the fact that it does not squeeze through the bottleneck: if
language is genetically given to as a universal cognitive capacity, we
should have somehow evolved to get there. But if language as an
innate cognitive capacity is universal and non-functional, infinitely
generative and static, its emergence in the life of the human species
cannot be explained in evolutionary terms. For this, of course, only
evolutionary theory should be blamed. A language of this type would
9780190256623-Dor.indb 184 03/06/15 6:18 PM

The Evolution of Language and Its Speakers • 185
make sense in other worldviews—in the one, for example, that has the gift of
language bestowed upon all human souls by superior power. But the replace-
ment of creationism with the evolutionary perspective carries certain implica-
tions that cannot be ignored: things in this world arrange themselves in complex
patterns of variability; they develop and evolve because they are functional; they
are always finite; and they are always dynamic. This is what evolution is all about.
Chomksy himself, in the first four decades after Syntactic Structures, consis-
tently refused to discuss the question of evolution. It is a mystery, he used to say,
not a scientific problem: trying to deal with it would be “as absurd as it would
be to speculate about the ‘evolution’ of atoms from clouds of elementary parti-
cles” (Chomsky 1968, p. 61). Some of his colleagues—Pinker and Bloom (1990),
Pinker (1994), Jackendoff and Pinker (2005), Jackendoff (1999), piattelli-
Palmarini (1989), and others—have claimed over the years that there is common
ground to be found between the generative program and evolutionary theory,
but Chomsky was actually right all along: the evolution of his language is indeed
a mystery. In the last decade, however, Chomsky seems to have relented. In
Hauser, Chomsky, and Fitch (2002), he offers a tentative conceptual solution to
the problem. The solution, however, is so strange, so unconvincing, that the arti-
cle actually seems to require a deeper interpretation—as an implicit statement of
resignation. To begin with, the three authors insist that the “acrimonious debates
in this field” have been launched by the failure to distinguish between “ques-
tions concerning language as a communicative system and questions concerning
the computations underlying this system, such as those underlying recursion.”
Questions about language as a communicative system, the authors state, are
not questions about language as such, but about “the interface between abstract
computation and both sensory-motor and conceptual-intentional interfaces.”
They then make a distinction between two different types of human faculties for
language. The first, FLB, the Faculty of Language in the Broad sense, includes “a
sensory-motor system, a conceptual-intentional system, and the computational
mechanisms for recursion, providing the capacity to generate an infinite range of
expressions from a finite set of elements.” All these are not problematic from the
evolutionary point of view—they may all have precursors in other species—but
they are not what makes human language what it is. The secret of language lies in
the FLN, the Faculty of Language in the Narrow sense, and the evolution of this
faculty presents the theory of evolution with its “deepest challenge.” What, then,
does FLN include? It turns out that it only includes one property—recursion:
The internal architecture of FLN, so conceived, is a topic of much cur-

rent research and debate. Without prejudging the issues, we will, for con-
creteness, adopt a particular conception of this architecture. We assume,
9780190256623-Dor.indb 185 03/06/15 6:18 PM

putting aside the precise mechanisms, that a key component of FLN is a

computational system (narrow syntax) that generates internal representa-
tions and maps them into the sensory-motor interface by the phonologi-
cal system, and into the conceptual-intentional interface by the (formal)
semantic system; adopting alternatives that have been proposed would
not materially modify the ensuing discussion. All approaches agree that a
core property of FLN is recursion, attributed to narrow syntax in the con-
ception just outlined. FLN takes a finite set of elements and yields a po-
tentially infinite array of discrete expressions. This capacity of FLN yields
discrete infinity (a property that also characterizes the natural numbers).
Each of these discrete expressions is then passed to the sensory-motor and
conceptual-intentional systems, which process and elaborate this infor-
mation in the use of language (p. 1571).
This is it, then: the potential capacity for discrete infinity—which is never ob-
served in reality, because certain properties of FLB, such as memory and process-
ing limitations, prevent it from ever materializing—is the one thing that makes
human language such a unique phenomenon in the biological world. And how
did this capacity evolve? To begin with, the hypothesis that FLN only includes
recursion “has the interesting effect of nullifying the argument from design,
and thus rendering the status of FLN as an adaptation open to question. Propo-
nents of the idea that FLN is an adaptation would thus need to supply additional
data or arguments to support this viewpoint.” Until otherwise demonstrated,
FLN is not adaptive. So, how did it evolve? Well, it “may have evolved for rea-
sons other than language,” such as “navigation, number quantification, or social
relationships”:
[W]e suggest that by considering the possibility that FLN evolved for rea-
sons other than language, the comparative door has been opened in a new
and (we think) exciting way. Comparative work has generally focused on
animal communication or the capacity to acquire a human-created lan-
guage. If, however, one entertains the hypothesis that recursion evolved
to solve other computational problems such as navigation, number quan-
tification, or social relationships, then it is possible that other animals
have such abilities, but our research efforts have been targeted at an overly
narrow search space. If we find evidence for recursion in animals, but in
a noncommunicative domain, then we are more likely to pinpoint the
mechanisms underlying this ability and the selective pressures that led to
it. This discovery, in turn, would open the door to another suite of puz-
zles: Why did humans, but no other animal, take the power of recursion
9780190256623-Dor.indb 186 03/06/15 6:18 PM

to create an open-ended and limitless system of communication? Why

does our system of recursion operate over a broader range of elements or
inputs (e.g., numbers, words) than other animals? One possibility, con-
sistent with current thinking in the cognitive sciences, is that recursion
in animals represents a modular system designed for a particular function
(e.g., navigation) and impenetrable with respect to other systems. During
evolution, the modular and highly domain-specific system of recursion
may have become penetrable and domain-general. This opened the way
for humans, perhaps uniquely, to apply the power of recursion to other
problems. This change from domain-specific to domain-general may have
been guided by particular selective pressures, unique to our evolutionary
past, or as a consequence (by-product) of other kinds of neural reorgani-
zation. (p. 1578)
This is where the solution ends. The potential for discrete infinity may have
evolved for other reasons, and it may have become penetrable, in humans, to cog-
nitive domains other than the one within which it originally evolved—which
means that it may be discovered in other animals, in non-communicative do-
mains, and if it were indeed discovered in other animals, the door would be
opened “to another suite of puzzles.” Now, in order to understand the unique
human capacity for language, all we would have to do is figure out why the ca-
pacity remained domain-specific in the other species.
Well, this is really not a solution. Not even a tentative one. There is nothing
here but a weary and desperate attempt to keep the essence of language (what-
ever is left of it) in the realm of mystery—away from the domain of evolution-
ary explanation. Of course, capacities may evolve for one function and then be
adapted for others, and they may also be by-products of other “kinds of neural
reorganizations,” but in such processes the capacities evolve and change to fit
their new functional contexts: they do not simply stay the same. What is even
more problematic is the capacity itself that is thus salvaged from explanation.
After fifty years of research, all that is left is the original assumption of infi-
nite generativity—the idea that everything we ever do and experience, which is
finite by definition, is always an arbitrary obstacle on our way toward the fulfill-
ment and understanding of our infinite linguistic potential. This is a philosoph-
ical assumption, actually a religious assumption, that goes against the very idea
of science. In this sense, the series of articles by Hauser, Chomsky, and Fitch
might be more favorably read as joint statements of resignation: we have tried
to find common ground between linguistics and evolutionary science; as far as
the periphery of language is concerned, we believe there is no real problem; at
its core, however, language still seems to defy the mode of explanation that is at
9780190256623-Dor.indb 187 03/06/15 6:18 PM

the core of evolutionary theory; maybe, only maybe, what we believe about the
core of language might be reconciled with something at the periphery of evo-
lutionary theory; but beyond that, we really have nothing to offer. The mystery
is there to stay.
There is, of course, something deeply ironic in all this. If it were not for
Chomsky’s insistence on the innateness of language, the very question of the
evolution of language (in its modern form) would probably not have emerged
in the first place—and it would definitely not have assumed the central posi-
tion it now holds in the linguistic discourse. It may safely be said that virtually
all the high-quality work in the field of language evolution—almost regardless
of theoretical inclination—has been motivated by the attempt to salvage lan-
guage from the Chomskian paradox of unevolvable innateness: to find a way to
make language (innate or not) evolvable again. Most of the theories in the field
show a serious commitment to the logic of evolutionary theory, a sophisticated
approach to the question of evidence, a deep understanding of multiple causal-
ity, a basic suspicion toward the idea that actual pieces of linguistic knowledge
are encoded in our genes, a much more serious understanding of learning, and
a firm belief in the very idea that Chomsky rejected: that language evolved be-
cause it was functional. Most importantly, many of the theories are committed
to a co-evolutionary view of the process, the idea that from a certain moment on,
the processes of cognitive evolution and cultural evolution were entangled in a
bidirectional spiral of influence (Deacon 1997, Tomasello 1999, 2008, Pinker
2003, Levinson and Jaisson 2006, Evans 2013, Evans and Levinson 2009, Hur-
ford 2007, 2011, Richerson and Boyd 2005).
At their core, however, many of the theories are still formulated as attempts
to answer the very question that Chomsky himself refused to deal with: what
was it that happened to the human mind (or the human brain) that eventually
allowed it to carry language? This is still a cognitive question, and the “mind”
is still singular. A generally accepted research strategy is to look for linguisti-
cally relevant cognitive capacities—in humans and other animals—that may
be shown to have been there before language, thus helping prepare the grounds
for its emergence. Fitch (2010), for example, shows that we share components
of the language capacity (in the FLB sense) with chimpanzees and other pri-
mates, whales and seals, birds and bees, and Hurford (2007, 2011) claims that
the minds of the apes, our closest relatives, are in many different ways already
language-ready. These explanations characteristically focus on those aspects of
language-related cognition that are also involved in other human activities, and
thus highlight the indisputable fact that the use of language requires many ca-
pacities that are cognitively general and multi-functional—and thus potentially
shared with other species. Language did not appear in a cognitive vacuum, and
9780190256623-Dor.indb 188 03/06/15 6:18 PM

in many ways it still reflects properties of pre-linguistic humans, some of which

may be shared with the apes, with other mammals, with other social animals, or
with other animals who have a nervous system.
There is much to be learned from this perspective, but it also suffers from
three interrelated problems. First, not all the cognitive challenges presented by
language are cognitively general and multi-functional. The most important chal-
lenges are actually specifically unique (Jackendoff and Pinker 2005). The specific
capacities that we have for lexical memory, for example, cannot be reduced to
our general capacities for memory (long or short term), and the general auditory
capacities that we share cannot explain the unique ways in which we compute
linguistic sound. We are also anatomically and physiologically adapted to fast
speech (Lieberman 1991, 2007). How could such capacities, uniquely dedicated
to language, emerge before language itself? Second, the theories have eventually
very little to say about the actual evolutionary dynamics that brought language
about. They only attempt to show how the evolution of language became possi-
ble, not how it actually happened. Third, while Chomsky’s view fails to squeeze
in through the bottleneck of evolution, the theories pass it too easily—so easily,
in fact, that they raise the opposite question: if so much is shared with the other
species, why haven’t they developed language? Why does Kanzi, for example,
with all his enormous achievements, only manage to get to the brink of language
(Savage-Rumbaugh and Lewin 1994)? As Bickerton (2008, p. 288–289) puts it
in his review of Hurford (2007), the belief that “the cognitive and communi-
cative capacities of great apes brought them to the brink of language” implies,
among other things, that “minor improvements in ape cognition and com-
munication gradually accumulated until some progenitor of humans became
‘language-ready,’ so that the actual transition to language was no big deal.” But if
this is the case, why did the improvements only accumulate in one species?
(W)hy not in all, or at least one or two others? Why are these improve-
ments not continuing in modern apes, so that we can observe them in
action? Why is it that while we have thousands of complex languages
with convoluted structures and tens or hundreds of thousands of words
in each, they have communication systems resembling those of birds and
fish? Why, while we are making moon-landings and sonatas are they still
fishing for termites and cracking palm-nuts? (p. 289)
The answer I would like to propose follows a major line of thinking in the general
field of human evolution (cf. Suddendorf 2013) that has recently been gaining
ground in language evolution too: the real story is social-technological, not cog-
nitive. What separates us from the apes is a sequence of social and technological
9780190256623-Dor.indb 189 03/06/15 6:18 PM

revolutions—one major change after the other in the life experiences of human
communities. The emergence of language was one such revolution, not the last
and definitely not the first. It was preceded by an entire history of revolutions,
all of which brought human communities to the point where they could invent
language—as yet another feat of collective genius. As language gradually estab
lished itself in human communities, individuals began to be selected for the capacity
to meet the growing challenges of the new technology. It was the collective inven-
tion that eventually shaped the cognitions of its users, not the other way around.
10.1 Re-Formulating the Questions of Evolution

The new perspective is founded on a major theoretical re-arrangement of the
two issues we are interested in. The question of the evolution of language is
no longer a cognitive question: it has to do with the evolutionary history of
the technology—its invention, development, propagation, and diversification,
the social contexts within which it emerged in ancient human communities,
the ways it changed society once it was established, and so on. It is a question
about the social-technological development of humanity. The question of the
evolution of human minds (in the plural) and their relations with the emergent
technology is thus secondary: it has to do with the involvement of individual
human minds in a technologically-driven process.
Unfortunately, most of what we would like to know about the social-
technological evolution of language is probably buried forever in the past. Where
was language invented? When exactly did that happen? How many times, and
in how many communities, was language invented, then forgotten, then re-
invented, before it stabilized in some communities as a regular element of social
life? How long did it take before language reached the moment of universal
spread? How long ago did we still have communities of non-speaking humans?
We may, however, make a number of assumptions, all of which are based on what
we know about the evolution of other technologies (Arthur 2007). We know
that first prototypes, the first versions of a new technology that actually work,
do not look anywhere close to the final version of the technology. The first pro-
totypes of language were much less complex, much less sophisticated, much less
efficient, than the language of the world as we know them today. We know that
necessity, not just capacity, is the mother of all invention: the absolute need to
solve a problem inspires an exploratory process that eventually stumbles upon a
good-enough solution, which is often identified as such only in retrospect (and
because of that, inventions always require a considerable amount of luck.) We
know that the understanding of an invention requires an understanding of its
9780190256623-Dor.indb 190 03/06/15 6:18 PM

social-technological context. Innovations never appear out of nowhere. They

always have a past—older technologies that came to be modified and combined
in new ways to solve new problems; old obstacles that had to be removed; or old
problems that came to be solved by new means. The context, moreover, heavily
determines the future of innovations: many pre-conditions have to be met for
innovations to eventually be accepted, stabilized, and propagated. And we know
that innovations further evolve together with their contexts: to the extent that
an invention proves to be as efficient as it promised to be, it actually begins to
change its own environment; the modifications to the environment call for new
functions, which effect changes in the technology itself; and so the technology
and everything around it come to be entangled in co-evolutionary spirals.
The first prototype of an invention, any invention, makes a difference because
it makes certain things possible that could not be done before—at a very rudi-
mentary level of success. As the first users begin to actually work with the pro-
totype, as they start to accumulate their experiences as users, they begin to learn
things about the interactions between the invention and the environment that
could not have been known before. They gradually understand more about the
capacities of the technology and the ways it should be used, and they discover
some of problems (totally new problems) that arise as the technology interacts
with its environment. For a long time, and very often unintentionally, different
users thus introduce many slight modifications to the prototype (and to the way
it is used), which in their turn have a cumulative and gradual effect on the general
efficiency of the system. The effect is quantitative: the system remains the same,
it just gets better. In some cases, this long line of accumulated improvements
eventually leads to the stabilization of the invention in its final form. In other
cases, the quantitative process eventually translates into a qualitative effect—a
technological revolution. The revolution may occur as a direct result of the grad-
ual accumulation of modifications. The system reaches a critical point at which
totally new patterns of usage are all of a sudden made possible. It may also occur
because of the accumulation of problems. New problems arise as the system
gradually improves, and a new set of modifications, which emerges as an attempt
to solve these problems, turns out to open up entirely new functional capacities.
Sometimes, the revolution is made possible only when the environment (phys-
ical, social, and technological) changes in a specific way. This way or the other,
the system that eventually emerges from all this should be properly thought of
as the first prototype of the next generation of the technology. It is qualitatively
different from the first generation in its architecture, and it can do things that
were outside the functional envelope of the first generation—again, at a very ru-
dimentary level of success. Consequently, the system now enters a new phase of
gradual evolution, in which new capacities and new problems arise and many
9780190256623-Dor.indb 191 03/06/15 6:18 PM

slight modifications are introduced. Then another revolution occurs, the next
generation of the technology appears, and so on and so forth. The process goes on
as long as it remains both necessary and possible. All the complex technologies
that we use today are the products of such evolutionary histories. They are all re-
lated to some original invention through a long and complex line of revolutions
and gradual modifications. There is no reason why language should be an excep-
tion. Finally, we know that technological systems, as they evolve, impose more
and more system constraints on the possible venues of their own future evolution.
The system acquires a certain specificity that gradually makes certain types of
changes more difficult to incorporate, and thus in effect participates in directing
its own evolution.
It should now be quite clear why the question of the evolution of language
as a social technology should be dealt with before the question of the evolution
of speakers can even be considered: throughout the entire process, at different
stages of the evolution of language, different human individuals, who occupied
different positions vis-à-vis the evolving technology and vis-à-vis their groups,
found themselves facing different cognitive challenges. Between the linguis-
tic leaders—the inventors and developers of the invention—and all those who
could not yet handle the technology, there were learners and imitators (with dif-
ferent capacities), co-operators (in different social positions), speakers and listen-
ers (more or less competent), passive listeners and eavesdroppers, and probably,
from a certain point on, multi-linguals, liars, and translators. Each had to be able
to do different things. We need to understand these different challenges, and
then ask: How did these individuals manage to cope? How did they recruit the
required capacities?
Once we frame the question this way, two points become obvious. First, the
cognitive challenges of language, throughout its evolution, grew and developed
along with it. The use of the first prototype in its original form required cer-
tain sets of capacities; the more advanced versions of the prototype gradually
required more of these capacities; the first revolutionary advance in the system
brought with it new challenges that required new capacities, which were then
gradually stretched as the new generation of the technology entered the phase
of gradual growth, and so on and so forth. The cognitive challenges of the first
speakers of language were not our challenges. The challenges themselves evolved.
Second, not everybody who should have met the challenges of language, at any
of the stages, actually did. Many along the way probably failed, and many others
only met some of the challenges, at different levels of success. This is what hap-
pens with every technology. From the very beginning, then, the existence of lan-
guage created new patterns of variability between individuals—in their roles as
language users—patterns of variability that asserted themselves as such for the
9780190256623-Dor.indb 192 03/06/15 6:18 PM

first time. Some of the variability probably emerged from the social positioning
of the different individuals vis-à-vis language: the level of their exposure to the
system, their status and rank, their social ability to actually use language to their
advantage. Some of the variability, however, must have been causally connected
to variability in cognitive capacity—some of which must have been related to
parallel patterns of genetic variability. Other things being equal (and they never
really are), we may assume that those who found the technology easier to learn,
handle, and improve—those whose cognitions and genetics were more suited for
the task—earned more dividends from the system (both individually and col-
lectively), and were gradually selected over those whose cognitions were less com-
patible with the technology.
The question, however, still remains: where did those individuals who did
manage to meet the challenges of language, at the different stages of its evolu-
tion, find the capacities to do it? The traditional, cognitively oriented answer
would be that the relevant capacities must have already been there before: behav-
ior is made possible by pre-existing capacity, and because of that, the explanation
for the capacity should be sought somewhere else, away from language and its
challenges—in the structures of the brain (or the mind) and the structures of our
genes. Recent advances in evolutionary theory, however, allow for a very differ-
ent answer: the capacities evolved together with language—for language. First we
invented language collectively, then language changed us individually. To see how
this is possible, we have to change something fundamental in our perception of
the general relationship between behavior, capacity, genetics—and innovation.
10.2 Innovation, Behavior, Capacity, and Genes

In the last decade or so, many of the most sophisticated new theories in evolu-
tionary biology, especially in the domain of Evolutionary-Developmental Theory
(evo-devo), have been informed by the understanding that behavioral innovation
plays a much more important role in the evolution of biological species than has
previously been assumed (West-Eberhard 2003, Jablonka and Lamb 2005). Let
us consider the simplest scenario of evolution: a number of individual members
of a biological species, which are fairly reasonably adapted to the conditions of
their environment, who all of a sudden find themselves losing ground because
the environment begins to change—bringing with it new problems to which
the individuals have not yet had the opportunity to adapt. In the traditional,
gene-centered view of evolution, these individuals have no choice but to go on
behaving as if the environment has not changed (their arsenal of possible behav-
iors has already been genetically fixed), wait (so to speak) for a helpful genetic
9780190256623-Dor.indb 193 03/06/15 6:18 PM

mutation, and let natural selection determine their fate. This, however, is not
what actually happens. Biological organisms react to environmental changes and
launch processes of exploration in which they try all kinds of behaviors they have
not been genetically adapted to before. Terrestrial mammals are not adapted to
swimming, but when they find themselves surrounded by water they neverthe-
less do what they can to keep afloat. When the environment gets colder, animals
look for shelter in places they have never entered (or even noticed) before. The
chaotic nature of this process is most clearly revealed when we think about it in
the most down-to-earth terms: the new environmental conditions raise the level
of the animal’s stress, and stress brings about behaviors that are unorderly, ex-
ploratory, accidental, and sometimes even frantic (depending on the severity of
the threat). Most behaviors do not help much, but from time to time an animal
stumbles upon a behavior that actually lets it survive, at least for a while. In these
cases, the animals survive not because they were already genetically prepared for
the new circumstances, but the other way around: they survive because they were
capable of behaving outside the confines of their genetically selected-for behav-
ioral envelope.
Biological organisms, then, are adapted in different ways to their environ-
ments, but way and above these adaptations they are also adapted—to different
degrees and in different ways—to the foundational fact that their environments
keep changing. They are capable of innovation. In the biological literature, this
capacity is referred to as the capacity for plasticity. There is, of course, a genetic
foundation for the capacity of plasticity, and different species are capable of dif-
ferent types and levels of innovation. This fact will come to play a major role in
our analysis. It is also crucial to understand, however, that this genetic founda-
tion has very little to do with the actual behavioral products of the exploration
processes. Innovating organisms are genetically prepared for the search for behav-
iors that break the specific, genetically selected-for mold of the regular patterns
of their lives. The innovations themselves emerge from the search process itself.
Luckily stumbling upon a successful behavior, however, is only the begin-
ning of a much longer process—in which the new behavior has to be stabilized
as part of the behavioral arsenal of the organisms. The organisms have to iden-
tify the successful behavior, isolate it from the other accidental behaviors that
were not helpful, understand something (at the relevant cognitive level) about
the causal connection between the behavior and its functional output, learn how
and when to initiate it systematically—and, eventually, get used to it. Different
species, again, are different from each other in terms of their ability to stabilize
a new behavior. Assuming, then, for the sake of simplicity, that in our scenario
the environment changes and then stabilizes again with a new set of conditions,
those individual members of the species that would survive the ordeal, regardless
9780190256623-Dor.indb 194 03/06/15 6:18 PM

of their genetic makeup, would survive on the basis of what they actually man-
aged to achieve with their innovations—everything that they managed to learn
and apply in their relationships with the new environment.
The next step in the argument takes things to an entirely different level of
complexity. Assume that some of our innovative individuals manage to survive in
the new environment—and eventually multiply. Their offspring will now be born
into a world in which the new stabilized behavior is simply there. They would
not have to re-invent. They would have to learn. None of the offspring would be
already genetically adapted to the task: the capacity for the innovative behavior of
their parents was not passed on to them in their genes. The fact that the behavior
is now already in their world, however, would radically change the way their genes
would express themselves in the process of their ontogenetic development. The
young organisms would have no choice, in the course of their development, but
to launch an exploratory process of their own, recruit as many of their genetically
given capacities as possible (capacities that evolved for other purposes), combine
them in innovative ways—and attempt to do whatever they can, with the tools
they have, in order to master the behavior. In the course of this effort, then, a to-
tally new pattern of cognitive and genetic variability among the learners would be
exposed—variability in the types and the qualities of the genetically given capaci-
ties that the different learners can recruit, and combine, for the new learning task.
This variability, in its turn, would assert itself in two complementary ways.
First, different learners would eventually adopt (and stabilize) different strate-
gies for the learning task. This is so, because the different learners would rely
on different capacities, of different qualities, and would thus have no choice but
to attack the learning problem from different angles. Second, different learn-
ers would eventually master the behavior to different degrees. Some would find
the learning task relatively easy, many would find it challenging but possible,
others would find it very difficult, even impossible. To the extent that the be-
havior remains obligatory for survival, these differences would reflect themselves
in the ability of the learners to multiply—which means that they would also be
reflected in the patterns of genetic distribution in the next generation. Specific
combinations of genes, which only came to be functionally related to each other
because they were recruited for the new task, would now be selected for—and
the next generation would actually be more genetically adapted to the behavior
(only more adapted, never totally). This process is called in the biological liter-
ature genetic accommodation: genes accommodating themselves to innovations
(West-Eberhard 2003), capacities accommodating themselves to behaviors. Not
the other way around.
In the simplest scenario, then, the process that leads from the change in the
external conditions of life to the change in the distribution of genes across the
9780190256623-Dor.indb 195 03/06/15 6:18 PM

population involves necessity, exploration, luck, innovation, stabilization, and

then learning, exploration again, recruitment and re-combination, exposure of
cognitive and genetic variability, strategy stabilization, and then, eventually, ge-
netic accommodation. It is a very different process indeed from the one envi-
sioned in the traditional, gene-centered conception of evolution. Note that none
of this implies that the more traditional mechanisms of evolutionary change
are no longer there. Mutations, and other molecular changes such as genetic re-
shuffling, still occur and remain crucial. The point is that the innovations have
a direct influence on the way the products of these genetic changes end up ex-
pressing themselves. They change the general pool of genetic variability, and are
selected for, or against, on the basis of their contribution to the effort of inven-
tion and stabilization.
Finally, as Jablonka and Lamb (2005) show, once a process of this type is
launched, and as long as certain conditions are met, the emergence of new ca-
pacities may lead to the further development and refinement of the innovation
itself. As noted, the capacities for plasticity manifested by different species (and
different members of different species) vary—but they all share a common prop-
erty: they are all finite. The capacity for invention, and for the learning process
that follows, is never completely open-ended. The further an innovation is from
the envelope of the already-adapted behaviors of the individual, the more dif-
ficult it would be for the individual to invent and stabilize it. Because of that,
individuals of later generations, who have by now adapted themselves, at least
partially, to the behavior invented by their ancestors, would now be able—if re-
quired by necessity—to invent and stabilize behaviors that were outside the ca-
pacity of their ancestors. Such additional innovations, to the extent that they
prove useful, would launch another process of learning, exposure of genetic var-
iability and eventually genetic accommodation, and then more innovation—as
long as necessity is there. Jablonka and Lamb refer to all this as the assimilate-
stretch dynamic: innovative behaviors become easy to accomplish because of ge-
netic accommodation, so individuals can stretch their behavioral envelopes by
further innovations, then assimilation occurs again, and so on and so forth. The
evolutionary paths of the innovation and its users find themselves entangled in
co-evolutionary spirals.
It is crucial to see, then, that in the course of this complex process new ca-
pacities emerge that are not just re-combinations of capacities that were already
there before. Behavioral innovation produces cognitive novelty. New behavioral
patterns are forced into existence by necessity; they are gradually carved by ex-
perience to approximate their specific functional goals; they become objects of
learning, and eventually mold capacities in their shape. It is thus not the case
that behavior is based on already existing capacity: capacity actually emerges
9780190256623-Dor.indb 196 03/06/15 6:18 PM

from behavior. We are never already capable before we begin. We gradually

become capable as we try. Skill emerges from practice, not the other way around.
Quite obviously, new capacities are never totally unrelated to their past: pre-
adaptations play an important role in the story. New capacities, however, emerge
from the interaction—made possible by plasticity—between old capacities and
new necessities, and because of that, they are never just reflections of their past.
They really are new.
Based on this general perspective, then, Eva Jablonka and I have developed
a principled model of the co-evolutionary dynamic of language and its speakers
(Dor and Jablonka 2000, 2001, 2010, 2014). As we show, pre-linguistic humans
must have already been ready (socially and cognitively) for the beginning of the
process of exploration that eventually gave birth to the first prototype (or pro-
totypes) of language—but they did not have to be cognitively language-ready
before language came into being. Language was, in all probability, invented
before its speakers were fully prepared for it. It was born out of necessity, be-
tween human individuals, on the basis of plasticity, and its technological evolu-
tion was the driving engine of the entire process. The cyclical collective dynamics
of invention, negotiation, propagation, habit formation, and conventionaliza-
tion (and then more inventions, and so on and so forth) remained ahead of the
individuals who were involved in it. Speakers were struggling to keep up with
language, and whenever they managed to adapt to it, it was already somewhere
else, further down the road of evolutionary development. In the course of the
process, language itself gradually developed into a highly specialized system,
with unique technological properties. Consequently, acquiring sufficient skill
with the system came to require unique capacities, which means that human
individuals as we know them today do indeed have a cognition that is partially
biased toward the acquisition and use of language. We do have innate capacities
for language, but these capacities are derivative, emergent, variable, and partial—
not constitutive, foundational, universal, and complete. Innateness is a poste-
riori, not a priori.
And it is not only cognitive—it is also emotional (Jablonka, Ginsburg, and
Dor 2012). The deepest indication that we are by nature destined to participate
in the social activity of language is not the fact that we can do it—but the fact
that we need it, that we crave it. Throughout the evolution of language, indi-
viduals were not just selected for the capacity to participate in the activity of
language—but also for their will to do it. Those who were more deeply attracted
to the evolving technology, who were more desperate to understand and to talk,
who longed for mutual-identification—all these simply spent more time around
language; instead of doing other things, they spent more time mutually identi-
fying from an earlier age, and invested more energy in the acquisition of their
9780190256623-Dor.indb 197 03/06/15 6:18 PM

linguistic skills. Their fascination with language thus increased their chances
of survival, and was thus partially genetically assimilated. We are already born
with (different levels of) this fascination. Our minds are language-craving. This
is why, as children, we actively look for language. Not because we already know it
(or parts of it), but because we want it. Obviously, we also crave experiential com-
munication, and in this sense, we are not very different from many other species.
Other animals, just like us, can feel lonely without social contact. Only humans,
however, are born with the hunger for the type of social contact that can only be
achieved by mutual-identification.
10.3 The Pre-History of Language

It is on the basis of this conception of the evolutionary dynamics that we may
now try to delve deeper into the actual process. Our first task is to figure out
as much as we can about the social, technological, and communicative lives of
pre-linguistic humans, to find clues to the context that made the invention of
language both possible and necessary. In 2011, Chris Knight, Jerome Lewis, and
I invited twenty-four scholars, from a wide array of disciplines, to discuss this
question in an intensive workshop in London. The volume based on the work-
shop, Dor, Knight, and Lewis (2014), is the first collective attempt in the liter-
ature to systematically construct a synthetic picture of pre-linguistic social and
technological life, rich and detailed enough to shed light on the social origins of
language. We, as modern humans, are so used to language that we find it difficult
to imagine social life without it, but the emerging picture of our pre-linguistic
ancestors reveals a level of social, technological, and communicative complexity
and sophistication much closer to our own than to ape societies. There is a sober-
ing lesson here: being human is not all about language.
For a million and a half years or more, before the emergence of Homo sa-
piens, archaic human communities had already been constantly moving away
from the ape-like societies of their ancestors toward human social life as we
know it. At least two hominin species were involved—Homo erectus and Homo
heidelbergensis—and although it is hard to figure out exactly which species con-
tributed what to the process, there was a clear common thread: throughout the
process, human survival gradually came to depend less on individual behavior
and more on collective co-operation. Individuals gradually came to depend
more on others. This dependency required more and more communication—
still at the level of experiential presentation, still within the here-and-now of the
communication event—and it required individuals to become more and more
socially sensitive. Higher levels of communication and sensitivity allowed for
9780190256623-Dor.indb 198 03/06/15 6:18 PM

further social and technological developments, and so on. Pre-linguistic human

societies were already deeply entangled in a unique spiral of social, technolog-
ical, communicative, cognitive, and emotional co-evolution—and it was from
this spiral that language eventually emerged. Different scholars in the field of
human evolution highlight different facets of this process, sometimes present-
ing them as singular explanations for the entire drama, but it is important to
remember that all facets were also connected to each other in what Sterelny
(2012) calls feedback loops—re-inforcing, directing, and shaping each other.
Hominin societies were changing in all possible ways, none of which can ex-
plain the process in isolation from the others.
Wrangham (2009) highlights the control of fire and the invention of cook-
ing. As he shows, cooking dramatically increased the amount of energy ancient
humans obtained from their food, and consequently changed human life at all
the relevant levels. We adapted to cooking anatomically and physiologically—
our digestive systems got smaller, our jaws weaker—and the reduced costs of
feeding left much more energy for brain growth. Our ape relatives spend around
six hours a day chewing raw food; our cooking ancestors had much more free
time. They ate around the fire, which turned into the central site of social and
cultural life. Most importantly for our present purposes, cooking created a new
division of labor between the sexes: men hunted; women gathered and cooked.
Released from the need to spend most of the day eating, the men could now con-
centrate on bringing more meat back to the camp. As a result, men and women
came to depend on each other for their subsistence. Based on a wide array of
evidence, Wrangham identifies the invention of cooking with the rise of Homo
erectus, about 1.8 million years ago.
Hrdy (2009) tells of another story that probably unfolded in the days of
Homo erectus. Erectus babies were already taking much more time to mature
than their ape peers, and they required constant feeding and protection for
a long time after weaning. From a certain point on, “human mothers began
to bear offspring too costly to rear by themselves” (p. 283). What evolved
was a social arrangement unique among the primates—alloparenting. Babies
began to be taken care of collectively, not just by the mother, but also by fa-
thers, grandmothers, and other family members. This created a complex web
of practical and emotional dependencies, between mother and alloparents,
between alloparents and children. As Hrdy shows, the emergence of co-
operative breeding must have been a major driving engine behind the evolu-
tion of human cognitions and emotions as we know them today—the unpar-
alleled will and capacity to figure out what others are thinking and feeling,
empathy, the uniquely human feelings of shame and guilt (cf. Jablonka, Gins-
burg, and Dor 2012).
9780190256623-Dor.indb 199 03/06/15 6:18 PM

Sterelny (2012) and Tomasello et al. (2012) concentrate on yet another rev-
olutionary development: the emergence of collaborative foraging, especially the
collaborative hunting of big game—a complex and risky endeavour that requires
high levels of group planning and co-operation on site, but also promises more
dividends for the individual hunter than he could expect to gain from hunt-
ing alone. Collaborative hunting is clearly evidenced in Homo heidelbergensis,
around four hundred thousand years ago, but it may also have earlier beginnings.
It created new types of dependencies, and it probably contributed much to the
emergence of the uniquely human sense of group identity that is based on the
sharing of food. Most importantly, it required much higher levels of skill, and
gradually came to depend on advances in tool manufacture, from stone knives
to projectile weapons. As Sterelny (2012) convincingly shows, all this must have
gradually created a totally new type of dependency: the survival of the group
came to depend on the ability of experts to pass their knowledge to the young.
What emerged was a regime of apprenticeship, in which adult experts began to
actively intervene in the learning processes of the young and organize their own
activities in ways conducive to learning. This was the beginning of pedagogy, a
uniquely human social activity.
It is easy to see how these dynamics and others, once launched, would spiral
together very quickly and begin to re-inforce and shape each other. It is also
clear that they all require more and more communication, more and more in-
formation sharing. What this suggests is that ancient human societies, for a
very long time before language, must have already been raising their levels of
communication in revolutionary ways. The best discussion of this side of the
drama is still Donald (1991). As he shows, Homo erectus societies must have
already turned from episodic to mimetic. Ape societies are episodic: commu-
nication between individuals, as sophisticated as it is, is still unreflective, con-
crete, and situation-bound. Erectus societies added an entirely new dimension
to the episodic modes of communication they inherited from their ancestors:
mimetic communication combines mimicry, imitation, gesture, tone of voice,
facial expression, bodily movement, and eye contact to produce intentional
and reflective communicative acts. It includes, quite simply, everything we use
in a game of charades. It was exactly this new mode of communication that al-
lowed human societies to meet the growing challenges of dependency. Donald
demonstrates this with the all-important development of pedagogy. The
manufacture of erectus tools of the more advanced, Acheulian type, requires
months of training. The skill cannot be learned just by imitation, and mimesis
is still, today, the most efficient way to teach it: going through the motions
more slowly, intentionally freezing at different points in the process, pointing
at this facet or other of the work, expressing frustration and satisfaction, and
9780190256623-Dor.indb 200 03/06/15 6:18 PM

so on and so forth. As we saw in chapter 2, this is also the case with all other
manual skills that we learn and teach. Donald also highlights the significance
of mimesis in the communication of emotions, private and social—in such
revolutionary forms of expression as pantomime, mimicry, music (mainly sing-
ing), dance, and ritual. All these, of course, are still extremely important in
our lives as modern humans. Donald’s argument strongly suggests that they
are much older than language. Much of the work reported in Dor, Knight, and
Lewis (2014) shows that, in all probability, this is indeed the case. Language
was born into a human social world already suffused with polymodalic com-
munication (Kendon 2014, Lewis 2014).
The new means of communication allowed for the transfer of knowledge and
identity between generations, and thus for the emergence of Tomasello’s (1999)
ratchet effect—the stable accumulation of innovations. They participated in the
achievement of the levels of trust required for collective work (Knight 2014), in
the establishment of normativity (Zlatev 2014, Lamm 2014), in the emergence
of new forms of play-and-display (Whitehead 2014), including the display of the
self, and in further, ritual-based developments in the relations between the sexes
(Power 2014, Watts 2014). Most importantly, they allowed for a revolutionary
rise in the very ability of human societies for collaborative innovation (Dor and
Jablonka 2014). Research shows that apes innovate too (McGrew 1992; Whiten
et al. 1999; Yamamoto et al. 2008), but they very rarely do it together. The
process of collaborative innovation, in which different individuals, with their
different experiential perspectives on the problem, work together to find a solu-
tion that none of them could solve alone—this is a uniquely human capacity
(Glăveanu 2011).
All this makes for an extremely complex and dynamic view of pre-linguistic
societies, but I would like to claim that at a more abstract level, everything that
was happening revolved around one thing: the collective effort of experiential
mutual-identification. Our pre-linguistic ancestors managed to achieve what
they did because they spent enormous amounts of collective effort in the strug-
gle for mutual understanding, mapping the differences and similarities between
their experiential worldviews, learning from each other and teaching each other.
They gradually spent more and more of their time doing things together, solving
problems together, sharing and comparing experiences. This was the first revolu-
tion that made us who we are. The apes have a theory of mind: they understand
that the other may have a different picture of the world, and they are capable of
reading the other’s intentions, follow the other’s gaze, and so on. This, however,
is still an individualistic capacity. The human breakthrough was the upgrade of
this capacity into a collective, mutualistic, dialogical capacity—and its establish-
ment as the single most important determinant of human life.
9780190256623-Dor.indb 201 03/06/15 6:18 PM

As far as the invention of language is concerned, the implications of this per-

spective are much more technical: it explains why, at a certain point in time,
the invention became both necessary and possible. The invention became pos-
sible because experiential mutual-identification is the machinery required for
the construction of language. Humans had already been systematically and effi-
ciently mutually identifying their experiences. What was required was a radical
change in the use of something that was already there. The invention became
necessary because the growing dependency on experiential mutual-identification
locked humanity in a vicious circle—an extremely beneficial circle but a vicious
one nevertheless: the ever-growing dependency of the community members on
mutual-identification required a constant rise in the amount and quality of the
information that could be shared and compared among the group; the rise in
information sharing, however, only contributed to the deepening of the depend-
ency. We are still trapped in this circle: the collective understandings that we
manage to establish are always a few steps behind the collective problems we
have to solve. This vicious circle, I would like to suggest, eventually forced pre-
linguistic humans—Homo erectus or Homo heidelbergensis—to begin their ex-
plorations into a new realm of communication. They had already expanded the
functional envelope of experiential communication to the maximum, but they
needed more. Everything that could be shown was already shown, but this was
no longer enough. They had to find a way to do what no other species could even
dream about: communicate what could not be communicated experientially.
The goal, then, was determined by necessity; the capacities and machinery were
already there. The only thing still missing was the functional strategy.
10.4 Crossing the Rubicon: From Experience to Truth

With all their revolutions, pre-linguistic humans were still living in a social
world defined by the here and now. Communicators could systematically nego-
tiate their experiences only if both or all of them managed to experience them
together. Handling situations in which the thing to experience was outside the
experiencing range of the interlocutors remained beyond the functional limits
of the entire system. With bodily and vocal mimesis, they helped each other
experience—you see, you point, I see what I hadn’t noticed myself, we look each
other in the eye and acknowledge—but this they still did as perfectly experiential
animals. Like all other species, they only knew how to follow their own senses.
So, in the simplest scenario, if individual A pointed at x (a prey, a predator, other
people, fire) and accompanied the pointing with some mimetic sound associated
with x, and if individual B saw A pointing, looked in the direction pointed at and
9780190256623-Dor.indb 202 03/06/15 6:18 PM

identified x, then all went well: mutual-identification has been achieved. But if x
was positioned outside B’s field of vision, the act failed. If anything, it widened
the experiential gap between them.
Turning such failures into success was exactly the pressing challenge. It must
have emerged slowly but consistently, in more and more severe instances of epi-
stemic dependency (Dor 2014), situations where (i) A experienced something that
called for action, but he or she could not act alone on the basis of the experience;
(ii) another individual, B, was in a position to act but had not experienced the
call for action; and (iii) the survival of both depended on A’s capacity to get B to
do what was needed. The challenge of epistemic dependency required a radical
change of attitude: the failure would turn into success if B managed to interpret
A’s communicative act not as an invitation to experience—but as an invitation
to imagine. B would have to understand (without words): “A is intentionally at-
tempting to turn my attention to something by pointing. His or her vocalization
indicates that it is of the type x. As for myself, I cannot see anything there. I will,
however, choose to go against my own experiential judgment, believe A’s experi-
ential judgment, imagine there is something there of the type x, and act upon my
imagination.” For me, this was the essence of the linguistic revolution: the emer-
gence of the will and capacity to imagine what you cannot see with your own
eyes, simply because you believe somebody else.
This, I would like to suggest, is what the inventors of language began to
experiment with: not the construction of a new system, but the use of the old
tools of experiential-mimetic communication for a new type of communicative
function—based on experiential trust. They were trying to use everything they
already had in order to refer to unshowable experiences as unseen instances of
what was already familiar: “We all know x, and there’s one of them, there, where
you cannot see.” Whenever they managed to pull this off, they actually turned
their mimetic signals into proto-linguistic signs, still holistic and analogue, but
already performing the task of instruction. This is why, in the beginning, there
was probably nothing perceptively different about the explorations: they looked
and sounded like regular events of experiential-mimetic communication. The
new function, however, must have asserted itself quite quickly as a revolution.
The epistemic reach of the inventors of language began to expand beyond what
they experienced themselves—alone or together. More and more elements of the
world began to penetrate their worlds from the outside: things they had not ex-
perienced by themselves (alone or together), but had been told about. For the
first time in the evolution of life, humans began to experience for others, and let
others experience for them. The consequences were enormous, both in terms of
the growing set of practical challenges facing the innovators’ communities, and
in terms of everything social. The Rubicon of experience was crossed.
9780190256623-Dor.indb 203 03/06/15 6:18 PM

This, however, did not come without a price. As long as communication is

experiential, the interlocutors always maintain the capacity to verify the commu-
nicated meaning at the time of the communication event, to see with their own
eyes. When immediate verification is always possible, the principled problem of
truth does not arise. It was exactly this sense of experiential confidence that the
inventors of language, in their explorations into the instruction of imagination,
had to sacrifice: the new function was based on the replacement of knowledge
with belief. It is only when we begin to count on what other people tell us that we
seriously begin to wonder: is this right? Are they telling us the truth? Could they
be mistaken? Is this a lie? The formal semantic understanding of truth as the
relationship between linguistic propositions and the world thus captures some-
thing very deep (as long as we remember to replace the world with the world
of experience): the problem of truth as we know it was born together with lan-
guage. It changed humanity forever. Among other things, it created new types
of social emotions (and new ways to manipulate them): truth-related anxieties
such as doubt and suspicion (Jablonka, Ginsburg, and Dor 2012). Hundreds of
thousands years later, modern humans began to come back to this foundational
relationship between language and truth, knowledge and belief, anxiety and
doubt—this time as reflective thinkers trying to understand themselves. This
was the birth of philosophy.
10.5 Toward a History of Language

If all this is on the right track, it allows for a very specific hypothesis concern-
ing the further evolutionary development of language and its speakers, from the
stabilization of the first exploratory beginnings to the stabilization of the tech-
nology of language as we know it today. The entire process was pushed forward
by the constant need to raise the levels of success in instances of instruction. The
function gradually shaped the technology, not the other way around: it forced
the emergence and further improvement of components and properties of the
technology; it created new problems that had to be solved; it directed the path
of development from beginning to end. Language as a specialized, autonomous
technology, with its fundamental characteristics, was the final result of this
process—but much of it was already in the cards at the moment of origin. In the
following, then, I would like to tell the story the way I see it. It is speculative, of
course, it must be, but I believe it captures something of essence about the causal
chain that took us from there to here.
Let us, then, get back to the inventors of language and their heirs, who had
already stabilized the new function and were now seeing their worlds begin to
9780190256623-Dor.indb 204 03/06/15 6:18 PM

expand. What did they have to do in order to increase the efficiency of their new
technology? Two major challenges clearly asserted themselves—on both ends of
the communication process. The first had to do with meaning. From the moment
they stabilized the new function, the users of language had a vested interest in
every minute advance they could achieve in the mutual-identification of their
experiences: the mutual-identification of another sign, pointing at a new type
of experience; the dissection of the experience into mutually identified compo-
nents; the mutual-identification of causal relations between experiences, and so
on. Every advance immediately allowed for the instruction of imagination into
further fields, in more precise ways. Very slowly at the beginning, and probably
faster later on, the first generations of language users must have begun to spend
yet more collective energy on the process of experiential mutual-identification,
dissection, and categorization—this time, however, for instructive communica-
tion. Gradually, it became clear that the instructive function demanded noth-
ing less than an entirely new outlook on the world of experience: looking at the
world in order to behave in it is not the same thing as looking at the world in
order to tell about it. Many experiential domains, which were of little interest to
experiential communication, began to enter the picture for the first time. The
physical terrain, for example, is always given as part of the context of experiential
communication. Everything in it can be pointed at. The direction of interlocu-
tors to places they have never been to, however, requires an entire project of clas-
sification, and eventually the creation of a new semantic field. All this, then, was
the beginning of the symbolic landscape.
The second challenge had to do with the old vocalizations and gestures used
by the first speakers for the instruction of imagination. On the analogue con-
tinuum of experiential communication (vocalic and manual, pre-mimetic and
mimetic), the physical and emotional variability between individuals is highly
functional. It is meaningful. The instructive function, however, demands that all
speakers mutually identify the same gestures and vocalizations for the same mu-
tually identified experiences—transcending individual differences. Under such
a demand, then, every minute change in the arsenal of vocalization and gestures
would be selected for if it provided higher levels of perceptual distinctiveness—
and thus minimized the probability of confusion. As Zuidema and de Boer
(2009) show, the accumulation of such changes would eventually produce a cat-
egorical and combinatorial phonetic system. Zuidema and de Boer stress the fact
that the process requires a significant level of noise: without it, the probability
of confusion is too low. This is important, because the entire process was indeed
embedded from the very beginning within the very noisy world of experiential-
mimetic communication. The challenge was not the construction of a sound and
gesture system out of nothing: it was the isolation of a distinct sound and gesture
9780190256623-Dor.indb 205 03/06/15 6:18 PM

system from the analogue continuum of experiential communication. What this

means, in simple words, is that instructive interactions gradually began to sound
differently. This was the beginning of phonetics.
For a very long time, then, the same thing was happening on both sides of
the communication process. At the level of meaning, the mutually identified
worldview of the symbolic landscape was beginning to demarcate itself from the
experiential worlds (private and collective) of its speakers. At the level of form,
increasingly phonetic vocalizations were beginning to demarcate the sounds
of language from the sounds of experiential-mimetic communication. On
both sides, the function of instruction was beginning to push language toward
autonomy.
At a certain point in time, then, everything already established must have
allowed for a new way to raise the general efficiency of the technology: some in-
novative speakers began to experiment with the concatenation of linguistic signs
into longer and longer strings. For proponents of combinatorial syntax, this is a
rather trivial development (Jackendoff 1999, Bickerton 2009), but I would like
to suggest that the emergence of concatenation was actually revolutionary—in
two complementary ways. To begin with, it presented listeners with a radically
new challenge: they were no longer required to bring up from their memories
clusters of experiences that were associated with mutually identified signs. They
were asked to imagine the experiences associated with the sounds, and then
calculate the intersection between them: concentrate on chasing-experiences,
and on rabbit-experiences, and then calculate the experience of rabbit-chasing.
This was revolutionary mainly because, to the extent that it worked, it allowed
for communication about the intersected cluster of experiences (the cluster of
rabbit-chasing) without the prior mutual-identification of the cluster itself.
Speakers could now communicate not just about the experiences they had mutu-
ally identified, but also about different combinations of these experiences. This
meant, among other things, that they could begin to invent imagined entities
and talk about them. The cultural consequences were enormous. All this must
have implied a great leap forward in the expressive power of the technology: the
function from the number of signs to the number of messages, which was up to
now a linear one, would turn into an exponential function. The dividends for
the mutual-identification of new signs grew much higher. As signs came to be
concatenated, again and again, only with certain signs but not with others, a
network of semantic connections began to emerge. Very gradually, the socially
constructed worldview of the symbolic landscape, which up to now included sets
of isolated experiences, began to turn into a categorized system.
The second revolutionary implication of concatenation was lineariza-
tion. The challenge of concatenation immediately allocated dividends to those
9780190256623-Dor.indb 206 03/06/15 6:18 PM

speakers who could produce longer strings, maintain the clarity and coherence
of their instructions, and do it faster. With the rise in speed, as signs came to
be pronounced closer and closer together, phonological relations at the utter-
ance level could begin to emerge, to allow for the swift move along the string
of sounds. The listeners, for their part, had to find ways to interpret the longer
strings, calculate the intersections between larger sets of experiences, and also do
it faster—to keep pace with the speakers. The emergence of concatenation, then,
began a developmental process that gradually forced the emergence of internal
complexity in the evolving technology—from the outside in. From the symbolic
landscape and the phonetic system that had already begun to evolve, semantic
and phonological structures began to emerge.
As the system grew in complexity, however, new types of problems began
to appear. Speakers were gradually producing longer, more complex utterances,
and these became more and more difficult to interpret. They were increasingly
ambiguous—the concatenated signs could be re-arranged in different ways to
produce different messages—and they would be increasingly more opaque: each
of the signs was still be mutually identified as such, but the intersections, more
and more complex, were not. This, together with other problems, must have
gradually begun to require a collective effort of a totally new type—that of the
mutual-identification of normative rules for the regulation of the actual process
of linguistic communication. This was the beginning of the protocol. Speakers,
in their constant attempts to understand and be understood, began to explore
different options: norms of linear order, for example, adjacency and iconicity.
Such innovations began to reduce the levels of misinterpretation, and sparked
a new dynamic of collective exploration and stabilization of more formalized
variations—more explicit standards for mutually identified behavior.
When they began to stabilize their protocols, speaking communities already
had all the components of the technology in their rightful place. From now on,
all the relevant evolutionary dynamics would spiral together. Every update in
each of the parameters required accommodations throughout the system: the
collective investigation of the world of experience; the ongoing expansion of the
symbolic landscape; the further construction of a social-semantic worldview;
the development of the dialectic relationship between this worldview and the
variable experiential worlds of the speakers; the growing formalization of the
sound system; the steady appearance of new communication problems; the con-
sequent rise in the complexity and generality of the mutually identified norms
invented to resolve them; the emergence of more and more complex utterances,
produced on the basis of more complex clusters of prescriptions; and then, on
the basis of all this, the discovery of more useful things to do with language, and
new expansions, constructions, formalizations, and regulations, new problems
9780190256623-Dor.indb 207 03/06/15 6:18 PM

and solutions—an endless process, inevitably leading human communities,

and human individuals, toward greater and greater dependency on their own
invention.
10.6 The Emergence of the Imaginative Species

The dating of the emergence of language is a highly contested issue in the dis-
course (see, for example, the fascinating debate between Watts 2014 and Dediu
and Levinson 2014). What all socially minded scholars seem to agree on is that
language began to emerge before Homo sapiens came onstage, in communities of
Homo erectus or Homo heidelbergensis. In terms of the narrative presented here,
this assumption makes perfect sense. These were the species that put us on track
as a co-operative, inventive, technological, mutually identifying animal. It stands
to reason that it was they who gradually found themselves confronted with the
challenge of instruction. Like the genius collective inventors they were, they also
found the solution.
In a very deep sense, however, the solution they found was probably already
out of their league. They were expert experiencers, probably much better than we
are, but language forced them to weaken their dependency on experience—and
develop a worldview based on imagination. They were expert experiential com-
municators, in all certainty much better than us, but language gradually forced
them to systematically suppress most of what they knew how to communicate.
At every given moment, it took the entire collective genius of their communities
to push the technology forward, but the individual speakers around language
only managed to use it to variable degrees. As they began to be selected for their
linguistic capacities—when language started to seriously change their selective
environment—individual speakers joined the evolutionary spiral, and began to
accommodate their cognitions, emotions, anatomies, physiologies, and genes to
the technology. From this process emerged a new species adapted to language:
Homo sapiens.
The new species adapted itself to language in the exact two ways predicted by
my theory. The first was the emergence of cognitions, anatomies, and physiolo-
gies specifically adapted to fast speech (Lieberman 1991, 2007). Tongues were
lowered into the pharynx, and brain circuits developed to control the increas-
ingly sophisticated processes involved in speech production. In terms of the hy-
pothetical narrative detailed above, this means that Homo sapiens either received
the challenge of fast concatenation from his ancestors, or invented it by him-
self. Either way, the specialized anatomies and physiologies evolved for the use
of a technology that was already there, strongly demanding higher capacities for
9780190256623-Dor.indb 208 03/06/15 6:18 PM

efficient usage. Lieberman dates the emergence of the full capacity for fast speech
between fifty to ninety thousand years ago.
It is around this time, maybe slightly earlier, that Mithen (2007) identifies the
first archaeological clues, from new social activities and new material artifacts,
for the second adaptation—full-fledged human imagination. Mithen begins his
story with the capacity of theory-of-mind that we share with the apes and defines
seven steps on the way from there to modern imagination. He mentions lan-
guage, of course, but almost in passing. From the point of view developed here,
however, language must have been the single most important determinant of the
emergence of human imagination as we know it.
Other animals probably have to use basic imagination for the planning of
action, and pre-linguistic humans developed the capacity further for their com-
plex activities. In all this, however, imagination is only activated on the spot, by
experiential problems that require the retrieval of past experiences. The imag-
ined experience combines real-time experience with materials from memory, and
the imagining animal has to calculate the relevant intersections between what it
experiences at the moment and what it remembers. Language, however, requires
something radically different: the construction of an imagined experience on
the sole basis of the creative assembly of pieces of experiential memory—in isola-
tion from real-time experience. It requires listeners to calculate the intersections
between sets of memories. For the first time, imagination is activated independ-
ently of experiencing.
Homo sapiens emerged as an answer to the two most pressing challenges of lan-
guage, but it was definitely not the only species that had to face them. Dediu and
Levinson (2014) analyze a wealth of evidence for the claim that our sister species,
Homo neandertalensis, had language too—a perfectly reasonable assumption if we
agree that language was invented by the ancestors of both species (which implies
that other descendants, like Homo denisova, may have also had language). Dediu
and Levinson also show that the Neandertals may have developed language fur-
ther: their sound-production anatomies seem to be more suited for the task, and
their cultures show clear signs of imagination. Like us, for example, they buried
their dead. Where they took their language, and what it looked like, we will prob-
ably never know. But if we assume that for a significant amount of time two (or
a few) human species spoke, and if we assume that they also maintained contact
in this way or the other, vestiges of the other species’ languages may actually be
still incorporated in our languages. As Dediu and Levinson suggest, the amaz-
ing variability of our languages may reflect the influences of the different species’
languages on each other: “just as for genetics, Neandertals and Denisovans (and
likely further archaic cousins) might be extinct as human lineages but continue
to live in us through their genes and perhaps speak through us as well” (p. 288).
9780190256623-Dor.indb 209 03/06/15 6:18 PM

10.7 The Darker Side of Imagination

The growing human capacity for creative imagination turned us into who we
are not just in positive terms. There was a darker side. At a certain point along
the way, some of the more intelligent speakers must have begun to realize that
the new technology could be used with a very different type of communicative
intent—the intent to deceive. This was a moment of enormous consequences: the
lie was born.
The intent to deceive as such was already there before language. Other ani-
mals deceive as well (Dawkins and Krebs 1978). Language, however, provided
deceivers with a tool so much more powerful than presentational communica-
tion that it changed deception forever. Three interrelated factors were involved.
First, experiential communication allows for the communication (honest or
deceptive) of a much narrower set of meanings than language—those mean-
ings that are anchored in the here and now of the communication event. With
language, the set of possible meaning-types explodes—for honest communica-
tion as well as for deception: everything that has ever been mutually identified
becomes a potential lie. Second, the nastiest characteristic of the lie is the fact
that it is functionally based on the very trust it betrays: you can only lie to those
with whom you share a language, and among those you can only lie to those who
trust you to tell the truth. The very logic of language and the very nature of the
process of socialization for language thus prepare the listeners for their unfortu-
nate role as the potential victims of deceptive communication. Third, and much
more important, is the fact that in presentational communication, communica-
tors can only present their interlocutors with something that is there for them
to experience. Communicators, for example, cannot threaten their interlocutors
unless they really are frightening. Whatever is communicated may be verified
or rejected by the others in real time, and because of that, presentational decep-
tion is a very difficult fit. (We still value it very much: the great actors that we
admire are the best presentational deceivers.) This is why, in terms of Zahavi’s
(1975) handicap principle, presentational signaling is heavy. Consequently, the
apes usually deceive by hiding something that is there—not by trying to present
something that is not. With language, however, the problem simply disappears.
It allows communicators to tell their interlocutors about things that they cannot
experience—and thus cannot verify or reject at the time of communication.
Language thus deprives the listeners of the single most important tool that they
could use to defend themselves against deception: the critical judgment of what
they just heard on the basis of what they experience with their own senses.
Taken together, the three factors actually carry a rather amazing implication:
the invention of language eventually did more to enhance the human capacity
9780190256623-Dor.indb 210 03/06/15 6:18 PM

for deception than it did to enhance the human capacity for honest communica-
tion. The functional envelope of presentational deception is narrower than that
of honest presentational communication, but the functional envelope of linguis-
tic deception is wider than that of honest linguistic communication: language
allows speakers to communicate mutually identified experiences external to the
here and now, but as long as they are honest, they may still only communicate, at
every given moment, those experiences they did experience: this is what honesty
is all about. Honest speaking is bound by the contingencies of the experiential
world of the speaker (both external and internal). In lying, however, the speaker
is for the first time truly released from the bounds of experience: everything that
can be said can be lied about. Language is deceivers’ heaven.
So much so, as a matter of fact, that it seems tempting to postulate that lan-
guage was originally invented for lying—that it was born as a tool of deception.
Everything said here so far indicates, however, that this could not possibly be
the case. The collective effort of the invention and stabilization of the new tech-
nology must have been based on high levels of reliability and trust between the
inventors: otherwise they would not have been able to get the system going. But
when language was stabilized, when certain levels of trust for language were
achieved, the door was opened—and some individuals rushed in. Because of
that, the entire history of the evolution of language, beyond the original inven-
tion, must have been closely tied up with the function of the lie.
Theoretical models of the evolution of language usually think about the lie
in terms of the more fundamental problem of the evolution of co-operative be-
havior. The argument runs as follows: The individuals involved in any collective
project should not just be willing to share the collective gains of the project—
they should also be committed to give their share of the effort. They should be
willing to pay the price. For the project to survive, of course, the gains should be
greater than the cost. The problem is that co-operative projects also invite free-
loaders to the table: if you manage to get your share of the gains without putting
in your share of the effort, you end up in an even better position than others.
This is a rational strategy, which means that it should in principle be adopted by
everybody. If it were, however, the entire project would collapse. Co-operation,
then, is a reasonable individual choice only to the extent that the others are also
willing to avoid freeloading. Everybody should agree to put some of their selfish
interests aside. To explain the emergence of co-operative systems in evolution, we
should find a way to theoretically control the phenomenon of freeloading. In the
case of language, freeloading is lying. Language is based on trust, but once the
trust is there, lying seems to be the most advantageous individual strategy. If eve-
rybody lied, however, the trust would collapse, dragging language down with it.
Different writers thus try to control lying in different ways: human societies and
9780190256623-Dor.indb 211 03/06/15 6:18 PM

individuals became more co-operative already before the emergence of language

(Tomasello, 2008, 2009); honest communication was ensured by conformist
learning and moralistic enforcement of norms (Richerson and Boyd, 2005); lan-
guage evolved on the basis of a rise in social trust and the emergence of the rule
of law (Knight, 1998, 2008); societies managed to win the war against individual
deception by the invention of the collective lie (Knight 1998); language evolved
as a kin-selected system, which ensured honest communication within the kin
group (Fitch, 2010); and more.
All these explanations are undoubtedly important, but they also seem to
betray an implicit universalist assumption: the option of freeloading is equally
open for all individuals. This, however, could not have been the case. At every
point in the evolution of language, individuals were different in their language-
related cognitive capacities, their emotional makeup, and their social status—
and each of these carried implications for the individual’s ability to either lie and
get away with it, or detect a lie and make sure that the liar was punished. Lying
requires more emotional control than telling the truth: liars have to prevent their
faces and bodies from betraying their intentions (Vrij, 2001). An individual’s
ability to lie and get away with it, as well as to punish a liar, is also dependent on
his or her social status: other things being equal, higher status guarantees more
immunity and more control. The consequences of getting caught lying are often
less intimidating than actually telling the truth (DePaulo et al. 1996). Most
importantly, lying is a more complex cognitive activity than honest speaking,
and lie detection is more complex than simple comprehension, both requiring
additional cognitive processing (Spence et al., 2004). This is also evidenced in
the very gradual development of children into full-fledged liars (Smith and La
Frenière 2013).
All this carries a simple implication: the drama of the lie should be read as a
variable story. Not everybody lied, not everybody lied efficiently, not everybody
lied to everyone else, and not everybody who lied got caught. More than any-
thing, the first liars must have been among the most imaginative speakers in their
communities. In honest linguistic communication, the speaker’s intent emerges
from his or her own experiences. The challenge is the translation of the intent
into the socially constructed terms of language. This challenge is also involved in
the lie, of course, but the major difficulty resides somewhere else: the speaker has
to artificially imagine an experiential intent in his or her mind which, from his
or her experiential point of view, is counter-factual. The speaker has to imagine a
world different from the one he or she actually experiences. All linguistic com-
munication requires imagining for understanding. The first liars found the ways
to imagine for speaking. They were probably good listeners as well: lying requires
a good understanding of the victim’s experiential world. And they found new
9780190256623-Dor.indb 212 03/06/15 6:18 PM

ways to control and suppress their emotions, and prevent their systems of pres-
entational communication from betraying their intentions: good liars have poker
faces. This is exactly what the apes simply cannot do. All this could not have been
easy. Patterns of variability among liars began to emerge: some were more imagi-
native than others, more controlled, more convincing, more cunning, quicker on
their feet. The better they were, the more they managed to freeload.
The victims of the liars, those lied to, must have made as variable a group as
the liars. Many of them may have never understood what was happening: their
skills were not good enough for the detection of lies. They were easy prey. The
liars lied and increased their share of the gains at the expense of their victims,
and the sense of trust on which language was founded remained intact. As long
as the lie was not exposed, the problem of instability never arose. Gradually, a
new relationship (a very special relationship) came to be formed between two
groups: the best liars and their most devoted believers. The division of labor was
clear: the liars described the world to their victims, turned their attention to cer-
tain experiences and away from others, invented collective lies, and constructed
the symbolic landscape to suit their goals. When those lied to began to look at
the world through the perspective spoken to them by the liars—precisely where
language took them to places they had no experience with—language turned
into the most effective tool of social coercion that ever was. It still is.
Not all those lied to, however, were easy prey. Some of them may have been
more experienced or more suspicious, better speakers and listeners, better readers
of presentational communication, or simply smarter. Many of them must have
been liars themselves—liars also lie to each other. They began to develop differ-
ent types of defense strategies—including those discussed in the literature—and
different individuals probably began to apply them to different degrees and in
different ways. One defense strategy was probably a retreat into the stronghold of
the safest, most intimate social bonds. The lie began to re-arrange societies along
new lines of suspicion. Secrecy was another strategy. Speaking the truth became
a moral issue.
At the same time, and as significantly, however, some individuals probably
began to developed new ways of lie detection. A more sensitive understanding
of speakers and the relationships between what they said and how they behaved
allowed for the more efficient detection of liars. Better memories helped listeners
keep track of what speakers were saying, for a longer time, and begin to compare.
New means gradually developed to critically judge the relationship between
the message and the world. Certain questions came up for the very first time:
is this reasonable? Does it make sense? Could it be? These contributed to the de-
velopment of language-based epistemology just as much as honest, co-operative
communication.
9780190256623-Dor.indb 213 03/06/15 6:18 PM

From a certain point on, then, a full-fledged arms race was launched between
the liars, with their unique capacities, and the lie detectors and decipherers, with
their own sets of skills. The liars were forced to work harder, sophisticate their
techniques, develop those linguistic behaviors that allowed them to convince:
this was the origin of rhetorics. Those lied-to were also forced to work harder
too, on all fronts: among other things, this was the origin of logical investigation.
Where the liars were strong enough, and especially where they learned to lie to-
gether, the levels of stability required for language were actually achieved by the
lie, in its collective form (Knight 1998). In other places, the levels of stability re-
quired for language were maintained and fractured, strengthened and betrayed,
again and again, in a constant battle. Freeloading was never controlled. The lie
has always been a key determining factor in the web of evolutionary relationships
between languages, their speakers, and their societies. A language that would
really evolve only for honest communication would probably be much simpler,
require much less from its speakers, and change society to a much less dramatic
degree.
10.8 Squeezing through the Bottleneck

This, then, is how my theory squeezes through the bottleneck of evolution.
Nothing is required beyond what we already seem to know about pre-linguistic
societies and their members, and what the theory of evo-devo tells us about the
dynamics of evolution—no additional stipulations at the social, cultural, behav-
ioral, communicative, cognitive, or genetic level.
The process has a very long pre-history, in which hominin communities grad-
ually re-invented themselves on the basis of the collective activity of experiential
mutual-identification. This is why the apes do not have language: they do not
mutually identify. The specific function of language was invented in explorations
into a new realm of communication, attempts to use the uniquely human tools
of experiential-mimetic communication for something completely new—when
the collective demands for information sharing began to exceed the collective
capacities of experiential-mimetic communication.
The system was then pushed forward by the constant need to raise the levels
of instructive success. The stabilization of the instructive strategy, and the fact
that it opened totally new horizons for human societies, dictated a constant flow
of innovative changes and developments, in the properties of the old tools them-
selves, in the communicative environment, and in the cognitive and emotional
lives of individuals. Some of the changes, most importantly the emergence of
concatenation, paved the way toward technological revolutions, which in their
9780190256623-Dor.indb 214 03/06/15 6:18 PM

turn dictated entire sets of new dynamics on all fronts. Technological problems
that appeared on the way required mutually identified solutions, and drove the
development of sets of normative rules for the regulation of instructive commu-
nication. Language emerged from the outside in, like a bridge constructed simul-
taneously from both ends of the experiential gap: it began with the first attempts
to connect experiential meaning and experiential-mimetic behavior for the func-
tion of instruction; gradually isolated the symbolic landscape from experiential
meaning and phonetics from experiential-mimetic communication; and then
gradually developed semantic and phonological complexity, morphology, and
linear syntax. The entire process was thus characterized by high levels of devel-
opmental determinism: if we agree to position the instruction of imagination at
the center of the story—with its unprecedented benefits—we find that much of
the way languages are today, much of the way we are today, was already there, as
potential, at the moment of origin.
Throughout the process, speakers were selected for their general ability to
work with the technology: our species, and maybe our sisters and cousins too,
emerged with unique adaptations to language. The adaptations, however, are not
foundational to language, and they are not universal. They emerged for a tech-
nology that was already there, and all along the way they were unevenly spread
across populations. The fact that almost all of us, modern humans, are capable
of acquiring and using our full-fledged languages does not imply that we are
less different from each other than our ancestors were. We have all climbed the
ladder of language together—generation after generation of variable capacities.
All along the way, certain capacities spread across entire populations, and some
of them were also partially genetically assimilated, and in this sense, variability
was indeed reduced. But at the very same time, languages kept evolving, new cog-
nitive challenges emerged—and new patterns of variability were exposed. The
best example of this is literacy (Jablonka and Rechav 1996): we have not adapted
ourselves genetically to the activities of reading and writing (there has not been
enough time and the selective pressure has not been there either), but literacy
nevertheless exposed a complex pattern of variability, some of which seems to be
partially genetically determined—from the quickest and most efficient readers
and writers, all the way to individuals with literacy-related “disorders” such as
dyslexia.
9780190256623-Dor.indb 215 03/06/15 6:18 PM

11 Conclusion: reassembling the puzzle
In this book, I have aimed to make the first steps toward a new general
conception of language as a social communication technology, col-
lectively constructed in a never-ending process of experiential mutual-
identification for the specific function of the instruction of imagination.
In preliminary and informal analyses of various pieces of the puzzle of
language, I demonstrated how this conception re-formulates old ques-
tions and raises new ones, re-arranges and interprets major empirical
findings, resolves theoretical difficulties, and replaces stipulations and
formal complexities with reasonable explanations. Whether the end
result is persuasive I do not know—I have lived with these ideas for so
long that I can no longer judge—but I hope to have shown at least that
looking at language through the lens of this theory might be an interest-
ing and fruitful intellectual pursuit.
For me, the most important point is this: if the theory is on the
right track, it shows the way toward the reassembly of the puzzle of
language, and thus, potentially, for the re-unification of the frag-
mentary linguistic sciences into a well-formed scientific discipline.
In principle, it should allow researchers from all different disciplines
interested in language—cognitive scientists and social scientists, for-
malists and functionalists, synchronically or diachronically oriented
researchers, neuroscientists and semioticians, all the way from formal
semantics to literature, history, and critical studies—to begin to see
that they are all working on the same phenomenon. In a very real
sense, I now realize, what I have been trying to do is exactly what lan-
guage does: help the disciplines construct a worldview between them
that would allow them to systematically mutually identify their scien-
tific experiences.
This position, between the disciplines, is a complicated place to
occupy. I found myself there because the scientific work I have done
through the years suffered from a certain lack of systematicity. I wrote
on the semantics of epistemic verbs, the syntax and semantics of ques-
tions (regular and concealed), the distribution of that-deletion, the
9780190256623-Dor.indb 216 03/06/15 6:18 PM

Conclusion • 217
relations between the problems of syntax and the relativity question, the prag-
matics of newspaper headlines, the emergence of a new sociolinguistic regime on
the Internet, the construction of political ideology in the language of the media,
and the cultural-biological evolution of language and its speakers. In each of
these excursions, I found a different way of speaking about language, different
theoretical apparatuses, different assumptions. Every field I visited convinced me
that some things must be right: syntactic complexity must be semantically based;
semantics must be linguistic; all interpretation must eventually be pragmatic;
language must be functionally specific; language must be a collective entity;
being a speaker must be a highly variable thing; the essence of being a speaker
must lie in the internal, dialectic relationship between private experience and the
socially constructed worldview of language; language must have emerged before
its speakers were ready for it; the essential story of language must be about the
instruction of imagination. It was around this last conviction that I eventually
found myself constructing a new synthetic picture.
Language resides in the social domain. It allows us to systematically chan-
nel complex sets of instructions for imagination from mind to mind, across the
experiential gaps that separate us, through the socially constructed worldview
of the mutually identified symbolic landscape. Our words function as discrete
instructors, superimposing sets of semantic categorizations on our private, var-
iable and analogue, experiential worlds. This is why words behave the way they
do: they allow for approximate definitions but very rarely for exhaustive ones,
show prototype effects, produce polysemies, determine relations of grammatical
selection, and much more. We do not all carry our words in the same way, and
we constantly struggle to develop the levels of mutual-identification required for
successful communication. As much as we try, we still spend much of our time
correcting the impressions our words have on others. This is why our languages
also prescribe complex sets of norms of linguistic communication, to make
sure that we combine our discrete instructors into complex messages in mutu-
ally identified ways, and then translate them into utterances that our listeners
would be able to interpret as correctly as possible. Interpretation, however, is still
always an approximation. Linguistic communication is a fragile and frustrating
business.
Much of the drama of language takes place between the two levels of mean-
ing that make language possible, the level of private-experiential meaning—
everything we manage to learn from our experiences, including those we share
with the others without speaking—and the level of social-semantic meaning.
Our words connect the two levels to each other, which is why their properties
require explanations involving both levels. Our protocols guide us through
9780190256623-Dor.indb 217 03/06/15 6:18 PM

the translation process, stage by stage, and the result is a hierarchical semantic
structure—which then determines the hierarchical properties of our utterances.
To be able to use the technology, we have to allow the level of social-semantic
meaning to teach us how to experience-for-instructing, which means that the
mutually identified norms of language force us to experience in certain ways
when we speak. The tension between the ways we experience-for-instructing and
the ways we experience in general drives a never-ending spiral of bi-directional
influence—as always affecting different speakers, and different children, in dif-
ferent ways. This is also the source of the symbolic power of language: it allows
those who maintain control over the construction of the collective imagination
to manipulate others into submission.
And language is formal, yes, but not in the mathematical sense. Like formal
dress, or a formal meeting, language is formal in the sense that it prescribes
sets of pre-established, mutually identified formalities that we have to obey in
the course of linguistic communication. Semantics is formal, and structural
formalities are tightly correlated with the formalities of semantics—not with
the non-formal, private experiential worlds of the speakers. This is why, when
we only compare structural formalities to private experiences, they seem to
be autonomous from meaning. All these formalities are emergent entities, the
products of long social processes of negotiation and struggle: they are not
based on pure logic. Consider formal semantics, for example. Among all the
innovators of language, through its entire evolution, there were also those
who were implicitly interested in logic. Their innovations reflected their at-
tempt to streamline the process of interpretation, to clean linguistic commu-
nication from ambiguity. Quantifiers, modalities, negation, scope—all these
were introduced into language by logically oriented inventors simply because
other speakers did not take them into account. They emerged from attempts
to understand: “Let me make it clear for myself: did they or didn’t they? All
of them or only some?” Such attempts to understand eventually produced
semantic formalities, which must still carry their emergent nature on their
sleeve. This is why we should expect the formal semantic systems of differ-
ent language to be significantly variable (Bach et al. 1995). A formal seman-
tics that ignores all this does not describe language; it prescribes a logician’s
utopia. What seems to be required is a change of attitude. Formal semantics
should investigate whether the relevant semantic formalities of language look
like pure logic; how and to what extent; how pieces of logic are conventional-
ized; how they are spread within communities; how they are actually used.
It should begin to take a closer look at communicative interactions between
logically oriented speakers and others. It should try to understand the urge to
avoid ambiguity.
9780190256623-Dor.indb 218 03/06/15 6:18 PM

Conclusion • 219
Language is formal, and thus conservative, but it also constantly changes.

Everything in and around it reverberates all the time: our experiences, the condi-
tions for their mutual-identification, the widths of the experiential gaps between
us, the levels of success that we manage to achieve in linguistic communication,
the levels of success that we need to achieve in order to survive the day, the levels
of innovation we bring into our conversations, the actual meanings of our words,
the prescriptions of our protocol, our conceptions of what a good conversation
should look like. In this crucial sense, the technology of language reflects the
foundational reality within which it emerged: it is plastic. What makes it so
special, however, is the duality—the fact that it is both so plastic and so con-
servative. It allows for the construction of thousands of languages, as different
from each other as human creativity permits, and still, for the speakers of each of
them, it feels like the only natural way to communicate.
Language resides in the social domain, which means that we, as individual
speakers, live around it as end-users. We download copies of the software from
the social net into our minds/brains. We manage to use the software with varia-
ble levels of skill, we are variably influenced by it, and we variably participate in
its further construction (to the extent that the software is open-code: social power
always works to keep it as close as possible). Once we put things this way, a new
theoretical relationship between linguistics and psycholinguistics may begin to
emerge: linguistics investigates the social software; psycholinguistics investigates
the download, maintenance, and usage of the software in the personal computers
around it. Language requires much of our minds/brains, and we have gradually
evolved, for hundreds of thousands of years, to adapt ourselves to the demand:
our anatomies and physiologies are variably adapted to fast speech, our brains
are adapted to the computations of language, our capacity for imagination deter-
mines much of what we are. None of this, however, means that we have language
coded in our genes. It only means that we are born ready for the participation
in the collective activity of language: if it does turn out that children spread the
burden of acquisition between them, in a collective effort revolving around their
own experiences, the level of innate readiness for language may be reduced even
further. What we are definitely born with is the craving for language: the desper-
ate need to constantly construct bridges across the experiential gaps between us.
In this sense, as in many others, language is a double-edged sword. When we
speak the same language, it brings our minds closer together, but we are in fact
more different from each other than our common language makes us seem—
especially when watched from the outside. When we speak different variations
of a language, or different languages, they immediately separate us in ways that
might or might not be related to the actual gaps between our experiential worlds.
Language works best when it efficiently bridges relatively narrow experiential
9780190256623-Dor.indb 219 03/06/15 6:18 PM

gaps, in closely tied communities of shared experience—which is exactly where

it is least needed. It is most needed when the gaps are wide, but there it very
often fails to deliver. Language allows us to imagine—the freest of all cognitive
processes—but it only allows us to do that on the basis of social consensus. It is
based on trust, but this only makes it the most dangerous tool of deception ever
invented.
All this, I believe, makes for a very energetic picture of language, and for me, at
the end of the day, this is a major issue.1 The claim is often heard that the level of
the social does not really exist—in the material world there are only individuals
and institutions, nothing beyond that. This means, in our case, that positioning
language in the social domain is a non-starter. This, among many other things,
has kept socially oriented researchers on the outskirts of theoretical linguistics
throughout the last fifty years. Language, however, along with everything social,
is perfectly material: it simply appears on the surface as energy. Unlike individu-
als and institutions, it is always everywhere, happening in real time, covering the
entire social domain in a web of endless conversation. To understand language,
we need to break away from static models and idealizations—and focus on the
energy. Maybe today, as we are gradually adapting ourselves to yet another form
of collective life on the Internet, this will no longer be so difficult to do.
9780190256623-Dor.indb 220 03/06/15 6:18 PM

Notes
chapter 1
1. I have also made minor contributions to this effort in Dor (1993, 1996, 1999, 2000,
2005a).
chapter 2
1. See Ginsburg and Jablonka (2007a,b) for an account of the evolutionary emergence
of the process of experiencing in lower organisms.
2. Barsalou (1999) terms the analogue categorizations that emerge in the process of
learning “perceptual symbols.” I accept much of his analysis of these entities, but
the term is problematic. The categorizations are indeed perceptual, but they are not
symbolic.
chapter 3
1. In the last two decades, computer scientists have made much progress in simulat-
ing very simplified versions of this process. Most important among these efforts
are Luc Steels’s talking heads experiments (Steels 2012, Steels and Hild 2012, and
references therein): a population of virtual agents, embodied in robots with cam-
eras, audio input-output and a processor, are exposed to a shared environment—a
set of different shapes on a whiteboard. The agents identify shapes on the board,
produce names to mark them, and engage in a guessing game in which they gradu-
ally zoom in on a common lexicon. Significantly, the experiments clearly show that
the process allows for the spontaneous emergence of lexically based grammatical
structure.
2. This explains Clark’s (1987) principle of contrast, i.e., the hypothesis that language-
acquiring children work with the default assumption that “every two forms contrast
9780190256623-Dor.indb 221 03/06/15 6:18 PM

222 • Notes
in meaning.” Clark’s hypothesis has been shown to capture a very wide range of lan-
guage acquisition data (Clark 2009).
3. There is an interesting relationship between this view of the sign and Frege’s (1892
[1952]) conception. Frege is usually remembered for his distinction between the
sense and the reference of the sign, but the distinction can only be properly under-
stood against the background of his conception of the idea. It is here that Frege
establishes the status of the sign as a social entity, and the experiential gap as its
foundational background: “the reference and sense of a sign are to be distinguished
from the associated idea. If the reference of a sign is an object perceivable by the
senses, my idea of it is an internal image, arising from memories of sense impressions
which I have had and acts, both internal and external, which I have performed.
Such an idea is often saturated with feeling; the clarity of its separate parts varies
and oscillates. The same sense is not always connected, even in the same man, with
the same idea. The idea is subjective: one man’s idea is not that of another. . . . A
painter, a horseman, and a zoologist will probably connect different ideas with the
name ‘Bucephalus.’” It is at the level of sense that different people have the same
meaning associated with the sign. As Evans (1982) shows, the sign thus conceived
has sense exactly because it is part of a public language. Crucially, then, Frege de-
clares the idea immaterial for the understanding of the sign, because he is ultimately
a prescriptivist: his entire project attempts to construct a logically based, ambiguity-
free language that would allow for correct thinking. “From the laws of [logic],” he
writes, “there follow prescriptions about asserting, thinking, judging, inferring”
(Frege 1918 [1956], p. 325, quoted in Steinberger 2013). My conception, on the
other hand, does not attempt to prescribe: it positions the idea at the center of the
theory. I thank Nick Enfield for discussing this point with me.
4. For a discussion of the fascinating Searle-Derrida debate, see Kenaan (2002).
5. Intents are also prior to language in two more essential ways that we shall come
back to later: they are there, in the minds of children, before they acquire language;
and they were there, in the minds of our ancestors, before language was invented. In
both cases, of course, the intents are different.
6. The relationship between this view of the production process and Levelt’s (1989,
1993, 1999) psycholinguistic model will be explicated in chapter 6.
chapter 4
1. This, for example, is the paraphrase for be happy as it appears in Goddard (2002);
other sources (cf. Wierzbicka 1992) provide slightly different ones:
X was happy =
X felt something because X thought something
Sometimes a person thinks something like this:
Some very good things happened to me
I wanted things like this to happen
9780190256623-Dor.indb 222 03/06/15 6:18 PM

Notes • 223
I don’t want other things now

When this person thinks this, this person feels something good
X felt something like this
Because X thought something like this
2. Note that this is not a problem for Jackendoff, who includes necessary features in
his model.
3. A very similar account is given by Traugott and Dasher’s (2002) theory of invited
inferencing in processes of semantic change: meanings are first invited by implica-
ture, then are gradually semantically coded. Traugott and Dasher’s theory also sug-
gests that such process of change proceed along unidirectional, semantic paths.
4. Consider, for example, two different exchanges between A and B, guests at a birth-
day party for ten-year old Maggie, who was adopted at the age of three:
A: (pointing at a woman) Is she the mother?
B: Well, not exactly. She’s been raising her since she was three.
A: Wow, tough . . .
or
A: (pointing at a woman) Is she the mother?
B: Yes.
A: I’ll go say hi then.
B’s answers in the two exchanges reflect different judgments on the prototypicality
of the mother, each of which relies on a different representation of prototypical
motherhood. In the first, what counts is the birth-giving and the genetics, which
makes the woman pointed at a non-prototypical mother; in the second, the crucial
property is the home-giving and the nurturing, and it positions the woman close
enough to the center of the category for B to be able to refer to her as a prototypical
mother.
chapter 5
1. Note that sometimes, for example inside a house, the terms do not project easily
onto the immediate environment. Nevertheless, “informants display a keen sense
of absolute orientation and direction. For example, from inside a house their ges-
tures to distant locations appear to be very precise, distinguishing perhaps 10 to 15
degrees of arc” (Brown and Levinson 1993, p. 51).
2. Recent research seems to suggest that the interaction of language with experiencing
while speaking takes place mainly in the left hemisphere. For a review, see Gilbert
et al. (2007).
chapter 7
1. Here, maybe even more than in the other chapters, it should be remembered that
the semi-technical discussions, which concentrate on examples from English, carry
9780190256623-Dor.indb 223 03/06/15 6:18 PM

224 • Notes
no universal implications whatsoever: many languages have very different parts of

speech systems; many have no subjects. What is totally grammaticalized in English
may be much more flexible in other languages, and the other way around. The only
thing that is really shared by all language in this domain is that grammatical pat-
terns (strong and weak) are determined by clusters of (strong and weak) normative
conventions.
chapter 8
1. See Levinson’s (2013) interesting claim that in interactive dialogues, “adjacency

pairs like question-answer, request-compliance, offer-acceptance, and so forth”
sometimes allow for remarkable center-embeddings—“apparently with little effort
and to a much greater depth than is exhibited in syntax” (157).
chapter 9
1. As Evans (2011a) shows in his discussion of verbal art in different cultural settings,
the impact of such exceptionally talented speakers and authors “does not stop at
the work itself, but flows on to the rest of the language system” (p. 194). They play
a more dominant role in the processes of experiential mutual-identification for
language, and thus also have more influence on the experiential worldviews of the
others.
2. Like every other generalization about language, this one also has counter-examples:
Warlpiri, for example, has a well-developed baby-talk register (Laughren 1984).
This, of course, does not harm the argument. Overall, the default cultural pattern
around the world seems to be active interaction between adults and children with-
out explicit teaching.
chapter 11
1. This is not exactly the sense of energy used by Humboldt in his famous saying that
“in itself, [language] is no product (Ergon), but an activity (Energeia)” (1999, p. 49),
but his fingerprints are obviously all over this book. He was Yehuda Elkana’s favor-
ite linguist.
9780190256623-Dor.indb 224 03/06/15 6:18 PM

References
Addis, D. R., A. T. Wong, and D. L. Schacter (2008). Age-Related Changes in the

Episodic Simulation of Future Events. Psychological Science, 19, 1, 33–41.
Ahlsén, E. (2006). Introduction to Neurolinguistics. Amsterdam: John Benjamins.
Aitchison, J. (1996). The Seeds of Speech: Language Origin and Evolution. Cambridge:
Cambridge University Press.
Aitchison, J. (2001). Language Change: Progress or Decay? Cambridge: Cambridge
University Press.
Anderson, B. (1983). Imagined Communities: Reflections on the Origin and Spread of
Nationalism. London: Verso.
Anderson, J. M. (1997). A Notional Theory of Syntactic Categories. Cambridge: Cam-
bridge University Press.
Anderson, J. M. (2006). The Non-Autonomy of Syntax. Folia Linguistica, 3–4,
223–250.
Arbib, M. (2003). The Evolving Mirror System: A Neural Basis for Language Readi-
ness. In: Christiansen, M., and S. Kirby (eds.), Language Evolution, 182–200.
Oxford: Oxford University Press.
Ariés, P. (1962). Centuries of Childhood: A Social History of Family Life. New York:
Vintage.
Armstrong, S., L. Gleitman, and H. Gleitman (1983). What Some Concepts Might
Not Be. Cognition, 13, 263–308.
Arnon, I., M. Casillas, C. Kurumada, and B. Estigarribia (eds.) (2014). Language in
Interaction: Studies in Honor of Eve. V. Clark. Amsterdam: John Benjamins.
Aronoff, M., I. Meir, C. A. Padden, and W. Sandler (2008). The Roots of Linguistic
Organization in a New Language. Interaction Studies: A Special Issue on Holophra-
sis vs. Compositionality in the Emergence of Protolanguage, 9, 1, 131–150.
Arthur, W. B. (2007). The Structure of Invention. Policy Research, 36, 274–287
Austin, J. L. (1962). How to Do Things with Words. Oxford: Oxford University Press.
9780190256623-Dor.indb 225 03/06/15 6:18 PM

226 • References
Bach, E., E. Jelinek, A. Kratzer, and B. Partee (eds.) (1995) Quantification in Natural
Languages. Dordrecht: Kluwer.
Baker, M. C. (2003). Lexical Categories: Verbs, Nouns and Adjectives. Cambridge:
Barsalou, L. W. (1999). Perceptual Symbol Systems. Behavioral and Brain Sciences, 22,
577–660.
Bates, E., I. Bretherton, and L. Snyder (1988). From First Words to Grammar: Indi-
vidual Differences and Dissociable Mechanisms. Cambridge: Cambridge Univer-
sity Press.
Bates, E., P. S. Dale, and D. Thal (1995). Individual Differences and Their Implica-
tions for Theories of Language Development. In: Fletcher, P., and B. MacWhin-
ney (eds.), Handbook of Child Language, 96–151. Blackwell Publishing.
Beattie, G. W., and B. L. Butterworth (1979). Contextual Probability and Word Fre-
quency as Determinants of Pauses and Errors in Spontaneous Speech. Language
and Speech, 22, 201–211.
Berlin, B., and P. Kay (1969). Basic Color Terms: Their Universality and Evolution.
Stanford: CSLI.
Berman, R. (ed.), (2004). Language Development Across Childhood and Adolescence.
Amsterdam: John Benjamins.
Berman, R. A., and D. Slobin (1987). Five Ways of Learning How to Talk about Events:
A Crosslinguistic Study of Narrative Development. Working Paper No. 46, Center
for Cognitive Studies, Oakland: University of California, Berkeley.
Bicchieri, C. (2006). The Grammar of Society. Cambridge: Cambridge University Press.
Bickerton, D. (1990). Language and Species. Chicago: University of Chicago Press.
Bickerton, D. (2008). Ancestors of Meaning. Language and Communication, 28,
282–290.
Bickerton, D. (2009) Adam’s Tongue: How Humans Made Language, How Language
Made Humans. New York: Hill and Wang.
Bierwisch, M. (1983) Formal and Lexical Semantics. Linguistische Studien 114, 56–79.
Bierwisch, M., and E. Lang (eds.) (1989) Dimensional Adjectives: Grammatical Struc-
ture and Conceptual Interpretation. Dordrecht: Springer.
Bloom, L. (1970). Language Development: Form and Function in Emerging Gram-
mars. Cambridge: Cambridge University Press.
Bloom, P., and F. C. Keil (2001). Thinking Through Language. Mind & Language,
16, 351–367.
Bock, J. K. (1986). Syntactic Persistence in Language Production. Cognitive Psychol-
ogy, 18, 355–387.
Boesch, C., and M. Tomasello (1998). Chimpanzee and Human Cultures. Current
Anthropology 39, 5, 591–614.
Boroditsky, L. (2003). Linguistic Relativity. In: Nadel, L. (ed.) Encyclopedia of Cogni-
tive Science. London: MacMillan Press.
9780190256623-Dor.indb 226 03/06/15 6:18 PM

References • 227
Botha, R. (1992). Twentieth Century Conceptions of Language: Mastering the Meta-

physics Market. Oxford: Blackwell Publishing.
Botha, R. (2003). Unraveling the Evolution of Language. Oxford: Elsevier.
Botha, R., and C. Knight (2009). The Cradle of Language. Oxford: Oxford University
Press.
Bourdieu, P. (2000). Pascalian Meditations. Cambridge: Polity Press.
Bowerman, M., and S. Choi (2003). Space under Construction: Language-Specific
Spatial Categorization in First Language Acquisition. In: Gentner, D., and
S. Goldin-Meadow (eds.), Language in Mind: Advances in the Study of Language
and Thought, 387–427. Cambridge: MIT Press.
Bresnan, J. (2001). Lexical Functional Syntax. Oxford: Blackwell.
Brown, P. (2014). The Interactional Context of Language Learning in Tzeltal.
In: Arnon, I., M. Casillas, C., Kurumada, and B. Estigarribia (eds.), Language
in Interaction: Studies in Honor of Eve. V. Clark, 13–28 Amsterdam: John
Benjamins.
Brown, P., and Levinson S. C. (1993). “Uphill” and “Downhill” in Tzeltal. Journal of
Linguistic Anthropology, 3, l, 46–74.
Brown, R., and D. McNeill (1966). The “Tip of the Tongue” Phenomenon. Journal of
Verbal Learning and Verbal Behavior, 5, 325–337.
Brugman, C. (1988) The Story of Over: Polysemy, Semantics and the Structure of the
Lexicon. New York: Garland.
Bruner, J. S. (1983). Child’s Talk: Learning to Use Language. New York: W. W. Norton.
Bu Bois, J. W. (1985). Competing Motivations. In: Haiman, J. (ed.), Iconicity in Syntax,
Typological Studies in Language, 6, 343–365, Amsterdam: John Benjamins.
Butterworth, B., and G. W. Beattie (1978). Gesture and Silence as Indicators of Plan-
ning in Speech. In: Campbell, R. N., and P. T. Smith (eds.), Recent Advances in
the Psychology of Language: Volume 4, Formal and Experimental Approaches, 347–
360. New York: Plenum Press.
Bybee, J. L. (2006). Frequency of Use and the Organization of Language. Oxford:
Oxford University Press.
Call, J., and M. Tomasello (1994). Production and Comprehension of Referential
Pointing by Orangutans (Pongo pygmaeus). Journal of Comparative Psychology,
108, 307–317.
Caramazza, A. (1997). How Many Levels of Processing Are There in Lexical Access?
Cognitive Neuropsychology, 14, 177–208.
Cassirer, E. (1923, 1955). The Philosophy of Symbolic Forms. I. Language. New Haven:
Yale University Press.
Cassirer, E. (1944). An Essay on Man: An Introduction to a Philosophy of Human Cul-
ture. New Haven: Yale University Press.
Cheney, D. L., and R. M. Seyfarth (1985). Vervet Monkey Alarm Calls: Manipulation
Through Shared Information? Behaviour, 93, 150–166.
9780190256623-Dor.indb 227 03/06/15 6:18 PM

228 • References
Choi, S., and M. Bowerman (1991). Learning to Express Motion Events in English and
Korean: The Influence of Language-Specific Lexicalization Patterns. Cognition,
41, 83–121.
Chomsky, N. (1955, 1975). The Logical Structure of Linguistic Theory. Dordrecht:
Springer.
Chomsky, N. (1957). Syntactic Structures. De Hague: Mouton.
Chomsky, N. (1965). Aspects of the Theory of syntax. Cambridge: MIT Press.
Chomsky, N. (1968). Language and Mind. Cambridge: Cambridge University Press.
Chomsky, N. (1973). Conditions on Transformations. In: S. Anderson and P. Kip-
arsky (eds.), A Festschrift for Morris Halle, 232–286. New York: Holt, Rinehart
& Winston.
Chomsky, N. (1981). Lectures on Government and Binding. Dordrecht: Foris
Publications.
Chomsky, N. (1995). The Minimalist Program. Cambridge: MIT Press.
Chomsky, N. (1998). On Language. New York: The New Press.
Chomsky, N., and H. Lasnik (1993). The Theory of Principles and Parameters. In:
von Stechow, J., A. Jacobs, W. Sternefeld, and T. Vennemann (eds.), Syntax: an
International Handbook of Contemporary Research, 506–569. Berlin: De Gruyter.
Christiansen M. H., and S. Kirby (eds.) (2003a). Language Evolution. Oxford: Oxford
University Press.
Christiansen, M. H., and S. Kirby (2003b). Language Evolution: The Hardest Prob-
lem in Science? In: Christiansen, M. H., and S. Kirby (eds.), Language Evolution,
1–15. Oxford: Oxford University Press.
Clark, E. (1987). The Principle of Contrast. In: MacWhinney, B. (ed.), Mechanisms of
Language Acquisition, 1–33. Hillsdale: Laurence Erlbaum.
Clark, E. (2009). First Language Acquisition. Cambridge: Cambridge University Press.
Clark, H. H. (1996). Using Language. Cambridge: Cambridge University Press.
Clark, H. H. (1998). Communal Lexicons. In: Malmkjaer, K., and J. Williams (eds.),
Context in Language Learning and Language Understanding, 63–87, Cambridge
University Press.
Clark, H. H., and S. A. Brennan (1991). Grounding in Communication. In: Resnick,
L. B., J. M. Levine, and S. D. Teasley (eds.), Perspectives on Socially Shared Cogni-
tion, 127–149. Washington: APA Books.
Clark, H. H., and J. E. Fox Tree (2002). Using uh and um in Spontaneous Speech.
Cognition, 84, 73–111.
Clarke, J. (2004). Histories of Childhood. In: Wyse, D. (ed.), Childhood Studies: An
Introduction, 3–12. New York: Wiley-Blackwell.
Comrie, B. (1989). Language Universals and Linguistic Typology. Chicago: Chicago
University Press.
Connolly, A. C., J. A. Fodor, L. R. Gleitman, and H. Gleitman (2007). Why Stereo-
types Don’t Even Make Good Defaults. Cognition, 103, 1–22.
9780190256623-Dor.indb 228 03/06/15 6:18 PM

References • 229
Corballis, M. (2003). From Hand to Mouth: The Gestural Origins of Language. In:
Christiansen, M. H., and S. Kirby (eds.), Language Evolution, 403–428. Oxford:
Coulmas, F. (1998). The Handbook of Sociolinguistics. New York: John Wiley.
Couper-Kuhlen, E., and M. Selting (2001). Studies in Interactional Linguistics. Am-
sterdam: John Benjamins.
Croft, W. (1991). Syntactic Categories and Grammatical Relations: The Cognitive Or-
ganization of Information. Chicago: University of Chicago Press.
Croft, W. (2000). Explaining Language Change: An Evolutionary Approach. New
York: Longman, Linguistic Library.
Croft, W. (2001). Radical Construction Grammar. Oxford: Oxford University Press.
Croft, W. (2002). Typology and Universals. Cambridge: Cambridge University Press.
Cruse, D. A. (1986). Lexical Semantics. Cambridge: Cambridge University Press.
Crystal, D. (2002). Language Death. Cambridge: Cambridge University Press.
Culicover, P., and R. Jackendoff (2005). Simpler Syntax. Oxford: Oxford University
Press.
Cuyckens, H., and B. Zawada (eds.) (2001). Polysemy in Cognitive Linguistics. Amster-
dam: John Benjamins.
Davidson, D. (1984). Inquiries into Truth and Interpretation. Oxford: Oxford Univer-
sity Press. Dawkins, R., and J. R. Krebs (1978). Animal Signals: Information or
Manipulation? In: Krebs, J. R., and N. B. Davies (eds.), Behavioural Ecology: An
Evolutionary Approach, 282–309. Oxford: Blackwell.
Deacon, T. W. (1997). The Symbolic Species: The Co-Evolution of Language and the
Brain. New York: Norton Publications.
Deane, P. (1991). Limits to Attention: a Cognitive Theory of Island Phenomena. Cog-
nitive Linguistics 2, 1–63.
Dediu, D., and S. C. Levinson (2014). Social Origins and the Time Frame of Lan-
guage Evolution. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of
Language. 184–195. Oxford: Oxford University Press.
Dell, G. S. (1986). A Spreading-Activation Theory of Retrieval in Sentence Produc-
tion. Psychological Review, 93, 283–321.
DePaulo, B. M., D. A. Kashy, S. E. Kirkendol, M. M. Wyer, and J. A. Epstein, J. A.
(1996). Lying in Everyday Life. The Journal of Personality and Social Psychology,
70, 979–995.
Derrida, J. (1977 [1988]). Limited Inc. Evanston, Il: Northwestern University Press.
Deutscher, G. (2005). The Unfolding of Language: An Evolutionary Tour of Mankind’s
Greatest Invention. New York: Metropolitan Books.
Dewey, J. (1938). Experience and Education. Indianapolis, IN: Kappa Delta Pi.
Donald, M. (1991). Origins of the Modern Mind: Three Stages in the Evolution of Cul-
ture and Cognition. Cambridge: Harvard University Press.
9780190256623-Dor.indb 229 03/06/15 6:18 PM

230 • References
Dor, D. (1993). Towards a Semantic Account of Concealed Questions. In: M. Bern-

stein (ed.), Proceedings of Eastern States Conference On Linguistics, 92, 56–67.
Dor, D. (1996). An Epistemically-Based Analysis of Lexical Selection. Ph.D. disserta-
tion, Stanford University.
Dor, Daniel (1999). From Symbolic Forms to Lexical Semantics: Where Modern Lin-
guistics and Cassirer’s Philosophy Start to Converge. Science in Context, 12, 4,
493–511.
Dor, D. (2000). From the Autonomy of Syntax to the Autonomy of Linguistic Seman-
tics: Notes on the Correspondence between the Transparency Problem and the
Relationship Problem. Pragmatics & Cognition, 8 (2), 325–356.
Dor, D. (2004). Intifada Hits the Headlines: How the Israeli Press Misreported the Out-
burst of the Second Palestinian Uprising. Bloomington: Indiana University Press.
Dor, D. (2005a). Towards a Semantic Account of That-Deletion in English. Linguis-
tics, 42, 3, 345–382.
Dor, D. (2005b). The Suppression of Guilt: The Israeli Media and the Reoccupation of the
West Bank. London: Pluto Press.
Dor, D. (2014). The Instruction of Imagination: Language, and Its Evolution, as
a Communication Technology. In: Dor, D., C. Knight, and J. Lewis (eds.), The
Social Origins of Language, 105–125. Oxford: Oxford University Press.
Dor, D., and E. Jablonka (2000). From Cultural Selection to Genetic Selection: A
Framework for the Evolution of Language. Selection, 1, 33–56.
Dor, D., and E. Jablonka (2001). How Language Changed the Genes. In: Tabant,
J., and S. Ward (eds.). New Essays on the Origin of Language, 149–175. Berlin:
Mouton de Gruyer.
Dor, D., and E. Jablonka (2010). Plasticity and Canalization in the Evolution of Lin-
guistic Communication: an Evolutionary Developmental Approach. In: Larson,
R. K., V. Deprez, and H. Yamakido (eds.) The Evolution of Human Language: Bio-
linguistic Perspectives, 135–147. Cambridge: Cambridge University Press.
Dor, D., and E. Jablonka (2014). Why We Need to Move from Gene-Culture Co-
Evolution to Culturally-Driven Co-Evolution. In: Dor, D., C. Knight, and J. Lewis
(eds.), The Social Origins of Language, 15–30. Oxford: Oxford University Press.
Dor, D., C. Knight, and J. Lewis (eds.) (2014). The Social Origins of Language. Oxford:
Dorian, N. (ed.), (1989). Investigating Obsolescence: Studies in Language Contraction
and Death. Cambridge: Cambridge University Press.
Dowty, D. R. (1979). Word Meaning and Montague Grammar. Dordrecht: Kluwer.
Dummett, M. (1996). The Seas of Language. Oxford: Oxford University Press.
Dunbar, R. (1998). Grooming, Gossip, and the Evolution of Language. Cambridge:
Harvard University Press.
Duranti, A. (1997). Linguistic Anthropology. Cambridge: Cambridge University Press.
Dörnyei, Z. (2005). The Psychology of the Language Learner: Individual Differences in
Second Language Acquisition. Hillsdale: Lawrence Erlbaum.
9780190256623-Dor.indb 230 03/06/15 6:18 PM

References • 231
Dörnyei, Z. (2006). Individual Differences in Second Language Acquisition. AILA

Review, 19, 42–68.
Eagleton, T. (1991). Ideology: An Introduction. London: Verso.
Eckert, P. (2000). Three Waves of Variation Study: The Emergence of Meaning in the
Study of Variation. www.stanford.edu/~eckert/PDF/ThreeWavesofVariation.pdf.
Eckert, P. (2005). Variation, Convention, and Social Meaning. Paper presented at the
Annual Meeting of the Linguistic Society of America.
Ehrman, M. E., Leaver B. L., and R. L. Oxford (2003). A Brief Overview of Individual
Differences in Second Language Learning. System 31, 313–330.
Elman, J., A. Karmiloff-Smith, E. Bates, M. Johnson, D. Parisi, and K. Plunkett
(1996). Rethinking Innateness: A Connectionist Perspective on Development. Cam-
bridge: MIT Press.
Enfield, N. J., M. Dingemanse, J. Baranova, J. Blythe, P. Brown, T. Dirksmeyer,
P. Drew, S. Floyd, S. Gipper, R. S. Gisladottir, G. Hoymann, K. Kendrick,
S. C. Levinson, L. Magyari, E. Manrique, G. Rossi, L. San Roque, and F. Torreira
(2013). Huh? What?—A First Survey in 21 Languages. In: Hayashi, M., G. Ray-
mond, and J. Sidnell (eds.), Conversational Repair and Human Understanding,
343–380. Cambridge: Cambridge University Press.
Enfield, N. J., and J. Sidnell (2014). Language Presupposes An Enchronic Infrastruc-
ture for Social Interaction. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social
Origins of Language, 92–104. Oxford: Oxford University Press.
Enfield, N. J., and T. Stivers (2007). Person Reference in Interaction: Linguistic, Cul-
tural, and Social Perspectives. Cambridge: Cambridge University Press.
Enfield, N. J., and A. Wierzbicka (2002). Introduction: The Body in Description of
Emotion. Pragmatics and Cognition, 10, 1, 1–24.
Erteschik-Shir, N. (2007). Information Structure: The Syntax-Discourse Interface:
Syntax and Morphology. Oxford: Oxford University Press.
Erteschik-Shir, N., and S. Lappin (1979). Dominance and the Functional Explanation
of Island Constraints. Theoretical Linguistics, 6, 43–84.
Ervin-Tripp, S. (2014). Conversational Input to Bilingual Children. In: Arnon, I.,
M. Casillas, C. Kurumada, and B. Estigarribia, Language in Interaction: Studies
in Honor of Eve. V. Clark, 13–28. Amsterdam: John Benjamins.
Evans, G. (1982). The Varieties of Reference. Oxford: Oxford University Press.
Evans, N. (1995). A Grammar of Kayardild. Berlin: Mouton de Gruyter.
Evans, N. (2011a). Anything Can Happen: The Verb Lexicon and Interdisciplinary
Fieldwork. In: N. Thieberger (ed.),The Oxford Handbook of Linguistic Fieldwork,
Evans, N. (2011b). Dying Words: Endangered Languages and What They Have to Tell
Us. Malden: Wiley-Blackwell.
Evans, N. (2013). Language Diversity as a Resource for Understanding Cultural Evo-
lution. In: Richerson, P. J., and M. H. Christiansen (eds.), Cultural Evolution: So-
ciety, Technology, Language, and Religion,233–268. Cambridge: MIT Press.
9780190256623-Dor.indb 231 03/06/15 6:18 PM

232 • References
Evans, N., and S. C. Levinson (2009). The Myth of Language Universals: Language
Diversity and Its Importance for Cognitive Science. Behavioral and Brain Sci-
ences, 32, 429–492.
Evans, V. (2009). How Words Mean. Oxford: Oxford University Press.
Evans, V., and M. Green (2006). Cognitive Linguistics: An Introduction. Edinburgh:
Edinburgh University Press.
Everett, C. (2013). Linguistic Relativity: Evidence Across Languages and Cognitive Do-
mains. Berlin: Mouton de Gruyter.
Everett, D. (2012). Language: The Cultural Tool. London: Profile Books.
Falkum, I. L. (2011). The Semantics and Pragmatics of Polysemy: A Relevance Theoretic
Account. Ph.D. Dissertation, University College London.
Fauconnier, G. (1994). Mental Spaces: Aspects of Meaning Construction in Natural
Language. Cambridge: Cambridge University Press.
Fillmore, C. J. (1968). The Case For Case. In: Bach, E., and R. T. Harms (eds.), Univer-
sals in Linguistic Theory, 1–88. New York: Holt, Rinehart, and Winston.
Fillmore, C. J. (1970). The Grammar of Hitting and Breaking. In: Jacobs, R. A., and
P. S. Rosenbaum (eds.), Readings in English Transformational Grammar, 120–
133, Waltham, MA: Ginn & Co.
Fillmore, C. J. (1982). Frame Semantics. In: Linguistic Society of Korea (ed.), Linguis-
tics in the Morning Calm, 111–137. Seoul: Hanshin Publishing.
Fillmore, C. J., and C. Baker (2009). A Frames Approach to Semantic Analysis. In:
Heine, B., and H. Narrog (eds.), The Oxford Handbook of Linguistic Analysis, 313–
340. Oxford: Oxford University Press.
Fitch, W. T. (2010). The Evolution of Language. Cambridge: Cambridge University
Press.
Fodor, J. A. (1975). The Language of Thought. Cambridge: Harvard University Press.
Fodor, J. A. (1981). The Present Status of the Innateness Controversy. In: Representa-
tions: Philosophical Essays on the Foundations of Cognitive Science, 257–316, Cam-
bridge: MIT Press.
Fodor, J. A., M. F. Garrett, E. C. T. Walker, and C. H. Parkes (1980). Against Defini-
tions. Cognition, 8, 63–367.
Fodor, J. A., and E. Lepore (1996). The Red Herring and the Pet Fish: Why Concepts
Still Can’t Be Prototypes. Cognition, 58, 253–270.
Frawley, W. (1992). Linguistic Semantics. Hillsdale: Lawrence Erlbaum.
Frege, G. (1892 [1952]). On Sense and Reference. In: Geach, P., and M. Black (eds.),
Translations from the Philosophical Writings of Gottlob Frege, 56–78, Blackwell.
Frege, G. (1918 [1956]). The Thought. Mind, 65, 289–311.
Fromkin, V. (ed.), (1973). Speech Errors as Linguistic Evidence. Berlin: Mouton.
Gardiner, A. (1932). The Theory of Speech and Language. Oxford: Clarendon Press.
Garrett, M. F. (1975). Syntactic Process in Sentence Production. In: Bower, G. (ed.),
Psychology of Learning and Motivation: Advances in Research and Theory, 9, 133–
177. New York: Academic Press.
9780190256623-Dor.indb 232 03/06/15 6:18 PM

References • 233
Garrett, M. F. (1988). Processes in Language Production. In: Newmeyer, F. (ed.), Lin-

guistics: The Cambridge Survey III. Language: Psychological and Biological Aspects,
Garrett, M. F. (1992). Disorders of lexical selection. Cognition, 42, 143–180.
Geeraerts, D. (1990). The Lexicographical Treatment of Prototypical Polysemy. In:
S. L. Tsohatzidis (ed.), Meanings and Prototypes: Studies in Linguistic Categoriza-
tion, 195–210, Routledge Publishing.
Geeraerts, D. (1993). Vagueness’s Puzzles, Polysemy’s Vagaries. Cognitive Linguistics,
4, 223–272.
Geeraerts, D. (1997). Diachronic Prototype Semantics: A Contribution to Historical
Lexicology. Clarendon Press.
Geeraerts, D. (2010). Theories of Lexical Semantics. Oxford: Oxford University Press.
Geertz, C. (1973). The Interpretation of Cultures: Selected Essays. Basic Books.
Gilbert, M. (1989). On Social Facts. New York: Routledge.
Gilbert, A. L., T. Regier, P. Kay, and R. B. Ivry (2007). Support for Lateralization of
the Whorf Effect Beyond the Realm of Color Discrimination. Brain and Lan-
guage, 105, 2, 91–98.
Ginsburg, S., and E. Jablonka (2007a). The Transition to Experiencing: I. The Evolu-
tion of Limited Learning. Biological Theory, 2, 3, 218–230.
Ginsburg, S., and E. Jablonka (2007b). The Transition to Experiencing: II. The
Evolution of Associative Learning Based on Feelings. Biological Theory, 2, 3,
231–243.
Ginsburg, S., and E. Jablonka (2014). Memory, Imagination and the Evolution of
Modern Language. In: Dor, D., C. Knight, and J. Lewis (eds.) (2014). The Social
Givón, T. (1985). Iconicity, Isomorphism and Non-Arbitrary Coding in Syntax. In:
Haiman, J. (ed.), Iconicity in Syntax, 187–217. Amsterdam: John Benjamins.
Givón, T. (1995) Functionalism and Grammar. Amsterdam: John Benjamins.
Glăveanu, V. P. (2011). How Are We Creative Together? Comparing Sociocognitive
and Sociocultural Answers. Theory & Psychology, 21, 4, 473–492.
Gleitman, L. R., and A. Papafragou (2005). Language and Thought. In: Morrison, R.,
and K. Holyoak (eds.), Cambridge Handbook of Thinking and Reasoning, 633–
662. Cambridge: Cambridge University Press.
Goddard, C. (1998). Bad Arguments Against Semantic Primitives.Theoretical Lin-
guistics, 24, 129–156.
Goddard, C. (2002). The Search for the Shared Semantic Core of All Languages.
In: Goddard, C., and A. Wierzbicka (eds.), Meaning and Universal Grammar—
Theory and Empirical Findings, I, 5–40. Amsterdam: John Benjamins.
Goddard, C., and A. Wierzbicka (eds.) (1994). Semantic and Lexical Universals—
Theory and Empirical Findings. Amsterdam: John Benjamins.
Goldberg, A. (1995). Constructions: A Construction Grammar Approach to Argument
Structure. Chicago: Chicago University Press.
9780190256623-Dor.indb 233 03/06/15 6:18 PM

234 • References
Goldberg, A. (2006). Constructions at Work: The Nature of Generalization in Lan-

guage. Oxford University Press.
Goldberg, A. (2009). Essentialism Gives Way to Motivation (commentary on Evans
and Levinson). Behavioral and Brain Sciences, 32, 455–456.
Goldberg, A., and R. Jackendoff (2004). The English Resultatives as a Family of Con-
structions. Language 80, 3, 532–568.
Grice, H. P. (1975). Logic and Conversation. Cambridge: Harvard University Press.
Greenfield, P. M. (1991). Language, Tools and Brain: The Ontogeny and Phylogeny
of Hierarchically Organized Sequential Behavior. Behavioral and Brain Sciences,
14, 531–595.
Gumperz, J. J., and S. C. Levinson (eds.) (1996). Rethinking Linguistic Relativity.
Harley, T. A. (1984). A Critique of Top-Down Independent Levels Models of Speech
Production: Evidence from Non-Plan-Internal Speech Errors. Cognitive Science,
8, 191–219.
Harley, T. A. (2008). The Psychology of Language: From Data to Theory. Psychology
Press.
Harris, J. R. (1998). The Nurture Assumption: Why Children Turn Out the Way They
Do. The Free Press.
Harris, J. R. (2006). No Two Alike: Human Nature and Human Individuality. W. W.
Norton.
Hauser, M. D., N. Chomsky, and W. T. Fitch (2002). The Faculty of Language: What
Is It, Who Has It, and How Does It Evolve? Science, 298, 1569–1579.
Hauser, M. D. (1996). The Evolution of Communication. Cambridge: MIT Press.
Hayashi, M, G. Raymond and J. Sidnell (eds.) (2013). Conversational Repair and
Human Understanding. Cambridge: Cambridge University Press.
Heidegger, M. (1927, 1962). Being and Time. London: SCM Press.
Hempel, C. (1952). Fundamentals of Concept Formation in Empirical Science. Chi-
cago: Chicago University Press.
Hengeveld, K. (1992). Non-verbal Predication: Theory, Typology, Diachrony. Berlin:
Mouton de Gruyter.
Herder, J. G. (1772, 2002). Treatise on the Origin of Language. In: Forster, M. N.
(ed.), Herder: Philosophical Writings, 65–165. Cambridge University Press.
Hockett, C. (1960). The Origin of Speech. Scientific American, 203, 88–96.
Hofmeister, P., and I. Sag (2010). Cognitive Constraints and Island Effects. Language,
86, 2, 366–415.
Hopper, P., and E. Traugott (2003). Grammaticalization. Cambridge University Press.
Hrdy, S. B. (2009). Mothers and Others: The Evolutionary Origins of Mutual Under-
standing. Cambridge: Harvard University Press.
Humboldt, W. von (1836 [1999]). On Language: On the Diversity of Human Language
Construction and Its Influence on the Mental Development of the Human Species,
M. Losonsky (ed.), Cambridge University Press.
9780190256623-Dor.indb 234 03/06/15 6:18 PM

References • 235
Hurford, J. (2007). The Origins of Meaning: Language in the Light of Evolution.

Hurford, J. (2011). The Origins of Grammar: Language in the Light of Evolution.
Indefrey, P., and W. J. M. Levelt (2004). The Neural Correlates of Language Produc-
tion. In: Gazzaniga, M. (ed.), The New Cognitive Neurosciences, 845–865 Cam-
bridge: MIT Press.
Itkonen, E. (2003). What is Language? A Study in the Philosophy of Linguistics. Publi-
cations in General Linguistics 6. University of Turku Press.
Itkonen, E. (2008). The Central Role of Normativity for Language and Linguistics.
In: Zlatev, J., T. Racine, E. Sinha, and E. Itkonen (eds.), The Shared Mind: Perspec-
tives on Intersubjectivity, 279–305. Amsterdam: John Benjamins.
Jablonka, E., S. Ginsburg, and D. Dor (2012). The Co-Evolution of Language and
Emotions.Philosophical Transactions of the Royal Society B: Biological Sciences, 367,
1599, 2152–2159.
Jablonka, E., and M. Lamb (2005). Evolution in Four Dimensions. Cambridge: MIT
Press.
Jablonka, E., and G. Rechav (1996). The Evolution of Language in Light of the Evolu-
tion of Literacy. In: Trabant, J. and S. Ward (eds.), Origins of Language, 70–88,
Collegium Budapest, Workshop Series 2.
Jackendoff, R. (1972). Semantic Interpretation in Generative Grammar. Cambridge:
MIT Press.
Jackendoff, R. (1983). Semantics and Cognition. Cambridge: MIT Press.
Jackendoff, R. (1990). Semantic Structures. Cambridge: MIT Press.
Jackendoff, R. (1999). Possible Stages in the Evolution of the Language Capacity.
Trends in Cognitive Science, 3, 272–279.
Jackendoff, R. (2007). Language, Consciousness, Culture: Essays on Mental Structure.
Cambridge: MIT Press.
Jackendoff, R., and S. Pinker (2005). The Faculty of Language: What’s Special about
It? Cognition, 95, 201–236.
Jakobson, R. (1960). Linguistics and Poetics. In: Seboek, T. (ed.), Style in Language,
James, W. (1890). The Principles of Psychology. New York: Dover Publications.
James, W. (1904, 1995). Pragmatism: A New Name for Some Old Ways of Thinking.
New York: Dover Publications.
Jitsumori, M., M. Ohkita, and T. Ushitani (2011). The Learning of Basic-Level Cate-
gories by Pigeons: The Prototype Effect, Attention, and Effects of Categorization.
Learning Behavior, 39, 271–287.
Kant, E. (1768, 1999). Critique of Pure Reason. Cambridge: Cambridge University Press.
Katz, J. J. (1972). Semantic Theory. New York: Harper and Row.
Katz, J. J., and J. Fodor (1963). The Structure of a Semantic Theory. Language, 39,
170–210.
9780190256623-Dor.indb 235 03/06/15 6:18 PM

236 • References
Katz, J. J., and M. Postal (1964). An Integrated Theory of Linguistic Description. Cam-
bridge: MIT Press.
Kegl, J. (1994). The Nicaraguan Sign Language Project: An Overview. Signpost, 7, 1,
24–31.
Kegl J., A. Senghas, and M. Coppola (1999). Creation Through Contact: Sign Lan-
guage Emergence and Sign Language Change in Nicaragua. In: M. deGraff (ed.),
Comparative Grammatical Change: The Intersection of Language Acquisition,
Creole Genesis, and Diachronic Syntax, 179–237 Cambridge: MIT Press.
Kenaan, H. (2002). Language, Philosophy and the Risk of Failure: Rereading the
Debate Between Searle and Derrida. Continental Philosophical Review, 35,
117–133.
Kendon, A. (2014). The “Poly-Modalic” Nature of Utterances and its Relevance for In-
quiring into Language Origins. In: Dor, D., C. Knight, and J. Lewis (eds.) (2014).
The Social Origins of Language, 67–76. Oxford: Oxford University Press.
Kiparsky, P. (1997). Remarks on Denominal Verbs. In: Alsina, A., J. Bresnan, and
P. Sells (eds.), Argument Structure, 473–99. Stanford: CSLI.
Kisch, S. (2008). “Deaf Discourse”: The Social Construction of Deafness in a Bedouin
Community. Medical Anthropology, 27, 3, 283–313.
Klepousniotou, E. (2002). The Processing of Lexical Ambiguity: Homonymy and
Polysemy in the Mental Lexicon. Brain and Language 81, 205–223.
Knight, C. (1998). Ritual/Speech Coevolution: A Solution to the Problem of Decep-
tion. In: Hurford, J. R., S. Kennedy, and C. Knight (eds.), Approaches to the Evolu-
tion of Language, 68–91. Cambridge: Cambridge University Press.
Knight, C. (2014). Language and Symbolic Culture: An Outcome of Hunter-
Gatherer Reverse Dominance. In: Dor, D., C. Knight, and J. Lewis (eds.) (2014).
The Social Origins of Language, 228–246. Oxford: Oxford University Press.
Knight, C., M. Studdert-Kennedy, and J. Hurford (eds.) (2000). The Evolutionary
Emergence of Language: Social Function and the Origins of Linguistic Form. Cam-
bridge: Cambridge University Press.
Kobayashi, H., and S. Kohshima (2001). Unique Morphology of the Human Eye and
Its Adaptive Meaning: Comparative Studies on External Morphology of the Pri-
mate Eye. Journal of Human Evolution, 40, 5, 419–435.
Koerner, E. F. K. (1982). Ferdinand de Saussure. Tokyo: Taishukan Publishing
Company.
Köyman, B., E. Lieven, D. A. Engemann, H. Rakoczy, F. Warneken, and M. Toma-
sello (2014). Children’s Norm Enforcement in Their Interactions With Peers.
Child Development, 85, 3, 1108–1122.
Labov, W. (1966). The Social Stratification of English in New York City. Washington,
D.C.: Center for Applied Linguistics.
Labov, W. (1973). The Boundaries of Words and their Meanings. In: Bailey, C. J., and
R. Shuy (eds.), New Ways of Analyzing Variation in English, 340–373, Washing-
ton DC: Georgetown University Press.
9780190256623-Dor.indb 236 03/06/15 6:18 PM

References • 237
Lakoff, G. (1987) Women, Fire and Dangerous Things: What Categories Reveal about
the Mind. Chicago: Chicago University Press.
Lakoff, G., and M. Johnson (1980). Metaphors We Live By. Chicago: Chicago Univer-
sity Press.
Lamm, E. (2014). Forever United: The Co-Evolution of Language and Normativity.
In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of Language, 267–
283. Oxford: Oxford University Press.
Langacker, R. W. (1987). Foundations of Cognitive Grammar, vol. 1, Theoretical Pre-
requisites. Stanford: Stanford University Press.
Langacker, R. W. (1999). Grammar and Conceptualization. Berlin: de Gruyter.
Larson, R. K., V. Deprez, and H. Yamakido (eds.) (2010). The Evolution of Human
Language: Biolinguistic Perspectives. Cambridge: Cambridge University Press.
Laughren, M. (1984). Warlpiri Baby talk. Australian Journal of Linguistics, 4, 1,
73–88.
Leavitt, J. (2006). Linguistic Relativities: Language Diversity and Modern Thought.
Cambridge: Cambridge University Press.
Lehmann, C. (2002). Thoughts on Grammaticalization. Arbeitspapiere des Seminars
für Sprachwissenschaft der Universität Erfurt, No. 9. Erfurt.
Levelt, W. J. M. (1989). Speaking: From Intention to Articulation. Cambridge: MIT
Press.
Levelt, W. J. M. (1993). Lexical Access in Speech Production. New York: Blackwell.
Levelt, W. J. M. (1999). Models of Word Production.Trends in Cognitive Sciences, 3,
223–232.
Levelt, W. J. M. (2013).A History of Psycholinguistics: The Pre-Chomskyan Era. Oxford:
Levin, B. (1993). English Verb Classes and Alternations: A Preliminary Investigation.
Chicago: Chicago University Press.
Levin, B., and M. Rappaport-Hovav (1995). Unaccusativity. Cambridge: MIT Press.
Levinson, S. C. (1996). Relativity in Spatial Conception and Description. In:
Gumperz, J. J., and S. C. Levinson (eds.), Rethinking Linguistic Relativity, 177–
202. Cambridge: Cambridge University Press.
Levinson, S. C. (1997). From Outer to Inner Space: Linguistic Categories and Non-
linguistic Thinking. In Nuyts, J., and E. Pederson (eds.), Language and Conceptu-
alization, 13–45. Cambridge: Cambridge University Press.
Levinson, S. C. (2000). Presumptive Meanings: The Theory of Generalized Conversa-
tional Implicature. Cambridge: MIT Press.
Levinson, S. C. (2012). The Original Sin of Cognitive Science.Topics in Cognitive Sci-
ence, 4, 396–403.
Levinson, S. C. (2013). Short Report: Recursion in Pragmatics. Language, 89, 1,
149–162.
Levinson, S. C., and N. Evans (2010). Time for a Sea-Change in Linguistics: Response
to Comments on “The Myth of Language Universals.” Lingua, 120, 2733–2758.
9780190256623-Dor.indb 237 03/06/15 6:18 PM

238 • References
Levinson, S. C., and P. Jaisson (2006). Evolution and Culture. Cambridge: MIT Press.
Levinson, S. C., S. Kita, D. B. M. Haun, and B. H. Rasch (2002). Returning the
Tables: Language Affects Spatial Reasoning. Cognition, 84, 2, 155–188.
Lewis, D. K. (1969). Convention: A Philosophical Study. Cambridge: Harvard Univer-
sity Press.
Lewis, J. (2014). BaYaka Pygmy Multi-Modal and Mimetic Communication Tradi-
tions. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of Language,
Li, P., and L. Gleitman (2002). Turning the Tables: Spatial Language and Spatial Cog-
nition. Cognition, 83, 265–294.
Li, C. N., and S. A. Thompson (1976). Subject and Topic: A New Typology of Lan-
guage. In: Li, C. N. (ed.), Subject and Topic, 457–461. New York: Academic Press.
Lieberman, P. (1991). Uniquely Human: The Evolution of Speech, Thought and Selfless
Behavior. Cambridge: Harvard University Press.
Lieberman, P. (2007). The Evolution of Human Speech: Its Anatomical and Neural
Bases. Current Anthropology, 48, 1, 39–66.
Liszkowski, U., M. Carpenter, T. Striano, and M. Tomasello (2006). 12- and
18-Month-Olds Point to Provide Information for Others. Journal of Cognition
and Development, 7, 173–187.
Lucy, J. (1992a). Grammatical Categories and Cognition: A Case Study of the Linguistic
Relativity Hypothesis. Cambridge: Cambridge University Press.
Lucy, J. (1992b). Language Diversity and Thought: A Reformulation of the Linguistic
Relativity Hypothesis. Cambridge: Cambridge University Press.
Lucy, J., and S. Gaskins (2001). Grammatical Categories and the Development of Clas-
sification Preferences: A Comparative Approach. In: S. Levinson and M. Bow-
erman (eds.), Language Acquisition and Conceptual Development, 257–283.
Lucy, J., and S. Gaskins (2003). Interaction of Language Type and Referent Type in
the Development of Nonverbal Classification Preferences. In: Gentner, D. and
S. Goldin-Meadow (eds.) Language in Mind: Advances in the Study of Language
and Thought, 465–492. Cambridge: MIT Press.
Lyons, J. (1977). Semantics. Cambridge: Cambridge University Press.
Lyotard, J. (1983 [1988]). The Differend: Phrases in Dispute. Minneapolis: University
of Minnesota Press.
Margolis, E., and S. Laurence (2003). Concepts. In: Stephen P. Stich (ed.), The Black-
well Guide to the Philosophy of Mind Concepts: Core Readings, 190–213. New
York: Blackwell.
McCloskey, M. E., and S. Glucksberg (1978). Natural Categories: Well Defined or
Fuzzy Sets? Memory & Cognition, 6, 4, 462–472.
McGrew, W. C. (1992). Chimpanzee Material Culture: Implications for Human Evo-
lution. Cambridge: Cambridge University Press.
9780190256623-Dor.indb 238 03/06/15 6:18 PM

References • 239
Mehler, J., E. W. Jusczyk, G. Lambertz, N. Halsted, J. Bertoncini, and C. Amiel-Tison

(1988). A Precursor of Language Acquisition in Young Infants. Cognition, 29,
143–178.
Meillet, A. (1921). Linguistique Historique et Linguistique Generale. Paris: H.
Champion.
Meir, I., M. Aronoff, W. Sandler, and C. Padden (2010). Sign Languages and Com-
pounding. In: Scalise, S., and I. Vogel (eds.), Compounding, 301–322. Amsterdam:
John Benjamins.
Meltzoff, A. N., and R. Brooks (2007). Eyes Wide Shut: The Importance of Eyes in
Infant Gaze-Following and Understanding Other Minds. In: Flom, R., K. Lee,
and D. Muir (eds.), Gaze Following: Its Development and Significance, 217–241,
Hillsdale: Lawrence Erlbaum.
Merleau-Ponty, M. (1945 [2005]). Phenomenology of Perception. London: Routledge.
Mervis, C. R., and E. Rosch (1981). Categorization of Natural Objects. Annual Review
of Psychology, 32, 89–115.
Mithen, S. (2007). Seven Steps in the Evolution of Human Imagination. Proceedings
of the British Academy, 147, 3–29.
Montague, R. (1973). The Proper Treatment of Quantification in Ordinary English.
In: Hintikka, J., J. Moravcsik, and P. Suppes (eds.), Approaches to Natural Lan-
guage, 221–242. Dordrecht.
Mufwene, S. (2013). Language as Technology: Some Questions that Evolutionary
Linguists Should Address. In: T. Lohndal (ed.), In Search of Universal Grammar:
From Old Norse to Zoque, 327–358. Amsterdam: John Benjamins.
Nettle, D., and S. Romaine (2000). Vanishing Voices: The Extinction of the World’s
Languages. Oxford: Oxford University Press.
Newmeyer, F. (1998). Language Form and Language Function. Cambridge: MIT Press.
Newmeyer, F. (2005). Possible and Probable Languages: A Generative Perspective on
Linguistic Typology. Oxford: Oxford University Press.
Oakeshott, M. (1933). Experience and Its Modes. Cambridge: Cambridge University
Press.
Ochs, E., E. Schegloff, and S. Thompson. (1996). Interaction and Grammar. Cam-
Osherson, D., and E. Smith (1981). On the Adequacy of Prototype Theory as a Theory
of Concepts. Cognition, 9, 35–58.
Östman, J. O., and M. Fried (eds.) (2005). Construction Grammars: Cognitive Ground-
ing and Theoretical Extensions. Amsterdam: John Benjamins.
Papafragou, A., J. Hulbert, and J. Trueswell (2008). Does Language Guide Event Per-
ception? Evidence from Eye Movements. Cognition, 108, 1, 155–184.
Pepperberg, I. M. (1992). Proficient Performance of a Conjunctive, Recursive Task by
an African Grey Parrot (Psittacus erithacus). Journal of Comparative Psychology,
106, 295–305.
9780190256623-Dor.indb 239 03/06/15 6:18 PM

240 • References
Peters, J. D. (2000). Speaking into the Air: A History of the Idea of Communication.
Chicago: Chicago University Press.
Piattelli-Palmarini, M. (1989). Evolution, Selection and Cognition: From “Learning”
to Parameter Setting in Biology and in the Study of Language. Cognition, 31, 1–44.
Pickering, M. J., and S. Garrod (2013). An Integrated Theory of Language Production
and Comprehension. Behavioral and Brain Sciences, 36, 04, 329–347.
Pinchevski, A. (2005). By Way of Interruption: Levinas and the Ethics of Communi-
cation. Pittsburgh: Duquesne University Press.
Pinker, S. (1994). The Language Instinct: How the Mind Creates Language. New York:
HarperPerennial.
Pinker, S. (2003). Language as an Adaptation to the Cognitive Niche. In: Christian-
sen, M. and S. Kirby (eds.), Language Evolution: States of the Art, 16–37. New
York: Oxford University Press.
Pinker, S., and P. Bloom (1990). Natural Language and Natural Selection. Behavioral
and Brain Sciences, 13, 707–726.
Pollard, C., and I. A. Sag (1994). Head-Driven Phrase Structure Grammar. Chicago:
Chicago University Press.
Pollock, L. H. (1983). Forgotten Children: Parent-Child Relations from 1500 to 1900.
Postma, A. (2000). Detection of Errors During Speech Production: A Review of
Speech Monitoring Models. Cognition, 77, 97.
Power, C. (2014). Signal Evolution and the Social Brain. In: Dor, D., C. Knight, and
J. Lewis (eds.), The Social Origins of Language, 47–55, Oxford: Oxford Univer-
sity Press.
Pustejovsky, J. (1995). The Generative Lexicon. Cambridge: MIT Press.
Quine, W. V. O. (1960). Word and Object. Cambridge: MIT Press.
Ravin, Y., and C. Leacock (eds.) (2000). Polysemy: Theoretical and Computational Ap-
proaches. Oxford: Oxford University Press.
Reddy, M. J. (1979). The Conduit Metaphor – a Case of Frame Conflict in Our Lan-
guage About Language. In: Ortony, A. (ed.), Metaphor and Thought, 284–297.
Richerson, P. J., and R. Boyd (2005). Not by Genes Alone: How Culture Transformed
Human Evolution. Chicago: Chicago University Press.
Riemer, N. (2006). Reductive Paraphrase and Meaning: a Critique of Wierzbickian
Semantics. Linguistics and Philosophy, 29, 347–379.
Robbins, P., and M. Aydede (eds.) (2009). The Cambridge Handbook of Situated Cog-
nition. Cambridge: Cambridge University Press.
Rosch, E. (1975). Cognitive Representations of Semantic Categories. Journal of Ex-
perimental Psychology, 104, 3, 192–233.
Rosch, E. (1978). Principles of Categorization. In: Rosch, E. and B. B. Lloyd (eds.),
Cognition and Categorization, 27–48. Hillsdale: Lawrence Erlbaum.
Ross, J. (1967). Constraints on Variables in Syntax. Ph.D. Dissertation, MIT.
9780190256623-Dor.indb 240 03/06/15 6:18 PM

References • 241
Rupert, R. D. (2009). Innateness and the Situated Mind. In: Robbins, P., and M.
Aydede (eds.), The Cambridge Handbook of Situated Cognition, 96–116. Cam-
Sacks, J. S. (1967). Recognition Memory for Syntactic and Semantic Aspects of Con-
nected Discourse. Perception and Psychophysics, 2, 437–442.
Sacks, H. (1995). Lectures on Conversation. New York: Blackwell Publishing.
Sakel, J., and E. Stapert (2010). Pirahã—in Need of Recursive Syntax? In: Hulst,
H.v. d. (ed.), Recursion and Human Language, 3–16 Berlin: de Gruyter.
Sandler, W., and D. Lillo-Martin (2006). Sign Language and Linguistic Universals.
Sandler, W., I. Meir, C. Padden, and M. Aronoff (2005). The Emergence of Grammar
in a New Sign Language. Proceedings of the National Academy of Sciences, 102, 7,
2661–2665.
Sapir, E. (1921). Language: An Introduction to the Study of Speech. New York: Har-
court, Brace.
Sausssure, F. de (1916 [1983]). Course in General Linguistics. La Salle, Il: Open Court.
Savage-Rumbough, S., and R. Lewin (1994). Kanzi: The Ape at the Brink of the Human
Mind. New York: John Wiley.
Schriefers, H., A. S. Meyer, and W. J. M. Levelt (1990). Exploring the Time Course of
Lexical Access in Language Production: Picture-Word Interference Studies. Jour-
nal of Memory and Language, 29, 1, 86–102.
Scruton, R. (2009). Imagination. In: Davies, S., K. M. Higgins, R. Hopkins,
R. Stecker, and D. E. Cooper (eds.), A Companion to Aesthetics, 346–350. New
York: Wiley-Blackwell.
Searle, J. (1969). Speech Acts: An Essay in the Philosophy of Language. Cambridge Uni-
versity Press.
Searle, J., and D. Vanderveken (1985). Foundations of Illocutionary Logic. Cambridge:
Senghas, A., and M. Coppola (2001). Children Creating Language: How Nicaraguan
Sign Language Acquired a Spatial Grammar. Psychological Science, 12, 4, 323–328.
Senghas, A., S. Kita, and A. Ozyurek (2004). Children Creating Core Properties
of Language: Evidence from an Emerging Sign Language in Nicaragua. Science,
305, 17.
Senghas, R. J., A. Senghas, and J. E. Pyers (2005). The Emergence of Nicaraguan Sign
Language: Questions of Development, Acquisition and Evolution. In: Langer, J.,
S. T. Parker, and C. Milbrath (eds.), Biology and Knowledge Revisited: From Neu-
rogenesis to Psychogenesis, 287–306. Mahwah, NJ: Lawrence Erlbaum.
Shegloff, E. A. (2007). Sequence Organization in Interaction, Volume 1: A Primer in
Conversation Analysis. Cambridge: Cambridge University Press.
Shen, Y., and D. Gil (to appear). Language, Thought and the Animacy Hierarchy: Ex-
perimental Studies of Hybrids in Hebrew, Indonesian and Minangkabau. Pacific
Linguistics.
9780190256623-Dor.indb 241 03/06/15 6:18 PM

242 • References
Sinha, C. (2014). Niche Construction and Semiosis: Biocultural and Social Dynam-
ics. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of Language.
Skutnabb-Kangas, T. (2000). Linguistic Genocide in Education, or Worldwide Diver-
sity and Human Rights? Hillsdale: Lawrence Erlbaum.
Slobin, D. (1996). From “Thought and Language” to “Thinking for Speaking.” In:
Gumperz, J. J., and S. C. Levinson (eds.), Rethinking Linguistic Relativity, 70–96.
Slobin, D. (1997). The Crosslinguistic Study of Language Acquisition: Expanding the
Contexts. New York: Psychology Press.
Slobin, D. (2003). Language and Thought Online: Cognitive Consequences of Lin-
guistic Relativity. In: Gentner, D., and S. Goldin-Meadow (eds.), Language in
Mind: Advances in the Study of Language and Thought, 157–192 Cambridge: MIT
Press.
Slobin, D. (2008). Breaking the Molds: Signed Languages and the Nature of Human
Language. Sign Language Studies, 8, 2, 114–130.
Smith, J. D., J. S. Redford, and S. M. Haas (2008). Prototype Abstraction by Monkeys
(Macaca mulatta). Journal of Experimental Psychology, 137, 2, 390–401.
Smith, R. M., and P. La Freniere (2013). Development of Tactical Deception from 4 to
8 Years of Age. British Journal of Developmental Psychology, 31, 30–41.
Spence, S. A., M. D. Hunter, T. F. Farrow, R. D. Green, D. H. Leung, and C. J. Hughes
(2004). A Cognitive Neurobiological Account of Deception: Evidence from
Functional Neuroimaging. Philosophical Transactions of the Royal Society, Bio-
logical Science, 359, 1755–1762.
Sperber, D., and D. Wilson (1986). Relevance: Communication and Cognition. Oxford:
Blackwell.
Steels, L. (2001). Grounding Symbols through Evolutionary Language Games. In:
Cangelosi, A., and D. Parisi (eds.), Simulating the Evolution of Language, 1–22.
Dordrecht: Springer.
Steels, L. (ed.), (2012). Experiments in Cultural Language Evolution. Advances in In-
teraction Studies 3. Amsterdam: John Benjamins.
Steels, L. (2014). Breaking Down Barriers to Understanding. In: Dor, D., C. Knight,
and J. Lewis (eds.), The Social Origins of Language, 336–349, Oxford: Oxford
University Press.
Steels, L., and M. Hild (eds.) (2012). Language Grounding in Robots. Dordrecht:
Springer.
Steinberger, F. (2013). How Tolerant Can You Be? Carnap on the Normativity of Logic.
Unpublished ms.
Stephens, G., and D. Matthews (2014). Referential Pacts in Child Language Develop-
ment. In: Arnon, I., M. Casillas, C. Kurumada, and B. Estigarribia (eds.), Lan-
guage in Interaction: Studies in Honor of Eve. V. Clark, 175–190. Amsterdam:
John Benjamins.
9780190256623-Dor.indb 242 03/06/15 6:18 PM

References • 243
Sterelny, K. (2012). The Evolved Apprentice: How Evolution Made Humans Unique.
Cambridge: MIT Press.
Stevenson, L. (2003). Twelve Conceptions of Imagination. British Journal of Aesthet-
ics, 43, 3, 238–259.
Suddendorf, T. (2013). The Gap: The Science of What Separates Us from Other Animals.
New York: Basic Books.
Szabolcsi, A. (2006). Strong vs. Weak Islands. In: Everaert, M., and H. van Riems-
dijk (eds.), The Blackwell Companion to Syntax, Volume 4, 479–531. New York:
Blackwell.
Taylor, J. R. (1995). Linguistic Categorization: Prototypes in Linguistic Theory. Oxford:
Taylor, J. R. (2003). Polysemy’s Paradoxes. Language Sciences, 25, 637–655.
Tomasello, M. (1999). The Cultural Origins of Human Cognition. Cambridge: Har-
vard University Press.
Tomasello, M. (2003). Constructing a Language: A Usage-Based Theory of Language
Acquisition. Cambridge: Harvard University Press.
Tomasello, M. (2008). Origins of Human Communication. Cambridge: MIT Press.
Tomasello, M. (2009). Why We Cooperate. Cambridge: MIT Press.
Tomasello, M., M. Carpenter, J. Call, T. Behne, and H. Moll (2005). Understanding
and Sharing Intentions: The Origins of Cultural Cognition. Behavioral and Brain
Sciences, 28, 5, 675–691.
Tomasello M., M. Carpenter, and U. Lizskowski (2007). A New Look at Infant Point-
ing. Child Development, 78, 705–722.
Tomasello, M., R. Strosberg, and N. Akhtar (1996). Eighteen-Month-Old Chil-
dren Learn Words in Non-Ostensive Contexts. Journal of Child Language, 23,
157–176.
Tomasello, M., A. P. Melis, C. Tennie, E. Wyman, and E. Herrmann (2012). Two Key
Steps in the Evolution of Human Cooperation: The Interdependence Hypothesis.
Current Anthropology, 53, 6, 673–692.
Traugott, E. C., and R. B. Dasher (2002). Regularity in Semantic Change. Cambridge:
Traugott, E. C., and B. Heine (eds.) (1991). Approaches to Grammaticalization. Am-
sterdam: John Benjamins.
Trudgill, P. (2011). Sociolinguistic Typology: Social Determinants of Linguistic Com-
plexity. Oxford: Oxford University Press.
Tuggy, D. (1993). Ambiguity, Polysemy, and Vagueness. Cognitive Linguistics, 4,
273–290.
Verhagen, A. (2009). The Conception of Constructions as Complex Signs: Emer-
gence of Structure and Reduction to Usage.Constructions and Frames, 1,
119–152.
Vrij, A. (2001). Implicit Lie Detection. The Psychologist, 14, 58–60.
Vygotsky, L. (1978). Mind and Society. Cambridge: Harvard University Press.
9780190256623-Dor.indb 243 03/06/15 6:18 PM

24 4 • References
Vygotsky, L. (1986). Thought and Language. Cambridge: MIT Press.

Watts, I. (2014). The Red Thread: Pigment Use and the Evolution of Collective Ritual.
In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of Language, 208–
227. Oxford University Press.
Weaver, W., and C. E. Shannon (1963). The Mathematical Theory of Communication.
Champaign: University of Illinois Press.
West-Eberhard, M. J. (2003). Developmental Plasticity and Evolution. Oxford: Oxford
University Press.
Wheeldon, L. R., and S. Monsell (1992). The Locus of Repetition Priming of
Spoken Word Production. Quarterly Journal of Experimental Psychology, 44A,
723–761.
Whitehead, C. (2014). Why Humans and Not Apes: The Social Preconditions for the
Emergence of Language. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social
Whiten, A., J. Goodall, W. C. McGrew, T. Nishida, V. Reynolds, Y. Sugiyama,
C. E. G. Tutin, R. W. Wrangham, and C. Boesch (1999). Cultures in Chimpan-
zees. Nature, 399, 682–685.
Whorf, B. (1956). Language, Thought, and Reality: Selected Writings of Benjamin Lee
Whorf. Cambridge: MIT Press.
Wierzbicka, A. (1992). Semantics, Culture and Cognition: Universal Human Concepts
in Culture-Specific Configurations. Oxford: Oxford University Press.
Wierzbicka, A. (1996). Semantics: Primes and Universals. Oxford: Oxford University
Press.
Wierzbicka, A. (1997). Understanding Cultures through Their Key Words: English,
Russian, Polish, German, Japanese. Oxford: Oxford University Press.
Wilson, D., and R. Carston (2007). A Unitary Approach to Lexical Pragmatics: Rel-
evance, Inference and Ad Hoc Concepts. In Burton-Roberts, N. (ed.), Pragmatics,
230–259, London: Palgrave.
Wittgenstein, L. (1953). Philosophical Investigations. Oxford: Blackwell. Wrangham,
R. W. (2009). Catching Fire: How Cooking Made Us Human. New York: Basic
Books.
Wyman, E. (2014). Language and Collective Fiction: From Children’s Pretense to
Social Institutions In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of
Language, 171–183. Oxford: Oxford University Press.
Yamamoto, S., G. Yamakoshi, T. Humle, and T. Matsuzawa (2008). Invention and
Modification of a New Tool Use Behavior: Ant-Fishing in Trees by a Wild Chim-
panzee (Pan troglodytes cirus) at Bossou, Guinea. American Journal of Primatol-
ogy, 70, 699–702.
Zahavi, A. (1975). Mate Selection—A Selection for a Handicap. Journal of Theoretical
Biology, 53, 205–214.
9780190256623-Dor.indb 244 03/06/15 6:18 PM

References • 245
Zlatev, J. (2008). The Dependence of Language on Consciousness. Journal of Con-

sciousness Studies, 15, 6, 34–62.
Zlatev, J. (2009). The Semiotic Hierarchy: Life, Consciousness, Signs and Language.
Cognitive Semiotics, 4, 169–200.
Zlatev, J. (2014). The Co-Evolution of Human Intersubjectivity, Morality and Lan-
guage. In: Dor, D., C. Knight, and J. Lewis (eds.), The Social Origins of Language,
Zlatev, J., T. Racine, C. Sinha, and E. Itkonen (eds.) (2008). The Shared Mind: Perspec-
tives on Intersubjectivity. Amsterdam: John Benjamins.
Zuidema, W. and B. de Boer (2009). The Evolution of Combinatorial Phonology.
Journal of Phonetics, 37, 2, 125–144.
9780190256623-Dor.indb 245 03/06/15 6:18 PM

9780190256623-Dor.indb 246 03/06/15 6:18 PM
Author Inde x
Addis, D. R., 27 Botha, R., 10, 12

Ahlsen, E., 10 Bourdieu, P., 102
Aitchison, J., 10, 29–30 Boyd, R., 10, 188, 212
Anderson, B., 27 Bresnan, J., 123
Anderson, J. M., 126 Brown, P., 172, 223
Arbib, M., 10 Brown, R., 120
Aries, P., 172 Brugman, C., 69
Armstrong, S., 67 Bruner, J. S., 10, 164
Arnon, I., 164–5 Butterworth, B., 121
Aronoff, M., 174 Bybee, J. L., 104, 123
Arthur, W. B., 190
Austin, J. L., 9, 13, 47–8, 53 Caramazza, A., 120
Aydede, M., 17 Carston, R., 71, 79
Cassirer, E., 89
Bach, E., 218 Cheney, D. L., 23
Baker, M. C., 9, 126 Choi, S, 91, 98–9
Barsalou, L. W., 221 Chomsky, N., 6–11, 13–15, 16, 90, 111,
Bates, E., 10, 12, 41, 164–5, 167 123–5, 146–9, 152–6, 161–2, 164,
Beattie, G. W., 121 174, 184–5, 187–9
Berlin, B., 91 Christiansen M. H, 10
Berman, R., 10, 93, 164 Clark, E., 10, 164, 221–2
Bicchieri, C., 48 Clarke, J., 172
Bickerton, D., 10, 28, 189, 206 Clark, H. H., 17, 36, 38, 46, 121
Bierwisch, M., 70, 81–2 Comrie, B., 10
Bloom, P., 92, 185 Connolly, A. C., 67
Bock, J. K., 122 Coppola, M., 174
Boesch, C., 172 Corballis, M., 10
Boroditsky, L., 91, 99 Coulmas, F., 12
9780190256623-Dor.indb 247 03/06/15 6:18 PM

248 • Author Index
Couper-Kuhlen, E., 10 Falkum, I. L., 71, 79

Croft, W., 9–10, 12, 104, 123, Fauconnier, G., 9
126–7 Fillmore, C. J., 9, 42, 69, 123, 132
Cruse, D. A., 9 Fitch, W. T., 10, 22, 111, 156, 185,
Crystal, D., 154 187–8, 212
Culicover, P, 123 Fodor, J. A., 19, 63, 67, 77, 90
Cuyckens, H., 70 Fox Tree, J. E., 121
Frawley, W., 134
Davidson, D., 11 Frege, G., 222
Deacon, T. W., 10, 29, 188 Fried, M., 9
Deane, P., 138 Fromkin, V., 56, 120
De Boer, B., 205
Dediu, D., 208–9 Gardiner, A., 11
Dell, G. S., 119 Garrett, M. F., 118–22
DePaulo, B. M., 212 Garrod, S., 117
Derrida, J., 47–8, 222 Gaskins, S., 91–2, 94
Descartes, R. 16, 21 Geeraerts, D., 60–1, 65–6, 68–9, 72–3,
Deutscher, G., 10, 154 76
Dewey, J., 16 Gilbert, A. L., 223
Donald, M., 10, 17, 19, 24, Gil, D., 95
200–1 Gilbert, M., 58
Dor, D., 10, 19, 27, 29, 89, 102, Ginsburg, S., 27, 197, 199, 204, 221
143, 197–204, 221 Givon, T., 104, 123, 154
Dorian, N., 154 Glăveanu, V. P., 201
Dornyei, Z., 167 Gleitman, L. R, 92–3, 98
Dowty, D. R., 9 Glucksberg, S., 73
Dummett, M., 11 Goddard, C., 64, 222
Duranti, A., 9 Goldberg, A., 9, 60, 104, 123, 132, 149
Grice, H. P., 9, 117, 145
Eagleton, T., 102 Gumperz, J. J., 9, 86, 91
Eckert, P., 9
Ehrman, M. E., 167 Harley, T. A., 10, 117–19
Elman, J., 10, 19, 65, 164 Harris, J. R., 165, 169–70
Enfield, N. J., 10, 56, 64, 222 Hauser, M. D., 22, 111, 156, 185, 187
Erteschik-Shir, N., 139 Hayashi, M, G., 56
Ervin-Tripp, S., 165 Heidegger, M., 16
Evans, G., 222 Hempel, C., 4, 11
Evans, N., 10, 90, 111, 147–9, 151, Hengeveld, K., 126–7
154, 157–9, 162, 188, 224 Herder, J. G., 88
Evans, V., 70, 123 Hockett, C., 22, 28, 149, 151, 158
Everett, C., 86, 102 Hofmeister, P., 138
Everett, D. 12, 102, 157 Hopper, P., 154
9780190256623-Dor.indb 248 03/06/15 6:18 PM

Author Index • 249
Humboldt, W. von, 3, 89, 224 Leacock, C. 68

Hurford, J., 10, 188–9 Leavitt, J., 86
Lehmann, C., 10, 154
Indefrey, P., 120 Lepore, E. 67
Itkonen, E., 11 Levelt, W. J. M., 10, 73, 117, 119–22, 222
Levin, B., 9
Jablonka, E., 10, 15, 27, 193, 196–9, 201, Levinson, S. C., 9–10, 21, 61, 86, 90–3,
204, 215, 221 95, 98, 111, 115, 134, 147–9, 151,
Jackendoff, R., 9, 19, 42, 64, 66, 75, 123, 157–8, 162, 165–6, 188, 208–9,
133, 180, 185, 206, 223 223–4
Jakobson, R., 13 Lewis, D. K., 11, 48, 198, 201
James, W., 16, 18 Lewis, J., 10, 19
Jitsumori, M., 76 Li, C. N., 129
Johnson, M., 97 Lieberman, P. 189, 208–9
Lillo-Martin, D., 148, 159
Kant, E., 21, 26, 88–9 Li, P, 92–3, 98
Katz, J. J., 63 Liszkowski, U., 35
Kay, P., 91 Lucy, J., 86, 91–4
Kegl, J., 173–4, 180 Lyons, J., 126
Keil, F. C., 92 Lyotard, J., 100
Kenaan, H., 222
Kendon, A., 201 Margolis, E., 67
Kiparsky, P., 9 Matthews, D., 165
Kisch, S., 179–80 McCloskey, M. E., 73
Klepousniotou, E., 83 McGrew, W. C., 201
Knight, C., 10, 19, 198, 201, 212, 214 McNeill, D., 120
Kobayashi, H., 37 Mehler, J., 46
Koerner, E. F. K., 12 Meillet, A., 11
Kohshima, S., 37 Meir, I., 174
Koyman, B., 165 Meltzoff, A. N, 9
Merleau-Ponty, M., 16
Labov, W., 9, 66, 73 Mervis, C. R., 65
La Frenière, P., 212 Mithen, S., 26, 209
Lakoff, G., 9, 69–70, 82, 97 Monsell, S., 120
Lamb, M., 193, 196 Montague, R., 9
Lamm, E., 201 Mufwene, S., 28
Langacker, R. W., 9, 123, 126, 129
Lappin, S., 139 Nettle, D., 154
Larson, R. K., 10 Newmeyer, F., 14, 153
Lasnik, H., 7, 153
Laughren, M., 224 Oakeshott, M., 18
Laurence, S., 67 Ochs, E., 10
9780190256623-Dor.indb 249 03/06/15 6:18 PM

250 • Author Index
Osherson, D., 67 Searle, J., 9, 13, 47–8, 53, 222

Östman, J. O., 9 Senghas, A., 174–8, 180
Seyfarth, M., 23
Papafragou, A., 92–3 Shannon, C. E., 54
Pepperberg, I. M., 158 Shegloff, E. A., 10
Peters, J. D., 11 Shen, Y., 95
Piattelli-Palmarini, M., 185 Sinha, C., 149
Pickering, M. J., 117 Skutnabb-Kangas, T., 154
Pinchevski, A., 49 Slobin, D., 10, 43, 86–7, 91, 93, 95, 160,
Pinker, S., 10, 90, 95, 148, 169–70, 185, 164
188–9 Smith, E., 67
Pollard, C., 123 Smith, J. D., 76
Pollock, L. H., 172 Smith, R. M., 212
Postal, M., 63 Spence, S. A., 212
Postma, A., 119 Sperber, D., 9, 36, 71, 116–17
Pustejovsky, J., 70, 81 Stapert, E., 158
Steels, L., 10, 221
Quine, W. V. O., 38 Steinberger, F., 222
Stephens, G., 165
Rappaport-Hovav, M., 9 Sterelny, K., 199–200
Ravin, Y., 68 Stevenson, L., 26
Rechav, G., 215 Stivers, T., 10
Reddy, M. J. 11, 57–9 Suddendorf, T., 189
Richerson, P. J, 10, 188, 212 Szabolcsi, A., 138
Riemer, N., 64–5
Robbins, P., 17 Taylor, J. R., 65–6, 68–9, 76, 83
Romaine, S., 154 Tomasello, M., 9, 10, 12, 17, 19, 24, 35–8,
Rosch, E., 9, 65–7 60, 164–5, 172, 188, 200–1, 212
Ross, J., 137 Traugott, E. C., 10, 154, 223
Rupert, R. D., 19 Trudgill, P., 9
Tuggy, D., 68
Sacks, H., 10
Sacks, J. S., 116 Vanderveken, D., 47
Sag, I., 123, 138 Verhagen, A., 60
Sakel, J., 158 Vrij, A., 212
Sandler, W., 148, 159, 174–5, 180 Vygotsky, L. 16, 20, 94, 102
Sapir, E., 11–12, 56, 89–90, 106
Sausssure, F. de, 11–12, 104 Watts, I., 201, 208
Savage-Rumbough, S., 189 Weaver, W., 54
Schriefers, H., 120 West-Eberhard, M. J., 193, 195
Scruton, R., 26 Wheeldon, L. R., 120
9780190256623-Dor.indb 250 03/06/15 6:18 PM

Author Index • 251
Whitehead, C., 171, 201 Yamamoto, S., 201

Whiten, A., 201
Whorf, B., 89–91 Zahavi, A., 210
Wierzbicka, A., 9, 64–5, 75, 222 Zawada, B., 70
Wilson, D., 9, 36, 71, 79, 116–17 Zlatev, J., 9, 11, 17, 19, 24,
Wittgenstein, L., 11–12, 16, 35, 38, 65 201
Wyman, E., 171 Zuidema, W., 205
9780190256623-Dor.indb 251 03/06/15 6:18 PM

9780190256623-Dor.indb 252 03/06/15 6:18 PM
Subjec t Inde x
acquisition, 3–4, 7, 10–12, 15, 35–6, bee dances, 2, 28, 188

38, 84, 92, 99, 104, 115, 131, behavior, 4, 22–3, 29, 32, 34, 37, 47–8,
150–3, 164–73, 175, 179–83, 50–1, 55, 67, 96, 130–1, 179, 182,
186, 197, 215, 219, 221–2 193–8, 209–11, 213–15
as collective project, 164–5, belief, 22–3, 26, 46, 100, 143–4, 157,
169–83 203–4, 213
alloparenting, 32, 199 brain, the, 1, 3, 7, 10, 17, 119–20, 150,
Al-Sayyid Bedouin Sign Language, 170, 193, 199, 208, 219
174–9
ambiguity, 68–9, 71–2, 83, 115, 157, childhood, history of, 172
179, 218, 222 children,
apes, 23–4, 26–7, 31–2, 35–7, 188–9, acquiring language, 7–8, 12, 15, 35,
201, 209–10, 213–14 38, 78, 116, 131, 150, 152–3, 155,
archaic humans (hominins), 19, 31–2, 164–73, 182–3, 219, 221, 224
184–215 inventing language, 173–83
Homo denisova, 209 in language evolution, 198, 199, 212
Homo erectus, 198, 202 and the relativity problem, 91–2, 94,
Homo heidelbergensis, 198, 200, 202, 96–100, 116, 131, 218
208 and their parents (and other
Homo neandertalis, 209 caregivers), 164–5, 169–73, 169,
Homo sapiens, 198, 208–9 176–7, 195, 199
argument structure (and thematic and their peers, 20, 165, 169–71, 178,
roles), 42, 70, 129, 132–4, 141 183
autonomy cognition, 1, 3–4, 6–12, 16–17, 21,
of language, 10, 62, 90, 204, 27, 37, 51, 54–7, 60–1, 87, 90–4,
206 102–4, 116, 119–20, 124–5,
of the lexicon, 60–3 133–5, 138–9, 146, 150, 157–8,
of syntax, 15, 124–46, 218 164–7, 220
9780190256623-Dor.indb 253 03/06/15 6:18 PM

254 • Subject Index
cognition (continued ) 101–3, 106, 110, 129, 134, 140,

in evolution, 4, 134, 184–99, 208–15 151, 155–61, 169–82, 190,
general, 8–10, 14–15, 19, 21, 60–2, 198–208, 212, 214, 218, 220
65–6, 69–71, 76, 134–5, 138–9, competence and performance, 7–8, 10,
146–7, 188 54, 104, 118, 130, 166
language-craving, 176, 182, 197–8, 219 componential analysis, see definitions
language-specific, 7, 9–10, 15, 60–1, (of word meanings)
85, 124, 146, 157–8, 184, 187–9, computers, 1, 3, 8, 13, 68, 72, 80, 103,
192, 197 219, 221
cognitive science, 1, 4–5, 9, 16–17, 21, computer science, 8, 184, 221
35, 146, 167, 175–6, 184, 216 comprehension, 15, 49, 51, 103, 105,
collaborative foraging, 200 114–17, 122–3, 139, 212
collaborative innovation, 201 concatenation, 206–8, 214
collective construction of language, 1, 4, conceptual structure, 19, 62, 70, 133,
15, 25, 32–48, 62, 108–10, 146–50, 157
164–83, 190, 193, 197–214, constructionism, 9, 123, 131–2
215–20 constructions, 50–1, 60, 83, 97, 131–3,
common ground, 36–7 174
communal lexicons, 38 conventions (and conventionalization),
communication 12, 23, 35, 38, 47–9, 71, 77, 82–3,
experiential (experience sharing), 2, 4, 103–17, 121–2, 127–30, 135, 145,
17, 23–4, 30–1, 50–1, 201, 203, 151, 157, 163, 171, 175, 178–81,
210, 220 218, 224
instructive, see Instructive conversational Analysis, 10, 49, 56
communication cooking, 32, 199
intentional, 2, 21–3, 50–1, 200, 203 cooperation, 12, 17, 199, 200
presentational, 22–5, 30–2, 35, 43, culture, 4, 8–9, 12, 19–21, 32, 64, 88,
50, 121, 210–11, 213 90, 92, 154, 157, 169–72, 209
re-presentational, 22–5, 28, 30
communication protocol, the, 25, 34, 38, deception, 46, 210–12, 220
46–9, 51, 54, 103–5, 108, 112, 114, definitions (of word meanings), 60,
118–23, 127–8, 134, 144, 150, 158, 62–7, 71, 76–7, 84, 217
161, 179, 182, 207, 217, 219 demonstration, 23–4, 30–1
communicative intent, 8, 22–4, 31, 34, descriptivism and prescriptivism, 49, 57,
45–7, 49–55, 101, 103, 105–7, 78, 146, 150–1, 155–6, 161, 218,
110, 115–18, 121, 130, 145, 151, 222
210–12, 222 development, 20, 21, 102–4, 164, 167–9,
communicative success and failure, 4, 35, 171, 195, 208–9, 212
38, 44–7, 53–9, 95, 105, 116, 131, diachronic and synchronic analysis, 8,
134, 179, 203–4, 217, 219 10, 71, 154, 216
communities, 1, 7–8, 12, 15, 25, 27, 29, dialogue, 23, 30, 32, 37–8, 201, 224
38–48, 60, 72–9, 83–4, 94–5, differend, the, 100
9780190256623-Dor.indb 254 03/06/15 6:18 PM

Subject Index • 255
discourse analysis, 49 148, 150, 158, 161, 165, 171–2,

displacement, 28 176–82, 197–8, 201–8, 210–11,
diversity (of languages), 11, 15, 90, 126, 214–19, 224
147–63 exploration and stabilization, 4, 38, 168,
dominance (in pragmatics), 139, 154 178–9, 190–1, 194–7, 203–7, 214
dyslexia, 215 eye contact, 24, 36–7, 200, 202
emotions, 3, 4, 21–3, 29, 31, 54, 100, facial expression, 29–30, 158–9, 200,
172, 197, 199, 201, 204–5, 208, 212–13
212–14 fast speech, 149, 189, 207–9, 219
empathy, 17, 199 feedback (in psycholinguistics),
empiricism, 88–9, 152 119–21
epistemic dependency, 203 feedback loops (in evolution), 199
epistemic solitude, 17, 37, 178, 182 feedback loops (in conversation), 121
eventualities, 41–3, 106–13, 126–8, formal analysis, 6, 8, 16, 34, 43, 47–8,
132–4, 140–4, 179 50–1, 63, 66–70, 72, 78, 81–4,
evolution of language, 2, 4, 10, 15, 19–20, 89–90, 94, 106, 109, 123–30,
27, 31–2, 76, 84, 110, 128, 134, 138, 152, 154, 186, 204, 207, 216,
147, 149–50, 152, 155, 158, 165, 218–19
184–215, 218 freeloading problem, the, 210–14
evolution of technology, 190–2 functional specificity of language, see
evolutionary theory, 184–5, 188, 193–7 instruction of imagination, the
experience, the nature of, 7, 16–32,
43–4, 94, 102, 106–7, 202–4 gender, 12, 21, 43, 68, 99–100, 102, 122,
experience of communication, 46–8, 179
103 gene-culture co-evolution, 4, 134, 188,
experiential anchor, the, 39, 45, 62, 191, 197–9
76–8, 84 genes, 4, 7, 85, 145, 184, 188–98, 208–9,
experiential cluster, the, 34, 41, 44–6, 214–15, 219
62, 78–85, 115, 127, 134, 163, genetic accommodation, 195–6
206–7 gestures, 23–4, 35–6, 38, 47, 121, 174,
experiential communication, see 178–80, 200, 205, 223
communication, experiential grammar, 3, 7–8, 11–12, 89, 91, 93,
experiential gap, the, 2, 19–24, 27–8, 99–100, 120, 122–5, 139–45, 151,
32, 34, 36–41, 44–8, 51, 53, 55–6, 157, 167, 174–5, 180, 217, 221, 224,
59, 62, 87, 104, 116–17, 151, 160, see also syntax
177–8, 181, 203, 215, 217, 219–20, grammatical functions, 129–31
222 grammaticality judgments, 7–8, 48, 62,
experiential mutual identification, 25, 67, 76, 79, 84, 96, 104, 125, 130,
29, 31–2, 34–5, 37–48, 59–62, 133, 136, 138–9, 143–5, 223
74–86, 94, 96, 98–100, 103–5, grammaticalization, 10, 154, 224
109, 113–14, 118, 130, 140, 146, grounding (in conversation), 38
9780190256623-Dor.indb 255 03/06/15 6:18 PM

handicap principle, the, 210 Kayardild, 157–8

Hockett’s design features, 22, 28, 149,
151, 158 langue and parole, see competence and
performance
iconicity, 154–6, 207 language change, 1, 4, 8, 10, 12, 82, 85,
identity, 4, 12, 48, 170, 179, 181–2, 87, 110, 191–2, 205, 214, 219, 223
200–1 language evolution, see evolution of
imagination, 26–7, 102, 148, 171, language
208–10, 218–19 language of thought, 19, 90
evolution of, 15, 27, 32, 202–4, 208–9 learning, 8, 10, 17, 23, 37, 94–5, 104,
instruction of, see instructive 152, 165–72, 177–8, 181, 188,
communication 195–6, 200–1, 212
imagined communities, 27, 102 lexical semantics, 3, 14, 60–85, 106, 115,
imitation, 17, 172, 200 132–7, 157, 180, 189, 217, 221
implicature, 105, 115–17, 223 linearization, 112–15, 120, 124, 130,
innateness, 7–8, 10, 12, 15, 19, 22, 35, 137, 141, 146, 206–7
61, 64–5, 72, 75, 85, 94, 104, 124, linguistic relativity, 15, 43–4, 62,
144–5, 149, 152, 164–5, 168, 170, 86–102, 217
172, 174–5, 180, 182–4, 188, 197, linguistics, 3, 5–12, 15, 49, 56–7, 64,
219 104, 118, 123–4, 129, 151, 154,
innovation (and invention), 1, 4, 21, 32, 160–2, 166–7, 173, 175, 187,
56, 87, 100–1, 151, 156, 158, 175, 216–20
177–82, 191–9, 201–2, 207–8, anthropological, 9, 12, 49, 88
210–12, 218–19 functional-cognitive, 9, 13, 69–70, 73,
instructive communication, 2–3, 14–16, 123–4, 126–7, 129–33, 138–9,
22–5, 27–35, 42–6, 61–3, 75–85, 154–6, 216
87, 94–101, 134, 151–2, 160, 166, generative, 6–8, 11, 62, 66, 70, 92,
171, 177, 203–8, 214–18 104, 111, 123–30, 137–49, 154,
as process of conversion, 49–57, 159, 162, 165–7, 184–8
103–22 interactional, 10
fragility of, 4, 35, 38, 44, 46, 53–9, psycholinguistics, 15, 56, 83–4,
105 104–5, 117–22
Internet, the, 3, 13, 106, 217, 220 socially-oriented, 1, 11–13
interpretation, 14, 34, 46, 49–53, 55–6, sociolinguistics, 9, 12, 48–9, 56
61, 63, 68, 70–1, 82, 85, 97, 101, unification of, 6, 10–11, 15,
103, 105, 114–17, 127, 151, 157, 216–20
207, 217–18 logic, 47–8, 111, 117, 161,171, 214,
intersubjectivity, 9, 16–17, 19–21, 24, 218, 222
35–7, 56 long-distance dependencies (island
constraints), 131, 137–45
joint attention, see intersubjectivity lying, 192, 211–14
9780190256623-Dor.indb 256 03/06/15 6:18 PM

Marbe’s law, 73 noise, 54–5, 205

Martian argument, the, 148, 166 normativity, 201
meaning, 3, 8–9, 12, 14, 22, 34, 39–44, norms (and prescriptions), 15, 25, 34,
47, 50–1, 55–7, 60–102, 105–7, 38, 46–9, 51, 54–6, 62, 84, 86,
110, 115–17, 120, 123–9, 146, 149, 95, 103–6, 109, 111, 113, 118–25,
154–6, 159, 162, 171, 177–9, 127–31, 135–42, 146, 151, 179,
204–6, 210–15, 217–19, 222–3 201, 207, 212, 215, 217–19, 224
lexical, see lexical semantics
private-experiential, 43–5, 50, 62, 78, parts of speech, 125–8, 174, 224
84–5, 125, 161, 217 nouns, 91–2, 121, 125–8, 138, 159,
social-semantic, see semantics 167
analogue vs. discrete, 18–19, 24–5, verbs, 42, 63–4, 69, 77, 125–8, 132,
29–31, 43–7, 75–8, 105–6, 115, 134–6, 157, 174, 180, 216
121, 125, 128, 133, 139, 203–6, philosophy, 11–12, 16, 88, 184, 204
117, 221 phonetics, 51, 54, 114, 147, 159, 205–7,
memory, 2, 21, 25–6, 32, 51, 54, 73, 215
79–81, 85, 93, 99, 139, 178, 186, phonology, 51, 113, 114, 119–20, 147,
189, 206, 209, 213, 222 149, 159, 180, 186, 207, 215
message, the, 51–2, 206–7, 213, 217 physiology, 4, 129, 189, 199, 208
message comprehension, 115–18 Pirahã, 12, 102, 157–8
message construction, 105–11, 113, play, 94, 165, 168, 170, 201
118–19, 121, 123–5, 129, 134–6, plasticity, 194–7
140–4, 146, 151, 158, 163 pointing, 24, 32, 35–9, 41, 74, 79–81,
mimesis and mimetic culture, 17, 19, 83–5, 159, 200, 202–5, 223
23–4, 32, 36, 121, 200–3, 205–6, polysemy, 60, 63, 68–72, 79–85
214–15 power (and politics), 4, 12, 43–4, 48, 87,
minimal specificity requirements, 110, 101–2, 171, 178, 218–19
134, 179 practical instruction, 29–31
miscommunication (and pragmatics, 9, 16, 49, 61, 71–2, 79, 82–3,
misunderstanding), 40, 46, 56, 84, 85, 105, 114–17, 138–9, 163, 217
111, 116, 145, 152, 178, 181 principle of contrast, 221
monkeys, 23, 76 principles and parameters, 7, 153
morphology, 51, 100, 112, 114, 147, 163, production, 14–15, 34, 49–52, 54, 83,
173, 180, 215 103–23, 135, 139, 141–2, 144–5,
mutual identification, see experiential 151, 178–9, 208–9, 222
mutual identification proficiency, 54–5, 97, 104, 114, 167, 181
propaganda, 87, 102
naming, 29, 31, 39–41, 73–6, 78–80, prototypicality, 60, 63, 65–9, 71–3,
83, 85, 91, 95, 121, 126–7 75–6, 78–9, 82, 84, 127–8, 223
Nicaraguan Sign Language, 15, 165, prototype (of technology), 4, 190–3,
173–82 197
9780190256623-Dor.indb 257 03/06/15 6:18 PM

psychology, 6–8, 11–12, 16, 20, 22, 49, symbolic landscape, the, 25, 34, 38,
56, 66, 72–3, 88, 100, 118, 184 39–46, 50, 54, 62, 75, 77–81, 84,
Ptolemy’s astronomy, 161 89, 99, 101, 104–12, 115, 117,
127–9, 134, 136, 146, 150, 161,
ratchet effect, the, 201 165, 163, 179, 205–7, 213, 215
rationalism, 8, 88–9, 148 syntax, 6–7, 9, 15, 43, 51, 67, 84, 105,
recursion, 111, 147, 156–8, 185–7, 224 109, 120, 122–47, 149, 151, 156–9,
relevance theory, 36, 71, 79, 85, 114, 183, 186, 206, 215–17, 224
116–17 syntactic selection, 131–7
repair, 7, 56, 79, 119, 121
rhetoric, 8, 214 teaching, 23, 29–31, 43, 80, 168–9,
ritual, 23, 28, 80, 201 172–3, 183, 200–1, 224
technologies (other than language), 1–2,
Sapir-Whorf Hypothesis, the, see 13, 24–5, 27–8, 49–50, 53, 103,
linguistic relativity 168, 190–3, 198–9
secrecy, 49, 213 theory of mind, 37, 201
semantic theory, 9, 42, 63–5, 70, 81–2, toolmakers’ paradigm, the, 11, 58
132–9, 204, 218, 222–3 tools, 32, 58, 111, 200
semantics, 34, 41–5, 94, 103, 105, 109, trust, 29, 46, 97, 100, 201, 203, 210–13,
112–13, 115–17, 120–7, 129, 220
132–43, 146–7, 149, 151, 154–5, truth, 3, 12, 61, 97, 174, 202–4,
157, 159, 163, 166, 169, 177, 185, 210–13
205–7, 217–18 typology, 10, 147
semantics, lexical, see lexical semantics Tzeltal, 92, 95, 98
semantic coherence principle, the, 132,
135 Universal Grammar, 7, 10, 149, 155
semiotics, 8, 11–12, 24, 91, 216 universals (and universality of
seriality (in language processing), 104–5, language), 3, 5, 10–11, 15, 56,
118–20 64–5, 75, 90, 94, 111, 123–4,
sign, the, 34, 39–46, 60–85, 205–6, 222 147–68, 170, 180, 184, 190, 199,
signified, the, 34, 41, 44–6, 51, 62, 76–8, 212, 215, 224
81, 83–5, 108, 111–16, 120, 122, utterance, the, 12, 15, 49, 50–2, 56, 82,
134–7, 141, 144, 163, 179 93, 103, 107, 109, 111–17, 119–22,
signifier, the, 34, 39–41, 44–6, 51, 80, 125, 130, 136–7, 144, 151, 154–5,
83–5, 111–15, 120, 131, 135, 137, 157–9, 207, 217–18
141, 159, 163
sign languages, 15, 158–60, 165, 173–82 variability
situated (embodied) cognition, 17 in experience, 21, 39, 43, 47, 139,
speech acts, 13, 27, 47–8, 107, 140–1, 155–6, 168, 205
145, 167 in individual cognition, 15, 192–3,
speech errors, 56, 118–21 195, 215
9780190256623-Dor.indb 258 03/06/15 6:18 PM

in innately-given capacity, 168, 181–3, and the relativity problem, 87,

193, 195–7 93–101, 120
between languages, see language vocalization, 28, 149, 202–3, 205–6
diversity
in knowledge of language, 15, 44, 62, word meaning, see semantics, lexical
68, 72–5, 80–4, 139, 191–2, writing and reading, 13, 17, 90–1,
212–13, 215, 218–19 115–16, 158, 166, 215
in language acquisition, 15, 20,
164–70 Yucatec Maya, 91–2, 95
9780190256623-Dor.indb 259 03/06/15 6:18 PM

9780190256623-Dor.indb 260 03/06/15 6:18 PM
9780190256623-Dor.indb 261 03/06/15 6:18 PM
9780190256623-Dor.indb 262 03/06/15 6:18 PM
9780190256623-Dor.indb 263 03/06/15 6:18 PM
9780190256623-Dor.indb 264 03/06/15 6:18 PM
9780190256623-Dor.indb 265 03/06/15 6:18 PM
9780190256623-Dor.indb 266 03/06/15 6:18 PM

The Instruction of Imagination Language PDF

Uploaded by

Copyright:

Available Formats

The Instruction of Imagination Language PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Instruction of Imagination Language PDF

Uploaded by

Copyright:

Available Formats

T h e I n s t r u c ti o n o f I m a g i n ati o n

9780190256623-Dor.indb 1 03/06/15 6:17 PM

Series Editorial Board

Recently Published in the Series

9780190256623-Dor.indb 2 03/06/15 6:17 PM

9780190256623-Dor.indb 3 03/06/15 6:17 PM

Oxford New York

Oxford is a registered trade mark of Oxford University Press

Published in the United States of America by

© Daniel Dor 2015

All rights reserved. No part of this publication may be reproduced, stored in a

You must not circulate this work in any other form

Library of Congress Cataloging-in-Publication Data

Printed in the United States of America on acid-free paper

9780190256623-Dor.indb 4 03/06/15 6:17 PM

9780190256623-Dor.indb 5 03/06/15 6:17 PM

9780190256623-Dor.indb 6 03/06/15 6:17 PM

Series Editor’s preface  ix

9780190256623-Dor.indb 7 03/06/15 6:17 PM

There would be no hope of understanding the foundations of human interac-

9780190256623-Dor.indb 9 03/06/15 6:17 PM

If it weren’t for good conversation, there would be no point to intellectual life,

9780190256623-Dor.indb 11 03/06/15 6:17 PM

9780190256623-Dor.indb 12 03/06/15 6:17 PM

9780190256623-Dor.indb 13 03/06/15 6:17 PM

This book attempts to tackle the fundamental question of the linguis-

9780190256623-Dor.indb 1 03/06/15 6:17 PM

9780190256623-Dor.indb 2 03/06/15 6:17 PM

Nothing in the social-technological approach to be developed here denies

9780190256623-Dor.indb 3 03/06/15 6:17 PM

phenomenon; the inherent difficulty and fragility of linguistic communica-

A scientific theory might therefore be likened to a complex spatial net-

9780190256623-Dor.indb 4 03/06/15 6:17 PM

Hempel’s characterization of scientific theory as a spatial network of terms, defi-

9780190256623-Dor.indb 5 03/06/15 6:17 PM

1.1 Setting the Theoretical Stage: Where Are We Today?

(a) Language is a generative system: it allows for the production of an infinite

9780190256623-Dor.indb 6 03/06/15 6:17 PM

9780190256623-Dor.indb 7 03/06/15 6:17 PM

9780190256623-Dor.indb 8 03/06/15 6:17 PM

Chomsky’s question was a major engine behind the stellar appearance of

9780190256623-Dor.indb 9 03/06/15 6:17 PM

9780190256623-Dor.indb 10 03/06/15 6:17 PM

1.2 Language as a Communication Technology

9780190256623-Dor.indb 11 03/06/15 6:17 PM

capacities, rather than innate pieces of linguistic knowledge, then language

9780190256623-Dor.indb 12 03/06/15 6:17 PM

9780190256623-Dor.indb 13 03/06/15 6:17 PM

(cognitively based) general-purpose system of communication (for a thorough

1.3 The Structure of the Argument

9780190256623-Dor.indb 14 03/06/15 6:17 PM

re-interpretation of the question of linguistic relativity, as a question about the

9780190256623-Dor.indb 15 03/06/15 6:17 PM

The question, then, is this: What does language do as a socially

2.1 Experience, Experiencing, and the Experiential Gap

9780190256623-Dor.indb 16 03/06/15 6:17 PM

9780190256623-Dor.indb 17 03/06/15 6:17 PM

Reflection on the character of experience finds in most men a mind filled

Second, the concept of experience carries a commitment to a dynamic and

Series Editor’s preface ix