Anshen Transdisciplinary Lectureships in Art, Science and the Philosophy of Culture


Previously Published: MoNoGRAPH ONE

REAL DISCOVERY OF AMERICA: MExic;o, NoVEMBER 8, 1519 by Hugh Thomas






Fred Hoyle


. . : and the philosophy culture monograph 3) 2. Queen Square. I.{)76-1 (pb) 1995 Printed in the United States of America Distributed in North America by Publishers Group West. LffiRARY OF CONGRESS CATALOGIN�IN-PUBUCATION DATA Chomsky. Thought ISBN 1-55921. science. - (Anshen of transdisci­ plinary lectureships in art. P. and thinking. without permission in writing from Moyer BeD. including photocopying. Language Series. . electronic or mechanical. Box 8843. recording or any infor­ mation retrieval system. 93-36148 CIP Includes bibliographical references (p. 1. Emeryville CA 94662. London NlO 2RN. Kymbolde Way. 3.means. Title. Muswell Hill. Rhode Island 02879 or 112 Sydney Road. em.O. 800-788-3123 (in California 511). II. Wake­ field. p. P37.658-3453) and in Europe by Gazelle Book Services Ltd.Published by Moyer Bell Copyright© 1993 by The Frick Collection All rights rese rved No part of this publication may be reproduced or transmitted in any form or by any.C546 1994 401'.9--dc20 and language&-Pbil050phy. Language and thought I by Noam Chomsky. Noam. Ulncaster LAl lRN Eng­ land 524-68765. Psycholinguistics.). Falcon House.


president of the Russell Sage Foun­ leader in various programs concerning behavioral sci- 7 . I am pleased to welcome you to The Frick Collection. Ruth Anshen's remarkable vision. and he no doubt will provide a forthright and i nvigorating investigation tonight . In order to facilitate the discussion foll owin g Dr. Eric Wanner. and the Philosophy of Culture. most provocative thinkers of our age.INTRODUCTION On the occ asion of this third Anshen Trans disci­ plinary Lecture in Art. be Dr. Science. Once again we are drawn by Dr. we have gathered on the stage several persons who will re ply to him. Noam Chomsky has frequently challenged us in his discussions of contemporary soc ial and political behav­ ior. Dr. Chom­ sky's talk. He remains a foremo st challenger regarding questions oflanguage and linguistics. this time to explore questions of langu age and th ought with one of the most celebrated. The lea der of this discussion will dati on a distinguished scho lar of cognitive science and a . It is an hono r to gree t you on behalf of our Trustees and our staff.


ences. He will be joined by three professors who are also present on the platform with us: Dr. Akeel Bilgrami, Professor of Philosophy at Columbia University; Dr. George Miller, Professor of Psychology at Princeton Uni­ v ersity; and Dr. James Schwartz, Professor of Neurology at

the Co llege of Physicians and Surgeons, Columbia Uni­ versity, also well known as a philosopher and a linguist.
The firsttwo lectures in this series are already in print: one by Lord Hugh Thomas, who spoke about The Real Discovery of America; and Sir Fred Hoyle's talk on The Origin of the Universe and the Origin of Religion, given a year ago, on April2, 1992, which has just been published. Dr. Anshen has guided these transdisciplinary lectures with indefatigable concern for every part of the process. Her own most recent book, Morals Equals Manners, was also published within the year, and has been received widely with acc laim She has been at the heart of all the

planning for this evening. Philosophy, mathematics, the sciences generally are held in her life with the same kind of grace, knowledge, intuition, and quickness of mind as are history, the social sciences, art, and music. Ruth Anshen is a rare-no, a unique-bearer in ou r own time of the ancient Greek spirit, of the Greek philosophical soul. We are most happy to have her open our proceed-Charles Ryskamp

" ings today.



Since we are all plunged in language, necessity is laid upon me to say the following few words;

If any single concern distinguishes present scholars in
all disciplines, it is a concentration of attention upon

language, but only iti each isolated discipline. Convinced as I

some of our worst contemporary muddles are due

to the general neglect of language as

instrument of

Language is far too interesting and important to be left

to the old-line philologist or contemporary reductionist. For there is

unity that springs from two assumptions.

One is that the connection between the human psyche and speech is far more subtle and complicated than one
realizes. The other is that language, though it is but one of the ways by which we communicate with each other, conveys not only thought but also emotion, so that a

merely rational analysis of language will no more explain

it than a chemical explanation .of a rose will define the rose.



phrase brings a blush

Those of us who are lovers of words, to. whom a fine


out, our salutation, our dis covery and our assent o what t happens to us from within and to what hap p ens to us
from without. Like all lovers we add as·�uch as we can to

merely one medium through which we expres s our crying

response, know that wor d s are

what we love. We are not l o vers for nothing, but for life
itself. I wi ll tell Tormented


you a short story:

Prophet wandered in a gloo my desert when an angel ' appeare d to him at the cross -road and crie d "Arise,

a spiritual thirst a traveller


w as a

Prophet, and go over land and sea and emblazone the Russian poet.) And as that. ancient Greek Epitome of wisdom, Heraclitus, said "Do not listen to me but to the Thought itself must be accompanied by a critical un­ Word." hearts of people with a Word." (So wrote Pushkin, the


deeper and most persistent intuition of man. It is by vittueofth,e provocative power of language which grasps


of the

relation of linguistic expression to the

shakes, and transforms that human' beings become hu­ man. For nothing really human can be so without this meaning, whether thelanguag!3 be uttered or silent. In thisJway, hinguage as the power ofuniversals is given to . . -us in order that we must transcend our environment, in orderthatwemay have aworld Thus, weare thrust into a state ofuniversal mobilitysince we attach the swift and
' : _-





restless force of our individuaL existence to the retarded only one source, namely, the very

and more massive poweroflanguage. And for this there is

nature of mind, If wor ds issued from an origin. other than mind; from one's native land, one's country, they would he born and


envisages. it is language severed from all practical. Language is a process of liberation from conceptual. to reveal them. and the intuitive gra sp in g of l angua ge is the primary act and function of that one and single power which is called reason. constructive. which entrusted to the writt en or spoken tradition defeats the voracity of time. tricked into an illusion. evoking powers hitherto hidden or inert but awaiting only that sum:rp. Thus we would be deceived. For it is then that we may pass from the passive acceptance of sense-data to a fresh. The p �o� lem of this wondrous gift which enables us to communicate with others. from the very inception of civilization. all immediate applicability which teaches us that words are the ultimat e symbols of ideas and that the power of life or death li es in the tongue. Language not only ar­ ticu lates connects and infers.PREFACE they would die with it. language. it also . It is language. Such was the power that. to invoke the p sychical properties of the mind and heart in invocation. full�panoplied. 1l . to foster their entranc e into existence and into time. bothclf which constitute the ulti" . issuing hot and pulsating from the human mouth. And it is power preci sely because it awakens to life secret and latent forc es Its work is that of . in e ntreat y through this unique instrument. to reveal to others the hidden secrets of our being. This is the problem concerning us. The word is power. taxing our minds and our hearts. in grief. or despair. man has �ttributed to the word.ons to br ing them into the· light. expressing our sorrows. logical or discursive rationalism. and alive. and spontaneous insight into the universe. Language thus becomes indispensable not only for the construction of the world of thought but also for the construction of the world of perception.

we remember. the word is given within us. It is the sp iri t made flesh." -Ruth Nanda Anshen .. has said. not only of commu­ nication and self-expression but of orientation in the universe.PREFACE mate nexus of an intelligible communion. are are that great philosopher of language. as Wilhelm von Humboldt. between all of us. sp iritual and moral. And finally. And it is the creation of ever­ widening horizons of human communication which is now coming to embrace all humanity that we are sum moned by an unbending i nner necessity to n o u rish ­ and honor-the vision of communication through commun­ ion. "We human not because we have language but because we language. an activity. the desperate isolation we e xperience finally breaks through in language. Language is an energy. We need not seek the word. 12 . The violent muteness.


They have elicited intricate and subtle inquiry. The thernes .little interest. what was then at all understood occu­ pies verylittle of today's curricula. That remains partially true today. accelerating in recent years. it has. Not m<1ny years ago. go bac k to the origins . Those days are long past. often meant displacement of pen(}trating insights in Javor of technical manipulation.become highly specialized.LANGUAGE AND THOUGHT Let me quickly all<1y any expectations that I might hope to do more than chip away at the rather grandiose themes suggested by the title of these remarks. Specialization is no proof of progress. and reach to the heart ofour nature. every faculty member in.· pf . it took no great effort to master the theoretical t:ontE)nt of linguistics and psychology. The empirically-oriented disciplines concerned with lan­ guage and thinking have .of recorded thought. though only partially. 15 . in my opinion. WhenJ was a graduate student forty years own department could take an active pru:t in every thesis defense.

stressing in advance that it is a personal and surely a :p1inority view." Nevertl. New questions are being posed that could not have been imagined a few years ago. Frege's basic a assumption is that "mankind possesses common trea­ sure of thoughts which is transmitted from generation to generation. Similar evaluations were given in the past. incorrectly in my view. and to which explanatory theory must be answerable.NOAM CHOMSKY Traditional questions are no longer forgotten or dismissed as absurd and senseless.teless. as they were during the heyday of "behavioral science" and the various brands of struc­ turalism. There has been explosive growth in the range of empirical phenom­ ena that are reasonably well understood. "a common science would be 16 . A standard starting point is the framework constructed by Gottlob Frege just a hundred years ago. seriously investigated. and they seem to be the right ones. which has proved a paradigm for much that followed. opening the way to new understanding. but you end up where you started. They have been reopened and in some cases. The gap between public relations success and relevant achieve­ ment often seems to me rather impressive. in some areas progress has been significant. it may be useful to recall Vol­ taire's thoughts about metaphysics: a dance with elegant moves. They should be regarded with a skeptical eye today as well." something that "cannot well be denied. I think. or about a "cognitive revolution. and unsuspected problems. I will try to sketch the landscape as it looks to me." Were it ·not the case. Even in the case of work of considerable care and sophistication. I have in mind claims about the enormous promise of neural net (con­ nectionist) models or artificial inJelligence.

and referent are external entities. it designates an object in the world. to understand an ­ true for every "well constructed" expression. and may not go far enough The Fregeau picture has technical problems that have inspired a great deal ofinsightful work. The idea that a sign picks out an object in the world to whiGh it refers "makes evident good sense. outside the 17 ." that will be . each person may have an indi vidual mental image connected with the objective sense. sign has a "sense" that fixes the reference and is "grasped by everybody" who knows the language. Akeel Bilgrami develops an account in terms of agents' concepti6ns with' a much narrower dependence on external object and shared public language." ." to which we may add related beliefs about indexicals and pronouns. its referent. suppose that we observe the moon through a telescope.Gareth Evans observes in one of the most important recent studies of reference. consi sting of sh ared signs. It has occasionally been questioned. mind/brain. Second. The basic picture has been widely adopted." These common thoughts are expressed in a shared public language. an object common to all observers. We may thiri. In addition. the individual retinal image is analogous to the individual mental image. in a "logically perfectlanguage. A sign has two asp ects First. but it is worth noting thal none ofits principles is obvious. a expression is to know its sens e in the shared public language. To adopt Frege's analogy. for examplei by Peter Strawson. In very recent work.k of Sign. as analo­ gous to the s ense . But such qualms are rare. Theissues that arise are too intricate artd wide-ranging to hope to · . sense. who warned 40 years ago of "the myth of the logically proper name. the real image of the moon projected in the·interior of the telescope.LANGUAGE AND THOUGIIT impossible.

" " etc. and philosophical se­ manti cs . Many would agree with Michael Dummett. intention.). the encyclopedist du Marsais and later Thomas Reid argued. and prov i des no grounds for positing "thoughts" to which John stan ds in a relation. in fact. the expressi on means only "John thinks" (desires. Consider the second . . grasp of the l ang u age." to adopt terms introduced whe n this" ordinary language" approach was reinvented 150 years later. But from this we can not move to s ay ing that there a re thoughts that they share. diamo nd." thoughts are expressed in a comm o n public la ngu age ··some version of this idea is presupposed by virtually all work in phil o so phy of l angu age . in cid ental ly a leadi ng Frege scholar that you and I not only share a public l a ngu age �English-exists " but that that la nguage "independently " of any particular speaker s . as we say that two p eo pl e live in the same place. etc. The b asic as sumption that there i s a common store of thou ghts surely can be denied. or a stor e of such thought s Philosophers have been misled by the ''surface gra mmar of a systemati . it had been plausibly denied a century e arl ier by critics of the theory of ideas who argued that it is a mistake to i nter pret the ex press ion "John has a thought (desire. This idea is completely 18 . In th e former case.NOAM CHOMSKY review.) on the analogy of "John has a . each of us has only a "partial. p erh aps so much alike that we even say they have the same thought. Argument is requir ed to show that though ts are entities that are "possessed. How solid the argument is may be questioned. . in my opi ni o n ." as di am onds are. To say that people have similar thoughts is to say t hat they think alike. Let me simply indicate a few doubts. " " ­ cally misle adi ng expression. assumption: " that the shared . and partial ly erroneous.

And there is no theoretical construct to replace them. we are told). except that the dimensi onality provided by interest and circumstance is . or explanatory gap that such a c ons truct might fill if it were devise d It is true . or that the worl d is divided into objective areas. we do not co nclude that there are common shapes that people share. as c ertai n authority figures ins ist? For the " emp irical study of language. In ordinary human life. the questions are mean i ng­ less. rather. or whether it means uninterested. as almost all speakers believe (ignorantly. of hopeless underspec ifi cation It is . languages " or " dialects " in ordi­ nary usage are complex amalgams determined by colors maps. oceans political institutions. or unbiased. also enter into various and shifting authority and defer­ en c e relations. Nor ha s " anyone ever indicated wh at sense it might have. What are called on . People not a matter of 19 . we find all sorts of shi fting communities and expectations . S imilarl y they may look a l ike and . From these facts. while neither talks like Deng Xiaoping. live near one another. far more diverse and complex. how do we decide. though neither looks like or lives near Deng.LANGUAGE AND THOUGHT foreign to the emp irical study of language " . that Peter and Mary may ta lk alike. for example. varying widely with individuals and groups. The problem is no t one of vagueness. whether the word disintereste d in the l anguage I partially ·know is pronounced as in Boston or in Oxford. and so on. even as idealiza­ tions. or languages and communities to which these lang u ages bel o ng . To ask whether Peter speaks the same language as Mary is like asking whether Boston is near New York but not Lon don or whether John is almost home. with obscure normative-teleological aspects. and no "right answer" as to how they shou l d be selected.

idered. replacing "brang" with "brought. in my study of meaning and intentionality relies uncritically on derstanding." True." a practice in a shared community language as Strawson ·puts the Wittgenstein­ ian point in his 1983 Woodbridge lectures. as a ventions. A good deal of themost serious work in the opinion.-Danish. only when there are standards of "correctness of use or application in terms ofcommon agreementin Hnguistic communication. or misuse oflanguage are general ly unproblem­ atic under conditions of normal usage as is "near New . there is no general as when we say that the language of south ern Sweden was once Danish but became Swedish a few years later without changing. It is.: a curiosity that thisis a doctrine of "ordinary language .'' contrary to "common agree­ ment. a language thatdiffers from inine in this among many other respects. perhaps. i nformal notions as Swedish�vs. a shared form of life. her internal language may change. "near B oston " or "looks like"." If it does not.NOAM CHOMSKY abstraction from diversity. Questions ofmeaning are usually 20 . any more than in the case of way to abstract. and must be seriouslyrecons." ·But they can hardly be expected to enter into attempts to reach theor et i cal un­ such notions." insofar as the word means anything. norms and con­ result ofmilitary conquest� Such York" or " l ooks like Mary. invoked to a cco unt for "rule-following" and for the fact of It is commonly assumed that such notions must be practice. and speaking it "correctly. Thus rule-following can be attributed philosophy. rather. though given particular interests we can. she'll be speaking."·sinceordinary language takes quite·a differ­ _ent course." we would not hesitate to say she is following the rule for "sing-sang-sung. If my granddaughter were to say ''l brangthe b ook.

Communication is a more. in fact. it seems. We might take note of the doctrine that attribution of rule-following. or impossible. As for communication. The task may be easy. unintelligible) notion of "access in prin cip le.LANGUAGE AND THOUGHT considered different. then introducing modifications as needed.or-less matter. This terminologi­ cal stipulation-it is no more than that-runs counter to common usage." as in John Searle's recent efforts to avoid obvious prob­ lems that have been raised. I will not pursue them here. beyond the level of c onscio usne s s . seek­ ing a fair estimate of what the other person said and has in mind. 21 . and accurate determination is rarely required for communication to succeed for the purpose at hand. and somehow more profound. A r eas onable speculation is that we tacitly assume that the other person is identical to us. that would be im interesting (and not surprising) empirical discovery. it does not require shared "public meanings" any more than it requires "public pronunciations. or mental states and processes generall y. It could turn out there really is seine­ thing like "public shared meaning." because the highly restrictive innate properties ofthe language faculty allow so little variation. but no different in relevant respects." Nor need we assume that the "mean­ ings" {or "sounds") of one participant be discoverable by the other. The doctrine becomes only more mysterious when supplemented by an unexplained (and. they seem merely more obscure. That has to be argued. has no place in inquiry into language and thought in the manner of the sciences. Jargelyreflexively. difficult. These matters have been discussed elsewhere. but there is no concep­ tual requirement that anything ·of the sort be true. requires accessibility to consciousness. and falls onto hopeless problems.

streets. still being London. though I did not use any word that picks out London by virtue of its meaning (conversely. We can regard London with or without regard to its population: from one point of view." But it does raise ques­ tions: are the technical innovations appropriate? True. Thus we allow that under some circumstances. if I say "London is or is not on the Thames. that begin in nothing. from another. that only want houses. and lead nowhere." with "spacious t avenues. it is a very curious thing. Frege had to invent technical terms. "I had an interview at BBC the other day and was shocked by the deterioration of the city. that I was referring to and talking about London. which only lack great thoroughfares to ornament" -but still Washington. is there an object London to which I am referring? If so. but it is quite a leap to conclude that the words refer to these things. we can say that London came to have a harsher feel to it through the Thatcher 22 . and inhabitants. people use words to refer to things and talk about them. not referring to London at all). public buildings that need but a public to be complete and ornaments of great thoroughfares. Furthermore. ··mile-long. that a sign selects an object in the world. roads. Suppose I tell you." I may be making a point of logic. London could be completely destroyed and rebuilt some­ where else 1000 years hence. for that reason.NOAM CHOMSKY What of Frege's third basic idea. theoretical discourse rarely adheres to "folk science. that same thing. Charles Dickens described Washington as "the City of Magnificent Intentions. it is the same city if its people desert it. in a manner determined by its sense? Note first that these ideas are not part of ordinary usage. accurately." Someone observing the exchange might say. That fact does not discredit these moves.

Suppose I say "that book. No object in the world could have this collection of properties. institutions. Fur­ thermore. which has the requisite understanding and structures.LANGUAGE AND THOUGHT years. but there neither are nor are believed to be things-in-the-world with the properties of the intricate modes of reference that a city name encapsulates. ugly. Referring to London. which I am referring is simultaneously abstract and concrete. to assume so would quickly lead to extreme paradox. which John wrote in his head. and refer to things. My internally stored language. weighs five pounds. the air above (but not too high). The most elaborate dictio­ naries provide no more than bare hints about the mean­ ings of words (or about the i r sounds).. a linguistic expression provides a com plex perspective from which to think about. an individual property. a comment on how people act and live. A single occurrence of the term can serve all these functions simultaneously. we can be talking about a location. talk about. Such terms as London are used to talk about the actual world. as when I say that London is so unhappy. and p oll uted that it should be destroyed and rebuilt 100 miles away. etc. not a thing in the world. in large mea­ sure independently of experience." The book to . we . the conclusion only becomes clearer as we move from the simplest case�proper names and common nouns-to words with inherent relational structures and more complex constructions.find much the same wherever we inquire into lexical properties. In general. just as the most elaborate traditional grammars provide only hints about the form and meaning of complex constructions-hints that may be adequate for a human intelligence. provides such perspec- 23 . buildings. or what we take to be things. in various combinations. people who sometimes live there.

NOAM CHOMSKY tives and ways for me to express my thoughts using them. To take a very simple example. consider the expressions "John ate an apple." as in the paired structions. we conclude that this sen­ tertce means that John. won't talk t�l Bill. similar in other respects.people won't.e. notbeingtalked to." yielding "John.. rather. "John is too stubborn to talk to Bill. and is largely independent of variations of range. here we apply the natural rule that if an expression is "missing. and insofar as they are similar. so doe s yours. being stubborn. 24 . But it does not. Their semantic properties are highly articulated and intricate . we can communicate more or The pervasive property of "poverty of stimulus" is striking even in the case of simple lexical items.'' Consi der next the sentence. Suppose we drop "Bill.. by analogy to the former case). being stubborn. and we are less well. is too stubborn to talk to. won't talk to sbmeone or other." meaning that John.:word con­ way we interpret "missing phrases." the latter understood to mean that John ate something or other.and known in detail that vastly tran sc en ds any relevant experience. not doing the talking. and the unspecified person i� doing the talking. it means that . To illustrate in a sli ght ly more subtle case. consider the sentence "Jones was too angry to run the meeting. John is being talked to.. ." "John ate." Applying the natural rule (i..talkto John (because ·of his stubbornness). That conclusion too becqmes more firmly estab­ experience and of spe cific neural structures over a broad· lished as we move to the meanings of multi." we interpret it to be "something or other. the interpretations of the former case are inverted." Who is understood to be running the meeting? There are two interpretations: The "silent subject" of''run" can be taken .

then we can be sure that the very phenomena will remain hidden from view. intuitively to mean that the crowd was suppose d to run the meeting unlike "which meeting was the crowd too angry for us to run?"-which has no ·�silent subject" that requires interpretation)." as part of our nature." interpreted counter . let alone any understanding of what lies behind them. without experience. probably because the facts are known "intuitively. applying to a wide range of phenomena in typologically diverse languages. Thereasons are well-understood in such cases as these. Even the relevant phenomena had escaped attention until recently.:how have: "which me eting was Jon es too angry to run?" Now the ambiguity is resolved. Or it can be taken to be unspecified in reference. If we are satisfied with the "expla­ nation" that things fall to their natural place or that our knowledge of form and meaning results from experience or perhaps na tural selection. so that the meaning is that Jones wouldn t run the meeting because of his anger. such as the fact that ai1 apple falls from a tree. in this case we say that the silent subject is "controlled" by Jones. Such understanding as we have of these matters does 25 . " .LANGUAGE AND THOUGHT to be Jones. " the a Jones refused to run the meeting (compare "which meet� ­ ing was the crowd too angry to run. so we . so that the meaning is that (say) we couldn't run the m eeting because of ' Jones's anger (comp are "the crowd was too angry to run me eting ) Suppose that we replace "the m eeting by qu estio n phrase. The crucial point is that all of this is known without experience and involves computational p ro c es se s and principles that are quite inaccessible to consciousness. Serious inquiry begins when we are willing to be surprised by s im ple phenomena of nature. or a phrase means what it does.

In this connection. goals and interests. common lan­ guages that we partially know. but there is no reason why it should. which pose all the same problems as arcane syntactic constructions: they are rule-governed. we may take note of the occasional suggestion that the problem of accounting for language use and acquisition is simplified or solved ifwe assume that these processes are somehow "based on meaning. or a relation of reference between words and things and a mode of fixing it. of our interactions with other persons and the external environment. sug­ gestions of Gerald Edelman's are a recent example. Though I won't pursue the issue here. but also not literally refutable or unintelligible." the idea being that semantics is soft and mushy. reflecting beliefs and perceptions. public pronunciations or meanings. within this skep­ tical view. An "internalist" approach that eschews such obscure notions seems to provide an adequate basis for the study of language and its use. and fixed in relative 26 . It provides no barrier against skepticism.NOAM CHOMSKY not seem to be enhanced by invoking thoughts that we grasp. and so on. I do not know of any (non-question begging) argument against a full-blown "internalism" of the Cartesian variety that doubts the existence of external things. algorithmic. sharply delineated. the concepts w ould (trivially) not be "of external things" and would not arise through experience or evolution. community practices. digital character of syntax and the apparent variability and continuous flux of individual experience and neural structure. The proposals seem to me of no interest. Such speculations cannot withstand the most casual look at simple semantic properties. Thus ''semantics-based" approaches • will not induce the "crisis" caused by the gap between the apparent rule-governed.

one might argue. his specific project is unaffected by a:ny of these consider­ ations. and its principles of organization and function are discovered. it is a typical problem of unification in the sciences: successful ex­ planatory theory at one level cannot be integrated with others. the belief that a semantics-based approach is �n alternative to one that is syntax-based derives from mi sreading of the literature so extreme as to defy brief comment. Recall that Frege was speaking of a "logically perfect language. but even "in principle incoherent. That will be true of any complex system. If so. That aside.LANGUAGE AND THOUGHT independence of experience and known aspects of neural structure. perhaps because the others have to be fundamen­ tally recast." He regarded natural language as not only imperfect. the problem is seriously mis­ stated. but it is not a "crisis" or "embarrassment" for cognitive psychology. Not many years ago. it was b elieved that knowledge of language is at most simple induction from extensive experience or even training and languages appeared to differ from one another as radically as neural structures do to many a trained eye today. Frege's project may be well-designed for his specific purposes. adapted to the present case. as sometimes alleged." Dummett argues. the goal is to develop a "perfect lan­ guage" to express a "common treasure of thoughts" with 27 . before it comes to be under­ stood. But we mayask what it has to do with the inquiry into language and thought. Perhaps very little. Furthermore. Voltaire's unkind thoughts may not be entirely unfair. . In the enterprise of science and mathematics." one that will allow for "a common science. Rather. The gap between what is known about language and about the brain sciences or experience is real enough.

working· scien­ tist in a reasonab le way.NOAM CHOMSKY terms that refer to actual things in the world. intentions are similar. thought and language: it would delineate an ideal to which t h ey approximate. The proposal seems to me to capture the commitments of the. merely "more perfect" than human language. understood as the kinds of nature. Suppose we take a Fregean "perfect language" to be a goal of science. ultim ately I presume that in the more central parts of the natural sciences. Pursuing the quest for theoretical understanding. or so it could be argued. The reasoning assumes. say. In the corner of science in which I work. some state or property of the brain." he means what he writes. that the scientific enterprise that proceeds ovor gonerations. that the symbolic systems constructed for science are languages. we will not intro­ duce a term such as "London. something like a frictionless plane." which can be used to refer in a wide range of perplexing ways. and uses the term "field" to refer. including natural kinds. lead­ ing to the deliberate and painstaking formulation of a . writes that such math­ ematical objects as electric and magnetic fields must be understoodto be "real physical 'stuff'" be c ause ofthe way they "push each other about. when I propose a technical concept embedded in a theoretical framework. 2M . and second. when Roger Penrose. I mean others to understand that I intend itto refer to something real-typically. It would be sheer (and unlikely) acddent if concepts of language and common sense understanding were to survive the transition to this enterprise. a concept foreign to natural language. Then the Fregean project would indirectly tell us something important about . Suppose we claim further that science is just refined common sense. first. even in a single occurrence.

Language acquisition is something that happens to a child placed in a certain environment. or as the child undergoes puberty at a certain stage of matura­ tion. It has led to deep confusion in modern linguistics and philosophy of language resulting from faulty structural analogies bet�een formal systems and natural language. The question whether "scientific languages" are lan­ guages is not a serious one. The symbolic systems created in the scientific enterprise differ radically from natural lan­ guages in their fundamental formal properties.LANGUAGE AND THOUGHT system of alleged truths. to human language.of whether machines can think. as in their semantic properties. T hese assumptions are hardly obvious. to formal arithmetic."ranging from cellular interaction to Shakes­ peare's sonnets to. or "com­ munication. "intelligent" extraterrestials. Let's consider them in turn. it appears. the problems only mount when we turn to questions of meaning and intentionality. any more than there is a study of "locomotion" ranging from amoeba to eagle to science-fiction space ship. The child's language "grows in the mind" as the visual system develops the capacity for binocular vision. We caul� not draw conclusions about the problems ofworkingAmericans by taking the physicist's concept of work to be an ideal that is approximated. There is no study of "language" ranging from ants. makes use of the mechanisms of common sense understanding. The move is no more appropriate in the present case. we might note that the same considerations hold for the much-debated question . To raise it is as pointless as asking whether airplanes really fly or cameras real}y see. opened in the modern period in a classic 1950 29 . to chimps. not something that the child does. In passing. To call them "language" is simply to adopt a metaphor.

In the latter case.whether a chess-playing program. though the classical discussion did not fall into the errors ofthe modern revival. he and his 10 . When Jacques dt:. . just as there is no empirical question of whether airp l anes can fly to London or whether submarines really set sail but do not swim.NOAM CHOMSKY paper by the British mathematician Alan Turing. The conclusion remains if we add further sensory conditions or cr iter ia beyond perfor· _ mance. Turing seems to have agreed with Wittgenstein as to the pointlessness of the discussion and debate that has ensued. play chess. he speculated. until today. A completely separate issue is whether simulation might teach us something about the process simulated. might teach us something about human thought. do long division. . in which he proposed what has since been called the "Turing test" for machine intelligence. not fact. conditions might have changed enough for us to alter our usage. I think Turing's stand was correct. over whether machines can (in principle) think. and so on. He regarded the question as "too meaningless to deserve discussion. in my opinion.l Vnucunson amazed observers with his remarkabln r:ontrlvuncos. for example. but in principle simulation certainly can provide much insight. just as some languages use the metaphor of flying for airplanes. etc . contrary to Turing's intentions. These are questions of decision about sharpening and altering usage. as has been proposed. the topic is very badly chosen." though in half a century. or whether robots can reach for objects and pick them up. The question has aroused lively discussion and controversy. and about how we could "empirically" establish that they do. murder. That much was well understood centuries ago. understand Chinese.

a direct gift of Nature. a precept of ." in my opinion. as . . His clockwork duck. artificial intelligence. not a facsimile that might fool his audience. and cogni­ tive science. At the origins of modern rationalist psychology. . Does the scientific enterprise employ the same mechanisms as ordinary thought and inquiry? Over a broad spectrum in psychology. A second question concerns thinking. was intended to be a model of the actual digestion of a duck. . it has been assumed that there are "mecha­ nisms of general intelligence. In this regard. That is the purpose of simulation generally in the natural sciences. the neuropsychologist John Marshall points out in a recent study. for example. 31 . there has been consid­ erable regression in the modern "cognitive revolution. Returning to the Fregean picture as a kind of friction· less plane. and surely no point in a debate over whether Vaucanson's duck really digests. though Th.LANGUAGE AND THOUGHT audience were concerned to understand the animate systems he was modelling. There is little if any role here for operational tests of one or another sort." general procedures that apply indifferently to various cognitive domains. the first question that arises concerns lan­ guage: is the picture presented in any meaningful sense an ideal to which human language might be an approxi­ mation? That has to be argued.ring himself was clear about the matter. Lord Herbert discussed the "principles or notions implanted in the mind" that "we bring to objects from our­ selves . it seems unlikely to withstand analysis. not merely presupposed. But we should take some care to sort out the components of the tradi­ tional beliefs. This "uniformity assumption" has deep roots. specu­ lative neurophysiology. philosophy.

" This " ep istemi c naturalism" ha s reappeared in the mo der n period in several forms . this has to do with the human a c tiv ity called· "science." and is· happy to dismiss beliefs "to which we are inescapably commit­ ted." but it has to be s hown that these are the same as those that . in my opinion." which has little regard for "the common sense of mankin d . su c h as Strawson's suggestion that we have a "general framework of beliefs to which we are inescapably committed" by virtue of our nature (so that debate with the skeptic is i d le ." "Thomas Reid held.NOAM CHOMSKY natural instinct".V. affairs of life. Doubtless scientific inquiry too is based on principles "imprinted on the soul by the d ictate s of Nature itself. these" common n otions" and "intellec­ tual truths" are " i mprinted on the soul by t he dictates of Nature itself. including W. Similar ic!eas were later developed in Cartesian psy chology and by Cambridge Platonists. he argues). and • in other versions. cons ists of "original· and natural judgments" that are "a "a part ofthat furniture which nature hath given to the human understanding" and "direct us in the common." which represents a sharp and unwarranted departure from the natural sciences." "The c ommon sense of mankin d ." such as the inescapable belief that the sun is dropping b elow the horizon or that spa ce is Euclidean. ·1 A separate question is what. In Burne's terms much later. including a version of Gestalt psychology. and is being rediscovered and given more substan c e in current empirical research. in a particularly r ich way. Quine's in fl uen tial "naturalized episte ­ mology. The e pi s temi c naturalism of early modern tho_ught ap pears to be quite reasonable. part of human knowledge derives "from the or iginal hand ofnature" as s p ecies ofinstinct." and th ough "stimulated by objects" are not " co nve yed " by them.

about how we handle "the common affairs of life." In domains where understanding goes. could they be deyised." or those uncommon affairs c al led "science. but that one can sense in important work and that students s omehow "learn by doing. has regularly found that the vari ous subcomponents of the mind function quite differently. no embryologist would be much interested in tQ. · Inquiry into particular ab iliti es aspects of knowledge and belief. an otherwise :unknown uniformity principle applies. " . say neural net or statistical appro ac hes to lan­ guage. and so on. There is little reason to suppose it to be con ducted with general skills and meth­ ods without spe cial adaptations. and come to appreciate. or engage­ ment in the craft. special structure is invariably built in. 33 .e proposal that unstruc­ tured systems with unknown properties might some day . indicates. when particular tasks are directly addressed. beneath the surfa c e we seem to find spe c i al structure and design." Living one's life is a creative activity too. It is still less dear why we should take seriously the extensive current discussion of the potential implica­ tions of unstructured alternatives to specific articulated theories. Proposals as to what may be common seem to reduc e to trivialities. or to believe that such devices apply as well to the specialized forms of creativ­ ity th�t humans engage in. similarly." Study of its history. That is far from obvious. such as "break a task into subpart s . that it is guided by curious concepts of intelligibility and insight that are very diffi­ cult to conv ey. It's not dear why we should take seriously the possibility that in just those domains where little :or n oth ing is understood. when they move beyond the "common affairs. Little is known.LANGUAGE AND THOUGHT "direct us in the common affairs of life/' as c ommonly assumed.

presumably. - where by "language. some imaginary Collapse of the " connectionist " system. perhaps. by now. If so. distinguishing parochial and culture-bound notions from tho olomunts of folk theo­ ries" that are a common humnn ondnwnumt. "organs of the body" has its properties They fall together. if true. at the level of cellular b iology but no "organ . surely the interactions are dense and close. we mean human language. on.1 . though we should be c areful to observe the practice of - serious ethnoscience. the study of language will neither provide a useful model for other parts of the study of the mind. hypothesis traditional uniformity should not come as a surprise. but that is another matter entirely." now. there will be no field of "cognitive science" dealing with the general properties of cognitive systems. this implies nothing about how language interacts with other mental faculties and sys­ tems. the cell s internal program pro duction of proteins. Each of these . theory" deals with the properties of organs in general . The various faculties and cognitive systems of the mind may be much the same. the kidney the circulatory system. We have. dedicated to language and its use . Specifically. and so ' . and others. "a direct gift " .NOAM CHOMSKY account for development of organisms without appeal to intricate theories involving concentration of chemicals. fairly substantial evidence that one of the components of the mind/brain is a language faculty . . nor draw from them significantly. Note that. not various metaphoric extensions of the term Other components provide "com­ mon sense understanding" of the world and our place in it what is often called "folk psychology" and the like. We find n othing like it in the study of other complex systems: t he visual cortex.

just as the symbolic systems constructed appear to differ radically from natural languages in their formal and semantic properties. we might take the mind/brain to be a complex system with a highly differentiated structure. Speculating beyond the little that is known. basically the view that things affect one anoth er by contact. the picture that has gu ided the most important work on these topic s in the last century may be seriously misconceived. The me th ods and goals of the scientific enterprise may tell us little ab out human thought in general. 35 . ignorance with a title. and one that is. He tried to show that in these terms one can explain every. Descartes offered a sketchy account of the physical world in terms of the "mechanical philosophy". those involved in moral and aes­ thetic judgment and in the special kind of rational in­ quiry undertaken in the natural scienc es and much else. not an easy problem." to dign i fy . These c ould be quite different in character from those that yield common sense under­ standing" in its various forms. It is. with separate "faculties. If so. I think. The history of modern science perhaps suggests that the distinctions are " not trivial." the Cartesian theory of body and mind. . useful to deepen the historical perspective beyond the Fregean paradigm and consider the "first cogn itiv e revolution.LANGUAGE AND THOUGHT of Nature". Other compon ents make it possible for human s to . and no dogmatism is in order. that is one way to interpret the startling conflicts that have arisen between common sense understanding and what scientific in quiry reveals." such as the language faculty. It is an open empirical question. at least. too lightly dismissed.conduct scientific and mathematical in­ quiry and sometimes to achieve remarkable insight: we may call them "the science-forming faculty. I think.

a second substan ce whose e ssen ce is thought. The Cartesian computational theory events. evoking thoughts that language use.NOAM CHOMSKY thing in the inorganic world. The "cognitive power is properly called mind. and a large part of the and animals and much of the functioning of humans. of a triangle. no artifact could exhibit the normal \n. Descartes laid the founda� to tions for modern neurophysiology. kin d of mind. The proposals have a contemporary flavor. But Descartes noticed that certain phenomena do not appear to fall within the mechanical philosophy. up organic world as well. ·• properties of language use : the fact that it is unb ounde d scope. internal computational resources p rod uce an image-say. among other contribu� tions.).rect contact. A series of physical stimuli. stimulate the reti na (�he hand. some new principle must be invoked.way-a c ollec � tion of properties that we may call "the creative aspect of ti o ns though riot caused by them. In the course of this sketch.' not ran d o m but coherent and appropriate to situa� thehearer might have expressed the same.. He also demolished the neoscholastic theory of perception. for the Cartesians." Descartes held. etc. always involving only di. or people walki ng in a crowd-on the basis of. and. when it manifests this . no t in accord philosophy. which a s sum e d that the form of a perceived object implants itself in the brain in some mysterious fash ion�obviously. not determined by external stimuli or internal st�te. Specifi� cally. these scattered of the with the mechanical alternativ e invoked a." Accordingly. he argued. including everything about plants e l ements of sensation and perception. and their impact was primarily in the areas that have flour� ished during the revival ofsuch notions since the 1950s: vision and language.

that excite many of the same question s · arises throughout the natural sciences: showing how ­ another object has.intelligent act (say. We also have the problem of determining whether If s ome creature passes the d evi se to test whether it expresses and interprets new thoughts as I do. Note that this is normal. like developing a li tmus test for acidity: the task is to determine whether one of the real components ofthe world is pre sen t in a certain case__. a mind like uurs. play chess).tiis basically thatofthe natural sciences. followers. the probl em of determining the nature of this res cogitans.artifacts that fascinated the 17th century imagi nation to those.nteract. 37 . notably Geraud de C ord e moy outlined experi­ mental tests thaf could be u s ed to answer this question. it would be "unreasonable" to doubt that it has a min d like mine. . It is interesting to. I think it is fair to speak of a conceptualregression since the cognitive revolution· of the 17th century. D es c artes and his. and we face the unification problem that We then have mind an d body i.acidity." beyo n d the c ontrol of sense or imaginati on or memory . Again. The approac. as it "forms new ideas" or "attends to those already formed. a change from reasonable (though incorrect) science to· an approach that is forei gn to the methods or concerns of the use. in the traditional formulation. and the re as oning is unaffected when we move from the complex. or a mind.LANGUAGEAND THOUGHT creative aspect. Gompare the Cartesian tests for the existence of othe r minds with the current reliance on the Turing test to determine "empirically" whether a ma­ chine can carry out some . garden variety science. C orde moy argues. and speculations today. focusing on language hardest experiments I can sciences.

Like many leading scientists of the day Newton found these results disturbing. that inanimate brute Matter s hould . the nature of consciousness). either the aspects just mentioned or others (say. the Cartes ian program c ollap sed within a generati on It is c ommonly derided today as the belief th at there is "a ghost in the m achine. However one jud ges the achievement. it does not deal with the traditional mind/ body problem. • 38 . " . not the gh ost Newton found that bodies had unex pected ghostly properties. Newton demonstrated that the Cartesian theory of the material world was fatally ina d equate. the theory of mind." .NOAM CHOMSKY The traditional mind /bo dy p robl em is often miscon­ ceived in the recent revival. agreeing with the Cartesians that "It is inconceivable. . But that conclusion mistakes what happ ene d It was the Cartesian . " . their "occult quality" of action at a distance trans cends the co mmon notion of body or material object. unable to account for the most element ary prop erties of motion. Thus Herbert Simon argues in his autobiography that a 1956 program for proving theorems of propositional calculus "solved the venerable mind/body pr oblem explaining how a system composed of matter can have the pro perties of mind by treating symbols as m aterial patterns. He had nothing to say about the ghost in the machine. theory of body that co llapsed. the idea of action at a distance through a vac uum is "so great an Absurdity that I believe no Man who has in philosophical matters a comp etent Faculty of thinking. su ch as it was. remained unaffected. As is well-known. without the Mediation of something else. can ever fall into it. which is not material. operate upon and affe ct other matter without mutual Contact". he exorcised the machine.

as does the view that "philosophical accounts of our minds." the domain of "the natural sciences. If study of Newton's occult quality leads to p ostulation of curved space-time. Con formity to common sense understanding is henceforth put aside. including those previously called "mental. 39 ." Hume wrote in his History of England. even if we cannot explain it in terms of the self-evident "mechanical philosophy. or perhaps ever. The world is what it is." None of these positions can be formulated coherently without a delimitation of the "ma­ terial world. Meta­ physical dualism becomes unstateable. and harmonious with." As many commentators have observed. however common sense may be offended thereby." in which the goal is " not to seek ultimate explanations.R." While "Newton seemed to draw off the veil from s ome of the mysteri es of nature.LANGUAGE AND THOUGHT Newton concluded that we must accept that universal gravity exists. Baldwin calls (approvingly) "meta­ physical naturalism. or yesterday. Bernard Co­ hen). this intellectual move "set forth a new view of science." Such notions as "physicalism" or "eliminative materialism" lose any clear sense. so be it. These moves also deprive us of any determinate notion of body or matter. with whatever strange properties it has. the natural sciences" (Daniel Den­ nett). "he shewed at the same time the imperfections of the me­ chanical philosophy. and thereby restored her ultimate secrets to that obscurity in which they ever did and ever will remain. as a criterion for rational inquiry. a view that T." but to find the best theoretical account we can of the phe­ nomena of experience and experiment (I." But what is that? Surely not what was called "physics" a century ago. our knowl­ edge. our language must in the end be continuous with.

0 . it would by rational inquiry ·in to reject the Periodic table. on basis for these conclusions. which the only way we know . except as a • grounds tional processes. apart from t hose com­ these grounds. as far as I can see. results of 19th century chemistry could not be accommodated by the physics of the day. which is 4. often in quite far-reaching wa y s. not 'continuous with' those a dv anced grasping by the natural sciences. part of the presumed truth about the proceeds. ." The reason lies in Frege's hy p othesis that a thought As Baldwin expresses the critique.NOAM CHOMSKY sciences in some way. the theory of organi c molecules. perhaps modifying them as inquiry judged legitimate. meaning in natural language conclusion that the remarkable properties of form are explained by computa­ no and on that contemporary biology offers apparent terminological device to distinguish various aspects of the We might turn. our criteria. to all of matter. it cannot be formulated in anything like the Carte� sian manner. that's a fact about the world. Ideas that yield understanding and insight are and to assimilate what we find to the core natural We seek to extend our understanding of the world world. to a stan dard criticism of natu ral world. If we mu st try to comprehend in the manner of the sciences. it is no less irrational to. and so have been an absur di t y Similarly. as understanding grows. of rationality and intelligibility may hu m an s have "ghostly properties" mon also change and develop. As fo� the mind/body distinc� tion. at this point. on. Frege s "hypothesis is ' held to violate the conditions of metaphysical naturalism. that is. the Fregeau paradigm for its avo w ed Platonism. or any other. dismiss the valence. if the.

lena that are se-· lected to <:�. At least. takes place is a. f. The critique assumes that "metaphysical naturalism" is an intelligible· doctrine. processes. We do not as sume a metaphysical divide when we speak of chemical events. curved space-time.aspects of the natural world. Though metaphysical naturalism appears to be unfor­ mulable. processes and states." massless partic les . incll)d­ ing what have traditionally been called mental.LANGUAGE AND THOUGHT on the very confines of the which for that reason cannot be completely understood from a purely psychological standpoint. based upon some delimitation ofthe domain of "the physiCal" that excludes Fregean "thoughts" in principle. and this process is perhaps the most mysterious of all. we sho ul d investigate these aspects oftll. but includes mathematical ob­ jects that " push each other about.dvance the search into deeper principles. attempting to const). Baldwin agrees.l. I cannot. lllent�l and which. For in grasping the law something comes into view whose nature is .e world as we do any ofhers. process Frege's position. events. which holds that study of the mind is an inquiry into longer mental in the proper sense. and by postulating a process of "grasping a thought" that cannot be incorporated within the natural sciences. we can formulate a kind of methodological naturalism. infinite one�dimensional strings in 10-dimensioria1 space. and the same 41 . c onflicts with the doctrine that "all fundamental fo r c es are physical" by positing a thought that is objective b ut not physical. and states.nd that. and whatever will be c ontri ved tomorrow. of phenoil. we cannot understand the critique. But until the delimitation is explai n e d.'uct ntelligible exp lan at ory theories that provide i insight and understanding . namely the thought.


should be true of the domain of the mental, if we borrow traditional terms for descriptive purposes. This "natural­ istic approach" will look forward to eventual integration with the core natural sciences, but whether that is pos­ sible in principle, or for a human intelligence, is a question of fact, not dogma. This approach-what I will henceforth mean by "naturalism"-should be unconten­ tious, though its reach remains to be determined. Plainly, such an approach does not exclude other ways of trying to comprehend the world. Someone committed to it (as I am) can consistently believe (as I do) that we learn much more of human interest about how people think and feel and act by reading novels or studying history than from all of naturalistic psychology, and perhaps always will; similarly, the arts may offer appre­ ciation of the heavens to which astrophysics cannot aspire. We ate speaking here of theoretical understand­ ing, a particular mode of comprehension. In this domain, any departure from a naturalistic. approach carries a burden of justification. Perhaps one can be given, but I know ofnone. Departures from this naturalistic approach are not uncommon, including, in my opinion, much of the most reflective and considered work in the philoso­ J2hY of language and mind, a fact that merits some thought, if true. A naturalistic approach will assume that like other complex systems, the human brain can be profitably viewed as an array of interacting subcomponents, which can be studied at various levels: atoms, cells, cell assem­ blies, neural networks, computational systems of the kind pioneered, at a primitive stage, by the Cartesians, and so on. We cannot know in advance which (if any) of these approaches will provide insight and understanding. In



several domains, including language the computational app roaches currently have the str ongest claim to sc ien
, .


tific status, at least on naturalistic grounds We might ask whether a study of the brain in such terms is improper or controversi al If not we then ask
. ,

whether the theories develop ed are true. Does the brain in fact have the architecture, s ubsystems states prop erties spelled out in some parti cular theory? As for the first query, it is hardly controversial to su p pose that the brain,
, , ,

like other complex s ystem s has subsystems with states

and properties. The properties attributed in computa tional theories are by and large well un derst ood. No general conceptual issues seem to arise, only questions of truth the secon d query, which we may put aside here. It has been common to try to relieve uneasiness about computational approaches by invoking computer models to show that we have robust, har d headed instances of the
­ , -

kind: p sy chology then studies software problems. T hat is a dubious move. Artifacts pose all kinds of questions that do not arise in the case of natural objects. Whether some object is a key or a table or a computer dep ends on designer s intent, standard use, mode of interpretation,

and so on. The

considerations arise when we ask whether the device is malfunctioning, following a rule, etc. There is no natural kind or normal case. The hardware-software distinction is a matter of interpreta­ tion not simply of physical structure, thoug h with fur­ ther assumptions about intention, design, and use we could sharpen it. Such questions do not arise in the study of org anic molecules, nematodes, the lang uage faculty or
, ,

other natural objects, viewed (to the extent we can achieve this stand point) as what they are, not in a highly intricate and shifting space of human interests and con-



cerns. The belief that there was a problem to resolve, beyond the normal ones, reflects an unwarranted depar­ ture from naturalism; the solution offered carries us from a manageable frying.pan to a fire that is out of control. We naturally want to solve the unification problem, that is, to re te studies of the brain undertaken at various l a levels. Sometimes unification will be reductive, as when much of biology was incorporated w'ithin known bio­ chemistry; sometimes it may require radical modification of the more "fundamental" discipline, as when physics was "expanded" in the new quantum theory, enabling it to account for properties that had been discovered and explained by chemists. We cannot know in advance what course unification will take, if it succeeds at all. If thereare answers to the questions we raise, there is no guarantee that we can find them; or that we. are capable of asking the right questions in the first place. Any organism has certain ways ·of perceiving and inter­ preting the world, a certain "Umwelt" or "cognitive space," determined in large part by its specific natureand by general <properties of biological systems. Given an organism with its special cognitive systems, we can identify a category of "problem situations" in which it might find itself: ari array of circumstances that it per­ ceives and interprets in ·a certain way by virtue of its , nature and prior history, including (for humans) ques­ tions that are posed and background belief and under­ standing that are brought to bear on them, and even problem situations that. are contrived on the .basis of theory-driven considerations and approached witha de­ gree of self-awaremess-:-the. activity that we call "sci­ ence." Some problem situations fall within the animal's c ognitive capacities, others not. Let us ·call these "prob-


in contrast. though ahuman might. a "prime humber maze. which has allowed some rays of light to penetrate· the general obscurity. and there are much eariier beliefs about our 45 . ·Philosophical questions would discipline: philosophy. Since Charles Sanders Peirce. is a permanent mystery. and mysteries that will be forever beyond our cognitive reach. the rat does not have the cognitive resources to deal with it." being t then be the "formulable mysteries" {for humans). not angels." respectively. In forthcoming work. The to be real. seek sol uti ons to mysteries. or even far simpler ones. chance convergence. We might think· of the natural scienc es as a kind of mysteries-for-humans. If humans are part of the natural world. There might even arise adiscipline·dedicated to this quest. wodd. the self-conscious and focused on problems. The concepts are point). natural world and properties. For a rat.. always failing. but they can hardly fail poses a· problem that a rat might solve quite well. of the hu m an mind/brain. As reflective beings we may· well . separating from the natural sCiences as they become increasingly same ·is true of them: there are problems that we might · hope to solve. and conversely.LANGUAGE AND. THOUGHT be only a p roblem for a monkey. A radial maze. deliberating about questions that chance convergence b etween aspects of the. Colin M�Ginn suggests that there indeed is such a appear to be of peculiar dep h and "hardness. there· have been prop os als about evolutionary factors that allegedly guarantee that we cari find the truth about the. distinctions need not be absolute. in that nothing in nature has "designed" us to deal with quandaries we face and can sometimes formulate. (turn right at every prime choiCe relative to an organism: what is a mystery for a rat might lems" and "mysteries.

" and "it would be absurd to doubt that of which we inwardly experience and perceive as existing within ourselves" just because it lies beyond our compre­ hension. We should not. That could be true. Drawing the natural conclusion. If central domains of the "mental" are cognitively inaccessible to us. in some way other than naturalistic inquiry. most forcefully by La Mettrie. consistent with anything we know about the natural world. Returning to Newton's demolition of the common sense theory of body. we face a series of questions: What exactly are these properties ofthings in the world? How do they arise in the individual and the speci es? How are they put to use in action and interpre­ tation? How can organized matter have these properties (the new version of the unificati on problem}? Certain aspects of these questions have been produc- 46 . That conclusion was drawn very soon. too qu ickly dismiss Descartes's specu­ lation that we may not "have intelligence enough" to comprehend the creative aspect of language use and other kinds of free choice and action. though "we are so conscious of the liberty and indifference which exists in us that there is nothing that we comprehend more dearly and perfectly. and so on. though neither at­ tempted to deal with the prop e rti e s of mind identified by the Cartesians. I think. like "powers of attraction and repulsion." electrical charge. But such speculations seem groundless. the natural c onclusion is that human thought and action are properties of organized matter. a· generation later by the eminent chemist Joseph Priestley.NOAM CHOMSKY unique access to the nature of our own minds and their products . we shall have to learn about humans. just as they have been put aside in the • revival of "cognitive sci en c e " since the 1950s. as best we can.

The fundamental Cartesian questions remain elusive. As Descartes concluded. and apparently unique to the species in essential respects. In the case of langu age . interpretation. it has been possible to study a number of traditional questions that had eluded serious inquiry. Pursuing a naturalistic course as far as it reaches. le ading to much new understanding of at least some central features of the mind and its ftmc­ tioning . the language faculty is. asking questions . In each state. matter and mind are not two categories of things . ex pressi on of thought. and nontrivial theories that go some distance in explaining the evidence. nothing remot ely similar has been detected else­ where in the biological worl d. for example. The performance s ystem s are g en erally assumed to be 47 . it passes th ro ugh a series of states and attains a relatively stable "steady state" at about puberty. but the y appear to pose entirely different kinds of quandaries for human intelligence. which use it for articul atio n . changing later only in periph eral respects. a c ommon human attr i bute . however. at least . The cognitive system stores informati on that is ac cesse d by the performance sy stems . to recast them significantly. we may di sting uish two components: a cognitive system and performance sys­ tems. and so on. we turn to the i nv e stigati on of parti c ular faculties of the mind. our knowledge ab out sound and meaning and their relation s over an unbounded range. There is by now a large mass of reliable data about these matters from a variety of typo­ logically different languages. Under the triggering and (marginally) shaping effect of experience. referring.LANGUAGE AND THOUGHT tively investigated. The cognitive sy s tem accounts for our infinite knowledge. to a good approximation. and more recently. Its "initial state" is deter­ mined by genetic endowment.

. ()f the way their elements are interpreted "at he interface". the language faculty of the brain. highly guage faculty develops in its no rmal course. We are studying a real object. for that matter that of my brother. as in studying embry­ ology. The reason is. The information prov ide d by lexical . as have a c quire d any language unless its fundame·ntal prop­ argued in the epistemic naturalism of early rationalist psychology. • 48 . say locomotion. external conditions are far tooinipover­ ishe d to have more than a marginal impact on the. in advance of experience. items and other expressions yields perspectives for thinking and speaking about the world hy ·virtue. not that of my parents. wife. entail­ ment. and surely . It is by virtue of the way the c o gn itive system is embedde d in performance systeins that the formal prop­ erties of expressions are interpr ete d as rhyme. or children._or.NOAM CHOMSKY fixed and invariant. and so on. enough is known to indicate that the differences among languages may not be very impressive co m p a red with the overwhelming commonal­ ity. they could serve as instructions for some other activity.. ign orance There is no evidence that this simplest :assumption is false The cognitive sys tems however. which has assumed a particular state ­ . at least from the standpoint we . the onset of puberty and oth e r aspects of growth and development. By now. of that we can be sure.adopt towards organ isms other than ourselves. basically. dO vary: my lan­ guage is not· that of a person in East Africa__:. The scientific problem is to establish explic . embed­ t ded in different performance systemsih some hypotheti­ cal (perhaps biologically impossible) organism. We could not articulated and intricate structures that arise as the lan­ erties were already in place. The variety cannot be great. ­ itly what we ass ume must be true. . .

) the language L when the cognitive component of Peter's language faculty is in the state L. More deeply. whatever). "without any grammatical instruction. and so on. Pursuing this course. . We may say that Peter has (knows.LANGUAGE AND TIIOUGHT that provides instructions to performance systems that play a role in articulation. written." Though 49 . speaks. Particular signs. The tradition persisted through the 19th century and beyond. the task is to notion of their structure which is def i nite enough to guide him in framing s ent enc e s of his own. ." As modern be­ havioral science and structuralist approaches were taking " form 70 years ago. The cognitive system is a generative proce­ dure that determines an infinite class of linguistic expres­ sions. we can progress towards under­ standing of some aspects of the Cartesian problem: the "infinite use of finite means. increasingly diverging from what was considered scientific linguistics. . the ability of each person to construct and understand "free expressions. speech acts are manifestations of lin­ guist ic expressions in a broader sense. will abstract some sound with a meaning." typically new. signed. telling stories. the topic is human language. are manifestations of linguistic expressions (spoken . . referring. expression of beliefs and desires. For such reasons. interpretation. a traditional conception. each a discover how the structures that underlie this ability "come into existence in the mind of a speaker" who. the language is a way to speak and understand. So regarded. the Danish linguist Otto Jespersen recognized that the central concern of the linguist must be free creation. in the Fregean sense. each a collection of instructions for the perfor­ mance systems." as Wilhelm von Humboldt rephrased it. from innumerable sentences heard and understood . .

The reopening of the traditional questions 40 years ago kept these assumptions intact. with options of variation restricted to subparts of the lexicon. later Greek grammar and its descendants. for example. The moder:n. with new formal insights that made it possible to pursue them seriously. The picture that is ·emerging is the first really signifi­ cant departure from a rich tradition extending to Indian grammar 2500 years ago. questions might be regarded as a confluenc� oftraditional Work of the past few years has to some extent suc­ ceeded in identifying general principles. 50 .least formul'able in a realistic. which were enormously influ­ ential. there is only one human lan­ guage.NOAM CHOMSKY important and basically correct. as a rational Martian observing humans would have assumed.• . unlike the far narrower Saussurean conceptions and behaviorist doctrines. The traditionalideas could not receive clear expression until the formal sciences provided the concept of genera­ tive (recursive) procedure. Acquisition· of a particular language is the process of fixing the lexical options on the basis of simple and accessible data. One goal of research. is to be able literally to deduce Hungarian or Swahili by fixing theoptions within the finit e lexical variety allowed . But mistakenly. way. which always. these ideas had little impact. ·study of these ideas that had been dismissed· as senseless or unwork­ able. now at . in this sense. The "com­ putational system" oflanguage that determines the forms and relations of linguistic expressions may indeed be invariant. of language that · can be attributed to in� tial endowment. took a language to be a system of rules that are specif�c to particular constructions in that language: rules for forming questions in English.

it predicts that language should be unusable. then. Grammatical constructions such as relative clause. pro perties that. This "minimalist" program also seeks to :reduce the descriptive technology to the level of virtual conceptual necessity. to a considerable degree. the array of phenomena derive from the interaction of principles of much greater generality. . their consequences reduc­ ing to more general and abstract propertie s of the compu­ tational system. they should yield much deeper insight into the computational processes that underlie our linguistic abilities. lexical opti. If these directions prove correct. processes that seem radic ally different from what was as:mmed only a few years ago. Still more recent work indicates that these principles may themselves be epiphenomenal. Such a. only scattered fragments are readily usable. pass iv e verbal phrase. An e?Cpression of the lan­ guage L. have a kind of "least effort" flavor. program faces an ex­ tremely heavy empirical burden. given the.ons for L. sharply restricting the devices available for description.LANGUAGE AND TIIOUGHT it now seems. of the class of "free expressions'' determined by the "notion of struc­ ture" in our "terrestial mammal" or "house­ hold pet".ed in terms of abstract principles of economy of · derivation and representation. which means that the complex phenomena of widely varied languages must be ex­ plaip. It is well known that language is b adly adapted to use". This conception of language critic ally introduces glo­ bal properties of computations of a kind that are known to yield extreme computational complexity. appear to be taxonomic artifacts. Even short and s imple expressions often cannot . and so on. The general conclusion is surely correct. would be a formal object that satisfies the universal interface conditi. . " 51 .ons in the optimal way.

NOAM CHOMSKY be handled easily by our performance systems. which appears superficially to be so "messy. the same is true of simple problems ofreasoning. That is a hard and interesting problem. however. usability cross-cuts deviance. many only recently discovered. some deviant expres­ sions are perfectly comprehensible. the other can inter­ pret. a harder but not completely recalcitrant problem. The tional systems of the mind/brain." though in fact. standing alongside the older problem of explaining a broad range of properties of sound and meaning. It is the mechanisms that enter into thought and action that have proven most amenable to inquiry. and about how perceptual-articulatory systems interpret and use the instructions they provide. It remains to be shown. over a large range. and approaching the unification problem-in brief. a fact that in itself may tell us little about our reasoning capacities. In the study of language." can produce something with the curious digital and computational properties of form and mean­ ing in natural language. so what one can produce. The unusability of language does not interfere with communication: speaker and hearer have similar languages and (perhaps identical) performance systems. A good deal is known about the acquisition of these systems. showing how the brain. Further­ more. just coming into focus. that the specific predictions about usability are correct. there is new understanding of the computa­ monly called "phonetic" or "semantic. all are "syntactic" in the broader sense that they have to do with mental representations. including those com­ 52 . while non-deviant ones often are processed inaccurately (with respect to the cognitive system). there are also many interesting ideas about the "conceptual-intentional" interface.

apart from the study of the mechanism that constitute it. to borrow the terms of the Spanish philosopher-physician Juan Huarte in the late 16th cen­ tury." All of these topics-including the ones that seem to be mysteries-still fall within the lower form of human intelligence. in my opinion. We should not necessarily regard that as an unhappy outcome. though not "compelled. at least for now. something that may involve "a mixture of madness. the nature of actions that are appro­ priate. been transcended." and the properties of free creation that are "properly called mind. those we may be "incited and inclined" to perform. Noam Chomsky 53 . however. which is foreign to "beasts and plants" but falls short of true exercise of the creative imagination. coherent.LANGUAGE AND THOUGHT limits of the 17th century revolution have not. The fundamental Cartesian issues still lie far beyond reach: the creative aspect of language use. perhaps forever. and more generally." Even the lower form lies beyond the reach of our theoretical understanding. They arise in connection with the "generative fac­ ulty" of ordinary understanding. and intelligible but apparently un­ caused.


It is p erhap s a symptom of what I probably should call the cognitive sciences-and not cognitive science-that we have as our panel a philosopher. On the other hand it rej ects various questions (such as the question: 57 . and a neuroscientist. On the one hand it ques tions a number of philosophical assumptions that frame a certain picture of language and thought. and it turns out accidentally. which he calls the Fregean picture. Professor Akeel Bilgrami: Chomsky's paper has a com­ plicated dialectic which is complicated in many ways. we will begin with philosophy. Eric Wanner (Moderator): I want to thank Noam Chomsky very much and pro ceed to the discus­ sion. Our first discussant is Professor Akeel Bilgrami of Columbia University. It is also surely a symptom of the range and impact of Noam Chomsky's ideas For alphabetical reasons only. proceed in decreasing levels of abstraction through psyc hology and neuroscience. .DISCUSSION ) Dr. a psychologist.

At the end of all this accumulated cri tique he offers some positive suggestions both methodological and the oreti cally substantive on his ch o sen theme of . a diagnosis which turns on its head this century's fashion for treating Descartes as a philosophical leper .DISCUSSION can machines think?) which philosophers. This is a n atur al conflation for those unfamiliar with tho dovolopments in this disciplines. in particular the discipline of the Philosophy of Languugo. Those in the audience who are not directl y involved in Cho ms ky s areas ofinterest will perhaps take the fierce ' seridusness and detail with which his arguments are ·mQunted and his prop os als offered as botroying a techni­ cal and narrow concep t ion of his subject. subjectivist Carte sian tendencies in such p hilosophers as John Searle and Tom Nagel . And in the midst of all this complex and d etailed business he slips in a rather beau­ tiful historical d i agnosis of what was right and what wrong in the Cartesian philosophy. But though it might bo natural. My brief comments on a paper of such rich persp ective and panoramic scope will of ne c essity have to be highly selective. the conflation is thoroughly unfair and m issos I ho point of one of his most central claims. Indeed I think it is fair to say that from the inside of these disdplinos. and it is worth point­ ing out that he manages to do this without the slig htest concession to recent rearguard. and cognitive sc ient ists have asked as not being well-formed questions and therefore leading to no part ic ul ar worth y intellectual pursuit. language and thought. It also fin d s certain attitud es that philosophers strike (such as for instance what he calls 'metaphysical naturalism') to be dogmatic and in some s ense even inconsistent. His Chom- 58 .

This is because in this community the meaning of 'arthri­ tis' is that it is a disease of the joints only. if Chomsky is right. which liberates the subject from artificial.DISCUSSION sky's position rather than many of the other current views that he is criticizing. It should be just as possible to say that that individual's term 'arthritis' m eans a wider class of ailment which afflicts both j oints and ligaments. On Chomsky's view however no clear s ens e can be made of the idea of what meaning is in a community and so no sense can be made of the idea that on an individual's lips the meaning of 'arthritis' or of any other term is constituted by com­ munal usage.. According to this assumption if someone living and speaking say. This denial of social or communal determinations of meaning is closely linked with his denial of a certain kind of normativity of meaning. What fills philosophers with qualm is that. is a true utterance. there will be no· scope to as sess our protagonist as saying something false or as mak ing a mistake: we can at best say that his individual language does not coincide in this lexical fragment with his neighbour's. This compulsion to find s p eakers correct or mistaken in this respect flows 59 . distorting and narrowing theories about the nature of meaning and the content of thought.D. a practising M. Th erefore on his view there is no particular compulsion to think ofthat utterance as false. Lets begin with the criticisms in his paper of the social conception of language as well as of the re lated idea that there is something essentially and intrinsically normative in the study of meaning. giv e n his meaning of arthritis. "I have arthritis in my thigh" he would be saying something false. in this upper-East side . If so his utterance "I have arthritis in my thigh". Let me exp la i n. say. community were to say out of medical ignorance.

to go beyond individual l anguages to a pretence of a more social and norm-governed phe­ nomenon. then I suspect he would be applauded for having liberated the study of language in precisely the way he has. Exactly the opposite is true for he has liberated the study of language from a set of unnecessary and ill-described constraints. and there­ fore his elevation of the authority figure in the commu­ nity (in this case. was just a sublimated form of the 'will to power'. This has given the imp re s s ion to many that he is narrowing the subject to leave out interesting social and normative elements.DISCUSSION precisely from a commitment to the normativity of mean­ ing which is missing in Chomsky's view. they have more to do with authority structures that lie outside of the notion of language. the doctor) as having some intrinsic norm�tive relevance for meaning. If Chomsky had made his point more flamboyantly and with the requisite hyperbole. And my suspicion is that the impression is largely a matt er of being misled by the presentation. Chomsky points out that there is nothing of theoretical interest in the idea of such norms of language because such norms are not intrinsic to the idea of meaning. The non-specialist is often heard to say that Chomsky is too scientistic about his subject of language and thought and ignores the fact that language is 60 . There is another p o int on which there is much uncom­ prehending complaint. and said that the philosopher's hankering to assess usage for a certain kind of correctness and mistakenness. The point is part of a soberly made critique based on a realistic sense of what of what is theoretically tractable and identifiable. as say Foucault might have. and it merely paralyses the prospect of giving an account of language and the cognitive aspects under­ lying our mastery of it.

This difference makes all the differ­ ence to what is at present scientifically tractable and what is not. language were best studied as referring to objects or classes of objects in the world. who have consis­ tently said that he has unnecessarily restricted scientific and formal theorizing to certain aspects of syntax and formal semantics leaving tho loxical aspect of semantics ("meaning") out in the cold.DISCUSSION a subtle craft. But again the plain fact is that Chomsky is m ore clear-headed than most other theorists and philoso­ phers of l an gua ge about how much of language as com­ monly conceived is not susceptible to scientific inquiry. This is in marked contrast to those philosophers who treat the items in the lexicon in terms of the concept of reference to things in the world. If one thought that the names and predicates in our temptation to think that meaning and semantics could be naturalized. and it is on ly because he has seen through the trumped theoretically ambitious philosophical accounts that he can afford to be clearer than others about this. That is to say the idea that the word 'cow'. then there would be a . one of the crucial points in his overall conception of language is that the lex ical aspects of language agent's perspective are on to be thought of as bringing in an things in the world. As he points out in his paper. refers to cows could be seen as constructible from and therefore reducible to the idea that individual occurrences of the representation 'cow' in an individual psychological economy stand in a regular (though of 61 . Evidence of this may be found in the fact that exactly the opposite up notions of language which are the object of other more complaint is made by the specialists. a mysterious art and a cultural instrument of great power. Let me explain again. for instance.

DISCUSSION this kind because more or less everybody stands in c ausal relations to more or less the same thi ngs in the world cows. chemically knowledgeable person's representation of 'w:ater' no doubt stands in causal relations'towater just as the chemical ignoramus's. Though I have obviously had to put it very crodely in ence to things iii the study of the lexicon and speaks Perspectives unlike reference are messy things. We can thus make semantics and inten­ tional psychology a scientific enterprise with. the time I have. if we are ever to say thatthese two people have the same meanings or concepts which explain their 62 . but their perspectives on the one has chemical bel�efs that the other does not. and' in particular they introduce an element that is too various and too shifting . That is we can have laws covering everybody who has representations of course not exceptionless) causal relations with actual around them. they introduce such 'things as beliefs as mediating the things in the world with which between one context of representation and another. I don't think I have done any violence whatsoever to Jerry Fodor's view of things in the exposi­ tion of thi s last paragraph. As far as meaning is concerned. conceived referentially along these lines. A · we stand in causal relations. Now it is precisely this conception of meaning that Chomsky has resisted' when he rejects the idea of refer­ instead of a linguistic agent's perspective on things. Since these are purely causal facts we can see meaning. So the universal laws covering all those who have these repre­ same substance are naturally entirely different because sentations are simply notto be had. universal generalizations. If there are to be any generalizations. as strictly subsum able under scientific laws.

field was only possible while one worked with a trumped up externalist. The answer has. A good question to ask is: why does the effort to see meaning in terms of reference and external causality seem too high a cost to pay for making .with the help of a contrast within Chomsky's own overall picture. I think. And it can be given . Those aspects of formal semantiCs and of syntax that Chomsky pla c e s within the ken of universal grammar and that fall easily within methodological naturalism's scientific ef­ forts are indeed asp e c ts of our cognitive make up That is. But the question really is: what under­ lies the stress on perspective that Chomsky takes to be essential to the lexicon. To say . to do with self-knowledge.that generalizations are highly localized in this way is pre­ cisely to say what Chomsky says in his lecture : that we should leave certain aspects of language and thought to the illumination that non-scientific enterprises such as commonsense psychology and history and literature at­ tempt to shed. The point is that the claim to science in this . referentialist picture of language as obj e c t of study. Here let me add a few remarks to Chomsky's own by way of diagnosis. Thus the non-specialist's impression of technical narrowness in Chomsky is unfair in just the proportion that the specialist's complaint and aspirations are misplaced. But the 1exical aspect brings in an element for which something more · . 63 .DISCUSSION linguistic and other behaviour it can only be in particular contexts or localities where the chemical beliefs and knowledge of the one are not relevant. the the ory of grammar here captures th i ngs we know (or 'cognize' as Chomsky puts it somewhere).the subject naturalistically and scientifically tractable? The short answer of course is because it leaves the perspec­ tival element out.

then we are in a position to see why purely referential and external causal relations will not capture what is essen­ tial. The reason is that to do so is to attribute to the former beliefs and 64 . and these are of course nothing over and above my conceptions and beliefs. I could. London are not on the forefront of my mind. But no such requirement of self­ knowledge is intuitively felt for the knowledge and beliefs attributed by the theory of the non-lexical and formal aspects of semantics or by the theory of syntax. but it would be quite wrong to form universal generalizations and attribute to the former (the ignorant) the same concept of water in all contexts as one would attribute it to the latter (the knowledgeable).to the lexical aspect of language. with enough attention and recall. bring them with more or less success to the forefront of my mind. even if I can't always verbalize them. If this is right.DISCUSSION than this may seem intuitively to be in play. there is an intuition that even when the perspective or conceptions invo lve d in my concept of. My term 'arthritis' or 'cow' or 'London' are caught up with certain perspectives on things in the world. Only something like an education in the linguistics department at the Massachusetts Institute of Technology will bring that about. So for example. But I think unlike the knowledge or cognizings attributed by the other aspects of the theory of syntax. chemically ignorant as well as chemi­ cally knowledgeable agents stand in causal relations with H20. say. that is if it is right that only the perspectival element that Chomsky stresses in the lexicon consists of knowledge or cognizings or beliefs that are themselves known to the agent in this way. recall and attention are not likely to bring anything substantial or precise to the forefront of our own minds. There.

according to this principle. 65 . despite orthographic coinci­ dence. self­ knowledge of this kind is taken for granted unless there .­ nosis I am offering for wh y these efforts at scientific treatment of lexical meani ng are bound to. But the perspectives . I think. Unlike syntax and the non-lexical and formal aspects of s eman ti c s when the subject is the lexicon and its pe rs p e ctival element. just as I might fail to know many things about other sorts of obj e cts non�mental ob j ect s such as cows or water or London. with Chomsky's resistance in his paper to the notion. must be individually treated.t when we attribute a thought to someone we are attribut­ ing some object to which he is related by his thinking it. and I should be curious to know what Chomsky would say in response to it. beli efs and c onceptions can be threatened by obstacles that come from non-psychological sources such as those posited by abstract philosophical theories of reference. fail at present is base d on an intuitively held principle. So the diag. tha. such as self-deception or inattention etc. is precisely what many philosophers have uncritically taken for granted: that self-knowledge of our perspectives.DISCUSSION intentional states that they could not possibly have self­ knowledge of thr ough recall and attention. owing to Frege. The princ ip le is just this. This principle fits in nicely. If there were such 'objects of thought' it would be possible for someone to say that I could fail to know my own thoughts because I could fail to know a lot of things about these objects of thought. to know many thing s about water that the chemically ignorant and the chemically knowledgeable have different perspectives on these things and their respective terms for water. are psychological ob s tac les to it. What we will not allow. It is because we can fail.

and that is why we cannot be ignorant about these perspectives. Let me conclude by changing the subject to the Turing test for intelligence and thought. barring internal psychological censors and obstacles. unless there were psycho­ logical obstacles of the sort I mentioned.l d also have descriptions under which we would have self-knowledge of them. what we w o:p. under those descriptions. All this I say by way of diagnosing what may underlie and underpin Chomsky's doubts about the scientific aspirations for the more common-sense aspects of lan­ guage that come with the lexicon. and raise a final ques­ tion for Chomsky.DISCUSSION themselves are not objects like London and water to which we are epistemologically related as Frege claimed. Both the idea that lexical meaning involves an external causal and naturalistically and scientifically tractable notion of reference. my point is that when we do such a theory. None of this is to suggest that we may not one day have a theory of perspectives which would describe perspec­ tives in a way that we would not as agents have self­ knowledge of them. But even so. as well as the idea of objects of thought. we cannot fail to have self-knowledge of them except of course if there are purely internal psycho­ logical censors such as self-deception etc. and if I'm right these doubts can now be integrated with what at first sight seem altogether independent doubts about the idea of Fregeau objects of thought. transparent to one. undermine the principle that one's own perspectives on the world are. Chomsky's view on contemporary uses of such tests is that they are based on an ill-conceived question: Can machines think? But I wonder if there is wouM have if we got it is a theory of something that 66 .

generally. The idea of such a list is not inconceivable of course since it is a finite list. but that seems to be a frivolous conclusion.DISCUSSION not a somewhat different and perfectly good question that they can be used to raise. a question to which his own conception of language provides a decisively negative answer. This is because its responses which perfectly mimics one of us are going to be finite since we ourselves are creatures with finite lives. suppose some would be tempted to say that all this shows is that for all we know we are not generative but merely following a long but finite list. that is proof that the machine which ex hypothesi pas se s the Turing test. thereby proving (as Popper might have said) that the question was a sharp and well-formed one. What this im agined example shows rather is that merely looking to external things like behavi our in re­ sponse to the world around us is not sufficient for the identification of thought and intelligence. And in saying it is frivolous one is agreeing with Chomsky about the importance of generat­ ivity. unlike us it is not a generative creature but is instead following a very long list of 'elaborate. Here's what I have in mind. This is the question whether the idea of intelligence and thought is something that is exhaustively characterizable in behav­ ioural and. Imagine a machine that produces all the verbal responses any one of us does to all the things it hears and sees and encounters. as Chomsky has himself always insisted. external terms that leave out any­ thing internal to the agent. The only difference is that all its responses are canned. we have also to 67 . counterfactually formulated instructions. which we may im agi ne is identical to what any one of us encounters. Now if generat­ ivity is basic to language and intelligence and thought. nevertheless has no more intelligence than my toaster I .

This talk tod ay was a typically rich pattern of themes· woven ·together in a typical graceful way. And so my question is: doesn't this show that Thring at least in one sense helps to pose a fairly sharp question and test? Dr. And so on. I'm a very lucky person because I've been able to listen to Noam ing experience. for a non-generative machine however we would have to add to its list of instructions. For a generative machine we might find we could revive it by adding to its memory-space. implicitly at least. For example. we may test for generativity versus canned list-following in a machine by seeing what is needed to be done to a machine's.DISCUSSION look theoretically at what is inside. given Chomsky's insistence on the criterion of generativity. Professor George Miller: Thank you. which he expressed. It's always areward­ couple that particu­ feeling that cognitive science· is headed in the wrong direction:. One was his 68 . at internal mecha­ nisms and the internal cognitive andfunctional architec­ ture.. Eric Wanner: Professor Chomsky has agreed to withhold his response until all three discussants have finished. only tests upon internal mechanisms of this kind will be sufficient to answer the question about thought and intelligence. functional makeup to revive it when it comes to a halt. but there were a Chomsky for almost fo:rty·years now. Eric. T worry about picking up any one of them lest the pattern fall apart. The general point is that if. at larly caught my attention and interested me. then the Thring test which makes only external demands is decisively refuted as a test for these things. . Our second discussant is Professor George Miller of Princeton University.

At any rate. Bruner. Chom�ky. have publicly announced that this thing we started that has been called the c ogniti ve revolution has been kidnapped and taken off in some direction that we disapprove of. ago__. I don't know..for about a decade. a narrow structuralism in linguistics and anthropology. We were just trying to make sense out ofthe obvious facts that we found in front of us. l ogic al ·positivism. The activities that we undertook jointly and separately have since been called by other p e ople a cognitive revolution. and cognitive scien­ tists generally.. It would be a fascinating thjng for me to get the three of us together and find out whether'we think the kidnappers have all .years ago-almost forty years . It may be that the only thing that united us forty years ago was our opposi­ tion to behaviorism.DISCUSSION several times The other is his warning that the human . and. to which I would like to subscribe it.. and I. the interesting thing to me is what Chom­ sky is trying to tell us wou ld be the right direction to go. iri so far as I understand The interesting thing to me is that within the p ast four or five years. done the same wrong thing. I interacted with Noam Chom­ sky and with Jerry Bruner in Cambridge. all three of us. I don't think any of us called it a cognitive revolution. forty.. intelligence may not be adequate to answer all the ques­ tions that we have been asking About the first one. many suggesti ons in this rich and stimulating talk that need to be picked up and looked at carefully by psychologists.. One that appeals to me is this warning that we may not 69 . linguists. a sort of anticipation of a view that I think Noam is now calling methodological naturalism. There are many.

The problem. of course. that I have with this is telling which is which. I believe that the federal government has. it would probably be far advanced over ours. is trying to solve the maze. At that point I remembered the psychologist looking at the rat. I summarize this as the principle that you can't conceive of what you can't conceive of. as we do the rat? Sagan never answered. and think of what all the wonderful things are that we could learn by interacting with such an advanced civilization. mysteries like 70 . I once tried to get into an argument with Carl Sagan about this. on the advice of some very distinguished scientists. and the rat. in which case. appropriated funds for listening for intelligent life from outer space. The argument for doing this is that if we could contact a civilization. but there are a lot of problems that the psychologist can solve that the rat doesn't even know about. but they may be fairly far along. I heard Noam talk about the psychologist looking at the rat. It's been very helpful to me over the years to keep that firmly in mind. I do very much like the distinction between the prob­ lems that we can. without the rats learning very much at all from this �uperior intelligence that we have. For example. Then. We are just barely out of the trees. perhaps eventually. even if I knew which were the mysteries that I might set aside. we'll have to settle for what we can achieve. Why is he so sure that they will not 1 regar d us as sort of a bipedal rat and try to exterminate us . I have to distinguish between which are the mysteries that I can ·safely accept. of course. And it's able to do that. solve and the mysteries that we can never solve. Years ago.DISCUSSION have sufficient mental capacity to achieve a reductionis­ tic account of the mind. How long have we inhabited this world with rats.

I am not a philosopher . Both ofthese new approach es to understand­ ing the mind have bee n extremely productive. movement. nor am I a linguist-my revolution is cell. for example.DISCUSSION action at a distance. and learning.and molecu­ lar biology. But Chomsky didn't promise to answer all of my questions. similar mapping of brain areas inv olving language is an e xtr emely active endeavor in rary neuroscience. and I am grateful to him for th at . and the five senses. I think he was really trying to point in the direction that a sensible cognitive scientist should try to go. Eric Wanner: Our thi rd discussant is Profe ssor James Schwartz of Columbia University. Schwartz: The 1960s saw two sci­ important to the study of the mind/ brain: the cogn itiv e revolution that we have been talking about and the reductionist revolution of cell. c o nte mpo­ 71 .and molecular biology. or human consciousness. In addition. Dr. despite what Mr. P erhaps more pertinently. the emotion. perhaps . mole cular and electrophysiological probes have confirmed and ex­ tended knowledge about brain areas that are critical to specific and mental functions. like God's will. I need only cite a few of the familiar adva nc e s made: an almost com plete descrip t ion of the molecules that mediate the conduction of the nerve impulse and of synaptic transmission. In my revolution-and . Professor fames entific re volutions H. Thank you. and the character­ iz a tion of neurotransmitters and neuromodulators thought to be involved in pain. Thus. memory. and col o r. and which are the one tha t I should be free to doubt . distinct cortical regions have been identified that are sp e cific for perceiving shape. Ryskamp says.

of which language is a fundamental and essential component in the human species. since he was one of the leaders of that revolution. I believe that is necessary to review its roots. my colleagues warn me to stick to molecules. "If we're so smart. As he recently wrote: "There is real grounds for considerable optimism about the prospects that lie ahead. (folk psychology like my grandmother's. why don't we have an answer to this fundamental question: "What. they say that the gap between molecules and mind is so vast that it is absolutely fruitless to think about bridging the gap. but also for the study of cognitive systems of the mind/brain. not only for the study of language proper. why aren't we rich?" If the state of information on the mind/brain is so rosy. At their politest. mind/brain." It is probably not necessary to say that Professor Chomsky is justly pleased with the cogni­ tive revolution. This school 72 . because it is the only hard-nosed approach. They solve the problem denying it. does the "/" mean in Chomsky's succinct formulation.) So what is all of this intellectual discomfort about? In order to understand the nature of the current discomfort in the neurosciences. But as my grandmother used to say. Unlike most textbook histories of science. Many of my colleagues are deeply pessimistic about the concept of mind. At their worst. which are told backwards from the vantage point of "the truth" of our present understanding through the hazy fog of losophy called Eliminative Materialism. ultimately must re­ duce to neuroscience anyhow.DISCUSSION Professor Chomsky also clearly considers his revolution-though he may deny it tonight-to be suc­ cessful. echoing the currently popular school of phi­ claims· that mentalist explanations of psychology.

to start around 1800 with the word of Franz Joseph Gall and trace the dialectical opinions ab o ut whether mental function is discretely local iz ed on the one hand. on balance. The conviction that the mind is situated in the brain and that mental functions are localized was extremely controversial. our biological place in nature. often thought of as a charlatan. I think. thes e ideas led to a materialistic view of mind that was as uncongenial to 19th-century thought as were the theories of Darwin. Why? First. 73 . As for Gall's other i dea. Their theories forced most members of society to recog­ nize three other apparently inescapable facts of life: Darwin. . Nonetheless. Marx. he asserted that the cortex is composed of special­ ized organs. history of i d e as For this p urpose it is convenient. Marx. ensemble property of the brain. b e c au s e his work led to the popular pseudoscience of Phrenology. Marx and Freud can be consid­ ered the fathers of 2oth-century thought and culture. or some . like Count Cagliostro and Franz Anton Mesmer. along with other fashion ­ able but obvious quacks of the late 18th century. the evidence accumulated so far favors the localizationists. on the other Gall is still . if Darwin. a reasonable case can be made for Gall as grandfather of the 20th century. This evi. both simple and com­ plex. Thus. each de di cate d to a given mental faculty. Inevitably. because Gall finally convinced most members of society that the brain is the organ of the mind. and Freud. . the funda­ mental importance of economic structure.DISCUSSION prior error the story should be told in the context of the . and Freud. are situated discretely over the cerebral context. He also proposed that menta l functions. the irrational comp on ent s of behavior. There is no doubt today that the brain is the organ of the mind.

such as positron emission tomography or PET. In addition. Gall was explicitly opposed to interfering with nature experimentally. and magnetic resonance imaging or MRI. In this regard. Even though an excellent neuro-anatomist. This clinical evidence for fimc­ tional localization is extended by the more modern stud­ ies of neurosurgeons who have tested function removing before parts of the cortex for intractable epilepsy or tumors. He used non­ experimental-what today would be called noninvasive-techniques.DISCUSSION dence includes the wealth of localization data from the so-called natural experiments observed by the great 19th­ century neurologists. of schoolmates with extreme character traits. another important aspect of Gall's influence bears directly on these questions. he reflected the romantic view expressed by William Wordsworth in his famous 74 . and examination of the cranial features of statesmen and criminals. that Gall decided where the specific cortical organs should be. this tradition is exemplified by phys� ognomy and by the description ofcharacter types. It was through observation. But the key question remains: How do you get a mental event from the brain? How do discrete "organs of the mind" work? What is the meaning of the "/" in "mind/brain"? As a matter of fact. continue to support localization. the work of experimenta l physiolo­ gists. In the study of personality and behavior. Gall was a psychologist. Gall developed his ideas in the long tradition of observa­ tion that could be traced to Aristotle. the modern techniques of brain imag­ ing. from David Ferrier's experiments at the end of the 19th century to experimental studies of the present day continue to provide convincing proof for localization of function. Finally.

I Close up these barren leaves. How different from the Age of Enlightenment. all of Marx' writings." Nonetheless. and social interaction." And: "Enough of science and of art. are based squarely on observation rather than on experiment. "We murder to dissect. Darwin's expression of emotions in man and animals (1872) and most of his other work. psychology is often thought of as a soft science. in his 1928 obituary notice for Farrier in the Royal Society. sacrificing the broad significance of function to a detailed description of physiological mechanisms. and all of Freud's work after his paper on hysteria written with Breuer (1893)." It is only a slight exaggeration to think that experimen­ tal neurophysiology. For example. but has not yet needed any direct physiological information. said that Farrier proved the concept of cerebral localization and provided the basis for a "scientific phrenology. Charles Sherrington . I Come forth. whose 75 . It has not yet needed any direct psychologi­ cal knowledge.DISCUSSION dictum. and Farrier who championed experimental neurophysiology. but it is curious that none of the acknowledged intellectual fathers of the 20th century -surely to be regarded as the age of technology-were experimentalists. Psychology analyzes brain function as related to character. and hurry with your heart I That watches and receives. the scientific discomfort that we are now feeling has its roots in this dialectic. In a very real sense. from Pierre Flourens's in the 1820s to Farrier's was developed to disprove Gall's psychology. personality. who represents the tradition of psychology. experimental neurophysiology analyzes decom­ posed parts of the brain. there is a sharp dissonance in approach to mind/brain between Gall. To the modern reductionist mind. In contrast.

Professor Chomsky. the primary example being Isaac Newton. [But] the very same innate factors that provide the richness and variety of mental life . maze. . conscious­ ness. . arena. and cognition may be mysteries that can never be understood by science. [and] we are fortunate to have this rich innate endowment .DISCUSSION fathers were experimentalists. (And perhaps vigor as well!) Can we be satisfied that the slash between mind/brain will be understood using a combination of the "new phrenology" and the modern theories of infor­ mation processing applied to the nervous system now called connectionism? In answer. Following the golden clue of the genetic material.ntial. Now.n't be a surprise that Plato's "Meno" is one of Professm Chom­ sky's favorite examples illustrating the innate st rue: I ure of 76 . or battlefiold for generating our actions using that potential. forum.* Our genetic maturation . . neuroscience has not yet attained an adequate synthesis of psychology's habit of observation and physi­ ology's experimental vigor-rather rigor. Clearly. leads us only to the structures that endow humans with pote. my reductionist scientific colleagues tend toward the philosophy of Eliminative Materialism and tho com­ puter logic of information processing. DNA. self-awareness. impose severe bounds on what the mind can achieve. As you have indicated. Professor Chomsky has written. It should. endowment provides for the growth and of special mental organs . neither by the methods of experimental physiology nor by observation. It is the natural world we live in that provides the playground.

the specific knowledge that Socrates shows to be innate is not of geometry. syntax. and. you seem to be like the stingray who stuns those who comes near him for my tongue is really paralyzed . and to many people. but at this moment I cannot even say what knowing is. "In your powers over others.DISCUSSION mind. Professor Chomsky. or phonology. Meno's comment to So­ crates seems to apply to you just as well. I have given an infinite variety of lectures about cognition before now. but rather the knowledge of the good. I thought they were pretty good. ." 77 . boy. Would it be unfair to ask whether your philosophy. Right before the philosopher demonstrates that knowledge is inborn. he says. semantics. In that dialogue. can be thought of as "eliminative mentalism" with the social logic akin to Plato's theory of the good? Professor Chomsky. .

in particular.CONCLUSION Professor Noam Chomsky: [Let me first respond to] Dr. I think the history of science is very useful in trying to figure out what is going on here about this. I think that the history of science teaches something radically different from what many contempo­ rary biologists and. Crick and Watson. I agree with him about one thing certainly. I don't think they're paying attention to the history of science. Schwartz' comment. there was one dramatic recent case of successful redu­ tionism. and so on succeeded in giving a reductionist account of large parts of biology in terms of relatively 79 . However. and Pauling. and I agree with him also about the problem of the "/" in the mind/brain. It's true that Crick and Watson. as far as I know. because they're kind of mesmerized with one extremely rare incident. neurophysiologists have drawn from it. Namely. I think it teaches the opposite of the conclusion they draw from it. In fact. about the only one: namely.

bu:t that's extremely rare in the history of science. · · And in fact-that was said: if you look b ack at 17th­ century debate. it's all in�onsistent with the mechanical philosophy. . And Newton hiniself-there was some self-ridicule-he was concerned. If we followed the prin­ ciple that we must have reductionism. it's unifi­ cation. The problem in science is not reductionism. let's just study the kind of thing we dogmatically assume to be true and not look at the phenomena we discover or the explanations we discover for them." he would have said: "Yes. It took a long time before people finally understood that that's the wrong approach. Well. there are no planetary orbits. "OK. Nothing moves on the earth. In fact. in fact. the classical moment. because.ber of very distinguished continentafscientists on essentially those grounds. Newton's results were ridiculed by a nuDJ. that's what the reductionists would have said. saying. But if he followed the position of which the equivalent today is eliminative materialism. he s aid as I quote d and as Huygens and . others agreed." Because he succeeded in demonstrating-that's what Volume II of the Principia is about-that you cannot explain planetary orbits or any other kind of terrestrial motion in terms of the mechanical philosophy.CONCLUSION known biochemistry. say. "Well. They work to 80 ." as he's shown. then Newton's conclusion would be. by it. planets don't have orbits. take the few cases that I mentioned. which is something quite different. Kepler's laws are false. Newton. There are different ways of looking at the world. which is self-evident. No person who is even sane can doubt that the mechanical philosophy is true-except that it's false! His conclusion-and he was left with this paradox. Take.

we throw it out. and the periodic table. there is no chemi­ cal bond. T he question is asked. well. Let's take now the slash between the mind and the 81 . we would like to integrate them. But how do we get p l anet ary orbits from mechanical philosophy? Well. fields. how do we get the chemical bond. How do we get them in terms of 19th-century physics? Answer: We don't. which weren't part of the earlier science. It's very commonly been the case that what we think of as more the funda­ mental science had to be radically revised. at least as I read it. We don t ' say. like that. let's say or the states of matter. How do we get electromagnetic phenomena from the motion of particles? The answer: We don't." Thue. no one knows much about the slash. but it doesn't exist. in fact. Therefore. most of the history of science is. which real ly weren't understood until quantum theory. a reductionist would say: "OK. except in rare cases. We've shown it doesn't exist because of the slash. we give up the mechanical philosophy. Well. and so on and so forth. we introduce new principles. Maxwell's equations. Throw out chemistry.CONCLUSION whatever extent they do. that has rarely been true. but reduction is only one way to integrate them. How do we get a mental event from the brain? Well. through the course of modern science. there is no magnetism because it can't be explained in terms of known physics. in fact. and valence. Or. And. Therefore. 19th-century physics is just incapable of dealing with it. it has not been achieved through strict reductionism. To the extent that unification has been achieved. In fact. the answer is: We don't. the chemists seem to be explaining all sorts of things in terms of these Kekule molecule models. namely. therefore.

say. and see ifyou can somehow dream up the mess. fifty years. and language. fact. It's perfectly the wrong what was called Prout's Law: Why do chemical el ement s seem to be roughly integral multiples of·the atomi c weight of hydrogen. It's entirely likely that the same is true in this case. In . and so on. 1They just had the wrong data. they didn't tell you a lot. and the way to overcome the sla s h is exactly the way it's always been: pursue the various approaches to understanding a way to 82 . If you look at the brain today you find the same thing was a plausible conclusion from the massive data they ." I'm virtually quoting� in fact. probably the reason for that is they just don't know the right things about the brain. there was a huge mass of information about l ang uage You go to the library. Linneaus . as was later discovered. linguists was ''languages cari differ . Well. By scientifi c standards. there was a huge discussion of tion and this and tha t They had tons of phenomena. Probably because you don't understand It lo o ks as -if nothing digi ta l and computational with these things. say. They achieve some pretty surprising things. that their conclusions don't mesh with what'they know ' about the brain. The standard view am ong you'd find big fat books about all kinds oflanguages. vision.CONCLUSION brain. ifyou go back.from one another in arbitrary ways. they just happened to facts . as the neurophysioiogists say. That s knew lots of facts. During the 19th century. And that had. at least. There is tremendous experimenta­ . That has often been true in the past. but they were the wrong phenomena. there are pretty successful computational theories of. not to say that they don't know a lot of facts. but weird properties that are discussed could possibly come out of this mess.

beyond the levelof big molecules. You don't know if you'll be able to do it. You say. For example. parti cu lar conditions of human evolution. and I think that people should change. well if something els� had happened that wasn't functional the organism wouldn't have reproduced and would have died off. It's striking �djust 83 . but there's virtu ally nothing of a theory there. but when you try to account for why particular organs develop. You can teach population genetics and Men del and so on. That space :might be extremely narrow. mind is vast. " . but something like it . and you get into hand waving. it might be so narrow that under the. .when you get above big molecules under._ standing tails off very quickly. There's plausible descriptive accounts of why snails get bigger shells and so on. about everything else in biology is also vast. I dori't propos e that. What Darwin achieved is of extraordinary importance. there isn't much in the way of theory: Let's take what's c a l le d the theory of evolution. but the explanatory force is limited. or species." To move to more far reaching explanation. you just don't under­ stand very mut::h. The problem is that. there's one possibility for something with 1011 neurons packed into something the size of a basketball: namely. or else just description. That's not be misled by the one really striking caseofreduction­ ism. and so on. you're going tobave to find something about the space of physical possibility within which s e lection operates. but the gap between molecules It's perfectly true that the· gap between molecules and to see how. is wave your hands. You don't kn o w what's going to have to been the course of science. a brain that has these computational properties. all you can do.CONCLUSION bring them together. There's nothing much to teach.

When you get b e yond p roteins understanding does tail off.Philosophers and scientists. Now. without much knowledge as far as I am aware.CONCLUSION could turn out to be true. Even terres. there is no way of answering that question. So they are real mathematical obj ect s Is that consistent with materialism? Every physi­ cist says it is. and nobody can tell you what matter is. that's true. Well. take fields. . and that they achieve binocu l ar vision at four months. . If you take p roblem s in embry­ ology. there can't be any such. among many . like why does a chicken limb develop just the way it does. you get a lot of very i nteres ting descriptive comment ary. . but since we have no concept of matter. that those thing s are determined by the genes s omehow but the gaps between those two facts and molecules is just as vast as the gap between generative grammar and molecules. Those topics are not well understood because so little is really known in any depth when you get beyond big molecules. 84 . that create a space within which reproductive success makes a difference. For example. thing as eliminative materialism. what about eliminative materialism? It's a com­ mon view. but very little in the way of gene ral theory. say D' ArcyThomp­ son and others talked about. There are two ways of loqking at eliminative materialism as far as I can see. Until somel:lody tells us what materialism is. Everybody assumes. those are facts. Just take the two examples I me ntioned : the fact that children undergo puberty at a certain age. There might be very narrow physical possibilities of the kind that. One is that it's total gibberish until somebody tells us what matter is. they are basically mathematical obj ects but physicists treat them as real because they push each other around as Roger Penrose put it. Well.

It might make about as much sense today as telling a neurophysi­ ologist: "Look. In fact. So." It wasn't going to work. look at quarks. Connectionism is a radical abstraction from what's known about the brain and the 85 . maybe they're implicated. The connectionist approaches that you mentioned are even ·stranger in my view. Wh�n people say the mental is the neurophysiological at a higher level. That makes about as much sense as saying to chemists 100 years ago: "You want to understand valence. it's just a kind of hope. You want to understand." Well. undergo­ ing puberty. Maybe there are other aspects of the brain that nobody has even dreamt of looking at yet. you want to understand neurophysiology. the belief that neurophysiology is even relevant to the functioning of the mind is just a hypothesis. within the mechanical philosophy. look at billiard balls bump­ ing into each other. we don't know what eliminative materialism is. The other way of looking at eliminative materialism is to observe the way it is actually practiced: just look at neurophysiology. That's often happened in the history of science. look around and you see neurons. That's one way of looking at it. We know a lot about the mental from a scientific point of view. We have explanatory theories that account for a lot of things. The belief that neurophysiology is implicated in these things could be true. maybe. say. We can't talk about it. Until someone comes along and tells us what matter is. they're being radically unscientific. but it doesn't help a lot.CONCLUSION trial motion is not consistent with what Newton thought matter was. look at neurophysiology. That's where the answer is. Who knows if we're looking at the right aspects of the brain at all. but we have very little evidence for it.

People know exac tly the developmental pattern. we can't be dogmatic about these things. It ·says often new uild�rstanding has come out in undreamt of ways. after all. the evidence for connectionist models is. which interact in all kinds of ways. but ta ke the studies of nematodes. which are useful organisms because their wiring diagram right.CONCLUSION brain sciences. I think Gerald Edelman is quite most trivial problems that have been addressed-like learning a few hundred words-have been total failures. but not thro ugh reduction. These are. In the case of language. through expansion if you want 86 . Here. but nob ody can connectionist models and they gave them up quickly because they just abstract too far away from the physical properties of the nervous system. There's n o reason to believe you're abstracting the right thing. for the moment. They have 300 ne urons and are very simple. The You can tell me about this bec ause you surely know is completely known.Zero. cellular structures. 'which has been trying to figure out why the stupid little worm does the thing it does. We know all of its neurology. about . not just the things abstracted in· synaptic Connections. It would have been crazy togive up itto known physics. That's a bout all it tells you. There's no evidence for it. They did try dogmatic. we didn't understand. be­ chemistry in the 19th century because you couldn't relate cause systems we thought we· understood. We know entirely about its developmental pattern. In fact. sciences had to be radica lly recast. There is a re s earc h group at MIT. Then you could bring tlwm together. It's a very radical abstraction. you had to change the known physics radically. The· history of science tells us you can't be figure out what the heck it's doing and why. more about itthan I do. My main point is. and the fundamental.

so. 'Heil . "Say. So I'd like to modify that and say we may not have enough capacity to achieve unification. therefore. That trivial fact is completely out of the range of any form of even bad science. If somebody came along and tr ained a machine gun on the audience and said.CONCLUSION to give it a name in fact. But those phenomena are as clear to us as any are. you just try to understand things better and maybe y ou ll learn more about the brain.Hitler. or peo pl e make pretenses about it. Can we learn what's a mystery for us? It's conceivable. There's no logical contradiction in supposing we might ' . "Heil Hitler. Nobody has the foggiest idea of what to make of that. thought he .'" and people. The only change I'd like to make is that ther e s no reason to expe ct that a reduc­ tionist account is the true one. my opinion The gap is real. You don t know how to overcome it. mayb e you l l l earn that you are looking at the wrong thi ng Pos s ibly the gap will be overcome or maybe we just never figure it out. changing everything. 87 . nobody pays attention to it. probably ever ybody would say. two ofthem. at least for now. That's quite p os sible A very normal phe­ nomenon of life is that our actions are under our control within the range of physical possibility. We could do something else. ' ' ' . ' . and we can t even ask the right questions about them That's p art of Descartes' point They seem to be out of tl1e range of our cognitive capacities. That leads to questions that George Miller raised. A bout not having enough capacity-! think what George said is that we may not have enough capacity-to achieve a reductionist account." but we know perfectly well that we don't have to. and maybe forever. but the way to overcome the gap is the same as in othe r cases. So the reductionist moment should not be taken as a model in . The action is under our control. meant it. .

about 95 percent of them we'd look with the same blank stare that they looked at them._I'lLgive you an example that's false of the kind of thing we might find. And the typical phenomenon is either we think of noth­ ing. If we take the problems raised by the Greeks. So maybe there's a way to study the bounds of our 88 . usually at about the same time.. Mostly they look with blank stares . You take a look at the history of science. some ideas come along. that's got to be right. but we could discover something like that. the transition from the problem situation to the ideas follows a very curious path.. Every once in a while. as a deterministic syste�. we know that there were infinitely many possible theories consistent with the facts in a problem situation.:. one might study problem situations and the ideas that people come up with. We would then know that any system of the world which doesn't fall within that bound is a mystery to us. and consistently. We cou�d discover something like that. Suppose that were what we discov­ ered. the same ideas come to the minds of many people. or as something which involves randomness. For example. or else everybody more or less thinks of the same thing or at least kind of recognizes it as plausible. nothing happens. or somebody happens to be first and the other people say. That indicates a high degree of structure in the cognitive system.CONCLUSION learn what our own cognitive capacities are.." Well. So. People look at certain problems. "Yeah. Suppose we discovered that every time we come up with a solution it's because we've been able to formulate it eithe. Just as a matter of set theory. And it is entirely possible that we could learn what that s tructure is. I'm not suggesting that's the right one. for example. there are plenty of problem situations.

and try to figure out what you know about that word. I have all been kind of grumpy about the course of cognitive science and have said that the sub­ ject's kidnapped. But let me put that a side. In the case of simple lexical meaning. take a look at our ages and think what people of our age tend to say about their children and the course that they're pursuing. we know very little. At the moment. Say. Now we don't have anything much in the way of theory about these things. artd you find that it is so radically beyond anything that's ever been described that you're just caught up in a big descriptive problem in the first place. Akeel's questions would take much more time to answer than I could if I had the time. here are an awful lot of things where we can't even think of the right questions. which I don't. all I think we can do is be descriptive. and has inherited all kinds of errors and. There is no theory to speak 89 . On the question of the lexicon and perspectives and accessibility to consciousness. The general conception of what it's all about in my opinion is badly wrong.CONCLUSION understanding. he and Jerry Bruner and. or knew what to answer. one could make an unkind comment: namely. like the study of vision and the study of language and so on. In my opinion at least. look. let alone the right answers. You take a look at a word like "London" or "table" or "house" or "dog" or something. what he says seems to me plausible-if I understood it-but I think there's another possibility we might think about. I think has not reached the level of the 17th century in many respects for reasons I mentioned. About George's other point that. in fact. some have been doing pretty reasonably. namely. Well. in the specific empirical areas of the so-called cognitive sciences.

and we all knowit just barely touches the surfaces. And we. do-we confirm or refute the Turing Test by considering the possibility of a machine that duplicates our finite behavior? Well. which. So there is an input-output system. it is all accessibleto consciousness. a topic that's rarely been studied. descr iptive commentary like most of evolutionary biol­ of about agent's perspectives. On the other matter. no. 90 . I'm not convinced. You can see when you think about even the trivial examples I mentioned. is the yourselves. it may well be beca use we just don't understand enough. Breathing is a thing that humans do. that we'd look at and see if it tea ches us anything about humans. Let's try an analog. This is. I think George Miller's work is some of the first study of it. we are just stuck with ogy. in my view. Roughly speaking what happens is air comes into the nose and carbon dioxide goes out after a lot of things go on. the machine would not be breathing fortrivial reasons.CONCLUSION right way to look at it. We could get a machine that duplicates that completely by so me crazy mechanism. What else we might discover that it is not accessible to us any more than the principles that are involved in thatslightly more complex sentence that I mentioned are accessible to us. That we don't know. If it does. As a just try ing to get a se n s e of what the descrip­ tive phenomena are for the first time. therefore. the ma­ chine isn'tbreathing. Would the machine be breathing? Well. We breathe. Where things seem accessible. If it doesn't teach us anything about humans. it's a good model of hu:mans. air to carbon dioxide. Is it a good model of hmnans? Well. send it to Hume's flames. When you have just descriptive commentary about could it be? If we had a theory about agent's perspectives.

think?" That's not the way we talk English. It's just like asking." the answer is no. the qu e stion whether a machine can think is too meaningless to deserve discus­ sion. look. J think Turing was right. 91 . a knife into somebody's heart. He said. Would it be playing chess? Well. the questions just Therefore. The same is true about this breathing device or about machines thinking and so on. It flies if you want to call that flying. If we asked. It's like asking in 1900 whether an airplane flies. any more than my legs take a walk. it doesn't seem to me possible to refute the Turing Test this way. made exactly the moves he would every time. Robots can't murder.CONCLUSION It seems to me exactly the same is true when we turn to thought and intelligence. It doesn't fly if you don't want to call that flying. Just for the same reason that a subma ri n e doesn't swim. A machine that duplicated the air-to-carbon diox­ ide exchange would not be breathing for trivial reasons. It's not an interesting point to discuss. It's a trivial point. my brain doesn't play chess or understand English. That's some­ thing humans do. Let's say somebody could come along with a chess-playing program that behaved exactly like Kasparov. My legs don't take a walk. no. but these are not substantive ques­ tions. "Does Kasparov's brain play chess. just as in the case of "breathing. but if you want to change the language you could say it. we could say they do. It's not a meaningful question. but his brain doesn't play chess. "Does my brain. not murdering him. Kasparov has a brain. If we want to extend the metaphor to submarines. For these reasons. Remember what Turing said." Playing chess is something that people do. English happened to pick a different metaphor. it's just as if a robot sticks don't mean anything. Swimming is something that fish do.

Simlllation that doesn't teach us anything is useless. good. and we'll talk about that as thinking as we talk about airplanes flying. scientifically. In my opinion. than a chess-playing program. For one thing because chess playing is not an interesting topic to 92 . John Searle's Chinese room and so on. or how do we empirically decide whether comput­ ers play chess. say. If it does.CONCLUSION What Turing suggested is. Turing's point was maybe this will teach us something about thinking. Well. But nothing substantive will have happened. That's like 250 years ago. but that tells you exactly where the field has gone off from the first moment. wliich as Herbert Simon once put it. These are not meaningful questions. computational models of intelligence. just the decision to use a metaphor. did the rock murder him? It's the same kind of question. like deciding to say that submarines set sail. if it doesn't. That's perfectly reasonable. There are few projects less interesting. and let's try to create models of intelligence. and we're not confused into thinking it." the idea around which everything converges. it seems to me just like asking: Does the brain think? Do my legs take a walk? If a rock fell off a roof and shattered someone's skull. He's sort of right descriptively. because maybe it will teach us something about ducks. throw it out. he also said that maybe 50 years from now we will have just changed our lan­ guage. We should drop them and just look at the serious questions like whether simulation teaches us anything. is the "'drosophila of cognitive science. It doesn't mean anything. de Vaucanson saying let's construct an automaton that does things kind of like a duck. let's drop the question of what thinking is. all the discussion that's gone on for the last ten years about. Take the whole business aboutchess"'-playing programs. I think.

Playing ches s is something way out on the margins of what people do-that's why it's a game. but Professor Chomsky has in timat e d as a hint. That just means it s rotten simulation. then mayb e someday we ll get to p ole ' vaulting. Anshen: I want also to thank the disc u s sants and. from the very first moment it became was to d evi at e radically from the way human beings do it and to use the capacities of computers. in my opinion as . right now. it's unlikely to help us learn anything about human beings. perhaps. Let's first fig ure out how the y move one leg in front of the other. Anshen for c onvening the meeting� Ruth.CONCLUSION study. If Carn egie Tech's computer program can beat Kasparov. ' clear that the way to win at chess that's ab out as interesti ng as the fact that a bulldozer can lift more than some weight lifter." That just wouldn't be a sane scientific endeavor. and someone s ai d . Noam Chomsky and Charle s Ryskamp. take the fun out of playing chess far as I ca n see. ab out its only interest is to . are a relevant example of the mystery 93 . Now the fact that a huge amount of effort and money from th e National Science Foundation-! ho pe not the Russell S age Foundation-has gone i nto this simply shows how conceptual errors have misled the field. I o n ly want to remind all of us that the poets have been ignored. We should be awa re of that. would you like to have the last word: Dr. the poets. Furthermore. Dr. Maybe. It's too remote from what we unders tand to make any sense to study. and it's ' of no scientific intere st In fact. . that they. Who c are s ? It doesn t teach you anything about the weight lifter. Warmer: Let us thank Dr. " L et s figure out how ' they pole vault. of course. It's as if we didn't understand how peop l e walk.

I think. Shall this mystery us? Unknown speaker: I hope not. Let us take the poet Rimbaud. 94 . consciousness. on me pe n se. reverence the mystery ever be revealedto of philosop hy. wisdom. is whether in art. Dr. of creativity. "C'est fam� de dire. Wanner: Thank you all very much for coming. it thinks me. some of you will understand it. of mind/brain/spirit. . on devrait dire. and language. music.CONCLUSION of the creative p rocess . " Now I translate: "It false to say. science. je pense. We must creativity. morals." This mystery of "it" is the sour c e in my opinion. even thought. one must say. I'll say it first in French.

the History of Science Society. and the recent volume Morals Equals Manners. and edits. founded. Credo Perspectives. Perspectives in Humanism. Bohr. PH. Chomsky. and Heisenberg to Rabi. She is also the author of Biography of an Idea.. her writings. Anshen's book Anatomy of Evil. her ability to attract a most important group of contributors to these series. Fellow of the Royal Society of Arts of London. and particularly through her close association with many of the great scientists and thinkers of this century. She has exerted remarkable influence through her worldwide lectures. the Science of Culture Series. Dr. Tillich. from Whitehead. several distinguished series of books. Einstein.BIOGRAPHICAL NOTES RuTH NANDA-ANSHEN.D. a study in the phenom­ enology of evil. the Tree of Life Series. including World Perspectives Religious Perspectives. plans. demonstrates the interrelationship between good and evil. and Wheeler. Dr. Anshen is a member of the American Philosophical Asso­ ciation. underlying the unitary principle of all reality. the International 95 . and Convergence.

New York City. Eric Wanner. Professor of Philosophy. Princeton University. Akeel Bilgrami. and the Metaphysical Society of America. George A. Science. James Schwartz. Columbia University. '• 96 . Columbia Uni­ versity. Miller. Professor of Psychology.BIOGRAPIDCAL NOTES Philosophical Society. Dr. and the Philosophy of Culture at the Frick Collection. Massachusetts Institute of Technology. Noam Chomsky. College of Physicians and Surgeons. Professor of Neuro-Biology. Founder of the Anshen Transdisciplinacy Lecture­ ships in Art. President of the Russell Sage Founda­ tion. Professor of Linguistics and Philosophy.

-] . but the temptation must be qualified: Chomsky's status derives solely from the free pursuit of knowledge and his only interest in "authority" is to challenge those aspects of received wisdom chat are intellectually or morally suspect.This brilliant essay attests co the aptness of Nro�swuk's identification of Professor Chomsky as one of the very few thinkers today whose genius can be compared to that of Freud. in his characteristic way. He is rewarding on first reading. ENGLAND Noam Chomsky.. here lights a flame char is illuminating in itself and that will ignite many other flames among his readers. yet immensely imponanr and creative. more rewarding on repeated readings. MARSHALL UNIVERSITY DEPARTMENT OF CLINICAL NEUROLOGY THE RADC LIFFE INFIRMARY. His "methodological naturalism. ii it avoids the dogmatisms that constantly tempt humankind. He demonstrates his remarkable gift both for reopening old questions roo often fotgotten O( dismissed and for exploring new ques­ tions. modest in the scope of its claims. Marx or Darwin. PRJNCETON UNIVERSITY · It is hard not to call Noam Chomsky the leading authority on the structure of language. The nuanced arguments put forward concerning the history and significance of the mind/brain juxtaposition reveal why. his ideas have helped transform the way we think about the nature of the world. Very Good ISBN 1-55921-076-1 ." carefully distinguished from metaphysical naturalism. j OHN C. Language and Thought (Anshe the Philosophy of cunure) Used. OXFORD. is bold in its ventures. He shows that human knowledge is a fragile bark in a myste­ rious sea of the unknown. -ROGER SHINN PROFES OR OF THEOL GY AND PHILOSOP!-JY UNION THEOLOGICAL SEMINARY Ill 111111111111111111111 111111111111 1111 X0002DJX45 . -DR. as with the three luminaries..jORGE I<LOR DE ALVA PROFESSOR OF ANTHROPOLOGY.

Sign up to vote on this title
UsefulNot useful