JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR

ON CHOMSKY'S REVIEW OF SKINNER'SVERBAL BEHAVIOR

KENNETH MACCORQUODALE1' 2

UNIVERSITY OF MINNESOTA

Skinner's book, Verbal Behavior, was pub-lished in 1957. Chomsky's review of it ap-peared in 1959. By the criterion of seminalinfluence in generating controversy and stimu-lating publication, both must be counted ma-jor successes, although the reputation andinfluence of the review are more widely ac-knowledged. It has been reprinted at leastthree times (The Bobbs-Merrill Reprint Seriesin the Social Sciences, No. A-34; Fodor andKatz, 1964; Jakobovits and Miron, 1967), andChomsky has recently written (in Jakobovitsand Miron, 1967, p. 142) that he would takeback little of it if he were rewriting it now.

Skinner's Verbal Behavior is an analysis ofspeech in terms of its "controlling relations"which include the speaker's current motiva-tional state, his current stimulus circum-stances, his past reinforcements, and his ge-netic constitution. Skinner has accepted theconstraints of natural science in his basic ana-lytical apparatus in that all of its terms areempirically definable. He intends to accountonly for the objective dimensions of verbal be-havior and to invoke only objective, nonmen-talistic and nonhypothetical entities to accountfor it. The notion of control, anathema to thepolitically oversensitive, means only "causa-tion" in its purely functional sense, and neednot alarm. It is not arguable nor criticizablethat behavior is an orderly, controlled datum,sensitive to the circumstances of the behaver;

1I am greatly indebted to Professor Stephen Winokurwho read an earlier version of this paper and mademany valuable suggestions.

2Preparation of this paper was supported in part bygrants to the University of Minnesota Center for Re-search in Human Learning from the National ScienceFoundation (GS-1761) and the National Institute ofChild Health and Human Development (HD-01136)and the Graduate School of the University of Minne-sota. Reprints may be obtained from the author, De-partment of Psychology, University of Minnesota, Elli-ott Hall, Minneapolis, Minnesota 55455.

this is simply a fact which has been amply con-firmed.Chomsky's review was, to put it mildly, dis-

pleased. It was also a virtuoso performancewhose echoes are still reverberating in psychol-ogy and whose dust has still not settled after10 years. It has two parts. The first is an ex-tended criticism of the basic analytical appa-ratus which Skinner brought to verbal behav-ior. So much occupies over one-half of thelengthy paper; the second part is a brief, actu-ally rather casual, criticism of the applicationitself, as if the demolition of the basic explana-tory apparatus had made serious discussion ofits relevance to verbal behavior superfluous.The fact that the review has never been sys-

tematically replied to (although partial replieshave appeared in Wiest, 1967 and Katahn andKoplin, 1968) has become the basis for an ap-parently wide-spread conclusion that it is infact unanswerable, and that its criticisms aretherefore essentially valid, a belief whichChomsky shares (Jakobovits and Miron, 1967,p. 142). There are, in truth, several sufficientreasons for the lack of rejoinder and none ofthem have anything to do with the merits ofeither Chomsky's or Skinner's case. First, be-cause not all S-R psychologists are sympatheticto Skinner's version many of them felt them-selves out of Chomsky's range and were notmoved to defend themselves or Skinner. Thisis somewhat ingenuous of them, however, sinceChomsky's actual target is only about one-halfSkinner, with the rest a mixture of odds andends of other behaviorisms and some otherfancies of vague origin. No behaviorist escapeduntouched. On the other hand, most Skinner-ians correctly concluded that their behavior-ism was not particularly the focus of the re-view, much of which they frankly did notunderstand. For example, the review devotessix utterly bewildering pages (Chomsky, 1959,pp. 39-44) to yet another refutation (they must

83

1970, 13, 83-99 NUMBER I (JANUARY)

KENNETH MacCORQUODALE

number now in the hundreds) of the drive-re-duction theory of reinforcement, which haslong since disappeared from everyone's behav-iorism, I believe, and which never character-ized Skinner's (Wiest, 1967, makes the sameobservation). Finally, and it must be said,probably the strongest reason why no one hasreplied to the review is its tone. It is ungener-ous to a fault; condescending, unforgiving,obtuse, and ill-humored. For example, theperfectly well-defined word "response" is con-sistently called a "notion" which creates, intime, an overwhelming atmosphere of dubietywith respect to the word. The review's onekind word is in a footnote (Chomsky, 1959,p. 32). It is almost impossible to reply to what-ever substantive points the review might havemade without at the same time sounding ei-ther defensive and apologetic, or as truculentas the reviewer. I have hesitated until now be-cause I am an editor for the Series in whichVerbal Behavior was published. Caveat lector.I believe that the review is, in fact, quite an-swerable. In spite of its length it is highly re-dundant; nearly all of Chomsky's seemingcornucopia of criticisms of Skinner's basic be-haviorism reduce in fact to only three, whichcan be addressed in finite, if necessarily some-what extended, space and time, and one canavoid the provocation to an ad hominem reply.This discussion will be organized about thesethree points, followed by a very brief commentconcerning Chomsky's criticisms of the appli-cation to verbal behavior per se.The reader should realize in advance that

there were and are no directly relevant factsto be brought to bear in this discussion. Al-though his thesis is empirical, Skinner's bookhas no experimental data involving the labo-ratory manipulation of verbal responses whichdefinitively demonstrate that the processes heinvokes to explain verbal behavior are in factinvolved in its production, although reinforce-ment has been shown to be effective in control-ling verbal responses (Baer and Sherman, 1964;Brigham and Sherman, 1968; Holz and Azrin,1966; Krasner, 1958; Lovaas, Berberich, Per-loff, and Schaeffer, 1966; Salzinger, 1959; Salz-inger, Feldman, Cowan, and Salzinger, 1965).Chomsky had no data to disprove the thesis ofVerbal Behavior, nor does he yet. This can besaid in the face of rather frequent statementssubsequent to the review which assert, for ex-ample, that "Chomsky's paper shows that ver-

bal behavior cannot be accounted for by Skin-ner's form of functional analysis (Fodor andKatz, 1964, p. 546)." Chomsky showed no suchthing; he merely asserted it. Chomsky's criti-cisms of Skinner are, then, necessarily method-ological. The disagreement is fundamentallyan epistemological one, a "paradigm clash" asKatahn and Koplin have put it (Katahn andKoplin, 1968). It is therefore most peculiarthat Chomsky nowhere refers to Skinner's ear-lier book, Science and Human Behavior (Skin-ner, 1953), the source to which Skinner specifi-cally sends the reader of Verbal Behavior forelaboration of general methodological matters(Skinner, 1957, pp. 11, 23, 130, 145, et seq.). Itmay be seen there, and in Cumulative Record(Skinner, 1959, 1961), that Skinner has neverbeen reticent about his methodological convic-tions nor vague as to his reasons for maintain-ing them. By omitting all reference to thesearguments Chomsky creates the highly erro-neous impression that Skinner has innocentlyand impulsively blundered along unmindfulof the difficulties inherent in what he wasdoing. This simply is not so. His application ofthe basic operant model to verbal behavior hasbeen evolving since 1934 (Skinner, 1957, vii).It has survived explication, and criticism byinformed but not universally convinced stu-dents, in the classroom intermittently sincethen, and in the William James Lectures atHarvard in 1947. The 1957 book is, then,hardly the result of a momentary enthusiasm.It deserves a more thoughtful hearing.

In what follows I shall consider Chomsky'sthree basic methodological criticisms in turnand compare each with what Skinner in factsaid. The reader should understand that theitalicized statements of Criticism are nowhereexplicit in Chomsky's review, which merelyadumbrates them.

Criticism 1: Verbal Behavior Is an UntestedHypothesis Which Has, Therefore, NoClaim upon Our Credibility

Neither Skinner nor Chomsky uses the word"hypothesis" to characterize Verbal Behavior,but it is one, in fact. Skinner avoids the wordbut is perfectly clear about what he is up to:"The emphasis [in Verbal Behavior] is uponan orderly arrangement of well-known facts, inaccordance with a formulation of behavior de-rived from an experimental analysis of a morerigorous sort. The present extension to verbal

84

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

behavior is thus an exercise in interpretationrather than a quantitative extrapolation ofrigorous experimental results (Skinner, 1957,p. 11)." And that, of course, is a hypothesis.The data to be accounted for are readily avail-able. As Skinner says: "The basic facts to beanalyzed [verbal behavior] are well known toevery educated person and do not need to besubstantiated statistically or experimentally atthe level of rigor here attempted (Skinner,1957, p. 11)." The explanatory apparatus heinvokes does indeed differ from that in mostpsychological hypotheses since it does not con-tain any fictional or hypothetical events ormechanisms, being composed instead of well-verified laws of behavior based upon observa-tion of non-verbal organisms emitting non-verbal responses. The hypothesis of VerbalBehavior is simply that the facts of verbal be-havior are in the domain of the facts fromwhich the system has been constructed. Skin-ner's stratagem is to find plausible referents inthe speech episode for the laws and terms inhis explanatory system: stimulus, response, re-inforcement, and motivation. The relevanceof these laws and their component variablesfor the verbal events is hypothesized only; it isnot dogmatically claimed (Chomsky, 1959,p. 43). The hypothesis may prove to be wrong,but our antecedent confidence in its correct-ness is at least enhanced by the fa-ct that thebasic laws which it invokes have become verysophisticated and impressively well-researched(see Honig, 1966). They have also been shownto be "surprisingly free of species restrictions.Recent work has shown that the methods canbe extended to human behavior without seri-ous modification (Skinner, 1957, p. 3)." Skin-ner also makes the cogent point elsewhere that"It would be rash to assert at this point thatthere is no essential difference between humanbehavior and the behavior of lower species;but until an attempt has been made to dealwith both in the same terms, it would beequally rash to assert that there is (Skinner,1953, p. 38)." Verbal Behavior is such an at-tempt for the case of speech.

Skinner's reasons for avoiding the word "hy-pothesis" in this connection can only beguessed. Psychologists readily confuse "hypoth-esis" with "hypothetical" in the sense of "fic-tional", and it is a strong point in Skinner'shypothesis that it contains no reference to fic-tional causal entities. All of the events, proc-

esses, and mechanisms invoked are themselvesempirical, and therefore the hypothesis con-taining them is in principle fully testable andpossibly disconfirmable. A more potent reasonfor his avoiding the word, however, is probablythat "hypothesis" has somewhat curiouslycome to imply the possibility of experimentaltest, which Skinner has not performed andwhich he does not seem to consider feasible,although Verbal Behavior is rich in observa-tional evidence. According to his hypothesisspeech is the product of the convergence ofmany concurrent and interacting variables inthe natural environment, which does not sus-tain the experimental separation and detectionof the relevant component variables. Yet any-thing less than concurrence and interaction ofmany variables would not, according to thehypothesis, generate speech. Skinner's situa-tion resembles that of the astronomer "ex-plaining" tides as the resultants of manyinteracting attractions. No one has ever ex-perimentally tested that hypothesis directly ei-ther, yet it is highly plausible and supportedby much observational evidence which is prob-ably the strongest conclusion we shall ever beable to make for it.Chomsky avoids the word "hypothesis" in

favor of more picturesque terms: "[Skinner]utilizes the experimental results [of laboratorystudies of infra-human, non-verbal behavior]as evidence for the scientific character of hissystem of behavior, and analogic guesses (for-mulated in terms of a metaphoric extension ofthe technical vocabulary of the laboratory) asevidence for its scope. This creates the illusionof a rigorous scientific theory with a very broadscope, although in fact the terms used in thedescription of real-life and of laboratory be-havior may be mere homonyms, with at mosta vague similarity of meaning ... with a literalreading (where the terms of the descriptivesystem have something like the technicalmeanings given in Skinner's definitions) thebook covers almost no aspect of linguistic be-havior, and . . . with a metaphoric reading, itis no more scientific than the traditional ap-proaches to this subject matter . . (Chomsky,1959, pp. 30-31. Italics added)." Which isreally only to say that the technical languageof Skinner's system is used in a hypothesisabout verbal behavior; all scientific terms inuntested hypotheses are necessarily "meta-phoric extensions" and "analogic guesses".

85-

KENNETH MacCORQUODALE

What is puzzling, therefore, is the pejorativeaspect which "metaphor" and "analogic" as-sume in the passage quoted.Even more puzzling is the giddy speed with

which the argument moves from its insightthat the terms in the hypothesis are for nowmetaphoric and analogic, proceeds to the pos-sibility that this may prove to be all they are,and concludes flatly with the verdict that thetechnical terms used do not describe verbalbehavior. This goes too fast! That remains tobe seen. Until the hypothesis is tested the lit-eral (non-metaphoric, non-analogic) applica-bility of its explanatory terms remains indoubt, at worst. Chomsky's only real argumentfor his conclusion that the terms of the theorydo not in fact apply to verbal behavior is givenin the quotation above. It depends upon theamazing possibility that "real-life" and labora-tory behavior may be different, as if somehownature maintains two sets of natural laws, onefor laboratories and the other for the rest ofthe world so that any law observed in the labo-ratory is prima facie suspect when applied toevents outside. Entrancing though this idea is,it seems unparsimonious to suppose it. Thatreally does not sound like nature.The fact is simply that we do not yet know

if verbal behavior is within the domain ofSkinner's system and whether the technicalterms stimulus, response, reinforcement are lit-erally applicable to verbal behavior and cor-rectly parse it into its functional parts ofspeech.Chomsky raises special considerations for

doubting that each particular term of the basictheory applies to the verbal case. These will bebriefly noted.The stimulus: Chomsky holds Skinner se-

verely accountable for hypothesizing certainstimulus-response relations in Verbal Behav-ior, such as "a piece of music" as a stimulus forthe response "Mozart", or a certain paintingfor "Dutch", and a red chair for "red" or"chair". "Since properties are free for the ask-ing, we can account for a wide class of re-sponses in terms of Skinnerian functional anal-ysis by identifying the 'controlling stimuli.'But the word 'stimulus' has lost all objectivityin this usage." He then goes on to say: "Stim-uli are no longer part of the outside physicalworld; they are driven back into the organism(Chomsky, 1959, p. 32)." This is a non sequi-tur. Stimuli are "free for the asking" only in

hypotheses. Their quid pro quo is payable inempirical demonstrations of the evokingpower of the putative stimuli. None of the pur-ported stimuli listed above seems outrageouslyimprobable for those responses, and not untilsuch an empirical test of their evocative con-trol has failed is anyone entitled to concludethat these are not stimuli for those responses.Chomsky's conclusion that a putative stimulushas lost its objectivity because it occurs in ahypothesis is merely muddled. Skinner did nothypothesize a (hypothetical) stimulus. Thestimulus is as real as ever. He hypothesizedthat there is a controlling relation between thereal stimulus and the real response. As for hisconclusion that the stimulus in a hypothesizedstimulus-response relation has somehow been"driven back into the organism", the rationaleis harder to reconstruct. Reading Chomsky onthe subject of the stimulus here and elsewherein his review arouses a growing suspicion thathe imagines that by naming one stimulus for averbal response we name its only stimulus, andthat one stimulus somehow preempts a re-sponse. He criticizes Skinner's characterizationof the responses "Eisenhower" and "Moscow"as proper names, controlled by the man or thecity, because one frequently says "Eisenhower"and "Moscow" when the man and the city arenot present (Chomsky, 1959, p. 32). Indeed onedoes, but this only shows, as Verbal Behaviorrepeatedly and clearly insists, that a verbal re-sponse may be controlled by different stimulion different occasions. Verbal behavior doesnot obey any "one response-one stimulus" ruleand it makes no sense to speak of the stimulusfor anything. "Eisenhower" and "Moscow" aresaid for many reasons, among which are thepresence of the man and the city. PerhapsChomsky's conclusion that Skinner's stimulifor verbal responses have receded into themind of the speaker is based upon this point:if I say "Eisenhower" when there is no Eisen-hower then he must be in my mind. Is that thedifficulty? Only if one is misguidedly deter-mined to preserve Eisenhower as the only stim-ulus for Eisenhower. It is really impossible tobe sure. However clear it is in its conclusions,the review is not much help on matters ofrationale.

Reinforcement. Inevitably Chomsky findsSkinner's functional definition of a reinforcerunsatisfactory (that it increases the strength ofany operant which precedes it), saying that it

86

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

is "perfectly useless . . . in the discussion ofreal-life [sic] behavior, unless we can somehowcharacterize the stimuli which are reinforcing... (Chomsky, 1959, p. 36)." He is complainingbecause reinforcers can only be postdictedfrom the fact of reinforcement, since they can-

not be "characterized" in terms of any univer-sal, independently knowable correlated prop-

erty such as drive-reducing power. Many psy-

chologists share this dissatisfaction. But thefault, if any, is in nature, not in our theories.Reinforcers seem in fact to have only one uni-versal property: they reinforce, and no amountof dissatisfaction will either add a correlatedproperty nor disprove the fact that they doreinforce.

To be quite correct, whether a specific stim-ulus will be reinforcing for the behavior of anyspecific organism can be predicted withoutactually trying it. That is, reinforcers can bepredicted, since all reinforcers are either spe-cies-characteristic (the unconditioned reinforc-ers) or they have, in the history of the behaver,been paired with an unconditioned reinforcer(the conditioned reinforcers). Both of theseclasses are knowable before any behavioral testof their effect upon behavior is made (al-though it is technically infeasible to enumer-

ate the members of the second class in thehuman case.) Furthermore, as Premack's datahave shown, all reinforcing stimuli are at leastpartially transituational; they will reinforceany operant whose initial probability is lessthan the consummatory or preconsummatorybehavior which the reinforcing stimulus itselfoccasions. Therefore, a prediction of future re-

inforcing effect must be made given a fact ofpast reinforcing effect for any stimulus as wellas information concerning the momentaryprobabilities of the operant to be reinforcedand the behavior occasioned by the reinforcer.These considerations, in addition to providingbases for prediction as to which stimuli willreinforce which responses, also act as con-

straints upon the illicit invocation of ad hocreinforcers. Together they remove the conceptof reinforcement from "perfect uselessness".Reinforcement is a real and powerful behav-ioral influence. Its inclusion in a theory ofverbal behavior is decided on the basis of itsown claim; it becomes a necessity whether it is''useful" in analyzing an instance of casualconversation or not.Chomsky seems convinced that Skinner

claims that "slow and careful" reinforcementapplied with "meticulous care" is necessary forthe acquisition and maintenance of verbal be-havior (Chomsky, 1959, pp. 39, 42 [twice], 43).Chomsky does not cite Verbal Behavior in thiscontext, and the fact is that Skinner does notsay or imply that the reinforcement for verbalbehavior must be carefully arranged or thatdifferential reinforcement must be "careful",applied with "meticulous care", and "slow andcareful" (Chomsky, 1959, p. 42). The idea ispreposterous and the implication that Skinnersaid it is both careless and false.

Skinner does not, in fact, explicitly claimthat any reinforcement is necessary for verbalbehavior, although Chomsky supposes he does(Chomsky, 1959, pp. 36, 37, 38). His referencesare to statements in Verbal Behavior which sayno such thing, and to Miller and Dollard(1941), who may. Skinner does claim that rein-forcement is a potent influence upon verbalbehavior, and, in fairness, he specifies no otherstrengthening operation for it. Nothing what-ever is at stake in excluding from the hypoth-esis such alternative response-strengtheningmechanisms as learning by imitation or bylatent (non-reinforced) learning, if theseshould become demonstrable. The systemwould not then be destroyed or disproved; itwould simply be supplemented by laws whichspecify the conditions under which these proc-esses occur. Chomsky suggests that it is well-known that much language learning in chil-dren proceeds by imitation (Chomsky, 1959,p. 43). So, in fact, does Skinner (1957, pp. 55-65) but he further specifies that the imitativerepertoire (which he calls echoic in the ver-bal case) is itself a product of reinforcement.The evidence for an innate imitative tendencyis very weak, so that the problem as Skinnersaw it was to explain echoism when it does oc-cur, and to account for the facts that the imita-tive tendency gradually restricts itself to thesmall segment of the vocal spectrum which theparent language uses, that its flexibility dis-appears with age, and that the echoic reper-toire contains quite different dimensions indifferent speech communities (such as pitch insome, and not in others). These are all con-sistent with a reinforcement interpretation ofthe echoic's origins.As for latent (unreinforced) learning, it is

certainly incorrect to conclude that "Few in-vestigators still doubt the existence of the phe-

87

KENNETH MacCORQUODALE

nomenon (Chomsky, 1959, p. 39)." The manystudies which Chomsky cites in support of theexistence of latent learning revealed mostlythat the methodological problems involved ina crucial experiment on that question areoverwhelming. The matter was not resolved.It was dropped.

Probability. Chomsky criticizes Skinner's"'extrapolation' of the notion [sic] of prob-ability" as being, "in effect, nothing more thana decision to use the word 'probability' (Chom-sky, 1959, p. 35)." This is the same objectionthat has been made to "stimulus" and "rein-forcer", i.e., the word occurs in a hypothesis,and therefore we need not reconstruct the ar-gument on either side. Chomsky says, in addi-tion, that "The term 'probability' has somerather obscure meaning for Skinner in thisbook (Chomsky, 1959, p. 34)." Small wonder,since he cites (Chomsky, 1959, pp. 29, 34)Hull's definition of probability (resistance toextinction) as Skinner's basic indicator ofprobability or "strength" rather than Skin-ner's, which is simply the likelihood of occur-rence of a response, measured as a rate wherepossible, but as a relative frequency in anycase. Skinner thus defines probability quite asany other natural scientist does. Much moreominously for Skinner's purposes, Chomskyseems not to grasp the difference between theoverall probability of occurrence of an item ina speaker's verbal repertoire, which is the fre-quency with which it occurs in his speech overtime without regard to his momentary circum-stances, and the momentary probability of agiven response in some specified set of circum-stances. (See, for example, Chomsky, 1959,p. 34.) The two probabilities are very different.The overall probability that any speaker willsay, for example, "mulct", is very low; it occursrarely in comparison with such responses as"the" or "of". The probability that he will say"mulct" may become momentarily extremelyhigh, as when he sees the printed word. Of thetwo, overall probability is a typically linguisticconcern, while momentary probability shiftsare, in a sense, the very heart of the psycholo-gists' problem, since they reflect the relationbetween speech and its controlling variables.Under what conditions does an organism speakan item from his repertoire? Simply knowingthe repertoire tells us precisely nothing aboutthat. If Chomsky really did not, in fact, see thisdifference it is impossible to imagine what the

rest of Verbal Behavior could have meant tohim, and no wonder that he regarded it withsuch astonishment and dismay.

Verbal behavior's momentary probabilitiesare difficult to assess in practice because themost sensitive experimental indicator in non-verbal research, rate, is not useful: strong ver-bal responses are not normally repeated sev-eral times. Skinner mentions some productioneffects which on occasion may reflect thestrength of a non-repeated, single utterance,such as loudness, speed of production, or repe-tition if it does occur. Skinner says quickly andexplicitly of these, however, that they are un-trustworthy: "It is easy to overestimate thesignificance of these indicators (Skinner, 1959,p. 25; additional warnings are given on pp. 27and 141)." It is somewhat shocking, therefore,that in spite of Skinner's disclaimers Chomskyimputes to his hypothesis an entailment that astrong response must be "shrieked (Chomsky,1959, p. 35)" or shouted "frequently and in ahigh-pitched voice (Chomsky, 1959, p. 52)."So much for the emigration of the system

out of the laboratory. Chomsky faults the argu-ment because it did so.

Criticism 2: Skinner's Technical Terms AreMere Paraphrases for More TraditionalTreatments of Verbal BehaviorThis point is at very high strength in Chom-

sky's review. It is made in turn for Skinner'sterms "stimulus" (Chomsky, 1959, pp. 32, 33,48, 50), for "deprivation" (Chomsky, 1959,pp. 46, 47), "reinforcement" (Chomsky, 1959,p. 38) and for "probability" (Chomsky, 1959,p. 35).As Chomsky's criticisms somewhat tend to,

this one has several quite independent facets.First is a sort of premise that the technicalSkinnerian vocabulary simply renames an oldnotion in a new but more prestigious way. Ibelieve this is obviously quite false. Second isa conclusion that, being a paraphrase, the tech-nical term is therefore no more objective thanits traditional counterpart, which I believe isneither a consequence of the first premise norcorrect. Both notions thread through the fol-lowing sentences from the review, though theymay also be found in other examples: "Hisanalysis is fundamentally the same as the tra-ditional one, though much less carefullyphrased. In particular, it differs only by indis-criminate paraphrase of such notions as deno-

88

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

tation (reference) and connotation (meaning),which have been kept clearly apart in tradi-tional formulations, in terms of the vague con-cept 'stimulus control' (Chomsky, 1959, p.48)." Let us see what can be done.Although Skinner did not do so, it probably

would be a service if a scientific and technicalparaphrase were given for such traditionalmentalisms as "refer", "denote", "meaning","wanting", "liking", and so forth, each ofwhich Chomsky says Skinner has in factpacked into one or another of his technicalterms. To do so, one would start with a tra(li-tional term, "refer", for example, and give afunctional account of the conditions whichcontrol its occurrences (roughly, its "use"). Ifone were to do so, however, he would quicklydiscover that not all instances of what we in-discriminately call "referring" involve any-thing like the same functional controlling rela-tions, and one therefore can find no consistentparaphrase among the terms in a functionalaccount for the reference notion. As we haveseen, one may "refer" to Eisenhower whereverhe is in relation to the speaker, but while thereference relation between the response andthe rnan remains thus constant, the control ofthe response may vary among such stimuli asthe physical presence of the man himself, orhis picture, or his printed name, or his namespoken by another, or some other verbal stimu-lus such as "Ike" or "Mamie's husband". Inonly one instance is the controlling stimulusfor the response also the person referred to.The remaining stimuli control the responsebut the response does not refer to them.

Reference and stimulation also differ dia-metrically in their direction of influence: astimulus acts from the environment upon thespeaker to control his verbal behavior, whilein reference the speaker's response acts uponthe environment to single out its stimuluscomponents. An analogy is with the ancienttheory of vision which supposed that vaporsemanate from the eye to contact the environ-ment as in the dynamics of reference, as op-pose(l to the mo(lern view that objects are seenwhen light from them controls the eye, as inthe stimulus role. To complete the full catalogof nonequivalence, we must only note thatmany verbal responses which are controlled bystimuli have no referents whatsoever (try "Ohdamn!") and also that the concept of "stimuluscontrol" involves causality, which is not in-

volved in "reference". Reference is simply arelation between the world and an item in thelanguage (as opposed to an item in a speaker'sactual behavior, the distinction which persistsin eluding Chomsky).

In brief, no one technical term in Skinner'scausal verbal analysis covers all instances ofreference (nor was any intended to), and theone term stimulus control covers much that isnot reference in the traditional sense. Thesame argument can be made for the nonequiv-alence of the other terms, deprivation, rein-forcement, and probability, to other more tra-ditional terms; if these are simply paraphrases,they do not map unequivocally and isomorphi-cally, term for term, into each other. Curi-ously, Chomsky seems to sense this too, and socriticizes the behavioristic paraphrase for blur-ring the traditional conceptsl Given all this, itseems quite obvious that the term "para-phrase" is simply inappropriate in this con-text. Skinner's analysis is no more a paraphraseof linguistic-philosophical mentalisms thanmodern physics is a paraphrase of pantheism.They merely converge, but from quite differ-ent directions and with quite different creden-tials, upon some aspects of the same domains.Whether it is a "mere" paraphrase of the

traditional account or not, Skinner's analysisis far more objective and less vague than thetraditional one and therefore scientifically pref-erable. Every term in Skinner's account namessome real thing which must be physically in-volved and locatable in any verbal event forwhich it is invoked. That is objectivity. If inhis hypothesis Skinner invokes a particularstimulus to account for the occurrence of aresponse, he is saying that at least some of theoccurrences of that response are due to thephysical presence of that particular stimulus.The discovery that the response also occurs atother times does not disprove the facts of stim-ulus control; it simply means that other con-trolling variables (usually themselves otherstimuli) must be discovered for these occur-rences. The notion of control as a relation isitself perfectly objective. To make a similarclaim for the objectivity of such terms as "ref-erence" (and "wishing", "wanting", "liking",and so forth) would presuppose first definingthem in terms of some physical dimensions.But that would at once be another "mere par-aphrase" of these terms in which, if we followChomsky, instead of the mentalism's gaininig

89

9KENNVETH MacCORQUODALE

in objectivity, the defining physical dimensionsare doomed, by some logical alcheniy, to losetheir objectivity. This is an odd thesis.This particular criticism in Chomsky's re-

view occupies a great deal of its total space andaccounts for much of its apparent thrust andits most vivid writing. It is not often that areviewer becomes so overwrought as to permithimself to characterize his author as entertain-ing "a serious delusion (Chomsky, 1959, p.38)." But there it is.

Criticism 3: Speech Is Complex BehaviorWhose Understanding and ExplanationRequire a Complex, Mediational,Neurological-Genetic TheoryChomsky expresses his surprise at "the par-

ticular limitations [Skinner] has imposed onthe way in which the observables of behaviorare to be studied, and, above all, the particu-larly simple nature of the 'function' [sic]which, he claims, describes the causation ofbehavior (Chomsky, 1959, p. 27)." Skinner'sbasic explanatory system is indeed simple incomparison to the complexity of the domainit is intended to cover. It is not so simple as toreduce to a single function however; it hasmany variables and at least as many functions.It is customary in scientific analysis to reducecomplex phenomena to their component proc-esses, each of which appears simple when de-fined in isolation by means of the controltechniques of the laboratory. In the naturalenvironment (curiously called "real" life byChomsky) the components recombine and in-teract to generate properties that none alonefully accounts for. According to the hypothesisin Verbal Behavior, one such quasi-emergentproperty is grammar, of which more later. Suf-fice to say that a theory of verbal behavior thatdoes not have special grammar-generating lawsin it may still be capable of generating out-comes which have grammatical properties.The general relationship of the domain ofverbal behavior to general behavior laws isreductionistic; the complex is explained interms of the simple. A system of simple lawswhich can generate complex outcomes is saidto have scientific elegance. As we move fromnonverbal to verbal behavior it is more parsi-monious to suppose that nature has not givenus a whole new set of behavior laws for justthis one aspect of behavior; not even geneticmutations account for that much invention.

This supposition that the laws of behavior arethus general enough to account for the verbalcase is not a claim that they are sufficient; it isa working assumption that they will prove'tobe.

It is a curious omission from Chomsky's re-view, considering that he wonders explicitlyat the simplicity of Skinner's account, that henowhere mentions the possibility that thesimple laws which the account contains mayact concurrently, and so interact as to modifyeach other's effects, converging upon a singleitem of verbal behavior, to make it somethingwhich is controlled or contributed by no oneof them alone. The omission is all the morepeculiar when one discovers that an entiresection of Verbal Behavior is devoted to elab-orating this possibility (Skinner, 1959, PartIII: pp. 227-309). References to the possibili-ties for special effects due to multiple causa-tion begin appearing in Verbal Behavior asearly as page 42, and follow frequentlythroughout the rest of the book. A carefulreading of the whole book shows that whenthe whole system is given its due, it is not atall limited to accounting for only simple be-havior.

In no area of psychology is the contrastbetween "simplemindedness" and "muddle-headedness" more poignant and clear than inthe case of verbal behavior. The S-R psycholo-gist is indeed at the simpleminded end ofthings, supposing as he does that verbal behav-ior can be reduced to its component processes,that these will be simpler functions thanspeech is, and that they will be familiar. Thisis the psychology of the nothing-but. In fact, ifhis analysis does not reveal simple and uni-vocal relationships in a new domain, the S-Rpsychologist tends to suspect that he has speci-fied the wrong variables on the input side orthe wrong dimensions on the output side, andhe will try again elsewhere. The alternative tosimplemindedness is muddleheadedness, whichfinds it inconceivable that complexity may becomposed of simplicities, and writes off thepossibility of simple explanations as "trivial","very unilluminating" or "not interesting",wanting a theory composed of something-more, and certain that it must be needed. Thehistory of science is probably on the side ofsimplemindedness. In the case of verbal behav-ior at present it is the disposition of ignorancewhich is at issue, as it may always be in a sim-

90

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

pleminded-muddleheaded confrontation. Skin-ner hypothesizes that speech will prove to belike other operant behavior when we under-stand it, and can be decomposed into compo-nent processes. Chomsky finds in its unana-lyzed mysteries a justification for presumingcausal innovation and complexity. He says:"In the present state of our knowledge, wemust attribute an overwhelming influence onactual behavior to ill-defined factors of atten-tion, set, volition and caprice (Chomsky, 1959,p. 30; italics added)." This is a very remark-able statement. The reader is encouraged tocontemplate it as a rationale for theory con-struction; here it is a substitute for knowledge.

Mediational TermsSkinner's laws are called functional because

they describe direct relations between each ofthe several controlling variables (evokingstimuli, reinforcement, motivational states)and momentary probabilities of different be-haviors in an individual's repertoire. That isto say, he does not invoke other events, proc-esses, or mechanisms which are hypothesizedor invented for the purpose of mediating be-tween behavior and its empirical determi-nants. This omission is sometimes miscon-strued as a denial that mediating mechanismsexist; they obviously do, they are obviouslyneurological and they are also obviously them-selves lawful (see Skinner, 1953, p. 28; 1957,p. 435). Because they are themselves lawful,these mediating events, processes, and mecha-nisms generate and maintain lawful functionalcovariations between the controlling variablesof molar behaviorism and the behavior theycontrol. The argument is simplicity itself, andSkinner has made it abundantly available (seeespecially Skinner, 1959, 1961, pp. 39-69). Heconsiders such theoretical terms unnecessary;they may generate research whose only useful-ness is to disconfirm the mediating entity orredefine it without increasing our knowledgeof behavior's controlling variables; they canbecome the absorbing focus of an inquiry andso deflect attention from behavior itself; andthey can become a "refuge from the data", asmotivation has tended to be in psychology. Itis often simply "what varies so as to accountfor otherwise unaccounted-for variability inbehavior."Chomsky does not refer to Skinner's discus-

sions of why he omits mediational constructs,

but he was apparently little affected by them,finding the violation of his own preconcep-tions a sufficient reason for ignoring them:"One would naturally expect that the predic-tion of the behavior of a complex organism (ormachine) would require, in addition to infor-mation about external stimulation, knowledgeof the internal structure of the organism, theways in which it processes information and or-ganizes its own behavior" (Chomsky, 1959, p.27). Perlhaps one would, but he need not. It isperfectly feasible and sufficient to know merelythat the speaker's "internal structure. .. proc-esses information" so as to generate lawfulrelations between the speaker's circumstances(past and present) and his speech. Unless oneis a neurophysiologist it is not necessary in theleast to know how the internal structure goesabout doing so nor which structures are in-volved. The psychologist's knowing how itdoes so would not improve the precision ofpredicting behavior from knowledge of thespeaker's circumstances, nor would this knowl-edge make existing functional laws of behaviorany more true, nor could it show them to beuntrue. It is simply false, of course, that onecannot accurately predict behavior, even com-plex behavior, without knowing and takinginto account the behaver's structure and inter-nal processes; we do it all the time. In point offact, our current knowledge of the functionallaws of behavior far precedes and outweighsboth our knowledge of, and even our theoriesabout, the mediating mechanisms involved.For example, so far as I can tell, the behavioralfacts of reinforcement are by now so well-known and dependable that theories of thedetails of its mediation are no longer of greatinterest. Where interest in the mediating struc-tures survives, it is behavioral data which il-luminate them, not the other way around.

Neurological-Genetic MediatorsAlthough Chomsky locates the missing me-

diator variously, now, in the organism's inter-nal structure, now in some prebehavioral proc-essing and organizing activity, or, sometimes,in deeper grammatical processes, it is clearfrom his most detailed examples that he in-tends to locate them precisely in the brain,and moreover that he supposed they got therein large part by genetic predetermination orpreprogramming.

91

KENNETH MacCORQUODALE

So far as I can see, he is almost certainly cor-rect on both counts, neither of which has theslightest relevance to the question of the valid-ity of Skinner's hypothesis, although they ap-parently have some crucial significance toChomsky however elusive it proves to be whenone tries to characterize it.With regard to neurological mediators in

general he says: "anyone who sets himself theproblem of analyzing the causation of behav-ior will (in the absence of independent neuro-physiological evidence) concern himself withthe only data available, namely the record ofinputs to the organism and the organism'spresent response, and will try to describe thefunction [sic] specifying the response in termsof the history of inputs (Chomsky, 1959, p.27)." The psychologist finds himself standinghere in a strange light, making-do with "theonly data available", behavior, but really onlymarking time until neurology can catch upand give him all of the real explanations ofbehavior. Aside from its condescension, thefacts and the logic in Chomsky's statement areboth wrong. The facts are that we are notmerely trying to "specify" behavior in termsof its past history and current cirmumstances(the "input" referred to), we are doing so, andwith increasing accuracy. The (functional) lawof reinforcement is an enormously powerfulpredictive (specifying?) device. At least it isfor nonverbal behavior; and no one can saythat it is not powerful for verbal behavior too.The logic, as we saw in the previous section,is that a valid functional law can be com-pletely established on the basis of the onlydata available, and does not need "indepen-dent neurophysiological evidence". The func-tional law of reinforcement, in addition tobeing powerful, is an established empiricalfact. It is not a theory awaiting neurologicalvalidation.The possibility that certain aspects of verbal

behavior may be genetically predeterminedseems to be loaded with special significancefor Chomsky. He appears to draw at least twoconclusions from the possibility; one is that ifthe brain is in fact genetically preprogrammedfor such behavior, it becomes all the more ob-vious that the structure of the brain must be"considered" in the explanation of that be-havior. The second is that the fact of geneticpredetermination is incompatible with thefacts of reinforcement. Are those valid infer-

ences? Consider the first. Recall that Skinnerexplained imitative verbal behavior as theproduct of reinforcement for "echoic" re-sponses. Chomsky says of this ". . . however, itis possible that ability to select out of the com-plex auditory input those features that arephonologically relevant may develop largelyindependently of reinforcement, through ge-netically determined maturation." He thengoes on to say: "To the extent that this istrue, an account of the development and cau-sation of behavior that fails to consider thestructure of the organism will provide no un-derstanding of the real processes involved(Chomsky, 1959, p. 44)." One hardly knowswhere to begin. First, it is not necessary to"consider the structure of the organism" inpsychological laws no matter how the braingot programmed. There is nothing uniqueabout the logical status of a genetically pro-grammed mediator. So long as the brain isprogrammed it will maintain lawful covaria-tions between "phonologically relevant stim-uli" and echoic behavior, and a functional lawreferring only to such stimuli and behaviorcan be written without reference to the brainand its "program". Second, if genetic prepro-gramming is a characteristic of the "real proc-esses involved" in echoic responding, that factwill be revealed through "considering" thebehavior of the organism, not its structure,and will appear normally as a reinforcementparameter. Genetically determined behavior iswhat does not have to be learned. Although atone level the brain explains behavior, in thetactics of scientific discovery it is behavior thatexplains the brain. And, yet once more, fromthe psychologist's point of view the "real"processes involved in echoics are the presenta-tion of a "phonologically relevant stimulus"and the occurrence of an echoic response.Chomsky's second neurological example ap-

pears to say simply that grammatical behaviormay be similarly preprogrammed. Its importis more complex than that, however, sincesome experience is obviously necessary forgrammatical behavior in addition to the ge-netic head start. The example is doubly im-portant because it seems to have been takenvery seriously by many psycholinguists (seeespecially Lenneberg, 1964, 1967). Chomskysaid: "As long as we are speculating, we mayconsider the possibility that the brain hasevolved to the point where, given an input of

92

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

observed Chinese sentences, it produces (byan 'induction' of apparently fantastic com-plexity and suddenness) the 'rules' of Chinesegrammar, and given an input of observedEnglish sentences, it produces (by, perhaps,exactly the same process of induction) therules of English grammar; or that given anobserved application of a term to certain in-stances it automatically predicts the extensionto a class of complexly related instances. Ifclearly recognized as such, this speculation isneither unreasonable nor fantastic . . . (Chom-sky, 1969, p. 44)." Nor, alas, it is particularlyrelevant. As we noted in discussing echoic orimitative verbal behavior, the mere fact thatthe brain has evolved does not force its intro-duction as a mediator into a functional law.

Neither does the fact that the brain hasevolved tell us anything useful about how it"produces" the "rules" of grammar. Whateverthat may mean, exactly, there can be no doubtthat the human brain has evolved to the pointwhere it has the capacity to mediate the acqui-sition of grammatical behavior. This says noth-ing in itself as to whether the acquisition proc-ess involves the sort of learning by imitationor observation supposed in the example or, in-stead, learning by reinforcement. The capacityto learn by imitation or observation is cer-tainly not a peculiarly or uniquely diagnosticsymptom of evolutionary advance and, as wehave seen, the possibility of some acquisitionprocess other than reinforcement is not itselfoverwhelming to Skinner's system.There is no lethal incompatibility or even

mild inconsistency between the principles ofgenetic evolution and the principle of rein-forcement. Reinforcement has many necessarypoints of contact with genetics. Reinforceabil-ity is itself a genetically determined character-istic; organisms are simply born reinforceable.They have evolved that way. The fact thatorganisms behave at all is due to genetic de-termination. Stimulus generalization and re-sponse induction are genetically determinedcharacteristics. The only incompatibility be-tween genetic determination and learning byreinforcement is that if some behavior iswholly genetically determined, as uncondi-tioned reflexes are, then no learning is neededto account for its occurrences. Such behaviorshardly "disprove" reinforcement theory, ofcourse.Whether, and if any, how much, grammati-

cal behavior is in fact due to genetic pre-determination of any great specificity is an-other matter. Obviously we have not inheriteda special set of grammatical neurons, so pre-spliced as to arrange verbal responses in cer-tain standard orders. At most we may be sup-posed to have inherited a predisposition tolearn grammatical behavior, and to do so ina certain way. The fact, if it is a fact, thatthere are grammatical universals hardly en-courages us to adopt an "inherited grammarnerve-net" hypothesis; if language learnerseverywhere share a common, somewhat simple,dynamic acquisition mechanism such as rein-forcement (which they do) we should expectthem to acquire complex behavior repertoires,both verbal and nonverbal, having many prop-erties in common (which they do). Any limita-tion in behavioral variety suggested by behav-ior universals may simply reflect a limitationimposed by the reinforcement process and pos-sibly some structural characteristics of thebrain such that it can only go about learningto order verbal responses in a distinctly lim-ited number of ways, due, indeed, to the sim-plicity of the reinforcement process and thefixity of the brain.The fact that some, but by no means all,

children acquire grammatical behavior at arather early age and rather suddenly (Chomskyfinds its rapidity "fantastic") does not requirea previously laid-down inherited grammaticalnerve net nor, even, anything much in the wayof a strong genetic prepotency for grammarlearning. As we have seen, nothing about thereinforcement process per se requires it to beslow and painstaking as Chomsky so insistentlyasserts it does (Chomsky, 1957, pp. 39, 42). Sim-ple responses appear, in fact, to condition inone reinforcement even in lower organisms,and the child is not a lower organism. Thedynamics but not the parameter values of thechild's reinforcement processes will resemblethose of the pigeon. The applicability of thelaw to the child is not in question merely be-cause the process proceeds at a more rapid rate.That a child learns certain orders, such as

adjective-noun, and actor-action sequences, onthe basis of a relatively small sampling fromthe enormous universe of such instances showssimply that a child is able to make complexabstractions and to generalize from them todiverse new instances. A parameter value maysurprise us, but it does not prove that the proc-

93

KENNETH MacCORQUODALE

esses of stimulus generalization and responseinduction are not applicable.

In brief, Skinner's hypothesis concerns howwhatever grammar acquisition genetic prede-termination leaves remaining to be done is infact done. The two sorts of determiners arecomplementary, not antagonistic. On the con-trary. It is distinctly inconsistent to argue thatwhile we may have inherited a disposition togrammatical behavior, we could not havelearned it by reinforcement. Both evolutionand reinforcement theory provide that whatsurvives behaviorally is what increases sur-vival chances, or, roughly, what reinforces.

Grammatical BehaviorChomsky's discussions of the controlling var-

iables for grammatical behavior suggest thathe views the necessity for postulating a media-tional mechanism for this particular aspect ofspeech as especially acute and, apparently, ob-vious. He says, for example, that "The childwho learns a language has in some sense con-structed the grammar for himself . . . thisgrammar is of an extremely complex and ab-stract character, . . . the young child has suc-ceeded in carrying out what from the formalpoint of view, at least, seems to be a remark-able type of theory construction (Chomsky,1959, p. 57)." This "grammar" is, then, a the-ory or, sometimes, "rules" and more recently"competence". It is a thing which the child,and later the adult, has and uses. It reveals it-self in two ways: as an understanding devicewhen its possessor listens, and as a generatingdevice when he speaks. As a consequence theword grammar is used in any number of ways.It is the name of the competence or rules ortheory which the speaker has constructed orinherited or learned; it is the name of a per-ceptual property of stimulus sentences he hearsor reads; and it is a property of his behaviorwhen he actually speaks himself. The first ac-tually underlies the second two and mediatesboth indifferently.As it functions during listening, the gram-

mar construction receives input in the form ofheard or read stimulus sentences, which itgoes to work upon so as to "distinguish sen-tences from nonsentences, to understand newsentences (in part), to note certain ambigui-ties, etc. (Chomsky, 1959, p. 56)." Presumablyit also communicates its verdicts to the rest ofthe person somehow, although what interest

he could have in a steady stream of diagnosesof sentencehood and nonsentencehood is some-what puzzling. At any rate the activity in-volved is readily recognizable simply as stimu-lus discrimination. There is nothing in that totax an S-R analysis, and nothing to force us tohypothesize an underlying theoretical con-struction. Nearly any set of diverse stimulusobjects, including the sentences one hears, canbe assorted into classes or subsets having someproperty in common and differing from othersubsets in some property. In this sense hearersprobably do learn to discriminate sentencetypes, but there is nothing unique to gram-matical stimuli in these discriminations; theydo not require any special, separate percep-tual mechanism or innovative perceptual proc-ess on the part of the hearer. Sentence dis-crimination, becoming highly sophisticated inlinguistics, probably exhausts the empiricalbasis from which inferences about the struc-ture of a speaker's underlying grammar con-struction can be made. That is, our knowledgeof any speaker's "competence" will necessarilybe a product of our perception of his actualemitted sentences, plus some empirically sur-plus inferences. It will be a highly discrimina-tive and abstract account of what he says, butit is speechbound and takes nothing else intoaccount.But what can speech alone tell us about ac-

tual causes of speech, including its grammati-cal determiners? Nothing unequivocal; thefact that there is a stimulus dimension in anindividual's observed verbal behavior identifi-able as "grammar" by no means entails thatthere is any unique causal variable called"grammar" at work in the production of hisverbal behavior. A simple causal system havingno pattern axioms at all may generate highlypatterned outcomes, with the pattern becom-ing represented as such only in the outcome,although, one hopes, predictable from an un-derstanding of the components and interac-tions of the nonpattern causal variables.So at least Skinner conceptualizes the auto-

clitic processes, defined in Verbal Behavior as"verbal behavior which is dependent upon orbased upon other verbal behavior (Skinner,1956, p. 315)." The formulation is abstruseand difficult and it takes some getting used to.It is certainly the most complex part of Skin-ner's hypothesis, although its complexities in-here in the interaction it supposes exist among

94

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

what are rather simple component processes.According to the formulation, speech may be-gin when the speaker has something to say, adisposition to respond due to his current stim-ulus and motivational circumstances. This"primary" speech is fragmentary, in that itdoes not include purely syntactic forms; it isunordered, in that many responses are concur-rently available, and it has no grammar. Givensomething to say, the speaker can then respondautoclitically to aspects of it, specifically to itsstrength and origins, by ordering and com-menting upon it as it appears in his speech.In terms of Skinner's analysis, such behavior issimply a complex kind of tact. The tact itself,is, however, not generically a grammaticalprocess at all, and it includes much that is notgrammar. The grammar does not come first,then, the elements of speech do. These insti-gate speech in which grammar emerges as theway these elements literally arrange them-selves.Chomsky's comment upon this hypothesis is

modestly placed in a footnote (Chomsky, 1959,ftn. 45, p. 54) which says: "One might just aswell argue exactly the opposite is true", andno doubt at this stage of knowledge one might.And so Chomsky does, supposing, after Lash-ley (1951), that syntactic structure is "a gen-eralized pattern imposed upon the specific actsas they occur (Chomsky, 1959, p. 55)." Thus,grammar is said to preexist outside verbal be-havior and exert a causal influence upon it.Lashley's conclusion was based upon, and isrelevant only to, an S-R analysis of grammati-cal ordering which hypothesizes that gram-matical behavior is a result of a left-to-rightprocess of intraverbal chaining. But Skinner'sautoclitic hypothesis involves no left-to-right.intraverbal chaining. It very adroitly (and al-most certainly in full knowledge of Lashley'spaper) puts the necessary controlling variablesin the interrelationships among the fragmen-tary "primary" verbal responses which aresimultaneously, not serially, available to thespeaker.

Skinner accounts for the instigation and de-termination of verbal behavior, grammar andall, in terms of variables external to the speechepisode itself, with a secondary, autoclitic stepadded once instigation is under way. Chomskyis totally silent, on the other hand, about whatmight be the form of input which would simi-larly engage the grammar construct when

speech is to be produced, and tell it what to begrammatical about, and how to select a possi-ble transformation to say it in, and so forth.So far as one can tell, Chomsky's one control-ling variable for speech production-grammar,rules, competence-rests locked away in thebrain somewhere, inert and entirely isolatedfrom any input variables which could ever getit to say something. Unless some external in-put is permitted one must suppose that thegrammar construct regulates itself, a repug-nant notion. No one speaks pure grammar. Allsentences have grammatically irrelevant prop-erties; they are, in addition, about something.Chomsky elsewhere in the review very firmlyrules out control by stimulus and motivationalvariables, as we have seen. One waits withbated breath to see what is left. The behaviorof the grammar construct must now be ex-plained. Until it is we are no further alongthan we were without it. It is simply "thatwhich controls grammatical behavior". Butthat, of course, is the question, not the answer.The speaker's cognitions will not do, sincethey too are theoretical constructions and mustin turn be explained. Sooner or later some-thing must enter the system. Guthrie com-plained that Tolman had left the rat "lost inthought" because he provided no relation be-tween the expectancy and behavior. Chomskyleaves the speaker lost in thought with nothingwhatever to say.

In sum, the verbally competent person candiscriminate a syntactic dimension in speechas a stimulus, and he can emit speech whichhas syntactic properties in the sense that ahearer can discriminate them. This does notprove in any way that some underlying theorygoverns both behaviors. A child learns both towalk and to discriminate walking. Nothing isgained by saying that therefore he has con-structed a theory of walking which he uses inhis perceptions and in his activities. So he maybe conceived to learn to speak and to perceivespeech, directly and without stopping to con-struct a theory or apply a rule.

The Extension to Verbal BehaviorChomsky's criticisms focus principally upon

Skinner's basic systematic apparatus, ratherthan its application to verbal behavior. Thathe feels there is relatively little left to say isrevealed in his introduction to his discussionof the application itself: "Since this system is

95

KENNETH MacCORQUODALE

based on the notions [sic] 'stimulus', 'response',and 'reinforcement', we can conclude . . . thatit will be vague and arbitrary (Chomsky, 1959,p. 44)." His treatments of the mand, tact,echoic, and so forth are therefore brief andadd little new in the way of specific criticism.A few details which have mostly to do withmisinterpretation of psychological fact or mis-reading of Skinner's text should be noted how-ever.

The MandIn Skinner's definition, a mand is a "verbal

operant in which the response is reinforced bya characteristic consequence and is thereforeunder the functional control of relevant con-ditions of deprivation or aversive stimulation(Skinner, 1957, pp. 35, 36)." "Characteristic"in the definition means a consequence havinga specific form, not a routine or inevitableconsequence as Chomsky misread it to. Chom-sky criticizes the definition because it is, as hesays, "generally impossible" to have informa-tion concerning the speaker's motivational cir-cumstances, and so the behavior analyst can-not make a correct diagnosis of whether a re-sponse is a mand or not. Similarly, as Chomskyreasons, the hearer, as reinforcement mediator,could not know whether or how to reinforce"relevantly." These are not real problems atall. The verbal behavior analyst must takeinto account whatever variables control be-havior, no matter how infeasible it is to detectthem in ordinary conversation. He will notundertake to test his theory in the drawingroom, after all, and since the speaker's moti-vational circumstances are objectively mea-surable they may in principle be known. Thetest of a good theory is not how verifiable it isby a casual and non-expert observer. Modernphysics would do very poorly by such a cri-terion.As for the reinforcement mediator, he need

not know anything whatever about the speak-er's motivation in order to play his role effec-tively as a mand conditioner. If a verbal re-sponse specifies characteristic consequences,for example, "pass the salt", "milk, please", or"get off my foot", and the hearer complies,then if the speaker has a relevant motivationalcondition, the reinforcement operation com-pletes itself, so to speak, and a mand, com-posed of that particular motivational condi-tion controlling that particular response, will

be strengthened. But if the speaker mandswhen he is not suitably motivated, the rein-forcer, although presented, is automatically in-effective and the response will extinguish forthe unmotivated state, thus sharpening moti-vational control. The reinforcement mediatorneed not concern himself with the speaker'smotivation; psychology will take care of it.

In questioning the possibility of ever dis-covering the relevant deprivations for suchmands as "give me the book", "take me for aride" or "let me fix it", Chomsky is forgettingthat reinforcers are not necessarily drive-re-ducers. A book may be a conditioned rein-forcer whose momentary effectiveness varieswith other motivational conditions: "I cannotfinish this paper and go to bed until I have thereference in that book", "I need something toprop this door open", "I hid ten dollars init". The effectiveness of conditioned reinforc-ers depends upon deprivation for somethingelse. Chomsky is quite correct in concludingthat deprivation is "relevant at most to a mi-nute fragment of verbal behavior (Chomsky,1959, p. 46)," but he is incorrect if he supposesthis is a defect in the system. One of the greatinsights of Verbal Behavior is that human re-inforcement-mediators can also reinforce non-motivated, non-mand, disinterested verbal be-havior. This fact, rather than a genetic muta-tion enjoyed only by his species, very likelyaccounts for the fact that only humans lhavedeveloped verbal behavior.Some sort of lapse appears to have occurred

where Chomsky erroneously detects an absurd-ity: "a speaker will not respond properly to themand 'Your money or your life' unless he hasa past history of being killed (Chomsky, 1959,p. 46)." The speaker? The speaker emits themand, he does not respond to it. He needs onlya history of having needed money. That israther common. There may be an absurdityhere but it is not in Verbal Behavior.

The TactSkinner defines the tact as "a verbal oper-

ant in which a response of given form isevoked (or at least strengthened) by a particu-lar object or event or property of an object orevent (Skinner, 1957, p. 81-82)." Chomsky'sprincipal objection to this treatment is its lackof congruence with the notions of referenceand meaning, which has already been dis-

96

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR

cussed, and which, although true, is a virtue,not a defect.

In addition, however, he criticizes Skinner'sformulation of why the hearer reinforces tact-ing: by doing so the hearer's potential contactwith the environment is functionally ex-tended. Once told that "dinner is ready" (atact) the hearer may behave in a way directlyreinforcing to himself. Although the dinnercan be seen by the tacter only in the diningroom, the tact may be heard by the reinforce-ment mediator around corners, upstairs, out-doors, or across town. Accurate tacters are,simply, very useful to have around. If theywere not, many of civilization's most cultivatedinstitutions such as schools, including profes-sional graduate schools, probably would notexist. Most knowledge about the world existsas talk. Chomsky's objection (1959, p. 48) thatthe parents of first-borns could not knowenough to teach them to tact (because they donot yet have the appropriate history of rein-forcement for hearing tacts) ignores the factthat parents already have a lifetime history ofhearing tacts from other speakers, mostlyadults. One does not wait until he has bornechildren to hear his first assertions about theworld and be reinforced for listening. Chil-dren's tacts are, in fact, useless and boring.One reinforces them all the same because ifthe children become good at tacting, theirtacts may become very useful indeed.Chomsky apparently finding the cash-value

explanation of reinforcement for tacting tooharsh, wonders if it would not be just as sci-entific to say that a parent has "a desire tosee the child develop and extend his capaci-ties (Chomsky, 1959, p. 48)." No, it certainlywould not, unless by "capacity" we mean atendency to tact accurately. Only well-discrim-inated (and therefore potentially useful) tact-ing gives the pleasure; babbling, jabbering,prattling, and outright lying may all be elabo-rately developed and extended as capacitiesbut they do not please parents nearly so much.Chomsky evidently misunderstood Skinner's

operant paraphrase of Bertrand Russell's re-spondent version of a hearer's response to thetact "fox". This paraphrase is most certainlynot Skinner's "own equally inadequate anal-ysis (Chomsky, 1959, p. 48)." It is merely a par-aphrase into the operant vocabulary of howthe events in Russell's example, if they oc-curred, would be analyzed. It was not intended

to suggest that those events are plausible, and,in fact, they are not.

Finally, Chomsky alludes to Skinner's treat-ment of how speakers are able to tact privateevents. Superficially this capacity seems mostmysterious from any point of view. How do welearn that the English name for this thing is"headache", for that thing "contentment", andfor those other things "thoughts"? For publicthings, like cows, someone who already knewthe name saw the cow we were looking at andtold us, and could reinforce our own response"cow" if the cow was there. The paradigm isimpossible for headaches and contentment andthouglhts because the reinforcement mediatorcannot share the relevant stimuli. Skinnermeets this problem head on; and, I think, bril-liantly. The interested reader should studythe explanation in Verbal Behavior (Skinner,1957, pp. 130 ff), and see also Skinner (1959,pp. 272-286). Skinner's interesting point is, infact, that only those internal stimuli that haveobvious external correlates which are observ-able by the reinforcement mediator can be-come discriminated, so that, as he says, it is thecommunity that teaches one to "know him-self". Chomsky dismisses the argument brisklyas a "heavy appeal to obscure internal stim-uli", a grossly inadequate characterization of avery sophisticated analysis.

The EchoicAn echoic is a response which "generates a

sound pattern similar to that of the stimulus(Skinner, 1956, p. 55)." Chomsky criticizes theaccount principally and afresh because it at-tributes the echoic repertoire to reinforcementrather than to instinctual imitative mecha-nisms. The significance of the genetic-rein-forcement aspect of this objection has been dis-cussed above.

The TextualA textual, which is a verbal response to a

written stimulus, and which makes no de-mands upon linguistic competence or gram-matical behavior, but is surely verbal andusually more grammatical than any other ver-bal behavior from the same speaker, is not dis-cussed in the review.

The IntraverbalIntraverbals, which are verbal responses un-

der control of other verbal behavior, are dis-

97

KENNfTH MacCORQUODALE

missed, but hardly discussed, by Chomskyalong with the very important role which Skin-ner argues they play in nearly all extended in-tervals of verbal behavior. Once verbal behav-ior begins we are able to continue speaking,almost endlessly, under the stimulus influenceof what we have already said. The role of in-traverbal stimuli in instruction is to combinewith echoics and textuals so as to produce aresponse which was not previously available.Chomsky wonders (1959, p. 52) in what sensethis can be true for someone who is told (anechoic stimulus) that "the railroads face col-lapse" since the hearer could have said thisbefore. But the point is that he could not havesaid it unless he had a momentary reason todo so. He could as readily have said "the rail-roads do not face collapse" or "there is a four-banded armadillo in the gazebo". Chomsky isonce again overlooking the difference betweena speaker's vocabulary as a response repertoire(what he is able to say) and speaking as a re-sponse (what he is able to say now). Psychologyis concerned with both, but principally thelatter. They are clearly different.

CONCLUSIONI conclude that Chomsky's review did not

constitute a critical analysis of Skinner'sVerbal Behavior. The theory criticized inthe review was an amalgam of some ratheroutdated behavioristic lore including rein-forcement by drive reduction, the extinctioncriterion for response strength, a pseudo-in-compatibility of genetic and reinforcementprocesses, and other notions which have noth-ing to do with Skinner's account. Chomskymisunderstood the intent of Verbal Behavior,evaluating it as an accomplished explanationof verbal behavior rather than a hypothesisabout the causes of verbal behavior. His reviewrejected in principle the products of Skinner'smethodology without having come to termswith his rationale, especially as it concerns thenecessity for theoretical mediating variables.The review took a view of extrapolation oflaboratory findings that would bring any scien-tist's methodology to a dead stop. It rejectedwithout discussion the logic of reductionism.It criticized Verbal Behavior for not havingbeen about something else, that is, a theoryof verbal behavior rather than verbal behavioritself. It redefined verbal behavior by equating

it exclusively with the items in the behavioralrepertoire rather than the speaker's momen-tary verbal responding, an obvious and crucialdistinction for psychology.The review completely ignored much that is

central to an understanding, application andassessment of Skinner's position. Most impor-tantly it failed to reflect Skinner's repeated in-sistence that the full adequacy of his explana-tory apparatus for complex cases, includingverbal behavior, cannot be assessed unless thepossibilities for interaction among its severalcontrolling variables acting concurrently wererealized; this is what is different between thelaboratory and the real world. In the labora-tory, variables are made to act "one at a time",for all practical purposes. The real world sim-ply puts the environment back together again.Multiple causality is never mentioned in thereview; it is mentioned throughout VerbalBehavior. The mystery of its omission from thereview is compounded by the fact that Chom-sky found it mysterious that Skinner thoughtsomething so complex as speech could be ac-counted for "by a simple function"!But the review, however approximate, has

had an enormous influence in psychology.Nearly every aspect of currently popular psy-cholinguistic dogma was adumbrated in it, in-cluding its warlike tone; the new look is afrown. It speaks of itself as a "revolution", notas a research area; it produces "confronta-tions", not inquiries. So far there have beenno telling engagements in the revolution. Thedeclaration of war has been unilateral, prob-ably because the behaviorist cannot clearlyrecognize why he should defend himself. Hehas not hurt anyone; he has not preempted theverbal territory by applying his methods toverbal behavior; he has not used up all of theverbal behavior nor has he precluded otherscientists from investigating it to their heart'scontent, with any methods and theories whichplease them; he need not be routed before theydo so.The behaviorist will not be roused to self-

defense by having a few new paradigms rattledat him. New paradigms in psychology are,bluntly, less than a dime a dozen. They comeand go. The most illuminating example in thepresent instance is Gestalt psychology, as Neis-ser has noted (1967). The behaviorist does, onthe other hand, understand new data. He willbe the first and best judge of whether they are

98

ON CHOMSKY'S REVIEW OF SKINNER'S VERBAL BEHAVIOR 99

incompatible with his own paradigms and hecan be trusted to take them into account eitherway. Psycholinguistics will do itself a disserviceby spending more of its time trying to destroybehaviorism, but if it is determined to do so,it should learn first exactly what the behavior-ists really said, and how behaviorisms differfrom one another. The same amount of timespent developing the positive aspects of itspoint of view will at least test whether it candefine itself as something coherent and posi-tive, rather than merely antibehavioristic. Thepsycholinguists are probably in a uniquelyfavorable position to make important advancesin speech perception and in discovering whatthe stimulus dimensions of syntactic structuresare. This knowledge will be invaluable to any-one interested in producing a causal system ortheory of speech production, but I do not be-lieve that he will be a psycholinguist.Meanwhile, it has been 10 years. One can

only agree with another observer of this scenewho recently said: "in the flush of their initialvictories, many linguists have made extrava-gant claims and drawn sweeping but unsup-ported conclusions about the inadequacy ofstimulus-response theories to handle any cen-tral aspects of language behavior . . . Theclaims and conclusions so boldly enunciatedare supported neither by careful mathematicalargument . . . nor by systematic presentationof empirical evidence to show that the basicassumptions of these [stimulus-response] the-ories are empirically false. To cite two recentbooks of some importance, neither theoremsnor data are to be found in Chomsky (1965) orKatz and Postal (1964), but one can find rathermany useful examples of linguistic analysis,many interesting and insightful remarks aboutlanguage behavior, and many ba(dly formu-lated and incompletely worked out argumentsabout the theories of language learning."(Suppes, 1968, pp. 1-2.)

Just so.

REFERENCESBaer, D. M. and Sherman, J. A. Reinforcement con-

trol of generalized imitation in young children.Journal of Experimental Child Psychology, 1964, 1,37-49.

Brigham, T. A. and Sherman, J. A. An experimentalanalysis of verbal imitation in preschool children.Journal of Applied Behavior Analysis, 1968, 1, 151-158.

Chomsky, N. Verbal Behavior. By B. F. Skinner. Lan-guage, 1959, 35, 26-58. Reprinted as item A-34 in theBobbs-Merrill Reprint Series in the Social Sciences;in J. A. Fodor and J. J. Katz, 1964; and in L. A.Jakobovitz and M. S. Miron, 1967.

Fodor, J. A. and Katz, J. J. The structure of language:readings in the philosophy of language. EnglewoodCliffs, N.J.: Prentice-Hall, Inc., 1964.

Holz, W. C. and Azrin, N. H. Conditioning humanverbal behavior. In W. K. Honig (Ed.), Operant be-havior: areas of research and application. NewYork: Appleton-Century-Crofts, 1966. Pp. 790-826.

Honig, W. K. (Ed.). Operant behavior: areas of re-search and application. New York: Appleton-Cen-tury-Crofts, 1966.

Jakobovits, L. A. and Miron, M. S. (Eds.). Reading inthe psychology of language. Englewood Cliffs, N.J.:Prentice-Hall, Inc., 1967.

Katahn, M. and Koplin, J. H. Paradigm clash: com-ment on "Some Recent Criticisms of Behaviorismand Learning Theory with Special Reference toBreger and McGaugh and to Chomsky." Psychologi-cal Bulletin, 1968, 69, 147-148.

Krasner, L. Studies of the conditioning of verbal be-havior. Psychological Bulletin, 1958, 55, 148-170.

Lashley, K. The problem of serial order in behavior.In L. A. Jeffress (Ed.), Cerebral mechanisms in be-havior. The Hixon symposium. New York: JohnWiley & Sons, Inc., 1951. Pp. 112-136.

Lenneberg, E. H. The capacity for language acquisi-tion. In J. A. Fodor and J. J. Katz (Eds.), The struc-ture of language. Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1964. Pp. 579-603.

Lenneberg, E. H. Biological foundations of language.New York: John Wiley & Sons, Inc., 1967.

Lovaas, 0. I., Berberich, J. P., Perloff, B. F., andSchaeffer, B. Acquisition of imitative speech byschizophrenic children. Science, 1966, 151, 705-707.

Miller, N. E. and Dollard, J. Social learning and imi-tation. New Haven: Yale University Press, 1941.

Neisser, U. Cognitive Psychology. New York: Apple-ton-Century-Crofts, 1967.

Salzinger, K. Experimental manipulation of verbalbehavior: a review. Journal of General Psychology,1959, 61, 65-94.

Salzinger, K., Feldman, R. S., Cowan, J. E., and Sal-zinger, S. Operant conditioning of verbal behaviorof two young speech-deficient boys. In L. Krasnerand L. P. Ullmann (Eds.), Research in behaviormtiodification: new developments and implications.New York: Holt, Rinehart & Winston, 1965. Pp. 82-105.

Skinner, B. F. Science and human behavior. NewYork: The Macmillan Company, 1953.

Skinner, B. F. Verbal behavior. New York: Appleton-Century-Crofts, 1957.

Skinner, B. F. Cumulative record. New York: Apple-ton-Century-Crofts, 1959, 1961.

Suppes, P. Stimulus-response theory of finite auto-mata. Technical report No. 133. Psychology Series.Institute for Mathematical Studies in the SocialSciences. Stanford University, 1968.

Wiest, W. A. Some recent criticisms of behaviorismand learning theory with special reference to Bregerand McGaugh and to Chomsky. Psychological Bul-letin, 1967, 67, 214-225.