verbal behavior. by b. f. skinner. (the century...

,

.I

!Ii.t-

~i!..

~~,::;:;~"~'~\'"J

I ';.;~i;:",c5

Verbal behavior. By B. F. SKINNER. (The Century Psychology Series.) pp~viii, 478 New York: Appleton.Centliry-Crofts, Inc., 1957...

Reviewed by NOAK CHOMSKY, MassachuseU& Institute of Technologyand Institute for Advanced Study

1. A great many linguists and philosophers concerned with language haveexpressed the hope that their studies might ultimatcly be embeddcd in a frame-work provided by behaviorist psychology, and that refractory areas of investi-gation, particularly those in which mooning is involvcd, will in this way be openedup to fruitful exploration. Since this volume is the first large-scale attempt toincorporate the major aspects of linguistic behavior within a behaviorist frame-work, it merits IUld will undoubtedly receive careful attention. Skinner is notedfor his contributions to the study of animal behavior. The book undcr revicw ~the product of study of linguistic behavior extending over more than twentyycars. Enrli~r versions of it hnvc l)C()n fairly widcly cireulat<.'{], tLnd thcro ufoquite a fcw references in the psychological literature to its major ideas.

Tho problem to which thil; I)ook il; add~ i" that, uf givillg a 'fUllctiUllUlanalysi!3' of verbal behavior. By functional analysis,.Skinner means identifica-tion of the variables that control this behavior and spooification of how thcyinteract to dctcrmine a particular vcrbal response. Furtbermore, the controllingvariables are to be described completely in terms of such notions as stimulus,reinforcement, deprivation, which have been given a reasonably clear meaningin animal experimentation. In other words, the goal of the book is to provide away to predict and control vcrbal bchavior by obscrving and manipulatinJt thophysical ellvironmcnt of tho spcakcr.

Skinllcr fcols thnt rcccnt advnncc/j in tll0 1n.l>oratory I!tudy of 1111imal bchl~viorpcrmit us W upprolwh this problom with a certain optimism, liincc 'the basicprocesses and relations which give verbal behavior its special characteristicsare now fuirly well undcrstood ...tho fCl!ults [uf this oxpcrimcnt.u.l wurk) havcbeen ijurprisingly freo of species restrictions. Recent work has shown that themcthods can be extcndcd to human bchavior without serious modification'(3).1

( 0:; :~,;

,..:

:f!;"

.~}~::. ~i1~~f: .:'

..I;~J.""""'1~'~~.J.,~

~E~~:a.W"»

r~~~, C"'~?'~

l ,fi4~:~'.,;1)

)A'.lji~::~",,'-:.."" ..,

~:~A"

~~~i""i!:~'i;~

I Skinner's confidence in recent achievements in the study of animal behavior and theirapplicability to complex human behavior does not appear to be widely shared. In manyrecent publications of confirmed bchaviorist¥ there is a prevailing note of skepticiam withregard to the scope of these achievements. For repreaentative comments, see the contribu~tion~ to Modern learninu theory (by l~tcII at. al.j New York, 1064); Bugajski, P8rchology 01learning (New York, 1956); Koch, in Nebraaka aympo8ium.on motillation 58 (Lincoln, 19M);Verplanck, Learned and innate behavior, Parch. re.. 52.139 (1955). Perhape the strongestview is that of Harlow, who has asserted (Mice, monkeYI, men, and motivee, PaycA. r...00.23-32 [19531) that '1\ I!tronK calla can 1>0 made for tho propOHition that the importAnco oftha psych<llogieal problalD8 ~I.udicd during the last 15 years has decreased as a negativelyaccelerated function approaching an asymptote of complete indifference.' Tinbergen, aleading representative of a different approach to animal behavior studiM (comparativeethology), concludes a discussion of 'functional analysis' with the commen' that 'we may

26Reprinted from LANGUAGe. Vol. 35, January-March, 1959

Copyright, 1959, Unguiltic Society of A_1ca .All right. ,.MfVed

:..~""

~

27REVIEWS

v

It is important to see clearly just what it is in Skinner's program and claimsthat makes them appcar HO bold und remarkable. It is not primarily the factthat he has I3Ct functional analysis I1tI his problem, or that he llinits himself tostudy of 'obsetvables', i.e. input-output ~lations. What is 80 surprising is theparticular limitations he has imposed on 'the way in which the o~n'ables ofbehavior are to be studied, and, above all, the particularly simple nature of tho'function' which, hc claims, describes the causation of behavior. One wouldnaturally expect that prediction of the behavior of a complex organism (ormachine) would require, in uddition to information about extcrnal stimulation,knowledge of the internal structure of the organism, the ways in which it proc-e~Minput information II.Jld organiZCH its own behavior. 'nlCtiO churactcristicti ofthe organism are in general a complicated product of ~born structure, thegcnctically detcrmined coursc of maturation, and past cxpcricncc. Insofar l~indcpcndcnt ncurophysiologicnl cvidcnce is not avnilnble, it is obvious thatinrcrcnCC!! coll(,'Crning the structure uf the orgnniHm nre bnscd on obscrvution ofbehavior and outside events. Nevertheless, one's estimnte of the relative im-J)()r~LnCC or cxtcmul fucwr!) ILIIU illtcrlllLI Kt.ructurc in tllC dctcrminution of lJ(!-fllLviur will havc an importullt effect 011 the direction of research on linguistic(ur uny utJlcr) bchuvi(,lr, UIIU (,III tJ1C kinuM of ullulogie!! frum animul bchuvi(,lrstudies that will be considered relevnnt or suggestive. .

Putting it diffcrently, anyone who scts himsclf the problem or anulyzing thc

causation of behuvior will (in tho ubsencc of independent neurophysiologicalevidence) concern himself with the only data available, nnmely the record ofinputs to the organism and thc orgunism's present responsc, and will try todescribe the function specifying the response in terms of the history of inputs.TlliK iK nutJling more tflun the ucfinitiun of his problcm. r!'Jlcrc urc no poMSible~rounu~ for arbrument. flcre, if one ucccpts tho problcm /\.'I lcgit.imate, though~killncr hM oftcn aUVCLIICcd unu dcfcndcd thill u(:finitic»1 Clf IL problem ILS if it.

wore u t.hcliiM which ut.hcr illvc!)tiguwr!) reject. rrhCrUilTC cncc!) thut ILrisc bctWOO11thoRO wht? ~m and those who deny the importlLII(~ tho Kpccific 'contributionur UIU urg:Llli~m' to luurllillg ILIIU l)(!rfcJrllulllc:(: ~u urll tile particullLr chuructcr:Llld co~plcxity of ~his function, IlllU tilc kinds of obscrvations and ~ILrchlIuce~y for arriving at u preci::;c I-;I)(!cificutioll of it. If the contribution uf theorgnnism is complex, the only hopc of predicting behavior even in a gross waywill be through a very indirect, progrum of research that gegins by studying thedetailed cha."8.cter of the behavior itsclf and the particularcapa~itiesof the~rg~\niSJJ. involved. --

Skinner's thesis is that extcrtlnl fnctors consisting of present stimulation andthe history of rcillforccmcllt (in particular the frequency, arrungement, andwithholding of reinforcing stimuli) are of overwhelming importance, and thatthc genernl princ:iplcs revcalcd in laboratory studies of these phenomena pro-viuc tho hlL...iK for tmdcrKt.:Lllding tho complexities of verbal bchavior. Hc confi-

now clraw thc conclusion that the CAusation of behavior is immcnscly more complex thanWUIi alJKumccl in the gcncralizationK or thc past. A number of internal and external factorsuct upon complex centrul nervous structurcl!. Sccond, it will be obvious thAt the facta at ourcliHpoKul arc vcry frugmentury inclcccl'-The Btudy oj in.linct 74 (OxCorcl, 1951).

,.i~~~~\.~a

:;,":r;"~ ;t "",~,",~'9!-';~~l~"'.""

',~~,,~(":;;~4:)'" .~tj,;r 't~;'::~i;;'~"i";t'!~'~;;-~::;',.'..-".":.,": ,~,.:.,~.,.,. .';::~ ..:-}f.{9:::t!fo;~,~~~c~;;,-o,:;:..:i$~

;':~~",'"c,'" c, , N'.

;'1

:"..c~,."';;"""\ ;~,~\

28 LANGUAGE, VOLUME 35, NUMBER 1 (1959)

dently and repeatedly voices his claim to have demonstrated that ~ t1nntri.y~~~ ~ker is Quito trivi~l and elementary, and that prcciHC prediction

of verbal behavior involves only specification of the few. external factors thathe has isolated experimentally with wWer organisms.

fC~ef1,ll study 9f this bQOk (and of the researc~ on which it draws) reveals,~~rJ. that these astonishing .claims are far from justifiedl It indicates, fur-thermore, that the insights that have been achieved in the laboratories of tllereinforcement theorist, though quite genuine, can be applied to complex humanbehavior, only in tlle most gross and superficial way. and that speculativeattempts to discuss linguistic behavior in these terms alone omit from con-sideration factors of fundamental importance that are, no doubt, amenable toscientific study, although their specific character cannot at present be preciselyformulated. Since Skinner's work is the most extensive attempt to accommodatchuman behavior involving higher mentW faculties within a strict behavioristschema of the type that has attracted many linguists and philosophers, as wellas psychologists, a detailed documentation is of independent interest. The mag-nitude of tho failure of .Jt!!jB attcmpt to account for ~~rJ?lI.l behavior ~rvcs ~ akind. of .~cnsure of t.h" importance of the ~actors omitted from. ooneideration,and an indication Of-.ho\vJjj,~!~_iB,~y known about tJ1is remarkably complexphenomenon.

The force of Skinner's argumcnt lies in tlle enormous wealth and range ofexamples for which he proposes a functional analysis. The only way to evaluatethe success of his program and the correctness of his basic assumptions aboutverbal behavior is to review these examples in detail and to determine the pre-cise character of the concepts in terms of which the functional analysis is pre-scntcd. 12 of this review describes the experimental context with respect towhich these concepts are originally defined. 16.1-:4 deal with the basic concepts'stimulus', 'response', and '(Qj!!i9mQlllopt', i§6-10 with the new descriptivcmachinery developed specifically for the description of !~l'~al ~havior. In §5we consider the status of the fundamental claim, drawn from the laboratory,whicll tlCrvcs as tho bj18is for tl\c nllulogic guCS8CS about human behavior thathave been proposed by many psychologists. The final section (ill) will considersome ways in which further linguistic work may playa part in clarifying someof these problems.

2. Although this book makes no direct reference to experimental work, itcan be understood only in terms of the general framework that Skinner hasdeveloped for the description of behavior. Skinner divides the responses of theanimal into two main categories. Re8p~ are purely refi,ex responses elicitedby particular stimuli. Operants are emitted responses, for which no obviousstimulus can be discovered. Skinner has been concerned primarily with operantbehavior. The experimental arrangement that he introduced consists basicallyof a box with a bar attached to one wall in such a way that when the bar iopressed, a food pellet is dropped into a tray (and die bar press is recorded). Arat placed in the box will soon press the bar, releasing a pellet into the tray. Thisstate of affairs, resulting from the bar press, increases the strength of the bar.

29REVIEWS

pressing operant. The food pellet is called a reinforcer; the event, a reinforcingevent. The strength of an operant iR defined by Skinner in terms of thc rote ofresponse during extinction (i.e. after the last reinforcement and before return tothe preconditioning rate). '

Suppose that release of the pellet is 'conditional on the flashing of a light.Then the rat will come to press the bar only when the light flashes. This is calledstimu~ ~~i.~~ The response is called a diacriminated operant and thelight is called the ~ for its emission; this is to be distinguished from elicita-tion of a response by a stimulus in the case of the respondent.' Suppose that theapparatus is so arranged that bar-pressing of only a certain character (e.g.duration) will release the pellet. The rat will then come to press the bar in therequired way. This process is called response differentiation. By successive slightchanges in the conditions under which the response will be reinforced it is pos-sible to shape the response of a rat or a pigeon in very surprising ways in a veryshort time, so that rather complex behavior can be produced by a process ofsuccessive approximation.

A stimulus can become .reinforcing b.vJ:epeated asiOCiation with an alroudyreinforcing stimulus. Such a stimulus is called a aecondary reinforcer. Like manycontemporary behaviorists, Skinner considers money, approval, and the like tobe secondary reinforcers which have become reinforcing because of their associa-tion with food etc.'- Sccond~' reinforcers can be generalized by 8BSOCiating themwith a variety of different primary reinforcers.

Another variable that can affect the rate of the bar-pressing operant ~v~,which Skinner defines operationally in terms othours of deprivation. His majorscientific book, !Behavior of organi8m8, is a study of the effects of food-depriva-tion and conditioning on tho strength of the bar-prcl:l8ing response of healthymature rats. Probably Skinner's most original contribution to animal behaviorstudiCB hfJ.8 boon hi.. invcHtigntion of tho effects of J!JJ!;!.!m~t.ont reilu~rcomcllt.,arranged in various different ways, presented in BeMuiM of organi8m8 and ex-tended (with pecking of pigeons as the operant under investigation) in the ro-cent,.s:~dulc6 of reinforcement by li'erster and Skinner (1957). It is apparently~

'In Behavior of organilma (New York, 1938), Skinner remarks that 'although a condi-tioned operant i8 the reault of the correlation of the rMponae with a particular reinforce-ment, a relation between it and a di8criminative 8timulu8 acting prior to the response is thealmost universal rule' (178-9). Even emitted behavior i8 held to be produced by 80me 80rt of'originating force' (51) which, in the case of operant behavior i8 not under experimentalcontrol. The di8tinction between eliciting 8timuli, discriminated 8timuli, and 'originatingforces' has never been adequately clarified, and becomes even more confu8ing when privateinternal events are considered to be discriminated stimuli (see below).

I In a famous experiment, chimpanzees were taught to perform complex tasks to receivetokens which had become secondary reinforcers because of aaaociation with food. rrhe ideathat money, approval, prestige, etc. actually acquire their motivating effects on humanbehavior according to thi8 paradigm t8 unproved, and not particularly plau8ible. Manypsychologists within the behaviorist movement are quite skeptical about this (cf. fn. 23).As in the case of most aspects of human behavior, the evidence about secondary reinforce-ment is 80 fragmentary, confiicting, and complex that alm08t any view can find some sup-port.

f;" ...!~~~:

.ffi


these studieS-that Skinner has in mind when he refers to the recent advances inthe study of animal behavior..

The notions ~~~ '~eponse', '~2~~~nt' are relatively well definedwith respect to the bar-pressing experiments and others similarly restricted.Before we can extend them to real.life behavior, however, certain difficultiesmust be faced. We must decide, first of all, whether any physical event to whichthe organism is capable of reacting is to be called a stimulus on a given occasion,or only one to which the organism in fact reacts; and correspondingly, we mustdecide whether any part of behavior is to be caHed a response, or only one con-nected with stimuli in lawful waye. Questions of this sort pose something of adilemma for the experimental psychologist. If he accepts the broad definitions,characterizing any physical event impinging on the organism as a stimulus andany part of the organism's behavior as a response, he must conclude that be-havior hUll not boon demon titrated to bc .lawful. In tho prescnt stntc of ourknowledge, we must attribute an overwhelming influence on actual behavior toill-defincd factors of attention, set, volition, and caprice. U we acccpt_t:.l!C!1M:j'uwcr uefinitionti, thon.~r ill !awful by definition (if it consistti of responses);but thiti fact is 9LlLmitc~ sig~ancel since most of what Ulo animal does wJ!L8:i~ply not be considered ~h~~inr --Hence the psychologist either must admitthat-oehavior is not lawful (or that he cannot at present show that it is-notat all a damaging admission for a developing science), or must restrict his atten-tion to those highly limited areas in which it is lawful (e.g. with adequate con-trols, bar-pressing in rats; lawfulness of the observed behavior provides, forSkinner, an implicit definition of a good experiment).

Skinner does not consistently adopt either COurse'~1;1t~~.th~perim§{ltll~sul~-~:iQQU~~~iQRtifio-c:harac~ .his .systeIn-9L ~avior, andanalogic guesses (formulated in terms of a metaphoric extension of the -t.ecnmcalvocabulary of the laboratory) as evidence for its scope. This creates the illusionof a rigorous scientific theory with a very broad scope, although in fact tbe termsused in the description of real-life and of laboratory bchavior may be merehomonyms, with at most a vague similarity of meaning. To substantiate this

.Skinner's remark quoted above about the generality of his basic results must be under-stood in the light oC the experimentallimitationa he bas imposed. If it were true in any deepI;cnljo that tho ballic proccsses in language are well understood and Cree oC 8peciell restric-tiona, i.t..a:'11-,1ti h~~tr~Jn~~~~gua.geJaJimi.w9~~ With the exception of afew scattered observations (cf. his article, A case history in 8cientific method, The America"psychologist 11.221-33 [l956J), Skinner is apparently basing this claim on the fact that quali-tatively similar results are obtained with bar-pressing oC rats and pecking oC pigeona underspecial conditions oC deprivation and various schedules oC reinforcement. One immediatelyquestions how much can be based on these facts, which are in part at least an artifact trace-able to experimental design and the definition oC 'stimulus' and 'responae' in terms oC'smooth dynamic curves' (see below). The da~&era-inhereat,.-ift-aDy-&~p~ ~ 'ext,ap;,k,te'~~~~vior from the study oCauoh-eimple FeapenseaA8 bar-pr~iD&,Heuld-b8..ob..~ous, and ~av~ orten been commented on (cf. e.g. Harlow, op.cit.). The generality of eventhe simplest results is'open to serious question. CC. in this connection Bitterman, WodiDlky,and Candland, Some comparative psychology, Am. jour. oj psych. 71.~110 (1958), whereit is shown that there are important qualitative differences in solution of comparable ele-mentary problems by rats and fish.

e

31REVIEWS

evaluation, a critical account of his book must show that with a literal reading(where the terms of the descriptive system have something like the technicalmeanings given in Skinner's definitions), the book ~1"_~ ~ost no asooct ~linguistic behavior, and that with a metaphoric reading, it is no more scientific~an-uietraaditional approaches to this subject matter, and rarely as clear andcareful,'

3. Consider first Skinner's use of the notions 'stimulus' and 'response'. InBehavior of organi81n8 (9) he comInits himself to the narrow definitions for theseterms. A part of the environment and a part of behavior are called stimulus(eliciting, discriminated, or reinforcing) and response, respectively, only if theyare lawfully related; that is, if the: 'dynamic laws' relating them show smoothand reproducible curves, Evidently stimuli and responses, so defined, have notbeen shown to figure very widely in ordinary human behavior,' We can, in theface of presently av~able evide.n9.e, ~tinue to maintain the lawfulness of therelation between~~nd~sp(;ln~:;. e:;. Qepri"iI.~ tb"...cl-t.heir objectivecltnracter. A typical cxn.mplc of 'stimulus control' for Skinner would be theresponse to a piece of music with the utterance Mozart or to a painting with theresponse Dutch. These responses are asserted to be 'under the control of extremelysubtle properties' of the physical object or event (lOB). Suppose instead of sayingDutch we had said Clashes with the waUpaper, I thought you liked abstract work,Never saw it before, Tilted, Hanging too low, Beautiful, Hideous, Remember ourcamping trip last 8Ummer', or ,vhatever else might come into our minds whenlooking at a picture (in Skinnerian trunslation, whatever other responses existin sufficient strength). Skinner could only say that each of these respo~s isunder the control of some other stimulus property of the physical object. If welook at a red chair and say red, the response is under the control of the stimulus'redness'; if wc say clwir, it is under the control of the collcctioli of properties(for Skinner, the object) 'chairness' (110), and similarly for any other response.This device is as simple as it is empty, Since properties are free for the asking

I An analogous argument, in connection with a di/Terent aspect or Skinner's thinking, isKivcn by Scrivon in It .ludy oJrodical behalliori.m -Unit!. oj Minn. .ludie. in ph.il030ph.~ oj8cience, Vol. 1. cr. Verplanck's contribution to Modern learning theOrJI (283-$) for moregeneral discussion or the difficulties in formulating an adequate definition of 'stimulus' and'rcllponsc'. Hc concludcs, quite corrcctly, that in Skinner's senso of the word, stimuli arcnot objectively identifiable independently or the resulting behavior, nor are they manipu-lable. YfII;pIAnck presents a clear discussion of many other aspects of Skinner's system, com-menting on the untestability of many of the so-called 'laws or behavior' and the limitedscope of many of the others, and the arbitrary and obscure character of Skinner's notion of'lawful rcllltion'j Ilnd, at thCHI,me time, noting thc importanco of thecxperimcntal data thatt;kinncr hllJj accumullltcd.

.In Behat!ioT 0/ organi8ms, Skinner apparently was willing to accept this consequence.He insists (41-2) that the terms of casual description in the popular vocabulary arc notvalidly descriptivc until the dcfining propcrtiell or stimulus and rcsponsc arc spccified, thccorrelation is demonstrated experimentally, and the dynamic changes in it are shownto be lawful. Thus, in describing a child as hiding from a dog, 'it will not be enough to dig-niry the popular vocabulary by appealing to essential properties or "dogness" or "hiding-ness" and 1.0 suppose them intuitively known.' But this is exactly what Skinner does in thebook under rcview, as we will see directly.

..

~~;

.~j;~:"

,-. ;;,"-,..11 "-,.,"- ,--

, ':"f'~~ ;,:,': ~ ';",:: ,f:':Ai;,":;":;.' j;-,...,," f"0;" ...~". P"":;"'~':;(:"""<:&'"";<:~':tl."",~,--- , , "" ~~ioi;p';,~':~r~c,

,'~~,

,;~."~~'~

32 LANGUAGE, VOLmm 35, NUMBER 1 (1959)

(\\.c have as many of them as we have nonsynonymous descriptive expressionsin our language, whatever this means exactly), we can account for a wide classof responses in terms of Skinnerian fu~<'ttional analysis by identifying the 'con-trolling stimuli'. But the word 'stimulus' has lost all objectivity in this usage.Stimuli are no longer part of the outside physical world i they are driven backinto the organism. We identify the stimulus when we hear the response. It isclear from such examples, which abound, that the talk of ~ulus control'simply disguises a complete retreat to mentalistic psycholog.f;1We cannot pre-dict verbal behavior in terms of the stimuli in the speaker's environment, sincewe do not know what the current stimuli are until he responds.! Furthermore,since we cannot control the property of a physical object to wh~ an individualwill respond, except in highly artificial cases, Skinner's claim that his system,as opposed to the traditional one, permits the practical control of verbal be-havior? is quite false.

Other examples of '8.timulus~ontrol' merely add to the general mystification.Thus a propcr noun is held to be a response 'undcr tho control of a specificperson or thing' (as controlling stimulus, 113). I have otten used the wordsb'isenhower and M 08COW, which I presume are proper nouns if anYthing is, buthave never been 'stimulated' by the corresponding objects. How can this factbe made compatible with this definition? Suppose that I use the name of a friendwho is not prcscnt. Is this an instance of a proper noun under the control of thefriend as stimulus"fflsewhere it is asserted that a stjmyl.\.J:S co~tro1s.a responsein, t.h~~.~~:1:~~~~.pf~~u~ inc~ the prob~~~QftJ;1~~,~~But it is obvious1~ J!n..!~e_that the probabtl1t-y-tnaf: a speaker will pro-ducc a lull name is incren"'Seo wncn its bearer faces the speakcr, Furthermore,

7253 f. and elsewhere, repeatedly. As an example of how well we can control behaviorufling the notion8 devolopod in this book, Skinner Ihows hero how ho would KO about evok-ing the response pencil. Tbe most efJective way, be luggeete, il to laY to the lubject 'Pleaselay pe~il' (our chances would, presumably, be even further improved by use of 'aversive8timulation', c.g. holding a gun to his head). We can also 'make suro t.hat no pencil orwrit.inginlltrument. ill availal)le, then hand our suuject a pad of paper appropriatc t.o pencil8ketch-ing, and offer him a handsome reward for a recognizable picture of a cat.' It would altO beulieCul to have voices saying pencil or pen and ...in the backgr:ound; ligns readin, pencilor pen and ...; or to place a 'large and unusual pencil in an unusual place clearly insight'. 'Under such circumstances, it is highly probable that our subject willeay pe~il.''The available techniques are all illustrated in this eample.' These contributions of behav-ior theory to the practical control of human behavior are amply illustrated elsewhere inthe book, 8.8 when Skinner shows (113-4) how we can evoke the response red (the devicesuggested is to hold a red object before the subject and say 'Tell me what color thil is').

In !airncss, it must be mentioned that there are certain nontrivial applications of 'op-erant conditioning' to the control of human behavior. A wide variety of experimente haveshown that t~~~nlur~~~-ple) produced by a subject will increaseif the experimenter says 'right' or 'good' when one is produced (similarly, positive attitudeson a certain issue, stories with particulnr content, etc.; cf. Krasner, Studie8 of the condi-tioning of verbal behavior, Pluch. bull., Vol. 65 [1958], for a survey of several doleD experi-ments of this kind, mostly with positive results). It i8 of some interest that the 8ubject i8usually unaware of the process. JU8t what in8ight this gives into normal verbal behavior isnot obvious. Nevertheless, it is an example of positive and not totally expected reault.a us-ing the Skinnerian paradigm.

~~:\~Jr'.

REVIEWS 33

.

how can one's own name be a proper noun in this sense? A multitude of similnrquestions arise immediately. It appears that$~-~o~ ~~~here is merely a.misleading p~aJ':hrase for the traditio:n~.de~~~..Qr.~. The assertion (115)that so far as th;~er-rsconcemed; tbe relation of reference is 'simply theprobability that the speaker will emit a response of a given form in tho presencoof a stimulus having specified properties' is surely incorrect if we take the words'presence', 'stimulus', and 'probability' in their literal sense. That they are notintended to be taken literally is indicated by many examples, as when a responseis said to be 'controlled' by a situation or state of affairs as 'stimulus'. Thus,the expression a needle in a haystack 'may be controlled as a unit by a particulartype of situation' (116); the words in a single part of speech, e.g. all adjectives,are under the control of n singlo set of subtle properties of stimuli (121); 'thosentence The boy runs a 8tore is under the control of an extremely complex stim-ulus situation' (335); 'He i8 not at all weU may function as a standard responseunder the control of a state of affairs which might also control He i8 ailing'(325); when an envoy observ~ events in a foreign country and reports upon his3tum, his report is under 'remote stimulus control' (416); the stntement This

i8 war may be a response to a 'confusing international ,situation' (441); thesuffix -ett is controlled by that 'subtle property of stimuli which we speak of asaction-in-the-past' (121) just as the -8 in The boy rum is under the control of

) such specific features of the situation as its 'currency' (332). No characterization" of the notion 'stimulus control' that is remotely related to the bar-pressing ex-

periment (or that preserves the faintest objectivity) can be made to cover aaet- d~plesliketJiese, in which, for example, the '~~trollin& stim-!!lqe'need not~y~p-~ing~.Q.r:t the responding organism.

Consider now Skinner's use-of the notion 'reSponse'. The problem of identi-fying units in verbal bchnvior has of coursc been a primnry concern of linguists,and it ~ms very likely that experimental psychologists should be able to pro-vide much-needed assistance in clearing up the many remaining difficulties insystematic identification. Skinner recognizes (20) the fundamental character ofthe problem of identification of a unit of verbal behavior, but is satisfied with~~~~ so vague and subjective that it does not really contribute to its solu-tion. ~e --umt-m-r-veffi"al~~~~E"the verbal operant-is defined as a class ofresponses 6f identifiable form functionally related to one or more controlJipgvari~b1es. No methooiS-Su-gg-e-st.edfordetermining in a particular instance what.--are the controlling variables, how many troch units have occurred, or where theirboundaries are in the total response. Nor is any attempt made to specify howmuch or what kind of similarity in form or 'control' is required for two physicalevents to be considered instances of the same operant. In short, no answers aresuggested for the most elementary questions that must be asked of anyone pro-posing a method for description of behavior. Skinner is content with what hecalls an 'extrapolation' of the concept of operant developed in the laboratory tothe verb~eid-:--rn~ typical Skinnerian experiment, the problem of identify-ing the unit of behavior is not too crucial. It is defined, by fiat, as a recordedpeck or bar-press, and systematic variations in the rate of this operant and its

34 L\XG UAGE , '"OLUME 35, NUMBER 1 (1959)

resistance to extinction are studied as a function of deprivatio11 and schedululgof reinforceme11t (pellet~). Thc opcraut is thus defined \vith respe-£!. ~_~PM.!JQY.:-

~~~_e~_P:e:ri~ental prOCc:dure~T~rfectlY-'reasonable; and""has led to manyInterestIng results. It IS, however, ~~~lpletely mea-uwgl..= to ape" oj: eKt:'.-.pe-la,~ins this concept ~perant to ordinary verba] ~hnyio.rA.Such 'extrapolation'leaves us with no \vay of justifying one or another decision about the units inthe 'verbal repertoire'. .

Skinner specifies 'response strength' as the basic datum, the basic dependentvariable in his fu11ctional analysis. In the bar-pressi11g experiment, responsestrength is defined in terms of rate of emission during extinction. Skinne.. hasargued' that this is 'the 'pnly datum that varies significantly and in the expecteddirection under conditions which are relevant to the "leanung process".' In thebook under review, rcSpo11se strc11gth is defined as 'probability of emission' (22).This definition provides a COmfOrtillg impression of objectivity, which, however,is quickly dispclled whe11 we look into the matter more closely. The term ~ba-bility' has some rather obscure meaning for Skinner in this book.' We are told,on tlac olle laand, that 'our cvidcnce for the contribution of each varia.blc [toresponsc strength) is bnsed on observtLtion of frequencies alone' (28). At thesamc time, it uppcars that frequency is a very misleading measure of strength,since, for example, the frequency of a response may be 'primarily attributableto the frequency of occurrence of controlling variables' (27). It is not clearho\v the frequellcy of a response can be attributable to anything BUT the fre-quency of occurrence of its controlling variables if we accept Skinner's viewthat the behavior occurring in a given situation is 'fully determined' by therelevant controlling variables (175, 228). Furthermore, although the evidencefor tlae contribution of each variable to response strength is bnsed on observationof frcquencies alone, it turns out that 'we base the notion of strength uponsc\"cral killds of cvidence' (22), ill particular (22-8): emission of the response(particularly in unusual circumstances), energy level (stress), pitch level, speedand delay of emission, size of letters etc. in writing, immediate repetition, and-t1 finul factor, rclcvullt but lniijlcuuillg-ovor-nll frequellcy.

Of course, Skinner recognizes that these measures do not co-vary, because(among other reasons) pitch, stress, quantity, and reduplication may have in-

.Are theories of learning necessary?, Peych. rIll. 57.193-216 (1950).

.And elsewhere. In his paper Are theories of learning necessary?, Skinner considers theproblem how to extend his analysis of behavior to experimental situations in which it isimpossible to observe frequencies, rate of response being the only valid datum. His answeris that 'the notion of probability is usually extrapolated to cases in which a frequencyanalysis cannot be carried out. In the field of behavior we arrange a situation in whichfrequencies are available as data, but we use the notion of probability in analyzing or for-mulating instances of even types of behavior which are not susceptible to this analysis'(199). There are, of course, conceptions of probability not based directly on frequency, butI do not lice how nny of thcsc apply to the cnscs thnt Skinner has in mind. I see no wayof interpreting the quoted passage other than as signifying an intention to use the word'probability' in describing behavior quite independently of whether the notion of proba-bility is at all relevant.

REVIEWS 35

.)

~rnal linguistic functions.1o However, he does not ho]d these conflicts to bevery important, since the proposed factors indicative of strength are 'fullyunderstood by everyone' in the cuJture (27). For examp]e, 'if we are sho\vn aprized work of art and exclaim Beautiful!, the speed and energy of the responsewill not be lost on the owner.' It does not appear totally obvious that in thiscase the way to impress the owner is to shriek Beautiful in a loud, high-pitchedvoice, repeatedly, and with no delay (high response strength). It may be equallyeffective to ]ook at the picture si]ent]y (long delay), and then to murmur Beau-tiful in a soft, low-pitched voice (by definition, very low response strength).

It is not unfair, I believe, to conclude from Skinner's discussion of rcsponscstrength, the 'basic datum' in functional analysis, that his 'extrapolation' ofthe notion of probability can best be in~rpreted ~, in effect~iB~ lfte:'-3-t.~an

..to use w d 'probability', with its favorable connotations of ob-j~y, as a cover term to parap rase suc ow-s us word~~-ii~-tCrest','intention', 'belief', and the like. This interpretation is fully justified by the\\'ay in which Skinner uses the terms 'probability' and. 'strength'. To ciw justonc example, Skinner defines the procct!8 of confirming an assertion in tlCicnccas one oC 'gencrating additional varinblcs to increase its probability' (425), andmore generlilly, its strength (425-9). If we take this suggestion quite literally,the degree of confirmation of a scientific assertion can be measured as a simplefunction of the loudness, pitch, and frequency with which it is proclaimed, anda genera] procedure for increasing its degree of confirmation wou]d be, for in-stance, to train machine guns on large crowds of peop]e who have been instructedto shout it. A better indication of what Skinner probably has in mind here isgiven by his description of how the theory of evo]ution, as an example, is con-firmed. 1'hiti 'single set of verbw responses ...is made more plausib]e-is strengthened-by sevcrw types of construction based upon verbn] responsesin geology, palcontology, gcnctics, and so on' (427). Wc urn no doubt to inwr-pret the terms 'strength' and 'probability' in this context as paraphrases ofmore Camilinr locutions such as 'justified be]ief' or 'warranted al(SCrtability', orsomctlling of the sort. Similar latitude of inwrprc~tion iti presumably expectedwhen we read that 'frequency of effective action accounts in turn for what wemay call tho listener's "belief" , (88) or that 'our belief in what someone tells

us is similarly a function of, or identical with, our tendency to act upon theverbal stimuli which he provides' (160).n

10 Fortunately, 'In English this presents no great difficulty' since. for example. 'relativepitch levels ...are not ...important' (25). No reference is made to the numerous studies ofthe function of relative pitch levels and other intonational features in English.

II The vagueness of the word 'tendency', &8 opposed to 'frequency', laves the latterquotation from the obvious incorrectness of the former. Nevertbelesa,a good deal of atreteb.ing is necessary. If 'tendency' bas anything like its ordinary meaning. the remark iaclearlyfalse. One may believe strongly the assertion that Jupiter has four moons. that many ofSophocles' plays have been irretrievably lollt, that the earth will burn to a crisp in tenmillion years, etc., without experiencing the slightest tendency to act upon these verbalstimuli. We may, of course. turn Skinner's assertion into a very unilluminating truth bydefining 'tendency to act' to include tendencies to answer questions in certain waye, undermotivation to say what one believes is true.

I

..


I think it is evident, thcn, that Skinner's use of the terms 'stimulus', 'control','response', and 'strength' justify the general conclusion stated in the last para-graph of §2 above. The way in which. tkese terms are brought to bear on theactual data indicates that we must interl>ret them as mere paraphrases for thepopular vocabulary commonly used to describe behavior, and as having noparticular connection with the homonymous expressions used in the descriptionof laboratory experiments. Naturally, this terminological revision adds no ob-j~ity to the familiar 'mentalistic' mode of description.

~ 4 .~The other fundamental notion borrowed from the description of bar-pre .g experiments is 'reinforcement'. It raises problems which are similar, and -

~~~eveJ1'more serious. In Behavior of organi8m8, '~ ogerat.inn nt' minfQrcAmAnt. i~defincd as ~~~tion of ace~nkind of stimu~_a tem~ral relatio~~it~ ('l~ra stimulus or resPOn~. A reinforcing stimulus is defined as such byits. -pQwer to proouce.-thA rp_~lllt.Lng "h"n:~ t;n &*~ngth). There is no circularityabout this: some stimuli are found to produce the change, others not, and theyurc cl~'i.'Iificd ~ J:.Q.!!1(orcillg ulld lIon-reinforcing accordingly' (62). This is a per-fectly appropriute dcflnitionJt fortfle'study of schcdulcs of reinforcement. It is.~r(e~tlv U~§§I 110wever, i~discussion of real-life behay!or. ~_~~~n~omehaw-~cterizetQ~t_imuli which are reinforcing (and the situations andCOnditIons Unde~~which--t1iey-are reinforcTn-g)-:--ConSi'der first of all the status ofthe basic principle that Skinner calls the '~~ ofcon~_~~ ~w of effect).It reM~:'i~-e.£~~-E-peran~_~ f~ed by ~rese~c~ of_a ~iof~rci!!g.timulus, the strength is increased' (Be~or oj organi8m8 21). As 'reinforce-

ment' was defin , -IS aw omesa tautolo~:11 For Skinner, learning is justchange in response strength." Although the statement that presence of reinforce-ment is a sufficient condition for learning and maintenance of behavior is vacu-ous, thc cluim that it is a neccssary condition may have some content, dependingon how the class of reinforcers (and appropriate situations) is characterized.~inn!!.~ does ~ke h,yery clear t!t~t_~J:I: hi~ view- ~orcement is a nOCC88lJ.ryponditiofi- for-lalllr;u"i!c_learni!!g. and for -the-can~~_availability of ~~!~

~e..;ponses!n.the ad~lt.li However, .t~~~-~f!.~~~.~.~~~~~~nt' as11 One should add, however, that it is in general not the atimulUB as such that is reinforc-

ing, but the stimulus in a particular situational context. Depending on experimental ar-rangement, a particular physical event or object may be reinforcing, punishing, or un-noticed. Because Skinner limits himself to a particular, very simple experimentalarrangement, it is not necessary for him to add this qualification, which would not be atall es.sy to formulate precisely. But it is of course necessary if he expects to extend hisdescriptive system to behavior in general....Iji)'fhis hs.s been frequently noted.

It See, for example, Are theories of learning necessary? 199. Elsewhere, he 8Uggests thatthe term 'learning' be restricted to complex situations, but these are not characterized.

1. 'A child acquires verbal behavior when relatively unpatterned vocalisations, aelec-tively reinforced, gradually assume forms which produce appropriate consequ~n~ in agiven verbal community' (31). 'Differential reinforcement shapea up all verbal fonDa, aDdwhen a prior stimulus enters into the contingency, reinforcement is responsible for itsresulting control... The availability of behavior, its probability or strength, dependa onwhether reinforcements C<mtinu, in eft'"ect and according to what schedules' (~). Else-where, frequently.

('

..I:;'$Z,

37REVIEWS

Skinner uses it in the book under revie\v makes it entirely pointless to inquireinto the truth or falsity of this cla~Ex.amining the instances of what Skinnercalls 'reinforcement', we find that not even the requirement that a reinforcer bean identifiable stimulus is taken seriously. In fact, the term is used in such 11.\vay that the assertion that reituorcement is necessary for learning and contin-ued availability of behavior is likewise empty.

To show this, we consider some e~ro.p.!~_<!L'.~~em.~nt'. First of all, wefind a heavy appeal to automatic self-reinforcement. Thus, 'a man talks tohimself .., because of the reinforcement he receives' (163); 'the child is reinforcedautomatically when he duplicates the sounds of airplanes, streetcars ..,' (164);'the young child alone in the nursery may automatically reinforce his own ex-ploratory verbal behavior when he produces sounds which he has heard in thespeech of others' (58); 'the speaker who is also an accomplished listener "knowswhen he has correctly echoed a response" and is reinforced thereby' (68) i think-ing is 'behaving which automatically affects the behaver and is reinforcing be-cause it does so' (438; cutting one's finger should thus be reinforcing, and anexample of thinking); 'the verbal fantasy, whether overt or covert, is automati-cally reinforcing to the speaker as listener. Just as the musician plays or composes,,-hat he is reinforced by hearing, or as the artist paints what reinforces himvisually, so the speaker engaged in verbal fantasy says what he is reinforced byhearing or writes what he is reinforced by reading' (439); similarly, care inproblem solving, and rationalization, are automatically self-reinforcing (442-3)..We,£an ~so reinforce som~~~~g_~~ha:vio~ aw:b. (since thisrurcsou-t~nss-oraversTVe stimulations, 1672, by not emitti_~~~!.~! -~~~(~~pl1"!~.!!U1!..~ng attention, 199), orby~acting appropnately on somefuture occasion (152: 'toe-strength or [the speaker's] behavior is dotcrminedmainly by tho behavior which tho listenor will exhibit with rcspcct to a givenstate of affairs'; this Skinner considers the general case of 'communication' or'letting the listener know'), In most such casas, of COUfsc, tho spookor is notpresent at the time when the reinforcement takes place, as when 'the artist ...is reinforced by the effects his works have upon ." others' (224), or when thewriter is reinforced by the fact that his 'verbal behavior may reach over cen-turies or to thousands of listeners or readers at the same time. The writer maynot be reinforced often or immediately, but his net reinforcement may be great'(206; this accounts for the great 'strength' of his behavior), An individual mayalso find it reinforcing to injure someone by criticism or by bringing bad news, orto publish an experimental result which upsets the theory of a rival (154), todescribe circumstances which would be reinforcing if they were to occur (165),to avoid repetition (222), to 'hear' his own name though in fact it was not men-tioned or to hear nonexistent words in his child's babbling (259), to cla.rify orotherwise intensify the effect of a stimulus which serves an important discrim-inative function (416), etc.

From this sample, ~ be~n~t~ notion o[mnforcement h~~lost whatever objective meaniit it;maye ;er have hadJRUnnfu-g-t;h~ these

-ex-amples;-We-see-that a person can be reinforced though he emits no response atall, and that the reinforcing 'stimulus' need not impinge on the 'reinforced person'

..


or need not even exist (it is sufficient that it be imagined or hoped for). When weread that a person plays ".hat music he likes (165), says what he likes (165),thinks what he likes (438-9), reads what books he likes (163), etc.,.BECAUSE~nds it reinforcing to do so, or ~~~~ books or inform otbers of facts

q,~_~E. ~~- W ~ope ~we ~l.tJ!ny~~~ ~f -~ ~d~r ~~ li~~n~)Y.e.-1:a.n only cOD,rt.lude ~t the.-~~!mQ~I!1~' .A~ a-~.!!~~~ The phrase\.~ reinforced b~~tlIDulus, ~~-fairs, event, e~)' is ~ng ~ as 0: co~~~m f-O:r-!X w.a.nts..Y', 'X ~.¥!s '~wishes_~~_)~~~~_.the~'. etc.)Invoking the term 'reinforcement' has no ex-planatory force, and any id-ea'tJiit this paraphrase introduces any new clarity orobjectivity into the description of wishing, liking, etc., is a serious delusion. Theonly effect i~!Q obscure the important differences among the notions beingparaphrasedJOnce we recognize the latitude~;tb w!lich!!ie term 're~rnent~is being used, many rather startling comments lose their initial effectJ-ror in-s ce, a vlor 0 a lve is 'co~~ --entire~ bl.~contingencies of reinforcement' }(150). What as nhoped for from tbe psy-~~88Orne indication how the casual and informal description of everydaybehavior in the popular vocllbulary can be explained or clarified in terms of tbenotions developed in careful experiment and observation, or perhaps replacedin terms of a better scheme. ~mere terminQIQgicaJ mviRinB, in ~bi~h II. t.ermborroWQ9._f~ the l~wy-_i_~~d yith the lull vam~ss o!~e QrdWiCY-:v~-"b,,1A~, is 2!-~~J!~_~ble interest.

m- seem~ that Skinner's claim- thaOt.-aUverbal behavior is acquired and main-tained in 'strength. through reinforcement is quite em~ ~~i~otion ofreinforcement h~~Q cLear~tent, f~nctioniDg...oDl.Y:-..Y:.a-r.over term foL.ADy

~~ac~ble or not, ~w ncquisiti.smor maiDtPcnan~~bAljJeha..~~~ Skini1er's use 6f"the term 'conditioning' suffers from a similar difficulty.Pavlovilln and opcrant conditioniug are processes about which psychologistshave developed real understanding. Instruction of human beings is not. Theclaim that instruction and imparting of information are simply matters of condi-tioning (357-6G) is pointless. rrhe claim is true, if we extend the term 'condition-ing' to cover these procc8SCs, but we know no more about them after havingrevised this term in such a way as to deprive it of its relatively clear and ob-jcctive character. It is, as far as we know, quite false, if \\.e use 'conditioning' inits literal sense. Similarly, when we say that 'it is the function of predication tofacilitate the transfer of response from one term to another or from one objectto another' (361). we have said nothing of any significance. In what se~ isthis true of the predication Whale8 are mammals? Or. to take Skinner's example,what point is there in saying that the effect of The telephone is out oj order onthe listener is to bring behavior formerly controlled by the stimulus out oj orderunder control of the stimulus telephone (or the telephone itself) by a process ofsimple conditioning (362)? What laws of conditioning hold in this case? Further-

If Talk of schedules of reinforcement here i8 entirely pointless. Bow are we to decide.for example, according to what 8chedule8 covert reinforcement is 'arranged'." ill Wnk:incor verbal fantasy, or what the 8cheduling is of 8uch factors 8.8 silence. speech. and ap-propriate future reaction8 to communicated information?.,

~::-4:i;[,f~~

REVIEWS 39

more, \vhat behavior is 'controlled' by the stimulus out oJordcr, in the abstrt~t?Depending on the object of ,,'hich this is predicated, the present state of motiva-tion of the listener, etc" the behavior may vary from rage to pleasure, from fixingthe object to throwing it out, from simplY' not using it to trying to use it in thenormal way (e,g. to see if it is really out of order), and 80 on. To speak of 'con-ditioning' or 'bringing previously available behavior under control of a newstimulus' in such a case is just a kind of play-acting at science, Cf, also foot-note 43,

5. The claim that careful arrangement of c~nt_ing~ncieso!~nforcement bythe verbal community is a necessary condition for language leiirnfnii:ria8 ap-peared, in one form or another, in many places}' Since it is based not on actualobscrvation, bu.Lon analogies to laboratory~udYQLIQ\vcLQ!:g~~ is im-'portant to detennine the status of the underlying assertion within experimentalpsychology proper, The most common c.hJu:1l~~ri "II ti98 Q£ rAin( """"meat.-{ one\vhich Skinner explicitly rejects, incidentally) js in tenns_.Qf drive redu~ion Thischaract.crizution CUll 00 given tiuust.unce by uefuling drivel; in ~me way inde-pendently of "'hat in fact is learned~f.Jl..dJ:~a...pa8iulAt~~ft~Re ~~--f.J.h.cf~~ le~t:!!iJ1~ t~1!;es PlAce..t t~e claim ~hat J'Pin(n~m8A~ ie~~y for-l~n~ill..yai~..~om.e ~-~~ 9,.sit~~..§kfriDeA~~~~here-is a-n-e-xtensive literaturc on tho questiQn of \\'hether thore can beJ.e.n,roi_~ with-out driv~:-~uction ~t lew'ning) , The 'classical' experiment of Blodgettind1ca~a-that rats who had CXPToIred-/i maze without reward sho\ved a markcddrop in number of errors (as compared to a control group which had not ex-plored the maze) upon introduction of a food reward, indicating that the rathad lcarncd thc structure of thc maze \vithout reduction of the hungcr drivc,Drive-reduction theorists countered with an exploratory drive \vhich was re.uuccd, during the prcreward lcarning, and claimcd that a sligJlt uccrcmcnt inerrors could be noted before food reward, A \vide variety of experiments, withsomewhat conflicting results, have been carried out with a similar design,"Few investigators still doubt the existence of the phenomenon, Hilgard, in hisgeneral review of learning theory ,It concludes that 'there is no longer any doubtbut that, under appropriate circumstances, latent learning is demonstrable,'

More recent work has shown that novelty and variety of stimulus are sufficient

IT See, for example, Miller and Dollard, Social learning and imitation 82-3 (New York,1941), for a discussion of the 'meticulous training' that they.seem to consider necessary fora child to learn the meanings of words and syntactic patterns. The same notion is implicitin Mowrer's speculative account of how language might be acquired, in Learning tJaeOT¥ Gndperaonalily dynamica, C.hapter 23 (New York, 1950). Actually, the view appears to bequite general.

II For a general review and analysis of this literature, see Thistlethwaitc, A criticalreview of latent learning and related experiments, P'ych. bull. 48.97-129 (1~1). MacCor-quodale and Meehl, in their contribution to Modem learning theor'J, carry out a seriOUI andconsidered attempt to handle the latent learning material from the standpoint of drive-reduction theory, with (as they point out) not entirely satisfactory results. Thorpe review8the literature from the 8tandpoint of the ethologist, adding also material on homing andtopographical orientation (Learning and inrtinct in animal, [Cambridge, 19M).

It Theoriu of learning 214 (1956).


to arouse (:uriosity in the rat and to motivate it to explore (visually), and infact, to learn (since on a presentation of two stimuli, one novel, one repeated,the rat will attend to the novel one) jlO that rats will learn to choose the armof a single-choice maze that leads to a complex maze, running through thisbeing their only 'reward'jI1 that monkeys can learn object discriminations andmwn~n tJ1eir performance at a high level of efficiency with visual exploration(looking out of a window for 30 seconds) as the only rewardjll and, perhaps moststrikingly of all, that monkeys and apes will solve rather complex manipulationproblems that are simply placed in their cages, and will solve discriminationproblem!; witJ1 only cxploration and manipulation as incentives.1I In these c~---,solving,th~nhlo~apparent1y its own 'Mward:... Results of this kind can behandled.by-reinf .if the are w"illin-t:o-~p c-;-'-.exglorll.t.i11n~ a~-:~~~ d~~es, o!-to~ula~ ~mehow aoo~~qu~-driv,.~ for whi~h t.herelS no evidpn~p outside n( t,he fact that l~';; takespl!l.l'.p- in t.hP;~;:~-=- s-¥&~: _a_- ._v -~~~

.Dcrlync, Novelty and curiosity aa determinants of exploratory behavior, Brit. jour. 01psvc.\. 41.6&-80 (1950); id., PerceptuBI curiosity in the rat, JO1£r. 01 compo pAlInol. p.vcA.48.238-46 (1955); Thompson and Solomon, Spontaneous pattern dilcrimination in the rat,ibid. 47.104-7 (1954).

II Montgomery, The role of the exploratory drive in learning, ibid. ~. Many otherpBpers in the same journal are designed to show that exploratory behavior is a relativelyindependent primary 'drive' aroused by novel external .timulation.

n Butler, Discrimination learning by Rhesus monkeys to vilUal-exploration motivation,ibid. 46.95-8 (1953). Later experiments showed that this 'drive' is highly persistent, asopposed to derived drives which rapidly extinguish.

SI Harlow, Harlow, and Meyer, Learning motivated by a manipulation drive, Jour. up.psych. 40.22&-34 (1950), and later invcstigations initiated by Harlow. Harlow has beenparticularly insistent on maintaining the inadequacy of physiologically baaod drivOi andhomcflHtatic nocd statCIi for oxplainiug tho porsistenoo of motivation and rapidity of learn.ing in primates. He points out, in many papers, that curiosity, play, exploration, and ma-nipulation are, for primates, often more potent drives than hunger and the like, and thatthcy show nono or tho charl~ct.criltics or uctjuirod drivOli. Hobb ullu prl*Qnt.llllehavioraJ andlIupporting ncurulogicul ovidcnce in support of the view that in higher animals there is apositive attraction in work, risk, puazle, intellectual activity, mild fear and frustration,etc. (Drives and the CNS, P,vch. rev. 62.243-64 [1955]). He concludes that 'we need Dotwork out tortuous and improbable ways to explain why men work for moDey, why childrenlearn without pain, why people dislike doing nothing.'

In a brief note (Early recognition of the manipulative drive in monkeys, Briti.hjournal01 animal behaviour 3.71-2 [1955]), W. Dennis calls attention to the fact that early investi-gators (Romanes, 1882; Thorndike, 1901), whose 'perception was reJatively unaffected bylearning theory, did note the intrinsically motivated behavior of monkeys', aJthouih, heR.8scrt.ll, no similar observations on monkeys have been made until Harlow's experiments.He quotes Romanes (Animal intelligence [1882]) as saying that 'much' the most striking fea-ture in the psychology of this animal, and the one which is least like anything met with inother animals, was the tirele81 spirit of investigation.' Analogous developments, in whichgcnuinc discovcries hBvc blinded syst.cmatic investigators to the important inaiihta ofearlier work, are easily found within recent structural linguistics as well.

S. Thus J. S. Brown, in commenting on a paper of Harlow's in Currenl theory and ruearchin motivation (Lincoln, 1953), argues that 'in probably every instance [of the experimentscited by Harlow] an ingenious drive-reduction theorist could find IOme fragment of fear,insecurity, frustration, or whatever, that he could insist was reduced and hence was rein.forcing' (53). The same sort of thing could be sald for the ingenious phlogiston or ethertheorist.

~:~'~

~f~

,;~~~t~

41REVIEWS

There is a variety of other kinds of evidence that has been offered..t.o.challeoge..the view that drive-reduction is necessary for learning. Results on sensory-sensory conditioning have been interpre¥ as demonstrating learning withoutdrive-reduction.'i Olds has reported reinforcement by direct stimulation of thebrain, from which he concludes that reward need not satisfy a physiological needor withdraw a drive stimulus.'1 The phenomenon of imprinting, long obscrved byzoologists, is of particular interest in this connection. Some of the most complexpatterns of behavior of birds, in particular, are directed towards objects andanimals of the type to which they have been exposed at certain critical earlyperiods of life.17 Imprinting is the most striking evidence for the innate disposi-tion of the animal to learn in a certain direction, and to react appropriately topatterns and objects of certain restricted types, often only long after the originallearning has taken place. It is, consequently, unrewarded learning, though theresulting patterns of behavior may be refined through reinforcement. Acquisitionof the typical songs of song birds is, in some cases, a type of imprinting. Thorpereports studies that show 'that some characteristics of the normal song huvcbeen loornt in the earliest youth, before the bird itself is able to produoo anykind of full song'." The phenomenon of imprinting has recently been investi-gated under laboratory conditions and controls with positive results..t

Phenomena of this general type are certainly familiar from everyday experi-ence. We recognize people and places to which we have given no particularattention. We can look up something in a book and learn it perfectly well withno other motive than to confute reinforcement theory, or out of boredom, or idlecuriosity. Everyone engaged in research must have had the experience of workingwith feverish and prolonged intensity to write a paper which no one else willread or to solve a problem which no one else thinks important and which willbring no conceivable reward-which may only confirm a general opinion thatthe rcscarcher is wasting his time on irrelevancies. The fact that rats and monkeysdo likewise is interesting, and important to show in careful experiment. In fact,studiCH of behavior of the typo mentioned above huve an indepcndcnt and posi-tive significance that far outweighs their incidental importance in bringing intoquestion the claim that learning is impossible without drive-reduction. It is notat all unlikely that insights arising from animal behavior studies with thisbroadened scope may have the kind of relevance to such complex activities asverbal behavior that reinforcement theory has, so far, failed to exhibit. In anyevent, in the light of presently available evidence, it is difficult to see how anyone

II Cf. Birch and Bitterman, Reinforcement and learning: The process of sensory inte-gration, Pa~ch. rev. 56.292-308 (1949).

18 See, for example, his paper A physiological study Qf reward in McClelland (ed.),,5tudiea in motivation 134-43 (New York, 1955).

IT See Thorpe. op.cit., particularly 115-8 and 337-76, for an exoellent discuasion of thisphenomcnon, which has bccn brought to prominence pa.rticularly by the work of K. Lorenz(cf. Der Kumpan in der Umwelt des Vogels, parts of which a.re reprinted in English transla-tion in Schiller (ed.), Instinctive behavior 83-118 (New York, 1957).

II Op.cit. 372. '

II See e.g. Jaynes, Imprinting: Interaction of learned and innate behavior, Jour. oj compoph~aiol. pa~ch. 49.201-6 (1956), where the conclusion is reached that 'the experimenta provethat without any observable reward young birds of this species follow a moving 8timulusobject and very rapidly come to prefer that object to others.' ,

~~~


can be willing to claim that reinforcement is necessary for learning, if reinforce-ment is taken seriously as something identifiable independently of the resultingchange in behavior. -,

Similarly, it seems quite beyond question that children acquire a good dealof their verbal and nonverbtl.1 behavior by .e~~a1 ~~serv@I) and im!-~~.Q_~-2fa.dults and other children. ~ It is simply not true that children can learn language

only through 'meticulous care' on the part of adults who shape their verbalrepertoire through careful differential reinforcement, though it may be that suchcare is often the custom in academic fa.milies,--It is a common obscrvation thata young child of immi rant aren s ma learn a second lan age in the st

,._r_om ren ~i~a.m~~~T and that his speec may be com-p.l~ly_611"nt nn-J CQJ:~':!~~!!~2~ -wIiiIeUe-subtleties ~t-become

Second nature to the child may elude his parents despite high motivation andcontinucd practicc. A child may pick up a large part of his vocabulary and 'feel'for sentence structure from television, from reading, from listening to adults,etc. J';vcn 11 very young child who hll.ll not yet ~quircd a minimal repcrtoire fromwhich to form new utterances may imitate a word quioo wcll on an early try,with no atoompt on tlle part of his parents to teach it to him. It is also perfectlyobvious that, at a later stage, a child will be able to construct and understandutooranccs which are quire new, and are, at the same time, acceptablc sentencesin his language. Every time an adult reads a newspaper, he undoubtedly comesupon countless new sentences which are not at all similar, in a simple, physicalscnsc, to any that he has heard before, and which he will recognize as sentencesand understand j he will also be able to detect slight distortions or misprints.Tnlk of 'stimulus gcncra.1ization' in such a case simply perpetuatetl the mysteryunder a new title. These abilities indicate that there must be fundamentalprocCI!SCH at work quire indcpclldclltly of 'food back , from tllC cJlvironment. Ihave bccn uble to find no support whutsoever for the doctrine of Skinner andothers that slow and careful shaping of verbal behavior through ditTcrential re-illfurccJncnt it! WI ubtiOlutc nec~ity..:j1.~mfQ~~~.~nt._t.!1~.~rewly_.mQ.Y~~

.Of course it i8 perfectly po88ible to incorporate thi8 fact within the Skinnerian frame"work. If, for example, a child watches an adult using a comb and then, with no instruction'trie8 to comb hi8 own hair, we can explain thi8 act by saying that he performs it becausehe find8 it reinforcing to do 80, or because of the reinforcement provided by behaving like aperson who i8 'reinforcing' (cf. 164). Similarly, an automatic explanation i8 available forany other behavior. It seem8 8trange at first that Skinner paya 80 little attention to theliterature on latent learning and related topics, considering the tremendous reliance thathe places on the notion of reinforcement; I have 8een no reference to it in his writinp.Similarly, Kcllcr and SchocnfeJd, ill what appcars to be the only tex.t written under pre-dominantly Skinnerian influence, Principles of psychology (New York, 1950), diamiaa thelatent-learning literature in one sentence 8.8 'be8ide the point', serving only 'to ob8cure,rather than clarify, a fundamental principle' (the law of effect, 41). However, thi8 neglecti~ pcrfccUy appropriatc in Skinncr'lj CalC. 'fo the drive-reductioni8t, or anyone else forwhom the notion 'reinforcement' h8.8 80me 8ub8tantive meaning, these experimeI&ts andobscrvations are important (and often embarr8.8sing). But in the Skinnerian sense of theword, neither the8e results nor any conceivable others can cast any doubt on the claim thatreinforcement is esscntial for the acquisition and maintenance of behavior. Behavior cer-tainly has lome concomitant circumstance8, and whatever they are, we can call them'reinforcement'.

REVIEWS 43

assumption that there be such meticulous care., it seems best to regard thissimply as,A.~~t,in M absurdum arg'.1Inent~ainst.this approach. It isaiso noteasy to ~4_~.I!X_bas~.(or, f~~ma_t~~..ftoAtach very much con~J&Jtb_eclaim t re or~!!l contingencies set. ~ h~ the verbal commW1i!Y ~ theei!l.gLe~~L~~ble for maintaining th~-.§t.rn!!~1!. Qf v~rb:l.l behaviol\ Thesources of the 'strength' of this behavior are almost a total mystery at present.

~orcempnt. 1JDdoubt&Ql!.~_~ significant role, but so do ~y~e_~y ~-v~oua1-faGt.D~~bo~t which ~othing-seno~.'knowri in the --£a8c ~r' hUrtlM

~~Asn fs. ...f Ian . ed . I h .ft:~r;""ftft~~ t.ar ~ acqulSltlon 0 guage lS concern J It seems c car t at r~\I"" d~...~n ,

casual observation, and natural inquisitiveness (coupled with a strong tendency""

to imitate) are important-factors,as is the remarkable capacity or tho child togerfeiiJize, hypothesize, and 'process information' in a variety of very Kpecial andapparently highly complex ways which we cannot yet dCHCribc or begin to under-btand, and which may be largely innate, or may develop through some IIOrt oflearnin~ or through maturation or the nervous Aystem. Thc manner in which HuchflWtorH opcrate and interact in language acquisition is completely unknown. It isclear that what is ncccssary in such a cnsc is research, not dogmatic arid pcr-fectly arbitrary claims, based on analogies to that small part of the expcrimentalliterature in which one happens to be interested.

The pointlessness of these claims becomes clear when we consider the well.known difficulties in determining to what exten~rn~,rlLC..tYre, ~~ra~,a~__l~arning are-xesponsib1e fQ_r:_~~~ulA.rf.Qrm of a skilled or comolex 00'--t:Q..rmance.JI To take just one example,a the gaping response or a nestling thrushif! at first released by jarring or the nest, and, at a later stage, by a moving objectof specific size, shape, and position relative to the nestling. At this later stagethe response is directed towards the part of Ule stimulus object correspondingt,o the purent's hcad, und churacterized by u complex configuration of Htimulithat can be precisely described. Knowing just this, it would be possible to con-Ktruct ,~ Rpcculative, learning-theoretic accotmt or how this HC<IU<1I1CQ of I>chuviorpattenaH might have developed through a process of differential reinforcement,,~nd it would no doubt be possible to troin ru.ts to do something similur. I-Jowever,there appears to be good evidence that these responses to fairly complex 'signf!timuli' are genetically dctermined and mature without learning. Clcarly, thepossibility cannot be discounted. Consider now the comparable case or a childimitating new words. At an early stage we may find rather gross correspond-ences. At a later stage, we find that repetition is of course far from exact (i.e.

II Tinl>ergen (op.cit., Chapter VI) reviews some aspects of this problem, discussing the

primary role of maturation in the development of many complex motor patterns (e.g.flying, swimming) in lower organisms, and the effect of an 'innate disposition to learn' incertain llpecific ways and at certain specific 'times. Cf. also Schiller, Inslinclille b'~or26&-88, for B dillcussion of thc rolo of mnturing motor patterns in ~ppnr~lItly inllight(ul 00-havior in the chimpanzee.

Lenneberg (Language, ellolu lion , and purpoaille behallior, unpublished) presents a veryinteresting discussicn of the part that biological structure may play in tho acquiaition oflanguage, nnd the dangers in neglecting this possibility.

12 From among many cited by Tinbergen, op.cit. (this on page 85).


it is not mimicry, a fact which itSelf is interesting), but that it reproduces thehighly complex configuration of sound features that constitute the phonologicalstructure of the language in question. Ag&iD, we can propose a speculative ac-count of how this result might have been 'obtained through elaborate arrange-ment of reinforcing contingencies. Here too, however, it is possible th~ abiljJ;y~~l~t~~e _C_~1I1 ~u:u~i tor¥~tho~~£~ ~~~pologica1lyrne.Y.an1-may develop largely independ~tly of reinforcement. th~nugh gAn_Ati-

0--,.-- callJ.9~~_~-~~..rro the extent that this is true, an account of the

development and causation of behavior that fails to consider the structUfe.of..theQ~~ism will providc no und~tanding of the real processesrinvolved.

It is often argued that experience, rather than innate capacity to handle in-fonnation in certain specific ways, must be the factor of overwhelming dom-inance in determining the specific character of language acquisition, since achild speaks tho languagc of the group in which he lives. But this is a superficialargument. As long as we are speculating, we may consider the possibility thattho brain hM evolvcd to the point whore, givcn an input of ob6Crved Chinese~ntences, it produces (by an 'induction' of apparently fantastic complexity andsuddcnncSH) the 'rulcs' of Chin~ grammar, and given an input of observedEnglish ~ntences, it produces (by, perhaps, exactly the same process of induc-tion) the rules of English grammar; or that given an observed application of aterm to certain instances it automatically pred~cts the extension to a clDo98 ofcomplexly related instances. If clearly recognized as such, this speculation isneither unreasonable nor fantastic; nor, for that matter, is it beyond the boundsof possible study. There is of course no known neural structure capable of per-formin~ this task in tho spccific ways that obscrvation of tho resulting bchaviormight lead us to postulate; but for that matter, the structures capable of ac-counting for even thc simpl~t kindH of lCll.rning h"vc Himilllrly dclicd detcctioJJ."

Summarizing this brief discussion, it seems that there is neither empiricalevidence nor any known argument to support any SPECIFIC claim about Ulercll,tive importnncc of 'fccdback' from tllC environment and tllC 'indepcJldentcontribution of the organism' in the process of language acquisition.

6. We now turn to the system that Skinner develops specifically for the de-scription of verbal behavior. Since this system is based on the notions 'stimulus','response', an~int'or.~~~n~9we can ~~I!.c:l~rom the preceding sections

~!!J!..I~~~d arbitr&IY.:.1:or reasons noted in §1, however, I think itis important to see in detail how far from the mark any analysis phrased solelyin these terms must be and how completely this system fails to account for thefacts of verbal behavior.

Consider first the term 'verbal behavior' itself. This is defined as ibehavior

II Cf. LashJey, In search of the engram, S"mpolium of the Societ" for EzperimcntalBiolog" 4.454-82 (1950). Sperry, On the neural buil of the conditioned response, Brit""journal of animal behaviour 3.41-4 (1955), argues that to account for the experimental resultlof LashJey and others, and for other facts that he cites, it is neCe88ary to auume thathigh-level cerebral activity of the type of insight, expectancy, etc. is invoJved even in simpJeconditioning. He states that 'we still Jack today a SAtisfactory picture of the underlyin,neural mechanism' of the conditioned response. -

~'

\~:,

REVIEWS 45

reinforced through the mediation of other persons' (2). The definition is clearlymuch too broad. It would include as 'verbal behavior', for example, a rat pressingthe bar in a Skinner-box, a child brushing'his teeth, a boxer retreating beforean opponent, and a mechanic repairing an automobile. Exactly how much ofordinary linguistic behavior is 'verbal' in this sense, however, is something of aquestion: perhaps, as I have pointed out above, a fairly small fraction of it, ifany substantive meaning is assigned to the term 'reinforced'. This definition is

entl refin e additional provision tnat the mediatin res nse of!:tIe"reinforcing ~~e 'listener')lD.us .yin..Qrdel: in rein_force tb~ behavior ~ the s~ker'(225, italics his). This still coversthe examples given above, If we can assume that the 'reinforcing' behavior oftho psychologist, the parent, the opposing boxer, and the paying customer aretho rcsult of appropriate trnining, which is perhapB not unrcl18Onablc. A signifi-cant part of the fragment of linguistic behavior covered by the earlier definitionwill no doubt be excluded by the refinement, however. Suppose, for example,that whilc cro88ing the stroot I hour someone shout Watch out lor the car andjump out of the way. It can hardly be proposed that myjumpmg (tho mediating,reinforcing response in Skinner's usage) 'Was conditioned (that is, I was trainedto jump) precisely in order to reinforce the behavior of the speaker. Similarlyfor a wide class of cnses. Skinner's assertion that with this refined definition 'wenarrow our subject to what is traditionally recognized as the verbal field' (225)appears to be grossly in error. .

7. Verbal operants are classified by Skinner in terms of their 'functional' rela-tion to discriminated stimulus, reinforcement, and other verbal responses. Amartd is defined as 'a verbal operant in which Ute response is reinforced by achnrlLCtcristic confiCqucncc and is thcrcforo under the functionnl control of rolc-VILllt conuitiull~ uf dcprivution or uvcrmvc stimulatiuJl' (35). 'fhis it! moont toinclude questions, commands, etc. Each of the terms in this dcfmition raises ahol;t of problcms. A mand such as PaB8 the Ball i~ a cll1K8 of rcHponllCH. Wc cannott~ll uy uul;Crving thc form of u roSpO111;C whcthcr it bclongli W this cl~ (Skirnlcris vcry clear about this), but only by identifying the controlling variables. Thisis generally impossible. Deprivation is defined in the bar-pressing experiment interms of length of time that the animal has not been fed or permitted to drink.In the present context, however, it is quite a mysterious notion. No attempt ismade here to describe a method for determining 'relevant conditions of depriva-tion' independently of the 'controlled' response. It is of no help at all to be told(32) that it can be characterized in terms of the operations of the experimenter.If wc define deprivation in terms of clapsed time, then at any moment a personis in countless states of deprivation." It appears that we must decide that therelevant condition of deprivation was (say) salt-dcprivation, on'the bll.8is of thefnct that the speaker asked for salt (the reinforcing community which 'sets up'

~4 Furthermore, the motivation or the speaker does not, except in the simplest cues,correspond in intensity to the duration or deprivation. An obvious counter-example iawhat Hebb bas called the 'salted-nut phenomenon' (Organization oj behavior 199 [NewYork, 1949). The difficulty is or course even more serious when we consider 'deprivations'not related to physiological drives.

.; .l~~:-~~~\ .'f:'M',,',;%~, ,',. -;~. "~il. '


the mand is in a similar predicament). In this case, the assertion that a mand isunder the control of relevant deprivation is empty, and we are (contrary toSkinner's intention) identif~g the respqnse as a mand completely in tenns ofform. The word 'relevant' in the definition above conceals some rather ~nouscomplications.

In tlte case of the mand Paa8 the 8alt, the word 'deprivation' is not out of place,though it appears to be of little use for functional analysis. Sup~ howeverthat tlte speaker says Give me the book, Take.me for a ride, or Let me [LX it. Whatkinds of deprivation can be associated with these mands? How do we determineor measure the relevant deprivation? I think we must conclude in tltis case, asbefore, either that the notion 'deprivation' is relevant at ~ost to a minute frag-ment of verbal behavior, or else that the statement 'X is under Y -deprivation' isjust an odd paraphrase for 'X.wants Y', bearing a misleading and uitjustifiablec,onnotation of objectivity.

The notion 'aversive control' is just as confused. This is intended to coverthreats, beating, and the like (33). The manner in which avorHivo BtimulAtionfunctions iH Himply deHCribed. If A speaker has had A history of Appropriate re-inforcement (e.g. if A ccrtain response was followed by 'cessntion of the tltreat ofsuch injury-of events which have previously been followed by such injury andwhich are therefore conditioned Aversivc stimuli') then he will tend to give theproper responsc when the threat which had previously been followed by theinjury is presented. It would appear to follow from this description that a speakerwill not respond properly to the mand Your money or your life (38) unless hehas a past history of being killed. But even if tlte difficulties in describing themechaniRm of aversive control are someho\y-removed by A more careful analysis,it will be of little use for identifying operants for reasons similar to th~ men-tioned in the cnsc of dcprivAtion.

I t scorns, then, tltAt in Skinner's terms there is in most cases no way to decidewhether a given response is an instance of a particular mand. Hence it is meaning.Icss, within the tenns of his systom, to speak of tilC clwraclerib'tic conscquencosof a mand, as in tlte dcfinition above. Furthermore, even if we extend tlte systemso tltat mands can somehow be idclltified, we will have to face the obvious factthat most of us are not fortunate enough to have our requests, commands,advicc, and so on cllaructcristicnlly reinforced (they muy nevertheless exist inconsiderable 'strength'). These responses could therefore not be consideredmands by Skinner. In fact, Skinner eets up a category of 'magical mands' (48-9)to cover the case of 'mands which cannot be accounted for by showing that theyhave ever had the effect specified or any similar effcct upon Himilur occallions'(the word 'ever' in this statement should be replaced by 'characteristically').In these pseudo mands, 'the speaker simply describes tbe-J:eiR£o~ccmQRt &P~~~!!.1e...t.o-.A..givell state ofdep~ivDotiDl1_Q!:..~'le st;~ti~n~n 9t9_er-words,givcn tho meaning that \...e havc !>ccn lcd to asl!ign to 'rcinforcemcnt' and 'depri.vation', the speaker asks for \\'hat he wants. The remark that 'a speaker appearsto create new mands on the analogy of old ones' is also not very helpful.

Skinner's claim that his new descriptive system is superior to the traditiollalone 'because its terms can be defined with respect to experimental operations'

~~

"',;'...I

~

REVIEWS 47

(45) is, we see once again, an illusion. The statement 'X wants Y' is not clarifiedby pointing out a relution between rate of bar-pressing and hours of food-depri-vation; replacing 'X wants Y' by 'X is deprived of Y' adds no new objectivityto the description of behnvior. His furtbe~ claim for the superiority of the newanalysis of mands is that it provides an objective basis for the traditional classi-fication into requestIJ, commands, etc. (38-41). The traditionnl claBSification isin te~ of the intention of the speaker. But ~tlon;-SkiDner liorQ8;can~reduCed to contingencies of reinforcement, and, correspondingly, we can explainthe traditional claBSification in terms of the reinforcing behavior of the listener.Thus a question is n mand which 'specifies verblll action, and thc behnvior of thelistener permits us to classify it as a request, a command, or a prayer' (39). It isa request ,if 'the listener is independently motivated to reinforce the speaker';a command if 'the listencr'tI behavior is ...reinforced by reducing u threat'j aprnycr if the mand 'promotes reinforcement by gcncrntin~ an cmotional dispOtli-tion'. "l'he mand is advice if the listener is positively reinforced by the conse-quenccs of medinting the reinforcement of the speaker; it is a warning if 'bycurrying out the bchnvior specified by the speakcr the litltencr Cl!Ca~ {romavcrsive stimulntion'; nnd 80 on. All this is obviously wrong if Skinner is usingthe words 'request', 'command', etc., in anything like the sense of the corre-sponding English words. The word 'question' does not cover commands. Pleasepa88 the &all is n request (but not n question), whether or not the listener happensto be motivated to fulfill it; not everyone to whom a request is addressed isfuvornbly disposed. A rcsponse does not cease to be a command if it is not fol-lowed; nor does a question become a command if the speaker answers it becauseof an implied or imagined threat. Not nIl advicc is good advice, nnd n responsedocs not cease. to bc advice if it is not followed. Similarly, a warning mny bemi~uicJcdj hc<..<Ji/lg it mtLY CtLUtIC tLVOJ'Hivo tltimulution, IU/d ignoring it might bcpoRitivoly reinforcing. In Rhort, tho ontirc classification is beside tile point. Amoment's thought is sufficient to demonstrate the impossibility of distinguishingbetwccn requosts, commands, u.dvicc, etc., on the basis of tho behavior or dis-pot;ition of tho particular listener. Nor can we do this on the basis of the typicalbehavior of all listeners. Some ndvice is never taken, is nlways bad, etc., andsimilarly with other kinds of mands. Skinner's evident satisfaction with thisanalysis of the traditional classificntion is extremely puzzling.

8. Mands are operants with no specified relation to a prior stimulus. A tact,on the other hand, is defined as 'a verbal operant in which a response of givenform is evoked (or at least strengthened) by a particular object or event orproperty of an object or event' (81). The examples quoted in the discussion ofstimulus control (§3) nrc nil tncts. The obscurity of the notion 'stimulus control'makes the concept of the tact rather mystical. Since, however, the tact is 'themost importnnt of verbnl opernnts', it is important to investigate the develop-ment of this concept in more detail.

We first ask why the verbal community 'sets up' tacts in the child-that is,how the parent is reinforced by setting up the tact. The basic explanation forthis behavior of the parent (85-6) is the reinforcement he obtains by the fact thathis contact with the environment is extended; to use Sirinner's example, the

.;'~i'"'.I:;~~~'

~"~~;~~~


child may later be able to call him to the telephone. (It is difficult to see, then,how first children acquire tacts, since the parent does not have the appropriatehistory of reinforcement). Reasoning in the same way, we may conclude that, ,the parent induces the child to walk so iJVIot he can make some money deliveringnewspapers. Similarly, the parent sets up an 'echoic repertoire' (e.g. a phonemicsystem) in the child bccauliC thiij makcij it casier to te1I.Ch him new vocabulary,and extending the child's vocabulary is ultimately useful to. the parent. 'In allthese cases we explain the behavior of the reinforcing listener by pointing to animprovement in the possibility of controlling the speaker whom he reinforces'(56). Perhaps this provides the explanation for the behavior of the parent ininducing the child to walk: the parent is reinforced by the improvement in hiscontrol of the child when the child's mobility increases. Underlying these modesof explanation is a curious view that it is somehow more scientific to attributeto a parent a desire to control the child or enhance his own possibilities for actionthan a desire to see the child develop and extend his capacities. Needless to say,no evidence is offered to support this contention.

Consider now the problem of explaining the response of the listener to a tact.Supposc, for example, that B hoars A sayfo% and reacts appropriat31y, loobaround, runs away, aims his rifle, etc. How can we explain B's behavior? Skinnerrightly rejects analyses of this offered by Watson and Bertrand Russell. His owneclually inadequate analysis proceeds as follows (87-8). We assume (1) 'that inthe history of [B] the stimulus fo% has been an occasion upon which lookingaround has been followed by seeing a fox' and (2) 'that the listener has somecurrent "interest in seeing foxes"-that behavior which depends upon a seen foxfor its cxecution is strong, and that the stimulus supplied by a fox is thereforereinforcing'. B carries out the appropriate behavior, then, because 'the heardstimulus fox is the occ~ion upon which turning and looking about is frcquentlyfollowed by tho reinforcement of liCCing a fox'; i.e. his bchuvior is a ditlCriminatodoperant. This explanation is unconvincing. B may never have seen a fox andmay have no current intcreRt in sccing one, and yet may rot\Ct ,~ppropriately tothe ijtimulus fox." Since exactly the same behavior may take place when neitherof the assumptions is fulfilled, some other mechanism must be operative here.

Skinner remarks several times that his analysis of the tact in terms of stimuluscontrol is an improvement over the traditional formulations in tenns of referenceand meaning. This is simply not true. His analysis is fundamentally the same asthe traditional one, though much less carefully phrased. In particular, it differsonly by indiscriminate paraphrase of such notions as denotation (reference)and connotation (meaning), which have been kept clearly apart in traditionnlformulations, in terms of the vague concept 'stimulus control'. In one tradi-tional formulation a descriptive term is said to denote a set of entities and toconnote or designate a certain property or condition that an entity must possess

" Just as he may have the appropriate reaction, both emotional and behavionol, to 8uchutterances as The volcano i, erupting or There', a homicidal maniac in the nezt room 1ritboutany previous pairing of the verbal and the physica1 8timulus. Skinner's discu88ion of Pav-lovian conditioning in language (164) is similarly unconvincing.

~

:~, .

:',.

,~

REVIEWS 49

or fulfil if the term is to apply to it." Thus the term vertebrate refers to (denotes,is true of) vertebrates and connotes the property 'having a spine' or somethingof the sort. This connoted defining property is called the meaning of the term.,Two terms may have the same referen~e.but different meanings. Thus it is ap-parently true that the creatures with hearts are all and only the vertebrates.If so, then the term creature with a heart refers to vertebrates and designates theproperty 'having a heart'. This is presumably a different property (a differentgeneral condition) from having a spine; hence the terms vertebrate and creaturewith a heart are said to have different meanings. This analysis is not incorrect(for at lcn.st one sense of meaning), but its many limitations have frequently beenpointed out.1f The major problem is that there is no good way to decide whethertwo descriptive terms designate the same property." As we have just ~n, it isnot sufficient that they refer to the same objects. Vertebrate and creature with a&pine would be said to designate the same property (distinct from tJlat designatedby creature with a heart). If we ask why this is so, the only answer appears to bethat the terms are synonymous. The notion 'property' tllus seems somehow

..Mill, A IYltem 01 logic (1843). Carnap lives a recent reformulation in Meanini andsynonymy in naturallanguagea, Phil. Itudiu 6.33-47 (1965), defining the meaning [lIlten-sion) of a predicate 'Q' for a speaker X &8 'the general condition which an object y mustfulfill in ordcr for X to be willing tu &8cribe the predicate "Q" to y'. The connotation of anexpression il often said to conatitute ita 'cognitive meanini' &8 opposed to it. 'emotivemeaning', which is, essentially, the emotional reaction to the exprealion.

Whether or not this is the best way to approach meaning', it is clear that denotation,cognitive meaning, and emotive meaning are quite different thinp. The differences are ofteno~cured in empiricalltudies of meaning, with much conaequent confusion. Thua Osg'oodhas sct himself the t~k of accounting' for the fact that a stimulus comoa to be a agn foranother stimulus (a buzzer becomcs a lig'n for food, a word for a thillg', etc.). Thil it clearly(fur linguistic Bigos) a proulom of donotation. Tho method t.hat he actually develops forquantifying and mealluring meaning (cf. Ollgood, Suci, Tannenbaum, The mealunment o(meaning (Urbana, 1957) applies, however, only to emotive meaning. Suppose, for example,that A hatell both Hitler and science inteneely, and considers both highly potent. and'activo', whilo 13, agreoing with A auout Hitler, likcs ecience vary much, although he eon-lIiders it rather ineffective and not too import.ant. Then A may assign t.o 'Hitler' and 'Ici-euce' the lIame poaition on the scmantic difforential, while B will asllign 'Hitler' the sameposition as A did, but 'science' a totally different position. Yet A does not think that 'Hit-ler' and 'science' are synonymous or that they have t.he same reforenco, and A and B mayagree precisely on the cognitive meaning of 'science'. Clearly it is the attitude toward thethings (the emotive meaning of the words) that is being measured here. There is a gradualshift in Osgood's account from denotation to cognitive meaning to emotive meaning. Theconfusion is caused, no doubt, by the fact that the term 'meaning' is used in all three lenses(and others). [See Carroll's review of the book by Osgood, Suci. and Tannenbaum in thisnumber of LANGUAGE.)

" Most clearly by Quine. See From a logical point ollliew (Cambridg'e, 1953), especiallyChapters 2, 3, and 7.

..A method for charactcrizing synonymy in torms of rcference is suggcsted by Good.man, On likeness of meaning, Analy,i, 10.1-7 (1949). Difficulties are dilcuesed by Goodman,On some differences about meaning, ibid. 13.90-6 (1953). Carnap (op.cit.) presents a verysimilar idea (§6), but somewhat miijleadingly phrased, since he does not brinK out the factthat only extensional (referential) notions are being used.

'~':::{"',

~~'"~'~.";" "~~ ~"'L"'" c,\~,;":~ !;..",~""~~"i!';", .~ r;;'~.,~";,,.~~~..,,'.

'" ",:; ;;,0. ,...' :;o,,!:~",,::i:i\;"


language-bound, and appeal to 'defining properties' sheds little light on questionsof meaning and synonymy.

Skinner accepts the traditional accowft in toto, as can be seen from his defini-tion of a tact as a response under con trot of a property (stimulus) of some physi-cal objcct or event. We have found that the notion 'control' has no real sub-stance, and is perhaps best understood as a paraphrase of 'denote' or 'connote'or, ambiguously, both. The only consequence of adopting the new term 'stimuluscontrol' is that the important differences between reference and meaning areobscured. It provides no new objectivity. The stimulus controlling the responseis determined by the respOllBe itself; there is no independent and objectivemethod of identification (see §3 above). Consequently, when Skinner defines'synonymy' as the CaI;C in which 'the same stimulus leads to quite different rc-sponses' (118), we can have no objection. The responses chair and red madealternatively to the same object are not synonymous, because the stimuli arecalled different. The responses vertebrate and creature with a spine would be con-siucr(,.u SYllonymous bccaWIC tllcy are controlled by the same proporty or theobjcct undcr investigation; in more traditional and no ICSij scientific terms,they evoke the ~me concept. Similarly, when metaphorical extension is ex-plained as due to 'the control exercised by properties of the stimulus which,though present at reinforcement, do not enter into the contingency rcspccted bythe verbnl community' (92; traditionally, accidental properties), no objectioncan be raised which has not already been levelled against the traditional account.Just as we could 'explain' the response Mozart to a piece of music in terms orsubtle properties of the controlling stimuli, we can, with equal facility, explain

-thc appenrnncc of tho rcsponsc sun when no sun is present, as in Juliet i8 [like]the BUn. 'We do so by noting thnt Juliet and the sun have common properties,a(, lc~t ill their el!(,'C(, on tJIC /$pcukcr' (93). Since any two ubjcct.!$ huvc indcfulitclymany properties in common, we can be certain that we will never be at a lossto explain n response of the form A i8 like B, for arbitrnry A nnd B. It is clenr,however, thut Skinner's recurrent cluim that his formulation i8 simpler undmore scientific thnn the traditional account has no basis in fact.

Tacts under the control of private 8timuli (Bloomfield's 'displaced speech')form a large and important class (130-46), including not only such responses asfamiliar ana beautiful, but also verbal responses referring to past, potential, orfuture events or behavior. For example, the response There was an elephant atthe zoo 'must be understood as a response to current stimuli, including eventswithin the speaker himself' (143)." If we now ask ourselves what proportion of

a'In general, the examples discussed here are badly handled, and the success of theproposed analyses is overstated. In each case, it is easy to see that the proposed analysis,which usually has an air of objectivity, is not equivalent to the analyzed expression. Totake just one example, the response I am looking lor my gloB'u is certainly not equivalentto the proposed paraphrases: 'When I have behaved in this way in the past, I have foundmy glasses and have then stopped behaving in this way', or 'Circumstances have arisen inwhich I am inclined to emit any behavior which in the past baa led to the discovery ofmy glasses; such behavior includes the behavior of looking in which I am now engaged.'One may look for one's glasses for the first time; or one may emit the same behavior in

J

REVIEWS 51

the tacts in actual life are responses to (descriptions of) actual current outsidestimulation, we can see just how large a role must be attributed to privatestimuli. A minute amoUnt of verbal behaNior, outside the nursery, consists ofsuch remarks as This ill red and There is a.man. The fact that 'fUnctional analysis'must make such a heavy appeal to obscure internal stimuli is again a measure ofi~ actual advance over traditional formulations.

9. Responses Under the control of prior verbal stimuli are considered under adifferent hending from the tact. An echoic operant is a response which 'generatesa sound pattern similar to that of the stimulus' (55). It covers only cases ofimmediate imitation.4O No attempt is made to define the sense in which a child'sechoic response is 'similar' to the stimulus spoken in the father's bass voice;it seems, though there are no clcar statements about this, that Skinner would notaccept the account of the phonologist in this respect, but nothing else is offered.111e development of an echoic repertoire is attributed completely to different.ialreinforcement. Since the speaker will do no more, according to Skinner, thanwhat iH dcmanded of him by the vcrbal community, the degree of accuracy in-sisted on by this community will determine the elements of tho repertoire,whatever these may be (not necessarily phonemes). 'In a verbal communitywhich does not insist on a precise correspondence, an echoic repertoire may re-main Rlack and will be ICM successfully applied to novel patterns'. There is nodiscussion of such familiar phenomena as the accuracy with which a child willpick up a second language or a local. dialect in the course of playing with otherchildren, which seem sharply in conflict with these assertions. No anthropologi-cal evidence is cited to support the claim that an effective phonemic systemdOCI; not develop (thil; is the subl;tance of the quoted remark) in communitiesthat do not insiHt on prccise corrcspondence.

A verolLI re.'ipOIIIiC to IL writtcn Htimulus (reading) is callcd 'textual bcI.avior'.-looking for one's glasses as in looking for one's watch, in which case I am looking for mvgla88e.4 I\lId 1 am looking fur ".1/ walcJ. ,~rc cquivalcnt, undcr the Skinllcriall parapl.ra¥e....ltc Ilinicult qucstio/lll of purpo!!ivc/lcHIl cannot be handlcd in this lIupcrficial manner.

e Ski/lncr t.akCl! groat. PQins, howcver, t.o dcny t.hc exi~t.cncc in human beings (or parrots)

of allY innQt.c facult.y or tcndcncy t.o imitat.c.lli~ only argument iB that. no one would sugge~tan innate tendency to read, yet reading and echoic behavior have Bimilar 'dYDamic proper-tics'. ThiK lIimilarity, howcvcr, simply indicates tho grossness of his dcscript.ive cate(ories.

In the case of parrots, Skinner claims that they have no instinctive capacity to imitate,but only to be reinforced by successful imitation (59). Given Skinner's use of the word'reinforcement', it is difficult to perceive any di&tinction here, since exactly the same thingcould be Raid of any other instinctive behavior. For example, where another scientist wouldKay that. a cert.ain bird inst.inctively buildt! R nest in a certain way, we could Bay in Skin-Iler'!! terminology (equivalently) that tho bird is inst.inctively reinforced by building thenest. in thit! way. One is therefore inclined to dismiss this claim as another ritual introductionof the word 'reinforce'. Though there may, under some suitable clarification, be some truthill it., it. iK cJinic\!It. t.o HCC It()W m:~IIY of tllc Calle!! report.cd lIy compct.cnt obscrvcrB can beh~lIdleu if 'reillforccment.' it! given some substantive meaning. Cf. Thorpe, op.cit. 353 f.jJ~renz, King Solomon', ring 85-8 (New York, 1952); even Mowrer, ~.ho tries to Bhow howimit.ation might develop through secondary reinforcement, cites a case, op.cit. 694, whichtIc app:~rcntly bclieves, but where thi!! could hnrdly be true. In young children, it seemsmO:it inll'l:~'lsible to explain imitation in terms of second"ary reinforcement.

~~

'"

;~.:,~

A~ £~:!\\~~~.~~g~i~"; ~'.


Other verba.! responses to verba.! stimuli are called 'intraverbal operants'.Paradigm instances are the response four to the stimulus two plU8 two or theresponse Paris to the stimulus capital of France. Simple conditioning may be suf-..ficient to account for the response four to two plU8 twO,.l but the notion of intra-verbal response loses all meaning when we find it extended to cover most of thefacts of history and many of the facts of science (72, 129); a.!l word associationand '1light of ideas' (73-6); all translations and paraphrase (77); reports of thingsseen, heard, or remembered (315). and, in general, large segments of scientific,mathematica.!, and literary discourse. Obviously the kind of explanation thatmight be proposed for a student's ability to respond with Paris to capital ofFrance, after suitable practice, can hardly be seriously offered to account forhis ability to make a judicious guess in answering the questions (to him new)What is the feat of the Fret'i.M government', ...the source of the literary dialect!, ...the chief target of the German blitzkrieg', etc., or his ability to prove a new theorem,translate a new passage, or paraphrase a remark for the first time or in a new

way.Tho proooss of 'getting someone to see a point', to see something your way,

or to understand a complex state of nffairs (e.g. a difficult politiCtl.l situation or amathematical proof) is, for Skinner, simply a matter of increasing the strength ofthe listener's already available behavior.a Since 'the process is often exemplifiedby relatively intellectual scientific or philosophical discourse', Skinner considers it'all the more surprising that it may be reduced to echoic, textua.!, or intraverbalsupplementation' (269). Again, it is only the vagueness and latitude with whichthe notions 'strength' and 'intraverba.! response' are used that save this fromabsurdity. If we use these terms in their literal ~nse, it is clear that understand-ing a statement cannot be equated to shouting it frequently in a high-pitchedvoice (high response strength), and a clever and convincing argumont cannotbe accountcd for on tho basis of a history of pairings of verbuJ responses."

41 Though even this possibility is limited. If we were to take theao paradiRm instaneC8seriously, iL 8hould follow LhuL a child who know8 how to oounL from one to 100 could learnan arbitrary 10 X 10 matrix with these number8 as entrie8 as readily as the multiplicationtablc.

U Similarly. 'the universality of a literary work refers to the number of potential readersinclined to 8ay the same thing' (275; i.e. the most 'universal' work is a dictionary of clich~sand greetings); a speaker is 'stimulating' if he says what we are about to say ourselves(272); etc.

U Similarly, consider Skinner's contention (362-5) that communication of knowledge orfacts is just the process of making a new reaponse available to the speaker. Here the anal-ogy to animal experiments i8 particularly weak. When we train a rat to carry out somepeculiar act, it makes sense to consider this a matter of adding a response to hi8 repertoirc.In the ease of human communication, however, it is very difficult to attach any meaning tothis terminology. If A imparts to B the information (new to B) that the railroads face col-lapse, in what scnse can the responsc The railroad. face collapae be 811.id to be now, but notpreviously, available to B? Surely B could have said it before (not knowing whether it wastrue), and known that it was a sentence (as opposed to Collapse J~ railroada the). Nor isthere any reason to assume that the response has increased in strength, whatever this meansexactly (e.g. B may have no interest in the fact, or he may want it 8UPPreseed). It it notclear how we can characterizc this notion of 'making a response available' without reducingSkinner's account of 'imparting knowledge' to a triviality.

..

~~~~~1~:~~

".-' ,~~~~,.,;:.". ..'Yo"~,..~;~;'I.~'~~';;C ~;", "~,

/':':,,~~~;~:;j.;~.:~~~ ',.;~~'~'

53REVIEWS

10. A final class of operants, called autocliticB, includes those that are involvedin assertion, negation, quantification, qualification of responses, construction ofsentences, and the 'highly complex manipulations of verbal thinking'. All tJ:1eseacts are to be explained 'in terms of behavior which is evoked by or acts uponother behavior of the speaker' (313). AutOOlitics are, then, responses to alreadygiven responses, or rather, as we find in reading through this section, they areresponses to covert or incipient or potential verbal behavior. Among the auto-clitics are listed such expressions as I recall, I imagine, Jor example, a&8'Ume,let X equal ~.., the tenns of negation, the is of predication and assertion, all,807ne, if, then, and, in general, all morphemes other than nouns, verbs, andadjectives, as well as grammatical processes of ordering and arrangement.Hardly a remark in this section can be accepted witJ:1out serious qualification. Totake just one example, consider Skinner's account of the autoclitic all in Allswam are white (329). Obviously we cannot assume Utat this is a tact to all swansas stimulus. It is suggested, tJ:1erefore, that we take all to be an autoclitic modify-ing the whole sentence Swam are white. AU can then be taken as equivalent toalways, or always it is possible to say. Notice, however, that the modified eentenccSwans are white is just as general as AU swam are white. Furthermore, Ute pro-posed translation of all is incorrect if taken literally. It is just as possible to saySwans are green as to say Swam are white. It is not always possible to say either(e.g. while you are saying something else or sleeping). Probably what Skinnermeans is that the sentence can be paraphrased' X is white is true, for each swanX'. But this paraphrase cannot be given within his sYstem, which has no placefor true.

Skinner's account of grammar and syntax as autoclitic processes (Chapter 13)differs from a familiar traditional account mainly in the use of the pseudoscientifictermR 'control' or 'evoke' in placc of the traditioDll.1 '.refer'. Thus in The boy runs,thc finals of runs is a tu.ct undcr control of such 'subtle propertics of a situation'as 'the nature of running as an activity rather than an object or property of anobjcct'." (Presumably, then, in Thc attempt Jails, The difflCUlty remains, Hisanxiety increases, etc., we must also say that the s indicates that the objootdescribed as the attempt is carrying out the activity of failing, ete.) In the boY'8gun, however, Ute 8 denotes possession (as, presumably, in the boy', arrival, ...8tory, ...age, etc.) and is under the control of this 'relational aspect of thesituation' (336). The 'relational autoclitic of order' (whatever it may mean tocall the order of a set of responses a response to them) in The boy rum the 8toreis under the control of an 'extremely complex stimulus situation', namely, thatthe boy is running the store (335). And in the hat and the &hoe is under Ute controlof the propcrty 'pair'. Through in the dog went through the hedge is under the con-trol of the 'relation between the going dog and the hedge' (342). In general,nouns are evoked by objects, verbs by actions, and so on.

Skinner considers a sentence to be a set of key responses (nouns, verbs, ad-jectives) on a skeletal frame (346). If we are concerned with the fact that Sam

ff 332. On the next page, however, the I in the same example indicates that 'the objectdescribed as the boy possesses the property of running.! The difficulty of even maintainiDIconsistency with a conceptual scheme like this is easy to appreciate.

;a;;~,

54 LANGUAGE, VOL~ 36, NUMBER 1 (1959)

rented a leaky boat, the raw responses to Ule situation are rent, boat, leak, andSam. Autoclitics (including order) which qualify these responses, express rela-tions between them, and the like, are theh added by a process called 'composi-tion' and the result is a grammatical sent:ence, one of many alternatives amongwhich selection is rather arbitrary. The idea that sentences consist of lexicalitems placed in a grammatical frame is of course a traditional one, within bothphilosophy and linguistics. Skinner adds to it only the very implausible specula-tion that in the internal process of composition, the nouns, verbs, and adjectivesare chosen first and then are arranged, qualified, etc" by autoclitic rcspon~ tothese internal activities..'

This view of sentence structure, whether phrased in terms of autoclitics,syncategorematic expressions, or grammatical and lexicnI morphemes, is in-adequate. Sheep provide wool has no (physical) frame at all, but no othor arrange-ment of these words is an English ~ntence, The ~quences furi0U8ly 8leep ~green colorless and friendly young dog8 8ee7n harmle88 have the same framos, buto/lly 0110 is u ~nt.cncc of Ellgliml (Himilarly, only one of the scquenccs formed byreadillg these from back to front), Struggling arti8l8 can be a nui&ancc has thesnme frame as marking paper8 can be a nui8ance, but is quite different in ~ntencestructure, as can be seen by replacing can be by i8 or are in both cases. Thereare many other similar and equally simple examples, It is evident that moreis involved in sentence structure than insertion of lexical items in grammaticnIframes; no approach to language that fails to take these deeper proce&qes intoaccoullt can possibly achieve much success in accounting for actual linguisticbehavior,

11. The preceding discussion covers all tho major notions that Skinner intro-duCOB in his descriptivo system, My PUI,JOBC in discussing tho concopts ono by0110 WUS to show tllat in each ~, if we take his terms in thoir literll.l meaning,t.hc..drN-liptiOJl oovere .almost no aspect of verbal- behavior, and if_lV.ej.ak~ themmetaphorically, the description offers no improvement over various trl.ditloniil-formulut.iuulO, 'rho terms borrowed from experimental psychology simply losetheir objective meaning with this extension, and take over the full vagueness ofordinary language. Since Skinner limits himself to such a amnII set of tenns forparaphrase, many important distinctions are obscured. I think that this analysissupports the view expressed in §1 above, that elimination of the independentcontribution of the speaker and learner (a result which Skinner considers ofgreat importance, cf. 311-2) can be achieved only at the cost of eliminating allsignificance from the descriptive system, which then operates at a level 80 grossalld crude that no answers are suggested to the most elementary questions,.'

..One might just as well argue that exactly the opposite is true. The atudy of heaitatiollpauses has shown that these tend to occur before the large categories-noun, verb, adJ~c:.t,ive; thi.. finding ill usually described by the statemcnt that the pauses occur where thereis mlLximum uncertlLinty or information. Insofar as hesitation indicates on-going uomposi-tion (if it does at 1111), it would appe~r that the 'key responses' are chosen only alter the'grammatical frame', Cf. C, E. Osgood, unpublished paper; Goldman-Eisler, Speech &lIaly-sis and mental processes, Language and speech 1.67 (1958).

..E.g. what are in f~ct the actual units of verbal behavior? Under what conditions willa physical event capture the attention (be 11 stimulus) CJr be a reinforcer? How do we decide

~~t'. ~~l~~~1~~;,-'I~~{:.

55REVIEWS

The questions to which Skinner has o.ddressed his speculations are hopelesslypremature. It is futile to inquire into the causation of verbal behavior untilmuch more is known about the specific character of this behavior; and there islittle point in speculating about the process of acquisition without much betterunderstanding of what is acquired.

Anyone who seriously approaches the study of linguistic behavior, whetherlinguist, psychologist, or philosopher, must quickly become aware of the enor-mous difficulty of stating a problem which will define the area of his investiga-tions, and which will not be either completely trivial or hopelessly beyond therange of present-day underl$tu.nding and technique. In selecting functionl.l.lanalysis as his problem, Skinner has set himself a task of the latter type. Inan cxtrcmcly intcrcsting and insightful paper,47 K. S. l.w!hlcy has implicitlydelimited a class of problems which can be approached in a fruitful WIlY by thelinguist und pt;ychologist, Wld which arc clearly preliminary to tJl011C with whichSkinner is concerned. Lashley recognizes, as anyone must who seriously con-siders tho dut.u, tJ1Ut tho compoHitioll and production of nIl uttcrunce is !lotsimply a matter of stringing together a sequence of responses under the controlof uutsido stimulation alld intrnverbal association, and that the Hyntactic orguni-zation of an utterance is not something directly represented in any simple wayin the physical structure of the uttcrance itself. A variety of observations leadhim to conclude that syntactic structure is 'a generalized pattern imposed onthe specific acts as they occur', and that 'a consideration of the structure of thesentence and other motor sequences will show... that there are, behind theovertly expressed sequences, a multiplicity of integrative processes which canonly bo inforred from the final rctiults of their activity'. He also comments onthe great difficulty of determining the 'selective mechanisms' used in the actualconHtruction of a purticulur uttcruncc.

Although present-dny linguit;tics cannot provide a precise account of theseintegrative processes, imposed pntterns, and selective mechanisms, it can atlcast 1ICt. it.sclf the problem of characterizing these completely. It is reusonableto rc~nrd the grammar of n lunguage L ideally as n mechanism that provides~

what 8timuli are in 'control' in a 8pecific case? When are 8timuli 'similar'? And so on. (Iti8 not intereeting to be told e.g. that we 8ay Stop to an automobile or billiard ball becausethey are 8ufficicntly similar to reinCorcing people [46].)

The U8e oC unanalyzed notion8 like '8imilar' and 'generalization' is particularly di8-turbing, since it indicate8 an apparent lack oC intereet in every significant aspect oC thelearning or the u8e oC language in new 8ituation8. No one has ever doubted that in somelIenlle, language i8 learned hy gencralization, or that novel utterances and situation8 are insome way similar to Camiliar ones. The only matter oC serious interest i8 the specific 'simi-larity'. Skinner has, apparently, no intere8t in thi8. Keller and SchoenCeld (op.cit.) proceedto incorporate the8e notion8 (which they identiCy) into their Skinnerian 'modern objectivep8ychology' by defining two 8timuli to be 8imilar when 'we make the 8ame 8ort oC reaponaeto tbcm' (J21; but wbcn I1rc rCIi!>UIIMCIl or tbc 'same lIort'?). Tbey do not &cem to notice thatthi8 definition converts their 'principle oC generalization' (116), under any reasonable in-terpretation oC thi8, into a tautology. It i8 obviou8 that such a definition will not be oCmuch help in the study oC language learning or con8truction of new respon8ee in appropriate8ituation8.

47 The problem of 8erial order in behavior, in Jeffre18 (ed.), Hixon 8'Jmpo8iutn on cerebral

mechani8m1 in behavior (New York, 1951).

1~

~

~,~."k~, '~~i~",;,r~,~'4


an enumeration of the *3ntences of L in something like the way in which a deduc-tive theory gives an enumeration of a ~t of theorems. ('Grammar', in this *3nseof the word, includes phonology.) Furt.hQrmore, the theory of language can beregarded as a study of the formal properties of such grammars, and, with aprecise enough formulation, this general theory can provide a uniform methodfor determining, from the process of generation of a given sentence, a structuraldescription which can give a good deal of insight into how this sentence is usedand understood. In short, it should be possible to derive from a properly formu-lated grammar a statement of the integrative processes and generalized patternsimposed on the specific acts that constitute an utterance. The rules of a gram-mar of the appropriate form can be subdivided into the two types, optional andobligatory; only the latter must be applied in generating an utterancc. Theoptional rules of the grammar can be viewed, Ulen, as the selective mechanismsinvolved in the production of a particular utterance. The problem of specifyingthese integrative processes and ~lective mechanisms is nontrivial and not be-yond the range of po8.'!ible invcstigation. Tho results of such A titudy might, ~Lashley suggests, be of independent interest for psychology and neurology (andconversely). Although Huch A Htudy, oven if succeaBful, would by no means answerthe major problems involved in the investigation of meaning and the causationof behavior, it surely will not be unrelated to these. It is at least possible, fur-thcrmore, that such notions ~ 'sema.ntic generaliZAtion', to which such heavyappeal is made in all approaches to language in use, conceal complexities andspecific structure of inference not far different from th~ that can be studiedand exhibited in the case of syntax, and that consequently the genera.l characterof the results of syntactic investigations may be a corrective to oversimplifiedapproaches to tile theory of meaning.

Thc behavior of the spcakor, listoner, and learnor of languagc co~titutcH, ofcourliC, the actual dat:.l. for Any study of language. The construction of a grammarwhich enumerates sentences in such a way that a meaningful structural descrip-tion can bc dctormincU for oach IiCnt.oncc d~ not in itliClC providc an accountof this actual behavior. It merely characterizes abstractly the ability of onewho htLS m~tered the language to distinguish sentences from no~ntenoes, tounderstand new sentences (in part), to note certain ambiguities, etc. These arevery rema.rkable abilities. We constantly read and hear new sequences of words,recognize them as sentences, and understand them. It is easy to show that thenew events that we accept and understand as sentences are not related to thosewith which we are familiar by any simple notion of formal (or semantic or sta-tistical) similarity or identity of grammatical frame. Talk of generalization inthis CaBo is entirely pointless and empty. It appears that we recognize a newitem ns u sentence not because it matches some familiar item in any simple way,but because it is gencrated by thc grammar that each individual hl\b' somchow~nd in ,"ume (orm internulized. And we understand a new sentence, in part,because we are someho,v capable of determining the process by which thissentence is derived in this grammar.

Suppose that we manage to construct grammars having the properties out-lined above. We can then attempt to describe and study the achievement ofthe speaker, listener, and learner. The speaker and the listener, we must assume,

;,p.~"";;c.;'f'if'~i,1...""", ,.".

':Ii,

57REVIEWS

have already acquired the capacities characterized abstractly by the grammar.Tha..spea.ker's task is to select a particular compatible set of optional rules. Ifwe know, from grammatical study, what ehoices are available to him -and""'whatconditions of compatibility the choices mu~t meet, we can proceed meaningfullyto investigate the factors that lead him to make one or another choice. Thelistener (or reader) must determine, from an exhibited utterance, what optionalrules were chosen in the construction of the utterance. It must be admitted thatthe ability of a human being to do this far surpasses our present understanding.The child who learns a language has in some sense constructed the grammar forhimself on the basis of his observation of sentences and nonsentences (i.e. cor-rections by the verbal community). Study of the actual observed ability of aspeaker to distinguish sentences from nonsentences, detect ambiguities, etc.,apparently forces us to the conclusion that this grammar is of an extremely com-plex and abstract character, and that the young child bas succeeded in carryingout what from the formal point of view, at least, ~ to be a remarkable typeof thcory coostruction. Furthermore, this task is accomplOOcd in an astonish-ingly short time, to a large extent independently of intelligence, and in a com-parable way by all children. Any theory of learning must cope with these facts.

It is not easy to accept the view that a child is capable of constructing anextremely complex mechanism for generating a set of sentences, some of whichhe has heard, or that an adult can instantaneously determine whether (and ifso, how) a particular item is generated by this mechanism, which bas many ofthe properties of an abstract deductive theory. Yet this appears to be a fairdescription of the performance of the speaker ,.listener, and learner. If this iscorrect, we can predict that a direct attempt to account for the actual behaviorof speaker, listener, and learner, not based on a prior understanding of the struc-ture of grammars, will achieve very limited success. The grammar must be re-garded as a component in the behavior of the speaKer-andl18~ner.-w~only be inferred, as Lashley has put it, from the resulting physical acts. Thefact that all normal children acquire essentially comparable grammars of greatcomplexity with remarkable rapidity suggests that human beings are ROmebowspecially designed to do this, with data-handling or 'hypothesis-formulating'ability of unknown character and complexity." The study of linguistic structure

..There is nothing essentially mysterious about this. Complex innate behavior patternsand innate 'tendencies to learn in specific ways' have been carefully studied in lower or-ganisms. Many psychologists have been inclil1ed to believe that such biological structurewill not have an important eft'ect on acquisition of complex behavior in higher organisms,but I have not been able to find any serious justification for this attitude. Some recentstudies have stressed the necessity for carefully analyr.ing the strateKies available to theorganism, regarded as a complex 'information-processing system' (cf. Bruner, Goodnow,and Austin, A ,tudy of thinking [New York, 1956); Newell, Shaw, and Simon, Elements of atheory of human problem solving, P,ych. r'lI. 65.151~ 11958]), if anything significantis to be said about the character of humAn learning. These mAY be largely innate, or devel-oped by early learning processes about which very little is yet known. (But lee Barlow, Theformation of learning sets, P,ych. r'lI. 56.51-65 (1949), and many later papers, where atrik-ing shifts in the character of learning are shown as ..result of early tre.ininc; ..110 Bebb,Organization of b,hallior 109 ft'.) Thcy are undoubtedly quite complex. Cf. Lenneberc, op.cit.,and Lees, review of ChomskY'8 Syntactic atroctur,a in Lg. 33.406 f. (1957), for dilcusaion ofthe topics mentioned in this 8ection.

.J.J ;;.j,~~ .~..";;;o~;"":;;;..;;;";';


may ultimately lead to some significant insights into this matter. At the momentthe question cannot be seriously posed, but in principle it may be possible tostudy the problem of determining what the built-in structure of an information-processing (hypothesis-forming) system \nust be to enable it to arrive at thegrammar of a language from the available data in the available time. At anyrate, just as the attempt to eliminate the contribution of the speaker leads to a

--.: 'mentalistic' descriptive system that succeeds only in blurring important tradi.-::;~.:-;;.,f~~n~ distinc~ons, a refusal to ~tudy the contribution of th~ ~ild ~ language

learning permIts only a superficIal account of language acqu1S1tlon, WIth a v~tand unanalyzed contribution attributeQ to a step called 'generalization' whichin fact includes just about everything of interest in this process. If the study oflanguage is limited in these ways, it seems inevitable that major aBpoots ofverba) behavior will remain a mystery.

verbal behavior. by b. f. skinner. (the century...

Documents