JOURNAL OF THE EXPERIMENTAL ANALYSIS OF BEHAVIOR

VERBAL BEHA VIORJACK MICHAEL

WESTERN MICHIGAN UNIVERSITY

The recent history and current status of the area of verbal behavior are considered in termsof three major thematic lines: the operant conditioning of adult verbal behavior, learningto be an effective speaker and listener, and developments directly related to Skinner's Ver-bal Behavior. Other topics not directly related to the main themes are also considered: thework of Kurt Salzinger, ape-language research, and human operant research related torule-governed behavior.Key words: verbal behavior, operant conditioning of vocal behavior, verbal behavior ac-

quisition, verbal behavior in apes, rule-governed behavior, verbal stimuli and the listener

A reasonable history might have been thefollowing: As a result of major contributionsby Darwin, Thorndike, Pavlov, Watson,and important contributions by manyothers, B. F. Skinner in the mid-thirtiesdeveloped the approach that has come to beknown as the experimental analysis of be-havior. At first this consisted largely oflaboratory research with lower animals-mainly rats and pigeons-on the relationsbetween behavior and its consequences.Rapid progress was made with these simplerspecies, whose behavior was found to be af-fected in lawful ways by reinforcement con-tingencies of considerable complexity.Laboratory work with other species, espe-cially monkeys and apes, revealed consider-able generality of the basic relations. Speciesdifferences had to be considered, of course,but turned out to be related chiefly to (1)sensory and motor capacities, (2) the specificevents that function as effective conse-quences, and (3) unlearned behavior relatedto mating and care of the young.

It soon became possible to extend this typeof research to human behavior, which hadbeen the principal goal all along. The moststraightforward extensions were obtainedwith simpler forms of human behavior not

This paper was written in tribute to Don Hake.Reprints are available from the author, Department ofPsychology, Western Michigan University,Kalamazoo, Michigan 49008.

involving language. The behavior of prever-bal children and of humans who for variousreasons fail to develop language was studiedunder laboratory conditions and found to begoverned by essentially the same laws thathad been discovered with rats and pigeons.

This laboratory-based science of humanbehavior gave rise to a vigorous applied ap-proach called behavior modification, dealingat first with young children and institu-tionalized psychotic and mentally retardedadults, but soon applying the science ofbehavior to all types of 'clients' and in alltypes of settings. In this latter respect, it wasclear that human language complicated thepicture considerably, but extensive labora-tory research on language was already underway, and new knowledge in this area wouldundoubtedly contribute greatly to our under-standing and control of human behavior.

Basic research on language was concernedwith several major issues: What does thelanguage repertoire consist of? How dochildren acquire this repertoire? Whenlinguistic behavior occurs, how do the result-ing stimuli affect listeners, and how dolanguage repertoires affect speakers as self-listeners?

But this is not the way things actually hap-pened. The first paragraph is reasonablycorrect. When it comes to the extensions tohuman behavior, however, there are twoclear discrepancies. First, the vigorous ap-plied approach called behavior modification

363

1984, 429 363-376 NUMBER 3 (NOVEMBER)

JACK MICHAEL

developed without the benefit of any ap-preciable laboratory research with humans.At the beginning (the late fifties and earlysixties), it was a direct attempt to useoperant, animal-laboratory methods andconcepts to alter socially important humanbehavior. The relevance of these concepts tohuman behavior was taken for granted bythe behavior modifiers, as it was in Skinner'sScience and Human Behavior (1953), which un-doubtedly played a major role in the devel-opment of this movement. There were somepublished studies demonstrating that humanbehavior is indeed affected by consequences,and in a manner not drastically differentfrom nonhuman behavior, but such demon-strations could hardly be considered a care-ful laboratory extension. Only with the re-cent explicit identification of an area referredto as "human operant research" has therebeen an effective call for human laboratoryresearch (e.g., Hake, 1982; also relevant isthe newsletter publication, ExperimentalAnalysis ofHuman Behavior Bulletin, edited byW. F. Buskist of Auburn University).The second departure from "what-might-

have-been" is that a complex and detailedbehavioral treatment of human language be-came available simultaneously with theexperimental analysis of behavior. WhileSkinner was working on the basic methodsand relations that would be reported in TheBehavior of Organisms in 1938, he was alreadyconvinced that these same principles werenecessary and sufficient for understandinghuman language. He dates the beginning ofhis systematic work on language as 1934 (thepersonal epilogue, "No Black Scorpion," re-ported in Verbal Behavior, pp. 456-460). By1945 much of the material that was to appear12 years later in Verbal Behavior had beencompleted, and was the basis of a summercourse at Columbia University in 1947.Ralph Hefferline attended that course and amimeographed version of his transcribedshorthand notes was distributed privatelythe following year. Another version of thismaterial was the basis of the William JamesLectures by Skinner in the fall of 1947, andseveral hundred mimeographed copies were

circulated subsequently. Skinner's analysisof verbal behavior differed quite remarkablyfrom what might have been expected -first,in the richness, complexity, and detail of thetreatment; and second, in its empirical basethat was either the animal laboratory data ofThe Behavior of Organisms (1938) or "basicfacts . . . well known to every educated per-son" (1957, p. 11). Skinner comments on thisissue in The Shaping of a Behaviorist (1979) asfollows:

I had collected a lot of experimental dataon verbal behavior-on how people learnstrings of nonsense syllables, or thenonsense names of nonsense figures, andI had my own results on verbal summa-tion, alliteration, and guessing. Theybegan to clutter up the manuscript with-out adding much by way of validation.They threw the book as a whole badly outof balance because I could not find exper-iments for the greater part of the analysis.I was still the empiricist at heart, but I didnot think it would betray that position ifmy book were not a review of establishedfacts. I was interpreting a complex field,using principles that had been verifiedunder simpler, controlled conditions. Ex-cept for certain aspects of the solar sys-tem, most of astronomy is interpretationin this sense, its principles being derivedfrom laboratory experiments. I decided toleave out all experimental data. (An in-teresting question then arose: what sur-vived to reinforce writing or reading thebook? Was not confirmation the be-alland end-all of science? It was a questionconcerning my own behavior, and Ithought I had an answer: "February 2,1945. What is motivational substitute forthing-confirmation? Pretty important inteaching method to graduate students.Resulting order instead of confirmation?"My reinforcers were the discovery ofuniformities, the ordering of confusingdata, the resolution of puzzlement.) (p.282)By the early fifties most of the relatively

small group of people who were practicing or

364

VERBAL BEHA VIOR

teaching about the experimental analysis ofbehavior knew of the existence of Skinner'ssoon-to-be-published major work on lan-guage, or had actually seen one of the earliermimeographed versions. The influential textby Keller and Schoenfeld (1950) devoted 25pages (pp. 376-400) to a consideration ofverbal behavior based largely on Skinner'swork; the forthcoming book was mentionedin a footnote (p. 210) in Science and HumanBehavior (1953); some attention to the largerwork was also undoubtedly inspired by thearticle, "The Operational Definition of Psy-chological Terms," which appeared in Psy-chological Review in 1945. Verbal Behavior wasfinally published in 1957 and was more thantwice as long as the mimeographed versions.

So by the end of its first 20 years, the ex-perimental analysis of behavior was well onits way as an approach to basic research inbehavior. It had given rise to a rapidly grow-ing applied branch called "behavior modifi-cation" (and later, "applied behavior anal-ysis") and also had available an extensiveand carefully worked out interpretation ofthat most complex behavior, human lan-guage. A most encouraging beginning. It isnow a quarter of a century later. What hashappened since that early period of rapid ex-pansion, and more specifically, what hashappened to the behavioral analysis of lan-guage? At the considerable risk of over-simplification, it seems possible to identifythree major lines of development plus somerelevant miscellanea.

THE OPERANT CONDITIONING OFADULT VERBAL BEHAVIOR

While at Indiana University in the lateforties, Skinner's graduate seminars weresometimes devoted to verbal behavior. JoelGreenspoon, a student in the clinical pro-gram, was a member of one of those semi-nars and as a result developed a method forthe experimental investigation of the effect ofreinforcement on verbal behavior (Green-spoon, 1955). He asked his subjects to saywords-but not sentences or phrases; whenthe subject said a word that could be

classified as a plural noun, the experimentersaid "mmm-hmmm"- the type of vocalresponse by a listener meant to encourage aspeaker to continue. This increased therelative frequency of plural nouns and wastaken as evidence that verbal behavior wassensitive to its consequences, just like theother behavior studied in the operant condi-tioning laboratory.Many similar studies soon appeared and

from 1958 they were the basis of severalreviews (Greenspoon, 1962; Holz & Azrin,1966; Kanfer, 1968; Krasner, 1958; Salz-inger, 1959, 1969). By the late sixties,however, this particular type of research hadbecome infrequent, and there is very little ofit taking place at the present time.Why did so much of this research take

place and why did it stop? Three factors mayhave initiated and maintained this research.Operant research had grown remarkablysince 1938, largely, it would seem, as aresult of the discovery of a good experimen-tal paradigm. Skinneres use of rate in thefree-operant setting, the cumulative recorderproviding a form of "already analyzed" data,the discovery of the various effects of inter-mittent reinforcement- all had led to arapidly increasing body of important re-search findings. The possibility of a break-through in the study of human language wascertainly part of the inspiration for the workin the operant conditioning of verbal behav-ior.The search for a good methodology forthe study of language was the main purposeof Greenspoon's own efforts in this area (per-sonal communication, August, 1984), andthe possibility that one had been found wasprobably important to many of those whofollowed his lead.

Dollard and Miller's Personality and Psycho-therapy (1950) provided an analysis of thepsychotherapy situation that emphasized thereactions of the therapist as reinforcementfor the client and thus an important factor inchanging the client's behavior. The operantconditioning of verbal behavior in theGreenspoon experiment was amusingly sim-ilar to popular presentations of the interac-tion between nondirective therapist and

365

JACK MICHAEL

client, a similarity that was not lost on theincreasing number of students of behaviorand professors of clinical psychology. So a

number of the studies during the late fiftieswere directly aimed at understanding or im-proving the clinical psychologist's pro-

cedures, and understanding the develop-ment of abnormal behavior.Many cognitively oriented psychologists

questioned the validity of the operant condi-tioning of verbal behavior. They argued thatin these studies human subjects were simplyaware of the nature of the experiment and,when it suited them, cooperated with the ex-

perimenter. From that point of view, the ex-

periments were not demonstrating operantconditioning, but rather a more complexcognitive process. The data with subjectswho were not aware of the contingencies,then, became the focus of much of the laterresearch, with behaviorists trying to demon-strate conditioning in unaware subjects, andcognitivists trying to show that only whenthe subjects were aware did any "condition-ing" take place.Why did this general line of research come

to an end? A simple answer is that the break-through did not take place, and beyondshowing (although this is still disputed bycognitivists) that verbal behavior is operantbehavior and thus affected by its conse-

quences, there was not much additionalyield. The development of behavior modifi-cation in the late fifties and early sixties,with its emphasis on direct manipulation ofthe environment, may have detracted froman interest in traditional psychotherapy. Italso detracted from the significance ofDollard and Miller's (1950) analysis, whichmerely reinterpreted psychodynamictherapy.The awareness controversy may have had

detrimental effects, in that the issue did notseem methodologically resolvable. A more

subtle difficulty was identified by Holz andAzrin (1966): "Perhaps, because of thepolemic regarding conditionability of verbalresponse, the experiments have been forcedinto strict parallels of basic motor condition-

ing experiments. As a result, no distinctly'verbal' characteristics or effects seem to haveemerged from these experiments" (p. 815).This is probably due to the fact that in theexperimental analysis of behavior the usuallaboratory dependent variable is the rate ofresponse, defined by operation of some me-chanical device, such as a microswitch. Ver-bal behavior achieves its uniquely verbal ef-fects by consisting of a sequence of multiplebut functionally separate topographies thatuntil recently could not be identified byautomatic recording equipment. It may notbe possible to force the verbal dependentvariable into the typical operant researchmold and have anything verbal left over. Ofcourse the sequence of multiple topographiescould have been studied as a dependent var-iable, but only by using a human observer toidentify the different response units. How-ever, this research procedure was unattrac-tive.There is a strange incongruity between

Skinner's elaborate analysis of verbalbehavior in terms of operant and respondentconditioning and an attempt to show onlythat verbal behavior is affected by its conse-quences. Skinner showed no doubt about theoperant nature of verbal behavior, and infact mentions Greenspoon's experiment inChapter 6 of Verbal Behavior (pp. 148-149),treating it as a special case rather than fun-damental to the analysis.

In summary, this thematic line did notlead anywhere and has died out. In labora-tory characteristics it resembles recent workon the effects of such things as rules and in-structions on human schedule performance,but the latter approach has a verbal indepen-dent rather than dependent variable. Withthe advent of computer technology, it shouldbe possible to overcome the difficulties ofstudying verbal behavior as an operantdependent variable, and this thematic linemay be revived. But the same computertechnology makes possible so many otherunanticipated ways to study verbal behaviorthat any new developments probably will notappreciably resemble the older research.

366

VERBAL BEHAVIOR

LEARNING TO BE AN EFFECTIVESPEAKER AND LISTENER

Having arrived at basic principles describ-ing the way environmental variables alterthe repertoires of individual organisms(respondent and operant conditioning, ex-tinction, stimulus control, etc.), the nextstep is to see what repertoires are developed.In humans the verbal repertoire is highlysignificant. Understanding the behavioralprocesses by which an individual becomesan effective speaker and listener is importantand will have great practical value in makingpossible an effective technology of educa-tion. This should include preventing andremediating defective verbal repertoires, aswell as improving on normal acquisition.There is a large body of literature on

language acquisition. Much of it describesthe chronology and sequencing of variousaspects of the verbal repertoire, and the rela-tions of this chronology and sequencing toother variables (e.g., birth order, IQ scores,socioeconomic status, personal and socialcharacteristics of parents). This information,though not unimportant, does not contributemuch to an understanding of the behavioralprocesses relevant to the chronology or se-quencing. Of course, if the verbal speakingand listening repertoire is innate, as somelinguists and psychologists believe, thenshort of genetic and physiological informa-tion, description of what appears and whenis all that is possible. Although innate deter-mination is not incompatible with a behav-ioral approach, the empirical evidence forextensive (though not necessarily exclusive)environmental participation is overwhelm-ing, and there is a good deal of research andspeculation about the nature of this partici-pation in the traditional literature onlanguage acquisition.From the perspective of an experimental

analysis of behavior, there has not beenmuch basic research in this area. Whentraditional approaches attempt a func-tional analysis rather than simply describ-ing developmental sequences, they typicallyrely on mentalistic-cognitive terms and

assumptions. This has discouragedbehavioral interest in this literature. The ex-perimental analysis of behavior and its ap-plications have emphasized manipulativerather than descriptive research; however,manipulative experimentation is difficult inthis area because of the ethical implicationsof "tampering" with the normal language-acquistion process.

Nevertheless, there has been some impor-tant research of this type. Not long after theoperant conditioning of adult verbal behav-ior began, several studies appeared demon-strating that infants' vocal behavior was af-fected by its consequences. After Brackbill's(1958) demonstration that infants' smilingwas affected by continuous and by intermit-tent presentation of adult smiling, vocaliz-ing, touching, and the like, Rheinhold,Gewirtz, and Ross (1959) showed similar ef-fects with vocalizations in 3-month-old in-fants. Subsequent studies by Weisberg(1963), Todd and Palmer (1968), Routh(1969), and Haugan and McIntire (1972)showed similar effects and in addition pro-vided information about other variables rele-vant to the reinforcement effect. For most ofthis work the dependent variable was simplyfrequency of vocal responses, but Routh(1969) was also able to alter the relative fre-quency of vowels versus consonants usingdifferential reinforcement.As the issue of awareness confounded the

operant conditioning of adult verbal behav-ior, so the effectiveness of adult socialresponses as eliciting stimuli has confoundedthe interpretation of the effects of suchresponses as reinforcement for infant vocalbehavior. These issues have been reviewedextensively (Hulsebus, 1973; Millar, 1976),studied more carefully with various controlprocedures (Bloom, 1975, 1977, 1979;Bloom & Esposito, 1975; Sheppard, 1969),and finally resolved in support of the originalfindings (Poulson, 1983). These studies areimportant in demonstrating that the socialreactions of adults can play a role in increas-ing the frequency of infant vocalizations andin shaping the specific topographies thatultimately become the child's speech. Of

367

JACK MICHAEL

course they do not prove that speech does infact develop as a function of reinforcement;however, that a frequently occurring type ofadult reaction functions as an effective con-sequence is sufficient to maintain an interestin the role of reinforcement in the develop-ment of verbal behavior.

Imitation is an important way of acquir-ing new verbal behavior, so the work on therole of reinforcement in the development ofgeneralized imitation is critical to the under-standing of normal acquisition (see Peter-son, 1968). The first studies in this area wereby Baer and Sherman (1964), Metz (1965),and Baer, Peterson, and Sherman (1965),and since then imitation has been the topicof extensive research in applied behavioranalysis. Another line of applied behavioralresearch contributing to this general themeconcerns the generation of specific kinds ofverbal repertoires-for example, the use ofdescriptive adjectives (Hart & Risley, 1968),generative use of the plural morpheme(Guess, Sailor, Rutherford, & Baer, 1968),prepositional usage (Sailor & Taman, 1972),acquisition and generalization of verbs(Campbell & Stremel-Campbell, 1982), andthe development of generative sentences(Lutzker & Sherman, 1974).

Within the context of remedial instruc-tion, other important research topics havebeen: receptive language training (e. g., Baer& Guess, 1971); transfer between receptiveand productive language (e. g., Keller &Bucher, 1979); developing a nonvocal (sign)repertoire (e.g., Faw, Reid, Schepis, Fitz-gerald, & Welty, 1981), and the relation be-tween verbal and nonverbal behavior (seeIsrael, 1978, for a review of this area).Although most of the research has been

with children whose verbal behavior is defi-cient (e.g., culturally disadvantaged, men-tally retarded, autistic), the behavioralmethodology, results, and implications areall relevant to the general topic of verbalrepertoire acquisition. In many cases thebehavioral methods used to improve the ver-bal behavior of these subjects also would beappropriate for improving the verbal behav-ior of normal children and for preventing

behavioral deficits. These methods andresults direct our attention to importantfeatures of the normal environment-child in-teraction that might well be overlooked wereit not for their role in creating or remediatingverbal deficits. An example is the procedurecalled incidental teaching (Hart & Risley,1975), which consists of systematicallyprompting, then requiring verbal behaviorfrom the child in the process of interactingwith the nonverbal environment. For exam-ple, in a playroom setting, a child's observ-able interest in playing with a toy that is on ashelf too high for the child to reach functionsas a basis for teaching the location, shape,color, etc. of the toy, with the receipt of thetoy reinforcing the child's appropriate verbalbehavior. The effectiveness of this trainingtechnique, especially in the continued occur-rence of new behavior outside the trainingsituation, suggests the importance of suchnatural reinforcement in the development ofverbal behavior under normal conditions,and invites further descriptive research inbehavioral terms.With the exception of the work on the ef-

fects of reinforcement on infant vocal behav-ior, this thematic line is not only a tech-nology for the remediation of defective ver-bal behavior but also an important source offacts, concepts, and theories relevant to theacquisition of a verbal repertoire and to theinteraction of that repertoire with nonverbalbehavior. It is concerned both with behavingas a speaker and as a listener.

This work seems to have resulted mostdirectly from the basic notions of operantconditioning and single-subject researchmethodology as described by Skinner(1938), from the broad implications of thiswork for all aspects of human behavior dis-cussed in Science and Human Behavior (1953),from the development of basic-researchmethodology for the study of children'sbehavior, and a strong societal need to dealmore effectively with developmentallydisabled, culturally disadvantaged, andother handicapped children.

Interestingly, this extensive body ofresearch makes almost no use of the con-

368

VERBAL BEHAVIOR

cepts, terms, and analyses that appear inSkinner's (1957) Verbal Behavior. Althoughthe term "verbal behavior" had become wide-spread, the recent trend is toward increaseduse of the traditional term, "language," inspite of its implication of a common processunderlying kinds of behavior that differ con-siderably from one another, such as speak-ing and listening. The terms for elementaryverbal relations- "mand," "tact," "echoic,"etc. -are used occasionally, but not to anyimportant purpose; the research could easilyhave been conceived without the benefit ofthe distinctions Skinner makes.At present this line of research and theory

is very active. There is good contact withand contribution to basic research, involvinga productive interaction between this workand the area described later as the study ofrule-governed behavior. Most of the studiesare with developmentally disabled children,but the merging of this thematic line withbehavioral medicine makes it likely thatbehavioral contributions will increase under-standing of the effects of brain injury on ver-bal behavior. This is a contribution that issorely needed.

SKINNER'S BOOK, VERBAL BEHA VIORA third thematic line consists of inter-

pretation and research arising more or lessdirectly from Skinner's own analysis of ver-bal behavior. Much of this work is, like Ver-bal Behavior itself, interpretation rather thanexperimental research, although some re-search has resulted from Skinner's analysisand this seems to be increasing.There has been much comment about the

seeming paucity of research generated byVerbal Behavior. Salzinger (1978) suggestedthat a major reason is that the book pre-sented no data. E. Vargas (in press) notedthat few instructors have taught much aboutSkinner's analysis, that it is diametrically op-posed to the widely prevalent common-senseinterpretation of language, and that the bookis not easy to understand. Another reasonmay be that the basic researchers whosetraining prepared them to appreciate Skin-

ner's analysis already had a highly produc-tive research methodology, involvingresponse rate as the dependent variable andautomated data collection; and in 1958 theJournal of the Experimental Analysis of BehaviorUEAB) began publication. Operant research-ers in the late fifties and early sixties werestrongly committed to behaviorism as a data-based science, and less interested in-or insome cases even embarrassed by- Skinner'sspeculative extensions to human affairs.Possibly as a reflection of this same orienta-tion, or as a result of conflict with nonbehav-ioral orientations, the applied behavior anal-ysis developing in this same period also hadan emphasis on data as the only valid basisfor procedure and policy, and from this per-spective Verbal Behavior did not seem par-ticularly useful.The book is really most appropriate for

language scholars who are also strongly pre-disposed to welcome a behavioral approachto their subject matter-at best a smallgroup. I have used it as a text for graduateand undergraduate students ever since it waspublished, and found that the students' maindifficulty is not with the "behavioralizing,"but rather with what is being behavioralized.For example, the following are included inthe first 20 entries listed in the index:abstraction; acrostics; agglutinated lan-guages; agnosia; agrammatism; alexia;allegory; alliteration; allusion; amanuensis;and ambiguity.

In 1959 Verbal Behavior was given a veryunfavorable review by N. Chomsky, a wellknown linguist. This review has been repub-lished in several sources, is widely quoted,and is often credited with having discouragedlanguage scholars from a possibly more fa-vorable approach to Skinner's analysis. Thisseems reasonable, but the effect is probablyoverestimated. There had long been astrongly antideterministic and antibehav-iorist sentiment at the core of the humanitiesand the behavioral and social sciences. It isdifficult to believe that Chomsky could havedone much to intensify this sentiment, al-though the critique was enthusiastically re-ceived by those who were already convinced

369

JACK MICHAEL

that Skinnerian behaviorism was a badthing. MacCorquodale's (1969) "Verbal Be-havior: A Retrospective Appreciation" ap-peared inJEAB; it was an "attempt to clarifywhy Verbal Behavior is vulnerable to somemisunderstandings" and a reconstruction of"the salient points of the book's argument"(p. 831). One year later MacCorquodale(1970) critically described Chomsky's re-view. Since 1970, a great deal has been writ-ten on the topic of verbal behavior from abehavioral perspective, including an anal-ysis of linguistic theory by Julia (1983) andtwo introductory-level books (Peterson,1978; Winokur, 1976).

In describing Verbal Behavior as "an exer-cise in interpretation rather than a quan-titative extrapolation of rigorous experimen-tal results" (1957, p. 11), Skinner adds:

The lack of quanititative rigor is to someextent offset by an insistence that the con-ditions appealed to in the analysis be, sofar as possible, accessible and manipula-ble. The formulation is inherently prac-tical and suggests immediate technologicalapplications at almost every step. (p. 12)

One of the earliest of such applications wasZoellner's (1969) method for teachingEnglish composition at the college level. J.Vargas (1978) offered a program for teach-ing composition to elementary school pupils.Sloane, Endo, and Della-Piana (1980) pro-vided an analysis of and some suggestionsfor facilitating creativity, and Skinnerdescribed "How To Discover What YouHave To Say" (1981) in an application toprofessional writing. Johnson and Chase(1981) used some basic concepts from VerbalBehavior to develop a typology of verbal tasksfor the purposes of instructional design.Glenn (1983) used Verbal Behavior to analyzeclient maladaptive behavior (e.g., lying,denial, demanding) in the clinical situation;and Burns, Heiby, and Tharp (1983) used itin an analysis of auditory hallucinations.Sundberg (1983) used it to assess the lan-guage deficits of the mentally retarded, andthere are other applications not mentionedhere.

The papers just described are for the mostpart technological applications, based ontheory and on empirical observation, but arenot research in the formal sense. More for-mal research also has been conducted. Thebasic classification system (mands, tacts,etc.) was used by Salzinger (1958) to studythe verbal communications of emotionallydisturbed adolescents, and by Horner andGussow (1972) to study mother-child inter-actions of two 3-year-old black children. Boeand Winokur (1978) investigated the controlof echoic behavior in college students; Lee(1981) studied the independence of thespeaking and listening repertoires and, later(Lee & Pegler, 1982), the independence ofspelling and reading. Likewise, Lamarreand Holland (in press) have investigated theindependence of mands and tacts.Although the amount of formal research

directly related to Verbal Behavior is stillsmall, Skinner's "exercise in interpretation"has had a profound influence. One continu-ation of this theme is in the philosophicaland theoretical area. The behavioral episte-mology that developed out of Verbal Behavioris an active and growing alternative to tradi-tional approaches. Perhaps the most encour-aging development is the increasing revisionand refinement by others of Skinner'sanalysis of verbal behavior.A different approach that does not fit

readily into the main themes identified inthis paper, but which should be noted, isthat by Salzinger and his associates (Salz-inger, 1959, 1968, 1969, 1973, 1978; Salz-inger, Portnoy, & Feldman, 1964, 1966).Their work is unique in having employed amuch broader methodology to study lan-guage than is seen in any of the precedingthemes (which are mostly concerned with thespeaker rather than the listener). Skinner(1957), for example, was mainly concernedwith the speaker and has been criticized forfailing to deal more extensively with the ef-fects of verbal stimuli on the listener (e.g.,Parrott, 1984). He justifies his emphasis onthe grounds that in many respects verbalstimuli serve the same function that nonver-bal stimuli do (e.g., as conditioned elicitors,

370

VERBAL BEHA VIOR

discriminative stimuli, conditioned rein-forcers) and have no uniquely verbal fea-tures (Skinner, 1957, pp. 33-34)'. There arecases in which functioning as a listenerseems uniquely verbal, but this seems to bebecause of the listener's repertoire as aspeaker.Much of Salzinger's research has dealt

behaviorally with conventional psycholin-guistic issues, usually related to the effects ofverbal stimuli on a listener or reader. Heuses the cloze procedure (Taylor, 1953),Osgood's semantic differential (Osgood,Suci, & Tannenbaum, 1957), and otherrelated techniques, both as dependent andindependent variables (Salzinger, 1978; seealso Salzinger & Feldman, 1973; Salzinger &Salzinger, 1967). Recently (1984) he arguedthat behavioral psychologists should broadentheir methodological scope. That the verbaldependent variable seems to require a multi-ple classification of topographies is probablyrevelant to the paucity of operant researchon verbal behavior. Salzinger has been ableto produce and to inspire a steady output ofbasic behavioristic language research, mak-ing use of a much broader methodologicalperspective than has been seen in much ofthe other experimental literature.

VERBAL BEHAVIOR INNONHUMAN ANIMALS

Efforts to teach language to nonhumananimals is a thematic line of research thathas derived very little from the experimentalanalysis of behavior. However, in earlyoperant research, behavior was developed innonhuman animals that was deliberatelysimilar to parts of the human verbal reper-toire-for example, some components ofarithmetic behavior (Ferster & Hammer,1966). The most ambitious project wasPremack's work (1970, 1976), which wasconcerned with the extent to which chimpan-zees could behave appropriately with respectto certain logical and/or semantic relations.

In contrast to the other ape-languageresearchers described below, Premack didnot attempt to develop a general-purpose

functional language, but instead studied theeffects of training and generalization of asmall number of "concepts." His book(Premack, 1976) reports the results of thatresearch and includes insightful general in-terpretations of verbal behavior and its con-trolling relations. The first report of thiswork was published in JEAB (1970) andPremack's earlier contributions are not farfrom the mainstream of behavioral thought,but the work described in the book seems toowe more to a cognitive orientation than to abehavioral one. The training methods wereclearly behavioral in the explicit use of rein-forcement and in the manipulation of stim-uli, but the interpretation of the results is incommon-sense, semitechnical, or logicalterms that usually imply internal processes.Many human verbal responses are, in fact,under the control of complex abstract rela-tions such as the ones Premack studied, and,in spite of various methodological criticisms(Terrace, 1979a), he would seem to haveshown that chimps' behavior can be broughtunder the control of similar abstract rela-tions.

Early efforts to develop human languagein chimpanzees were directed at vocal be-havior, but probably because the ape's vocalapparatus does not make human noises eas-ily, if at all, such efforts were unsuccessful.Gardner and Gardner (1969) avoided thevocal difficulty by teaching a chimpanzee(Washoe) hand movements and positions(using the conventional signs of theAmerican Sign Language of the deaf) as ver-bal responses, a strategy that was soonfollowed by other researchers (Fouts, 1974;Patterson & Linden, 1981; Terrace, 1979b).When human adults teach children to

talk, they do not have to give any considera-tion to the complexities of language, viewedeither from a cognitive-linguistic perspectiveor from a behavioral one. They simply talkto the children, tell them the words forthings, actions, relations, etc., prompt themto use these words in appropriate situations,and seemingly without much instructionaleffort, the children become verbally fluent.By the time a child is 3 or 4 years old, there

371

JACK MICHAEL

is usually no question about the linguisticstatus of the vocal repertoire.Ape-language research could have had

similarly unambiguous results. The research-ers essentially tried to teach the apes to signthe same way children are taught to talk.There was considerable initial success in thatthe apes learned to make the individual signsseemingly under the same kinds of control-ling variables relevant to a child's earlywords. The vocabulary of signs grew rap-idly- in some cases at a rate similar to achild's early vocabulary growth- and therewas no clear reason to believe that the apeswould not soon be discussing the details ofthe project with the researchers. Had thishappened, the research would have been anenormous scientific success, easily fundable,and undoubtedly of great popular interest.The researchers themselves were properlyconservative, but early press coveragestrongly implied that dramatic results wereimminent. When Terrace (1979b) enteredthe field, it was easy to believe that progresswould be even more rapid.

Unfortunately, this did not happen. Theapes continued to learn more signs, but theirlanguage did not "take off' as a child's does.The work came under considerable criticismfrom ethologists, linguists, and psycholin-guists who questioned whether the apes'behavior was "really" language, and the ape-language researchers began to criticize eachother's work. The research was very expen-sive, and when grant funds dwindled, someof the projects had to be discontinued.Rumbaugh (1977) overcame the vocal in-

adequacy of his chimpanzee subject, Lana,by teaching her to select particular keys on aconsole. Each key had a distinctive embossedfigure (lexigram) on it, and the different lex-igrams were systematically related to envi-ronmental objects, properties, actions, orrelations. This approach has the advantageof replacing topography shaping with sim-pler training procedures, and of automaticrecording of responses rather than depen-dence on human observation and judgment.

Except for the type of language and theautomated management of the system with a

computer, the early work with Lana wasmuch like the work with the signing apes.Instead of being "taught signs," she was"taught lexigrams." On the other hand, se-quencing and grammatical relations were apart of the training contingencies: The com-puter that processed the input from theanimal's console would not accept (and thuswould not reinforce) inputs that were notproper according to the rules of the artificiallanguage that was being used. And as withthe signing apes, Lana might have devel-oped a repertoire of such complexity and ef-fectiveness that no one would have doubtedits linguistic status. She did acquire someverbal behavior according to behavioralcriteria, but again it did not reach the levelthat even quite young children reach, or alevel any more impressive than that achievedby the signing apes.More recent work by Savage-Rumbaugh,

Rumbaugh, and Boysen (1978a, 1978b,1980) with Lana and two additional chimps,Sherman and Austin, resembles to some ex-tent the work by Premack. The goal has notbeen to develop a general-purpose verbalrepertoire, but rather to test for specificcognitive capacities that are presumed to beessential to language. In a series of ingeniousexperiments involving tool usage, tool "nam-ing," categorical sorting, and cooperativebehavior between chimps, these reseachersseem to have demonstrated more complexand human-like forms of verbal behaviorthan appear systematically in any of theother projects.

It is difficult to judge the ape-languageresearch from a behavioral point of view,because the rationale, procedural descrip-tions, and interpretations of results areheavily cognitive in form. The animal is saidto be "associating" objects and their "names"and thus learning the "meanings" of the signsor lexigrams. When they sign or touch lex-igrams, they are said to be "using" the signsor lexigrams to "express these meanings, orto express their emotions or needs." If theirbehavior is true communication, their "use"of signs or lexigrams is done with "inten-tionality"; otherwise it is merely "responding

372

VERBAL BEHA VIOR

to stimuli in a rote manner." They "sym-bolically encode" various environmentalevents and relations, which permits themlater to "recall these representations." Whenthey react to the signs or lexigrams producedby others, they are responding "receptively,"that is, receiving "information," which theymay or may not "process" adequately. Thus,ape-language research provides many exam-ples of what Skinner has called "internal sur-rogates of the contingencies" (1977, p. 1), or"internalization of the environment" (p. 5).These hypothesized internal events could bedefined behaviorally, but it would be mucheasier if the language were more behavioralto begin with. A hopeful sign is the recentpaper by Savage-Rumbaugh (1984) inJEABin which she interpreted the projects men-tioned above (Savage-Rumbaugh et al.1978a, 1978b, 1980) in terms of the conceptsand general approach of Skinner's VerbalBehavior.

There seem to be two future directionsthat this type of research might take withrespect to the experimental analysis ofbehavior. First, the current ape-languageresearchers might become more familiarwith behavioral theory and research-pos-sibly as a result of supportive interest and of-fers of cooperative participation on the partof behavioral psychologists who are in-terested in this area. This would seem to beoccurring to some degree already. Anotherfuture development might be behavioralpsychologists doing their own research in thedevelopment of human language in non-human animals. Most of the accomplish-ments of the apes studied so far do not seemvery different from nonverbal behavior thathas already been demonstrated in laboratorysettings with smaller brained and much lessexpensive animals. From the perspective ofSkinner's Verbal Behavior, many of theelementary aspects of behaving as speakerand listener should be demonstrable withmonkeys. Some work, although not verybehavioral, is under way with the parrott(Pepperberg, 1981) and with sea lions(Schusterman & Krieger, 1984). Some of thebasic verbal functional relations may even

be studied with pigeons (Sundberg, 1984).To what extent such less expensive animalscan be usefully studied with respect to in-teresting aspects of verbal behavior is simplynot known at this time, but their use wouldcertainly be a reasonable approach. And ofcourse, it would be ideal if behavioral psy-chologists with extensive familiarity withVerbal Behavior and related research could dotheir own research with apes.

RULE-GOVERNED BEHAVIOR ANDHUMAN OPERANT RESEARCH

Another recent research development isone that in many respects resembles theoperant conditioning of adult verbal behav-ior described above. It started with attemptsto understand why human schedule perfor-mance (e.g., on a fixed-interval schedule)did not resemble that of other organisms thathad been studied. Instead of the typicalfixed-interval scallop, humans on fixed-interval schedules in a laboratory setting(e.g., pushing buttons for monetary rein-forcement) either showed a low rate patternwith only a few responses occurring duringthe interreinforcement interval, or respondedat a high rate throughout the entire interval.A good deal of recent research has beenaimed at understanding these human devia-tions from lower animal schedule perfor-mance. It has become clear that they arerelated to instructions given by the ex-perimenter, implied from the experimenter'sbehavior, or somehow occur as a part of thesubject's own verbal repertoire as a form of"self-instruction" (Harzem, Lowe, &Bagshaw, 1978; for a review, see Lowe,1979). The topic is sometimes considered interms of sensitivity to schedule parameters,with instructional variables affecting sub-jects' sensitivity (Shimoff, Catania, & Mat-thews, 1981).Although this research examines human

operant performance (repeated respondingunder the control of a reinforcement sched-ule) and a type of factor that affects it (in-structions that are given, implied, or self-generated), it can also be conceptualized as a

373

374 JACK MICHAEL

convenient way to study the effects of verbalstimuli, including those that are self-gen-erated, on nonverbal behavior. There is con-siderable current interest in rule-governedbehavior, but not much research. Here is away to study rule-governed behavior with aconvenient and well understood dependentvariable. A similar approach is involved inthe study of the effects of verbal behavior ondelayed matching-to-sample performance(Parsons, Taylor, & Joyce, 1981), and re-peated acquisition (Vaughan, in press). Thisapproach could also be taken with respect tofeature-value effects, generalized matching,and stimulus equivalence.

This would seem to be the beginning of anexperimental analysis of the effects of verbalstimuli on nonverbal behavior, as well as away of studying the effects of one's own ver-bal behavior on other behavior occurring atthe same time. This same topic is also underinvestigation in the applied field as the studyand the generation of correspondence be-tween verbal and nonverbal behavior, and itis to be hoped that the two lines of investiga-tion will merge.The area of verbal behavior is the link be-

tween principles of nonverbal behavior thatwe share with other species and our uniquelyhuman social and intellectual accomplish-ments. Our efforts to understand this areafrom a behavioral perspective are just begin-ning, but it seems to me that we have an ex-cellent start and our progress is accelerating.

REFERENCES

Baer, D. M., & Guess, D. (1971). Receptive train-ing of adjectival inflections in mental retardates.Journal ofApplied Behavior Analysis, 4, 129-139.

Baer, D. M., Peterson, R. F., & Sherman, J. A.(1965). Building an imitative repertoire by programmingsimilarity between child and model as discriminative forreinforcement. Paper presented at the meeting of theSociety for Research in Child Development, Min-neapolis, MN.

Baer, D. M., & Sherman, J. A. (1964). Reinforce-ment control of generalized imitation in youngchildren. Journal of Experimental Child Psychology, 1,37-49.

Bloom, K. (1975). Social elicitation of infant vocalbehavior. Journal ofExperimental Child Psychology, 20,51-58.

Bloom, K. (1977). Patterning of infant vocal behav-

ior. Journal of Expenrmental Child Psychology, 23,367-377.

Bloom, K. (1979). Evaluation of infant vocal con-ditioning. Journal ofExperimental Child Psychology, 27,60-70.

Bloom, K., & Esposito, A. (1975). Social condition-ing and its proper control procedures. Journal ofEx-perimental Child Psychology, 19, 209-222.

Boe, R., & Winokur, S. (1978). A procedure forstudying echoic control in verbal behavior. Journalof the Experimental Analysis of Behavior, 30, 213-217.

Brackbill, Y. (1958). Extinction of the smilingresponse in infants as a function of reinforcementschedule. Child Development, 29, 115-124.

Burns, C. E. S., Heiby, E. M., & Tharp, R. G.(1983). A verbal behavior analysis of auditoryhallucinations. Behavior Analyst, 6, 133-143.

Campbell, C. R., & Stremel-Campbell, K. (1982).Programming 'loose training" as a strategy tofacilitate language generalization. Journal ofAppliedBehavior Analysis, 15, 295-301.

Chomsky, N. (1959). Review of B. F. Skinner's Ver-bal Behavior. Language, 35, 26-58.

Dollard, J., & Miller, N. E. (1950). Personality andpsychotherapy. New York: McGraw-Hill.

Faw, G. D., Reid, D. H., Schepis, M. M., Fitzgerald,J. R., & Welty, P. A. (1981). Involving institu-tional staff in the development and maintenance ofsign language skills with profoundly retarded per-sons. Journal of Applied Behavior Analysis, 14,411-423.

Ferster, C. B., & Hammer, C. E., Jr. (1966).Synthesizing the components of arithmeticbehavior. In W. K. Honig (Ed.), Operant behavior.Areas of research and application (pp. 634-676). NewYork: Appleton-Century-Crofts.

Fouts, R. S. (1974). Capacities for languages ingreat apes. In Proceedings of the 18th International Con-gress of Anthropological and Ethnological Sciences (pp.1-20). Mouton: The Hague.

Gardner, R. A., & Gardner, B. T. (1969). Teachingsign language to a chimpanzee. Science, 165,664-672.

Glenn, S. S. (1983). Maladaptive functional rela-tions in client verbal behavior. Behavior Analyst, 6,47-56.

Greenspoon, J. (1955). The reinforcing effect of twospoken sounds on the frequency of two responses.Amencan Journal of Psychology, 68, 409-416.

Greenspoon, J. (1962). Verbal conditioning andclinical psychology. In A. J. Bachrach (Ed.), Ex-perinmntal foundations of clinical psychology (pp.510-553). New York: Basic Books.

Guess, S., Sailor, W., Rutherford, G., & Baer, D. M.(1968). An experimental analysis of linguisticdevelopment: The productive use of the plural mor-pheme. Journal of Applied Behavior Analysis, 1,297-306.

Hake, D. F. (1982). The basic-applied continuumand the possible evolution of human operant socialand verbal research. Behavior Analyst, 5, 21-28

Hart, B. M., & Risley, T. R. (1968). Establishinguse of descriptive adjectives in the spontaneousspeech of disadvantaged preschool children. JournalofApplied Behavior Analysis, 1, 109-120.

Hart, B. M., & Risley, T. R. (1975). Incidental

VERBAL BEHA VIOR 375

teaching of language in the preschool. Journal ofAp-plied Behavior Analysis, 8, 411-420.

Harzem, P., Lowe, C. F., & Bagshaw, M. (1978).Verbal control in human operant behavior.Psychological Record, 28, 405-423.

Haugan, G. M., & McIntire, R. W. (1972). Com-parison of vocal imitation, tactile stimulation, andfood as reinforcers for infant vocalizations.Developmental Psychology, 6, 201-209.

Holz, W. C., & Azrin, N. H. (1966). Conditioninghuman verbal behavior. In W. K. Honig (Ed.),Operant behavior: Areas of research and application (pp.790-826). New York: Appleton-Century-Crofts.

Horner, V. M., & Gussow, J. D. (1972). John andMary: A pilot study in linguistic ecology. In C. B.Cazden, V. P. John, & D. Hymes (Eds.), Functionsof language in the classroom (pp. 155-194). New York:Teachers College Press.

Hulsebus, R. C. (1973). Operant conditioning ofinfant behavior: A review. In H. W. Reese (Ed.)Advances in child development and behavior (Vol. 8, pp.111- 158). New York: Academic Press.

Israel, A. C. (1978). Some thoughts on correspon-dence between saying and doing. Journal ofAppliedBehavior Analysis, 11, 271-276.

Johnson, K. R., & Chase, P. N. (1981). Behavioranalysis in instructional design: A functionaltypology of verbal tasks. Behavior Analyst, 4,103-121.

Julia, P. (1983). Explanatory models in linguistics: Abehavioral perspective. Princeton, NJ: PrincetonUniversity Press.

Kanfer, F. H. (1968). Verbal conditioning: A reviewof its current status. In T. R. Dixon & D. L. Hor-ton (Eds.), Verbal behavior and general behavior theory(pp. 254-290). Englewood Cliffs, NJ: Prentice-Hall.

Keller, F. S., & Schoenfeld, W. N. (1950). Principlesofpsychology. New York: Appleton-Century-Crofts.

Keller, M. F., & Bucher, B. D. (1979). Transferbetween receptive and productive language indevelopmentally disabled children. Journal ofAppliedBehavior Analysis, 12, 311.

Krasner, L. (1958). Studies of the conditioning ofverbal behavior. Psychological Bulletin, 55, 148-170.

Lamarre, J., & Holland, J. G. (in press). Thefunctional independence of mands and tacts.Journalof the Experimental Analysis of Behavior.

Lee, V. L. (1981). Prepositional phrases spoken andheard. Journal of the Experimental Analysis of Behavior,35, 227-242.

Lee, V. L., & Pegler, A. M. (1982). Effects onspelling of training children to read. Journal of theExperimental Analysis of Behavior, 37, 311-322.

Lowe, C. F. (1979). Determinants ofhuman operantbehaviour. In M. D. Zeiler & P. Harzem (Eds.),Advances in analysis of behaviour: Vol. 1. Reinforcementand the organization of behaviour (pp. 159-192).Chichester, England: Wiley.

Lutzker, J. R., & Sherman, J. A. (1974). Producinggenerative sentence usage by imitation and rein-forcement procedures. Journal of Applied BehaviorAnalysis, 7, 447-460.

MacCorquodale, K. (1969). B. F. Skinner's VerbalBehavior: A retrospective appreciation. Journal of theExperimental Analysis of Behavior, 12, 831-841.

MacCorquodale, K. (1970). On Chomsky's reviewof Skinneres Verbal Behavior. Journal of the ExperimentalAnalysis of Behavior, 13, 83-99.

Metz, J. R. (1965). Conditioning generalized imita-tion in autistic children. Journal ofExperimental ChildPsychology, 2, 389-399.

Millar, W. S. (1976). Operant acquisition of socialbehaviors in infancy: Basic problems and con-straints. In H. W. Reese (Ed.), Advances in childdevelopment and behavior (Vol. 11, pp. 107-140). NewYork: Academic Press.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H.(1957). The measurement of meaning. Urbana:University of Illinois Press.

Parrott, L. J. (1984). Listening and understanding.Behavior Analyst, 7, 29-39.

Parsons,J. A., Taylor, D. C., &Joyce, T. M. (1981).Precurrent self-prompting operants in children:"Remembering."Journal of the Experimental Analysis ofBehavior, 36, 253-266.

Patterson, F., & Linden, E. (1981). The education ofKoko. New York: Holt, Rinehart & Winston.

Pepperberg, I. M. (1981). Functional vocalizationsby an African grey parrott (Psittacus erithacus).Zeitschriftfur Tierpsychologie, 55, 139-160.

Peterson, N. (1978). An introduction to verbal behavior.Grand Rapids, MI: Behavior Associates.

Peterson, R. F. (1968). Imitation: A basic behav-ioral mechanism. In H. N. Sloane, Jr., & B. D.MacAulay (Eds.), Operant procedures in remedial speechand language training (pp. 61-74). Boston: HoughtonMifflin.

Poulson, C. L. (1983). Differential reinforcement ofother-than-vocalization as a control procedure inthe conditioning of infant vocalization rate. Journalof Experimental Child Psychology, 36, 471-489.

Premack, D. (1970). A functional analysis of lan-guage. Journal of the Experimental Analysis of Behavior,14, 107-125.

Premack, D. (1976). Intelligenc in ape and man. Hills-dale, NJ: Erlbaum.

Rheingold, H. L., Gewirtz, J. L., & Ross, H. W.(1959). Social conditioning of vocalizations in theinfant. Journal of Comparative and PhysiologicalPsychology, 52, 68-73.

Routh, D. K. (1969). Conditioning of vocal responsedifferentiation in infants. Developmental Psychology, 1,219-226.

Rumbaugh, D. M. (Ed.). (1977). Language learning bya chimpanzee: The Lana project. New York: AcademicPress.

Sailor, W., & Taman, T. (1972). Stimulus factors inthe training of prepositional usage in three autisticchildren. Journal of Applied Behavior Analysis, 5,183-190.

Salzinger, K. (1958). A method of analysis of theprocess of verbal communication between a groupof emotionally disturbed adolescents and theirfriends and relatives. Journal of Social Psychology, 47,39-53.

Salzinger, K. (1959). Experimental manipulation ofverbal behavior: A review. Journal of GeneralPsychology, 61, 65-94.

Salzinger, K. (1968). On the operant conditioningof complex behavior. In J. M. Shlien & H. Hunt(Eds.), Research in psychotherapy (pp. 122-129).

376 JACK MICHAEL

Washington, DC: American Psychological Associa-tion.

Salzinger, K. (1969). The place of operant condi-tioning of verbal behavior in psychotherapy. In C.Franks (Ed.), Behavior therapy: Appraisal and status(pp. 375-395). New York: McGraw-Hill.

Salzinger, K. (1973). Some problems of responsemeasurement in verbal behavior: The response unitand intraresponse relations. In K. Salzinger & R.S. Feldman (Eds.), Studies in verbal behavior: An em-pirical approach (pp. 5-15). New York: PergamonPress.

Salzinger, K. (1978). Language behavior. In A. C.Catania & T. A. Brigham (Eds.), Handbook of ap-plied behavior analysis: Social and instructional processes(pp. 275-321). New York: Irvington.

Salzinger, K. (1984, May). The maintenance of verbalbehavior. Address presented at the meeting of theAssociation for Behavior Analysis, Nashville, TN.

Salzinger, K., & Feldman, R. S. (Eds.). (1973).Studies in verbal behavior: An empirical approach. NewYork: Pergamon Press.

Salzinger, K., Portnoy, S., & Feldman, R. S. (1964).Verbal behavior of schizophrenic and normal sub-jects. Annals of the New York Academy of Sciences, 105,845-860.

Salzinger, K., Portnoy, S., & Feldman, R. S. (1966).Verbal behavior in schizophrenics and some com-ments toward a theory of schizophrenia. In P. H.Hoch & J. Zubin (Eds.), Psychopathology of schizo-phrenia (pp. 98-128). New York: Grune & Stratton.

Salzinger, K., & Salzinger, S. (Eds.). (1967).Research in verbal behavior and some neurological implica-tions. New York: Academic Press.

Savage-Rumbaugh, E. S. (1984). Verbal behavior ata procedural level in the chimpanzee. Journal of theExperimnntal Analysis of Behavior, 41, 223-250.

Savage-Rumbaugh, E. S., Rumbaugh, D. M., &Boysen, S. (1978a). Symbolic communicationbetween two chimpanzees (Pan troglodytes). Science,201, 641-644.

Savage-Rumbaugh, E. S., Rumbaugh, D. M., &Boysen, S. (1978b). Linguistically mediated tooluse and exchange by chimpanzees (Pan troglodytes).Behavioral and Brain Sciences, 1, 539-629. (Includescommentary)

Savage-Rumbaugh, E. S., Rumbaugh, D. M., &Boysen, S. (1980). Do apes use language?American Scientist, 68, 49-61.

Schusterman, R. J., & Krieger, K. (1984). Cali-fornia sea lions are capable of semantic comprehen-sion. Psychological Record, 34, 3-23.

Sheppard, W. C. (1969). Operant control of infantvocal and motor behavior. Journal of ExperimentalChild Psychology, 7, 36-51.

Shimoff, E., Catania, A. C., & Matthews, B. A.(1981). Uninstructed human responding: Sen-

sitivity of low-rate performance to schedule con-tingencies. Journal of the Experimental Analysis ofBehavior, 36, 207-220.

Skinner, B. F. (1938). The behavior of organisms.New York: Appleton-Century-Crofts.

Skinner, B. F. (1945). The operational analysisof psychological terms. Psychological Review, 52,270-277.

Skinner, B. F. (1953). Science and human behavior. NewYork: Macmillan.

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

Skinner, B. F. (1977). Why I am not a cognitivepsychologist. Behaviorism, 5(2), 1-10.

Skinner, B. F. (1979). The shaping of a behaviorist.New York: Knopf.

Skinner, B. F. (1981). How to discover what youhave to say-A talk to students. BehaviorAnalyst, 4,1-7.

Sloane, H. N., Endo, G. T., & Della-Piana, G. M.(1980). Creative behavior. Behavior Analyst, 3(1),11-21.

Sundberg, M. L. (1983). Language. In J. L.Matson & S. E. Breuning (Eds.), Assessing the men-tally retarded (pp. 285-310). New York: Grune &Stratton.

Sundberg, M. L. (1984, May). Teaching verbalbehavior to pigeons. Paper presented at the meeting ofthe Association for Behavior Analysis, Nashville,TN.

Taylor, W. L. (1953). "Cloze procedure": A new toolfor measuring readability. Journalism Quarterly, 30,415-433.

Terrace, H. S. (1979a). Is problem-solving lan-guage? Journal of the Experimental Analysis of Behavior,31, 161-175.

Terrace, H. S. (1979b). Nim. New York: Knopf.Todd, G. A., & Palmer, B. (1968). Social reinforce-

ment of infant babbling. Child Development, 39,591-596.

Vargas, E. A. (in press). Intraverbal behavior. InL. J. Parrott & P. Chase (Eds.), Behavioral analysis ofverbal behavior: Issues and extensions. Springfield, IL:Thomas.

Vargas, J. S. A behavioral approach to the teachingof composition. Behavior Analyst, 1(1), 16-24.

Vaughan, M. E. (in press). Repeated acquisitionand rule-governed behavior. Journal of the Experimen-tal Analysis of Behavior.

Weisberg, P. (1963). Social and nonsocial condition-ing of infant vocalizations. Child Development, 34,377-388.

Winokur, S. (1976). A primer of verbal behavior: Anoperant view. Englewood Cliffs, NJ: Prentice-Hall.

Zoellner, R. (1969). Talk-write: A behavioralpedagogy for composition. College English, 30,267-320.