[go to other interviews]

Dave Palmer, who teaches Behavioral Psychology and Statistics at Smith College and Verbal Behavior at Simmons College, generously shares with us his deep understanding and appreciation of Skinner’s masterpiece –unique to delve into the comprehension of human behavior.

Which part of Skinner’s Verbal Behavior still intrigues you the most, even after many years of teaching it and studying it?

It is a remarkable book, with riches scattered throughout it. I cannot refrain from mentioning a number of passages that have affected me strongly.

It was in this book that he first identified the distinction between experimental analysis and interpretation (p. 11), a distinction that I have found to be fundamental in making sense of behavioral phenomena.

His discussion of "meaning" (p. 14, 112, and elsewhere) is wholly at odds with the traditional notion that meaning can be found in the structure of verbal behavior.

On p. 22, 254, and elsewhere, he discusses the important concept of latent behavior and the probability of response.

On p. 56-63 he first discusses how echoic behavior (and other minimal, or atomic, repertoires) "short-circuit the process of successive approximations" in the acquisition of behavior –a simple point but one with dramatic implications for complex behavior and the transmission of cultural practices. Skinner interpreted recall as behavior under control of current stimuli, not past stimuli (p. 143), a formulation quite different from that of traditional theories.

The section on the speaker as listener is important for its coverage of automatic shaping (pp. 163-166 and for introducing the topic of transfer of function (p. 155).

I love the off-hand remark on p. 189 that the classification of verbal operants is useful only for clarifying various kinds of controlling relations. Pure verbal operants are probably rare.

The chapters on multiple control are extremely important and under-appreciated. Almost all verbal behavior is under multiple control. On p. 228 he alludes to the additivity of stimulus control, another concept that I find indispensable when trying to understand problem solving and many other examples of complex behavior.

On p. 269 he points out that the listener is behaving verbally as well, usually by saying the same thing as the speaker and at about the same time. In effect, the speaker induces the listener to engage in the same verbal behavior he is emitting. But this is more than echoic behavior; a listener sometimes jumps ahead of the speaker, or he can sometimes supply a word or phrase for a faltering or forgetful speaker.

His subsequent discussion of the behavioral changes that take place as we come to understand a complex text, as well as the kinds of writing that we enjoy reading is brilliant (pp. 271-280).

But to answer the question as posed, I am particularly interested in his discussion of the conditioning of the behavior of the listener (pp. 357-367). It is surely one of the most important features of verbal behavior that it can alter the repertoire of the listener rapidly and in remarkable ways, and it is among the least understood. Skinner alludes to a long and complex history, but he stops short of a detailed account of how this conditioning happens. It is clear that listener behavior must be an important part of any explanation, but unfortunately much relevant listener behavior is covert. As a consequence, our accounts are speculative. But we must not shrink from the interpretive task. The conditioning of the behavior of the listener is not only one of the most commonplace of phenomena, it is among the most important. Behavior analysis must offer a cogent account if it is to claim to be a comprehensive science of human behavior.

Which part of Skinner’s Verbal Behavior do you consider as having the deepest implications for understanding human behavior?

As I foreshadowed in my previous answer, I believe that two aspects of Skinner's analysis are particularly relevant to those aspects of behavior that are distinctively human.

The first is his concept of atomic repertoires, or as he sometimes calls them, minimal repertoires. An atomic repertoire is a set of elementary responses, each evoked by a distinctive stimulus, that can be combined in virtually any permutation simply by the arrangement of corresponding discriminative stimuli. Examples include echoic responses, imitative responses, textual responses, sight-reading of music, Morse Code, and rule-governed responses. As I argued in a paper on the topic, atomic repertoires permit the extremely rapid transmission of adaptive cultural practices. An adaptive response shaped in one organism through slow and tedious shaping can be transmitted to someone else in a single trial. Tens of thousands of years must have elapsed before one person discovered that planting a seed ultimately leads to a plant that produces many seeds, but once discovered, the strategy could be acquired by a second person virtually immediately. Such repertoires are not wholly unique to humans, but humans have a variety of "alphabets" of atomic responses, many of them verbal, and our exceptional facility with them enables the complexity of human behavior to leave that of all other organisms far behind. Skinner lays the groundwork for this analysis in his discussion of the echoic and extends it in his subsequent "Operant Analysis of Problem Solving."

The second aspect of Skinner's analysis that is particularly relevant to the distinctiveness of human behavior is the conditioning of the behavior of the listener. I can announce to a listener that snakes with triangular heads are venomous, and he will, perhaps much later, avoid snakes with triangular heads. The conditioning happens with no conspicuous behavior on the part of the listener and no conspicuous reinforcement. Relevant behavior and reinforcement does in fact occur but is difficult to observe, measure, or manipulate. As a consequence, our accounts must be speculative, but the importance of such conditioning for human behavior is beyond doubt.

Which part(s) of Skinner’s book have you found to be the most challenging for your students? Do you have an idea as to why this is the case?

The chapters on the autoclitic are without a doubt the hardest part of Skinner's analysis for students. The reason for this is that the autoclitic is a single name given to a heterogeneous set of phenomena, many of which are poorly understood. Descriptive autoclitics, such as "I hesitate to say..." or "I heard that...." are easily understood, but it is much harder to specify the controlling variables for subtle grammatical distinctions like the conjugation of verbs. The subject of grammar is vast and extremely complex, and Skinner could only point to relevant variables. He does not offer a detailed account. Students are usually only dimly aware of the subtleties of grammar; few of them are able to pick up the threads of Skinner's discussion and weave them into an explanation for a novel example.  

Are there common or recurrent observations from your students concerning Skinner’s book, that you can share with us and that might be considered as an indicator of your students’ appreciation for Skinner’s work?

Students like the topic of multiple control, particularly as it has been extended in the topic of joint control. Other favorites are the many faces of response probability, ranging from latent behavior with virtually no probability through latent behavior with high probability, covert, low-amplitude behavior, to overt, high-amplitude behavior. The topic of automatic reinforcement and Skinner's discussion of the ways a verbal community shapes of tacts of private events are also popular. Finally, they all seem to appreciate the interpretation/experimental analysis distinction, and they understand that a behavioral interpretation based on laboratory principles is superior to accounts, however sophisticated they appear, that rest on hypothetical structures or processes.

Can you remember any unusual or particularly thought-provoking question a student has posed as a result of their reading of Skinner’s book?

Under the heading of multiple control, how do we account for the temporal gradient in the following everyday phenomenon: If someone mentions an unusual place or event – for example, if someone says he met his wife in Kazakhstan – and if you come across another reference to Kazakhstan a little later, you will have an "Aha!" response ("Kazakhstan again! What a coincidence!") as if that response were rendered "hot" by the previous reference. The response will remain "hot" all day, probably all week, perhaps longer, but it will recede to normal eventually, Why? One student suggested that the panorama of controlling stimuli from one day to the next tend to be more similar than from one week to the next, one month to the next, one year to the next. Perhaps the gradient arises from such changes, including, perhaps especially, intra-organismic changes. I don't have a well developed response to such questions myself, but these are the sorts of discussions Skinner's book occasions.

[go to other interviews]