Word meaning: A verbal behavior account
Barry Lowenkron
California State University, Los Angeles
Return to Table of Contents
Presented at the annual convention of the Association for Behavior Analysis, Washington DC, May, 2000

    I think it can be fairly said  that one of the most significant impediments to the general acceptance of behavior analysis as a complete account of human functioning is the problem of word meaning. The notion of word meaning is still one of the most cognitive of notions, but some account is mandatory if the behavioral approach is ever to provide a comprehensive description of thought and language.  To see some possible solutions to this problem, we may begin by appreciating the fact that current usages of the term meaning refers to two different kinds of behavior.  One kind is  the expressive behavior of the speaker who expresses meanings, and the other is the receptive behavior of the listener who gets, or comprehends meaning.  Skinner’s account of the expressive behavior of the speaker is fairly complete, and so it is the latter topic, the receptive behavior of the listener, that most needs to be considered.


    Let us begin this examination by considering  a simple form of receptive behavior: selecting stimuli in response to other stimuli. as in a conditional discrimination. In this task an array of comparison objects are presented,  here a triangle and a square, and the subject must select one object  in response to the sample object – here a square. On the usual account of this behavior, henceforth referred to as unmediated selection., it is said that due to a history of reinforcement, the sample serves as a conditional stimulus, making one of the comparisons function as a SD for a selection response. Thus, here, the sample square functions to make the comparison square act as an SD.  This produces an increase in the momentary rate of the pointing response, thereby causing the square comparison to be selected.


    Where one of the stimuli is a word, the same account applies unchanged.  Thus, as we see in this next picture (Figure 2),  we may explain the selection of a square, in response to the spoken word “square” and likewise the selection of  the printed word “square” in response to the object, in exactly the same terms:  The sample, as a conditional stimulus, makes one of the comparison into an SD  for the pointing response.

    But the problem with this unmediated selection is that it can’t account for the emergence of untrained relations.  This is because the same mechanism operates for both relational and arbitrary matching.  And so as we see in Figure 3, given the proper reinforcement history, just as the sample square may cause the comparison square to function as an SD for the pointing response, so may the sample square cause the color blue to function as an SD for the pointing response. Unmediated selection does not appreciate that a special relation, identity, exists in one case but  not the other.
    Likewise,  in describing relations between words and objects we find similar problems. Thus, if the only association between a word and an object is based on the unmediated selection I have just described, then in word-object matching the subject can respond to no other relation between the stimuli except what was trained  (Figure 4).   Thus, after a subject is trained to select the appropriate shapes in response to the words square and circle alone, and to respond to stimuli containing the words under and over with other shapes; unmediated selection cannot account for generalization to novel combinations of these words so that, for example, subjects select one and the same stimulus in response to the phrase “square over circle” and “circle under square”, and select a second stimulus in response to both the phrase “circle over square” and the phrase “square under circle”.
    In summary here, the notion of unmediated selection appears to be intrinsically unable to account for accurate responding to generalized  relations. Clearly, an alternate account is needed.
    What follows proposes just such an account: an alternate based on the notion of joint control.  As discussed in recent articles (see papers and articles cited on this web site) joint control involves nothing more than the usual kind of operant stimulus control, except that under joint control two verbal stimuli jointly exert stimulus control over a single, common verbal topography.   Now while this may not sound like much, this form of stimulus control seems to have exceedingly powerful and important effects.

    As a demonstration of joint control consider the task shown in Figure 5 in which you must locate a particular 6-digit number in an array in response to the spoken sample 939173.  Take a moment to find 939173 in this array.  Now finding the correct number  required joint control. This next slide illustrates what this is.


    Look at the top of Panel A. in  Figure 6.  First off, after you read the spoken sample 939173, you began to peruse the array looking for that number.  As you did so, you repeated the sample number.  In the language of Skinner’s verbal operants, this would be described  as echoic rehearsal of the sample, and it is illustrated along the top of Panel A.
    Second, as illustrated in the bottom of Panel A, as you perused the array, you also attempted to tact each 6-digit number, and you continued doing this until, as illustrated here, you encountered a particular  6-digit number that you could emit both as a tact, and simultaneously as the rehearsed self-echoic.
That is, at some point you could say 939173 both as a tact of the printed number, and jointly as a self echoic rehearsal of the spoken number.  And so, at this point you were repeating the number 93173 under joint tact/echoic control. This event, this onset of joint control, is a unique source of stimulus control because it only happens when the specified comparison is encountered.  Indeed, the onset of this joint self-echoic/tact control is in fact the only possible way you can identify which number in the array was the one I said.
    Now it is also important to note here that even though this is a matching to sample task,  you were all able to recognize the specified number without actually emitting a pointing response to the number itself.  Rather, you simply relied upon the onset of joint control to determine when you had located the correct number.   It would thus appear that the onset of joint control is the event one would identify as the event of recognizing the specified number set.  You could then report this recognition event, by emitting the autoclitic  “I found the number”.
    But, as illustrated in the lower part of the figure, in Panel B, had there been a request that you point to the specified number set,  you could have easily conformed to my mand, and pointed to it..  As discussed elsewhere, such a pointing  response may be described as a selection-based  autoclitic response, a response reporting to others which number in the array entered into joint stimulus control with the self-echoic you were already rehearsing.  Let’s look at this more closely.


    In Figure 7) the sample is: Black dot in a smaller pentagon.  Now again, as you peruse the comparisons you rehearse the sample phrase as a self-echoic.  And when a comparison is found that evokes this same topography, both as a tact of that comparison,  and jointly as an echoic rehearsal of the sample, you report the source of this joint control by an autoclitic pointing response.
    Together, this example and the prior number-finding task, illustrate that the joint-control event is generic, and thus independent of any particular stimuli.  Thus, in these two tasks it didn’t matter whether it was the names of numbers or the names of colors, shapes, or relative sizes.  In all cases,  whenever the self-echoic comes under joint control, this event, this onset of joint control, is itself a common generic event.  And it is this generic event that evokes a report of recognition and/or a pointing response.      As we shall see next, the generic nature of joint control makes it surprisingly ubiquitous in our verbal behavior: producing a variety of behavioral phenomena typically attributed to the semantic notion of word meaning.


    We shall begin this examination with a consideration of how words are tied to the objects they specify. That is, exactly how do words specify objects and events?
    Lets look at the experiment shown in Figure 8.  In this experiment, prior to learning a matching to sample task, we trained retarded children to make the handsigns shown in Panel A to each of the corresponding shapes when these were shown alone.  In the next phase of the experiment, -- shown in Panel B-- the children were trained to make the correct handsign to the sample, and then to rehearse the handsign over a delay interval, and then to continue to rehearse the handsign as they matched it to the correct comparison.  Subjects thus learned to use the handsigns to mediate identity matching to sample behavior.  Subsequently, when identity matching was tested with the novel stimuli shown in Panel C,  there was no generalized matching.  But after only training handsigns to these novel stimuli, subjects immediately showed unreinforced generalized identity matching with these new stimuli, thus illustrating that this form of stimulus control, joint control over matching, somehow incorporates relations between the stimuli, so that one subsequently novel stimulus may indeed specify the selection of an identical other.  That is, generalized identity matching to sample.


    The way this works is not hard to see.  Thus, as we see in Figure 9 the handsign to the two-dot sample is a tact, and maintaining the handsign over the delay interval is a self-duplic.  And selection of the comparison occurs when the two sources of  control operate jointly.  Thus we see that the handsign is jointly evoked both as a self-duplic rehearsal, and simultaneously as a tact, only with one particular comparison.   Thus the two dots, as a sample, evoke the handsign illustrated, and if this handsign is maintained over the interval (as a duplic), it is jointly evoked by the two-dot comparison.  Likewise, in the lower part of the figure, in Panel B the line, as a sample, evokes the handsign illustrated, and if this handsign is maintained over the interval, it is jointly evoked by the line comparison.
    The way one stimulus specifies another is thus revealed. The sample stimulus specifies a comparison stimulus because the sample stimulus evokes a particular handsign, and this handsign can only enter into joint control with one particular comparison.  Thus the two-dot sample evokes the  particular handsign shown here, and that handsign can occur under joint control only with the-two dot comparison.   Likewise in the bottom figure, the sample line specifies the line comparison because the sample line evokes the illustrated handsign, and that handsign only enters into joint control with the line comparison.

    Now with virtually no modification, this account provides a simple explanation of how names specify objects. Thus, as we see in Figure 10, suppose the shape to be selected was named by the experimenter providing a handsign..  Thus, as we see here, by providing  a handsign, the experimenter is telling the subject which shape to select.  That is, the shape that enters into joint control with the given  handsign.  Again, the sample stimulus specifies a particular comparison stimulus because only that comparison evokes the same handsign as the sample.
    And of course the jump to vocal language is pretty obvious. Thus as we see in Figure 11)  the subject is still selecting the comparison tact that enters into joint control with  the rehearsed phrase “two dots”.  The only difference is that the speaker’s response is vocal rather than a handsign. Thus we see how a word heard by the listener specifies an object.  That is, the heard word evokes the topography to be rehearsed in the listener, and that topography, in turn, acts to specify one particular comparison stimulus: the stimulus that itself evokes the same topography.
    Given this we may now behavioralize that aspect of word meaning usually called reference. We might say that for the listener, the referent of a particular word  is that object, event, or relation that enters into joint control with that word.
    Of course this is not restricted to single words.  Thus as we see in Figure 12, by saying the phrase Black dot in the smaller pentagon, the speaker specifies to the listener that topography for rehearsal and thus that object, and no other, for selection.
    Given these examples, I would propose that this specification function serves the role of what is ordinarily called DESCRIPTION.  I would propose that in Figure 13, the words square on circle function as a description of one these stimuli, but not the other, because these word enter into joint control with a tact of one stimulus but not the other.
    Attendant to the role joint control plays  in specification and description, we can also see a role for joint control in what is ordinarily called RECOGNITION from a description. To see this, consider your behavior earlier when you were asked to find the 6-digit number. In that case I know you all could have reported that you did indeed recognize the specified number, even if none of you actually emitted a selection response such as pointing to the screen.  But without an overt pointing response what is meant behaviorally by the term recognition?
    The notion of joint control provides a simple answer.  Can you find the circle in a square? You know you have recognized something from its description when your rehearsal of that description enters into joint control with a tact of the object recognized.  Thus at the moment you are able to emit the entire phrase circle in square both as a self-echoic and also as a tact of a particular stimulus, you would say you recognized that stimulus as the one described by saying something like “there it is” or “I  found it”.
     We should note however, that this only happens the first time you select in response to the description. If you are subsequently asked to point to that stimulus again, you no longer need repeat the description and respond under joint control.  Indeed, given the immediate history of reinforcement for pointing to that stimulus in the presence of the description, one would think the behavior is now just unmediated responding.
    Finally, one other aspect of word meaning that we need to cover is the notion of comprehension.  What does it mean to comprehend a description? Here is an example.  Given you the description point oh bar point, you would  first say that you don’t  really comprehend it.  And if you see the shapes shown in Figure 15, your initial reaction is not much changed.  But as you peruse these comparison , you notice that the bottom set  can be tacted in such a way as to fit the description.  That is, the description point oh bar point, can function jointly as a self-echoic repetition of my spoken sample, and simultaneously as a tact of that one comparison.  You can say point oh bar point both as a self-echoic of what you read, and also, however weakly, as a tact of the bottom line of the figure. This would suggest that we say that we comprehend a new description only when the self-echoic enters into joint control, but not before.

    It can only be briefly noted here that this account goes beyond the comprehension of nouns, but extends to the description of abstract relations as well  Thus, we are able to identify box on circle and circle under box  with one stimulus, and box under circle and circle on box with another only because of the manner in which each description enters into joint control with tacts of the various stimuli and certainly not because of any simple correspondence between the words of the descriptions and the elements of the object described.
    Let  us conclude with an observation,  Viewed through the notion of joint control, we see that specification, description, recognition and comprehension seem to be different aspects of a single process of stimulus control.  Joint control thus  provides a simple behavioral mechanism to account for many allegedly cognitive and semantic functions. And in doing so it increases the utility, and thus the plausibility, of Skinner’s account of verbal behavior.  Though research on this topic is not easy, I think the conceptual path is quite clear and should be explored.