The search for stimulus equivalence in nonverbal organisms.
These results have led to considerable recent theorizing and speculation of the role of language in the formation of equivalence classes (e.g., Cerutti & Rumbaugh, 1993: this issue; Herman, 1989; McIntire, Cleary, & Thompson, 1989; Saunders, 1989; Schusterman & Gisiner, 1989; Vaughan, 1989; Zentall & Urcuioli, 1993: this issue). Some have suggested, for example, that linguistic competence may be necessary for subjects to demonstrate the formal properties of equivalence (Dugdale & Lowe, 1990). A study by Devany, Hayes, and Nelson (1986) was particularly provocative because they reported equivalence class formation in mentally retarded children with some expressive language skills, but not in mental-age-matched retarded children who were nonverbal. However, Mackay and Sidman (1984) and Sidman (1986) have suggested that equivalence relations may underlie some aspects of language. Sidman (1990) has gone on to propose that stimulus equivalence may be a basic stimulus function that is not reducible to more elementary behavioral processes. Questions about the relation between equivalence and linguistic competence would be answered, of course, by a demonstration of equivalence class formation in nonverbal experimental subjects.
Previous Findings with Nonhumans
One study with primates (Oden et al. 1988) and one with dolphins (Pack, Herman, & Roitblat, 1991, novel transfer tests) demonstrated generalized identity matching under conditions comparable to the reflexivity tests used with human subjects. Several other studies provide supporting data, although methodological variation may complicate interpretation (D'Amato & Colombo, 1985; Herman, 1988; Herman, Hovancik, Gory, & Bradshaw, 1989; Schusterman & Gisiner, 1989; Washburn, Hopkins, & Rumbaugh, 1989; see Dube, McIlvane, & Green, 1992, and Saunders & Green, 1992 for discussions of reflexivity test procedures). In contrast, generalized identity matching has been repeatedly questioned in laboratory animals such as the rat and pigeon (e.g., Carter & Werner, 1978; D'Amato, Salmon, Loukas, & Tomie, 1986). Wright, Cook, Rivera, Sands, and Delius' (1988) demonstration of identity matching with nonconditional stimulus functions, however, suggests that limitations in experimental procedures may have contributed to previous failures.
With both pigeons and primates, negative results have been the rule in tests of symmetry and transitivity (D'Amato et al., 1985; Gray, 1966; Kendall, 1983; Lipkens, Kop, & Matthijs, 1988; Rodewald, 1974; Sidman et al., 1982; but see Zentall & Urcuioli, 1993: this issue). There are occasional reports of individual performances that may indicate symmetry or transitivity in primates (e.g., Boysen & Berntson, 1989; D'Amato et al., 1985; Premack, 1986; Savage-Rumbaugh, Rumbaugh, Smith, & Lawson, 1980; Tomonaga et al., 1991). Other reports of transitivity with pigeons (Kendall, 1983) and symmetry and transitivity in monkeys (McIntire, Cleary, & Thompson, 1987) have depended on directly conditioned chains of mediating behavior and thus did not document transitive stimulus control under conditions comparable to those required by standard equivalence tests (Hayes, 1989; Saunders, 1989).
We will contrast two interpretations of the widespread success in demonstrating equivalence with humans and the continued difficulties with nonhumans. The first interpretation is that the difference in human vs. nonhuman performance is a qualitative one reflecting fundamentally different behavioral processes. An implication of this interpretation is that equivalence requires associative neural systems or networks that develop only in the human central nervous system. One research direction suggested by a qualitative interpretation is specification of the underlying structure of such networks by biobehavioral studies incorporatinq neuroimaging or modern electrophysiological techniques (e.g., Posner, Petersen, Fox, & Raichle, 1988).
A second interpretation is a quantitative one: Failures to demonstrate equivalence in nonhumans may be related to inadequate preparation, inadequate testing, or both. One implication of a quantitative interpretation is that a positive outcome with the standard equivalence procedures involves behavioral prerequisites that humans, but not laboratory animals, are likely to have acquired through preexperimental experience. One research direction suggested by a quantitative interpretation is a behavior-analytic approach. The initial assumption of such an approach is that equivalence is behavior resulting from a specifiable history of reinforcement. The research path is to identify the behavioral prerequisites for equivalence by developing an instructional program that will teach equivalence by supplying those prerequisites. Research in human mental retardation has shown that some behavioral deficits apparently caused by cognitive limitations could be overcome by a programmed-instructional approach (Dube, Iennaco, Rocco, Kledaras, & McIlvane, 1992; McIlvane, Dube, Kledaras, Iennaco, & Stoddard, 1990; McIlvane, Kledaras, Dube, & Stoddard, 1989; Saunders & Spradlin, 1990; Sidman & Stoddard, 1966; Stoddard, 1982). In recent years, quantitative approaches, including many that are not nominally behavior analytic, have extended the range of behavior demonstrated in nonhumans to include areas such as exclusion (Schusterman, Gisiner, Grimm, & Hanggi, 1993; Tomonaga, 1993), rudimentary numerical competency (Rumbaugh, Hopkins, Washburn, & Savage-Rumbaugh, 1989), and procedural verbal behavior (e.g., Rumbaugh, 1990).
Functional Stimulus Classes and Stimulus Equivalence Classes
This section will examine briefly the relationship between equivalence classes and functional stimulus classes. In functional stimulus classes, (a) all class members share a common stimulus function, and (b) variables applied to one class member may affect the others without explicit conditioning (Galloway & Petrie, 1968; Goldiamond, 1962, 1966; Skinner, 1935). In one type of experimental procedure for studying functional stimulus classes, the stimuli in one class are discriminative for reinforcement and thus function as S+, and those in the other class are discriminative for extinction and function as S-. Because such classes are defined by the prevailing reinforcement contingencies, they have been referred to as contingency classes (Sidman, Wynne, Maguire, & Barnes, 1989).
The relation between contingency classes and classes of stimuli that are equivalent according to Sidman's criteria is relevant to the present discussion for two reasons: First, Vaughan (1988) has demonstrated contingency class formation in pigeons (see also Urcuioli & Zentall, 1993: this issue); thus, this type of behavior seems within the capabilities of nonverbal organisms. Second, recent research with human subjects suggests that contingency classes and equivalence classes may have many of the same behavioral prerequisites (Dube, McDonald, & McIlvane, 1991; Sidman et al., 1989). If so, then the early stages of a program designed to establish the prerequisites for equivalence may benefit from the inclusion of procedures that establish contingency classes.
Contingency classes in the pigeon. In Vaughan's (1988) experiment, 40 different slides of trees were assigned randomly to two sets of 20 slides each. At the beginning of the experiment, stimuli in Set 1 were designated S+ and stimuli in Set 2 were S-, and the two sets were intermixed in presentation. Order of presentation was randomized for every session, and the complete set of 40 slides was presented twice in each session. All slides were presented for variable intervals that ranged from 11 to 41 sec. Following the interval, S+ slides remained displayed until the subject pecked twice on the designated response key within 2 sec, followed by access to grain; and S- slides remained displayed until the subject did not peck the response key for 2 sec, followed by termination of the slide and the intertrial interval.
When response rates during the first 10 sec of each slide presentation were high for S+ slides and low for S- slides, the discriminative functions of all stimuli were reversed; all slides in Set 1 became S- and those in Set 2 became S+. When the reversed discrimination was established, the contingencies were reversed once again, and the original discrimination was reestablished. Repeated discrimination reversal training proceeded in this manner for a substantial period of time, with reversals ultimately programmed every fifth, sixth, or seventh session.
After many reversals, the birds' behavior began to conform to reversed contingencies within the first 40 trials of a reversal, that is, during the first presentation of the 40-slide set. For example, when the Set-2 slides presented early in a session became S+ after a recent history as S-, the birds responded to the remaining Set-2 slides on later trials with the higher response rates typical for S+, even though the most recent history with those particular slides was extinction. Similarly, when pecking in the presence of Set-1 slides on the early trials of a session no longer produced reinforcement, pecking in the presence of the remaining Set-1 slides was at the low rates typical for S-, even though the birds' most recent history with those slides was reinforcement. Exposure to changed contingencies for set members presented early in the session resulted in changes in the discriminative functions of other members presented later in the session. The slides were assigned randomly to the two sets and counterbalanced across subjects, and they could not be discriminated on the basis of common features (e.g., presence vs. absence of trees). The absence of any physical basis for classification distinguishes this work from other studies of relational control that have received considerable attention in the fields of animal learning and cognition, such as transposition, identity, oddity, abstraction, and categorical classification (e.g., Bhatt, Wasserman, Reynolds, & Knauss, 1988; Herrnstein, Loveland, & Cable, 1976; Pisacreta, Gough, Kramer, & Schultz, 1989; Wright et al., 1988). Vaughan's experiment was a convincing demonstration of functional stimulus class formation in the pigeon. It was a landmark study because such class formation had previously been demonstrated only in humans (e.g., Lazar, 1977; Mackay & Sidman, 1984; Silverman, Anderson, Marshall, & Baer, 1986).
Although Vaughan characterized his study as a demonstration of stimulus equivalence, it has not satisfied everyone as such. ln mathematics, partition into sets implies equivalence. However, as Hayes (1989) has pointed out, Vaughan's study did not conduct the requisite tests for reflexivity, symmetry, and transitivity, On the face of it, this concern could be taken only as one of definition, but it actually involves a more fundamental issue. A critical aspect of the various tests for equivalence is the emergence of performances that are not traced to a history of direct reinforcement. In the case of Vaughan's pigeons, one can reasonably argue, as Hayes did, that the contingency-class performances had been explicitly trained through the repeated reversal procedures. That is, the pigeons had been taught, via direct exposure to contingencies, to reverse the discriminative functions of later set members when responses to early ones went unreinforced.
Experimentation with human subjects has shown that membership in contingency classes is frequently--but not always--accompanied by membership in equivalence classes (de Rose, McIlvane, Dube, & Stoddard, 1988; McIlvane, Dube, Kledaras, Iennaco, & Stoddard, 1989; Sidman et al., 1989; cf. Dube et al. 1991). In Sidman and colleagues' study, for example, contingency classes were established by the repeated reversals technique, and then tests for stimulus equivalence were conducted. Equivalence was documented for two of three experimental subjects; for the third subject, contingency class membership was not accompanied by equivalence class membership. Such demonstrations help to put Vaughan's (1988) results in perspective. If contingency class membership can, under some circumstances, be independent of equivalence class membership, then the former cannot be accepted as a demonstration of the latter. On balance, because contingency class membership is frequently accompanied by equivalence class membership, the two forms of class membership may have common behavioral prerequisites. If so, then the chances for documenting equivalence classes may increase in a context where contingency classes have already been established.
Preliminary Studies of Stimulus Classes in Rats
The final section of this paper will describe some preliminary studies that are part of an ongoing attempt to develop programmed equivalence methods for laboratory animals. This project is exploratory and on a small scale, and only a few rats have been studied. Thus, this report is one of work in the early stages of progress, and it is described here as an example of a quantitative approach to asking about the potential for equivalence in a nonverbal organism. The initial goals of the project are to develop reliable and efficient methods for establishing concurrent discriminations and reversals and to systematically replicate Vaughan's findings in another species. The long-term goals are to develop methods for testing equivalence according to Sidman's criteria in a context of verified contingency classes. Even if the ultimate goal of documenting equivalence is not achieved, however, there appear to be advantages to a quantitative approach. One such benefit is that the systematic identification and evaluation of potential behavioral prerequisites is likely to contribute to a more complete specification of qualitative differences. Another potential benefit is the development of new methodologies. Reliable, efficient procedures for producing stable functional stimulus classes will contribute behavioral baselines generally useful in the study of variables that may affect abstract relational processes (e.g., in neuropsychology, psychopharmacology, and animal cognition).
Successive vs. simultaneous contingency-class procedures. Vaughan's study was conducted with successive discrimination procedures; the contingency-class studies with human subjects described above used simultaneous procedures (every trial presented S+ and S- stimuli simultaneously in different locations). There appear to be several advantages in successive procedures for research with laboratory animals (cf. Wasserman, 1976): Vaughan's experimental procedures are available as a basis for systematic replications; successive procedures are appropriate for both visual and auditory stimuli; and successive procedures may require less elaborate behavior by the subject because they present only one stimulus per trial.
There is another, perhaps less obvious, potential disadvantage in simultaneous procedures. Because stimuli from both contingency classes are present on every trial, the simultaneous procedure opens the door to the possibility of different stimulus control topographies on different trials (see McIlvane & Dube, 1992). That is, responses that meet the prevailing reinforcement contingencies may sometimes involve selection of the nominal S+ stimulus and at other times rejection of the nominal S-, analogous to sample/S- control in matching to sample (Carrigan & Sidman, 1992). With two-choice simultaneous procedures, discrimination reversals may even encourage control by S- stimuli; accurate reversal performance need not require a shift in the controlling stimulus but merely a shift in response topography (Carrigan & Sidman, 1992). For example, the experimental context may be one where a particular stimulus sometimes controls touching the key where it is displayed, and other times (in reversals) controls rejecting the key where it is displayed and touching a different key displaying any other stimulus. The single-stimulus presentation in successive procedures seems preferable because it restricts the range of potential stimulus control topographies.
Auditory discrimination procedures. The apparatus is a standard operant conditioning chamber controlled by a Macintosh computer with software and an interface developed in our laboratory. A reinforcer magazine centered on one end wall dispenses 45-mg food pellets or 1-ml drops of 12% (by weight) sucrose solution. The chamber is equipped with two response levers: Lever A is centered on the wall opposite the reinforcer magazine, and Lever B is next to the magazine on the same wall.
Auditory stimuli were selected for study because of rats' capacity for difficult and complex auditory discriminations (e.g., D'Amato & Salmon, 1982, 1984). Auditory stimuli are presented through a speaker mounted in a metal funnel; the small end of the funnel is mounted in a hole drilled through the housing of Lever B. Thus, the sound source is at the response location, an arrangement that has been shown to facilitate the acquisition of auditory discriminations in rats (Harrison, 1988, 1992; Neill & Harrison, 1987).
After experimentation with several alternatives (for details see Dube, Callahan, & McIlvane, 1993), the following trial procedure was adopted: Pressing Lever A produced, on a variable ratio schedule (VR2), an auditory stimulus presented through the speaker at Lever B. If the stimulus was S+, then it was presented continuously until Lever B was pressed and a reinforcer followed. If the stimulus was S-, then it was presented continuously for 5 sec regardless of the subject's behavior and no reinforcer followed. For all stimuli, a response was defined as pressing Lever B within 5 sec of stimulus onset.
This two-lever procedure was established by hand shaping and backward chaining in 2 or 3 pretraining sessions that presented S+ only. When discrimination training began and S- stimuli were introduced, subjects typically pressed Lever B in the presence of both S+ and S-. Discrimination of S+ from S- was demonstrated when subjects pressed Lever B within 5 sec of S+ presentation and did not press Lever B for the 5-sec duration of S- (when S- was presented, they typically continued to press Lever A).
Repeated reversals of concurrent discriminations. In initial attempts to systematically replicate Vaughan's procedure, rats were given repeated reversals of concurrent discrimination among six tones, three S+ and three S-. The goal was to determine whether, with smaller stimulus sets than Vaughan's, the rate of learning successive reversals would improve to the point where tests for contingency classes would be feasible. If new reversals were acquired rapidly, then the tests could be conducted by omitting some stimuli, both S+ and S-, from the initial trials of the reversal. Then, when response rates for the presented stimuli had changed in accordance with the reversed contingencies, the omitted stimuli could be reintroduced: Would initial responding to these stimuli be in line with their most recent prereversal functions, or would responding be consistent with the reversal? The latter outcome would be evidence for contingency classes.
Our initial efforts to develop the necessary baselines of repeated reversals of concurrent discriminations have been described elsewhere and so here will be summarized only (Dube et al., 1993). The stimuli were six tones produced by the Macintosh ROM square-wave tone generator. Slightly different sets of stimuli were used with different subjects. For subjects whose contingency-class test data will be described below, the software activated the tone generator with frequency values of 500 (A1), 970 (A2), 1730 (B1), 3180 (B2), 5980 (C1), and 11100 (C2), and amplitude values were adjusted to produce the tones at 75 dB at Lever B. The procedures were designed to establish two three-member contingency classes consisting of A1-B1-C1 (Set 1) and A2-B2-C2 (Set 2). To eliminate confounds attributable to primary stimulus generalization, potential class members were tones with nonadjacent frequencies; the frequency of Stimulus A2, for example, was between that of A1 and B1.
A differential outcomes procedure was used throughout training (Goeters, Blakely, & Poling, 1992; Peterson, 1984; Trapold, 1970). When Set-1 stimuli were S+, responses produced sucrose/water solution; when Set-2 stimuli were S+, responses produced a food pellet. Differential outcomes procedures have been shown to enhance conditional-discrimination learning in rats; the procedure was used for these initial studies to increase the chances of success in training a difficult discrimination.
As reported in Dube and colleagues (1993), five rats were given concurrent discrimination training with six tones, and all five acquired the initial discrimination. When the reinforcement contingencies were reversed for all stimuli, four of the five rats acquired the reversed discriminations. With exposure to additional contingency reversals, two animals showed savings; they met the acquisition criterion more quickly on later reversal problems. Only one of these five rats eventually acquired reversals rapidly enough to permit contingency class tests; by the eighth reversal, it typically met the learning criterion for successive reversals within only three or four sessions (Subject R5 in Dube et al., 1993).
The results of initial contingency class tests are shown in Figure 1. Panel A shows the data for the 12th reversal of the full six-stimulus baseline; Set 1 is positive and Set 2 is negative. For the first test (Panel B), Stimuli B1 and B2 were removed from the baseline and the discrimination reversal proceeded with A2 and C2 positive and A1 and C1 negative. The two points in the center row of Panel B show response rates for the first 15 trials with B1 and B2 when they were reintroduced in the second half of the fourth session. The results were consistent with contingency classes. The rat pressed the lever on 100% of the trials with B2 and on 13% (2/15) of the trials with B1; following exposure to reversed contingencies for the A and C stimuli, the discriminative functions of the B stimuli were also reversed. Next, the B stimuli again were removed from the baseline and contingencies were reversed for the A and C stimuli (Panel C). When the B stimuli were reintroduced, their functions had once again reversed; response rates were comparable to those in Panel A. Similar results were obtained in subsequent tests with the C stimuli (Panels D and E) and, to a lesser extent, the A stimuli (Panels F and G). When test stimuli were presented, the rat behaved as though it had been exposed to reversal training with those stimuli, and not as it had in its most recent experience with them. This subject went on to complete nine additional tests like those in Figure 1 with similar results.
Repeated reversals of concurrent discriminations: Two-tone procedure. Two other rats were given a variation of the six-tone procedures designed to speed acquisition rates by reducing the number of stimuli presented in each block of trials. The stimuli and trial procedures were the same as those described above. Instead of distributing trials with all six stimuli evenly throughout the session, three two-tone discriminations were presented successively in separate blocks of trials in each session. For example, the first session following preliminary training consisted of 100 trials of A1/C2 (S+/S-), followed by 100 trials of B1/A2, and finally 100 trials of C1/B2. In the immediately following sessions, the stimulus pairings remained the same (A1/C2, B1/A2, C1/B2), but the order of the 100-trial blocks varied. When the three two-tone discriminations were acquired, the discriminative functions of all stimuli were reversed and new S+/S- pairs were assigned, A2/C1, B2/A1, and C2/B1, and training continued.
With the two-tone procedure, the first day of each new reversal provided an opportunity for performance consistent with contingency class formation. If the repeated-reversals procedure established functional stimulus classes, then a change in the function of one or more class members would be sufficient for a change in the functions of remaining class member(s). That is, after encountering reversed contingencies for the first pair (or first two pairs) of stimuli in the session, the subject may respond immediately to the remaining pair(s) of stimuli in a manner consistent with a reversal.
Two subjects were given the two-tone procedures. One required 23 sessions (300 trials each) to acquire the initial discrimination, and we discontinued the procedure after 69 sessions of the first reversal without progress. The second subject, however, acquired the initial discriminations in 17 sessions and showed substantial savings by the fourth reversal, which was accomplished in 7 sessions.
Evidence consistent with contingency classes was obtained in the first session of the fifth reversal. In the first two reversed discriminations, A2/A1 and C2/C1, the subject responded to the new S- stimuli on a majority of trials. In the third discrimination, however, responding was immediately and almost completely consistent with the reversed contingencies: In the first 40 trials of the B2/B1 discrimination, there was only one error, a single response to B1 on its eighth presentation.
As training continued, the speed of reversal learning continued to increase, and within-session improvement on the first day of a reversal became more common. To help separate the effects of within-session learning attributable to differential reinforcement contingencies from those that may be attributable to contingency class formation, Figure 2 shows accuracy scores for the first 10 trials of each discrimination (i.e., Trials 1-10, 101-110, and 201-210) on the first day of a reversal. For the original discrimination and first four reversals, all performances are at chance accuracy levels. The first performance consistent with contingency classes is shown in the fifth reversal. As training continued, a similar pattern emerged in Reversals 9-12: Chance accuracy for the first one or two reversed discriminations, then immediately high accuracy on the initial trials of the discrimination(s) that followed. For example, on the first day of the tenth reversal, accuracy was 60% for the first 10 trials of C1/B2 at the beginning of the session, then 90% for the first 10 trials of B1/A2, and finally 100% for the first 10 trials of A1/C2.
Tests for conditional control by differential outcomes reinforcers. One interpretation of the data shown in Figures 1 and 2 is that contact with reversed contingencies for stimuli presented earlier in the session resulted in a change in the discriminative functions of the stimuli presented later--an indication of contingency class formation. A possible alternative explanation concerns the differential outcomes procedures. Because responses to Set-1 S+ stimuli produced sugar/water and responses to Set-2 S+ stimuli produced food pellets, the discriminative functions of the tones may have become conditional upon the reinforcer recently presented--a type of successive conditional discrimination. That is, if sugar/water, then A1, B1, and C1 are positive and A2, B2, and C2 are negative; if pellet, then A1, B1, and C1 are negative and A2, B2, and C2 are positive.
As a test for conditional control by the reinforcing stimuli, the "negative-class" reinforcer was presented on some S+ trials; that is, the reinforcer was the one previously presented only when the current S-class was positive. If reinforcers exerted conditional control over lever pressing, then the subject should be highly likely to respond on an S-trial that immediately followed presentation of a negative-class reinforcer (i.e., if sugar/water, then A1 is positive, etc.).
For the subject that received the six-tone procedure, training with the full six-tone baseline continued without a reversal until performance was accurate and stable. Then, negative-class reinforcer probes were presented on 20% of the S+ trials during the second halves of several sessions, 4 sessions when Set-1 stimuli were positive and 5 sessions when Set-2 stimuli were positive. When Set 2 was positive, presenting the negative-class reinforcer (sugar/water) had little effect: Prior to the probes (i.e., in the first halves of the probe sessions), the rat responded to all S+ stimuli and to 40% of those S- stimuli that immediately followed reinforcer presentations. When probes were introduced, it continued to respond on all S+ trials, on 43% of the S- trials following presentation of the positive-class reinforcer, and on 47% of the S- trials following the negative-class reinforcer. In contrast, when Set-1 stimuli were positive, presenting a negative-class reinforcer (a pellet) had a large effect: Prior to the probes, the rat responded to all S+ stimuli and to 33% of the S- stimuli that followed reinforcers. When probes were introduced, it continued to respond on all S+ trials but also responded on 84% and 96% of the S- trials following positive- and negative-class reinforcers, respectively. Thus, there was little change in performance when sugar/water replaced pellets, but discrimination was disrupted when pellets replaced sugar/water.
Similar tests were conducted with the subject that received the two-tone procedure. A total of 14 reinforcer probe sessions were conducted over six reversals. The subject responded on virtually all S+ trials. In the blocks of trials immediately prior to the probes, responses occurred on 20% of the S- trials immediately following reinforcers. Within probe blocks, responding on the S- trials that followed negative-class reinforcers was more likely (37%) than on the S- trials that followed positive-class reinforcers (24%). These results are consistent with some conditional control by the reinforcer, although discrimination accuracy remained fairly high.
The changes in performance following reinforcer probes suggests that the outcome-specific reinforcers exerted conditional control of tone discrimination in some instances. The limited degree of change in each case, however, suggests that control by reinforcers was shared with other variables--variables with effects that are consistent with membership in contingency classes. These data invite further research on the role of outcome-specific reinforcers as potential stimulus class members. As we have argued elsewhere, conditional control by stimulus-stimulus relations involving discriminative and reinforcing stimuli may also indicate stimulus class formation (Dube, McIlvane, Mackay, & Stoddard, 1987; Dube, McIlvane, Maguire, Mackay, & Stoddard, 1989; see also Zentall & Urcuioli, 1993: this issue).
Directions for continued experimentation. As may be expected in a new area of investigation, our preliminary studies indicate the need for procedural refinements and further methodological development before definitive experiments in contingency class formation are conducted. Although only two of seven rats in these initial studies progressed to the point of contingency class tests, the data for those subjects are consistent with Vaughan's findings with pigeons: mastery of repeated reversals, savings across reversals, and apparent emergence of discriminations that are not consistent with the subject's most recent conditioning history. Ongoing studies in our laboratory are aimed at improving the effectiveness and efficiency of the contingency-class training procedures. Preliminary results suggest (a) more rapid concurrent discrimination learning when there is greater variety in the auditory stimuli (digitized music or human speech, white noise, clicks, etc.), and (b) that the outcome-specific contingencies are not necessary for concurrent discrimination learning.
If continued experimentation successfully demonstrates contingency class formation, the methods may provide a context in which we can go on to ask whether rats are capable of emergent behavior consistent with stimulus equivalence. The next step will be to establish a baseline of conditional discrimination among contingency-class members. Successive conditional discrimination procedures seem best suited for auditory stimuli. Such procedures are go/nogo analogs of matching to sample (variants of Konorski, 1959; e.g., Cohen, Grassi, & Dowson, 1988; Wasserman, 1976; Wilkie & Wilson, 1977). On each trial, pressing Lever A produces "sample" and "comparison" stimuli successively. On trials where both stimuli are from the same contingency class (e.g., A1 followed by B1), pressing Lever B is designated the correct response and followed by reinforcers. On trials where the stimuli are from different classes (e.g., A1 followed by B2), no reinforcers are available (subjects that acquire the discrimination are likely to continue pressing Lever A, as in the present studies). Conditional discrimination is demonstrated by high rates of responding to Lever B on same-class trials and low rates on different-class trials. The stimulus equivalence paradigm could be implemented by training a subset of the possible conditional discriminations among contingency class members (e.g., AA, AB, and BC) and then evaluating performance on the remaining discriminations (BB, CC, BA, CA, etc.). If a history of emergent conditional discrimination among contingency class members can be established, the final stage of an equivalence demonstration will be to relate new stimuli to some members of existing classes and then to test for emergent relations between the new stimuli and the remaining class members (Sidman et al., 1989).
Clearly, such a course of investigation is ambitious, and there will be more procedural obstacles to overcome. As we noted before, the development of new methods is one potential advantage of a quantitative approach to the problem. We expect that the answers to continuing questions about the relation between verbal repertoires and stimulus equivalence will be resolved only with the continued accumulation of data from experimentation with nonhumans and nonverbal humans.
BHATT, R. S., WASSERMAN, E. A., REYNOLDS, W. F., & KNAUSS K. S. (1988). Conceptual behavior in pigeons: Categorization of both familiar and novel examples from four classes of natural and artificial stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 14, 219-234.
BOYSEN, S. T., & BERNTSON, G. G. (1989). Numerical competence in a chimpanzee (Pan troglodytes). Journal of Comparative and Physiological Psychology, 103, 23-31.
CARRIGAN, P. F., Jr., & SIDMAN, M. (1992). Conditional discrimination and equivalence relations: A theoretical analysis of control by negative stimuli. Journal of the Experimental Analysis of Behavior, 58, 183-204.
CARTER, D. E., & WERNER, T. J. (1978). Complex learning and information processing by pigeons: A critical analysis. Journal of the Experimental Analysis of Behavior, 29, 565-601.
CERUTTI, D. T., & RUMBAUGH, D. M. (1993: this issue). Stimulus relations in comparative primate perspective. The Psychological Record, 43, 811-822.
COHEN, J. S., GRASSI, J., & DOWSON, P. (1988). The role of within-trial location of the retention interval in rats' delayed conditional discrimination performance. Animal Learning and Behavior, 16, 40-66.
D'AMATO, M. R., & COLOMBO, M. (1985). Auditory matching-to-sample in monkeys (Cebus apella). Animal Learning and Behavior, 13, 375-382.
D'AMATO, M. R., & SALMON, D. (1982). Tune discrimination in monkeys (Cebus apella) and rats. Animal Learning and Behavior, 10, 126-134.
D'AMATO, M. R., & SALMON, D. (1984). Processing of complex auditory stimuli (tunes) by rats and monkeys (Cebus apella). Animal Learning and Behavior, 12, 184-194.
D'AMATO, M. R., SALMON, D., LOUKAS, E., & TOMIE, A. (1985). Symmetry and transitivity of conditional relations in monkeys (Cebus apella) and pigeons (Columba livia). Journal of the Experimental Analysis of Behavior, 44, 35-47.
D'AMATO, M. R., SALMON, D. P., LOUKAS, E., & TOMIE, A. (1986). Processing of identity and conditional relations in monkeys (Cebus apella) and pigeons (Columba livia). Animal Learning and Behavior, 14, 365-373.
DE ROSE, J. C., MCILVANE, W. J., DUBE, W. V., & STODDARD, L. T. (1988). Stimulus class formation and functional equivalence in moderately retarded individuals' conditional discrimination. Behavioural Processes, 17, 167-175.
DEVANY, J. M., HAYES, S. C., & NELSON, R. O. (1986). Equivalence class formation in language-able and language-disabled children. Journal of the Experimental Analysis of Behavior, 46, 243-257.
DUBE, W. V., CALLAHAN, T. D., & MCILVANE, W. J. (1993). Serial reversals of concurrent auditory discriminations in rats. The Psychological Record, 43, 429-440.
DUBE, W. V., IENNACO, F. M., ROCCO, F. J., KLEDARAS, J. B., & MCILVANE, W. J. (1992). Microcomputer-based programmed instruction in identity matching to sample for persons with severe disabilities. Journal of Behavioral Education, 2, 29-51.
DUBE, W. V., MCDONALD, S. J., & MCILVANE, W. J. (1991). A note on the relationship between equivalence classes and functional stimulus classes. Experimental Analysis of Human Behavior Bulletin, 9, 7-11.
DUBE, W. V., MCILVANE, W. J., & GREEN, G. (1992). An analysis of generalized identity matching-to-sample test procedures. The Psychological Record, 42, 17-28.
DUBE, W. V, MCILVANE, W. J., MACKAY, H. A., & STODDARD, L. T. (1987). Stimulus class membership established via stimulus-reinforcer relations. Journal of the Experimental Analysis of Behavior, 47, 159-175.
DUBE, W. V., MCILVANE, W. J., MAGUIRE, R. A., MACKAY, H. A., & STODDARD, L. T. (1989). Stimulus class formation and stimulus-reinforcer relations. Journal of the Experimental Analysis of Behavior, 51, 65-76.
DUGDALE, N., & LOWE, C. F. (1990). Naming and stimulus equivalence. In D. E. Blackman & H. Lejeune (Eds.), Behaviour analysis in theory and practice: Contributions and controversies. (pp. 115-138). Hillsdale, NJ: Erlbaum.
GALLOWAY, C., & PETRIE, R. D. (1968). An analysis of the functional equivalence of stimulus class members. Journal of Experimental Child Psychology, 6, 384-390.
GOETERS, S., BLAKELY, E., & POLING, A. (1992). The differential outcomes effect. The Psychological Record, 42, 389-411.
GOLDIAMOND, I. (1962). Perception. In A. J. Bachrach (Ed.), Experimental foundations of clinical psychology (pp. 280-340). New York: Basic Books.
GOLDIAMOND, I. (1966). Perception, language, and conceptualization rules. In B. Kleinmuntz (Ed.), Problem solving (pp. 183-224). New York: Wiley.
GRAY, L. (1966). Backward associations in the pigeon. Psychonomic Science, 4, 333-334.
HARRISON, J. M. (1988). Control of responding by sounds of different quality: An evolutionary analysis. Journal of the Experimental Analysis of Behavior, 50, 521-539.
HARRISON, J. M. (1992). Avoiding conflicts between the natural behavior of the animal and the demands of discrimination experiments. Journal of the Acoustical Society of America, 92, 1331-1345.
HAYES, S. C. (1989). Nonhumans have not yet shown stimulus equivalence. Journal of the Experimental Analysis of Behavior, 51, 385-392.
HERMAN, L. M. (1988). The language of animal language research. The Psychological Record, 38, 349-382.
HERMAN, L. M. (1989). In which Procrustean bed does the sea lion sleep tonight? The Psychological Record, 39, 19-50.
HERMAN, L. M., HOVANCIK, J. R., GORY, J. D., & BRADSHAW, G. L. (1989). Generalization of visual matching by a bottlenosed dolphin (Tursiops truncatus): Evidence for invariance of cognitive performance with visual and auditory materials. Journal of Experimental Psychology: Animal Behavior Processes, 15, 124-136.
HERRNSTEIN, R. J., LOVELAND, D. H., & CABLE, C. (1976). Natural concepts in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 2, 285-302.
KENDALL, S. B. (1983). Tests for mediated transfer in pigeons. The Psychological Record, 33, 245-256.
KONORSKI, J. (1959). A new method of psychological investigation of recent memory in animals. Bulletin de L'Academie Polonaise des Sciences, 7, 115-119.
LAZAR, R. (1977). Extending sequence-class membership with matching to sample. Journal of the Experimental Analysis of Behavior, 27, 381-392.
LIPKENS, R., KOP, P. F. M., & MATTHIJS, W. (1988). A test for symmetry and transitivity in the conditional discrimination performances of pigeons. Journal of the Experimental Analysis of Behavior, 49, 395-409.
MACKAY, H. A., & SIDMAN, M. (1984). Teaching new behavior via equivalence relations. In P. H. Brooks, R. Sperber, & C. MacCauley (Eds.), Learning and cognition in the mentally retarded. Hillsdale, NJ: Erlbaum.
MCILVANE, W. J., & DUBE, W. V. (1992). Stimulus control shaping and stimulus control topographies. The Behavior Analyst, 15, 89-94.
MCILVANE, W. J., DUBE, W. V., KLEDARAS, J. B., IENNACO, F. M., & STODDARD, L. T. (1989). Stimulus-consequence relations and stimulus classes in mental retardation. Experimental Analysis of Human Behavior Bulletin, 7, 17-20.
MCILVANE, W. J., DUBE, W. V., KLEDARAS, J. B., IENNACO, F. M., & STODDARD, L. T. (1990). Teaching relational discrimination to individuals with mental retardation: Some problems and possible solutions. American Journal on Mental Retardation, 95, 283-296.
MCILVANE, W. J., KLEDARAS, J. B., DUBE, W. V. & STODDARD, L. T. (1989). Automated instruction of severely and profoundly retarded individuals. In J. Mulick & R. Antonak (Eds.), Transitions in mental retardation, Vol. 4. Norwood, NJ: Ablex.
MCINTIRE, K. D., CLEARY, J., & THOMPSON, T. (1987). Conditional relations by monkeys: Reflexivity, symmetry, and transitivity. Journal of the Experimental Analysis of Behavior, 47, 279-285.
MCINTIRE, K. D., CLEARY, J., & THOMPSON, T. (1989). Reply to Saunders and Hayes. Journal of the Experimental Analysis of Behavior, 51, 393-396.
NEILL, J. C., & HARRISON, J. M. (1987). Auditory discrimination: Konorski's quality-location effect. Journal of the Experimental Analysis of Behavior, 48, 81-95.
ODEN, D. L., THOMPSON, R. K. R., & PREMACK, D. (1988). Spontaneous transfer of matching by infant chimpanzees. Journal of Experimental Psychology: Animal Behavior Processes, 14, 140-145.
PACK, A. A., HERMAN, L. M. & ROITBLAT, H. L. (1991). Generalization of visual matching and delayed matching by a California sea lion (Zalophus californianus). Animal Learning and Behavior, 19, 37-48.
PETERSON, G. B. (1984). How expectancies guide behavior. In H. L. Roitblat, T. G. Bever, & H. S. Terrace (Eds.), Animal cognition (pp. 135-148). Hillsdale, NJ: Erlbaum.
PlSACRETA, R., GOUGH, D., KRAMER, J., & SCHULTZ, W. (1989). Some factors that influence transfer of oddity performance in the pigeon. The Psychological Record, 39, 221-246.
POSNER, M. I., PETERSEN, S. E., FOX, P. T., & RAICHLE, M. E. (1988). Localization of cognitive operations in the human brain. Science, 240, 1627-1631.
PREMACK, D. (1986). "Gavagai!" or the future history of the animal language controversy. Cambridge, MA: MIT Press.
RODEWALD, H. K. (1974). Symbolic matching-to-sample by pigeons. Psychological Reports, 34, 987-990.
RUMBAUGH, D. M. (1990). Comparative psychology and the great apes: Their competence in learning, language, and numbers. The Psychological Record, 40, 15-39.
RUMBAUGH, D. M., HOPKINS, W. D., WASHBURN, D. A., & SAVAGE-RUMBAUGH, E. S. (1989). Lana chimpanzee learns to count by "Numath": A summary of a videotaped experimental report. The Psychological Record, 39, 459-470.
SAUNDERS, K. J. (1989). Naming in conditional discrimination and stimulus equivalence. Journal of the Experimental Analysis of Behavior, 51, 379-384.
SAUNDERS, R. R., & GREEN, G. (1992). The nonequivalence of behavioral and mathematical equivalence. Journal of the Experimental Analysis of Behavior, 57, 227-241.
SAUNDERS, K. J., & SPRADLIN, J. E. (1990). Conditional discrimination in mentally retarded adults: The development of generalized skills. Journal of the Experimental Analysis of Behavior, 54, 239-250.
SAVAGE-RUMBAUGH, E. S., RUMBAUGH, D. M, SMITH, S. T., & LAWSON, J. (1980). Reference: The linguistic essential. Science, 210, 922-925.
SCHUSTERMAN, R. J., & GISINER, R. C. (1989). Please parse the sentence: Animal cognition in the procrustean bed of linguistics. The Psychological Record, 39, 3-18.
SCHUSTERMAN, R. J., GISINER, R., GRIMM, B. K., & HANGGI, E.B. (1993). Behavior control by exclusion and attempts at establishing semanticity in marine mammals using match-to-sample paradigms, In H. Roitblat, L. Herman, & P. Nachtigall (Eds.), Language and communication: Comparative perspectives (pp. 249-274). Hillsdale, NJ: Erlbaum.
SIDMAN, M. (1986). The measurement of behavioral development. In N. A. Krasnegor, D. B. Gray, & T. Thompson (Eds.), Advances in behavioral pharmacology, Vol. 5: Developmental behavioral pharmacology. Hillsdale, NJ: Erlbaum.
SIDMAN, M. (1990). Equivalence relations: Where do they come from? In D. E. Blackman & H. Lejeune (Eds.), Behaviour analysis in theory and practice: Contributions and controversies. Hillsdale, NJ: Erlbaum.
SIDMAN, M., RAUZIN, R., LAZAR, R., CUNNINGHAM, S., TAILBY, W., & CARRIGAN, P. (1982). A search for symmetry in the conditional discrimination of rhesus monkeys, baboons, and children. Journal of the Experimental Analysis of Behavior, 37, 23-44.
SIDMAN, M., & STODDARD, L. T. (1966). Programming perception and learning for retarded children. In N. R. Ellis (Ed.), International Review of Research in Mental Retardation, Vol. 2 (pp. 151-208). New York: Academic Press.
SIDMAN, M., & TAILBY, W. (1982). Conditional discrimination vs. matching-to-sample: An expansion of the testing paradigm. Journal of the Experimental Analysis of Behavior, 37, 5-22.
SIDMAN, M., WYNNE, C. K., MAGUIRE, R. W., & BARNES, T. (1989). Functional classes and equivalence relations. Journal of the Experimental Analysis of Behavior, 52, 261-274.
SILVERMAN, K., ANDERSON, S. R., MARSHALL, A. M., & BAER, D. M. (1986). Establishing and generalizing audience control of new language repertoires. Analysis and Intervention in Developmental Disabilities, 6, 21-40.
SKINNER, B. F. (1935). The generic nature of the concepts of stimulus and response. Journal of General Psychology, 12, 40-65.
STODDARD, L. T. (1982). An investigation of automated methods for teaching severely retarded individuals. In N. R. Ellis (Ed.), International review of research in mental retardation. New York: Academic Press.
TOMONAGA, M. (1993). Tests for control by exclusion and negative stimulus relations of arbitrary matching to sample in a "symmetry-emergent" chimpanzee. Journal of the Experimental Analysis of Behavior, 59, 215-229.
TOMONAGA, M., MATZUSAWA, T., FUJITA, K., & YAMAMOTO, J. (1991). Emergence of symmetry in a visual conditional discrimination by chimpanzees (Pan troglodytes). Psychological Reports, 68, 51-60.
TRAPOLD, M. (1970). Are expectancies based upon different positive reinforcing events discriminably different? Learning and Motivation, 1, 129-140.
URCUIOLI, P. J., & ZENTALL, T. R. (1993: this issue). A test of comparison-stimulus substitutability following one-to-many matching by pigeons. The Psychological Record, 43, 745-760.
VAUGHAN, W. (1988). Formation of equivalence sets in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 14, 36-42.
VAUGHAN, W. (1989). Reply to Hayes. Journal of the Experimental Analysis of Behavior, 51, 397.
WASHBURN, D. A., HOPKINS, W. D., & RUMBAUGH, D. M. (1989). Video-task assessment of learning and memory in macaques (Macaca mulatta): Effects of stimulus movement on performance. Journal of Experimental Psychology: Animal Behavior Processes, 15, 393-400.
WASSERMAN, E. A. (1976). Successive matching-to-sample in the pigeon: Variations on a theme by Konorski. Behavior Research Methods and Instrumentation, 8, 278-282.
WILKIE, D. M., & WILSON, C. S. (1977). Control of pigeons' pecking by trace stimuli. Journal of the Experimental Analysis of Behavior, 27, 293-299.
WRIGHT, A. A., COOK, R. G., RIVERA, J. J. SANDS, S. F., & DELIUS, J. D. (1988). Concept learning by pigeons: Matching-to-sample with trial-unique video picture stimuli. Animal Learning and Behavior, 16, 436-444.
ZENTALL, T. R., & URCUIOLI, P. J. (1993: this issue). Emergent relations in the formation of stimulus classes by pigeons. The Psychological Record, 43, 795-810.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Special Issue: Stimulus Equivalence|
|Author:||Dube, William V.; McIlvane, William J.; Callahan, Thomas D.; Stoddard, Lawrence T.|
|Publication:||The Psychological Record|
|Date:||Sep 22, 1993|
|Previous Article:||A test of comparison-stimulus substitutability following one-to-many matching by pigeons.|
|Next Article:||Similarity- and nonsimilarity-based conceptualization in children and pigeons.|