Equivalence classes were established using a stimulus- pairing/yes-no procedure. On all trials, two stimuli were presented with the second introduced 250 ms ...
The Psychological Record, 1997, 47, 661-686
EQUIVALENCE CLASS FORMATION USING STIMULUS-PAIRING AND YES-NO RESPONDING LANNY FIELDS, KENNETH F. REEVE, ANTONIOS VARELAS, DEVORAH ROSEN, and JAMES BELANICH Queens College/CUNY, and The Graduate School and University Center/CUNY
Equivalence classes were established using a stimuluspairing/yes-no procedure. On all trials, two stimuli were presented with the second introduced 250 ms after termination of the first. Pressing one of two keys labeled YES or NO was required during the second stimulus. AB, BC, and CO relations were established by reinforcing the YES response on within-set trials (A 1-B 1, A2-B2, B1-C1, B2-C2, C1-01, and C2-02), and the NO response on between-set trials (A 1-B2, A2-B1, B1-C2, B2-C1, C1-02, and C201). For 10 of 18 college students, the YES response occurred on all within-set emergent relations probes for symmetry (B1-A 1, C1B1, 01-C1, B2-A2, C2-B2, and 02-C2), transitivity (A1-C1, B1-01, A1-01, A2-C2, B2-02, and A2-02), and equivalence (C1-A1, 01B1, 01-A 1, C2-A2, 02-B2, and 02-A2), and the NO response occurred on all between-set probes (B1-A2, C1-B2, 01-C2, B2-A 1, C2-B1, 02-C1, A 1-C2, B1-02, A 1-02, A2-C1, B2-01, A2-01, C2A 1, 02-B1, 02-A 1, C1-A2, 01-B2, and 01-B2). Thus, a stimuluspairing/yes-no procedure established the prerequisite conditional relations of equivalence classes, assessed symmetry, transitivity, and their combined effects, and documented the emergence of equivalence classes. After class formation, emergent relations probes presented in a matching-to-sample format occasioned class-consistent responding, showing the maintenance of the classes across testing formats and contingencies. Because stimulus pairing/yes-no and matching-to-sample formats differ, matching procedures are not needed to assess the emergence of equivalence classes or form their prerequisites.
In a set of N stimuli that are perceptually disparate, there are N 2 ordered pairs (Fields & Verhave, 1987; Saunders & Green, 1992; Sidman, 1994). After directly training (N-1) of these stimulus pairs (called baseline relations), a subject may respond in a class-consistent manner to the remaining (N 2 - N + 1) untrained stimulus pairs (called emergent relations probes) even though the performances do not produce reinforcement. These probe performances, which are indicative of reflexivity, symmetry, transitivity, and the combined effects of symmetry and transitivity, demonstrate that the stimuli within a set are functioning as an equivalence This research was supported partly by ARI Contract MDA903-90-C-0132, RACO Contract DASW01-96-X-0009, and PSC/CUNY Grants 661287 and 668457, and by fellowship to Devorah Rosen from NYS/OMRDD. Reprints can be obtained from Lanny Fields, Queens College/CUNY, 65-30 Kissena Boulevard, Flushing, NY 11367.
662
FIELDS ET AL.
class (Fields, Adams, Newman. & Verhave, 1992; Fields & Verhave, 1987; Fields, Verhave, & Fath, 1984; Sidman & Tail by, 1982). This definition of equivalence, however, does not specify the use of a particular procedure for the training of the baseline relations or for the assessment of responding occasioned by the emergent relations probes. In most studies of equivalence class formation, the baseline conditional relations that are the prerequisites for equivalence classes have been established using arbitrary matching-to-sample procedures (e.g., Bush, Sidman, & de Rose, 1989; Devany, Hayes, & Nelson, 1986; Fields, Adams, & Verhave, 1993; Fields, Adams, Verhave, & Newman, 1990, 1993; Fields, Reeve, Adams, & Verhave, 1991; Lazar, Davis-Lang, & Sanchez, 1984; McDonagh, Mcllvane, & Stoddard, 1984; Saunders, Saunders, Kirby, & Spradlin, 1988; Schusterman & Kastak, 1993; Sidman, 1971; Sidman, Kirk, & Wilson-Morris, 1985; Sidman & Tailby, 1982; Watt, Keenan, Barnes, & Cairns, 1991). In these experiments, a sample was presented and remained on for the duration of a trial. When subjects emitted an observing response in the presence of a sample, it produced the concurrent presentation of all comparisons. Subjects then selected a comparison by touching it or by pressing a key that was displaced from, but corresponded to, the position of a given comparison. The conditional relations that are the prerequisites of equivalence classes have also been established using variations of the abovementioned matching-to-sample procedure. In one variation, the presentation of the sample was temporally separated from the presentation of the comparison stimuli. The comparisons were then presented on a concurrent basis (Bentall, Dickins, & Fox, 1993; Dickins, Bentall, & Smith, 1993; Stromer & Mackay, 1992, 1993a; Stromer, Mcllvane, Dube, & Mackay, 1993). In another variation, Dube, Green, and Serna (1993) established equivalence classes consisting of auditory stimuli alone by use of a novel successive matching-to-sample procedure in which only one stimulus was presented at a time. Each trial consisted of the presentation of four dictated nonsense syllables. Each stimulus was produced by pressing the same button. The first press produced a sample, the second press prbduced one comparison and a box located in one position on a computer screen both of which terminated concurrently, the third press produced the same sample, and the fourth press produced a second comparison with a box located at a different position on the computer screen. Thereafter the two boxes that accompanied the comparison stimuli were presented concurrently and a comparison was selected by touching one of the boxes. In other studies, the baseline relations for equivalence classes have been established using a constructed-response training procedure (Calcagno, Dube, DeFaria Galvao, & Sidman, 1994; Dube, McDonald, Mcllvane, & Mackay, 1991; Mackay & Sidman, 1984), or sequence training (Lazar, 1977; Lazar & Kotlarchyk, 1986; Sigurdardottir, Green, & Saunders, 1990; Stromer & Mackay, 1993b) which has also been called simultaneous chaining (Terrace, 1983; Terrace, Straub, Bever, &
STIMULUS PAIRING AND EQUIVALENCE
663
Seidenberg, 1977}. In addition, two-choice discrimination reversal procedures have also been used to establish discriminative repertoires that were prerequisite to the emergence of equivalence classes (Sidman, Wynne, Maguire, & Barnes, 1989; Vaughan, 1988). Although it would seem that a broad range of procedures has been used to establish the prerequisite relations for equivalence classes, that breadth may be more apparent than real because the constructedresponse, sequence training, and discrimination reversal procedures can each be viewed as a variation of matching to sample. In the constructedresponse procedure, a sample is presented along with a pool of letters. The letters can be selected in sequence to construct a variety of words. When a sample stimulus is presented, before any letters have been selected, the first letter of the matching word functions as the positive comparison and all other letters function as the negative comparisons. In the presence of the sample and the first letter, the second letter of the matching word serves as the positive comparison in the remaining letter pool and the remaining letters in the pool function as negative comparisons. This analysis can also be applied to all remaining letter selections while constructing the word that matches the sample stimulus. Using a similar analysis, it can also be argued that sequence training is a variant of matching to sample. Specifically, after the selection of the first stimulus in the sequence, the remaining stimuli function as comparisons; the next stimulus in the sequence acts as the positive comparison and the remaining stimUli act as negative comparisons. Finally, Sidman (1994) has argued that the discrimination reversal procedure used by Vaughan could well be a delayed matching procedure in which the stimuli on one trial functioned as samples that signaled the functions served by the stimuli on the next trial (the "comparisons"). From the analysis presented above, it appears that all of the procedures used to establish the baseline relations for equivalence classes are variations of the matching-to-sample format. There is, however, one recent exception. Leader, Barnes, and Smeets (1996) established the baseline relations using a respondent-type sensory preconditioning procedure to pair stimuli that were potential members of the same class. For example, A 1 was presented and terminated after which B1 was presented; no response was required in the presence of either stimulus. Then, the B2-C2 pair might be presented. After multiple presentations of the (N-1) stimulus pairs needed to link the stimuli in a set, the emergence of equivalence classes was assessed with symmetry and equivalence probes presented in a standard matching-to-sample format. These tests were used to determine when subjects would select comparisons that came from the same set as the samples. Most subjects demonstrated the emergence of equivalence classes after repeated exposures to the training and testing trials. When one considers the procedures that have been used to measure the emergence of equivalence classes, the situation is even more restrictive than that found with training procedures. In all
664
FIELDS ET AL.
experiments that have been conducted to demonstrate stimulus equivalence, the tests of reflexivity, symmetry, transitivity, and equivalence have been presented in the traditional matching-to-sample format. If the emergence of equivalence classes can be measured by only one method, that method would have to become a defining component of equivalence classes. Such a constraint would severely limit the theoretical utility of using equivalence class formation to model the emergence of a broad range of complex human behaviors in a wide variety of training and testing conditions (Sidman, 1994). Thus, using a procedure that differs from the traditional matching-to-sample format to establish the conditional discriminations that are the prerequisites of an equivalence class and to measure the emergence of equivalence classes is of practical and theoretical importance. A stimulus-pairing/yesno procedure is one such option Stimulus-pairing/yes-no procedures are also referred to as a "go-left go-right," "yes-no" (D'Amato & Worsham, 1974), or "same/different" procedures (Edwards, Jagielo, & Zentall, 1982). Each trial in a stimuluspairing/yes-no procedure involves the presentation of only two stimuli, the first of which is terminated before the presentation of the second. On some trials, the stimuli are from the same set; in others, the stimuli are from different sets. The YES response occasions reinforcement when a trial contains two stimuli from the same set (within-set trials), and the NO response occasions reinforcement when a trial contains two stimuli from different sets (between-set trials). Stimulus-stimulus relations established with stimulus-pairing/yes-no procedures have been used to study a broad range of psychological phenomena that include stimulus priming (Salota & Lorch, 1986; McNamara & Altarriba, 1988; McNamara & Healy, 1988), the scaling of time by humans (Wearden, 1995), the scaling of stimulus similarity by humans and pigeons (Fetterman & Killeen, 1995), categorization based on numerosity and stimulus mixtures by birds (Honig & Matheson, 1995), and the formation of stimulus classes of "same" and "different" with primates (Neiworth & Wright, 1994; Wright, Santiago, & Sands, 1984; Wright, Shyan, & Jitsumori, 1990) and pigeons (Cook, Cavato, & Cavato, 1995; Edwards et aI., 1982; Wasserman, Hugart, & Kirkpatrick-Steger, 1995). To date, no study has used a stimulus-pairing/yes-no procedure to establish equivalence classes. In the present study, a stimuluspairing/yes-no procedure was used to establish the baseline conditional relations that are the prerequisites to equivalence classes. In addition, performances occasioned by symmetry, transitivity, and equivalence probes presented in a stimulus-pairing/yes-no format demonstrated the emergence of equivalence classes. These results demonstrate that traditional matching-to-sample procedures are not necessary to establish the prerequisites of equivalence classes or to measure the emergence of equivalence classes.
STIMULUS PAIRING AND EQUIVALENCE
665
Method Subjects Eighteen undergraduate students at Queens College/CUNY were recruited from an introductory psychology class. The students had no prior familiarity with experiments dealing with equivalence class formation. They received partial course credit upon completion of the experiment, regardless of performance. Seventeen subjects completed the experiment in one experimental session, and one subject completed the experiment in two sessions separated by 4 days. All subjects who formed classes took between 81 and 161 min to complete the experiment. Apparatus and Stimuli The experimental stimuli were presented to the subjects using an IBM compatible microcomputer with a monochrome monitor. Each subject was seated in a cubicle at a table facing the computer. Responses were made by touching specific keys on the computer keyboard. All responses were automatically recorded by the computer. The experiment was conducted using software specifically developed for , the establishment of equivalence classes. Each of the two classes contained four nonsense syllables which are presented here with their letter-number designations. In Class 1, they were LEQ (A 1), HUK (B1), POV (C1), and BAF (01 ); in Class 2, they were MEV (A2), GUQ (B2), ZOJ (C2), and YAR (22). Procedure Initial instructions. When a subject entered the experimental cubicle, the following instructions were on the computer screen: Thank you for volunteering to be a subject in this experiment . PLEASE DO NOT TOUCH ANY OF THE KEYS ON THE KEYBOARD YET. In this experiment you will be presented with many trials. Each contains two CUES; these will be common words, or three-letter nonsense words such as lEa or WUv. YOUR TASK IS TO DISCOVER WHETHER THE WORDS GO TOGETHER. IF THEY DO, PLEASE PRESS THE KEY LABELLED 'YES.' IF THEY DON'T, PLEASE PRESS THE KEY LABELLED 'NO.' Initially there will also be INSTRUCTIONS that tell you how to respond to the cues, as well as LABELS that will help you to identify the cues on the screen. The labels and the instructions that tell you which KEYS to press will slowly disappear. Your task will be to RESPOND CORRECTLY to the CUES and the INSTRUCTIONS by pressing a key on the computer's keyboard. The experiment is conducted in phases. When each phase ends, the screen will tell you how you did. If you want to take a break at any time, please call the experimenter. Press = = = > SPACE BAR < = = = to continue.
FIELDS ET AL.
666
Trial structure, contingencies, and responses within a trial. Figure 1 provides a schematic representation of a typical trial and the trial contingencies. Each trial began when "Press ENTER" appeared on the screen. Pressing the ENTER key removed this message and displayed the first stimulus which remained on the screen until the subject pressed the space bar. This response cleared the screen for 250 ms and was followed by the presentation of the second stimulus on that trial. It remained on screen until the subject pressed either the number 1 or the number 2 key on the keyboard, which were labeled YES and NO,
Press ENTER
1
I
Press >
Figure 1. The order of stimulus presentation and response requirements in the stimuluspairing/yes-no procedure.
667
STIMULUS PAIRING AND EQUIVALENCE
respectively. The labels were .affixed above the numbers on the surface of the keys and remained there throughout the experiment. When the two stimuli in the trial were from the same set, pressing the YES key was correct. If the two stimuli in the trial were from different sets, pressing the NO key was correct. Either response cleared the stimulus from the screen and produced one of three feedback messages. When informative feedback was scheduled, if the correct response occurred, the message "RIGHT" appeared on the screen and remained there until the subject pressed the "R" key. If the incorrect response occurred, the message "WRONG" appeared on the screen and remained there until the subject pressed the "W" key. On some trials, noninformative feedback was scheduled, in which case, neither RIGHT nor WRONG appeared on the screen Instead, a dashed line, -----, appeared on the screen as soon as the subject selected either comparison and remained there until the subject pressed the "E" key (the noninformative feedback stimulus E stands for "end of trial." The "E" key was used because it is between the Rand W keys on a QWERTY keyboard). After pressing the R, W, or E key, the screen cleared and the next trial began. Keyboard familiarization. After the presentation of the instructions, a subject was taught to emit the appropriate keyboard responses in the presence of each cue used in a trial. The 16 stimulus pairs used in the keyboard familiarization trials are listed in Table 1. Each trial contained a different stimulus pair. The pairs were presented in a random order without replacement for a total of 16 trials. Half of the stimulus pairs were words that were semantically related (e.g., RED COLOR), and the words in the other pairs were not related semantically (e.g., DUNGEONS PENCIL). The relatedness of the words in a trial served as semantic prompts for correct responding on each trial. The presentation of Table 1 Stimuli Used in Keyboard Familiarization REINFORCED RESPONSE YES
NO
1st Stim
2nd Stim
1st Stim
2nd Stim
CANARY RED DOG PAPER CAT MUD THIS MOTHER
SPARROW COLOR WOLF WRITE MOUSE PIG THAT FATHER
EGGS KINGS ANT DUNGEONS ALCOHOL PIANO SOAP COMETS
SNEAKER CAMELS COW PENCIL MOUSE SOCK DOOR FATHER
Note. The word pairs in the columns under YES occasioned reinforcement of the YES response, and the word pairs in the columns under NO occasioned reinforcement of the NO response. Each word pair on a row under the YES column or the NO column was presented on the same trial. The words in the column labeled "1 st Stim" were presented first in a trial , and the words in the column labeled "2nd Stim" were presented second in a trial.
668
FI ELDS ET AL.
semantically related words occasioned reinforcement of the YES response, and the presentation of semantically unrelated words occasioned reinforcement of the NO response. Initially, the trials contained the stimulus pairs, labels for each stimulus, and instructional prompts that specified the responses to be made at each point in a trial. The labels and prompts were deleted serially in six steps, as illustrated in Figure 2. The frames presented in each row of Figure 2, when viewed from left to right, illustrate the SAMPLE
Pre.. ENTER to continue
SA RIGHT/WRONG Pres. SPACE BAR to continue
Compari.on CO Pres. YES or NO
Press R/W to continue
SAMPLE
Pres. ENTER to continue
SA RIGHT/WRONG Pr... SPACE BAR to continue
Press ENTER to continue
SAMPLE SA RIGHT/WRONG Press SPACE BAR to continue
Press ENTER to continue
D D DO
RIGHT/WRONG
CO Press .,.
SA
Press SPACE BAR to continue
Press ENTER to continue
Comparison CO Press •••
SA
Press SPACE BAR to continue
Press ENTER to continue
Compari.on CO Press YES or NO
RIGHT/WRONG
CO
RIGHT/WRONG
Figure 2. The serial removal of prompts and labels during pretraining to teach the responses that were required to successfully complete a trial in the stimulus-pairinglyes-no procedure. Each box indicates a screen display. The boxes on a given row indicate the series of screen displays presented within a trial from onset (on left) to termination (on right).
STIMULUS PAIRING AND EQUIVALENCE
669
stimulus changes that occurred within a trial. The order in which the labels and prompts were eliminated is indicated by the successive rows. A minimum of 12 trials was required to complete the serial elimination of all labels and prompts if all trials occasioned correct responses. Finally, four trials were presented with no prompts. The trial configurations depicted on Rows 1-6 were presented 3, 3, 3, 2, 1, and 4 times, respectively. Any keyboarding errors resulted in the presentation of the relevant keyboard prompt for the next 3 trials. All subjects completed pretraining within the 16 programmed trials. This procedure is similar to that described by Fields et al. (1990). 3-member class formation using the simple-to-complex protocol. Once keyboard familiarization was completed, two 3-member classes were established using the stimulus-pairing/yes-no format as described above. The simple-to-complex protocol was used to control the order of introducing baseline relations and emergent relations tests (Adams, Fields, & Verhave, 1993; Fields et aI., 1991). Each stage of training and testing in this protocol was conducted with blocks of trials. All trials in a block were presented in a random order without replacement. Any repetition of a block resulted in the presentation of the trials in a different order from previous presentations of that block. Table 2 displays the stimulus pairs used and the number of trial presentations in each training and testing block. First, AB was trained. The symmetrical property of AB was then assessed with a block of BA probes. Next, BC was trained after which the symmetrical property of BC was assessed with a block of CB probes. After assessing the maintenance of the symmetry performances with a block containing BA and CB relations together, transitivity was assessed with a block of AC probes. Then, equivalence was assessed with CA tests. Finally, the symmetry, transitivity, and equivalence probes used to assess the emergence of the 3-member classes were presented along with baseline trials in a mixed test block (3M IX) to confirm the formation of two 3member classes. In all of these test blocks, prerequisite baseline relations were also presented as reviews. Reflexivity tests were not included because it was assumed that college students had a generalized identity matching repertoire. Expansion of class size. After subjects passed the 3MIX test, thus, indicating the formation of two 3-member classes, class size was expanded to four members using the same stimulus-pairing/yes-no format. After training the CD baseline relations, a 4-member mixed test (4MIX) was conducted. This mixed test contained all symmetry, transitivity, and equivalence probes for assessing the emergence of 3and 4-member classes along with review of the CD baseline trials. Mastery criterion. The mastery criterion for training and testing was defined as class-consistent responding on more than 97% of the trials within a block. To establish AB, BC, and CD, each training block was repeated until the mastery criterion was reached. All responses during these blocks were followed by informative feedback. Thereafter,
670
FIELDS ET AL.
Table 2 Symbolic Representation of Stimuli Used to Form 3- and 4-member Classes BETWEEN SET TRIALS/BLOCK
WITHIN SET BLOCK FXN
REL
1st/2nd
1st/2nd
1st/2nd
1st/2nd
Train
AB
BL
A1
B1
A2
B2
A1
B2
A2
B1
Test
BA
BL
A1
B1
A2
B2
A1
B2
A2
B1
100 75 25 0
8 8 8 16 B1 A1 B2 A2 B2 A1 Sym B1 A2 16 Train BC BL A1 B1 A1 B2 A2 B1 16 8 8 8 A2 B2 B2 C1 B1 C1 B2 C2 B1 C2 16 8 8 8 Test CB BL A1 B1 A2 B2 A1 B2 A2 B1 8 B1 C1 B2 C2 B1 C2 B2 C1 8 Sym C1 B1 C2 B2 C1 B2 C2 B1 16 Test BAiCB BL A1 B1 A2 B1 A2 B2 A1 B2 8 B2 C1 B1 C2 B1 C1 B2 C2 8 B2 A1 Sym B1 A1 B2 A2 B1 A2 8 C1 B1 C2 B2 C1 B2 C2 B1 8 Test AC BL A1 B1 A2 B2 A1 B2 A2 B1 8 B1 C1 B2 C2 B1 C2 B2 C1 8 Trans A1 C1 A2 C2 A1 C2 A2 C1 16 Test CA BL A1 B1 A2 B2 A1 B2 A2 B1 8 B1 C1 B2 C2 B1 C2 B2 C1 8 Equiv C1 A1 C2 A2 C1 A2 C2 A1 16 Test 3 MIX BL A1 B1 A2 B2 A1 B2 A2 B1 8 B1 C1 B2 C1 B2 C2 B1 C2 8 Sym B1 A1 B2 A2 B1 A2 B2 A1 4 C1 B1 C2 B2 C1 B2 C2 B1 4 A1 C1 A2 C2 A1 C2 A2 C1 4 Trans C1 A1 C2 A2 C1 A2 C2 A1 4 Equiv Train CD BL A1 B1 A2 B2 A1 B2 A2 B1 4 4 4 4 B1 C1 B2 C2 B1 C2 B2 C1 4 4 4 4 C1 01 C2 02 C2 01 12 4 4 4 C1 02 C1 01 C2 02 C1 02 C2 01 4 Test 4 MIX BL B1 A1 B2 A2 B1 A2 B2 A1 8 Sym C2 B1 C1 B1 C2 B2 C1 B2 8 01 C1 02 C2 01 C2 02 C1 8 Trans A1 C1 A2 C2 A1 C2 A2 C1 8 B1 01 B2 02 B1 02 B2 01 8 A2 01 A1 01 A2 02 A1 02 8 Equiv C1 A1 C2 A2 C1 A2 C2 A1 8 01 B1 02 B2 01 B2 02 B1 8 02 A1 01 A2 01 A1 02 A2 8 Note. BLOCK FXN indicates the training or testing function of each block and the symbols represent the stimulus pairs that were the primary focus of the block. The entires in the REL column indicate the type of conditional relation that was measured in the trials listed on that row. BL indicates the baseline conditional relations that were established by training. WITHIN-SET and BETWEEN-SET stimulus pairs used to assess symmetry, transitivity, and equivalence are indicated on the rows labeled Sym, Trans, and Equiv, respectively. The headers labeled "1 stl2nd" designate the first and second stimuli presented in a trial. For example in the "A 1 B1" pair, the A 1 stimulus is presented first and B1 stimulus is presented second. Rows that do not have labels in the REL column serve the function of the last label on a preceding row. The numbers beneath the "TRIALS/BLOCK" header indicate the four levels of informative feedback scheduled in various phases of training. The values in the columns under each feedback level indicate the number of trials in a block for each level of informative feedback. 16
STIMULUS PAIRING AND EQUIVALENCE
671
as long as performance within a block remained at or above the mastery criterion, the percentage of trials that occasioned the presentation of informative feedback was reduced from 100% to 75% to 25% and finally to 0% over successive blocks. When the 75% and 25% feedback levels were used, the particular trials that occasioned informative feedback were randomly determined. If performance within a training block dropped below the mastery criterion, the percentage of feedback was increased to the previous higher level on the next block of trials. The establishment of each baseline relation was confirmed when the mastery level of performance occurred in one block of trials with no informative feedback. During all test blocks, each response was followed by noninformative feedback. The block for a given test was repeated either until the mastery criterion was reached or for a maximum of six successive presentations. If responding within a test block did not reach the mastery criterion after six presentations, the subject was presented with a training block of prer8quisite baseline relations with 100% feedback. Thus, failure of the BA test led to retraining of the AB baseline relations and then retesting of BA. Failure of the CB test led to retraining of the AB and BC baselines and retesting of CB. Failure of the combined test of BA and CB led to retraining of the AB and BC baselines followed by retesting of CB. Failure of the AC test led to retraining of the AB and BC baseline and then retesting of AC. Once the mastery criterion was met, the procedure moved forward in the sequence described above. Failure to reach criterion on a second sixblock cycle of any test led to termination of the experiment for that subject. Confirmation of equivalence class formation with matching-tosample tests. Subjects who met criterion on the 4MIX test in the stimulus-pairing/yes-no format were then exposed to a delayed matching to sample test. This contained trials which assessed the maintenance of the baseline relations established earlier in the experiment as well as probes that assessed the symmetrical, transitive, and equivalence relations that emerged in the prior tests. Trial configurations are illustrated in Table 3. Instructions. The following instructions for the matching-to-sample test were presented on the computer screen: PLEASE DO NOT TOUCH ANY KEYS ON THE KEYBOARD YET. In this part of the experiment many new trials will be presented. Each trial will now contain THREE nonsense words you've seen before. First, ONE will appear at the TOP of the screen then, TWO will appear at the BOTTOM. Your TASK is to select the BOTTOM word that goes with the TOP word. Initial trials will also include INSTRUCTIONS that tell you which keys to press. These instructions will eventually disappear. Thank you for your cooperation! Press to continue.
= = = > SPACE BAR < = = =
672
FIELDS ET AL.
Table 3 Symbolic Representation of Stimuli Used in the Delayed Matching-to-Sample Emergent Relations Test
# of Trials
Type
SET 1 TRIALS Sa Co+ Co-
SET 2 TRIALS Sa Co+ Co-
BL
C1
D1
02
C2
02
01
4
Sym
B1 C1 01
A1 B1 C1
A2 B2 C2
B2 C2 02
A2 B2 C2
A1 B1 C1
8 8 8
Trans
A1 B1 A1
C1 01 01
C2 C2 C2
A2 B2 A2
C2 02 02
C1 C1 01
8 8 8
Equiv
C1 D1 01
A1 B1 A1
A2 B2 A2
C2 D2 D2
A2 B2 A2
A1 B1 A1
8 8 8
Total
76
Note. BL indicates the conditional relations that had been trained to expand class size; Sym indicates symmetry probes, Trans indicates transitivity probed, and Equiv indicates equivalence probes. Sa indicates sample, Co+ and Co- indicate positive and negative comparisons, respectively. For each line, the trials are divided equally between Sets 1 and 2. For trials in a given set, the positive comparison was presented equally often on the right and left.
Trial format and within-trial response requirements. Each trial was presented in the delayed matching-to-sample format illustrated in Figure 3. The sample stimulus appeared centered at the top of the screen. When the space bar was pressed, the sample disappeared and the screen remained blank for 250 ms. At that time, two comparison stimuli appeared at the bottom of the screen, on the same horizontal plane, centered from left to right and separated by 10 cm. One comparison came from the same class as the sample on that trial and the other comparison came from the other class. Both comparisons on a given trial were always of the same letter designation (A 1 and A2 , B 1 and B2, C1 and C2, or 01 and 02). The comparisons remained on the screen until the subject pressed the number 1 or 2 key to indicate selection of the comparison on the left or right, respectively. No informative feedback was scheduled on any trial. Therefore, the selection of either comparison cleared the screen and resulted in the presentation of the trial-ending stimulus, ------, which remained there until the subject pressed the E key. Pressing the E key terminated the -----stimulus and resulted in the presentation of the prompt, "Press space bar to continue." Content of the matching-to-sample test trials. The matching-tosample test consisted of a block of trials that included the CO baseline relations, and all of the 3- and 4-member symmetry, transitivity, and equivalence probes that could be derived from the two previously established 4-member classes. Each type of relation was assessed with the presentation of trials containing stimuli that are represented symbolically in Table 3. The trials listed in Table 3 were presented in a
STIMULUS PAIRING AND EQUIVALENCE
Preas ENTER
673
z
2
2
W (!J
~ 1 W
> «
o
0 AB
75
25
BA
Be
75
25
CB
SA, CB
AC
CA
3MIX
CO
75
25
0
4IIIIX
RELATION TYPE Figure 5. The average number of blocks needed to learn the baseline relations and to pass each emergent relations test presented during the formation of 3-member classes and the expansion of class size to 4 members for the 10 subjects who formed equivalence classes. The unfilled bars indicate the trained stimulus relations, and the filled bars are for single emergent relations probes. The top and bottom of each "I" beam indicates plus and minus one standard deviation.
Table 4 quantifies the process of class formation for each of the 10 subjects who formed classes by listing the percentage of classconsistent responding in each block. There were seven different types of emergent relations tests presented in the experiment: SA, CS, the combined SA and CS test, AC, CA, 3MIX, and 4MIX. Taken as a group, the 10 subjects were presented with a total of 70 different emergent relations tests. Of the 70 different emergent relations tests, 43 occasioned 100% class-consistent responding on the first presentation of a test block, 21 were passed on the second presentation of a test block, and the remaining 6 tests were passed in three to six blocks. Thus, most emergent relations tests were passed quite rapidly. When more than one test block was required to pass an emergent relations test, the baseline trials and the emergent relations probes evoked very high levels of class-consistent responding throughout testing. When expansion of class size was assessed, 6 of the 10 subjects required two blocks to meet criterion. The performances on the initial blocks were at least 95% class consistent. When more than one block was needed to demonstrate the expansion of class size, both the baseline trials and the emergent relations probes evoked high levels of class-consistent responding throughout testing. Matching-to-sample tests. All subjects who formed classes under the stimulus-pairing/yes-no procedure responded in a class-consistent manner during the emergent relations probes and the baseline review trials presented under the matching-to-sample format. Thus, classes remained intact across testing formats. The last column of Table 4 indicates the percentage of trials in a block that occasioned the selection of the comparison from the same class as the sample. All 10 subjects
677
STIMULUS PAIRING AND EQUIVALENCE
Table 4 The Percentage of Trials That Occasioned Class-Consistent Responding in Each Block Presented in Each Phase of the Experiment MTS
STIMULUS PAIRINGNES-NO RESPONDING RELATION TYPE AND PERCENTAGE OF FEEDBACK
AB
75
25
BA
BC
75
25
CB BA,CB
AC
CA
3MIX
CD
75
25
32
32
20
20
20
4MIX 4MIX
TRIALS PER BLOCK SUBJ 16
32
16
32
32
32
100.0
20
76
76
90.0 95.0 100.0 100.0 100.0 92.1 100.0 100.0 98.7
842 100.0
855
37.5 87.5 100.0 100.0 100.0 87.5 100.0 100.0 100.0 100.0 100.0 96.9 100.0 100.0 100.0 100.0
100.0
100.0
90.0 100.0 100.0 100.0 100.0 96.1 95.0 97.4
844
56.3 100.0 100.0 100.0 100.0 87.5 100.0 100.0 100.0 100.0 100.0 96.7 100.0 81.3 75.0 100.0 93.8 100.0
96.7 96.7 96.7 100.0
100.0
95.0 100.0 100.0 95.0 96.7 100.0 100.0 98.7 95.0 95.0 100.0
837
62.5 100.0 100.0 100.0 100.0 81 .3 100.0 100.0 100.0 96.9 96.9 93.8 87.5 100.0 100.0 100.0 100.0 100.0
100.0
100.0
95.0 95.0 100.0 100.0 100.0 97.4 95.0 100.0 95.0 100.0
839
37.5 100.0 100.0 100.0 100.0 75.0 100.0 100.0 100.0 100.0 96.9 96.9 81.3 100.0 90.6 100.0 100.0 90.6 93.8 100.0
96.9 100.0
96.9 100.0
95.0 100.0 100.0 100.0 94.7 96.1 95.0 97.4 98.7 100.0
838
56.3 100.0 100.0 100.0 96.9 75.0 87.5 100.0 87.5 100.0 100.0 96.9 43.8 96.9 93.8 100.0 87.5 100.0 100.0 100.0 100.0 61.3 87.5 100.0
96.9 100.0
969 96.9 96.9 100.0
95.0 100.0 95.0 95.0 98.7 94.7 85.0 95.0 95.0 94.7 95.0 100.0 95.0 98.7 100.0 100.0 100.0
851
87.5 100.0 100.0 100.0 96.9 93.8 100.0 100.0 100.0 100.0 96.9 93.8 100.0 100.0 100.0 93.8 100.0
84.4 71.9 78.1 75.0 78.1 100.0
100.0
100.0
85.0 100.0 100.0 100.0 94.7 96.1 90.0 98.7 98.7 100.0
841
31.3 100.0 100.0 100.0 100.0 87.5 93.8 100.0
87.5 87.5 100.0 100.0 100.0 100.0 96.9 100.0 62.5 100.0 75.0 81 .3 93.8 93.8 100.0
96.9 100.0
100.0
90.0 95.0 100.0 100.0 96.1 100.0 95.0 100.0 98.7 100.0
840
62.5 100.0 100.0 100.0 100.0 93.8 100.0 100.0 100.0 100.0 100.0 84.4 43.8 100.0 100.0 93.8 100.0
100.0
100.0
95.0 100.0 100.0 100.0 97.4 100.0 100.0 100.0 95.0 100.0 100.0 100.0
857
43.8 100.0 87.5 100.0 100.0 75.0 100.0 100.0 100.0 100.0 96.9 93.8 62.5 87.5 100.0 96.9 100.0 100.0 43.8 87.5 62.5 100.0 62.5 100.0 93.8 93.8 100.0
100.0
100.0
95.0 100.0 100.0 100.0 96.1 94.7 75.0 98.7 96.1 100.0 98.7
Note. Data are included for the 10 subjects who formed classes, each of whom is identified in the first column. The column headers "75," "25," and "0" indicate the percentage feedback for the relation to the immediate left of the "75" header. The data in the right-hand column were obtained in the matching-to-sample test.
678
FIELDS ET AL.
responded at or near criterion on all test blocks. Of the 6 subjects who did not reach the mastery criterion on the first block, 4 reached criterion by the second block and the remaining 2 subjects did so on the third block. These percentages are based on the 76 trials in the 4MIX test block which contained 4 review trials for the previously trained CD relations and 72 emergent relations trials. Therefore, the overall percentages are reflective of emergent relations performances. Failures of equivalence class formation. The 8 subjects who failed to form 3-member classes all learned the AB and BC relations during Table 5 Percentage of Each Type of Trial in Sets of Test Blocks That Occasioned Yes Response for Subjects Not Forming Classes under Stimulus-PairingNes-No Procedure Baseline Relations AB BC W/IN
BET
W/IN
Emergent Relations Probes AC TRANS BA SYM CB SYM BET SIC
# A1 A2 A1 A2 SBJ BLK B1 B2 B1 B1
B1 B2 B1 B2 C1 C2 C2 C1
858 6 859 6
67 75 8 8 75 83 42 25 58 42 33 17
6
50 83 26 33 58 83 17 25 585080 8 17 0 25
92 92 100 100
8 0
0 0
H H
848 6*
0
100 100
0
0
H
0
849 6 100 100 0 0 92 100 58 0 H 61001000010010080 H 61001000010010000 C 854 6 6
100 83 100 100
8 8
8 0
100 100 100 83
0 0
BET
W/IN
BET
C1 C2 C1 C2 B1 B2 B2 B1
W/IN
BET SIC
A1 A2 A1 A2 C1 C2 C2 C1
71 75 42 58
L L L
859 6 100 92 2* 100 25 25 100
W/IN
B1 B2 B1 B2 A1 A2 A2 A1
83 58 8 38 58 63 21 17 75 50
0 0
8 0
92 92 0 8 100 75 25 25
H
o 100 o 17
0
0
100 83
0
0
H
0
0
0 83 21 60 o 8
0 0 0
0 4 0
N L N
83 100
L
96 4100 0 961310017
S S
8 0
H H
846 6 100 100 0 0 83 92 0 0 61001000010010000
H C
o o
0 0
0 0
0 0
N N
845 6 6
83 92 100 100
8 0 0 17
67 92 35 92
8 0
0 0
H H
o
4
853 3#
83 100
0 17
100 100
0
0
H
8
8 4 0
0 0 0
4 4 0
N N N
8486 83100 0 0 100100 0 0 61001000010010000
H C
o o
0 0
0 0
0 0
N N
Note. W/IN indicates within-set test trials. BET indicates between-set test trials. The letternumber pairs listed in the third and fourth rows indicate the initial and second stimuli presented in a given type of trial. The first column indicates the subject identification number. The second column indicates the number of blocks from which the percentages are computed. The column header SIC lists the form of stimulus control implied by the pattern of responding for the within and between set relations to the left of the code letters: C = complete stimulus control, H = high level of conditional discriminative control, L = low level of conditional discriminative control, R = unsystematic or random responding, N = preference for the NO response, and S = control by sample stimuli. * The combined BAiCB symmetry tests were passed on the ninth and seventh presentations of the test block by Subjects 859 and Subject 848 respectively. All of the remaining subjects did not pass the indicated emergent relations tests with repetition of the test blocks. # Subjects 853 left the experiment after the third AC transitivity test block.
STIMULUS PAIRING AND EQUIVALENCE
679
training. Failures occurred during the emergent relations test blocks. Table 5 illustrates the responding occasioned by the between- and within-set trials presented in the test blocks used to assess the emergence of the 3-member classes. The AB and BC baseline trials in the test blocks yielded 18 sets for analysis. The baseline trials occasioned 100% class-consistent responding which showed maintenance of control in three sets (labeled C), 81 to 98% class-consistent responding indicating partial disruption of stimulus control in 12 sets (labeled H), and 62-80% class-consistent responding indicating low levels of conditional stimulus control in three sets (labeled L). When performances occasioned by the emergent relations probes are considered, low levels of conditional stimulus control were observed in the presence of two sets of symmetry probes (labeled L for Subjects 849 and 858), one combined symmetry (labeled L for Subject 859), and two transitivity probe sets (labeled L for Subject 859). Emission of the NO response regardless of stimulus content was the most frequent error pattern (labeled N). This occurred in the presence of one symmetry one combined symmetry, and seven transitivity probe sets. The transitivity probes occasioned the most errors. In two cases, the within- and between-set transitivity probes evoked differential levels of responding indicating low levels of conditional discriminative control. In seven cases, the NO response was evoked regardless of trial content. In two other cases, one response occurred in the presence of the A 1 sample and the other response occurred in the presence of the A2 sample, indicating control by sample stimuli alone. Discussion
Equivalence class formation under the stimulus-pairing/yes-no procedure. Using a stimulus-pairing/yes-no procedure, six within-set and six between-set conditional relations were directly trained. Interspersed among the relations that were trained, subjects were presented with a variety of within- and between-set emergent relations probes. The occurrence of YES responses to all of the within-set emergent relations probes and NO responses to all of the between-set probes demonstrated the immediate emergence of two 4-member classes. The YES responses occasioned by the within-set probes demonstrated the interchangeabilty of the stimuli in each stimulus set (Sidman, 1994; Spradlin & Saunders, 1986). The test performances also documented the properties of symmetry, transitivity, and the combined effects of symmetry and transitivity among the stimuli within each stimulus set. The NO responses occasioned by the between-set probes demonstrated a discrimination of stimuli from different sets (Dinsmoor, 1995a, 1995b; Fields et aI., 1995; Mackay, 1991; Sidman, 1992). The emergent performances, then, provided the first demonstration of the immediate emergence of equivalence classes which used a training and testing
680
FIELDS ET AL.
procedure that differed in format and contingencies from those defined by a matching-to-sample procedure. The results of the current experiment expand incrementally the methods that can be used to establish the baseline relations that are the prerequisites to equivalence classes; in addition to matching to sample, the baselines can be established with a respondent-type sensory preconditioning procedure (Leader et aI., 1996), or a stimuluspairing/yes-no procedure. The current experiment, along with the results of the Leader et al. (1996) study, then, show that the use of a matchingto-sample procedure is not necessary to train the relations that are the prerequisites of an equivalence class. In all prior experiments, the probes used to assess symmetry, transitivity, and their combined effects, and, thus, the emergence of equivalence classes, were presented in the matching-to-sample format. The results of the current experiment showed that the emergence of equivalence classes can also be measured with probes presented in a stimulus-pairing/yes-no format. Thus, the results of this experiment double the methods that can be used to assess equivalence class formation. These findings make the phenomenon of equivalence class formation less dependent on the use of specific procedures for their establishment or for their definition (Sidman, 1994). In the present study, equivalence classes were formed when no two stimuli were ever presented concurrently in training and in testing. The only other demonstration was provided by Dube et al. (1993). The classes established in the present study consisted of visual stimuli and were formed using a stimulus-pairing/yes-no procedure for training and testing; the classes established by Dube et al. (1993) consisted of auditory stimuli alone and were formed using a successive delayed matching-to-sample procedure for training and testing. Because different stimuli and different procedures were used, the results of both experiments demonstrate the generality of equivalence class formation in which there is no temporal overlap among the stimuli that come to function as class members. The techniques described in these experiments have potentially important implications for studying the neural substrate of equivalence class formation which might be measured using ERP (Polich,1993) or fMRI (Cohen, Noll, & Schneider,1993) imaging technologies. Such measurements are facilitated by the presentation of temporally isolated stimuli. For the same reason, studying the interactions of equivalence classes and priming (Balota & Lorch,1986; McNamara & Altarriba,1988; McNamara & Healy, 1988) would be facilitated by the procedures described herein. Substitutability. An essential property of an equivalence class is that it functions as a transfer network for new responses and stimulus functions (Barnes, Browne, Smeets, & Roche, 1995; Barnes & Keenan, 1993; de Rose, Mcllvane, Dube, Galpin, & Stoddard, 1988; Dougher, Augustson, Markham, Greenway, & Wulfert, 1994; Fields et aI., 1993;
STIMULUS PAIRING AND EQUIVALENCE
681
Fields et aI., 1990; Hayes, Kohlenberg, & Hayes, 1991; Sidman, 1994). Fields et al. (1990) and Fields et al. (1993) noted that the expansion of class size can be viewed as a transfer test because a stimulus linked by training to one member of a class occasioned the selection of the other class members. Indeed, many experiments have demonstrated such an effect for classes established with the matching-to-sample procedure. In the present experiment, after forming two 3-member classes, CD relations were trained and emergent relations probes that contained the D stimuli in combination always led to the selection of the A and B stimuli that were members of the previously trained ABC classes. The performances during the tests for expansion of class size, then, demonstrated the interchangeability of previously established members of the equivalence classes. Thus, the classes formed under the stimuluspairing/yes-no procedure functioned in the same manner as equivalence classes established using matching-to-sample procedures. Transfer to the matching-to-sample format. After establishing classes using the stimulus-pairing/yes-no procedure, the maintenance of classconsistent responding was demonstrated when class members were presented in the typical matching-to-sample format. These performances confirmed that the classes established and expanded in size using a stimulus-pairing/yes-no procedure continued to function as such when presented in the format that is traditionally used to demonstrate equivalence class formation. To summarize, the test performances that showed the substitutability of class members and transfer of function across testing formats support the view that stimulus classes formed with the stimulus-pairing/yes-no procedure have the same behavioral properties as equivalence classes formed using matching to sample procedures. Equivalence classes were thus established using the stimulus-pairing/yes-no procedure. Failure of equivalence class formation. Of the 18 subjects, 8 did not form classes in this experiment. One source of failure was disruption of conditional discriminative control by the baseline relations included in the emergent relations test blocks. The likelihood of equivalence class formation using the stimulus-pairing/yes-no format, then, could probably be increased by the introduction of variables that would maintain baseline performances during emergent relations tests and not promote discriminations between the trained relations and the probes. Another frequent source of failure involved selecting the NO response on most of the symmetry or transitivity probes trials in a test block. Although the NO response was "appropriate" on the between-set probes, it was not appropriate on the within-set probes. The occurrence of the NO response on all or most of the within-set probes, however, makes sense if viewed as a report of the novelty of the within-class stimulus configurations. Because the stimuli in the transitivity probes had not been previously seen together and the stimuli in the symmetry probes had not been presented in the same order, the NO response could be viewed as being appropriate. This interpretation received partial support from verbal reports of two subjects
682
FIELDS ET AL.
during informal postexperimental interviews. Thus, the identification of variables that suppress such alternative sources of control during probe trials should increase the likelihood of equivalence class formation (Sidman, 1994; Wulfert, Dougher, & Greenway, 1991) using the stimuluspairing/yes-no procedure. Potential effects of language-based parameters. In the present experiment, verbally sophisticated subjects were presented with three language-based cues. (a) The instructions directed the subject to identity whether the words "GO TOGETHER." (b) The Keyboard Familiarization procedure involved the presentation of semantically related stimuli to prompt the occurrence of correct within-trial keyboard responses. (c) The words YES and NO were the discriminanda that identified the response keys. It is possible that these cues occasioned language-based repertoires that might have influenced the outcome of the current experiment. Saunders, Saunders, Williams, and Spradlin (1993) found that the addition of instructions enhanced likelihood of class formation by individuals with moderate levels of mental retardation. In contrast, Spencer and Chase (1996) found that likelihood of equivalence class formation by college students was not influenced by variation in instructions. Thus, the extent to which the language-based cues influenced likelihood of class formation in the current experiment is not clear. Additional research is needed to determine the effect of languagebased variables on equivalence class formation under the stimuluspairing/yes-no protocol. Equivalence class formation in natural settings. In natural settings, an individual will, of necessity, be exposed to a variety of contingencies. To what extent will equivalence classes emerge under such a diversity of conditions? Prior laboratory demonstrations of equivalence class formation used matching-to-sample contingencies in a variety of guises. To date, laboratory experiments have provide limited support for the view that equivalence classes can emerge with exposure to the variety of contingencies that would be experienced in natural settings. The present experiment demonstrated the emergence of equivalence classes using a procedure other than matching to sample. These results, then, bolster the view that equivalence classes can emerge in a natural setting through contact with a variety of training and testing contingencies (Commons & Rodriguez, 1993; Dougher et aI., 1994; Watt et aI., 1991).
References ADAMS, B. J., FIELDS, L., & VERHAVE, T. (1993). Effects of test order on intersubject variability during equivalence class formation. The Psychological Record, 43, 133-152. BALOTA, D. A., & LORCH, R. F. (1986). Depth of automatic spreading activation: Mediated priming effects in pronunCiation but not in lexical decision. Journal of Experimental Psychology; Learning, Memory, and Cognition, 12, 336-345.
STIMULUS PAIRING AND EQUIVALENCE
683
BARNES, D., BROWNE, M., SMEETS, P., & ROCHE, B. (1995). A transfer of functions and conditional transfer of functions through equivalence relations in three- to six-year-old children. The Psychological Record, 45, 405-430. BARNES, D., & KEENAN, M. (1993). A transfer of functions through derived arbitrary and nonarbitrary stimulus relations. Journal of the Experimental Analysis of Behavior, 59, 61-82. BENTALL, R. P., DICKINS, D. W., & FOX, S. R. A. (1993). Naming and equivalence: Response latencies for emergent relations. The Quarterly Journal of Experimental Psychology, 46B, (May), 187-214. BUSH, K. M., SIDMAN, M., & DE ROSE, T. (1989). Conditional control of emergent equivalence relations. Journal of the Experimental Analysis of Behavior, 51, 29-46. CALCAGNO, S., DUBE, W. v., DE FARIA GALVAO, 0., & SIDMAN, M. (1994). Emergence of conditional discriminations after constructed-response matching-to-sample training. The Psychological Record, 44, 509-520. COHEN, J. D., NOLL, D. C., & SCHNEIDER, W. (1993). Functional magnetic resonance imaging: Overview and methods for psychological research. Behavior Research Methods, Instruments, & Computers, 25, 101-113. COMMONS, M. L., & RODRIGUEZ, J. A. (1993). The development of hierarchically complex equivalence classes. The Psychological Record, 43, 667-698. COOK, R. G., CAVATO, K. K., & CAVATO, B. R. (1995). Same-different texture discrimination and concept learning by pigeons. The Psychological Record, 21,253-261. D'AMATO, M. R., & WORSHAM, R. W. (1974). Retrieval cues and short-term memory in capuchin monkeys. Journal of Comparative and Physiological Psychology, 86, 274-282. DE ROSE, J. C., MCILVANE, W. J., DUBE, W. v., GALPIN, V. C., & STODDARD, L. T. (1988). Emergent simple discrimination established by indirect relation to differential consequences. Journal of the Experimental Analysis of Behavior, 50,1-20. 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. DICKINS, D. W., BENTALL, R. P., & SMITH, A. B. (1993). The role of individual stimulus names in the emergence of equivalence relations: The effects of interpolated paired-associates training of discordant associations between names. The Psychological Record, 43, 713-724. DINSMOOR, J. A. (1995a). Stimulus control: Part I. The Behavior Analyst,18, 51-68. DINSMOOR, J. A. (1995b). Stimulus control: Part II. The Behavior Analyst, 18, 253-270. DOUGHER, M. J., AUGUSTSON, E., MARKHAM, M. R., GREENWAY, D. E., & WULFERT, E. (1994). The transfer of emotional respondent eliciting and extinction functions through stimulus equivalence classes. Journal of the Experimental Analysis of Behavior, 62, 331-351. DUBE, W. v., GREEN, G., & SERNA, R. W. (1993). Auditory successive conditional discrimination and auditory stimulus equivalence classes. Journal of the Experimental Analysis of Behavior, 59, 103-114. DUBE, W. v., MCDONALD, S. J., MCILVANE, W. J., & MACKAY, H. A. (1991). Constructed-response matching-to-sample and spelling instruction. Journal of Applied Behavior Analysis, 24, 305-317.
684
FIELDS ET AL.
EDWARDS, C. A., JAGIELO, J. A., & ZENTALL, T. R. (1982). Acquired equivalence and distinctiveness in matching to sample by pigeons: Mediation by reinforcer-specific expectancies. Journal of Experimental Psychology: Animal Behavior Processes, 8, 244-259. FETTERMAN, J. G., & KILLEEN, P. R. (1995). Categorical scaling of time: Implications for clock-counter models. Journal of Experimental Psychology: Animal Behavior Processes, 21, 43-63. FIELDS, L., ADAMS, B. J., NEWMAN, S., & VERHAVE, T. (1992). Interactions of emergent relations during the formation of equivalence classes. The Quarterly Journal of Experimental Psychology, 45B, 125-138. FIELDS, L., ADAMS, B. J., & VERHAVE, T. (1993). The effects of equivalence class structure on test performances. The Psychological Record, 43, 697-713. FIELDS, L., ADAMS, B. J., VERHAVE, T., & NEWMAN, S. (1990). The effects of nodality on the formation of equivalence classes. Journal of the Experimental Analysis of Behavior, 53, 345-358. FIELDS, L., ADAMS, B. J., VERHAVE, T., & NEWMAN, S. (1993). Are stimuli in equivalence classes equally related to each other? The Psychological Record, 45, 85-105. FIELDS, L., LANDON-JIMENEZ, D. V., BUFFINGTON, D. M., & ADAMS, B. J. (1995). Maintained nodal distance effects after equivalence class formation. Journal of the Experimental Analysis of Behavior, 64, 129-146. FIELDS, L., REEVE, K. F., ADAMS, B. J., & VERHAVE, T. (1991). The generalization of equivalence relations: A model for natural categories. Journal of the Experimental Analysis of Behavior, 55, 305-312. FIELDS, L., & VERHAVE, T. (1987). The structure of equivalence classes. Journal of the Experimental Analysis of Behavior, 48, 317-332. FIELDS, L., VERHAVE, T., & FATH, S. J. (1984). Stimulus equivalence and transitive associations: A methodological analysis. Journal of the Experimental Analysis of Behavior, 42, 143-157. HAYES, S. C., KOHLENBERG, B. S., & HAYES, L. (1991). The transfer of specific and general consequential functions through simple conditional equivalence relations. Journal of the Experimental Analysis of Behavior, 56, 119-137. HONIG, W. K., & MATHESON, W. R. (1995}. Discrimination of relative numerosity as stimulus mixture by pigeons with comparable tasks. Journal of Experimental Psychology: Animal Behavior Processes, 21,348-363. LAZAR, R. M. (1977). Extending sequence-class membership with matching-tosample. Journal of the Experimental Analysis of Behavior, 27, 381-392. LAZAR, R. M., DAVIS-LANG, D., & SANCHEZ, L. (1984). The formation of visual stimulus equivalences in children. Journal of the Experimental Analysis of Behavior, 41 , 251-266. LAZAR, R. M., & KOTLARCHYK, B. J. (1986). Second-order control of sequenceclass equivalences in children. Behavioural Processes, 13, 205-215. LEADER, G., BARNES, D., & SMEETS, P. M. (1996). Establishing equivalence relations using a respondent-type training procedure. The Psychological Record, 46, 685-706. MACKAY, H. A. (1991). Conditional stimulus control. In I. Iversen & K. Lattal (Eds.), Experimental analysis of behavior (Part 1) (pp. 301-350). New York, NY: Elsevier Science Publishers BV. MACKAY, H. M., & SIDMAN, M. (1984). Teaching new behavior via equivalency relations. In P. H. Brooks, R. Sperber, & C. McCauley (Eds.), Learning and cognition in the mentally retarded (pp. 493-513). Hillsdale, NJ: Erlbaum .
STIMULUS PAIRING AND EQUIVALENCE
685
MCDONAGH, E. C., MCILVANE, W. C., & STODDARD, L. T. (1984). Teaching coin equivalences via matching-to-sample. Applied Research in Mental Retardation, 5,177-197. MCNAMARA, T. P., & ALTARRIBA, J. (1988). Depth of spreading activation revisited: Semantic mediated priming occurs in lexical decisions. Journal of Memory and Language, 27, 545-559. MCNAMARA, T. P., & HEALY, A. F. (1988) . Semantic, phonological, and mediated priming in reading and lexical decisions. Journal of Experimental Psychology: Learning, Memory, and Cognition, 14,398-409. NEIWORTH, J., & WRIGHT, A. A. (1994). Monkeys (Macaca mulatta) learn category matching in a nonidentical same-different task. Journal of Experimental Psychology: Animal Behavior Processes, 20, 429-435. POLICH, J. (1993). Cognitive brain potentials. Current Directions in Psychological Science, 2, 175-178. SAUNDERS, K. J., SAUNDERS, R. R, WILLIAMS, D. C., & SPRADLIN, J. E. (1993). An interaction of instructions and training design on stimulus class formation: Extending the analysis of equivalence. The Psychological Record, 43, 725-744. SAUNDERS, R L., & GREEN, G. (1992). Tr,A nonequivalence of behavioral and mathematical equivalence. Journal of the Experimental Analysis of Behavior, 57,227-241. SAUNDERS, R. R, SAUNDERS, K. J., KIRBY, K. C., & SPRADLIN, J. E. (1988). The merger and development of equivalence classes by unreinforced conditional selection of comparison stimuli. Journal of the Experimental Analysis of Behavior, 50,145-162. SCHUSTERMAN, R, & KASTAK, D. (1993). A California sea lion (Zalophus californianus) is capable of forming equivalence relations. The Psychological Record, 43, 823-840. SIDMAN, M. (1971). Reading and auditory-visual equivalences. Journal of Speech and Hearing Research, 14, 5-13. SIDMAN, M. (1992). Equivalence relations: Some basic considerations. In S. C. Hayes & L. J. Hayes (Eds.), Understanding verbal relations (pp. 15-27). Reno, NV: Context Press. SIDMAN, M. (1994). Equivalence relations and behavior: A research story. Boston, MA: Authors Cooperative, Inc. SIDMAN, M., KIRK, B., & WILLSON-MORRIS, M. (1985). Six-member stimulus classes generated by conditional-discrimination procedures. Journal of the Experimental Analysis of Behavior, 43, 21-42. SIDMAN, M., & TAILBY, W. (1982). Conditional discrimination vs matching-tosample: 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. SIGURDARDOTTIR, Z. G., GREEN, G., & SAUNDERS, R. R. (1990) . Equivalence classes generated by sequence training. Journal of the Experimental Analysis of Behavior, 53, 47-64. SPENCER, T. J., & CHASE, P. N. (1996). Speed analyses of stimulus equivalence. Journal of the Experimental Analysis of Behavior, 65, 643-659. SPRADLIN, J. E., & SAUNDERS, R. R. (1986) . The development of stimulus classes using match-to-sample procedures: Sample classification versus comparison classification. Analysis and Intervention in Developmental Disabilities, 6, 41-58.
686
FIELDS ET AL.
STROMER, R., & MACKAY, H. A (1992). Spelling and emergent picture-printed word relations established with delayed identity matching to complex samples. Journal of Applied Behavior Analysis, 25, 893-904. STROMER, R., & MACKAY, H. A. (1993a). Delayed identity matching to complex samples: Teaching students with mental retardation spelling and the prerequisites for equivalence classes. Research in Developmental Disabilities, 14, 19-38. STROMER, R., & MACKAY, H. A. (1993b). Human sequential behavior: Relations among stimuli, class formation, and derived sequences. The Psychological Record, 43,107-131. STROMER, R., MCILVANE, W. J., DUBE, W. V., & MACKAY, H. A. (1993). Assessing control by elements of complex stimuli in delayed matching to sample. Journal of the Experimental Analysis of Behavior, 59, 83-102. TERRACE, H. S. (1983). Simultaneous chaining. In M. L. Commons, R. J. Herrnstein, & A R. Wagner (Eds.), Quantitative analyses of behavior, Vol. 4: Discrimination Processes. Cambridge, MA: Ballinger. TERRACE, H. S., STRAUB, R., BEVER, T. G., & SEIDENBERG, M. (1977). Representation of a sequence by pigeons. Bulletin of the Psychonomic Society, 10, 269. VAUGHAN, W., Jr. (1988). Formation of equivalence sets in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 14, 36-42. WASSERMAN, E. A., HUGART, J. A, & KIRKPATRICK-STEGER, K. (1995). Pigeons show same-different conceptualization after training with complex visual stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 21,248-252. WATT, A., KEENAN, M., BARNES, D., & CAIRNS, E. (1991). Social categorization and stimulus equivalence. The Psychological Record, 41, 33-50. WEARDEN, J. H. (1995). Categorical scaling of stimulus duration by humans. Journal of Experimental Psychology: Animal Behavior Processes, 21, 318-330. WRIGHT, A. A, SANTIAGO, H. C., & SANDS, S. F. (1984). Monkey memory: Same/different concept learning, serial probe acquisition, and probe delay effects. Journal of Experimental Psychology: Animal Behavior Processes, 10, 513-529. WRIGHT, A. A , SHYAN, M. R., & JITSUMORI, M. (1990). Auditory same/different concept learning by monkeys. Animal Learning and Behavior, 18,287-294. WULFERT, E., & DOUGHER, M. J., & GREENWAY, D. E. (1991). Protocol analysis of the correspondence of verbal behavior and equivalence class formation. Journal of the Experimental Analysis of Behavior, 56, 489-504.
