Printer Friendly

The effects of sexually explicit words on the formations of stimulus equivalence classes.

Stimulus equivalence can be defined as a relationship between two things such that one of those things may be substituted for (or come to take the place of) the other in a particular setting and not significantly alter the situation; that is, they come to evoke the same or nearly the same response. Stimulus equivalence is a behaviorally based approach to exploring and understanding how humans process the relationships between these things - these symbolic stimuli - a process traditionally viewed as cognitive in nature but for which behavioral response patterns do emerge.

Stimulus equivalence is concerned with both verbal behavior and rule-governed behavior of humans (Plaud, 1995). Utilizing a conditional discrimination procedure, in stimulus equivalence studies a subject is typically shown a stimulus (e.g., A), called a sample stimulus. Different stimuli are then presented, called comparison stimuli (e.g., B and c). The subject learns through explicit reinforcement of "correct" choices, for example, to pick stimulus B when shown stimulus A. The subject also learns to pick stimulus C when shown stimulus B as the sample. Subjects learn these responses by direct reinforcement contingencies, therefore, it is not unexpected that subjects will choose stimulus B when first shown stimulus A, or choose stimulus C when shown stimulus B. because direct reinforcement has set up explicit (matching-to-sample) pairing of these stimulus combinations.

Stimulus equivalence, however, refers to the unexpected finding that if a subject was shown stimulus A as a sample stimulus, and given the choice between comparison stimuli C and D (controlling for stimulus placement and the number of stimulus presentations), the subject will most likely choose stimulus C, even though there were no reinforcement contingencies in effect for that choice. Therefore, equivalence of responding is shown in the absence of explicit reinforcement contingencies; in other words, such stimulus combinations may come to act as members of a single class (Plaud, 1995) without any direct training to respond to the stimuli in this way.

Sidman and Tailby (1982) defined three basic elements to stimulus equivalence. The first condition is called reflexivity, and it refers to the phenomenon that a stimulus can be matched to itself. For example, if the sample stimulus were X, the subject matches it to the comparison stimulus X. Symmetry is defined as the property of reversing conditional discrimination relationships. For example, if shown sample stimulus X, and trained to choose comparison stimulus Y, without explicit training the subject then correctly chooses stimulus X when presented with sample stimulus Y. Transitivity is shown if the relationship given X choose Z is established without explicit training.

According to Hayes (1991), relations among stimuli have three logical characteristics. The first characteristic Hayes defines as mutual entailment: responding to one stimulus in terms of another stimulus (e.g., if X is related to Y, then Y is related to X; if X is better than Y, then Y is worse than X, etc.). Expressed as a logical argument, mutual entailment has the following form:

[Mathematical Expression Omitted] (1.1)

where r stands for relation, x and y denote the specific aspects of the relation relevant to the two stimuli involved, A and B. Mutual entailment means that a person is responding relationally to A and B as well as to B and A. For example, "if a person is responding to an object because of its relation to a word, the person must also be able to respond to the word because of its relation to the object" (Hayes, 1991, p. 22, bold original).

The second characteristic of relations is termed combinatorial entailment. This property is different from mutual entailment in that the exact nature of the entailed relations may not be known. The logical form of this argument is as follows

[Mathematical Expression Omitted] (1.2)

and the combinatorial nature of the characteristic is seen on the right side of Expression 1.2. As specified in this expression, a relation is detailed between stimuli A and B, and between stimuli B and C (left side of Expression 1.2). Further, mutual entailment means that the relations between C and B and between B and A must also be specified (left side of Expression 1.2). However, the relation between stimuli A and C and stimuli C and A may not be specified or known completely (right side of Expression 1.2).

If stimulus A is different than stimulus B, and B is different than stimulus C, the relation between A and C and between C and A may not be known. For example, "if Harry is bigger than Sam, and Sam is faster than Joe, who is prettier, Harry or Joe?" (Hayes, 1991, p. 23). However, the importance of combinatorial entailment lies in the fact that it states that if relational responding exists between A and B and between B and C, then some type of relational responding should also exist between A and C. Relational responding should exist between A and C, according to combinatorial entailment, because of the logical form "A is related to B is related to C" (i.e., all terms are related through the mediation of term B between terms A and c), even if the exact nature of this relation is not specified.

The third characteristic of relational control theory is labeled transfer of functions. Expressed logically, this characteristic states that given mutual entailment and combinatorial entailment between stimuli A, B, and C, the psychological function of A entails both the psychological functions of both B and C in terms of underlying relations:

[Mathematical Expression Omitted] (1.3)

where f refers to a stimulus function that is psychologically relevant and r refers to the type of relational responding that is occurring. Hayes (1991) explains:

Suppose A is specified as smaller than B, and B is smaller than C. Suppose that A is given conditioned reinforcement functions and that these functions are relevant to the relation of size between A, B, and C. We might expect that B and C will have ordinally more value as conditioned reinforcers than A, to the degree that a response to B or C is an aspect of a relational response to A, B, and C. As a practical example, consider a child who has learned to work for pennies. Pennies are now conditioned reinforcers. As the child learns to relate events in terms of the abstract relation of amount, the child will work harder when promised a nickel or a dime than a penny without direct experience of getting paid with nickels or dimes (p. 24).

Mutual entailment specifies that one stimulus may be responded to in terms of another stimulus. Combinatorial entailment further specifies that new relations may emerge, even if their exact nature is unknown (i.e., even if the underlying psychological dimension of relation is not exactly known). Transfer of functions specifies that new emergent combinatorial relations among stimuli are not actually unknown, but rather function on a psychologically relevant dimension. These characteristics form a basis for an explanation of stimulus equivalence phenomena.

Therefore, according to relational control theory of stimulus equivalence, given stimuli A, B, and C, certain new combinations of stimuli may come about through mutual and combinatorial entailment. Further, transfer of functions (or equivalences) may result if the stimuli share an underlying psychologically relevant context, such as in the penny, nickel, and dime example, in which these three stimuli covary along the psychological dimension of secondary monetary reinforcers, all of which could be exchanged for primary reinforcers (i.e., exchanging money for goods).

As summarized by Plaud (1995), stimulus equivalence research to date indicates that the formation of equivalence relationships is primarily a property of human (or verbal) behavior; that reinforcement is required for initial training of discriminative relationships, but reinforcement only expedites the emergence of equivalence relations; equivalence relations have been shown to appear in children as young as 3 years old, as well as in children 3-5 years old who are hearing and/or verbally impaired as well as mentally retarded; and equivalences may be formed among many different stimuli, with the production of large numbers of untrained conditional relations. Further, stimulus equivalence testing may prove to be another technique to assess psychological factors generally studied via cognitive assessment techniques.

When equivalence between stimuli is established, the related stimuli are considered to be in a class (Fields et al., 1990). Members of a class resemble one another enough to evoke similar responses in the absence of direct training.

Stimulus equivalence research has relevance for behavioral research in at least two areas (Hayes, 1991): (a) verbal behavior and other language phenomena, such as a child substituting an object for a written word without explicit training (cf. Plaud & Newberry, 1996) and (b) basic behavioral research concerning the three-term contingency (i.e., discriminative stimulus, response, consequence).

The investigation of stimulus equivalence formation has covered a wide range of stimuli and subjects. Early studies on the formation of stimulus equivalence classes used symbols such as Greek or Latin characters or Mandarin characters with nonsense syllables in order to show that humans can form connections between targeted stimuli with no formal training to do so (Fields et al., 1990; Lazar, Davis-Lang, & Sanchez, 1989).

Stimuli from other modalities have also been used to explore the human capability to form equivalence classes. Subjects have been able to form equivalence classes using verbal stimuli and gustatory stimuli (Annett & Leslie, 1995; Hayes, Tilley, & Hayes, 1988); tactile stimuli (Plaud, 1995; Tierney, DeLargy, & Bracken, 1995); auditory stimuli (Dube, Green, & Serna, 1993) and musical stimuli (Hayes, Thompson, & Hayes, 1989). A consistent finding is that humans can form equivalence classes both within and across modalities. The exploration of equivalency and how humans form these symbolic relationships has included stimuli which have already acquired meaning for the individual. Studying clinically anxious and nonanxious subjects, Leslie, Tierney, Robinson, Keenan, and Watt (1993) found that seven of eight anxious subjects were not able to form equivalence classes when anxiety-provoking situational words such as "job interview" or "exam" were used as the sample stimuli while six of eight nonanxious subjects were able to do so. Plaud (1995) found that subjects took more time to form equivalence classes with fear words (words relating to snakes) than neutral words (words relating to flowers).

Watt, Keenan, Barnes, and Cairns (1991) demonstrated the applicability of these principles to a social phenomenon significant in Northern Ireland. The stimulus sets were Catholic names (A), nonsense syllables (B), and Protestant symbols (C). English subjects responded by forming equivalence classes. Northern Irish subjects matched the novel Protestant stimuli only with the Protestant sample stimulus, unable to form equivalence classes with the Catholic stimulus. Maxon, Keenan, and Hine (1993) used as stimuli female names, and occupations either female-stereotypic or male-stereotypic. A number of subjects were unable to form equivalence classes with female names and male-stereotypic jobs. The researchers speculated that the subjects did not respond equivalently because the experimentally trained relations between stimuli were in competition with previously established relations.

Plaud (1995, 1997) utilized the stimulus equivalence paradigm to study responding differences when fear-related words were contrasted with a fear-irrelevant stimulus category of words. Two stimulus conditions were studied. In the first condition, snake-related stimuli were tested; and in Condition 2 flower-related stimuli were examined. Stimuli in the two "meaning" groups were matched in terms of the number of letters in each stimulus, letter placement, and word frequency. In both studies Plaud found that there was a significant interaction between the formation of stimulus equivalence classes and anxiety-related stimulus sets. Plaud (1995) found that stimulus set performances were achieved more slowly for stimuli composed of fear-relevant stimuli. In other words, snake-related stimuli generally impeded the formation of stimulus equivalence classes. Plaud et al. (in press) also studied the relationship between the formation of stimulus equivalence classes and stimulus "meaning" using terms derived from Rational Emotive Behavior Therapy (REBT) such as "should" and "must." The researchers found that most subjects consistently biased their responding in favor of forming stimulus equivalence classes more slowly for REBT-related and non-REBT emotionally charged words.

These studies served as the empirical background to a research question concerning sexually explicit stimuli. Given the strong relationship of sexually charged words (or audio cues) to sexual arousal and behavior (Gaither & Plaud, 1997), can it be experimentally demonstrated that there is a natural relationship between sexually explicit words and the formation of stimulus equivalence classes? In other words, can specifically trained relationships among certain combinations of sexually explicit words lead to the formation of other relationships among such words when compared with ambiguously sexual words and neutral words? The stimulus equivalence paradigm allows us to study this proposed relationship directly. Given prior research conducted by Plaud (1995, 1997) and Plaud et al. (in press), it would seem logical to hypothesize that sexually explicit words would inhibit the formation of new relationships (i.e., without reinforcement) when testing for stimulus equivalence. It is hypothesized that sexually charged words will inhibit the formation of new relationships because they are very salient stimuli, and as the prior research of Plaud and Plaud et al. has demonstrated, salient stimuli in other domains (i.e., fear, emotional charge) inhibited the formation of new stimulus classes. This hypothesis will be investigated directly by comparing sexually explicit words with two other word groups: sexually ambiguous words and neutral words chosen from a prior study conducted by Gaither, Weller, and Plaud (1997) described below. If it can be reliably demonstrated that sexually charged words are related to or inhibit equivalence class formation, there will be important implications for behavioral models of the relationship of sexually related verbal stimuli on sexual arousal and behavior. For example, if sexually explicit words interfere with the formation of emergent relationships with other sexually explicit words, it may be more difficult to condition new responses to sexual stimuli when compared with nonsexual or neutral stimuli. It is to this basic behavioral investigation that we now turn.



As part of another study (Gaither, Weller, & Plaud, 1996), 200 undergraduate students in psychology courses at the University of North Dakota completed a packet of questionnaires in small groups. After completing the questionnaires, each student was given a sheet of paper that contained information regarding the current study and asked to indicate whether he or she would like to participate. Forty one students served as participants in the present study (13 males and 28 females). The participants ranged in age from 18 to 34 (M = 20.9, SD = 3.88), were all Caucasians, and predominantly single (n = 32, 78%). Regarding religious affiliation, 19 (46%) were Lutheran, 10 (24%) were Catholic, 7 (17%) were Protestant, and 5 (12%) considered themselves agnostic, atheist, or of no religious affiliation. Participants received 4 hours of research credit toward their psychology course.


Three sets of words served as the stimuli in the present study. Similar to the procedure reported by Plaud (1995), stimuli were chosen for the "meaning" categories [sexually explicit (SE), sexually ambiguous (SA), and a third category of neutral (Neut) words]. SE words consisted of the following stimuli: naked, condom, erotic, lover, passion, and genital. The SA words were: bush, cream, score, nuts, french, beaver. The second comparison stimulus class of neutral words was composed of the following stimuli: relish, idolize, feather, sailor, gutter, speech.

Stimuli were chosen from a list containing 400 words originally developed by Gaither et al. (1996). Two hundred participants were instructed to rate how sexual he or she found each word to be on a 5-point Likert scale (1 = not at all sexual, 5 = extremely sexual). Thirty of these words came from the Word Association Test (WAT; Galbraith & Mosher, 1968): These words have been found to have double meanings, one of which is sexual (e.g., "screw," "bust," and "prick"). The six words in each of the three stimulus categories were identified by choosing the words ranked in the highest range for sexuality with the lowest standard deviation (SE words), the middle range (SA words), and the low range with the highest standard deviation (Neut words). Stimulus word combinations, including means and standard deviations obtained for the stimuli from Gaither et al. (1996) are presented in Table 1.



Each participant was seated facing a computer monitor. Each subject was then given a code number entered via the use of standardized keyboard input. At the beginning of each subject's experimental session, the following message was displayed on the computer monitor:

Thank you for agreeing to participate as a subject in this experiment. PLEASE DO NOT TOUCH ANY OF THE KEYS ON THE KEYBOARD UNTIL INSTRUCTED TO DO SO. In this experiment you will be presented with many trials. Each trial contains three words that you will see on the screen. YOUR JOB IS TO FIGURE OUT WHICH WORDS GO TOGETHER.

In the first phase of the experiment, the computer screen will display instructions that tell you how to respond as well as labels that help you to identify which keys to press. After a while, these instructions will go away so that you may be able to participate in more trials. Your task will be to RESPOND CORRECTLY to the instructions by pressing the proper key on the computer keyboard. If you would like to take a break at any time, call for the experimenter, who will be in an adjoining room that you will be shown prior to beginning the experiment. When you are ready to start, press the ENTER key, and the experiment will be under way.

Phase 1. A trial-by-trial computer program utilizing ToolBook for Windows (Asymetrix, 1994) was implemented following the introductory instructional set. A message was shown on the computer screen instructing subjects to "Press Any Key to Begin." Once initiated by the subject a sample word appeared at the top of the screen with prompts to press the space bar (A nondifferential observing response; Fields et al., 1990). Once the space bar was pressed, two comparison words appeared below the sample word forming a triangle of words. The subject was then prompted to press the "1" key if the left comparison word was to be chosen, and the "2" key if the right comparison word was to be chosen. One comparison word was designated correct (C+) and one incorrect (C-) for a particular sample word.

If the subject chose the correct comparison stimulus, the computer screen displayed the phrase "You Were Right... Congratulation!" If the incorrect stimulus was chosen, the screen flashed the phrase "Wrong." After one of these two messages appeared, three additional letters were displayed at the bottom of the computer screen in the following order: RX-W. The subject was instructed via a screen prompt that if he/she was correct in the preceding trial then he/she should press the "R" key (for "Right"). If the subject was wrong, he/she was prompted to press the "W" key (for "Wrong"), and if no feedback was given by the computer (a condition implemented during later phases of the study), the subject was prompted to press the "X" key on the computer keyboard.

Twenty-four sample trials were presented to each subject, with all instructions and screen prompts, to familiarize the subjects with the procedure and also to fade out the screen prompt instructions. These practice stimuli are shown in Table 2. Each bogus sample stimulus was presented four times in a randomized order controlled for left and right C+/C- comparison stimulus locations. After the 24 trials were completed, the subjects were shown the following message:

The Actual Experiment Will Now Begin. After the Sample Word appears at the top of the screen, press the space bar to view the two other comparison words. If you think the left word is correct, press the "1" key. If you think the right word is correct, press the "2" key. If the computer tells you that you were right, press the "R" key. If the computer tells you that you were wrong, press the "W" key. If you get no message, press the "X" key to continue. Remember that your task will be to RESPOND CORRECTLY by pressing the proper key on the computer keyboard, and that YOUR JOB IS TO FIGURE OUT WHICH WORDS GO TOGETHER. Press any key to begin, and good luck! Descriptive feedback and a message to press any key to begin the next trial will be the only other screen messages presented after completion of the sample trial presentations.


Phase 2. Once each subject was allowed practice with the computer program in Phase 1, he/she completed participation in Phases 2-5 in three different stimulus conditions: the SE, SA, and Neut conditions (i.e., Phases 2-5 in SE condition, Phases 2-5 in SA condition, and Phases 2-5 in Neut condition). Therefore, all subjects participated in three distinct stimulus conditions, and did not move on to the second condition until the first condition Phase 5 was completed, and did not begin the third condition until the conclusion of testing in Phase 5 of the second stimulus condition. Order of the conditions (SE, SA, and Neut words) was determined randomly for each subject.

Each of the actual sample stimuli (SE, SA, and Neut conditions) were presented four times in a randomly determined order, controlling for the placement of C+ and C- (i.e., C+ was presented both on the right side and on the left side for each sample stimulus) for a total of 16 trials per block. A block of trials was repeated for each subject until 100% accuracy was obtained in responding during a trial block. As presented in Table 1, there were six stimulus words in each condition (SE, SA, and Neut): A1, A2, B1, B2, C1, and C2. When A1 was presented as the sample stimulus the subject's behavior of choosing the comparison stimulus B1 was reinforced, whereas choosing B2 was not reinforced, and when A2 was presented as the sample stimulus the subject's behavior of choosing B2 was reinforced, whereas picking B1 was not reinforced. In the SE condition, for example, this means that when the word "naked" (A1) was presented, the subject would be verbally reinforced only if she/he picked "condom" (B1) but not "erotic" (B2). Likewise, when the word "lover" (A2) was presented as a sample stimulus in the SE condition, choosing "erotic" (B2) was reinforced, and choosing "condom" (B1) was not reinforced. Refer to Table 2 for a detailed presentation of the stimulus word combinations used in this study.

Phase 3. Once the A1/A2 sample stimuli reached criterion responding in Phase 2 (i.e., 100% correct matching of stimulus words in each stimulus condition), the B1/B2 sample stimuli were trained using the same method employed during Phase 2, until each subject reached a responding level of 100% accuracy in a given block. Sixteen-trial blocks were used in this training session, controlling for placement location as in Phase 2. Refer to Table 2 for the stimulus word combinations. In the SA condition, for example, when the word "cream" (B1) was presented, the subject would be reinforced if she/he picked "french" (C1) but not "beaver" (C2). Likewise, when the word "score" (B2) was presented as a sample stimulus, choosing "beaver" (C2) was reinforced, and choosing "french" (C1) was not reinforced.

Phase 4. Once criterion responding was reached for both sets of sample stimuli (i.e., both A and B stimuli), the stimuli were then tested together in a randomized 1:1 ratio. In other words, Phase 4 was a combination of Phases 2 and 3 in order to ensure that subjects had reliably mastered the stimulus word combinations shown in Table 2. The number of trials per block remained at 16, however, each sample stimulus was reduced from four to two presentations, until 100% accuracy in responding was reached for aggregate sample. During the four phases just described, feedback reinforcement messages were progressively faded, from 100% during the sample stimuli to 75% for the A1/A2 trials in Phase 2. Informative feedback was further reduced to 50% for the B1/B2 trials during Phase 3, decreasing to 25% for the aggregate Phase 4 A/B trials. During Phase 5 no evaluative feedback was given.

Phase 5. Once criterion was established in Phases 2 to 4, symmetry and transitivity testing were initiated, in order to probe for the formation of stimulus equivalence classes. A block was composed of 24 trials. Phase 5 testing consisted of the following stimuli in randomized order, controlling for left/right responding: eight trials using original NB stimuli as in Phases 2 to 4, eight trials of transitivity testing (i.e., given A choose C, and given C choose A), and eight trials of symmetry testing (i.e., given B choose A, and given C choose B). Table 3 presents all the stimulus combinations in Phase 5. As previously stated, subjects received no informative feedback during this phase of the experiment. Therefore, this was the phase used to determine whether untrained or emergent relationships would be shown for the different stimulus conditions. For example, in the Neut word condition, if "relish" (A1) was shown as the sample stimulus and the subject picked "gutter" (C1), this is evidence of transitivity, because in Phases 2-4 this relationship was not trained, only A1[right arrow]B1 and B1[right arrow]C1 relationships were trained. Likewise, if the subject chose "relish" (A1) when shown "idolize" (B1) this would be evidence of symmetry, because the trained relationship A1[right arrow])B1 became B1[right arrow]A1 without any specific training for this relationship.
Table 3

Emergent Relations Testing Comparison Stimuli Combinations: Stimulus
Equivalence Testing

Sample Correct Comparison (C+) Incorrect Comparison (C-)

A1 C1 C2
A2 C2 C1
C1 A1 A2
C2 A2 A1

 Symmetry Testing
Sample Correct Comparison (C+) Incorrect Comparison (C-)

B1 A1 A2
B2 A2 A1
C1 B1 B2
C2 B2 B1

Once correct responding was shown on 100% of the trials in a block during Phase 5, the experiment was completed for that subject. However, if subjects failed to complete this part of the experiment in 1 hour, the experiment was terminated. Those who completed this part of the experiment were shown the following message:

The Experiment is now concluded for today. Please indicate to the experimenter that you have completed this phase of the experiment, and thank you for your participation.

Once the subject completed his/her participation in the five phases described above, he/she was invited back the following week in order to participate in the second and then the third stimulus class phases of the experiment.


The number of blocks to criterion responding for each phase (measured from the beginning of each phase) in each condition, the number of error responses in each phase, and reaction time (in milliseconds) from the presentation of the stimuli to the first keyboard press for each subject in the SE, SA, and Neut stimulus conditions served as the dependent variables in the present study. All subjects reached criterion responding during Phase 5 stimulus equivalence testing in each of the three stimulus conditions. Table 4 lists the number of blocks of trials each subject took to reach criterion responding in each phase of the SE, SA, and Neut stimulus conditions.

It was hypothesized that subjects would take more blocks to show the formal properties of stimulus equivalence classes in the SE stimulus condition compared to the SA and Neut stimulus conditions. This prediction was evaluated by conducting a series of 4 (Phase) x 3 (Condition) within-subjects multivariate analysis of variance (MANOVA). The dependent variables were number of blocks to criterion responding, number of errors, and reaction time for Phases 2 to 5 in each of the three stimulus conditions (Phase 5 being the test for the emergence of stimulus equivalences). Each of the three dependent variables will be analyzed separately.
Table 4

Mean Number of Blocks to Criterion Responding in Each Phase of Each

 Phase 2 Phase 3 Phase 4 Phase 5
Condition (A/B) (B/C) (A/B; B/C) (Emergent)

Sexually Explicit 1.88(b) 2.07(b) 1.46(b) 5.54(a,b)
Sexually Ambiguous 1.42 1.71 1.51 1.71(a)
Neutral 1.56 1.78 1.27 1.63(a)

Mean Number of Errors Per Block in Each Phase of Each Condition

 Phase 2 Phase 3 Phase 4 Phase 5
Condition (A/B) (B/C) (A/B; B/C) (Emergent)

Sexually Explicit 0.43(d) 0.70(d) 0.27(d) 2.46(c,d)
Sexually Ambiguous 0.24 0.38 0.29 0.35(c)
Neutral 0.26 0.41 0.10 0.30(c)

Mean Reaction Time (milliseconds) in Each Phase of Each Condition

 Phase 2 Phase 3 Phase 4 Phase 5
Condition (A/B) (B/C) (A/B; B/C) (Emergent)

Sexually Explicit 1633 1885 2094 2344(c)
Sexually Ambiguous 1674 1876 1911 2189
Neutral 1619 1969 1811 2031(c)

Note. Means that share the same superscript are significantly
different at p [less than] .05.

Number of Blocks to Criterion Responding in Phases 2 to 5 for SE, SA, and Neut Stimulus Conditions

The MANOVA revealed a significant difference among the three stimulus conditions on the number of blocks required for subjects to reach criterion responding in Phases 2 to 5 [Wilks' Lambda (8, 33) = 0.53, p [less than] .01], indicating that responding differences across phases were affected by stimulus condition. Data showing mean performance across all stimulus conditions and phases are presented in Table 4. This finding was further evaluated using Scheffe post hoc tests. The Scheffe tests indicated that subjects took significantly more blocks to reach criterion responding in Phase 5 (stimulus equivalence testing) for the SE words when compared with both the SA and Neut words, though the latter two were not significantly different from each other. One way analyses of variance (ANOVA) were also performed on each of the three stimulus conditions in order to determine whether there were differences within each stimulus condition between Phases 2 to 5. Results of the ANOVAs on each stimulus condition indicated that only for the SE words were there significant differences among the experimental phases [F(3, 120) = 10.63, p [less than] .001]. Scheffe post hoc analyses indicated that Phase 5 (stimulus equivalence testing) took significantly more blocks to reach criterion responding than all other phases, though Phases 2 to 4 did not differ from each other.

Number of Errors in Responding in Phases 2 to 5 for SE, SA, and Neut Stimulus Conditions

The MANOVA again revealed a significant difference among the three stimulus conditions on the number of errors in responding in Phases 2 to 5 [Wilks' Lambda (8, 33) = 0.55, p [less than] .01], indicating that there were differences in error responding across phases in the three stimulus conditions. Data showing mean performance across all stimulus conditions and phases are presented in Table 4. Scheffe tests indicated that subjects made significantly more errors in Phase 5 (stimulus equivalence testing) for the SE words when compared with both the SA and Neut words, though again the latter two were not significantly different from each other. One way analyses of variance (ANOVA) in each stimulus condition between Phases 2 to 5 indicated that only for the SE words were there significant differences among the experimental phases [F(3, 120) = 8.28, p [less than] .001] in the number of errors emitted by subjects. Scheffe post hoc analyses again indicated that Phase 5 (stimulus equivalence testing) took significantly more blocks to reach criterion responding than all other phases, though Phases 2 to 4 were not significantly different from each other.

Reaction Time in Responding in Phases 2 to 5 for SE, SA, and Neut Stimulus Conditions

The MANOVA did not reveal significant differences in reaction time in responding among the three stimulus conditions, although Scheffe tests indicated that subjects took significantly more time to respond to the SE words compared with the Neut words.


Stimulus equivalence is an approach to understanding how new relationships form among different stimuli in the absence of explicit reinforcement contingencies. Understanding stimulus equivalence processes allows the behavioral researcher to study the parameters under which new relationships among stimuli may emerge without training. Consider the following example:

A primate may learn to hide in a thicket given the presence of a lion. We could think of this as a conditional discrimination: given lion select a thicket over open savanna. The lion is a conditional discriminative stimulus, in the presence of which the thicket is a discriminative stimulus for an approach response such as running toward it. Such a contingency arrangement provides no grounds to suppose that reversing the functions of the stimuli involved will be reinforced. The value of approaching a thicket given a lion does not imply the value of approaching a lion given a thicket. "Given thicket approach lion" makes little sense (Hayes, 1991, p. 20).

For humans, at least, we do see the formation of interrelationships among stimuli without training. Take the example given above. Given the presence of a lion, the reinforced choice is running toward a thicket. Therefore, an explicit contingency is set between the lion (Stimulus A) and the thicket (Stimulus B). Suppose that when in a thicket, reinforcement is given for hiding behind one of the many trees grouped together. Therefore, given thicket (Stimulus B), choose a tree to hide behind (Stimulus C). Say the animal is in the open, and there is no thicket, just a savanna, and a lone tree. Given the relationships above, it would be expected that the primate would go for the tree over the savannah. This new relationship may have formed as a function of the above explicit contingencies, even though choosing tree in isolation was never itself explicitly reinforced.

An interesting extension of the stimulus equivalence paradigm is whether certain a priori defined classes of stimuli influence how equivalence relations are formed. For example, in the previous studies by Plaud (1995, 1997) and Plaud et al. (in press), it was found that fear-relevant stimuli and nonsexually explicit emotionally charged stimuli inhibited the formation of stimulus equivalence classes. If there was no effect for stimulus class, then it would be the case that equivalences would form equally among different stimulus classes using a modified matching-to-sample procedure to program stimulus combinations, as was employed in the present study and in all other studies of stimulus equivalence to date. In the present study it was shown that subjects took more blocks to reach criterion responding, made more errors, and took more time to respond in the sexually explicit stimulus condition when compared with sexually ambiguous and neutral words. What is the significance of this finding? It would appear, first, that not all stimuli are alike or prepotent in the formation of equivalence classes, which is important data in its own right. It is clear from the present results that sexually explicit words tend to make it much more difficult to form new relationships with other sexually explicit words. This finding has further implication for theories of human sexuality.

As a general theoretical model, the effects of sexually explicit stimuli on conditional discrimination (Phases 2 to 4) and stimulus equivalence (Phase 5) trials may be conceptualized as follows: During the first three phases of testing using a standard matching-to-sample format with informative feedback, the effect of sexually explicit words on criterion behavior acted like friction, providing an inhibitory force that was overridden by the positive forward momentum of the informative feedback for correct and incorrect choices. When this source of forward momentum was removed during Phase 5 (testing for the emergence of untrained relationships without explicit feedback), the force of the friction had a more pronounced effect on the formation of novel (untrained) relationships, as the results of the MANOVAs and post-hoc analyses have shown.

These results have direct implication for the ability of sexually explicit categories or "meanings" to generalize to other sexually explicit categories. Much of the studies involving the classical and operant conditioning of human sexual arousal are concerned with the transfer of function from one neutral or sexual domain with another. In the case of transfer from a neutral to a sexual domain, such as a study conducted by Plaud and Martini (in press) found that human males showed physiological responding to a picture of a penny jar that was contingently paired in a classical conditioning procedure with sexually explicit stimuli. This transfer of response from an unconditioned stimulus to a conditioned stimulus involved generalization from a sexually explicit stimulus to a nonsexually explicit stimulus. It is usually not possible to study this potential transfer from one sexually explicit stimulus to another because both stimuli by definition themselves initially elicit sexual responses. The present study allows for examination of this factor because arousal is not the dependent measure: Rather, transfer of function in the present case is measured through the emergence of new stimulus equivalence classes. What this means is that it is more difficult (in context of sexually ambiguous and neutral words) to transfer to one sexual stimulus the response made to another sexual stimulus. In other words, like anxiety-related stimuli, sexually explicit stimuli are not easily generalizable to other sexual stimuli, though not impossible given that eventually all subjects formed stimulus equivalence classes in the sexually explicit stimulus category. One clinical implication of these findings, for example, is that if someone emits sexual behavior (such as paraphilic behavior) towards one object or person, this behavior may not generalize as quickly to other behaviors or stimuli as we might expect given the parameters of stimulus equivalence and stimulus and response generalization. It may also mean that the relationship between certain sexual stimuli and behavior once established is a strong behavioral phenomenon, not as easily changed as other behaviors. These implications from the present results should be explored in future studies which link sexual stimuli directly to sexual behavior (rather than computer key depressions as studied here).

The results of the present study are a further step in analyzing the effects of stimulus "meaning" on the formation of stimulus equivalences, ultimately tying the behavioral results found in the present study to advances made in cognitive theories in this area. These results provide confirmatory evidence that when sexually explicit stimuli are used as sample and comparison stimuli, the formation of stimulus equivalence classes is significantly inhibited. Results of this study show that consideration should be given to the effect of stimulus "meaning" on the formation of stimulus equivalences. The phenomenon that emerges in these data is that sexually explicit words do not easily generalize to other sexually explicit words - they tend to retain their own uniqueness.

As Plaud (1995, 1997) and Plaud et al. (in press) have discussed in relation to fear-relevant responding, cognitive assessment techniques have been used to test for the effects of fear-relevant stimuli. For example, results of the modified Stroop procedure (Hope, Rapee, Heimberg, & Dombeck, 1990) have shown that subjects usually take longer to color-name fear-related stimuli in comparison with their neutral counterparts. This study has demonstrated a novel behaviorally based approach to the assessment and effects of stimulus "meaning" on behavior. The results obtained here are consistent with results obtained in Stroop-like procedures, however, the stimulus equivalence procedure is embedded within a behavioral paradigm, relating particular discriminative stimulus conditions to a specified pattern of responding results.

Further studies could measure the relationship of stimulus equivalence formation when sexually explicit and neutral words are used together to examine whether new relations are more readily formed when neutral stimuli are paired with sexually related words (similar to the experimental methodology used in classical conditioning studies of sexual arousal, as discussed above). Further studies addressing the influence of "meaning" on the formation of new relationships could also manipulate stimulus salience and uniqueness in order to investigate whether words, such as sexually explicit words, are perceived as more emotional as well as more salient when compared with nonsexual words or other classes of words such as anxiety stimuli. More complex forms of equivalence testing involving five or more member classes could also be investigated. Also, it would be important to test whether the sets of stimuli used here also differ on other more cognitive tests, such as the modified Stroop procedure, in order to detect whether the three groups of stimuli are processed in similar or different fashions. The present results strongly support the conclusion that stimulus "meaning" has relevance for research attempting to understand how humans form emergent relationships without explicit feedback for doing so.


ANNETT, J. M., & LESLIE, J. C. (1995). Stimulus equivalence classes involving olfactory stimuli. The Psychological Record, 45, 439-450.

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.

FIELDS, L., ADAMS, B. J., VERHAVE, T., & NEWMAN, S. (1990). Are stimuli in equivalence classes equally related to each other? The Psychological Record, 43, 85-105.

GAITHER, G. A., & PLAUD, J. J. (1997). The effects of secondary stimulus characteristics on male sexual arousal. Journal of Sex Research, 34, 231-236.

GAITHER, G., WELLER, L. A., & PLAUD, J. J. (1996). Gender differences in the sexual rating of words. Manuscript submitted for publication.

GALBRAITH, G. G., & MOSHER, D. L. (1968). Associative sexual responses in relation to sexual arousal, guilt, and external approval contingencies. Journal of Personality and Social Psychology, 10, 142-147.

HAYES, S.C. (1991). A relational control theory of stimulus equivalence. In L. J. Hayes & S.C. Hayes (Eds.), Dialogues on verbal behavior (pp. 19-40). Reno, NV: Context Press.

HAYES, L. J., THOMPSON, S., & HAYES, S.C. (1989). Stimulus equivalence and rule following. Journal of the Experimental Analysis of Behavior, 52, 275-291.

HAYES, L. J., TILLEY, K. J., & HAYES, S.C. (1988). Extending equivalence class membership to gustatory stimuli. The Psychological Record, 38, 473-482.

HOPE, D. A., RAPEE, R. M., HEIMBERG, R. G., & DOMBECK, M. J. (1990). Representations of the self in social phobia: Vulnerability to social threat. Cognitive Therapy and Research, 14, 177-189.

LAZAR, R. M., DAVIS-LANG, D., & SANCHEZ, L. (1989). The formation of visual stimulus equivalences in children. Journal of the Experimental Analysis of Behavior, 41, 251-266.

LESLIE, J. C., TIERNEY, K. J., ROBINSON, C. P., KEENAN, M., & WATT, A. (1993). Differences between clinically anxious and non-anxious subjects in a stimulus equivalence training task involving threat words. The Psychological Record, 43, 153-161.

MAXON, R D., KEENAN, M., & HINE, L. (1993). Gender-role stereotyping and stimulus equivalence. The Psychological Record, 43, 381-394.

PLAUD, J. J. (1995). The formation of stimulus equivalences: Fear-relevant versus fear-irrelevant stimulus classes. The Psychological Record, 45, 207-222.

PLAUD, J. J. (1997). Behavioral analysis of fear-related responding using a modified matching-to-sample procedure. Manuscript submitted for publication.

PLAUD, J. J., GAITHER, G. A., WELLER, L. A., BIGWOOD, S. J., BARTH, J., & FRANKLIN, M. (in press). Rational-Emotive Behavior Therapy and the formation of stimulus equivalence classes. Journal of Clinical Psychology.

PLAUD, J. J., & MARTINI, J. R. (in press). The respondent conditioning of sexual arousal. Behavior Modification.

PLAUD, J. J., & NEWBERRY, D. E. (1996). Rule-governed behavior and pedophilia. Sexual Abuse: A Journal of Research and Treatment, 8, 143-159.

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.

TIERNEY, K. J., DELARGY, P., & BRACKEN, M. (1995). Formation of an equivalence class incorporating haptic stimuli. The Psychological Record, 45,431-437.

WATT, A., KEENAN, M., BARNES, D., & CAIRNS, E. (1991). Social categorization and stimulus equivalence. The Psychological Record, 41, 33-41.
COPYRIGHT 1998 The Psychological Record
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1998 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Plaud, Joseph J.; Gaither, George A.; Franklin, MIchael; Weller, Louise A.; Barth, Jeannie
Publication:The Psychological Record
Date:Jan 1, 1998
Previous Article:Superstitious math performance: interactions between rules and scheduled contingencies.
Next Article:Stimulus familiarity and the delayed emergence of stimulus equivalence or consistent nonequivalence.

Related Articles
The effects of equivalence class structure on test performances.
The formation of stimulus equivalences: fear-relevant versus fear-irrelevant stimulus classes.
Establishing equivalence relations using a respondent-type training procedure.
Equivalence class formation using stimulus-pairing and yes-no responding.
Stimulus familiarity and the delayed emergence of stimulus equivalence or consistent nonequivalence.
Prior learning and equivalence class formation.
Stimulus equivalence and the blocking effect.
Competing contingencies over derived relational responding: a behavioral model of the implicit association test.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters