Printer Friendly

Identity and delay functions of meaningful stimuli: enhanced equivalence class formation.

Traditionally, meaningfulness of stimuli has been defined in terms of their denotative and connotative properties. The denotative properties of a stimulus are their defining features. For example, some of the denotative properties of a cat are that it has four legs, purrs, scratches furniture, and chases mice. The connotative properties of a stimulus are its emotive and judgmental features. For example, a cat is snuggly and makes you feel good, but can also be scary. Presumably, the depth of meaningfulness of a stimulus is directly related to the number or denotative and connotative terms with which it is associated.

Meaningful stimuli can also influence the learning of other behavior. For instance, the inclusion of a meaningful stimulus in a set of otherwise meaningless stimuli will influence the likelihood of those stimuli coming to function as members of a conceptual category called an equivalence class amongst these stimuli. An equivalence class is a finite set of perceptually disparate stimuli that initially are unrelated to each other, but through a minimal amount of training, come to be substitutable for each other. One example of this is the five words for the number 7, as written in Cyrillic, Chinese, Urdu, Korean, and Arabic. To most members of an English speaking society, before class formation, these stimuli would not be substitutable for each other; would not occasion the mutual selection of each other. After class formation, they would function interchangeably as witnessed by their mutual selection of each other.

The effect of meaningful stimuli on equivalence class formation has been studied using a variety of procedures, one of which is called the simultaneous protocol. In this protocol, all baseline relations are trained to mastery after which all derived relations probes are presented together (Nartey et al. in press). Class consistent responding evoked by all of these probes documents the formation of the equivalence class.

When the simultaneous protocol is used for training and testing, and all of the members of a potential equivalence class consist of meaningless stimuli, it is very unlikely that the stimuli will come to function as members of an equivalence class (Buffington et al. 1997). Indeed, it is the low basal yields produced using the simultaneous protocol that makes it a sensitive measure of variables that can enhance class formation.

Using the simultaneous protocol, when all of the members of a potential equivalence class were meaningful but unrelated to each other, there is a very high likelihood of forming a class among them if they have compatible connotative valences (e.g., Dickins et al. 1993). In contrast, much lower likelihoods of class formation were observed if the meaningful stimuli had connotatively contradictory valences (Grehan 1998; Leslie et al. 1993; Plaud 1995; Watt et al. 1991). Finally, an intermediate percentage of participants formed equivalence classes if only one of the stimuli in the set was meaningful (e.g., a picture) and the remaining stimuli were meaningless (e.g., Arntzen and Lian 2010; Fields et al. 2012). These class enhancement effects are typically attributed to the connotative/emotive properties of the meaningful stimulus, the stimuli with which the stimulus is related and the responses elicited by the stimulus (Travis et al. in press).

There is, however, an alternative account of the class-enhancing properties of meaningful stimuli, which is based on their acquired behavioral stimulus control functions (Fields et al. 2012; Tyndall et al. 2004). A meaningful stimulus can evoke the occurrence of any number of productive responses, such as saying or writing its name; thus, it would be functioning as a discriminative stimulus for each of these responses. A meaningful stimulus can occasion the selection of another stimulus such as a printed definitional statement or a printed statement of its emotive content. In each instance, the meaningful stimulus would be functioning as a member of at least one conditional discriminative relation. Finally, a meaningful stimulus can also be used interchangeably with many other stimuli; thus, it would be acting as a member of an equivalence class or a naturally occurring category composed of these stimuli. It follows, then, that some proportion of the class-enhancing effect of a meaningful stimulus could be accounted for by the presence of any combination of these stimulus control repertoires, apart from and/or in addition to their connotative or denotative properties.

Initial support for this view has been provided by Fields et a1. (2012). Using meaningless stimuli designated as A, B, C, D, and E, the experiment began with the establishment of the C stimuli in different sets as discriminanda in simple successive and simultaneous discrimination paradigms. Thereafter, ABCDE equivalence classes were formed by 50 % of participants. In contrast, the same classes were formed by 0 % of participants for whom the C stimuli had no discrimination training history. In addition, the use of familiar pictures as the C stimuli resulted in the formation of equivalence classes by 80 % of participants in a group. Fields et a1. (2012) concluded that the discriminative functions served by meaningful stimuli could account for a portion of their class-enhancing properties. In this experiment, the abstract stimuli that had acquired discriminative functions did not have either identifiable denotative or connotative properties.

Therefore, these properties could not be invoked to account for class enhancement.

Meaningful stimuli also serve conditional discriminative functions in natural settings. These conditional discriminative functions could have been in the form of identity or arbitrary relations. In an identity conditional relation, an individual is presented with a range of sample stimuli along with a set of comparison stimuli, one of which is identical to the sample stimulus; the selection of the comparison that is identical to the sample stimulus would document the presence of an identity conditional relation. Thus, if a meaningful stimulus such as pictures of a person's son or daughter were presented as sample stimuli, the selection of the identical picture would instantiate the presence of the identity conditional discrimination. Because previously established discriminative functions enhanced class formation (Fields et a1. 2012), and other functions can also be established prior to class formation, the present experiment sought to determine whether the pre-class formation establishment of identity conditional discriminations with an abstract stimulus designated as C, might also enhance the likelihood of forming equivalence classes when it is included in a set of other meaningless stimuli designated A, B, D, and E.

Identity conditional discriminations can be established on a simultaneous or a delayed basis.

In both cases, the acquisition of the conditional function has been studied in the context of trials administered in a matching--to-sample format (MTS). When establishing an identity conditional discrimination in the simultaneous format, one sample stimulus is presented along with at least two comparison stimuli, one of which is identical to the sample stimulus. Reinforcement is provided for the selection of the comparison that is the same as the sample stimulus.

This procedure is designated "Id-S-MTS". When established in the delayed format, the sample stimuli arc presented but the sample is eliminated either by the occurrence of an observing response or by the passage of time (in this case 6 s), before the presentation of the comparison stimuli, and the participant selects a comparison in the absence of the sample. This procedure is designated "Id-6sD-MTS".

Some research has shown that the likelihood of forming identity conditional discriminations is an inverse function of the duration of the delay interval, the time separating sample termination and comparison presentation (e.g., Berryman et al. 1963; Blough 1959). In addition, asymptotic accuracy of responding in the DMTS procedure has also been reported to be an inverse function of delay duration. On some occasions, however, terminal performances sometimes have shown to be enhanced by the occurrence of sample labeling behavior that occurs during the delay interval (Constantine and Sidman 1975).

Findings such as these, however, are not universal. In some cases, the opposite has been found; for example, Amtzen (2006) has reported that training of the baseline relations of an equivalence class using delayed matching enhanced the likelihood of class formation. In addition, Bortoloti and de Rose (2009, 2012); Bortoloti and De Rose (2012) reported increases in the "relatedness" of stimuli in an equivalence class when the baseline relations were trained using a delayed matching-to-sample trial format. Thus, the existing data do not provide a clear basis for predicting how histories of identity conditional discriminations formed on simultaneous or delayed bases will influence the likelihood of subsequent equivalence class formation.

The present experiment determined whether the prior establishment of an identity relation with one meaningless stimulus using simultaneous or delayed matching procedures would influence the likelihood of forming an equivalence class that contained that stimulus and other meaningless stimuli. in all four groups, participants attempted to form 3 three-node five-member equivalence classes that contained stimuli designated as A, B, C, D, and E. Training and testing were conducted under the simultaneous protocol. In three groups, the stimuli were abstract or meaningless shapes. In the fourth, the classes contained the same A, B, D, and E stimuli, but the C stimuli were familiar and meaningful pictures. In the ABS group, participants attempted to form the classes with five abstract stimuli, where there was no prior training in identity conditional discrimination training with any of the stimuli. In the PIC group, participants attempted to form classes with abstract A, B, D. and E stimuli and a meaningful picture as the C stimulus. In the Id-S-MTS and Id-6sD-MTS groups, identity conditional discriminations were formed with the C stimuli. Once acquired, the C stimuli were included in the ABCDE sets from which equivalence classes could emerge. A comparison of outcomes across groups documented how the prior formation of identity conditional relations in simultaneous and delayed MTS formats influenced subsequent likelihood of equivalence class formation. Finally, those results also clarified whether those functions could account for some portion of the class enhancing properties manifested by meaningful stimuli.



Forty students enrolled in the University of Ghana (13 men and 27 women), who varied in age from 19 to 31 years old and averaged 22.2 years, participated voluntarily in this study. None of these participants had any prior knowledge of stimulus equivalence research and methodology.

Seven other participants who started the experiment either quit or were dismissed because they did not acquire the baseline relations after 2 h of training.

In each case, a new participant was recruited to replace the person who did not complete the experiment. Participants were assigned on a block-randomized basis to one of four experimental conditions, ten participants in each group.

All participants had learned English from the start of elementary school, along with their native language.


Setting The experiment was conducted in the graduate seminar room of the Department of Psychology at the University of Ghana, Legon.

The seminar room measured approximately 5 m2 and was furnished with tables and chairs.

Hardware The experiments were conducted on an HP Compaq nc6320 laptop computer that used an 1828 MHz Intel Centrino [R]processor, and had a screen with a 16.8 in.

diagonal length and a 16 * 9 horizontal-to-vertical ratio. An external mouse was used by participants to control the position of the cursor throughout the experiment.

Software All sessions for training and testing of conditional discriminations for all participants were conducted with a software program version 3.12 made by Psych Fusion Software in collaboration with first author. The software controlled the presentation of all stimuli and also made recordings of data including the trial number, number of training trials, reaction time to sample and comparison stimuli, whether or not participants made the correct/incorrect comparison choice, and whether or not programmed consequences were delivered.

A summary of symmetry and equivalence tests as well as the duration of the experiment was also provided by the software.

Stimuli Stimulus sets used in the experiment are shown in Fig. 1. The two sections of the figure show stimuli used as members of the equivalence classes, which consist of 15 abstract and the three familiar picture stimuli. The C-stimuli were also used in the pretraining. The abstract stimuli were displayed in black and the picture stimuli in colors, all on a white background.


Small plastic-laminated pictures sized 3.8 [cm.sup.2] were made from the 15 abstract stimuli and the pictures to be used in the experiment.

The size of the touch sensitive areas on the screen was 9.4 cm * 3.4 cm.


Overview The experiment was conducted in six phases: Initial screening, pre-class formation categorization testing, instructions prior to training, pre-class formation identity conditional discrimination training, equivalence class formation, and post-class formation categorization test.

Each phase will be described below.

Initial screening Upon arrival at the experimental setting, participants were made to take a seat and given an informed consent document to read. In the consent document, participants were told that they were about to participate in an experiment in the field of behavior analysis, and that it would last approximately 1.5 h. They were also informed that they were required to respond to certain stimuli on the screen of a computer with mouse clicks and that there are no known harmful effects of participating in the study.

They were also told that they were free to withdraw from the experiment at any time without any negative consequences.

After reading, those who agreed by signing the forms began the experimental protocol.

Categorization test After signing the informed consent document, the participants remained seated in the experimental cubicle and were given the 15 plastic-laminated cards that corresponded to the stimuli to be used in the condition to which the participant was assigned and told to "put them into groups". Participants in the C-abstract, C-based Id-SMTS and Id-6sD-MTS conditions were presented with 15 abstract stimuli cards, while those in the meaningful C stimulus group were given 12 abstract stimuli and the three picture C stimuli. Finally, the categorization or sorting test was re-administered at the completion of the experiment.

Instructions prior to training Upon the completion of the categorization test, participants remained seated behind the computer and were presented with the following instructions on the computer screen:
  "In a moment a stimulus will appear in the middle of the screen.
  Click on this by using the computer mouse. Three stimuli will
  then appear in three corners of the screen. Choose one of them
  by clicking on it with the mouse. If you choose the stimulus
  we have defined as correct, words like "very good", "excellent",
  and so on will appear on the screen. If you press a wrong
  stimulus, the word "wrong" will appear on the screen. At
  the bottom of the screen, the number of correct responses
  you have made will be counted. During some stages of the
  experiment, the computer will NOT tell you if your choices
  are correct or wrong. Please do your best to get everything
  right. Thank you and good luck!"

No further instructions were given before and after the experiment started.

Trial structure and contingencies All participants were exposed to the simultaneous protocol to form equivalence classes, and some received identity conditional discrimination training. In all cases, however, all training and testing were done using trials presented in a matching--to-sample format that had the following parameters.

Each trial began with the presentation of the sample stimulus in the middle of the screen. Responding to the sample stimulus by a mouse click on it was immediately followed by the presentation of the three comparison stimuli at three of the corners of the screen, while the sample stimulus still remained on the screen. When the participant moved the mouse cursor to it and pressed the left button on the mouse, a comparison was selected. Correct responses, in the form of choosing the correct comparison stimulus according to the experimenter designated classes resulted in the removal of the sample and comparisons stimuli and the production of the words correct, very good, super, or excellent on the screen. Any other response produced the word wrong on the screen. If a programmed consequence followed a selection response, it was displayed in the middle of the screen for 1,000 ms. Termination of the programmcd consequences message was followed with a 500 ms inter-trial interval. Between trials, the mouse cursor was returned to the center of the screen.

Equivalence class ibrmation The simultaneous protocol (Buffington et al. 1997) was used to establish equivalence classes. This was a three-stage process, each of which consisted of blocks of trials. First, the baseline relations for the equivalence classes were trained in a concurrent manner until the achievement of a mastery criterion. Second, the baseline relations were maintained in the presence of blocks that contained a decreasing proportion of reinforced trials. Third, all of the baseline relations as well as all derived relations were presented randomly in one emergent relations testing block.

Acquisition of baseline trials All baseline conditional discriminations were trained according to a linear series training structure, and were presented concurrently. Thus, every trial type was trained from the beginning in one block. The trials presented were: A1/B1-62-B3, A2/B1-B2-B3, A3/B1-B2-B3, B1 /C1-C2-C3, B2/C1-C2-C3, B3/C1-C2-C3, C1 /D1-D2-D3, C2/D1-D2-D3, C3/D1-D2-D3, Dl/E1-E2-E3, D2/E1-E2-E3, and D3/E1-E2-E3 (the sample is the first letter in each string followed by the three comparisons; the correct comparison in each string is underlined). They were, however, randomly presented (See Table 1 for an overview of the experimental phases). Each of the trial types was presented three times with one from each class, consequently making it 36 trials per block. A mastery criterion of at least 90 % correct choices of comparisons was required to progress to the next stage of the experiment. If the participants did not reach this criterion, the block was repeated until the criterion was met. In this block, programmed consequences were scheduled for all comparison choices.
Table 1 This table provides an overview of the experimental phases
in the conditional discrimination training and testing

Experimental              Trial Types        %Programmed    Min.
Phases                                       Consequences   #of

Acquisition of       A1 B1, A2B2, A3B3,B1C1       100       36
baseline relations              ,B2C2, B3C3
All trial types            CID1,C2D2, C3D3,
presented                   D1E1.D2E2, D3E3

Maintenance:         A1B1, A2B2.A3B3.B1 C1,        75       36
Thinning of                      B2C2, B3C3
programmed                      C1D1, C2D2,
consequences           C3D3,D1E1,D2E2, D3E3
                     A1B1. A2B2.A3B3.B 1C1,        50       36
                                 B2C2, B3C3
                           C1D1, C2D2,C3D3,
                            D1E1,D2E2, D3E3
                    A1 B1. A2B2, A3B3.B 1C1        25       36
                                .B2C2. B3C3
                       C1D1, C2D2, C3D3,D 1
                              El,D2E2, D3E3
                     A1B1. A2B2, A3B3,B1C1,         0       36
                     C1D1, C2D2, C3D3.D 1E1
                                ,D2E2, D3E3

Test for derived            Baseline trial
relations           A1B1, A2B2, A3B3, B1C1,         0       36
All trial types      B2C2, B3C3 C1D1. C2D2,
presented            C3D3, D1E1, D2E2, D3E3

                            Symmetry trials
                    B1A1, B2A2, B3A3, C1B1,         0       36
                                 C2B2. C3B3
                    D1C1, D2C2, D3C3, E1D1,
                                 E2D2, E3D3

                              1 Node trials
                    A1C1, A2C2, A3C3, C1A1,
                                C2A2, C3A3,
                    B1D1, B2D2, B3D3, D1B1,         0       54
                                D2B2, D3B3,
                    C1E1, C2E2, C3E3, E1C1,
                                E2C2, E3C3,

                              2 Node trials
                    A1D1, A2D2, A3D3, D1A1,         0       36
                                D2A2, D3A3,
                    B1E1, B2E2, B3E3, E1B1,
                                 E2B2, E3B3

                              3 Node trials
                    A1E1, A2E2, A3E3, E1AI.         0       18
                                 E2A2, E3A3

Maintenance of baseline trials During the maintenance component, all phases were conducted in blocks of trials. The trial types presented here were the same as the trials presented in the acquisition block. Programmed consequences for trials in a block were presented on 75 %, 25 %, and finally 0 % of trials in a block, in that order. Reduction of programmed consequences occurred as soon as a participant responded with at least 90 % accuracy on all of the trials in a block at a given level. A block at a given programmed consequence level was repeated until it produced the criterion level of responding. For each level, the trials that produced programmed consequences were randomized in a block. When the participants reached the mastery criterion on the last block with no programmed consequences, the test for emergent relations was introduced.

Test fir emergence ofderived relations The last block with no programmed consequences was followed by an emergent relations test block containing 180 trials. These 180 trials comprised of 36 baseline trials; 36 symmetry trials; 54 one-node trials; 36 two-node trials; and 18 three-node trials. All of the trials were randomly presented and were without programmed consequences.

For purposes of analysis, the 180-trial test block was divided into two 90-trial test blocks, with test block 1 containing trials 1-90 and test block 2 containing trials 91-180. The formation of equivalence classes was defined by the selection of the comparisons that were consistent with experimenter-defined classes on at least 90 % of the trials in the block. Immediate emergence was defined as meeting mastery on test block 1, with maintenance assessed by performances on the second test block. Delayed emergence was defined as the selection of correct comparison in at least 90 % for the trials in test block 2. Failure to form classes was defined by the selection of class indicative comparisons on less than 90 % of the trials in each test block.

Training of identity conditional discriminations As mentioned above, participants in the Id-S-MTS and the Id-6sD-MTS conditions acquired C-based identity conditional discriminations before the administration of the simultaneous protocol used to establish the ABCDE equivalence classes. For the participants in the Id-S-MTS and the Id-6sD-MTS groups, the three C stimuli were presented as trials; Cl/C1-C2-C3, C2/C1-C2-C3, and C3/C1-C2-C3. Training was done in a minimum of three blocks with 100 % programmed consequences (See Table 2 for details). Three more blocks of fading of programed consequences (75 %, 25 %, and 0 [degrees]A)) were required to be completed before a test block brought the procedure to an end.
Table 2 Overivew of phases in the idenity simultaneous and
layed matching-to-sample training.

Experinmental Phases  Trial Types                   % of      Min.# of
                                                 Programmed    Trials

Acquisition of bascline trials (Randomized)

                       C1/C1-C2-C3.C2/C1-C2-C3      100          36

Maintenmece: Thimming of programmed cconsequences

                       C1/C1-C2-C3.C2/C1-C2-C3       75          18

                       C1/C1-C2-C3.C2/C1-C2-C3       25          18

                       C1/C1-C2-C3.C2/C1-C2-C3        0          72

Each programmed consequence lasted 1,000 ms and was followed by an inter-trial interval lasting 1,000 ms.

In each block, at least 90 % correct responses were required before participants could proceed to the next stage in the experiment.

Each block was repeated until the mastery criterion was attained, but if it is not attained after three blocks, the level of programmed consequences was increased to the previously used value in the next block.

In the Id-6sD-MTS group, it was arranged for a fixed 6 s delay between the offset of the sample and the onset of the comparisons.

Thus, with a mouse click on the sample stimulus in each trial, the sample disappeared 6 s before the onset of the comparison stimuli.


Pre-class formation sorting pelf-ormances Sorting performances for individual participants are shown in the left side of Table 3.

Prior to class formation training, no participants sorted the stimulus cards into piles that corresponded to the three experimenter-defined classes.

Instead, they sorted the cards into stacks from two to .seven stimuli, none of which contained all cards from the experimenter-defined classes. These "zero yields" are also presented as a summary value in the leftmost bar in each cluster in Fig. 2.

Equivalence class formation Figure 2 shows the percentage of participants in each group who did and did not show class indicative performances in the pre- and post-training sorting tests, and the formation of classes during the simultaneous protocol. Each cluster of three bars is for a different experimental condition.


All but two participants maintained the same number of responses according to the experimenter-defined classes in the first and second test halves. Therefore, some showed the immediate emergence of classes and their maintenance with test block repetition, and others did not show mastery levels of responding in either block, which documented the failure of class formation.

The two exceptions, Participants 4622 and 4619, responded differently in each test block. Specifically, each showed an increase in class consistent responding from the first to second test block, from 80 to 100 % of trials in a block that were in accordance with experimenter-defined classes. These performances then showed the delayed emergence of the classes. Therefore, overall yields (class formation regardless of delay) will be used to evaluate the effects of the independent variables on class formation.

During the simultaneous protocol (middle bar per cluster), when no pre-training was used (the ABS group), a very small proportion of participants formed the equivalence classes that contained abstract stimuli only. In contrast, 80 % of participants formed the equivalence classes when the C stimuli were meaningful pictures. Thus, the inclusion of a meaningful picture as a potential class member enhanced the likelihood of class formation.

When the Id-S-MTS procedure was used to establish the C-based identity conditional discriminations, no participants formed classes. Thus, the likelihood of class formation was similar, but even lower than that found when classes were formed with no prior training. The prior training of simultaneous identity conditional relations with the C stimuli, therefore, either did not influence, or could have suppressed the likelihood of the subsequent formation of equivalence classes.

For the Id-6sD-MTS group, where the C-based identity conditional discriminations were established on a delayed basis, yields were much greater than that obtained after Id-S-MTS training. In addition, the yields approximated those produced when the classes contained meaningful picture as the C stimuli.

Fisher's Exact Test (FET) showed that the yield obtained in the PIC condition was significantly different from the yields produced by the ABS (p=0.0001) and the zero-delay condition (p=0.0001). In addition, the yields produced by the SIM and 6 s conditions were significantly different from each other (FET p=0.011). The difference in yields produced by the ABS and 6 s conditions approached significance by FET p=0.057, and was significant by Chi Square = 0.019.

Post-class formation sorting perfbrmances The rightmost bar per cluster shows the results of the sorting tests conducted after the administration of the simultaneous protocol (see the right side of Table 3). The yields in the ABS, Id-S-MTS, and Id-6sD-MTS groups were the same as those seen during the simultaneous protocol. In the PIC condition, however, all of the participants who formed classes under the simultaneous protocol continued to do so in the sorting test, and one more participant (4637) who did not show class formation in the derived relations tests subsequently did so in the sorting test. Thus, the data obtained from 39 of the 40 participants in the experiment showed perfect concordance between the performances evoked by the derived relations and the sorting tests.
Table 3 pre- and Post-Sorting Test Outcomes

                 Cls-1  Cls-2  Cls-3  Cls-1  Cls-2  Cls-3  Cls-1  Cls-2

           4610    4      0      0      1      3      5      0      2
           4600    1      0      2      1      1      0      1      1
           4619    2      0      1      1      0      2      1      1
           4609    1      2      0      1      1      1      1      0
           4616    2      1      0      l      0      2      2      1
Abstract   4633    2      2      1      2      0      3      1      3
           4604    3      1      0      1      0      1      1      2
           4615    3      0      0      1      1      2      1      1
           4644    4      1      0      1      1      1      0      2

           4640    1      2      0      2      1      1      1      0
           4620    3      0      2      1      1      1      1      2
           4639    2      1      0      1      0      2      1      1
           4621    3      l      1      1      1      1      1      3
           4606    2      0      1      1      1      1      2      1
C as PKs   4603    3      l      0      1      1      1      1      1
           4611    3      l      1      1      1      1      1      3
           4602    4      0      0      1      I      1      0      4
           4622    3      1      0      1      1      1      1      1
           4637    1      0      1      1      1      0      1      1
           4611    1      2      1      2      1      1      1      1

           4613    4      0      4      1      5      1
           4612    3      0      1      2      2      1      0      3
           4624    2      1      2      2      2      1      1      2
Abs C      4605    2      0      1      3      0      0      0      3
ld-6s-MTS  4614    3      0      1      1      1      1      1      1
           4626    3      0      0      1      2      2      1      1
           4607    1      0      1      2      1      2      1      2
           4608    3      l      0      1      0      2      1      2
           4646    2      0      1      1      1      0      1      1
           4642    2      0      1      1      1      1      1      1

           4628    2      1      0      1      0      1      2      1
           4630    3      1      1      2      4      4
Abs C      4632    4      2      1      1      1      3      0      2
ld-S-MTS   4629    1      0      1      2      1      0      1      1
           4631    2      0      1      1      0      1      1      1
           4634    4      2      1      1      3      4
           4635    2      1      2      3      1      0      0      3
           4636    2      2      1      1      1      3      2      2
           4643    4      1      1      1      4      4
           4645    2      0      1      1      1      1      2      1

                 Cls-3  Cls-1  Cls-2  Cls-3  Cls-1  Cls-2  Cls-3  Cls-1

           4610    0
           4600    1      1      1      0      1      2      2
           4619    0      1      3      0      0      1      2
           4609    2      2      0      1      0      2      1
           4616    0      0      2      1      0      1      2
Abstract   4633    1
           4604    2      0      2      2
           4615    0      0      2      1      0      1      2
           4644    2      0      1      2

           4640    1      1      1      1      0      1      2
           4620    1      0      2      1
           4639    1      1      3      2
           4621    3
           4606    0      0      2      1      0      1      2
C as PKs   4603    1      0      2      3
           4611    3
           4602    4
           4622    0      0      2      2      0      0      2
           4637    0      1      1      0      1      1      0      0
           4611    1      1      :      2

           4612    3
           4624    2
Abs C      4605    0      0      1      2      0      1      2
ld-6s-MTS  4614    2      0      3      1
           4626    2      0      2      1
           4607    1      1      2      0      0      0      1
           4608    3      0      2      0
           4646    0      1      0      1      0      1      1      0
           4642    0      1      0      2      0      2      0      0

           4628    0      0      1      1      0      1      2      0
Abs C      4632    1
ld-S-MTS   4629    0      1      1      1      0      2      1      0
           4631    2      1      4      1
           4635    3
           4636    1
           4645    0      0      1      1      0      1      1      0

                 Cls-2  Cls-3  Cls-1  Cls-2  Cls-3  Cls-1  Cls-2

           4610                                       5      0
           4600                                       3      0
           4619                                       5      0
           4609                                       2      0
           4616                                       1      0
Abstract   4633                                       2      1
           4604                                       2      1
           4615                                       2      0
           4644                                       1      0
           4640                                       2      1

           4620                                       5      0
           4639                                       5      0
           4621                                              0
           4606                                       5      0
C as PKs   4603                                       5      0
           4611                                       5      0
           4602                                       5      0
           4622                                       5      0
           4637    l      1      0      0      3      5      0
           4611                                       3      0

           4613                                       5      0
           4612                                       5      0
           4624                                       5      0
Abs C      4605                                       5      0
ld-6s-MTS  4614                                       5      0
           4626                                       5      0
           4607                                       2      1
           4608                                       3      0
           4646    l      1      0      1      1      4      0
           4642    l      1                           5      0

           4628    1      1                           3      0
           4630                                       3      0
Abs C      4632                                       2      0
ld-S-MTS   4629    0      2                           2      0
           4631                                       4      0
           4634                                       3      1
           4635                                       3      2
           4636                                       3      1
           4643                                       3      2
           4645    1      1                           2      3

                 Cls-3  Cls-1  Cls-2  Cls-3  Cls-1  Cls-2  Cls-3  Cls-1

           4610    0      0      5      0      0      0      5
           4600    0      2      0      0      0      3      0      0
           4619    0      0      5      0      0      0      5
           4609    1      1      0      2      1      1      1      1
           4616    5      4      0      0      0      3      0      0
Abstract   4633    2      2      2      1      1      2      2
           4604    0      3      0      1      0      4      0      0
           4615    2      2      2      0      1      0      3      0
           4644    2      3      0      1      1      1      2      0
           4640    0      2      0      3      1      1      1      0

           4620    0      0      5      0      0      0      5
           4639    0      0      5      0      0      0      5
           4621    0      0      5      0      0      0      5
           4606    0      0      5      0      0      0      5
C as PKs   4603    0      0      5      0      0      0      5
           4611    0      0      5      0      0      0      5
           4602    0      0      5      0      0      0      5
           4622    0      0      5      0      0      0      5
           4637    0      0      5      0      0      0      5
           4611    2      0      5      0      2      0      3

           4613    0      0      5      0      0      0      5
           4612    0      0      5      0      0      0      5
           4624    0      0      5      0      0      0      5
Abs C      4605    0      0      5      0      0      0      5
ld-6s-MTS  4614    0      0      5      0      0      0      5
           4626    0      0      5      0      0      0      5
           4607    0      1      1      1      2      0      1      0
           4608    0      2      0      0      0      2      1      0
           4646    0      1      1      0      0      4      0      0
           4642    0      0      4      1      0      1      4

           4628    0      2      0      0      0      2      0      0
           4630    0      2      0      0      0      2      1      0
Abs C      4632    0      3      1      0      0      3      0      0
ld-S-MTS   4629    0      2      0      1      1      0      1      0
           4631    0      1      3      2      0      2      3
           4634    0      2      0      0      0      4      0      0
           4635    0      2      3      0      0      0      5
           4636    0      2      0      0      0      4      0      0
           4643    2      2      3      3
           4645    2      3      2      0      0      0      3

                 Cls-2  Cls-3  Cls-1  Cls-2  Cls-3  Cls-1  Cls-2  Cls-3

           4600    2      1      0      0      4
           4609    2      0      0      2      1
           4616    2      0
Abstract   4633
           4604    0      4
           4615    3      0
           4644    0      4
           4640    2      0      0      1      1

C as PKs   4603

Abs C      4605
ld-6s-MTS  4614
           4607    3      0      0      0      3
           4608    3      0      0      0      2      0      0      2
           4646    0      5

           4628    2      0      0      1      1      0      0      4
           4630    3      0      0      0      2      0      0      2
Abs C      4632    1      3      0      0      2
ld-S-MTS   4629    4      0      0      1      1      0      0      2
           4634    0      5
           4636    0      5

           4610  YES  YES
           4600  NO   NO
           4619  NO   NO
           4609  NO   NO
           4616  NO   NO
Abstract   4633  NO   NO
           4604  NO   NO
           4615  NO   NO
           4644  NO   NO
           4640  NO   NO

           4620  YES  YES
           4639  YES  YES
           4621  YES  YES
           4606  YES  YES
C as PKs   4603  YES  YES
           4611  YES  YES
           4602  YES  YES
           4622  YES  YES
           4637  NO   YES
           4611  NO   NO

           4613  YES  YES
           4612  YES  YES
           4624  YES  YES
Abs C      4605  YES  YES
ld-6s-MTS  4614  YES  YES
           4626  YES  YES
           4607  NO   NO
           4608  NO   NO
           4646  NO   NO
           4642  NO   NO

           4628  NO   NO
           4630  NO   NO
Abs C      4632  NO   NO
ld-S-MTS   4629  NO   NO
           4631  NO   NO
           4634  NO   NO
           4635  NO   NO
           4636  NO   NO
           4643  NO   NO
           4645  NO   NO

Note: The first column indicates the experimental conditions
studied in the experiment. Each row indicates data for a
given participant. The table consists of boxes each of which
has three cells. Each cell contains information for
experimenter-defined Classes 1,2, or 3. The entries in each
cell indicate the number of stimuli in each
experimenter-defined class that the participant assigned to
one subject-defined class. Thus, each box designates the
content of a subject-defined class. Participants were
separated into those who did and did not form equivalence
classes (ECE) based on emergent relations test performances,
as indicated by Y and N in the next to the right-most
column. The right-most column indicates the presence of
sorting clusters that correspond to the presence of all
three experimenter-defined classes. ECF = Equivalence Class
Formation, ABS - abstract stimuli, Id-S-MTS - identity
simultaneous matching-to-sample, Id-6sL)-MTS = identity 6
s delayed matching-to-sample, and PIC - Pictures

Acquisition of baseline relations and class lbrmation In addition to influencing the likelihood of equivalence class formation, the experimental conditions could also influence the acquisition of the baseline relations for the equivalence classes. This possibility was evaluated in the upper panel of' Fig. 3, which depicts the median/average number of trials needed to acquire the baseline relations in each group, for participants who did and did not form classes.

Acquisition of the baseline relations was significantly faster for participants who formed classes than for those who did not, t (38)=-3.861, p<0.05. Across all conditions, then, acquisition speed of the baseline relations was inversely correlated with class formation.

For the ABS, PIC, and Id-6sD-MTS groups, similar rates of acquisition occurred for those who then formed classes. In addition, slower but similar rates of acquisition occurred for the participants who did not form classes. A comparison of the data in the two right most pairs of bars for participants who did not form classes, indicated that the acquisition of the baseline relations occurred more rapidly after preliminary training with a delay than without a delay. In addition, the number of trials needed to acquire the baseline relations after the preliminary establishment of the identity conditional relations was far greater than when observed after any other condition, regardless of subsequent class formation.

The lower panel of Fig. 3 depicts the percentage of correct trials during the emergent relations tests blocks for participants who did not form classes in each experimental condition. The percentage of errors was greater in the ABS and Id-S-MTS conditions than for the PIC and Id-6sD-MTS conditions. A one-way ANOVA showed that the Id-S-MTS and the ABS groups had a significantly higher likelihood of making errors during the derived relations tests when compared to the PIC and the Id-6sD-MTS groups (F =11.65, p= 0.0006).


Discussion Equivalence classes were formed by 10 % of the participants when all of the class members were abstract stimuli (ABS group), and by 80 % of participants when the middle node was a familiar picture and the other class members were abstract stimuli (PIC group). These findings replicate those reported by Fields et al. (2012), and show that the inclusion of a meaningful stimulus enhances equivalence class formation.

In addition, 60 % of participants formed equivalence classes with sets of abstract stimuli when the middle nodes had previously become members of identity conditional discriminations that were established on a delayed basis (Id-6sD-MTS group). Further, when the identity function was established with no delay, 0 % of the participants formed the equivalence classes. Thus, the delay that separated the offset of sample stimuli and the onset of the comparison stimuli in identity conditional discrimination training played a critical role in the enhancement of equivalence class formation. By implication, these results support the view that some proportion of the class-enhancing properties of meaningful stimuli can be attributed to their presumed membership in identity relations established on a delayed basis.

Fields et al. (2012) and Travis et al. (in press) have shown that likelihood of equivalence class formation increased by the prior acquisition of simultaneous and successive discriminative functions. In addition, Travis et al. (in press) showed that class enhancement was a positive linear function of the amount of pre-class formation overtraining of simple successive discriminations. The present experiment showed that the establishment of identity conditional relations on a delayed basis also enhanced the likelihood of subsequent equivalence class formation. Taken together, all of these data support the more general view that enhancement of class formation produced by the inclusion of a meaningful stimulus in that class can be accounted for by the many stimulus control functions served by the meaningful stimulus as well as, or in addition to, their connotative and denotative properties.

Effects of identity and/or delay functions The results of the present experiment demonstrated that class enhancement was directly related to the delay that separated the termination of sample stimuli and the presentation of the comparison stimuli in identity conditional relations. Thus, the enhancement effect could be due to the delay function itself, or the delay function as constrained by the context of identity conditional discriminations. The relative influence of the type of conditional relation alone, however, cannot be discerned from the present experiment. The isolated and interactive effects of these two variables can be determined with additional research in which delay value and type of conditional relations are varied in a factorial manner.

Specifically, conditional relations can be identity based, where the samples and positive comparisons are the same (Z

[right arrow] Z) or arbitrarily based where the samples and positive comparisons are perceptually disparate (Z [right arrow] X).

Each of these relations can be established using simultaneous or delayed matching trial formats. If the same delay effects are found with identity and arbitrary conditional relations, the delay function would be responsible for class enhancement and not type of conditional relation. If different delay effects on class enhancement were to be found with the two types of conditional relation, both delay and type of conditional relation would be responsible for class enhancement. The validity of either of these conclusions awaits the conduct of such an experiment. In the present experiment, the abstract stimulus that acquired conditional control before class formation was subsequently used as the middle node in three-node five-member equivalence classes. Recent research (e.g., Anitzen, 2004; Fields et a1., 2012) has shown that the position of a meaningful stimulus in an equivalence class, as well as its temporal order of inclusion in training, substantially influence class enhancement of stimuli. Thus it is possible that abstract stimuli that acquire identity functions can have a similar effect. The evaluation of these suppositions awaits future research.

Quantification of the delay effect The present experiment found that a 6s delay had a major effect on class enhancement, while no delay did not produce any class enhancement. To what extent then will the likelihood and magnitude of class enhancement depend on the duration of the delay interval? Experiments on working memory typically find that delays greater than 3 s result in greater recall (e.g., Baddeley 2007), presumably because participants have time to encode stimuli and imbed their representations in long term memory for use in recall. By implication, enhancement should be substantial for delays of as little as 3 s and then decline markedly below that value. Thus, the effects of the 0 and 6 s delays would be placed in a parametric context that also make contact with the mechanisms of memory and relate them to the likelihood of equivalence class formation.

As noted in the results, after the establishment of simultaneous identity relations in preliminary training, no participants formed equivalence classes. Indeed, that yield was even lower than the already low yield obtained in the ABS condition (0 % vs. 20 % yields, respectively). The fact that none of the participants had a history of forming simultaneous identity relations does not mean that these individuals never responded correctly during the derived relations tests. On average, they selected class consistent comparison stimuli on 52 % of the derived relations probes.

Furthermore, likelihood of errors in the derived relations tests were greater after the formation of identity relations on a simultaneous instead of a delayed basis. Thus, the establishment of identity relations with the C stimuli in the simultaneous condition had an effect both on the speed of acquisition of the baseline condition discriminations and the errors made during the derived relations tests.

The apparent suppression of yield might have been driven in the following manner. When forming simultaneous identify relations, a sample stimulus is presented with one comparison that is identical to the sample, and another that is not. Reinforcement depends on selection of the comparison that is the same as the sample stimulus rather than the selection of the comparison that differs from the sample stimulus. In contrast, the baseline relations for the equivalence classes are arbitrary conditional relations where both comparisons differ from the prevailing sample stimulus; thus, reinforcement depends on the selection of a comparison that differs from a sample stimulus. This stimulus control repertoire then is in conflict with the stimulus control repertoire established during identity training. Overcoming that repertoire would be reflected in many more trials than would be needed to form the baseline relations for the classes, as indicated by the trials to acquisition data. In addition, the processing activity involved in overcoming the identity training might have interfered with attending to the relations among the stimuli in different baseline relations, which would obviate the emergence of derived relations, and result in the absence of equivalence class formation.

The effects of delay and coding Typical studies show slower or less likely formation of conditional discriminations with increasing delay (White 2013).

In the present experiment, the opposite occurred with the acquisition of the baseline relations. In addition, the likelihood of equivalence class formation was a direct function of delay duration in pre-class formation identity conditional discrimination training.

Both of these findings may be related to the emergence of coding (Urcuioli 2013) or mediating behavior in the delay interval. As has been argued by Amtzen (2006), the DMTS procedure encourages the emergence of mediating behavior that bridges the gap between sample offset and comparison onset and leads to the more likely selection of the correct comparisons in DMTS trials.

In the context of the present experiment, such a repertoire could have been induced during the establishment of the C-based identity conditional discriminations in the delay condition, after which they would generalize to the emergent relations tests, where the mediating behavior would enhance the coding of the sample stimuli in the derived relations test trials set and facilitate the selection of a comparison from the same class.

That possibility would also increase the likelihood of class formation. By contrast, when establishing the C-based identity conditional discriminations in the simultaneous matching-to-sample context, both sample and comparisons were presented at the same time, so there is no need for response mediation to solve the identity problem. Because a response mediation repertoire did not occur during identity training, that repertoire was not available to generalize to the stimuli in the emergent relations tests, and enhance the emergence of the classes.

Connotative properties and conditional discriminations Traditionally, the class-enhancing effect of including a meaningful stimulus as a member of a to-be-formed equivalence class has been attributed to the connotative properties of the meaningful stimulus. Connotative property is defined by the stimuli that have come to be associated with the meaningful stimulus, i.e., the stimulus-stimulus relations of which the meaningful stimulus is a member. In such an account, it is assumed that the connotative property has been established pre-experimentally, and by procedures unknown to the present researchers.

In the present experiment, a meaningless stimulus (C) became a member of the simplest form of stimulus-stimulus relation--an identity conditional discrimination (C [right arrow] C)--by an explicitly defined pre-class formation training procedure. It follows, then, that the C stimulus acquired a minimal connotative property. Thus, the enhancement of class formation by the inclusion of a nominally meaningless stimulus that previously acquired a delayed identity function is akin to saying the enhancement of class formation can be attributed, in part, to an explicitly produced connotative property of that stimulus. Conversely, the class enhancement effect can be attributed, in part, to the procedures used to establish the identity conditional discrimination established on a delayed basis. Regardless of the phraseology, the results of the present experiment extend our knowledge of procedures that enhance the formation of equivalence classes when training and testing are conducted using the simultaneous protocol, a protocol that typically does not support high likelihoods of equivalence class formation, and also provides an account of the behavioral stimulus control functions served by meaningful stimuli that are partially responsible for the enhancement of equivalence class formation that contain such stimuli.

Sorting pedbrmances After exposure to the simultaneous protocol, the participants who formed classes then showed class indicative performances in the sorting test, but not so for those who did not form classes. These results documented the maintenance of the classes in a different test format, replicated Fields et a1. (in press), and complemented related findings reported by Bortoloti and de Rose (2009, 2012). This concordance of tests results in the present experiment was obtained under all experimental conditions.

Therefore, prior identity conditional discrimination training and the inclusion of pictorial stimuli in equivalence classes did not influence the concordance of the two test performances.

Summary The inclusion of a meaningful stimulus in a set of abstract stimuli enhanced the likelihood of forming an equivalence class with the set. A similar effect was obtained when classes were formed with all abstract stimuli after one of them came to function as a member of an identity conditional discrimination established on a delayed basis. No enhancement was obtained when the identity relation was established on a simultaneous basis. By implication, the enhancement of class formation by the inclusion of a meaningful stimulus can be driven by the presumed delayed identity function served by the meaningful stimulus.

Conflict of Interest There is no conflict of interest to declare concerning the three authors.

Part of this paper was presented in a symposium at the 10th International Congress on Behavior Studies, Lisbon, Portugal.


Amtzen, E. (2004). Probability of equivalence formation: Familiar stimuli and training sequence.

The Psychological Record, 54, 275-291.

Retrieved from

Amen, E. (2006). Delayed matching to sample and stimulus equivalence: Probability of responding in accord with equivalence, as a function of different delays.

The Psychological Record, 56, 135-167. Retrieved from

Amtzen, E., & Lian, T. (2010). Trained and derived relations with pictures as nodes. The Psychological Record, 60, 659-677. Retrieved from

Baddeley, A. (2007). Working memory, thought, and action. New York, NY: Oxford University Press, Berryman, R., Cumming, W. W., & Nevin, J. A. (1963). Acquisition of Delayed Matching in Pigeon.

Journal of the Experimental Analysis of Behavior 6, 101.

doi:10.1901/jeab.1963.6-101. Retrieved from ISI:A19634769B00003.

Blough, D. S. (1959). Delayed matching in the pigeon. Journal of the Experimental Analysis of Behavior 2,151-160. doi:10.1901/jeab. 1959.2-151.

Bortoloti, R., & de Rose, J. C. (2009).

Assessment of the relatedness of equivalent stimuli through a semantic differential.

The Psychological Record, 59, 563-590. Retrieved from

Bortoloti, R., & De Rose, J. C. (2012).

Equivalent stimuli are more strongly related after training with delayed matching than after simultaneous matching: A study using the implicit relational assement procedure RAP). The Psychological Record, 62,41-54. Retrieved from

Buffington, D. M., Fields, L., & Adams, B. J. (1997).

Enhancing equivalence class formation by pretraining of other equivalence classes. The Psychological Record, 47, 69-96.

Retrieved from

Constantine, B., & Sidman, M. (1975). Role of naming in delayed matching-to-sample. American Journal of Mental Deficiency, 79, 680-689. Retrieved from http://www.ncbi.nlm.nih.govientrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt=citation&list_uids=3799735.

Dickins, D. W., Bentall, R. R, & 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 associates between names. The Psychological Record, 43, 713-724.

Fields, L., Arntzen, E., Nartey, R. K., & Eilifsen, C. (2012). Effects of a meaningful, a discriminative, and a meaningless stimulus on equivalence class formation. Journal of the Experimental Analysis of Behavior; 97, 163-181. doi:10.1 901/jeab.2012.97-1 63.

Fields, L., Arntzen, E., & Moksness, M. (in press).

Stimulus Sorting: A quick and sensitive index of equivalence class formation. The Psychological Record.

Grehan, P. M. (1998). Depressed subjects' formation of mood congruent and incongruent equivalence relations. Unpublished doctoral dissertation. Hofstra University, Hempstead, New York.

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 works.

The Psychological Record, 43, 153-161. Retrieved from

Nartey, R. K., Arntzen, E., & Fields, L. (in press). The functions of meaningful stimuli disclosed by equivalence class formation:

conditional discriminative functions. The Psychological Record.

Retrieved from

Plaud, J. J. (1995).

The formation of stimulus equivalences: Fear-relevant versus fear-irrelevant stimulus classes. The Psychological Record, 45, 207-222. Retrieved from

Travis, R. W., Fields, L., & Annzen, E. (in press). Discriminative functions and over-training as class-enhancing determinants of meaningful stimuli. Journal of the Experimental Analysis of Behavior.

Tyndall, 1. T., Roche, B., & James, J. E. (2004). The relation between stimulus function and equivalence class formation. Journal of the Experimental Analysis of Behavior 81,257-266. doi:10.1901/jeab.2004.81-257.

Urcuioli, P. (2013). Stimulus control and stimulus class formation. In G. Madden, W. V. Dube, T. D. Hackenberg, G. P. Hanky, & K. A. Lanai (Eds.), APA Handbook of Behavior Analysis (Vol. 1, pp. 361386). Washington, DC: American Psychological Association.

Watt, A., Keenan, M., Barnes, D., & Cairns, E. (1991). Social categorization and stimulus equivalence. The Psychological Record, 41, 33-50. Retrieved from

White, K. G. (2013).

Remembering and forgetting. In G. Madden, W. V. Dube, T. D. Hackenberg, G. P. Hanley, & K. A. Lattal (Eds.), APA Handbook of Behavior Analysis (Vol. 1, pp. 411-437).

Washington, DC: American Psychological Association.

Erik Arntzen * Richard K. Nartey * Lanny Fields

E. Amtzen * R. K. Nartey

Oslo and Akershus University College, Oslo, Norway

L. Fields

Queens College, The Graduate School of the City University of New York, Flushing, USA

E. Arntzen (E)

Department for Behavioral Science, Oslo and Akershus University College, PO Box 4, St. Olavs Plass, 0130 Oslo, Norway

COPYRIGHT 2014 The Psychological Record
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Arntzen, Erik; Nartey, Richard K.; Fields, Lanny
Publication:The Psychological Record
Article Type:Report
Date:Sep 1, 2014
Previous Article:Motivation Behavior Analysis A Critique.
Next Article:Local response rates on variable-ratio schedules of wheel-running reinforcement are relatively invariant compared to sucrose.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters