Acquisition of Children’s Relational Responding: The Role of the Intradimensional and Interdimensional Abstract Tact and the Autoclitic Frame

Abstract

The acquisition of verbal behavior is complex and requires the analysis of myriad variables. Ernst Moerk estimated that by the time a child has reached 4 years of age they have experienced nearly 9 million language learning trials with mothers using at least 14 categories of maternal teaching interactions. The interactions provide a foundation for children learning the tact, mand, echoic, intraverbal, autoclitic, and other relations, described by Skinner in Verbal Behavior. Here we examine two relations that have been overlooked to some extent and arguably account for many of the generative features of verbal behavior and shared meaning: the abstract tact, or more precisely the interdimensional abstract tact, and the autoclitic frame. We describe Goldiamond’s treatment of stimulus control in its many forms; dimensional, abstractional, and instructional, and how it can be used to understand the acquisition of both intradimensional and interdimensional abstract tacts and autoclitic frames that guide seemingly complex relational responding and meet consequential contingency requirements. We argue the development of complex relational responding in children can be explained parsimoniously without mediating variables or hypothetical constructs.

Ernst Moerk (1976, 1983a, 1990, 1992, 2000), using his data and data contributed by others, wrote comprehensively about the early childhood acquisition of verbal behavior. He found that verbal behavior training is extensive and continuous (also see, Roy, 2009; Suskind, 2015). Beginning at birth the child begins to form basic verbal constituents that are then shaped by the social interactions with the parents and the familial contingencies. Moerk (2000) noted,

A brief calculation shows that these numbers work out to approximately 3,000,000 instances of major sentence constituents per year. Much of it must obviously be quite redundant. Wells (1986) reported between about 80 and 720 utterances per hour for children between the ages of 15 and 60 months, indicating both the wealth of the input and large variations between families (p. 63).

Earlier, Moerk (1976) observed that parental response to these verbal instances begin simply and grow in complexity

. . . . the mother obtains an answer, hears a spontaneous statement of the child or an imitation of her own utterance, and tests it by comparing it with her standards. If it is found acceptable, no OPERATION is needed. If not, the mother supplies linguistic information by means of a correction, expansion, etc. Thereafter she tests again to see whether the OPERATION was registered by the child and whether it had the desired effect. If the discrepancy is eliminated, the EXIT mode can be chosen; if not, another subroutine containing an OPERATE and a TEST phase can be added (p. 1077).

Moerk (1990) went on to describe complex sets of three-term contingencies operating to shape framing and relations.

From the perspective of language learning, which has focused largely on the controversy of syntax acquisition, the frequent occurrences of mapping in the second position of this three-term contingency pattern are of considerable importance. Mapping refers to the child's encoding of environmental relations and events in syntactic form—a complex tact in Skinner’s terminology. The mother approves and improves this filial construction by repeating and expanding it in Step 3 of the contingency pattern. Analyses in Moerk (1983b) showed that mapping followed by expansion is the third most frequent (n = 100) two-step pattern in child–mother (C–M) sequences. . . . (p. 300; emphasis added)

Moerk’s data are important for two reasons: First, they illustrate the extraordinarily large number of “teaching-trials” experienced by a typically developing child, and second; observed interactions within the family can account for the establishment of many complex verbal relations, including the complex tact relations we will discuss here. We shall explore two important verbal relations described by Skinner (1957), the abstract tact (as noted by Moerk) and the autoclitic frame, and describe their importance to understanding how children acquire complex relational responding. We contend that much of the seeming generative qualities of verbal behavior can be analyzed and accounted for using features of stimulus control that have a been a part of the experimental and applied analysis of behavior for some time (see, for example, Goldiamond, 1966).

The Abstract Tact

Skinner (1957) defined the tact as "a response “evoked by a particular object or event or property of an object or event” (p. 115). The stimulus is nonverbal, the response is verbal without a one-to-one correspondence between stimulus and response as there is in textual responding. Thus, seeing the printed word French and saying “French” is a textual response. Seeing the printed word French and saying “English,” responding to the language in which the word is written, is a tact.

What is less discussed in Skinner’s (1957) definition of the tact is the section that states ". . . or property of an . . . event." When the response is occasioned by one feature or one set of features of an object or event and not by any other features of the stimulus, Skinner calls this an abstract tact. Many everyday tacts are abstract tacts. One’s seeing a shoe and saying “shoe” is a tact, but when we say “shoe" to many different shoes and not to boots or sandals we are engaging in an abstract tact. That is, there are some features of shoe that result in our being able to say shoe to many different looking shoes, and when one or more of those feature are absent we do not.

Few behavior analytic studies involving the tact have explicitly distinguished between simple tacts and abstract tacts (for exceptions, see Robbins et al., 1995; Twyman, 1996). In a procedural sense, this is an important distinction. This difference is illustrated by the following:

Put the flowers on “the” table

Put the flowers on “a” table

Tacting "the" table requires the response is controlled by one specific table. What features are controlling the behavior are unimportant. One simply must discriminate this one particular table from other objects in the room. The flowers are to be placed on one, and only one, table. The table is considered a single stimulus occasioning the response "table." One will not place the flowers on a chair, a bench, a bookcase, or, for that matter, other tables. The tact “a table” now brings an array of tables into consideration. Here the flowers will not be placed on a chair, a bench, or bookcase, but may be placed on any table. What is guiding the tact and flower placement are the features of a table that distinguish it from other furniture, in other words, a ". . . property of an . . . event,” or a subset of properties is guiding the response. All other features of the table are free to vary. One can find any table, no matter its size, color, material, and so forth, as long as it meets the criteria of being a table. This is not the case for the table, where there is one and only one stimulus that can occasion the response. Common abstract tacts include, tree, apple, leaf, bush, vehicle, bicycle, tire, person, dog, fish, glass, spoon, cup, table, chair, art, beauty, photo, lamp, pencil, pen, pillow, bed, blanket, book, magazine, shelf, clock, box, fence, opposite, open, closed, inside, outside, same, larger, smaller, near, distant, over, above, under, behind, me, you, and them, to name but a few.

The Intradimensional Abstract Tact: Dimensional and Abstractional Control

Using common behavior analytic terminology, a discriminative stimulus (S^D) is designated when a stimulus is presented, and a specified behavior is reinforced in its presence. An S∆ is designated when other stimuli are presented and the specified behavior is not reinforced. For example, the response “table” is reinforced only in the presence of a table, but not in the presence of a chair. An individual may thereafter reliably say “table” in the presence of that table and not in the presence of other furniture.

We do not, however, know which specific features of the table as S^D (e.g., height, color, material) occasion the response, "table." Ray and Sidman (1970) noted that there may be a difference between the experimenter-designated S^D and the actual controlling "stimulus–response relation.” For example, we can present a triangle and reinforce a pigeon's pecks to the key on which it is displayed. When the triangle is absent, pecks go unreinforced. But what precisely is guiding the bird's pecking the triangle? Is it the color of the triangle, the wide base, the sharp peak? In essence, there are many features of the triangle that might, if presented alone, occasion pecking.

For this discussion, the stimulus and all its features or dimensions, not simply the ones guiding the response, will be designated as S^Dd. This term describes to what we respond. We will designate any specific features of the S^Dd that occasion a response as S^Da, or abstractional control. These features or properties (S^Da) control how we respond to an S^Dd. It is important to note that S^Da occasions saying “table” even if it is one with dimensions such as a new material that we have never before encountered. For simple tacts, S^Dd and S^Da are treated as the same; the features controlling the response are not specified. We are concerned with one specific table. Here the term S^D will suffice. We simply provide the S^D without reference to any guiding features and reinforce saying “table” in its presence. With this type of relation, particular features of the table guiding the response go unspecified and may vary from individual to individual or from response to response (see Ray & Sidman, 1970).

Where S^Dd and S^Da are not treated as the same, that is, reinforcement is contingent on a response guided by a specified subset of features of the stimulus and not other features, abstractional control is defined (see Goldiamond, 1966; Goldiamond & Thompson, 1967/2022; Layng et al., 2011 for more detailed discussion of S^Dd and S^Da). The verbal response, guided by the features (S^Da), is defined as an abstract tact.

This type of abstractional control has been studied and applied extensively in the context of concept analysis and teaching (see Layng, 2019; Tiemann & Markle, 1991). The study of abstractional control overlaps most closely with what Bruner et al. (1956) described as conjunctive concepts, a class of stimuli, with class membership defined by a set of common, unvarying features, whereas other features vary across instances of the class (see Layng, 2019). For example, the critical features of chair defining S^Da include that it seats one person, has a back, and one’s feet supported at about a 90° angle; whereas any one chair can be made of a variety of materials, have or not have arms, may or may not have four legs, and can come in many colors. In this literature, critical features are described that are always present (must have features), even though a novel example is presented where other (can have features) vary. A child says chair in the presence of a never-before-seen chair. This definition of concept is synonymous with Skinner’s definition of abstract tact, and years of concept analysis and teaching research and application (see Layng, 2019, for a summary) support Skinner’s description of the abstract tact.

In teaching abstract tacts, the features comprising S^Da are held constant whereas the other features are varied from example to example. In essence, S^Dd changes whereas S^Da is unchanged. The response “table” is made to a wide range of tables which may differ in color, size, and material, for example. Equally important is distinguishing tables from nontables. This requires a form of multiple exemplar training that also includes nonexamples, which are essential to ensuring only the must have features control the response. That is, close-in nonexamples of table, such as a backless bench, which lack one or more of the critical features, are essential to ensuring that only the must have features control the response. This interplay between examples and nonexamples that lack one or more critical features is what results in only the must have (S^Da) features guiding the response “table” and defines the verbal operant as an abstract tact.

Goldiamond (1966) and Layng et al. (2011) also noted that S^Da can change whereas the S^Dd remains unchanged. For the same S^Dd there can be different S^Da control depending on the situation. Further, there can be changes to both S^Dd and S^Da. Table 1 depicts a 2 X 2 matrix of the possible relations that emerge from the various combinations of S^Dd and S^Da. One cell describes the simple tact, the other cells describe the various changes in the abstract tact as guiding stimulus features change.

As shown in Table 1, a child may learn to:

Say “chair” to a specific chair when it is in the living room or if moved to the patio: unchanged S^Dd, unchanged S^Da, a simple tact;

Say “chair” to the overstuffed living room chair, to the wooden dining room chair, or to a rocking chair: changed S^Dd, unchanged S^Da, an abstract tact;

Now say “wood” in the presence of the wooden chair: unchanged S^Dd, changed S^Da, abstract tact;

And then say “wood” in the presence of a wooden table: changed S^Dd, changed S^Da changed, an abstract tact.

Table 1.

Depiction of Possible Relations Obtained with S^Dd and S^Da Unchanged and Changed

	SDa Same (Unchanged)	SDa Different (Changed)
SDd Same (Unchanged)	Simple Tact: Stimulus and response unchanged Say “chair” to one particular chair	Abstract tact: Change in guiding features as stimulus stays the same Say “chair” then say “wood” to a wooden chair
SDd Different (Changed)	Abstract Tact: Maintain guiding features as stimulus changes Say “chair” to a novel example of chair	Abstract tact: Change in guiding features as stimulus stays the same, then maintain those guiding features as stimulus changes Say “chair” to a wooden chair, then say “wood” to a wooden chair, then say “wood” to a wooden table

As more abstract tacts are acquired with both the same and different S^Dds, the greater verbal complexity may be observed Figure 1 further illustrates these relations.

Fig. 1.

Pictorial Illustration of the Various Combinations of Unchanged and Changing S^Dd and S^Da Stimulus Control that Differentiates the Simple Tact from the Abstract Tact

There are two distinct categories of abstract tact. One is where the control is provided by features within a single stimulus and another where control is provided by a specific relation between two or more stimuli. We call the former an intradimensional abstract tact and the latter an interdimensional abstract tact (cf. Layng, 2014, 2019). The intradimensional abstract tact requires but a single stimulus to be present. A subset of features (S^Da) of this single stimulus (S^Dd) occasion the response. All other features are free to vary. So far, we have been describing common intradimensional abstract tacts such as chair or table.

The Interdimensional Abstract Tact

The interdimensional abstract tact requires at least two stimuli to be present. A set of (must have) features (S^Da) describes a consistent relation between these two or more stimuli (S^Dd) and occasion the response. All other (can have) features are free to vary. This includes the specific stimuli that demonstrate the relation. Bruner et al. (1956) referred to these arrangements as relational concepts. Figure 2 shows an example of a match to sample exercise where the match requires an interdimensional abstract tact. The task requires the participant to select a comparison the reflects a 90° rotation from the sample. The arrow rotation that is correct for one sample, a vertical arrow, is incorrect for a different sample, a horizontal arrow, as the 90° rotation remains constant. Here the S^Dd are the two stimuli; the S^Da is the specified, must have, relation between them that meets the contingency requirement for reinforcement. It is this relation, 90° rotation, that guides this abstract tact. The stimuli that demonstrate the relation are free to vary.

Fig. 2.

The Circle Indicates a Correct Match. Note. The S^Da (rotated 90°) remains the same as the S^Dd (arrows) are rotated

As depicted in Fig. 3 we can substitute other stimuli that can reflect a rotation and occasion the abstract tact “rotated 90°.” The stimuli compared are varying (can have) features, just as the material or the color of a chair are varying features. The S^Dd changed, S^Da unchanged relation applies to both intradimensional and interdimensional abstract tacts; no special treatment is required for relations between stimuli.

Fig. 3.

The Circle Indicates a Correct Match. Note. The stimuli (S^Dd) depicting the relation are free to vary; rotated 90° (S^Da) guiding the response stays the same

There are many important interdimensional abstract tacts. For example, opposite, same, different, now, past, future, near, distant, larger, smaller, here, there, me, you, over, under, steeper, above, below, inside, outside, and believe are all common interdimensional abstract tacts for which the guiding features can be determined. An analysis of the abstract tact "opposite" reveals the relation governing its use has two critical features comprising S^Da, (1) events on a given dimension in an array, that are (2) positioned approximately equidistant from a midpoint along the array. Figure 4 illustrates this relation. Depicted is an array of shades of gray from black to white. Although it is typically obvious that each end of the array, black versus white, is opposite, what can be overlooked is that any two shades equidistant from a midpoint are also opposite (see Fig. 4).

Fig. 4.

Array of Gray Shades Depicting the Abstract Tact Opposite. Note. Events on a given dimension that are positioned about equidistant from the midpoint along an array. The white arrow shows a midpoint, the red arrows indicate the shades that are opposite of one another

As noted, the stimuli comprising the array are free to vary. They can be political parties: Democrat is the opposite of Republican, size: small is the opposite of large, distance: near is the opposite of far and so on. Guidance by the critical features (S^Da) occurs regardless of the stimuli comprising the array. Where there is disagreement as to what is opposite of something else, such arguments are often about where the events fall along the array. Instead of Democrat being opposite of Republican, one might argue that no, the Socialist Workers Party is the opposite of Republican.

The same category of interactions that establish intradimensional abstract tact control, saying chair to novel examples of chair, govern interdimensional abstract tact control (see Engelmann & Carnine, 1982; Layng, 2019, for examples of the same teaching routines being applied for each). Further, there is no requirement that a person is able to describe the S^Da features that comprise opposite, just as few can describe the features of a chair yet can say "chair" in the presence of a novel example of chair. Thus, it is likely that both intradimensional and interdimensional abstract tacts are acquired similarly through interactions with examples and nonexamples encountered in everyday life within the verbal community. We shall return to this point below.

From Abstractional (S^Da) to Instructional (S^Di) Control

The word opposite is guided by certain critical features and can be used to establish such control where it might be absent. That is, when we say X is opposite Y, we are essentially saying respond to X and Y as events along an array that are about equidistant from a midpoint. We call the use of the word opposite to establish control by the critical features as an instructional S^D or S^Di. We define S^Di as a supplementary stimulus that restricts response alternatives to a set of features (S^Da) of the S^Dd that meet a contingency specified requirement. Instructional control produces the same control as if S^Da control were contingency shaped. That is, S^Da control can be extended to new stimuli without having to be exposed to contingency shaping. However, S^Di control cannot occur unless the abstract tact has been previously established.

The word “opposite” becomes an S^Di when it is paired with the critical, must have, features of opposite (S^Da) not individual or arbitrary stimuli (S^Dd). That is, it makes no difference what is opposite of what as long as the S^Da requirements are met. This is what Tiemann and Markle (1991) describe as a two-way paired associate, or others describe as symmetry (Sidman, 1994). A speaker’s use of the word opposite first comes under the control of certain critical features, and its subsequent use is reinforced by listeners, who have a history of such S^Da control. Coming under the control of those critical features now occasion the use of “opposite” by the speaker. Stated in a different way, S^Da occasions the speaker’s use of the word “opposite" as a response; and the word "opposite" functions as an S^Di to restrict listener response alternatives to a specific set of features, S^Da. This results in control by that S^Da and no other features of the stimuli. The word “opposite” or any stimulus paired with S^Da can then be used as an S^Di to extend abstractional control by the critical features to a new situation or set of stimuli.

To be clear, the child is not abstracting, nor forming an abstraction. Nor is the child relating or forming a relation between stimuli. Abstractional control is in the stimulus control relation established by the contingency, the features of which may be analyzed and specified. As Skinner (1957) noted, “To discover the 'essence' of a chair, we should examine . . . the actual contingencies of reinforcement in a given community” (p. 126). For the abstract tact, S^Da does more than occasion the word “chair.” It restricts response alternatives to activities that are common to the community, thus giving the occasion its shared meaning in the community. Likewise, the spoken S^Di “chair” is not simply a match to stimulus; it restricts response alternatives to a set of shared activities within a verbal community that can only occur with a chair. That is, the word “chair” does not represent, stand for, substitute for, or take on relational properties of a chair, it simply restricts response alternatives to activities that can only occur with a chair. When one says, “bring me a chair,” the speaker is not communicating a mutually entailed match. They are saying, “bring me an object where I alone can sit with support for my back, and where my legs will be supported at about 90°.” To the extent one brings an object that meets that requirement, there is a shared meaning of chair. Such consequential governance is critical to the establishment and maintenance of abstractional control. The word chair is not a picture of the object chair. The word chair derives its meaning by a set of shared practices (cf. Wittgenstein, 1953). As Sue Markle once said (S. Markle, personal communication, October, 1972), “concept [abstract tact] analysis is an exercise in cultural anthropology” (brackets added).

The Interdimensional Abstract Tact and the Autoclitic Frame

An autoclitic is verbal behavior “which is based upon or depends upon other verbal behavior” (Skinner, 1957, p. 315). It “clarifies or alter(s) the effect of verbal behavior upon the listener” (p. 332). The earlier example of the chair versus a chair illustrates this effect. A speaker’s utterance of either “the” or “a” modify the effect of saying “table” on the listeners behavior. The autoclitic frame is another type of autoclitic effect. It allows for relations to be described between stimuli.

Once instructional control is established, one can say: X is the opposite (S^Di) of Y. This type of verbal unit comprises an autoclitic frame. An autoclitic frame requires that an interdimensional abstract tact, in this case opposite, is part of the repertoire of both the speaker and the listener. That is, both speaker and listener are under the same S^Da control. The abstract tact when used as an S^Di defines the "frame," which means it instructs the listener to “respond to a particular S^Da relation,” not to the individual stimuli or other possible relations within or between stimuli. For example, autoclitic frames featuring the abstract tact opposite result in X and Y being treated as stimuli positioned about equidistant from the midpoint along an array without ever before having to encounter X or Y. Nothing must have been paired with either X or Y. X and Y become examples of varying (can have) features of the interdimensional abstract tact opposite. This is no different from treating a previously unseen chair as a chair.

X and Y can be abstract tacts themselves or members of a class such that any member becomes opposite. For example, we can say, solid (S^Da) is the opposite (S^Di) of liquid (S^Da). Now *any* example of a solid will be the opposite of *any* example of a liquid as long as the abstract tacts solid, opposite, and liquid have been established. This creates an autoclitic frame where any example of a solid becomes the opposite of any example of liquid. It is not arbitrary stimuli that have a history of being paired with other stimuli that are responded to as opposite, but the critical features (S^Da), which are the same regardless of the particular stimulus (S^Dd). The solid or liquid, even if never previously encountered, or paired with any other stimulus, will be responded to as opposite. To emphasize, this type of relational extension is possible only if the abstract tacts solid, liquid, and opposite have been previously established, thus making possible the autoclitic frame (typically the form X S^Di Y) that links different S^Da control in a consistent way. This is of great importance to the remarkable generativity of verbal behavior1 and may account for many of the generative features observed in the rapid attainment of verbal behavior in children (Luke et al., 2011). We shall return to this point below.

In the autoclitic frame X S^Di Y, X and Y need not be abstract tacts, they can be arbitrary (S^Dd = S^Da) stimuli; that is, the stimuli share no common features. We can say ¥ is opposite (S^Di) of Ω. There are no critical features other than the entire stimulus (S^Dd) itself. These arbitrary stimuli will be treated as though they are stimuli positioned about equidistant from the midpoint along an array. Given that control by the interdimensional abstract tact "opposite" has been established, and a stimulus––a printed or spoken word or other stimulus, has been established as S^Di, the stimuli will be responded to as opposite.

An autoclitic frame can link S^Dd to S^Da in other ways, with S^Di not being a word at all. The minimal abstract tact indicating possessive (“ ‘ “), for example, can be used to establish the autoclitic frame, X's Y.2 The autoclitic frame Joe’s nose is an example. Joe is not an abstract tact but nose is. The autoclitic frame (X's Y) indicates that the nose belongs to Joe. S^Di is simply a stimulus that restricts response alternatives to those occasioned by an already established S^Da.

Further, once well-established, there can be S^Di control in the absence of the S^Dd/S^Da stimuli that originally occasioned it. This may provide the basis for what may often be labeled as imagination, planning, and so on. As Layng et al. (2011, p. 5) described,

Verbal stimuli, spoken and textual, can provide S^Di guidance over repertoires established when both the S^Dd and the S^Di were present. One can ask someone to pretend they are driving a car. We may see the individual grasp an invisible steering wheel, adjust their feet, and extend their arms, though no steering wheel is present. We might also observe subtle movements of the hands as if one is adjusting the car’s direction while driving. If we shout, “There’s a cat in the road!” we may see a sudden movement as to indicate a rapid turning of the wheel. We may even notice an “emotional” response indicated by raised eyebrows and widening eyes. But there is no steering wheel and no cat is present. What we have done is to potentiate S^Di guidance over behavior similar to that which may exist when there is also S^Dd guidance—that is, when we are actually driving. S^Di guidance in the absence of S^Dd guidance is similar, but not the same as that which occurs with S^Dd guidance. This difference has been noted in discussions of rule-established behavior, where behavior is primarily under S^Di guidance (Skinner, 1966). This is an important distinction. Behavior under S^Di guidance in the absence of an S^Dd is different than when the S^Dd is present, even when the observed topography may initially look the same. Responding guided by S^Di in the absence of S^Dd may cause some to conclude that we are responding to a private stimulus, when in fact we are simply responding to S^Di guidance.

Space limitations prevent a full discussion here of the relation between S^Di control and putative private stimuli, though, a more detailed discussion can be found in Layng et al. (2011). A child’s verbalizations while playing, however, may reflect such S^Di control, and once spoken and heard become stimuli that may further restrict or alter response alternatives. The actual S^Di control provided by the verbal stimuli will be a function of the previously established S^Da control.

It is through S^Di control that a speaker and a listener (or speaker as listener) share words that have a “common meaning.” The speaker, in the presence of a table, says “table.” A listener, who is guided by the same S^Da features, hears the S^Di “table,” and their response alternatives are restricted to those features guiding the speaker’s behavior. Stated in a different way, it is the joint history of S^Da control occasioned by the spoken and heard S^Di that accounts for much of what may be described as shared understanding.

On the other hand, the utterance of a word heard by one child may have a different S^Di effect from that same word heard by another child. That is, a word acquired as a simple tact is in an entirely different relation to the environment than is the same word acquired as an abstract tact. Hence, the difference in developing repertoires is not simply a matter of the presence of more or fewer words as has been suggested by some investigators (Hart & Risley, 1995; Suskind, 2015). As Layng (2022) noted,

This distinction between the acquisition of many vocabulary words as simple tacts and the acquisition of complex abstract tacts may alone have a profound effect on later performance. In homes with high language interaction, in which many examples and nonexamples presumably are available, children with the exact same words in their vocabulary may develop entirely different relations to their environment. We may be tempted to account for subsequent differences in academic performance as indicating differences in brain development when in fact we may be observing the effect of topographically similar but functionally quite different entering repertoires. Similarly, brain development may be impacted by which behavioral relations are in place at what time. (p. 344)

The type and complexity of autoclitic frames would likewise be affected, and such difference may appear when certain autoclitic frames are relied upon to teach more complex relations in school, or in measures of reading comprehension (see below, and Layng et al., 2011).

Interlocking Autoclitic Frames

Autoclitic frames can combine so as to meet new contingency requirements that restrict response alternatives such that novel categorization may occur. We can use the interdimensional abstract tact “same” to establish ¥ same (S^Di) as Ç. And we can use the interdimensional abstract tact “opposite” to establish ¥ is opposite (S^Di) of Ω. We can use the autoclitic frames as follows:

• Because ¥ is opposite (S^Di) of Ω

• And ¥ is the same (S^Di) as Ç

• Ç is the opposite of Ω

Because Ç is the same as ¥, Ç takes ¥’s place on the continuum that defines opposite. Thus, as a result of a history of S^Da control it is treated as opposite of Ω without training, just as it (Ç) would if it shared a common S^Da control with ¥. Such relational responding is not arbitrary but is a function of features governing interdimensional abstract tacts. No application of an abstraction nor act of relating is required. Thus, autoclitic frames can be used to establish control by objects other than spoken or printed words. For example, as long as the abstract tact (S^Da) “means” is in one’s repertoire, (X means (S^Di) Y), we can say ∆ means opposite. When ∆ is present X will be opposite of Y. Likewise we can say ø means same. When ø is present X will be the same as Y. Symbols can be considered extensions of S^Di control based on a history of S^Da control (cf. Saussure, 1916/1983).

Discerning readers may see the obvious relevance of this discussion to issues involved in areas described as stimulus equivalence and relational frame theory (RFT). A full discussion is beyond the scope of this article. In brief, however, much of the recent investigation of relational responding has been based on the stimulus equivalence paradigm. Here stimuli that share no contingency defined features are arbitrarily grouped in the context of fixed alternative forced choice match-to-sample procedures. Often left unaccounted for are the interdimensional abstract tact repertoires of the participants in these experiments. This becomes even more critical when some of the relations involve interdimensional abstract tacts, such as same, opposite, differ, and so on, which are often the subject of RFT investigations. In reviews of the relation between these formulations and Skinner’s (1957) analysis of verbal behavior, the implications of the abstract tact and in particular the interdimensional abstract tact tend to be overlooked (Barnes-Holmes et al., 2000; Barnes-Holmes et al., 2018; but see Hall & Chase, 1991). Where participants may lack the interdimensional abstract tact repertoires, pretraining procedures are often employed to ensure their presence (Steerle & Hayes, 1991; Quezada Velázquez et al., 2018). The relational responding observed in these experiments may be considered derivative of a history of interdimensional abstract tact control that serves as a basis for training autoclitic frames using the stimuli presented in the context fixed alternative forced choice match-to-sample procedures, much the same as described above.

Interdimensional Abstract Tacts and Relations between Relations

Abstract tacts can be used to describe relations between relations. In instructional design parlance these are called principles (Tiemann & Markle, 1991), or for simpler relations joined concepts (Engelmann & Carnine, 1982).

An example of a simple principle or joined concept is, “round things roll.” Three abstract tacts are required, round, things, and roll. Once these are established, we can sequence them as a series of S^Di statements to create a new stimulus control relation. This joint or combinatorial S^Di effect becomes the mainstay of bringing the child’s verbal behavior under the control of a wide range of complex relations without having to have direct contact with the relations other than having the component abstract tacts in their repertoires. Seeing a barrel for the first time a child may push it and have it roll without ever having observed it move. It is important to understand that the critical features that comprise the various S^Da relations may go undescribed. The S^Da control does not require the child, or anyone, be able to state the critical features controlling the response. It only requires that the verbal community respond in a differentiated way when those features do or do not control behavior.3

More complex relations can be specified such as, “for every action there is an equal and opposite reaction.” Again, for these relations, the interdimensional relations described depend on the abstract tacts previously established, these include:

• ––every

• ––action

• ––reaction

• ––equal

• ––opposite

If object A pushes on object B, then object B pushes on A with the same force but in the opposite direction. We do not have to have any direct contact or history with object A or B. They are varying, can have, features.

Analogies and Metaphor

The abstract tact/autoclitic frame combination is also essential for understanding the effect of analogies. In an analogy, an autoclitic frame comprises the relation upon which the analogy is based, that is, the shared S^Da/S^Di control. Many analogy problems are solved where such control is discovered, such as, where one must generate a word that completes the analogy. For example: Feathers are to birds, as ________ is to cats. A strategy described by Whimbey et al., (1999) using a talk aloud problem solving strategy (also see Robbins, 2011) illustrates how S^Da control may be revealed. The strategy focuses on discovering a common autoclitic frame and its interdimensional abstract tact. Learners are taught to generate autoclitic frames to determine the abstract tact that is the same for both. For example: Feathers cover (S^Di) birds; fur covers (S^Di) cats. The autoclitic frame X covers Y is guided by the abstract tact, covers. Thus, feathers are to birds, as fur is to cats. The task is to make explicit the implicit autoclitic frame that is the basis for the analogy.4

Below is another example.

• pack is to wolves as __________ is to ___________.

• a. wheel: spokesc: alphabet: letters

• b. garage: carsd. aquarium: fish

Here control by close-in abstract tacts control the selection. We can say Xs are part of Y. Cars are not part of garages, that’s out. Fish are not part of aquariums, that’s out. Spokes are part of wheels, yes. That may be it. Let’s check, wolves are part of a pack. But are they? Wolves, “plural” are the pack. Wolves compose the pack, spokes do not compose a wheel. A wheel is still a wheel, even without the spokes. A pack without wolves is not a pack. Letters, however, compose an alphabet. An alphabet without letters is not an alphabet. The answer is c. Xs compose a Y fits both, so the autoclitic frame guided by the abstract tact “compose” determines the analogy. Whether or not the S^Da control of the abstract tact “part of” is distinguished from the S^Da control of the abstract tact “compose” will determine the selection of the correct answer. The extent to which the required abstract tact is part of one’s repertoire and is differentiated from other close-in S^Da control, will determine whether or not a solution is found (cf. Cordeiro et al., 2021; Lin et al., 1996; Meyer et al., 2019; Miguel et al., 2015). After practicing, the control may be such that the comparisons will be performed and the leaners may report privately hearing the words that they might have previously spoken out loud (see Layng et al., 2011, for an analysis of how such patterns may be established; cf. Miguel, 2018).

Metaphor can be treated similarly. Creating an autoclitic frame using established S^Di control in reference to stimuli not typically included in the frame, so as to bring the listener’s behavior under the control of a new S^Da relevant to that stimulus, describes much of what can be considered as metaphorical extension (Skinner, 1957). We will illustrate with an example from Ted Cohen’s (2009) book, Thinking of Others. The abstract tact “utensil” was used by Winston Churchill to describe Mussolini’s relationship to Hitler, when he said, “Mussolini was a utensil.” In other words, Mussolini was but a useful tool. The S^Di, utensil, extends those features to Mussolini.5 In essence, the S^Di control of utensil is transferred to Mussolini, resulting in a what may be described to be a transfer of stimulus function. Space precludes a greater discussion of the relation between S^Da, S^Di, and S^Dd control and metaphor; however, for an expanded, related treatment see Butcher (2021).

How Are Abstract Tacts and Autoclitic Frames Acquired?

One of the strengths of the account offered here is that the relations described can readily be observed to be established in everyday interactions between parents and children. Though a well-established laboratory or educational procedure may reliably establish complex relations, this does not mean such procedures or relations occur outside the laboratory. For instance, errorless establishment of discriminative responding and its transfer can readily be established in the laboratory. However, it is highly doubtful that instances of errorless teaching arise naturally in the environment outside the lab. Thus, when looking to account for early childhood learning, errorless learning will not likely enter into such an account. Thus, extrapolation from highly controlled laboratory arrangements must be done with caution. We are confident that interactions observed in the home by Moerk and others, including the direct learning through nonexample example discrimination as observed in “fast-mapping” (Carey & Bartlett, 1978; Swingley, 2010), are sufficient to establish the relationships discussed here.6

Further, there is evidence that programs that carefully establish these relations may greatly benefit children who have difficulty acquiring them through typical home interactions. For example, in an early study (Engelmann, 1968) directly teaching interdimensional abstract tacts in the context of autoclitic frames to disadvantaged learners resulted in a gain on average of over 24 IQ points (ending IQ average of 121), whereas control participants gained on average about 5 points (ending average 99.61).

Summary

We contend that much of the seeming generative qualities of verbal behavior can be analyzed and accounted for using features of stimulus control that have a been a part of the experimental and applied analysis of behavior for some time (see, for example, Goldiamond, 1966). A careful examination of two fundamental processes described by Skinner (1957), the abstract tact and the autoclitic frame, provide powerful analytic tools for adding to our understanding verbal behavior complexity. The contingencies that are responsible for S^Dd, S^Da, and S^Di control and their operation within the autoclitic frame, become the basis for understanding important aspects of the continually developing relational responding observed in children. Further, though not discussed here, procedures exist for teaching these relations where they may be absent (see, for example, Engelmann & Carnine, 1982; Layng, 2019; Tiemann & Markle, 1991; Twyman, 2021).

As noted earlier, there is no need to resort to hypothetical constructs that posit the child forms abstractions, processes information, or relates one event to another. The abstracting and relating are found in the contingency requirements (after Layng et al., 2022), not in the child. The relations discussed here are not to be considered exhaustive, but instead, part of the larger investigation of verbal behavior development. Whereas this account might be considered post-Skinner, it is decidedly not post-Skinnerian; it demonstrates the ongoing contributions of Skinner’s (1957) analysis of verbal behavior along with Goldiamond’s (1966; Goldiamond & Thompson, 1967/2022) extended treatment of stimulus control.

Data Availability

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Funding and/or Conflicts of interests/Competing interests

The authors have no relevant financial or non-financial interests to disclose.

This article was prepared without financial support.