Teaching Intraverbal Behavior
Mark L. Sundberg
January 25, 2006
DRAFT VERSION
Much of our day-to-day language involves emitting words, phrases and sentences that are in response to the words, phrases, and sentences of others. For example, if someone asks How are you, the tendency to say I’m fine is a type of language where the words of others function as the primary discriminative stimulus for other words. That is, How are you evokes I’m fine due to a verbal conditioning history. Skinner (1957) identifies this type of verbal behavior as intraverbal behavior, and suggests that it plays a major role in our daily language interactions. He also suggests that the intraverbal, like the other verbal operants are functionally independent from each other. By “functionally independent” Skinner means the ability to emit response related to one of the verbal operants (e.g., a tact) does not guarantee the ability to emit that same response in another (e.g., intraverbal). For example, a child may be able to say Ball when he wants a ball (mand), say Ball when he sees a ball (tact), and say Ball when someone else says Ball (echoic), but not able to say Ball when someone says What can you bounce? (intraverbal). A growing body of empirical research with both typical children and children with developmental disabilities, supports Skinner’s conceptual analysis of the independence of the intraverbal repertoire (e.g., Braam & Poling, 1982; Miguel, Carr, & Poindexter, 2005; Luciano, 1986; Partington & Bailey, 1993; Watkins, Pack-Texteria, & Howard, 1989).
In technical terms, the way the verbal operants differ from each other is mainly in what functions as an antecedent for each verbal response. The antecedent for the mand is motivation (MO), the antecedent for the tact is the physical environment (nonverbal discriminative stimuli), the antecedent for the echoic consists of words that match the words of the relevant response (verbal discriminative stimuli with point-to-point correspondence with the verbal response), and the antecedent for the intraverbal consists of words that do not match the words of the relevant response (a verbal discriminative stimulus that does not have point-to-point correspondence to the response). The intraverbal relation is closest to the echoic relation in that both verbal operants involve antecedents that are verbal discriminative stimuli, however, in echoic relation the response matches the antecedent, whereas in intraverbal relation the response does not match the verbal stimulus (see Table XXX, See appendix).
Intraverbal behavior can be observed in many typical verbal interactions between people. Perhaps some of the most obvious types of intraverbals involve answering the questions of others. For example, if a parent asks What did you do at school today? And a child answers A fireman came into our classroom and talked to us. The answer to the parent’s question is intraverbal because the antecedent that evokes the child’s response was a verbal stimulus that did not match the child’s verbal response (the question asked by the parent would be classified as a mand). Some intraverbal behavior is very simple and even trivial such as producing the sounds of animals when asked to do so (e,g., saying Meow when someone says A kitty says....). While other intraverbal interactions are extremely complex such as answering questions with multiple components and advanced content (e.g., Can you explain a modified quadratic equation?). The current chapter will describe a variety of procedures for teaching intraverbal behavior and suggest a general order for training. The sequence of intraverbal skills (the curriculum) presented in this chapter follows the same type of analysis presented for tact training in that the focus is on the increasing complexity of the relevant types of stimulus control and the increasingly complex responses that they evoke. This analysis suggests that verbal stimulus control and the related intraverbal responses can, to some degree, be viewed on a continuum with simple types of control and responses being easier to establish than more complex types of control. As with tacting, establishing easier forms of stimulus control seems to set the stage for the establishment of more complex types of control.
Establishing early intraverbal behavior
The acquisition of intraverbal behavior by typically developing children can provide a general guide for the sequencing of an intraverbal curriculum for children with language delays. The first question is when do typical children begin to acquire intraverbal behavior, and the second question is what types of intraverbal behavior do they acquire? Clearly, intraverbal behavior is not the first type of verbal behavior acquired by typical children. It was suggested previously that the mand is the first type of verbal behavior acquired (usually in the form of crying), followed by echoic, receptive, and tacting skills (not necessarily in that order). Intraverbal stimulus control is often a more difficult type of control to establish (see below), and it is reasonable that verbal responses would first come under of motivation and nonverbal visual stimuli. However, some types of intraverbal control are relatively simple and some children acquire a few intraverbal responses shortly after a few spoken mands and tacts are acquired (typically around 12-14 months). For example, many children learn to provide the sounds that favorite animals make when asked to do so, fill-in-the blanks of favorite songs, or fill-in-the-blanks related to favorite activities. The word “favorite” suggests that motivate variables and perhaps automatic reinforcement may be at work here and share control over the behavior with the verbal stimulus (i.e., multiple control). Nonetheless, these early types of verbal stimulus control can begin to set the stage for more advance types of verbal stimulus control.
Starting intraverbal training for children with autism
In general, formal intraverbal training should not be a part of an intervention program for a child with autism until the mand and tact repertoires are established, and the child can emit good echoic, imitation, receptive, and matching-to-sample behavior. That is, the first six basic language and learning skills should be the focus of an intervention program prior to adding formal intraverbal training. There certainly are exceptions, and an individual analysis of each child can provide guidance as to when intraverbal training should begin and what form it should take. For example, some children do acquire intraverbal responses before tacting, and intraverbal prompts can be used to establish tacts (Sundberg, Endicott, & Eigenheer, 2001). While it is true that some children begin to acquire a few intraverbal responses shortly after one year of age, they are relatively trivial (e.g., animal sounds) and often emerge with very little formal training (perhaps due to the effects of pairing and automatic reinforcement). The recommendation here is that there is no harm in trying to establish these same types of intraverbal behaviors and they may be of ultimate value to a child with language delays, but if they are hard to establish for an individual child intraverbal training should be put aside until the six basic repertoires are stronger (see the section on “red flags” below).
Formal intraverbal training seems to work best for a child who can easily emit a number of different mands, tacts, and receptive discrimination skills. As far as how many, it is difficult to put a specific number on these, but as a general guideline the mand repertoire should involve a variety of different mands (10 might be a reasonable number, but some children may be successful at intraverbal training with fewer mands), and come easily for a child. The tact repertoire should involve at least 50 different tacts, some of which should include verbs (again, however, some children may acquire intraverbal behavior without this number of tacts, while other may do better with more tacts). An individual analysis is required to determine what is best for a specific child, and that child’s successes and failures (red flags) will provide the relevant data for decision making. In addition to mand and tact skills, echoic (or imitation skills for those who use sign language) should be strong, as well as matching-to-sample skills. Finally, it may be helpful, but not necessary for early intraverbal training, that the child be able to show some success with RFFC procedures. Certainly, success with RFFC procedures will ultimately play a major role in advancing a child’s intraverbal skills, but may not be necessary for early intraverbal training.
Level 1 intraverbal training: Teaching the first intraverbal responses
The simplest types of intraverbal behaviors are the same as those that occur early for many typical children. Providing the sounds that animals make and filling-in the missing words from songs and common phrases are often easy to teach, and have the effect of beginning to teach a child to make a non-echoic response to a verbal stimulus. There are several ways to teach intraverbal behavior, but most involve the same basic teaching procedures described for the other forms of verbal behavior (i.e., behavior modification procedures). For these early intraverbal responses, echoic prompting and fading procedures may work best (Table 8-1). First present the target verbal stimulus such as A kitty says..., then provide an immediate echoic prompt (Meow), and pre-present the target verbal stimulus and delay the echoic prompt (the transfer trial). Reinforce a correct response (Meow) with praise or an opportunity to mand. Incorrect responses (e.g., the child echos A kitty says...) should be followed by repeating the trial with the slightly delayed prompt and attempting to more carefully time and fade the echoic prompt (e.g., partial prompts, reduce volume, reduce delays). If errors continue, try reversing the intraverbal to Meow says a... and use a tact prompting and fading procedure. Specifically, instead of echoic prompts, use a picture of a cat as a prompt (tact) and fade out that prompt using the standard prompting, fading, and transfer of stimulus control procedures. If errors still continue, try a few other animal sounds, and if still the child is not successful discontinue this type of intraverbal training for a while (perhaps a month or two). Some children, however, may be successful with song fill-ins, despite failing to acquire animal sounds and the procedures described below should also be tried.
Table 8-1
Echoic to intraverbal transfer with animal sounds.
________________________________________________________________________
Antecedent Behavior Consequence
Target verbal stimulus (A kitty says...) Meow Praise
Echoic prompt (Meow)
It is often quite easy to teach a child to intraverbally respond to part of a song with a fill-in type response, especially if it is a reinforcing song such as one that comes from a favorite TV show or video. For example, if a child finds Sponge Bob reinforcing and watches the TV show or video frequently, he may be able to correctly fill-in Sponge Bob Square.... with Pants. If this response can be evoked in the absence of the nonverbal stimulus of Sponge Bob and his square pants it can be classified as intraverbal. Again, it may not be a very sophisticated type of intraverbal relation, yet it is intraverbal nonetheless. In addition, it is possible that a child could get this intraverbal correct, but not be able to tact pants or receptively discriminate pants for other items, especially clothing items. However, this activity can still have value in that it begins to teach a child to differentially respond to verbal discriminative stimuli. If a child fails to provide the correct intraverbal response, an echoic or tact prompting procedure should be tried. Obviously, the tact procedure will only work if the child can tact Bob’s pants. For both procedures present the target verbal stimulus (Sponge Bob square....) with a delay between the verbal stimulus and the target response and with a facial expectation prompt to the child (e.g., the trainer has a questioning look on her face), then say Pants. On the next trial (transfer trial) further delay the Pants prompt (but keep the facial prompt) and reinforce a correct response with praise of the opportunity to mand. An incorrect response should be followed by reducing the delay between the two verbal stimuli, and by using more careful prompting and fading steps. Correct responses should be followed by slowly fading the all the prompts. If the child continues to fail, try the procedure in the context of the TV show or movie (multiple control by adding in nonverbal prompts), or try a few other songs. If these interventions fail, put this type of training on hold for a period of time (perhaps one or two months) then try again later.
Table 8-2
Echoic or tact to intraverbal transfer with a song fill-in.
________________________________________________________________________
Antecedent Behavior Consequence
Target verbal stimulus (Sponge Bob square... pants Praise
Echoic or tact prompt (Pants/pants)
Fun activities can also be used as a source of verbal stimuli for early intraverbal development. For example, if a child enjoys playing peek-a-boo during the game the trainer can present the verbal stimulus Peek-a and delay boo, followed by a echoic prompt, transfer trial, and essentially the same teaching and correction procedure described above for song fill-ins. If a child responds Boo it is most likely part mand and part intraverbal (it could also be part tact, in that the child might be attending to the hands covering the trainer’s face). The intraverbal source of stimulus control will become more established when a second game is added that involves a second verbal stimulus and a different response. For example, if the child enjoys being spun in a chair the trainer could say Spin the... and delay the word chair. Using the echoic prompt and transfer procedure it may be possible to evoke the response chair from the child. Successful switching between Spin the and Peek a will begin to make verbal stimulus control more apparent. It may not be unexpected for the child say boo when the training says Spin the... This type of error indicates that verbal stimuli have not yet acquired a discriminative function and further training is required. A similar type of “overgeneralization” (and failure to discriminate) error might be observed in early tact training when a child identifies a new item with a response related to a previously trained item (e.g., after learning to tact shoe, he calls a car shoe). Additional fun activities can be added and interspersed with the other intraverbal activities that the child has shown some success with. It is important to note that these early intraverbal responses may be multiply controlled, especially with the fun activities exercises, but at some point it will be important to break the response free from these additional sources of control. A measure of intraverbal development is necessary and can be provided by periodic cold probes (first trial data) on just the intraverbal relations.