Week 4 Fisher Chapter 16 Verbal Behavior PDF

## Chapter 16: Teaching Verbal Behavior **Matt Tincani, Caio Miguel, Andy Bondy, and Shannon Crozier** Lori is eating popcorn in a classroom while she prepares a lesson for preschool students. Lily walks in, sees the popcorn, and says *"Popcorn!"* Lori smiles and gives her popcorn. Jack walks in a...

## Chapter 16: Teaching Verbal Behavior **Matt Tincani, Caio Miguel, Andy Bondy, and Shannon Crozier** Lori is eating popcorn in a classroom while she prepares a lesson for preschool students. Lily walks in, sees the popcorn, and says *"Popcorn!"* Lori smiles and gives her popcorn. Jack walks in and sees the popcorn. He grabs for the bag without saying anything. Lori asks *"What do you want?"* Jack says *"Popcorn!"* Lori smiles and gives him popcorn. Char walks in, sees the popcorn, and grabs for the bag while remaining silent. Lori says *"What do you want?"* but Char continues to grab for the bag while remaining silent. Lori then says *"Say popcorn."* Char immediately says *"Popcorn!"* and Lori smiles and gives her popcorn. If Lori were to complete a checklist based on these interactions, including the question *"Can each child say popcorn?"*, the correct answer would be yes. However, if we were to ask *"Does each child do the same thing?"*, the correct answer would be no. Although the form of the response *"Popcorn"* is the same, each example involves a different controlling relation. The difference in perspective between the first and second questions is at the heart of a functional analysis of verbal behavior. Although knowing the form of verbal behavior is useful, understanding its functional control is more important. This chapter reviews how a functional analysis of verbal behavior can provide guidance about the complex issues associated with teaching communication and language to those with limited or no verbal repertoire. ### Skinner's Analysis of Verbal Behavior In his seminal book *Verbal Behavior*, Skinner (1957) showed how we can analyze language by using the principles of operant and respondent conditioning. Skinner chose the term *verbal behavior* rather than speech, because he did not want to restrict his analysis to vocal behavior. Traditionally, language was considered an ability or system in an organism responsible for generating speech. The source of control for language or verbal behavior, respectively, was a major difference between the traditional approach and Skinner's approach. According to the traditionalists, control for language originates within the organism. According to Skinner, by contrast, control for verbal behavior originates from contingencies of reinforcement acting upon the organism's behavior. Skinner's terminology emphasized that verbal behavior is like any other operant behavior. The term *verbal behavior*, which at that time was "relatively unfamiliar in traditional modes of explanation" (Skinner, 1957, p. 1), set the stage for a departure from traditional explanations of language. Skinner (1957) defined verbal behavior as "behavior reinforced through the mediation of other persons" (p. 2). In other words, a speaker's behavior is reinforced through or mediated by the behavior of a listener. The topography of the speaker's behavior, such as vocal, gestural, or visual, is irrelevant within this definition. Skinner stated his unique orientation to verbal behavior explicitly: "In defining verbal behavior as behavior reinforced through the mediation of other persons, we do not, and cannot, specify any one form, mode, or medium" (p. 14). Sometimes our behavior influences the environment in direct ways. For example, a little boy ties his own shoelaces and can run around immediately. At other times, our behavior influences other people whose actions lead to reinforcement. For example, the boy asks his brother to tie his shoes and runs around after his brother does so. Although the first example does not meet the definition of verbal behavior, the second one does. The verbal community selects specific forms of verbal behavior to function in certain ways. The behavior forms may be vocal, such as spoken words; graphic lines, such as writing; or hand postures and movements, such as sign language. These behavior forms produce an effect on the environment. Skinner added an important refinement to the definition of verbal behavior when he wrote, "The listener must be responding in ways that have been conditioned precisely to reinforce the behavior of the speaker" (p. 225, emphasis in original). An individual learns listening behavior as a member of a verbal community. Thus the listener learns to react to the speaker's verbal behavior, and this is a requirement for behavior to be verbal. In the example above, when the little boy asks his brother, "Will you tie my shoes?", the auditory product of this behavior, the sound pattern, serves as a stimulus that evokes his brother's shoe tying. But notice that the brother can react appropriately as a listener by performing the task only if he has learned how to respond to this request previously. In lay terms, the listener must understand the speaker. We contrast the behavioral approach Skinner (1957) advocated with the traditional approach to language development and intervention of Chomsky (1965), Brown (1973), and Piaget (1951), who conceptualized language by the form or topography of the learner's verbal repertoire (i.e., vocabulary, grammar, syntax), with little regard for function. From the traditional perspective, language development is the function of hypothesized innate developmental, neurological, and cognitive structures. We use the term language deficit or delay when a learner's verbal repertoire is deficient in comparison to the verbal repertoire of a same-age, typically developing learner. Proponents of a traditional-language approach often relate language deficits to genetic or neurological abnormalities (Lord, Cook, Leventhal, & Amaral, 2000). Intervention tends to focus on the acquisition of forms, from sounds to words to larger structures, with less attention to the behavioral function of such forms (American Speech-Language-Hearing Association, 2016). By contrast, the behavioral approach focuses on contingency or functional analysis of language or verbal behavior. Specific environment-behavior relations or contingencies of reinforcement are responsible for language development, according to the behavioral approach. Although Skinner (1957) did not review language deficits extensively in his book, researchers have applied his analysis to teaching language to children and adults with language difficulties (Barbera & Rasmussen, 2007; Carr & Miguel, 2013; Frost & Bondy, 2002; LaFrance & Miguel, 2014; Sundberg, 2008; Sundberg & Partington, 1998). However, some behaviorally oriented language training programs (e.g., Lovaas, 2003; Maurice, Green, & Luce, 1996) have not used Skinner's analysis as a framework for teaching verbal behavior. In our view, Skinner's analysis is essential to developing successful programs for training verbal behavior. In the sections that follow, we illustrate the basic tenets of Skinner's approach to verbal behavior and illustrate how we may apply it in specific training protocols. ### Primary Verbal Operants Skinner (1957) identified and named six types of functional relations between controlling variables and verbal responses. These consist of the mand, tact, intraverbal, textual, echoic, and audience relations. He named two more in the section on transcription: copying a text and taking dictation. Skinner referred to these relations as verbal operants, because he classified them by the antecedents and consequences that control their form (i.e., operant behavior). A mand is a verbal operant that a characteristic consequence reinforces, and the relevant conditions of deprivation or aversive stimulation control the response. Deprivation or aversive stimulation are motivating operations, or events that alter the value of a reinforcer. Establishing operations increase the value of a stimulus as a reinforcer. Abolishing operations decrease the value of a stimulus as a reinforcer (Michael & Miguel, 2020; Michael, 1993). For example, behavior that produces water, such as touching a communication card that signals the therapist to deliver water, is likely to increase after consumption of salty snacks, which is the motivating operation in this example. The learner is likely to emit behavior that has produced water in the past under these conditions. Newborn babies and crying provide another example. Although newborns cry reflexively, they learn to cry when hungry when crying produces food. In this example, food deprivation is the motivating operation, which will control crying if crying produces food. These examples show that mands develop when specific response forms, such as touching a communication card or crying, produce specific consequences, such as water or food, respectively. Mands are unique among verbal operants, because mands are controlled by relevant motiving operations. By contrast, discriminative stimuli control other verbal operants, such as tacts and intraverbals. Another difference between mands and other verbal operants is that specific stimulus forms reinforce mands. By contrast, nonspecific, generalized stimuli reinforce other verbal operants. The tact is a verbal operant in which a response of a given form, such as vocal, sign, or writing, is controlled by a nonverbal stimulus or "a particular object or event or property of an object or event" (Skinner, 1957, p. 82). The presence of a car, for example, increases the likelihood of the learner's emitting the vocal, signed, or written response car. The object evokes the response because the English-speaking verbal community reinforced this specific verbalization in the presence of this object. An object may serve as a discriminative stimulus for various response forms. A toy car can evoke not only the response car, but also other tacts such as vehicle, red, and fast. Environmental stimuli are likely to control many verbal responses, and we discuss this issue more fully below. Skinner (1957) also illustrated that tacts can occur in the presence of novel objects or events to which the speaker has not been exposed previously if the novel object shares physical properties with the original stimulus. For example, a child may give the verbal response car in the presence of a novel exemplar because it shares common physical properties with the stimulus that was present when the child learned to say, sign, or write car. In his book, Skinner referred to this type of stimulus generalization as generic tact extension. Other types of extensions include metaphorical and metonymic extensions. A metaphoric extension occurs when the novel stimulus shares some, but not all, characteristics with the original stimulus. The word surfing when referring to the Internet is an example. A metonymical extension occurs when the novel stimulus does not share any physical similarity with the original one. Rather, they just happened to appear together during the acquisition of the tact. For example, the sentence The White House released a statement is equivalent to The President released a statement, because the President and the White House usually appear together. Skinner (1957) identified other verbal operants whose forms, such as what a learner says or writes, are all evoked by verbal discriminative stimuli (i.e., the products of someone else's verbal behavior). The echoic relation is one of these operants in which the speaker's behavior is controlled by the auditory stimulus arising from someone else's vocal behavior. In the case of the echoic, the response bears formal similarity to the stimulus. For instance, a girl says, "Ball," after her teacher says, "Ball," when there is no ball in view. Copying a text is like the echoic in that the response form also bears formal similarity with the stimulus, and reinforcement may also depend on close correspondence between the stimulus and response. For instance, a girl writes the word ball after seeing the printed word ball. Taking dictation is writing what someone says. Writing ball when hearing someone say, "Ball," is an example. There is still correspondence between the stimulus and response, because specific sound patterns control specific hand movements. There is no physical similarity between them, however, because the stimulus is auditory, and the response produces a printed word or visual stimulus. Like taking dictation, the textual relation consists of reading printed words. The intraverbal relation is a response form evoked by a verbal discriminative stimulus. The stimulus and response are different words and do not resemble each other. For example, a parent asks, "What day is it today?", and the child responds, "Thursday." Or the teacher says, "Two, four, six...," and the student responds, "Eight." In these cases, there are no parts or subdivisions of the stimulus controlling parts or subdivisions of the response. Even though all verbal operants except the mand are maintained by generalized conditioned reinforcement, Skinner suggested that "the action that a listener takes with respect to a verbal response is often more important to the speaker than generalized reinforcement" (p. 151). In other words, verbal behavior is sensitive to listener behavior, such as a listener's nodding as a speaker provides directions to a restaurant. This observation suggests that if we want to establish an effective speaker repertoire, we must establish listener reactions as effective forms of reinforcement (e.g., Maffei, Singer-Dudek, & Keohane, 2014). ### The Listener Emphasizing the role of the listener in these primary verbal relations, including the sources of control for the listener's behavior, is important. According to Skinner (1957), an analysis of both speaker and listener behavior is necessary to understand the "total verbal episode" (p. 36). The mand primarily benefits the speaker, such as when a girl requests and receives water. By contrast, other primary verbal operants, such as the tact, largely benefit the listener. For example, if a speaker says, "It's going to rain," in response to a dark sky, the listener may contact reinforcement that he or she would not have contacted otherwise. The stimulus may evoke listener behavior such as carrying an umbrella or delaying a walk outside. The reinforcer for the listener is avoidance of an aversive stimulus: getting wet. Thus competent speakers extend a listener's contact with the stimulating environment, and this effect on the listener may serve as an important yet subtle source of reinforcement for the speaker's verbal behavior. Additionally, effective programs for training language must explicitly teach the speaker to engage the listener. For example, after beginning speakers learn to make requests to attentive listeners, they may need to learn to recruit attention from inattentive listeners. In applied settings, increasing the consistency of the communicative partner's attention and responses to the speaker's communicative attempts is an important intervention goal (Goldstein, Kaczmarek, Pennington, & Shafer, 1992). This is necessary to support the efforts of the speaker in his or her communicative attempts. Importantly, effective communicators can act as both speakers and listeners, which allows them to understand (i.e., react to) their own verbal behavior. Unfortunately, young children or those with developmental disabilities may not speak readily after learning to react to words as listeners, or vice versa (Petursdottir & Carr, 2011). This suggests that when attempting to establish functional verbal skills, we must teach speaker and listener behavior simultaneously (e.g., Fiorile & Greer, 2007). ### Multiple Control The verbal operants discussed thus far involve control by a single variable, such as a motivating operation or a verbal or nonverbal stimulus. However, most verbal behavior involves different topographies under control of multiple variables (Michael, Palmer, & Sundberg, 2011). Multiple control occurs when a single response is controlled by more than one variable, or a single variable controls more than one response (Bondy, Tincani, & Frost, 2004; Skinner, 1957). The first type of multiple control, convergent control, occurs when the verbal community arranges reinforcement for a response form in the presence of more than one stimulus. For example, a girl receives reinforcement for saying, "Ball," in response to the printed word ball, a picture of a ball, and the question "What do you throw?" The second type of multiple control, divergent control, occurs when a given variable strengthens multiple responses. An example is when liquid deprivation strengthens verbal responses such as Water please, May I have a drink?, and I'm thirsty. Multiple control that produces impure verbal operants, (i.e., those that more than one variable strengthens simultaneously) may be present when teaching verbal responses (Skinner, 1957, p. 151). For example, a boy who is liquid-deprived is more likely to say, "Juice, please," if we present him with a cup of juice than when deprivation or the cup of juice is presented alone. The motivating operation and nonverbal stimulus have a combined effect, producing an impure verbal operant, which is the mand-tact. Several variables may combine to produce a response. When a mother asks, "Would you like some juice?" with a cup of juice present, and her daughter says, "Yes, please," we may call the girl's response a mand-tact-intraverbal if it is multiply controlled by a motivating operation (deprivation), nonverbal stimulus (cup of juice), and verbal stimulus (mother's question). In this manner, we can identify the potential sources of control and establish multiply controlled operants. The relation among the different verbal operants is important in teaching verbal skills. For example, does acquisition of a response form under one set of variables lead to emission of the same response form under a different set of variables? There is considerable evidence that verbal operants are functionally independent under certain conditions (Kelley, Shillingsburg, Castro, Addison, & LaRue, 2007; LaFrance, Wilder, Normand, & Squires, 2009; LaMarre & Holland, 1985; Petursdottir, Carr, & Michael, 2005; Twyman, 1996). That is, a response topography (i.e., word) learned with one set of controlling variables will not necessarily occur in the presence of different variables unless it is explicitly taught. A response topography taught as a mand, therefore, will not occur automatically as a tact, or vice versa. For example, a boy who is taught to say, sign, or write tree in response to the question "What is it?" and a picture of a tree may not reply with tree in the presence of the picture of the tree by itself (tact), because the learned response involves a different controlling relation-the picture of the tree plus the question. The analysis of multiply controlled verbal operants becomes particularly relevant in the design of communication training programs. Even though many behaviorally oriented language training programs seek to establish complex verbal operants (e.g., Leaf & McEachin, 1999; Lovaas, 2003; Maurice et al., 1996), they do not describe procedures to transfer control from the question plus object to the object itself. Without explicit procedures for transfer of stimulus control, the learner is likely to develop a highly selective repertoire in which tacts will occur only in the presence of objects accompanied by questions. By contrast, a more functional tacting repertoire involves response topographies that occur in several stimulus combinations, including presentation of the object alone. Intraverbals are usually under the control of multiple verbal stimuli. For example, the response toast in the presence of the question "What do you eat for breakfast?" must be under control of both the verbal stimuli eat and breakfast. Although the stimulus eat may evoke several responses (divergent control), including, pasta, pizza, toast, and broccoli, the addition of the stimulus breakfast will serve to strengthen the response toast (convergent control). In the absence of specific instructions to distinguish between both features of this compound verbal stimulus, eat and breakfast, a learner may attend to only one component (i.e., eat) and continue to respond with toast to any question that includes the stimulus eat (e.g., Q: "What do rabbits eat?" A: toast). Behavior analysts can prevent this kind of failure, or rote responding, by teaching learners to respond to both features of the complex verbal stimulus at the onset of instruction (Axe, 2008; DeSouza, Fisher, & Rodriguez, 2019; Sundberg & Sundberg, 2011) The determination of controlling variables helps direct teaching protocols for several relevant situations. For example, a seemingly simple verbal skill, such as Martin's learning to say, "Swing," may have many possible sources of control: 1. Martin learns an echoic. He says, "Swing," when the teacher says, "Swing," and the teacher provides praise. 2. Martin learns an echoic-mand. When his teacher says, "Swing," he says, "Swing," and his teacher puts him on the swing. 3. Martin learns a mand. Martin says, "Swing," to his teacher, and his teacher puts him on the swing. 4. Martin learns a mand-tact. Martin sees a swing and says, "Swing," to his teacher. His teacher puts him on the swing. 5. Martin learns an intraverbal-mand. His teacher asks, "What do you want to do?", without a swing in sight. Martin says, "Swing," and his teacher puts him on the swing. 6. Martin learns a pure tact. Martin sees a swing, and he says, "Swing." The teacher says, "Yes, I see it too," but does not put him on the swing. 7. Martin learns an intraverbal-tact. When his teacher points to a swing and asks, "What is that?", Martin says, "Swing." He teacher provides praise, but does not put him on the swing. 8. Martin learns an intraverbal-mand-tact. Martin sees a swing, and his teacher asks, "What do you want to do?" Martin says, "Swing," and his teacher puts him on the swing. Martin's teacher must determine which variables are relevant and explicitly arrange each variation to teach Martin to say, "Swing." ### Selecting a Response Modality Skinner (1957) wrote that modality is not a determinant of whether a behavior can function as a verbal operant. Therefore, gestures, sign language, picture-based communication, and digitally based modalities of communication may all function as verbal behavior (Tincani & Zawacki, 2012). There is little reason to suggest that one modality of communication is inherently better than other ones (Bondy et al., 2004; Tincani, 2004). Rather, some learners may demonstrate higher rates of acquisition with one modality, and others with a different modality (Lorah et al., 2013). Of course, the most common modality of verbal behavior is speech. In typical development, infants acquire nonvocal mands, such as pointing or gesturing to an object while looking back and forth between the object and parent, before developing specific spoken words (Mundy, 1995). Speech arises out of babbling sounds that appear to be species and not culturally specific. That is, young children around the world tend to produce similar sounds, some of which produce direct or automatic reinforcement in particular language groups (Miguel, Carr, & Michael, 2002; Werker & Tees, 1999). Many issues may interfere with typical language development, from structural problems associated with oral functioning to difficulties in acquiring imitative repertoires (Fey, 1986). When learners do not speak, most interventionists try to promote speech first (Mirenda, 2003). When learners do not acquire speech via the typical pattern, researchers have identified several promising strategies to promote verbal behavior, though none is universally effective (Wankoff, 2005). Broadly speaking, such strategies encourage speech production, often without regard to function. The therapist engages the learner in various playful and reinforcing activities to increase the learner's production of sound. If sounds occur, the therapist attempts to reinforce their frequency. Next, the therapist teaches an echoic repertoire. The therapist makes a sound and provides reinforcement if the learner makes the same sound. The therapist then teaches the learner to blend sounds together in increasingly complex patterns, forming words and then short phrases (e.g., "Want cookie"). Several factors may make speech a difficult modality to acquire for many young learners, particularly those with intellectual or developmental disorders. Speech production requires refined coordination of many actions, including breath and oral-motor movements. A generalized imitative repertoire is particularly critical to speech development (Garcia, Baer, & Firestone, 1971; Young, Krantz, McClannahan, & Poulson, 1994). Teaching a learner an echoic repertoire is difficult until the learner displays generalized imitative responding. If the learner does not display generalized imitative responding, he or she is not likely to acquire a comprehensive vocal repertoire. The onset of speech via babbling appears to be relatively time-restricted and may not be available when some learners with disabilities begin communication training (Werker & Tees, 2005). We should consider training verbal behavior via other modalities if speech is a low-probability option. Most speakers use body language and gestures, which Skinner (1957) called autoclitics, to modify the meaning of their statements. We can shape these behaviors into sign language, either conforming to the grammar of a specific language or involving unique grammatical rules. One potential advantage of this approach is that sign language involves topographically different movements of the hands, thus requiring no external support (Sundberg & Partington, 1998). Modalities that use visual icons can include pictures, symbols, or print media. Some systems are low-tech, such as the Picture Exchange Communication System (PECS; Frost & Bondy, 2002); others are high-tech or digital, such as speech-generating devices (Alzrayer, Banda, & Koul, 2014; Tincani & Boutot, 2005). Learners can acquire writing skills with either a keyboard or a writing implement without prior use of speech (Lovaas & Lovaas, 1999). A behavior analyst should consider the ease of and the necessary prompts for response acquisition and the role of the verbal community's responses to a learner when selecting a communication modality (Mirenda, 2003). Although we currently have little empirical evidence to guide modality selection (cf. McLay et al., 2017), the behavior analyst should consider the following issues. First, a learner who does not have a generalized vocal imitation repertoire may not learn speech readily. In such instances, the behavior analyst should consider an augmentative and alternative communication system. Sign language also may be a viable alternative for learners who lack a vocal imitative repertoire. However, a learner must have a generalized motor imitation repertoire to acquire a functional sign-language vocabulary because of its topography-based nature. The behavior analyst should consider an aided or device-based alternative and augmentative communication system if the learner does not display vocal and motor imitation skills. Aided systems include picture-based systems, such as PECS (Frost & Bondy, 2002), and speech-generating devices (Lorah, Parnell, Whitby, & Hantula, 2015). Relevant factors for selecting an aided system may include (1) the availability of a device or system; (2) the ease of use for the primary listeners, such as parents, siblings, or teachers; (3) the potential for the device or system to accommodate several communication symbols and vocabulary; and (4) the capability of the device or system to produce a repertoire of independent, functional verbal behavior for the listener (Tincani, 2007; Tincani & Boutot, 2005). ### Teaching the Mand The mand directly benefits the speaker; thus assessment of the people, items, and events that function as reinforcers for the learner is critical before teaching the mand. Researchers have developed systematic strategies to assess learners' preferences for reinforcers, though a detailed description of these is beyond the scope of this chapter (see Saini, Retzlaff, Roane, & Piazza, Chapter 10, this volume). Extensive research has focused on teaching the mand as a component of functional-communication training (Durand & Merges, 2001; Greer, Fisher, Saini, Owen, & Jones, 2016; Mancil, 2006; Saini, Miller, & Fisher, 2016). The primary purpose of functional-communication training is to reduce challenging behavior by teaching alternative responses (i.e., mands) that produce the same reinforcing consequences as challenging behavior (see Fisher, Greer, & Bouxsein, Chapter 20, this volume). Other researchers have evaluated procedures to establish independent or spontaneous mand repertoires for learners who do not engage in challenging behavior. Several studies have established the efficacy of time-delay prompting procedures for teaching mand repertoires to learners with developmental disabilities (e.g., Halle, Baer, & Spradlin, 1981; Halle, Marshall, & Spradlin, 1979; Kratzer, Spooner, & Test, 1993; Landa, Hansen, & Shillingsburg, 2017; Shillingsburg, Marya, Bartlett, & Thompson, 2019). For instance, Halle et al. (1981) used a 5-second time-delay prompting procedure to teach learners with intellectual developmental disorder and language delays to mand in a naturalistic setting. The teacher arranged a cue likely to evoke a mand, such as approaching the learner with a cup of juice. The teacher waited 5 seconds for the learner to perform the mand before delivering a prompt. Progressive time-delay prompting or errorless teaching (Karsten & Carr, 2009; Touchette & Howard, 1984) is a variation of the time-delay procedure, in which the teacher presents the cue simultaneously with the prompt and gradually increases the duration between the cue and prompt as the learner makes independent responses. Such time-delay procedures use establishing operations (Michael, 1993) to promote mand acquisition. The interrupted-chain procedure (Hall & Sundberg, 1997) also uses establishing operations, in which the teacher prevents the learner from completing a behavior chain until the learner performs an appropriate mand. For example, a teacher might hide a step stool from a learner who typically uses the stool to access a game on a shelf. The teacher would use this situation to teach the learner to request help in locating the missing stool. A common feature of these procedures is a systematic manipulation of the learner's environment to promote a functional manding repertoire (Carnett et al., 2017; Ingvarsson & Hollobaugh, 2013; Landa et al., 2017; Lechago, Carr, Grow, Love, & Almason, 2010; Shillingsburg, Bowen, Valentino, & Pierce, 2014; Shillingsburg et al., 2019). Three basic response forms that beginning communicators need to learn are asking for a break, asking for help, and saying (or otherwise communicating) no to an offered item or activity. These simple skills are important, because they allow the learner to exert control over his or her environment and bring the learner into contact with contingencies that will be critical for developing other skills (Hixson, 2004). Asking for a break or assistance and communicating no are under functional control of aversive events such as demands, uncompleted or difficult tasks, or unwanted items. Communicating no allows the learner to escape or avoid an unwanted item or activity. It functions as a qualifying autoclitic in a mand function (see Skinner, 1957, p. 322). When a learner requests an alternative item or activity in the context of an escape or avoidance situation, the request can also function as a form of rejection maintained in part by negative reinforcement. Such choice-making responses may have the collateral effect of reducing challenging behavior associated with escape or avoidance contingencies. Researchers have validated several strategies for teaching requesting and rejecting behaviors. Best practice integrates instruction of communicative responses into daily routines by systematically identifying all potential opportunities for a learner to engage in target behaviors and embedding instruction into naturally occurring events (Carnett et al., 2017; Sigafoos, Kerr, Roberts, & Couzens, 1994; Ylvisaker & Feeney, 1994). Strategies for creating opportunities for verbal behavior focus on teaching communication in real-life activities. These strategies include (1) delaying access to an item or activity that is present until the learner makes a request (Halle et al., 1981); (2) withholding an item necessary to complete a preferred activity (Cipani, 1988; Lechago et al., 2010); (3) blocking a response or interrupting an activity to create the need for a request (Carnett, Bravo, & Waddington, 2019; Shafer, 1995; Sigafoos et al., 1994); (4) providing only part of what the learner has requested to create a new need (Duker, Kraaykamp, & Visser, 1994); (5) intentionally giving the learner the incorrect item (Choi, O'Reilly, Sigafoos, & Lancioni, 2010; Sigafoos et al., 1994); or (6) delaying offers of assistance until the learner makes a request for help (Rodriguez, Levesque, Cohrs, & Niemeier, 2017; Sigafoos & Roberts-Pennell, 1999). The learner escapes or avoids nonpreferred items or activities in a socially acceptable manner if he or she can make a rejection response. Learners who do not communicate this need may adopt an idiosyncratic behavior that is difficult to interpret (Iacono, Carter, & Hook, 1998), or may learn to escape or avoid activities through challenging behavior (Carr, 1994). The learner is more likely to acquire socially appropriate escape or avoidance communication in situations where he or she is motivated to escape or avoid an item or activity. The behavior analyst can create a context for teaching the learner to reject an offer appropriately if the analyst can identify items or activities the learner is motivated to escape or avoid (Chezan, Drasgow, Martin, & Halle, 2016; Sigafoos & Roberts-Pennell, 1999). There are several techniques for teaching rejection or refusal behavior. First, the behavior analyst can strengthen an existing appropriate behavior to make it more effective, specific, or consistent (Warren, Yoder, Gazdag, Kim, & Jones, 1993). Second, the behavior analyst can teach new communicative behaviors by chaining the new response to an existing, inefficient behavior (Keen, Sigafoos, & Woodyatt, 2001) or by prompting (Drasgow, Halle, Ostrosky, & Harbers, 1996). Third, the behavior analyst can replace socially unacceptable communicative behaviors with acceptable, functionally equivalent behaviors. For example, the behavior analyst could teach a learner to point to the word stop on a communication board to end a task instead of throwing a tantrum (Carr, 1994). Fourth, the behavior analyst must teach the learner that escape or avoidance will not always be possible even when the learner has requested it appropriately (Sigafoos, 1998). Even though we are suggesting that we should initially teach communicating no as a mand, more advanced learners will benefit from acquiring no and yes as tacts and intraverbals. For example, we should teach a learner who can speak to say, "No," if the teacher asks, "Is this a giraffe?" when she presents a picture of a cow (tact), and to say, "Yes," when the teacher asks, "Does a cow says moo?" (intraverbal; see Shillingsburg, Kelley, Roane, Kisamore, & Brown, 2009). ### Teaching Other Verbal Operants Although much research has focused on the acquisition of mands, research also has shown that the tact repertoire is foundational for the development of other verbal and nonverbal behaviors, such as mands, intraverbals, stimulus categorization, and analogical reasoning (e.g., Finn, Miguel, & Ahearn, 2012; Greer & Du, 2010; Miguel et al., 2015; Miguel & Kobari-Wright, 2013; Miguel, Petursdottir, & Carr, 2005; Sprinkle & Miguel, 2012). Additionally, tacts may not readily emerge after receptive-discrimination instruction (Contreras, Cooper, & Kahng, 2020; Petursdottir & Carr, 2011), especially if a child lacks a generalized echoic repertoire (Horne & Lowe, 1996). For this reason, a behavior analyst should prioritize tacts when attempting to expand a learner's vocal repertoire (e.g., Greer & Du, 2010). Tact training should focus initially on preferred, familiar, and functional three-dimensional stimuli readily found in the learner's environment; it should then move toward two-dimensional complex stimuli, including functions, features, relations, and private events (LeBlanc, Dillon, & Sautter, 2009). Behavior analysts should teach echoics early in programming when teaching vocal tacts, because therapists then can use echoic prompts for other verbal operants (e.g., Kodak & Clements, 2009). Although researchers have studied generalized vocal imitation (e.g., Kymissis & Poulson, 1990), more recent investigations have focused on procedures to establish vocalizations in learners who are otherwise nonvocal. One of these procedures, called stimulus-stimulus pairing, establishes vocalizations as conditioned reinforcers. Thus the response-produced auditory stimulus may function as reinforcement for the vocalizations that produced them (Lepper & Petursdottir, 2017; Shillingsburg, Hollander, Yosick, Bowen, & Muskat, 2015). Other more naturalistic procedures evoke vocalizations by capturing the learner's interest while modeling vocal sounds (Charlop-Christy, LeBlanc, & Carpenter, 1999). Behavior analysts can teach simple intraverbals in the form of fill-in-the-blanks at the same time they are teaching simple mands and tacts. By contrast, the behavior analyst should teach more complex intraverbals after the learner can respond to the same topography or word as a listener (i.e., receptive discrimination) and speaker (i.e., tact). For example, before the analyst teaches a learner to say, "Carrot," when the learner hears, "What does a rabbit eat?", he or she should first learn to tact carrot and eat and respond as a listener when hearing these words (Petursdottir, Ólafsdóttir, & Aradóttir, 2008