As our focus is primarily on functionality, we will use the terms language and communication as synonyms. In actuality, we prefer the term verbal behavior (Skinner 1957), as it shifts attention to the variable clinicians can manipulate: the environment. The term verbal behavior was first coined by Skinner (1957) in his seminal book by the same name. In it, he conceptualized language for the very first time from a behavioral perspective, taking into consideration both context and the importance of functional consequences. Skinner’s account provided a potentially pragmatic way of teaching language and addressing impairments. It gave scientists and practitioners alike a taxonomy by which to classify language units as well as a systematic approach to measuring and, most importantly, changing language.

In his analysis, Skinner (1957) defines verbal behavior and distinguishes it from nonverbal behavior. Specifically, verbal behavior is defined as the behavior of an individual, which is reinforced through the mediation of another person’s behavior (p. 2). Otherwise stated, reinforcement can only be obtained through the actions of another individual. Skinner further refines this definition by stipulating that the behavior of the person mediating a speaker’s response (i.e., the listener) must have been specifically trained to serve that function. For example, if a thirsty child (i.e., speaker) says “water” in the presence of an adult, the adult (i.e., listener) will likely reinforce that child’s response by providing the requested item: water. The listener reacts in this manner precisely because he/she has already learned to get a glass of water when someone asks for one. It is through such means that the verbal community comes to establish control over, and influences its members’ (verbal) behaviors .

In sharp contrast, nonverbal behavior achieves its effect on the environment directly. It does not require the mediation of another person (Skinner 1957). Utilizing the same example, a thirsty child could obtain water independently through several means (i.e., getting water out of the refrigerator, pouring himself/herself a cup of water, drinking water from a fountain or faucet, etc.), none of which would necessarily require the presence of another individual.

Furthermore, Skinner (1957) makes the distinction between verbal and vocal behavior. In other words, verbal behavior may be vocal, in the sense that it engages the vocal chords, but this is not a requisite condition for behavior to be considered verbal. To illustrate this point, consider a child with little to no spoken “language,” who points to a desired item that is out of reach. If that child is in the presence of another individual (i.e., a listener), he/she will likely obtain that item the same way he/she would have had that child used spoken words. Thus, for the remainder of this chapter, the term verbal will not be used as a synonym of vocal, but to define behavior that produces mediated consequences. In other words, verbal means communicative .

Verbal behavior can be broken down into teachable units according to the specific relations between controlling variables and response forms. These relations are called verbal operants, precisely because of the way they “operate” on the environment. The elementary verbal operants consist of the mand, the tact, the echoic, the intraverbal, textual behavior (i.e., reading), and transcription (i.e., dictation and copying text). Skinner also discusses the role of the audience in the control of large portions of our verbal behavior. Of these relations, there are four which are especially relevant to the initial development of an adequate vocal-verbal repertoire. These are the mand, tact, intraverbal, and echoic, and are defined below.

The mand, according to Skinner (1957), is a type of verbal operant in which the response is reinforced by a characteristic consequence. Using more colloquial terms, this relation may be described as “requesting.” Unlike all other verbal operants, however, the mand occurs under the control of relevant conditions of deprivation or aversive stimulation (i.e., motivating operations). These conditions greatly affect the likelihood of the occurrence of the mand. More specifically, a motivating operation (MO; Michael 1982) constitutes any event in the environment, which alters both the reinforcing effectiveness (i.e., the relative value) of a given stimulus, and the relative frequency of a behavior, which in the past has been reinforced with that particular stimulus. Sticking to the example used above, a child who has not had water for some time would respond in ways that in the past has resulted in water. If in the presence of a listener, the child may request (mand for) water, and the reinforcement for the response would consist of receiving the requested item. Thus, it is said that the mand specifies its own reinforcer (Skinner 1957) .

The tact, on the other hand, is a verbal operant under the control of a prior object, action, or property (i.e., nonverbal stimulus), and whose response is, at least when learning this behavior, reinforced by some generalized form of reinforcement (i.e., praise). For example, when learning to label objects in his/her environment, a child might see a glass of water (nonverbal stimulus), and say “water,” which may result in some form of approval.

The intraverbal is a type of verbal operant under the control of a prior verbal stimulus, and whose response form does not resemble the preceding stimulus. As with the tact, the intraverbal response is initially reinforced by some generalized form of reinforcement. An adult may ask a child, for example, what his/her favorite animal is, and the child may respond by naming any animal within his/her repertoire. The adult’s original question evoked the child’s response, but neither vocal utterance (adult or child’s) resembled the other.

Finally, the echoic consists of a verbal operant under the control of a prior verbal stimulus, with which the response is identical. As with the tact and the intraverbal, the echoic response is reinforced by some form of generalized reinforcement. An individual may say “Hi” to another individual passing by, to which that passerby may respond “Hi.” These verbal stimuli are identical in form, and maintained by some type of socially mediated reinforcement such as the other’s response, or a smile.

It is of primary importance to note that the acquisition of any one type of verbal operant (e.g., tact) does not necessarily lead to the acquisition of the other verbal operants (e.g., mand). In other words, learning to say “cookie,” as when a child sees a cookie, does not necessarily mean that the child will be able to ask for a “cookie” when he/she wants one. This phenomenon is known as functional independence and has been empirically demonstrated by numerous researchers (Hall and Sundberg 1987; Lamarre and Holland 1985; Partington et al. 1994). Thus, although transfer between operants may occur readily in verbally sophisticated and typically developing learners, or under certain teaching arrangements (Finn et al. 2012; Petursdottir et al. 2005), specific training may be necessary to establish each type of verbal operant in individuals with disabilities. This is a critical factor related to the notion of understanding, as we will discuss below .

Language interventions developed according to Skinner’s (1957) taxonomy (Bondy et al. 2004; Greer and Ross 2008; Sundberg and Partington 1998; Williams and Greer 1993) have been shown to be highly effective at teaching the verbal operants described above. This model focuses on the function rather than the form of spoken language, and emphasizes the importance of context (i.e., verbal operants). The verbal operants defined above must each be learned as language develops in order to acquire an adequate basic vocal-verbal repertoire.

Since its original publication, Skinner’s work has had a significant impact on the way(s) in which language skills are taught. By placing an emphasis on context and function, his taxonomy has given behavior analysis a powerful tool by which to teach and shape language, especially in instances where the communicative repertoire is significantly impacted .

When developing a plan for teaching basic communication skills for a child with autism, one needs to first assess the learner’s current repertoire . The results obtained allow for the development of a comprehensive and parsimonious intervention program, by providing important information pertaining to areas of strength and deficit. Unfortunately, despite the abundant availability of language assessments, very few break language down into teachable units such as the verbal operants. This lack of specificity renders most standardized assessments impractical in helping to identify what needs to be specifically taught to the child (Carr and Miguel 2013; Esch et al. 2010). Furthermore, standardized assessments are limited in their time-sensitive nature (i.e., most are valid only when administered according to their intended timelines, such as every 6 or 12 months), making them inadequate in evaluating a child’s progress over time, especially when more repeated and frequent measurement is required, as is often the case with intervention programs designed for children with ASDs.

There are at least two assessments developed from a behavioral perspective which have proven to be useful curriculum guides, as they can be used not only to assess the child’s initial repertoire, but also his/her progress over time. These assessments are: (1) The Assessment of Basic Language and Learning Skills—Revised (ABLLS-R; Partington 2006) and (2) The Verbal Behavior Milestones Assessment and Placement Program (VB-MAPP; Sundberg 2008). Both are criterion-referenced assessments that evaluate a child’s learning, verbal behavior, and academic repertoires in several domains such as mands, tacts, echoics, intraverbals, play skills, etc. Each yields scores based on direct observation, probing, and data collection. Evaluation, sometimes subjective in nature, as well as the number of skills requiring observation and probing can result in a time consuming process. It should be noted that no data on these assessments’ psychometric properties have been published (Carr and Miguel 2013), and research related to these properties is needed. However, given the lack of standardized verbal behavior assessments, and the inadequacy of traditional assessments, behavioral practitioners seem to be making great use of both the ABLLS-R and the VB-MAPP as a way of identifying targets for language intervention .

Once specific targets for intervention are identified, behavior analysts develop lessons using well-researched teaching methodologies derived from behavior analysis (Cooper et al. 2007). Because language is viewed as behavior and behavior is a function of environmental variables, the goal is to create the necessary conditions for the child to learn how to communicate effectively. Some of these environmental arrangements include the use of reinforcing consequences to increase behavior and the manipulation of antecedent variables to make it more likely that the desired behavior will occur.

However, the results of any particular assessment tool often yield a greater number of skills requiring intervention and teaching than can be feasibly targeted within a given period of time (e.g., on a quarterly basis, in a 6-month period of time, from one yearly individualized education plan (IEP) meeting to the next, within the context of service provision over a period of 2–3 years). Additionally, the sheer amount of delay with respect to skills domains, and related to language particularly, often compounds this problem. Therefore, clinicians must carefully prioritize which behaviors will be addressed first, both within and across domains, while simultaneously planning out a more long-term strategy to address whichever skills and behaviors will be remaining to teach. Further, the way in which skills are taught can greatly influence the future need for teaching. For example, intentionally exposing the learner to several different examples of a particular item to be taught (e.g., a blue cup, a red cup, a small cup, a large cup, a cup made of glass, a cup made of plastic) will lead to better generalization and concept formation, than if this strategy were not employed, thereby circumventing a potential future loss of time which would otherwise have to be spent teaching the generalization of skills specifically (Greer and Ross 2008; Stokes and Osnes 1989). This aspect of target selection and prioritization poses a special problem and is a critical component of a comprehensive intervention program, as it ensures that the time spent in intervention is utilized in the most effective and efficient manner possible, by maximizing the amount of learning taking place. This particular issue is addressed by ensuring that the teaching strategies and methodologies incorporated into intervention are empirically supported and well vetted. The problem that remains, however, pertains to which skills to begin teaching. To this end, the concept of behavioral cusps (Rosales-Ruiz and Baer 1997) becomes especially helpful and relevant .

Though similar to the notion of developmental milestones, behaviorally speaking, the concept of cusps is quite different. Cusps are defined as behavior changes that systematically lead to either widespread further changes or to important further changes (Bosch and Fuqua 2001; Rosales-Ruiz and Baer 1997). Importance, in this sense, refers to a change which is valued by either the individual acquiring the cusp himself/herself, significant others in the individual’s environment (e.g., the child’s family members or teachers), the relevance of the behavior change in response to the environment (e.g., asking for a favorite toy is much more likely to be reinforced and maintained and is much more desirable in general in the social community than tantruming), or a combination of all or any of these. Unlike the notion of stages of development or milestones, the concept of cusps frees the clinician from a perspective that necessarily follows a set and predetermined sequence. Additionally, a very small change may constitute a cusp for the organism, depending upon whether that change will lead to additional change(s) .

The focus of a behavioral account of development therefore consists of the individual organism’s experience with naturally occurring contingencies in his/her environment. In other words, when the acquisition of a particular behavior brings the organism into contact with different contingencies that will be critical to the subsequent development of more complex or sophisticated behaviors, this is considered a cusp. Additionally, the concept of cusps relies on the relevance and application of contingencies throughout the individual’s lifespan. Therefore, what makes a behavior change a cusp is that it exposes the individual to new contingencies, including and especially as this pertains to accessing new reinforcers and/or punishers, new environments (such as enrollment in a preschool), new sources of stimulus control (such as the generalization of skills to the teacher), and new communities that will either contribute toward the maintenance (such as peer groups or the classroom environment), or elimination of certain contingencies (such as rehabilitation centers for individuals engaging in substance abuse). Cusps, as a result, are often identified by their effects, in that they lead to generative and/or novel behavior and facilitate future learning by functioning as “either a prerequisite or a component of more complex responses” (Bosch and Fuqua 2001, p. 124). The use of the term refers to the potential for the expansion, refinement, and recombination of other, less developed repertoires into more sophisticated ones. The net result necessarily leads to both more as well as accelerated learning and competing with the emission of inappropriate responses .

Some general examples of cusps almost every human being will contact throughout the course of his/her lifetime include learning to sit upright, learning to crawl and then to walk, learning through observation of others, and learning to read. Additional examples may include experiences of a more undesirable nature, such as being exposed to addictive substances for the first time. Although this may seem counterintuitive, such experiences inevitably expose the individual to new forms of reinforcement and punishment, as well as different contingencies he/she may otherwise never have been exposed to. The behaviors mentioned here, though non-exhaustive, are all considered cusps as they lead the individual to contact new experiences which would otherwise be outside of reach (Hixson 2004).

Within the context of behavior analytic intervention programs, some specific examples of cusps for young children may include: the reduction of behavior excesses, attending, cognitive, motor, imitation, academic, and language skills. In this sense, cusps which are of greatest relevance for a younger demographic correspond relatively closely to early developmental milestones. Examples of cusps for older children may include some of the ones listed above in addition to: social skills, self-help, daily living, and vocational skills. With this population, the focus often becomes teaching behaviors that will lead to further independence. In general then, if a child is demonstrating a large proportion of skills in a certain domain at or near age level, other domains will be prioritized, and a greater number of lessons will be developed to address this deficit area. Additionally, skills which constitute early prerequisites for the subsequent development of appropriate age-level skills, will also be prioritized (Hixson 2004) .

A multitude of other variables also enter into the equation when making decisions related to prioritization and planning. For example, some additional considerations which must be taken into account when selecting which skills to target include: the importance of the development of a particular behavior (i.e., what is the extent of the behavior change, what later changes will the behavior change enable?), whether the development of a particular skill will expose the individual to new cusps as a result of the behavior change (e.g., when an increase in attending and a decrease in behavior excesses allows the child to contact an additional number of learning opportunities), whether the individual (assessed whenever possible) and other important individuals in his/her environment consider the behavior change important, as well as whether a behavioral trap (i.e., this refers to the transfer of stimulus control to a community of reinforcement existing in the organism’s environment which can ensure the ongoing maintenance of skills as well as shape up new behaviors) can be set up to maintain the behavior change (e.g., attendance to preschool will help ensure the maintenance and development of more complex academic, play, social, and language skills). Bosch and Fuqua (2001) offer a systematic approach to both the identification and prioritization of cusps and the reader is referred to their article for a more detailed discussion.

Language and the acquisition of a sufficient and effective communicative repertoire is, quite arguably, the most important cusp, and the same problem of prioritization applies to teaching language. Following assessment, especially in the case of young children diagnosed with developmental disabilities, it is more often than not the case that a very significant deficit exists across all necessary skills required to become an effective speaker (i.e., expressive language) and listener (i.e., receptive language). Both constitute separate repertoires, and both are incredibly complex as they relate to the development of other higher-order skills (e.g., categorization, mathematics, academic skills, play and social skills, problem solving, behavior excesses, and independence) from which the vast majority of skills related to other developmental domains will be rooted. Thus, a more careful analysis of cusps must be undertaken when teaching language (Skinner 1957; Sundberg 2008). In the sections below, we will discuss the issue in further detail and offer up some practical suggestions for guiding these types of programming decisions .

As previously mentioned, the first goal of communication training is to give the child a functional way to express his or her needs. The mand is perhaps the most important of the verbal operants, as it benefits the speaker directly. It is usually one of the first to emerge developmentally, and can often lead to the acquisition of further language as well as immediate decreases in interfering behaviors. For this reason, the mand should be prioritized, emphasized, and taught early on in intervention. This can be accomplished by using a variety of communication modalities such as vocalizing, signing, the use of augmentative devices, or the use of picture exchange systems (Mirenda and Dattilo 1987). Vocal behavior is always the preferred method (Shafer 1993), but some children show great difficulties with vocal production and articulation, as well as vocal imitation (Smith and Bryson 1994), which could greatly delay the acquisition of functional speech and delay the child’s cognitive progress. In these cases, either sign language or a picture-based system could be employed. Deciding whether pictures or signs should be used has been a matter of controversy among researchers and clinicians alike for a number of years now (Michael 1985; Shafer 1994; Sundberg and Michael 2001; Tincani 2004; Ziomek and Rehfeldt 2008). The debate lies in determining whether one system is superior to another in terms of rates of acquisition and the potential for emergence of vocal language (i.e., a cusp skill). While some research has pointed to the superiority of sign language over symbolic systems in terms of acquisition and accuracy (Potter and Brown 1997), other studies (Adkins and Axelrod 2002; Tincani 2004; Vignes 2007; Ziomek and Rehfeldt 2008) have suggested that children with autism acquire symbolic systems such as the Picture Exchange Communication System (PECS; Bondy and Frost 2001) quickly and with a high degree of accuracy. Thus, when choosing a mode of communication, clinicians should pay careful attention to a number of variables, including the child’s motor skills, scanning abilities, accessibility and portability of the communication system itself, as well as the likelihood of having a verbal community that would be able to communicate with the child fluently utilizing the selected modality (Carr and Miguel 2012). It is important to note that the use of either signs or symbolic systems does not prevent a child from continuing to learn vocal behavior or speech. Thus, the use of symbolic systems may be an efficient and effective way of teaching functional communication to a child who otherwise lacks the ability to express his/her needs in an appropriate way (LeBlanc et al. 2009).

Once the mode of communication is selected, clinicians should decide what specific topographies or words to teach. Because many problem behaviors initially displayed by children diagnosed with developmental disabilities may have social and communicative functions (Durand and Carr 1985), it makes sense to select mand targets that would replace or serve the same function as problem behavior (Hagopian et al. 1998).

Many recent studies have focused on the selection of the communicative response and how this selection can influence the efficacy of training (Harding et al. 2009). Variables to consider when selecting the alternative response topography include: the amount of effort required, novelty, history of reinforcement with previous topographies, and preference. Functionally equivalent alternative communicative responses that require less effort than the problem behavior to obtain reinforcement are more likely to achieve four specific results: (1) they are likely to be acquired quickly, (2) they are likely to be maintained, (3) they are likely to lead to a more rapid decrease in behavior excesses, and (4) they are likely to contribute to the maintenance of low rates of problem behavior (Hernandez et al. 2007; Horner and Day 1991; Richman et al. 2001). The likelihood of an individual choosing to engage in the alternative communicative response is inversely proportional to the amount of effort required to engage in that response, relative to the problem behavior. In other words, the less effortful the response, the more likely the individual is to engage in that response. Therefore, the alternative response must require less effort compared to the problem behavior in order for the former to replace the latter. For this reason, mand topographies previously displayed by an individual should take precedence over novel mand topographies, even when existing mand topographies are associated with higher rates of problem behavior (Winborn et al. 2002). Since this intervention is highly individualized, results of preference assessments of mand topographies are idiosyncratic, with some individuals preferring to engage in vocal behavior or signs (Harding et al. 2009) and others indicating preference for methods such as pressing a microswitch (Falcomata et al. 2010).

Once functional mands have been taught, the client should then learn to mand for a variety of preferred items using single words across many settings and people. This is because it is more advantageous to the child, in terms of control over his/her environment to have a variety of one-word mands, than to have a limited number of mand sentences (LeBlanc et al. 2009). Clinicians should also avoid teaching abstract generalized mands such as “more” and “please” in the beginning of an intervention program since these mands are nonspecific. When learning these mands, the child learns to request whatever preferred item he or she wants, but only when the item is present. Therefore, the child becomes dependent on having the item in sight to mand. Additionally, this may discourage children to learn to mand using the item’s name (Barbera 2007).

Several studies on how to teach mands for items and information have been conducted (e.g., Sundberg et al. 2002; Lechago et al. 2010). One of the most studied procedures is called “interrupted chain” (Hall and Sundberg 1987) in which after learning to engage in a chain of behaviors (e.g., making a snack), materials needed to complete a step in the chain are made unavailable, therefore generating the ideal antecedent condition for teaching the behavior. For example, one may teach a child to serve himself/herself a bowl of ice cream, but withhold the spoon. This procedure has been recently adapted to teach mands for information such as who, which, where, and how (Lechago et al. 2010). For example, Lechago et al. taught participants to mand for a spoon utilizing an interrupted chain procedure targeting one skill. They subsequently assessed for the generalization of the specific mand taught (i.e., asking for a spoon) across multiple establishing operations² (EOs; Laraway et al. 2003) by evoking the response in two different behavioral chains. Their results not only showed that all participants generalized the mand across EOs when the specific mand was evoked by chains of responding which required provision of a spoon for completion, but also that it is possible for new mand topographies to emerge following this type of training. Using this experimental arrangement, the authors were able to demonstrate that a specific mand form can be taught and generalized across different EOs without the need for additional training, and that the teaching of a specific mand topography may also lead to the development of additional mands.

Mand training opportunities can also be captured as opposed to contrived (Sundberg and Partington 1998, 1999; Barbera 2007). More specifically, when opportunities to mand are set up in such a way that the antecedent conditions are structured by the interventionist, we call these contrived. Examples include placing highly preferred items in clear containers that the child cannot open independently, setting these items in sight but out of reach, or providing only some of the pieces required to complete a preferred activity (e.g., puzzles). In these situations, opportunities are contrived as they are specifically arranged to set the occasion for the child to mand and receive reinforcement by establishing an EO. This method is exceptionally helpful in teaching a foundational repertoire consisting of mands for preferred and desired items. When naturally occurring opportunities are capitalized upon, however, these are considered captured. An example of this type of arrangement consists of blocking a child’s access to an item after he/she has already initiated reaching for that item. This method allows one to guarantee an EO is in place and utilize the opportunity to teach manding. It is often regarded as being much more natural than the former strategy, as it is embedded throughout the course of whatever activities would typically already occur and appears to be more fluid. Mands for the cessation/removal of undesired or aversive stimuli are most appropriately taught using this method, so as to prevent the inadvertent establishment of inappropriate behavior excesses as a more effective means of achieving this outcome (LeBlanc et al. 2009). Usually, a combination of these two methods is utilized to help maximize the total overall number of learning opportunities presented and ensure that manding is strengthened in the behavioral repertoire.

In certain instances, a child may not necessarily have already learned the words required to mand, or may not yet produce these particular vocalizations. As mentioned above, the use of a symbolic system such as sign language or PECS can often act as a “bridge” to vocal production (Charlop-Christy et al. 2002). However, if a child is observed to produce vocalizations and to imitate, vocal mands can be shaped up through the use of transfer of stimulus control and errorless teaching procedures (Green 2001). Generally speaking, in these types of procedures, prompts that reliably evoke the desired behavior are used to occasion responding in a certain condition. A delay is then typically inserted between the presentation of the antecedent stimulus (which is the one intended to occasion responding) and the prompt (which is the one controlling responding). The delay is systematically and gradually increased until such a point that the response solely occurs in the presence of the antecedent.

So, when teaching a child to mand for a missing item necessary to make a snack (e.g., a spoon), the teacher can prompt the child to “say spoon” and deliver the spoon contingent on the correct vocalization. In this case, the child’s response was solely occasioned by the clinician’s prompt, but it also occurred in the presence of the right motivational variable (wanting the spoon), and produced the specific desired consequence (the spoon). With careful programming, the clinician can slowly delay his/her echoic prompt which would cause the child to eventually respond prior to the prompt, ensuring that the child’s responding eventually comes under sole control of his/her motivation to obtain the item. This systematic delaying of a prompt is referred to as a “prompt delay procedure” (Cooper et al. 2007), and has been used to teach a variety of behaviors (Touchette and Howard 1984), and there are specific guidelines on how to use it to maximize its effectiveness (Karsten and Carr 2009). More generally speaking, these types of procedures can be used to “jumpstart” verbal operants that are weak in the repertoire by taking advantage of other operants that are consistently and reliably demonstrated.

Otherwise stated then, echoics can be capitalized upon as a prompting procedure to teach intelligible spoken mands. Once a robust repertoire of echoic behavior is established, spoken language eventually replaces the use of communication systems, as it generally requires less effort both in production and in portability, and is much more likely to come into contact with continued reinforcement across environments.

As alluded to in the section above, the ability to imitate someone else’s vocal behavior plays an important role in the development of other forms of verbal behavior (Hixson 2004; Horne and Lowe 1996; Lowenkron 1998). Typically developing children often learn to label items (tacts) when imitating their caregivers’ labeling of a particular item in its presence (Horne and Lowe 1996). Additionally, when a child with disabilities is able to successfully imitate, then his/her vocabulary can be greatly expanded. Instead of having to slowly shape sounds into full words using differential consequences, the clinician is able to simply prompt the child vocally, by telling him/her what to say in a specific condition.

Because echoic behavior serves an important function in expanding the child’s functional verbal repertoire, its absence is concerning to those attempting to develop vocal communication or speech. Over the years, many studies on how to teach vocal imitation were published (Kymissis and Poulson 1990; Young et al. 1994) and described in teaching manuals (Lovaas 2003; Sundberg and Partington 1998).

When children are nonresponsive to these procedures, clinicians may want to consider switching to a different mode of communication (symbolic or signs) while still attempting to teach vocal behavior in this context (Charlop-Christy et al. 2002). Alternative procedures such as stimulus–stimulus pairing (SSP; Smith et al. 1996) as well as the natural language paradigm (NLP; LeBlanc et al. 2006) can also be employed to jumpstart vocal behavior.

In the SSP procedure, the therapist models a target sound which can be a full word, or just a phoneme (e.g., “ba”) and then gives the child a preferred item regardless of whether or not the child imitates. Repeated sequences of the target sound and access to the preferred item establish a close relationship along a temporal dimension (i.e., contiguity), turning the sound into a conditioned reinforcer which may lead the individual to produce that sound more frequently (Miguel et al. 2002; Smith et al. 1996). Although the method does not seem to produce consistent results (e.g., Carroll and Klatt 2008; Esch et al. 2005; Normand and Knoll 2006), some authors (Esch et al. 2009) have suggested that by ensuring that participants are attending to the experimenter, presenting the vocal response immediately followed by the edible, and presenting a nontargeted vocalization not followed by the presentation of the edible, may enhance the effects of the procedure.

As described by Charlop-Christy et al. (1999), the NLP procedure consists of having the clinician play with the child in a naturalistic setting such as setting up a multitude of toys and/or games on the floor. During play, the clinician offers a choice of three objects/toys and allows the child to make a selection. Once a selection is made, the clinician subsequently models appropriate play actions and related vocalizations (e.g., “vroom, vroom!” while rolling a car on the ground). Any child-directed attempt to vocally imitate results in immediate reinforcement; typically in the form of gaining brief access to the item selected. This same process is repeated, with the clinician modeling actions and vocalizations following each interchange. Over time, the clinician fades his/her vocalizations so that the stimuli which come to evoke the child’s vocal responding consist of the objects and actions in play. Additionally, the intelligibility and accuracy of the child’s utterances are typically shaped gradually through the use of echoic prompts (see the description of transfer of stimulus control procedures in the section on teaching mands above).