A lot of talk exists in the world of autism services about different methodologies, different approaches, and different theories of what works. Most noticeably, there have recently been many discussions about particular behavioral interventions and their differences. The acronyms relevant to this site that are most often discussed are "ABA," "AVB," "DTT," and "NET." So what is AVB? In actuality, there is no such thing, not if having its own acronym implies that it is something separate from the science of behavior analysis. The term "AVB" has apparently become a shorthand for a program of applied behavior analysis that focuses on teaching verbal behavior through a collection of highly effective teaching procedures taken from the science of behavior analysis. In that case, AVB is ABA, plain and simple. Most, if not all, good ABA programs incorporate most, if not all, of the effective teaching procedures described elsewhere on this site. While it is important to specify that there are ABA programs that do and do not incorporate the teaching of verbal behavior based on Skinner's analysis of language, it would be unfortunate if people thought of teaching VB as anything other than ABA. So if you come across the term "AVB," know that it probably refers to ABA with a focus on teaching verbal behavior, but is truly simply ABA.
That said, what is ABA and what is verbal behavior? In 1938, Skinner published The Behavior Of Organisms, which described operant conditioning, or the process by which learning occurs as the result of selection by consequences of behavior. Skinner also discussed how antecedent stimuli, when correlated with the function altering effects of consequences, also alter future occurrences of that behavior. This is known as a three-term contingency (A-B-C), the basic unit of analysis of behavior, and was the first description of the discrete trial. In addition to describing the instructional trial, Skinner detailed the basic experimental methodology that led to his findings, which he termed the experimental analysis of behavior (EAB). Later applications of this science to education, and to other matters of socially significant behavior, by behavior analysts led to what is now known as Applied Behavior Analysis (ABA).
In 1957, as applied behavior analysis was developing and research on ABA was being published, Skinner published Verbal Behavior, which detailed a functional analysis of verbal behavior. What Skinner's text did was to extend operant conditioning to verbal behavior in order to fully account for the range of human behavior. Since the publication of Verbal Behavior, many applied behavior analysts, including Jack Michael, Mark Sundberg, Jim Partington, and Vince Carbone, have conducted and published research on verbal behavior, much of which can be found in The Analysis of Verbal Behavior journal. This body of research serves as the basic and applied foundation of teaching VB as part of an ABA program, or what is now sometimes referred to as AVB, as discussed above. The science of applied behavior analysis now has a solid empirical foundation to support it, due largely in part to Skinner and Ivar Lovaas.
In Verbal Behavior, Skinner outlined his analysis of VB, which describes a group of verbal operants, or functional units of language. Skinner explained that language could be analyzed into a set of functional units, with each type of operant serving a different function. He coined terms that didn't exist (to separate these operants from anything described by traditional linguistics) for these operants. AVB is ABA with a focus on Skinner's analysis of verbal behavior; it is the application of the science of behavior analysis to teaching verbal behavior. While there is some debate among behavior analysts as to the merits of the "AVB" acronym, most of the concern revolves around the potential for an artificial distinction between ABA in general and VB in particular. To me, as long as it's understood that they are the same science, it's reasonable, though unnecessary, to discuss the application of Skinner's functional analysis of VB as "AVB," much the way we discuss the application of Skinner's science of behavior analysis as ABA. Whatever it's called, this application can look quite different from that of language instruction through discrete trial teaching (DTT); some of those differences are discussed in the next section. The primary verbal operants, which are most often initially discussed in relation to teaching children with autism, are echoics, mands, tacts, and intraverbals. Here I will briefly explain the functions of these operants and how they may be taught. I also want to touch on the establishing operation and how it relates to teaching language.
In order to learn any skill, a child must have an imitation repertoire. Without imitation it is nearly impossible to teach anything. This is especially true for teaching language. To learn to sign, for example, a child needs to develop a good motor imitation (mimetic) repertoire; to learn to speak, the child needs a strong vocal imitation (echoic) repertoire. The echoic is the verbal operant that relates to vocal imitation. An echoic is verbal behavior whose form is controlled by someone else's verbal behavior with point-to-point (1:1) correspondence. What this means is that the child echoes exactly the speech of the teacher. For example, the teacher says, "Cookie" and the child says, "Cookie." In order for speech to be reinforced, it must occur. The echoic provides us with a mechanism for evoking speech such that we may reinforce it. If, for example, we wish to teach the child to say, "Mommy," but there is no echoic repertoire, we would have to wait until the child said, "Mommy" on his own and then reinforce it strongly. If instead we teach the child to develop a strong echoic repertoire, we can repeatedly say, "Mommy," the child can echo, "Mommy," and we can reinforce it many times, thus increasing the probability of the behavior (the word Mommy) in the future. Thus, you can see how a strong echoic repertoire is critical in teaching new language, since the child's ability to imitate vocally allows the teacher to create many opportunities for the child to use and be reinforced for speech. Echoics are key in teaching the other verbal operants as well.
The mand is verbal behavior whose form is controlled by states of deprivation and aversion; it is often said to "specify its own reinforcer." What this means loosely is that the function of a mand is to request or to obtain what is wanted. So if a child says "Cookie," and it is functioning as a mand, that means the child is requesting the cookie. Think of mand as short for "demand" or "command." The way to reinforce a mand is to deliver the item manded for. So if a child says "Cookie," you'd give him a cookie. This positive consequence (reinforcement) of the mand will make it more likely that the behavior will occur in the future, i.e., that the next time the child wants a cookie, he will say cookie. So you can equate a mand with a request. We mand for a great many things every day without really thinking of them as mands: Desired items ("I want pizza for dinner"); information ("What time is it?"); assistance ("Can you help me"); missing items (given a bowl filled with cereal and milk, the child says "I need a spoon"); actions ("Play with me"); attention ("Mommy, look what I did"); negative reinforcement (removing something undesired/aversive) ("Turn off that loud music!"), etc., etc., etc. Manding is typically a first step in teaching language because it's based in the child's motivation. Manding typically increases language in general because, through the positive reinforcement delivered as a consequence for the mand, the child comes to associate the sound of his/her own voice with positive consequences.
Tied inextricably to the mand is the motivative/establishing operation (MO/EO). Technically, the MO/EO (as per Jack Michael, 1982) is a set of environmental events that temporarily alter the value of other stimuli/events as reinforcers and therefore evoke all behaviors that have produced these events in the past. The MO/EO relates to conditions of deprivation and aversion. When the child is deprived of something, the MO/EO for the item is high because the "not having" makes the item more attractive. However, once the child has access to the item, he becomes satiated and the MO/EO is low. For example, if a child who loves cookies has not had any for weeks, the MO/EO (desire) for cookies is probably very high. If you take a platter of cookies and offer one to the child, you could likely teach the mand for cookie fairly easily. You would hold up a cookie and say, "Cookie." If the child has a strong echoic repertoire, he will probably echo, "Cookie," which you then reinforce by giving the child the cookie. Once this has transpired several times, the child will begin to mand "Cookie" in the presence of the cookie when the MO/EO for it is strong because saying, "Cookie" has historically led to access to cookies. However, after the child has eaten the platter of cookies, the MO/EO is gone and the mand will probably not occur. Thus, as you can see, MOs/EOs are dynamic, not static, and are temporary. When teaching mands, you want to teach in a condition of deprivation, when the MO/EO for the stimulus is high. In mand training, there is an MO/EO for the target stimulus, which is also the reinforcer that will be delivered. The MO/EO is probably the single most important motivative variable in teaching children language, although it is typically not discussed outside the circles of verbal behavior.
Once the child has an echoic repertoire and has acquired a number of consistent mands, you can begin to teach the tact. The tact is verbal behavior that is under the control of the nonverbal environment and includes nouns, actions, adjectives, pronouns, relations, and others. This one you can think of as a label of something in the environment or vocabulary. The word tact, another of Skinner's intentionally "nonsense" words, comes from the notion of the child's making "conTACT" with the nonverbal environment. Tacting is functionally very different from manding. If a child sees a cookie and says "Cookie," but maybe has just had dinner or a bunch of cookies and is satiated (there is no or a weak MO/EO), his saying, "Cookie" is not functioning as a mand, but as a tact. He could just as easily say "Hey, there's a cookie." We also do this all the time, in so many ways it's hard to enumerate, but think of it essentially as labeling. The way to reinforce a tact is NOT by delivery of the item named, because a tact does NOT specify its own reinforcer, as a mand does. You reinforce tacts with generalized reinforcers, essentially anything other than the item named. Naturally, praise or confirmation are typical means of reinforcement (i.e., to the child labeling "Airplane!" the mother says "You're right, it IS an airplane" and maybe ruffles the kid's hair). You can also reinforce with a primary/tangible reinforcer: "You're right, it's an airplane. Here's a cookie." Tacting is, in a way, most of vocabulary and makes up a huge portion of everyday language. It is usually the focus of many DTT programs, although echoics and mands are arguably far more important, especially when first teaching language. When teaching tacts, you want to teach in a condition of satiation, when the MO/EO for the stimulus is low. This is the exact opposite of mand teaching. To teach a tact, you would choose a stimulus for which there is no or a weak MO/EO and give the echoic, "Cookie" (after he's had his fill). When the child echoes, "Cookie," you could say, "Right, it's a cookie!" and reinforce with chips, or something else for which there's an MO/EO. In tact training, there is no MO/EO for the target stimulus, but there still must be a strong MO/EO for the reinforcer that will be delivered. Thus, the MO/EO is still critical in tact training, although it relates to the reinforcer that is now different from the target stimulus.
Requests and vocabulary are obviously very important in language acquisition. Equally important is another operant, the intraverbal. The intraverbal is verbal behavior that is under the control of other verbal behavior and is strengthened by social reinforcement. Intraverbals are typically thought of in terms of conversational language because they are responses to the language of another person, usually answers to "wh-" questions. There are two classes of intraverbals, fill-ins and wh- questions. So if you say to the child "I'm baking..." and the child finishes the sentence with "Cookies," that's an intraverbal fill-in. Also, if you say, "What's something you bake?" (with no cookie present) and the child says, "Cookies," that's an intraverbal (wh- question). Intraverbals allow children to discuss stimuli that aren't present, which describes most conversation. With an intraverbal, what the child says in response to the adult's/peer's language does not match what the adult or peer originally said. Intraverbals can be reinforced in a number of ways, with praise, generalized reinforcers, or, naturally, with a continuation of the conversational exchange, i.e., "Wow, cookies! They smell great!" (to which the intraverbal response could be "Thanks" or "You can have some when they're done."). To teach an intraverbal, you would ask a question and prompt the response with an echoic, reinforcing based on the MO/EO when the child echoed the correct response. Obviously, it's unnecessary to explain why teaching conversation is important.
Hopefully, it is clear how all of these verbal operants, along with the MO/EO, come together in language teaching. Once the child has a strong echoic repertoire (or mimetic/motor imitation repertoire for sign) you can teach across all the functions of language by prompting echoically and reinforcing differently. We target these operants and teaches them through errorless learning (prompting and immediately fading prompts/transfer trials), mixing and varying targets, interspersing easy and hard tasks, and teaching to fluency. This combination of curriculum and teaching procedures has led to great success for many children with autism. It is important to point out that teaching VB can be effective for any student whose language is delayed or disordered, whether they will speak, sign, or use PECS (or other forms of augmentative communication), and whether they are early, intermediate, or advanced learners. The examples above are mostly for early learners, but all learners who need language instruction can benefit from ABA with a focus on teaching VB.
FURTHER THOUGHTS ON THE "IS AVB STILL ABA?" QUESTION