To learn a language is to learn the activities, practices, actions and reactions within characteristic contexts in which the rule governed use of words are integrated (Hacker, 2003). It is to do things with words such as asking, telling, naming, directing, promising, describing, explaining, cajoling, negotiating, refuting, refusing, agreeing, correcting, teasing, tattling, inviting, etc.
To learn a language is to learn to manipulate symbols (words, gesture) according to the conventions for their use; to learn their meaning. To learn a language is to learn a practice.
Having a language grants many benefits: to be able to reason, deliberate, to talk about our pasts and futures, hopes, wishes, disappointments and joys, etc. It only makes sense that the goals within a language-based intervention program be those things we do in language.
Conceptual Systems
The conceptual system for teaching children a language, whether children with autism or not, is not found in theories. The ways we use language are a part of a cultural practice and as such, is best understood as an anthropological phenomenon and accounted for descriptively and historically. Practices are not accounted for in theory.
Theories are a part of science. Science constructs theories, which enable us to predict and explain events. But any "scientific" attempt to account for what we say or when we say things, as in theories such as Verbal Behavior or Relational Frame Theory are misdirected. Do you think its possible predict or explain, based on a scientific theory, what or why someone might say, "I forgot something", "To watch TV", "Mom told me to", "To get ice cream", "To get you a present", "I'm tired", "To be alone" if they were asked "Why did you go upstairs?". The possibilities are almost limitless: ... What in science would allow us to predict or explain such utterances?
Practitioners of behavior analysis often invoke scientific principles to account for what we do in language. 'Response generalization' is a knee jerk accounting for how it is possible to have many different responses 'controlled' by the same antecedent stimulus. However, such attempts fall by disagreement about what is meant by 'response generalization' (Skinner, 1953, Mayer, Sulzer-Azaroff, and Wallace, 2011, Carr, 1988). Some definitions require that responses be similar to each other (In what are any of the responses above similar?). Others argue that stimulus equivalents are at play (None of the statements above are interchangeable in the way that dog, pooch or canine are held up as examples of the power of the stimulus equivalents argument). Still, others argue that in order to say that there is response generalization, functional equivalence is operating (Escape, automatic reinforcement, attention?) or it requires that response classes are shared (All are intraverbals...but what does that add to our understanding?)
Behaviorist will also invoke 'histories of reinforcement' as a way to account for the unlimited number of possible verbal responses that might be uttered in the presence of a single 'antecedent controlling stimulus'. But such histories are unverifiable. Variability is the bedrock of science. But "unverifiable science" is oxymoronic yet stands as science in VB and Relational Frame circles (consider the double meaning of the term "circles" here.)
But ultimately, invoking scientific principles to account for cultural phenomenon is a anemic in the face of obvious insufficiency to account for daily use and counter-examples found in language. What we say in language is not reducible to science and moves in language take place in a practice which are not predicated on science and controlling agents.
The reality that there can be an indefinite number of responses in the presence of the same stimulus is the nature of our linguistic practice. Our mothers, fathers, sisters, brothers, aunts and uncles etc. showed us, guided us in learning to do and recognize the kinds of things we do in our practice...they didn't teach tacting, manding or intra-verbaling. They didn't ask us how many intraverbals, tacts or mands we 'had' during school but whether we 'told' or 'asked' our teacher something, whether we 'lied' to Johnny, 'promised' Suzy, 'reminded' Billy, or 'invited' Harry. They taught us how to make appropriate moves in language. These are the kinds of 'targets' which need to be considered for language instruction.
The 'Science" does not support the science"
This is a fundamentally different conceptual framework than the current and popular 'Verbal Behavior' approach with it's concomitant commitment to teaching mands, tacts, echoics, autoclitics and intraverbals. While teaching language within an Early Intensive Behavior Intervention (EIBI) framework employs behavioral tools, procedures, strategies and techniques, theory is not required. EIBI, when implemented expertly and appropriately, leads to effective, comprehensive and efficient learning, sans 'language theory'. The best long term outcome studies demonstrate this (Lovaas, 1987, Eldevik, S., Hastings, R.P., Hughes, J.C., Jahr, E., Eikeseth, S., & Cross, S. 2010, Smith, T., Green, A., & Wynn, J. 2000, McEachin, J. J., Smith, T., & Lovaas, O. I. 1993, Howard, Jane S. , Sparkman, Coleen R., Cohen, Howard G., Green, Gina, & Stanislaw, Harold, 2005). Applied Verbal Behavior (AVB) was not employed in those studies but the use of applied behavior analysis was. In fact, there are no studies comparing the use of AVB to methods used in the groundbreaking studies demonstrating the benefits of EIBI. There is no science... there are no outcome studies on the scale of those cited above which demonstrate the superior benefits of VB or Relational Frame Theory in helping children to become language users.
Language. Not behavior
Proponents of AVB argue that AVB should be used because it is "behavioral" (Sundberg and Partington, 1998). However, simply because something is "behavioral" does not make it better. Nor does it make sense. Teaching children to become language users is to see that children learn to "use" language. Language is something we use. Not do. It is a tool we use to communicate. We don't have theories for the use of tools; we don't have theories of hammer use or knife use. Tool use does not require a theory, but instruction in technique. Our job as interventionist is to teach technique. No inscrutable, incoherent theory needed. We learn technique.
In the case of teaching children language, the proponents of VB disregard one significant fact, VB is not about language although they conflate it with language. This is simply misleading. Skinner was clear that that VB was not about language but about behavior (Skinner, 1957, p.2). And, since AVB is not about language, mastery of a language necessarily occurs outside the constraints of such popular conceptual approaches which insist on the accumulation of tacts, mands, intraverbals, etc...the accumulation of verbal operants.
Learning a language is like learning to play a game
Learning a language is much like learning to play chess. When learning to play chess, first, one learns the names of the pieces (Simply knowing the names of things e.g. 'tacting' is to do nothing in a chess game or in language). Then one learns the rules for how pieces move. Then players learn to recognize rudimentary patterns of activity and basic moves, and later to recognize more complex patterns and more sophisticated technique. (For more on this, readers are encouraged to read about a Wittgensteinian approach to language by first reading anything by Baker and Hacker, PMS Hacker or Anthony Kenny.) The pieces in chess are used as tools and are used according to rules for their use for different purposes. A pawn may block, take other pieces, be sacrificed etc. Technique is employed in the game as need arises. Mastery of technique with pawns, will help win or lose a game. Similarly, words and sentences can be used as tools that can be used to different effect in different circumstances...to appease, cajole, tell etc.
It takes a great deal of time, training and careful engineering for a child to begin to master a linguistic practice. There are constituents which need to be first acquired and elements need to be drilled, even before training in the practice can begin. Children need to learn the names of things, to acquire concepts (ask, tell, before, after, quantitative concepts, pronouns, attributes etc.) and to recognize basic patterns of activity (negotiation, cajoling, refuting etc.) that are often a part of activity (playing games, making things, searching for things etc.) and in which words are put into play. Behavioral strategies and tools are invaluable in this effort e.g., shaping, discrete trial, task analysis, etc. But, no theory of language will advance our ability to teach children to be "language users".