What we mean by language


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. 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. Our job is simply to clarify the ways terms are used and how they can be taught. This is what informs an 'ordinary language' approach to intervention.

language within an Early Intensive Behavior Intervention framework employs behavioral tools, procedures, strategies and techniques and, when applied expertly and appropriately leads to effective, comprehensive and efficient learning. The best long term outcome studies demonstrate this. No Applied Verbal Behavior (AVB) required. In fact, no long term studies comparing AVB to better known studies have been done. Mastery of a language necessarily occurs outside the constraints of this trendy (e.i., AVB) conceptual approach which insist on the accumulation of tacts, mands, intraverbals, etc

What we know about language is not found in theories
The conceptual system for teaching children a language, whether children with autism or not, is not found in theories but rather, we are socialized into a certain way of being, that we learn how to do such things as making claims, raising questions, conducting arguments, sensing disagreements, recognizing agreements, and so on (Shotter, 1984).  Our mothers, fathers, sisters, brothers, aunts and uncles etc. showed us, guided us in learning to do and recognize these kinds of things...not to tact, mand or intra-verbal. They didn't ask us how many intraverbals we engaged in 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 the words that express the concepts that live under words; "ball", 'car", "when", "why", "thinking", "wishing", "hoping" etc. They taught use how to use them; they taught us their meanings. 

Our job as interventionists is similar. We guide by first providing training in the rudimentary constituents of a language. Then we must contrive situations and circumstances and introduce activities which allow for pointed practice and guidance around particular circumstances in which appropriate linguistic forms and accompanying words can be used. As a result, children are introduced to and subsequently learn to recognize patterns of activity in which it is appropriate to use particular linguistic forms and words. It must be stressed early in this discussion, the kinds of things we might say across circumstances are not "fixed",  not specified not 'determined' or 'controlled', as behavioral theory/ philosophy insists.

Learning a language is to learn moves in a game or more accurately many kinds of language games. It is much like learning to play chess.  When learning to play chess, first we learn 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 in response, 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 operate according to rules for their use...and they are used 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 used to different effect, using different techniques for different effects. Moves in chess and language are done with purpose. There is a point in moving one way or another. Not so in a verbal behavior scheme.

The premise of VB has nothing to do with learning a language
Verbal behavior is concerned with behavior. Language is not behavior. We don't language. Language does not have frequency, duration or intensity. Behavior does. Speaking is behavior. But, that which is said or written is not behavior. When I say, "I promise to give you the car". The utterance is behavior. It is 'clockable'. But the promise itself, my promise to give a car, exists beyond the utterance lives in language...not behavior


Skinner's analysis of verbal behavior is not an analysis of language...but of behavior; the behavior of speakers.  It is a theory which attempts to account for the controlling variables which result in the behavior of uttering sounds in the presence of other sounds or things. The 'response' and its controlling antecedent stimuli and reinforcing events are under investigation. The 'response' is the goal in applied verbal behavior, i.e, to accumulate them. To acquire operant units (the controlling antecedent and the response); Not to act in any way other than to utter responses; not to do things with words (plead, promise, accuse, compliment etc.) 


 Formats for teaching in a verbal behavior approach
One popular format for teaching within a verbal behavior paradigm is intra-verbal webbing. Here is an example:

  • Teacher: "What is something you bounce?"Child: "You bounce a ball" 

  • Teacher: "What is something you kick?"Child: "You kick a ball" 

  • Teacher: "What is something that is round?"Child: "A ball is round" 

These are the kinds of operant units taught in 'intraverbal webbing'. The VB units relate to the response "ball"; uttering "ball" is the targeted response in the presence of the verbal antecedent.  But this effort fails to consider what it takes to get children to learn what words mean (how they are used).  Word meaning, the appropriate use of words is central to learning a language. In the cases above, the point of language learning is lost. Language learning needs to also address the meaning of the other words, 'kick', 'round', 'bounce', 'you', 'what' that are used in the utterances above. What would it take and what needs to be considered if children are to learn the meaning of all of these words, i.e., to use and respond to words appropriately according to the rules for their use. These questions are not part of the calculus in a Verbal Behavior approach.

In VB, the words "bounce", "kick", "round", "you", etc. are relegated to the static realm of antecedent stimuli. Word meaning itself in VB has a rather bizarre nature. Moore, (2002) states that, in behavior analytic terms, word meaning for the 'listener',"is to be found to the extent to which an utterance enters into contingencies affecting the listener’s behavior. The most common way for the utterance to affect the listener’s behavior is to function as a discriminitive stimulus".  This means that the meaning of the word "ball" is the question, "Tell me something round". This also means that meaning of the word sun, tire, face, globe, bubble, coin, ferris wheel is "something that is round? Does this then mean a football and rugby ball are not balls? This is just absurd and misguided and has nothing to do with what we do when learning a language.


As English speakers we know the meaning of the words, coin, sun, bubbles, face as demonstrated in our ability to use and respond to them. This is not confusing for us. We also know that if these words had the same meaning, we would be able to use them interchangeably just as we use 'large' and 'big' interchangeably.  But, we don't say, "Give me that sun", when asking for a coin. But we don't do that, because we know what they mean.

What does "intra-verbal webbing" actually teach children about the meanings of words they hear or say? Nothing. But, in order to participate in a linguistic practice, learning word meaning is foundational. Thus, when asked, 'What is something you kick?', appropriate responses might include 'a ball', 'a door', 'your brother', 'the wall'. (In behavior analysis, the meaning of "wall", "brother" or "door" is 'something you kick'.) Knowing if someone understands the meaning of a word is manifest in appropriate use and response to a word. To illustrate this further,  if a person said that they, "kick a can down the road", "kick ideas around", kick up our heels, "kick someone out of a game", "kick a habit","kick a dead horse", or 'I get a kick out of you", this would be further evidence of their understanding of the word 'kick'; that they know what the word means; they know how and when to use it.  Moreover, these phrases in which the word 'kick' is present are used for different purposes, and such varied uses eventually need to be considered in efforts within comprehensive language instruction. 

So again, what is the point of these intra-verbal webbing exercises? They do nothing to account for the rich uses found in our practice. This VB strategy of 'webbing' has nothing to do with language instruction. The goals in VB miss the point of language instruction. Moreover, as you will see below, webbing and VB in general, fails to consider that things we say have a point.

Goals for language instruction; to do things with words/ not simply

to accumulate 'responses'
In an ordinary language approach to teaching, rudimentary word meaning is acquired. The meaning of the word 'ball' will not be acquired simply by asking things like, "Tell me something that is round". Rudiments need to be put in place. Then words need to be systematically integrated within countless other kinds of activities such as when searching for a ball, throwing a ball, hitting, giving, rolling a ball. But then, the word 'ball' needs to be put into play across circumstances in which different linguistic technique across different language games is employed. In the examples below, we see the activities of complaining, asserting, joking, commanding. They all contain the word ball. What is important to note here is that the word "ball" is simply put into play as a part of the various language games. The use of the term is not the point of the game itself. It is not clear how 'webbing' will lead to learning to do things with words or what words mean; to learn technique.

  •  Sally: "Why do we always have to use my balls?" (Game: Complaint? Request? Bid for explanation?) Johnny: "I promise to bring the balls to our next game" (Move: Mollify, promise) 

  • Sally: "Your ball hit the net, again." (Game: Assertion? Accusation?, Cajoling?) Johnny: "It did not. My serves are always perfect.". (Move: Playful retort, Refutation)

  • Sally: "Call Johnny and ask if we can use his balls please" (Game: Command, Request ) Johnny: "You call him" (Move: Refusal)

More on word meaning
 The examples above illustrate kinds of language games in which the use of 'ball' (the thing used in games) is integrated into activities such as promising, complaining, accusing etc. But such examples only touch on one use of the word and do not account for the different uses of the word.  We say that someone knows the meaning of a word when they use it and respond to it appropriately. So to really know a word is to know its multiple uses. Many words have multiple uses/meanings (e.g., bank, band, bark, interest, type, etc, etc, etc) . Let's look further at what we do in language.


Take the word 'ball' . We use the word "ball'  to refer to something we play with, to refer to an event (a formal affair) or to describe our experience . There are other uses as well. If I say, "I had a ball", determining if this utterance is a 'tact', 'mand' or 'intraverbal' tells us nothing about this statement's place in language; its meaning or how and when to use it and how to respond to its use. Nor does it matter if there is "joint control" or if an utterance is a "pure tact", an intraverbal, a "magical" or "a defective mand"? 

These categories of operants have nothing to do with the use of a word or sentence; with its meaning.  If someone asked, "How was the party?" and the answer given is, "I had a ball",  it's use here means, "I had a great time". Even though this utterance may be classified as an 'intraverbal', it is of no consequence. If it is determined there is multiple causation, so what? It is
use we need to teach children...to use words appropriately under the appropriate circumstance and to how to do things with words. Most other considerations are pointlessly academic, misguided and pedantic.

A ordinary language approach alerts us of the need to consider that words (the same word/symbol) often have multiple uses. It is our job to see that children learn the circumstances, transactions, situations in which appropriate use comes into play. And it must be restated, these differences in use reflect norms for use. Not generalization and not science.

It's all about "Use"

This ordinary approach to language eliminates the need to sit through painfully obtuse, incomprehensible power point presentations that offer fantastical verbal analyses but little guidance. In a VB analysis offered by Vince Carbone, which included a 90 slide 'analysis' of "yes" and "no" (once available online, but now offered for CEUs in some other interation), his 'verbal analysis' lead to the conclusion that, "The behavior of saying 'no' was brought under the discriminative stimulus control of "tacting the absence of joint control". I have no idea what this means. How do you tact something that is not there?  (The definition of a tact precludes such things.)  And, even if it were possible (which it's not), wouldn't the ability to tact the absence of something first require an ability to tact it? Does this mean that children who utter 'no' are able to tact 'joint control'? 


More important is the question, "How is this analysis helpful to those trying to teach children about how to use and respond to the word 'no'. How does this analysis guide and inform teaching. If we want children to learn about the word 'no', all we need to do is teach children how to use it.  We all know that it is used to refuse, reject, warn, refute, to disallow etc. And if any analysis is needed it's not of the VB kind. It's simpler: It's to ask what is involved on a practical level. Concerning the use of "no"is not about joint control , but rather it's to recognize that what sits behind the use of yes/no is 'correspondence' or 'agreement'. E.g., It is to ask, for example, if there is correspondence/agreement between one's desire and what is offered, i.e., do you want a cookie?, (Yes, I do want a cookie, "No, I don't want a cookie) "Please clean your room" (I don't want to) or if there correspondence between facts, e.g., Is a ball an animal, or "Did you go to the store?" No convoluted VB analysis is necessary. All we have to concern ourselves with when teaching a children to use 'no' is to ensure that constituent abilities are in place and to contrive arrangements (along with rigorous practice) in order to get appropriate use in play under the appropriate circumstances.   

Saying that Verbal Behavior is language misleads

Setting goals for language instruction simply requires that we account for what we do in language and then teach to those goals. But, goals within an AVB program, as informed by Skinner's analysis of verbal behavior, are different. The goal is to acquire 'verbal units. But this is not what we do in language. We are not response emitters but rather are word users. To say that Applied Verbal Behavior is "language intervention" (Sundberg and Partington, 1998; 2010; Partington, 2010; Sidner, 2010) is simply wrong and misleads. No amount of analysis about 'what kind' of verbal operant something "is" will inform what needs to be done when teaching language.  In language there is point in doing something with words - whether it be to assert something, to refute something, to negotiate, to mollify, agree, etc. Operating within a VB paradigm misses the point that when we utter words it is to do something.

lisa teaching.jpg

Drilling down a bit more: 'Language' vs. Verbal Behavior

Skinner's analysis of verbal behavior is very popular. Its theoretical underpinnings are deterministic and sets out to achieve 'prediction and control'. Verbal responses are said to be determined by controlling relations; joint, multiple or otherwise. Thus, verbal responses are automatically 'emitted or evoked' in the presence of 'controlling' antecedent stimuli in whose presence they've been reinforced. Within 'verbal behavior' based intervention, hefty lists of "verbal operants" (according to response class) are accumulated and graphed. Graphs with upward trends allow for the conclusion, "Children are learning".


Within this scheme, children do learn. This is not in question. However, what children learn in VB intervention is not language, but rather, "verbal behavior'; to emit automatic"responses" under the control of antecedent stimuli. "Having responses" has nothing to do with one's ability to do things with words... (greet, report, invite, direct, refute, etc.)  "Having verbal responses" says nothing about being able to recognize patterns of activity in which certain moves are appropriate or learn how to use words/their meanings.

By the way, if there is any question about the goal of AVB, leading figures within the VB community suggest that adult speakers 'have' upwards of four hundred thousand (Sundberg and Partington, 1998) units of verbal behavior. It is not clear how this number was derived but, it certainly was not by science. Nevertheless, 'number of responses' is what drives AVB ; the more that can be taught, the better. 


To make moves within myriad language games is to learn language

Let's consider'greetings'.  Everyone understands what a greeting is. Within a VB paradigm, "Hi" is said to evoke "Hello" (or some other learned response) in another speaker. Thus, in VB,  the "Hi", controls the response "Hello". But, in our linguistic practice, there are any number of appropriate things someone may say within this activity, in this greetings game. We see this every day. Instead of "hello" someone might respond, "Where've you been?", "I'm glad you're feeling better","How's your Mother?", "I'm so happy to see you again", "Sorry I'm late", "I love your skirt", "You won't believe what just happened to me on my way here!"or say nothing at all and just greet by kissing, hugging, shaking hands, smiling warmly, snarling, etc.  Any such moves would be appropriate within a greeting activity within our culture as a part of our linguistic practice. No explanation is required. No circuitous, tautological and impenetrable 'analysis' required. This is what we do.

But, in a Verbal Behavior paradigm, the possible responses in the presence of "Hi" assumes that someone would have to have 'learned' that when they hear "Hi", they are to emit, "Hello", "I love your new hair style", "I'm glad you're feeling better" or "You won't believe what just happened to me on my way here!"etc. While all of these 'responses' may be controlled by "Hi" (taught to be emitted in the presence of "Hi"), not all of them make sense in the presence of "Hi" all the time. But, there is no such thing as making 'sense' in VB. In VB, a words place in grammar, in the web of words is not relevant. There are only controlling events and what is 'to be' emitted in their presence. It can't be any other way within a VB calculus. Verbal responses, the things that are to be emitted  are a function of operant learning! What are the implications of such a worldview? 

Everything ever said has been learned/ and is under stimulus control 

Every utterance in every interview, debate, argument...all learned...and maintained by 'histories reinforcement'. Every utterance in every discussion or conversation you've ever had...all learned and 'controlled'.  Any wish wished, every plea pleaded, every promise made. Learned. Reinforced. Every word uttered in every game ever played. All learned. All utterances persons 'emit' are emitted  because of the presence of controlling stimuli.  There would be no 'weighing what we say', no 'formulating responses". It is on this proposition that 'verbal behavior' (and necessarily, AVB) stands.


To say that this is the foundation of a science called to account for all of what we say is a hard sell. I wonder about the feasibility of performing an analysis of the verbal behavior of the utterances on this page. Is what that is written before you an "intraverbal"', "tact", or "mand"? Is there joint control? What are the controlling stimuli? What 'verbal unit' can you identify? Is it even a meaningful question?


Adopting this world view requires that one accept that the words on this page are things I learned' to emit in presence of some controlling stimulus and maintained by histories of reinforcement. Do readers believe it is possible to trace events leading up to these words "emitted" in relation to some specific antecedent evoking them? or to identify the controlling variables! It seems unlikely. It seems unverifiable. Yet verify-ability is the bedrock of science. What does this mean concerning Skinner's theory of 'why we say what we say'... of his analysis of verbal behavior? Invoking vague 'histories of reinforcement', unverifiable histories of learning and questionable "stimulus control" is not a compelling scientific argument. (Readers are encouraged to read Chomsky's critique of Skinner's VB. One doesn't have to subscribe to Chomskian linguistics in order to recognize the merits of his critique.)

"The 'verbal behavior' of the person producing the 'verbal antecedent'? 

Let's stay with 'greetings' a bit longer. In the case cited earlier, the topic under discussion was a speaker's response to the verbal antecedent, "Hi". But efforts to chase down the 'controlling stimuli' related to the utterance of the person initiating "Hi",  is left open. What exactly produces the initiating greeter's "Hi". Why doesn't the initiating greeter emit 'tacts', so that rather than saying "Hi",  the greeter says something like, "Red hat", "Pants" or "Man" in response to the person standing in front of them? Why "Hi"? What controls the "hi". It can't be a 'tact'. Perhaps its a mand. I'm sure some behavior analysts in the academy can offer some lengthy analysis for what controls someone saying, "Hi". I imagine they might conclude that it's a mand... which is to say the initiating person is "Manding" for "hello"which reflects a state of deprivation of "hellos"... assuming that there is equivalence between "Hi" and "Hello" (remember...the mand names its reinforcer).  While this is all rather silly, it is not an unlikely 'analysis'. Perhaps someone has already published a paper on it. Perhaps it is now a part of the VB canon and all newly minted behavior analysts can feel secure in their roles if this question ever comes up. However, a more parsimonious explanation is that"Hi" is part of a practice... a social / linguistic practice. It's simply something we may do (say) in our culture under these circumstances when we first see someone. Similarly, we might also first compliment them, reprimand them, congratulate them, interrogate them, etc. No agonizingly arcane analysis required. No explanation needed. No theory necessary. It's all part of an accepted cultural practice.


Response generalization; Attempting to account for how it is possible to have varied responses to the same 'controlling antecedent'. 


Proclamations of 'response generalization' are a common fall back position when attempting to explain how we come to say so many different things in the presence of the same stimuli. I can't tell you how many times I hear professionals exclaim, "Oh... that's response generalization".  But, oddly enough, there is little agreement about what is meant by 'response generalization'. Some definitions require that responses be similar to each other. Others argue that stimulus equivalents are at play. Still, others argue that in order to say that there is response generalization, functional equivalence is operating or it requires that response classes are shared.  

With these 'definitions' in mind, consider the question, "Why did you go upstairs?" and the possible responses that could follow; "I like it there", "My boss told me to", "I forgot something", "To get ice cream", "To wash my hands", "To watch t.v.", "None of your business". How these and countless other possible responses might conform to what is considered to be 'response generalization'' as defined within ABA is difficult to reconcile, both because of the confusion about what the term means, and because, when considering language, regardless of the definition one adopts, it is difficult to see how response generalization applies.  Let's explore this briefly.


1. In order to say there is response generalization, responses need to be similar to each other: There is nothing similar in the responses above, ("None of your business", "To get ice cream", "I forgot something",etc.) Thus, when considering 'similarity' as a requirement for response generalization in language, this principle doesn't hold up well.

2. In order to say there is response generalization, stimulus equivalents are at play. I can't see what equivalence exists between the responses above. In none of the responses above could one be substituted for another and mean the same thing in the way that 'dog', 'canine' and 'pooch' are held up as evidence for response generalization being a function of stimulus equivalence. The transitivity of A = B / B = C  A=C, makes no sense in the language samples above and this simplistic calculus has little extension in language.


3. In order to say there is response generalization, functional equivalence is at play Escape? Attention? Access to tangibles? Automatic reinforcement? Words we use might serve those functions. It is possible to find words which function equivalently to serve the same function as other behaviors. Of course... this undergirds the whole "functional communication" paradigm. For example, tantrums (escape function) can be replaced with requests for breaks (escape function). And, is it is possible that a functional analysis will reveal that saying any of these statements might have any of one these functions? Surely. But is it possible to say that same function would apply across all statements provided  to the question, "Why did you go upstairs?", surely not. So, is there response generalization here?

4. In order to say there is response generalization, responses need to be in the same response class. This is less clear. According to Jack Michael (1985), response class in verbal behavior are either topological or selection based. But generally, response class refers to topography. Speech, would be one topology, writing another, signs another. They might all have the same effect on the environment if, for example, the word was "stop". But this analysis doesn't take us very far and is useless when considering the responses listed above. Michael's explication is pretty much unreadable. You should look at it though and see if you can make sense of it.

As illustrated above, no matter how you apply it, the principle of response generalization, as applied to language, doesn't hold up well. Ultimately this principle fails because it is a principle used to account for behavior within a scientific framework. But, as stated earlier, to investigate language is an investigation of the practices of a verbal community...it's an ethnological endeavor.  Such practices are not reducible to science.  Any related scientific principles called to account for cultural phenomenon will be rendered useless. You can't explain cultural practices using science; you can only describe and historicize them.

It's fuzzy, it's fun, it's language

As we rule out the explanatory power of 'response generalization' as an account for how we come to say different things in the presence of 'controlling' antecedent stimuli...it seems, that within an VB world view, we are left only with the brute force of controlling antecedents. That means, each and every antecedent has a corresponding response...as adduced before.


But, when words and sentences are viewed as technique  used to different effect... the need to learn 400,000 (?) operant units is negated. What a relief. Absurdity is put to rest. As humans, we are tool users... this applies in language as well. The ways in which tools and technique may be  be applied in language is inexact and fuzzy and does not lend itself to scientific investigation or explanation. Scientific passion in the case of language is misplaced. Language doesn't abide theories...language is messy... but if you just accept that it follows its own normative rules, it's actually fun...just follow the words to learn about how they are used


And, you can say about the things we do in language, without needing to explain, that we do those things, "just because...that's how we, in our (linguistic) culture do it... 'that is how it's done'. For example, why do we say that things 'fall down'? Why don't they 'fall up?' Why do we say we 'fill forms out" and 'fill forms in' to mean to complete a form? How can both be acceptable?' And why do we say that alarms "go off" and "go on" to mean that an alarm sounded? And as Wittgenstein liked to ask, "Why can't something be red and green all over?" These are not questions of logic or science, but of language. Our practice dictates how words are used. A Verbal Behavior approach? What does it add to our understanding of our linguistic practice or what it takes to indoctrinate persons in the practice?

Language instruction is more than accumulating verbal operants

Teaching children a language is not to teach children to automatically emit words (sounds); that we 'are to emit "x" when they hear or see "y". To be a language user, is to learn a practice or as Kenny says "A technique of language". To teach a language requires that we teach children to recognize patterns of activity and appropriate moves that can be made within those activities. This means teaching children how to do things with words and the meaning of words...how they are used according to convention. Finally, it requires early training in the rudiments and constituents of language, using behavioral tools to do so, and constructing intervention that is engineered toward progressive mastery of a technique of language.

Learning Language while Solving Problems 
With all of this in mind... there are a number of platforms we use to help get a language into place. One platform we employ is our active learning platform. In an active learning platform, situations are contrived so that children learn to solve practical problems. This platform allows for teaching many things within the context of solving a ‘problem’... So that children also learn the names of things, the properties of the things , how precisely to use things, where to get the things they need should the problem come up again, to learn to ask for help as problems arise, to describe the problem, explain things, and to consider alternate solutions. Examples of how we do this are found in the ASAT article and our book.