What's hardest for me in all of this is learning the language. It's very precise, and not intuitive for me. Okay, enough whining.
Questions:

1. Autoclitic frames?
Autoclitic: a unit of verbal behaviour that depends on other verbal behaviour for its occurrence and that modifies the effects of that behaviour on the listener. (Catania)
Ex:if-then?
So in Skinner's quote on pg 15 he's pointing to relational framing as a behaviour without explicitly delineating it?

2. Language hypothesis-the idea that differences between instructed and uninstructed performances could be accounted for by human language.
...the behaviour is verbal in Skinner's approach because a specially conditioned listener mediates reinforcement of this behaviour. (p.16)
I'm still struggling with this. So, for Skinner, if a tree falls in the forest and no one hears it, it doesn't make a sound?
Another question: For Skinner, can the listener and the speaker be the same person?

How can listening not be verbal? "The role of the listener in any verbal episode was thus "not necessarily verbal in any special sense.""
I think I understand the unworkability of this definition, I just want to understand Skinner's conceptualization.
Please don't tell me I have to read Verbal Behaviour.

I'm not quite understanding Skinner's dilemma regarding a functional definition of "specifying". He cannot refer to reference. I understand that, but am not getting the inability to refer to verbal behaviour. Is this why? "(Skinner)did not distinguish between verbal rules and regularities observed in other complex antecedents." p17

thanks
Joanne

Here is my summary of chapter one. I know i am a little behind schedule :(
but would be grateful if anyone could point out any fundamental misunderstandings before i get too far with ch 2. Unfortunately the formatting of headings has disappeared but i guess that doesn't make much difference to how much sense it makes.

Chapter 1

Why a behavioural approach is a good idea.

This chapter opens by (briefly) making a case for a behavioural approach to language. It points to the progress made the behavioural approach in other areas, and to the difficulty the cognitive approach has had in stepping outside common sense assumptions (e.g. that words can be understood as representing ideas). The cognitive approach is said to provide a reasonable account within its own assumptions (i.e. that the task is to describe how words and sentences etc. are perceived, encoded, produced etc.). However, a behavioural approach may have something different and useful to offer.

The chapter continues by saying that language is, or should be, of fundamental importance to behaviour analysts, pointing out that the early focus on the behaviour of non-verbal organisms was only ever intended to be a starting place.

What sort of behavioural approach we are talking about.

The authors outline the assumptions of their particular brand of behaviour analysis, i.e. functional contextualism. This is really very simple. If you start from the goals (predict and influence) everything else follows. These goals (predict and influence) mean that the independent variables must be in the environment of the person/system/group being studied (else, obviously, the theory cannot tell you how to change the person/group/system). Hence the importance of context or environment. Also, ‘truth’ has to be what works in achieving those goals. There is no other criterion. That’s all there is to it. (??)

Why the previous attempts of behaviourists (even Skinner’s) to provide an account of language didn’t work and why this one will be better.

The next few sections of the chapter are concerned with outlining previous attempts by people working within the behavioural tradition to develop accounts of human language. Since these appear to have come unstuck in some way or other, it is important to understand what went wrong, and perhaps to explain the ways in which the current approach is different and stands a better chance of success.

The first to be considered is Kantor’s interbehavioural approach, which apparently did not generate a viable programme of research. The case is made that this is because it was based on descriptive rather than functional contextualism. In other words, the lesson is look what happens if you do not have your goals in place. (?) Some valuable and influential aspects of his work are also outlined: specifically, the importance of context and the idea that the stimulus functions of an event should be seen as part of the response to the event and that stimulus functions can be transferred from one event to another.

The chapter then looks at Skinner’s approach to language, and two attempts (those of Willard Day and Kurt Salzinger) to obtain the empirical data that would be needed to test and develop this model. The main point of this section is to argue that Skinner’s work, although an enormous step forward at the time, is not up to the job of moving us forward now. Basically if it was going to work we’d be a lot further ahead by now. The authors say that if you are happy to accept this conclusion you can skip the next few pages, but if you need to be convinced then you have to read them. I guess people who are familiar with Skinner’s Verbal Behaviour would find this section easy to follow, but I’m not and I didn’t. However the main point seems to be that there is problem with his definition of verbal behaviour: you decide whether or not a piece of behaviour is verbal by seeing how it is reinforced. If it is reinforced by someone who has been trained to reinforce it in this way by the verbal community then the behaviour is verbal. However, the behaviour of the listener/reinforcer is not itself necessarily verbal (it could be reinforced in some other way). This is a problem, because trained ‘listeners’ who deliver the reinforcement could be reinforcing all sorts of different behaviours that we wouldn’t generally consider verbal, such as the behaviour of rats in operant conditioning experiments. This makes the definition both too broad (it includes rats in operant conditioning experiments) and too narrow (it excludes the behaviour of the listener).

The research has led to RFT

The last part of the chapter outlines the work in behaviour analysis that paved the way for RFT. There are two main strands of research.

The first to be considered is research on rule-governed behaviour. Basically, instead of having to learn contingencies by trial and error you can just be told them and that way you get them much quicker, but you tend to take longer to notice if they turn out not to be what you were told they were. This would seem to be something to do with language (the “language hypothesis”). However, an account of rule-governed behaviour as verbal behaviour could not be accommodated within Skinner’s approach because a) the function of a stimulus as a rule depends on the history of the listener rather than the speaker and Skinner’s definition of verbal behaviour concerned only the speaker, and b) he could only say a rule was a discriminative stimulus which specifies a contingency, but this is inadequate (because it does not distinguish rules from other types of discriminative stimuli) unless you can specify what is meant by ‘specify’ which he was not able to do (following his own rules).

The second is derived relational responding. This section starts by outlining a famous paper by Sidman (1971) in which he reports that a person (child?) with learning difficulties who had been taught to match spoken words to pictures and separately taught to match the same set of spoken words to printed words was then able to match pictures to the corresponding printed words without actually being trained to do this. [NB I think there is an error in book near the top of page 18 in the sentence that begins, “Sidman’s (1971) …”?]. Because the picture-written words relationships had not been trained they are called derived. The next paragraph summarises how equivalence relations were studied using matching to sample procedures. If, having learnt that selecting X from X, Y and Z is correct in the presence of A, A is then selected from an array from stimuli such as A, B and C in the presence of X, then this is said to show symmetry. Further, if having learnt that selecting X is correct in the presence of A, and selecting P is correct in the presence of X, when then presented with an array of say P, Q and R in the presence of A, you pick P, well then that shows transitivity.

This was exciting because it seems very like language (e.g. the relationships between words and things are symmetrical). It was also puzzling because this symmetry does not generally apply in other situations, and the example of approaching thicket in the presence of a lion, but not a lion in the presence of a thicket is given.

The chapter ends by claiming that taking this idea of derived relational responding further (in the next chapters) leads to a more adequate definition of verbal behaviour, allows rule-governed behaviour to be understood as verbal, and lays the ground for a ‘vibrant research agenda’.

Thanks,

Janet

Okay, help. Can someone give me a real life example that will help me distinguish a mand from a tact?
Tact: A verbal operant in which a response (from the person emitting the operant or from the environment) of a given form is evoked.. by a particular object or event or property of an event of an oject or event."

There is an excellent summary of the core ideas of Verbal Behavior in Kohlenberg and Tsai's 1991 FAP Book. The whole of chapter 3 is about these issues and is very interesting as is the whole book. I copy a few lines from it since they tell it much better than I could :

A tact is defined as a verbal response that is under the precise control of discriminative stimuli, and that is reinforced by generalized secondary reinforcers. For example, if you are shown a red ball and asked, "What is this?" and you say "A red ball," you would be tacting because the form of your response ("red ball") is controlled by the object and is reinforced by a conditioned generalized reinforcer such as "uh-huh," "right" or "thank you," or any of hundreds of reactions that indicate you were understood. Notice that the contingency or reinforcer is borad and general, whereas the prior discriminative stimulus (Sd) is specific. The tact is thus brought about by the presence of a particular stimulus (e.g., a red ball) and an audience (the therapist or parent). Tacts, in this sense, are similar to the notion of labels or names ( p.54)

1.1.3 Skinner's Approach p9
Mand: A verbal operant in which the response, (whose? the person emitting the operant or the environment?) is reinforced by a characteristic consequence and is therefore under the functional control of relevant conditions of deprivation or aversive stimulation.

Stil from the FAP book : Mands are the speech involved in demands, commands, requests, and questions. A mand is behavior with the following characteristics: (1) it occurs because it was followed by a particular reinforcer, (2) its strength varies with the relevant deprivation or aversive stimulation, and (3) it appears under a very broad range of discriminative stimuli. Thus, if you were to say, "I would like some water" because you were thirsty, this would be a mand because it would be reinforced by a very specific reinforcer - someone hearing you and giving you water or showing you were to get some. Your "I want some water" response would not be reinforced by a generalized secondary reinforcer such as someone saying "That's right," or "Thanks for sharing that with me," or "I understand what you said". It's strength would also vary with how water deprived you were. Your mand for water can occur in almost any setting where you are thirsty and there is another person who can hear you. (p.56)

1.2.1 The definition of verbal behavior is not functional.p12

I'm not sure I understand this sentence in the last PP on the page
The behavior is not superstitious: the contingency is non arbitrary and is produced by the rat's behaviour.

Hope this is correct behavioralese : If you give arbitrary reinforcement to a pigeon, the frequency of the behavior closely preceding the reinforcement (it could be odd !) will rise. Thus, the probability that this particular behavior will occur short before the next instance of arbitrary reinforcement will rise, so it will be reinforced again and so on. As a final result, the pigeon will in the end emit this particular behavior with a high frequency, although the behavior in itself has no effect whatsoever on the reinforcing contingency. Which is not the case in the described example : As I understand it, by pressing the lever, the rat slightly shakes the feedbag and as a result, every five presses in average, a food pellet is jarred loose. This reinforcement is not arbitrary and is produced by the rat's behavior.

Here's another one on p13.
...Leighland cited Skinners ...theorising that the restricted contingencies required for abstraction ( a highly precise form of stimulus control) could only arise from an extensive history of social mediation.
The whole next paragraph just compounds the murk.

I find it difficult too. Maybe the important thing is to understand the critic made to Skinner's VB : Making appeal to the history of the listener in order to understand the behavior of the speaker is said to be a «conceptual error» making the design of fruitful experimental strategies extraordinarily difficult.

Okay enough for now. My mind is acting up.
Joanne

So does mine
Philippe

Below is a conversation regarding these questions:

RE: 1.2.1 The Definition of verbal behavior is not functioning. (Submitted by JT Blackledge on June 7, 2006 - 8:42pm.)

Hi Joanne & all--

This can be a tough one to convey and grasp, but I'll give it a shot anyway. The issue is this: Skinner (1957) defined verbal behavior as any behavior emitted by a speaker that is reinforced through the mediation of a listener trained by a verbal community to mediate such reinforcement. In other words, a given behavior is verbal if it is reinforced by someone else who has been specifically trained by a "verbal community" to reinforce such behavior in certain ways.

From this perspective, the lever-pressing behavior of every rat & pigeon that has gone through an operant conditioning experiment is verbal, because the reinforcement they receive (often in the form of a food pellet, for example) for pressing the lever depends on the mediation of the experimenter (specifically trained by a "verbal community" to mediate such reinforcement) who programs when the machine dispenses reinforcers. An odd (but true) statement, and one that makes you wonder "why define verbal behavior in such a way that it includes the behavior of apparently 'non-verbal' organisms?"

But that's not the biggest problem with the definition--Skinner's definition of VB actually violates one of the central tenets of behaviorism: What matters is the FUNCTION of a behavior, which refers to a reliable relationship between the emission of a response and the delivery of a reinforcer. From a functional perspective, it doesn't matter what the response looks like. My good friend Pete Linnerooth describes it this way: You can walk to the fridge to get a beer, run to the fridge to get a beer, or crawl to the fridge to get a beer. Either way you get a beer. The form of the behavior doesn't matter. What holds a variety of different-looking behaviors together in the same class is their ability to receive the same kind of reinforcement--and we're interested in changing classes of behavior, not just specific instances (e.g., stop someone from running to the fridge to get a beer, and they can still walk, crawl, or swing from the rafters to get one...). From a functional perspective, it also doesn't matter how the reinforcement is delivered (e.g., if it magically drops from the Heavens, is delivered by Sue or Joe, or results from more 'natural' contingencies). What matters is that, in a given context, there is a reliable and predictable relationship between behavior and subsequent reinforcement. Hence the two examples on p. 12. If a rat gets a food pellet after pressing a lever 5 times (on average), it doesn't matter if this result is mediated by an experimenter, or if it simply occurs due to a more 'natural' set of cirucmstances (i.e., a hole in the food bag that releases a pellet after the bag is jarred every five times or so by a lever). The lever-pressing behavior will be shaped up and maintained regardless of the 'source' of the reinforcement, as long as the contingencies are relatively stable. To ascribe importance to anything other than this basic functional issue in a definition of verbal behavior doesn't make sense from a behavioral perspective, and places Skinner's definition of VB in contrast to virtually every functional definition within behavior analysis. (That being said, I still dig Skinner for the brilliant contributions he made....).

Hope this helps.

Best,
JT

JT Blackledge
Lecturer
University of Wollongong
New South Wales, Australia

The functional nature of Tacts and Mands. (Submitted by Aidan Hart on June 9, 2006 - 5:23am.)

I’m still attempting to understand many of the concepts in the RFT book and Skinners definition of verbal behaviour. So far, my contact with Skinners system has been primarily within the chapter contained within Kohlenberg and Tsai (1991).

However, it seems to me that Skinners system, (and in particular Tacts and Mands as verbal response classes) is indeed a functional one. How one distinguishes a tact from a mand is not based on the content of the verbal statement but on the function it serves. Does this not make it functional?

For example, a client recently remarked to me that they found aspects of our sessions difficult (I can’t remember the exact phrase that they used and I suppose the exact phraseology is not important). Skinners system, as explained by Kohlenberg and Tsai, provides a framework in which I can I attempt to understand the function of the clients statement.

Is the client’s verbal behaviour a Tact? If so, in this case the clients verbal behaviour is under the discriminative control of me as a therapist, the questions I might be asking (and have asked) and possibly the setting events of the therapy session itself, the clients description may then be reinforced by my response (“Yes, these are difficult things to talk about”) If the client has a history of being under assertive or does not like to admit when they are struggling, this Tact may also serve as a CRB2 and my response to them may help reinforce and build their ‘self-tacting’ and ‘tacting’ to others behaviour.

However, I also considered that the clients verbal statement may have been a Mand, albeit a disguised one. In tacting that the sessions were sometimes difficult, the client may also have been Manding that they wished to talk about something else. The client may have wanted recognition or praise for their efforts. So even though the content remains the same, the distinction between the tact and the mand in this case surely lies in the function of the verbal statement itself which itself can not be understood independent of my (and others) history of responding to such statements.

I think that there is merit in Skinners position concerning appealing to the history of the listener in order to understand the behaviour of the speaker. I think potentially what Skinner may have been arguing for here is an ‘appeal to the contingencies’. In attempting to understand the function of any verbal interaction and in particular the function of the speakers verbal behaviour we must understand the context in which the verbal behaviour has occurred. The listener is part of this context. Understanding the function of the listener in that context will indeed require an understanding of the history of the listener. If the speaker’s behaviour is under the discriminative or consequential control of the listener and the listener’s response, then we need to understand the behaviour of the listener and their history of interacting with the speaker, if we are to understand the function of the speaker’s behaviour in the first place.

In the example above, distinguishing between a tact and mand and determining the function of the mand in terms or appetitive or aversive control will require understanding our history of interaction. This requires understanding my history of responding to such descriptions or request from clients or this client in particular. Therefore understanding the clients (speakers) behaviour is surely contingent upon a functional understanding of my (the listeners) behaviour.

Just some musings.

re: the functional nature of tacts and mands. (Submitted by JT Blackledge on June 10, 2006 - 8:24pm.)

Hi Aidan--

Very clever observation, me thinks. Tacts and mands may point to functional distinctions as they can describe responses made in the presence of different discriminative stimuli which function differently (e.g., a mand is a response reinforced by the acquisition of a tangible reinforcer; a tact reinforced by attention in the form of social praise). The problem is, use of these terms doesn't add anything functionally new to the analysis given Skinner's core definition of verbal behavior. In the bar-pressing examples used in the book, for example, the rat's bar-press on a reinforcement schedule controlled by the experimenter would be considered a mand; the rat's bar-press on a reinforcement schedule controlled by the 'hole in the feed bag' would simply be a response made in the presence of a discriminative stimulus (or SD) which functions to acquire a tangible reinforcer. But either way, the SD (in this case, say the bar or lever) functions exactly the same way (to elicit a bar-pressing response) and exactly the same schedule of reinforcement results. In other words, describing the rat's behavior as verbal or non-verbal adds nothing functionally useful to the analysis--the rat will continue to respond as generally expected for a rat receiving food pellets on a VI5 schedule, and no new terms based on the learning history of the listener or experimenter need be forwarded. In fact, by implying that the learning history of the listener/experimenter needs to be taken into account in order to determine 'what kind of behavior' the rat is engaging in, Skinner (perhaps inadvertantly) departed from a core focus in behavior analysis: It is the learning history of the individual under analysis, and how this history ascribes differential functions to current contexts, that is solely of interest. Plus, the way Skinner defined VB turned out not to address the differences in responding so far empirically observed between humans and non-humans (the fact that we can talk about a rat emitting verbal behavior using Skinner's definition points to part of the problem).

Stimulus equivalence theory and RFT manage to draw functionally useful distinctions between verbal and non-verbal behavior, distinctions that have repeatedly held up under empirical scrutiny. RFT in particular draws the dividing line between VB and non-VB solely at a process level (i.e., derived relational responding in the case of verbal behavior; operant behavior based on direct contingencies and generalization in the case of non-verbal behavior) that can be witnessed by focusing solely on the learning history of the individual under analysis.

And virtually all of this particular thread of the discussion is probably unimportant to those of you simply trying to understand RFT better. I think it only really becomes important when you are trying to sort through different behavioral accounts of verbal behavior and determine which one has a better smell to it.

Fun stuff, anyway.

Best,
JT

The role of the listener. (Submitted by Jacqueline A-Tjak on June 14, 2006 - 1:54am.)

Hi Aidan and all,
John and Francis are right: we do not need to understand Skinners view on VB or the corrections RFT made to understand the RFT stuff. Still there is some fun in it and I sure am glad with the questions asked, all the comments, examples and definitions that were given by others (on another childpage).
The role of the listener, Aidan, seems very important to me too. And he would not have his role as listener if he did not have a history of learning to respond/reinforce the words of the speaker.
But as I get it, focussing on the history of the listener brings about so many problems that verbal behavior cannot be studied in a way that helps us to understand human behavior better. John explained some of these problems. So RFT changed the definition of verbal behavior. This does not mean that studying the role of the listener suddenly does not seem important at all any more. It is just not a part of the definition any more.
Still, I do think you are very right that it is the interaction that 'needs' to be studied: behavior in context. The listener is (part of the) context for the speaker. The listeners behavior (which has a history) can be reinforcing for the speaker. The listeners behavior can now be seen as verbal behavior and the listener can now be seen as a speaker and the speaker as a listener, who reinforces the respons of the speaker who was a listener at first etc. Which is a complicated way to say that therapist and client mutually influence each others behavior.

Jacqueline

The role of the listener. (Submitted by Marco Kleen on June 14, 2006 - 4:16am.)

Hi everybody,

I think you hit an important point here, Jacqueline. The clinical implications of mutual influence of speakers and listeners are huge. This subject is very related to things like therapeutic relation, motivation, 'nonspecific variables' and so on. Just focussing on the intrapsychological processes would mean ignoring the role of the listener (and observer in experiments) as important contextual factor for the speaker.
A question: does that mean that the only thing a therapist can theoretically do is influencing the context because he's nothing more than a (relevant) contexual factor, or are there more theoretical possibilities?

Marco

The role of the listener. (Submitted by JT Blackledge on June 14, 2006 - 9:18pm.)

Hi Marco--

The issue you mention (the mutual influence of therapists and clients) is critically important--I wholeheartedly agree that we can and should monitor both therapist and client behavior, and that we should be interested in monitoring how therapist and client interact and in changing the behaviors of both parties for the better. And the function issue as it relates to verbal behavior is actually a different issue than that. The behavioral assumption is that stimuli take on different functions for an individual solely because of the learning history of the individual. When I'm interacting with a client in therapy, whatever I do functions in specific ways for the client solely because of the learning history of the client. Once we've started interacting, the client's interactions with me become part of his learning history--but it's still solely his learning history (intersecting with current context) that determines functions of stimuli in his environment, and thus determines his behavior. And certainly, many of the stimuli 'presented' to the client by me will be dependent on my learning history as the therapist. But the functions those stimuli take on for the client still depend entirely on the interaction between the client's learning history and the current context (not on ontological claim--just a pragmatic one). In other words, it's the client or speaker's learning history--not the therapist or listener's learning history--that determines what functions stimuli take on for the speaker from moment to moment.

The biggest take-home message from chapter 1, I think, is that it sets up the notion that RFT adds a clear description of how verbal behavior allows stimuli to take on different functions for an individual in a way (through derived relational responding) that's a bit different from non-verbal processes like direct contingency operant conditioning. That's one critically important thing RFT does that Skinner's analysis doesn't (Skinner assumed that verbal behavior and nonverbal behavior lead to stimuli taking on functions in the same ways). The second thing is that RFT does this without stating that we have to focus on the learning histories of both the client (e.g., the speaker) and the therapist (e.g., the listener) simultaneously in order to determine what function a given stimulus has for the client/speaker. It's the speaker's learning history that determines how a stimulus (even a stimulus provided by a "listener") functions for the speaker at any given moment--not the listener's learning history.

Anyway, fun to talk about.

Best,
JT

Mands and Tacts. (Submitted by JKesselring on June 7, 2006 - 4:20am.)

I'll have a go at the mand vs tact distinction. Consider this example: Bill and Joe are walking together down a street when they see a mutual acquantance ahead of them. Joe says to Bill, "What's that guys name?" and Bill replies "Jack". In this case the specific form of Bill's response (saying "Jack")was controlled by the specific properties of the stimulus (the guy calls himself Jack). This is a tact. [Joes question is a mand that sets the occasion for Bills response, but lets not worry about that now.]

Since Bill and Joe walk faster than Jack they gradually catch up to him. When they get a few feet behind him Bill says "Jack" which leads Jack to stop, turn toward them, and start to interact. In this situation a response with the same (or similar) formal characteristics as in the tact example (saying "Jack") has different functional characteristics. This is a mand since Bill said it "to get Jacks attention." Or to put it more behaviorally, he responded this way because in the past he has been reinforced by social interaction in similar contexts when he has said a person's name.

Another example: Joe says to his friend, Moe, who just baked a cake, "That cake looks tasty." This response would probably be a tact in the case where Joe has eaten a big meal and is throughly satiated. He is describing stimulus characteristics of the cake because in the past describing positive characteristics of the food that Moe prepares has been reinforced in various ways.

On the other hand, Joe's verbalization would be a mand in the case where he is food deprived and statements of this type in Moe's presence have been reinforced by food in the past. In this case Joe is says it because in similar contexts he has a history of being given a piece of the food after making positive remarks about the food.

Nontechnically speaking, the same words might be a compliment or a request for food (or both). (If his words did function as a request for food some would call them a "disguised mand" since they formally sound like a tact -- i.e., like a statement). The definition of a tact emphasizes the form of the response being controlled by of the properties of a contextual stimulus (in this case the cake or features of the cake), while the definition of the mand emphasizes the characteristics of the reinforcing stimulus relative to a deprivation condition (or other establishing operation).

I hope I didn't further confuse things.

John Kesselring

Mands. (Submitted by samtully on June 6, 2006 - 10:00pm.)

Correct me if I am wrong. I understand mands to be behavior, as outlined by Phillippe, though not limited to speech. i.e. if I wave at my friend to get him to come over this would be a mand. When a baby cries for food, this would be a mand. If a child spits to get attention, this is a mand. And of course, textual mands as well as sign language.

Cheers!
Shelly