A new group paper

[From Bill Powers (2011.05.10.0205 MDT)]

Hello to all of an extended list.

The experience some of us had with a group-written and -edited paper was encouraging, although the result has left editors unimpressed. I think it may be possible to use the same medium, Google Docs, to generate another paper that will much harder to dismiss without review as "not the sort of thing we publish." A good part of the rejections that fell in that category would have been better expressed by saying "The things we publish concern phenomena that are defined in the theories of behavior we believe in, and your paper doesn't discuss those phenomena." An obvious example would be a paper sent to JEAB which doesn't discuss stimulus control of behavior or operants and instead speaks of intentions, goals, and control of input. That's not the sort of paper published in JEAB.

What I propose is a frontal assault, rather than an attempt to sneak up on the mainstream. If PCT is correct in its essentials, then no other theory now in use is correct. I suggest that we write a paper that takes on each major theory now accepted and shows that it is incorrect and how PCT handles the same phenomena better. Of course anyone in PCT who doesn't agree that this is the case will not be required to contribute and can join the other side to argue against us.

I contributed to the last group attempt at writing a paper by writing a version of what I thought we should publish. This time I can't do that. I don't have the knowledge, the resources, or the standing in psychology. The first thing we have to do is select the theories we are going to demolish and state them fully and authoritatively so we can avoid setting up straw men to knock over. We have to find the strongest and most widely-accepted version of each theory, stated so clearly that if it were correct we ourselves would become convinced of it. It might be that communications with the major proponents of each theory might be required, or at least serious explorations of their writings that can result in a fair and complete representation. This is a job for academics with a lot of experience and students who are learning this stuff anyway, who can do or have already done literature searches with the facilities at hand. I can join in the PCT part of the project but I simply don't have the familiarity with the field required to play Devil's Advocate.

Tim Carey reported that when he first demonstrated MOL to Warren Mansell, Warren's reaction was to say "You go right for the schema, don't you?" That's what I'm proposing here. I'm saying that we should try to get psychology to go up a level -- at least one -- and, aiming right for the schema, talk about the real issues that separate PCT from other theories.

If we do this right it can have two kinds of results. First, by trying to express existing theories in as convincing a way as we can, we may discover places where PCT needs to adopt some existing ideas or deal more seriously with phenomena we haven't considered. And second, we will challenge PCT in the same way proponents of those theories would challenge it, thus anticipating and finding ways to deal with counterarguments.

My proposal is pretty general and lacks useful details. I invite Bruce Nevin to start a Google Doc that is a brief outline of this project, which participants can edit until it looks like a definition of a writable paper. Then, in a month or two or three, we can start writing the paper that will put PCT on the map for good before I stop writing and start rotting.

Best,

Bill P.

[Martin Taylor 2011.05.10.09.24]

Bill's idea seems right to me. However, I think one of the problems with the original version was that it was not targeted toward any particular journal. What Bill proposes sounds to me like exactly the sort of thing Behavioural and Brain Sciences does best. I suggest that if Bill's proposal is taken up, the target journal should specifically be BBS, and we don't try to hedge it to make it suitable for other journals if BBS rejects it.

For myself, I don't feel I am up to date in other approaches sufficiently to provide an effective critique. And this illustrates a problem. Because we have to be very careful not to attribute to other theories things that their proponents don't believe, the writing does have to be done by "bitheoretic" authors, people equally at ease in discussing the PCT view and the view from a different theoretical background. How many such people are there, considering that when one does integrate PCT into one's thinking, the orthodox theory is seen to be a hollow shell?

Something like what Bill proposes does seem required. Could it be done in an evolutionary way? By "evolutionary" I mean by erecting straw man versions of other theories and getting some of their practitioners to knock down the straw men so that a wood man version could be constructed to be knocked down again, and so that finally an iron-man version can be built against which the PCT version could legitimately be compared? That would be a lot of work, if it is even possible. But is there another way?

One of the great virtues of PCT, which some people take as a vice, is its wide applicability to both individual and social applications. The problem is to show that PCT offers more than do the current approaches based on correlational statistics, which are seen to offer cookbook methods of addressing specific issues -- whether those methods are truly effective or not.

I quote from [From Bill Powers (2011.05.08.0915 MDT)]: "the believer needs only to assert whatever facts are needed to support the belief -- they don't have to be real facts. So the believer can present a facade of competence and confidence as if there can be no doubt whatsover, convincing naive listeners simply by force of sincerity, charisma, threat of humiliation, or verbal skill. In the long run, science always wins. But the length of the run can be a lifetime, and the opposition can lengthen it even more."

Martin

···

On 2011/05/10 5:12 AM, Bill Powers wrote:

[From Bill Powers (2011.05.10.0205 MDT)]

Hello to all of an extended list.

The experience some of us had with a group-written and -edited paper was encouraging, although the result has left editors unimpressed. I think it may be possible to use the same medium, Google Docs, to generate another paper that will much harder to dismiss without review as "not the sort of thing we publish." A good part of the rejections that fell in that category would have been better expressed by saying "The things we publish concern phenomena that are defined in the theories of behavior we believe in, and your paper doesn't discuss those phenomena." An obvious example would be a paper sent to JEAB which doesn't discuss stimulus control of behavior or operants and instead speaks of intentions, goals, and control of input. That's not the sort of paper published in JEAB.

What I propose is a frontal assault, rather than an attempt to sneak up on the mainstream. If PCT is correct in its essentials, then no other theory now in use is correct. I suggest that we write a paper that takes on each major theory now accepted and shows that it is incorrect and how PCT handles the same phenomena better. Of course anyone in PCT who doesn't agree that this is the case will not be required to contribute and can join the other side to argue against us.

I contributed to the last group attempt at writing a paper by writing a version of what I thought we should publish. This time I can't do that. I don't have the knowledge, the resources, or the standing in psychology. The first thing we have to do is select the theories we are going to demolish and state them fully and authoritatively so we can avoid setting up straw men to knock over. We have to find the strongest and most widely-accepted version of each theory, stated so clearly that if it were correct we ourselves would become convinced of it. It might be that communications with the major proponents of each theory might be required, or at least serious explorations of their writings that can result in a fair and complete representation. This is a job for academics with a lot of experience and students who are learning this stuff anyway, who can do or have already done literature searches with the facilities at hand. I can join in the PCT part of the project but I simply don't have the familiarity with the field required to play Devil's Advocate.

Tim Carey reported that when he first demonstrated MOL to Warren Mansell, Warren's reaction was to say "You go right for the schema, don't you?" That's what I'm proposing here. I'm saying that we should try to get psychology to go up a level -- at least one -- and, aiming right for the schema, talk about the real issues that separate PCT from other theories.

If we do this right it can have two kinds of results. First, by trying to express existing theories in as convincing a way as we can, we may discover places where PCT needs to adopt some existing ideas or deal more seriously with phenomena we haven't considered. And second, we will challenge PCT in the same way proponents of those theories would challenge it, thus anticipating and finding ways to deal with counterarguments.

My proposal is pretty general and lacks useful details. I invite Bruce Nevin to start a Google Doc that is a brief outline of this project, which participants can edit until it looks like a definition of a writable paper. Then, in a month or two or three, we can start writing the paper that will put PCT on the map for good before I stop writing and start rotting.

Best,

Bill P.

[From Bill P.]

[Martin Taylor 2011.05.10.09.24]

Bill's idea seems right to me. However, I think one of the problems with the original version was that it was not targeted toward any particular journal. What Bill proposes sounds to me like exactly the sort of thing Behavioural and Brain Sciences does best. I suggest that if Bill's proposal is taken up, the target journal should specifically be BBS, and we don't try to hedge it to make it suitable for other journals if BBS rejects it.

BP: BBS is a possibility -- I was a commentator for that journal several times, but never could get Harnad to accept a target article. Not scholarly enough. The new editor might be better, but who knows? And of course my coauthors will not be as unknown or unscholarly as I am.

MMT: For myself, I don't feel I am up to date in other approaches sufficiently to provide an effective critique. And this illustrates a problem. Because we have to be very careful not to attribute to other theories things that their proponents don't believe, the writing does have to be done by "bitheoretic" authors, people equally at ease in discussing the PCT view and the view from a different theoretical background. How many such people are there, considering that when one does integrate PCT into one's thinking, the orthodox theory is seen to be a hollow shell?

BP: Perhaps we could offer a variation on the usual format of BBS, if the editor is amenable. The usual format is a target article that is send out for commentary, and when the commentaries are received they are sent to the author who gets to add a reply to all of them, and then the whole package is published.

To avoid the straw-man accusation, we need to get ratification of our representations of the theories to be discussed. So we might ask for two rounds, the first to critique our way of presenting the other theories, and the second after we have finished the article and received the usual commentaries and replied to them.

MMT: Something like what Bill proposes does seem required. Could it be done in an evolutionary way? By "evolutionary" I mean by erecting straw man versions of other theories and getting some of their practitioners to knock down the straw men so that a wood man version could be constructed to be knocked down again, and so that finally an iron-man version can be built against which the PCT version could legitimately be compared? That would be a lot of work, if it is even possible. But is there another way?

BP: Actually this part of your post gave me the above idea. I think one round of knocking down would be enough to test this approach the first time it's tried, but that round would be a great improvement over the one-shot method. One has to be cautious about launching an infinite series. I expect that if our article is accepted, there will be plenty of repercussions in the form of counter-articles so the infinite series might happen anyway, but with luck and cool heads, it might converge.

Perhaps we could frame the opening round as an outline specifying what aspects of a theory we will focus on so the critiques will be pertinent to what we intend to write.

MMT: One of the great virtues of PCT, which some people take as a vice, is its wide applicability to both individual and social applications. The problem is to show that PCT offers more than do the current approaches based on correlational statistics, which are seen to offer cookbook methods of addressing specific issues -- whether those methods are truly effective or not.

BP: What I'm hoping we can get across is that PCT is a completely new concept of what behavior is and how behavior works. By organizing the article around demonstrations, we can show that these are not simply abstract concepts, but testable hypotheses and accurate predictors of observations, while the theories it replaces have never been tested. We should be able to point out that a fundamental revision of our understanding of behavior is going to have practical consequences everywhere!

Best,

Bill P.

Note: Please "Reply to all" so everyone on the list gets the replies. If everyone were to subscribe to CSGnet we could just use that convenient vehicle. Or some smart person might figure out another way for us to stay in touch outside the Google Docs format.

···

At 10:08 AM 5/10/2011 -0400, Martin Taylor wrote:

Bill,

Note: Please "Reply to all" so everyone on the list gets the replies. If everyone were to subscribe to CSGnet we could just use that convenient vehicle. Or some smart person might figure out another way for us to stay in touch outside the Google Docs format.

You sent your message (so far as I saw) only to CSGnet, so I replied there. Your reply with this note is also only on CSGnet, so far as I can see.

Martin

This post and the previous one was sent "To"
CSGNET@listserv.uiuc.edu, warren.mansell@manchester.ac.uk, wmansell@gmail.com, Sara Tai <sara.tai@manchester.ac.uk>, Tim.Carey@flinders.edu.au, mcclel@grinnell.edu, Henry Yin <hy43@duke.edu>, Cindy Richardson <mathwerkx@yahoo.com>

Perhaps your "To" list was truncated in the display. Or because you got it through the CSGnet server, the other items weren't there. I'll move the other addresses to the CC field this time -- see if they are there.

Bill

···

At 03:14 PM 5/10/2011 -0400, Martin Taylor wrote:

Bill,

Note: Please "Reply to all" so everyone on the list gets the replies. If everyone were to subscribe to CSGnet we could just use that convenient vehicle. Or some smart person might figure out another way for us to stay in touch outside the Google Docs format.

You sent your message (so far as I saw) only to CSGnet, so I replied there. Your reply with this note is also only on CSGnet, so far as I can see.

[From Adam Matic, 2011.05.12 19.00 GMT+1]

I would be happy to contribute in any way. Being a psychology student, perhaps in finding

relevant definitions, formulations, theories accepted in the mainstream and so on.

Best

Adam

Hi, Tim –

I like this idea alot, within
one small reservation … I think PCT is at a different level of
explanation to most theories that are out there. PCT (as I think of it)
is a ‘meta-theory’ at the level of S-R theory and S-R theory is a theory
that’s implicit in research designs and models but almost never
explicitly stated. It’s things like circular causality and control of
input that sets PCT apart, not the actual bits of the theory like error,
references, perceptions, and so on.

BP: Yes. That’s pretty much what I have in mind, though at some point we
have to justify the claims by presenting a model that predicts correctly.
After all, one of our main points is that theories need to be challenged
and pass tests before they are accepted. Stimulus-response theory has
never, to my knowledge, been tested, much less passed a test. The same
goes for the theory of cognitively planned actions. Theories are used as
explanations and seem to be judged mainly by how easily they can be made
to fit (verbally) what has already been observed. Theory-specific
predictions don’t ever seem to be made (i.e., if this prediction is
wrong, so is the theory that made it).

One way we can get the discussion up a level is to organize the paper
around demonstrations of control phenomena and especially the principles
that they illustrate. We could show a task, for example, in which it
seems that there are plans of action based on perceptions of
disturbances, and show that the actual behavior does not follow those
plans although the consequences of the behavior do.

One demonstration I’ve had in mind for a long time is illustrated by the
“pulleys”, which I haven’t got around to showing anywhere yet,
though it’s running.

Emacs!

The cursor is located at the right end of the green string. Moving the
cursor left and right pulls that end left and right. The task is to keep
the left end of the left-hand green string exactly in the gap between the
red target lines. The middle six pulleys wander slowly and independently
up and down, so when the right end of the green string is stationary, the
end over on the left side moves left and right (imagine there is an
invisible string pulling on the left end of the green string all the time
to keep it taut).

Explaining this demo using the cognitive plan-of-action theory can be
done, at the risk of making a lot of dubious assumptions about what the
brain can do. No matter what they are, they can be disproven immediately
by hanging a piece of cardboard over the part of the screen where the
pulleys are. When the pulleys can’t be seen, control continues just as
before.

TC: The frontal assault needs to
be on our notions of causality not any particular
theory.

BP: I think the pulleys do that.

TC: The others smoldering issue
I have is the “so what?” question. We can dismantle theories
all we like but until we demonstrate some practical relevance of PCT I
don’t think we’ll grab anyone’s attention. That’s not meant to sound
pessimistic or demeaning but just to say I think we have to be able to
demonstrate how PCT will make a researcher’s job much better than it is
now. You sort of get at this when you say we need to demonstrate that PCT
can handle the phenomenon better than existing theories but, again, I
think we need to have the ‘so what?’ question front and center.

BP: That sounds like the second section of the paper, or the second
installment of a two-part paper. Also, don’t forget that S-R and
cognitive theories don’t have any demonstrated relevance, either, because
they misrepresent what is observed. And we mustn’t overlook the fact that
both SR and cognitive approaches seem aimed at correcting behavior, which
in itself is a mistake since it’s not behavior that needs
fixing.

TC: PCT, for example, has made
my clinical work unimaginably better than it was before. Clearer,
cleaner, more efficient, etc.

The recent conversation with Steve Hayes, though, should be a good
example of how difficult it’s going to be to dismantle people’s cherished
theories. From the little I know of Steve Hayes’ theory - it’s not even
expressed in a way that could be dismantled!

BP: Just wait. He hasn’t replied yet to the last post in which I
predicted that he will give up the behavioristic approach.

We are not going to convince die-hard old-timers who are determined not
to be convinced. Let’s just forget about them. There are plenty of
reasonable people out there who can be convinced by a well-constructed
argument that engages their interest. Let’s aim at them. Sooner or later,
the balance is going to shift and we will stop being in the
minority.

Best,

Bill

(Attachment 637f11.jpg is missing)

···

At 03:45 AM 5/10/2011 -0700, Tim Carey wrote: