A psychologist joins CSGnet.
Unedited posts from archives of CSG-L (see INTROCSG.NET):
From time to time a new participant appears on CSG-L.Each time a variety of basic questions are asked and answered. This threadshows such questions, answers and challenges in an unusually thorough andspirited debate between Jeff Vancouver and CSG regulars. You will findnumerous comments on PCT and research and at the same time get a feel for theflavor of discourse on CSGnet.
The thread incorporates several thought provoking posts, such as PCT ANDITS CRITICS [Bill Powers (940922.1035 MDT)].
I think this thread reflects CSGnet well. All participants expose ideasand learn from the exchanges. This discussion took place in the past. Iexpect that ideas and opinions expressed here have continued to evolve. I knowmine have.
I have done my best to include a complete sequence of questions and answersin this thread. Posts are not edited for content.
Dag Forssell March 17, 1995
======================================
Date: Mon Jun 06, 1994 1:44 pm PST
Subj: Introduction
Hello CSG_net,
My name is Dr. Jeffrey B. Vancouver. I am an assistant professor inIndustrial/ Organizational Psychology at New York University. My research andteaching are in the areas of motivation and group processes. However, myprimary mission (reference signal) is promoting interdisciplinarycollaboration. To that end, I have come to adopt the philosophy of vonBertalanffy. That is, seeking a higher-orderparadigm that will allow scientists working in different areas to collaborateand learn from one another.
I have been enamored with PCT for some time (about 7 years) and discoveredthis net about 3 years ago. I have been lurking on it ever since. Initially, Idid not participate because my exposure to PCT had been through people in myfield (e.g., Carver & Scheier, 1981; Lord & Hanges, 1987).
These renditions piqued my interest, but left much unclear. I have sinceread Powers' book (1973), which I concur is profound, and his first collectionof works. In addition, I have read most of the second half of Hershberger'scollection; most of the American Behavioral Scientist special edition (1991);and the Richardson (1991) book on feedback thought. As well as a lot more in myfield that is using the basic PCT model.
Even after having "digested" much of this I was reluctant to participate inthis list as it is such a time sucker. Cziko's recent Penis post sums up thatproblem. I have often deleted or save large numbers of posting to read later(which I have done about 25% of the time). But summer is here, I recentlysubmitted an article about this stuff for my field, and I feel ready toparticipate more interactively.
One of the reasons I want to jump in now is in response to the recentposting Dag Forssell made to the TQM, reengineering, and quality lists. [See BPR-L.THD]I should have jumped in earlier, for there is a problem I see with that postthat will turn off many on those lists. This is a problem that permeates thislist. It is an us against the world attitude.
There are several reasons for this attitude, not the least of which is thetruth to it. BUT, I do not think it is bad as portrayed here. For instance,Forssell, as Marken did in 1991, begins by articulating PCT as a science ofpurposiveness. They go on to say that purposiveness is considered an evil inpsychology. There is no doubt that it WAS. But contemporary psychology isembracing purposeful behavior in a big way. For instance, the most popularmotivational theory in organizational psychology currently is Locke's goal-settingtheory. That theory is based on many of the same self-regulatoryideas as PCT. There is no question that Locke and his school (along withBandura) has not appreciated PCT, but they have ended up specifying models thatare purposeful.
In fact, I see many contemporary theories and applications based on thesame underlying principles of PCT. My argument is much like Richardson's(1991), which I was amazed to see on your reading list, particularly given therecent thread with the systems/cybernetic person whose name I cannot recall,that attempted to distance PCT from systems.
Much of the rhetoric from all the schools of thought remind me of a jokethat would be funny if it were not so telling. If someone could tell me theorigin or any teller I would appreciate it because all I have is a vaguememory, but it goes something like this.
A man approaches another who was about to jump off a bridge to certaindeath. In an attempt to talk the jumper down the conversation turned toreligion. "I am a protestant," said the jumper. "So am I," said the first man,"what denomination?" "Baptist." "So am I, orthodox or reformed?" "Reformed." Soam I, Eastern or traditional?" "Eastern." "So am I, Tririllian or Sectarian?""Tririllian." "I am sectarian, you heretic!" and pushes the jumper off thebridge.
The point, common in religion, is that those who are closest in beliefs areoften at greatest odds. The PCT school suffers this same solipsism that doesnot serve the greater aims of its members. This was made public recently in anexchange with Marken, who apparently has a bad boy image on this net. Whenpointed out to him, it appears an error signal was generated that changed hisbehavior. I am assuming that the change in behavior reflects an referencesignal in him for healthy debate. I also extrapolate this reference signal toothers on the net, and healthy debate is needed.
The most recent example of this on the net is the Paul Revere thread. BillC. seems to represent the broader psychological community in his description ofuncertainty and decision-making.His interest in the focus of control and resources parallels my own. Furtherhis understanding of the current psychological literature seems to reflect abroader understanding in psychology than many on the net. This is not to callpeople on the net stupid, just limited. We all suffer from this problem. But,the rejection of psychology out-of-handis dangerous. Perhaps Mary P. is right when she says PCT and the DME havenothing to do with anything psychology has dealt with, but I doubt it. She usesthe chemistry/oxygen paradigm shift as her analogy. I prefer the Newtonianparadigm shift. Euclidian geometry is not "wrong," just limited in scope. Sotoo are many of the decision-makingmodels that use subjective expected utility (SEU). But, they are still usefulif one understands the scope -an understand PCT can give. For example Beach's Image theory (book by thattitle 1990) uses both old SEU and new control theory concepts (although heseems to make some of the mistake Miller, Galanter & Pribium made).
But the point is that PCT netters don't often walk the talk. They wantothers to see their way but reject anything not from them. This leads tomisunderstandings and straw images of the each others theories.
Another example from Dag's post. He says that psychology (presumablecognitive) articulates a model of blind execution of internal plans. My readingof the literature is that the plans are a set of reference signals, just likePCT talks about. Further, an error signal that is larger that the plan expectswill cause a focus of attention to that point in the hierarchy (see Vallacher& Wegner, 1987, psych review). There is some very interesting stuff comingfrom many corners of psychology.
Another unfortunate impression one has from the Dag post is that PCT solvesthe human element problem for TQM. Any quality/TQM/BPR_L list members lookingfor PCT to solve these problems will be disappointed. This is not because PCTis a poor model, but because it is a long way from solving the applied "humanelement" problems. I think that the PCT model is VERY helpful for looking atthe issues. BUT, each individual in the applied setting is controlling theirown unique set of reference signals, using their own perceptual functions, andtheir own behavior repertoires. Dealing with all this complexity, diversity,and interaction is not easy, period. The one response I have come across mostfor not adopting systems theory and PCT is that it has not fulfilled itspromise. It would help if PCT stopped promising so much so soon.
One final thing, I have been teaching PCT to undergrads and grads, so youcan increment the number of teachers teaching it. I hope to participate morefully in the future, but forgive me if I am not as responsive as others on thelist. As the assistant in my title implies, my am trying to get tenure.
Jeff Vancouver
P.S. I just had a student ask if PCT deals with non-optimizingreference signals. I believe so, but cannot give a primary reference. This isone of Bandura's issues with PCT and we want to squelch it. Anysuggestions?
Date: Mon Jun 06, 1994 8:04 pm PST
Subj: Re: Introduction
Tom Bourbon [940606.1631]
> Hello CSG_net,
Hello, Jeff.
> I have been enamored with PCT for some time (about 7 years) anddiscovered this net about 3 years ago. I have been lurking on it eversince.
Now there is a virtual person to go with the name I've seen every time Ireviewed the list of subscribers. I hope you have been watching long enough torecognize my remarks and questions in this reply as friendly.
> Initially, I did not participate because my exposure to PCT had beenthrough people in my field (e.g., Carver & Scheier, 1981; Lord &Hanges, 1987).
> These renditions piqued my interest, but left much unclear. I havesince read Powers' book (1973), which I concur is profound, and his firstcollection of works. In addition, I have read most of the second half ofHershberger's collection; most of the American Behavioral Scientist specialedition (1991); and the Richardson (1991) book on feedback thought. As well asa lot more in my field that is using the basic PCT model.
You seem to have read a large portion of the collected works on PCT, aswell as some other sources with which we might not be familiar. Could you givesome citations of work in your field (Industrial/Organizational Psychology)that uses the basic PCT model? . . .
> One of the reasons I want to jump in now is in response to the recentposting Dag Forssell made to the TQM, reengineering, and quality lists. Ishould have jumped in earlier, for there is a problem I see with that post thatwill turn off many on those lists. This is a problem that permeates this list.It is an us against the world attitude.
Remember, you are seeing us when we talk among ourselves --_after_ we have put on our best faces and tried to communicate with "theworld." When we try to publish work on PCT modeling, we usually go out of ourway to avoid such an "attitude" --not that it seems to help.
> There are several reasons for this attitude, not the least of which isthe truth to it. BUT, I do not think it is bad as portrayed here.
Hmm. That doesn't seem to jibe with the contents of my file of reviews andrejections for PCT-relatedmanuscripts.
> For instance, Forssell, as Marken did in 1991, begins by articulatingPCT as a science of purposiveness. They go on to say that purposiveness isconsidered an evil in psychology. There is no doubt that it WAS. Butcontemporary psychology is embracing purposeful behavior in a big way. Forinstance, the most popular motivational theory in organizational psychologycurrently is Locke's goal-settingtheory. That theory is based on many of the same self-regulatoryideas as PCT.
It is certainly the case that there are many putatively self-regulatorymodels in psychology these days, especially in organizational psychology, or soI believe from my limited acquaintance with org. psych. But therein turns partof the tale: The model in PCT is not a self-regulatorymodel and PCT is not about self-regulation.
> There is no question that Locke and his school (along with Bandura)has not appreciated PCT, but they have ended up specifying models that arepurposeful.
Or so they say. I've never seen either of them demonstrate that theirmodels will behave (purposefully or otherwise) in simulation. Instead, I'veseen them assert that things work in a particular way, then they gathervoluminous correlational data in which they look for associations (low, butsignificant, correlations) between measures they *assert* are related to theprocess of self-regulation.Their research strategy doesn't really produce the kinds of data we need inorder to determine if their "models" work at all, much less if they work in thealleged manner. (See more on this below.)
I say this as a description of the state of affairs, not as a criticism. Ibecome critical only when adherents of that style of research begin to assertthat they understand control theory better than we do, and that they know ourways of using it and testing it are inadequate. You will rarely see thosecomments from the self-regulatorycamp in print, but they are very common at the stage or reviewing and rejectingarticles on PCT modeling. Maybe that's one reason we seem to come across asplaying us-against-the-world --you never see the other side of the argument in print.
> In fact, I see many contemporary theories and applications based onthe same underlying principles of PCT.
Are they based on the idea that behavior is the control of perception andthat most of what an observer sees when watching one who controls is irrelevantor unknown to the controller? Or that the control of perception is usually notthe same as what is often called self-regulation?Or is it that they say (but do not test in simulation) that feedback (ingeneral, or perhaps negative feedback, or perhaps both negative and positivefeedback) is important. I do not ask these questions rhetorically orsarcastically; I am continually on the lookout for new material in which peoplereally do use PCT, whether or not they call it by that name. . . .
> The point, common in religion, is that those who are closest inbeliefs are often at greatest odds. The PCT school suffers this same solipsismthat does not serve the greater aims of its members
> The most recent example of this on the net is the Paul Revere thread.Bill C. seems to represent the broader psychological community in hisdescription of uncertainty and decision-making.His interest in the focus of control and resources parallels my own. Furtherhis understanding of the current psychological literature seems to reflect abroader understanding in psychology than many on the net. This is not to callpeople on the net stupid, just limited.
Thanks. Some of our reviewers are not as kind as you! :-))
> We all suffer from this problem. But, the rejection of psychology out-of-handis dangerous.
That would certainly be the case, were we to do it. Again, it is a pitythere are ethical constraints on our simply publishing all of the reviews andrejections of our manuscripts on PCT modeling. Those documents might give you abetter feel for who is the rejector and who the rejectee.
> Perhaps Mary P. is right when she says PCT and the DME have nothing todo with anything psychology has dealt with, but I doubt it.
But she was merely saying what many of my reviewers have said. Let me usesome of my own experience as an example. You mentioned that you have read bothWayne Hershberger's book and the special PCT issue of American BehavioralScientist. Perhaps you saw my two published articles on PCT modeling of socialinteractions; In part, both articles were about instances in which two peoplesimultaneously perform a tracking task and the actions of one or both of theminterfere with a variable controlled by the other. Those are the only things Ihave in print on social interaction, but I have a file of unpublished relatedmanuscripts and data, some going on eight years old. Whenever I submitted thatwork to traditional journals, I always cited people whose work might be seen as"related," even when I knew that was not the case. I was careful to say that Iknew some aspects of the work were unconventional --I sampled continuous data from two people, I ran models in simulation todetermine if the models would reproduce then predict later instances ofperformance by the two people, and so on. I presented the manuscript as anexample of a different way to do social research, a way that was different frommethods in the conventional literature, but that in no way as intended as achallenge to or rejection of traditional methods, and on and on. The result?Rejections in which people said such things as, "This is not like the researchwe are accustomed to seeing." "Why continuous variables? Surely the author(s)could have recast the experiment to provide discrete data." And so on. In everycase, I was told that I was dealing with something different --something they weren't interested in. So you see, Mary had it right.
. . .
> But the point is that PCT netters don't often walk the talk. Theywant others to see their way but reject anything not from them. This leads tomisunderstandings and straw images of the each others theories.
We are often said to offer "straw images" of other people's models, but inour defense I offer the fact that the people who posit "models" of self-regulationor self-controltypically do not provide anything resembling a working model for their ideasand they certainly do not test their ideas by requiring their "models" tobehave in simulation. Absent any working models from those theorists, we oftentry to turn their words into working models, in order to test them insimulation. Perhaps the fact that those "straw models" so often fail insimulation is a sign of something other than a deliberate attempt by us to makeother people look bad. After all, *anyone* --anyone at all --who objects to our "straw images" can, at any time, provide their own (non-straw)version of a model and demonstrate that it *does* behave the way they say itwill. (Of course, when I have foolishly suggested that possibility, inmanuscripts that I submitted, my suggestion has drawn comments that I wasengaging in a cute, cheap ploy, intended to make my own presentation lookbetter. And all the while I thought I was inviting people to shoot me down andmake themselves look good. Silly me!)
> Another example from Dag's post. He says that psychology (presumablecognitive) articulates a model of blind execution of internal plans. Myreading of the literature is that the plans are a set of reference signals,just like PCT talks about.
This is a crucial point, Jeff. Do the *writers* of that literature say thatpeople act to produce and control their own perceptions, with their actionsserving as unintended means to that end? (In the PCT model, we use referencesignals to represent those kinds of intentions.) Or is it, as you literallysay, that you *read* the literature that way --perhaps reading into it something you *want to see* --something you believe *ought to be there*?
> Further, an error signal that is larger that the plan expects willcause a focus of attention to that point in the hierarchy (see Vallacher &Wegner, 1987, psych review). There is some very interesting stuff coming frommany corners of psychology.
I'm not sure what you mean by " an error signal that is larger than theplan expects." Could you say a little more about that idea?
> BUT, each individual in the applied setting is controlling their ownunique set of reference signals, using their own perceptual functions, andtheir own behavior repertoires. Dealing with all this complexity, diversity,and interaction is not easy, period. The one response I have come across mostfor not adopting systems theory and PCT is that it has not fulfilled itspromise. It would help if PCT stopped promising so much so soon.
PCT doesn't promise anything, but some of its adherents do. What should"we" do, instead? (And who are we, anyway?) :-)
> One final thing, I have been teaching PCT to undergrads and grads, soyou can increment the number of teachers teaching it.
Great. You can take my place --I recently stopped teaching and fell from the list!
> I hope to participate more fully in the future, but forgive me if I amnot as responsive as others on the list.
Watch out; this net has a way of taking over your life!
> As the assistant in my title implies, my am trying to gettenure.
A brave person, indeed!
See you on the list.
Later, Tom
Date: Mon Jun 06, 1994 9:55 pm PST
Subj: Welcome, Jeff Vancouver
[From Bill Powers (940606.1820 MDT)]
Jeff Vancouver (940606, I presume) --
Welcome to the land of the speaking, Jeff. I'm glad you have decided tospeak up, because you have some important observations to make.
I'll let Dag Forssell speak for himself. I can tell you, however, thatthere are often things going on behind the scenes that you wouldn't hear about.For example:
> But contemporary psychology is embracing purposeful behavior in a bigway. For instance, the most popular motivational theory in organizationalpsychology currently is Locke's goal-settingtheory. That theory is based on many of the same self-regulatory ideas as PCT. There is no question that Locke and his school (alongwith Bandura) has not appreciated PCT, but they have ended up specifying modelsthat are purposeful.
What they have done is to admit that behavior is purposeful, which peoplehave been claiming at least since McDougall in the 1920s, and they may evenconclude that behavior is caused by the difference between the a goal andreality, but they (Locke and Bandura, for example) explicitly reject controltheory as an explanation of these well-knownphenomena. This is extremely puzzling, so puzzling that both Mary and I havewritten to Locke in an attempt to clear up some of his misconceptions aboutcontrol theory (in response to an article by him violently attacking controltheory). The problem is that Locke rejects control theory precisely because ofhis mistaken idea of what it is, and he refuses to change his understanding ofit. Our correspondence with him has netted exactly zero by way of any improvedunderstanding. He has picked a position and is not going to change it.
The appearance of George Richardson's book on our reading list is not sostrange. If you read the book carefully, you will see that he puts PCT on athread of its own, which is neither the cybernetics thread nor the systemsthread. I was one of the prepublication reviewers of his book, and I gave it athumbs up because he had a deep enough understanding of PCT to see that itdiffered in essential ways from the two mainstream ideas.
You have to have some experience with individuals from other disciplines tograsp just how ridiculous the rejection of PCT ideas can get. There is fierceacademic competition going on out there, and the tactics used for trying tokeep opposing ideas in abeyance are sometimes every bit as self-servingand underhanded as one might expect in the sleazier parts of the businessworld. You and I think of the ideal scientist as a person who might be dismayedat discovering that his ideas are refuted by some new approach, but never as aperson who would deliberately try to suppress such ideas simply to maintain hisreputation of rightness. Ideal scientists, alas, are few and far between, andthey do not rise to positions of power and influence. Those who are at the topare, like any person endowed suspiciously well with the rewards of life, at thetop because that is where they want to be and intend to stay.
You say
> In fact, I see many contemporary theories and applications based onthe same underlying principles of PCT.
This I seriously doubt. You may find a lot of people talking about purposesand goals, but that is not theory; it is simply observation. I doubt very muchthat you will find _any_ other theory that contains an explanation of purposiveand goal-directedbehavior like that of PCT --unless the explanation is in fact PCT, acknowledged or not. For example, RobinVallacher writes a lot about very PCT-ishideas, but that is not surprising. In about 1974 or 75, he read my book andinvited me to give a seminar at ITT, where he was at the time. He got athorough personal introduction to my brand of control theory, but you will havea hard time finding any acknowledgement of that in his writings. Even then,incidentally, I was saying that the focus of attention goes to where thelargest error signals are, the only way I could think of to confinereorganization to the areas of the brain that actually neededreorganization.
Carver and Schier, who are a little better at acknowledgements, asked formy criticisms of their first book while it was being written, and I gaveconsiderable time and effort to them, including a critique of one of the mainpoints in their book about self-awareness --which they ignored, although the last time I spoke to Carver (to invite him toa CSG meeting, which he begged off from), he mentioned that they gave up onthat approach.
I have corresponded with literally hundreds of people about PCT, and haveshared my ideas and explanations without stint when asked for them, holdingnothing back. Whenever I have seen some piece of work that looked as if it wereon the same track or close to it, I have written to the authors inviting theirattention to PCT, explaining how it might be useful to them or apply in theirwork. The return on this effort has been miserable, although I have seen anincrease in allusions to PCT-likeideas in the literature as the years have gone by, and have wondered just howmuch of that reflected my efforts. The least satisfactory responses haveuniformly come from the most famous and admired people. Gerald Edelman, forexample, was condescending and insulting in his reply. Roger Penrose ignored myletter completely. Jeremy Campbell never replied. Joseph Engleberger concludedthat there was nothing in Perceptual Control Theory that was of any interest inhis work with robotics. The list is long.
The biggest problem I see is that people simply don't realize thedifference between observing that behavior is purposive and explaining how itcan possibly be that. The conclusions of PCT get across, but rarely doesanything of the theory come through. Even your student obviously doesn'trealize that there is a theory beneath the trappings of PCT:
> P.S. I just had a student ask if PCT deals with non-optimizingreference signals.
The words "non-optimizingreference signal" don't strike me as meaning anything. A reference signal issimply a specification for the state of a perception: it is a signal that has aparticular value, against which a perceptual signal is compared. There isnothing to optimize about it, as far as I can see, and a reference signal cancertain DO no optimizing of anything.
----------------------------
I don't go in much for bashing the opposition, except when particularlyfrustrated and needing to blow off some steam among friends. I would not dreamof using most of the tactics that have been used on me. I think that the faceof PCT that you see on CSG-Lis quite different from its public face. I do agree with you wholeheartedlyabout promising what we have never in fact delivered; I am as reluctant to emitjunk mail as I am to receive it. But I have learned long ago that people do thebest they can, and if my standards differ from theirs so be it. I just want tomake sure that if a person uses the term PCT, it is in fact PCT that is beingtalked about. Even that isn't always easy to do.
Best Bill P.
Date: Mon Jun 06, 1994 10:53 pm PST
Subj: Re: Introduction
[From Rick Marken (940606.1800)] Jeff Vancouver (940606)
> Hello CSG_net
Hi Jeff!
> I have been enamored with PCT for some time (about 7 years)
Don't talk of love. Show me! ;-)
> There is no question that Locke and his school (along with Bandura)has not appreciated PCT, but they have ended up specifying models that arepurposeful.
Two little nits. First, to my knowledge, Bandura and Locke have neverrecognized the phenomenon of purposeful behavior as control. Their idea of"purposeful behavior" is a lot like the idea of resting state of a dynamicvariable--no control involved at all. Second, they have never specified a model (in thePCT sense, ie. a model that actually behaves) of anything. Other than that,they have indeed done what you said.
> In fact, I see many contemporary theories and applications based onthe same underlying principles of PCT.
This is fairly vague. One "principle" of PCT is that motor output is afunction of perceptual input --o = f(i) --and there are many theories based on this principle. But in PCT this"principle" occurs in a closed negative feedback loop in which, ultimately,input is a function of an internal reference variable: behavior then is thecontrol of a perceptual variable. Is there really any other contemporary theory(besides PCT) that says this? If so, then I would imagine that there would be agreat deal of research being done on the types of perceptual variables thatorganisms control. Where's the research on controlled variables?
> healthy debate is needed.
What's a healthy debate? What was "unhealthy" about the debates that havebeen going on on the net?
> the rejection of psychology out-of-handis dangerous.
Who has rejected psychology out of hand? I think some of us (Bill P., TomB. and I) have rejected some individual tenets of psychology rather handily,though ;-)
> Perhaps Mary P. is right when she says PCT and the DME [DecisionMaking Entity] have nothing to do with anything psychology has dealt with, butI doubt it.
I think Mary has been suggesting that the DME probably has nothing to dowith PCT. PCT, however, has everything to do with psychology.
> But the point is that PCT netters don't often walk the talk. Theywant others to see their way but reject anything not from them.
We explain the phenomena and model of control. We reject what isdemonstrably false (information about the cause of variation in perception,control of contrasts in speech, social control, etc etc). We don't reject"anything" that does not come from "us"; we reject what's wrong. In fact, thereis a great deal of work that comes out of conventional psychology on which werely; especially the work on perception. Is there something, in particular,that you think we have unfairly rejected?
> This leads to misunderstandings and straw images of the each otherstheories.
PCT is tested by comparing the performance of working models to actualbehavior. In order to compare alternative theories to PCT we often have totranslate verbal descriptions into working models. When our implementations ofother theories fail miserably, the proponents of these theories yell "strawman". We have asked the proponents of other theories to show us how toimplement their theories correctly; that's usually the last time we hear fromthem --as they walk away infuriated, still yelling "straw man" and mumbling about howwe reject theories just because they are not ours. Do you have any suggestionsabout how to deal with this problem?
> [Dag] says that psychology (presumable cognitive) articulates a modelof blind execution of internal plans. My reading of the literature is that theplans are a set of reference signals, just like PCT talks about.
Really? And these reference signals specify the required states ofperceptual variables? Then why are there no studies of the perceptual variablesthat are controlled during the execution of these plans? Are the plansthemselves a perceptual variable? If plans are controlled, then a perception ofthe plan would be the controlled variable, in PCT? Where are the studies of thecontrol of the perception of a plan?
> Further, an error signal that is larger that the plan expects willcause a focus of attention to that point in the hierarchy (see Vallacher &Wegner, 1987, psych review).
How does a plan "expect" a certain level of error signal? Is the plancontrolling the error signal? If so, then how does "focusing attention" movethe error signal to its (possibly non-zero)reference level? Do they base their model on data showing control of errorsignals? What data is their model based on? How well does it account for thedata?
> There is some very interesting stuff coming from many corners ofpsychology.
Yes. It looks very interesting. Let's discuss the Vallacher & Wegnermodel, by all means. But, before we get started, I've got to know: if theirmodel produces no quantitative results, if it predicts average behavior oversubjects or over trials, if it doesn't work at all, do I still have to like itand not criticize it in order to have a "healthy debate"? If so, then we cansave a lot of time since I can give you my evaluation of their model right now --excellent, most illuminating, marvelous ;-).
> The one response I have come across most for not adopting systemstheory and PCT is that it has not fulfilled its promise.
PCT only promises the correct basic model of purposeful behavior. It worksextraordinarily well in the limited circumstances where it has been tested. Itmakes clear, falsifiable predictions that, so far, have not been falsified. Theconventional psychological models and data that fill the journals rarely makeclear, falsifiable predictions about anything except statistical properties ofbehavior --so they are never really rejected though they rarely work well. I'd say thatPCT has fulfilled its promise in spades --the promise of providing a strong FOUNDATION for the study of behavior.Apparently, few people want to build on that foundation because it's not yetthe Taj Mahal. They would rather keep playing in the shack built on theshifting desert dunes. Nu? What can we do?
> P.S. I just had a student ask if PCT deals with non-optimizingreference signals. I believe so, but cannot give a primary reference.
What in the world is a non-optimizingreference signal? Optimization is a judgment an observer makes about thebehavior of a control system; a control system just controls.
> This is one of Bandura's issues with PCT and we want to squelch it.Any suggestions?
Boy, Bandura goes right to the periphery of the issue, doesn't he? Mysuggestion for dealing with it? Admit it. Say "Yes, you're right Al. PCT doesNOT deal with non-optimizingreference signals. You are one sharp cookie Al. If only those PCT guys knewwhat they were missing... er ... what are they missing, Al??";-)
Best Rick
Date: Tue Jun 07, 1994 1:48 am PST
Subj: Re: Introduction
<[Bill Leach 940606.23:41 EST(EDT)]
>Jeff Vancouver 10030 on Mon, 06 Jun 1994 10:52:58 -0400
Jeff;
Being both a "non-professional"and really a novice seeker of understanding of PCT, I'll comment to a"lurker".
First before being too critical, I'd like to remark that you likely couldcontribute a great deal what goes on here. You are quite obviously not one to"shoot to quickly from the hip" and can add a great deal of experience to thediscussions that take place here.
I personally agree that there is something to the "Us against the world"attitude that is present here. Rick is, as most (including Rick) are willing toadmit, the most direct about "attacking" that which appear "not toconform".
It is my opinion however, that it was this very insistence on purity thathas helped me to understand PCT better than I might otherwise have done. Rickwas often quite wrong about where I was erring or often even that I was erringBUT, the result of trying to explain or "defend" something that I saidinvariably resulted in my learning something new about PCT. Quite often thelearning actually came from a posting of Bill Powers in his attempt to clear upa difference between Rick's perspective and my own but again, such a postingwould likely NOT have occurred if Rick had not also been so insistent.
I agree that there is a 'danger' that some may be driven away by the"puritan" approach of many that post here frequently. I am not sure what to sayabout it though. I know, for example, that Dag is truly appreciative of seriousconsideration of his work. I feel that he is able to make excellent use of thecomments that are made.
People such a Bill, Mary, Tom and Rick just about have to stick to thepurest possible form of PCT. Even the HPCT discussion need to contain thereservations that Bill so often makes. These are the people that "are" PCT andthey really do need to be careful if they are not to become like some many"popular" scientists. Bill's attitude about how PCT will stand ONLY FOR SO LONGAS REALITY ALLOWS is almost frightening to one accustomed to physical sciencesbut it really IS mandatory.
It is this very demand that Bill has demanded that makes PCT so vastlydifferent from all other attempts at a behavioral science. All other behavioralscience is too willing to dismiss variance as noise. PCT is unwilling to be so"loose". But that also means that PCT MUST remain very precise in terminologyand in what IS and IS NOT actually a part of the theory.
The interest of many of us is in the practical application of PCT to "reallife situations". Dag is especially one of those and of course Ed Ford isanother. Personally, my interest is first to gain a real understanding of whatPCT is and then how it can be used in "real world" situations.
I think it is the responsibility of those of us that have a "real world"concern with PCT to recognize that those that are the "bearer of the torch" CANNOT permit "loose" use of PCT. Those "bearers" may well make errors in theunderstanding of the posting from such as myself but it is MY responsibility toattempt to clear up possible misunderstandings.
You have mentioned that a number of researchers have produced work where itappears that they have grasped PCT principles. Personally I would agree. Indeed(though I now need to study him again) the man that initiated the entire ideaof "pop-psychology"(Dale Carnegie) in the classic book "How to win friends and influence people"was obviously espousing PCT principles. It is, on the other hand, probably verycritical that he did not actually recognize the fundamental essence of what hewas saying.
In the same manner, those that talk about "purposeful" behavior withoutrealizing that humans are negative feedback control systems controllingperception ALL the time no matter what happens around them, are missing themain point (I think).
The point is (again I think) that the environment DOES NOT cause behavior --ever!
Does the environment affect behavior (as perceived by others)? Of coursebut the only really important element is the subject's perceptions and thereference to which the perceptions are being controlled. This does not"simplify" anything (as you mentioned) but it does mean that often those tryingto help or otherwise deal with others are making things more complicated byintroducing issues that are not...
> Religion
This IS a fine point. The PCT "torch bearers" have to be very careful toavoid 'religion'. Several of Bill's postings have dealt with just that subject.He has frequently stated that he does not want a bunch of "supporters" thatblindly believe the "high priests". On the Other Hand, heretics (yes Rick, Istill believe that heretics can exist) must not go unanswered. PCT is in it'sinfancy and certainly there may well be many improvements and refinements.Martin's insistence upon evoking Information Theory could at sometime in thefuture prove to be necessary (at some time if or when experimental results failto match reality).
I don't think that Bill (and others) are being "religious" when they insistthat any "enhancements" stand the test of necessity.
> The point, common in religion, is that those who are closest inbeliefs are often at greatest odds.
Again, I agree that this is true but...
I have often felt that Rick was wrong in his perception of what Dag(specifically) was saying. I really don't believe that it hurt either Dag ormyself to try to come to an understanding with Rick. Indeed, it is likely thatboth Dag and I benefitted from the exchange.
> The PCT school suffers this same solipsism that does not serve thegreater aims of its members. This was made public recently in an exchange withMarken, who apparently has a bad boy image on this net.
I have not had time to review the postings on the net while Rick was"silent" (I was also not on the net during that time) but my understanding wasthat Rick stayed away to see if the discussions were "freer" without him. Idon't believe that they were but will defer to others since I have not reviewedthe postings.
> This is not to call the people on the net stupid, just limited.
Here is a place where I again tend to agree with you except that I feelthat my trying to understand the PCT "viewpoint" as espoused by Rick, Bill, Tomand others generally helps me more than trying to convince them that theirviewpoint is limited.
I have several times myself been a bit frustrated when trying to integratemy limited understanding of PCT with my limited experience in human behavior. Ibelieve that the frustration comes from failure to understand the significanceof PCT and from failure to recognize that the theory itself is not able toexplain may aspects of human behavior in a rigorous fashion (rigor as definedfor PCT not psychology).
Thus, very soon after leaving the realm of direct experiment in PCTprinciples we are in an area where "opinions carry the weight". This is toughterritory for the newcomer (such as myself). Bill and others have given a greatdeal of thought to many of the assertion that they make. Not only that, butthey have had the experience of trying to "second guess" the theory in areaswhere experiments were eventually performed and "learned a thing ortwo".
> Another example from Dag's post. He says that psychology (presumablecognitive) articulates a model of blind execution of internal plans. Myreading of the literature is that the plans are a set of reference signals,just like PCT talks about. Further, an error signal that is larger that theplan expects will cause a focus of attention to that point in the hierarchy(see Vallacher & Wegner, 1987, psych review). There is some veryinteresting stuff coming from many corners of psychology.
I suspect that this again misses the point. Anyone that really tries toobserve human behavior will conclude that there is "purpose" in at least mostbehavior. What PCT says is not that there is "purpose" in behavior but that allbehavior is the result of attempting to control perception. There is NOdifference between the two assertion IF "purpose" is defined to mean "controlof perception" but that is seldom the case. "Purpose" is usually taken to meansome sort of "higher" goal (that maybe the subject has even written on a pieceof paper). Such a goal may or may not be related to behavior and this ISsignificant.
> TQM
Again, I agree and believe that Dag agrees too. However, the problem isthat most TQM programs fail to consider how and why humans function the waythat they do. It is not that PCT will provide any instant magic answers butrather that understanding PCT will keep one from wasting time trying to dealwith matters and principles that have nothing to do with the problems that oneis facing.
> promises
I have to consider B:CP as far a promises. I think that Bill did a good jobof pointing out that PCT could help a great deal in understanding the behaviorof living things. He has also stated that he personally doubts that anyone willever model a human mind well enough to use it for exact prediction ofindividual behavior.
You have raised a number of issues that are a real importance to anyoneinterested in PCT. I certainly don't speak for the net and certainly caught myshare of flack for fuzzy thinking and other errors. I really do believe thatthe "mechanics" PCT is not "where it is at" for those of us not directlyinvolved in PCT research but rather in trying to really understand theimplications of PCT.
I also understand your difficulty in keeping up with the net activity. Ihave not made a serious posting for close to two months and even this one isfar more hurried than it deserves. Please do try to comment when you can, ifyou stir me to trying to think, I can just imagine what you must be doing toothers here.
-bill
Date: Tue Jun 07, 1994 2:22 am PST
Subj: Afterthoughts on PCT versus other theories
[From Bill Powers (940607.0200 MDT)]
Trying to get adjusted for GMT, so far with the opposite results. I am somewhere in Australia at the moment.
Jeff Vancouver (940606) --
Afterthoughts from re-readingyour post. You say
> However, my primary mission (reference signal) is promotinginterdisciplinary collaboration. To that end, I have come to adopt thephilosophy of von Bertalanffy. That is, seeking a higher-orderparadigm that will allow scientists working in different areas to collaborateand learn from one another.
It's not that easy to put into words what your "primary" reference signalis. All you have to do is ask yourself, "Suppose I did succeed in promotinginterdisciplinary collaboration. What would that get me?" This will show youthat achieving interdisciplinary collaboration is only a means to satisfying aneven higher-levelreference condition --and not just one single condition.
Actually, interdisciplinary collaboration is a lot easier to achieve thancollaboration within a discipline like psychology. The stickiest situationsarise when PCT comes up against people offering different theories about the_very same phenomenon_. Bandura and Locke reject control theory whiledescribing the very phenomena that control theory is designed to explain. Evenwithout PCT, you have cognitive scientists, Skinnerian behaviorists (there areplenty of them alive and kicking --er --responding) and personality theorists (like Lord and Hanges, Hyland, Bandura,and Locke) all operating completely separately and completely at odds with eachother. Just consider the recent split of the whole field of psychology rightdown the middle, clinicians against self-proclaimed"scientists."
I find the usage of the term "theory" in various branches of the behavioralsciences rather odd. When Bandura says he has a "theory" that there are"proactive" behaviors, he seems to think that simply announcing this phenomenonamounts to offering a theory. But to me, it is only a description of somethingthat we either can or can't observe. If the phenomenon is replicable, we haveto accept it as real --but that leaves the job of theory, as I have always understood the term,undone.
PCT is not simply a description of purposive behavior dressed up in a newvocabulary. We don't just substitute "reference signal" for "goal," "perceptualsignal" for "stimulus," and "output" for "response." And when someone makes thetranslations in the opposite direction, the result is not an understanding ofPCT, it is only switching words and continuing to apply them in the context ofthe same old model. In fact, a facile adoption of the terminology of PCT is anexcellent way to avoid getting the point.
The point of PCT is to explain how it is that an organism can select somephysical condition that does not currently exist and bring it into existence byacting on the environment. The first step in getting psychologists tounderstand the explanation is to get them to accept that organisms can in factdo this sort of thing. There have always been a few psychologists who acceptedthat fact, although usually with poor justifications. But for most of mycareer, at least, the vast majority of "behavioral scientists" didn't evenaccept that as a valid description of behavior. This meant that I couldn't eventake the first step toward explaining PCT --what good does it do to offer an explanation of a phenomenon that yourlisteners believe to be illusory? I have spent a large part of my career justtrying to demonstrate that the phenomenon itself exists.
Now that more people are coming around to the view that behavior ispurposive, goal-directed,intentional, etc., it is at least a little easier to start explaining what PCTis about. Or it should be. Unfortunately, too many people STILL think that whenyou have described a phenomenon, you have explained it. So when Bandura andLocke say that people pursue goals, and that goals cause behavior through amethod of discrepancy reduction, they think they have explained goal-seekingbehavior. This illusion could be shattered in an instant if you could just getthem to focus on the question, "What is a goal, that it can have such aneffect?" Or, "How is it that a discrepancy between a goal and the actuality canproduce just those detailed motor activities that have consequences tending toreduce the discrepancy?"
PCT is sitting here with detailed answers to such questions, and has beensitting here for nigh unto 40 years, but to no avail: that sort of questiondoesn't seem to come up among the likes of Locke and Bandura. In explainingthat behavior is goal-directed,they seem to think they have reached the foundations. I hope you can understandhow frustrating that is to me. I want to say, "Good, now you understand theproblem. How about listening to my solution?" But they seem to think that theirdescription of the problem IS the solution.
In fields like AI, neuroscience, and the new incarnation of AI, ArtificialLife, almost the opposite problem exists for PCT. Here we have explanationsgalore in terms of neural circuits and system designs, all good stuff, butpractically no appreciation of the phenomena of ordinary purposive behavior.Thus many of these people are using their armory of explanatory tools toexplain phenomena that don't occur in nature. Look at the neural net people.What behaviors are they trying to explain? Responses to stimuli! Unfortunately,when you approach the organism at the level of the functions of its components,as PCT does, you can construct gazillions of models that do _something_. Butunless you can tie the models to the sorts of behaviors that organisms actuallyproduce, the whole effort is just an exercise in imagination; mathematicalOnanism.
PCT consists of two equally important parts: the definition of the problem(people seem to behave in very specific purposive ways) and a model that solvesthe problem (how they must be constructed in order to do that). It is necessaryto take both parts of PCT seriously to understand PCT. In fact, you can'treally understand either part of PCT without putting some serious effort intounderstanding the other part. A lot of our problems on the net have come frompeople with a fair understanding of one facet of PCT, but hardly anyunderstanding of the other.
The PCT model, with its hierarchical control-loopstructure and mathematical properties, helps us to recognize purposive controlbehavior when we see it; seeing examples of real behavior helps us to selectpossible architectures and discard others that are equally plausible oncomputational grounds alone. This interplay requires approaching PCT alwaysfrom both sides, giving neither side a disproportionate emphasis. Observationkeeps us honest; computation keeps us rational. Which one should we dowithout?
Best, Bill P.
Date: Tue Jun 07, 1994 12:01 pm PST
Subj: Comments from Mary
[Mary Powers 940607] Jeff Vancouver:
The reason I used the phlogiston/oxygen example was because the flip-flopfrom something emitted by a burning substance to something being added to itresembled to me the flip-flopfrom behavior being an outcome or consequence of external or internal forces tobehavior simply being a means by which something else entirely is accomplished -the control of perception. "Rejection of other psychologies out-of-handis dangerous" perhaps (in fact carries a considerable risk of career-blight),but it was not entered into lightly. I'd say that of the 40 or so years PCT hasbeen in existence it took 15 or 20 to come, reluctantly, to the conclusion thatPCT was revolutionary and would continue to meet for many years to come thekind of resistance and rejection that Thomas Kuhn described.
Some people scold us because they think we are knocking other psychologiesunnecessarily and making life more difficult for ourselves than we have to. Iagree that "those who are closest in beliefs are often at greatest odds".However, many of the people you cite (Carver, Lord) are further away from PCTthan they (or you?) think. Truly, they are talking about phlogiston and we'retalking about oxygen.
But Bill has answered your point about purpose, pointing out that admittingor asserting that it exists is not the same as having an explanation of how itworks.
To add to Bill's litany of active lack of interest in control theory, Iwrote to Lord and Hanges years ago, and received no reply. Some time later,Lord called Bill (about something to do with the Volitional Action book). Whilehe was on the phone, I asked Bill to ask him why he never replied to me. Theanswer: " didn't know what to make of it". The scientific mind at work ;-)
[added later]
Before this is sent, I saw the posts from Bill, Rick, Tom and BillL.
Do you feel like a quarterback sacked by the entire defensive line? That iscertainly not the intention. The criticisms you made were very important toeverybody who answered you -more hostile versions are very familiar (Tom Bourbon's famous rejection file,etc.)
I just want to clarify one point -theory versus practical applications. I'm wondering if you have gone beyondreading about PCT to getting the computer demos and simulations. I think theybring home the point about PCT being an explanatory theory, as opposed todescriptive or statistical. Going back to Kuhn, this suggests that pre-PCTpsychology is pre-scientific.Lots of data, lots of ideas about how it ties together, but no theory in thesense that physics has theories (part of the difficulty here is the wordtheory, meaning a body of principles OR a guess or conjecture). If we criticizeother psychological theories, it is from the point of view of the first kindlooking at the second kind. Certainly as far as the basic model is concerned.We aren't so much psychology-bashingas theory-bashing.
Lots of therapists and educators and social workers and organizationaldevelopers do good work -no matter what theory they believe underlies what they do. The fashions comeand go in these fields and don't make much difference. We like to think thatthe good ones are intuitive control theorists who probably don't need any help -and that the bad ones can learn to be better by being taught what the good onesfigured out for themselves. Psychologists teach useful rules and techniques,but these often have no connection with one another, and are rationalized bycompeting and incompatible theories as to why they work. PCT is almost inanother universe, demanding of, and at least sometimes providing, the "whys" aplausible theory of "how".
Mary P.
Date: Tue Jun 07, 1994 12:42 pm PST
Subj: Us against the world
[From Rick Marken (940607.1020)] Jeff Vancouver (940606)
> It is an us against the world attitude.
Tom Bourbon (940606.1631)--
> Remember, you are seeing us when we talk among ourselves --_after_ we have put on our best faces and tried to communicate with "theworld." When we try to publish work on PCT modeling, we usually go out of ourway to avoid such an "attitude" --not that it seems to help.
This is an excellent point. I think if you saw how Tom, Bill P. and I havebeen treated by reviewers and editors you would see who really has the "usagainst the world" attitude. CSG-Lis the only place we get to talk about PCT honestly and candidly, withouthaving to worry about whether what we say about the model is "palatable" tothose who are in power (those who control access to the journals, grants,tenure).
We are only against cant and arm waving. If there is an alternative to theview that 1) behavior is control and 2) PCT is the model that explains thatphenomenon then we are happy to consider it. So far, no viable alternative hasbeen presented. Is that our fault? Was it Copernicus' fault that a heliocentricmodel ultimately was simpler and more accurate than the geocentric one?
Do the people who reject PCT have an "us against the world" attitudetoo?
Best Rick
Date: Mon Jun 13, 1994 9:25 am PST
Subj: Responding to welcomes -Jeff
Hello again,
I am pleased that my introduction got such a response. Being ignored wouldhave been the ultimate insult. I did not feel unduly rebuffed, but, of course,I will respond back. My response centers around themes, not necessarilyindividuals. Specifically, I consider the following issues: self-regulation,modeling, some theorists in my field, and some minor issues:
Self-regulation:
Forgive my ignorance, but I think Tom Bourbon and Bill Powers think thatthe difference between self-regulationmodels and PCT is that PCT is about controlling perceptions and self-regulationis about controlling behavior. This is no doubt because the promoters of manyself-regulationmodels _say_ they are about regulating behavior. However, if the modelincorporates a test of the difference between a perceived state and a desiredstate, the result of which drives behavior, then the model is describing thecontrol of perceptions. If the modeler says the model describes how individualscontrol behavior, they either don't realize their error or they are trying toreach a certain audience. But the model still describes the control ofperceptions. D. Ford's Living Systems Framework is that type of model. I amstill trying to get a handle on the consequences of the error for those models.So, to answer Marken's (940606.1800) post, yes there are others that describethe CSG loop.
Modelling:
A major theme in the responses to my introduction was the role of modellingas a test of PCT and the other models I mentioned (indirectly throughproponents -e.g., Locke, Bandura). I think something that people on this net understandthat many in psychology do not understand is that the ultimate goal of scienceis to create a comprehensive model of the phenomenon they study. (Don'tmisunderstand me here. That goal will never be achieved. Indeed, the scienceshould be constantly refining, filling in gaps, and occasionally rejecting themodel -if one is to believe Kuhn -which I do.), But psychology, burned by previous attempts at grand theories andcomprised of individual, autonomous control systems is leery of such models. AsI understand it, this is not unique to psychology. Other sciences have gone orare going through similar periods (the pre-paradigmor pre-paradigmshift period, Mary Powers 940607). As a result, grand theories are eschewed formini-theories.Mini-theories,because of their reliance on exogenous variables, do not lend themselves tomodeling. That is, they won't work because importance relationships andvariables are left out of the model -usually feedback relationships. Nor do the architects of these theoriesunderstand how to model or what it can do for them. Further, these theories areempirical generalizations of relationships between latent variables, notstructural models of underlying architecture (Powers, 1973). The question, thatwe might answer differently, is whether these mini-theoriesare of any scientific value. (Practical value is a related, but separatequestion). I think they are because I think there are gaps in PCT that havebeen addressed, certainly not resolved, by these mini-theories. More on these gaps later.
But the previous argument is mostly on an intellectually level. The realityof rejection letters and lack of collaborative spirit from the powers-that-beis painful, dispiriting, and financially and occupationally challenging. Iapologize for raising any of those feelings. I have numerous colleagues withless than kind rejection letters based on the same reasons that many of youalluded to. On the other hand, I know others who have had little trouble onthat score. Of course, their work is probably not completely in line with the"core." Most notably, they do not require a working model as a requirement fortheir work (there is also an emerging appreciation for which journals are"friendly" and which are not). Herein lies one of my concerns with thesolipsism among the netters. Marken (940606) says "PCT is tested by comparingthe performance of working models to actual behavior. In order to comparealternative theories to PCT we often have to translate verbal descriptions intoworking models." This requirement confounds the test of a theory with thetheory. Marken is applying his criteria for a theory on the other's theory. Ido not condemn the criteria, merely the blind application. If empirical data isalways required, Einstein's GTR would never have lasted the 20 (?) years beforeit could be tested.
Now I realize that these mini-theoriesare not GTR. Indeed, many are very problematic. But I try to separate the chafefrom the wheat. Carver & Scheier's self-awarenessconstruct was clearly wanting, but they exposed many to control mechanisms thathad not known them before. To quote Bill P. (940607) "Now that more people arecoming around to the view that behavior is purposive, goal-directed,intentional, etc., it is at least a little easier to start explaining what PCTis about." No doubt there are errors in the views perpetuated by the non-PCTmodels, but it is a step. Further, theories like Carver & Scheier's dealwith many social processes which PCT does not deal with as thoroughly.Identifying the discrepancies between these models and PCT, and developingempirical tests the results of which both parties can take as evidence one wayor another is, I believe, a reasonable next step. And let me be perfectlyclear, I think sometimes they have a point and PCT needs to incorporateit.
The problem with Marken's approach is that they don't buy/understand thedata you use. Most don't understand modeling and/or don't trust its results.The rigorous requirements, central to your understanding of science, are seenas too rigorous for a science in it's infancy. You say it need not be that wayand I TEND to agree, but one must lead, not force the transition. (Some willnever change, leave them behind). Where I do not agree, perhaps because I wasraised as one of them, is at the higher-levelprocesses. Here I will talk about gaps. (I suspect I will learn of myignorance, but that is one reason for being here.)
Gaps & theorists in my field
First, as Bill P. said in his intro to Living Control Systems, there is nocontent. What are the intrinsic signals and where do they come from? What are,if not the higher-orderreferences signals, likely candidates and where do they come from? How areactions chosen and how does reorganization proceed? I see stabs at the firstquestion on this net (biological needs) but another type of intrinsic signal isprobably system efficiency variables. The recent discussion of uncertainty isone such variable. You are absolutely right to attempt to figure out how theorganism could perceive uncertainty and reject it if it can't (but beware, justcause you cannot figure it out does not mean it doesn't exist). The point isthat another perspective has inform the guess that uncertainty is a referencesignal.
My reading of your literature and posts is that the identification of higher-orderreference signals is truly a gap (although I fear I misunderstand the postshere). Many in psychology are seeking to study candidates for reference signal,including Little, Deci & Ryan, Emmons, Pervin (See edited books by Pervins,1989, _Goal Concepts in Personality and Social Psychology_, and Ford &Ford's, 1987, _Human's as Self-ConstructingLiving Systems.). These are self-regulationtheorists, and like I said above, I don't care what they think they are doing,I find what they are doing useful and interesting and I think some of youshould too.
To the third question, dealing with reorganization (and other matters likestress and leadership) are people in my field. Proponents of PCT-likemodels (forgive them their transgressions), including Tsui & Ashford (e.g.,Journal of Management, 1994), Edwards (e.g., Academy of Management Review,1992), Cropanzano, James, & Cetera (Research in Organizational Behavior,1993), Lord & Levy (Applied Psychology: An international review, 1994) andManz (Academy of Management Review, 1986) to name a few. Other interestingthinkers include the Germany action theorists (Kuhl, Hechausen) and others(Vallacher & Wegner). Many or most of these authors or theories may havealready been read and rejected. I do not agree with everything _any_ of themsay. But herein lies the challenge: finding which loops conflict, proposingalternatives, and testing the them.
Marken (940606) says you have rejected social control. Well, I am notexactly sure what is meant by social control, but my impression from earlierthreads (and B:CP) is that you rejected social influence. I find this veryproblematic. Marken (same post) says you have rejected "information about thecause of variation in perception." Again, I don't know exactly what Markenmeans by that phrase, but I get the impression from my readings that PCTeschews attributions and other beliefs (like self-efficacy)as relevant. This I find very problematic. Now I even find that the DME is notpart of PCT. You mean that all conscience thought is irrelevant? This is VERYproblematic. Reorganization is not simply random. A model that shows that itcould be (e.g., E. coli, which some of you have contented) does not make it so.But I think I must be misunderstanding you here, too. Planning and thinking ispart of PCT. Figuring out how that planning and thinking translates intochoosing and doing is where contemporary psychology can help. PCT has beentested in only "limited circumstances" (Marken, 940606), where it has not beentested improvements can be articulated and empirical tests constructed(hopefully).
Finally, in reference to Vallacher and Wegner, apparently Bill P. thinksthey are using PCT, just not acknowledging it. So Marken, when you say
> Yes. It looks very interesting. Let's discuss the Vallacher &Wegner model, by all means. But, before we get started, I've got to know: iftheir model produces no quantitative results, if it predicts average behaviorover subjects or over trials, if it doesn't work at all, do I still have tolike it and not criticize it in order to have a "healthy debate'? If so, thenwe can save a lot of time since I can give you my evaluation...
This is what I mean by unhealthy debate.
Some minor issues:
I was largely responding to Dag's post when I warned of solipsism. Thatpost was meant for those outside the net. So my comments regardingpresentation were legitimate. I did, however, appreciate the distinctionbetween messages on the net and messages to the outside world. It is callimpression management and the research in it has some interesting points forPCT to consider (although they do seem to miss the point that one can onlyattempt to manage the perceptions of impressions).
Some mentioned I should send my recent paper to them and others forcomment. I appreciate that, copies are on their way. One comment: The paperargues for the adoption of a PCT-likemodel as a paradigm for organizational behavior. The impetus for the argumentcame from a recent article in my field calling for a paradigm. Just before Ifinished my article a third article appeared condemning the paradigm approachbecause it lead to elitism and gate-keeping.This is a problematic argument because we have elitism and gate-keepingnow, as you know. I think that once PCT come to it's own (and it looks like Iam betting my career that it will), the new gate-keeperswill be you people. Now do you understand my fear regarding solipsism as itappears on the net.
Regarding the non-optimizingreference signal. I realize now that it was not a well-thought-outquestion. Nor is it worth considering further, I can deal with it. Thank youBill P. for at least trying to address the question. Marken, your comment wasvery helpful, NOT! :-)
Regarding my "primary reference signal." (promoting interdisciplinarycollaboration). Bill P. (940607.0200) took me too literally. Promotinginterdisciplinary collaboration is an espoused, higher-order,professional, self-concept,reference signal (EHOPSCRS). Why I really do what I do is beyond me (althoughattempting to find out is important to me). My EHOPSCRS identifies a niche thatI hope to convince others needs filling and that I can help fill. And yes Bill,cross-disciplinecollaboration may be easier than within, as my religion analogy was attemptingto say. It may be harder as the modeling issue shows (social sciences does notuse the methods of physical sciences).
I do need to get the demo's and simulations. I have Marken's Lotus program,which was when I found out about CSG_net (I also heard about it from Lord). Idid not understand it then and I don't think it worked when I tried it since.(Rick, you might remember, I am the one who had Quattro Pro, it worked fine onthat version, but I have since updated and I think that may be theproblem).
Prologue:
There are some very ridiculous criticism to PCT out there. Clearly, thebehavior we observe among the dominant coalition is based somewhat on the errorthey anticipate when their life's work is rendered as superfluous by PCT (e.g.,Locke). Another is the mistaken implications associated with the word"control." I think its tendency to be overplayed and the connotation it evokeshas caused many to misunderstand PCT, which has evoked a strong counterresponse from PCTers that is also misinterpreted (for example, I think I havemis-interpretedyour social control posts). But the negative PR factor is a real problem. Icannot see the average citizen accepting the theory as they have accepted thegeneral theory of relativity simply because of the name. Call this superficial,but I now refer to my work a part of living systems theory (of course, it isnot core PCT, so you probably appreciate that -although I do cite PCT references often).
In conclusion, there are many roles to be played here. To evoke thereligious analogy again (understanding that science seeks data moreemphatically than religion, but it is so appropriate). The high priests willattempt to maintain the purity of their belief system, as well they should. Theclergy will attempt to bring the message to the people, even if the messageloses some purity by incorporating local beliefs. But eventually, some of thoselocal beliefs bubble up to the high priests, who, seeing their merit,incorporate them into their the religion's belief system. Conflict and acrimonywill mark the way, but such is the nature of social interaction. Shall we geton with it?
Later, Jeff
P.S. It was Cliff Joslyn who made many of the same arguments I am making,only he is in ST/Cyb and I am in psychology. Also, he is more literate than Iin your methods. Cliff, if you are out there, I think we are kindredspirits.
Date: Mon Jun 13, 1994 1:27 pm PST
Subj: Re: Responding to Welcomes
[From Rick Marken (940613.1030)] Jeff Vancouver (940613)
> Forgive my ignorance, but I think Tom Bourbon and Bill Powers thinkthat the difference between self-regulationmodels and PCT is that PCT is about controlling perceptions and self-regulationis about controlling behavior. This is no doubt because the promoters of manyself-regulationmodels _say_ they are about regulating behavior.
Yes, partly. But, for me, the main problem is that their research shows noevidence that they understand "control of perception"; there is no testing forcontrolled variables --or anything like it. Self-regulationpromoters have described PCT-likemodels pretty well; but they don't evidence any understanding of these modelsin terms of how they do their research. The proof is simply in the pudding; self-regulationresearch is NOT PCT pudding.
If all you care about is whether the theory described by the self-regulationpeople looks (in diagrams) and sounds (in words) like PCT then I will concedethat self-regulation theories are very similar to PCT. As I recall, the first twochapters in the Carver/Scheier "Self-regulation"book give an excellent description of PCT --really good. It's only by reading the rest of the book --the research "based on the theory" --that you can see that they didn't "get it" at all. Actually, the Carver/Scheierbook is a good example of how people can "talk the talk" of PCT with ratherremarkable fidelity and still not be able to "walk the walk" --ie. do PCT.
Again, I think the basic problem is that people can describe the theorypretty well; they are just not aware of the phenomenon that they theoryexplains; the phenomenon of control. I think people really have to experiencethe phenomenon up close and personal, in all its manifestations --from controlling the position of your hand to controlling where you worship god --to really "get" PCT. Anyone can give a good account of the model, but it's apretty empty exercise unless you know what in the world the model is there for;and it's there for explaining something that's all around you all the time --control.
> Marken is applying his criteria for a theory on the other's theory. Ido not condemn the criteria, merely the blind application. If empirical datais always required, Einstein's GTR would never have lasted the 20 (?) yearsbefore it could be tested.
But SOME empirical data is required or you have no basis for evaluating atheory at all. I think Einstein's theory was constrained by one hell of a lotof data.
> The problem with Marken's approach is that they don't buy/understandthe data you use.
What would you suggest that I do?
> The rigorous requirements, central to your understanding of science,are seen as too rigorous for a science in it's infancy.
I know. I've heard people say that research like that done by Tom, Bill andmyself is trivial because we always get what we expect --often to the third decimal place. A cognitive psychologist friend of mine (whothinks PCT is hogwash, by the way) thinks we could make a better impression ifwe would get results more like what psychologists are used to --the kinds where you really need a statistical test to find out what happened.I'll leave it to Tom Bourbon to comment on this (if he's around and wants to),since the guy who said this is one of my dearest friends; we obviously don'ttalk about PCT much.
> These are self-regulationtheorists [studying higher level variables], and like I said above, I don'tcare what they think they are doing, I find what they are doing useful andinteresting and I think some of you should too.
Great. Tell us what they are doing and how they do it. It's hard for me toget access to all this stuff. I think the net would be a great place to discussit. Believe me, if these people have some good hypotheses about higher levelcontrolled variables and some good ways to test those hypotheses we will beinterested --VERY interested!
> Marken (940606) says you have rejected social control...I find thisvery problematic. Marken (same post) says you have rejected "information aboutthe cause of variation in perception."... This I find very problematic. Now Ieven find that the DME is not part of PCT...This is VERY problematic.
Well, I think you misunderstand what is meant by each of theseterms.
1. We reject "social control" in the sense that we reject the idea thatthere is an entity (called "society") that is a control system exerting controlover groups of individuals. Individuals control and when this occurs ininteraction with other individuals, this is "social control".
2. There is no information about the disturbance to a controlled perceptionin the variance of the controlled perception itself because this variance 1) isa simultaneous result of both disturbance(s) and output and 2) there are anindetermine number of variables that may be a disturbance to the perceptualvariable. This is an important point because it rules out models of controlthat view the controlling system as one that calculates the appropriate"disturbance canceling" output based on perceptual input.
3. The DME is simply an open question; there are no data suggesting itsnecessity, beyond what can already be handled by mechanisms suggested in BCP.So it's not currently part of PCT but it certainly could become part of PCTwere this demanded by the data.
> Reorganization is not simply random.
This may or may not be true. This is why data is important; if you want toconvince me that this is true then you have to present me with data thatsupport this contention. If the data is convincing, I will certainly acceptthis claim. The current model of the reorganizing system is random, not becausedata demand that that be the case, but for logical reasons; ie. how do you knowhow to change control parameters when things go wrong? There may indeed bebiases in reorganization; but this is something to be determined by experiment,not reason.
> Finally, in reference to Vallacher and Wegner, apparently Bill P.thinks they are using PCT, just not acknowledging it. So Marken, when yousay
>> Yes. It looks very interesting. Let's discuss the Vallacher &Wegner model, by all means. But, before we get started, I've got to know: iftheir model produces no quantitative results, if it predicts average behaviorover subjects or over trials, if it doesn't work at all, do I still have tolike it and not criticize it in order to have a "healthy debate'? If so, thenwe can save a lot of time since I can give you my evaluation...
> This is what I mean by unhealthy debate.
And it's what I mean by "no debate at all". Why don't we just discussVallacher & Wegner; if the debate get's unhealthy, we can consult aphysician ;-)
> I do need to get the demo's and simulations. I have Marken's Lotusprogram, which was when I found out about CSG_net (I also heard about it fromLord). I did not understand it then and I don't think it worked when I triedit since.
Get Dag's demo disk!!! The spreadsheet model comes with it's own 1.2.3compatible spreadsheet program. Everything on Dag's disk works. If you don'thave this disk, GET IT. No wonder you think that conventional psychology stillhas something worthwhile to say ;-)
Get thee to Dag's Diskery!
Best Rick
Date: Mon Jun 13, 1994 1:39 pm PST
Subj: Re: Responding to welcomes -Jeff
[Martin Taylor 940613 14:40]
>Jeff Vancouver, (apparently Mon, 13 Jun 1994 11:05:50 -0400)
A very nice posting, that should cause substantial disturbance somereaders. There are a few things with which I might disagree, but they are lostin the wealth of comments with which I do agree.
One main disagreement:
> Now I even find that the DME is not part of PCT. You mean that allconscience thought is irrelevant?
To my mind, these two sentences are unrelated. The question does not followfrom the statement at all, so far as I can see.
Consciousness is not considered, one way or the other, in what I havecalled "core," "classical," or "conventional" PCT. But several people,including Bill Powers, Bob Clark, and I, have speculated about it. I side withthe Powers' in saying that the DME seems to require a mechanism that isessentially a duplicate control hierarchy, and therefore an indefiniterecursion of explanation; for what is it that allows the DME to choose whatdecisions it is to make? A DDME? Mind you, Bob Clark has many times argued thatthis is a wrong understanding of the DME, but I have not yet properlyunderstood his position.
For me, decisions seem to be a property of the program level (which I'm notso sure can be identified with the perception of conditional relationships,despite my acceptance of that idea when Marken posted it). If so, they areunrelated to consciousness except as a side effect of what I speculateconsciousness "really ;-)"is--astate of possible switching among which perceptual signals are to becontrolled. The perceptions that are, and the perceptions that may be,controlled are in consciousness. None else.
Anyway, to reject the DME is not to reject the notion ofconsciousness.
Another point:
> Reorganization is not simply random. A model that shows that it couldbe (e.g., E. coli, which some of you have contented) does not make it so. ButI think I must be misunderstanding you here, too.
I think you are. Last year I presented a taxonomy of possible loci oflearning in a control hierarchy. If I remember, there were 12 kinds. Someinvolve topologically continuous neighborhoods, in which gradient search ispossible (and thus not necessarily random), and some involve spaces withouttopological continuity (and thus necessarily random). For example, in itssimplest possible form, the perceptual-sensoryconnections in a control hierarchy are exactly those of a multilayer perceptron(perhaps a TDNN), and ANY learning algorithm that applies to an MLP or TDNNwould apply to learning what to perceive. That's non-random,usually.
But the main point of control is CONTROL, and failure to control means thatperceptual signals grow when they should shrink, and vice-versa.To change this is to change the sign of the loop gain, a discontinuous step.And to change the sign of any one link might well throw other loops out ofcontrol that use the same link. There's no way for the hierarchy to know whichsets of link signs lead to stability with the present condition of the realouter world, and when the outer world changes, the appropriate sets of signsmay change. That's part of the job of higher-levelcontrol systems--theyhave perceptual signals that include conditional aspects of the outer world(see for example J.G.Taylor's experiments in the 1950's and 60's onconditioning to inverting and other distorting spectacles--there'sa spectacular (!) movie of Seymour Papert falling many times off a bicycle thatillustrates that development of higher control systems). Changes have to berandom, though the rough location in the network where they occur may well benon-random.
Some may think that PCT is a religion requiring high priests (or offeringthe opportunity for them to take their places). I don't think that's a usefulview. It's a useful way of looking at the living world. It may even be true.But then, so may some of our "laws" of physics. Nobody will ever know.Meanwhile (i.e. forever), we use the best tools we have, as they are suited forthe problems that interest us. For much of psychology, PCT is the best tool Iknow. Which is why I spend a lot of my limited effort trying to do what I canwith it and for it.
Martin
Date: Tue Jun 14, 1994 3:32 pm PST
Subj: Re: Jeff, on reorganization
From Tom Bourbon [940614.1658]
Reply to some of the topics addressed by Jeff Vancouver: 13 June1994
> I am pleased that my introduction got such a response. Being ignoredwould have been the ultimate insult.
There was never any danger of your being ignored here! Judging from thereplies to your initial questions and remarks, your self-introductionseemed to be welcomed by everyone.
> I did not feel unduly rebuffed, but, of course, I will respond back.My response centers around themes, not necessarily individuals. Specifically,I consider the following issues: self-regulation, modeling, some theorists in my field, and some minor issues:
> Self-regulation:
> Forgive my ignorance, but I think Tom Bourbon and Bill Powers thinkthat the difference between self-regulationmodels and PCT is that PCT is about controlling perceptions and self-regulationis about controlling behavior.
Yes, I do think that is a big difference (not the only one) between self-regulationmodels and PCT.
> This is no doubt because the promoters of many self-regulationmodels _say_ they are about regulating behavior.
Yes, that is one of the reasons I think their ideas are different fromours: they say as much.
> However, if the model incorporates a test of the difference between aperceived state and a desired state, the result of which drives behavior, thenthe model is describing the control of perceptions.
Not if a "modeler" says her or his model is about how people control theirown behavior. If that is what a modeler says, then the odds are immense thatsooner or later in the paper he or she will make serious mistakes when they saymore about control. I do not allow myself to divorce an author's _diagram_ of acontrol system from what he or she _says_ about the model. I have learned thatI must look beyond the often accurate depiction of a PCT model at the beginningof a paper on self-regulation,to see what the author(s) say in the middle of the paper and at the end. In aremarkable number of cases, a promising beginning collapses into a recountingof traditional control-of-behaviortheories, with a bit of control theory terminology grafted on. (I do not saythis is true of _all_ cases, and I reiterate my eagerness to see any examplesto the contrary that you might provide.)
> If the modeler says the model describes how individuals controlbehavior, they either don't realize their error or they are trying to reach acertain audience.
Then in the first case, they do not understand the control of perceptionand they are not talking about it; in the second, they are giving theiraudience an incorrect interpretation of the control of perception that will bedifficult or impossible to correct, later on.
> But the model still describes the control of perceptions.
I am not as generous as you on this count. In my perhaps prejudiced book,if a modeler doesn't _talk_ about behavior as the control of perception(whether the "modeler" does that as a strategy, or out of ignorance), then themodeler's model isn't about behavior as the control of perception.
> D. Ford's Living Systems Framework is that type of model. I am stilltrying to get a handle on the consequences of the error for those models. So,to answer Marken's (940606.1800) post, yes there are others that describe theCSG loop.
Do they call it the "CSG loop" --the PCT loop? Do they say that behavior is the purposive control of perceptionand that the behaviors that control perception are themselves unintended anduncontrolled? Or are those things you (and I) _would like to see themsay?_
> Modelling:
> A major theme in the responses to my introduction was the role ofmodelling as a test of PCT and the other models I mentioned (indirectly throughproponents -e.g., Locke, Bandura). . . . But the previous argument is mostly on anintellectually level. The reality of rejection letters and lack ofcollaborative spirit from the powers-that-beis painful, dispiriting, and financially and occupationally challenging. Iapologize for raising any of those feelings.
No need to apologize! It just happens that, whenever someone says PCTmodelers are out of line to reject psychology, PST modelers are _very_ likelyto reply that the rejection seems to run the other way.
> I have numerous colleagues with less than kind rejection letters basedon the same reasons that many of you alluded to. On the other hand, I knowothers who have had little trouble on that score. Of course, their work isprobably not completely in line with the "core." Most notably, they do notrequire a working model as a requirement for their work (there is also anemerging appreciation for which journals are "friendly" and which arenot).
Fine. Not everyone _does_ modeling; but do the people to whom you referknow about the modeling that _is_ done? Do they try to incorporate it intotheir writing, perhaps as empirical verification of theoretical points they tryto develop in their writing? Are they careful to see that what they say intheir theorizing is consistent with what has been learned form themodeling?
> Herein lies one of my concerns with the solipsism among the netters.Marken (940606) says "PCT is tested by comparing the performance of workingmodels to actual behavior. In order to compare alternative theories to PCT weoften have to translate verbal descriptions into working models." Thisrequirement confounds the test of a theory with the theory.
Say what? I really don't follow you here. How else are we to test whetherthe core causal assumptions in another theory (as best we can understand thoseassumptions, which are often left unanalyzed by advocates of the othertheories) can really behave as their advocates claim they can?
> Marken is applying his criteria for a theory on the other'stheory.
Of course he is. And so is Bill Powers. And so am I. If someone tells ustheory X explains behavior Y, but does not tell us how or why that should beso, and especially if the someone in question does not demonstrate that theassertion is warranted, then are we merely to say, "OK. Whatever you say osright. Sorry to have asked for evidence that your theory does what you say itdoes?" And if the person goes further to say that PCT cannot do any of thethings we say it does, are we to acquiesce, saying "Of course. We were wrong toapply our criteria to your theory. Go right ahead and apply your criteria toours --reject us by nothing more than assertion, rather than by a demonstration thatPCT does not work." I'm sorry. Jeff, but that is not my understanding of howscience us done --or at least how it _ought to be done_. That's one of the (many) things that Ilike about doing PCT science --I am _required_ to show that _every_ claim I make can be backed up indemonstration --in modeling and simulation. No endless flights of puffery and bald-facedassertion allowed.
> I do not condemn the criteria, merely the blind application. Ifempirical data is always required, Einstein's GTR would never have lasted the20 (?) years before it could be tested.
You lost me here. Shouldn't we expect people at least to show why theythink the lineal cause-effectmodel lurking inside their theory does work, after all? Or shouldn't we expectthem to tell us why control of behavior, rather than control of perception, isthe best way to explain behavior? If they don't pass muster on basic questionslike those two, there is no need for those of us trying to understand and modelperceptual control to look at the rest of the other theory, no matter howsophisticated and inspiring it might seem on other counts.
> Now I realize that these mini-theoriesare not GTR. Indeed, many are very problematic. But I try to separate thechafe from the wheat. Carver & Scheier's self-awarenessconstruct was clearly wanting, but they exposed many to control mechanisms thathad not known them before.
Yes. And some of those so introduced have progressed beyond the Carver andScheier stage. But most have not. Are the many in the latter class any betteroff for the experience, so far as their understanding of perceptual control? IsPCT any better off for the presence of people in that latter group on thereview panels of journals?
> . . . No doubt there are errors in the views perpetuated by the non-PCTmodels, but it is a step.
Some steps along some paths are demonstrably wrong --at least they are if the parties accept demonstrations as evidence.
> Further, theories like Carver & Scheier's deal with many socialprocesses which PCT does not deal with as thoroughly. Identifying thediscrepancies between these models and PCT, and developing empirical tests theresults of which both parties can take as evidence one way or another is, Ibelieve, a reasonable next step.
> And let me be perfectly clear, I think sometimes they have a point andPCT needs to incorporate it.
But up to now the efforts at "bridging" these gaps have run in onedirection. Bill Powers told you if his experience when he tried, many timesover many years, to "reach out" to people who allegedly already occupy the muchprized "higher ground" of higher-levelprocesses. With very few exceptions, they ignored him. That, or, like Bandura,they unleashed a special kind of wrath upon him in print. I have had similarexperiences, on a much smaller scale, with people like Carver and Lord, who areoften cited as examples of theorists who "really understand PCT," no matterwhat they actually say in print.
> The problem with Marken's approach is that they don't buy/understandthe data you use. Most don't understand modeling and/or don't trust itsresults. The rigorous requirements, central to your understanding of science,are seen as too rigorous for a science in it's infancy.
What "science," in which "infancy?" Certainly not psychology, theexperimental study of which is as old as science itself, in the West. The olddodge that "psychology is too new as a science to have any real laws ortheories" won't wash.
And which problem with "Marken's approach?" Have you noticed any others onthis net who espouse similar criteria and a similar approach? ;-))
Now that I have my "E. coli style" models or adaptive control workingagain, I'll say a few things about the section of your post on reorganizationand reference signals, but I'll do that tomorrow.
Later, Tom
Date: Wed Jun 15, 1994 6:16 pm PST
Subj: Re: Responding to welcomes -Jeff
Tom Bourbon [940615.1322]
Back to your reply to our replies to your self-introduction.In particular, back to some of your comments related to what you called "gaps"and to reorganization.
> Gaps & theorists in my field
> . . . How are actions chosen and how does reorganization proceed?I see stabs at the first question on this net (biological needs) but anothertype of intrinsic signal is probably system efficiency variables. The recentdiscussion of uncertainty is one such variable. You are absolutely right toattempt to figure out how the organism could perceive uncertainty and reject itif it can't (but beware, just cause you cannot figure it out does not mean itdoesn't exist). The point is that another perspective has inform the guessthat uncertainty is a reference signal.
Concerning perceptions of "uncertainty." In light of recent discussions onthe net, it would be inappropriate for _us_ (PCT modelers) to try to figure outhow the organism could perceive uncertainty; to do that would be for us topresuppose that the organism _does_ perceive uncertainty and that we shouldtest _our_ understanding of what _others_ mean when they say organisms perceiveit. I believe the process should work differently. Those who say organismsperceive and control uncertainty bear the burden of performing The Test For TheControlled Variable, then showing us that, yes indeed, organisms controluncertainty. Were _we_ to do the testing, then we would be subject to the oftheard criticism recently leveled by Bill Leach (to whom I replied yesterday)that we are guilty of constructing straw versions of the ideas and modelsproposed by other people, then shooting them down unfairly. No more of that forme. From no on, I expect people who propose new and different "models" and"controlled variables" to do the dirty work of identifying, quantifying,testing, and modeling. Then if their models and variables fail, the creditfalls where it belongs, not on us. And if the proposers are not willing to goto that trouble, they can go talk to someone else. (I've finally learned tostop testing other people's ideas.) Up to now, I have not seen "anotherperspective inform a guess that uncertainty is a reference signal," but I_have_ seen an _assertion_, by advocates of another position, that it is. I'mstill waiting to see them perform "the test" and the modeling.
> My reading of your literature and posts is that the identification ofhigher-order reference signals is truly a gap (although I fear I misunderstand theposts here). Many in psychology are seeking to study candidates for referencesignal, including Little, Deci & Ryan, Emmons, Pervin (See edited books byPervins, 1989, _Goal Concepts in Personality and Social Psychology_, and Ford& Ford's, 1987, _Human's as Self-ConstructingLiving Systems.). These are self-regulationtheorists, and like I said above, I don't care what they think they are doing,I find what they are doing useful and interesting and I think some of youshould too.
But I _don't_ find their work useful and interesting, which in no wayimplies that _you_ should not or could not. I've tried to take it seriously, asan "applied" version of PCT, but I can't. For one thing, I don't see themstudying "candidates for reference signals." Rather, I see them performingresearch in the tradition of lineal cause-effect,in order to test the null hypothesis that there is no relationship between twosets of scores, one of which (they assert) is a measure of "self efficacy" or"self acceptance" or some other "self" construct. I've never seen the self-regulationpeople talk about and test for the presence of controlled variables, a testthat, if successful, might support an assumption that reference signals are atwork. On the other hand, while reading their work I've _often_ found myselfthinking, "If only they had tested for controlled variables. If only they hadnot abandoned the PCT model after the first few pages of the article." And soon.
> To the third question, dealing with reorganization (and other matterslike stress and leadership) are people in my field. Proponents of PCT-likemodels (forgive them their transgressions), including Tsui & Ashford (e.g.,Journal of Management, 1994), Edwards (e.g., Academy of Management Review,1992), Cropanzano, James, & Cetera (Research in Organizational Behavior,1993), Lord & Levy (Applied Psychology: An international review, 1994) andManz (Academy of Management Review, 1986) to name a few. Other interestingthinkers include the Germany action theorists (Kuhl, Hechausen) and others(Vallacher & Wegner). Many or most of these authors or theories may havealready been read and rejected. I do not agree with everything _any_ of themsay. But herein lies the challenge: finding which loops conflict, proposingalternatives, and testing the them.
Purportedly, they and we are all talking about the phenomenon of control.Incontrovertibly, we have a model of control that works, but they do not. Whoshould look more closely at whose work? (I don't say this to assert a kind ofsuperiority or pureness, but to show you how I think science _should_ work. I'mdescribing my system-levelpreferences and intentions for how science is done. I learned some time agothat my system-levelideas about science are out of whack. :-) )
> Marken (940606) says you have rejected social control. Well, I am notexactly sure what is meant by social control, but my impression from earlierthreads (and B:CP) is that you rejected social influence. I find this veryproblematic.
Indeed, that would be problematic; but no PCT modeler has ever rejected theidea that one person's actions can affect variables controlled by another andthat, consequently, the results of one person's actions can act as disturbancesthe effects of which are canceled by the actions of a second person. We _have_rejected the idea that a "society" is an agent --a control system that acts purposively to control the actions of a givenindividual person. . . .
> You mean that all conscience thought is irrelevant? This is VERYproblematic. Reorganization is not simply random. A model that shows that itcould be (e.g., E. coli, which some of you have contented) does not make it so.But I think I must be misunderstanding you here, too. Planning and thinking ispart of PCT. Figuring out how that planning and thinking translates intochoosing and doing is where contemporary psychology can help. PCT has beentested in only "limited circumstances" (Marken, 940606), where it has not beentested improvements can be articulated and empirical tests constructed(hopefully).
Rick already addressed your concerns that over the idea that PCT (that isto say, PCT modelers) says conscious thought is irrelevant. I second hisremarks to you, in which he said there was no need to worry about that point. Ido want to say a few things about your remarks on reorganization.
You have _asserted_ that reorganization "is not simply random." Fine. Youcan assert anything you wish, but I believe you might have misread ourdiscussions about reorganization. Never did any PCT modeler say thatreorganization is simply random. I believe we have said that, for a human whois functioning more or less normally, when present strategies fail to producethe intended results, the person may engage in a formal program-levelsearch for alternative strategies. The person, for example, might follow awidely-known(perhaps commercially promoted) technique to "brain storm" for new strategies,or might rely on techniques and procedures learned through personalexperiences. (When I use words like "procedures, programs, and strategies," Irefer to logic-level"actions" that occur when the person attempts to control perceptions at otherlevels; not to programmed actions.) These program-levelattempts at control often include formal "if-then"branching decision trees.
Before program-levelcontrol is established (in general, or for a particular perception), or whenthat level of control fails, or when an instance of perceptual control simplyis not programmatic, what do we do? And how do we account for changes in thecontrol behavior of creatures that show no evidence of a program level? Thoseare some of the instances in which we have suggested that a random ("E. coli")process of reorganization might occur. Further, we have developed several waysto model such a process and have shown, in simulations of our models, that"ecoli reorganization" can work in a number of applications. From that start,we are willing to speculate (not assert as an established fact) that themechanisms modeled in ecoli reorganization might serve as an explanation forreorganization in many other circumstances. As always, those speculations aresubject to confirmation, or to rejection, whenever they are put to thetest.
(In some of my recent modeling of various interactions between two controlsystems [which can mean between two people], I use an "ecoli reorganizationprocedure" in the model for one of the systems. The modeling is difficult todescribe in a few words --I'll demonstrate it in Wales next week and write about it in Martin Taylor'sissue of IJMMS. I use the ecoli procedure to make one of the interacting models"adaptive," in that it randomly --and I assume "unconsciously" --changes its own gain, or its own reference signal, or the sign of its ownfeedback loop --all as part of its "attempt" to control of its perceptions of the othersystem.)
We do not contend that ecoli reorganization explains all changes in controlbehavior. As you say, the uttering of such contentions would not in itself makethem so; that is one reason we never utter them. But it is also true that ecolireorganization _does_ work in the instances where we have shown it to work --a fact that will not go away just because some people assert that it should.(Not that _you_ are making that assertion now, but assertions very much like ithave come up in the past.)
> Some mentioned I should send my recent paper to them and others forcomment. I appreciate that, copies are on their way. One comment: The paperargues for the adoption of a PCT-likemodel as a paradigm for organizational behavior.
If I was not on the original list to receive a copy, please add my name.I'd like to see it. :-). . .
> In conclusion, there are many roles to be played here. To evoke thereligious analogy again (understanding that science seeks data moreemphatically than religion, but it is so appropriate). The high priests willattempt to maintain the purity of their belief system, as well they should.The clergy will attempt to bring the message to the people, even if the messageloses some purity by incorporating local beliefs. But eventually, some ofthose local beliefs bubble up to the high priests, who, seeing their merit,incorporate them into their the religion's belief system.
Ouch! I know you are using the analogy to religion in the loose sense, butwhenever I see it applied to PCT, I cringe. It too easily supports the ideathat PCT modelers are "high priests" defending a belief-based"faith," with a social hierarchy trailing away beneath them --a hierarchy of clergy, acolytes, unwashed believers and (I suppose) heathens,heretics, and other rotten types. I don't like that image at all! I'd rathersee _everyone_ jump in over their heads and play the game of PCT science forall it's worth.
> Conflict and acrimony will mark the way, but such is the nature ofsocial interaction. Shall we get on with it?
Come on in! I think there will be far less conflict and acrimony that youexpect. :-)
Later, Tom
Date: Wed Jun 22, 1994 6:09 pm PST
Subj: self-regulation
[From Jeff Vancouver 940622)] Tom Bourbon 940614.1658
Suspecting that I am one of these self-regulationpeople, it is important that I understand what discrepancies that model createsfor you all (or what discrepancies PCT creates for me). I should preface thiswith my understanding of the work of others. No one has it right. They are allwrong. You are wrong. I am wrong. Everybody is wrong. Given that, I usually tryto find what is useful to me. Useful in filling a gap in my understanding oruseful in provided a testable hypothesis I can get a pub from, etc.
Your (Bourbon 940615) lack in interest in uncertainty seems to be like mylack in interest in some of the issues of interest to some PCTers. We cannot beinterested in all of it, it is too much (do you buy that). But I thoughtidentifying reference signals was a primary concern. One of the first steps, ifnot the first, is to guess at a possible reference signal. Do you guessrandomly? Using other's ideas might help to make an educated guess. That iswhat I thought the uncertainty guess was about.
Now, onto self-regulation.I get the impression that PCT likes to think of actions as unintended. Icertainly think they can be. But I also think they can be intended as well.That does not mean intent equals success. Now one thing I had noticed amongsome self-regpeople is an exclusive interest in intended behaviors. I see this as a boundaryof their theories. Some, I think are aware that this is a boundary. Others areoblivious to the limits of their model. Attempting to point out the boundarygives me something to do, so I don't mind too much (I have not gotten rejectionletters for PCT-likeideas yet). But are you, Tom, telling me that PCT does not included intendedbehaviors in its model? I know the net well enough to know that there are nobehaviors beyond PCT's description, so it must be the intentions that disturb.Why do intentions disturb?
Do the people I mentioned and know use or know of PCT modeling? I cannotanswer for all of them, but I think many do not. Gosh, I have been reading thisstuff for a long time and can't distinguish what I know from the modeling fromwhat I know from the verbal descriptions of the theory.
I referred to D. Ford's living systems theory (LST) as very close to PCT.You asked if he defines behavior as the control of perception. Hard to tellbecause he calls perception a kind of behavior. You ask if he calls his loopthe PCT loop or the CSG loop. No, he does not seem to like your jargon (so itgoes), but judging by the references, he gives Powers some credit (I am stillreading it). In fact, M. Ford's book using LST to discuss motivation givesPowers (of course, he lumps in Lord and Carver & Scheier and other) a lotof credit (I did read that one). M. Ford's complaint is the jargon and the lackof coverage of emotion (they probably don't know that Powers' chapter onemotions was cut). When Bill gets back we will have to ask if the Ford's areone of those who ignored him.
You say (Bourbon 940615) that you have a model that works and those otherdo not. I would contend that you have a model that works in a limited context.What is going on at the program levels and above seems much more speculative.These others have equally speculative models. You simply don't appreciate theirmethod of testing their models. I think that is largely because you havedifferent aims than them. What remains a question is whether there is anyoverlap. Given that my aims are much more like theirs (establishing a career byshowing something useful to the world) than yours (developing a model of humanbehavior), and I am paying some attention to your work, I must believe there isan overlap.
I guess my interest in this net is to learn more and to help others learnmore. The learning is to improve the designs of our experiments and models. Iam particularly interested in identifying the gaps in understanding, whetherthose gaps are within PCT or between PCT and other models (by this last I meanmostly to get the others up to speed. I am outlining a continuation of Bill'sblunder's paper, so their problems with the control of behavior and other gapsinterest me).
Later, Jeff
Date: Wed Jun 29, 1994 5:20 pm PST
Subj: Re: self-regulation
From Tom Bourbon [940629.1222] >[From Jeff Vancouver 940622)]
> Suspecting that I am one of these self-regulationpeople, it is important that I understand what discrepancies that model createsfor you all (or what discrepancies PCT creates for me). I should preface thiswith my understanding of the work of others. No one has it right. They areall wrong. You are wrong. I am wrong. Everybody is wrong.
OK, I get it. We are all wrong. ;-)
Now we can get down to the business of examining the performance of PCTmodels --and the performance of other kinds of models, if the necessary informationhappens to be available. (One problem with my reading of the self-regulationliterature is that nowhere in it have I ever seen a working model for self-regulation--and I'm afraid I've lost all patience with the idea that any of us shouldaccept a lot of talk about something as constituting a "model.")
I'm more interested in some of the specific things you say about PCTmodeling, later in your post, but first a few remarks about your initialideas.
> Your (Bourbon 940615) lack in interest in uncertainty seems to be likemy lack in interest in some of the issues of interest to some PCTers. Wecannot be interested in all of it, it is too much (do you buy that).
Maybe. Let's see if we are thinking about the same thing here. I buy theidea that if I decide to say I am interested in information theory, where"uncertainty" has (or once upon a time had) a precise technical meaning, I hadbest pay attention to (show an interest in?) the things information theoristssay about uncertainty. If I am not interested in (or informed about)uncertainty, as _they_ use the term and its associated measures, perhaps Ishould hold back from telling them they don't know what they are talking about,or from asserting that it's all a matter of preferences and personal interests.On the other hand, if an information theorist asserts that informatic measuresof uncertainty are foundational to the workings of perceptual control systems,then I buy the idea that I might be able to say something to inform thediscussion.
Is that the kind of thing you meant?
> But I thought identifying reference signals was a primary concern.One of the first steps, if not the first, is to guess at a possible referencesignal. Do you guess randomly? Using other's ideas might help to make aneducated guess. That is what I thought the uncertainty guess was about.
I'm sorry, Jeff. I don't follow you here. I thought this thread began withsome assertions that uncertainty can be perceived directly, hence controlled. Ididn't think it had to do with anything as mundane as, for example, my"uncertainty" about (simply not knowing about) which reference signal a controlsystem might be controlling. In fact, during all of the discussions on csg-labout information and uncertainty as they might relate to PCT, I've been alittle concerned that the advocates of information theoretic terms were usingthem in the common sense manner, rather than the technical one. If all we aretalking about is "not knowing which", rather than about quantified bits ofuncertainty, then the discussion takes on a different form.
> Now, onto self-regulation.
Good! The real meat!
> I get the impression that PCT likes to think of actions asunintended.
Well, aside from the fact the PCT is a theory and as such it doesn't think,;-)I think I agree --under the right circumstances, such as when actions are not intended --which is the case _every time_ the _results_ of actions are intended. If the_results_ are intended, the actions _cannot_ be --no ifs ands or buts about it. That's not a matter of theory or conjecture orpreference; it's just the way control of perception works.
> I certainly think they can be.
We agree on that.
> But I also think they can be intended as well. That does not meanintent equals success.
Of course they can be intended. And of course intent does not _necessarily_equal success.
Not too long ago I wrote a paper about how a person can control his or herperception of his or her own actions. The paper was:
Mimicry, repetition, and perceptual control.
It was in the famous journal, _Closed Loop_, Fall 1993, Vol. 3, No. 4, pp.55-71.
One point of the paper was that people can act to feel the movements oftheir hands matching the seen movements of a target. I called that condition"mimicry." That's a kind of "control of movement", isn't it? A person can dothat. So can a simple PCT model of the person.
For me, an interesting thing about the results is that when a person or PCTmodel makes perceptions of self-movementsmatch a seen movement, any other variables affected by the self-movementsgo "out of control" and the results look like those produced by a purelyreflexological S-Rsystem. Control your perceptions of your own actions, as in mimicry, and youlose control of your perceptions of other consequences of your actions; controlyour perceptions of other consequences of your actions, and you lose control ofyour perceptions of your own actions. In _either_ case, the person and modelact to control their own perceptions; it's all a matter of _which_perception(s) they intend to control.
In the same paper I showed how a person and a PCT model can control theirperceptions of their own actions in another way. People can act to make theirmomentary felt hand movements match a remembered pattern of movements. A simplePCT model can duplicate the results. I called that condition "repetition."Isn't it another kind of control of actions? Incidentally, the results looklike those when a purely "cognitive," plan-drivenor command-drivensystem controls its own actions as commanded outputs; perceptions of theactions are controlled and perceptions of the consequences of those actions goout of control. As I said above in the discussion of mimicry:
"Control your perceptions of your own actions . . . and you lose control ofyour perceptions of other consequences of your actions; control yourperceptions of other consequences of your actions, and you lose control of yourperceptions of your own actions. In _either_ case, the person and model act tocontrol their own perceptions; it's all a matter of _which_ perception(s) theyintend to control."
> Now one thing I had noticed among some self-regpeople is an exclusive interest in intended behaviors.
Then we see the same things in their literature. I hope you see why we (PCTmodelers) have so often said negative things about the self-regulationpseudo-modelswe have seen; none of those word-modelscan serve as a model for living systems, unless, of course, it does not matterthat a "self-regulating"organism has no control whatsoever over anything other than its own actions. Ifthe consequences of actions matter, then no system that controls only its ownactions, by whatever means it achieves that control, can survive.
> I see this as a boundary of their theories. Some, I think are awarethat this is a boundary. Others are oblivious to the limits of their model.Attempting to point out the boundary gives me something to do, so I don't mindtoo much (I have not gotten rejection letters for PCT-likeideas yet).
Don't worry. You will get them, and soon, if you take on that task!
> But are you, Tom, telling me that PCT does not included intendedbehaviors in its model?
I wasn't telling you that (see the discussions above, about mimicry andrepetition); maybe something about my earlier wording or style led you awayfrom what I was trying to say.
> I know the net well enough to know that there are no behaviors beyondPCT's description, so it must be the intentions that disturb. Why dointentions disturb?
I don't understand this, Jeff. Can you tell me more about what you meanhere?
> Do the people I mentioned and know use or know of PCT modeling? Icannot answer for all of them, but I think many do not.
I haven't seen one of them do it. Heck, how could they; we can't publish intheir journals. The editors and reviewers keep saying, "We've never seenanything like this before and we don't want to see it now." :-)
> I referred to D. Ford's living systems theory (LST) as very close toPCT. You asked if he defines behavior as the control of perception. Hard totell because he calls perception a kind of behavior.
His use of the terms might (or might not) be a sign of some common ideasabout what is controlled and what does the controlling. I'd need tolook.
> You ask if he calls his loop the PCT loop or the CSG loop. No, hedoes not seem to like your jargon (so it goes), . . .
Jargon is _always_ a problem. That's one reason some modelers are so big onusing models to make their ideas less ambiguous, and then to test for whetherthe relationships and interactions expressed in those ideas really produce thecontrolled results they (the modelers) believe they will. PCT is not about thewords and the jargon; it is about an empirically observable (not conjectured)phenomenon of control by living systems, and about a working (not conjectured)generative model that literally (not hypothetically) duplicates many examplesof control by living systems.
Does Ford use working (behaving, generative) models to make his jargon lessambiguous and to test his ideas about self-control?Or does he simply offer his own jargon, which he understandably likes more thanours?
> but judging by the references, he gives Powers some credit (I am stillreading it).
Jeff, _many_ people "give Powers credit" in the form of citing him. Someeven go further and pretend to summarize or interpret Powers's ideas. Most ofthose people cite him out of context or inappropriately (once in a great whilethey cite another PCTer or two), and many who pretend to interpret Powers orPCT do it terribly and in a misleading way. (I can cite examples of all ofthese ways of "giving credit," but not right now.)
> In fact, M. Ford's book using LST to discuss motivation gives Powers(of course, he lumps in Lord and Carver & Scheier and other) a lot ofcredit (I did read that one). M. Ford's complaint is the jargon . ...
The fact that Ford lumps Powers with Lord, and Carver and Scheier makes meeven more suspicious of Ford's grasp of PCT --I mean PCT as the science of perceptual control, not "PCT" as the idea that,somehow or other, people control themselves. In what way(s) do you see Ford"giving Powers a lot of credit?" Is it by citing Powers a lot, or maybe bysaying," I give Powers a lot of credit for doing ...?"
> You say (Bourbon 940615) that you have a model that works and thoseother do not. I would contend that you have a model that works in a limitedcontext.
I'll stand by the range of applications of the PCT model and I'll bet thatit compares quite favorably with the range of any other _working_ model ofbehavior you can find on the market today. I'll go even further and make astatement I do not believe is at all rash: There is no other generative modelof behavior that can be applied successfully in as many settings as the PCTmodel. Period. Prove me wrong and I'll admit it publicly, right here on csg-l.
> What is going on at the program levels and above seems much morespeculative.
Hmm. My presentation in Wales was titled, "Program-levelcontrol of a sequence of relationships." The task and the accompanying modelgot to the program level and the model included three different levels ofperceptual control. Kent McClelland talked about (but did not model) socialcooperation in terms of the control of perceptions at the levels of systems,principles, and programs. There have been a few demonstrations of control ofperceptions at the systems level. The degree to which our remarks about higherlevels can be called "speculative" is shrinking, but not nearly so fast as itwould were more people to stop saying, "you people haven't done much at thehigher levels," and join in to help do the work. ;-)(That's a hint, and a plea for help.)
> These others have equally speculative models.
I disagree. Their models are far more speculative than the model in PCT.Their "models" don't even work at the _lowest_ levels, so they _cannot_ workfor the higher levels, which _must_ depend on the lower ones. That's one of thebiggest problems with their non-models--they don't behave at all.
> You simply don't appreciate their method of testing theirmodels.
Agreed. I do not appreciate their method of testing models. I am trying tohelp develop a science of behavior that can explain the actions of individuals.The requirement that a "model" need only "sound plausible" is far too lax tosuit me. That's all they require --that, and perhaps some statistically significant differences between means ofgroups, or some statistically significant correlations, either of which isuseless for explaining or predicting how a particular individual will act andwhat will be the consequences of those actions. I do not appreciate thosemethods of testing "models."
> I think that is largely because you have different aims thanthem.
You are right. I want to help study a phenomenon that is seen everywhere innature, not just in statistical abstractions, and to develop a generative modelthat works as an explanation for that ubiquitous phenomenon.
> What remains a question is whether there is any overlap. Given thatmy aims are much more like theirs (establishing a career by showing somethinguseful to the world) than yours (developing a model of human behavior), and Iam paying some attention to your work, I must believe there is anoverlap.
Hey, the fact that I may soon be out of work again shouldn't enter intothis discussion. ;-)I was doing PCT even when I was a legitimate working person with tenure (if_any_ person with tenure can be called legitimate). I want to understand andexplain one of the things living systems do all of the time and very well --control. I'm especially interested in the hierarchy of control and ininteractions between independent control systems. Who ever said I didn't wantto make a living at the same time? It just isn't working out that way, that'sall.
As for what is useful to the world, I'll go way out there and assert thatmost of what psychology claims to have learned about behavior is not onlyuseless, but is also dangerous to the world. No joke. :-(
Later, Tom
Date: Thu Jun 30, 1994 8:52 am PST
Subj: rightness, reorganization, self-regulationand alerting
Jeff Vancouver [940629]
on the "rightness" of PCT:
[Bill Leach 940622.23:59, Martin Taylor 940622, & Tom Bourbon940629]
I am claiming PCT is wrong in the sense of Kuhn's paradigms and Socrates'"all knowledge is tentative." Eventually, assuming it gains favor, PCT will be_replaced_ by another paradigm because it is no longer _adequate_. By adequateI mean it no longer points to ways of closing or identifies new gaps betweenour model and our observations of reality. By replaced I mean it will either becompletely tossed or incorporated into the new paradigm. My guess is thelatter, but this is crystal ball stuff.
In the meantime, many gaps are articulated by PCT and much work needs to bedone. You see, science is about identifying and filling gaps. So when Bill whenyou say:
> Unlike ALL other psychological work, PCT is NEVER a matter ofopinion.
You are missing the point. We seek opinion, so that we can shoot is downwith data or logic (models being one form of logic). That is what we do. Aperfect theory (no gaps) would be useless to scientists (although very usefulto engineers) because it would give them nothing to do. I have no fear of this.The question is not whether PCT is right or wrong for it is surely the latter,but whether it is the best model out there at this time. I think it is. Ofcourse one's definition of best depends on one's constituency. For theengineer/applied scientist, best means it can help in designing controls or atleast predict behaviors or events (i.e., it is likely to be useful). For thescientist, best is gap identification and filling (even more specifically, itallows one to produce a publishable/fundable program of research, wherefundable often relates to the applied scientists' criteria).
I felt vindicated when Martin (940623) said to Bill L.:
> I see PCT as a framework theory, like Quantum Electrodynamics inphysics.
Of course, Martin is likely to get jumped on for that comment (probablyalready has -I suffer a severe lag on the net).
The point of this is that some things PCT (or any theory) cannot and evenwill not handle. There is no reason to get too worked up about this. It is partof the game. ALL KNOWLEDGE IS TENTATIVE.
on reorganization:
[Martin Taylor 921005 and 940622]]
Thanks. Your posts on reorganization confirmed my understanding of thepossible reorganization process. I agree when you say "this aspect [howreorganization works] is not securely developed." I understand there is a bigdifference between theoretically possible and empirically tested. I also thinkthere is a big difference between theoretically possible and modeled. Bytheoretically possible I mean that nothing in the theory precludes thepossibility. It does not mean the theory requires all the types you suggest(although it does require some minimum types). But until all this is worked out(i.e., modelled and tested against the real world), we can assume with cautionall 12 (and maybe more) as possible. That is what I wanted to hear. (Of coursethe negative side of all this potential is trivalization. This is theneoFreudian problem -anything is possible within the theory and thus nothing is predictable or muchless controllable.)
To my question [Jeff 940622] "Is the process of deciding a course ofaction, creating a new or using a different reference signal?" you said:
> The tenor of this question suggests the control of output, whereas PCTis based on the idea that what is controlled is perception...
Yes, I get that (see more below). But you go on to say:
> "deciding a course of action" can be rephrased as "deciding referencelevels for lower-levelperceptions" and when it is phrased like that, the answer becomes self-evident. There is no "new of different" references signal, but the values of existingreference signals are affected.
Apparently it is not as self-evidentas you suggest. First, by "different" I meant different _values_ of existingreference signals. If there are different values for a reference signal, somemethod must exist for select the value that is sent. That something ispresumably the effector function. Changes to that function is one of your 12types of reorganization. Second, "new" means new ECS. Can't a new ECS becreated as part of a conscious process, as exemplified by Bandura's modelingconcept? Surely this is a reasonable question? I think this is what Tom meansbe the term "mimicry" (Bourbon 940629).
on self-regulation:[same people]
I suspect I am wasting my fingers here, but I think the main problem isthat many netters have their gain set too high on this behavior thing. I see itin Bill P. writing from early on, which I think accounts for it. Specifically,Bill was fighting the behavioralists. It was a very important fight back then.What seems missing is the understanding that the fight was won. PCT was maybeonly a small platoon in the war, but the war is over.
That is, contemporary psychology assumes we are talking about perceptions,not behavior. I realized this after my reaction to most posts I received backwas "of course." When contemporary psychology says we attempt to controlbehavior, that is simply a shorthand for saying we attempt to control ourperception of behavior. Most know we cannot "know" our behavior, merely ourperception of the results of it. Usually, there is little difference, so likephysics often ignores friction, psychology often ignores biases and mis-perceptionsbecause they are of minor importance. On the other hand, some concernsthemselves with the question of when biases and mis-perceptionsare likely to be important.
You can call this imprecise use of the language, but one can also call itprudent and expedient. The norm on the net is to call it imprecise, so I willtry to respect that lest I get more "modify your behavior" messages.
But there is one more thing before I leave behavior and self-regulation. The meaning and role of intentions. Bill L. and Martin refers toreference signals as intentions. This may be what others on the net think of asintentions. When I say intentions, I am referring to contemporary psychology'suse of the word, which is useful for PCT as well. That use is as thephenomological experience of wanting to engage in behavior. (A PCTer might saythey wish to change the perception of one of their behaviors, but since it is aphenomenological phenomenon, it is perfectly legitimate to say "wanting tocontrol behavior"). Usually these intentions are as a response to a goal theperson "wants" to accomplish. In this sense both intentions and goals are notreference signals. They may be fair representations of references signals, orthey may not. Our conscious experience of our goals and intentions is probablyonly an inference of those reference signals. Some in the self-regschool don't make this distinction, many do.
But, the question is, how does this conscious processing interact with theperceptual control hierarchy. Most interesting is how the phenomonalogicalexperience of a goal and a desire to behave in some way to achieve that goaltranslates into PCT reference signals. Because most of the time, consciousdesires are translated into behaviors, which affect the environmental variablesas intended (except when the environmental variable is your children ;-)),which affects your perception of the environmental variable. Further, yourbeliefs and attributions about all that process affect subsequent goals andintentions. I think these are all legitimate phenomenon for psychologists tostudy and should be of interest to some PCTer as well.
Indeed, when Martin says:
> ...I control a perception of typing this message as part of thecontrol of my intrinsic variable "blood sugar" (supposing that to be anintrinsic variable). How? I see typing this as a way of perceiving myself tobe better informed on the subject I write about, and as a way of helping otherpeople to be better informed ...
Part of what self-regulationpsychologists are trying to do is understanding why typing, etc., would controlblood sugar.
Meanwhile, psychologists need to get a better handle on the underlyingmechanism of behavior and PCT provides a very good model for doing that. Many,seeing the limitation of a purely conscious model are beginning to focus onconsciously unintended behavior (that is, unintended perception control). Theyare missing the framework (psychologists are paradigm shy -having been burned in the past).
You all will probably hate me for this. But I am beginning to see themodeling and PCT v. self-regdebate as a form of the reductionist argument. To use the computer analogy.Many in my field (I/O psychology) are interested in the higher-levelsoftware. They want to create macros and engage in other resource savingtechniques. Those in the more higher-levelmental processes on also interested in the higher-levellanguage, but are more into the code. Those in the lower-mentalprocesses are more into assembly and machine language. Neuro and bio people aremore into the electronics that make it work, etc. etc.
Now when I was studying computer science in college, I took the usualsequence of programming courses and learned how to program. But I did not feelI understood how the computer worked. Then I took the computer architecturecourse. It covered the link between the hardware and the software. Ahhh, now itwas making sense. I could explain things to my level of satisfaction. I newthat if I could understand it even better if I took some electrical engineeringand physics courses, but I was satisfied with my level of understanding. I didnot feel the need to know at those levels. Now, I program some in basic and Iam learning visual basic. My programming courses have helped the most in mylearning, but the architecture course has helped as well (maybe more than Iknow). Nonetheless, I can focus my learning efforts on that higher-orderlevel (learning the programming language), and not the architecturallevel.
PCT is like that architecture course. My experience was very similar -now I have a much better feel for how it all works. Their is still some fuzzyareas, but some are not as critical to "writing efficient code" of the type Iwould write as others. A large part of the trick is to figure out what oneneeds to know to proceed efficiently. For me, I feel knowing something of thearchitecture is helpful. Many do not, they may be right, they may be wrong. Theproblem is that they cannot judge because they do not even know that a model ofit exists. But I digress. The point is that all the players, from the physicsto the engineers to the various levels of programmers can work somewhatindependently -not completely independently nor completely interdependently. Finding the mixis the source of the conflict here. Perhaps we (or someone) should model it,because currently it is all a matter of opinion.
on alerting experiment [Martin Taylor 940623]
Martin, you and I are much closer on this than you appear to think.
First, by ECS I assume you mean Effector, Comparator, Sensor.
Second, I said
>> But the real question is the phenomenon you are trying to examine.Perhaps you have addressed this in previous posts, but what I think of when yousay "alerting" is the allocation of attentional (or some other) resources tosome part of the loop.
You said.
> Quite the opposite. Firstly, the alerting perception is specificallyNOT part of the loop of any controlled perception, just as the beep is not partof the perception of the line-locationthat is to be controlled. Secondly, the whole notion of alerting is to avoidthe requirement of allocating attentional resources.
So I am correct that this is largely an attentional resource allocationexperiment. I never meant to imply the alerting loop was part of the line-locationloop. Indeed, my original fear is that the alerting loop might take attentionalresources away from the line-locationloop(s). Your main idea is that the beep will free up attentional and visualresources, thus making control of the line-locationmore efficient (understanding control efficiency is key to selling the PCTmodel to my constituency).
Your hypothesis assumes the system would seek (have a reference signal) toincrease efficiency (otherwise the beeps would be ignored). I think that is aperfectly reasonable assumption. Although specifying it would help (andmodeling it would be even better)
Problems that still trouble me:
1) the unmodeled efficiency ECS (which may be intrinsic)
2) What resources exactly are involved (attentional, visual). The problemhere is their lack of independence, must less our understanding the wasattentional resources means.
3) The changing focus problem I mentioned in my previous post. I stillworry that, in the name of efficiency, one only addresses line-locationwhen the beep sounds. This way, visual resources can focus on correcting andnot searching. Attentional resources can be devoted to sound/columnidentification. This process could be identified if, after the subject performsfor awhile in the complex condition, complexity is lowered by slowing thetarget movements. If I am correct, no attempt will be made to correct a line-locationuntil a beep is heard, even if the previous deviants was corrected. In otherwords, the subject will be slow to return to a search mode even thoughresources required for correcting could be spared.
I few points I did not understand.
> There is no argument IN PRINCIPLE about the availability of degrees offreedom in the working of the brain ... If my guess is right about attention,that what is attended are those perceptions to which or from which control isto be given or taken, then the output df limit the number of attentional dfthat can usefully be deployed.
I got lost. Whose principle? PCT's? There are definitely theories whichpostulate limited attentional processes. I buy this concept. And I think it iscritical to your experiment. You are focusing on output df, but the criticalconstraint in your experiment might be the limit to attentionalprocesses.
When you say:
> This is what I called the "search mode;" a perception that "should be"controlled cannot be effectively controlled because of data lack. Otherperceptions are controlled in such a way as to remedy that lack...
You are discussing the other ECSs that may take various resources. Thisresources may by physical (with the 125 df) and/or they may be attentional(like worrying). Your hypotheses is that providing an alert in the environmentwill obviate the need to search thus free more resources to reducediscrepancies. I am saying it may also obviate the need to worry. In fact, Ithink it is more likely to be the latter rather than the former, since the beepwill still require a search for the offending column, although less so as tonesbecome associated with columns.
This post told about four hours to compose, how do you people do it?
Tom, I will get back to your specific questions later.
Later Jeff
Date: Thu Jun 30, 1994 3:50 pm PST
Subj: uncertainty, & self-regulation
[From Jeff Vancouver 940630] Tom Bourbon [940614]
I will try to keep this brief.
on uncertainty:
I fall somewhere in between information theorists and "not knowing which",but I am winging it somewhat. The formal meaning of uncertainty in IT is notknowable by humans, therefore they are not controlling it and "informaticmeasures of uncertainty" are not likely to be of much use when consideringcontrol. We are much more likely to control the mundane sense of uncertainty.But the mundane sense is merely a poor measure of the IT sense because that isall the better we can sense it. But don't pay attention to me here. I don'treally know. Uncertainty was just an example of a possible reference signal Ithought people on the net were interested in and that that interest had comefrom another theory.
on mimicry:
It sounds like your subjects could not control all the variables at once.Why do you suppose that is?
on modeling and psychology:
"Useless and dangerous" -that is a limb. I must get back to it, watch out.
Later Jeff
Date: Thu Jun 30, 1994 3:56 pm PST
Subject: Back in service; a few comments
[From Bill Powers (940630.1330 MDT)]
Jeff Vancouver (940629) --
> We seek opinion, so that we can shoot it down with data or logic(models being one form of logic). That is what we do. A perfect theory (nogaps) would be useless to scientists (although very useful to engineers)because it would give them nothing to do. I have no fear of this.
Neither do I fear for science running out of something to do. On the otherhand, I wouldn't avoid making a theory as good as possible just to keepengineers from taking over science. I think we have to work as if we're tryingto perfect theories, especially by subjecting them to most direct challenges wecan devise --to do otherwise is simply to encourage sloppiness.
PCT is indeed wrong in the terms you state --but the catch is that we don't know where it's wrong. It has survived all theexperimental challenges we have thrown at it lately, although of course it gotinto its present shape by failing a lot of experimental challenges first. Idon't worry about perfectibility anyway. If we think of a test that the theorycan't handle, we'll just get a better theory out of it. I'm not too worriedabout a total catastrophic failure at this point. Even Newtonian physics stillworks perfectly well in its own little niche (most ordinary affairs, includingspace travel).
> That is, contemporary psychology assumes we are talking aboutperceptions, not behavior. I realized this after my reaction to most posts Ireceived back was "of course." When contemporary psychology says we attempt tocontrol behavior, that is simply a shorthand for saying we attempt to controlour perception of behavior. Most know we cannot "know" our behavior, merelyour perception of the results of it.
This is certainly news to me. What field do you refer to as "contemporarypsychology?" I guess I thought that behaviorists, who are still around despitepremature funeral announcements, and cognitive psychologists as well, stilltried to account for the actions an organism performs, not the perceptions itcontrols. Have psychologists suddenly stopped arguing with each other?
Best to all, Bill P.
Date: Fri Jul 01, 1994 7:19 am PST
Subj: Re: uncertainty, & self-regulation
From Tom Bourbon [940701.0908] >[From Jeff 940630] >Tom[940614]
> I will try to keep this brief.
I'm disappointed. :-(You left out most of the things I really wanted to hear about. After readingyour earlier post addressed to me [From Jeff Vancouver 940622)], I tried toreply to each of your questions and to ask some questions of my own. I washoping for a thoughtful answer, not a brush off. ;-((
> on mimicry:
> It sounds like your subjects could not control all the variables atonce. Why do you suppose that is?
That's it? Shucks.
I was trying to see what _you_ thought about the topic. Why _can't_ aperson control actions, and the disturbed consequences of actions, both at thesame time? Or do you think they should be able to control both at the sametime? I'd like to see your thought on the topic --then something about how the mimicry-repetitionstudy relates to that idea. As I recall, I was replying to a post from you, inwhich you said the following:
==============================
> Now, onto self-regulation.I get the impression that PCT likes to think of actions as unintended. Icertainly think they can be. But I also think they can be intended as well.That does not mean intent equals success. Now one thing I had noticed amongsome self-regpeople is an exclusive interest in intended behaviors.
> But are you, Tom, telling me that PCT does not included intendedbehaviors in its model?
===================
So I took you seriously and drafted a reply.
And what about Ford, and my questions to you about whether the people yousay are doing the same thing as PCT do any modeling? That's not an idlequestion. If they don't do modeling, and if they don't explicitly say theybelieve people control their own perceptions, then there is no direct evidencethat they are doing "the same thing" as PCT. I had hoped to learn _your_ ideasabout that. If you think they are saying and doing the same things as PCT, thenwhat are _your_ criteria, what evidence do you see, for deciding that is so?I'm not challenging you or attacking you; I'm a simple inquiring mind and Iwant to know. :-)
> on modeling and psychology:
> "Useless and dangerous" -that is a limb. I must get back to it, watch out.
I'm watching! I can't wait! Don't disappoint me this time! ;-)
Later, Tom
Date: Thu Jul 07, 1994 10:20 am PST
Subj: On a limb
From Tom Bourbon [940707.1217]
Back in June, Jeff Vancouver replied to a post from me. He concluded hispost with a remark that psychologists were trying to do things that are usefulto the world. I made this reply to Jeff.
==============================
From Tom Bourbon [940629.1222]
As for what is useful to the world, I'll go way out there and assert thatmost of what psychology claims to have learned about behavior is not onlyuseless, but is also dangerous to the world. No joke. :-(
===================
To which Jeff replied:
==========
[From Jeff Vancouver 940630]
"Useless and dangerous" -that is a limb. I must get back to it, watch out.
===============
And I said:
==================
From Tom Bourbon [940701.0908]
I'm watching! I can't wait! Don't disappoint me this time! ;-)
==============
Jeff, I'm still waiting out here on the limb. Come on out --but before you start sawing, be sure to check and see which of us is fartherfrom the trunk! ;-)
Tom
Date: Thu Jul 07, 1994 3:22 pm PST
Subj: Re: uncertainty, & self-regulation
[from Jeff Vancouver 940707] Tom Bourbon [940701.0908]
and Tom Bourbon [940614] and [940707] and [Marken 940707] while I
was writing this.
I am glad you are curious about what I have to say. I am sorry I am notfast enough or long enough in the tooth for you. Remember, I am trying to gettenure.
ME:
>> on mimicry:
>> It sounds like your subjects could not control all the variablesat once. Why do you suppose that is?
Tom [940701]
> That's it? Shucks.
> I was trying to see what _you_ thought about the topic. Why _can't_ aperson control actions, and the disturbed consequences of actions, both at thesame time? Or do you think they should be able to control both at the sametime? I'd like to see your thought on the topic --then something about how the mimicry-repetitionstudy relates to that idea.
My thoughts are that it depends on the resources and the interactionbetween the perceptions that are trying to be controlled. If by "disturbedconsequences of action" you mean, for example, that I knock over a lamp whilemimicking some movement with my hands, I will not be able to prevent the lampfrom hitting the floor using my hands and still mimic. The degrees of freedomin my hands don't allow it. But if I use my foot, which is not involved in themimicking, then maybe I can continue to mimic and correct the falling lampsimultaneously, if I can divide my attentional resources appropriately.
Tom again
> I was replying to a post from you, in which you said thefollowing:
> Now, onto self-regulation. I get the impression that PCT likes to think of actions as unintended. Icertainly think they can be. But I also think they can be intended as well.That does not mean intent equals success. Now one thing I had noticed amongsome self-regpeople is an exclusive interest in intended behaviors.
> But are you, Tom, telling me that PCT does not included intendedbehaviors in its model?
>===================
> So I took you seriously and drafted a reply.
I think the issue about intended behaviors is my misspeaking about intendedresults. I posted about this on 940630.
> And what about Ford, and my questions to you about whether the peopleyou say are doing the same thing as PCT do any modeling? That's not an idlequestion. If they don't do modeling, and if they don't explicitly say theybelieve people control their own perceptions, then there is no direct evidencethat they are doing "the same thing" as PCT. I had hoped to learn _your_ ideasabout that. If you think they are saying and doing the same things as PCT, thenwhat are _your_ criteria, what evidence do you see, for deciding that is so?I'm not challenging you or attacking you; I'm a simple inquiring mind and Iwant to know. :-)
I don't know yet. I told you what he said about behavior and perceptions.Somehow Bill P. saw PCT light in the description Ford made about behavior(although I cannot find that post).
>> on modeling and psychology:
>> "Useless and dangerous" -that is a limb. I must get back to it, watch out.
> I'm watching! I can't wait! Don't disappoint me this time! ;-)
> Later, Tom
My child awaits. Priorities. A teaser before I go. The paper I mentioned inmy introductory post has come back as revise and resubmit. Naturally many ofthe reviewers did not like the PCT flavor of it. I am trying to formulatequestions for those of you on the net to help address the issues. But thiswill take awhile. I am trying to be thoughtful :-)
Later Jeff
Date: Thu Jul 07, 1994 5:04 pm PST
Subj: CONTROL
From Tom Bourbon [940707.1729]
Just to tweak up the interest level a bit, one reason I say most claims toknowledge by psychology are dangerous is that innocent people might be temptedto believe those claims --some might even go so far as to appeal to psychological "knowledge" when theyattempt to establish new social policies, laws, and the like.
Later, Tom
Date: Fri Jul 08, 1994 8:13 am PST
Subj: Tenure, Publication
[From Rick Marken (940708.0820)] Jeff Vancouver (940707)
> Remember, I am trying to get tenure.
PCT (REAL PCT, not the Carver, Scheier, Hyland etc variety) and tenure donot mix. It's true that Tom Bourbon and I were able to get tenure while doingPCT, but we had to masquerade as "real" psychologists in order to do this. And,besides, we are _amazing_. Do not try this at home ;-)
> The paper I mentioned in my introductory post has come back as reviseand resubmit. Naturally many of the reviewers did not like the PCT flavor ofit. I am trying to formulate questions for those of you on the net to helpaddress the issues.
This could get pretty confusing for you; having papers rejected by journalsbecause they have a PCT flavor and then getting lashed by loonies (like me) onthe net who say your papers don't have nearly enough PCT flavor.
I suggest that you not assume that PCT is the reason why reviewers arerejecting your papers. If you are persistent (as I was) you will get some ofyour papers published in some journals. The fact of the matter is that journalsare very selective --they have to reject at least 80% of the papers submitted. The REAL criterionfor acceptance and rejection is probably political, more than anything else.One of my advisors in grad school could publish nearly anything he wrote (allof which was pure horseshit) because he was 1) a known entity and 2) goodbuddies with the editors of all the big journals.
I think PCT papers are often rejected for what are basically politicalreasons; but some are surely rejected because they really could be better. It'svery hard to tell what's going on. The fact that the review process is sosubjective means that political clout will count for a great deal of whatactually gets published. Since PCT people have zero political clout, its verytough to get papers published, regardless of the merits of those papers.
The only thing that works is persistence; keep trying. Getting published islike an e. coli walk up a chemical gradient; you don't know what to do to getup the gradient (to publication city) so you, basically, make random changes inthe paper and see if it gets looked on more favorably on the next submission.If you persist, you MAY eventually get published; if you give up, of course,there is no chance.
I'm not sure that what one does to get published is really random; one doestry to take the reviewers comments into account when revising. But doing so isno guarantee that the paper will be looked on more favorably. I don't know whatit was about some of my papers that got them published right away while otherstook (literally) years or were never published. It's really a crap shoot. Oneimportant consideration, of course, is the journal to which you submit; thereshould be a strong match between the contents of the paper and what the readersof the journal care about --at least at that moment. I had a lot of trouble with the "Degrees of freedom"paper, for example, until I sent it to Psychological Science; apparently theyjust happened to be interested in "motor control" and "new ideas" at the time,and it got in pretty easily. The paper on "Random walk chemotaxis", on theother hand, took years (and MANY different journals) until it was finallyhidden away in the never-readBehavioral Neuroscience. The problem with that paper (I think) was that we(Bill P. and I) were too overt in pointing out how the experiment and modellingresults were inconsistent with any notion of "reinforcement". Saying that thereis no such thing as "reinforcement" is just not "politically correct" inpsychology circles; we got the paper published, finally, by billing it as astudy of how control systems can navigate without steering (and hide the"reinforcement" conclusion in an aside in the middle of the discussion).
So don't blame non-acceptanceof your papers on PCT (even if there might be some truth to it). It is possibleto publish real, live PCT papers; it's not easy and it requires persistencebut, unless you are working with a politically powerful ally, this is true fornon-PCTpapers too.
Best Rick
Date: Mon Jul 11, 1994 12:38 pm PST
Subj: PCT as paradigm
[Jeff Vancouver 940711]
Okay, the rubber has hit the road and I am seeking your input. As Imentioned the paper I referred to in my introductory post has been invited forrevision and resubmission. Basically, the paper argues that my field(organizational behavior) needs a paradigm and that living systems theoryshould be that paradigm. Living systems theory (LST) is very PCT-like(this post is about some of the differences as I see them). In fact, many ofthe reviewers think LST is simply PCT with a different name. Given theprominence of Powers in my paper, this assessment is not far off base. However,I am advocating some things I do not see PCTers advocating, so I assume you allappreciate my use of another name.
1. I have focused a great deal of attention to the feedforward process. AsI understand it and explain in my paper, feedforward is the process ofanticipating discrepancies via memory of effector and perceptual signals. Ibelieve this is in line with PCT. However, I go on to say that plans are madeand even choices to engage are made based on those anticipated discrepancies.That the anticipated discrepancies are used to assess potential environmentaldisturbances and head them off. This is consistent with Ashby, if not PCT. Iguess my question is: why this is _not_ consistent with PCT? Let me anticipatethe possible answers:
a) anticipated _disturbances_ are not possible, regardless of how imperfectthey may be (this last phrase counters the argument that the anticipateddisturbances are merely conjectures on the part of the individual, becauseactually knowing the disturbances is beyond the system);
b) we have not gotten around to modeling it yet because it is too difficultto model or other things have taken priority;
c) anticipated disturbances and their effects on the system have been modelin PCT, see (cite).
2. I think the main counter-argumentto the PCT way of doing science (i.e., modelling), is that social interaction,meaning, and other higher-levelprocesses, do not lend themselves to precise, quantitative equations. This iswhat I mean in b) above. Also, this is the intrigue of neural nets and fuzzylogic for many. They are quantitative, but not precise. I too lend toward thoserepresentations as possibly necessary when modeling the higher-levelprocesses. Where does PCT stand on the issue of precision in it'smodels?
3. The distinction between the perceptual hierarchy and an individualsconcept of that hierarchy seems to be a critical issue that separates PCT frommost of psychology. That is, most of psychology concerns itself with conceptslike self-conceptand beliefs, but not with the actual system concepts, principles, etc. thatdrive behavior. Some do, but many do not. If I understand the psychologist'sposition, self-conceptsand beliefs are available to the conscious. If I understand PCT, referencesignals are not directly available, but perceptions are. What I am not sureabout is whether the perceptions that are available to the conscious are beforeor after filtering through the input function and do those perceptions from themodel that the psychologist study?
When I warned to I was going to make this post, Marken [940708.0820]replied with a comical post. Attempting to understand the behavior ofreviewers, editors, etc. he vacillated between random and political, with alittle writing style and content thrown in for good measure. (Marken, can yousimulate this oscillating behavior -a control system trying to predict the behavior of other controlsystems).
Ironically, I agree, that all the factors are involved (weights change toprotect the innocent). But one specific thing I want to highlight. Markensaid:
> This could get pretty confusing for you; having papers rejected byjournals because they have a PCT flavor and then getting lashed by loonies(like me) on the net who say yours papers don't have nearly enough PCTflavor.
Building communication links between schools of thought that have not beenable to communicate well before is exactly the niche I am carving for myself.It is not an easy task, but if it was, there would be no need.
Later Jeff
Date: Mon Jul 11, 1994 4:56 pm PST
Subj: Re: PCT as paradigm
From Tom Bourbon [940711.1655] >[Jeff Vancouver 940711]
> Okay, the rubber has hit the road and I am seeking your input. As Imentioned the paper I referred to in my introductory post has been invited forrevision and resubmission. . . . However, I am advocating some things I donot see PCTers advocating, so I assume you all appreciate my use of anothername.
You assumed right. :-)
You do have an interesting job in store, trying to satisfy thosereviewers.
> 1. I have focused a great deal of attention to the feedforwardprocess. As I understand it and explain in my paper, feedforward is theprocess of anticipating discrepancies via memory of effector and perceptualsignals. I believe this is in line with PCT. However, I go on to say thatplans are made and even choices to engage are made based on those anticipateddiscrepancies. That the anticipated discrepancies are used to assess potentialenvironmental disturbances and head them off. This is consistent with Ashby, ifnot PCT. I guess my question is: why this is _not_ consistent with PCT? Letme anticipate the possible answers:
> a) anticipated _disturbances_ are not possible, regardless of howimperfect they may be (this last phrase counters the argument that theanticipated disturbances are merely conjectures on the part of the individual,because actually knowing the disturbances is beyond the system);
Do you mean _anticipation of disturbance_ is not possible? I think that'swhat you intended and will reply as though that's the case. If I'm wrong,disregard everything I say, which you might do anyway. ;-)
To _know_ disturbances in advance of their occurrence is not possible. Toanticipate (imagine) them in advance of their occurrence is possible, but theanticipating occurs in the present and that is where all of the actions to wardthem off also occur --in the present. If that is the case, then actions I take _now_ to create theperceptions of preparedness that I intend to experience _now_ can be modeled aspart of a present-timeprocess of negative feedback control: no future event is involved, only present-timeimagination and intention and perception and action. Carrying an umbrella whenI leave the house after hearing a weather report that predicts rain is present-timeperceptual control, not feedforward. The fact that it is not now raining seemsto lead some people into thinking that my actions are directed forward into thefuture, when they are in fact happening right now.
Also, in hierarchical PCT systems, higher levels have longer time constantsthan lower ones. An observer who notices the actions of lower-levelloops sometimes "sees" those loops "taking action in advance of environmentalevents:" a person leans forward before taking a step forward --a clear case of feedforward, is it not? Or before taking a long trip a personplans an itinerary, decides what private "stuff" to pack and carry, and makesarrangements for the care and feeding of pets, plants and other dependentcreatures --a clear case of feedforward, is it not?
In either case, I believe the answer is, "It is not." What is easilyoverlooked in either case is the "bigger picture" of what the person iscontrolling. The body does not "lean" independently of the person "taking astep." Leaning and stepping are not two discrete, isolated events; one thinghas happened and an observer has treated it as though it were two (or more)things.
Part of the problem for an external observer who watches the behavior of ahierarchical, high-gain,negative-feedbackcontrol system is that activity at the lower end of the hierarchy of perceptualcontrol occurs on a time scale that is easy to see in a glance; when thingslook different to us in successive glances, we easily see different "things"happening, then give them different labels, then explain them by differentmechanisms. When actions happen close in time to what the external observeridentifies as the purpose of the actions, the observer often says they involvefeedback control; when actions happen in advance of what the observer says isthe purpose of the actions, at least some observers say they involvefeedforward. All the while, hierarchical, high-gain,negative-feedbackcontrol is probably lurking in the background, ready to confuse the innocentobserver.
> b) we have not gotten around to modeling it yet because it is toodifficult to model or other things have taken priority;
See my reply to a).
> c) anticipated disturbances and their effects on the system have beenmodel in PCT, see (cite).
See my reply to a).
> 2. I think the main counter-argumentto the PCT way of doing science (i.e., modelling), is that social interaction,meaning, and other higher-levelprocesses, do not lend themselves to precise, quantitative equations.
You have identified a frequent comment from reviewers and editors. Tocounter them, I think you (all of us who try to spread the news about PCTscience) need to redirect their attention to the phenomenon of control. Forexample, if you can show (empirically, not theoretically or in a model) that aparticular social interaction includes controlled variables and that theactions of each social actor affect variables controlled by other actors, thenyou establish the fact of control in a social setting. Once you establish thefact --the phenomenon --of control, the nature of the game changes, or at least it should. Now anyonewho wishes to explain the observed social interaction must demonstrate that, atleast in principle, the suggested explanation can explain the phenomenon ofcontrol. Any "explanation" that cannot in principle explain control _should be_dismissed. (All of this is easier to say than to do --if you discover a way to make our interactions with reviewers and editors workthe way they _ should_ work, share the news immediately!)
> This is what I mean in b) above. Also, this is the intrigue of neuralnets and fuzzy logic for many. They are quantitative, but not precise. I toolend toward those representations as possibly necessary when modeling the higher-levelprocesses.
This is where some of us part company with you, not necessarily because youare wrong and we are right, but because we don't want to allow ourselves anyway out. We want to see just how far we can get using nothing other than thePCT model. Our activity is driven by our belief that we see evidence of controlin phenomena where others think they need to talk about "higher-levelprocesses." Our way of working has a lot in common with that of scientists whorefuse to allow "the hand of God" as part of their explanations of nature; thechallenge they set for themselves is to see how far they can get without"giving up" and invoking principles or powers from outside their scientificmodel.
> Where does PCT stand on the issue of precision in it's models?
Precision is our guiding p-star.
Oops! Time to run. We are finishing the plans for our daughter's wedding onSaturday and I'm almost late for a fitting! I'll try to get back to theremainder of your post tomorrow.
Later, Tom
Date: Mon Jul 11, 1994 8:49 pm PST
Subj: Re: feedforward: planning perceptions
[From Bill Powers (940711.2115 MDT)] Jeff Vancouver (940711)
> As I understand it and explain in my paper, feedforward is the processof anticipating discrepancies via memory of effector and perceptualsignals.
But if you anticipate discrepancies, all you have done up to that point isto extrapolate from the present and perceive a calculated future. What happensnext? A perceived discrepancy, whether it be in present time, calculated, orimagined, does not tell you what action you will have to take to correct it.When the time comes to act, you will still have to deal with the world as it isat that instant.
What you're overlooking is that ALL behavior, at ALL levels, is control ofperception. You can't plan actions; you can only plan the perceivableconsequences of actions.
You can plan to stop at the newsstand on the way home to buy a paper.Superficially, that might seem like planning actions: stopping at thenewsstand; buying a paper. But the actions that lead you to perceive yourselfas being at the newsstand are not predictable. You may park your car in frontof it, if there's a space empty, or five spaces away in either direction, or inthe next block. You may have to walk if your wife just drove off with the car.Wherever you start, you will have to walk toward the door of the newsstand fromthe exact spot in which you find yourself and not from where you vaguelyimagined you would be, past the people standing around and going in and out,not on the empty stage you imagined. You must actually carry out every lastdetail of every muscle contraction that is required to get you through the doorand up to the counter. When you "buy a paper", you can't plan where theproprietor will be standing, or whether the newspapers are sold out, or whetherthe proprietor has change for a ten, or which hand the proprietor will reachwith to receive your money, or where the newspaper will be --on the counter, or handed to you. You can't plan the actions needed to buy apaper well enough even to end up inside the newsstand and not crashing throughits window.
The best you can do is form a very incomplete picture of the generalsituation you hope to experience; being somewhere at the newsstand, walkingaway with a newspaper held in one hand or the other, then being back home withit. You don't plan the means of achieving any of these perceptual goals. Youcan't. The world is too unpredictable. All you can do is plan _goals_, andleave it up to your control systems to bring them about in real-timeperception, dealing with the world as it actually is, in all its detail.
Best, Bill P.
Date: Mon Jul 11, 1994 9:34 pm PST
Subj: Replies to Jeff
[From Rick Marken (940711.2200)] Jeff Vancouver (940711) --
> 1. I have focused a great deal of attention to the feedforwardprocess.
> I guess my question is: why this is _not_ consistent with PCT? Let meanticipate the possible answers:
> a) anticipated _disturbances_ are not possible,
You got it right off the bat --though it would be more correct to say "the actions that will compensate fordisturbances cannot (and need not) be anticipated". I see Tom B. and Bill P.explain this rather nicely.
> 2...Where does PCT stand on the issue of precision in it'smodels?
Precise is nice. But the main point of PCT is that behavior is CONTROL.Since conventional behavioral science data provides no precise evidence of thevariables people control, PCT does not apply to this data of conventionalbehavioral science.
> 3. The distinction between the perceptual hierarchy and an individualsconcept of that hierarchy seems to be a critical issue that separates PCT frommost of psychology.
I think this is really irrelevant. The critical issue that separates PCTfrom most (all?) psychology is the issue of control. Conventional psychology isabout the control of behavior; PCT is about the behavior called "control".Conventional psychology tries to determine the variables that controlorganisms; PCT tries to determine the variables that organisms control.Conventional psychology and PCT are not talking AGAINST one another; they aretalking PAST one another.
> When I warned to I was going to make this post, Marken [940708.0820]replied with a comical post.
Thanks for thinking it was funny, but I was actually trying to give serioussuggestions about how to get published. What was so funny about it? I wasreally trying to help --and encourage you to be persistent in your efforts to publish. I rooting foryou to get tenure; once you're safe and secure THEN you can go ahead and do PCTright.
I said:
> This could get pretty confusing for you; having papers rejected byjournals because they have a PCT flavor and then getting lashed by loonies(like me) on the net who say yours papers don't have nearly enough PCTflavor.
Jeff says:
> Building communication links between schools of thought that have notbeen able to communicate well before is exactly the niche I am carving formyself. It is not an easy task, but if it was, there would be no need.
I'm glad that you want to build communication links but I also want to besure that you're putting the right message over the wires. Carver and Scheierand their ilk have gotten conventional psychologists to listen to the messageof PCT by providing the wrong message; this is called lying. I'm happy tocommunicate with other schools of thought; but if they will only listen whenyou tell them that PCT is something other than what it is then whatcommunication has there been?
Bill Powers (for one) has been communicating to the other schools ofthought quite clearly for many years; it's seems to me that the other schoolsof thought have not shown any serious interest in the message of PCT --and for good reason; it would mean the end of psychology as they know it, avery unpleasant experience for people who have built careers on psychology aswe know it. Revolutions are not fun --just ask Galileo. Should Galileo have placed the earth at the center of themodel solar system just because it would have gone down better with the "otherschools of thought" of the day. I don't think so.
Best Rick
Date: Tue Jul 19, 1994 12:02 pm PST
Subj: free will, feedforward, misc.
[From Jeff Vancouver 940718]
Well, getting to my office once a week certainly puts me behind the times.I can only skim the posts. Several posts where RE: Replies to Paul and Jeff,but were actually only to Paul. Please be careful, the subject headings on oneway I reduce information overload. Incidently, Paul, glad to see a kindredspirit on the net. Our kindred is not in our professions (I am a psychologist),but in our message (seek similarities with others, not differences).
Ironically, to seek that goal I am still seeking to understand some of thedifferences. Tom [940711.1655] thanks for the post regarding feedforward, ithelped somewhat. But I am beginning to worry that Locke is right, or we arearguing semantics.
You said "Carrying an umbrella when I leave the house after hearing aweather report that predicts rain is present-timeperceptual control, not feedforward." Does it matter if the weather report wasgiven that morning or the night before? Surely you are not saying theperception of impending rain, compared to a reference signal would produce a"get umbrella" output. Instead, you are saying the weather report triggers amemory of the perception of walking in the rain which is compared with a desirenot to get wet to produce the output "get an umbrella when you are going out."The use of a memory store, an not real time perceptions is what I mean byfeedforward.
Bill P. [940711.2115] says "You can't plan actions; you can only plan theperceivable consequences of actions." and later "All you can do is plan_goals_, and leave it up to your control systems to bring them about in real-timeperception, dealing with the world as it actually is, in all itsdetail."
This is exactly what I mean by feedforward. We do not propagate theplanning too far down the hierarchy. Such detail would get us into the troubleof over specifying a situation that is too variable. In a private post fromCharles Tucker, he seemed to be saying that Bill P. has a precise definitionfor feedforward that 1) requires the propagation and 2) is therefore notrelevant. He suggests using the term planning. How does this sound?
P.S. That the plan are goals (reference signals), not actions is often,though not always missed by cognitive psychologists.
Back to Tom. You said you would address the rest of my questions afterfixing your tux, but I never saw a followup. Did I miss something?
Marken [940711.2200]
on the feedforward question you say: "the actions that will compensate fordisturbances cannot (and need not) be anticipated."
I say that we (complex systems) need not, cannot completely, but cangrossly anticipate disturbances. The fundamental issue is the role of consciousprocessing in humans. I am trying to say that complex systems use anticipatorymechanisms to more effectively maintain their essential (intrinsic) andcontrolled (perceptual) variables. Further, I am saying that consciousprocesses are heavily involved in that process. What role does consciousprocesses have in PCT?
As I reread some of Locke and other psychologists I see the issue of freewill raised as central to their problem with PCT. They often go too far -action is a function of conscious will. Do you take the other extreme -action is random? Locke seems to think you do. I think the answer is inbetween. Action is indeterminable, but related to the perceptions one iscontrolling. This moves the free will debate to the question of will overperceptions controlled. I have no idea where PCT stands here (I certainly havemy own ideas).
I certainly agree with you Marken [940711.2200] "Conventional psychologyand PCT ... are talking PAST one another." You say because they don't studycontrol. Humor me, I say they do sometimes study it, just not always exactlylike you. I say a difference is the emphasis on an individual's concept oftheir place in the world versus the actual hierarchy. The concept is a perverserendition particularly because it rarely uses a PCT image. But, psychologist,me among them, say that concept is relevant to how they act _and_ perceive. Ithink some in PCT think that as well. Again I ask, where does PCT stand on theself-concept?
By the way, the humor in your tenure post was the vacillation on the causesof paper acceptance. Reread it, given the right frame of mind you too may findit funny. I do appreciate the thought, though.
I did have a strong objection to your calling Carver & Scheier liars.The may have been mistaken, but not intentionally so (at least, I would givethem the benefit of the doubt). Using a word like "lying" is inappropriate. Idon't buy that you are just talking among yourselves. As Dag just noted, youhave a number of lurkers. This is a public net. Beside, _I_ expectcivility.
On precision:
Your faith in the TEST is great. Yet, even Runkel said it is hard todetermine controlled perceptions during reorganization. There may be otherways. PCTers need not pursue them, but can't they let others without saying itis useless?
On the E.coli model:
Thanks Dag, I got the simulations. I have run most. Unfortunately, theE.coli, where you "showed" reinforcement does not work confirms my straw modelargument. Reinforcement is operationalized as "use the last turn that resultedin denser foodstuffs" according to your documentation. Why not remember thelast vector. Why not remember the last set of vectors? I am not proficient inthis type of modeling, but your model does not convince you have givenreinforcement theory a fair shot. (I am not bring it up to argue forreinforcement theory, but to note the use straw models).
[Bill Leach 940707.23:28]
You say: "I don't think that anyone on the net has a problem with the ideathat people can control more than one operation at a time as long as it isphysically possible to do so. Though I don't see 'an issue' here between PCTand any other 'model'."
The issue is one of cognitive resources not physical. Some psychologicalmodels are concerned with the limitation of our cognitive processes/attentionalresources. This is why I want to understand the role of cognitive processes inPCT. BTW, the distinction between cognitive and conscious not always made inthese models, but I don't want to revisit that either, unless it is relevant tothe question at hand.
in conclusion:
I have printed a number of the posts between Paul and the PCTers. I willtry to read them before my next visit to my office. I like the model. It givesa good reference for discussion. For example, is D what psychologists try tostudy? Here's a juice one, if numerous individuals F4's produce Hs thatrepresents F in nearly identical ways and given that F3 is constant acrosspeople can I model (in Bandura's sense) someone's D to affect H, which I thinkis affecting F? The point of the question is that many of the components in themodel can be ignored without much cost to predictive ability. I am going outto shoot myself now :-)
Later Jeff
Date: Wed Jul 20, 1994 1:36 am PST
Subj: Jeff
[From Rick Marken (940719.1400)] Jeff Vancouver (940718)
> (seek similarities with others, not differences).
Along with Mary Powers (940717) I think it is a mistake (well-intentioned,but a mistake nevertheless) to assume that you can move others (and yourself)toward PCT gradually, by seeking similarities with what they (and you) alreadybelieve. As Mary says:
> There really are no baby steps to take between behavior as outcome,consequence or result, and behavior as the control of perception
Jeff says:
> I certainly agree with you Marken [940711.2200] "Conventionalpsychology and PCT ... are talking PAST one another." You say because theydon't study control. Humor me, I say they do sometimes study it, just notalways exactly like you.
Humor me back, Jeff. Give me one example of a conventional psychologicalstudy in which control is the object of study.
> I say a difference is the emphasis on an individual's concept of theirplace in the world versus the actual hierarchy.
This has nothing to do with control as it is defined in PCT. I wrote a longpost on the nature of control; perhaps we can start from there. Once you haveread and understood that post, please give me an example of a conventionalpsychological study of control, even if that study is not exactly like one ofours.
> Again I ask, where does PCT stand on the self-concept?
I believe the term typically refers to a high level perceptual variable orset of variables. I think I control many different perceptions that have to dowith my "self" --simple perceptions like where I am at any particular time and complexperceptions like perceptions of myself following certain principles and systemconcepts. I think there is room for what has been called the "self concept" atall levels of the perceptual control hierarchy.
> I did have a strong objection to your calling Carver & Scheierliars.
Sorry. I re-readmy post where I said this and I see that I expressed myself poorly. I didn'tmean to say that C&S are liars. I meant that I would be a liar if Ipresented PCT as Carver and Scheier do. C&S are NOT liars because they haveno idea what control is or how PCT explains it; they are not being duplicitouswhen they make PCT seem compatible with what conventional psychologists arealready doing because they have no idea how PCT differs from what conventionalpsychologists are doing. I, however, do know what control is and how PCTexplains it. I also know how PCT differs from what conventional psychologistsare doing (so do Tom B., Bill P. and Mary P.). Therefore, if I made PCT seemcompatible with what conventional psychologists are doing, I would be a liar.I'm not ;-)
> Your faith in the TEST is great. Yet, even Runkel said it is hard todetermine controlled perceptions during reorganization. There may be otherways. PCTers need not pursue them, but can't they let others without sayingit is useless?
You'll have to expand on this one a bit. What is "faith in the Test"? TheTest is based on the definition of a controlled variable; it works --amazingly well (ever do the "Mindreading" demo?). Where's the faith part? Andwhat does reorganization have to do with it; if you are reorganizing --and not in control of any particular variable --then there IS NO variable under control and the Test will reveal that fact. Thephrase "it is hard to determine controlled perceptions during reorganization"makes no sense; controlled perceptions are controlled perceptions. It doesn'tmatter at all whether they are controlled during reorganization or during athunderstorm in Houston; if they are there, the Test will pick them up; noquestion. So I'm sure that Phil Runkel never said anything like that becausePhil makes a lot of sense.
> Unfortunately, the E.coli, where you "showed" reinforcement does notwork confirms my straw model argument. Reinforcement is operationalized as"use the last turn that resulted in denser foodstuffs" according to yourdocumentation. Why not remember the last vector. Why not remember the lastset of vectors?
Why not try that model? I'm REALLY tired of hearing this stuff about "strawmen" theories and then getting verbal descriptions of how it could "really" bedone --and never being shown that it CAN really be done that way. Talk about faith!How stupid do you think Bill Powers and I are, anyway? We tried every model wecould think of that was consistent with reinforcement principles as weunderstand them; none worked. Knowing that reinforcement theorists would saythat we only tried "straw men" versions of their theory, we asked --begged, pleaded with --reinforcement theorists to show us how to do it RIGHT; we asked to see how ANYversion of their theory could produce the operant behavior in the E. colistudy. All we get from reinforcement theorists is the same line you deliver inthe next sentence:
> your model does not convince you have given reinforcement theory afair shot.
What would convince you? What model would be fair? If we come up with themodel, it's a straw man (as long as it doesn't work, and none of our "strawmen", so far, have worked); but the reinforcement theorists won't show us howto do it "correctly". They just say (like you) that reinforcement theory CANexplain our results. Pretty cute. Apparently, reinforcement theory cannot berejected. It's proponents do not feel like it's necessary to do anything moreabout our data than be unconvinced and say that we are not fair. Well,excuuuuuuse me for daring to challenge the wisdom of reinforcementtheory.
> (I am not bring it up t argue for reinforcement theory, but to notethe use straw models).
What are you arguing for then, Jeff, the unrejectability of a theory? Yousay you are not arguing FOR reinforcement theory but you are also saying (notshowing, SAYING) that the theory CAN explain the operant behavior in the E.coli situation, where the consequences of actions are random. I think you'remissing a bet, here, Jeff. If I were you, I'd be arguing FOR reinforcementtheory; then I couldn't possibly be wrong.
Best Rick
Date: Wed Jul 20, 1994 9:12 am PST
Subj: Reinforcement theory vs E. coli
[From Bill Powers (940720.0745 MDT)] Jeff Vancouver (940718)
> Unfortunately, the E.coli, where you "showed" reinforcement does notwork confirms my straw model argument. Reinforcement is operationalized as"use the last turn that resulted in denser foodstuffs" according to yourdocumentation. Why not remember the last vector. Why not remember the lastset of vectors?
Because none of these interpretations results in a working model, either.We also tried "previous time rate of change of concentration" and "average rateof change of concentration over past n episodes" and "difference in averageconcentration between episodes before and after a tumble." The basic problem isthis: the next direction that will result from a tumble is always selected atrandom. So no matter what criterion you use as the basis for tumbling, it can'tbe used to select a more favorable direction of the next tumble. There is nocontrol over direction, no way to predict what the next direction willbe.
The point of this experiment was not to show that "reinforcement doesn'twork." It was to show that the organism can move in the right direction evenwhen there can be no basis for differentially reinforcing a "correct response"on the basis of past consequences of responses. Thus the explanation for thebehavior we see has to be something other than reinforcement theory. The paperoffered a PCT model that _does_ work, the model demonstrating its owncorrectness by producing behavior very much like that of E. coli. In thismodel, E. coli senses the time rate of change of concentration of theattractant (due to its swimming speed and direction in the chemical gradient),and varies the delay before the next tumble according to whether the _current_sensed rate of change is above or below a reference setting. The result is thatthe organism tumbles sooner when swimming the wrong way and delays the nexttumble while swimming the right way. Previous consequences of tumbles don'tenter into the model at all.
There is no consequence of the _previous_ direction of swimming that hasany systematic relationship to the _next_ direction of swimming, becausetumbles are truly random. Since reinforcement theory depends on _past_consequences of behavior to shape _future_ behavior, all attempts to producinga working reinforcement model run into the snag that the tumbles result in newdirections unrelated to the old direction. When all behaviors are randomlyrewarded and punished, reinforcement theory must predict either random behavioror superstitious behavior, neither of which will result in the observedbehavior of E. coli.
I should mention that the model is based on direct observation of thebehavior of E. coli tethered in a flow of liquid in which the concentration ofperfused attractants could be varied experimentally. The spacing of tumbleepisodes was found in this way to be proportional to the time rate of change ofconcentration relative to some particular rate of change (which we took asevidence of a reference setting). The relevant work is reported in Koshland, D.(1980); _Bacterial Chemotaxis as a Model Behavioral System_, New York: RavenPress.
The referees who rejected the paper could not reconcile the fact that thetumbles resulted in random new directions of swimming with the fact that E.coli found its way very efficiently up the gradient. Most of them simplyrefused to believe that the new directions were random (Koshland cited actualmeasurements showing a uniform and random distribution in space). Thealternative explanations offered by the referees all depended on NON-randomnessof tumbling, on the existence of discriminative stimuli which they made up outof thin air (and wouldn't have helped anyway), or on geometric arguments thatsimply ignored the fact of randomness. Their faith was being sorely tested, butthey kept to it in the usual way: by changing the facts or making up facts.They rejected our paper because according to their beliefs neither E. coli norour working model could possibly have behaved they way they did. In fact, theirreactions were a clear-cutproof that reinforcement theory could not handle this phenomenon (although thatdidn't help with getting the paper published).
The other factor that all the referees shared was that their arguments werepurely verbal; not one of them offered a refutation in the form of a workingmodel that actually behaved as they claimed such a model could behave. Not oneof them found any flaw in the construction of our working model. Not one ofthem commented on the striking similarity between the plots of the behavior ofour model and the behavior of E. coli. I don't think that any of them actuallyunderstood the concept of a simulation as a way of demonstrating thepredictions of a theory. Like most psychologists, they simply assumed that aworking model built to fit their explanations would actually work as theyclaimed it would. In this case, it would not.
Best, Bill P.
Date: Thu Jul 21, 1994 3:25 pm PST
Subj: Reinforcement theory and free will
[from Jeff Vancouver 940721]
First I want to apologize for my inability to read all the posts on thenet. I just skimmed it for references to my posts. I found two [Marken940719.1400 and Powers 940720.0745]. To which I respond below. Meanwhile I amcontinuing to develop a strategy to respond to my reviewers. Sometimes thistakes my questions to different levels, as you will see below.
[Marken 940719.1400]
Study of control by Psychologists:
Campion, M. A., & Lord, R. G. (1982). A control systemsconceptualization of the goal-settingand changing process. Organizational Behavior and Human Processes, 30, 265-287.
Kernan, M. C., & Lord, R. G. (1990). Effects of valence, expectancies,and goal-performancediscrepancies in single and multiple goal environments. Journal of AppliedPsychology. 75, 194-203.
Hollenbeck, J.R. (1989). Control theory and the perception of workenvironments: The effects of focus of attention on affective and behavioralreactions to work. Organizational Behavior and Human Decision Processes, 43,406-430.
___________________
Can you give me the reference for your "long post on the nature ofcontrol?" Most of your posts seem to fit this criteria :-)
By the way, I was reading your Degrees of Freedom paper. You cite neuralnetworks as promising for modeling perceptions. Have you changed your mind ordid I misunderstand something you said?
_________________________________
On reinforcement [both Marken & Powers]:
I want to take this to a higher level, but I need to respond to Marken& Powers. You are correct that I cannot give a counter model. I do nothow to model! I sympathize with your inability to get your detractors to workwith you on such a model. That is the most one can ask. Your efforts areimpressive. I no long think that you "straw maned" reinforcement theory.Instead, I think your description of E.coli _as you model it_ can beinterpreted as reinforcement theory. I should reread Powers response to thebehaviorists in earlier writings. But let me quote Bill P's post(940720.0745):
"In this model, E. coli senses the time rate of change of concentration ofthe attractant..., and varies the delay before the next tumble according towhether the _current_ sensed rate of change is above of below a referencesetting."
Forgive me, but cannot that be restated "E. coli senses the time rate ofchange of concentration [stimuli] of the attractant [reinforcer]..., and variesthe delay [a response] before the next tumble according to whether the_current_ sensed rate of change is above of below a reference setting."
That is, the stimuli "causes" a behavior, where the behavior is length ofdelay!
If this argument, or some form of it, has been rebutted in published work,just cite it -no need to repeat here. Unlike Paul, I have amassed most of it over the years.Unfortunately, it is beginning to fade from my mind.
________________
Here is the lasted bottom line:
The problem of "reinforcement theory" in your models is important (to me)because of Locke's claim PCT is neobehaviorism. If I understand it, the usualresponse is that the model includes internal [the real crux of control] as wellas external variables (e.g., behavior). Since behaviorism did not acknowledgeany of these internal variables, PCT is not behavioristic. This always workedfor me.
But as I try to reconstruct my own counter arguments (the editor wants meto take the debate further), I am beginning to see a point that I cannot getaround. If the values of reference signals (higher-orderoutput signals) are arrived at randomly, which is the central process forreorganization, then your model does seem -on one level -difficult to distinguish from neobehaviorism (by which I think they mean S-O-R). Stimulus (S) leads to response (R) with the organism's (O) perception as amediating variable and discrepancy (of perception from reference signal) andthe ultimate reinforcer. (Ironically, Bandura was one of those who popularizedS-O-R).
Adding the O takes a lot of wind out of the PCT "is not neobehaviorism"argument I had used in my mind. Maybe I have been too generous in thinkingneobehaviorism includes O. If it does, than I can accept that, argue that PCTis neobehaviorism, so what? It still serves functions goal theory and socialcognitive theory do not. But I suspect S-O-Ris not acceptable to PCT either. Plus the argument takes us to interesting newdimensions.
It is Marken's insistence that behavior is out-of-controlthat seems to support the neobehaviorism argument. As I understand it,controlling a perception variable occurs in two ways. Once the appropriatelower-orderreference signals have been discovered, the loop simple sends those signals.Although this might result in different behaviors (due to differentcircumstance each lower-orderloop must deal with to meet their sent reference signals), the flavor of theresponse to a stimuli (which was translated to a perception that sends thefocal loop's perception off its reference signal), it the same. That is, theresponse is reoccurring.
I anticipate red flags going off here. PCTers will say, "but the behavioris different because the circumstances are different. Only that special casewhere the circumstances are exactly the same, or close enough for yourstatistic analysis, will a reinforcement model work. And what is the _flavor_crap?"
This _flavor_ crap is the other way Locke might mean neobehaviorism.Regardless of Locke or Bandura, however, I can describe a concept called"flavor of a response" which means set lower-orderreference signals the same way. The result is the R in S-O-Ris now flavor of response. It is completely internal because lower-orderreference signals are internal. (Perhaps Locke and Bandura would not acceptthis meaning for R and I can stop here. But this argument has been made byBill P. and I want to take it further. When one takes it further, it gets veryinteresting.)
If one can accept the new definition of R, then I can move to the next wayof controlling --that is, before a flavor of a response is developed. This is thereorganization process. During this process varying reference signals are sentto various lower order loops in a random, but localized fashion. As theserandom changes begin to reduce the discrepancy, the set of reference signalsfor the lower-orderloops begins to be defined. That is, the probability that a lower-orderreference signal will have a certain value increases as the discrepancy isreduced (reinforced). Recognize behaviorism?
Now, will a simple adjustment counter the argument? If the set ofreference signals is developed discontinuously (either it eliminates thediscrepancy or it does not. If not, try completely different configuration),than the phrase cannot be probabilistic (it becomes: a set of lower-orderreference signals will be adapted if the discrepancy is made zero). But thisis not much of a change and becomes problematic for meeting complexperceptions. Also, the concept of "localized" reorganization would need to beexamined.
No, the real problem is the method of reference signal selection. Let usnot consider hardwired, which some consider, but has not been modeled.Hardwired will not work for controlling most complex perceptions anyway, solets only concern ourselves with the only other process considered by PCTadvocates -random. This is the only central difference between PCT and the Locke andBandura self-regulationmodels where the Locke & Bandura arguments make some sense to me. (Anotherdifference is that they talk of controlling behavior, which is clearly wrong ontheir part, so we need not discuss it further with regards to Locke &Bandura).
Here is where we get really philosophical. Locke & Bandura are arguingthat reference signal selection is our source of free will. We consciouslychoose our goals (reference signals). I think they would concede that thisdoes not happen all the time, but at least some of the time. They claimconscious choice translates to free will, because it is not completelypredictable or determinable. I don't buy the free will argument for a second(although lack of predictability and determinability is much easier to accept).The free will argument requires that humans are fundamentally different fromthe rest of nature (Sappington, 1990?, Psych Bull). A conclusion I cannotaccept. (Ironically, Sappington uses Bill P. and Bandura to argue for freewill, but it seems Sappington was assuming that reference signal selection wasultimately non-random.)
What I do accept is that conscious processes are often involved inselecting the (set of) reference signals. I am a soft determinist, which meansconsciousness enters the process of behavior selection (which strictly speakingis better said reference signal selection), but that conscious choices aredetermined by some other factors. This is what I like about Bandura's model,the other factors are abstractly described (e.g., self efficacy). Of course,as Bill P. pointed out (1991, Amer Psych), the specifications are somewhatproblematic if one were to try to model them. I agree. What I don't know ifthese _types_ of factors are considered in PCT? Specifically, Bill P. says"beliefs about one's actual effectiveness in achieving a given goal [Bandura'sself-efficacy]"is a perception (1991). By that do you mean self-efficacyis just another controlled variable? If is it a controlled variable than 1)what is "F" in your model and 2) what is F2? If F is an internal variablerelated to the focal goal and if F2 affects references signals related to theloop (goal/task) under consideration then we have just resolved a majorconflict among the models.
One more problem and I will have resolved two of my biggest conflicts withPCT. If a function type for F2 includes using an external address for areference signal, then outside influences are available for constructing ahierarchical control system. Is it possible?
The messy models in psychology begin to describe many of these influences.Together with the structure of PCT and the rigor in your methodology, I thinkpsychology can make great strides.
Postscript: I think one can persuade others that PCT has something tooffer. The exact function is debatable, but I have moved closer and closer toPCT as I have come to understand more and more of it. Bill P. claimed mostpeople take 2 years. Ed Ford and Dag report making converts slowly. Thisdata, anecdotal though it is, indicate the process is not completelydiscontinuous, although certain leaps my occur on the way. I suspect thatLocke and Bandura will only be convinced by PCT when they perceive rewards arecontingent upon accepting it, which will probably not happen in what remains oftheir careers. But I will first seek to convince myself, then the reviewersand editor. If I am not convinced, I will take what I am convinced about andmake clear the distinctions between PCT and my view. I would appreciate BillP.'s sanction on the manuscript to assure our respective views are properlyrepresented.
A parallel question is can I convince PCTers that psychology has somethingto offer PCT? The argument is parallel. Some are already convinced. Otherswill never be. But, unfortunately, in my system, that loop has little gain.PCTers won't grant me tenure.
Later Jeff
Date: Thu Jul 21, 1994 4:47 pm PST
Subj: out on a limb, part 2
[From Jeff Vancouver 940721.1808]
I hesitate to post this given that my previous post is my main concern, butI forgot to ask Marken where "the Blind men and the elephant" waspublished.
But while here, I wanted to rebute the psychology is useless and dangerousnotion. In my field we develop tests of cognitive ability and other predictorsof job performance. Focusing on job samples, which rarely have adverse impact(where scores differ depending on race or sex), these tests can predict jobperformance at around .30 to .60, depending on the job mostly. The alternative,doing nothing, would correlate .00 with job performance, or the job interview(before we improved it) .11. The differences might seem trivial to those wholook for correlations in the upper 90s, but the difference can save a companyhundreds of thousands of dollars (we have data on this). From the individualsstand point it will improve the fit between their skills and their job, whichusually make the individual happier and more secure. Without the work ofpsychologists, organizations would be less productive and individuals wouldwander from job to job looking for a good match.
There is work in psychology that is dangerous and other work that isuseless.
My colleague next door does it (just kidding). But there is also useful andhelpful work. May we be the ones who decide which is which.
Later Jeff
Date: Thu Jul 21, 1994 11:17 pm PST
Subj: Re: out on a limb, part 2
<[Bill Leach 940722.00:26 EST(EDT)]
>[From Jeff Vancouver 940721.1808]
> But while here, I wanted to rebute the psychology is useless anddangerous notion. ...
> There is work in psychology that is dangerous and other work that isuseless. My colleague next door does it (just kidding). But there is alsouseful and helpful work.
Sounds like your work may indeed be useful. You work with data that doesnot explain (or try to explain) how people function but rather with data thatis statistical and correctly so. There is no doubt the occasional error whereyour data is not correct for a particular individual but like the "mortalitytables" your data "is the best we can do.
PCT is probably not at a point where is could be reliably or economicallyapplied to such a task (even if there were enough PCTers to try to doso).
The quality of your data would however benefit from a deep understanding ofPCT. A serious problem with a great deal of the data collected in thebehavioral sciences is that critical information needed to relate the data topeople based upon how people actually function is not taken.
-bill
Date: Thu Jul 21, 1994 11:30 pm PST
Subj: Re: Reinforcement theory and free will
<[Bill Leach 940721.23:14 EST(EDT)] >[Jeff Vancouver 940721]
I recognize, Jeff, that you must primarily pay attention to the writings ofthe PCT researchers but I would like to comment nonetheless:
> Forgive me, but cannot that be restated "E. coli senses the time rateof change of concentration [stimuli] of the attractant [reinforcer]..., andvaries the delay [a response] before the next tumble according to whether the_current_ sensed rate of change is above of below a reference setting."
> That is, the stimuli "causes" a behavior, where the behavior is lengthof delay!
This would be ok except that there is no know way to postulate a "sense oftime" for the E.coli. Indeed, even for humans, it seems that elapsed time isavailable only through indirect perceptions. In addition, it is probablypossible to prove that the trigger is a concentration issue and not a timeissue (though I don't know that such proof has been attempted).
> But as I try to reconstruct my own counter arguments (the editor wantsme to take the debate further), ... If the values of reference signals (higher-orderoutput signals) are arrived at randomly, which is the central process forreorganization, then your model does seem -on one level -difficult to distinguish from neobehaviorism. ...
I think that there are several different "opinions" concerningreorganization. One is that it is a "catastrophic" phenomenon that only "kicksin" when intrinsic errors exist that are either of a "large" magnitude, ormaybe exist for an extended period of time.
The second triggering condition also fosters the idea that reorganizationmight be a nearly continuous process. There is also the possibility thatreorganization is not only local but that the magnitude of its effect isrelated to the magnitude of the error.
As you are no doubt aware, observation of "learning" is strong evidence insupport of both the reorganization concept and its' "random" nature.
A serious problem that you face is that control theory itself is not wellunderstood in the first place. Even people with extensive control systemsexperience are often astonished at just how much power a closed loop controlsystem can exhibit. Indeed, the long time PCTers will, in truth, admit thatthey did not really realize the capability of just this one paradigm AND arestill discovering new abilities that such systems display.
> It is Marken's insistence that behavior is out-of-controlthat seems to support the neobehaviorism argument.
"Out-of-control"might be a poor choice of words (in some sense). Behavior is not controlled iscorrect. The sense of this is that behavior is driven by the presence of anerror between perception and reference. The behavior that results is a functionof physical laws including the physical capability of the person and thephysics of the portion of the environment being acted upon. The "person" is"causing" the behavior in that all behavioral is a direct result of somethingthat the person does but what is important is that the something is controlperception where control has the same meaning as it does in the field ofcontrol theory.
A loose definition of the term control allows one to say that we controlour own behavior but such a statement is using an entirely different meaning ofthe term control from the precise meaning used in control theory andPCT.
> lower-orderreference signals the same way. The result is the R in S-O-Ris now flavor of response. It is completely internal because lower-orderreference signals are internal.
This is fine except that it does not explain control action. A stimuluswhich directly affects a controlled perception will look like a form of"stimulus-response"to someone unaware of the control system nature of living beings. The problemoccurs when an applied stimulus only partially affects (or does not affect atall) any controlled perception of the subject. Then you get a "66%" of thesubject, etc. type report with the researcher ignoring the other 34% percentthat are also very real subjects too.
> If one can accept the new definition of R, then I can move to the nextway of controlling --that is, before a flavor of a response is developed. This is thereorganization process. During this process varying reference signals are sentto various lower order loops in a random, but localized fashion. As theserandom changes begin to reduce the discrepancy, the set of reference signalsfor the lower-orderloops begins to be defined. That is, the probability that a lower-orderreference signal will have a certain value increases as the discrepancy isreduced (reinforced). Recognize behaviorism?
I have to admit that at first, this one really "hit me" as a problem butupon thinking for a bit...
What if there is no stimulus? What if there is a stimulus but the behaviorthat develops is eventually completely unrelated to the original stimulus? Thatchallenge seems to ignore purpose (or the idea that the references are set upby the organism and not the environment.
> Now, will a simple adjustment counter the argument? If the set ofreference signals is developed discontinuously (either it eliminates thediscrepancy or it does not. If not, try completely different configuration),than [then?] the phrase cannot be probabilistic (it becomes: a set of lower-orderreference signals will be adapted if the discrepancy is [not?] madezero).
I'm not sure I am following what you are trying to say there well enough toeven comment.
> But this is not much of a change and becomes problematic for meetingcomplex perceptions. Also, the concept of "localized" reorganization wouldneed to be examined.
"Becomes problematic" is an opinion that is not well supported by thefacts.
I think that you will find that the PCTer agree that "localized"reorganization needs to be examined.
> No, the real problem is the method of reference signal selection. Letus not consider hardwired, which some consider, but has not been modeled.Hardwired will not work for controlling most complex perceptions anyway, solets only concern ourselves with the only other process considered by PCTadvocates -random.
A difficulty here is again the assumption that "hardwired" will not work.There is no evidence that the highest level perceptions to which all others areultimately related may indeed NOT be "hardwired" (at least over the lifetime ofa specific instance of an organism). I believe that the problem withreorganization is that not enough is actually known about control systems ingeneral and random changes to effect control.
> The free will argument requires that humans are fundamentallydifferent from the rest of nature (Sappington, 1990?, Psych Bull).
> A conclusion I cannot accept. (Ironically, Sappington uses Bill P.and Bandura to argue for free will, but it seems Sappington was assuming thatreference signal selection was ultimately non-random.)
If by deterministic you mean that given sufficient knowledge about aspecific organism and the environment for that organism one could predictexactly what that organism will do, then that is likely true. However, when youcount the apparent randomness of physics at the particle interaction level suchpredictions are not likely. New evidence that electron orbital energy releasesare random has recently been obtained (as if there was a great deal of doubt tostart with).
> One more problem and I will have resolved two of my biggest conflictswith PCT. If a function type for F2 includes using an external address for areference signal, then outside influences are available for constructing ahierarchical control system. Is it possible?
If I understand what you are saying there, then the answer is no.
-bill
Date: Fri Jul 22, 1994 1:03 am PST
Subj: Re: free will, feedforward, misc.
From Tom Bourbon [940721.1803]
Still catching up after the wedding. >[Jeff Vancouver 940718]
> Ironically, to seek that goal I am still seeking to understand some ofthe differences. Tom [940711.1655] thanks for the post regarding feedforward,it helped somewhat. But I am beginning to worry that Locke is right, or we arearguing semantics.
I'm afraid I don't catch your allusion to Locke. I know his work, but /idon't catch what you mean when you say you "worry that Locke is right?" Rightabout what? Why does whatever he said worry you?
> You said "Carrying an umbrella when I leave the house after hearing aweather report that predicts rain is present-timeperceptual control, not feedforward." Does it matter if the weather report wasgiven that morning or the night before?
Not a bit, except that the accuracy of weather reports drops offexponentially with time! Carrying the umbrella is present-timebehavior, no matter how long ago I heard the weather report.
> Surely you are not saying the perception of impending rain, comparedto a reference signal would produce a "get umbrella" output.
I'm not? Maybe I am; maybe I'm not. I _am_ curious about why you wouldn'twant me to say that --why my saying that would perturb you. Can you let me know?
> Instead, you are saying the weather report triggers a memory of theperception of walking in the rain which is compared with a desire not to getwet to produce the output "get an umbrella when you are going out."
But aren't you saying here that, if there is a discrepancy when present-timeperceptions are compared with present-timereference signals, then there will be a change in actions? That was what I wassaying --it's all in the present. The past does not exist now; the future does not existnow; now exists now.
> The use of a memory store, an not real time perceptions is what I meanby feedforward.
Oops. Where did feedforward come from? "Memories" are present-time"things," aren't they? And so are perceptions, and comparisons, anddiscrepancies (error signals) and actions. Maybe it would help me understandyou better if you were to give me your definition of "feedforward."
> Bill P. [940711.2115] says "You can't plan actions; you can only planthe perceivable consequences of actions." and later "All you can do is plan_goals_, and leave it up to your control systems to bring them about in real-timeperception, dealing with the world as it actually is, in all itsdetail."
> This is exactly what I mean by feedforward.
This example doesn't help me very much, as a clue about your definition offeedforward. I see Bill talking about present-time"planning" of the expected results of unplanned actions. I plan to experiencesome dinner soon --I've had nothing to eat since breakfast, which was long ago. I "plan" toexperience myself logging off, locking the office, taking the elevator down,waiting for my wife to stop by on her way home, and so on. All of theseimaginings are right now; all of my actions (which I cannot plan in advance)will be happening "now;" all of my perceptions will be controlled "now."
> We do not propagate the planning too far down the hierarchy.
I don't think we propagate a plan for _actions_ at all.
> Such detail would get us into the trouble of over specifying asituation that is too variable.
The variability is precisely the problem. ;-)
> In a private post from Charles Tucker, he seemed to be saying thatBill P. has a precise definition for feedforward that 1) requires thepropagation and 2) is therefore not relevant. He suggests using the termplanning. How does this sound?
Planning perceptions, yes. Planning actions, no.
> P.S. That the plan are goals (reference signals), not actions isoften, though not always missed by cognitive psychologists.
I'll drink to that! Or maybe I'll eat a few bites of dinner tothat.
> Back to Tom. You said you would address the rest of my questionsafter fixing your tux, but I never saw a followup. Did I misssomething?
I'm still looking for them. I'm not yet caught up on the past week ofmail.
> Marken [940711.2200]
> on the feedforward question you say: "the actions that will compensatefor disturbances cannot (and need not) be anticipated."
> I say that we (complex systems) need not, cannot completely, but cangrossly anticipate disturbances.
Jeff, _why_ do you say that? Can you give some examples we could model,quantitatively? How would we test for whether a person is doing what you sayhere?
> As I reread some of Locke and other psychologists I see the issue offree will raised as central to their problem with PCT. They often go too far -action is a function of conscious will. Do you take the other extreme -action is random?
Did you ever see anyone (PCT modelers) say action is random? (You mighthave --I'm just asking.) If so, what were the conditions under which they said actionswould be random --or at least _appear to be_ random, to an observer?
> Locke seems to think you do.
Can you quote Locke on that? I have a small sampling of his writing and I'dlike to see if I have that quote, or if I can locate it for my file of mistakenideas about PCT.
> I think the answer is in between. Action is indeterminable, butrelated to the perceptions one is controlling.
You are onto an important theme here. Indeterminable from whoseperspective? That is the crux of the matter.
> I certainly agree with you Marken [940711.2200] "Conventionalpsychology and PCT ... are talking PAST one another." You say because theydon't study control. Humor me, I say they do sometimes study it, just notalways exactly like you.
Again, all you need to do to convince us on this point is show us, chapterand verse, where they themselves say they study the phenomenon in whichindividuals control their own perceptions and, incidentally, control variablesin their environments. I think Rick is saying there isn't evidence thatpsychologists study that phenomenon.
> I say a difference is the emphasis on an individual's concept of theirplace in the world versus the actual hierarchy. The concept is a perverserendition particularly because it rarely uses a PCT image. But, psychologist,me among them, say that concept is relevant to how they act _and_ perceive. Ithink some in PCT think that as well. Again I ask, where does PCT stand on theself-concept?
Rick and Mary have already answered you on this point. I concur.
Time for dinner! Now, what was that plan? First twitch the extensorsattached to that bone over there ... . ;-))
Later, Tom
Date: Fri, 22 Jul 1994 11:08:58 -0600
Subj: Misc replies RE PCT
[From Bill Powers (940722.0800 MDT)]
Jeff Vancouver (940721) --
In studies of control by psychologists, the main missing ingredient is anunderstanding of what happens when outputs affect inputs at the same time thatinputs are affecting outputs. You are perfectly correct in seeing that acontrol system is just an S-Rsystem with a reference signal that introduces a bias. In fact, you couldprobably find S-Rstudies in which the reference signal was recognized, as the "effective zero"of the stimulus. Way back when, I was convinced that this was the wedge thatwould get control theory into psychology; in fact, it turned out to be a"wedge" more in the modern usage of a sticking place.
There are two reasons for this wedge. One is that psychology is heavilybiased toward seeing stimuli and responses as discrete events: first thestimulus goes "ping" and then the response goes "pop." I think that one reasonfor doing this (and designing experiments this way) was precisely to keepresponses from interfering with the administration of stimuli. If you could getthe stimulus over with quickly enough, the response couldn't modify it beforeyou were through manipulating it. That meant that the stimulus could beconsidered an independent variable and the response a dependent variable, asrequired by the statistical analyses (and the theories) that were used. Also,for those who did recognize the closed loop, this meant that they could treatthe loop as a sequence of events: S-R-S-R-and so on.
The second reason is that psychologists knew nothing about control theoryand therefore didn't understand that closed loops of causality would haveemergent properties that were not obvious from the sequential event-orientedviewpoint. As rumors from engineering began to spread into psychology, in the1950s, psychologists picked up a few ideas, such as the fact that controlsystems tended to be slow and unstable, and would run away if they were toosensitive. This smattering of ignorance convinced them that feedback phenomenacouldn't be very important in behavior, because organisms could act quickly andstably, and didn't show runaway behavior.
PCT is based on the actual properties of sensitive closed-loopsystems. Most psychological attempts to analyze goal-seekingsystem are based on incorrect rules of thumb, or completely misunderstandingsof how such systems work.
-------------------------------------
> I think your description of E.coli _as you model it_ can beinterpreted as reinforcement theory.
> ... "E. coli senses the time rate of change of concentration [stimuli]of the attractant [reinforcer]..., and varies the delay [a response] before thenext tumble according to whether the _current_ sensed rate of change is aboveor below a reference setting."
The problem with saying it this way is that it isn't reinforcement theory.What is happening is that the _current_ time rate of change is affecting the_current_ delay, thus varying the time of the _next_ tumble. According toSkinner, reinforcements are consequences of _past_ behaviors, and when theyoccur, they tend to increase the probability that the same behavior will occuragain in the future.
When the next tumble occurs, the result is a new time-rate-of-changeof attractant. That is the consequence of the _current_ amount of delay beinggenerated --as you say, "the response". Unfortunately, the next time-rate-of-changeis completely unrelated to the current one. Whether the current response be ashort delay or a long one, the chances that after the next tumble theconcentration of attractant will be increasing are equal to the chances that itwill be decreasing. Long and short delays are not differentially rewarded bythe results of the next tumble.
There is in fact no strategy based on past rewarding outcomes that can beused as a way of selecting a better future response. Each segment of E. coli'stravel is an entity unto itself; whatever results in systematic progress up thegradient must happen during each segment independently of all others before orafter.
I don't object to the words of reinforcement theory. If you want to callinput variables "stimuli" and "reinforcers", and outputs "responses," that's OKwith me. What I object to is the _organization_ of reinforcement theory, whichclaims that past rewards determine future behaviors: that behavior iscontrolled (meaning determined) by its consequences, to put the thesis exactlyin Skinner's words.
--------------------------------------
> The problem of "reinforcement theory" in your models is important (tome) because of Locke's claim PCT is neobehaviorism.
How about quoting us some quotes? I can see that this might give youreditor some problems. From what I've seen of Locke, he attributescharacteristics to control theory that just aren't true; maybe we can find somespecific statement to refute. One of Locke's complaints about control theory isthat it isn't based on any experimental data! I guess Locke carries a lot ofweight in your field.
> Maybe I have been too generous in thinking neobehaviorism includes O.If it does, than I can accept that, argue that PCT is neobehaviorism, so what?It still serves functions goal theory and social cognitive theory donot.
Attaboy. It doesn't matter what you call it. Except, of course, to youreditor.
-------------------------------------
> As I understand it, controlling a perception variable occurs in twoways. Once the appropriate lower-orderreference signals have been discovered, the loop simply sends thosesignals.
Actually, the higher level of control is continuously monitoring its owncontrolled variable, and continuously varying the reference signal of the lowersystem as a means of controlling the higher variable. The reference signals arenot just sent as blind outputs to the lower system. The result of sending themis always being reflected in the state of the higher perception, so control iscontinuous. This is true at all levels. The higher system can't just decide ona good output and stick with it, because there are always many influencestending to alter its perception. It has to vary the output that becomes thelower-levelreference signal according to the current state of error in the higher system. ------------------------------------>No, the real problem is the method of reference signal selection. >Letus not consider hardwired, which some consider, but has not >been modeled.Hardwired will not work for controlling most >complex perceptions anyway, solets only concern ourselves with >the only other process considered by PCTadvocates -random.
Careful, here. In PCT, control systems are what you might call soft-wired.That is, reorganization can slowly and randomly change the wiring, but on thetime-scaleof ordinary behavior the system is, for all practical purposes, hard-wired.We handle control of complex perceptions by dividing them into orthogonal sets,each of which varies in magnitude only, in one dimension only. We see aseparate control system for each possible dimension of variation (that is undercontrol). What is normally treated as a single complex perception then becomesa collection of perceptions, each representing one dimension of the complexperception.
This seems very wasteful, but actually you end up with extremely simplecontrol systems for any one dimension, simple enough that they could beimplemented with a few neurons. The other way of approaching it looks morecompact, but requires an enormous number of computations, so I don't think youend up saving any neural capacity.
Reorganization involves a random component, but the operation of thecontrol systems in the main hierarchy doesn't.
------------------------------------
> Locke & Bandura are arguing that reference signal selection is oursource of free will. We consciously choose our goals (referencesignals).
Well, that's a feeble step toward a hierarchical model, isn't it? Afterthey have pushed this idea of internal goal-selectionfor a few years, maybe one of them will wonder _why_ a particular goal ischosen --what does it do for the person to freely chose that goal instead of another?Then they might realize that there is another level of goals and perceptions,which are achieved by selecting the lower level of goals.
Free will isn't as simple a concept as Locke and Bandura make it out to be.I'm perfectly free to move my hands any way I please, until I am using them tosteer a car. Then I must vary my hand position as required by physical laws tokeep the car where I have freely chosen it to be on the road. If I wish toavoid running into a culvert, on the other hand, then I must choose positionsfor the car on the road that do not intersect the culvert, and if I choose toleave one road and turn onto another, I have to choose positions for the carthat will achieve that goal --a very limited set of positions, in comparison to what free choice mightallow.
So every goal is chosen as a means of achieving a higher goal, and as soonas that is recognized, the lower-levelgoal can no longer be chosen freely. It has to be chosen so that achieving itwill achieve the higher goal. Is there a highest-levelgoal? Where does it come from? Questions to be answered experimentally, not byphilosophy.
--------------------------------------
> Specifically, Bill P. says "beliefs about one's actual effectivenessin achieving a given goal [Bandura's self-efficacy]"is a perception (1991). By that do you mean self-efficacyis just another controlled variable? If is it a controlled variable than 1)what is "F" in your model ...
F is the set of lower-levelperceptions on which you base your perception of self-efficacy.They might consist of such perceptions as one's own degree of skill, other'sopinions about one's effectiveness, memories of successes and failures, and soforth. You can control the perception of self-efficacytoward a high level by developing more lower-levelskills, by persuading others to admire you, or whatever means will alter theperceptions on which you based the perception of self-efficacy.You can also control for a low-degreeof self-efficacy:you can make clumsy mistakes, antagonize others, refuse to learn any skills,and so forth (depending on what perceptions add up to self-efficacy for you). Why would you choose a low level of perceived self-efficacy? Perhaps to avoid being given responsibilities --after all, nobody gives the hard important work to a klutz. ------------------------------------>If a function type for F2 includes using an external address for >areference signal, then outside influences are available for >constructing ahierarchical control system. Is it possible?
Uh-uh.Reference signals in the hierarchy are strictly the outputs of higher-levelsystems. There is no way the environment can directly set any reference signalinside the person. You can set up circumstances in which a person might wellchoose to set a given reference signal in a given way, but that is always up tothe person, not the environment. You can be told "Ride this horse to lose therace or I will kill your daughter." After weighing your goals regarding losingraces and losing your daughter, you might decide to ride the race to lose. Butyou might also weigh other goals, and kill the person who is threatening you onthe spot. Or you might decide that you can always make another daughter, butyou have only one reputation as Dick Francis to lose. Higher considerationsalways come into play, and they always come into play inside the personactually doing the controlling. All the outside world can do is presentcircumstances and connections. It can't force a person to choose any settingfor any reference signal.
-------------------------------------------------------
> I would appreciate Bill P.'s sanction on the manuscript to assure ourrespective views are properly represented.
At your service.
--------------------------------------------------
> In my field we develop tests of cognitive ability and other predictorsof job performance. Focusing on job samples, which rarely have adverse impact(where scores differ depending on race or sex), these tests can predict jobperformance at around .30 to .60, depending on the job mostly. Thealternative, doing nothing, would correlate .00 with job performance, or thejob interview (before we improved it) .11. The differences might seem trivialto those who look for correlations in the upper 90s, but the difference cansave a company hundreds of thousands of dollars (we have data on this).
Now this gets ticklish, because if I don't just say that your field iswonderful and useful, you will get all prickly and start thinking up defenses,and we won't get anywhere. So try to remain calm.
Why do you suppose it is that companies and other large organizations arewilling to put out serious money to get potential employees tested, while jobapplicants fear these tests and avoid them wherever possible? The answer liesin the correlations you cite above. Obviously, if a company uses screeningtests, even with as low a correlation as 0.3, it will in the long run avoidhiring quite so many unsuitable people. That can save it a lot of money andgrief.
But now consider the testing procedure from the standpoint of the persontaking the test. If the correlations are as high as 0.6, this means that the"coefficient of uselessness" (sqrt(1-r^2),more traditionally called the coefficient of alienation) is 0.8. If I rememberGary Cziko's explanation correctly, this means that you would do 80% as well inpredicting performance simply by taking the mean of the group performance. Ifthe correlation is as low as 0.3, the coefficient is 0.95. In any event, thismeans that many people who test low actually belong in the high group and viceversa.
From the individual's point of view, this means that there is a very highprobability of being misjudged --either being accepted for a job at which you will fail, or being rejected fromemployment which you could easily handle. Where the large company can even outthe statistics by using the test on hundreds of people per year, the individualapplicant gets only one chance every five or ten years. Furthermore, the payoffmatrix for the company is weighted oppositely to that for the individualapplicant. If the company makes an occasional mistake, it loses little, andoccasionally gets even more than it bargained for. The individual, however, isfaced with the alternative between making a good living and a poor living (ornone at all). A misjudgment is far more serious for the individual than for thecompany. The usual justification for using these tests is that "over the longrun" they are quite reliable. But for the individual, there is no "longrun."
I realize that I am taking on a multi-billion-dollarindustry here, and have about as much chance of reforming it as the proverbialsnowflake of surviving in hell. But am I not speaking the truth? The harm doneby psychological testing in industry is not to its beneficiaries, the companieswho commission such testing. It is to those who are tested.
You say " From the individuals stand point it will improve the fit betweentheir skills and their job, which usually make the individual happier and moresecure." But that is a myth. Over the long run, what you say is true from thecompany's standpoint --but it is false for a very large proportion of the individuals, particularly ifyou include all the individuals tested, not just those selected.
Best to all, Bill P.
Date: Fri, 22 Jul 1994 13:01:10 CST
Subj: Re: out on a limb, part 2
From Tom Bourbon [940722.1105]
>[From Jeff Vancouver 940721.1808]
> I hesitate to post this given that my previous post is my mainconcern, but I forgot to ask Marken where "the Blind men and the elephant" waspublished.
It is in our ghetto journal, _Closed Loop_, 1993, vol. 3, no. 1 --the same issue that contains "Models and their worlds," by Bill Powers and me.Copies are available from Mary Powers and I have several extras.
> But while here, I wanted to rebute the psychology is useless anddangerous notion.
Jeff, I'll start my reply by going to the conclusion of your post.
> There is work in psychology that is dangerous and other work that isuseless. My colleague next door does it (just kidding). But there is alsouseful and helpful work. May we be the ones who decide which is which.
I've stopped using the label in most settings, but I am also one of "we;" Iam --gulp --a psychologist.
I am still a card-carryingmember of the American Psychological Association, the American PsychologicalSociety (a charter member), and the Psychonomic Society (than which there is nopurer group of experimentalists ;-).I taught psychology courses, undergrad and graduate, for over 28 years. I speakfrom _inside_ psychology.
What I am about to say is not "merely" a PCT issue, but an issue thatshould concern all behavioral (life, medical, cognitive, etc) scientists, nomatter their theoretical stripes.
> In my field we develop tests of cognitive ability and other predictorsof job performance. Focusing on job samples, which rarely have adverse impact(where scores differ depending on race or sex),
Does this mean scores that lead to wrong predictions, but do not differaccording to the race or sex of the test taker, rarely have "adverseimpact?"
> these tests can predict job performance at around .30 to .60,depending on the job mostly. The alternative, doing nothing, would correlate.00 with job performance, or the job interview (before we improved it).11.
In my reply, I am assuming that the scores you report are correlationcoefficients, calculated from scores on the tests compared with measures ofperformance on the job. If my assumption is wrong, never mind what I say next.:-)I know, Jeff, that you already know many of the things I am about to say, butthere is a diverse audience looking over our shoulders and I want to be surethe same vocabulary is available to everyone. And please don't think I amtrying to impugn your motives or your character.
A common interpretation of a correlation coefficient (r) is in terms of thepercentage of the variance it "accounts for" in the relationship between thetwo variables (for example, between test scores and measures of jobperformance). A common estimate of the variance accounted for is the value of r-squared.In your example, r-squaredranges from .3^2 = .09, 0r 9%, to .6^2 = .36 or 36%. Now it is certainly truethat either of those correlations and variances accounted for is greater thanzero. But how well do they work, and does the use of the tests really have noadverse impact on people? From whose perspective is the effectiveness oradverse impact of the tests decided?
Concerning the effectiveness of the tests, there are other informative waysto interpret the relationships "captured" by correlation coefficients. One ofthose ways is to calculate k, the "coefficient of alienation," which is acoefficient of inefficiency or failure of the correlation as a predictor. Thecalculation is simple: k = square root( 1 -r-squared).
In your example, for r = .3, k = .954; for r = .6, k = .800.
What do those numbers mean? First, if you did not know the correlationbetween test scores and job performance, but you did know the means of bothsets of scores, then given a particular person's score on the test, your bestestimate of the person's job performance would be the mean score on the scaleof job performance. This is always the relationship between predictor andpredicted scores, _if_ you do not know the correlation, or if the correlationis zero. (When r = 0, k = 1.00 --maximal alienation or "uselessness" as a predictor.) Any non-zerocorrelation should reduce the coefficient of alienation, indicating that thecorrelation improves your ability to predict performance from the test. But thegain in predictive ability is low, until the correlation coefficient is verylarge.
For example, your correlation of .3 leaves the chance of a failure in yourprediction of job performance 95.4% as great as it was when you did not havethe test. And for r = .6, the chance of failed predictions is 80% as great asbefore the test.
I grant, right up front, that even a 4.6% success rate is non-zeroand that it might appear to be of some use to an employer, but what about thepeople to whom the test is applied and whose lives are thereby affected? Therace, gender, age, height, and sock size of those individual persons areirrelevant; the fact is that many more people will be harmed by the applicationof such a test than will be helped, unless, of course, the people we aretalking about are the employer or people on the employer's "team."
> The differences might seem trivial to those who look for correlationsin the upper 90s,
For now, let's just say they are demonstrably small and they are very poor,as predictors.
> but the difference can save a company hundreds of thousands of dollars(we have data on this).
I grant you that.
> From the individuals stand point it will improve the fit between theirskills and their job, which usually make the individual happier and moresecure.
From the stand point of _which_ individuals? Certainly not from that of the95.4% to 80% of people who are misclassified by such a test. Themisclassifications will include both the very large numbers of people who aredenied a job and the ones who get the "wrong" job.
> Without the work of psychologists, organizations would be lessproductive and individuals would wander from job to job looking for a goodmatch.
I think both of those claims can be challenged, especially the one thatseems to portray humanity as lost and wandering in the wilderness, but for theintervention of psychologists. Most of the psychologists I have known seemedpretty badly lost, themselves. ;-)
I am _not_ saying the quest for efficiency and fairness is wrong. I _am_saying that the psychological "knowledge" brought to that quest is poor andthat its application is dangerous, if and when the application is toindividuals.
=======
><[Bill Leach 940722.00:26 EST(EDT)]
>>[From Jeff Vancouver 940721.1808]
>> But while here, I wanted to rebute the psychology is useless anddangerous notion. ...
>> There is work in psychology that is dangerous and other work thatis useless. My colleague next door does it (just kidding). But there is alsouseful and helpful work.
> Sounds like your work may indeed be useful. You work with data thatdoes not explain (or try to explain) how people function but rather with datathat is statistical and correctly so.
Ah, Bill, the tests _do_ allegedly say something about how people function;individuals are the ones who are given jobs, or denied jobs, on thepsychologist's assumption that a single person's test score says somethingabout how that particular person would function in a specific job. The test maybe _administered_ to people en masse, and scored in bulk on a computerizedsystem, but its application and effects come down one misclassified person at atime.
> There is no doubt the occasional error where your data is not correctfor a particular individual
Occasional error? From 80% tp 95.4% errors don't rank as "occasional" in mybook. As a psychologist, I am ashamed that we use such lousy and dangerouspredictors. But then, as a psychologist, I'm a pretty lousy representative ofthe field.
> but like the "mortality tables" your data "is the best we cando.
And "the best we can do" is not good enough, in my private opinion.
By the way, the mortality tables do represent descriptive statistics ofhigh quality --we have pretty good data, from large numbers of cases, on the proportions ofpeople in various age groups who die during a given period of time and onwhether those proportions are stable or changing. We have nothing near thatquality with regard to the validity of screening tests.
> PCT is probably not at a point where is could be reliably oreconomically applied to such a task (even if there were enough PCTers to try todo so).
As I said up top, this is not a PCT issue, but an issue of the adequacy ofdata and predictions in behavioral science. I'll crawl even farther out on mylimb: The predictive power of psychological assessment "instruments" will neverimprove very much, so long as the causal model behind psychological researchand test construction is lineal and so long as psychologists continue to mis-applystatistical procedures, making statements and predictions about _individuals_,when they have used statistical procedures that (when used properly) only allowyou to speak about groups.
Later, Tom
Date: Sat, 23 Jul 1994 08:56:00 -0600
Subj: More on disturbances
[From Bill Powers (940723.1650 MDT)]
Jeff Vancouver (940721.1808)--
> But while here, I wanted to rebute the psychology is useless anddangerous notion. ...
I am quite taken by your "creative typo." The word "rebute" combines thefunctions of "refute" and "rebuke"; it means, obviously, "to prove beyond doubtthat your argument is faulty and at the same time chastise you for havingpresented it."
By the way, to "deny" a statement is to say that it is wrong. To "refute" astatement is to _prove_ that it is wrong. . . . . .
Best, Bill P.
Date: Wed, 27 Jul 1994 16:28:10 CST
Subj: Re: limbs, papers, models
From Tom Bourbon [940802.1418]
>[from Jeff Vancouver 940728]
>Tom Bourbon [940725.1200]
> Briefly, 1) organizations are going to select (discriminate)regardless of psychologists providing them tests (they will do somethingbecause they must)
Fine, so let them do it. Just don't expect me to participate in themisapplication of poor psychological data in a manner that unjustifiably harmsthe people who are tested and discriminated against. People who do such thingsdo so because they intend to do so, not because they must. If psychologists aresatisfied to earn a fat fee by helping employers in that discriminative task,power to them. For me, I've taken the PCT poverty vow.
> 2) prior to providing those tests organizations tended to discriminateunfairly (the popular notion of discriminating) and poorly, that is,organizations used methods that predicted performance very poorly.
And now they discriminate fairly? Using "instruments" that are wrong infrom 80% to 95% of the cases? Sorry, but I don't buy into that. The tests harmmany more innocent people than they help.
> 4) now _some_ organization use method that predict performance muchbetter (particular when used together) and thus save the organizations largeamounts of money
Much better? In your original post on this thread, you said the correlationbetween interviews and job performance was about .1 (any data on that?) andthat the correlation between psychological tests and job performance rangesfrom .3 to .6 --correlations that yield the percentages of incorrect predictions I mentionedabove. Even I ;-)can see that the proportion of failed predictions went from .995 (99.5% of themwere wrong) when only the interview was used, to .80 or .954 with the tests. Ican see the difference, but I can also say, as a psychologist expressing apersonal opinion, the difference isn't something I would be proud of.
Much better? Could you describe your criteria for making thatstatement?
As for saving the organizations money, if they say so, I accept it. Afterall, we are talking about _their_ bottom lines.
You say that when tests are used in combination, the results are even moreimpressive. But if the tests are independent, as the tester would want them tobe, then the results are multiplicative. If you use two independent tests, eachof which correlates .3 with job performance, then each of them "explains" .09of the variance --when used alone. When you use them together, they do not explain .09 + .09 =.18; but .09 * .09 = .008. Use them together and you are worse off than witheither used alone, and neither was very good when used alone.
> 5) individuals who are not selected by these tests are often betteroff because they would have been fired eventually or not done well, which isusually frustrating and debilitating.
Hmm. That's _very_ interesting. Let me try to get this straight, becauseyou seem to be alluding to a breakthrough in predictions that is of majorproportions. At a correlation of .3, a test would misclassify 95.4% of thetakers with regard to their performance on the job for which they applied. Yet,you are saying many of the people were in fact _correctly_ identified andfurther that those who were rejected would indeed have done sufficiently poorlythey would have been fired. Can you tell us about how someone would decidewhether any particular person who was rejected would have > been one of thesure-firefired failures? We could probably make a fortune by applying your technique. ;-)
> 6) the general public is often better off (we do not want airplanepilots that cannot fly very well, which we might not be able to tell exceptunder adverse -or in this case -simulated adverse conditions).
Agreed, and I'm damned pleased the pilots who took me to the meeting andback were good at their profession. But you were talking about something else --tests that lead to wrong conclusions in from 80% to 95.4% of theiradministrations. Pilots aren't selected that way.
> 7) individuals can use the results of tests to clue them intodeficiency and competencies -and often do.
Sorry, I don't follow you here. Can you help me?
> bottom line: tests give use more information than no tests. We mustuse that information responsibly (and we have associations that attempt to seethat we do).
Once more, I do not deny that there is a difference between 99.5% errorsand 95.4% errors. Speaking for myself, I think the only way to use suchinformation responsibly is to warn the public and do all we can to eliminatethe present abuse of innocent test takers.
> But psychologists help develop the tests and the methods for using theinformation gained from them responsibly. (e.g., we have always advisedagainst using the MMPI for selection purposes -it was to designed to aid in diagnosis).
Yes, psychologists often do try to prevent applications of their testsoutside the settings for which they were designed. I respect (some of) thoseefforts. However, I'm afraid my concerns also extend to applications in theoriginal, intended settings. Poor correlations are poor correlations, no matterwhere they occur; abuses that arise from the application of poor correlationsare abuses, no matter where they occur.
> I just picked up Runkel's casting nets book. He acknowledges the usesof the method of frequencies. This is what I have described above.
I'm glad to hear you got Phil's book. I would recommend it to _everyone._However, I don't think the tests you described illustrate the method describedby Phil. In fact, I believe Phil would identify most uses of psychologicalassessment as _inappropriate_ applications of the method of relativefrequencies. When it is used properly, the method of relative frequencies tellsyou that certain proportions of people are found in certain categories. Asimportant as that result can be sometimes, that is all the method offrequencies tells you. It leaves you in a situation where you can make_absolutely_ _no_ statement about any particular individual. Any application ofgroup data (even of properly collected group data) to specific individuals isunjustified.
> The method of specimens (and PCT) is we our profession needs tobetter understand humans and thus construct better tests (instruments is thebetter word, but too long).
Yes! On this, I believe Phil Runkel would agree, as well. And the _only_way to design better tests for predicting what a given individual will do is tostudy people one at a time and, paradoxically, thereby learn something abouthow all of them "tick." Phil called that kind of research the method ofspecimens. PCT is an example of a science that studies individuals, asspecimens of the species, or more generally as specimens of life.
Let us know what else you think of Runkel's book.
>Tom Bourbon [940725.1633]
> See my address above for sending the models paper. Appreciateit.
Great. A copy of "Models and their worlds" will be in the mail tomorrow. Ifyou read it, that will make a total of five or six people in all the world. ;-)
> I am still waiting to hear your reply to the rest of my post. I donot, nor does Bandura or Locke, interpret the S-O-Rsymbol as requiring lineal causality (although I see why it is easilyinterpreted that way).
But it _is_ lineal, Jeff. It includes two assumed end points, withcausality moving from the beginning to the conclusion. It doesn't matter a whitthat they put something between the beginning and the conclusion --causality still works in one direction with two end points. The same can besaid of _every_ information processing "model" that speaks of Input-Processing-Output.Every such model is a variation on the same lineal theme --and that theme is inadequate as an explanation of the behavior of livingthings.
> Bandura spends much time in his recent work to the reciprocaldeterminism idea (cyclical causality), which I use frequently (but have aproblem with the looseness of words -given I know PCT)
Ah, but the fact you know PCT should make Bandura's cute little word gamesall the more unacceptable to you --well, I can't defend that kind of prescription for you, but it certainlyapplies to me. The phenomenon of control is not an example of "cyclicalcausality," as Bandura defines that term, but it is an example of a continuous,simultaneous relationship between an organism and some particular part(s) ofits environment. If Bandura knows the difference, then he would serve sciencebetter were he to speak clearly and draw the distinctions crisply. But Ibelieve there is ample evidence he does not know the difference; what hebelieves, he says.
It is one thing to believe there is something "reciprocal" about therelationship between person and environment; it is quite another to understandhow such a relationship can come about and persist. Up to now, from all I haveseen, Bandura hasn't a clue about how it can happen. In fact, Bandura has madea special point of rejecting, out of hand, both (1) descriptions of thephenomenon of control and (2) the PCT model. He is clueless.
> But Bandura and Locke's models are flow charts (Powers 940507.1420),not system diagrams. That is why they cannot model their theories (and why PCTis fundamentally better than their theories).
Agreed, on the first and final points, but not quite agreed on one in themiddle. They use flow charts by design --by intention --not out of necessity. They have no intention of modeling their ideas. Theyverify their ideas by assertion, by citations of data that are lousy but arestatistically significant, and by appeal to authority, not by demonstratingthat the ideas work. It is by their own design that they do not model theirtheories.
> However, there are practical application of their flow charts that PCTis not capable of making. Like, if performance is low, check self-efficacy,if it is low, try to increase it, performance often improves (which makesEVERYONE happier).
Again, Jeff, I believe I understand why someone, a psychologist forexample, would want to know about or talk about such things, but the constructsare just too inexact for me. They define "self-efficacy"operationally --in terms of test scores that correlate with --with what? And why do they accept correlations that, while statisticallysignificant, suffer error rates as high as those for the pre-employmentscreenings you described earlier? I have seen no evidence from them that "self-efficacy"exists, as they define it, much less that it can act to cause behavior.
They use poor data and untested theories as justifications for theirstatements about "big" topics. The fact that PCT modelers often refrain fromspeaking about many of those topics should not be taken as evidence that thosewho do speak, speak from a base of scientific knowledge.
> What I want to know is how does a belief like self-efficacyplug into PCT? (My previous post began to talk about that).
I think Rick gave a good answer to this question.
> One final question (for now).
Wow! You have really fired off a salvo of questions! I have been at thislonger than I should have been and must run to the lab for a while. I promiseto come back to the final questions. Note the plural --you didn't stop with just one! :-))
Later, Tom
Date: Tue, 2 Aug 1994 17:20:50 -0600
Subj: Re: free will (from Mary)
[from Mary Powers 9408.02]
>[Jeff Vancouver 9407.18]
> As I reread some of Locke and other psychologists I see the issue offree will raised as central to their problem with PCT. They often go too far -action is a function of conscious will. Do you take the other extreme -action is random? Locke seems to think you do. I think the answer is inbetween. Action is indeterminable, but related to the perceptions one iscontrolling. This moves the free will debate to the question of will overperceptions to be controlled.
I really don't know what Locke and other psychologists are saying aboutfree will, but if they insist that it is about producing actions and that it isalways conscious then they will indeed have a problem with PCT.
The PCT alternative to the conscious production of action is not to assertthat actions are random. Actions must be free to vary in order to controlperceptions in a changing environment (and we've discussed the fact that theenvironment is always changing because factors like muscle fatigue mean that noactions can ever be exactly the same). But freedom to vary does not meanrandom. Actions, although free to vary, are limited by the environmentalconstraints which one perceives -I always leave a room by the door because I haven't figured out how to walkthrough walls, nor would a door in the ceiling be much use to me. In any case,I can usually get out of the room, which is my intention, even if theenvironment prevents me from doing it in certain ways, and certain actions aremore likely to be used than others, like walking and twisting the doorhandle.
We leave rooms all the time without conscious intent. We may becomeconscious of the problem if the door is locked, or if there is a fire in thehall and leaving by the window seems like a better idea. We intend to be out ofthe room, and do what we must to have that perception.
As for choosing which perceptions are to be controlled: eventually as yougo up the hierarchy you come to very important perceptions like "the self"which Locke and others seem to believe are conscious. But we act to controlperceptions all the time which, if you trace the reference signals up, domaintain a desired state of the self without there ever being consciousnessthat that is the goal. I may be leaving the room to go to the bathroom, but Ido not have to be, and therefore am not, conscious of maintaining my self assomeone who does not wet her pants in the room where the conference ismeeting.
Another angle on free will as coming from a conscious, self level is theidea that consciousness reaches the highest level there is. I prefer the ideathat it does not: that concepts like creativity, intuition, and having a higherpower, are a few of the ways of referring to properties of the human organismwhich are not accessible to consciousness, and which are the source ofreference levels which, when maintained at certain values, are experienced asthe self. But they do not come from the self.
I think that Locke and others see the self as not only conscious, but alsofixed at the highest level they can think of, which seems to be the logic level -and of course as performing all these huge calculations all the time to produceoutputs. Thus they see choosing, or intending, or having free will, as beingconscious and as being expressed as calculating those outputs. This concept hasto be a product of where they get to when they introspect as to how theythemselves function -informed, of course, by a particular model of how the brain must work.
I believe that "having free will" is simply how it feels to control one'sperceptions. You don't feel you have it when the environmental constraints aretoo severe, and you don't feel you have it when you have reference signals thatconflict with one another -when you can't control one set of perceptions because, awaredly or not, youalso want the opposite thing, and you are stuck in a position of being unableto achieve either. But these are special cases.
Mary P.
Date: Wed, 3 Aug 1994 00:32:50 -0400
Subj: Re: limbs, papers, models
<[Bill Leach 940802.23:36 EST(EDT)] >Tom Bourbon[940802.1418]
Tom;
I really dislike taking stabs at you after your previous post to me but aperception of an error must not be ignored...
> want them to be, then the results are multiplicative. If you use twoindependent tests, each of which correlates .3 with job performance, then eachof them "explains" .09 of the variance --when used alone. When you use them together, they do not explain .09 + .09 =.18; but .09 * .09 = .008. Use them together and you are worse off than witheither used alone, and neither was very good when used alone.
Seems to me that this is a matter of how the "results" are viewed. If"employ" results from any one test are interpreted to mean "hire this person",the calculation is additive (assuming that the test really are'independent').
I'll leave it to Jeff to explain how the results of multiple tests arehandled.
-bill
Date: Wed, 3 Aug 1994 12:11:32 -0600
Subj: free will;
[From Bill Powers (940803.0930 MDT)]
Bill Leach (940802.23:53 EDT) --
(horning in)
> Mary, are you postulating that the 'highest level' referencesestablish in a deterministic way the goals of the immediate leveldown...
Higher level reference signals determine only the higher level perceptualsignals. The reference signals actually sent to lower systems depend on whatdisturbances are acting. To achieve a higher goal you must VARY lower goals,not set them to fixed states.
Suppose you have the freely-chosengoal of holding a steering wheel centered. This means that you must setreference signals for applied force to WHATEVER LEVEL IS NEEDED to cancel anyexternally-originatingforces that also affect the position of the steering wheel. If the externalforces vary, you must vary the goal-settingfor applied force in exactly the opposite way, provided you still want to keepthe steering wheel centered. So the environment determines what forces you willhave to produce if you want to achieve the goal of a certain steering wheelposition.
Now, suppose you decide (freely) that instead of wanting to keep thesteering wheel centered, you want to keep the car on the road. This goal nowrequires that you VARY the position of the steering wheel in any way requiredto counteract crosswinds, tilts in the road, and curves in the road. Now thephysical properties of the world determine from moment to moment the angle ofthe steering wheel that you must set as a subgoal in order to achieve thehigher goal of keeping the car on the road.
You can continue this: you must vary the position on the road in order toturn off onto another road; you must vary your choice of roads in order to getto your destination despite detour signs; you must vary your destination if youwant to close a deal with another person who has moved to a different town; youmust vary which person you close the deal with if you want to get the bestdeal; and so forth. Each goal that seems freely chosen at the moment becomesdetermined by external circumstances when you take the next higher goal intoconsideration.
Even at the highest level, whatever it may be, you must VARY that highestgoal if you want to remain alive in the manner that best suits a human being --which is determined by your intrinsic reference signals, which are inherited.The only thing you are truly free to do arbitrarily is to be human --and only the species, not you as an individual, can do that, or choose to ceasedoing that.
So we have to conclude that free will does not consist of freedom to choosegoals arbitrarily. Arbitrary choice of goals would disable part or all of thehierarchy of control above the level of the choice.
Where, then, could truly free will come into play? Obviously it must comeinto play where it will not disrupt any higher control systems (which willautomatically resist the disruption). This suggests that it will come into playwhere there are no higher control systems --in branches of the hierarchy that remain unfinished. But reorganization willautomatically come into play there, too, if intrinsic reference levels areviolated. So the range of free will is not completely infinite, given the goalof staying alive.
However, for every control problem there are multiple solutions which areequivalent in terms of their effects on intrinsic state. A person can resolvecertain problems either by becoming a mathematician or by becoming a truckdriver. Either way, the person will survive adequately to prevent intrinsicerror. So we can suspect that what is called free will comes into play under"don't-care"conditions. Where there are multiple routes to a successful organization, onecan choose any route freely, without constraint because of any knowableconsequences at higher or lower levels.
It is sometimes claimed that free will is evident in the fact that we canvolitionally produce any arbitrary action we please at any time, for no reason.This is tantamount to saying that anything can be reorganized. But that ideaconsiders the volitional act only at the microsecond of its execution. All actshave consequences, and we are not free to like or dislike any particularconsequence chosen at random. If a volitional act, an act of reorganization,has consequences that cause errors in existing control systems, or in thereorganizing system, there will soon be a need for another volitional act,compensating for or undoing the effects of the original one. In other words, ifwe do not like the consequences of a volitional act, the only way we cancorrect the problem is to nullify the volitional act: we have no choice, if wewant to go on experiencing what we like rather than what we dislike.
At bottom, what we like and what we dislike are built into us: pain hurts,pleasure feels good. What we experience as pain and pleasure are consequencesof actions of control systems which exist beyond our ability to be conscious ofthem: all we know is that certain consequences seem in themselves to have aquality that is given, beyond analysis. What we call feeling "good" or feeling"bad" is the criterion against which we judge all experiences. Solving adifficult set of equations feels good; failing feels bad. That is our built-innature speaking to us, through intrinsic error signals or diminutionthereof.
So where does this leave "free will?" I think it leaves these words in thesame class as "intelligence" or "traits" or "phlogiston." These are words thatsound or once sounded, and were used, as though they must have some profoundmeaning. This impression proved ultimately to be mostly an illusion: there wasno experience to go with the words.
---------------------------
Your objection to Tom's remarks is valid:
Tom: >When you use them together, they do not explain .09 + .09 =>.18; but .09 * .09 = .008.
Variances are like probable errors squared. They add.
What Tom is saying applies to the truth-valueof assertions. I suppose there is a way of converting from "percent varianceexplained" to "probability that the hypothesis is true." If you have twohypotheses, each with a probability of truth of 0.6, then the probability thatboth are true at the same time is 0.36. So if the hypotheses are each "probablytrue", any conclusion that is drawn from assuming they are both true is"probably false." This is what dooms a psychology built on low-probabilitystatistics to remaining a simple science, if it can be called a science atall.
Of course any hypothesis that explains only 9% of the variance is "probablyfalse" to begin with.
Best to all, Bill P.
Date: Thu, 4 Aug 1994 16:04:13 CST
Subj: Re: free will
Tom Bourbon [940804.1519]
>[From Bill Powers (940803.0930 MDT)]
. . .
>Bill Leach (940802.23:53 EDT) --
. . .
> Your objection to Tom's remarks is valid:
> Tom:
>> When you use them together, they do not explain .09 + .09 = .18;but .09 * .09 = .008.
> Variances are like probable errors squared. They add.
I was talking about proportions (or, with a slight change, percentages),not variances. The correlations originally described by Jeff Vancouver (betweenpre-employmenttests and job performance for the few people who were hired after theemployers' psychologists finished with them) ranged from r = .3 to r = .6. Fora Pearson-rcorrelation, the "proportion of variance explained or accounted for" is r-squared;the percentage of variance "explained" is (100 * r-squared).For r = .03, the proportion of variance explained is = 09. For r = .6,"proportion of variance explained" = .36. Neither test is very good, as isshown by their respective coefficients of alienation, k, where k = squareroot(1 -r-squared).For r = .3, k = .954; for r = .6, k = .80.
Pretty soon, this turns into number salad, but something devastatinglyimportant for traditional behavioral (medical, cognitive, etc) research lurksamong the olives and lettuce leaves. Scientists often use correlations tojustify their attempts to predict one kind of performance (on the job, forexample) from some measure of another kind of performance (scores on a paper-and-penciltest, for example). If you do not know the correlation between the two sets ofscores, or if the correlation is zero, the best you can do, given that a personmakes any possible score on the test, is predict the mean value from the scaleof performance on the job. When you do that, the proportion (percentage) oferrors in your predictions will be the greatest possible. Any non-zerocorrelation between the two sets of scores reduces the proportion (percentage)of errors in your prediction of job performance from test scores. For any givencorrelation between the two sets of scores, the proportion of reduction inerrors of prediction is given by (1 -k); the percentage reduction is (100 * (1 -k)). For r = 1.0, k = 0, and the proportion of reduction in errors is (1 -0) = 1.0; for r = 0, k = 1.0 and the proportion of reduction in errors is (1 -1) = 0.
For the correlations reported by Jeff, r = .03, k = .954, and proportion ofreduction in error of predictions = 1 -.954 = .006; for r = .6, k = .8, and proportion of reduction in error ofprediction = 1 -.8 = .2. Here, we are talking about error-ladenpredictions of who will do well, or poorly, on a job. We are talking aboutpsychologists using inadequate predictors, then defending those predictors as"good" and "beneficial to humanity." Nasty business, that. Harmful to morepeople than it helps.
Number salad. Oil and vinegar, anyone?
> What Tom is saying applies to the truth-valueof assertions. I suppose there is a way of converting from "percent varianceexplained" to "probability that the hypothesis is true." If you have twohypotheses, each with a probability of truth of 0.6, then the probability thatboth are true at the same time is 0.36. So if the hypotheses are each "probablytrue", any conclusion that is drawn from assuming they are both true is"probably false." This is what dooms a psychology built on low-probabilitystatistics to remaining a simple science, if it can be called a science atall.
Doom and gloom, and not a minute too soon.
> Of course any hypothesis that explains only 9% of the variance is"probably false" to begin with.
And that's the truth. ;-))
Later, Tom
------------------------------
Date: Thu, 4 Aug 1994 17:44:29 -0400
Subj: free will
[Jeff Vancouver 940804]
Mary, Bill P, & Bill L.
I think that covers it. My e-mailis down, so I am not fully functional here (as if I ever was).
What I am trying to understand with my free will discussion is these fuzzyconcepts that Bill P. used like "decide" and "want." If ultimately the organismis attempting to maintain its intrinsic signals, then how can an organismdecide to die? That is, decide to commit suicide?
This is just the pen-ultimateexample. For me, this is not a problem in PCT because, in fact, the intrinsicsignals refer to much more specific things than "life". Nonetheless, the ideathat the organism can override all the intrinsic signals that are going offduring a hunger strike is difficult to explain.
What I am looking for (P) is mechanisms that 1) allow beliefs to influencethe choice of a reference signal, and 2) mechanisms of ignoring or toning downerror signals (particularly from intrinsic and higher-orderloops). For the first mechanism, I need more than simply to view beliefs as akind of perception (Marken). That is, where in the hierarchy is a belief typeloop, and how does it affect output functions. For it seems that the decisionmechanism is at the output function. An error signal goes to an output functionthat has more than one set of reference signals available to it (e.g., "I couldhit this guy or just smile and pretend to ignore him"). This condition evokes athinking mode (switching output and input gates), the results of whichdetermine the choice and give us the experience of some level of self efficacyand valuing. Free will, if it exists, which I doubt, is at the choice point. Atleast the experience of free will (e.g., "I choose not to be the type that hitspeople, no matter how offensive they are.")
The all-elseequal/doesn't matter condition strikes me as very unlikely. There are too manyloops involved. Every choice one makes will adversely affect higher-orderperceptions, at least for an adult. Hence (2).
In terms of (2) -toning down an error signal, or preventing one from happening -may relate to the lag of the loops. Is this experienced as "I just did notthink of that at the time?" But what is the experience "I swallowed my prideand just walked away"?
I am trying to reconcile subjective experience, which influences Locke andBandura, with PCT. The subjective experience must be just as explainable asbehavior. It is just a matter of translating PCT to explain this experience(and the causal appearance of this experience on behavior).
I guess I will address the testing stuff later, except to thank Bill P. andL. for correcting Tom.
Later Jeff
Date: Fri, 5 Aug 1994 17:19:00 CST
Subj: Re: free will
Tom Bourbon [940805.1702]
>[Jeff Vancouver 940804]
. . .
> I guess I will address the testing stuff later, except to thank BillP. and L. for correcting Tom.
I'm eagerly watching for your next comments about the goodness andusefulness of psychological testing when it is used as a crutch to justifymajor decisions about the lives of people who are required to take thetests.
Jeff, you must be referring to the post in which Bill P. corrected what hemistakenly thought were statements by me about the additive or mulitplicativeproperties of variances. As I indicated in Tom Bourbon [940804.1519], I wastalking about the *proportion* *of* *variance* *"explained"*, not aboutvariances per se. The distance from proportions to probabilities is not asgreat as that from probabilities to variances.
Besides, the onus is still on you to defend the use, by psychologists, ofpre-employmenttests that correlate no higher than .3 or .6 with performance on the job. Aboveall else, I am curious about why you think testing as poor as that has anythingsignificant to add to the theory of behavior in PCT. In behavioral science,correlations that poor should be viewed as sources of *noise*, not as sourcesof *facts* to be explained. For correlations that low, there are no "facts" tobe explained.
Later, Tom
PS Copies of the papers you requested are in the mail to you.
Date: Mon, 8 Aug 1994 13:23:54 -0400
Subj: Re: useless Psych
[from Jeff Vancouver 940808] >Tom Bourbon [940805.1702]
> I'm eagerly watching for your next comments about the goodness andusefulness of psychological testing when it is used as a crutch to justifymajor decisions about the lives of people who are required to take thetests.
I am not sure people are using it the "justify" major decisions, but theydo help provide information for those decisions. This seems to be one of thedifferences in our debate. You describe the often inaccurate attributions layindividuals give to the meaning of the test scores. I am referring to carefuluse of the scores.
> Besides, the onus is still on you to defend the use, by psychologists,of pre-employmenttests that correlate no higher than .3 or .6 with performance on the job.Above all else, I am curious about why you think testing as poor as that hasanything significant to add to the theory of behavior in PCT. In behavioralscience, correlations that poor should be viewed as sources of *noise*, not assources of *facts* to be explained. For correlations that low, there are no"facts" to be explained.
I think I see our problem here. I am not suggesting these tests addanything significant to the theory of behavior. Quite the contrary, PCT canprobably add something to the construction and use of these instruments. Theinstruments, because they are attempting to predict results of behavior in thefuture, will always be 1) statistical and 2) low by testing specimenstandards.
Let me ask you, have you every hired anyone? Selected a roommate? Chosen anauto-repairshop? Voted for an politician? What information would you use/not use?
Later Jeff
Date: Mon, 8 Aug 1994 15:55:25 CST
Subj: Re: useless Psych
Tom Bourbon [940808.1544] >[Jeff Vancouver 940808]
>>Tom Bourbon [940805.1702]
>> I'm eagerly watching for your next comments about the goodness andusefulness of psychological testing when it is used as a crutch to justifymajor decisions about the lives of people who are required to take thetests.
> I am not sure people are using it the "justify" major decisions, butthey do help provide information for those decisions. This seems to be one ofthe differences in our debate. You describe the often inaccurate attributionslay individuals give to the meaning of the test scores. I am referring tocareful use of the scores.
Jeff, I don't really intend to come across as a "heavy" on this topic. Ido, however, have some pretty serious problems with the uses of psychological"assessment instruments" *within* the profession, where everyone assumes theuses are careful. Test scores that correlate .3 and .6 with real life do notjustify most of the conclusions reached by careful professionals, no matter howhigh their intentions might be.
. . .
> I think I see our problem here. I am not suggesting these tests addanything significant to the theory of behavior. Quite the contrary, PCT canprobably add something to the construction and use of these instruments. Theinstruments, because they are attempting to predict results of behavior in thefuture, will always be 1) statistical and 2) low by testing specimenstandards.
And it is precisely their very poor statistical nature that renders theminappropriate for reaching conclusions about individuals.
> Let me ask you, have you every hired anyone?
Yes.
> Selected a roommate?
Yes.
> Chosen an auto-repairshop?
Yes.
> Voted for an politician?
Yes.
> What information would you use/not use?
In every case you mentioned, the *last* thing I would think of using wouldbe a "standardized psychological instrument." The MMPI for selecting amechanic? The Wechsler Adult Intelligence Scale for selecting a senator? TheDraw-A-PersonTest for selecting a room mate?
Later, Tom
Date: Mon, 8 Aug 1994 17:19:00 -0400
From: Jeff Vancouver
Subj: Re: useless Psych
>Tom Bourbon [940808.1544]
>> Let me ask you, have you every hired anyone?
> Yes.
>> Selected a roommate?
> Yes.
>> Chosen an auto-repairshop?
> Yes.
>> Voted for an politician?
> Yes.
>> What information would you use/not use?
> In every case you mentioned, the *last* thing I would think of usingwould be a "standardized psychological instrument." The MMPI for selecting amechanic? The Wechsler Adult Intelligence Scale for selecting a senator? TheDraw-A-PersonTest for selecting a room mate?
What did you use? Past experience? Past behavior? Typing speed (if you werehiring a secretary), position of a stereo (for the roommate). My guess is thatyou used some of these pieces of information. The role of the industrialpsychologists is to see if one's belief about the relationship between thesepieces of information (particularly past behavior) actually relates to futurebehavior (or in your words, future results). We look for matches between needs(e.g., a stereo) and measures of haves (imprecise though they may be). We tryto get at the skills (e.g., reading) learned in experience (e.g., school)rather than relying on the simply having the experience.
I would never use the MMPI to select anyone for a job. I do not rememberthe Draw a Man test. I would like to look into the relationship between TheWeschler and senate performance -I suspect we would find a relationship there. I would be more likely to want toknow that than whether they have a family, which seems the criteria of choicethese days.
The point is you must make decisions on imperfect information. If one canprovide at least some information, which our net (Runkel) has shown mightpredict behavior, I might very much want it.
later Jeff
Date: Tue, 9 Aug 1994 17:06:47 CST
Subj: Re: useless Psych
Tom Bourbon [940809.1554]
Jeff Vancouver [940809?]:
> What did you use? Past experience? Past behavior? Typing speed (if youwere hiring a secretary), position of a stereo (for the roommate). My guess isthat you used some of these pieces of information. . . .
I don't think I used any "information" at all. (Maybe I did, and I justdidn't realize that was what I was doing ;-)None of the things you mentioned seem to fit my idea of what "pieces" of"information" would look like.
> The role of the industrial psychologists is to see if one's beliefabout the relationship between these pieces of information (particularly pastbehavior) actually relates to future behavior (or in your words, futureresults). We look for matches between needs (e.g., a stereo) and measures ofhaves (imprecise though they may be). We try to get at the skills (e.g.,reading) learned in experience (e.g., school) rather than relying on the simplyhaving the experience.
Jeff, I don't argue what industrial psychologists say they try to do. Thatis not my "gripe." You began this line of discussion by describing tests thatcorrelate .3 to .6 with performance on the job. Correlations that low don'tallow one to improve very much over simply "playing a hunch" or "trusting tointuition" when it comes to hiring a new employee. On the down side, relianceon tests as poor as those *will* lead to many incorrect decisions that will bedefended as correct, with all sorts of "scientific" justifications to supportthe defense. If there are tests that perform better than r = .3 to .6, theywill improve the picture a little, but not very much --not unless they correlate .98 or more with performance on the work place. Whenr = .98, % variance "explained" = .96, coefficient of alienation = .199, and %reduction in errors of prediction (compared to not knowing the correlation) =80.1. Those are simple facts of correlational relationships and are not mattersof my opinion, or of PCT the theory. . . .
> The point is you must make decisions on imperfect information. If onecan provide at least some information, which our net (Runkel) has shown mightpredict behavior, I might very much want it.
When Phil Runkel (Casting Nets and Testing Specimens, Praeger, 1990 --highly recommended to everyone) wrote about statistically-basedsurveys and experiments, he carefully showed that the results of suchprocedures *cannot* be used to say things about any particular individual.Period. No exceptions. End of subject.
I believe Phil would be saddened to learn that his work is cited in anattempt to justify applications of (poor) group statistics to the selection ofindividuals for employment.
Later, Tom
Date: Thu, 11 Aug 1994 15:59:53 -0400
Subj: hiatus
[Jeff Vancouver 940811]
I am leaving for a conference and vacation for 3 1/2 weeks, so I will notbe responding. I do not want anyone to think I am ignoring them. I havecomments for Bill L and Tom B. specifically, but no time before I leave (sorrya paper and presentation take precedent).
BTW, the presentation is on my application of PCT to Person-organizationfit. Many would probably hate it; some might like it. I will be at the academyof management conference in Dallas. The presentation is 8:00 am if anyone isinterested.
Marken, thanks for the Closed Loop. I sent two papers your way. There maybe more in September.
See you after labor day.
Jeff
Date: Wed, 14 Sep 1994 17:33:40 -0400
Subj: return of a fringe element
[Jeff Vancouver 940913]
Well, I am back from giving my paper and visiting family. Add a week ofprepping for classes and the beginning of the term folderol and it comes toabout a month off the net. That is a lot of catching up. I have been skimmingthe posts and have noticed a couple of things I should probably address.
First, the Ayn Rand thread was interesting. Upshot, she is very close toPCT in principle, but errors on some of the details. This is the basic argumentI have been making to the Locke et al camp.
Second, regarding the work of Lord and associates. One of Lord's majorerrors is in placing a decision mechanism (function) between the comparator andthe output function. It is an attempt on his part to include a DME withincontrol theory. I give him credit for trying. Choices are made, incorporatingthat into the control theory model is paramount on my list of things to do. Hetook a stab at it.
Third, the issue of the use of statistics (conventional behavioral sciencemethods) continues to bother many on the net. I had been advised to read Runkeland have gotten most of the way through it. There was one particular figurethat hit home to the argument. Runkel discussed a "drawing" program thatrequired the subject to make slight deviations with the mouse around a circleto make a cursor draw whatever is was the subject wanted. He noted that anexamination of the circle or the deviations about that circle told one verylittle (or nothing) about the desired drawing. He showed a picture of a squarewith a triangle in it and a circle. Both were jagged, but their shapes wereclear. He pointed out that if one examined the circle (which represent theactions of the individual) or the deviations from the circle, one would neverbe able to derive the square and triangle (which was the actual goal of thesubject). A very interesting point, because much of psychology spends its timeexamining the deviations from the circle, thus missing the point of thebehavior.
However, the square and triangle were not very well drawn. That is, thelines were not straight, but jagged. Runkel was not arguing that the desire ofthe subject was to draw a jagged square and triangle, it just happened that themechanism for drawing was fairly difficult for the subject, so the results wasjagged. For us in the applied side of psychology, the jaggedness of thosesquares and triangles is exactly the error we wish to study. We are interestedin individual differences, training, or devices that related to the jaggednessof results. It is extremely important that we study the jaggedness of thesquares and triangle and not the circle, but jaggedness we study.
Finally, to continue on the psychology as useless thread, all I am tryingto say when it comes to using instruments to gather information for makingselections, is that the information will reduce the probability of making atype I or type II error or both. That probability may be small, but it is typeI or type II or both that we all abhor (which is worse depends on the decisionand who one asks).
I also thought of a set of psychological experiments that many would notthink useless or dangerous (although McPhail beat me to it with the Sherifstudies). Specifically, the studies that Thurgood Marshall used to argue theBrown v. Topeka BoE case. They were the studies were the black childrenidentified white dolls as better than the black dolls. He used those studies toargue that the black children's self-esteem(self-conceptin today's parlance) was lower than whites.
I also always thought that Asch's conformity experiments were veryinteresting.
The paper I mentioned early was a presentation (15 minutes) at the Academyof Management Conference on fitting persons and organizations using asystems/cybernetic/control theory framework. I see those as subsets of eachother. Anyway, 15 minutes is not long enough to say much and I did not get muchof a response from the audience. The organizational theorists were morereceptive than the organizational behaviorists. The discussant claimed to likeit -"after getting through the jargon" -I had terms like input function, equifinality, and requisite variety init.
I did interact with others at the conference who were sympathetic with PCTor some variation (most have only been exposed to Carver & Scheier, Lord,or some other "heretic"). I was exposed to more on action theory (aka, Germanpsychology), too much TOTE, but some nice stuff. I chatted with Locke, askinghim if it bothered him that his performance cycle (a positive feedback loop)had no negative loop (in the systems dynamics sense). He did not seem tocare.
Anyway, back to trying to get tenure.
Jeff
(P.S. I got the closed loop articles but have not read them yet,thanks)
Date: Thu, 15 Sep 1994 10:18:39 -0600
Subj: Re: intensional; the beak of the finch
[From Bill Powers (940915.0830 MDT)]
. . . . .
Jeff Vancouver (940913) --
Welcome back. I hope I can persuade Mary to give a report on her attemptsto communicate with Locke; you should find them interesting. . . . . . Best toall,
Bill P.
Date: Mon, 19 Sep 1994 10:22:53 -0700
Subj: Beyond the fringe
[From Rick Marken (940919.1015)] >Jeff Vancouver (940913) --
> Upshot, she is very close to PCT in principle, but errors on some ofthe details. This is the basic argument I have been making to the Locke et alcamp.
Ah, but in those details is where the god of PCT lives. The differencebetween PCT and conventional behavioral science is based on one, tiny detail; apreposition. In conventional behavioral science, behavior is controlled _by_perception; PCT shows that behavior is the control _of_ perception. Just adetail.
> incorporating that [DME] into the control theory model is paramount onmy list of things to do.
Why? Doesn't it seem like the first thing to do would be to obtain somenice, clear, reliable data and THEN decide what kind of model might be neededto explain it?
> Third, the issue of the use of statistics (conventional behavioralscience methods) continues to bother many on the net.
No. It's the MISuse of statistics that bothers us; in particular, usingaggregate data as the basis for conclusions about individuals.
> systems/cybernetic/control theory framework. I see those as subsetsof each other.
Me too. "Cybernetics" is a subset of "systems" because the latter is a typeof perceptual variable and the former is a state of that type of variable(other states include "S-R","information theory" and "control theory" itself). Control theory is thesuperset of both "cybernetics" and "systems" because it explains how and whypeople maintain (control) systems variables in states like "cybernetics" or"control theory".
Best Rick
Date: Mon, 19 Sep 1994 12:57:45 -0600
Subj: Re: Locke and PCT
[from Mary Powers 940919]
Jeff Vancouver 94013
Your comment on Locke's indifference to the implications of having apositive feedback loop was interesting ("he did not seem to care")
I will probably not hear back from him, because in his last to me he calledme presumptuous (for suggesting that he should learn something about controltheory before criticizing it). He said
I do not plan to read the 1973 book you cite [BCP], partly, because basedon what you have said in your letter, it will be more of the same, and partlybecause the title itself is invalid. People do not behave to controlperceptions but to achieve values, ie, to live. In short, I believe that yourpremises are _fundamentally_ mistaken. Given this, there is little pointhaggling over details.
Maybe Bill should retitle the book "Behavior, the control of
perception in order to achieve values". In any event, it should be
noted that when Locke talks about control theory it is not just
from ignorance -it is from deliberate ignorance.
Mary P.
Date: Tue, 20 Sep 1994 09:05:44 CST
Subj: Re: Locke and PCT
Tom Bourbon [940920.0835] >[Mary Powers 940919]
>>Jeff Vancouver 940913
> Your comment on Locke's indifference to the implications of having apositive feedback loop was interesting ("he did not seem to care")
> I will probably not hear back from him, because in his last to me hecalled me presumptuous (for suggesting that he should learn something aboutcontrol theory before criticizing it). He said
Wonderful, Mary. Henceforth we cannot be accused of being nasty when we saythat, on the subject of control theory, there is no reason to take Lockeseriously. He is a popular dispenser of --who knows what, but it is not control theory as a theory that explains thephenomenon of control.
We can add Locke to a growing list of experts on control theory whodeclined our earnest offers to communicate on the subject: Carver, Lord,Bandura. Can anyone else suggest other names for the list? How many more mustwe add before it becomes obvious why there is a lack of "bridges" betweenPCTers and other experts on control theory?
You quoted Locke:
> I do not plan to read the 1973 book you cite [BCP], partly, becausebased on what you have said in your letter, it will be more of the same, andpartly because the title itself is invalid. People do not behave to controlperceptions but to achieve values, ie, to live. In short, I believe that yourpremises are _fundamentally_ mistaken. Given this, there is little pointhaggling over details.
This beautiful piece (which deserves framing on the wall above every PCTmodeler's computer) demonstrates that we have been right, all along, when wetold people it was pointless to look for similarities between PCT and whateverthat stuff is that Locke dishes out. When it comes to the phenomenon ofcontrol, it seems there is no point to looking for people whose ideas are"close, in principle," to PCT. After witnessing the antics of many almost-PCTers,I am just about convinced there is no such animal as an expert on controltheory whose ideas are close to PCT, "in principle." It has always turned outto be all or nothing.
> Maybe Bill should retitle the book "Behavior, the control ofperception in order to achieve values". In any event, it should be noted thatwhen Locke talks about control theory it is not just from ignorance -it is from deliberate ignorance.
It is noted. In my biased book, innocent ignorance is no vice, but thedeliberate ignorance I have seen in many so-called"almost-PCT"experts is another matter. Thanks, Mary.
Later, Tom
Date: Tue, 20 Sep 1994 10:53:56 -0400
Subj: Details, Decisions, and Locke
[from Jeff Vancouver 940920]
First, I would like to congratulate Rick for a terse response. Itfacilitates discussion
>[From Rick Marken (940919.1015)]
>> Me
>> Upshot, she is very close to PCT in principle, but errors on someof the details. This is the basic argument I have been making to the Locke etal camp.
> Rick
> Ah, but in those details is where the god of PCT lives. The differencebetween PCT and conventional behavioral science is based on one, tiny detail; apreposition. In conventional behavioral science, behavior is controlled _by_perception; PCT shows that behavior is the control _of_ perception. Just adetail.
I think that Ayn Rand is arguing against behavior is controlled byperception as well, but their are fundamental problems with many of herarguments that prevent good science. A detail that Locke is clearly missing(and I thank Mary Powers [940919] for another compact response) is that valuesare merely a type of perception (the perception one wishes to perceive). BothLocke and Rick seem think the differences in their models are much larger thanI think. On the other hand, the differences in their approaches arefundamental. Locke believes in grounded research (inductive reasoning) and Rickbelieves in deductive reasoning. Again, though, both do more of the other thanLocke wishes to admit (Rick admits his need for data to inform theory). In bothcases, their approaches have served them well, in that it has provided themwith a method to accomplish their goals.
Mary, I would not worry about convincing Locke directly. He does have ablind spot toward control theory. My goal is merely to convince many of thosewho use his theory that a more elaborate and well specified theory exists thatcan take them much further than goal-settingtheory.
However, that more elaborate, well specified theory (PCT) needs toincorporate decision making:
ME:
>> incorporating that[ DME] into the control theory model isparamount on my list of things to do.
Rick:
> Why? Doesn't it seem like the first thing to do would be to obtainsome nice, clear, reliable data and THEN decide what kind of model might beneeded to explain it?
I am not sure what constitutes "nice, clear, reliable data" without sometheoretical filter for interpreting it. Reliability has it own theoreticalprecepts. Nice and clear require others.
The data/observation that compels me and others to explain DM is theobservation that we make choices. Given 2 alternatives, we chose one. What thismeans with a PCT framework is important. For example, when does the individualgo into thinking mode to evaluate alternatives? How to higher-ordercontrol systems enter the equation? I could go on, but I am trying to beterse.
Me:
>> Third, the issue of the use of statistics (conventional behavioralscience methods) continues to bother many on the net.
Rick:
> No. It's the MISuse of statistics that bothers us; in particular,using aggregate data as the basis for conclusions about individuals.
The applied researcher is much more interested in casting nets. What youlabel misuse, others label useful.
Me:
>> systems/cybernetic/control theory framework. I see those assubsets of each other.
Rick:
> Me too. "Cybernetics" is a subset of "systems" because the latter is atype of perceptual variable and the former is a state of that type of variable(other states include "S-R","information theory" and "control theory" itself). Control theory is thesuperset of both "cybernetics" and "systems" because it explains how and whypeople maintain (control) systems variables in states like "cybernetics" or"control theory".
I got confused in the multiple uses of the phrase "control theory" in thispassage.
Later Jeff
Date: Tue, 20 Sep 1994 11:32:15 -0700
Subj: Locke'd Out
[From Rick Marken (940920.1130)] Mary Powers (940919) --
> I will probably not hear back from him [Locke], because in his last tome he called me presumptuous (for suggesting that he should learn somethingabout control theory before criticizing it). He said
>> I do not plan to read the 1973 book you cite [BCP], partly,because based on what you have said in your letter, it will be more of thesame, and partly because the title itself is invalid. People do not behave tocontrol perceptions but to achieve values, ie, to live. In short, I believethat your premises are _fundamentally_ mistaken. Given this, there is littlepoint haggling over details.
As my daughter would say, it looks like you've been "dis-ed",Mary, my dear. I suppose it has never occurred to Locke that the "mistaken-ness"of your premises (and the correctness of his) could be tested byexperimentation.
> it should be noted that when Locke talks about control theory it isnot just from ignorance -it is from deliberate ignorance.
Beautifully put.
I forget who this Locke fellow is. It seems to me that there is this bunchof people, including Carver, Scheier, Locke, Karoly, etc who are fighting overthe virtues of variants of a "control theory" model of human nature. But sincethey all seem to be clueless about 1) the nature of behavior 2) the nature ofcontrol 3)the nature of control theory 4) the nature of modelling and 5) thenature of science itself, it's hard for me to remember who is ostensibly "for"and who is "against" control theory. To me, it looks like a confederacy ofdunces.
Best Rick
Date: Wed, 21 Sep 1994 07:03:29 -0600
Subj: Decisions; self-esteem
[From Bill Powers (940921.0400 MDT)]
Jeff Vancouver (940913) --
> One of Lord's major errors is in placing a decision mechanism(function) between the comparator and the output function. It is an attempt onhis part to include a DME within control theory. I give him credit for trying.Choices are made, incorporating that into the control theory model is paramounton my list of things to do. He took a stab at it.
My problem with a DME is that I don't know what it is supposed to do. Youpresent me with a piece of candy and a dime and tell me I can take one or theother but not both. After I pick up one of them, I am said to have made a"decision" or a "choice." But that is only the outcome; what is the process?How did I get from the state of looking at both items to the state of havingpicked one of them up? Can you describe that process without using "decide" or"choose" or other synonyms?
"Decision" and "choice" are words we learned when we were very young. Theyare "natural language" terms with no formal meanings --that is, meanings derived from any systematic understanding, any model. Likemost such terms, we use them without giving much thought to them; they seem tooobvious to define further. We are just supposed to know what they mean.
As a parallel problem, consider the term "comparison." We all know that acontrol system needs a comparator, a way of comparing the perceived state ofaffairs with a reference state. We all know what "compare" means. But how do weconstruct a comparator for a computer model of a control system? If you look atthe instructions available in any low-level programming language for any computer, you will not find an operationcalled "compare." Of course we all have a natural-languagemeaning for the term; if I show you a lemon and a grapefruit and ask you whichis larger, you compare one size with the other and pick the larger. But how didyou do that? What operation did you perform on the two sizes that resulted inknowing which was the greater? There is only one operation you can find in thecomputer that will create this result, and that is _subtraction_. You form aperception of size for each one, subtract one size-measurefrom the other, and assign the terms "larger" and "smaller" according to thesign of the result. Are there any other operations that will qualify as"comparison?" Perhaps there are some logical operations, but as far as I cansee they reduce in the end to subtracting one measure from another. And Ihaven't found a case where this is not true. That doesn't mean that no suchcase exists; I simply haven't found it.
So wherever we see the term "compare," we can substitute "subtract" after asuitable analysis of the situation. Now we have a meaning for compare that wecan put into a model that will run on a computer.
What I am asking with regard to "choose" or "decide" is what operation Ineed to program into a computer in order to do these things. If there is nosuch operation, then I need a definition of what the operation should be,whether or not it can run on any existing computer. If there is such anoperation, then we can hope to incorporate choice and decision into a workingmodel because now we know what operations to perform that will lead to havingmade a decision or choice.
The only such operations that I have ever seen proposed are arithmetical orlogical in nature. You consider alternatives, you weigh them against eachother, and you subject them to some rational process that results in anoutcome. But as soon as you lay out all the details in this way, there is nolonger any decision or choice to be made. You consider that you could spend thedime on something you like better than the candy, but also that the candy isright here and requires no trip to a store to buy a dime's worth of somethingelse. You compare the relative effort involved and let the course of actionthat involves the least energy expenditure in comparison with the expectedsatisfaction take place. There is no decision to be made, once you have pickedthe grounds on which the decision is based. That is, "decision" involvesnothing beyond perceiving the elements of the situation and applying analgorithm that yields an answer based on the elements and logic orarithmetic.
If that is all that decision means, then we don't need the term decisionexcept as a convenient shorthand. And we don't need to add anything to the PCTmodel, because all the processes and perceptions involved in making a decisionare already part of the model, with a specific level already assigned to doingthe operations that result in decisions. The only thing that would make us adda separate box labelled "decision" would be to find something in the processthat can't be explained by applying rules or algorithms to perceptions. Ofcourse if "decision" means something else, then we have to specify what thatsomething else is, and we have to add a level to the model capable of doingthat sort of thing.
PCT resulted from spending a lot of time, a very LARGE amount of time,analyzing common-languageterms that are used to describe behavior and experience. It was during thisprocess that I realized that there were large numbers of words that I usedevery day without ever having asked myself what they are supposed to mean. Iused to think of myself as quite the intellectual; I could argue on any subjectwith anybody and as far as I was concerned, win. Words, words, words, and noneof the important ones had any meanings that I had thought through. When I thinknow about some of my youthful pontifications, I want to avoid meeting anyonewho knew me then.
It seems to me that Locke, Bandura, Lord, and others in their field relyalmost exclusively on words with no meanings that have ever been thoughtthrough. They throw common-languageterms around as if merely saying them is enough. To pick a central term, whatis a "goal?" This word is used as if everybody knows what it means, but none ofthe writers who use this term ever stops to ask what a goal is, how it can haveany effects, why the actions we perform have the result of achieving a goal,how we know when there is a "goal-discrepancy."
Bandura is very proud of the term "self-efficacy",but what does it mean? It's a bastard construction to begin with, which leadsto confusion with the forms of "self-starting"and "self-explanatory"and the like. Does it mean "efficacy of the self in creating actions?""Efficacy of actions in producing effects pertaining to the self?" And whatdoes "efficacy" mean? What does "self" mean? When you hear sentencesconstructed with terms like these you get a sense of knowing what the speakermeans, but it is a very vague sense and subject to much idiosyncraticinterpretation. The discourse is composed exclusively of unanalyzed natural-languageterms. As long as you just let the words flow by, you get a feeling ofunderstanding what the speaker is going on about, but the moment you pause toexamine any part of it, and ask what it means, the sense of meaning disappears.Self-efficacyis simply a conjunction of two terms, one sort of referring to the aspect ofexperience we call ourselves, and the other sort of meaning being able to haveeffects or be effective or accomplish things. It's much like saying "foodnice." Say self. Say efficacy. Now say them together. Get it?
PCT puts a layer of formal meanings underneath the natural-languageterms in which we talk about behavior and experience. Of course these are_proposed_ meanings, and are always open to modification. But when we talk inways that have no such layer of formal meanings beneath them, we are justtalking; we aren't being theorists. We're conveying meanings as best we canwithout having any formal understanding of what terms are to mean. This is howpeople naturally converse with each other, how children talk and how adultstalk who have never thought deeply about what their own words mean to them. Itis somewhat miraculous that people are no worse at grasping what others aretalking about, and at conveying to others what they mean, than they are.
Science can happen only when people get together and say "Now look, we justhave to reach some agreements about what certain terms are going to mean." Whenthey actually get down to doing this, they find that they were far less inagreement than they thought they were. They find that they have to work theirway down and down through layers of language until they can find commonexperiences so simple that misunderstanding is next to impossible. "I say thatmeter reads 3.0." "Well, I say it reads 3.1." "Move over so you're lookingstraight down at it." "OK, it's 3.0." Once that's settled, they can startworking out the question of 3.0 _whats_.
Rick Marken's example of the three blind men and the elephant has anotherside to it. The three blind men, if they just rely on words, can think they aretalking about the same thing when they're really talking about differentexperiences. The one at the tail says "I feel a cylindrical appendage --let's call it a pachyoid." The one at the legs says, "OK, right, I feel fourpachyoids," and the one at the trunk says "There's another one here." When theywrite their report, they say that elephants are made of six pachyoids. And theygo their separate ways thinking they are talking about the same thing justbecause they are using the same word.
---------------------
New subject:
> I also thought of a set of psychological experiments that many wouldnot think useless or dangerous (although McPhail beat me to it with the Sherifstudies). Specifically, the studies that Thurgood Marshall used to argue theBrown v. Topeka BoE case. They were the studies were the black childrenidentified white dolls as better than the black dolls. He used those studies toargue that the black children's self-esteem(self-conceptin today's parlance) was lower than whites.
This illustrates another problem I have with much (not all) psychologicalresearch. What the studies you cite showed was that black children treatedwhite dolls as being better than black dolls. From this, it is informallyinferred that therefore black children think that white children are betterthan black children, and from that we conclude that black children consider_themselves_ worse than white children, and therefore that black children havelower self-esteemthan white children.
All that seems reasonable enough to base a decision on in the terms inwhich most decisions are actually made, but it's not enough to call it science.The problem is that _plausibility_ is substituted for _proof_. If the proposedlogical connections were actually valid, then the final conclusion would bevalid. But not otherwise.
Proof in this case would amount to finding some independent way ofdetermining the black children's degree of self-esteem.This is the only way to validate this procedure for determining self-esteem.Before this procedure is applied to a general population, we must establishthat when black children say white dolls are better, they are also experiencinglow self-esteem.We can see which dolls they choose, but how can we see their self-esteem?What we're trying to establish is that the choice of dolls, which is easy tomeasure, is a reliable indication of self-esteem,which is not so easy to measure. If we could establish that connection, wecould then substitute the easy measure for the hard measure in testing ageneral population. So the problem of showing that the test means somethingboils down to measuring self-esteemin some direct and independent way.
But here we are stymied. We can't see the self-esteemdirectly, unless we happen to be the particular black child under study, andhave the sophistication to think in such terms. We can't just ask a three-year-old child how much self-esteemhe or she feels; that's a meaningless term to a three-year-old.
The only recourse is to try to devise a completely different test for self-esteem.We see how black children behave when getting into lines with white children.We see how "assertive" they are in competing for toys with white children. Wesee what roles they play in pretend games. But in each case we run up againstthe same problem: there is no direct measure of self-esteemavailable for validating any of the tests. In each case, we simply assume thatthe behavior we can observe is indicative of self-esteem.
So what do we end up with? A series of tests which are validated againsteach other, and none of which is actually known to indicate self-esteem.This thing called self-esteemremains hidden from observation; in fact we don't know what any of these testsis indicating, if anything.
This problem became evident to me when I began taking psychology courses incollege. And it led me to wonder about something that psychologists didn't seemto care to think about: did the things that these tests supposedly measureexist at all? In the above example, I would have asked, is there really such athing as self-esteem?Is this something that people have? Or is it just a way of talking?
Now I think I would say that yes, there is such a thing as self-esteem,but that term is only a rough indicator of what is involved. To get any closerto a useful understanding, we have to ask what "self" means, and what "esteem"means, so we can see some sort of process going on. Obviously, self-esteemis a perception that can be had only by the person involved. It is a perceptionof one's own characteristics. And this perception, to have any value, must becompared with a reference level for the each state of those characteristics onehas chosen as a target. If you perceive yourself as less than you want to be ineach relevant respect, then you are probably experiencing what people call lowself-esteem.
But there is another side to this, which becomes visible only when you usethe PCT model. What if your perception of your self matches the reference levelwith respect to each variable involved in characterizing yourself, but you haveset your reference levels very low for characteristics that other peopleusually set to a high level? Your perception of self-esteemwould then be exactly what you have chosen in each respect. It would not seemlow to you, but others would still judge you as having low self-esteem.You, as a black child, would go behind the white children in the line, handover a toy to a white child without protest, and play the black victim ratherthan the white victor in pretend games. Perceiving yourself doing these thingswould fit perfectly with your reference levels for how you should be.
How could such reference levels come to be set so low? Obviously, becausethese are the settings that work, all things considered. If you are raised in asociety that punishes you for assuming equality or ascendance in the company ofwhite people, you can avoid that punishment by choosing ways of being that donot cause punishment. You remember how you were when you were getting along aswell as possible, and you pick those memories as reference levels. This doesnot require any thought; all it requires is reorganizing until the pain isminimized.
So now we are beginning to get a handle on self-esteem.We can distinguish between a person who perceives a self that is less than whatis wanted in various dimensions, and a person who has come to want less inthese same dimensions. And while we're at it, we have arrived at somedefinitions of the dimensions that pertain to self-perception.With this understanding, we're ready to apply control theory to the problem ofdetermining a person's self-esteem.
The procedure goes this way. You pick a person, and start interacting withthat person. What you're looking for are controlled variables that pertain toself-perception,as you define it. You see how this person reacts to disturbances of thevariables you have chosen; if there is no reaction, you strike that variablefrom the list. Eventually, after a lot of interaction, you arrive atdefinitions of variables relevant to self-perception,as you see it, that you can demonstrate to be under control by this person. Andyou also find the reference levels relative to which each variable iscontrolled, so you can tell if low self-esteemis due to a large error or to a low setting of the reference levels.
Then you pick a second person and go through this procedure again. And athird, and a fourth, until you have done all the people in your test group, oneat a time. By the end of this time you will know 50 or 100 or 200 people verywell indeed, with respect to what you see as their self-concepts.
Then you can start looking for common factors, for variables that indicatecommon problems, for deviations of individuals' reference levels from those ofmost of the group. You look for indirect indicators that show highly reliablecorrelations with either felt lack of esteem or low levels of aspiration foresteem. Now you have a direct measure of self-esteem,as you have carefully defined it, for a large enough group of individuals toallow validating other kinds of tests that are easier to apply. Of course theseother tests will NEVER yield the kind of reliability that the basic study ofone person at a time will yield, and no important decisions should be madeabout any one person on the basis of such tests. But such tests can be used asthe basis for modifying policies that apply to many people or for assessing theoverall effectiveness of a program, a teacher, or an administrator.
Compare where we started with where we ended. We began with childrenplaying with dolls, and we ended with a study of individuals aimed atdiscovering how their self-conceptscompare with their desires, and what those desires are. We started with a vaguecommon-languageterm, self-esteem, and ended by talking about specific variables associated with self-perception.Which approach is going to yield solid knowledge of how people perceivethemselves, how children learn to perceive themselves? I submit that the answeris obvious.
It's much, much easier to hand out a set of questionnaires to 30 peoplethan it is to sit down with each one of them for hours and hours, learning howeach person ticks. There are pressures of time and competition and pride andfunding that make psychologists look for quick and easy ways to get informationabout important aspects of human nature. But the results reflect the amount ofeffort and thought expended. If anyone wants to conduct psychologicalinvestigations in a way that will yield reliable knowledge about human nature,there is really only one way to start: one person at a time. Without that kindof basic understanding, all generalizations are fuzzy and empty and hardlyworth the trouble.
Best, Bill P.
Date: Wed, 21 Sep 1994 08:55:41 CST
Subj: Re: Details, Decisions, and Locke
Tom Bourbon [940921.0806]
Breaking in on Jeff's reply to Rick. Out of respect for Jeff's expressedpreference, I'll try to be "terse."
>[from Jeff Vancouver 940920]
Re: [From Rick Marken (940919.1015)]
. . .
Jeff:
> . . . Both Locke and Rick seem think the differences in theirmodels are much larger than I think. On the other hand, the differences intheir approaches are fundamental. Locke believes in grounded research(inductive reasoning) and Rick believes in deductive reasoning. Again, though,both do more of the other than Locke wishes to admit (Rick admits his need fordata to inform theory). In both cases, their approaches have served them well,in that it has provided them with a method to accomplish their goals.
Yes, Jeff. Locke believes in, and stakes his reputation on, research usingthe traditional experimental designs in behavioral science. The kind ofresearch in which the experimenter creates Hypothethetical NomotheticAndrogynous Persons (HYNAPs) --Neutered Average Persons possessed of all manner of drives, traits, qualitiesand other causal demons--then talks about then as though they (the HYNAPs and the demons) are real. Thekind of research in which the "experimental hypothesis" is _never_ evaluated bythe design or the analysis of data --the occurrence of a "statistically significant difference" between the averageHYNAPs in two or more groups of subjects in no way confirms the experimenter'shypotheses about the demons that made the hypothetical persons act as they did.The statistical procedures are constructed so that nothing whatsoever can besaid about any specific person. That's one of the differences Rick was talkingabout.
Of course you are correct when you say Locke's use of that kind of researchhas served him well. It is the ticket into respectable mainstream "behavioralscience." But does it _ever_ address the truth or falsehood of theexperimenter's hypotheses? To be terse, no, never.
> Mary, I would not worry about convincing Locke directly. He does havea blind spot toward control theory. My goal is merely to convince many ofthose who use his theory that a more elaborate and well specified theory existsthat can take them much further than goal-settingtheory.
Brave soul! Good luck.
Jeff:
>>> incorporating that[ DME] into the control theory model isparamount on my list of things to do.
Rick:
>> Why? Doesn't it seem like the first thing to do would be to obtainsome nice, clear, reliable data and THEN decide what kind of model might beneeded to explain it?
. . .
> The data/observation that compels me and others to explain DM is theobservation that we make choices. Given 2 alternatives, we chose one. Whatthis means with a PCT framework is important. For example, when does theindividual go into thinking mode to evaluate alternatives? How to higher-ordercontrol systems enter the equation? I could go on, but I am trying to beterse.
Jeff, I don't think Rick was asking if people ever make decisions. It seemsto me that he was asking if the occurrence of decisions in itself requiresadditions to or modifications of the PCT model. Have you tested the model todetermine if, in its present form, it can produce the "phenomenon" of "making adecision?" Or do you simply like the idea of DMEs and want to include them inthe model for reasons that are more aesthetic, or political, or whatever? Youwould not be the first to do such a thing. At one time or another all of ushave probably done something like that.
Me:
>>> Third, the issue of the use of statistics (conventionalbehavioral science methods) continues to bother many on the net.
Rick:
>> No. It's the MISuse of statistics that bothers us; in particular,using aggregate data as the basis for conclusions about individuals.
> The applied researcher is much more interested in casting nets. Whatyou label misuse, others label useful.
Yes. And when methods are used in ways which violate all of the underlyingcriteria that determine their appropriateness and relevance, that is alsomisuse and it is the reason we question or reject the majority of research inthe social-behavioral-lifesciences. The fact that traditional behavioral scientists are "interested in"their work does not alter the fact that they often (practically always?)violate the rules that might make their work legitimate as a kind of netcasting. (Phil Runkel speaks eloquently to this issue.)
We play our different games.
The rules of terse talk require that I quit.
Later, Tom
Date: Wed, 21 Sep 1994 11:35:56 -0400
Subj: Locke, all or nothing
[From Jeff Vancouver 940921]
to [Tom Bourbon 940929.0835][Rick Marken 940920.1130 & .1800]
I am glad to see that you all have the corner on truth. I suppose we shouldall take the perspective that our way and our understanding is the correct one.We would not be bothered by gaps in our understanding or differences ofopinion.
I could argue that you are as guilty as Locke for deliberate ignorance, butyou will say you have read (at least some) of their stuff. Of course, so hashe, and he did not fully grasp it.
I could argue that your goal (developing a model of humans) and theirs(developing applications to help individuals and organizations) are slightlydifferent and require different methods (casting nets v. testing specimens touse Runkel's language), but you seem to be ignoring that argument.
I could say that if you don't have something good to say, don't sayanything. But I do not even believe that (or I would not be writing thisresponse). I prefer, if you don't have something constructive to say, do notsay anything. But you will respond that you do have something constructive tosay, namely, don't waste your time looking at the work of these researchers.Yet, that is the deliberate ignorance you just railed against. Ah, we come fullcircle.
Einstein did not believe in quantum physics. I tend to agree with him inprinciple, I do not believe uncertainty is a physical property (although I amno theoretical physicist). Nonetheless, much has been achieved both practicallyand theoretically by adopting a quantum framework. If you believe Hawking, thecombination of general relativity and quantum physics will lead to the end ofphysics because all the questions will be answered. Hawking has certainly madecontributions combining the perspectives.
I see science as a big hierarchical control system. Conflicts betweenreference signals should be addressed, not ignored. Otherwise we are wastingresources. I don't think everyone needs to be addressing the conflicts, but donot begrudge those of us that attempt it.
Later Jeff
Date: Wed, 21 Sep 1994 11:28:17 -0700
Subj: Terse Replies
[From Rick Marken (940921.1100)] Jeff Vancouver (940920)
> First, I would like to congratulate Rick for a terse response. Itfacilitates discussion
Thanks. For you, it will be terse all the way.
> Locke believes in grounded research (inductive reasoning) and Rickbelieves in deductive reasoning.
Could you tersely explain what this means?
> In both cases, their approaches have served them well, in that it hasprovided them with a method to accomplish their goals.
True. My goal is understanding control; Locke's goal is gettingprofessional recognition.
> However, that more elaborate, well specified theory (PCT) needs toincorporate decision making:
It already does.
> I am not sure what constitutes "nice, clear, reliable data" withoutsome theoretical filter for interpreting it.
There are many examples in my "Mind Readings" book and in Bourbon andPowers' "Models and Their World's" article.
> The data/observation that compels me and others to explain DM is theobservation that we make choices.
But a choice-makingmechanism may not be needed to explain this. People can also be observedresponding to stimuli but an SR mechanism is not needed to explain this.
> The applied researcher is much more interested in casting nets. Whatyou label misuse, others label useful.
Yes. And what some label "shit", others label "caviar". But it is stillwhat it is. The goal of applied PCT is to improve individual control. I can'tsee how this goal can possibly be achieved by "casting nets".
Tersely Rick
Date: Wed, 21 Sep 1994 18:54:25 -0700
Subj: Locke, stock and barrel
[From Rick Marken (940921.1900)]
Ah, I am getting posts on the day they were posted, for the time being,anyway.
Jeff Vancouver (940921) --
> I could argue that you are as guilty as Locke for deliberateignorance, but you will say you have read (at least some) of theirstuff.
Not only that, but the two articles you sent are right here in front ofme.
> if you don't have something constructive to say, do not say anything.But you will respond that you do have something constructive to say, namely,don't waste your time looking at the work of these researchers. Yet, that isthe deliberate ignorance you just railed against.
In fact, I say "look at the work of these researchers carefully and youwill see why, what they are doing, has nothing to do with understandingpurposeful behavior". What is wrong (from a PCT perspective) with the work ofLocke, et al is what is wrong with conventional psychology in general. The PCTliterature is one long, careful, model and research-basedexplanation of why the kind of research and "theory" found in Locke's articlestells us nothing about the nature of control. Believe me, if there wereanything of even the slightest value in Locke's papers --anything that could be used to leverage a PCT based approach to research andtheory in this area (which seems to be called "goal setting")--I would run with it immediately. But there isn't; in fact, there isn't evenanything to argue against (with models and research); it is just, plainirrelevant.
The only thing that distinguishes Locke's work from the rest of theresearch in conventional psychology is his attempt to use the vocabulary of PCT(or, at least, of control theory). But it is all just words; there are no testsfor controlled variables, no one-person-at-a-time research, no working models of control, nothing but the SOS (same oldfecal material). This has nothing to do with access to "the truth". It has todo with understanding the nature of purposeful behavior, how to model it andhow to study it. Locke et al are clueless about this --as is most of the rest of the conventional psychological community. Clearly,these people are happy with the kind of research they are doing and what theyare learning from it. I'm glad that they are happy. You are the one who seemsto think that this research has some relevance to PCT, or vice versa. We havebeen trying to tell you (with clearly diminishing levels of patience on Tom andmy part) that this is simply not true. You don't seem to agree with us, so,fine. I think we're pretty explained out.
Tom and I get impatient with your claims about the relevance of Locke toPCT, not only because we have bad values (but I promise to peek though BillBennett's book so that I can learn the right values to have) but because wehave been through this 1000 times before, with reviewers, psychologists andeven some people in CSG who are sure that we can and should build "bridges" tothose doing work that seems "close" to PCT. It just doesn't work. PCT is PCT;it directly contradicts (or differs completely from) every other model ofbehavior that has ever been proposed (that we are aware of).
> Conflicts between reference signals should be addressed, not ignored.Otherwise we are wasting resources. I don't think everyone needs to beaddressing the conflicts, but do not begrudge those of us that attemptit.
But there is no conflict. The articles by Locke that are sitting in frontof me have nothing to do with what I am controlling for; an understanding ofthe nature of purposeful behavior through research and modelling. The onlything in these articles that could be of any use to PCT is the data and models.There are no models (except for an incorrect diagram of a control system) andthe data is useless because it is a summary of group performance and isextremely noisy at that.
I feel no conflict with Locke et al; we are clearly not trying to controlthe same variables. His research is no more of a disturbance than most of theother work in conventional psychology. It is completely and utterly irrelevant.The only disturbance occurs when people (like you) say that it is relevant. Itain't. You seem to think we (Tom and I) are in conflict with Locke; that iswrong. What we are saying is that Locke's work is useless to us --as useless, it appears, as ours is to him.
Untersely Rick
Date: Thu, 22 Sep 1994 08:35:29 CST
Subj: Re: Locke, all or nothing
Tom Bourbon [940922.0833] >[From Jeff Vancouver 940921]
>to [Tom Bourbon 940929.0835][Rick Marken 940920.1130 &.1800]
I read your post, Jeff. This will be terse.
Read [From Bill Powers (940921.0400 MDT)], in which you will see once againthe way PCT modelers go about the business of trying to understand humanbehavior and to help them. After you read those ideas one more time, I hope youwill give some thought to abandoning your ad hominem style when you are on thisnet.
Later, Tom
Date: Thu, 22 Sep 1994 10:28:07 -0400
Subj: Locke tersely
[from Jeff Vancouver 940922]
[Rick Marken 940921.1100 & 1900] & [Tom Bourbon 940921.0806]
Deductive reasoning is when the theorists focuses on the model and makesinferences about the observations (i.e., reasoning from the general to thespecific). Inductive reasoning is when one focuses on observations and attemptsto draw general premises (i.e., reasoning from the specific to thegeneral).
When you (Tom & Rick and others) focus on creating your simulation and_then_ testing it against observation, you are doing deductive research. A verylegitimate enterprise, particularly in the long run.
When Locke and others are collecting data and then trying to makeinferences of the underlying causes, they are doing inductive research. Veryuseful at early stages of phenomenon examination, but limited in theend.
When one has a good model from which to base deductive research, theyshould do it until their hearts are content. You have such a model and that iswhat you are doing. Great. You really do have little reason to bother with theothers' work (except, as Runkel maintains and I concur, to provide clues forcontrolled variables; I also believe there are gaps that could use some helpfrom other places, but you have located the most useful places -like perception psych). I do think you are wasting _your_ time keeping up onthe likes of Locke.
When you have a flawed model, like Locke's, deductive research is often aless than useful and can lead one to believe deductive research as a method isflawed. I think this is one of the things that is going on in Locke's head.Ironically, Locke's model is close to being like your model (actually, not soironically, because the observations are of the same phenomenon, so that themodels are similar is not surprising). But close does not cut it, _for thepurposes of understanding control_. Probably true.
But you also note that it is difficult to get tenure doing the PCT thing(deductive research using the PCT model). This is because the larger scientificcommunity (or at least a critical mass) has not accepted aspects of the model,the method, or the attitude of the proponents of PCT. So what does one do? 1)Give up convincing the scientific community if they can afford to, or 2) trythe convince some of them. You have chosen the former because you can afford to(or perceive that you can, which I do not question). I however, cannot affordto (or perceive that I cannot). Thus I am trying to convince.
All I am asking is that you do not make my job harder than it alreadyis!
Obviously, their are other differences in our position than this (I am muchless negative about averaging across people, etc.), but you have made thesepoints (to me anyway) and we are agreeing to disagree.
One final point, I found Mary's post about Locke very informative. I do notwant to stifle that kind of post. I gives me a better appreciation of where thedistractors are, what kinds of arguments might work, what might not, and whomight not be worth trying to convince. For example, I know the method argumentswill not work very well. I still make them, but put the thrust of my argumentselsewhere.
I will address Bill's post regarding DM and self-esteemafter getting some work done.
Later Jeff
Date: Thu, 22 Sep 1994 12:26:40 -0600
Subj: PCT and its critics
[From Bill Powers (940922.1035 MDT)] Jeff Vancouver (940921)
> I am glad to see that you all have the corner on truth. I suppose weshould all take the perspective that our way and our understanding is thecorrect one. We would not be bothered by gaps in our understanding ordifferences of opinion.
Anyone who understands the PCT model is welcome to tear it to shreds ifthey see something wrong in it. Unfortunately, the people who try hardest totear it to shreds are those who have given it the least study and who, as aconsequence, have leaped to a lot of erroneous conclusions about it. They'recriticizing a creature of their imaginations, not PCT. But PCT theorists sufferthe effects nonetheless when they try to make their views known.
I had written to Locke some time ago; with his response he sent a preprintof a paper titled "The Emperor is Naked", to be published in _AppliedPsychology: an international review_. Here is his theme paragraph from page3:
I must conclude at this point that control theory, as presented by thosewho claim to be control theorists, is so diverse in meaning, so all-encompassingin scope, and so devoid of specific, consistent content that it is everythingin general and therefore nothing in particular. That which is nothing inparticular is: nothing. I feel obliged, therefore, to play the role of thelittle boy in the children's story and declare the emperor to be naked. Ibelieve we would do well to abandon control theory altogether.
What Locke terms control theory is what he has picked up from readingCarver and Scheier, Hyland, Lord, and others writing in his own field. Withthat understanding, I tend to sympathize with his opinion. It would be betternot to mention control theory at all than to present it in a distorted andsuperficial way. The outcome of such presentations was evidently not impressiveto Locke and others in his coterie. I was not impressed, either.
However, PCT now stands discredited by a major player in the field ofpersonality research (if that is the right term for Locke's field). If pastexperience is any guide, we can predict what will now happen if a PCTresearcher sends a paper to Applied Psychology. The referees, having read whatLocke and Bandura and Binswanger have said about control theory, will realizethat control theory has no content and has been abandoned by right-thinkingscientists; they will cite these scientists, add a few irrelevant criticisms,and reject the paper. With luck, the referees will not be Locke et. al.themselves, but if they are not, they will still have no knowledge of controltheory of their own, and will take the word of experts in their field. Sincefew scientists read much if anything outside their own journals, that will bethat for PCT --once again.
Jeff, don't forget that you have just started learning PCT. You have a yearor so to go before your understanding of it will be complete enough to see whatis wrong in the way both Locke and the people he lambastes are treating it. Itis to your credit that coming from a different field and feeling stronglychallenged by PCT, you have stuck around to learn more and do your own thinkingabout it. But that simply puts you in the same position that nearly all currentPCTers were once in. Nobody on this net was conscripted and forced to read allthis stuff. Every person was self-selectedand came from fields built on ideas that clash in one way or another withPCT.
Tom Bourbon and Rick Marken did not begin life as PCTers; they were bothconventional psychologists who learned and taught all the things thatconventional psychologists learn and teach, including statistical methods. Theywere granted PhDs by mainstream psychologists. Rick Marken even wrote atextbook on statistical methods in experimental psychology. Almost all of theold PCTers except me were once just like you, your brethren in psychology andsociology and so on. All of them had the same problems you are having,accepting parts of PCT with great excitement but finding other parts inconflict with important things that they had accepted as truth.
PCT requires re-thinkingessentially everything in the behavioral and life sciences, because itintroduces a new kind of organization that was never part of those sciences.The phenomenon of control has always been present in the behavior of livingsystems, yet for at least 300 years the life sciences developed without onemention of it. How, we have to ask, did scientists manage to explain behaviorwhile ignoring the most central aspect of it? The only way we can answer thisquestion is to look at parallels in other fields. How did people explain theway ships disappear over the horizon while they believed the world is flat? Howdid they explain the movements of the sun, the moon, the planets, the stars,and comets while they believed that the universe revolved around the earth? Howdid they explain combustion while they believed that combustion required givingoff phlogiston rather than combination with oxygen?
What's most important to realize is that THEY DID FIND EXPLANATIONS.Furthermore, the best scientific minds of the times found these explanations tobe complete, plausible, and convincing. Scientists wrote in learned journalsabout these explanations, and applied the explanations to all sorts of aspectsof life and nature. We know little about dissenters from these mainstreamideas; our experiences with PCT may give some insights into why. There weresurely always dissenters, but dissenters do not write the history of scienceuntil they have prevailed. And then they, too, become the stodgy mainstreamers,fending off new dissenters.
Remember that life scientists DID FIND EXPLANATIONS of behavior withoutever considering the phenomenon of control. This means that no matter howplausible their arguments seem, how water-tighttheir reasoning sounds, how much experimental evidence they have amassed,_there is something wrong with what they say_. They have explained behaviorwithout mentioning its basic organizing principle. The more we learn about PCT,the more things about behavior that we see it explaining simply and clearly,the more obvious it is that the life sciences have not been playing with a fulldeck. A great deal of self-delusionmust have been taking place. Important problems must have been ignored,explanations must have been offered which are as flawed as those the old flat-earthersused to explain why you can see only the topmast of a ship eight miles out tosea, yet can see the whole ship again if you just climb a hill. How simple itis to propose that the earth is a ball, and how stoutly people insisted thattheir own complicated explanations were better!
And how simple it is to say that organisms behave in order to control whathappens to themselves. Yet look at all the complex, fuzzy, hard-to-test ideas that people have offered in place of this simple idea. Look at allthe terms like motivation and aspiration and traits, tendencies, propensities,proclivities, and habits that have been offered as explanations. Look at theunending attempts to trace the causes of behavior back to external stimuli, toinborn characteristics, to influences from situations and events and cues, tocomplex mental calculations. Look at all the experimentation in which even thevaguest relationships are seized upon as if they gave us the secrets of life.The sciences of behavior have been floundering in confusion since they firstbegan, all because they did not notice a simple, and now easily explainable,phenomenon.
The problems that PCT has had with conventional scientists are nothing newto science. We can only be encouraged that there are 150 or so scientistsinterested enough in PCT to keep tabs on what is going on, and even toparticipate in discussions about it and attend meetings about it. Who knows?There may be many more from whom we never hear, who are waiting to see whichway the ball bounces before committing themselves. There are over 5000 copiesof B:CP floating about somewhere in the world, not subtracting the ones hurledangrily into trash bins. Even though PCT has not figured prominently in theliterature (it has been mentioned more often in refutation than in support),there is obviously some pressure felt in the scientific community from thisidea. Why else would an author writing about something else insert a gratuitousobjection to PCT in a paper? The only thing that stands in the way ofwidespread acceptance is the magnitude of the changes that are implied by PCT.It is not what PCT has to offer that is the problem; it is what has to be givenup in order to understand it.
Best, Bill P.
Date: Thu, 22 Sep 1994 14:49:59 -0400
Subj: Re: Decisions; self-esteem
[from Jeff Vancouver 940922] >[Bill Powers (940921.0400)]
> My problem with a DME is that I don't know what it is supposed to do.You present me with a piece of candy and a dime and tell me I can take one orthe other but not both. After I pick up one of them, I am said to have made a"decision" or a "choice." But that is only the outcome; what is the process?How did I get from the state of looking at both items to the state of havingpicked one of them up? Can you describe that process without using "decide" or"choose" or other synonyms?
I think one problem with this discussion the use of a special designate:DME. I think the question of DM is the question of which reference signal ispassed down from the control unit that has as a reference signal "spend thisdime."
Once that reference signal is determined, the process is the usuallypropagation of reference signals down the hierarchy (which might have decisionpoints -I see eat candy, from which store do I get it?). Let's focus on one point inthe hierarchy.
> The only such operations that I have ever seen proposed arearithmetical or logical in nature. You consider alternatives, you weigh themagainst each other, and you subject them to some rational process that resultsin an outcome. But as soon as you lay out all the details in this way, there isno longer any decision or choice to be made. You consider that you could spendthe dime on something you like better than the candy, but also that the candyis right here and requires no trip to a store to buy a dime's worth ofsomething else. You compare the relative effort involved and let the course ofaction that involves the least energy expenditure in comparison with theexpected satisfaction take place. There is no decision to be made, once youhave picked the grounds on which the decision is based. That is, "decision"involves nothing beyond perceiving the elements of the situation and applyingan algorithm that yields an answer based on the elements and logic orarithmetic.
Couple of issues in this paragraph. First, "there is no decision to bemade, once you have picked..." what is picking if not deciding. You are sayingthere is no decision once you decided.
Second, I have no trouble (as a working hypothesis) with "nothing beyondperceiving the elements of the situation and applying an algorithm that yieldsan answer based on the elements and logic or arithmetic." This strikes me as areasonable description of the DM process. But the next question is whichelements are perceived? What is the nature of the algorithm? That is, what arethe parameters? Does the organism need to go into thinking mode to accessparameters for the algorithm? Can individuals control for perceptions thatcause (even in their own world view) other higher-orderperceptions to go out of line (i.e., why are some perceptual elements notrepresented in the algorithm)? Where is the algorithm?
I think the answer to the last question is the output function and henceLord's error. I think the structure for the answer is present in your model. Itmay be more fleshed out than I know because I have not read everything orforgotten some things. By the way, why do we forget?
I think that one of the elements in the algorithm, that is in the outputfunction is an estimate of the probability of attaining a reference signal(what Bandura calls self-efficacy).I believe the value for that estimate is partially a function of self-conceptslike self-esteem(more on that later) and perceptions related to achieving the perception in thepast (via actual interaction or perceptions of others doing or saying it). Ibelieve it is reasonable to attempt to influence those parameters. I believeBandura and others have shown us the reasonableness of these propositions. Ibelieve Bandura's model is flawed and incomplete and could be substantiallyimproved by your model (with the appropriate changes to the Bandura's model andpossible further specification of your model).
> It seems to me that Locke, Bandura, Lord, and others in their fieldrely almost exclusively on words with no meanings that have ever been thoughtthrough. They throw common-languageterms around as if merely saying them is enough. To pick a central term, whatis a "goal?" This word is used as if everybody knows what it means, but none ofthe writers who use this term ever stops to ask what a goal is, how it can haveany effects, why the actions we perform have the result of achieving a goal,how we know when there is a "goal-discrepancy."
We spend a great deal of time haggling over the meaning, measurement, andeven existence of our constructs. Sometimes this involves mathematical symbols,sometimes it does not. But we should not be fooled into believing that a greeksymbol is any more closer to the truth than an English one.
> PCT puts a layer of formal meanings underneath the natural-languageterms in which we talk about behavior and experience. Of course these are_proposed_ meanings, and are always open to modification. But when we talk inways that have no such layer of formal meanings beneath them, we are justtalking; we aren't being theorists. We're conveying meanings as best we canwithout having any formal understanding of what terms are to mean. This is howpeople naturally converse with each other, how children talk and how adultstalk who have never thought deeply about what their own words mean to them. Itis somewhat miraculous that people are no worse at grasping what others aretalking about, and at conveying to others what they mean, than they are.
As I said above, we look for deep meaning more than you give us credit. Iuse your BCP chapter on theory in my motivation class. Just taught ityesterday. Your division of extrapolation, abstraction, and modeling I finduseful, but your examples of personnel selection as extrapolation were off themark (mostly, some instruments are purely empirically based). And theabstraction you describe you never get published in our theoretical journals.There needs to be at least one layer lower of why we should see theabstractions we hypothesize. But you are right about the modeling. We would dowell to become that sophisticated.
> Science can happen only when people get together and say "Now look, wejust have to reach some agreements about what certain terms are going to mean."When they actually get down to doing this, they find that they were far less inagreement than they thought they were.
Often that is true about the definitions of their constructs, but oncepassed that they often find the processes they were talking about were thesame.
> New subject:
>> I also thought of a set of psychological experiments that manywould not think useless or dangerous (although McPhail beat me to it with theSherif studies). Specifically, the studies that Thurgood Marshall used to arguethe Brown v. Topeka BoE case. They were the studies were the black childrenidentified white dolls as better than the black dolls. He used those studies toargue that the black children's self-esteem(self-conceptin today's parlance) was lower than whites.
> This illustrates another problem I have with much (not all)psychological research.
Maybe we should focus on the not all part :-)
> What the studies you cite showed was that black children treated whitedolls as being better than black dolls. From this, it is informally inferredthat therefore black children think that white children are better than blackchildren, and from that we conclude that black children consider _themselves_worse than white children, and therefore that black children have lower self-esteemthan white children.
> All that seems reasonable enough to base a decision on in the terms inwhich most decisions are actually made, but it's not enough to call it science.The problem is that _plausibility_ is substituted for _proof_. If the proposedlogical connections were actually valid, then the final conclusion would bevalid. But not otherwise.
Only a poor scientist would call this proof. Runkel is correct that weshorthand our writing often, but skepticism is a value communicated to allresearchers. We our much less accepting than Runkel or you seem to think. Iwould begin your next sentence with "Strong supporting evidence..." instead ofproof, 'cause I do at least strive for accuracy in my writing.
> Proof in this case would amount to finding some independent way ofdetermining the black children's degree of self-esteem.This is the only way to validate this procedure for determining self-esteem.Before this procedure is applied to a general population, we must establishthat when black children say white dolls are better, they are also experiencinglow self-esteem.We can see which dolls they choose, but how can we see their self-esteem?What we're trying to establish is that the choice of dolls, which is easy tomeasure, is a reliable indication of self-esteem,which is not so easy to measure. If we could establish that connection, wecould then substitute the easy measure for the hard measure in testing ageneral population. So the problem of showing that the test means somethingboils down to measuring self-esteemin some direct and independent way.
But here we are stymied. We can't see the self-esteemdirectly, unless we happen to be the particular black child under study, andhave the sophistication to think in such terms. We can't just ask a three-year-old child how much self-esteemhe or she feels; that's a meaningless term to a three-year-old.
The only recourse is to try to devise a completely different test for self-esteem.We see how black children behave when getting into lines with white children.We see how "assertive" they are in competing for toys with white children. Wesee what roles they play in pretend games. But in each case we run up againstthe same problem: there is no direct measure of self-esteemavailable for validating any of the tests. In each case, we simply assume thatthe behavior we can observe is indicative of self-esteem. So what do we end up with? A series of tests which are validated against eachother, and none of which is actually known to indicate self-esteem.This thing called self-esteemremains hidden from observation; in fact we don't know what any of these testsis indicating, if anything.
This problem became evident to me when I began taking psychology courses incollege. And it led me to wonder about something that psychologists didn't seemto care to think about: did the things that these tests supposedly measureexist at all? In the above example, I would have asked, is there really such athing as self-esteem?Is this something that people have? Or is it just a way of talking?
I was asking that question several paragraphs before. Part of thatskepticism we scientist have. If your psych teacher did not mention it, thatwas his/her bad.
> Now I think I would say that yes, there is such a thing as self-esteem,but that term is only a rough indicator of what is involved. To get any closerto a useful understanding, we have to ask what "self" means, and what "esteem"means, so we can see some sort of process going on. Obviously, self-esteemis a perception that can be had only by the person involved. It is a perceptionof one's own characteristics. And this perception, to have any value, must becompared with a reference level for the each state of those characteristics onehas chosen as a target. If you perceive yourself as less than you want to be ineach relevant respect, then you are probably experiencing what people call lowself-esteem.
> But there is another side to this, which becomes visible only when youuse the PCT model. What if your perception of your self matches the referencelevel with respect to each variable involved in characterizing yourself, butyou have set your reference levels very low for characteristics that otherpeople usually set to a high level? Your perception of self-esteemwould then be exactly what you have chosen in each respect. It would not seemlow to you, but others would still judge you as having low self-esteem.You, as a black child, would go behind the white children in the line, handover a toy to a white child without protest, and play the black victim ratherthan the white victor in pretend games. Perceiving yourself doing these thingswould fit perfectly with your reference levels for how you should be.
> How could such reference levels come to be set so low? Obviously,because these are the settings that work, all things considered. If you areraised in a society that punishes you for assuming equality or ascendance inthe company of white people, you can avoid that punishment by choosing ways ofbeing that do not cause punishment. You remember how you were when you weregetting along as well as possible, and you pick those memories as referencelevels. This does not require any thought; all it requires is reorganizinguntil the pain is minimized.
I tend to agree with your analysis. It is an excellent example of why PCTis so important. It also shows why self-reportquestionnaires would not have been appropriate for they can only hope to get atdiscrepancies, and not necessarily very well. It also shows why the doll methodis reasonable, if not perfect. But I think what you miss is that the researcherhad some vague notion that the level of SE was more critical, not theperception of low SE. I think he (I believe it was a he) did a pretty good jobgiven he was probably not exposed to your model. The darkness that psychologistare stumbling around in is not as dark as you think.
[you then describe the test...]
> Then you can start looking for common factors, for variables thatindicate common problems, for deviations of individuals' reference levels fromthose of most of the group. You look for indirect indicators that show highlyreliable correlations with either felt lack of esteem or low levels ofaspiration for esteem. Now you have a direct measure of self-esteem,as you have carefully defined it, for a large enough group of individuals toallow validating other kinds of tests that are easier to apply. Of course theseother tests will NEVER yield the kind of reliability that the basic study ofone person at a time will yield, and no important decisions should be madeabout any one person on the basis of such tests. But such tests can be used asthe basis for modifying policies that apply to many people or for assessing theoverall effectiveness of a program, a teacher, or an administrator.
> Compare where we started with where we ended. We began with childrenplaying with dolls, and we ended with a study of individuals aimed atdiscovering how their self-conceptscompare with their desires, and what those desires are. We started with a vaguecommon-languageterm, self-esteem, and ended by talking about specific variables associated with self-perception.Which approach is going to yield solid knowledge of how people perceivethemselves, how children learn to perceive themselves? I submit that the answeris obvious.
> It's much, much easier to hand out a set of questionnaires to 30people than it is to sit down with each one of them for hours and hours,learning how each person ticks. There are pressures of time and competition andpride and funding that make psychologists look for quick and easy ways to getinformation about important aspects of human nature. But the results reflectthe amount of effort and thought expended. If anyone wants to conductpsychological investigations in a way that will yield reliable knowledge abouthuman nature, there is really only one way to start: one person at a time.Without that kind of basic understanding, all generalizations are fuzzy andempty and hardly worth the trouble.
Have you done the analysis you suggested? Have you looked for SE using theTest? If you have, it is because of all the psych research on it that clued youinto it. If you have not, why do you believe it is a dimension of self-perception?Because psychologists, like clinicians, have interviewed people...that isinformally applied the test? (I am guessing of course -I do not know you mind.)
Is it a dimension of everyone's self-perception?We cannot definitively answer that unless we give everyone the test. But isthere some critical mass? That would be an empirical approach. What if we runmodels that showed complex systems with certain properties like those ofhumans, always formed a reference signal that acted somewhat like SE isdescribed? This is my guess. SE is slightly different for everyone, but theprotection of one's sense of self is a critical feature of a self-organizing,self-aware,system.
This was not terse, but important I think. I hope you agree.
Later Jeff
------------------------------
Date: Thu, 22 Sep 1994 16:14:10 -0400
Subj: Re: Locke, all or nothing
[from Jeff Vancouver 940922.1610]
On Thu, 22 Sep 1994, Tom Bourbon wrote:
> I read your post, Jeff. This will be terse.
> Read [From Bill Powers (940921.0400 MDT)], in which you will see onceagain the way PCT modelers go about the business of trying to understand humanbehavior and to help them. After you read those ideas one more time, I hopeyou will give some thought to abandoning your ad hominem style when you are onthis net.
If you think that I apply an ad hominem style than we do havemiscommunication. I have a problem with the way you refer to non-PCTers,not with you per se. I apologize if you see that as an attack on you. That isthe last perception I am trying to maintain.
BTW, I did (non-tersely)reply to Bill P.
Later Jeff
Date: Thu, 22 Sep 1994 19:57:52 -0700
Subj: Simply PCT
[From Rick Marken (940922.2000)] Jeff Vancouver (940922)
> When Locke and others are collecting data and then trying to makeinferences of the underlying causes, they are doing inductive research.
We use both "inductive" and "deductive" research in PCT; we favor neither.The combination of the two is called "science".
Locke and others are, indeed, "collecting data and then trying to makeinferences of the underlying causes" but their data (as we know from many PCTstudies of human behavior) simply represent averaged side effects of thecontrolling done by groups of people. Their inferences of the underlying causesof this behavior are bound to be wrong; that was the point of my "Blind men andthe elephant" paper. Locke et al are playing a losing game. This is all wellknown and clearly demonstrated (with .99 precision) in numerous PCT demos andexperiments. Locke et al are studying irrelevant side effects of control andthey don't know it. It's not like we've been trying to hide this fact from them(or anyone else). If Locke et al were astronomers they would still be lookingfor new epicycles. There's not much one can say to people like this except"goodby and good luck".
> you also note that it is difficult to get tenure doing the PCTthing
I did it. Tom did it.
> So what does one do? 1) Give up convincing the scientific communityif they can afford to, or 2) try the convince some of them. You have chosenthe former because you can afford to (or perceive that you can, which I do notquestion).
I don't think we have given up trying to convince the scientific community(or anyone else) that PCT is the best scientific model of purposeful behavior.The only thing I've given up is the expectation that I will be successful in mylifetime.
> All I am asking is that you do not make my job harder than it alreadyis!
We don't want to make your job hard. All we want to do (here on CSG) isprovide an accurate representation of PCT. If you want to get tenure by sellingwhat amounts to a false representation of PCT to journal editors and yourtenure committee, that's up to you (or, as my grandpa would say, gezunterhait). I would certainly understand why you might want to do it, but pleasedon't ask me for help.
Truth be told, I don't think it is beyond one's capability to get publishedor to get tenure doing PCT. Bill Powers described an approach to individual,applied PCT that could certainly be turned into an acceptable research programby someone who understands the basics of PCT.
I appreciate your concerns about getting tenure, but they are not the sameas the ones that drove me from academia. I didn't have a problem because peoplerejected PCT; I had a problem of conscience because I didn't want to teachconventional psychology but, at the same time, I didn't want to be unfair to mystudents by teaching PCT in a non-PCT curriculum. But other people have been able to teach an accurate version ofPCT and maintain tenure and their academic position as well. Dick Robertsoncomes to mind as one stand-outexample. Clark McPhail and Chuck Tucker are sociologists who seem to have beenable to do it. Gary Cziko has managed to as well. I'm sure there are severalothers. Perhaps you could ask these people (those who are on the net) how theyhave been able to survive as academic PCTers in a non-PCTworld.
Best Rick
Date: Thu, 22 Sep 1994 19:21:39 -0600
Subj: PCT and psychology
[from Mary Powers 9409.22] Jeff Vancouver 940921
Well, for heaven's sake, what provoked that outburst?
You _could_ argue that we are as guilty as Locke for deliberate ignorance,etc., but you would be arguing from false premises. It is because we _have_read Locke and Lord and Carver and so on that we feel entitled to comment ontheir failure to grasp control theory.
You _could_ argue that developing a model is different from developingapplications. It's more fundamental, for one thing. How can you develop goodapplications if your model is a) absent or b) wrong?
Apparently the constructiveness or destructiveness of what we have to sayis debatable (since you are debating it). Is knocking down a theory thatdoesn't work really a bad thing to do? We have wasted a _lot_ of time lookingat what these researchers do, and we are unimpressed -not ignorant.
Finally, although PCTers spend a vast amount of time on this net teachingand explaining, that is not the sole function of csg-l.This extremely small corner of cyberspace was established so that PCTers, whoare scattered rather thinly over the planet, could talk to each other aboutcontrol theory. Listeners, questioners, students, and critics are welcome, BUT,the basic premise of csg-l is that PCT is a viable, productive theory of the organization of livingsystems. In some quarters, that qualifies it, operationally, as a reasonableapproximation of "the truth" (whatever that is). Since most of psychology hasignored control theory, and the rest misunderstands it, we do indeed have a"corner" on it. But it's not because we _want_ to have a monopoly of it, whichis what a corner means -it's more like a sack of kittens. We can't seem to give it away. Meanwhile,this is the only place where PCTers can talk freely to one another and expresstheir feelings about what, in the context of control theory, looks like badpsychology and worse science. And if we are so terribly, terribly thoughtlessas to say, on the net we established, for our own use, that our theory isbetter than anything else going in psychology, then what exactly is yourproblem with that?
Mary P.
Date: Fri, 23 Sep 1994 10:01:40 CST
Subj: Re: Locke, all or nothing
Tom Bourbon [940923.0946] >[Jeff Vancouver 940922.1610]
> On Thu, 22 Sep 1994, Tom Bourbon wrote:
>> . . I hope you will give some thought to abandoning your adhominem style when you are on this net.
> If you think that I apply an ad hominem style than we do havemiscommunication. I have a problem with the way you refer to non-PCTers,not with you per se. I apologize if you see that as an attack on you. That isthe last perception I am trying to maintain.
Understood, and I apologize for my remark. After sending it, and afterreading some of the subsequent posts from you and others on the Locke thread(which includes ideas about research, theories and statistics in behavioralscience), I have been wondering what led me to interpret your earlier remarksas "ad hominem." I have a few ideas on the topic and will say more about thatin a later post, after I've thought some more about it, and after I finish twoproposals to the local Human Subjects Committee --for studies of PCT tracking tasks as quantitative measures of neurologicdeficit.
Remember that I am not saying bad things about non-PCTresearchers, as people. However, I certainly do criticize, with passion, thenear-universalpractice of using statistical mush (mush in fact, not merely in my opinion) tosupport all manner of "plausible" psychological constructs.
The deadline for proposals draws near, and I have reams to write.
More later, Tom
Date: Fri, 23 Sep 1994 14:22:32 -0700
Subj: Decisions; self-esteem
[From Rick Marken (940923.1400)] Jeff Vancouver (940922) --
> I think the question of DM is the question of which reference signalis passed down from the control unit that has as a reference signal "spend thisdime."
What you seem to be saying (given my understanding of HPCT) is that DMrefers to the output of the control system that is trying to produce theperception of a "spent dime" --the perception specified by the reference signal. The output of this controlsystem sets (varies, actually) the reference inputs to the lower level systemsthat are the means by which the perception of a spent dime is produced. To seehow this process works, get a copy of my spreadsheet model, which is on DagForssell's PCT demo disk. There is no DM (other than subtraction andintegration) involved in hierarchical control.
> I have no trouble (as a working hypothesis) with "nothing beyondperceiving the elements of the situation and applying an algorithm that yieldsan answer based on the elements and logic or arithmetic." This strikes me as areasonable description of the DM process. But the next question is whichelements are perceived?
A question that can only be answered by doing The Test.
> What is the nature of the algorithm? That is, what are theparameters?
A question that can only be answered by building a hierarchical controlmodel.
> Does the organism need to go into thinking mode to access parametersfor the algorithm? Can individuals control for perceptions that cause (even intheir own world view) other higher-orderperceptions to go out of line (i.e., why are some perceptual elements notrepresented in the algorithm)? Where is the algorithm?
All questions that can only be answered by appropriate testing andmodelling. When you start to do this, you will find that some of thesequestions (like the one about "thinking mode") are probably not really thatwell conceived.
> I think that one of the elements in the algorithm, that is in theoutput function is an estimate of the probability of attaining a referencesignal (what Bandura calls self-efficacy).
Why don't you try to implement this in a model? I think what you are sayinghere makes no sense but I don't think you will believe me until you actuallytry to build a model of the kind of behavior you have in mind. One reason formy suspicion that this makes no sense is that a reference signal is not"attained"; it is determined. Perhaps you meant that the algorithm estimatesthe probability of "attaining" the perceptual signal --ie. getting it to match the reference signal. This sounds like a feedforwardprocess -where outputs are selected on the basis of estimates of the likelihood thatthey will produce the desired results. People do go through these predictionexercises in imagination but there is no evidence I know of that suggests thatsuch predictions are an important component of on-going controlling.
> I believe the value for that estimate is partially a function of self-conceptslike self-esteem
What is the data that leads you to this belief; what does the model looklike that implements this idea.
> I believe Bandura and others have shown us the reasonableness of thesepropositions.
To paraphrase George S. Kaufman (who, on his deathbed, said "Dying is easy;comedy is hard" --a favorite expression of my comedy-writerfather-in-law,who died) belief is easy; knowledge is hard. What leads you to believe thatBandura has shown us the "reasonableness" of these propositions. I judge"reasonableness" in terms of observation (is there evidence of the phenomenonto be explained; if so, it's reasonable) and modelling (does the proposedmechanism actually behave and does it behave like the observed phenomenon; ifso, it's reasonable). I have never seen anything produced by Bandura that metmy reasonableness criteria.
> I believe Bandura's model is flawed
A model that neither works nor predicts the phenomenon it is designed toexplain is, indeed, flawed.
> why do you believe it (self-esteem)is a dimension of self-perception?
I'm willing to call some perceptions I have "self esteem". That is why Ibelieve that there might be such dimensions of perception. Whether anyone elseis controlling such perceptions is a job for The Test.
> Because psychologists, like clinicians, have interviewed people...thatis informally applied the test?
Yes, hypotheses about controlled variables can certainly come from thesesources. They can even come from looking at conventional psychology data.You've gotta start somewhere. The "doll" study (as Bill pointed out) can be thestart of a systematic attempt to test for the variables that these kids areactually controlling.
What you should be clear about is that conventional psychology data(including that from Locke et al) is neither a better nor a worse source of_initial_ hypotheses about controlled variables than simple, causalobservation. But you can't find out what people are actually doing unless youdo The Test --that means iteratively testing (and eliminating) hypotheses about what a personis controlling. This is the step that conventional psychologists have nottaken.
If you and Martin Taylor were only looking to the conventional literatureas a source of hypotheses about controlled variables (hypotheses that arealmost certain to be wrong, but are at least a start) I would have nocomplaint. But you guys seems to think that conventional psychological researchhas already discovered something about behavior. It is possible that there issome evidence for controlled variables lurking somewhere in conventionalpsychological data --maybe people do control a variable that we could all agree could be called "self-esteem"; maybe people do control the equivalent of the energy output of a bandpass filter (in signal detection studies); but we can't know this until westart testing hypotheses about what variables individuals control --that is, by doing The Test. Conventional psychologists don't do The Test sothere is no possibility that they have discovered anything about purposefulbehavior.
Best Rick
Date: Sat, 24 Sep 1994 11:01:37 -0600
Subj: Data; decisions; tempting proposal
[From Bill Powers (940924.0829 MDT)]
Rick Marken, Jeff Vancouver --
Jeff:
> I believe Bandura and others have shown us the reasonableness of thesepropositions.
Rick:
> To paraphrase George S. Kaufman (who, on his deathbed, said "Dying iseasy; comedy is hard" --a favorite expression of my comedy-writerfather-in-law,who died) belief is easy; knowledge is hard. What leads you to believe thatBandura has shown us the "reasonableness" of these propositions.
I'm beginning to get the feeling of watching trench warfare here. One sidelobs a general reference to unnamed literature at the other; the other lobsback statements that nothing in that literature is any good. To one who hasread little of the literature, the whole argument is content-free.It's like hearing two people arguing about what a friend of theirs said,without ever mentioning what was said.
If we're going to make any pronouncements about Bandura's scientificfindings, how about someone summarizing the data? Even just one experiment. Isee nothing terribly unreasonable about saying that people will choose goals onthe basis of their assessment of their own ability to reach them, if that iswhat "self-efficacy"means. That is just what common sense and personal experience would say. Iwould like to experience space flight, but I don't think I could make itthrough the astronaut training program at my age, so I have regretfully deletedthat goal from the list of those I actually try to achieve.
The only question I have is whether we can take this informal sort ofstatement as having been formally established as a scientific truth. Aseveryone must be aware by now, I am concerned about general statements that areuntrue of a substantial part of a tested population, yet which are treated asif they were true of everyone. I'm concerned with the truth-valueof such statements, not so much one statement at a time but as these statementsare used in systems of reasoning that depend on the simultaneous truth ofseveral or many statements. I don't need to go through that argument again(unless asked by someone who missed it).
So if Bandura has made statements about the relationship of self-efficacy to other factors, I would like to know, aside from what the statementswere, how well his data support them. How many individuals counted as positiveinstances of the statements, how many as negative instances, and how many wereindeterminate, as a fraction of the total population studied? Before I acceptexperimental evidence as supporting any general statement about people, I wouldreally like to know that essentially all of the people in the study behaved ina way that supports the general statement. I vastly prefer statements thatapply to every individual.
The reason I ask is that if Bandura or Locke or anyone else has noticedreal phenomena of goal-seekingbehavior, these phenomena should be grist for the PCT mill. If we think thatthe phenomena are interesting but that the method of establishing them waslacking in some respect, then we should be able to design a PCT experiment toinvestigate further, and carry it out. Even if Bandura and Locke aren'tinterested in the result, some people might be interested.
-----------------------------------------
Jeff:
> I have no trouble (as a working hypothesis) with "nothing beyondperceiving the elements of the situation and applying an algorithm that yieldsan answer based on the elements and logic or arithmetic." This strikes me as areasonable description of the DM process. But the next question is whichelements are perceived?
Is any general answer to that question possible? There are so manydifferent bases for "decision making" and so many algorithms, ranging frommathematical to crazy, that it would be hard to come up with "the" elementsthat are perceived when decisions or choices are made. Better to start withexamples, and see what is generalizable and what is not.
When Mary goes grocery-shopping,she decides which product to buy, for certain products but not all, on thebasis of the price-per-weightsticker that grocery stores now carry. Of course quality also figures in; carobcookies are cheaper than chocolate ones, but she knows that I turn my nose upat carob. There are many variables to consider, but the overall result ispretty much ordained by the weights given to the variables and the logic thatis applied. The final result in each case is to pick up this package ratherthan that, from a set of similar products. As Mary has said on the net, theonly time she thinks of this process as "decision-making"is when there is a conflict: when the data and the reasoning do notautomatically select just one result, and there are several possibilities withno basis for choosing among them.
My shopping methods are very different. I know what kinds of cookies Ilike, so I scan along the shelves until I find one of them, and I buy it. If Ihad scanned further I might have found another kind I like better, but my loopgain for cookies is not very high so that doesn't bother me. If I don't findany cookies I like I don't buy any.
Mary no longer allows me to go shopping with her. I am in and out of therein 10 minutes, about 50 minutes before Mary is ready to go. She considers a lotmore variables than I do, and likes to have in mind what all the choices arebefore she lands on one. And she buys groceries for a lot less money than Iwould.
Now, both of us go into a store, pick items from a collection of similaritems, and buy them. Just looking at the behavior, an onlooker would say thatwe are both making decisions and therefore must have a Decision Maker inside ofus. But there is not very much in common about our methods for makingselections. I do it by setting up a few target items and scanning along theperceived items until there is a significant dip in the error signal, a serialsearch process coupled with monitoring for a reasonable degree of match withany of several reference signals. Mary does it in parallel, keeping memories ofavailable items in mind, assigning weights (I presume) to them based onmultiple characteristics of each item, and looking for the item with thelargest score (I suppose). These are both methods for "making decisions," butthe algorithms are clearly very different (hers requires a lot more brain powerand time).
Also, if you look at the detailed processes involved in either method,there isn't any one process that amounts to making a decision. Making adecision --that is, actually picking up one of the items --is the outcome of applying a lot of processes, none of which is a decision-making process, and any one of which could be used for all kinds of otherprocesses as well.
I'm working on a paper that contains a little essay on two kinds of models(descriptive and generative) and the problems that arise from trying to mixthem. I think that the concept of decision-makingbelongs among descriptive models, while the PCT explanation of decision-making,which does not include any explicit decision-maker,belongs among generative models. Decision-makingis a phenomenon; the processes that create the appearance of decision-makingconsist of underlying functions none of which is a decision-makingprocess. Note that in the control-system model itself, there is no box labeled "controller." Control is a processat the descriptive level which arises from connecting functions in ahypothetical underlying reality none of which can be called "controlling." I'llpost this essay when it's in a little better shape. -----------------------------------------------------------------------Francisco (Arocha?) (940924.0325) --
> I've always been very suspicious and critical of conventionalpsychology (Myself, with some friends, wrote a book about it)
Full reference, please!
> Forget about trying to convert the children of the Lockes and theBanduras of this world. They have already made up their minds. They alreadyknow what science is. They already know what psychology is. Let them live theirdelusions.
A very tempting proposal. After all, how much time do they spend patientlytrying to understand _us_? I have come very close more than once to reachingthis conclusion. The only thing that stops me is knowing that practicallyeveryone who is now on our side came from conventional fields. There are alwaysa few who get the message. The problem with appealing ONLY to dissenters isthat you get not only the free spirits and the independent thinkers, but allthe crazies, too. At least the people in conventional fields have someexperience with disciplined thinking, even if what they think about isn't sogreat.
Psychologists seem to have a great deal of loyalty to psychology --my science, right or wrong. For me this is a constant conflict that makes itdifficult to find a direction and pursue it. It doesn't seem possible to have asimple discussion of things like statistics without arousing instantopposition, from those who happen to use it a lot. Actually, it really doesn'tmatter what the subject is; if PCTers offer even an implied criticism, thedefenders of the faith will oppose it even before they start thinking about it.If we must continually worry about what we say for fear of threateningsomeone's beliefs, we will always be diverted into unprofitable arguments andaway from our own pursuit of truth, beauty, science, and chocolate-chipcookies.
But the other side of the coin is not only seeming to become a cult, butactually doing so. If we ever start thinking that we are no longer subject tothe requirement of persuading others through logic and demonstration, we willstart acting exactly like our opponents, assuming automatically that we areright and everyone else is so obviously wrong that debate is no longernecessary. After that, Saint Bill is just around the corner.
I know what we should do! We should train young people who haven't beencontaminated yet by the behavioral sciences and make sure they understand PCTfrom the ground up. Then we should require them, right after their training, toput on blue suits, white shirts, and neckties, and go in pairs from door todoor all across the country, carrying copies of B:CP and equipped with openerssuch as "Do you know that Bill loves you and your control systems?" We couldprepare little pamphlets ....
I guess that this conflict will just go on, until we find a way out of itthat doesn't lead to a worse situation. Maybe the answer will come out of ourattempts to form a Center for the Study of Living Control Systems. We'll thinkof something.
-----------------------------
> As for Rick M., well, I like him. Stop complaining about his "badmanners". He has done great experimental work and is the first defender of thebasic ideas of PCT. Besides, for me, he is to CSG-Lwhat chilies is to mexican food. It may make me sweat, but I sure likeit.
Yes, people do tend to forget that Rick has done some of the bestexperiments in PCT-world.When he demands that others come up with some real evidence, he speaks as onewho has DONE it. When people argue with him, I think they're obligated toproduce results of equal quality and specificity before they're entitled to betaken seriously. I am getting mighty tired of people putting their 0.3correlations up against his 0.95+ correlations. It gets ridiculous.
Thanks for spicing up the conversation.
Best to all, Bill P.
Date: Thu, 29 Sep 1994 16:12:20 -0400
Subj: goals, models, and salutation.
[from Jeff Vancouver 940929]
This is going to be my last post for awhile. I cannot maintain myinvolvement and still maintain what I thing is sufficient progress towardtenure. Also, perhaps more importantly, I echo Bruce Nevin's position. I do notknow how to model and it seems I need to learn to continue the debates I amparticipating in. Tom, Mark, and Bill (and probably others) cannot be expectedto respect my position until I respect theirs. I don't think the relevance ofKernan & Lord's data is explainable until the "lobs" we are making can bereduced. That will happen by either me adopting your mind set or me expandingyour mind set through your mind set (i.e., modeling).
I think Bruce Buchanan [940927] made some excellent points (too bad aboutthe behavior/perception mix-up,but it is a closed loop -I am very forgiving). I just gave my conference presentation from August (atthe Academy of Management) locally (NYU I/O Psych colloquium) the other day.The feedback was most informative. Basically, I need to show how PCT helps theindividual researcher do their work. I cannot argue that the science needs thismodel as I have been, only that you (i.e., some individual scientist) needs it(or it will help the scientist do his/her work). Just like with the individualson the net, I need to do that from their perspectives/beliefs etc., by showingthem where their believe/perspectives are limiting their progress and providethe PCT alternative. (Bruce B.-send me your address and I will send my revised paper that I think you willappreciate -it is almost done.)
I have learn a lot from this interaction and loath to give it up, but thetime drain to just too much. I have one specific, practical suggestionregarding the goals/future of CSG-L.Perhaps a short prefix in the subject field can clue readers and lurkers intothe domain of the message. For example, posts on models having to do withlevels 1-3begin M1-3:.Philosophy of science post begin PS: etc. A post that lists the agreed uponprefixes could be accessible or made part of the intro post. It is just athought.
Two more notes, I saw Locke's paper in Applied Psychology this lastweekend. It was appalling. You [I cannot remember who] did not quote the trulywretched sentences where Locke says:
"...I received a letter from a prominent advocate and founding father ofcontrol theory .... This theorist was not exactly clear on just what controltheory really was..."
This is from the guy who refused to read the book, which I assume Mary (orBill) cited in their letter to Locke. Correct me if I am wrong. BTW, I amtrashing Locke and Bandura in my revision (I am side-swipingLord)
2nd note is a quote from my wife after describing the situation and the netwith her:
"It sounds like Locke should be shut out, Lord is not God, and that Powersdefines right."
I thought that summed it up pretty well.
Much Later Jeff
Date: Fri, 30 Sep 1994 10:36:51 -0600
Subj: Replication;
[From Bill Powers (940930.0655 MDT)]
Jeff Vancouver (940929) --
I'll send this direct to you also in case you signed off already.
We'll be sorry to see you go, as you are a link to the mainstream world. Ihope your experiences with PCT so far have not spoiled anything for you!
I hope you had a chance to read Chuck Tucker's wonderful post of 940929.[PCTDOCS\REPLICAT.ION] Chuck is in much the same position you are in, notbeing a programmer and not having had any previous experience with modeling. Heis the perfect example of how non-technicalCSG members work with those who can write programs and set up models. Chuckdecided that he would simply run our models and get some experience with howthey work; as can be seen in his post, this has led to a tremendous increase inunderstanding of PCT, and the development of abilities to see the possibilitiesof PCT research of many kinds. He can now spot previous research in which therewere approaches compatible with PCT even if their data and analyses were notpresented in a way we could use directly. This is going to lead, eventually, toa whole methodology that sociologists can use to apply PCT in their own fields --without ever requiring them to become modelers or programmers.
I sympathize completely with your need to pursue tenure; what a world! ButI hope you will stay in touch with Chuck, because he can help you see how toapply PCT without having to be a techie.
Best to all, Bill P.