···
On Wed, Oct 21, 2015 at 4:20 PM, Bruce Nevin bnhpct@gmail.com wrote:
RM: DRoP … is aimed at showing people how to do PCT research. But I think it might be a little too abstract, even for people who are already involved in psychological research. I think what I will have to do is write another book that is less formal and includes lots of suggestions for the kind of research one might do to test PCT; a book that puts a little more “flesh on the bones” of PCT research.
BN: Music to my ears! I have been hoping you would take this up. Yes!
RM: Actually, I’m not sure I know how to do it. But I’ll try. I’ll need all the help I can get.
BN: In laboratory experiments, input, output, and disturbance quantities can be precisely measured, and extraneous disturbances can be perceived and controlled to zero by the experimenter. Methodological help is also needed for naturalistic work in conditions that cannot be so precisely controlled, for practical reasons and sometimes for ethical reasons. A huge proportion of such situations involve collective control, especially where elements of the environmental feedback paths are collectively controlled. Kent’s chapter for LCS IV lays out the issues and principles brilliantly. (I’m reading part 2 now.) But methodological help is needed with research into collective control.
RM: I am available for methodological consulting; special rates for CSG members;-)
RM: If we can get a critical number of researchers doing PCT … and … providing models for other researchers to imitate, I think we will then see PCT start to have a greater influence in the life sciences.
BN: People’s interest perks up with socially significant variables. Baseball catching has greater intrinsic interest than pursuit tracking, but its grounding in the game of baseball only brings it just over the horizon of social significance. There is enormous need for understanding our participation in collective control, and for understanding social change, because there is a great deal of social change going on, and much more of it in months and years shortly to come. We know that a grasp of PCT brings with it a radical change in our perception of rewards and punishments and the nature of motivation. Such change is needed in our institutions and social norms. An understanding of collective control, and how pervasive it is, and how necessary, similarly shrinks the narcissistic fantasy called ‘independence’ to a perspective that is more modest and more capable of autonomous control. (The difference between independence and autonomy.)
RM: I agree that sexy, socially significant research could attract people to PCT. But I’m still more interested in seeing more of the “balls rolling down inclined planes” kind of research done on PCT, the kind that really got physics off the ground. I think that one of the things that has kept my experimental psychologist peers – many of whom, like my PhD adviser, are very smart and mathematically compentent – from getting into PCT research is that they don’t know what such research would even look like; they don’t know what questions to ask or how to answer those questions. They have been trained (as I was) to ask questions about the perceptual/cognitive processes that intervene between input and output. My book would have to “turn their heads” away from this perspective and show with examples that have obvious real world implications – that’s what I meant by putting “meat on the bones” – how to go about doing such research.
RM: Maybe the book will be about how to use the basic findings from the tracking task as the basis for studying more “real world” examples of behavior. Anyway, another research book, hopefully one that more clearly explains the relationship of PCT to “real world” examples of behavior, is on my “to do” list.
Best
Rick
/Bruce
–
Richard S. Marken
www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble
On Wed, Oct 21, 2015 at 4:15 PM, Richard Marken rsmarken@gmail.com wrote:
[From Rick Marken (2015.10.21.1315)]
On Tue, Oct 20, 2015 at 2:00 PM, Bruce Nevin bnhpct@gmail.com wrote:
BN: Feedback and feedforward seem to have been reduced to synonyms for afferent and efferent, with no sense of a loop. This terminological bushwhack needs to be recognized and countered explicitly in any presentation, publication, or conversation. (Same with controlled variable vs. the several senses of ‘control variable’.)
RM: I agree. But I think the best way to “combat” this kind of misunderstanding of perceptual control to reduce the time spent countering these ideas and increase the time spent doing research (and application) based on PCT. I say this as one who is perhaps the most guilty of having spent way too much time countering these “open loop” ideas (in presentations, publications, and conversations) and far too little time doing PCT research.
RM: The question about the future of PCT made me realize that one of the big impediments to the adoption of PCT (by researchers anyway) – one that I hadn’t thought of before – is that people who might be interested in doing such research don’t really know how to do it. People trained in experimental or physiological psychology have been shown the kind of research that is done in the field; they have lots of models for what counts as a good research. But there are few such models for doing research based on PCT.
RM: PCT research is going to be very different than the kind of research that is familiar to conventional experimental/physiological psychologists. The questions a PCT researcher asks about behavior are completely different than those asked by a conventional researcher. So when a conventional psychologist gets interested in PCT he is either put off by the fact that he doesn’t know how to study it or assumes that the way you study it is the way you have been studying behavior all along and the result is the Carver/Scheier approach to PCT.
RM: DRoP (which you so kindly reviewed) is aimed at showing people how to do PCT research. But I think it might be a little too abstract, even for people who are already involved in psychological research. I think what I will have to do is write another book that is less formal and includes lots of suggestions for the kind of research one might do to test PCT; a book that puts a little more “flesh on the bones” of PCT research. If we can get a critical number of researchers doing PCT(what that number is, I don;'t know, but I’m pretty sure it’s > 3) and, thus, providing models for other researchers to imitate, I think we will then see PCT start to have a greater influence in the life sciences.
Best
Rick
There is a third sense of the term ‘internal model’, along the lines that Martin was talking about. The structure of the control hierarchy–the interconnections and weights of connections, the perceptual input functions, the reference input functions, and tje output functions–constitute a kind of model of the environment in which that control hierarchy controls successfully.
These folks seem to be trying to implement the ‘cognitive maps’ kind of notion promulgated in Cognitive Psychology: sensory data come in, are correlated with cognitive maps, ‘information processing’ takes place, and the computer in the skull issues commands to muscles.
/BN
–
Richard S. Marken
www.mindreadings.com
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble
On Sat, Sep 13, 2014 at 2:58 PM, Richard Marken rsmarken@gmail.com wrote:
[Rick Marken (201409.13.1200)]
Rupert Young (2014.09.13 13.00)–
RY: What my question was leading to was, is the PCT structure of
reference signals ‘internal models’, as the paper authors would
recognise them?
RM: I think “internal models” can refer to two different things. One is what Adam referred to: models of the physical environment that allow “feedforward” predictions of the effect of system output on the controlled variable. This kind of internal model can improve control somewhat in a completely predictable (and, thus, artificial) environment but not so much in a real one where disturbances are upredictable. This is why, as Adam puts it succinctly “Simulations [with internal models:RM] run fine, robots worse”. There is really nothing in PCT that corresponds to this kind of “internal model”.
RM: The other kind of “internal model” corresponds to a higher level controlled perception in PCT. Although I haven’t read the Saunders “Feedforward and prosthesis control” paper carefully, my impression is that the improved gripping with vision results from the subjects’ ability to control a higher level visual variable that could not be controlled when the prosthesis was out of sight.
RM: The desire to show that control results from output calculation (feedforward) is baffling to me. But it’s there and it makes it very difficult to get people to pay attention to PCT, particularly because (as Adam also notes) "the published criticism [of the calculated output approach to control:RM] comes from the Equilibrium Point hypothesis (EP) crowd (Feldman, Latash, …) " which, as Adam politely notes “doesn’t work very well” (since it is not a control model:RM).
RM: I think what is common to both the calculated output and EP approaches to control is fear of purposiveness; both aim to “account for” control (apparent purposiveness) in terms of lineal causal laws. Unfortunately, the purposiveness of control is not apparent; it is a fact. Ergo my continued harping on the fact of control. I really think that is the big deal about PCT: PCT shows us that the behavior of living system is control (purposeful), in fact, not just in theory. PCT shows how to demonstrate this fact and how to explain it.
Best regards
Rick
Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble
–