# Imitation vs Modelling **URL:** \ **Category:** CSG1992\ **Created:** [August 12, 1992, 8:25pm UTC](http://discourse.iapct.org/t/imitation-vs-modelling/5557 "1992-08-12T20:25:44Z")\ **Posts on this page:** 1\ **Page:** 1

**Author:** ![Human\_Factors\_Lab](http://discourse.iapct.org/letter_avatar_proxy/v4/letter/h/b3f665/32.png) [@Human\_Factors\_Lab](http://discourse.iapct.org/u/Human_Factors_Lab)\ **Post date:** [August 12, 1992, 8:25pm UTC](http://discourse.iapct.org/t/imitation-vs-modelling/5557/1 "1992-08-12T20:25:44Z")

[From Rick Marken (920812)] Oded Maler (920811) says: > But for engineers (unlike reverse-engineers) this > problem does not exist, because they build the boxes, they know the > I-O ("S-R") characteristics of (some of) the components they put in > the loop, and all your protests against S-R, which might be valid in > the context of psychological experiments, seem like fighting > non-existent windmills. They don't know and don't care about ID-IV > tests and the only arguments you may have with them is whether it is > useful or not to use the language of hierarchical servo loops, > reference signals etc. in order to achieve interesting performance (I > refrain from using the word "behavior"..). The argument I have with roboticists is the same as my argument with psychologists (both groups, incidentally, believe that I am "fighting non-existent windmills"). Both groups are trying to find a way to imitate behavioral appearances rather than model the underlying processes that cause those appearances. It all comes down to understanding the difference between behaving (what we see people -- or artifacts -- "doing") and controlling (what we can't see people -- or artifacts -- doing, viz. keeping perceptual variables at reference levels using whatever means is demanded by the situation). This is much more than a matter of getting people to "use the language" of PCT. It is a matter of understanding what it means to control. I agree that roboticists often use control systems in their efforts to build machines that put on "interesting performances". What I object to is the orientation towards the performance -- rather than the cause of the performance; the control of perception (at least in living systems). The difference to me is between imitation (the current goal of psychological theorists and roboticists) and modelling (finding the underlying organization that produces the appearance). When you successfully do the latter, the behavior of your system will also imitate that of the real system -- but it will do it for the right underlying reasons (and, thus, will imitate the behavior in new situations in which the "imitation only" device will fail). But I don't really care whether the roboticists ever understand the difference between imitation and modeling; their goals (as I understand them ) are simply not interesting to me. I don't need an artifact running around my house bringing me the paper or mopping up the floor. Robotics is interesting to me only as an implementation of a model of real living systems;robots are of value to me only insofar as they further my understanding of human nature. Whether a roboticist succeeds in making his or her toy behave in some interesting way is of no particular interest to me; whether they do it by just imitating behavior or if they do it by developing a model that accidentally includes some of the principles of control, I don't care. I'm not interested in seeing cute behavior (unless I'm at Disneyland). I care about the underlying principles -- not the superficial appearances -- and I see no evidence that roboticists understand the principles of organization of living systems any better than do psychologists. While I could care less what roboticists do (except insofar as their creations can serve as existence theorms for a paricular model of behavioral organiza- tion), I do care what psychologists do and think because they are presenting a model to the world of "how people work" and people assume that these folks know what they are talking about. The fact that psychology is currently based on the wrong model is not just unfortunate in terms of the progress of our understanding of human nature; it is also unfortunate because wrong models of people lead to wrong ways of dealing with people in the the real world. And this produces REAL life problems for all of us. People typically deal with other people in cause- effect terms (as sr devices or output generators). This leads to a focus on overt behavior (the kind that roboticists want to imitate) and efforts to get that behavior to look the way it "should" look. So rather than focusing on what people are trying to control -- and helping them control it so that the means they select are not so disturbing to others -- people tend to try to solve problems manipulatively -- with the attendant social disasters. These social disasters are just not necessary at all. I want a better world for my children and it won't happen unless people start to understand that their basic assumptions about how people work are FALSE. Force won't work; inflexible system concepts won't work, repression won't work; CONTROL WON'T WORK (in the long run) precisely because people ARE controllers. That's the message I care about; and it's why I think an understanding of PCT is so important. If the roboticists already do understand the the principles of control (as you imply) and are just using different language, and if they understand that these principles apply to people as well as robots, then they would be most helpful allies in our efforts to help psychologists (and lay people) understand the nature of human nature. Best regards Rick > **···** > > \*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\*\* > > Richard S. Marken USMail: 10459 Holman Ave > The Aerospace Corporation Los Angeles, CA 90024 > E-mail: marken@aero.org > (310) 336-6214 (day) > (310) 474-0313 (evening)