More Mind Readings

Contents

Foreword, Philip J. Runkel

You have in your hands one of the foundation documents in the construction of a new scientific psychology. It is a firm addition to the steadily growing experimentation concerning Perceptual Control Theory. This book, like its predecessor, Mind Readings: Experimental Studies of Purpose (1992), contains both reports of experimentation and some theoretical comments.

Most books reporting psychological research tell you the conditions or circumstances under which you are likely to find a greater-than-likely frequency of this or that sort of behavior. The reports typically take this form: Twenty-three percent of the people of this sort (or in this condition) did what we predicted, whereas only eighteen percent of the people of that sort (or in that condition) did so.

Marken's experimentation departs radically from that sort of investigation. Marken does not predict acts. He predicts only that the person will do anything necessary
to maintain a perception the person wants to maintain, varying action as necessary to maintain the perception. He predicts, too, that the person will do that continuously.

When I say that Marken predicts "only" the continuous control of perceptual variables, I do not mean that such a prediction is trivial or easy to investigate. To carry out this kind of investigation, Marken observes each subject continuously over a long enough period so that he can obtain thousands of data-points. In this feature alone, Marken's experimentation stands in startling contrast to typical psychological experiments—in which it is common, even usual, for each subject to yield only one or two data-points.

Marken predicts that the person will maintain a sensed variable (a condition or quantity) despite influences from the environment that would alter the variable were the person not there holding the variable steady. (And the person does so, in every case.) That is, the experiment does not "hold constant" all "unwanted" environmental variables. Marken does not even try to guess at what those unwanted variables might be like. He allows the environment to affect the variable the person wants to hold steady so that the behavior of the person in holding the variable steady can be observed.

The investigator's questions are "How can the person do this" and "How can we ascertain whether the person is indeed doing this?" The first question has been investigated in some piecemeal, fragmented ways neurologists, but no one had investigated it in the behavior of the whole organism until the advent of Perceptual Control Theory. The second question has been rejected by many psychologists as impossible to test. Marken, using the radical assumptions of Perceptual Control Theory, shows both how the intentions of the person (of one person at a time) can be ascertained and how the person can be capable of maintaining the intended perception.

You will find here the shape of a new psychology powerful in method and breathtaking in scope.

Philip J. Runkel
Eugene, Oregon
January, 2001

Introduction

More Mind Readings is my second collection of papers describing research based on the control theory model of purposeful behavior developed by William T. Powers (1973). Since the publication of my first collection of papers (Mind Readings, 1992) Powers' theory has come to be known as perceptual control theory (PCT) to distinguish it from other applications of control theory in psychology. The name "perceptual control theory" describes what distinguishes Powers' use of control theory from all others: the idea that purposeful behavior is the control of perception.

The papers in the present collection describe what control of perception means, how it works and what can be learned from a theory that views organisms as controllers of their own perceptual experience. But it is more than a change in the name of Powers' theory that motivates this new collection of papers. In the introduction to Mind Readings I noted that the publication of that collection would mark "... the end of an era in which my research focused largely on what is wrong with current theories of behavior and the beginning of an era in which my research will focus almost exclusively on what is right with control theory." Nearly ten years after making that

A Science of Purpose

William James got scientific psychology off to an excellent start by describing, as well as anyone before or since, the nature of purposive behavior:

Romeo wants Juliet as filings want a magnet; and if no obstacles intervene he moves toward her by as straight a line as they. But Romeo and Juliet, if a wall be built between them, do not remain idiotically pressing their faces against its opposite sides like the magnet and the filings with the [obstructing] card. Romeo soon finds a circuitous way, by scaling the wall or otherwise, of touching Juliet's lips directly. With the filings the path is fixed; whether it reaches the end depends on accidents. With the lover it is the end which is fixed, the path may be modified indefinitely. (James, 1890, p. 7)

A paragraph like this brings a tear to a control theorist's eye, not just for the beauty of the prose, but for the insight into the nature of behavior. James understood the difference between purposive and non-purposive behavior and tried, unsuccessfully, to launch psychology as the science of purposive behavior. He failed because, like other visionaries, his ideas were slightly ahead of their time.

James was trying to start a science of purpose at a time when "purpose" was a scientific profanity (it still is, to some extent, but we live in more permissive times).

From American Behavioral Scientist, 1990, 34 (1), 1-15

The Blind Men and the Elephant

Behavior has been described as a response to stimulation, an output controlled by reinforcement contingencies and an observable result of cognitive processes. It seems like these are descriptions of three different phenomena but they are actually descriptions of three different aspects of the same phenomenon—control. Control is like the proverbial elephant studied by the three blind men; what one concludes about it, and how one tries to explain it, depends on where one stands. It is suggested that the best place to stand is where one has a view of the whole phenomenon, be it elephant or control.

The behavior of living organisms (and some artifacts) is characterized by the production of consistent results in an unpredictably changing environment, a phenomenon known as control (Marken, 1988). Control can be as simple as maintaining one's balance on uneven terrain or as complex as maintaining one's self-esteem in a dysfunctional family. Control is a pervasive aspect of all behavior yet it has gone virtually unnoticed in psychology. What has been noticed is that behavior appears to be a

From Closed Loop, 1993, 3, 37-46.

"Mind Reading":
A Look at Changing Intentions

Methods adapted from control engineering can be used to discriminate intended from unintended consequences of an organism's actions. By continuously monitoring a quantity called the stability factor it is possible to observe changes in intentions, which are not visible in overt behavior.

We often find ourselves wondering what people are doing although their behavior is completely visible. This is a problem for those who consider behavior an objective phenomenon. What is not always obvious about behavior is its purpose. To understand the purpose of behavior one must determine the intentions of the actor. This involves `mind reading,' something current methods in experimental psychology are not equipped to handle. The solution has been to ignore intentions; behavior (what a person is doing) is defined as whatever the experimenter says it is.

Control theory provides an alternative approach. Intentions are viewed as internal models of the desired or reference states of perceptual inputs (Powers, 1973). The purpose of action is to keep perceptions matching refer-

From Psychological Reports, 1983, 53, 267-270. Reprinted with permission of Psychological Reports.

The Dancer and the Dance:
Methods in the Study of Living Control Systems

This paper describes methods for studying behavior when organisms are viewed as living control systems. These methods are aimed at determining the variables that organisms control when they are engaged in various observable behaviors. Controlled variables are perceptual representations of the environment that are protected from the effects of disturbance by the responses of the organism. It is possible to detect controlled variables by applying disturbances to aspects of the environment that might be under control and looking for lack of an effect of these disturbances. This Test for Controlled Variables (TCV) makes it possible to see the "dance" of behavior from the perspective of the behaving organism (the "dancer") rather than from that of the observer, who sees only the dance.

Powers (1989) has suggested that what we call "behavior" is actually a visible side-effect of a process called "control". Control is what organisms do to keep their own perceptual experiences in preferred states. This approach to behavior is based on control theory which views behavior as the control of perception (Powers, 1973). From

From Psychological Methods, 1997, 2, 436-446. Reprinted with permission of the American Psychological Association.

The Hierarchical Behavior of Perception

This paper argues that the coincidental development of hierarchical models of perception and behavior is not a coincidence. Perception and behavior are two sides of the same phenomenon—control. A hierarchical control system model shows that evidence of hierarchical organization in behavior is also evidence of hierarchical organization in perception. Studies of the temporal limitations of behavior, for example, are shown to be consistent with studies of temporal limitations of perception. A surprising implication of the control model is that the perceptual limits are the basis of the behavioral limits. Action systems cannot produce controlled behavioral results faster than the rate at which these results can be perceived. Behavioral skill turns on the ability to control a hierarchy of perceptions, not actions.

Psychologists have developed hierarchical models of both perception (e.g. Bryan and Harter, 1899; Palmer, 1977; Povel, 1981) and behavior (e.g. Albus, 1981; Arbib, 1972; Greeno and Simon, 1974; Lashley, 1951; Martin, 1972; Keele, Cohen and Ivry, 1990; Rosenbaum, 1987). This could be a coincidence, a case of similar models being applied to two very different kinds of phenomena. On the other hand, it could reflect the existence of a common basis for both perception and behavior. This paper argues for the latter possibility, suggesting that perception and behavior are two sides of the same phenomenon:

From Closed Loop, 1993, 3, 33-54.

Controlled Variables:
Psychology as the Center Fielder Views It

Perceptual Control Theory (PCT) views behavior as the control of perception. The central explanatory concept in PCT is the controlled variable, which is a perceived aspect of the environment that is brought to and maintained in states specified by the organism itself. According to PCT, understanding behavior is a matter of discovering the variables that organisms control. But the possible existence of controlled variables has been largely ignored in the behavioral sciences. One notable exception occurs in the study of how baseball outfielders catch fly balls. In these studies it is taken for granted that the fielder gets to the ball by controlling some visual aspect of the ball's movement. This paper describes the concept of a controlled variable in the context of research on fly ball catching behavior and shows how this concept can contribute to our understanding of behavior in general.

The publication of John B. Watson's (1913) Psychology as the Behaviorist Views It signaled the beginning of an era of psychological research dominated by the search for controlling variables; the variables that control behavior. Behavioral psychologists started looking for these variables in the organism's environment. Cognitive psychologists are now looking for these variables in the organism's mind (or

From American Journal of Psychology, 2001, 114(2), 100-110. Reprinted with permission of University of Illinois Press.

The Ethology of Purpose

In their target Behavioral and Brain Sciences article, R.A. and R.T. Gardner (1988) propose a "unified feedforward model of the learning of adaptive and maladaptive behavior" that emphasizes the importance of ethology in operant learning. Unfortunately, the "model" turns out to be little more than a restatement of the observations on which it is based. However, the observations themselves are of interest. First, we are shown evidence that the responses typically used in operant conditioning experiments (key-pecking, bar-pressing) are actually variants of species-specific behaviors evoked and maintained by food and water without any contingency at all. Second, we find that increases in deprivation can lead to decreases in the effectiveness of contingency-based training—surprising if one imagines that deprivation increases the "positiveness" of the consequences that shape behavior. Taken together, these observations are inconsistent with the notion that "arbitrary" behaviors can be trained by making positive consequences contingent on production of the behavior (Skinner, 1981).

Gardner and Gardner (1988) have taken aim at one of the most venerable principles of psychology and fired with both barrels. As a practicing non-fan of reinforcement theory (Marken, 1985) I applaud the goal and feel that some solid hits have been scored. The target article provides strong evidence that the role traditionally ascribed to response consequences is wrong. In operant con-

From Behavioral and Brain Sciences, 1988, 11(3), 460-461. Reprinted with permission of Cambridge University Press.

H. Economicus:
A Perceptual Control System Model of the Economy

This paper describes a model of the economy that is based on Treval C. Powers' (1996) historical analysis of economic data found in the Statistical Abstract of the United States. T.C. Powers' analysis is surprisingly (and unintentionally) consistent with the perceptual control theory model of individual behavior developed by his son, William T. Powers (1973). Powers père views the economy as a circular flow of money between producers and consumers. The behavior of this circular flow can be explained in terms of a perceptual control system model such as that described by Powers fils. A perceptual control system controls a perceptual representation of some aspect of environment that is shaped by the outputs of the system itself. The perceptual control theory model of the economy that is described in this paper controls a monetary representation of aspects of the economy that are shaped by the outputs of the model itself.

Circular Flow Analysis

T.C. Powers' analysis of the economy assumes that the basic economic process is a circular flow of money.

Paper presented at the 2000 Meeting of the Control Systems Group, Boston University, Boston. MA

PERCOLATe:
Perceptual Control Analysis of Tasks

This paper describes a new approach to task analysis based on perceptual control theory (Powers, 1973; 1990; Marken, 1992). Conventional task analysis (e.g. Kidd and Van Cott, 1972; Kirwin and Ainsworth, 1992) views the operator as an input-output device. The goal of conventional task analysis is to describe operator inputs, outputs and the rules that relate them. Perceptual control theory-based task analysis views the operator as a perceptual control system; the goal of the analysis is to determine the variables that the operator is to keep under control and the means the operator must have to effect this control.

The approach to task analysis described in this paper is called perceptual control analysis of tasks (PERCOLATe). PERCOLATe is an interview procedure that is designed to extract information from domain experts about how tasks are performed. PERCOLATe is based on the idea that all tasks involve control. A task involves control if the operator has some goal to achieve by carrying out the task. Thus, tasks like "searching" and "monitoring", which are not typically seen as control tasks, fit into the PERCOLATe

From The International Journal of Human-Computer Studies, 1999, 50 (6), 481-487. Reprinted with permission of Academic Press

References

Albus, J. (1981) Brains, behavior and robotics. Petersborough, NH: Byte Books

Anderson, J.R. (1983) The architecture of cognition. Cambridge, MA: Harvard University

Arbib, M. (1972) The metaphorical brain. New York: Wiley

Atwood, R. and Poison, P. (1976) A process model of water jar problems, Cognitive Psychology, 8, 191-216

Babler, T. and Dannemiller, J. (1993) Role of acceleration in judging landing location of free-falling objects, Journal Experimental Psychology: Human Perception and Performance, 19, 15-31

Black, H.S. (1934, January). Stabilized feedback amplifiers. Electrical Engineering, pp.114-120

Bordens, K.S. and Abbott, B.B. (1991) Research design and methods: A process approach, Mountain View: Mayfiel