The Need for a New Approach
In the attempt to develop new approaches to old problems, inevitably questions arise regarding the unanalyzed and, therefore, unchallenged presuppositions tacitly or explicitly accepted by the field. To establish a new brand of psychological theory necessarily requires that several fundamental tenets of traditional, contemporary psychology be challenged and, if possible, overthrown to make way for the new approach. However, before presenting the tenets of the ecological approach to the problem of knowledge, it will be useful to consider some of the major assumptions and hypotheses that underlie current cognitive psychology, since this is the area of psychology which has assumed responsibility for solving the problem of knowledge.
The man-machine analogy. Current cognitive theory may best be characterized as the information-processing approachâan approach that feeds upon the manâmachine, or brain-computer, analogy. This analogy has been the staple of most psychological theoristsâ diets for the last decade or so.
The information-processing approach, more than any previous one, seems to recognize the need for precision when characterizing psychological processes. In their search for ways to construct models that explain how knowledge might be represented, stored, and retrieved from memory, it was natural for psychologists to turn for help to the fast developing field of computer science, once the promise of behaviorism was seen to be overblown. After all, what seemed to be analogical problems were being explicitly addressed by computer scientists in their attempts to understand complex information-processing devices. How natural it was to extend the concept of computer to include the psychological machinery of humans in order to see if analogs of the information-processing solutions might not be found to hold there as well. Thus, there is no great mystery in explaining how current cognitive theory came to rely so heavily on computer science concepts and its modeling techniques. For many cognitive theorists their prayers for precision seemed to have been answered.
There should be little quarrel with the contention that this metaphor has in many ways served our field well, for it has helped to free many psychologists from the excessive restrictions of radical behaviorism. Similarly, this mechanistic metaphor has reinstated a serious and worthwhile concern for understanding how knowledge is represented and organized within humans and animals as self-organizing, knowledge-gathering systems. Moreover, the acceptance of the analogy of cognitive structures to computational structures (for example, networks, hierarchies, or heterarchies of control functions) has the additional virtue of fostering an appreciation for exactitude in theory that ultimately demands mathematically rigorous models (Cunningham, this volume). All this is surely an improvement over the simplistic serial processing models offered by stimulus-response theory over the preceding decades.
Yet we psychologists can not be content to rest on our theoretical laurels, especially when garnered so easily from the successes of another field. It would be both unfair and unwise to expect the solutions given by computer scientists to their problems to also qualify mutatis mutandi as solutions to our own problems, since serious questions can be raised whether or not the two sets of problems are even abstractly the same (Shaw & McIntyre, 1974).
It is wise to remember that computer scientistsâ systems are artificially contrived while ours are naturally evolved; their systems are passive while ours are active; their systems are purposeless (except in a second-hand way) while ours pursue primary goals of self-survival and adaptation; their systems, as complex as they are, are still astronomically simpler than those psychologists must understand. And what of the role of emotional, personality, and social factors in determining the questions appropriate to humans but as yet undefined for artificial systems? Indeed, to be realistic, we should ask who should be the tutor and who the tutored? Consequently, it is not unreasonable to suggest that at this juncture in the history of these two young sciences the most relevant concepts computer science may offer psychology are still in the offing. Thus, not pessimism but prudence counsels psychologists to assume a wait-and-see attitude before pronouncing sentence upon the manâmachine analogy. Indeed, the profound question of what constitutes a machine hangs in the balance over and against the question of what is a human. The answer may surprise us all.
On the other hand, there is the danger that premature application of the manâmachine analogy to psychology, to the exclusion of other approaches, may unnecessarily restrict our thinking about problems whose solutions demand greater theoretical flexibility than strict adherence to the analogy allows. For instance, as psychologists, there is a natural and understandable tendency to ignore certain complications of our phenomena for the sake of expediency in either doing experiments or constructing models. Rarely do we clearly enunciate any method by which theoretical principles designed to explain animal and human behavior in the laboratory might be generalized back to the natural contexts in which the behaviors originated. We tend to forget that humans and animals are active, investigatory creatures driven by definite intents through a complex, changing environment replete with meaning at a variety of levels of analysis.
Thus we feel no tinge of theoretical compunction in blithely comparing such active, knowledge-seeking beings with unconscious, static machines that lack a wit of natural motivation. Unlike humans and animals, who perceptually mine the world for information on a need-to-know basis, artificial systems can only soak up information passively by being spoon-fed batches of alpha-numeric characters that have been conceptually predigested by human programmers. In sharp contrast to natural information, such symbolic information is devoid of any natural meaning for the machine and, therefore, can play no role in either its adaptation or survival in a relatively dust-free, humidity controlled, air-conditioned environment. In such a sterile model for man, perceiving becomes a passive process and knowing a purposeless one, and as for action (that is, purposive behavior), it remains nonexistent.
Of course, it would be grossly unfair to criticize the computer science area for the rather myopic view many psychologists have taken of such problems, since too often the best concepts that computer science offers are much better than what psychologists elect to borrow. The theory of locomoting, special-purpose computers (robots), as well as current attempts by cyberneticists to devise prosthetic devices based on computer components (for example, artificial limbs and helmets with video inputs to the tactile system to aid the blind) are among the avant garde projects which may inspire new life in the nearly dead manâmachine metaphor, and thereby open fresh vistas for psychological theorists. Again it is wiser to counsel patience than pessimism regarding the ultimate worth of the metaphorical extension of the techniques of either field to the province of the other. Rather let us critically scrutinize the current coinage of the metaphor so as to assay what pragmatists would call its âcash valueâ as a working hypothesis.
Some light might be shed on the nature of the quest for alternatives to the manâmachine analogy if a sharp line of delineation is drawn around those epistemological questions explicitly asked, or tacitly assumed by the field, and those equally significant questions given short shrift. Typically, the questions given short shrift are those concerning the nature and origins of the information processed; how ecologically significant information is selected; and how its value to the animal or human might be determined. The latter questions focus on what is processed whereas current theory focuses on the how of processing. Attempts to reduce questions regarding what is processed to questions regarding how what is processed is processed are doomed to failure because the two questions fail to address the same issue.
More specifically, the manâmachine analogy is overdrawn when human perception is likened to the decoding process by which computers compile data, and human knowledge to the concept of retrieval of information from storage. The manâmachine analogy becomes a hindrance rather than an aid to psychological theory when it derails our thinking about how living creatures gather and act upon knowledge in dynamic natural contexts. Such questions can in no way be reduced to questions of how information is represented, stored, or retrieved from storage by static devices in artificially controlled environments. Rather the question should be how information âstoredâ in the world can be perceptually extracted by active, investigative creatures in their relatively successful effort to survive calamity and achieve well-being.
If the ecological approach merits any serious consideration as an alternative approach to psychological theory, then it might best be revealed by contrasting this approach with the more popular one of information processing psychology. In fairness, however, it should be remembered that although the ecological approach may at this time be only programmatic, in the sense that it offers embryonic theory sketches in the place of fully adequate theories of the knowledge-gathering activities of humans and animals, the information-processing approach is no less programmatic. So far it too has failed to offer adequate solutions to any of the major problems of psychology (for example, perception, memory, attention, action, language, motivation, emotion, personality). The timetable for success of any approach cannot be predicted, although the probability of success might be, if we but consider which of the approaches poses the significant questions most sensibly. The judicious allocation of the resources of our field demands that we avoid, if we can, the vain pursuit of will-o-the-wisp questions.
Let us not, however, delude ourselves: Our field is much too young, our theories too sketchy, the mettle of our techniques too untried, to condemn with certitude any approach without fair trial. Consequently, the purpose of this introductory chapter (and the whole book for that matter) is not to reject out of hand any existing program of research, but to make way for an alternative programâthe ecological approach.
In the next sections of this chapter, we address several fundamental issues of psychological theory from the vantage point of the ecological approach, questions such as: What is the nature of knowing? Is perception direct or mediated? What is the role of inference and memory in perceiving the true properties of the world? And, finally, what are the recalcitrant issues that most sharply distinguish current information processing approaches to psychological theory from the alternative ecological approach adopted by many of the authors (although not all) of this volume. Since the last question is the guiding thread of this essay, it will be discussed most fully.
The Ecological and Information-Processing Approaches Compared
A wise adage asserts that man proposes but nature disposes! This aphorism elegantly expresses the fundamental tenet of the ecological approachâthat the nature of humans is inextricably intertwined with the nature of a world in which they live, move, and have their being. In short, the ecological attitude is founded on the fundamental belief that man is indeed the mirror of nature. This view does not suggest, however, that psychological phenomena may be reduced to physical ones, but that physical phenomena be interpreted in the light of their relevance to psychological phenomenaâan attempt, if you will, to âecologizeâ physics so as to understand better how the world provides a habitat for humans as knowing-agents, and not just as physical or biological objects. Thus the ecological attitude germinates in the minds of theorists who come to the stark realization that humans, as must all living things that survive, depend upon their natural and cultural environments for knowledge as well as victuals. For such theorists it is impossible to accept any longer models for man whose only virtue is that they were contrived to function efficiently in artificial contexts.
The environment in which animals and humans have evolved, and upon which they depend for their well-being can be construed from both the material and functional viewpoints: The material environment of a species provides resources in the form of âhardâ goods, such as air, water, food, shelter, mates and tools, while the functional environment provides resources in the form of âsoftâ goods, such as the requisite knowledge-gathering experiences for planning and executing adaptive actions. By adaptive actions we mean, rather redundantly, those perceptual and behavioral activities by which the hard or soft goods needed for survival and well-being are secured. Taken together, the sum total of the material and functional aspects of the general environment required by a particular species in order to achieve optimal adaptation, that is, the requisite hard and soft goods, defines an ecological niche for that species. Hence the ecological niche for an animal is not where it lives (its habitat) but how it lives. (See Gibson, this volume.)
The ecological approach calls for nothing less than a complete understanding of the complex and everchanging relationship of person-as-knower to the en-vironment-as-known. Hence a complete understanding of the informational aspects of the ecological niche, as well as the behavioral consequences of such information, is of paramount concern to the ecological approach, regardless of the animal under studyâhumans not excluded. But, as we shall see, the concept of ecologically significant information (Mace, this volume; Shaw, McIntyre, & Mace, 1974) is as radically different from the concept of information used in communication theory, computer science, or some psychological theories currently in vogue, as the concept of ecologically significant action patterns is radically different from the behavioristsâ concept of unintentional animal movements (Turvey, Weimer, this volume).
It is natural, however, when attempting to clarify a new approach to problems in an area, that the old way of asking the questions come under fire. The subtitle of the next chapter by William Mace is a dra...