A Study of Thinking

Anton Zijderveld

• Editorial: Routledge
• Año: 2017
• ISBN:
  9781351534451
• Edición: 2
• Categoría: Psicología
• Categoría: Historia y teoría en psicología

A Study of Thinking is a pioneering account of how human beings achieve a measure of rationality in spite of the constraints imposed by bias, limited attention and memory, and the risks of error imposed by pressures of time and ignorance. First published in 1956 and hailed at its appearance as a groundbreaking study, it is still read three decades later as a major contribution to our understanding of the mind. In their insightful new introduction, the authors relate the book to the cognitive revolution and its handmaiden, artificial intelligence.

The central theme of the work is that the scientific study of human thinking must concentrate upon meaning and its achievement rather than upon the behaviorists' stimuli and responses and the presumed connections between them. The book's point of departure is how human beings group the world of particulars into ordered classes and categories-concepts-in order to impose a coherent and manageable order upon that world. But rather than relying principally on philosophical speculation to make its point, A Study of Thinking reports dozens of experiments to elucidate the strategies that people use in penetrating to the deep structure of the information they encounter.

This seminal study was a major event in the cognitive revolution of the 1950s. Reviewing it at the time, J. Robert Oppenheimer said it "has in many ways the flavor of conviction which makes it point to the future."

CHAPTER 1

Introduction

We begin with what seems a paradox. The world of experience of any normal man is composed of a tremendous array of discriminably different objects, events, people, impressions. There are estimated to be more than 7 million discriminable colors alone, and in the course of a week or two we come in contact with a fair proportion of them. No two people we see have an identical appearance and even objects that we judge to be the same object over a period of time change appearance from moment to moment with alterations in light or in the position of the viewer. All of these differences we are capable of seeing, for human beings have an exquisite capacity for making distinctions.

But were we to utilize fully our capacity for registering the differences in things and to respond to each event encountered as unique, we would soon be overwhelmed by the complexity of our environment. Consider only the linguistic task of acquiring a vocabulary fully adequate to cope with the world of color differences! The resolution of this seeming paradox—the existence of discrimination capacities which, if fully used, would make us slaves to the particular—is achieved by man’s capacity to categorize. To categorize is to render discriminably different things equivalent, to group the objects and events and people around us into classes, and to respond to them in terms of their class membership rather than their uniqueness. Our refined discriminative activity is reserved only for those segments of the environment with which we are specially concerned. For the rest, we respond by rather crude forms of categorial placement. In place of a color lexicon of 7 million items, people in our society get along with a dozen or so commonly used color names. It suffices to note that the book on the desk has a “blue” cover. If the task calls for finer discrimination, we may narrow the category and note that it is in the class of things called “medium blue.” It is rare indeed that we are ever called upon to place the book in a category of colors comprising only the unique hue-brightness-saturation combination it presents.

The process of categorizing involves, if you will, an act of invention. This hodgepodge of objects is comprised in the category “chairs,” that assortment of diverse numbers is all grouped together as “powers of 2,” these structures are “houses” but those others are “garages.” What is unique about categories of this kind is that once they are mastered they can be used without further learning. We need not learn de novo that the stimulus configuration before us is another house. If we have learned the class “house” as a concept, new exemplars can readily be recognized. The category becomes a tool for further use. The learning and utilization of categories represents one of the most elementary and general forms of cognition by which man adjusts to his environment. It was in this belief that the research reported in this volume was undertaken. For it is with the categorizing process and its many ramifications that this book is principally concerned.

Identity and Equivalence Categories

The full moon, the moon in quarter, and the crescent moon all evoke the same nominative response, “moon.” From a common response made by a person to an array of objects we infer that he “has” an equivalence or identity category. The similar responses from which we draw such an inference need not be verbal. An air-raid siren, a dislodged piton while climbing, and a severe dressing-down by a superior may all produce a common autonomic response in a man and by this fact we infer that they are all grouped as “danger situations.” Indeed, the person involved may not be able to verbalize the category. While this is in itself interesting, it is not crucial to our point, which is simply that an equivalence range is inferred from the presence of a common response to an array of discriminably different events. This leaves many technical questions unsettled (cf. Klüver, 1933), but it serves to get the inquiry under way.

Two broad types of categorizing responses are obviously of interest. One of them is the identity response, the other the equivalence response, and each points to a different kind of category.

Without belaboring the obvious, identity categorization may be defined as classing a variety of stimuli as forms of the same thing. What lies behind the identity response is not clear, save that it is obviously a response that is affected by learning. It does not do to say simply that an object is seen as the identical object on a later encounter if it has not “changed its characteristics too much.” The moon in its phases varies from a sliver to a circle, in color from luminous white to the bronzed hunter’s moon. Sheldon (1950) collected a series of photographs of the same individual over a period of 15 years, the person standing in the same position against a uniform background. The photographs span the period from early boyhood to full manhood. As one riffles through the stack, there is a strong and dramatic impression of the identical person in the process of growth. Yet the pictures go through a drastic metamorphosis. Because such identity responses are ubiquitous and because they are learned very early in life, we tend to regard them somehow as a different process from other forms of categorizing—the recognition of two different people as both being people. Yet both depend upon what Michotte (1950) speaks of as the presence of a cachet spécifique or essential quality. They are both forms of categorizing. What differs is the nature of the inference: in the one case we infer “identity” from the presence of the cachet, in the other case “equivalence.”

How one comes to learn to categorize in terms of identity categories is, as we have said, little understood. Too often we have succumbed to the Kantian heritage and taken identity categories as given. Piaget’s recent work (1953) and the work of Michotte (1946) leave the question open. Piaget speaks of the learning of identity as corresponding to the mastery of a principle of conservation as in the conservation of energy in physics. At certain stages of development, an object passed behind a screen is not judged by the child to be the same object when it emerges on the other side. Hebb (1949) proposes that certain forms of neural growth must precede the capacity for the maintenance of identity. Whether the capacity is “innate” and then developed by being extended to new ranges of events or whether the capacity to recognize identity is itself learned is not our concern here. It suffices to note that its development depends notably upon learning.

That there is confusion remaining in the adult world about what constitutes an identity class is testified to by such diverse proverbs as plus ça change, plus la même chose and the Heraclitan dictum that we never enter the same river twice. Indeed, in severe psychotic turmoil one sometimes notes an uncertainty about the identity category that is the “self” in states of depersonalization, and a rather poignant reminder that the identity of self is even equivocal in normal states is provided by the sign behind a bar in the Southwest:

I ain’t what I’ve been.
I ain’t what I’m going to be.
I am what I am.

We speak of an equivalence class when an individual responds to a set of discriminably different things as the same kind of thing or as amounting to the same thing. Again we depend for our knowledge of the existence of a category upon the presence of a common response. While there is a striking phenomenological difference between identity and equivalence, both depend upon the acceptance of certain properties of objects as being criterial or relevant—again Michotte’s cachet spécifique—and others as being irrelevant. One may distinguish three broad classes of equivalence categories, each distinguished by the kind of defining response involved. They may be called affective, functional, and formal categories.

Certain forms of grouping appear to depend very heavily upon whether or not the things placed in the same class evoke a common affective response. A group of people, books, weather of a certain kind, and certain states of mind are all grouped together as “alike,” the “same kind of thing.” Further inquiry may reveal that all of them were experienced during a particularly poignant summer of childhood. What holds them together and what leads one to say that some new experience “reminds one of such and such weather, people, and states” is the evocation of a defining affective response.

Characteristically, categories marked by an affective defining response are not amenable to ready description in terms of the properties of the objects comprising them. The difficulty appears to lie in the lack of correspondence between affective and linguistic categories. As Schachtel (1947) and McClelland (1951) have suggested, categories bound together by a common affective response frequently go back to early childhood and may resist conscious verbal insight by virtue of having been established before the full development of language. For categorizing activity at the preverbal stage appears to be predominantly nonrepresentational, depending not so much on the common external properties of objects as on the relation of things encountered to internal needs, to follow Piaget’s argument (1951), or, to follow SchachteFs, on idiosyncratic and highly personalized impressions. Dollard and Miller (1950) argue persuasively that much of psychotherapy consists of the verbal labelling and resorting of such preverbal categories, so that they may become more accessible to the forms of symbolic or linguistic manipulation characteristic of adult problem-solving. Indeed, it is not difficult to imagine that the effectiveness of poetry often rests on its ability to cut across our conventional linguistic categories in a way evocative of more affective categorizations. Archibald MacLeish (1939) catches well this esthetic need for freedom from conventional verbal categories in his lines,

A poem should be palpable and mute
As a globed fruit
Dumb
As old medallions to the thumb
Silent as the sleeve-worn stone
Of casement ledges where the moss has grown—
A poem should be wordless
As the flight of birds.

The problems of specifying the properties of objects that mediate a common categorizing response become less arduous when the category is a functional or utilitarian one. Rather than an internal state rendering a group of things equivalent, now equivalence is based on an external function. The objects of a functional category fulfill a concrete and specific task requirement—“things large enough and strong enough to plug this hole in the dike.” Such forms of defining response almost always have, as Bartlett (1951) suggests, a specific interpolative function (“gap filling”) or a specific extrapolative function (“how to take the next step”). The experiments by Maier (1930, 1931, 1945) represent an outstanding instance of research on the conditions which facilitate and inhibit the recognition of “requirements” necessary for correct identification of an object as fulfilling specific functions in a particular task situation, such situations, for example, as how to bridge a gap between two objects given certain limits and certain properties in the materials provided.

Formal categories are constructed by the act of specifying the intrinsic attribute properties required by the members of a class. Such categories have the characteristic that one can state reliably the diacritica of a class of objects or events short of describing their use. The formal properties of science are a case in point. Oftentimes the careful specification of defining properties even requires the constructions of special “artificial” languages to indicate that common-sense functional categories are not being used. The concept “force” in physics and the word standing for the functional class of events called “force” in common sense do not have the same kind of definition. What is accomplished in effect by formal categories is that one is able to devise classes whose defining properties are not determined by the suitability of objects to a specific task. The emphasis of definition is placed more and more on the attribute properties of class members and less and less on “utilitanda properties,” to borrow a term from Tolman (1932). The development of formalization is gradual. From “things I can drive this tent stake with” we move to the concept “hammer” and from there to “mechanical force,” each step being freer of definition by specific use than the former.

The development of formal categories is, of course, tantamount to science-making and we need not pause here to discuss this rather impenetrable problem. It suffices to note that formal categories and formal category systems appear to develop concurrently with methods for representing and manipulating them symbolically. What impels one to formalization we cannot say. That the urge is strong is unquestionable. Indeed, it is characteristic of highly elaborated cultures that symbolic representations of formal categories and formal category systems are eventually developed without reference to the classes of environmental events that the formal categories “stand for.” Geometry provides a case in point, and while it is true that its original development was contingent upon the utilitarian triangulation systems used for redividing plots after floods in the Nile Valley, it is now the case that geometers proceed without regard for the fit of their formal categories to specific empirical problems.

It is obvious that there are close relationships between affective, functional, and formal categories and that they are often convertible one into the other. About the conversion of functional categories into formal ones—finally rendering the category of “things good for postpartum mothers” such as ground bone and certain chalks into the formal class “calcium”—we have already taken some notice. It is interesting that the gifted mathematician often speaks of certain formal categories in terms that are also affective in nature. G. H. Hardy in his delightful “apology” (1940) speaks of the class of things known as “elegant solutions” and while these may have formal properties they are also marked by the fact that they evoke a common affective response. The distinction between the three types is, we would suggest, a useful one and it may well be that the process whereby they are learned is informatively different. It is suggestive, for example, that the brain-injured patients described by writers like Goldstein (1940) and Head (1926) seem quite capable of utilizing functional categories but are precipitated into a crisis when faced with the need of locating or forming or using categories divorced from the immediate function to be served by their exemplars.

The Invention of Categories

To one raised in Western culture, things that are treated as if they were equivalent seem not like man-made classes but like the products of nature. To be sure, the defining criteria in terms of which equivalence classes are formed exist in nature as potentially discriminable. Rocks have properties that permit us to classify them as rocks, and some human beings have the features that permit us to categorize them as handsome. But there exists a near infinitude of ways of grouping events in terms of discriminable properties, and we avail ourselves of only a few of these.

Our intellectual history is marked by a heritage of naive realism. For Newton, science was a voyage of discovery on an uncharted sea. The objective of the voyage was to discover the islands of truth. The truths existed in nature. Contemporary science has been hard put to shake the yoke of this dogma. Science and common-sense inquiry alike do not discover the ways in which events are grouped in the world; they invent ways of grouping. The test of the invention is the predictive benefits that result from the use of invented categories. The revolution of modern physics is as much as anything a revolution against naturalistic realism in the name of a new nominalism. Do such categories as tomatoes, lions, snobs, atoms, and mammalia exist? In so far as they have been invented and found applicable to instances of nature, they do.^* They exist as inventions, not as discoveries.

Stevens (1936) sums up the contemporary nominalism in these terms: “Nowadays we concede that the purpose of science is to invent workable descriptions of the universe. Workable by whom? By us. We invent logical systems such as logic and mathematics whose terms are used to denote discriminable aspects of nature and with these systems we formulate descriptions of the world as we see it and according to our convenience. We work in this fashion because there is no other way for us to work” (p. 93). Because the study of these acts of invention is within the competence of the psychologist, Stevens calls psychology “the propadeutic science.”

The recognition of the constructive or invented status of categories changes drastically the nature of the equivalence problem as a topic for psychological research. The study of equivalence becomes, essentially, a study of coding and recoding processes employed by organisms who have past histories and present requirements to be met. The implicit assumption that psychological equivalence was somehow determined by the “similarity” or “distinctive similarity” of environmental events is replaced by the view that psychological equivalence is only limited by and not determined by stimulus similarity. The number of ways in which an array of events can be differentiated into classes will vary with the ability of an organism to abstract features which some of the events share and others do not. The features available on which to base such categorial differentiation, taken singly and in combination, are very numerous indeed. As Klüver (1933) so well put it more than two decades ago, the stimulus similarity that serves as a basis for grouping is a selected or abstracted similarity. There is an act of rendering similar by a coding operation rather than a forcing of equivalence on the organism by the nature of stimulation.

Two consequences immediately become apparent. One may ask first what are the preconditions—situational and in the past history of the organism—that lead to one kind of grouping rather than another. The characteristic forms of coding, if you will, now become a dependent variable worthy of study in their own right. It now becomes a matter of interest to inquire what affects the formation of equivalence classes or systems of equivalence coding. The second consequence is that one is now more tempted to ask about systematic individual and cultural difference in categorizing behavior. In so far as each individual’s milieu and each culture has its own vicissitudes and problems, might one not expect that this would reflect itself in the characteristic ways in which members of a culture will group the events of their physical and social environment? And, moreover, since different cultures have different languages, and since these languages code or categorize the world into different classes, might it not be reasonable to expect some conformance between the categories normally employed by speakers and those contained in the language they use?

Consider now the scope of the problem, the generality of categorizing, and the benefit to be derived from studying its various manifestations.

The Generality of Categorizing

The first benefit to be derived from a closer study of categorizing behavior is a gain in generality for psychological theory. Categorizing is so ubiquitous that an understanding of its psychological nature cannot help but shed light on a wide range of problems within psychology. Most of the examples we have given thus far have, perhaps for simplicity’s sake, been drawn from the field of perception. This is misleading. For the act of categorizing, operationally d...