OPERATIONALISM

"Operationalism" is a program that aims at linking all scientific concepts to experimental procedures and at cleansing science of operationally undefinable terms, which it regards as being devoid of empirical meaning. Scientists adopted the operational approach to their subject before the principles of operationalism were made articulate. Operationalist theory was erected not on the basis of independent philosophical considerations but upon what was already implicit in the working practice of scientists. P. W. Bridgman, the Nobel Prize–winning physicist who is commonly regarded as the founder of operationalism, emphasized this point when he said, "it must be remembered that the operational point of view suggested itself from the observation of physicists in action" ("The Present State of Operationalism," in The Validation of Scientific Theories, edited by Philipp Frank, Boston, 1956, p. 79).

A fairly nontechnical illustration of the kinds of development in science in which one can discern an implicit operational point of view is the manner in which physicists treated the concept of physical length. In the nineteenth century it was discovered that Euclid's geometry was not logically unique and that other geometries based on different axioms were not necessarily internally inconsistent. The question was raised about the nature of physical space. Do lines and figures in physical space obey the theorems of Euclid?

At first sight this seems a perfectly sensible question to which there must be a definite answer. Even today some amount of sophistication is required to ask whether we have a clear notion of what could be done to find out whether space has a certain set of properties. Unless we can give an affirmative answer to this question, we should not take it for granted either that space has or that it lacks certain geometrical properties. By the end of the nineteenth century, however, scientists had accepted the view that if we cannot devise operations that would disclose whether or not space was Euclidean, then no definite geometrical properties can be assigned to space at all.

It is clear that in order to determine the geometrical properties of physical figures we must be able to compare distances. If we are unable to say whether distance AB is greater, smaller, or equal to distance CD, where AB and CD do not lie alongside one another, then we cannot even begin to investigate the geometrical nature of space. We take it for granted, however, that in order to compare distances we need a rigid measuring rod, that is, a rod which can be relied upon not to change in length while being transported from place to place. But the question whether the lengths of transported rods are preserved cannot be settled unless we presuppose the possession of some other standard of measurement to which these rods could be compared, but it is agreed that the sole standard of length is a rigid rod. Thus, there are no rigid rods except by fiat, and distances consequently cannot be spoken of as being objectively equal or unequal to one another, and the nature of space cannot uniquely be determined. From an operational point of view, therefore, space has no intrinsic metric, and it is a matter of convention whether we say space obeys this or that set of geometrical axioms.

The Operationalist Thesis

Although the idea that physical entities, processes, and properties do not have an independent existence transcending the operations through which we may ascertain their presence or absence played an influential role in the thoughts of scientists before the 1920s, it was not until 1927 that Bridgman, in his celebrated Logic of Modern Physics, stated operationalism as an explicit program, made an articulate case for it, and undertook extensive operational analyses of the foundations of numerous physical concepts.

Bridgman soon had to retreat from his first extreme statement of operationalism. He had maintained that every scientifically meaningful concept must be capable of full definition in terms of performable physical operations and that a scientific concept is nothing more than the set of operations entering into its definition. The untenability of this view was quickly noticed—for example, by L. J. Russell, who in 1928 pointed out that in science one often speaks of certain operations as being better than others and that one cannot do so except in relation to something existing over and above them. Moreover, useful physical concepts do not as a rule lend themselves to an exhaustive definition. Any connection they have with instrumental operations may be loose and indirect: statements in which the concepts appear may, in the context of a set of other statements (but not on their own), entail statements describing physical operations. Consequently, in his later writings Bridgman freely permitted "paper and pencil operations," by which he meant mathematical and logical maneuverings with the aid of which no more is required of a concept than that it should be "indirectly making connection with instrumental operations."

It is not hard to see how by taking as one's model a physical concept like the length of a body one arrives at Bridgman's original position. But suppose someone objected that the stepping-off procedure carried out by measuring rods is not the only way to compute the length of a body. We may, for example, define it equally well in terms of the result obtained by timing the body's oscillation when it is allowed to swing as a pendulum and by using the well-known equation connecting the length with the period of oscillation. Length, after all, may enter into all sorts of relationships with other physical parameters, some of which we perhaps have not yet discovered.

To this objection it would have been replied that there is a fundamental difference between the ways in which the two sets of operations are related to the concept of length. The length of a body is "synonymous" with the number of times one can lay a rigid standard of length alongside it; when we speak of the length of a body we mean no more nor less than the number obtained through the stepping-off procedure performed by a measuring stick. When, however, we time a pendulum and then make the appropriate calculations, we merely measure length indirectly, via the relationship of length to other physical parameters. The second approach does not define length but rather inserts the already defined concept of length into an equation accepted as representing a genuine physical relationship.

It is much more difficult to maintain this distinction in the case of such concepts as temperature. One way to give an operational definition of temperature is in terms of measurements made by a mercury thermometer; another way is in terms of measurements made by a platinum-wire thermometer. The first way relies on the theory that the length of bodies varies with temperature; the second, on the theory that electrical resistance varies with temperature. It is easy to see that the concept of temperature is no more than partially interpreted through each of these, and doubtless other, sets of operations to which it is linked by relevant theories. This same position has become generally adopted toward all physical concepts.

We may thus distinguish three stages in scientific theorizing. In the first, preoperational stage, the universe was thought to contain many things and processes that transcend our theories about them and the operations and manipulations through which we may catch a glimpse of them in the mirror of experience. In the second, "naive" operational stage, the other extreme was taken, and all the terms of science were regarded as no more than abbreviations for our experimental results. In the third stage, scientific terms are still not regarded as standing for things and processes having an independent existence of their own, but the meaning of scientific terms is given by a more or less elaborate system of empirical theories in which the terms appear, together with the observations on which the theories embodying the terms are grounded. It is recognized that the concepts of science can never be fully grasped as long as the theories which contain them are open to further development.

The three stages in scientific theorizing are perhaps more dramatically accentuated in psychology than in the physical sciences. Until the early twentieth century the prevailing view was that psychology is a unique discipline dealing with a very special class of events, processes, and entities: the constituents of the realm of consciousness, to which no one but the experiencing individual has access. Although this realm is out of the reach of objective public operations and experimentations, many theorists regarded it as real—indeed, as more real than anything else—and believed that it should be studied by a unique method, introspection.

The radical behaviorism that replaced this mentalistic psychology is a form of naive operationalism and is based on the tenet that psychology is the study not of mental events, processes, or entities but of behavior. Psychologists were not to be concerned with publicly unobservable phenomena, and introspection—at best a private method of inquiry—was completely outlawed.

Today, in the third stage, sensations, images, and thought processes are no longer regarded as beyond the reach of scientists. They are studied through overt behavior, just as in physics nonobservables are studied indirectly through what is observed. The situation in psychology is very much like that in physics. That which is conceptualized need not be completely defined in terms of operations, although it must make contact with the world of public experience.

Operationalism and Verificationism

Operationalism is a movement within the philosophy of science. It is instructive to study its development in conjunction with a parallel movement in general philosophy: logical positivism, or logical empiricism. Central to logical positivism is the principle of verifiability, according to which any statement that is not a tautology must be verifiable or else is meaningless. It was thought that through the extensive employment of this principle it would be possible to show that many of the traditional unsolved problems of philosophy could be dealt with by demonstrating that they are simply meaningless. It was soon found, however, that the principle as originally conceived would get rid not only of troublesome problems but also of much useful discourse. The principle consequently underwent a number of revisions in rapid succession.

Rudolf Carnap's paper "The Methodological Character of Theoretical Concepts" embodies all the significant revisions. Carnap clearly exhibits a desire not to prescribe what should be regarded as meaningful from some metascientific or philosophical point of view but rather to describe what is commonly and usefully regarded as empirically meaningful. As mentioned earlier, operationalism from the beginning sought to explicate an approach already implied in the work of practicing scientists. Whereas verificationists previously tried to embrace all human discourse, they now, like the operationalists, confine their attempts to designing a criterion that will faithfully reflect what is meaningful discourse within empirical science. It has been realized that meanings are contextual and that one is therefore not to inquire whether a given sentence or word has or lacks meaning by itself but rather whether it has or lacks meaning relative to a specified system of theoretical, observational, and mixed statements.

A third important change, also clearly enunciated for the first time in Carnap's paper, is the departure from the original policy of inquiring directly into the meaningfulness of whole sentences. Instead, like the operationalists, Carnap deals with individual terms. He distinguishes between logical and empirical terms and also between observational-empirical and theoretical-empirical terms. Theoretical-empirical terms are not admitted into empirical discourse unless they can be shown to be anchored in observation. They need not be completely defined observationally, but a sentence must be constructible that, in conjunction with other sentences, logically implies that certain observations take place. A theoretical-empirical term is then regarded as having passed the test of empirical meaningfulness. The empirical significance of a sentence is now made dependent on the possession of significance by the terms it contains: Any syntactically well-formed sentence in which every term is significant (that is, is either a logical, an observational-empirical, or a theoretical-empirical term which has passed the test of empirical meaningfulness) is itself significant in the context of the group of sentences forming our system of science.

The only issue that divides operationalism from logical positivism is that operationalism seems to associate meaningfulness with linkability to experimental activities, whereas the principle of verifiability is satisfied if an expression is anchored to mere passive observation. However, this particular requirement of operationalism can safely be discarded, leading to a complete merger of these two contemporary offshoots of empiricism.

Criticism

Even in its present form, operationalism has not gone uncriticized. The chief complaint is that in the course of weakening its demands in order to accommodate highly theoretical but useful terms that would otherwise have been excluded from science, it has become so watered down as to lose all significance. Operationalism, according to its critics, says nothing we did not know all along. Even in a discipline less precise than physics—for example, in the social sciences—and in a period when standards of rigor had not reached their present stringency, if anyone had advanced a theory employing concepts which had no bearing whatsoever on observables, his theory would have been rejected. It is admitted that operationalism as originally conceived did have practical impact; there are concrete results, especially in psychology, whose production was motivated by the naive operationalistic distrust of anything remote from experience—for example, results obtained in the investigation of subaudible speech. Psychologists came to this area of inquiry chiefly through their search for objective, nonmentalistic alternatives to thought processes. But now, with the liberalization of the criterion for empirical significance—so the complaint goes—when all that is stipulated is that no term qualifies for membership in the vocabulary of science unless it is in some way connected to the universe of operations, observables, and experience, the principle of operationalism is merely platitudinous.

In attempting to reply to this, we must not forget that the scope of operationalism is not confined to the weeding out from scientific vocabulary of terms devoid of empirical significance. Once we have adopted the operational point of view, we have formed in our own minds a particular image of the nature of scientific concepts, which colors our expectations and influences in all sorts of ways our practical approach and methodology.

The world of experience and observation was at one time looked upon as containing mere dim reflections of the world that is conceptualized in physics and whose real existence was on a transcendental plane ultimately beyond our reach. Admittedly, that which is without any observable manifestations whatsoever, which, so to speak, casts no shadow onto the plane of experience, would never have been considered as being of any use to science. Nevertheless, it is not unimportant whether we regard our operations as capturing at most the shadows of the furniture of the universe or as dealing with the furniture itself. Objects totally dissimilar in substance and even in size and shape may under particular circumstances cast identical shadows. Therefore, from the similarity of shadows one cannot infer a similarity in the corresponding objects or even that these objects always cast similar shadows. Similarly, so long as we regard as mere reflections the observations to which physical concepts are linked, the finding of resemblances between some of them will not give rise to the expectation that they resemble in all particulars. On adopting the operational point of view, on the other hand, we think we are looking not at reflected shadows but at the very entities and processes that are conceptualized in science, and our attitude changes accordingly.

To give an illustrative example, the properties of gravitational force and the laws governing it had been exhaustively investigated in the seventeenth and eighteenth centuries. Electromagnetic forces were comparative newcomers in science. Were they to be expected to behave like mechanical forces? There are excellent grounds for saying no: the sources from which electromagnetic forces arise, the systems with which they are associated, and the means by which they are generated are totally different from those involving mechanical forces. However, operationalists tend to see in the product of mass and acceleration (that is, in the measure of force) the very substance of force, although others might see in it no more than force's most immediately apparent reflection. Indeed, as soon as it was observed that electromagnetic phenomena are accompanied by the forcelike effect of accelerating masses, it was taken for granted that they are fully governed by all the laws of Newtonian mechanics, even though the latter was developed to deal with an effect of totally different origin.

An important aim of operationalism besides the practical one is philosophical. For philosophical purposes, it is far from sufficient to state generally that every empirically significant term must somehow be linked to observables—one must precisely articulate the nature of this link and construct in full detail a criterion of meaningfulness. Therefore, many concepts in the various sciences were analyzed in detail in order to clarify the exact role instrumental operations and observations play in the definition or explication of them. Believers in the ultimate formalizability of empirical significance hoped that the results would be generalized and expressible in a philosophically satisfactory way. It is, however, by no means clear that such work has been entirely successful. In fact, some philosophers are of the opinion that such efforts are altogether in vain and that although when faced with any individual term we are able quite easily to judge whether it is empirically significant, we shall never succeed in explicating the general criterion distinguishing meaningful from meaningless utterances.

There is thus unquestionably much scope for operationally clarifying basic concepts. The skeptic might try to show that just as there are no formal criteria by which to distinguish a fertile from a sterile theory, so there is no criterion by which to distinguish the empirically significant from the meaningless. One who believes that the contact empirical concepts must make with operations or experience in general can be precisely formalized might try to show that if our demands are modest enough and we do not expect the criterion of empirical significance to provide guidance for future scientific research, there are in principle no obstacles in the way of such formalization. Their next step would be to execute this formalization in a manner that would stand up to all criticism.

See also Bridgman, Percy William; Carnap, Rudolf; Logical Positivism; Scientific Theories; Verifiability Principle.

Bibliography

Bridgman, P. W. The Logic of Modern Physics. New York: Macmillan, 1927.

Bridgman, P. W. The Nature of Physical Theory. Princeton, NJ: Princeton University Press, 1936.

Bridgman, P. W. The Nature of Some of Our Physical Concepts. New York: Philosophical Library, 1952. Many individual concepts carefully analyzed.

Bridgman, P. W. The Nature of Thermodynamics. Cambridge, MA: Harvard University Press, 1941. Carefully analyzes many individual concepts.

Bridgman, P. W. Reflections of a Physicist. New York: Philosophical Library, 1950. A collection of papers.

Bures, C. E. "Operationism, Construction, and Inference." Journal of Philosophy 37 (1940): 393–401.

Carnap, Rudolf. "The Methodological Character of Theoretical Concepts." In Minnesota Studies in the Philosophy of Science, Vol. 1, edited by Herbert Feigl and Michael Scriven, 38–76. Minneapolis: University of Minnesota Press, 1956.

Carnap, Rudolf. "Testability and Meaning." Philosophy of Science 3 (1936): 419–471, and 4 (1937): 1–40.

Crissman, P. "The Operational Definition of Concepts." Psychological Review 46 (1939).

Dingle, Herbert. "A Theory of Measurement." British Journal for the Philosophy of Science 1 (1950). An extreme operationalist viewpoint.

Feigl, Herbert. "Operationism and Scientific Method." Psychological Review 52 (1945). A lucid and fair assessment of operationalism. Written as a contribution to "Symposium on Operationism," presented in the same issue.

Frank, Philipp. Modern Science and Its Philosophy. Cambridge, MA: Harvard University Press, 1949. On p. 44 the work of Bridgman is likened to that of Carnap.

Frank, Philipp, ed. The Validation of Scientific Theories. Boston: Beacon Press, 1956. Various writers assess the significance of operationalism in the 1950s.

Hearnshaw, L. J. "Psychology and Operationalism." Australasian Journal of Psychology and Philosophy 18 (1941).

Hempel, C. G. "A Logical Appraisal of Operationalism." Scientific Monthly 79 (1954): 215–220. Reprinted with modifications in his Aspects of Scientific Explanation, 123–133. New York, 1965.

Lindsay, L. B. "A Critique of Operationalism in Science." Philosophy of Science 4 (1937). Important and fair criticism of operationalism.

Margenau, Hans. The Nature of Physical Theory. New York, 1952. On p. 232, expresses impatience with operationalism.

Pap, Arthur. "Are Physical Magnitudes Operationally Definable?" In Measurements, Definitions, and Theories, edited by C. West Churchman. New York, 1959. Argues for the abolition of the demand for active operations.

Peters, Richard. "Observationalism in Psychology." Mind 68 (1959).

Reichenbach, Hans. Experience and Prediction. Chicago: University of Chicago Press, 1938.

Russell, L. J. Review of Bridgman's The Logic of Modern Physics. Mind 47 (1938).

Schlesinger, Georg. Method in the Physical Sciences. London: Routledge and Paul, 1963. Chapter 4 is devoted to a discussion of the practical scope of operationalism.

Skinner, B. F. "Behaviorism at Fifty." Science (1963).

Skinner, B. F. "The Operational Analysis of Psychological Terms." Psychological Review 52 (1945): 270–277.

Spence, K. W. "The Postulates and Methods of 'Behaviorism.'" Psychological Review 55 (1948): 67–78.

Stevens, S. S. "The Operational Basis of Psychology." American Journal of Psychology 46 (1935).

Stevens, S. S. "The Operational Definition of Psychological Concepts." Psychological Review 42 (1935). A major spokesman for operationalism in psychology.

G. Schlesinger (1967)