Fact, Theory, and Hypothesis - Sociology of Science

The terms fact, theory, and hypothesis are sometimes treated as though they had clear meanings and clear relations with one another, but their histories and uses are more complex and diverse than might be expected. The usual sense of these words places them in a relationship of increasing uncertainty. A fact is usually thought of as a described state of affairs in which the descriptions are true or highly supported. A highly corroborated or supported hypothesis is also a fact; a less well corroborated one is still a hypothesis. A hypothesis which is not supported by or corroborated by other evidence would not be a fact, but could become a fact if it came to be corroborated to a high degree of certainty by other evidence. Similarly, a theory, which is a logically connected set of hypotheses, could come to be a fact if the hypotheses in the theory were to be highly corroborated by the evidence.

The Conceptual Character of “Facts”

Even with this simple picture of the relationship between these terms, one can see a number of potential difficulties and raise a number of difficult questions. Begin with the notion of corroboration. If a fact is a highly corroborated hypothesis, this would seem to mean that there is a level that is prior to facts which supplies the evidence that goes into corroboration. If the corroborating evidence consists of other facts, one would want to know how these facts were corroborated. So it is more common to talk about some more fundamental level of evidence, such as data. “Data” literally means “given.” But the idea that there is something in the world that is simply given, and true or valid as such, has its own difficulties.

However, when we collect data we have already described them or have a conceptual category for them. Since the “data” are already in a predefined category, we are not dealing directly with the world but with an already categorized world. This idea of fact as already conceptual had a long history in writing about science and is particularly associated with the nineteenth century philosopher William Whewell. Whewell said the following: “Fact and Theory correspond to Sense on the one hand, and to Ideas on the other, so far as we are conscious of our Ideas; but all facts involve ideas unconsciously; and thus the distinction of Facts and Theories is not tenable, as that of Sense and Ideas is” (Whewell 1984: 249). And this raises the questions of where the categories themselves come from and what their status is. In 1932, L. J. Henderson wrote an article which was cited by the sociologist Talcott Parsons (Parsons 1968: 41) which defined a scientific fact as an empirically verifiable statement about phenomena in terms of a conceptual scheme” (Henderson 1932). What this implied, especially for Parsons, was that to be a fact it was necessary to be a part of or to depend on a conceptual scheme. And conceptual schemes were not givens but were, like theories, invented for the purpose of enabling us to make statements such as the statements in theories.

The question of where categories come from and how something becomes a fact has been a major concern of sociologists of scientific knowledge. An important book by Ludwig Fleck (1979), a physician scientist, provided the basic framework for this study. Fleck argued that to be accepted as a fact required something social, which he called a thought collective,” in terms of which a concept is transformed from idea into accepted truth. The emphasis was on the social phenomenon of acceptance, something which, Fleck showed, did not merely result from the accumulation of evidence, but rather from the activity of a community of persons, with a common thought style, exchanging ideas. This implied that discovery” was never an individual act, but rather collective; and that conceptual content was part of the collective thought of the community, which developed in the course of exchange. Only retrospectively, once the discovery had been fit into the collective thought of the community, could the significance of discoveries be fully understood.

Inductivism vs. Hypothetico-Deductivism

The methodological understanding of science that fits best the insight that facts are already conceptual is hypothetico-deductivism, which contrasts to a different view of methodology called inductivism. Inductivism was the traditional understanding that science consists of generalizations which can be built up on the basis of the collection of information or data which can then be arranged into generalizations. The problem with inductivism is that there is no logical way to get from a collection of finite singular pieces of information to a generalization which goes beyond the particulars that have been collected. Hypothetico-deductivism deals with this limitation by turning the problem upside down by beginning with hypotheses that are generalizations and asking whether the observable particulars are consistent with (because they are implied by) the generalization. The hypothesis all crows are black” has the potential to be contradicted every time we see a crow. Thus, each particular crow can be used as a test of the hypothesis and the more stringently we test the hypothesis, the more secure we are in our belief that the hypothesis is true.

Hypothetico-deductivism has an advantage over inductivism as a method in that hypothetico-deductivism can be used to corroborate theories where the concepts in the theories are not themselves directly observable. The wave hypothesis in physics is a traditional example of this. The hypothesis logically implied generalizations for which observations could be collected. Because the theory correctly predicted these and other facts that could be observed, the claims about what could not be observed were themselves corroborated. This is an especially important possibility in sociology because many of the concepts in sociology do not directly apply to observable facts in the world, but instead to grounding concepts such as “society,” or “role,” or “attitude.” These concepts can be understood as having observable manifestations, but are not limited to or equivalent to observable manifestations.

In physics the term observation and the notion of the logical relations between claims in a theory had a more straightforward meaning. The logical relations were mathematical. The way in which an implication was derived from a theory was by deriving it mathematically through a proof. The theories of sociology, in contrast, rarely if ever have this structure, although in many cases theories are presented with verbal formulations which have ”logical connections” in a looser sense, namely that the claims in the theory provide a good reason, in context, for expectations that can be tested or applied to cases. Sociological theories thus resemble physical theories in the hypothetico-deductive sense in some ways, but differ in others. A ”theory” may be a part of a theoretical structure, such as a system of conceptual categories which enable description. But it may instead be a description of unobservable forces or unobservable mechanisms, such as the mechanism that reinforces social hierarchy by selectively excluding members of the lower classes from the paths which lead to positions of wealth and power.

Sense Making In Theories

The major difference between sociological and physical theory is that the concepts in sociology are typically sense making: they serve to enable a fact described in its terms to be more fully intelligible. Making a fact more intelligible will usually make its consequences more predict able. If I even do something as simple as characterizing an action as a product of the agent’s beliefs and positive attitudes towards some outcome specified by the agent’s beliefs, I have improved the prediction over alternative descriptions or over chance. This is not the same thing as a prediction in physics, but it is predictive nevertheless.

If the sociologist can add to this simple situation of explaining in terms of beliefs and attitudes by characterizing the set of beliefs that support the particular belief that relates directly to the action, for example by understanding a religiously motivated action in terms of a typology of religious belief, and if the sociologist can explain how those beliefs come to be distributed in particular groups, she will have something that begins to look like a theory that explains those actions sociologically, that is to say at some level beyond the level of the individual. Similarly, for characterizations about such things as role, for example. If an individual’s behavior can be characterized in terms of the roles which they are fulfilling, this explanation can be extended by accounting for the process of socialization into the role in question and the ways in which role behavior is enforced as normative, or enacted and supported by the expectations of other agents.

These descriptions of mechanisms are more general claims than the explanation of the individual’s action; they are ”social” in the sense that they serve to organize the behavior of individuals in relation to a limited category of individuals. The characterizations are sense making in that they explicate the beliefs and expectations of the people involved, and predictive in the sense that they improve our own expectations about what people will actually do and what role conduct is likely to persist or appear in different social settings. Similarly, a good categorization scheme using intelligibility enhancing concepts (e.g., Weber’s categories of legitimate authority) will enable the sociologist equipped with it to improve expectations as well as achieve understanding.

Sociologists have traditionally differed with respect to the emphasis they place on different aspects of these kinds of loose theories.

Parsons, for example, was particularly concerned with the elaboration of conceptual distinctions which could be used to organize comprehensively the concepts of sociology and relate them to one another and to the concepts of other disciplines. Parsons placed little emphasis on making individual actions or beliefs intelligible and little emphasis on prediction; although he envisioned future possibilities of prediction he also believed that many of the central variables of sociology were unquantifiable and that this was an inherent limitation on sociological theories approximating physics.

Diversity in Theory in Sociology

Some sociologists have tried strictly to adhere to the idea of deductive theorizing as modeled on physics. Typically, these sociologists have attempted to devise experimental settings in which limited variables or sets of variables can be measured in relation to other variables in such ways that predictions can be made and confirmed. This strategy has the potential of illuminating basic concepts which can then be applied to social life outside the laboratory as fundamental theories which approximate the more complex realities of actual social life. One problem with this strategy is that there are often alternative theories which are equally effective or ineffective as means of making sense of and predicting in the more complex actual settings of the real world.

Some theories, generally called interpretive theories, are focused primarily on intelligibility itself. For these theorists, providing a more fully realized and rich interpretation of the actions, attitudes, and beliefs of individuals is the appropriate and most productive strategy for dealing with a world of agents, that is to say a world of individuals who are themselves interpreters of one another and who act in terms of these interpretations.

Other theories, such as rational choice theory, borrow the theoretical structure of decision theory, game theory, or economics to provide a particular kind of intelligibility to actions of individuals who are treated in abstraction from considerations about the specific actual beliefs and attitudes of the individuals, and the explanations are evaluated in terms of their ability to predict the choices of these individuals. In one sense this represents the highest level of intelligibility, namely rational choice. In another sense it is removed from the subjective experience and ongoing interpretive activities of individual agents and thus serves as a poor guide to these aspects of experience.

References:

Fleck, L. (1979 [1935]) Genesis and Development of a Scientific Fact. University of Chicago Press, Chicago.
Henderson, L. J. (1932) An Approximate Definition of Fact. University of California Studies in Philosophy 14: 179-99.
Parsons, T. (1968 [1937]) The Structure of Social Action, Vol. 1. Free Press, New York.
Whewell, W. (1984) Selected Writings on the History of Science. Ed. Y. Elkana. University of Chicago Press, Chicago.

Back to Sociology of Science