The concept of ”positivism” was originally used to denote the scientific study of social phenomena, but today the term positivism has become vague. Most often, it is used as a pejorative smear for certain kinds of intellectual activity in the social sciences, sociology in particular. Most frequently, at least within sociology, positivism is associated with such undesirable states as ”raw empiricism,” ”mindless quantification,” ”antihumanism,” ”legitimation of the status quo,” and ”scientific pretentiousness.” With few exceptions (e.g., Turner 1985), sociologists are unwilling to label themselves ”positivists.” Yet, the titular founder of sociology—Auguste Comte—used this label as a rallying cry for developing formal and abstract theory that could still be used to remake society; so, the current use of the term does not correspond to its original meaning. If anything, the term connotes almost the exact opposite of Comte’s vision (1830-1842). It is proper, therefore, to review Comte’s original conception of positivism and its use in early sociology, and then we can discover how and why the meaning of positivism changed.
In Cours dephilosophie positive, Comte began by asserting that ”the first characteristic of Positive Philosophy is that it regards all phenomena as subject to natural Laws” (1830-1842, p. 5). Moreover, he emphasized that ”research into what are called causes, whether first or final,” is ”in vain” (1830-1842, p. 6); and by the time he was well into Cours de philosophie positive, he stressed that a ”great hindrance to the use of observation is the empiricism which is introduced into it by those who . . . would interdict the use of any theory whatever” because ”no real observation of any kind of phenomena is possible, except in as far as it is first directed, and finally interpreted, by some theory” (1830-1842, p. 242). Rather, the goal of positivistic sociology is to ”pursue an accurate discovery of … Laws, with a view to reducing them to the smallest possible number,” and ”our real business is to analyze accurately the circumstance of phenomena, to connote them by natural relations of succession and resemblance” (18301842, p. 6). Comte’s exemplar for this advocacy was Newton’s law of gravitation, an affirmation of his early preference to label sociology as ”social physics.” Moreover, such laws were to be used to reconstruct society; and while Comte went off the deep end on this point, proclaiming himself, late in his career, to be the ”high priest of humanity” (Comte 1851-1854), it is difficult to see Comte’s positivism as antihumanistic, as conservative, or as legitimating the status quo.
How, then, did Comte get turned on his head? The answer to this question cannot be found in nineteenth-century sociology, for the most positiv-istic sociologists of this period—Herbert Spencer (1874-1896) and Emile Durkheim ([1893] 1947; [1895] 1934)—could hardly be accused of ”raw” and ”mindless” empiricism, nor could they in the context of their times be considered antihumanistic, conservative, and apologists for the status quo (the label ”conservative” for these thinkers is imposed retrospectively, through the refraction of contemporary eyeglasses). Moreover, early American sociologists—Albion Small, Frank Lester Ward, Robert Park, William Graham Sumner, and even the father of statistical methods and empiricism in American sociology, Franklin Giddings—all advocated Comtean and Spencerian positivism before World War I. Thus, the answer to this question is to be found in the natural sciences, particularly in a group of scientist-philosophers who are sometimes grouped under the rubric ”the Vienna Circle,” despite the fact that several intellectual generations of very different thinkers were part of this circle.
Before the ”circle” was evident, the nature of the issues was anticipated by Ernst Mach (1893), who argued that the best theory employs a minimum of variables and does not speculate on unobservable processes and forces. Mach emphasized reliance on immediate sense data, rejecting all speculation about causes and mechanisms to explain observed relations among variables. Indeed, he rejected all conceptions of the universe as being regulated by ”natural laws” and insisted that theory represent mathematical descriptions of relations among observable variables. Although Mach was not a member of the Vienna Circle, his ideas framed the issues for those who are more closely identified with this group. Yet, his ideas did not dictate their resolution. Many in the Vienna Circle were concerned primarily with logic and systems of formal thought, almost to the exclusion of observation (or, at least, to the point of subordinating it to their primary concerns). A split thus developed in the Vienna Circle over the relative emphasis on empirical observation and systems of logic; a radical faction emphasized that truth can be ”measured solely by logical coherence of statements” (which had been reduced to mathematics), whereas a more moderate group insisted that there is a ”material truth of observation” supplementing ”formal truths” (Johnston 1983, p. 189). Karl Popper, who was a somewhat marginal figure in the Vienna Circle of the 1930s, is perhaps the best-known mediator of this split, for he clearly tried to keep the two points of emphasis together. But even here the reconciliation is somewhat negative (Popper 1959, 1969): A formal theory can never be proved, only disproved; and so, data are to be used to mount assaults on abstract theories from which empirical hypotheses and predictions are formally ”deduced.”
Why did the philosopher-scientists in the Vienna Circle have any impact on sociology, especially American sociology? In Europe, of course, sociology had always been firmly anchored in philosophy, but in American sociology during the 1920s and 1930s, the rise of quantitative sociology was accelerating as the students of Franklin Giddings assumed key positions in academia and as Comtean and Spencerian sociology became a distant memory. (It should be noted, however, that Marx, Weber, and Durkheim had yet to have much impact on American sociology in the late 1920s or early 1930s.) But American sociology was concerned with its status as science and, hence, was receptive to philosophical arguments that could legitimate its scientific aspirations (Turner and Turner 1990). Mach was appealing because his advocacy legitimated statistical analysis of empirical regularities as variables; and Popper was to win converts with his uneasy reconciliation of observation and abstract theory. Both legitimated variable analyses; and for American sociologists in the 1930s and later from the 1940s through the early 1960s, this meant sampling, scaling, statistically aggregating, and analyzing empirical ”observations.” Members of the Vienna circle had even developed an appealing terminology, logical positivism, to describe this relation between theory (abstract statements organized by a formal calculus) and research (quantitative data for testing hypotheses logically deduced from abstract statements). The wartime migration of key figures in the late Vienna Circle to the United States no doubt increased their impact on the social sciences in the United States (despite the fact that the ”logical” part of this new label for ”positivism” was redundant in Comte’s original formulation). But logical positivism legitimated American empiricism in this sense: The quantitative data could be used to ”test” theories, and so it was important to improve upon methods of gathering data and analyzing methodologies in order to realize this lofty goal. Along the way, the connection of theory and research was mysteriously lost, and positivism became increasingly associated with empiricism and quantification, per se.
There was a brief and highly visible effort, reaching a peak in the late 1960s and early 1970s, to revive the ”logical” side of positivism by explaining to sociologists the process of ”theory construction.” Indeed, numerous texts on theory construction were produced (e.g., Zetterberg 1965; Dubin 1969; Blalock 1969; Reynolds 1971; Gibbs 1972; Hage 1972), but the somewhat mechanical, cookbook quality of these texts won few converts, and so the empiricist connotations of positivism were never successfully reconnected to abstract theory. Even the rather odd academic alliance of functional theory with quantitative sociology—for example, Merton and Lazarsfeld at Columbia and Parsons and Stouffer at Harvard—was unsuccessful in merging theory and research, once again leaving positivism to denote quantitative research divorced from theory.
Other intellectual events, anticipated by various figures of the Vienna Circle, created a new skepticism and cynicism about the capacity to develop ”objective” science, especially social science. This skepticism stressed the arbitrary nature of symbols and signs and hence their capacity to represent and denote the universe independently of the context in which such signs are produced and used. Such thinking was supplemented by Kuhn’s landmark work (1970) and by the sociology of science’s emphasis (e.g., Whitley 1984) on the politico-organizational dynamics distorting the idealized theory-data connection as advocated by Popper (1969). Out of all this ferment, a new label increasingly began to appear: postpositivism. This label appears to mean somewhat different things to varying audiences, but it connotes that Comte’s original vision and Popper’s effort to sustain the connection between empirical observations and theory are things of the past—just as ”rationalism” and ”modernity” are giving away to ”postmodernism.” Thus, one hears about a ”postpositivist” philosophy of science, which, despite the vagueness and diversity of usages for this label, is intended to signal the death of positivism. Curiously, this postpositivism is meant as an obituary for the older Comtean positivism or its resurrection as logical positivism by the Vienna Circle, where abstract logic and observation were more happily joined together.
The result is that the term ”positivism” no longer has a clear referent, but it is evident that, for many, being a positivist is not a good thing. It is unlikely, then, that ”positivism” will ever be an unambiguous and neutral term for sociological activity revolving around the formulation and testing of theory and the use of plausible theories for social engineering (or in more muted form, for ”sociological practice”). Other labels are likely to be employed in light of the negative connotations of positivism in an intellectual climate dominated by ”post-isms.”
Despite this apparent eclipse of positivism by various post-isms, positivistic sociology remains a vibrant activity, albeit by other names. Because of the pejorative use of the label ”positivism,” few are willing to embrace it, but many practice positivistic sociology. What, then, are the main tenets of positivism? This question can be answered under ten general points.
First, positivism assumes that there is a ”real world” that can be studied scientifically. The social world is not an illusion, or a total fabrication of sociologists’ imaginations. It is there; it has properties amenable to investigation.
Second, positivism assumes that there are fundamental properties of the social universe that are always operative when people act, interact, and organize. While the properties can manifest themselves in a wide variety of forms in varying contexts, they nonetheless exist; and they are what drive the dynamics of the social universe. The goal of positivism is to uncover these fundamental properties, to see how they work, to develop theories on their operation, and to test these theories with systematically collected data.
Third, the theories developed by positivists should strive for some degree of formality. The making of formal statements need not invoke mathematics or some other system of formal argument; rather, all that is necessary is that concepts denoting processes be explicitly defined and that relations among concepts be stated clearly. These goals can be met with ordinary language, although if they can be converted into mathematics, this is seen by most positivists as useful though not absolutely necessary.
Fourth, in defining concepts formally, these definitions should denote aspects of the social universe such that what is encompassed by the concept is clear and, equally important, what is not is also explicit. In stating relations among concepts denoting fundamental properties of the social world, these relations can be stated in three basic ways. One is functional (in the mathematical sense), whereby variation in one concept is seen to be related to another (e.g., the level of differentiation in a population is a positive function of its size). A second way to state relations is through analytical models that specify the direct, indirect, and reverse causal effects among those forces of the universe that are seen as connected. A third procedure is historical in which events at earlier points in time are seen to cause directly, or in combination with other events, an outcome. A fourth, though less desirable (and at best, preliminary), procedure is to find the place of particular forces in an abstract category system that juxtaposes phenomena (e.g., the periodic table in chemistry or Parsonian four-functions analysis).
Fifth, the goal of all positivistic theories statements is parsimony. Reducing theories to their simplest form is always desired, whether this be a simple equation, an analytical model, a historical sequence of cause, or even a simple set of categories.
Sixth, at the same time that statements move toward parsimony, they should become ever more abstract and should seek to explain as large a portion of reality as is possible. The goal is always to explain as much of the social universe with as few principles and models as can do justice to the dynamics of the social world.
Seventh, all theoretical statements must be testable, at least in principle. Some statements can be tested directly with existing methodologies; others must be transformed (e.g., from deductions to hypotheses); and still others may have to wait for new methodologies or for specific classes of events to occur. The critical criterion is that theories be testable, now or in the future.
They must suggest by their formulation ways of operationalization.
Eighth, theories can be tested by all relevant methods: historical, comparative, experimental, survey, observational, and even simulational. No one method identifies positivism; all are useful in assessing the plausibility of theories.
Ninth, tests must always be used to assess the plausibility of theories. When tests do not support the theory, the theory must be rejected and/or revised.
Tenth, theories that remain plausible constitute, for the time being, the best explanations of the social universe. And the more theories remain plausible, the more they are made parsimonious, and the more new theories are developed to explain what has not yet been explained, the more knowledge of the nature and operative dynamics of the social universe accumulates.
