Set in the context of the philosophy of the social sciences, this paper draws out polarities between competing tendencies in contemporary academic economics. It explores how naturalistic and non-naturalistic frameworks of thinking about economics differ at the fundamental levels of their philosophical presuppositions, the manner in which their theories are formulated and appraised, and the ways in which they relate to other social sciences. "Naturalistic" and "non-naturalistic" are familiar terms in the philosophy of social science, encompassing divergent ways of depicting and explaining human behavior and relationships. Although not commonly used in economics, the terms provide an organizing schema within which to view disagreements between rival ways of "doing" contemporary economics. Constructing the argument in this way also allows it to be situated in the context of the few hundred-year-old debate between naturalism and non-naturalism in the broader field of the social sciences (economics, sociology, anthropology, psychology, and political science).
Initially in the paper, terms are defined, followed by a description and history of the two main philosophical positions in social science, naturalism and non-naturalism, related particularly to economics. Then, bold conjectures are proffered of how naturalistic and non-naturalistic social sciences, including economics, differ from each other in philosophy and methodology. Three of the conjectures are considered more closely to illustrate the differences in economics. First is the nature of theories in economics, using Becker's  economic theory of human behavior as the example. Second is the existence of causal laws in economics, and third, the degree of openness/closedness of the systems with which economics deals. Since naturalistic positions are dominant currently in these three areas (and therefore are more familiar), each is evaluated from a non-naturalistic perspective to emphasize the disparities between the two philosophical positions.
Naturalistic is a common, long-used, and generic term in the philosophy of the social sciences depicting the model of explanation for human behavior drawn from approaches in the physical sciences. It is associated usually with formalistic modelling and with empirical research in which hypotheses are tested formally via quantitative data [e.g., Lessnof 1974; Rosenberg 1988; Bohman 1991; Clark 1992]. In its explanations, emphasis is placed on the scientific representation of human behavior and relationships, at least as the term "scientific" was understood up to about 30 years ago.
In direct opposition, non-naturalism contests the relevance of traditional natural science models, positing an alternative emphasis on understanding and interpreting human action, its meanings and intentions. In these interpretations, the roles of values, judgment, and individual insight are emphasized [e.g., Rabinow and Sullivan 1979; 1987].
The dichotomy between naturalism and non-naturalism is sometimes considered no longer adequate to represent the diversity of philosophical positions in the social sciences today [e.g., Redner 1993]. One complication confronting the bifurcation has been the emerging strength in the last 25 years of the philosophical position in the social sciences of transcendental and critical realism [e.g., Bhaskar 1989]. This position construes itself as a "new critical naturalism" that has resolved the dispute between conventional naturalism and non-naturalism [Bhaskar 1993, 313]. Realism has explicit advocates in economics [e.g., Hausman 1992; Maki 1993; Lawson 1994a; Hoover 1994]. The implications of the resolution claim are considered below, where it will be argued that the traditional dual schemata is a useful first approximational heuristic for clarifying fundamental divergences between contemporary positions in economics, even though it has been little used in recent methodological debate in the area. …