Causes, Explanation, and Nonpartitionability

Cause and explanation are vexed terms in interpretations of science (Woolgar 1981; Latour 1988a); in fact, they are in social science more generally (Lloyd 1986; Miller 1991). I am interested enough in the “heterogeneous constructionist” sense of causality and explanation reflected in Taylor (2005, Chapter 4) to explore it through the rest of the book. But I also recognize the importance of getting other interpreters of science interested in paying attention to scientists’ diverse resources (Taylor 2005, Chapter 5). This post, therefore, is intended to help readers position my account in relation to their own and other positions.

It is a lengthy note, but to shorten or omit this discussion would have reduced the chance of my attracting some sympathetic philosophers to relate my position to the terms of established philosophical ones, e.g., Mackie’s INUS conditions, Harré and Madden’s (1975) causal powers. Yet, this note avoids technical philosophical terminology, aiming to meet the needs of some general readers as well as those of philosophical specialists. (One specialist reference with particular affinity to the formulations to follow is Hart and Honoré 1959; see, in particular, the introduction and first two chapters.)

Descriptions favor certain explanations over others; by selection and juxtaposition they give weight to different factors and imply that some things happen because of previous and ongoing things. And, if there is a “because,” a how or why question is being answered, that is, an explanation is being given.

This said, I need to distinguish my formulation of causes and explanation from many others. I am not referring to the big-C Causality underlying grand trajectories of social development—e.g., in modernization theory or versions of Marxism—or to causes as the source of deviations in specific societies from those essential trajectories (McLaughlin 1989). My model of causes is not the physical sciences’ fundamental causes exposed one by one through suitably designed, repeatable, controlled experiments; nor do I want to promote explanation in the sense of statistical regularities. (The cases in Taylor 2005, Chapter 4 should indicate my willingness to explain singular, nonrepeatable situations; see Lloyd 1986 and Miller 1991). Finally, I am not interested in covering law generalization-abstractions for interpreting science.

This last formulation is the focus of Latour’s (1988a) polemic against explanation in general, in which he opposes tying a range of outcomes in one variable realm (science) to some feature of a relatively stable realm (society). Latour does not see social life as stable or as a realm separate from science. He wants to highlight the novel coalitions and outcomes involved in the production of science and society. Although I agree to a large extent with this perspective, the very network accounts he advocates build on multiple, diverse causes (Latour 1987).

Causal analysis of heterogeneous webs, as I formulate them (Taylor 2005, Chap. 4), proceeds in the spirit of historical explanation. That is, the analysis must attempt to identify numerous causes, no one overshadowing the others, but each in context making a difference and all together providing a composite or conjunction of conditions sufficient for readers to see why or how the situation or outcome under consideration happened and not some other possibilities (Miller 1991; Taylor 1994). In this sense, explaining is synonymous with giving an account of causes, that is, of conditions that would, if changed, make a difference to the outcome.

Some elaboration and refinement are called for. Causes and counterfactuals are bound together (see also future post on Counterfactuals); explainers choose to address certain contrasting possibilities and not others, and to allow for certain conditions to be changed and other conditions to be backgrounded because they are fixed or taken for granted. An audience-specificity tends to be imported by such explanatory choices. That is, although the hypothetical “if changed” need not be construed as the condition actually being changeable, acceptance of an explanation is enhanced, in practice, if an audience can be found or enlisted that considers that the condition is (or was) changeable. (Causes and realizable changes, not just hypothetical counterfactuals, tend to be bound together.) Explanations in the sense of sufficient composites should, in light of explanatory choices and their audience-specificity, always be viewed as provisional. They are subject to competition from other composites and likely to be superseded if the categories and detail of one of the others allows the reader to imagine more intimately how the agents were acting in the given situation (Foucault 1981; Humphries 1990).

Nonpartitionability of causes is an important feature of heterogeneous constructionist explanations (see feature i of the list at the end of Taylor (2005, Chapter 4). Given a number of focal causes, each of which make a difference in context, that is, in conjunction with other causes and background conditions, any analysis of the effect of a cause must operate jointly, not cause by cause. But, taken to the extreme of “everything is contingent on everything else,” this joint causation would be impossible for explainers and their audiences to handle. In practice, explainers discount some conditions as incidental, focusing, for example, on the modeler’s choice of software but not of, say, his use of baking soda for toothpaste. A variety of factors can take simplification further. For example, explainers can combine focal causes or background conditions into synthetic or structural conditions (e.g., centralized government policy-making); they can minimize the number of focal causes by shifting some or most to the background or incidental categories; and they can focus on causes of a similar kind (e.g., ones relating to the farmer’s decisions and not ones that spans the international economy and politics). This homogenization can even go so far as to lead to gross dichotomous categories of causes, such as natural and social, which are then held to “interact”—a bland term, often a smoke screen for the belief that one category is important, the other incidental. Heterogeneous constructionism, however, does not want to homogenize to that extent. While simplification is necessary for explanation, there is no logical or empirical reason for not choosing an intermediate level of simplicity or complexity, where the causes remain multiple and of diverse kinds. (The idea of diverse kinds assumes that we still permit ourselves the linguistic convenience of classifying into distinct kinds causes that are intricately interlinked.) At this level, it is difficult to “forget” that causes are causes-linked-in-context, and thus it is difficult to partition relative importance or responsibility for an outcome among separate types of cause. The character and implications of intermediate complexity (Taylor 2005, Chapter 5, section C), in which explanations preserve heterogeneity of causes and their interlinkages, warrant more attention from philosophers.

Finally, note that nonpartitionability does not mean all causes are equally important. Instead it means that causes are linked in context, so that the size of a cause’s effect is conditional on the given interlinkages. Similarly, something is not a resource until linked or mobilized. In principle, this conditionality includes backgrounded conditions, but, in practice, as long as change in the background conditions is not being considered, the conditionality tends to be omitted from accounts.

Excerpted from the Notes section of Taylor, Peter J. 2005. Unruly Complexity: Ecology, Interpretation, Engagement. U. Chicago Press.

Foucault, M. (1981). “Questions of Method: An Interview with Michel Foucault.” I&C 8: 3-14.
Harré, R. and Edward H. Madden (1975). Causal Powers: A Theory of Natural Necessity. Oxford: Basil Blackwell.
Hart, H. L. A. and A. M. Honoré (1959). Causation in the Law. Oxford: Clarendon Press.
Humphries, J. (1990). “Enclosures, common rights, and women: The proletarianization of families in the late eighteenth and early nineteenth centuries.” The Journal of Economic History 50(1): 17-42.
Latour, B. (1987). Science in Action: How to Follow Scientists and Engineers through Society. Milton Keynes: Open University Press.
—— (1988a). “The politics of explanation: an alternative,” in S. Woolgar (Ed.), Knowledge and Reflexivity: New Frontiers in the Sociology of Knowledge. London: Sage, 155-176.
Lloyd, C. (1986). Explanation in Social History. Oxford: Basil Blackwell.
McLaughlin, P. (1989). “Obstacles to a new sociology of agriculture: The persistence of essentialism.” Working Paper, Department of Rural Sociology, Cornell University.
Miller, R. W. (1991). “Fact and method in the social sciences,” in R. Boyd, P. Gasper and J. D. Trout (Eds.), The Philosophy of Science. Cambridge, MA: MIT Press, 743-762.
Taylor, P.J. (1994). “Shifting frames: From divided to distributed psychologies of scientific agents.” Proceedings of the Philosophy of Science Association 2: 304-310.
—— (2005). Unruly Complexity: Ecology, Interpretation, Engagement. Chicago: University of Chicago Press.
Woolgar, S. (1981). “Interests and explanation in the social study of science.” Social Studies of Science 11: 365-394.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s