Illustration 1: the formalistic modelling of social processes
Consider first the ease of method in modern economics. The dominant feature here is the widespread reliance upon the practice of formalistic modelling.10 This approach has certainly been univcrsaliscd within the economics discipline, and with little if any grounding or argument, and despite its record of failure. Feminists have also criticised it as masculinist. I think it is (albeit an approach that is also perhaps race and class, etc., specific). But what follows for the feminist critic? Is it that all other approaches, including any preferred by feminists, be given greater emphasis?11 Or should more feminists do formalistic modelling?12 Perhaps both responses follow; or arc they largely incompatible? Is it, say, that the scarcity of feminist modellers entails that the set of questions currently addressed is unnecessarily limited, or is it, perhaps, that formalistic methods themselves are undesirably limited in their usefulness, and possibly even debilitating of revelatory and emancipatory progress? I low do we begin to decide?
Answering such questions, questions that may or may not involve a false dichotomy, requires at the very least that the revelatory potential of formalistic methods be investigated. And this, I will argue, necessitates an attention to ontology. Specifically, I want to indicate that by briefly examining both the nature of social material and the ontological presuppositions of the procedures of formalistic modelling, it can be demonstrated that the latter procedures are not at all well equipped for illuminating the social realm. This conclusion is easily established, but rarely is so precisely because of the widespread reluctance to engage in ontology. It is just this reluctance that I wish to call into question.
The particularity’ of formalistic modelling
But is formalistic modelling really so restricted in its usefulness? As I say, I think it is and that it can he shown to he so. Let me briefly sketch my argument. First consider the sort of conditions under which formalistic modelling has relevance. Basically such modelling attempts to relate one (measurable) set of events or states of affairs to others. It presupposes correlations in surface phenomena, that is strict (possibly including probabilistic) regularities of the form ‘whenever event (or state of affairs) x then event (or state of affairs) y Let me refer to situations in which such regularities occur as closed systems. Formalistic modelling, to have general relevance, presupposes the ubiquity of such closures.
Now an observation often recorded but rarely reflected upon is that outside astronomy such event regularities, or at least those found to be of interest and significance in science in general, are mostly confined to situations of well-controlled experiments. An additional observation is that the results of controlled experiments are regularly successfully applied outside the experimental laboratory where event regularities arc not in evidence.
A recognition of this situation, then, already casts doubt on the rationality of ploughing ever more resources into producing yet additional formalistic economic models. Certainly the failure of the econometrics project over the last fifty years or so is indicative that the social world is open, that event regularities in the social realm arc far from ubiquitous. The usual response, of course, is to pronounce that we must try harder: to formulate ever more complex models using larger data sets, or to dig deeper in the expectation of finding the sought-after invariances at a more micro, or anyway different, level. However, the recognition that even in the natural realm significant event regularities arc systematically restricted, and largely found only in situations of experimental control, encourages a suspicion that significant successes in the social realm may not be possible even in principle. It is clearly essential at this stage that the observed patterning of event regularities be explained.
But how can we make sense of the observed confinement of most event regularities to the experimental set-up? Notice first that this observation generates immediate tensions for any program that insists that such event regularities are essential to science (as the indispensable objects of scientific laws including laws of nature, or some such). For it follows that science (if thought to necessitate the elaboration of event regularities) is after all not only far from universal but, outside astronomy, mostly confined to experimental set-ups; it is actually fenced off from most of the goings-on in the world. Moreover, one is bound to conclude from this that (many) laws of nature (if event regularities arc essential to them) depend upon human actions (in setting up the experimental situation), which is at least counterintuitive. But it also follows that the further familiar observation that science is efficacious outside the experiment, where event regularities do not occur, is unintelligible.
How, then, are we to make sense of these considerations? How is it that scientists, in their experimental activities, can (frequently) codetermine a particular pattern of events that would not have come about but for their intervention? And how can we make sense of the successful application of science outside of the experimental laboratory, and specifically in conditions in which event regularities do not necessarily occur? What must the world be like for such experimental practices, results, and their successful noncxpcrimcntal application to be possible?
