Social categories and theoretical descriptions
Discriminating between theoretical descriptions
The recognition that all social categories are theoretical descriptions, or, as Pawson (1989: 287) puts it, that 'all measurement is an act of translation', also raises the question of how to discriminate between the often-competing accounts provided by these descriptions. Our second claim is that social researchers must provide a defense involving an explicit and conscious scrutiny- of the theoretical descriptions they employ. (We assume at the very least that researchers believe that what they produce offers a different understanding from the understandings of participants, as non-researchers, of whatever aspect of the social world is being investigated. They may, in the end, produce understandings that turn out to be similar to those of non-researchers, of course, but this would still be for different reasons.) In other words, as we argued earlier, researchers must at least claim some form of epistemic authority.
Theoretical descriptions necessarily draw on other theoretical descriptions, and these other descriptions, as World 3 entities, are necessarily independent of the people who deploy them, in the important sense that they did not produce them. As propositions about reality, these descriptions can be assessed in terms of their truth or falsity, their consistency, their evidence and so forth. The importance of a realist ontology is critical here. It provides us with a firm view of the reality to which our theoretical descriptions refer; our descriptions are descriptions of something. Indeed, along with Devitt and Sterelny (1987: 190), we would argue that one cannot theorize about anything, least of all language, without implicit commitment to a view of the world. How things seem to us, in other words, depends both on the world and on our descriptions of it.
Combining a World 3 view of knowledge as theoretical descriptions produced by other minds, and upon which subsequent descriptions to some extent rest, with a robust recognition that entities do not depend for their existence or nature on our theories about them, allows for the defense of a modest notion of objectivity within the social sciences.
