I promised a surprise for today’s post. It’s a nasty one.
Philosophical analysis is a search for the essential natures of such things as knowledge, justice, and causality. I’ve been defending analysis on two fronts. First, I’ve argued that it its inputs—the case judgments delivered by our “starter theories” of the categories in question—are sufficient to direct analysts toward the truth about essential natures. Second, I’ve argued that its outputs, the essential natures themselves, are substantive matters of fact that are well worth getting to know.
But as anyone who’s played the analysis game knows, essential natures are rather hard to come by. In spite of all our efforts, we have yet to find an adequate analysis of knowledge or of singular causation. Not without reason, Timothy Williamson has written that philosophical analysis is “a degenerating research program.”
Should we worry? In Thinking Off Your Feet, I concede that we should. Analysis has come up short for the most discouraging possible reason: many categories of philosophical interest have no essential natures. There is no single thing, in other words, in virtue of which specimens do or do not qualify as category members—not even a very complicated thing.
The reason is that many philosophical categories (and indeed many basic natural kinds) function as “secondary explanatory organizers“. A primary explanatory organizer groups together things that share a single, but possibly complex, explanatory property. A secondary organizer, by contrast, groups together things that share a cluster of similar but distinct explanatory properties. There is no recipe that prescribes a strict criterion for belonging to the cluster. There is only a complex, multifaceted explanatory theory that we use, invoking inference to the best explanation, to decide category membership.
In the case of swans, for example, the theory lays down (among other things) a number of causal explanatory generalizations connecting swanhood to swans’ characteristic observable properties: something about swans causes whiteness of plumage, something about swans causes redness of beak, etc. We decide membership of the swan category by reasoning backward along the theory’s causal links. This organism has white feathers, a red beak, and so on. What could possibly explain this conjunction of features? That the organism is a swan. There are no competing explanations in the offing, so most likely, it is a swan. The swans are simply things that pass this explanatory test, for a fully informed categorizer.
Does that rule out the existence of a unified criterion for category membership? You might think not, reasoning as follows. Compile a complete specification of the circumstances under which an organism passes the explanatory test. Isn’t that an exceptionless criterion for swanhood, and so a plausible candidate for the essential nature of the swan category?
My answer (very roughly): It is indeed extensionally adequate, but we will nevertheless find it implausible as an essential nature because of its inductive character. It is manifestly a procedure for inferring swanhood; indeed, to all appearances, a defeasible procedure. So it can hardly be regarded as the definitive criterion for being a swan. But nothing else is in a position to play that role. No criterion, then, will strike the philosophical analyst as a convincing candidate for swanhood’s essential nature.
If most philosophical categories are, like swan, secondary explanatory organizers, then most will lack essential natures. Curtains for philosophical analysis?
I don’t think so. The analysis of a category can be highly rewarding even if there is no nature to be found. The analyst will, of course, fail to achieve their ultimate goal. But they will gain enough substantive philosophical knowledge along the way to have made the journey worthwhile. That knowledge may concern the category itself: we can learn that truth is necessary for a belief to count as knowledge, and that truth and justification are not sufficient. Or it may concern the underlying explanatory landscape: attempting the analysis of singular causation, we stand to learn a lot about event explanation.
The byproducts of analysis are so rich because philosophical categories, even when they lack essential natures, are—like basic natural kinds such as gold and water—potent explanatory organizers. In other words, they group together things with many explanatory similarities, that is, things, many of whose properties are explained in similar ways. Close scrutiny of the reasons that specimens fall or fail to fall into such a category is therefore close scrutiny of the explanatory landscape.