The story so far: Concepts of philosophical categories such as knowledge or justice, are, or gain their cognitive significance from, explanatory theories of the relevant domain (involving epistemic explanation in the case of knowledge and moral explanation in the case of justice). Thanks to the way that concepts semantically hook on to their corresponding categories, such theories provide reliable starting points for inquiry into the categories’ essential natures. In this post I return to the question whether the fruits of such inquiry are worth the long and often tedious effort of analysis.
Why worry? I gave one reason in my first post this week, that our starter concepts are mere “first drafts” of a useful taxonomy for the domain in question. A first draft is not likely to be anywhere near as good as the final draft.
Another reason is formulated by Machery in Philosophy Within Its Proper Bounds. Experimental philosophers have shown that there is great variation in the judgments made by ordinary people concerning important thought experiments such as TrueTemp, Gödel/Schmidt and even Gettier cases. One explanation for the variation is that different minds harbor different concepts, united only by their attachment to a shared natural language word. It would follow that your analysis of, say, knowledge would divulge the essential nature of your knowledge category, but not of the categories of thousands of other upstanding citizens. Such an analysis can be of only parochial interest.
A third reason: decades, in some cases centuries, of analysis have shown that the essential natures of such things as knowledge and singular causation are rather baroque. The chances of such a finicky construction mapping significant boundaries in the underlying explanatory landscape is surely slight.
In response, let me begin by pointing out that many of the same disconcerting arguments might be made about certain natural kind terms, such as “fruit.” Both the human race’s, and individual humans’, starter concepts of fruit are presumably biologically rather uninformed. Ordinary people may differ in their judgments about “edge” cases such as avocados, olives, and rhubarb. And attempts to capture such judgments in a set of necessary and sufficient conditions are unlikely to look pretty. Despite all of this, botanists will tell you exactly, what makes something a fruit, namely, its being a plant’s seed-bearing structure.
It might be said in this case that there are two concepts of fruit, one culinary and one botanical. Regardless, it is clear that botany began with a fruit concept as rudimentary as any starter concept, and converged on an explanatorily meaningful essential nature.
Can philosophical analysts hope to do the same? Their concepts, on my view, have the right kind of structure: they are, like basic natural kind concepts, collections of explanatory hypotheses, open to correction in the light of further information. In contrast with the case of basic natural kinds, however, sensory experience cannot be relied on to offer much help. (Well, not always: some categories of philosophical interest, such as “physical probability” and “concept” itself, are susceptible to considerable revision in the light of empirical information. But once again, I’ll put these cases aside.)
Is there any prospect of improving our philosophical starter theories in other ways? There must be: we can, for example, make moral progress, renouncing torture and slavery and embracing equal rights regardless of race, gender, sexual orientation, and so on. No doubt this progress is spurred by lived experience, but it is justified by arguments that are available in the armchair. Why shouldn’t reflection allow us equally well to make epistemic and metaphysical progress?
One worry is that our theories may be unable in principle to represent explanatory connections in the more esoteric domains. For that matter, our ability to represent explanatory connections like causation in the material domain has been subjected to searching scrutiny. I must concede that I have simply assumed, without comment, that we have the power to think explanatorily about the material, epistemic, and moral worlds.
A second worry concerns negative outcomes. What if the analyst concludes that there is no such thing as knowledge? Is that a conclusion of great interest, of great substance? Not, perhaps, if the category of knowledge is a purely parochial creation—then, it would hold the same interest as finding that nobody in your department shares your birthday. But I think that, because a positive outcome to analysis—a correct conclusion concerning knowledge’s essential nature—would be interesting, a negative conclusion would also worthy of note. It would tell us that there are no significant explanatory features in the vicinity.
A final worry attracts even more attention in Thinking Off Your Feet: When I decide that a classic Gettier case is not knowledge, I feel tremendous confidence in my judgment. No amount of new information, I believe, could prompt me to reconsider. That suggests that the parts of my knowledge concept guiding the judgment are not revisable. They are nailed down by the very nature of the concept, like a definition. But in that case, there are parts of my starter theory of knowledge that are fixed. The worries about parochialism and insignificance come flooding back.
I don’t have space here for an adequate treatment of this difficulty. Suffice to say that the near unrevisability of the elements of the starter theory in question owe their fixed aspect to a feature that makes them anything but parochial or baroque. They might be mistakes, but if so we can discover this and it would be really quite fascinating to do so.