Inconsistent Thoughts

A good model for conventionalism is that found in the work of Alan Sidelle (e.g. in “A Sweater Unravelled” 1998) and Iris Einheuser (e.g. in “Counterconventional Conditionals” 2006). The model is roughly as follows. The world of individual objects we are familiar with is the byproduct of two things: a substratum which is not by nature divided up into individuals (but which is divisable), and a carving which divides the substratum up into familiar individual objects.

That the substratum of the world is divisible in more than one way naturally prompts the question why there are any restrictions at all on how to correctly describe the world. More carefully: why are there restrictions on how to correctly identify individuals? The conventionalist answer is that the work of carving out individual objects from the substratum is done by our conventions of concept application. Our concepts are loaded with conditions for correct application that demand when to categorize some of the substrate of reality as an individual member of a sort. The semantic normativity of concept application is supposed to do the work of explaining why identifying anything as a gollyswoggle or an incar is strictly incorrect.

Such conventionalism is intriguing and worth investigation, but it is not what I want to talk about. I only introduce the view as a jumping-off point to talk some more about Composition as Identity (CID). When Frege and Lewis talk about the ‘external phenomenon’ or the ‘portion of reality’ which is divisible in more than one way, I take them to have in mind something like the conventionalist picture. The difference is that I take Frege and Lewis to be agnostic with respect to whether the dividing up and counting of individuals is conventional. On the other hand, I take Baxter (e.g. “Many-One Identity” 1988) to be more than agnostic, positively rejecting the sort of conventionalist story which makes individuation conditions something mind-dependent. Here is a story about how this wacky brand of realism might go.

First of all, we need to take some lessons from Sidelle. He uses the terminology of ’stuff’ or as others have said ‘world-stuff’ to talk about the substrate. What is implicit in this use of terminology is that the present conception of reality requires that ’substrate’ or whatever term we use in its place functions as a mass noun and not a count noun. For another thing, as Sidelle says “the world is not an undifferentiated blob”. It is not like we have to work too hard to figure out where boundaries of individual objects can exist. They are apparent to us as contrasts in appearance, texture, etc. This differentiation is what makes the carving out of individuals possible (what I previously alluded to as the ‘divisibility’ of the substrate). Now one way to accept a lot of this picture without going the conventionalist route is to just deny that there need to be restrictions on which possible individuations are the ‘right’ ones (the felt need for restrictions on individuation is at least one reason to find the conventionalist view plausible). The picture is that there is a substrate of reality which is divisible, but not by nature uniquely divided up in just one way, so that each possible division of the substrate into numerically distinguished, bounded individuals is as ‘right’ as every other.

On this view, the composition relation is surprisingly uninteresting. Some individuals x,y,z compose the individual A just in case the plurality x,y,z are carved out of the same portion of substrate that the singular A is carved from. It immediately follows that in one plausible sense x,y,z are identical to A. Bear in mind that they are not numerically identical, but identical in some more primitive sense that has nothing to do with ‘number’ or individuation at all. Take an example: we have a chess board, it has 64 tiles, there are 32 black tiles and 32 white tiles, there is one big disjoint criss-crossed swath of black and one big disjoint criss-crossed swath of white. According to someone like Baxter, any of these descriptions captures the same portion of substrate which is what makes them identical, i.e. the chess board is the 64 tiles. Each description also depends on a different carving, and each carving divides up substrate into individuals in a way that is distinct from each other carving. But these carvings are part of the fabric of mind-independent reality, not something imposed on reality by our transcendental egos. I think that this is at least one helpful picture of reality as the advocate of CID must conceive of it.

If this is right, it does prompt the further question why we find gollyswoggles and incars to be so jarring. Baxter sometimes suggests that there are certain carvings we care more about for our purposes, so that perhaps our language and customs have been oriented to those carvings at the neglect of all others. This would explain away the tension between our ontological intuitions and the unrestricted metaphysics on offer.

Any feedback would be greatly appreciated!