Wolfgang Schwarz

Humean Necessitarianism?

So I was given a replacement computer now until the other one arrives. If you're waiting for a sign of life from me, I'll probably contact you soon.

But first some philosophy. I want to argue that necessitarianism is compatible with Humean recombinatorialism because powers aren't intrinsic in the sense relevant to this. I also want to suggest that in an ontology of powers, what's fundamental aren't really the powers, but the causal or nomic relations.

Necessitarianism is the view that properties like mass and spin have their causal or nomic role essentially: if a property doesn't behave like mass, it isn't mass. It follows that the laws about mass are metaphysically necessary. (There are many different views in the vicinity here, maybe more about this later.)

This is usually taken to contradict a certain Humean principle of recombination according to which what happens in one part of a world is metaphysically independent of what happens in other parts:

Another use of the principle [of recombination] is to settle -- or as opponents might say, to beg -- the question whether laws of nature are strictly necessary. They are not [...]. Episodes of bread-eating are possible because actual; as are episodes of starvation. Juxtapose duplicates of the two, on the grounds that anything can follow anything; here is a possible world to violate the law that bread nourishes. So likewise against the necessity of more serious candidates for fundamental laws of nature. (Lewis, Plurality, p.91)

I believe that this is wrong.

What exactly does the Humean principle demand? Not that any properties of one thing are compossible with any properties of other things: an instance of roundhood cannot coexist with an instance of inhabiting a world without round objects, and arguably an instance of being Saul Kripke cannot fail to coexist with a distinct instance of being an ancestor of Saul Kripke. That's why Lewis puts the principle either in terms of intrinsic duplicates: intrinsic duplicates of any parts of possible worlds can be patched together in another possible world, or in terms of fundamental intrinsic properties: every distribution of fundamental intrinsic properties is possible. Inhabiting a world without round objects and being an ancestor of Saul Kripke are not intrinsic (let alone fundamental), so the above counterexamples fail: you can't have Kripke without an ancestor of Kripke, but you can have a duplicate of Kripke without any duplicate of an ancestor of Kripke.

So the question is whether, on the necessitarian view, the fundamental powers are intrinsic. If not, we don't get a violation of the Humean principle (or at least not in any straightforward way).

Many necessitarians (incl. Ellis, Mumford, Heil, Molnar and Bird) say that the powers really are intrinsic. Indeed, one of the five essential characteristics of powers listed in Molnar's Powers is that "powers are intrinsic properties of their bearers".

But what do they mean by "intrinsic"? The term is used in many slightly different ways in philosophy (see the Stanford Encyclopedia entry), and necessitarians often aren't very explicit about their usage. Whatever they mean, they probably wouldn't agree to this:

If F and G are intrinsic properties, then whether or not something x has F is metaphysically independent of how G is distributed over things distinct from x.

Otherwise powers could hardly (all) count as intrinsic: you cannot freely distribute starvation and bread-eating, or mass and velocity over distinct objects in a world.

However, this independence of what happens elsewhere is precisely what Humeans like Lewis have in mind when they speak of intrinsic properties: if one thing's having F is metaphysically incompatible with another (distinct) thing's having G, then F and G cannot both be intrinsic. (I'm not entirely sure this is strictly entailed by all of Lewis's five definitions of "intrinsic". If not, I think that's a problem with the definition.)

The necessitarians seem to have a different notion in mind, on which a property is intrinsic iff things have it somehow "in virtue of themselves", rather than in virtue of other things and their relations to them. On this conception, being Saul Kripke is probably intrinsic, whereas being a duplicate of Saul Kripke is not. For Lewis, it's exactly the other way round.

So we have to distinguish the Humean notion of intrinsicness ('H-intrinsic') from the necessitarian notion ('N-intrinsic'). Correspondingly, we get a Humean version of the Humean recombination principle and a necessitarian version.

Necessitarianism threatens the necessitarian version of the principle. But this version is almost certainly false anyway, even disregarding necessitarianism: being Saul Kripke can't be freely recombined with other properties, unless Kripke's essential properties are all H-intrinsic, which seems unlikely.

By contrast, the Humean version of the principle is not threatened by necessitarianism at all, because powers aren't intrinsic in the relevant sense. (In fact, the principle is arguably analytic when "intrinsic" is explicated in the Humean way. That isn't a problem, I think. What was it supposed to be: empirical? synthetic a priori?)

So necessitarians and Humeans don't disagree about recombination. Rather, they disagree about which properties are H-intrinsic: Humeans count mass and charge and spin as (H-)intrinsic, necessitarians don't.

Can we put it like this: for Humeans, the fundamental properties are all (H-)intrinsic, whereas for necessitarians, they are all (H-)extrinsic? It depends on what we mean by "fundamental".

In Lewis's framework, it doesn't make sense to suppose that fundamental properties might not be intrinsic. They are so more or less by definition: the central job of fundamental properties is to provide a basis for intrinsic sameness and difference between things, especially worlds. Extrinsic properties can't do that.

But suppose the necessitarian holds not only that real properties have their causal/nomic role essentially, but that they are actually individuated by this role: no two properties could possibly share exactly the same causal/nomic profile.

Then if all facts in a world are determined by the distribution of 'fundamental' powers, and the distribution of those powers is in turn determined by the distribution of causal/nomic profiles, everything ends up being determined by the distribution of some kind of causal/nomic relation. As this relation is second-order -- it holds between properties rather than particulars --, it is hard to say whether it should count as intrinsic or not. But a necessitarian could (and probably should) accept that arbitrary distributions of this relation are possible. Then it would come very close to being H-intrinsic.

There is of course something very unHumean about taking causal/nomic relations as fundamental. Also, for a Humean, everything is determined by the distribution of first-order properties over particulars, not by the distribution of second-order relations over properties that, apart from this distribution, have no intrinsic such-ness, just as for the Humean, the particulars have no intrinsic this-ness over and above their ordinary properties. It also seems that the mere distribution of nomic second-order relations will leave open how many instances the first-order properties have. Maybe it becomes a bit more transparent if one replaces the second-order relations by plural relations: x1,...,x7 stand in N to y1,...,y110; etc.?

Anyway, there certainly is a conflict between Humeanism and necessitarianism. But it has less to do with recombination than with the question whether causal/nomic properties are reducible to ordinary first-order properties or vice versa.