Confessions of a Modal Realist, Part 1

gakuran.com

by Leonard Finkelman

Like Massimo, I attended the 2012 American Philosophical Association Eastern Division conference the week before last. Unlike Massimo, I didn’t attend the “problem of counterfactual philosophers” session. I could have; in that sense, I suppose that I am one of those counterfactual philosophers. Or perhaps I’m not, but I might have been. It’s all very complicated, and so we philosophers make work for ourselves.

There’s a kind of poignancy in the conference ending with a discussion of counterfactuals. As Massimo noted, the APA traditionally holds its Eastern Division conference on December 27-30, and so the last session ends with mere hours left in the calendar year. That’s generally a time that people spend looking back on the past twelve months, taking stock of all that did happen, all that didn’t happen, and all that might have happened — that is, factual events and counterfactual ones.

Talking about counterfactuals is especially poignant for me. You see, I am a modal realist. By taking this position, I place myself in a small minority of philosophers; by publicly admitting to it, I place myself in an even smaller subset of that minority [1].

To call modal realism controversial would probably be generous. Most philosophers to whom I’ve spoken see it as akin to belief in Santa Claus: probably harmless, but definitely naive. If they would suspend their condescension long enough to consider the theory in earnest, though, I think they might find it as a nice complement to the ongoing effort to naturalize philosophy.

So here we go: my two-part examination of one of the skeletons in Philosophy’s closet. First up is a description of the theory; next will be my take on how it can be more palatable than respectable philosophers admit.

It all comes down to the deceptively simple act of thinking about how things might have gone differently. I had eggs for breakfast, but I could have had cereal instead; since I do have a box of Cheerios in my cupboard and a carton of milk in my refrigerator, surely that latter part is true [2]. But what makes it true? Certainly not me, who didn’t actually eat cereal; certainly not the cereal, which actually remains uneaten; certainly not anything else in this world, all of which seems somewhat irrelevant.

If you’re a naturalist, or a metaphysical realist, or an empiricist, there’s a problem with reference in a counterfactual proposition like “I might have eaten cereal.” It seems like the statement is true of me, but not by virtue of anything I’ve actually done. This is a question of what some philosophers call truthmakers: since a counterfactual proposition is about something that did not actually happen, it seems that there’s nothing in the actual world that makes the proposition true. We take simple counterfactual propositions to be true, but their failure of reference to anything actual makes it difficult to evaluate their truth. It can be a bit of a paradox.

In the past century, logicians have done quite a bit to work out the truth-valuation of apparent failures of reference. For example, a proposition like “the tyrannosaur beside me is totally awesome” doesn’t have a truth value prima facie — there is no dinosaur beside me that is totally awesome and no dinosaur beside me that isn’t totally awesome — and so seems to contradict the Law of the Excluded Middle, which is one law that even a financier couldn’t violate. Bertrand Russell developed a way of avoiding the contradiction: when I say “the tyrannosaur beside me is totally awesome,” I mean that “there exists a tyrannosaur beside me that is totally awesome,” and that latter proposition is clearly false. A good number of logicians now endorse this “theory of descriptions,” or at least some modified form of it, as the best way to evaluate factual propositions.

There’s a similar way of elaborating counterfactual propositions so that we can give them truth values. I may say that I might have had cereal, but what I really mean is something like, “there exists a possible world on which I had cereal.” The truthmaker of the counterfactual proposition, then, is not me; the proposition is made true by a possible world different from the actual world. This insight is the basis for various systems of modal logic, which evaluate arguments made using counterfactual claims [3].

This is not the controversial part. Most logicians agree about what they call “possible world semantics.” No, the controversy comes in when we consider what is meant by “possible world.”

Thinking about possible worlds comes naturally. It’s a fairly standard trope in science fiction. But fiction seems insufficient: we don’t (and shouldn’t) think of the movie Shrek as a truthmaker for the possibility of talking donkeys [4]. But (as long as you accept one of the theories linked above) stories don’t make propositions true; reality does. So why should our ability to imagine possible worlds amount to anything?

Ready for the controversial part now? Modal realism is the view that counterfactual propositions are made true by possible worlds and that these possible worlds really exist. (Balderdash!) So when I say that I might have had cereal, what I mean is that there is a set of spacetime coordinates — i.e., a universe — completely disconnected from ours that includes an alien being that almost exactly resembles me who had cereal for breakfast on this day in his universe. (Poppycock!) When a modal realist says, “there exists a possible world wherein...” the modal realist means it. (Foofaraw!)

It’s a relatively simple theory. It resolves the truthmaker problem easily. Almost everybody to whom I’ve ever spoken thinks it’s incredibly stupid. I did not mistype that last part.

Okay: I’ll admit that there are some obvious problems. First of all, this doesn’t seem to resolve the problem of reference since counterfactual claims about me are made true by some other-worldly alien who just happens to be a lot like me [5]. It seems like we’ve exchanged failure of reference for mistaken reference. Additionally, to say that other possible worlds are spatio-temporally disconnected from ours is to say that they’re completely unobservable in principle, which is suspiciously convenient (like certain other concepts that come to mind).

These objections are spurious, and the second more so than the first. Yes, possible worlds are unobservable, but so too are free will and causation and identity overtime. That’s why this is metaphysics rather than, say, physics. If you don’t like it, then I’m sure that Sam Harris or Lawrence Krauss would love to sell you a book. For my part, I stay away from fiction.

Unlike me, a majority of philosophers do prefer fiction, or at least the idea that possible worlds are fictitious constructs. We get these fictional worlds through the imaginary rearrangement of actual objects, properties, and/or events. So: I didn’t have cereal this morning, but Joe McMadeupguy did and we can easily imagine me in Mr. McMadeupguy’s place. Philosophers tend to prefer this account, but this is hardly a better way of avoiding the first objection to modal realism. If anything, it should seem less plausible that Joe McMadeupguy is a truthmaker for a claim about me; hell, I’ve never even met the man, and at least my otherworldly counterpart bears a nearly-complete resemblance to me [6].

Worse than that: on the fictionalist account, any counterfactual proposition should be strictly false, since (by definition) there would not exist any possible worlds on which the counterfactual proposition is true. I’m sorry, modal fictionalists; surely you expected me to think that the words you use really mean what they do, in fact, mean [7].

To be fair, a more serious objection can be raised against the traditional presentation of modal realism. Following David Lewis, modal realists tend to assert the existence of an infinitude of possible worlds: there exists a possible world for every true counterfactual proposition and we may construct infinitely many counterfactual propositions. But this seems to get the correspondence theory of truth backwards. It’s like saying that true propositions make possible worlds real, rather than real possible worlds making propositions true. That would indeed be hogwash.

I’ll admit it: I have no good response to that objection. Semantics is a poor guide for metaphysics. If it weren’t, then you should be able to physically separate the color black from my hair, or have a left side of a square without a right; after all, these concepts are semantically separable from one another. So it would indeed be a problem for modal realism if its metaphysics depended on semantics.

I do think there’s a way to rescue the theory. The theory that gets rescued may not be the classical version of modal realism, but it will at least preserve the core. I daresay it might even prove appealing to the empirically-minded among us. But that will have to wait until the sequel to this post...
___

[1] Understand: to utter these words in polite academic company is to invite the sort of scorn normally reserved for the Anthem Foundation table in the APA conference book publisher’s hall. The idea is so offensive that iOS simply won’t allow me to place the words “modal” and “realism” next to one another (“modal” is consistently auto-corrected to “moral”). When I finally worked up the courage to share this secret with a close colleague, he lowered his voice an octave and intoned, “You must never say that during a job interview.”

[2] I would ask you to sit down, determinists, but it’s not like you have a choice in the matter.

[3] These systems of logic are more useful than I’ve made them sound. It’s very difficult to speak meaningfully of things like laws of nature and causation without helping oneself to counterfactual claims.

[4] At 6:13 pm on 17 October 2012, I got an unprecedented number of colleagues — three! — to agree on this point. Quod erat demonstratum!

[5] It’s an interesting fact about modern philosophers that a good proportion of us can name the losing Democratic candidate for president in the 1968 election. Most of us have read the book wherein Saul Kripke famously proclaimed that the actual Hubert Humphrey probably wouldn’t get much comfort from his counterpart’s success.

[6] If you want to get technical about it, there are more subtle problems with this “counterpart” relation. To wit: Lewis says that I may have multiple counterparts on a single other possible world, and that this relation is ultimately stipulative; however, if we’re just going to stipulate the counterpart relation, then we might as well stipulate the truth of counterfactual claims. But stipulating the truth of propositions is one hell of a slippery slope. Why not just stipulate the truth of everything I say? Actually, that sounds like a good idea. Never mind.

[7] For more, see John Divers’ 1995 paper on modal fictionalism, which would have been an important influence on me if I had known about it ten years ago.