Are Unicorns Possible?

I think we should be skeptical.

Here’s why.

First, what we think of as a unicorn is really just a mishmash of different concepts—a Frankenstein-like fusion of being a horse, being horned, being winged (sometimes?), and whatever else. But simply imagining this combination doesn’t make it a real possibility. Nor is mere conceivability a reliable guide to possibility.

For what it’s worth, that final sentence isn’t redundant. Traditionally, a distinction is made between imagining something—literally forming a mental image—and conceiving something, which involves grasping a concept, even if we can’t picture it. There are things we can conceive but not imagine, like a polygon with an enormous number of sides (a chiliagon—bless you!). But there are also things we can conceive that are nonetheless impossible—like a square circle. That might sound controversial, but it shouldn’t be. I can conceive of a square circle in the sense that I grasp the concept, but in doing so, I also recognize that it entails a contradiction, which means it cannot exist. (For what it’s worth, deriving a contradiction is a reliable guide to real impossibility. But mere consistent conceivability—where a concept appears free from contradiction—is not a reliable guide to real possibility.¹)

The usual stock objections against conceivability as a guide to possibility are actually quite strong—at least strong enough to warrant initial skepticism. I can conceive of a man without a height, but that obviously doesn’t mean such a man could exist. I can also conceive (in some broad or loose sense) of ammonia-based water, yet such a thing is impossible. In fact, the latter example highlights what I take to be the central—and, I would argue, devastating—problem with the idea that conceivability reliably tracks real metaphysical possibility: namely, that there are hidden de re necessities that “overflow” the usual meanings of our concepts or terms. By hidden, I mean necessities that exist but remain unknown until revealed by empirical or scientific progress. Water was always H₂O, but it was only much later in human history that we came to know that water necessarily has a particular chemical composition.

There may be all sorts of biological de re necessities that make something like a unicorn impossible. What are they? Well, your guess is as good as mine. But the mere fact that such necessities exist in nature should make us at least mitigated modal skeptics—at least, that is, skeptical of “far-out” possibilities, ones that aren’t local to us and that we can’t ground in what we know about actual beings and their powers.

For example, I don’t need to be skeptical about my ability to become a sandwich artist at Jersey Mike’s. Humans—as experience shows—have that power, and I’ve actualized enough of my own abilities (speech, hand-eye coordination, etc.) to make it a real possibility. I could probably give a decent job interview. I already know how to ask, “Do you want this Mike’s way?” So I don’t need to be skeptical about possibilities that, for the most part, are practically significant.

But a unicorn? That’s not practically significant—it’s pretty far out. And for that reason, I should be skeptical about whether such a being could actually exist. (Of course, that doesn’t mean unicorns definitely couldn’t exist. I’m just saying you should be skeptical about the possibility, not outright deny it—or assume it’s possible just because you can imagine it.)

Just as I should be skeptical about brains in vats and conceivability arguments for dualism—which I am.

1. Nor will appealing to ideal conceivability—that is, what a conceiver free of cognitive limitations would grasp, etc.—help here (unless we’re talking about omniscient God). The problem remains that even an ideal conceiver could be deceived by hidden de re necessities. In other words, modal truths (truths about what must be the case) aren’t always discoverable simply by reasoning about concepts. Reality has baked-in necessities that can only be uncovered empirically—if even that.

Moreover, I think we actually have to know the essence of something before we can determine what is truly conceivable, rather than just bringing to mind a faulty or deficient concept. Thus, much of what happens in modal arguments based on conceivability is really just putting the metaphysical cart before the horse.