Modal Epistemology: Knowledge of Possibility & Necessity

Author: Bob Fischer
Categories: Epistemology, Metaphysics
Word Count: 998

Alice hits Betty, and Betty gets mad. Is her anger justified? Betty thinks so. After all, Alice didn’t need to hit her; Alice could have controlled her temper.

This justification for Betty’s anger seems reasonable. However, it isn’t reasonable unless we know that Alice could have behaved differently, and we don’t know that unless we know that it was possible for her to have behaved differently.

Do we?

"Imagine The Possibilities" by Carol Groenen — “Imagine The Possibilities” by Carol Groenen

We’re now in the realm of modal epistemology, which concerns how we know what’s possible and necessary. Let’s break that down.

First, “modal.” Many sentences can be true or false. For example, “I had tea with breakfast” is a false sentence (I had coffee); “2 + 2 = 4,” of course, is a true sentence. But sentences don’t just differ in terms of their truth or falsity: they also differ in terms of their mode of truth or falsity. Although it’s false that I had tea with breakfast, it might have been true: I could have made tea. However, “2 + 2 = 4” is true and couldn’t have been false—there is no way for 2 and 2 not to equal 4. In other words, though I didn’t have tea with breakfast, it’s possible that I did; by contrast, it’s necessary that 2 + 2 = 4. These two claims—about what’s possible and what’s necessary—are modal claims.

Now, “epistemology.” Very roughly, epistemology is the study of the scope and sources of our knowledge. In other words, epistemologists ask how much we know, as well as how we know it. So, for example, epistemologists ask whether you know that you aren’t dreaming right now. (How could you tell?) And, if you do know that you aren’t dreaming, epistemologists wonder how you know it. What explains your knowing that you aren’t dreaming?

Modal epistemology, then, is the study of the scope and sources of our modal knowledge. So: which modal claims do we know, and how, exactly, do we know them? Applied to the question about Betty’s anger: do we know that Alice could have done otherwise, and if so, then what explains our having that knowledge?

There are two popular answers to the scope question: the conventional answer and the moderate skeptic’s answer.

The conventional answer is that we know all sorts of things about what could and couldn’t be: for example, that there could be naturally purple cows, that there could be a device that instantly transports you to Mars, that you couldn’t be a peanut. The conventional answer to the scope question is linked to the conventional one about sources: namely, that we know what is and isn’t possible based on what we can imagine. According to this line of thinking, you know that there could be naturally purple cows because you can imagine them, and you know that you couldn’t be a peanut because you draw a blank when you try to imagine yourself as a legume. On this view, we definitely know that Alice didn’t need to hit Betty.

The moderate skeptic isn’t so sure. She might agree that we should answer the source question with reference to the imagination, but she worries that we rarely imagine things in sufficient detail. This leads her to doubt that we know as much as the conventional answer suggests. After all, what are we imagining when we say that we’ve imagined a teletransportation device? Are we picturing someone stepping into a box in one location, disappearing behind some falling glitter, and then appearing in another location? If we don’t imagine the mechanism that makes this possible, then why not think that we’re just imagining some special effects? (Bad answer: because we stipulate—that is, just decide and proclaim—that it’s a teletransportation machine. But we can stipulate that we’re imagining a circular square behind a curtain, and surely there can’t be any circular squares.)

Unfortunately, though, the moderate skeptic is at risk of becoming a complete skeptic. Recall Betty’s claim that Alice could have done otherwise. Can we imagine that in sufficient detail? Do we understand how the world would need to be different for Alice to behave differently? If not, then perhaps we’re imagining a nicer scene—one where, say, Alice gives Betty a hug—but not one that’s possible.

The moderate skeptic seems to raise an important worry about the conventional answer, but complete skepticism seems untenable: we do know that Alice didn’t need to hit Betty. What’s going on?

Perhaps we went wrong when thinking about the source question. Granted, imagining seems to be behind many modal judgments. But maybe this mental operation isn’t the source of our modal knowledge, but instead draws on another source of modal information. This is the idea behind essence-based modal epistemologies. According to these views, we figure out what things are essentially—what it is to be a rhododendron, or a human being, or a quasar; the properties that make those things the things they are, and without which they wouldn’t exist—and then that knowledge informs our imaginings. So if we know that every detail in a person’s history isn’t essential to her, we can know that she could have done otherwise.

The obvious trouble with essence-based views is that they push the problem back: how do we know about essences? This seems no easier a question than: how do we know what’s possible and necessary?

Such worries lead some philosophers to start with more mundane ways of gaining knowledge: namely, analogical and inductive arguments. For example, I’ve seen other houses repainted, and mine doesn’t seem to be relevantly different; so, mine could be repainted. Likewise, I know that people make many different decisions about how to behave. Alice doesn’t seem unlike those people in any important respect. So, probably, she could have made a different decision. These simple arguments probably won’t take us very far: for example, they might not tell us what to say about teletransporters, where various analogies might point in various directions. But they might deliver a lot of ordinary modal knowledge, and that may be good enough.