Imagination is one primary way in which we can get outside our own skin to get a sense of other people’s internal lives. When my friend tells me what it’s like to enjoy roller coasters (which I don’t), I can imagine the features she mentions—an enjoyable thrill, the expectation of the turns around the bend, the wind in her hair—to get a take on the feelings that would involve. On a more serious note, when I read accounts of traumatic experiences I may never have—e.g. Kang Chol-Hwan’s stunning account of his ten years in a North Korean prison camp—my access to how that felt for him, ‘from the inside,’ is by way of the imagination. Not only do we sometimes enjoy imagining what it would be like for our lives to be very different; we really should and must do this to understand the conditions of our friends across the world.

Getting this right, in many cases, seems like leaving behind certain aspects of myself in thought. For instance, if I personally am trying to imagine enjoying a rollercoaster, I need to set aside and ignore the fact that I feel a pit in my stomach just thinking about getting on a coaster. Or if I’m focusing on Kang’s memoir just to imagine what it was like for him, I need to leave out the basic conditions of my own life—including my political freedoms and my security in my culture—to even begin to perform my imaginative transformation into him.

I might be good at doing this kind of thing—that is, ignoring all the irrelevant details of your own life to inhabit the perspective of another in imagination. But even if I can do that, there’s a leftover question to ask about all your imaginative travels into other people’s worlds: must I always imagine myself being different—or can I simply imagine being different, full stop?

The difference might sound merely grammatical, but it’s not. It is about whether the content of all our imaginations really has to involve the concept “I” or “me.” In other words, it has to do with de se content, which essentially involves the first-person perspective.

Several philosophers—notably, Kendall Walton—think that all imagining is imagining of myself. This applies universally, even when it seems on the surface like what I’m imagining doesn’t have to do with me or any other person at all. Take a silly case: imagine that pigs can fly. I can do that. (So can you.) When I do, am I imagining anything of myself at all? Walton, surprisingly, says “yes”: imagining that pigs can fly is just a way of pretending, of myself, that I believe that pigs can fly. I bring my concept “I” or “me” along even when I’m ignoring all sorts of features of myself in the imagined scene.

This view can lead to odd places, though. There are things I can imagine that shouldn’t seem possible to imagine if all imagination is of myself in that sense. For instance, I can imagine what it would be like if all animals went extinct. (A barren, quiet earth.) This would be a world with no first-personal perspectives in it at all. 

But if all imagining is of myself, what would this imaginative project involve? Imagining, of myself, that I believe that all animals have gone extinct? That would involve sticking a first-personal perspective—a believer—precisely where she doesn’t belong. It would, surprisingly, involve attributing a false belief to myself in imagination. That doesn’t seem to be a fair description of what it takes to imagine—as I really can—that all animals have gone extinct.

Moved by these thoughts, other influential philosophers have denied that all imagination is imagination of myself. Bernard Williams, for example, claimed that one kind of imagination—the “participation” kind—does not have to be imagination of myself at all. In particular, this applies to imagining being someone different, like Napoleon. It seems possible to imagine being Napoleon. I might imagine fighting on horseback, suffering a crushing military defeat, and so forth. But I literally could not have been Napoleon. It’s impossible, as it is impossible for me to have been anyone else. 

That’s part of why it’s uncomfortable to say that this imaginative exercise would involve imagining, of myself, that I am Napoleon. To do that would be to imagine something completely and utterly impossible. So, Williams suggested, we should be more careful to say that such imagination is not imagining of myself that I am Napoleon; it’s just imagining being Napoleon. That involves pretending (mentally) to do all sorts of Napoleon-like things. Although in fact it’s me who’s imagining, I disappear from the content of the imagination on this view. I am not myself in the world imagined when I imagine being Napoleon; there’s just Napoleon and his own world.

This debate matters because there’s a relationship between what we can imagine and what is really possible. Philosophers often use a supposed link between imaginability and possibility to come to conclusions about alternative possibilities. The idea is that our ability to imagine some scenario—at least with a certain degree of explicit detail—should imply that the scenario is metaphysically possible. (That means that it could have been the case if the world was different than it is in relevant ways.)

But if all imagination is imagination of myself, and I can imagine of myself something that is manifestly impossible, then imagination starts to look worse as a means of access to other possibilities than the one in which we live. That’s one reason we might buy the Williams line on imagining being Napoleon. It’s not really imagining something impossible at all, if you accept what he says about it. That might let us save the link between imaginability and possibility.

I’m tempted to say, with Williams and against Walton, that not all imagination is first-personal. I don’t have to bring myself along when I imagine my way into other people’s lives; indeed, I don’t need to bring any first-personal perspective into a world that I imagine. You can simply decline to use your concept “I” or “me” inside the world of your imaginative fantasy.