A post by Sophie Grace Chappell.

Part 2 of 3

In my first post I introduced the notion of imaginative identification, and said how crucial I think it is for ethics. I also suggested, by reference to examples, that for the most part, modern moral philosophy has not made good sense of imaginative identification. That raises the question why not.

I think there are about six reasons why there’s been this failure. (I don’t mean that anyone has articulated all six of them; since these are reasons for a silence, some of them may never have been clearly expressed at all.) Three of these reasons are to do with the nature of imaginative identification itself. There is an eliminativist worry about intentional psychology; a Wittgensteinian one about the truth-conditions and the verification of claims about “what it’s like for someone else”; and a Leibnizian worry about the logical coherence of the very idea of adopting someone else’s viewpoint. And three of the reasons are about the nature of moral philosophy: its impersonality, its rationalism, and its moralism. In this post I’ll look at the three reasons to do with the nature of imaginative identification; in the next I’ll look at the reasons to do with the nature of moral philosophy. 

The eliminativist worry about imaginative identification just notes its deep involvement with intentional psychology. If you are dubious about intentional (or “folk”) psychology, as eliminativists are, then you have to be dubious about imaginative identification too. For imaginative identification means identification with others’ intentional psychology; and this presupposes that others (and I myself) have intentional psychologies for me to identify with. Conversely, you might take eliminativism’s worries about intentional psychology to be more a reason to reject eliminativism than to reject imaginative identification. Since this latter is closer to my own attitude to eliminativism, I say no more about it here.

The Wittgensteinian worry is a problem about how I know that my attempts to imaginatively identify with you are successful or unsuccessful. But note: a problem about how, not about that. Saying exactly how A can usually tell that B has given an accurate and perceptive account of how things are for C, and that D has not, may be tricky and complex. But that A can usually know this, and that in general we can usually tell good answers to the question “What is it like to be that other person?” from bad answers, seems to me beyond doubt. Thomas Nagel agrees:

There is a sense in which phenomenological facts are perfectly objective: one person can know or say of another what the quality of the other’s experience is… [though] the more different from oneself the other experiencer is, the less success one can expect from this enterprise.[1]

It’s just a datum of our life as social beings that we can—at least usually, at least to some extent—mind-read. Other people’s minds are not, in general, opaque to us. As Wittgenstein himself might have pointed out, it might even be truer to say that publicity is the mark of the mental than privacy.[2]

For my money, the Leibnizian worry is the most interesting of the three. One way into the puzzle that drives this third worry is the ordinary advisory locution “If I were you”. When I tell you what I would do “if I were you”, you may naturally ask how far I expect you to take the proposed identification; since if I were literally you, then it is just trivial that I would do whatever you in fact do. The scope of my identification with you, we may say, has to be weaker than that if we are to avoid this sort of triviality. Perhaps it will be as restricted as a game of cards; maybe all I mean to talk about when I use “if I were you” is what card I would play next if I were sitting where you are now sitting, playing the bridge hand that you are now playing. In other cases, however, a wider and deeper identification seems possible. I may think about what I would do, or (wider) how I would feel, if I were Anna Karenina having just seen Vronsky’s fateful riding accident; or I might wonder what it feels like to be Darcey Bussell, having just danced Swan Lake to triumphant applause at Covent Garden. This kind of imagining is (as Adam Morton and others point out) central to our experience of fiction. It is also, I think, central to the practice of imaginative identification. (More about the links between fiction and imaginative identification below.)

The puzzle is that, at the limit, the imaginative identifications seem to be complete; and we might be puzzled about how they can happen at all if they happen completely. Suppose I am not merely imagining myself, Sophie Grace Chappell, occupying the space or the clothes or the name or even the body that Anna or Darcey is actually occupying. I am imagining that I, Sophie, am that particular woman. Does that mean: my mind in her body? Not necessarily even that. Maybe part of what I am imagining is that I share Anna’s moody, melancholy psychology, or Darcey’s stoical ability to dance through the pain in her toes. Maybe part of what I am imagining is even that I share Anna’s or Darcey’s phenomenology. But in that case, just what am I supposing when I suppose that I am Darcey Bussell? What is added to the picture—which after all is ex hypothesi a picture of Darcey Bussell—by adding me to it? And what is the me that is so added? As Bernard Williams puts the difficulty in his remarkable and still seriously-undervalued paper “Imagination and the self”:

[S]uppose I conceive it possible that I might have been Napoleon—and mean by this that there might have been a world which contained a Napoleon exactly the same as the Napoleon that our world contained, except that he would have been me. What could be the difference be between the actual Napoleon and the imagined one? All I have to take to him in the imagined world is a Cartesian centre of consciousness; and that, the real Napoleon had already. Leibniz, perhaps, made something like this point when he said to one who expressed the wish that he were King of China, that all he wanted was that he should cease to exist and there should be a King of China. (Problems of the Self pp.42-3)

Now the fact that the real Napoleon already has a viewpoint on the world does not imply, as Williams seems to suggest, either that there can be as it were “no room in his head” for my viewpoint on the world. (As well? Instead?) Nor that if my viewpoint on the world were to get into his head, then it would simply become the same thing as his viewpoint on the world. What we need for it to be possible that “there might have been a world which contained a Napoleon exactly the same as the Napoleon that our world contained, except that he would have been me” is that my viewpoint on the world and Napoleon’s should be two different things; and that, in the imagined situation, my viewpoint should replace his.

So on this conception, the answer to the question “How complete can imaginative identification be?” is apparently this: in cases of the most complete imaginative identification, when I imagine myself as Darcey Bussell, what I do is imagine Darcey’s character and past and location and body and psychology and mood and phenomenology with me, Sophie Grace, not her, Darcey, as the viewpoint on the world from which all of these are experienced. The limit of my imaginative identification is the limit of my viewpoint on the world. It is possible, on the conception that I am exploring here, for me to go all the way to that limit without my simply becoming Darcey, in the way that Leibniz’s interlocutor, according to him, simply became the King of China. It is only when I go beyond that limit, when it is not merely my body or my location or my psychology etc. that becomes hers, but my viewpoint on the world too, that I simply melt into her. So the intuitive conception has a very clear answer to my other question, “What is added to the picture by adding me to it?”. Its answer is, simply, that what is added is my viewpoint on the world.

Williams apparently thinks that this intuitive conception of how imaginative identification works is deeply mistaken. He does not say clearly what exactly he thinks the mistake is. But we can hazard a guess at what he might say. He might deny that Napoleon’s and my viewpoints on the world are, as I put it two paragraphs back, two different things, so that one can replace the other. He might deny this not on the grounds that my and Napoleon’s viewpoints are the same thing, but on the grounds that neither is a thing at all. The problem with talking about individual consciousnesses as “viewpoints on the world” is (he could say) that it suggests a false analogy with ordinary, external-world viewpoints like, say, Chesil Beach, or the top of Magdalen College Tower. In that case there is a clear distinction between viewpoint and occupant. Different people can go on the Beach or up the Tower, and anyone who does will see what there is to be seen from that viewpoint, while remaining a separate person from any other possible occupant of the viewpoint. What someone brings to the Beach or the Tower, when she brings herself, is her body, her memories, her abilities, her dispositions, and the rest of it; and these are the things that make it possible to tell her from anyone else. But what can someone bring to Napoleon when she brings her viewpoint, and nothing but her viewpoint, to him? Trying to suppose that that is possibly transferable from me to Napoleon, and that we can tell the difference between Napoleon’s viewpoint on the world and my viewpoint on the world when there are no other accompaniments at all to these bare viewpoints, seems very close to incoherence. At this point Leibniz’s retort really does become attractive.

I think this line of thought probably is what Williams has in mind in his rather telescoped discussion in “Imagination and the Self”. I agree with Williams that it puts a serious obstacle in the way of anyone who thinks it coherent to suppose that imaginative identification with someone else can be complete, in the sense I’ve been exploring. The price of that supposition, apparently, is that we will have to reify viewpoints on the world, treat them as existents in their own right; and anyone who is not already a paid-up Cartesian is very likely to baulk at this price.

The line of thought is worth tracing out in full for at least two reasons. First, it is intrinsically interesting. It gives one fairly clear answer, for instance, to the question what it might mean to believe in a soul. One fairly clear answer; others are possible. But this conception, on which the soul is the reified consciousness, the viewpoint-on-the-world hypostatised, and nothing else, certainly answers to some uses that people have had for the concept of a soul: Descartes himself apparently among them.

Secondly, what we discover by tracing it out is that, in fact, the notion of imaginative identification faces no real threat from the Leibnizian argument whatsoever. The difficulty we have developed is, at most, the point that complete imaginative identification is incoherent, or is an uncomfortably Cartesian notion, or is both. (“At most”: and maybe not even that much. Even if I cannot coherently imagine that I become Napoleon while retaining my own viewpoint on the world, I can certainly imagine that I become Napoleon.) So the difficulty is a difficulty for the notion of complete imaginative identification. But then the way out of this difficulty has been obvious all along. All we need do is not insist on the word complete. Imaginative identification can just be a matter of my imagining what it would be like to be you in this or that partial respect, or during this or that limited time-span, or whatever.  

So when I imaginatively identify with you, this is what happens: I am careful to specify exactly what things are like within my imaginative identification only up to a point. Some details are necessarily left unspecified; for if we specified all the details of the imaginative-identification scenario, then that scenario would face awkward questions about how it can fit in to how things actually are. We keep the imaginative identification afloat by not pressing too hard on the details, and also by keeping things fluid: by not insisting on sticking with any one particular level of identification, but moving back and forth between different levels. (One familiar stylistic marker in literature of this kind of moving back and forth is free indirect discourse.) In these respects imaginative identification with other people is, I suggest, strikingly reminiscent of two other things we do with our minds that have interested philosophers: counterfactual conditionals, and fiction. In these cases too it is often true that we can only keep the discourse going at all by deliberately being unspecific about certain background details. (If you want to engage in the fiction that Sherlock Holmes lives at 221B Baker Street, do not attempt to visit him; if you want to explore the question “What if George Washington had lost and the British Government had won?”, do not get stuck on the rather basic objection that Washington didn’t lose.) The connections between the three activities, and between them and the notions of story-telling—and pretending—that apparently underlie all three, seem likely to me to run deep.

However that may be, the point for now is that neither the eliminativist, nor the Wittgensteinian, nor the Leibnizian worries show that the notion of imaginative identification is so hopelessly conceptually flawed that no one can use it without deep confusion. Indeed, exploring these worries serves on the whole just to bring out the emptiness of the supposed threat posed by the worries. So there is no insurmountable philosophical problem about the very idea of imaginative identification; there may be things in the notion of imaginative identification that explain why philosophers have neglected that notion, but nothing in the notion justifies this neglect.

[1] Nagel, Mortal Questions p.172.  

[2] Cp. Timothy Chappell, “What have I done?”, in Diametros special issue on intention: online at http://www.diametros.iphils.uj.edu.pl/?l=2&p=anr25&m=25&if=0&ii=40&ik=38.