A post by Greg Currie.

I have argued for a certain view about what goes on when, while sitting in the theatre, we “want Desdemona to be saved” as we might unguardedly put it. On my view, this is a case of what is called desire in imagination or sometimes i-desire. And on that view i-desire is not desire. I-desires stand to desires as imaginings stand to beliefs.[1]

In a recent post on The Junkyard, Peter argues that my own view gets me into the same sort of difficulty that I took to be a reason for rejecting an alternative view which he accepts and which makes do with desire only and has no need of i-desire. The alternative view is that what is really going on is that I desire that, in Othello, Desdemona be saved. I accused that view of attributing needless contradiction to audiences, since they (or some of them) know how the play ends and wanted the play to end that way—they chose to see it because it is a tragedy which ends with Desdemona’s death. If they had not wanted that they would have gone to see a comedy instead.

In the course of arguing I gave the example Death of Prime Minister, an imagined drama according to which Mrs Thatcher is pursued by an assassin. While being a supporter and admirer of her in real life, the gripping drama puts me on the side of the assassin. “I want her to be killed” I say as the drama unfolds. On my view that should be understood as requiring that I i-desire that she be killed, not that I desire it. Indeed I don’t desire it, and when a real assassin chases Mrs Thatcher into the theatre where I am sitting and shoots her, I am horrified. That was not what I wanted! I wanted something the satisfaction condition of which was that it be so according to a representation (the drama), not that it be so. That’s enough, on my account, for something to be an i-desire.

Peter then imagines that I am engaged with two fictions, the second being narratively similar to Death of a PM but so different in tone that it provokes me to say “I don’t want her to be killed”. Imagine that I have watched the first half of both the previous evening and want to conclude both this evening. It seems I have to say that I have (persisting) contradictory i-desires: one that she be killed and one that she not be killed. So I am just as guilty of needless contradiction as my opponents are.

I might try to get out of this by saying that we properly specify the i-desire that she not be killed in fiction A, and i-desire that she be killed in fiction B. I don’t want to say that, because I hold that these are cases of i-desiring which are character-focused rather than fiction focused; they belong to what goes on within the scope of the imagining we engage in when we engage with the story and in such cases we don’t want the fact that this is a fiction to figure in the content of our imaginative project (except in odd cases where we are dealing with Pirandello-type fictions).

Note that problems of exactly this kind arise without needing to consider the controversial case of i-desires. Take imagining in the ordinary sense. Given the variations in plot between these two assassination narratives, there will be things it is appropriate for me to imagine relative to the one which contradict things I am supposed to imagine relative to the other. It would seem I am imagining contradictory things. Peter sees this and suggests, towards the end of his post, that “ordinary” imaginings require specification of the fiction in their contents. As we will see, I don’t believe we need to do that.

From my point of view, the following is central to Peter’s line of thought:

Another response Currie might offer is that the satisfaction conditions of each state relate to distinct fictions and that this is why there is no conflict between the two.  But this would require that something other than an i-desire’s content and attitude determines whether it conflicts, or not, with another i-desire.  At that point, the notions of content and attitude are no longer doing the explanatory work we expect of them.

I take it that the response he is objecting to is this: in the case of the ascription of i-desires, when we specify what states someone is in we should not say merely that they i-desire such and such; we should say that, relative to a certain fiction, they i-desire such-and-such. That is not the same as the strategy I just rejected; i-desiring that P-according-to-F is not the same as i-desiring at P, relative to fiction F. The latter is the proposal that I want to elaborate on here.

Let me introduce at this point another kind of case that may be familiar: the case of acceptance in a context. Bratman distinguished belief and acceptance on the grounds that belief is context-free while acceptance is context-bound.[2] The captain who believes that the ship is close to the rocks simply believes that; she does not believe that while guiding the ship and fail to believe it later while safely on shore; she continues to believe what she previously believed, though she will now express it in the past tense.[3] But the captain who accepts and does not believe that the ship is close to the rocks and who, in consequence, contrives to think and act as if the ship were close to the rocks—taking extra care with navigation—ceases to accept it when the danger is past.

What I suggest is that, in this, i-desires are like acceptances. Perhaps they actually are acceptances, though different in kind perhaps from the ones Bratman was thinking about; I take no view on this.

Note that Bratman’s captain may persist in accepting that the ship is close to the rocks outside of moments when she is actually guiding the ship; the ship may be in a tricky position over several hours or days and it makes sense to say that she continues to accept that it is close to the rocks throughout that period, though at various points others take the controls. Indeed, the situation may be sufficiently complex that the captain adopts different policies as regards the steering of the ship moment by moment, putting into effect different policies of acceptance as seem appropriate to the changing circumstances. In such a situation it seems to me right to say that she is in various acceptance states through that whole period: accepting (but not believing) that the ship is in one kind of danger and that it is in a danger of some other, incompatible kind. Is the captain in contradictory states? It seems to me better to say that she accepts P relative to one context and accepts ¬P relative to another and that there is no contradiction in doing so. After all, we want to distinguish cases like this which don’t involve any irrationality from cases of irrational acceptance, as when the captain becomes so confused that she accepts P and ¬P in the same context, something she would be well advised to avoid.

This seems to me a helpful analogy for understanding the situation of the viewer of the two rival dramas that feature assassinations of Mrs Thatcher. I i-desire that she be killed relative to one fictional context and i-desire that she not be killed relative to another. Both cases involve the same attitude and the same content. Peter’s objection is that “At that point, the notions of content and attitude are no longer doing the explanatory work we expect of them”. I agree that they are not doing the explanatory work. I don’t agree that this is a failure. Something else is doing the explanatory work: the idea of relativity to a context. This is an idea we have need of to understand acceptances. Since we need it, why can’t I make use of it? 

[1] See my Tragedy, Analysis, 70, 1-7, 2010. “Desire in imagination” is my term (Desire in Imagination, in T. Gendler and J. Hawthorne (eds) Imagination, Conceivability and Possibility, Oxford University Press, 2002). The term “i-desire” is due to Andy Egan and Tyler Doggett (Wanting what you don’t want. Philosophers’ Imprint 7: 1–17, 2009).

[2] Bratman M. (1992) Practical reasoning and acceptance in a context. Mind 101: 1–15

[3] She may of course cease to believe it because new evidence comes in. The point is that belief does not change simply because context changes.