Forthcoming in
Mind.
Recreative
Minds, by Gregory Currie and
Ian Ravenscroft. Oxford: Oxford University Press, 2002. Pp. x +
233.
The imagination is enjoying a burst of
scholarly attention. Over the last decade, a number of researchers at the
intersection of philosophy and psychology have explored the nature of the
imagination. Gregory Currie has played a central role in these discussions, both
as an insightful critic and as an apologist for a ‘simulation’-based account of
the imagination. As a result, Recreative Minds, the new book by Currie
and Ian Ravenscroft has been widely anticipated by those in the field. The book
delivers. It stakes out an interesting, detailed, and provocative view of the
imagination. The account that emerges has broad ramifications for central issues
in aesthetics, philosophy of psychology, and philosophy of mind. The book is
also distinguished throughout by fair and balanced readings of their
opposition.
Like other work in this tradition, Currie and
Ravenscroft do not attempt to give an account of the creative imagination, the
capacity that led Einstein to discover relativity and Borges to invent Menard.
Rather, Currie and Ravenscroft offer an account of what they call the recreative imagination, the imaginative
capacity for putting oneself in another person’s place or ‘perspective shifting’
(p. 9). There are, according to Currie and Ravenscroft, a number of different
aspects of another person’s perspective that might be occupied through
recreative imagination. We can shift to take on beliefs, desires, perceptions.
Crucially, these recreative states have the same ‘character’ as their
counterpart states. So, belief-like imagining has the same character as belief,
desire-like imagining has the same character as desire, and vision-like
imagining has the same character as vision. Currie and Ravenscroft never define
the notion of character very precisely, but they do make quite clear that
character needs to be distinguished from content. What makes something count as
belief-like imagining or vision-like imagining, they maintain, is not the content of the imagining, but the character of the imagining. This allows
them to skirt familiar problems that face other accounts. For instance, visual imagining is
sometimes characterized as imagining that has, as part of its content, I see X (e.g. Walton, Mimesis as Make-Believe, Harvard
University Press, 1990, p. 28). This leads to obvious puzzles about whether, for
example, it is coherent to visually imagine an unseen landscape. Currie and
Ravenscroft easily evade these puzzles since they maintain that what makes
something count as visual imagining is not that it has content that tags it as visual, but
rather a visual character (pp. 30-1).
Recreative Minds offers a far reaching account of the
recreative imagination. One chapter sets out a state-of-the-art overview of
empirical work on visual imagery, making a powerful case for close parallels
between vision and visual imagery. However, the bulk of the book focuses on
propositional imaginings, and in particular belief-like imaginings and
desire-like imaginings. After setting out their account of the imagination,
Currie and Ravenscroft explore the possibility that the recreative imagination
is damaged in some clinical psychopathologies. They consider the now familiar
view that autism involves a deficit in imaginative capacities. The evidentiary
situation on this issue is now rather complicated, since a basic ability to
pretend, as well as mathematical abilities seem to be spared in autism. Indeed,
there is some reason to be optimistic that the capacity for counterfactual
reasoning is relatively intact in autism. Currie and Ravenscroft carefully
navigate this complicated literature to carve out their own interpretation of
the findings. More innovative, though, is their proposal that schizophrenia is
characterized by a different disorder of the imagination, a ‘failure to identify
imaginings’ (p. 164). Currie and Ravenscroft deftly use this idea to explain a
wide variety of symptoms characteristic of schizophrenia.
Not surprisingly, much of the book concerns
contemporary discussions of simulation theory. Simulation theory was originally
proposed as an account of how we predict and explain others’ behavior. The then
prevailing view, the ‘theory theory’, maintained that we predict and explain
behavior by relying on a tacit body of psychological information. The ‘off-line
simulation’ account held that in predicting another person’s decisions, I
exploit no specialized body of psychological information but rather I insert
‘pretend’ versions of the target’s beliefs and desires into my practical
reasoning mechanism, which then generates a ‘pretend’ decision which I use to
predict the decision of the target. Since those early years, research on
simulation and theory theory has been gravitating towards the middle. Most
theorists now maintain that both simulation-like and theory-like processes are
implicated in the everyday prediction and explanation of behavior (see e.g.,
Alvin Goldman, ‘The Mentalizing Folk’ in
D. Sperber (ed.) Metarepresentation, Oxford: Oxford
University Press, 2000; Shaun Nichols and Stephen Stich, Mindreading, Oxford: Oxford University
Press, 2003). Currie and Ravenscroft too embrace a hybrid theory. Perhaps the
central feature that places their hybrid on the simulation end of the spectrum
is their appeal to desire-like
imaginings. These, of course, are the Pretend Desires from past incarnations
of simulation-based accounts. The idea that there are desire-like states of
imagination is present in Currie’s work as far back as 1990, but Currie and
Ravenscroft offer the fullest presentation and the most extensively developed
exploration of the idea. Desire-like imaginings play a central role at several
parts of their theory. The appeal to desire-like imaginings explains how we
recreate practical reasoning (chapter 1), how simulation is to be distinguished
from theorizing (chapter 3), how supposition differs from full-fledged imagining
(chapter 2), and how tragic emotions get generated (chapter 9).
Recreative Minds deserves much more critical attention than I
have space to devote. I will focus my allotment on desire-like imaginings, since
desire-like imaginings play such an important role in the book. There is a
broader issue at stake here too, of course. If there are desire-like imaginings,
this greatly expands the available resources for explaining the phenomena
surrounding fiction and imagination.
What are the arguments for desire-like
imaginings? One familiar argument for desire-like imaginings comes from the
early simulation debates. Roughly, the argument is that if we predict others’
practical reasoning by running our own practical reasoning mechanism off line,
then we must use pretend desires as well as pretend beliefs. For to reach
accurate predictions, the simulation-system must accommodate the divergent
desires of the target. This argument is, of course, convincing only to those who
think that we predict other’s decisions by running our own practical reasoning
mechanism off line. As a result, this argument itself is entirely contentious in
the simulation debates. Currie and Ravenscroft present this argument briefly (p.
20) and move quickly on to two new arguments.
The first new argument is that in order to
explain our emotional responses to fiction we must invoke desire-like
imaginings. In normal emotion production, Currie and Ravenscroft maintain,
emotions that depend on beliefs also depend on desires: ‘where emotions depend
on beliefs, they depend on desires as well; if I didn’t desire your success or
success relevantly like it I would not be envious’ (p. 20). Hence, they argue,
the same must be true for emotions produced by the imagination: ‘And so it must
be with emotions that depend on belief-like imaginings; to be genuinely
belief-like, these imaginings ought to have emotional consequences only in
conjunction with states that are desire-like’ (p. 20). So the argument seems to
be that emotions that are generated by beliefs also must implicate desires;
hence, emotions that are generated by belief-like imaginings must similarly
implicate desire-like imaginings.
There are serious problems with this argument. First, it’s not at all
clear that emotions that are generated by beliefs must implicate current desires
as well. The central scientific tradition on emotion processing, the affect
program tradition, does not typically appeal to an agent’s current desires in
the explanation of the process that leads to an emotional response like fear,
sadness, or disgust (see, e.g., Paul Griffiths, What Emotions Really Are, Chicago:
University of Chicago Press, 1997). Rather, the standard explanation appeals to
appraisal processes that do apparently draw on information from the subject’s
beliefs about the situation. For instance, the appraisal of the details of a
described scenario can produce the basic emotions of fear, sadness, and disgust.
And, importantly, the responses can be triggered whether the subjects regard the
scenario as fact or fiction (see e.g., Paul Harris, The Work of the Imagination, Oxford:
Blackwell, 2000). Of course, it is possible that desires really do play a
crucial role in generating all these emotions, but that the role of desires has
not been adequately articulated in contemporary affect theory. However, even if
it turns out that basic emotional processing does crucially implicate an agent’s
current desires, it still doesn’t follow that one need appeal to desire-like
imaginings to explain imagination-based emotion. For it’s quite possible that
one’s actual desires can do the
necessary work.
Although contemporary affective science does not appeal to desires to
explain the processing of emotions like fear, there are emotions for which one’s
current desires do seem to be crucial, including envy and jealousy. However, for
those emotions, there’s a real question about how they get generated by the
imagination. One possibility is that for emotions like envy and jealousy, the
imagination only generates emotions that fit with one’s current real
desires. So, I can easily generate jealousy-affect by imagining my spouse
being unfaithful. It’s not implausible that this response depends on desires I
have about my spouse. By contrast, however, I cannot easily generate
jealousy-affect by imagining that Anna Karenina (or Laura Bush) is unfaithful.
One explanation for this is just that I don’t have the relevant real desires
when it comes to Anna. I can, of course, try to figure out how Alexey Karenin
feels by imagining how I would feel were my spouse to be unfaithful. But this
need not involve any desire-like imagining. As a result, even though there are
emotions that plausibly do depend on desires, it’s not clear that these emotions
are ever generated by desire-like imaginings.
When I am sorry and upset about the fate of
Desdemona, I am not sorry that this fiction has it that an innocent and
good-hearted girl suffers a cruel fate. One might be sorry about that, deploring
that there are fictions with such unhappy outcomes. This is not what at least
many of us are sorry about; we are glad that Shakespeare’s fiction has it this
way (p. 21).
Currie and Ravenscroft are surely right that
we wouldn’t want Othello to be rewritten so that it spared Desdemona. But
this doesn’t force us to abandon the real-desire proposal. To see why, consider
how Currie and Ravenscroft continue their discussion of their Othello example:
Part of the inner tension one experiences on
watching the play derives from the fact that we experience a desire-like
imagining that Desdemona flourish, combined with a (genuine) desire that the
play be one which will ensure that that desire-like imagining is unsatisfied
(pp. 21-2).
Currie and
Ravenscroft acknowledge here a tension in our experience of Othello. They maintain that this tension
derives from the combination of a real desire for a tragic narrative and the
desire-like imagining that Desdemona be saved. But this explanation relies on an
unargued presupposition – that we experience tension when we have the desire
that p and the desire-like imagining
that not p. Notice that the parallel
does not hold for belief-like imaginings. We feel no tension in having the
belief that Lolita was written by
Nabokov and the belief-like imagining that it was not. Why then should we
experience a tension when we have the desire that the narrative be tragic and
the desire-like imagining that it not?
Of course, Currie and Ravenscroft can simply add a stipulation to the
theory here and say that it is characteristic of desire-like imaginings that
they (unlike belief-like imaginings) come into tension with their counterparts.
But consider another alternative. The tension we experience in watching Othello arises because we have
conflicting real desires about the
play. One desire is that the fiction have it that Desdemona be saved; the other
desire is that the play be tragic. I both want it to be the case (fictionally,
of course) that Othello not kill Desdemona, and I also want it to be the case
that the narrative be tragic. This proposal preserves the idea that we have a
real desire that Desdemona be saved (in the fiction), and it also explains the
tension we feel without relying on any additional stipulations.
The appeal to real desires can explain many of the reactions to fiction
that are charted by Currie and Ravenscroft. But real desires needn’t do all the
work here. In some cases, our reports about fiction reflect our emotional
responses to the fiction. In the case of Desdemona, this seems especially
plausible. When we say, with Currie and Ravenscroft, ‘I am sorry and upset about
the fate of Desdemona’, we might merely be reporting the affect that follows
imagining her murderous end. As we saw above, there is no reason to assume that
emotional responses to fiction require desire-like imaginings. Having the
belief-like imagining that Desdemona is being unjustly murdered might be
sufficient to generate the sorrow and upset.
Given these resources for explaining our reactions to fiction, it’s
unclear that there are any cases left that would require desire-like imaginings.
But is there any way to test whether or not desire-like imaginings exist? If
there are desire-like imaginings, then we should be able to generate desire-like
imaginings that will modulate our responses to the same belief-like imaginings.
So, if we see Othello once
with desire-like imaginings that match our own desires and once with desire-like
imaginings that match the desires of, say, Iago, then we should have, according
to Currie and Ravenscroft, vastly different emotional and moral experiences of
the play. However, as the intriguing literature on imaginative resistance
suggests, we have a very hard time
doing such things (e.g., Richard Moran, ‘The Expression of Feeling in the
Imagination,’ The Philosophical
Review, 1994; Tamar Gendler, ‘The Puzzle of Imaginative Resistance,’ The Journal of Philosophy, 2000). That
is, it’s notoriously difficult to occupy the emotional and moral perspective of
someone with alien desires. We can’t get ourselves to respond the way we would
if we had the desire-like imagining that Desdemona meet a grisly demise. That
fits well with the hypothesis that there are no desire-like imaginings. Currie
and Ravenscroft have a different interpretation of the phenomena: ‘it is harder, much harder, to get people to
desire in imagination against the trend of their own real desires than it is to
get people to believe in imagination against the trend of what they really
believe’ (p. 23).
By adding the qualification that it’s hard to generate desire-like
imaginings that diverge from one’s own desires, the desire-like imagining
proposal becomes dangerously insulated from evaluation. It also leads to a
puzzle for Currie and Ravenscroft’s broader views. If it’s so difficult to have
desire-like imaginings against the trend of our real desires, how then do
desire-like imaginings play such a vital role in predicting the practical
reasoning of people with divergent desires? In fact, we seem to be tolerably
good at predicting the practical reasoning of those with nefarious desires. When
Iago acquires Desdemona’s handkerchief from his wife, we have no trouble
predicting important aspects of his practical reasoning – that he’ll incorporate
the handkerchief into his plan to frame Desdemona and that he will take care not
to let Othello discover how Desdemona lost the handkerchief. If we find it so
hard to generate divergent desire-like imaginings, then our success at these
predictive tasks can’t be explained by appeal to desire-like imaginings. Of
course, the account of desire-like imaginings might be qualified further to
accommodate this discrepancy. But as the qualifications grow, the plausibility
surely ebbs.
Department of Philosophy
SHAUN NICHOLS
College of Charleston
Charleston, SC 29424
USA