In this penultimate section, we would like to briefly illustrate how the fate of important claims about imagination made by philosophers and scientists depends on something like our taxonomy of experiential imagination. Although we believe that this taxonomy has philosophical value in its own right, we also would like to show that it is connected to central issues in philosophy and cognitive science. These issues concern, respectively, modal epistemology (section 7.1), cognitive resonance (section 7.2), mindreading (section 7.2), and imaginative identification (section 7.4). Our discussion in what follows, though, can only be rather programmatic in contrast to the rest of the essay.[22]
Imagination has been traditionally construed as providing evidence for modal claims. For instance, many philosophers since Descartes have suggested that what can be imagined is metaphysically possible. On the other hand, imagination has been shown to produce various sorts of modal illusions (Kripke 1980; Gendler & Hawthorne 2002). The main challenge faced by proponents of an internal relation between imagination and possibility (perhaps via conceivability) is thus to distinguish proper and improper uses of imagination, i.e., those uses that provide, and those that do not provide, evidence for modal claims. One might suggest, for instance, that proper uses of imagination require a certain format that other uses lack (Nichols 2006; Weinberg & Meskin 2006b).
In our view, there is an additional criterion that must be taken into account in these debates, which concerns the type of non-imaginative state that is re-created by the relevant imaginings. It might be that only some types of imagination are internally related to modal properties. For instance, it is not clear that cognitive imagination is essentially related to possibility. Assuming the correctness of our claim that cognitive imagination re-creates belief, the fact that one can cognitively imagine that p is no more evidence that p is possible than the mere fact that one believes that p. After all, one can believe all sorts of metaphysically impossible states of affairs (such as that Hesperus and Phosphorus are distinct celestial bodies).
The challenge is then to identify the types of imagination, if any, that are essentially or at least reliably related to what is metaphysically possible. One hypothesis, voiced by Dokic (2008), is to focus on types of imagination that re-create states of (actual or potential) knowledge. On this hypothesis, some uses of imagination are guides to possible contents because they are guides to the possibility of knowing. To the extent that sensory perception is commonly thought to be a source of knowledge, sensory imagination could be reliably linked to the possibility of what is imagined in this way (see also Williamson 2008).
This is not to say that cognitive imagination has no role to play in providing evidence for modal claims. Just as belief can be grounded on sensory perception and thereby be counted as knowledge, a single imagining might re-create not only belief and perception separately, but the complex mental state of believing that p on the basis of suitable sensory evidence (see Dokic 2008). The resulting imagining would be neither purely sensory nor purely cognitive, but to the extent that it re-creates a non-imaginative state of knowledge, its content might be bound to what is metaphysically possible.
If we are right, there is a phenomenologically accessible distinction between objective and subjective imagination. What it is like to visually imagine an action or a painful experience is typically different from what it is like to subjectively imagine acting or having pain. However, this distinction is rarely made explicit in the scientific literature on the neural underpinnings of imagination. Let us consider the case of action. It has been a remarkable discovery that observing and executing an action involve (at least sometimes) the same resonance system in the brain, and more precisely the same “mirror neurons,” corresponding to types of action such as grasping, reaching, or eating (Rizzolatti et al. 1996; Rizzolatti et al. 2001). What about imagining an action? Marc Jeannerod claims that “imagining a movement relies on the same mechanisms as actually performing it, except for the fact that execution is blocked” (Jeannerod 2006, p. 28). Does this claim concern objective imagination, subjective imagination, or both? On the one hand, his notion of “motor imagery”, defined as “the ability to generate a conscious image of the acting self” (p. 23), strongly suggests that he is talking about subjective imagination. Motor imagery seems to underlie the imaginative recreation of an internal experience of action, such as the intimate experience we have while executing an action or controlling our bodily movements. On the other hand, Jeannerod makes clear that the “action representations” involved in motor imagery can also operate during action observation (p. 39). To the extent that visually imagining an action is analogous to observing an action, one may surmise that objective imagination too involves the relevant action representations.[23]
What we would like to know, of course, is which action representations are common to both objective and subjective imagination of an action, and which action representations are specific to subjective imagination. Here as elsewhere, we think that phenomenological considerations can at least guide scientific investigations into the neural underpinnings of our ability to imagine actions, whether imaginatively observed or imaginatively executed.
We also think that much of the once-hot debate between the “theory theory” and the “simulation theory” of mindreading has missed the distinction between objective and subjective imagination, or at least its significance. Mindreading is often described as involving ways of “putting oneself in another person’s shoes” (Goldman 2006). However, as many have observed, that colloquial phrase can be used to refer to two different projects. One might try to understand either what one would do if one were in the other’s situation or what the other will do. The difference between these meanings has been conceived as depending on whether one performs the right “egocentric shift” and succeeds in mimicking the other’s mind (Gordon 1995). If we are right, there is another distinction thatis crucial to simulation-based mindreading, namely the objective versus subjective imagination distinction. We might perform the right egocentric shift but imaginatively re-create only the other’s external experiences. For instance, we might imaginatively adopt the other’s visual point of view and try to understand what he or she is actually seeing. In doing so, though, we imaginatively adopt a perspective that is not necessarily the other’s perspective. Visual perspectives can be shared. It is only if we re-create at least some of the other’s internal experiences that we imaginatively adopt a perspective that can only be that of the agent. Unlike external experiences, internal experiences cannot be shared.
Why is it important for the success of mindreading that the mindreader re-creates also internal experiences of the other person? Let us consider the case of pain. To the extent that both objectively and subjectively imagining another person in pain may trigger the same resonance (affective) mechanisms, we can argue that they are on par with respect to the imaginer’s understanding of the other’s experience (Gallese 2003). We surmise that the relevant difference between objectively and subjectively imagining the same painful experience concerns the dynamics of mindreading. Recreating an internal perspective on pain will spontaneously give rise to other subjective imaginings involving the recreation of the mental consequences of pain in the other. Objective imagination of another person in pain will likely develop in different directions. For instance, if we re-create a visual experience as of someone in pain, we will be inclined to re-create other visual experiences of the consequences of pain. More generally, someone who would be able to re-create only external experiences of pain would be blind to the internal consequences of pain. In contrast, subjective imagination promises to yield a better view of the other’s inner life as it unfolds in time.
In this essay, we did not explicitly mention an intriguing phenomenon in the field of imagination, namely our ability to imagine being someone else, or imaginative identification. For instance, we can imagine being Napoleon seeing the desolation at Austerlitz and being vaguely aware of one’s short stature (Williams 1976, p. 43). Recanati calls such cases “quasi-de se imaginings”:
I will, therefore, coin the term ‘quasi-de se’ to refer to the first person point of view type of thought one entertains when one imagines, say, being Napoleon. The type of imagining at stake is clearly first personal, yet the imaginer’s self is not involved […]. The properties that are imaginatively represented are not ascribed to the subject who imagines them, but to the person whose point of view she espouses. (Recanati 2007, pp. 206–207)
How can an imagining be both first-personal and not genuinely (but only “quasi”) de se? If we can imagine being Napoleon just by recreating his visual experience of the desolation at Austerlitz, it is not obvious that quasi-de se imagination is necessarily first-personal. Since visual perspectives can be shared, our visual imagining can re-create anyone’s perspective. In other words, objective imagination (i.e., the recreation of external perspectives) would not be sufficient to generate quasi-de se imaginings. Perhaps Recanati implicitly ties quasi-de se imagination to subjective imagination so that imagining being someone else involves the recreation of at least some internal experience. Again, in contrast to external perspectives, internal perspectives cannot be shared. For instance, a subject imagining to be Napoleon might, on the one hand, see in imagination the desolation at Austerlitz (i.e., an external perspective is re-created) and, on the other hand, be vaguely aware of his short stature and his hand in his tunic (i.e., an internal, proprioceptive perspective is re-created).
In what sense would subjective imagination be first-personal, then? One view is that the quasi-de se case somehow derives from the genuine de se case, in which we imagine ourselves having various external and internal experiences. On this view, there is an asymmetrical dependence between quasi-de se and genuine de se imagination: even if the former is not merely a type of the latter, imagining being someone else having such-and-such experiences depends on the ability to imagine oneself having these experiences.
However, our account of subjective imagination suggests an alternative view, according to which the identity of the subject need not be built into a subjective imagining. Consider the case of action again. The constraint imposed on subjective imagination, that the imagined perspective on the action can only be that of the agent, leaves open whose self is involved. That the action is my action, or someone else’s, is an additional fact in the imaginary world. In other words, subjective imagination can be neutral as to the identity of the self that occupies the relevant internal perspective. As a consequence, the same neutral imagining can give rise to either quasi-de se or genuine de se imagination, depending on the imaginary project at stake.[24] Subjectively imagining oneself swimming and subjectively imagining another person swimming both rest on the same type of imagining, i.e., the recreation of an internal experience of the action of swimming. We take this neutrality to be a potential advantage for our analysis of subjective imagination. Subjective imagination can be seen as a basis for the introduction of a notion of self that is conceptually on a par with other selves. In this respect, imagination acts as an antidote to solipsism.