Imagination, Transparency, and Attention

Cain Todd is Senior Lecturer in Philosophy at Lancaster University, and his research centres on issues in aesthetics, perception, and philosophy of mind. His most recent work has focussed on attention, on the content and phenomenology of emotion and imagination, on the representational capacities of olfaction, and on affective distortions of temporal perception.    

Cain Todd is Senior Lecturer in Philosophy at Lancaster University, and his research centres on issues in aesthetics, perception, and philosophy of mind. His most recent work has focussed on attention, on the content and phenomenology of emotion and imagination, on the representational capacities of olfaction, and on affective distortions of temporal perception.    

A post by Cain Todd.

It seems plausible to think that visual experiences are in some way transparent. There are various ways of articulating the claim, and various theoretical motivations for defending it. But the basic idea is that, when introspecting our visual experiences, reflecting on what it is like to have them, our descriptions primarily – or perhaps solely – report on the objects and properties that those experiences are experiences of. In other words, few or no properties intrinsic to the experiences themselves are revealed to us, and we are only aware of their representational content. One helpful way of putting the claim has been given recently by Matthew Soteriou, who, like Amy Kind, makes a distinction between weak and strong transparency:  

  • Strong Transparency: “introspection of one’s perceptual experience reveals only the objects, qualities, and relations one is apparently perceptually aware of in having the experience.” (Soteriou 193-4)
  • Weak Transparency: “when one introspectively attends to what it is like for one to be having a perceptual experience, it seems to one as though one can only do so by attending to the sorts of objects, qualities, and relations one is apparently perceptually aware of in having that experience.” (Ibid.)

Given that visual mental imagery has often been taken to be, in many respects, very similar to, and perhaps parasitic on visual experience, we might wonder to what extent, if any, our experiences of imagery are transparent. The question is not merely scholastic, since if imagery – and the imagination more generally – is implicated in ordinary visual experiences, as some philosophers hold, then the transparency of imagery might have implications for how we conceive of perceptual (and other types of) experiences involving it. 

One immediate obstacle, however, is that one finds very different notions of imagery in the literature. Here are two:

1) Imagery that is more or less voluntary, in being subject in principle to the will; not constrained by current perceptual experience; the result of memory or imaginative activity; and that is essentially conscious.

2) Imagery that plays an essential role in perceptual experience, and that may be relatively involuntary and unconscious. This is the notion of imagery that figures in discussion of, for example, amodal perception and has played a role in Kant, Husserl, and Strawson, amongst others.

If we focus for the moment on the first type, we can ask whether the voluntariness of such imagery figures in the phenomenal character of such experiences; that is, is the voluntary nature of such imagery phenomenologically evident to us? I contend that it is, or at least can be when attended to, partly because it seems to me that we can often, if not always, introspectively tell the difference between imagistic and perceptual experiences, and indeed other states, solely on this basis. This has a bearing on how we think about imaginative transparency. 

If we reformulate Soteriou’s distinction in terms of imagining rather than perceiving, then strong imaginative transparency would mean that when I attend to my experience of imagery, I am only aware of my imagery. Weak transparency, on the other hand, would seem to mean that when I attend to the experience of my imagery, I can only do so by attending to my imagery.

Given that most philosophers seem to agree that images themselves are not the intentional objects of imagining, however, we need to proceed carefully here. It is not obvious how to apply the idea of perceptual transparency to imagination. On the reading just given, it would seem that if imagining is strongly transparent, then when I visually imagine X, my imagining is transparent to the image by which I imagine X. If imagining is weakly transparent, then when I visually imagine X, my experience may involve awareness of the image in virtue of which I imagine X, as well as certain features of my experience of having the image of X.

I think it is difficult to make sense of the strong version just outlined, because it’s not clear what sort of awareness of imagery is involved. Putting that aside, one reason for thinking that imagining might be strongly transparent (to imagery) is given by O’Shaughnessy:

“in imagining that p, one has as the focus of one’s mental attention only the state of affairs picked out by ‘p’. It follows that, among other things, mental attention to one’s epistemic relation to ‘p’ is excluded for the duration of the imagining. Hence in occurrently imagining that p, one is not occurrently thinking of the fact that one only imagines that p, or does not believe that p, or that ‘p’ is not true.’ (362)

However, three things must be noted about this. First, he’s talking only of propositional imagining, and so it is not obviously applicable to sensory imagining employing imagery. Second, the ability to fully and successfully imagine p is subject to degree – like attention – and hence can be more or less successful. Third, even if “mental attention to one’s epistemic relation to ‘p’ is excluded for the duration of the imagining”, this seems compatible with some (perhaps peripheral) awareness of imagining. Otherwise, as I just noted, there’d be no clear way of distinguishing, phenomenologically, between perceiving and imagining p. There is, however, a phenomenological way of doing so that is compatible with weak transparency: this is the awareness of voluntariness when imagining with imagery. So, it seems plausible to suggest that imagination qua experience of imagery is at best weakly transparent (to imagery). Weakly because we are aware of what I call ‘phenomenal voluntariness’.

This clearly applies to the first notion of imagery outlined above, but what about the second one? Nanay, who is an advocate of the second, argues that imagery and perception have the same content. So, if perception is strongly transparent, then so too is imagery. Perceptual content is constituted by the properties that are perceptually attributed to the perceived scene, and imagery attributes various properties to various parts of the imagined scene. One aspect of this is that attention, he argues, makes the content of each – specifically the properties attributed in each – more determinate. The only difference, then, is that the determinacy of perceptual content comes from sensory stimulation, while the determinacy of imagery comes from my memories (or what I take to be my memories) or my beliefs or expectations.

However, I don’t think that attention can make imagery more determinate, or at least not in the same sense. Rather, it either adds more detail in terms of extra properties:  this is the familiar thought that imaginings notoriously lack ‘completion’ (and/or saturation) e.g. when imagining a speckled hen there need be no determinate number of speckles that I imagine. Or, attention makes imagery more vivid, where this admittedly vague notion seems to imply something more like ‘bright’ than determinate.  In either case, it is doubtful that attention to imagery can be selective in the same way as it is perceptual experience, where there are features of the visual field that can come into and out of focus.

If these thoughts are on the right track, then imagery does not have the same content as perception. It has a different structure and is related very differently to attention. If this is also present in the phenomenology of imagining, as I would maintain, then the role of attention also suggests that imagining is weakly transparent. Moreover, it seems to show that the second notion of imagery proposed above might really be referring to an entirely different type of state.

As I noted at the beginning, how we conceive of the transparency of imagining has implications for experiences in which imaginings play an essential role. I happen to think that all types of evaluative experience – aesthetic, moral, emotional, and others – involve a kind of imaginative construal. If that is right, then such experiences should, at least on reflection, strike us as being only weakly transparent, and possessing experiential features of which we are aware and that differ from cases of ordinary non-evaluative perceptual experience. I have outlined this idea elsewhere, but there remains a great deal of further work, both philosophical and empirical, to do.