Transparency and imagining seeing

Abstract

In his paper, The Transparency of Experience, M.G.F. Martin has put forward a well-known – though not always equally well understood – argument for the disjunctivist, and against the intentional, approach to perceptual experiences. In this article, I intend to do four things: (i) to present the details of Martin's complex argument; (ii) to defend its soundness against orthodox intentionalism; (iii) to show how Martin's argument speaks as much in favour of experiential intentionalism as it speaks in favour of disjunctivism; and (iv) to argue that there is a related reason to prefer experiential intentionalism over Martin's version of disjunctivism.

Keywords:

• disjunctivism; hallucinations; imagination; intentionalism; object awareness; perception; transparency; visualising

Acknowledgements

I thank Malcolm Budd, Peter Goldie, Rob Hopkins, Mike Martin, Kevin Mulligan, Matt Nudds, Gianfranco Soldati, Matt Soteriou, Juan Suarez and the members of the philosophy research colloquium at the University of Fribourg for many beneficial discussions about intentionalism and disjunctivism. In addition, I am very grateful to two anonymous referees and the editor of this symposium, Marcus Willaschek. My work on this article was generously supported – in the form of a Fellowship for Advanced Researchers – by the Swiss National Science Foundation (grant no. PA00P1-126157).

Notes

See, especially Martin (2003, 417ff.). McDowell (1998) and Fish (2009) defend versions of disjunctivism, which – in many relevant respects – come close to that defended by Martin and discussed here.

See Dorsch (2010a), as well as Dorsch and Soldati (forthcoming), for more detailed discussions of these differences. Among the current proponents of intentionalism, which have written explicitly on visualising and experiential imagination, are Hopkins (1998), Noordhof (2002), Currie and Ravenscroft (2003) and Burge (2005). Others are Dretske (1995), Tye (1995) and Speaks (2009).

I use the term ‘experience’ to denote not only sense experiences, but all instances of conscious object awareness, including episodes of visualising and of experiential imagination. The expression ‘perceptual experience’ is meant to cover both (veridical) perceptions and hallucinations, and I distinguish between the latter two independently of whether they differ in nature or not. Besides, when I speak of ‘hallucinations’, I mean to refer, if not stated otherwise, to perception-like hallucinations – that is, hallucinations which are subjectively indistinguishable from perceptions (Dorsch 2010a).

See Martin's distinction between the semantic and the stative conception of representation in Martin (2003).

The resulting characterisation of intentionalism is meant to be neutral between its various versions, including both naturalistic and non-naturalistic conceptions of its nature. That is, the first-personal presentational and attitudinal aspects of perceptions may – but need not be – further elucidated in terms of mental representation and functional role. In Dorsch and Soldati (forthcoming), we argue that at least reference to representation is not very helpful, since it ignores the essential connection between intentionality and consciousness or subjectivity. A similar issue with functional role is whether it is the same as rational role, or whether it is non-normative in nature. If the latter, we again contend that intentionality cannot be separated from normativity.

This conclusion is not necessitated, since perceptions and hallucinations might possibly differ in aspects of character that are not linked to how they make us aware of objects and features. But it is not clear what aspects that could be, or why it should be plausible to assume their existence.

This difference is sometimes also spelled out in terms of perceptions ‘acquaintaining’ us with things, or making them ‘manifest’ to us (though intentionalists can presumably adopt at least the first manner of talking). Besides, it can be ignored here what disjunctivists do, or should, say about hallucinations and their subjective indistinguishability from perceptions. It is perhaps defensible to argue that they still make us intentionally aware of objects (Smith 2002). But it seems more natural for disjunctivism to conclude that hallucinations do not make us aware of objects at all – they just seem to do so.

Indeed, it may very well be that the transparency at issue is, in fact, inseparably linked to the presentation of objects, rather than the entertainment of propositions. When introspecting experiences, the external things that we find are shown to us; and no internal objects are given to us in this way. The issue of whether a given thought is transparent in this sense does not arise then, since thinking is not an instance of object awareness.

Martin (2003, 37 and 39) mentions the example of mentally rotating a piece of furniture in a shop in order to see whether it is possible to get it through one's front door at home.

The claim may become trivial, if it can be established that object awareness is, by its very presentational nature, transparent and commital (see footnote 8). But to establish this is not a trivial task.

Note that Martin uses ‘experience’ here as short for ‘sense experience’, such as perception or bodily sensation. As already mentioned, my own use is less narrow in also including, say, visualising or other imaginative instances of object awareness.

A further possible locus of misunderstanding is perhaps that (ii) understands both visualising and imagining a perception as instances of object awareness, given that some intentionalists tend to construe visual experience in terms of thought-like contents (see, for instance, Dretske (1995) and Tye (1995)).

It is true, though, that others – such as Peacocke (1985) and O'Shaughnessy (2003) – have put forward stronger versions of the claim, extending even to all kinds of sensory imagining. And Martin is clearly sympathetic with this more general conclusion, as can be witnessed in Martin (2003, 404ff.) and Martin (2001).

See section 3 in Martin (2003), which is mostly occupied with the development of his argument in favour of thesis (ii), the restricted version of the Dependency Thesis. In what follows, I draw heavily on this section of Martin's paper. I am also very grateful to the challenging questions about this section raised by Marcus Willaschek, which helped me greatly to get clearer about certain details of Martin's argument, as well as about my own addition to it at the end of this section.

It does not really matter for Martin's main argument whether we are concerned here with two different sets of spatial properties of objects (e.g. one objective, and the other subjective), or instead with two modes of presentation of one and the same set. What is relevant here is primarily the fact that our perceptual access to spatially located objects is perspectival and, in particular, presents them as orientated towards us, rather than in more objective terms. But many of the points involved in the argument can be described more easily by reference to egocentric properties. Besides, the postulation of subjective orientations is not much different from the postulation of subjective modes of presentation (cf. the similar issue raised below with respect to the aspect of leftishness and similar phenomenal aspects).

In what follows, I concentrate on the fact that perceptions present objects in actual space and mention the temporal dimension only when it becomes relevant.

It should become clear very shortly that there is a third possibility: the experience may represent another experience as instantiating or presenting the quality.

Using the expression ‘being to the left’ to denote a monadic property is not ideal, since this expression clearly has some connotations of relationality. But it is not easy to come up with another formulation, without altogether loosing the connection to the perceived property of being to our actual left. I am grateful to Marcus Willaschek for making me aware of this issue.

Very similar issues arise, for instance, with respect to the status of the quality of ovalness – another perspectival aspect of perception – which figures in our experience when we are looking at objects from an angle and perceive them as round. Again, we typically draw round objects by tracing elliptical shapes on the canvas. But it is debatable whether our experiences present round objects as elliptical in addition to presenting them as round (see, for instance, Peacocke's (1983) discussion of what he calls sensational properties). One significant difference from egocentric orientation is, however, that, while roundness is an objective property, being to the left of us is not.

Of course, Martin cannot assume in his argument that an experience's instantiation of the aspect of leftishness is also sufficient for the existence of something to our actual left. This would follow only if the experience is a perception, and if perceptions are always factive – something that intentionalists deny.

There is also the issue of whether visualising always locates objects relative to us, rather than to some imagined subject. The default case seems to be that we visualise objects as orientated towards ourselves, and that imaginative projects involving subjects different from us require the additional identification of our imagined point of view with that of those other subjects (Wollheim 1984; Martin 2003, 411).

The possibility of experiencing a phantom limbs as hurting is no exception. The only difference is that, in this case, the existence of the hurting body part is subjective as well. But, in any case, nothing here depends on whether the presented view on pains and pain experiences is correct. The analogy is merely meant to further illustrate Martin's treatment of the involvement of subjective elements in imaginative experience.

Compare also the famous example of two incongruent hands with different orientations in Kant (1992).

The property of being to our left shares both discussed aspects with the property of being here if ascribed to ourselves. The latter, too, cannot be specified in purely objective terms. And we cannot go wrong in being aware of ourselves as being here. One difference between our awareness of us as being here and our awareness of other objects as being to our left is, however, that the latter may concern hallucinated objects. This is not problematic for the present argument since it relies only on the claim that the instantiation of egocentric orientations requires the occurrence of a perception of them, but not on the reverse claim. Besides, as already suggested in note 22, there are perhaps other ways of reconciling the subjectivity of a property with the possibility of hallucinating an instance of it.

Currie and Ravenscroft (2003, 28) also do not address the issue of subjectivity when they briefly sketch Martin's motivation for endorsing thesis (ii).

See also Peacocke's (1985) discussion of Wittgenstein's example of King's College on fire.

See also the similar point against the Dependency Thesis made by Currie and Ravenscroft (2003, 27ff.).

The process of photo-copying, which does not involve any such perspective onto the reproduced piece of paper, is perhaps an even better illustration of the kind of representation pertaining to experiential imagination (as proposed by Martin in a personal discussion about how best to understand Hume's Copy Principle).

This idea is not new. In particular, Hume's Copy Principle may be read as claiming pretty much the same if applied to the case of imagining.

Reference to the kind of representation at issue promises also to illuminate why episodes of visualising often possess a lesser degree of repleteness, determinacy or intensity than episodes of seeing. Just as the reproduction of a painting may lead to the loss of some of these qualities, imagining perceiving something may have this effect. Martin's employment – in Martin (2001) – of the Dependency Thesis in his account of the phenomenological differences between seeing, visually remembering and visualising provides another example of the explanatory force of treating at least some instances of visualising as an instance of experiential imagination.

It is interesting to ask to which extent imagining a subjective perspective involves imagining a subject. It seems that we need not be in any way specific about the identity of such a subject when visualising – although we may also be very specific in the form of additional suppositions (Martin 2003, 411).

He does not explicitly comment on how to characterise experiential imagination. But there is no reason to suspect that he would reject premiss (iii.2).

Again, this is false for disjunctivism, according to which perception-like hallucinations are non-commital because of their lack of any presentational elements. This means, among other things, that the intentionalist proposal of an object-independent non-neutrality of perceptions cannot be used to argue against the disjunctivist view that the non-neutrality of perceptions is due to their being constituted by their objects (and, indeed, vice versa). What we have here are two rival accounts of perceptual commitment, which are so far diactically on a par. The last section of the paper is meant to change the balance concerning this issue in favour of intentionalism.

See Martin (2003, 2004) and Dorsch (2010a). Similar ideas can be found in the intentionalist views inspired by the phenomenological tradition, such as that put forward in Smith (2002).

But see the end of Dorsch (2010a) for a line of reasoning which supports the supplementation of experiential intentionalism with structural disjunctivism.