Amy M. Schmitter ([email protected]), University of New Mexico
[Very
preliminary draft – do not cite, and ignore footnotes.]
Proponents of direct realist theories of
sense-perception often see themselves as championing a common-sense view of our
relation to the external world against a long and mistaken philosophical
heritage of representationalism. On some accounts, this heritage is rooted in
the early modern “way of ideas,” which – perhaps with the view of defeating
skepticism on its own turf and perhaps from confusion between the causes and
objects of perception – inserted representing ideas into the relation between
perceiver and perceived. Recent scholars, however, have suggested that many
early modern philosophers, (especially of the rationalist stripe) are not
nearly as committed to representationalism as it may once have seemed. Although
I agree with this assessment, I do not think that it follows that they were
committed to direct realism. Instead, I suggest that the distinction between
the two families of theories is harder to make than is often assumed. Indeed, I
argue that if we conceive of what is at issue as a matter of perceptual reference,
we may find it impossible to distinguish between the two kinds of theories. So,
it turns out to be a good thing that – as I also argue – many early modern
debates about the nature of sense-perception simply side-step the rivalry
between direct realism and representationalism altogether.
The main part of my paper is devoted to
making this case for Descartes by offering an alternative account of his main
concerns about our perceptual access to objects – concerns that are independent
of, and prior to, questions about directness or indirectness in
sense-perception. I argue that Descartes introduces operations of
representation to explain how we have access to meaningful, or intelligible
objects through a welter of unconceptualized sensations. Representations thus mediate between the mind and its
objective contents. But this is not merely causal mediation. Neither is it
representationalist in character. For representation does not hold between
inner, already accessible percepts and external physical objects; it operates
internally to perception to give conceptual shape to our perception – in short,
to “schematize” it.
Consider the difference between seeing Winston Churchill in a photograph and seeing him in the flesh. The difference seems obvious: seeing the man in the flesh, we see him directly. Seeing him in a photograph, or in a mirror, or in a crystal ball, we see him indirectly. Obvious this difference may be; a little less obvious is what has been made out of it – two competing theories (or families of theories) about the nature of sense-perception. On the one hand are direct realist theories, which maintain that we genuinely perceive external material things, and that at least some of our perception of them is direct. On the other hand, we find theories endorsing some form of indirect realism, or representative realism, or (as I shall call it) representationalism. These theories hold that our sense-perception of external, material things can only be indirect. Each theory purports to be global – perfectly general accounts of the nature of sense-perception and of our take on the physical world afforded by sense-perception. Any comprehensive view of sense-perception should stake a claim to one or the other of these theories. And together, they appear to exhaust the (realist) alternatives.
So it is that Thomas Reid disparages the“ideal,” or “Cartesian” system for being thoroughly representationalist in nature and proposes his own common-sense, direct realism in its place. On his account, the ideal system builds “upon consciousness as its sole foundation, and with ideas as its materials.”[1] Ideas become the internal and immediate objects of thought – a view Reid takes to be rooted in the twin beliefs that there must be “immediate concourse” between the mind and its objects, and that for any act of thought, an object must really exist.[2] For these reasons, the system introduces an unnecessary idea-object into every individual act of thought, and reduces all mental acts to simple perception. The result, Reid tells us, is that what little credibility the theory of ideas has is bought with an inevitable skepticism about the external world.
Reid’s characterization presents representationalism in its simplest form, which makes external things accessible only through representing ideas. Couched in more sophisticated terms, independent of the way of ideas, the rivalry between direct realism and representationalism continues unabated in contemporary philosophy. So it may seem like pure hubris even to suggest that the distinction may not be all it’s cracked up to be. Nonetheless, I suspect that is so. Here, however, I will restrict myself – mostly – to a more modest claim: that there are important issues about our perceptual access to objects that are independent of – indeed prior to -- questions about directness or indirectness in sense-perception. And those, I will argue, are the issues that most occupied Descartes. Reid describes the “ideal system” as “Cartesian.” He is not alone in considering Descartes to be the arch-representationalist of all representationalists, although a good deal of recent scholarship has challenged this interpretation.[3] But if I am right, the difficulties conscientious commentators have may signal that the wrong questions are being asked. To make my case, I will look first at Reid’s version of “Cartesian” representationalism, and then propose a somewhat different formulation of the contrast with direct realism. Although I do not think that even this different formulation fits Descartes, I hope it will indicate what he is really up to.
Reid rests his characterization in large part on a causal claim: representationalists hold that there must be “immediate concourse” between the mind and its objects.[4] Perhaps. But it is no easy matter to understand the significance of this claim. True, many of the philosophers Reid identifies rejected the possibility of action at a distance – but the “immediate concourse” is a matter of mental causation, for which no requirement of spatial contiguity has clear application. Reid may simply mean that causation can only take place between entities of the same kind, be they mental or extended. That, however, is a point of enormous controversy among philosophers of the way of ideas -- and it is a point Descartes openly denied. Descartes’s position here presents difficulties of its own. But it still seems pretty clear that he considered brain or pineal events to be the proximate causes of sense-perceptions, appetites and passions, as well as taking the mind to be the cause of voluntary movements in the body.
Even supposing Descartes were a closet-occasionalist, Reid’s claim runs into difficulty, for it seems particularly directed at the causal relations holding between the mind and its ideas, in particular, at how an idea can act on the mind.[5] Now, the clearest cases Descartes offers of ideas – or more specifically, the objects of ideas – exercising causal force on the mind are found in sense-perception and the passions. A particularly violent sense-perception, e.g., a painful poke in the eye, or a consuming passion, e.g., the fear of killer bees, can literally force itself on my attention.[6] But what seems prima facie the distinctive feature of these ideas, the feature that allows them to exercise force on my mind, is that the objects of such ideas are identical with some element in the external causal chain leading up to the perception.[7] Because of the particular structure of my body, a poke in the eye leads to a great rush of animal spirits overwhelming my unsuspecting pineal gland, and thus (somehow) my mind. But if we insist -- à la Reid -- that ideas are purely mental entities, then Descartes’s story of how they can exercise force on the mind crumbles.
Descartes, however, admits certain properties of ideas that seem capable of acting on the mind. The clarity and distinctness of an idea should compel the will to affirm its truth.[8] But this “compulsion” is surely not efficient causation, for it is free and spontaneous. Clarity and distinctness operate normatively, and it is the mind’s responsiveness to norms that prompts it spontaneously to affirm the truth of such ideas. Moreover, that some ideas are normatively binding on thought is taken by Descartes as evidence that they are “not invented by me or dependent on my mind” (AT VII 64, CSM II 45),[9] but are of true and immutable natures. The objects of such ideas may not be material things existing “outside of me,” but neither are they exhausted by my thinking of them. If these ideas are of things existing independently of me, which thus could exist “outside” me, we face some puzzles in maintaining that being an object for the mind requires immediate presence.[10]
Reid identifies another feature of the ideal system that is just as basic, but independent of the causal claim: there must be an object for any act of thought. Now if this supposition is to do any work, not just any object will do: it ought to be a mental entity, and it ought to have those properties we perceive it to have. If we think that the cat is on the mat and it is pink, then even if the cat is silver and in the Agora, there is something that is cat-like, on-the-mat-like, and pink. That thing is what we perceive, a mental proxy mistakenly supposed to resemble the external state-of-affairs.
But does this characterization fit Descartes? Opinion is divided about whether Descartes requires all ideas to have objects.[11] And as I’ve already suggested, ideas might have objects that are not simply dependent mental entities. But the real difficulties arise in accounting for misperceptions. On the view Reid offers, the explanation of why we ascribe “secondary properties” to external things is that we perceive some mental entity that has those properties. And that cannot be correct. Descartes does indeed hold that it is a conceptual confusion to think that extended things have properties of being colored, etc. But it is (at least) equally confused to ascribe these properties to mental entities. Where we can locate secondary properties is a vexed question, but it is not solved by saying that ideas are what are red, or hot, or smell like honey.
Moreover, saying
that for any idea an object must exist is not to claim that ideas are
themselves objects. Locke sometimes says that; Descartes does not.[12]
Ideas have objects for Descartes; it may even be the case that all
ideas have objects, without it following that those objects are mental things.
This claim is, in fact, compatible with a number of views, both
representationalist and not.[13]
Likewise, maintaining that every idea has an object does not entail that those
objects must be as they are perceived.[14]
For these reasons, the mantle of Reidian representationalism does not sit well
on Descartes.
Of course, such an admission hardly entails that Descartes embraces some form of direct realism. That would only follow if Reid’s were a completely satisfactory version of representationalism, and if representationalism and direct realism exhausted the available alternatives. Neither is true, I think, but I will challenge the latter claim by presenting somewhat different – and I hope adequate – versions of both representationalism and direct realism. To begin, let me note several points we should keep in mind. First, the contrast between direct and indirect seeing could be generalized to perception understood more broadly – perception in the comprehensive Cartesian sense that makes any receptivity a kind of perception. The perception of an “inner” object should not ipso facto qualify as indirect, for we analogous questions can be raised about it.
On the other hand, we do seem to need some sort of contrast between kinds of perception. Such contrast may be merely implied, e.g., by taking directness to be the default option, but it does seem necessary to give some substance to the competing theories. So it seems unlikely that there could be any global account of perception understood broadly (i.e., in the Cartesian sense) that makes it all indirect – or all direct, for that matter. [Notice that when I use “direct” and “indirect” here, I do not mean immediate and mediate.] That fact seems of little consequence so long the theories are restricted to sense-perception or its near relatives. Reid’s characterization of representationalism, for instance, makes perceptions of independent objects indirect, while allowing perceptions of non-independent objects (e.g., sensations) to remain direct. But as I’ve already argued, segregation by object is not enough to maintain the required contrast between kinds of perception. Whether a theory says yeah or nay to a division between perception directed at one class of objects and perception aimed at another class, it still owes us an account of how the perceptions differ – of what directness or indirectness amounts to in perception. My worry is that many accounts of Descartes simply assume that distinguishing between objects that are internal or mind-dependent and those that are not suffices to make the distinction between kinds of perception and thus kinds of theories. And that risks making the theories into mere window-dressing for decisions about how to divide objects into the internal and the external – decisions that themselves may be simply circular or arbitrary.[15]
Many accounts of Descartes’s representationalism – indeed of representationalism in general – hinge on the role given to causation. On the face of it, that also seems a real blunder: whatever else they may be about, direct realism and representationalism concern intentionality, and efficient, mechanical causation does not give intentionality. But even if intentionality cannot be reduced to the causal underpinnings of sense-perception, causal connections may still have some role to play. The account I will offer tries to take causal issues seriously, but not reductively. It also tries to respect what I take to be a general criterion for philosophical distinction-mongering: that both alternatives be prima facie live options. No account should start off characterizing direct realism in ways that seem so naïve as to be impossible, or making representationalism so unattractive that it loses any claim to being a theory of perception. And if the theories are to stand in opposition, there should be some shared territory on which they can disagree. Many defenses of one or the other position may fail to meet these criteria, and thus appear question-begging.
Now both theories require that there be some connection with external, physical things. And that connection should be one established in perception itself, not merely an inferential one running out from perception.[16] So we need a distinction between connections that qualify as representationalist, and those that count as direct. The minimal connection with external things is some sort of reference. What can make the difference is whether reference is achieved through something like immediate acquaintance or by causal transmission (supposing that noetic rays and the like are non-starters). This distinction can be cashed out to describe the relation between the (immediate) contents of a perception and its reference.
The representationalist, on this approach, can be understood as holding that we must distinguish between the contents of a perception and the reference the perception makes to whatever external, physical thing is their cause, while supposing that that thing forms no immediate part of the content. She might, for instance, distinguish between what contents “unaided perception” delivers to us, and those contents to which we gain access only by way of aids and props that produce the reference. This is a strategy adopted by Harold Brown, who takes both the use of instruments “that alter the causal chain between a perceiver and a physical object” (e.g., telescopes) and the bringing of revisable concepts to bear on our perception as the reference-producing aids and props. His view can be taken to task on a number of grounds, not least of which is the pristine isolation in which it supposes perception operates. Still, it seems to answer the question of how we make reference to something that is not part of the contents of unaided perception.[17] Whatever route is taken, the representationalist might have a workable position if she can maintain at least two things: that we do indeed perceive the thing at which our perception ultimately aims (the physical thing) – even that the thing or reference forms some part of the content – but that it is perceived in some way that is mediated by other, distinct contents of the perception, which are themselves taken as an object and perceived as such, in a way different from the ultimate target of reference. Some version of this view has, in fact, been attributed to Descartes by a number of commentators, who contrast what perception presents us with, and what we take it to refer to.[18] On this view, the external thing causes the presentational features of the perception, and that causal history in turn establishes the reference.
In contrast, a direct realist would hold that we are able to refer to physical things simply on the basis of the contents of our perception – even more that the content of our perception, or at least some important component of it (e.g., the object), just is the external thing to which the perception refers. Here too the external reference is taken to be the cause of the perception. We manage to refer to that external thing because we can identify it with some of the contents of our perception that it causes. But we may still wonder how this identification suffices to produce a genuine relation of perceptual reference. In fact, both the representationalist and the direct realist are saddled with two versions of the same problem: how to get reference out of some other, rather recalcitrant relation that holds between a perception and an external thing. Thinking of the perception and its contents as just things in the world does not yet establish any relation of reference, for (even) contents alone do not refer. We need to add something to the stew – something like an intent to refer. Then it looks as if we can generate two distinct positions simply by way of an intention that travels along whatever relation between the contents of the perception and the external thing is supposed to obtain: either “I intend to refer to whatever is the cause of that” (where ‘that’ is the objective content of the perception), or “I intend to refer to that” directly.
Maybe we now have a workable distinction between direct and indirect perception. ‘Direct’ perception occurs when the (objective) contents of the perception and its reference coincide; ‘indirect’ perception when they do not. Directness here becomes something like immediacy of acquaintance, and the representationalist and direct realist can disagree about the general nature of our perception of external things – without equivocating on the sense of ‘direct.’ The direct realist can even help herself to a distinction between direct and indirect perception, where the objects of both are external things. Seeing Winston Churchill in a mirror is indirect just because the contents of our perception is the mirror, yet its reference remains Winston himself.
This approach allows the representationalist to explain the possibility of perceptual deception quite handily. Simply by granting that different causes can produce the same perceptual contents, the representationalist has the resources to account for how we take a sense-perception to be of a round tower, when it is really of a square tower, or think we are perceiving a dagger, when there is no such item in the vicinity. We might, for instance, have a perception with dagger-like contents that is really the consequence of a fit of the vapors, or consuming a bit of less-than-fully-fresh fish. We still refer to those causes in perception, but we (perceptually) mistake their natures.
Misperception has seemed more of a stumbling block for direct realism. We can set aside hallucinations and the like for the moment: the direct realist may simply count them another species of mental events quite different from sense-perceptions.[19] Garden-variety misperceptions, though, are another matter – be they misperceptions that sucker us into false beliefs (the far-off towers) or those that fool no-one (elliptical appearances of table-tops), with qualia falling somewhere in between. Such misperceptions seem genuine perceptions, yet they may also be well-nigh inescapable. If so, direct realism seems doomed either to denying the evidence of qualia available to the merest slip of an philosophy student, or to supposing a very strange physical world indeed – the sort of world Cartesian science rejected.[20] But if the direct realist admits that some of those contents are not attributable in any robust way to physical things, then (the story goes) the floodgates are open for representative entities of various stripes to rush in.
These unattractive alternatives, however, are not incumbent on the direct realist, on the approach I’ve sketched. The direct realist holds that what we refer to is simply some component of the contents of our perception. But it does not follow that we perceive those contents correctly. We intend to refer to that, but we may get what that is dead wrong. The direct realist can even allow that the contents of our perception provide the opportunity for mistakes; she can, for instance, manage unruly qualia and other perceptible contents not readily attributable to physical things simply by removing them from the proper, objective contents of our perceptions – those identified with that. As D.M. Armstrong suggests, the re-filing of such sense-data can take two forms: one is the “disappearance view,” which holds that they modify nothing at all (and thus not physical things). The other is the “adverbial view,”[21] which takes qualia to modify acts of perceiving, rather than their objects.[22] The direct realist might, for instance, allow that when we perceive some X, we also perceive, dimly and through a glass darkly, the way in which we perceive. That would admit qualia by supposing a reflexive component to all our sense-perceptions. In both cases, the direct realist will treat the perception of qualia as such as little more than a psycho-perceptual quirk -- mere perceptual baggage and noise.[23]
The question of qualia is important, but I don’t think it is solved by treating them either as referential proxies or as mere noise. Understanding the alternatives depends on how (and whether) sensations further the ends of perception. But before moving to this topic, let us note some remaining problems for establishing reference to external things. The representationalist, in particular, seems to face a difficulty. For the gap between the contents of the perception and its reference is produced by intending to refer to the cause of the contents. But that drastically underdetermines the reference. What we would normally think of as the cause to which our perception refers – the cup, this page of type – may well be located in the causal chain leading to our perception. But it is only one link in that chain, neither first, nor last. Indeed considering only the causal chain, those ordinary, middle-sized objects to which we hope to refer have no special status at all. What’s more, any perception could be caused by various, quite different states of affairs – as Descartes’s favorite examples of phantom pains illustrates – some causal antecedents are standard, some not so standard, but they are still causes. We might think to fix reference by specifying that we are interested in that element in the chain correlated to the contents of our perception. But correlation fares no better than causation at least on Descartes’s account. There are a number of relay points in the causal chain to which we can correlate all the same features of perceptual content – the impact of light rays on our retinas, the patterns of brain events, etc. do just as well as do ordinary, middle-sized objects.
The problem is that it remains completely mysterious how an intent to refer can be transmitted along a causal chain when that chain involves no further intentions. The kind of transmission that the representationalist supposes is crucially different from socially transmitted intents to refer. With socially transmitted intents to refer, I can point to an I-know-not-what specified as the object of your reference, and so on through an entire chain of social links. Merely causal impacts do not do the trick of specification, which is why we had to introduce my intent to refer in the first place. One last hope might be left, which is to limit my intent to refer to the proximate cause of that content: this is a view adopted by some commentators, e.g. Ann Wilbur-Mackenzie, who holds that Descartes takes our sense-perceptions to refer to brain states (but not to any of their intrinsic properties). But I remain unsure why we should think that we stand a better chance at referring successfully to the proximate cause than we do to slightly more distant ones. If there is nothing to pin down our intended references when we seek to extend them beyond what we are immediately acquainted with, then it seems that the best we could hope for is an utterly vague and undifferentiated reference to what is ‘out there.’ The direct realist might then seem to win a victory by reason of vagueness. For it looks as if the only way we can perceptually refer to a particular state of affairs is by immediate acquaintance – that is, directly. Of course, that may leave us with no sort of realism at all, for reference of this sort does not guarantee that the contents of our perception are indeed identical to external things. And even if they are identical, one might wonder whether that truly establishes that we refer to external things when we refer to those contents – that is, whether reference, at least perceptual reference, is safeguarded by extensionality. I will return to this point.
Most importantly, we should realize that there is an issue prior to deciding between the two positions about reference. The direct realist says “I intend to refer to that,” where “that” is the objective content of the perception; the representationalist says, “I intend to refer to whatever is the cause of that.” But then both need to explain how we determine what that is: that is, how we distinguish the content of a perception from other features of the perception, or better, how we distinguish the objective content, both from other contents and from other features of the perception in general. How do we even get access to “that?” This issue is far more basic than any question about reference and its ilk. It is also, I suggest, the question Descartes addresses when considering representation in sense-perception – how our perception becomes meaningful, or significant – or intelligible. Meaning, then, not reference, is Descartes’s central concern, and he may even take meaning to fix reference, or simply fail to distinguish between the two.
According to Reid, the “ideal” system has a central metaphor – ideas are paintings, or more generally, pictures. And indeed the Meditations declares that ideas are tanquam imagines [“as-if images”]. Many commentators have taken the relation of picturing here to be analogous to the relation holding between a photograph of Winston Churchill and Winston Churchill. But there is another possibility: the relation in question could be like that holding between the photograph and the man depicted in it – or if you like, between a painting (that is not a portrait) and the man-in-the-painting. Now, part of what makes sense-perception distinctive for Descartes is its brute presentation of impingements on our bodies and to our minds.[24] But there’s the rub. For the world of extension is a plenum of matter in motion. It does not easily lend itself to the forms of our perception. Here I do not mean simply that we ordinarily perceive the external world through color, sounds, tastes, etc. that are not properly speaking properties of extension. Rather, I mean that the way we get deliverances from the world is through a welter of impacts on and in our bodies – impacts that are constantly changing, that do not carry determinate information about their causal origins, and that come to us in a mostly undifferentiated bundle. [25] And yet we somehow perceive a world that is at least prima facie intelligible – however many mistakes we make about its nature. Intelligibility in this sense does not require bridging a gap between inner and outer in order to connect an already-grasped mental object with the external world. Instead, it is a matter of somehow gaining access to an intelligible, or “meaningful,” world through a welter of sensory impacts – colors, sounds, even “puffs of air.”
There are two pervasive features of Descartes’s accounts of sense-perception that I think will support this gloss. The first, though, provides mainly negative evidence. Descartes denies early and often that our ideas resemble their causes (or objects). Now, it may seem that even to ask about resemblances betrays a concern with the relation between content and external reference, rather than with meaning in my sense. Only the man in the painting is a candidate for resembling Winston Churchill, not the painting, which no-one would ever think bears much resemblance to any man. But that is just my point. Whenever he considers a relation that is like the relation between pictured content and external referent, Descartes allows some degree of resemblance. Whether it is a matter of information-carrying patterns, distributions of light or motions, or the simple “look” of things, Descartes admits various kinds of resemblances between external things and retinal images, between retinal images and brain states, between brain states and vibrations in the pineal gland. His denials of resemblance come when he turns to how sensory ideas make us think of their causes and objects. Those denials make utter sense if the relation between ideas and their objects is like the relation between a picture and what it depicts. For then, resemblance is simply a non-starter.
Why then does Descartes even bother to raise the issue of resemblance? We may find an historical reason in the Scholastic theory of forms. This theory takes the content of a perception to be determined by a form derived from and shared with external things. The form does a lot of work: it specifies what the perception is of, what kind of perception it is, and what caused it. In performing its many tasks, the form establishes various resemblances, and tracing formal resemblance serves to answer questions of both meaning and reference – indeed of meaning above all, since forms are the very principle of intelligibility. Descartes’s turn to the issue of meaning is marked not only by the rejection of the Scholastic theory and its reliance on resemblances, but by raising problems undreamed of in that philosophy.
Let me turn to the second and more important piece of evidence for the claim that Descartes is primarily concerned with issues of perceptual meaning. Descartes makes frequent – and on the “referentialist” view, surprising – mentions of language perception:[26]
Words . . . bear no resemblance to the things they signify, and yet they make us think of these things, frequently even without our paying attention to the sound of the words or to their syllables. Thus it may happen that we hear an utterance whose meaning we understand perfectly well, but afterwards we cannot say in what language it was spoken. (AT XI 4, CSM I 81).
Despite the somewhat ambiguous mention of how words signify, the denial of resemblance is clear. So too is the theme of the passage; it is how we perceive sounds as words, the perception of their meaning.
The perception of sounds as meaningful provides an illustration of our ability to glean intelligible sense-perception through the sensory impacts on and in our bodies. But saying that may seem to commit a gross equivocation. We can understand these sensory impacts in two ways: quite literally, as impacts on our bodies, impacts that cause sense-perceptions, or as sensations that may or may not play some role in intelligible perception.[27] But important as these differences may be for metaphysics, they are largely irrelevant to our purposes. The opening of the World[28] makes just this point:
“perhaps you will say that our ears really cause us to perceive only the sound of the words . . . and that it is our mind which, recollecting what the words and the countenance signify, represents their meaning to us at the same time. I could reply that by the same token it is our mind which represents to us the idea of light each time our eye is affected by the action which signifies it.”
Descartes goes on to contrast our perception of “the sound of some words, [when we do not attend] to their meaning” with what causes them, namely the actions of the mouth, tongue and breath of the speaker.[29] Both somehow convey meaning, but not because of relations of resemblance or causation that they bear. We suppose other kinds of sense-perceptions (rightly or wrongly) to give us some insight into how external things stand, because we take those things to be both the objects and causes of our perceptions. But we make these connections by way of both impacts on the body and mental sensations that are very different in kind from those objects/causes. That sensations and the bodily impacts are also very different in kind from each other, however, does not prevent them from sharing the same role: they both function as means of representation allowing intelligible sense-experience.
Language perception is not, of course, the only model Descartes uses. Pictures also figure in his accounts of sense-perception. But Descartes emphatically denies that we need little pictures inside our head in order to sense-perceive. The Optics takes issue with just such views, and with those “philosophers” who assume that “in order to have sensory perceptions the soul must contemplate certain images transmitted by objects to the brain.” Instead, “we must conceive the nature of these images in an entirely different manner from that of the philosophers.” (AT VI 112, CSM I 165). That is just what Descartes does when he turns to sense-perception of actual pictures – in particular, of engravings:
. . . consisting simply of a little ink placed here and there on a piece of paper, [engravings] represent to us forests, towns, people, and even battles and storms; and although they make us think of countless different qualities in these objects, it is only in respect of shape that there is any real resemblance. And even this resemblance is very imperfect, . . . in accordance with the rules of perspective they often represent circles by ovals better than by other circles, squares by rhombuses better than by other squares, and similarly for other shapes (AT VI 113, CSM I 165).
Now, the engravings are themselves part of the external, extended world; nonetheless, they might seem to provide an explanatory analogy for how mental “pictures” make us perceive external things. But Descartes refuses to make any such analogy. Rather, he declares that “we must think of the images formed in our brain in just the same way, and note that the problem is to know simply how they can enable the soul to have sensory perceptions of all the various qualities of the objects to which they correspond – not how they can resemble those objects” (AT VI 113, CSM I 166). In fact, Descartes insists that “we must hold that it is the movements composing this picture which, acting directly upon our soul in so far as it is united to our body, are ordained by nature to make it have such sensations” (AT VI 130, CSM I 167).
This last point addresses the problem of the causal connection between mind and body, but it is not Descartes’s only concern. He moves from how marks on paper represent, to how images in the brain stimulate our sense-perception, to how the movements in our brain (which produce brain images) make our soul have sensations. What links these moves is that they each concern the means of representation, how very different kinds of things can serve as vehicles of information.
Descartes uses his examples of engravings to draw our attention to the ovals that represent circles and the rhombuses that represent squares. Unlike the movements in our brain, these are certainly something we can perceive if we attend to the composition of the engravings. In this respect, they function as do sounds in language perception. Sounds are mere “raw” sensations, yet they play a crucial role in enabling perception: they are contents of sense-perception – contents that represent an object or objectified content. Now, when we look at the engravings, Descartes makes no suggestion that we see the rhombuses directly as rhombuses and indirectly as table-tops; nor does he suggest that we directly hear words as sound, and only indirectly as meaningful bits of language. If anything, the effort goes in the other direction, at least for those fluent in the conventions of perspective or fluent in a language. This is not because the rhombuses or sounds are imperceptible, but because they represent transparently in letting us perceive what they represent.
In contrast, the neophyte language learner may well perceive the sounds of words – perceive them so loudly and obviously that meaning is drowned out. Likewise, we may be able to determine shapes and colors in a badly drawn picture without quite seeing what it represents. In general, the relation between sensations and what they represent runs a continuum from transparency to opacity, even to noise. It is always possible to turn our attention to those contents in our sense-perception that are doing the work of representing some object to us – to the sounds of the words, or the shapes on the page – although more often than not it takes quite an effort.[30] But the ease with which we can ignore the representing means and media when we are fluent speakers or practiced in pictorial conventions should not disguise the philosophical problem Descartes is investigating: how in sense-perception, our representations allow for intelligible experience.
Descartes proposes a doubled content and a direction of focus in sense-perception:[31] a content of representing sensations, correlated with the causal impacts on our bodies, and the objectified content to which we attend. Both contents are in principle perceptible. But in normal sense-perception, we pay attention only to the intelligible, objectified content. What represents that intelligible, objectified content is not itself part of it – and if it is intelligible in its own right, it may be only insofar as it is understood as a kind of pointer, an entity exhausted by its representing function.
All I have attempted to do is to characterize what Descartes sees as the problem for representation in sense-perception, not how he thinks we go about performing the rather remarkable feat of extracting intelligible experience from a collection of raw sensations. To answer that would take us deep into the heart of Descartes’s theory of ideas – to the status of their objective properties, and especially to the work performed by innate ideas in organizing and schematizing our thinking. And that is a whole other kettle of fish. But let me finish with a couple of points about how to understand Descartes’s approach. As I set it up, the debate between direct realism and representationalism revolves largely around reference. I think that emphasis is correct (not that reference is the only issue, but that it is deeply involved in the central issues). But if Descartes has anything by way of a semantic theory, it seems unlikely to be a primarily extensional one. For this reason, I think Descartes might deny – on metaphysical grounds – that we can refer very precisely to parts of the physical world. Our clear and distinct ideas equip us to understand the nature of body as extended. That may well serve to fix the extension of the idea. But the nature of extended substance does not allow it to be individuated in any way that runs very deep. Extended substance functions rather as does a mass term, without (I think) presupposing already differentiated individuals. I see no reason why Descartes should deny that we manage to refer to what is “out there,” but what chunk of the stuff “out there” we refer to remains vague – at least if we consider particular ideas or sense-perceptions in isolation. We may make decisions about how to individuate and differentiate the stuff out there, but when we do so, they will be founded more on convenience and usefulness to human purposes than on the metaphysics of the natures giving our ideas content. For this reason, I think that fixing a determinate reference will require something above and beyond the internal structure of our ideas – something that will allow us to triangulate on a piece of the world. This may be a matter of relations among our ideas – perhaps even of relations to the ideas of others (un-Cartesian as that thought may sound).
For Descartes,
representation is in the business of producing intelligibility: representations
provide the sort of mediation through which we can grasp intelligible,
meaningful objects. Our access to objects is always mediated, and so the ease
or transparency of access cannot be identified with immediacy. But neither does
mediated access mean representationalist access. There is a continuum of
possibilities ranging from transparency to opacity. In many cases, transparency
becomes possible only because we bring concepts or learned skills to bear on
our perception. Of course, the wrong concepts or skills may block access, but
then the nature of our access is not to be decided by invoking notions of
mediacy or immediacy, directness or indirectness tout court.
[1]An Inquiry into the Human Mind on the Principles of Common Sense in Inquiry and Essays, ed. R. Beanblossom and K. Lehrer (Hackett, 1983), p. 114
[2]Essays on the Intellectual Powers of Man, in Inquiry and Essays, p. 227
[3]Some examples
of commentators who take Descartes to be a direct realist include Steven
Nadler, Arnauld and the Cartesian Philosophy of Ideas (Princeton, 1989),
and earlier, Brian O’Neal, Epistemological Direct Realism in Descartes
(University of New Mexico Press).
[4]The causal import of this criterion is clear in Reid’s text: “there must be some immediate intercourse between the mind and its object, so that the one may act upon the other,” ibid, p. 227.
[5]Reid might simply mean here that ideas must be mental entities to be perceptible by the mind. But then all we are left with so far is the claim that Descartes thought that only ideas as mental entities are perceived, because only mental entities are perceptible –not a claim that gives us much in the way of either explanation or evidence for this position. Nonetheless, some of the arguments given below will have consequences militating against this position.
[6]The adventitious nature of sense-perceptions, i.e., that they seem to have some force on the mind, is, after all, the crucial step in the argument given in the Sixth Meditation for the existence of material things.
[7]This use of “perception” may seem to equivocate between the act of the mind’s taking something in and the “object,” or idea. But that is typical of much of Descartes’s own language, and it may not represent mere equivocation.
[8]The material falsity of an idea is sometimes treated as a “cause” for error. But Descartes is fairly explicit in restricting this claim to material causation only.
[9]All quotations from Descartes's works are taken from Ch.Adam and P. Tannery, eds. Oeuvres de Descartes (Paris, France: J. Vrin, 1996) (henceforth AT); English translations not my own are from J. Cottingham, R. Stoothoff and D. Murdoch (Cambridge: Cambridge U. Press, 1984) (henceforth CSM).
[10] Perhaps Reid simply means that there must be interaction with some mental item for me to think of an object. But then the rest of his account will not follow, for the more modest claim does not suppose that that mental item is representational, a mental proxy standing in for external things.
[11]Margaret Wilson, for one, argues that raw sensations probably do not have objects, see “Descartes and the Representationality of Sensation” in Kulstad and Cover, Essays in Early Modern Philosophy (Hackett)
[12]Perhaps the closest Descartes gets to such a claim is in the definitions given in the Second Replies: i.e., an idea is “the form of any given thought, immediate perception of which make me aware of the thought,” and objective reality of an idea is “the being of the thing which is represented by an idea, in so far as this exists in the idea” (AT VII 160-1, CSM II 113). Although the notion of objective reality (or being or perfection, etc.) comes close, Descartes stops short of saying that an idea is the (or a) object for the mind. Indeed his repeated insistence in Definition II that an idea is the” form of any given thought” (my emphasis) seems to belie taking ideas to be objects in any straightforward sense.
[13]. Spinoza, for instance, thought that for every idea, an object must exist, which has the positive properties the idea has objectively. Yet since that object is directly the human body, this seems to allow for a kind of direct realism, at least about bodily states. His refusal to countenance mind-body causation might, however, be judged incompatible with direct realism. I do not think so, but in any case, the point is to sketch possible positions – not to determine Spinoza’s acual one.
[14]Again the Spinozistic view allows that ideas may be quite inadequate. In fact, Descartes’s description of material falsity would suggest that he at least wants to allow that we may have ideas so irremediably confused that they do not allow us to perceive any object as it truly is. This gloss may strike an odd note, for material falsity is sometimes taken to demonstrate a clear commitment on Descartes’s part to representational mental entities. But as Arnauld’s criticisms show, Descartes would then be left with no account of material falsity in perception, but only of mistakes in judgment.
[15]There are a number of other problems plaguing the characterization of the distinction. Many proponents of either direct realism or representationalism attempt to formulate their positions by way of examples – like those I gave at the beginning. But the intuitive appeal of such cases is often not clear, and appeal to example may generate simple disagreement. Some philosophers count perceiving a far-away X through a telescope or an itty-bitty X through a microscope as prime examples of indirect, and even representational seeing, e.g., Harold Brown, et al. For others, neither telescopes nor microscopes (at least of the optical variety) impugn the directness with which we see that X, e.g., Clark, et al. These disagreements about whether lenses and other instruments suffice to re-direct our attention might simply reflect differing views about the reliability or alienating effects of such instruments. But I think that what they really reflect is uncertainty about when forms of perceptual mediation belong to the causal order or to the order of objects, or both.
[16] I will take it that this sort of inferentialism is a move of last resort, leaving us only an anemic contrast between whatever perception of an internal entity is the default kind and non-perceptual inference to external things.)
[17]Another route a representationalist might take is to distinguish between the causally efficacious portions of the content of our perceptions and the causally inert parts – that is, to distinguish between those parts of our perception that are causally responsible for our having the perception in the first place, and those that are not. This is a position suggested by D.M. Armstrong; although he wants to use it for a direct realist position, it can be modified for representationalist use when we take the causation in question to be the having of a mental event of a perceptual kind, or if we allow the representative entity to be some component or quality of an external thing, distinct from the thing itself.
[18] E.g., Margaret Wilson, Ann Wilbur-Mackenzie.
[19] So the direct realist can allow that sense-perception does indeed connect us directly to things in the world, while denying this status to hallucinations (the so-called “disjunctive” view). Here the difference between the direct realist and the skeptic-ridden representationalist) lies solely in the classification of mental events; the direct realist may, for instance, argue that hallucinatory events bear their difference in kind on their sleeve, or can be kept safely quarantined in some other way.
[20] Thus Howard Brown takes it to be an argument for representationalism that we do not truly perceive external, physical things as they “intrinsically” are – although we do have genuine access to those things. Our perceptions inevitably mischaracterize or misrepresent (and hence represent) the intrinsic natures of external, physical things. This form of representationalism thinks to trump direct realism by the-not-implausible claim that we do perceive external things, just none of their intrinsic properties. See also Wilbur-Mackenzie. Access is provided by scientific investigation, and it is because our best physical theories provide insight into the intrinsic properties of external, physical things that we can evaluate the way our perceptions represent those things. Brown’s view sounds true to Descartes’s views about the vagaries of sense-perception, as well as our abilities to correct our assessment of the intrinsic natures of external, physical things.
[21]Although this is not the exclusive property of direct realists cf. Hirst and Armstrong).
[22]We might still admit that when we perceive physical things we ipso facto perceive colors – but that will seem representational only because of an equivocation in the structure of the second “perceive.” We can rewrite the claim to read that when we perceive physical things we ipso facto perceive them in some determinate way or another. And if that seems to deny the evidence, we can even admit that when we perceive physical things we ipso facto have another sort of perception as well, a perceiving (as if) of colors (where the “of” looks more intentional than in the gloss offered above).
[23] One last virtue of this approach worth mentioning is its ability to accommodate some of the epistemological issues associated with each position. When we visually inspect what we see, and state that this cup is half-full, or this cup is on the table in front of me, or this cup is clean despite the shadows that fall on it, what provides our warrant for these judgments? If the warrant is the state of the cup itself, we seem to have a view most readily couched in the terms of direct realism. If, however, the warrant lies in some proxy, e.g., the perceptual contents caused by the cup in its present state, then we seem to be in representationalist territory. This does not mean that the representationalist must hold that such judgments are unwarranted, or that their warrant in no way derives from the state of affairs in the world. The difference lies in how we think the contents of perception carries authority for our judgments. The direct realist takes perceptual contents to be authoritative simply because they are identical with what we judge. But the representationalist thinks perceptual contents are authoritative by proxy. This can be a perfectly good authority, but it derives from the state-of-affairs in the world as if by power of attorney. In fact, for the representationalist, states-of-affairs, brute facts in the material world, may simply be incapable of exercising authority over judgments. They are like the state in civil contracts, requiring a representative to articulate the content of those judgments to which they extend authority. And as the metaphor suggests, those representatives can act in ways for which they are not authorized, and judgments can be pronounced by those not authorized to make them. The direct realist, of course, is not bound to take every judgment based on the evidence of perceptual contents to be authoritative, but when they are not, it is less a matter of abuse of authority than of failing to recognize authoritative pronouncements (for what they are).
[24]The Sixth Meditation argument for the existence of material things suggests this view – at least if we concentrate more on the claim to experience causal force than on the seemingly inferential move to material things.
[25]It would help to bear in mind that much of Descartes’s work in optics, physiology and the nature of perception is devoted to explaining three points: 1. that sense-perception in all its forms is enabled by a series of events external to my mind, which can be explained mechanically, that is, in terms of the causes proper to the external, material world; 2. that what we would normally count as the object of a sense-perception lies in the causal chain of that perception; 3. that that object’s causal role is insufficient to determine the character of our perception; indeed that the whole causal story is insufficient to explain important aspects of our perception – including, I would argue, its objective content. That insufficiency is (one) source of the possibility of misperception. Phantom pains are a case in point; what makes a pain a phantom pain is that I take the pain to be a pain in my arm (” or “about” the state of my arm). However, since I lost my arm in a particularly nasty round in the WWF, this is a misperception. Rather, the cause of the pain is a pinched nerve in my shoulder. Nonetheless, we can still maintain that the pain is about this state of my body, but this state is not the state of my arm, but of my shoulder. The misperception does not result from perceiving some sort of merely apparent object, which refers to nothing at all, but of the categorization (based on causality and natural institution) of what the perception points to. Because sense-perception underdetermines its causes and “the natural institutions” that govern how bodily events cause perceptions can operate under non-standard circumstances, this categorization can go astray. The skeptical doubts of the First Meditation are so familiar that it often seems as if Descartes must always be required to suspend judgment as the narrator does there. But there is really no question for Descartes in his non-meditative, non-skeptical moments that a pain, even a phantom one, does point to some event or state in my body, that is to some external happening that lies in the causal chain. If it did not do that, then it would not be a pain at all. What makes this, of course, a phantom pain, rather than a more garden-variety mistake, is that the natural institution makes it extremely difficult to perceive the cause of the pain as its object. But neither here nor in the more garden-variety sorts of mistakes does Descartes countenance a use of “feels a pain” or “sees-a-pink-elephant” that is indifferent about whether the perception is adventitious (comes to me from external causes) or fictitious (has only internal causes). And Descartes-the-scientist/physiologist (in contrast to Descartes-the-neophyte-Meditator) does not really countenance the possibility that we might mistake fictitious perceptions for adventitious ones – that we might, that is, take some purely mental proceeding for those perceptions that have external causes.
[26] There are many such references; see the World AT XI 4-5, CSM I 81-2, Optics AT VI 112-3, CSM I 165, Principles AT VIIIA 320-1, CSM I 284 Passions of the Soul, AT XI 369, CSM I 348. Yet few of them seem to have been noticed by commentators.
[27]Some commentators argue that sensations do not represent anything at all; see, e.g., Laura Keating, “Mechanism and the Representational Nature of Sensation in Descartes,” Canadian Journal of Philosophy 29, (1999), 411-30, and (to some extent) Mackenzie (1990). Although I do not tackle these arguments directly, my treatment of Descartes’s interest in language perception is supposed to show that sensations do indeed represent, but not in the sense those commentators reject. We might say that they represent without being representations themselves.
[28]This is where Descartes introduces the analogy with language perception cited just above.
[29] ibid, p. 82. The target in these passages is the difference between the idea we have of light and “what it is in the objects that produces this sensation,” and he describes the action of the latter on our eyes as a “sign” for our idea of light.
[30]Sometime, however, training allows us to perceive just these representing media. Certain kinds of artistic training aim at increasing our power to perceive these media.
[31]This view has much in common with the account of double perception Richard Wollheim has proposed to explain pictorial representation in, e.g., Art and its Objects.
. Consider again what it is like to see Winston Churchill in a photograph or on television. It seems that I see Winston indirectly, while seeing the photograph (or its surface), or the television (or its screen) directly. But in what sense do I “see” the photograph or the television screen? Both surfaces possess particular perceptible properties – properties not shared by Winston Churchill – that pose absolutely no challenge to my visual equipment. Yet when I am engrossed in old photos or yet another documentary about WWII, I am utterly unaware of those properties marking the texture of the photographic paper, or the light layer of dust that clings to the curved screen of the TV. In some perfectly ordinary sense of the term, I do not “see” them at all. There is little surprising here: watching television does not usually involve paying much attention to the television itself, although we can easily enough turn our attention to it.[31] We are most likely to become aware of such properties when they present a difficulty for seeing an object: seeing the reflection of the Statue of Liberty in the gray waters of NY harbor can be impeded by the perceptual qualities of the water. The water is rippled, wavering, broken by wavelets, and frankly, rather murky. And so is the statue’s reflection. But does this mean that we see the Statue of Liberty only indirectly (and the reflection directly)? We are seeing under conditions of some difficulty or at least oddity -- that is, under conditions that require some extra effort to extract information about what we are seeing (whether we are interested in the Statue of Liberty or the water). Yet it seems quite likely that with a bit of practice, one could become quite adept at seeing the Statue of Liberty in the reflection – just as some artists become quite skilled at seeing the objects represented by anamorphoses, however much they defeat the recognitional skills of the untrained. But these cases do not seem so much to mark special kinds of perception as special and difficult conditions for sense-perception; they are analogous to seeing something in non-standard lighting conditions. And surely the distinction between direct and representationalist perception is not adequately captured by the difference between seeing something in full daylight and seeing it in twilight.