To be or not to be in Descartes' shadow?
The title of my paper stems directly from the title of Professor Von Wright's book, In the Shadow of Descartes, published a couple of years ago in 1998. Much of Professor Von Wright’s work in the past has been devoted to the explanation of action. In the Shadow of Descartes, not only does he pursue and deepen some of his earlier insights, but he also faces some of the ontological perplexities prompted by his views on the explanation of action. In addition to topics in the theory of action, in his recent book, Professor Von Wright also addresses fundamental issues in the philosophy of perception. In the present paper, I will first examine some of the ontological questions raised by Professor Von Wright’s views on the explanation of human actions. Then I will scrutinize some of his views about visual perception. But before proceeding, I cannot resist the urge to share briefly with you some of the epistemological and ontological puzzles which the very title of Professor Von Wright’s recent book raises for me.
I seem to remember that Newton wrote somewhere that he discovered the theory of universal gravitation while sitting on the shoulders of such giants as Copernicus, Kepler, Galileo and Huygens. It is not insignificant, I think, that while physicists can think of themselves as discovering the ultimate laws of nature and as doing so by operating on the shoulders of earlier physicists, philosophers tend to think of themselves as doing conceptual analysis in the shadow of some or other towering figure. In fact, I find the metaphorical title of Professor Von Wright's book fascinating and perplexing: its ontological implications are somewhat tantalizing. I am not sure how seriously Professor Von Wright intends these implications to be. In fact, I suspect that he would be happy if his title was deprived of any ontological significance whatsoever and was given a non-literal interpretation.
Shadows, I think, are puzzling. I mean shadows of physical objects are puzzling. Shadows arise from the interception of the trajectory of light by an opaque physical object. If and when light rays originating from a light source fall upon some physical object, then the physical object casts a shadow upon the surfaces of neighboring physical objects.
Shadows of physical objects raise both epistemological and ontological puzzles. On the one hand, perceptual psychologists tell us that shadow information is very useful for the detection of visual attributes of physical objects: shadows contribute to the processing of the size, orientation, distance and motion of objects that cast them. On the other hand, psychology tells us that whereas shadow information contributes powerfully to the automatic processing of visual attributes of physical objects, shadows themselves are processed pre-attentively (i.e., very early on) so that rarely if ever do they become accessible to visual awareness when one is visually aware of some physical objects. Only through the effort of spatial attention do we become visually aware of shadows of physical objects. Shadows also raise ontological puzzles. What kind of things are they? It seems as if given a light source, given a physical object lying on the trajectory of light rays emanating from this light source and given some neighboring physical objects, then you have all the ingredients necessary to produce shadows. It seems as if physical objects will cast shadows whether or not any minded creature is around to perceive them. Shadows are not mental images. Nor are they thoughts or ideas. If they are not mind-dependent, then presumably they are mind-independent, i.e., physical objects. If so, then they are rather queer physical objects. For example, you can walk through them.
Be that as it may, we have little doubt that physical objects do cast shadows. What about mental shadows? Can thoughts and other mental entities cast shadows? If thoughts are immaterial entities, then it may seem hard to see how they could: if the process whereby mental entities cast their shadows is quite unlike the process whereby physical objects cast theirs, then why call the result of the former process “shadows”? Along Cartesian dualistic lines, one might take human reason to be a source of mental light. Suppose thought P lies on the trajectory of some mental light rays. P might then cast its shadow on some neighboring thoughts. Whatever one thinks of this neo-Cartesian story, it is unlikely that physicalism will be more hospitable to the idea that thoughts may cast shadows. Current techniques of brain imaging may allow us to visualize the areas of the brain involved in some cognitive task. Arguably, if thoughts are brain states (or brain processes), then perhaps brain imaging may make us able to visualize aspects of thought processes (or thinking under some aspect). However, unless light rays could travel through an individual’s skull, it is very unclear how brain processes — hence thoughts — could cast any shadow. And if light rays could travel through an individual’s skull, upon which surface could a brain state cast its shadow?
I will not investigate the ontology of shadows any further in this paper. I will now examine some of the ontological implications raised by some of Professor Von Wright’s insights into the explanation of human action.
First of all, given the towering presence of Descartes’ work in the background of Professor Von Wright’s recent reflections on action concepts, I want to pause to reflect on the differences and the connections between two kinds of dualism: ontological dualism and methodological dualism.
Ontological dualism is the view, famously held by Descartes, according to which things can be divided into two kinds: physical (or material) things (e.g., bodies) and mental (or immaterial) things (e.g., ideas). Whereas it is constitutive of physical things that they occupy three-dimensional space, thought (or thinking) is the constitutive property of mental things. As Professor Von Wright puts it in his book (p. 125), “mental things Descartes called ‘thoughts’. They have no extension”. Much of philosophy of mind since Descartes has been devoted to reflecting on ontological (or substance) dualism. Monist physicalists widely hold it against ontological dualism that it turns the problem of mental causation into an unsolvable mystery. Suppose my left hand goes upwards as a result of my intention to raise it. A mental event (my intention) caused a physical event (the motion of my left hand). If my intention is not itself a physical event, how could it produce any physical effect at all? This is the mystery mental causation raises for ontological dualism.
As I understand it, Professor Von Wright squarely rejects ontological dualism. He wants to resist what he calls (p. 108) “the temptation to substantialize the mind”, i.e., the temptation both to “attribute to the mind a kind of ‘shadowy existence’ as an immaterial and yet somehow material, ‘ethereal’, thing” and to think that “there can be such a thing as a ‘disembodied mind’”. He too seems to think that ontological dualism founders on the puzzle of the causal interaction between the mental and the physical.
The contemporary issue of methodological dualism arose in the context of the late nineteenth century division between the Naturwissenschaften and the Geisteswissenschaften. To a first approximation, methodological dualists deny and methodological monists assert that human mental and social phenomena can be studied according to the same standards of rationality as any other natural (e.g., atomic, cosmological, astronomical, chemical, geological or biological) phenomena. According to methodological monists, standards of rationality prevalent in the natural sciences should also prevail in the study of mental and other human phenomena — in particular in the explanation of human actions. The Diltheyan idea that the explanation of human actions must proceed via Verstehen (i.e., the empathetic understanding of an agent’s reasons) is, I take it, a clear example of a methodological dualist claim.
Professor Von Wright does, I think, subscribe to a weak version of methodological dualism: like many philosophers, he thinks that actions explanations are typically reason explanations. Wittgenstein seems to have held the view that reason explanations cannot be causal explanation on the grounds that reasons are not causes. Perhaps however, as Davidson (1963) has argued, reasons may be causes and reason explanations may be causal explanations after all. But if they are — as I am inclined to believe —, then they are sui generis in the following sense: reason explanations are supposed to make their explanandum intelligible. For example, a reason explanation ought to reveal how what the agent did was the thing to do in the circumstances. Not any causal explanation of an action will make the action intelligible because not any causal explanation will reveal the agent’s reasons. I call Professor Von Wright’s version of methodological dualism weak because, as far as I can see, nothing in his writings precludes the possibility that in the future the natural sciences may give rise to entirely new concepts which may be able to licence satisfactory causal explanations of human actions. Professor Von Wright’s view of the matter seems to be that so far, the explanations of human actions — reason explanations — are sui generis.
On Descartes’ own view, what I call anachronistically “metholological dualism” is connected to ontological dualism. On Descartes’ view, human intentional action is unlike the behavior of anything else in the universe because human beings have the ability to freely choose one among several courses of action. Descartes links the notion of human free will (i.e., the freedom of choice) with the indeterminacy of human action. As he put it in his Principles of Philosophy (I, 41), “we are so conscious of the liberty and indifference which exists in us that there is nothing that we comprehend more clearly and perfectly” so much so that “it would be absurd to doubt that of which we inwardly experience and perceive as existing within ourselves just because we do not comprehend a matter from its nature we know to be incomprehensible”. Hence, human action is not determined. In Descartes’ own view then, free will is one source of ontological dualism: unlike the behavior of any machine that falls under the principles of mechanical philosophy, human action is indeterminate. On Descartes’view, only minds — only things not falling under the principles of mechanical philosophy — could give rise to actions that result from free will.
One of the historical roots of the view that methodological dualism is consistent with the rejection of ontological dualism can, I think, be traced back to Kant’s compatibilist attempt to reconcile free will with determinism. In the Critique of Pure Reason, Kant’s project is to limit the explanatory scope of what he calls “theoretical reason” and make room for what he calls “practical reason”. Kant seems to hold the view that one can adopt two distinct standpoints on any human intentional action: the descriptive standpoint of theoretical reason and the normative standpoint of practical reason. From the standpoint of theoretical reason, any bodily motion is explainable by natural causal laws. From the standpoint of practical reason, one ought to be able to see the action’s agent as capable of free will so as to make room for the requirements of morality. The goal of Kant’s epistemological dualism between theoretical and pratical reason is to severe the link between free will and ontological or metaphysical dualism. Notice that Kant’s maneuver has a cost: from the standpoint of explanation, the normative is divorced from the descriptive or, as John McDowell has recently put it, “the realm of reasons” is divorced from the “realm of natural laws”.
Contemporary advocates of the view that distinctively human phenomena are not amenable to the standards of rationality that prevail in the natural sciences have built upon Kant’s insight. Professor Von Wright himself adopts a deflationary or therapeutic attitude towards free will. In In the Shadow of Descartes (p. 96), he writes that “causal determination of events in nature cannot constitute a ‘threat’ to the freedom of human action (‘free will’). To think that it could is to be guilty of conceptual misunderstanding”. In this respect, Kant’s insight allows him to stand outside Descartes’ shadow. Methodological dualists claim that reason explanations of human actions do not reduce to other kinds of causal explanations, e.g., to physiological explanations of bodily motions. On their view, the former stand to the latter in the same relation as the concept of an action stands to the concept of a bodily motion. The former no more reduces to the latter than mental concepts reduce to non-mental (e.g., to behavioral or neurophysiological) concepts. As the recent history of the philosophy of mind suggests, one can hold this anti-reductionist epistemological position (i.e., with respect to concepts and explanations) while sticking to physicalism, i.e., without endorsing ontological dualism. This is what both functionalism (of the Putnam-Dennett-Fodor variety) and Davidson’s anomalous monism suggest.
In In The Shadow of Descartes, Professor Von Wright argues convincingly that behavior cannot be equated with bodily motion. If I understand it correctly, his view is that behavior may consist in (or be constituted by) bodily motion, but the former is not identical with the latter. As he puts it (p. 98), when I raise my arm (e.g., because I have decided to do so), then my arm rises, but I raise my arm. If and when I get up from my chair, my body’s shape may change but I get up. If and when I walk across the room, my body moves across the room but I do the walking. I do share Professor Von Wright’s intuitions. My left hand may rise either as a result of my raising it or as a result of your raising it. In both cases, my left hand may follow exactly the same trajectory in space; its physical motion may be spatially indistinguishable in both cases. Nonetheless, in one case, I moved it. In the other case, you moved it. So far so good.
Now, as Professor Von Wright recognizes (p. 98), the question arises: “is behavior then something ‘over and above’ its ‘visible manifestation’”? Professor Von Wright’s solution to this query is to appeal to the notion of intentionality: “Normally”, Professor Von Wright writes (p. 98-99), “when we say that a (living) being behaves in a certain way and do not wish to identify what we say with a statement about changes in that being’s body, we think of the behavior as intentional, as something the being does, and of the bodily movements as having been intentionally performed”. Professor Von Wright’s view then seems to be that what distinguishes behavior from mere bodily motion is that the former, unlike the latter, involves an intentional origin. As he puts it (p. 104), “behavior is movement conceptualized under the aspect of intentionality”. Since the attribution of intentionality to a compass is at best metaphorical, the movement of the needle of a compass does not qualify as behavior. Not do the movements performed by plants — even by carnivorous plants (see p. 99) — since the attribution of intentionality to plants too is at best metaphorical.
Here I would like to distinguish two issues. One is the ontological question whether the distinction between behavior and bodily motion vindicates ontological dualism. The other question is the role of intentionality in the distinction between behavior and bodily motion. I turn to the latter first.
Here, I think I would like to resist Professor Von Wright’s move. There is, I think, an alternative to Professor Von Wright’s way of cutting the pie. First of all, we do not have merely two contrastive concepts but three: bodily motion, behavior and action. I agree with Professor Von Wright that it would be wrong to identify behavior with bodily motion for roughly the reasons he adduces. System S’s behavior is the process whereby some bodily motion m of some of S’s parts is being produced by some cause c internal to S. Neither the internal cause c nor the effect (i.e., the motor output) m is identical with S’s behavior. Rather, both c and m are components or constituents of S’s behavior. Behavior is the process whereby some motor output is internally caused. What, I think, lies at the heart of the distinction between motion and behavior is the possibility of analyzing a complex system into parts so that one event happening in the system may cause some other parts to move. If this is correct, then the distinction between bodily motion and behavior can be drawn without appealing to intentionality. In fact, I would argue that intentionality is relevant, not to the distinction between behavior and bodily motion, but to specifying among behaviors those that deserve to be called actions.
There is certainly a distinction to be drawn between things I do and things that happen to me. If I am hit or bitten to a dog, clearly the hitting and the biting are not things I do, but things that happen to me. Their causes are not internal to me. Granted, there are borderline cases such as growing up, getting sick or having an erection. Also there are things done by parts of me that are not to be confused with me: my hair grows, my heart beats, my immune system reacts to bacterial infections. None of these things are things I do. Among the things I do, there are things I do for reasons and things I do for no reason. The latter — the things I do for no reason — are nonetheless things I do. For example, salivating, blushing, shivering, perspiring or urinating are things I do for no reason. Unlike things I do for reasons, things done for no reason are not to be explained by the fact that I have intentionality, i.e., by the fact that I have representations in general and beliefs in particular. So intentionality demarcates a subset of behaviors that deserve to be called actions. If and when, for example, I move my right hand to the vicinity of a glass of water in front of me because I believe that it contains water and I want a sip of water, then what I do is an action — it is something done for a reason. It is explainable by what I want and believe. In this sense, an action is a piece of behavior explainable (or caused) by beliefs and desires.
On the view sketched here, intentionality matters to action, not to behavior per se. Plants and machines may not act out of their beliefs and desires since they have none, but they may behave nonetheless. Turning the furnace on is typical of thermostat behavior. Sheding its leaves in the winter is typical of deciduous tree behavior. Consider now one implication of Professor Von Wright’s ontological question: if behavior is not identical to bodily motion — if it is something over and above bodily motion —, then is it not something non-physical? Is not ontolological dualism vindicated? On p. 108 of In the Shadow of Descartes, Professor Von Wright expresses his worry that the view that behavior is something over and above bodily motion might pave the way to ontological dualism, i.e., that it might encourage the temptation to “substantialize the mind”.
Although I cannot fully address the issue here, I would like to suggest that the view that behavior is not identical to bodily motion does not lead inevitably to ontological dualism. As I said, on my view, intentionality is crucial to action, not to behavior per se. So I need not address the ontological status of intentionality (or mental content) here. The point is that behavior is a process, not an event: it involves a bodily motion, its internal cause and the causal relation holding between the cause and its effect. So far as I can see, the view that behavior is a process involving a bodily motion, its internal cause and the causal relation is compatible with a purely physicalist ontology.
In the second chapter of In the Shadow of Descartes, “On Sensations and Perceptions”, Professor Von Wright addresses some fundamental issues in the philosophy of perception to which I now turn. Here again, I think Professor Von Wright tries to distance himself from the shadow cast by Descartes. In addition to Wittgenstein, I also seem to detect the influence of John Austin’s views from Sense and Sensibilia. I will start out by laying out some crucial Cartesian assumptions that seem to me to lie in the background of Professor Von Wright’s discussion.
As I already said, Descartes assumed that the fundamental property of minds (or mental things) is thought, as opposed to sensory experience. As is well-known, he held on epistemological grounds that the latter, unlike the former, is a fundamental source of error. The assumption that thought (by which he meant conceptual representations) is crucial to minds had two important consequences. First, it led him to the transparency assumption, i.e., to the view that the mind must be transparent to itself. As Descartes wrote in Part Two of his Response to Arnauld’s Objections to The Meditations on First Philosophy:
As for the fact that there can be nothing in the mind, in so far as it is a thinking thing, of which it is not aware, this seems to me self-evident. For there is nothing that we can understand to be in the mind, regarded in this way, that it not a thought or dependent on a thought. If it were not a thought or dependent on a thought it would not belong to the mind qua thinking; and we cannot have any thought of which we are not aware at the very moment when it is in us. In view of this I do not doubt that the mind begins to think as soon as it is implanted in the body of an infant, and that it is immediately aware of his thoughts, even though it does not remember this afterwards because the impressions of these thoughts do not remain in the memory.
According to the transparency assumption, one cannot think a thought unless one is aware that one is thinking it. As Professor Von Wright remarks (p. 125), Descartes used the term ‘thought’ in a broad sense. I take it that when he writes that “the best modern word for Descartes’ ‘thought’ or ‘thinking’ is perhaps consciousness”, he is mostly thinking of this Cartesian emphasis upon the higher-order awareness whereby thinking involves thinking that one is thinking.
Secondly, the view that thought is constitutive of minds led Descartes to embrace what I wish to call an intellectualist or a judgmental view of visual experience and/or perception. Remember what Descartes famously wrote in the Second Meditation about his perception of the piece of wax:
We say that we see the wax itself, if it is there before us, not that we judge it to be there from its color or shape; and this might kead me to conclude without more ado that knowledge of the wax comes from what the eye sees, and not from the scrutiny of the mind alone. But then if I look out of the window and see men crossing the square, as I just happen to have done, I normally say that I see the men themselves, just as I say that I see the wax. Yet do I see any more than hats and coats which could conceal automatons? I judge that they are men. And so something which I thought I was seeing with my eyes is in fact grasped solely by the faculty of judgment which is in my mind.
The essence of Descartes’ intellectualist or judgmental view of visual experience is that seeing is judging or that all seeing is seeing that something is the case. It follows from what Descartes writes in this passage that unless a creature believes that he or she is seeing men out his or her window he cannot see men. I further take it that unless a creature possesses the required concept, she cannot form the appropriate belief and hence judgment. Of course, it is important to remind oneself that Descartes’ judgmental view of visual experience arises in the context of his epistemological project of providing a response to scepticism, not in the context of an investigation of the phenomenology of visual experience.
In the second chapter of In the Shadow of Descartes, Professor Von Wright makes a clear distinction between perception (or perceptual experiences) and perceptual reports (or the attribution of perceptual experiences). The distinction is important. What is true of perceptual reports is not ipso facto true of perceptual experience. Against the phenomenalist tradition in epistemology, Professor Von Wright argues (p. 45) both that “sensation language is secondary to perception language” and that “physical language is prior to perception language”. As he puts it (ibid.), “I cannot say that I seem to see a bird unless I can identify something as a bird”. I entirely agree with him. The judgmental or intellectualist theory is not true of perceptual experience; it is true of perceptual reports. I would not say that my visual experience (or impression) is as of a bird unless I judged it to be, i.e., unless I believed that I am seeing a bird. And I could not believe myself to be seeing a bird unless I could identify a bird — something I could not do unless I possess the concept of a bird.
We can, I think, go one step further in the linguistic analysis of reports of visual experiences. A person would not say that she saw a bird unless she believed it — i.e., unless she believed that what she saw was a bird or unless she had identified a bird by visual means. But now Moore’s paradox tells us that first-person belief reports interestingly differ from third-person belief reports. So I may say non-paradoxically: “It is raining but John does not believe it”. But I may not similarly say: “It is raining but I do not believe it”.
A similar asymmetry arises in the case of perceptual reports. I cannot truly say that I saw a bird unless I believe it, i.e., unless I believe that I saw a bird. In other words, I cannot truly say that I saw a bird without having identified what I saw as a bird. But what about third-person perceptual reports? Suppose I say that Anna saw a bird. Does it follow that Anna believes that she saw a bird or that she identified one? I think not.
In fact, on the Gricean view, which I favor (and which, on my view, has been persuasively pursued by Fred Dretske in his 1969 book, Seeing and Knowing), saying that Anna saw a bird pragmatically suggests — in Grice’s terms, it carries the implicature — that she identified what she saw as a bird. But no more. On Grice’s view, when what one says carries an implicature, the truth of what is said is consistent with the falsehood of the implicature. In other words, one cannot conclude that what was said — the proposition literally expressed — is false from the falsehood of an implicature. Consider the following case. Anna is a 9 months old human baby. She is comfortably sitting in her baby chair. Fifty centimeters from her head is a lap top computer with a blue monitor lit up. Suppose Anna’s visual system is in good working condition. The lighting is good. Now suppose that Anna moves her left hand towards the monitor of the lap top. I say: “Anna saw the lap top”. I think that what I said can be true, even though it may misleading imply (in Grice’s sense) that Anna believed that what she was seeing is a lap top or that she identified the target of her hand movement as a lap top. It may be true that Anna saw the lap top. This is why she moved her hand towards it. But she has no word for it. Nor does she possess the concept of a lap top. Although saying of Anna that she saw the lap top may misleadingly suggest that she identified it as such, it may nonetheless be perfectly true.
So perceptual reports are one thing. Perception is something else. To take an example of Professor Von Wright’s, suppose I misidentify a flower as a bird. If asked, I would say that I saw a bird. What did I really see? According to Professor Von Wright (pp. 45-6), what I saw was a flower because I stood in relation to a flower, not to a bird. I saw a flower that I mistakenly took to be a bird. Interestingly, as Professor Von Wright notices, I had a visual impression of a bird. But the impression was something I had, not something I saw. To take another example of Professor Von Wright’s, suppose now that I saw a bird with a red head. Suppose I cannot tell the color of its head. “Perhaps”, as Professor Von Wright insightfully puts it (p. 47), “I saw the bird’s head without noticing what color it was”. Perhaps I saw the color but it did not register, as one says. “This surely can happen”, writes Professor Von Wright. “Someone might then say: You must have seen the color, only you did not notice it”. Here, I think, Professor Von Wright is hitting the right nail head on. Take two other examples: you tell me that I must have seen your mother, since she was the only woman at 10:00AM at the baker shop and I was at the baker at 10:00AM and I did see a woman there. Indeed, I must have seen her. But having never met, I could not have known that the woman I saw there was your mother. I could not have identified the woman I saw as your mother. Suppose Anna (a five year old) sees a kangaroo. But since she does not know what kangaroos are, she cannot form the belief that she saw a kangaroo. She nonetheless saw one.
There are two complementary anti-Cartesian (anti-intellectualist or anti-judgmental) points to be made here. First, it is a psychological fact that most human adults cannot refrain from conceptualizing their sensory experiences. A normal human adult will automatically apply his or her concept of a kangaroo upon seeing one in good viewing conditions. It would, however, be wrong to conclude that unless a creature can apply the concept of a kangaroo, she will not be able to see one. Secondly, I may be an authority about what I believe, but I am not always an authority about what I see. Whether or not I believe so and so about what I saw depends upon my conceptual resources, my beliefs and background knowledge. You may inform me about what I saw if you are in possession of conceptual resources and background knowledge that I am lacking.
Professor Von Wright’s insistence upon the fact that I may not be the ultimate authority about what I see reveals the deep connection between the Cartesian intellectualist assumption that seeing is judging (or believing) and the transparency assumption. Let me remind you that according to the transparency assumption, one cannot think a thought unless one is aware that one is thinking it. Granting that I am an authority about what I believe (something that might be challenged on Freudian grounds), it would only follow that I am an authority about what I see if seeing was believing. If not, then even if I am an authority about what I believe, it does not follow that I am ipso facto an authority about what I see.
As I said, I entirely agree with Professor Von Wright’s distinction between perception and perceptual reports. The latter is constitutively linguistic and conceptually loaded. The former is not. Sensation language is derivative upon perception language. The latter in turn is derivative upon physical language. Not implausibly, Professor Von Wright claims that with respect to perception, the dependency holds upside down: “without visual impressions there is no seeing”, writes Professor Von Wright (p. 46). Although it sounds plausible to
In a sense, I think this is uncontroversially true. Perhaps however, recent discoveries in neuropsychology and the neuroscientific study of the visual system may cast doubt on