This paper has three goals. First, I want to look at Aristotle's views concerning self-knowledge. Aristotle, it seems, gives two contradictory accounts of how self-knowledge is possible, and one of the accounts has substantial intrinsic interest. Second and relatedly, I want to look at Aristotle's doctrine concerning the identity of the intellect of the knower and the object which is known. The view is, it seems, unmotivated by Aristotle's realism and contradictory to the interesting account of self-knowledge. Third, I will argue that the position developed to that point helps to solve two other problems, the problem of individuation of selves and the problem of epiphenomenalism.
The underlying concern in the first two parts is with the metaphor of the mirror. Aristotle's doctrine about the identity of knower and known seems to be underlain by a misplaced mirror metaphor, while the interesting account of self-knowledge seems to be underlain by a well-placed appeal to the same metaphor.
Let us ask how self-awareness is possible. First, assume that at least a decent chunk of the self is a set of acts of awareness of objects. If the act of awareness is distinct from the object of awareness, then it would appear that self-awareness is impossible. To be aware of these acts, which (at least) partially constitute the self, one needs further acts which will then constitute an enlarging self, infinitely. Aristotle solved this problem by making the act and object of awareness identical: "Thought thinks itself because it shares the nature of the object of thought..." (1072b20-21). Because the knower becomes identical with the object of knowledge by way of knowing the object, the knower is identical with the known. Since the known is known, and the knower is the known, the knower is known. Let us, not without prejudice, call this the trivial account, because if we accept it, self-knowledge appears to be a trivial side-effect of knowledge of objects. It's easy to know oneself, because oneself is identical to the known, and (here's the triviality) the known is already known. On Aristotle's account, it is impossible for someone to be knowledgeable but not self-knowledgeable.
This account seems intuitively to be false and to be liable to a wide variety of counter-examples. Here are three such examples. First: we usually do not think of children as being self-aware in a rich way, but we do not think of them as totally ignorant, either. Indeed, child prodigies might be aware of a great many objects and their natures, and be almost totally unreflective. Second: we are all familiar with the idea of the workaholic or genius who makes a mess of her personal life. Such a person knows a great deal about her work or studies, but has little idea how she ought to live. Even if ethics were her chosen area of study, she might not be able to apply general principles to her own life, not through ignorance of principles but through ignorance of self. Third: One might know a variety of facts about events which have an impact on one's own life, but not realize the nature of the impact itself. For instance, there is a musician, Al Stewart, whose music means a great deal to me. I know that I first heard his music in 1991, that his most well-known album came out in the mid-1970's, that my tape copy of one of his albums acquired a warble through overplaying, and a great deal more. But none of this tells me why the music matters to me; it does not tell me what it is about me that his music touches. Self-knowledge, it seems, is not a trivial consequence of the knowledge of objects.
There is an additional problem with the trivial account. Presumably Aristotle's doctrine about the self is at least modestly commonsensical in the following sense: you and I are, or have, different selves. But if the self is identical with the objects of its knowledge, then two persons which know all and only the same things would be the same person.
How might two persons know all and only the same things? Let us assume identical twins who follow exactly identical life-plans and who are inseparable. It seems that there is no way for them to know all and only the same things, because the difference in spatial perspective they bring to their experiences would cause them to have slightly different experiences. However, the self in this discussion is, at least largely, the intellect. The intellect becomes identical to its objects, but these objects are universals. And it seems not at all difficult to arrange for two persons to know all and only the same universals; our twins may have different experiences, but their experiences might well be similar enough to bring it about that neither of them experiences tokens of a type not experienced by the other, so that they will be aware of the same universals. If the self (or at least a large chunk of it) is identical to its objects, then the twins will be the same self, or at least more or less token-identical. (How could they be 'more or less' token-identical? If the twins each know the same universals, then their intellects will be token-identical with one another. If the self is identical to the intellect, then they are token-identical. But if the self is not identical to the intellect, but rather the intellect is a substantial fragment of the self, then the twins would have token-identical intellects, though their selves extend beyond their intellects. So the relation of each self to the other would be partial overlap, like the overlap between the top two-thirds of this piece of paper and the bottom two-thirds.) This is plainly wrong.
Not only do there appear to be good reasons to reject the trivial account of self-knowledge, Aristotle himself implies its falsity. If one becomes identical to the objects of one's knowledge, then one knows oneself by contemplating oneself. However, in a passage I will be discussing at length, Aristotle claims that "...we are not able to see what we are from ourselves..." (1213a15-16). This implies that something about the trivial account is false, because in that account one does know oneself through oneself. But Aristotle does not simply (and implicitly) reject the trivial account. He provides a different argument, which I will call the mirror argument, in the Magna Moralia :
Since then it is both a most difficult thing, as some of the sages have said, to attain a knowledge of oneself, and also a most pleasant (for to know oneself is pleasant) now we are not able to see what we are from ourselves (and that we cannot do so is plain from the way in which we blame others without being aware that we do the same things ourselves; and this is the effect of favour or passion, and there are many of us who are blinded by these things so that we judge not aright); as then when we wish to see our own face, we do so by looking into the mirror, in the same way when we wish to know ourselves we can obtain that knowledge by looking at our friend. For the friend is, as we assert, a second self. If, then, it is pleasant to know oneself, and it is not possible to know this without having some one else for a friend, the self-sufficing man will require friendship in order to know himself. (1213a13-26)
The passage requires some interpretation, and the interpretation some defense. As I understand it, the argument should begin with a disjunctive premise: (1) one knows oneself either through self, or through others. Aristotle doesn't say this, but he does consider only the two possibilities, so he is accepting it at least tacitly. He then introduces a negative claim: (2) one cannot know oneself through oneself. This is true, Aristotle argues, because we are biased judges in our own case. Let me call this latter claim the distortion thesis. (2) contradicts the trivial account, because the trivial account is an account of how one knows oneself through oneself. Luckily, Aristotle makes a positive claim: (3) one can know oneself through others. That this is true is suggested by the mirror analogy, and will be further justified below. Now, since one can know oneself through others but not through oneself, and these are the only ways one might know oneself, we can conclude: (4) one can know oneself only through others. Presumably such self-knowledge is a value to anyone, not just the already good.
This conclusion requires interpretation. When Aristotle says that one knows oneself through others, what does he mean? What is the self that is known? Is the known self the other self (the 'second self'), one's own self, or both (because they are substantially identical)? It seems that it must be one's own self, and not that of the other. This is because the reason we have difficulty knowing ourself through ourself is given by the distortion thesis. But the cause of the distortion is the object of knowledge. Thus, if our self were numerically identical to the second self, then the distortion thesis would show that (3) is false, just as it shows that (2) is true. So Aristotle appears to reject the trivial account.
What might (3) [and thus (4)] amount to? How might observing other people aid me in my quest for self-understanding? The mirror analogy is striking but only modestly helpful. What Aristotle says in a similar passage in the Nicomachean Ethics may help. Aristotle grounds an argument that we will gain pleasure through contemplating the good acts of others on the premise "we can contemplate our neighbours better than ourselves and their actions better than our own..." (1169b33-35) Let us call this the distance thesis: we can contemplate others more easily than we can ourselves, and we can thus form true judgments about them more easily than we form such judgments about ourselves. The distance thesis seems to justify (3). But if the distance thesis is true, then it is beneficial to the others that we form such judgments, for we may express such judgments to them, either explicitly or in action. By observing our responses to them, others learn about themselves; thus we aid them in their effort to know themselves. Thus it is to their benefit to have us as friends, because we act as mirrors to them.
Now, this interpretation may appear to go awry. Twice in NE IX,9, (at 1169b31-1170a4, and 1170a20-1170b18) Aristotle gives what I shall call the contemplation argument:
...if a thing's being one's own is one of the attributes that makes it pleasant, and if we can contemplate our neighbours better than ourselves and their actions better than our own, and if the actions of virtuous men who are their friends are pleasant to good men (since these have both the attributes that are naturally pleasant) if this be so, the blessed man will need friends of this sort, since he chooses to contemplate worthy actions and actions that are his own, and the actions of a good man who is his friend have both these qualities. (1169b31- 1170a4)
As I understand it, this argument runs as follows. (A) The acts of one's friend are one's own. This premise assumes that one's friend is oneself, in some sense, probably through type-identity with respect to goodness (or badness) to oneself. Call this the second self thesis. (B) The acts of one's friend are good. This premise assumes that one is, oneself, good. Call this latter claim the good self thesis. This thesis is required if we are to avoid a contradiction between the second self thesis and (B) which would render the argument unsound. Assume the second self thesis, (B), and the negation of the good self thesis. Then one is bad. In that case, by way of the second self thesis, we know that one's friend is bad. But if one's friend is bad, then (B) is false, because the acts of a bad person are typically bad. So we must reject one of (B), the second self thesis, or the negation of the good self thesis. If we reject the second self thesis, then we demotivate (A) and render the argument, for all we know, unsound. If we reject (B), then the argument will not go through. So we must reject the negation of the good self thesis and accept the thesis.
To continue with the main thread. (C): It is easier to contemplate the acts of one's friend than one's own. This is the distance thesis which I suggested justified (3) from the mirror argument. Thus, (D): It is easier to contemplate acts which are both one's own and good by contemplating the acts of one's friend than those which one performs oneself. The upshot is that, since such contemplation is good better than contemplating acts which are just good or one's own but not both one should seek out friends whose worthy acts one can contemplate.
My interpretation of the MM passage may appear to go awry because it may appear that Aristotle is there again giving the contemplation argument, and not the mirror argument as I have suggested. This is because of the second sentence of this excerpt from the passage, and especially the word 'for' which begins it: "when we wish to know ourselves we can obtain that knowledge by looking at our friend. For the friend is, as we assert, a second self." (1213a21-24) It seems more easy to interpret the passage as simply giving the contemplation argument again. We can gain knowledge of the self by looking at our friend, because our friend is us.
How might this work? One might well take Aristotle to be suggesting the following. One lacks self-knowledge, but has friends who are good. One does not know, initially, why one has befriended them. Over time, one acquires a sense of their goodness and realizes that this goodness is what initially drew one to them. But then, through acquiring the ability to recognize goodness, in others, one has also acquired the ability to recognize it in oneself. If one learns that acts of one's friends are worthy, then one is in a position to notice that one's own similar acts are similarly worthy. Let us call this the sharing account; it is the account of self-knowledge present in the contemplation argument. I will discuss it below.
If Aristotle nowhere gives the mirror argument, then the contemplation argument is nowhere sound in his work. This is because one premise of the contemplation argument is the good self thesis. But one cannot know that one is good by reflection, because of the distortion thesis. So one could not apply the contemplation argument to oneself and learn that one ought to search out good friends until one knew that one is good.
It is important that one seek out good friends specifically as a consequence of one's acceptance of arguments for the excellence of friendship, because if one does so otherwise than as a consequence of the best practical reasoning available to one, one's pursuit of friendship will not be truly excellent. Aristotle says that, "...in the moral part [of us] there are two types, natural excellence and excellence in the strict sense, and of these the latter involves practical wisdom." (1144b15-17) So only a friendship engaged in as a consequence of practical reasoning will be excellent in the strict sense. But since the practical reasoning involved in the contemplation argument is sound only if the good self thesis is true, one can know that the argument is sound only if one already knows that one is good. (The mirror argument, on the other hand, does not rely on the assumption that one is good.) Thus the self-knowledge afforded by the contemplation of one's good friends will only be known to be correct if the contemplation argument has already been accepted, and this acceptance requires that one know that one is good on the basis of something other than the contemplation of one's friends. This would appear to scotch the sharing account, because the sharing account requires that one already have good friends, but the act of having acquired them will have been worthy only if one did so due to practical reasoning, and the practical reasoning required in this case can only be done by someone who already knows that she is good.
Moreover, Aristotle does plainly accept the premises of the mirror argument. The distortion thesis is given in the MM passage, and the distance thesis in the NE passage. The lines of inference from these claims to (2) and (3) of the mirror argument are obvious.
The interpretation of the MM passage as giving the mirror argument turns on the interpretation of the metaphor of the second self. When Aristotle gives the contemplation argument, the second self is plainly another person who is (at least) type-identical with oneself with respect to goodness. If that is what Aristotle means by 'second self' in the MM passage, then the MM passage gives the contemplation argument. But the fact that 'second self' is a metaphor allows for a certain interpretive looseness. Aristotle need not mean the same thing by 'second self' in different contexts; indeed, a metaphor is rich and suggestive just to the extent that it can take on different meanings in different contexts. I suggest, then, that by 'second self' in the mirror argument Aristotle is referring to one's reflection. When I look into a mirror, I see, metaphorically, a second self. Likewise when I observe the responses of others to my actions: the other person is a second self not insofar as she is type-identical to me in some respect, but insofar as her responses to me indicate to me what kind of person I am. Aristotle himself suggests this reading, when he mentions that we try to discover our own nature in a mirror. The analogy would not be apt were the thing which we discover our own nature in literally like us or a copy of us, like a twin.
Let us return to the sharing account in this light of this interpretation of the mirror argument. Note that on the present account, the responses of others indicate to me what kind of person I am. The response by a person who is respected and thought worthy by me will probably have more weight in my self-reflection than the response by someone less respected. That a person who is good should seek to befriend me, for instance, might have great weight in my self-evaluation, but that a person who is not good should seek to befriend me seems insignificant. On my account, the benefit one receives from a knowledge of one's friend's goodness is that their friendly evaluation of one will help one achieve a knowledge of one's own goodness because one believes that one's good friend would not have befriended one were one not oneself, good.
But the sharing account presents a starkly different mechanism by which I infer my own goodness from that of my friend. On the sharing account, I identify my friend as good and then infer my own similarity to the friend. But this is appallingly arrogant, as an example might show. Let us say that my friend and I each pursue excellence in writing literature. If I acknowledge my friend's excellence, and then 'realize' that I, too, am good because my own writing is on a par with that of the friend, then I am not achieving a worthwhile kind of self-knowledge; the distortion thesis shows that my inference might well have been unsound. But if I acknowledge that my friend's writing is excellent, and she acknowledges the same about mine, then I have good reason to believe in my own excellence. If I am, in fact, excellent, then self-knowledge has been achieved because I am excellent and have good reason to believe so and do believe so. But the sharing account does not provide the same justification.
Let me hazard a guess which, if correct, would serve to encourage us in accepting the mirror argument. I imagine that the first sense of self-respect acquired by a child is a consequence of the cheerful and encouraging acceptance of that child's acts by her parents. The parents, in this case, play the role of the mirroring friend, and their goodness is accepted by the child on grounds of her complete dependence on them and their displays of apparent excellence in abilities not remotely possessed by the child herself. (Why would the child accept that the parent is not only powerful but good? Because of the dependence of the child on the parent, for a child to believe that her parents are vicious would be for the child to believe herself utterly doomed. Such a conclusion would, it seems to me, be repressed were it to emerge into the child's mind; she will believe her parents good because otherwise the world would be intolerable. Hopefully she will be right.) If this guess is correct, then the mirror argument seems to provide a powerful interpretation of the acquisition of self-respect: one respects oneself because one is respected by those one respects. And the parent-child relationship provides a paradigm case of this inference.
The present interpretation of these arguments derives largely from a discussion by Cooper (Cooper, 1980). Cooper likewise distinguishes the contemplation and mirror arguments (though he does not so label them) and finds the latter more compelling. However, our reasons for this are somewhat different. For Cooper, the contemplation argument begs the question it seeks to answer (the question whether the good life will include friendships), whereas for me, it begs a different question (the question of whether one is good). The overall point of both the mirror and contemplation arguments is to establish that the best life will be one with friends. For Cooper, though, the contemplation argument simply assumes that the good person will have friends. That is because the contemplation argument, on Cooper's interpretation,
...comes to the following:
(1) For a good person, life itself is a good and pleasant thing; it is always pleasant to be aware of oneself as possessing good things; therefore, the good person's awareness of himself as being alive is very pleasant and desirable to him (1170b1-5).
(2) A man's friend is to him a 'second self,' so that whatever is good for him as belonging to himself will also be good for him when possessed by his friend (1170b5-8).
(3) Since the good man's life and his awareness of it are pleasant and desirable to him, he will find the life of his 'second self' and his awareness of it also pleasant and desirable (170b8-10).
(4) But he cannot satisfy this desire to be aware of his friend's existence except by living in company with him, so he will need his friend 'to live with and share in discussion and thought with for this is what living together would seem to mean for human beings, and not feeding in the same place, as with cattle' (1170b10-14). (Cooper 1980, 318; internal citations Cooper's)
Cooper's complaint is that line (2) of this argument is unmotivated. For Cooper, Aristotle in line (2) is simply assuming that the good person will have friends. But since that is supposed to be the conclusion of the argument, the argument isn't very impressive. Now, here is what Aristotle says that Cooper interprets as line (2): "...as the virtuous man is to himself, he is to his friend also (for his friend is another self)..." (1170b5-6). It's not plain to me that Cooper's reading is the only one allowable or is even as plausible as one I shall present in a moment. On Cooper's reading, Aristotle's claim is an existentially quantified conjunction: the good person has friends and those friends have a certain feature, that of being the second self of the good person. But it seems that Aristotle could be giving a universally quantified material conditional instead; he could be specifying a trait which friends, if there are such things, in general possess, that of being second selves. Aristotle is saying that, if one has a friend, that friend will be a second self. Since there would be an advantage in contemplating the acts of oneself, but one can't do so directly, there is an advantage in contemplating the acts of one's friend, whose acts are indirectly one's own. And so we get the familiar contemplation argument again. The contemplation argument, on my reading, is rendered sound only by the previous acceptance of the good self thesis, while on Cooper's reading, it is rendered sound only by the previous acceptance of its conclusion, the thesis that one ought to pursue friendships. Aristotle is giving a non-circular argument on my account, so it seems more charitable to read Aristotle as I do. Nevertheless, the contemplation argument does not stand alone on my reading either; one can realize that it is sound when applied to oneself only if one has already accepted the mirror argument and learned one's goodness from one's friends.
But the distortion thesis has an additional effect on Aristotle's philosophy. Recall that Aristotle gives the abortive trivial account of self-knowledge. But the trivial account is not optional for Aristotle. Once he accepts that the knower becomes the known through knowing, the trivial account follows. Since the distortion thesis shows that the trivial account cannot be right, Aristotle is forced to reject the thesis that the knower becomes the known. I want now to defuse the impression that this would be a disaster for Aristotle's philosophy. I'll start by fleshing out a bit this 'knower becomes the known' mantra.
Aristotle adopts a model of conscious awareness in which the mind of the knower must reflect the known without distortion in order to know it. We can experience the world by way of our sense organs only because our sense organs can take on the perceptible form of the sensed objects; that is, they can become type-identical with those objects. Likewise for the mind: when the mind contemplates a universal, it becomes, in part, a token of that universal, and is type-identical with other things which exemplify that universal. Aristotle explains this view with respect to the sense organs by saying that, "...what has the power of sensation is potentially like what the perceived object is actually; that is, while at the beginning of the process of its being acted upon the two interacting factors are dissimilar, at the end the one acted upon is assimilated to the other and is identical in quality to it." (418a4-6) With respect to the intellect, he says that, "Actual knowledge is identical with its object..." (431a1).
Why does Aristotle say this? Aristotle accepts this doctrine because he accepts what I shall call (following Kelley 1986, esp. pp. 35-38) the diaphanous model:
If thinking is like perceiving, it must be either a process in which the soul is acted upon by what is capable of being thought, or a process different from but analogous to that. The thinking part of the soul must therefore be, while impassable, capable of receiving the form of an object; that is, must be potentially identical in character with its object without being the object.
Therefore, since everything is a possible object of thought, mind in order... to dominate, that is, to know, must be pure from all admixture; for the co-presence of what is alien to its nature is a hindrance and a block: it follows that it can have no nature of its own, other than that of having a certain capacity. Thus that in the soul which is called thought (by thought I mean that whereby the soul thinks and judges) is, before it thinks, not actually any real thing. For this reason it cannot reasonably be regarded as blended with the body: if so, it would acquire some quality, e.g. warmth or cold, or even have an organ like the sensitive faculty: as it is, it has none. (429a13-27)
The diaphanous model looks at cognition as occurring, as it were, in a mirror. To accurately reflect objects, a mirror must be free from distortion. But being anything but perfectly smooth and clear would be distortive. Thus, the ideal mirror has only smoothness and clarity as features. Likewise for the knowing subject. If the knowing subject had a nature of its own, then this nature would block its awareness of the world; the knower would already be , self-blindingly, the object of her awareness and so could not become various external objects. The diaphanous model assumes that, unless experience of objects is as it seems from the first-person point of view in no way affecting the appearance of the object then experience could not be of external objects at all. Aristotle's epistemic optimism, in combination with his acceptance of the diaphanous model, forces him to acknowledge that the intellect is identical with its objects. This forces the trivial account of self-knowledge, but the distortion thesis shows that this account is false. So Aristotle must give up his epistemic optimism, the diaphanous model, or the distortion thesis. It's clear, I think, that the diaphanous model is the problem.
On (as it were) reflection, it seems strange that anyone would ever accept the diaphanous model. Consider, for instance, the inverted spectrum thought experiment. Different persons, for some reason (probably having to do with the physical realization of their perceptual faculties), have different felt experiences under identical external circumstances, such that when one person has an experience like the one I have when I look at something blue, the other has an experience like the one I have when I look at something orange. The people in question would never notice the divergence and would both refer to the same patches of color in the world with the same words. Does this possibility, which is the possibility of perceptual relativity, suggest that either person is seeing wrongly? To the contrary, the fact that each person sees an object differently does not suggest that either sees wrongly. Rather, they each see the same object, differently. While the experience is different, the object is the same. The experiences are connected to their objects by way of causality, not similarity. Unless we already think that the felt experience is the object of the experience, no problem emerges. But that the felt experience must be identical to its object was the point at issue.
Recall the critical discussion of the trivial account. There I pointed out a consequence of the view that the intellect becomes identical to its objects: there might be fewer selves than there ought to be, because different persons might be aware of all and only the same universals. But now we can see that Aristotle accepts the idea that the knower becomes the known only because of his acceptance of the diaphanous model. Once we see that the diaphanous model is not necessary for epistemic optimism, we can reject it.
Now, one consequence of the knower becomes known doctrine was that the intellect could not, unlike sensation, be physically realized: "...that in the soul which is called thought... cannot reasonably be regarded as blended with the body..." (429a22-25). This gives rise to a problem of individuation for selves: to whatever degree selves are 'that which in the soul is called thought', such selves which experience tokens of the same types will be identical with one another. That this is possible on the trivial account was one of my motivations for rejecting that account. And that the trivial account was implied by the knower becomes known doctrine was my motivation for rejecting that doctrine.
Usually, Aristotle appeals to matter to individuate type-identical substances. But he cannot do this for intellects (and thus for selves, to whatever degree selves are intellects), because he regards intellects as immaterial. But now that we have rejected the knower becomes known doctrine, we can allow the intellect to be physically realized without giving up epistemic optimism. So the matter of the different physical realizations of intellects will individuate those intellects.
Allowing intellects to be physically realized turns out to be necessary if Aristotle is to avoid epiphenomenalism. How might Aristotle have landed himself in epiphenomenalism? By being what we nowadays call a functionalist. There is much in De Anima to support a functionalist reading (see Nussbaum and Putnam 1992). Consider this suggestion of Aristotle's: "...a physicist would define an affection of soul differently from a dialectician; the latter would define e.g. anger as the appetite for returning pain for pain, or something like that, while the former would define it as a boiling of the blood or warm substance around the heart." (403a29-403b1) The physicist, Aristotle explains, seeks material causes, while the dialectician seeks formal causes. That is, the physicist seeks causes of entities construed as material things, while "attributes not considered to be of this character he leaves to others..." (403b13-14) Note that the attributes are not said not to be of the material character. Rather they are not considered to be of this character. That is, Aristotle is not denying the materiality of the attributes studied by the dialectician, he says only that the dialectician does not study them insofar as they are material. Under what aspect, then, might the dialectician study them? Since it is the soul de anima which is being studied, the dialectician would be concerned with these attributes insofar as they are ensouling or animating attributes. The soul is "...substance in the sense which corresponds to the account of a thing. That means that it is what it is to be for a body of the character just assigned..." (412b11-12), where the character just assigned is "natural organized body" (412b5) The dialectician will study attributes of living things insofar as they are attributes of living things, even if the same attributes might also be attributes of non-living things.
What's essential here is that the dialectician need not focus on material concerns in her study. She studies attributes which might, at least in principle, be realized in any of a variety of materials. For instance, Aristotle's dialectical account of anger, that it is "the appetite for returning pain for pain, or something like that", might well be true despite the patent falsity of his physical account of anger as "a boiling of the blood or warm substance around the heart."
Functionalism in analytic philosophy has traditionally been regarded as a form of property dualism. Identifying mental properties with properties physically characterized was rejected by functionalism on the basis of the multiple realizability argument: mental states can be realized by entities which lack the physically characterized properties with which the identity theorist had sought to identify them. Thus mental states must be identified with functional states of entities, which states are not physical (not characterizable in the language of the physical sciences). Thus mental properties are non-physical.
However, any such form of property dualism is scotched by what I'll call the mental causation argument. Actions are physical events, with physical properties. All such events, and all of their physical properties, are caused only by other such events and like properties. No explanation of any such event, or any physical feature of such an event, will ever cite a non-physical cause. Thus, if mental properties are functional and non-physical, then they will never be cited in explanations of action. This epiphenomenalist conclusion threatens any form of property dualism and is plainly intolerable.
When faced with the multiple realizability argument against identity theory, and the mental causation argument against property dualism, how can we respond? Jaegwon Kim (in Kim 2000) responds by suggesting that functionalism should never have been construed as a form of property dualism. Rather, functionalism allows mental states to be identified with physical states disjunctively. Anger is either the boiling of the blood around the heart, or... whatever anger is in Martians, computers, and so forth.
But if we read further in the DA passage noted earlier, then we see that this may be what Aristotle had in mind for those functions of soul other than intellect; he continues to suggest that "...the genuine physicist..." will study "...that form in that material with that purpose or end..." (403b7-9, emphasis mine). So Aristotle has no problem with epiphenomenalism about states of soul other than intellectual states. But thanks to the mental causation argument, we know that functional states of organisms, even physically realized, are not as such causally efficacious unless they are identical with physical states of those organisms. So if intellect is to be causally efficacious (and let us hope that it is, for we would like to live excellently excellence occurs only in accordance with reason), then it must be identical to some material component of the person.
The argument of this paper turns on the means of self-knowledge. In the mirror argument, Aristotle suggests the distortion thesis, which means that knowledge of self must come by way of interaction with others. But if the knower is identical with the known, then, since one's intellect is identical with the known objects, one's intellect would be knowable otherwise than by interaction with others. Even if the intellect is not the self but merely a substantial fragment of the self, this means of self-knowledge is denied by the distortion thesis. But this means of self-knowledge which I dubbed the trivial account is forced on Aristotle by his acceptance of the diaphanous model and his epistemic optimism. I tried to suggest that the diaphanous model is simply unjustified, and that epistemic optimism can exist without diaphanousness. But Aristotle's motivation for denying that the intellect is physically realized had to do with the preconditions of diaphanousness; once we no longer seek diaphanousness, we can allow the intellect to be physically realized. This has two advantages: it allows for material individuation of intellects, and it eliminates a worry that the Aristotelian intellect might be epiphenomenal.
To put the main theses of this paper in terms of the mirror metaphor: Aristotle is wrong to say that the mind mirrors the world, but right to say that we each mirror each other.
Aristotle. 1984. The Complete Works of Aristotle . Barnes, Jonathan, ed. Princeton: Princeton UP.
Cooper, John. 1980. "Aristotle on Friendship". in Rorty, Amélie Oksenberg, ed. 1980. Essays on Aristotle's Ethics . Berkeley: UC Press. pp. 301-340.
Kelley, David. 1986. The Evidence of the Senses . Baton Rouge: Louisiana State UP.
Kim, Jaegwon. 2000. Mind in a Physical World . Cambridge: Cambridge UP.
Nussbaum, Martha, and Putnam, Hilary. 1992. "Changing Aristotle's Mind". in Essays on Aristotle's De Anima . Nussbaum, Martha, and Rorty, Amélie Oksenberg, eds. 1992. Oxford: Oxford UP. pp. 27-56.