University of Chicago
Phil333,A01-lecture notes 5 (took two classes)
‘The Cartesian Illusion’
Descartes; Wittgenstein, Blue Book; Strawson, The Bounds of Sense, part 3.II; Evans, Varieties of Reference; Nagel, The View from Nowhere
Strawson’s master claim is that Descartes’ belief that he can conceive of himself existing without any property except thought stems from a peculiar feature of introspection. It will take some work to put us in a position to understand this claim.
Warning: my reconstruction of Strawson’s argument may end up sharing the defect many see in Strawson’s reconstruction of Kant’s argument: namely, excessive charity (assuming I’m right to want him to have the argument I will ascribe to him) at the expense of rigorous textual fidelity.
1. Introspection and immunity to error through misidentification
‘Introspection’ is sometimes used by philosophers to refer to a controversial account of how introspection, understood in the above general sense, works. (Namely, that it’s quasi-perceptual.)
Notable features of introspection:
Authoritative. I.e., there is
a presumption of truth for any belief arrived at by introspection.
Now, introspection is not the only way in which we can arrive at beliefs about our mental states: Sometimes we observe our behavior and draw conclusions.
We just noted some examples. And consider Robert Motherwell’s remark, quoted by Davidson: “Most good painters don’t know what they think until they paint it.”
A third feature of introspection:
3. Beliefs acquired through introspection are immune to error through misidentification.
What does that mean?
There are two ways I could go wrong in forming a simple belief on the basis of certain information (“simple” meaning of the form that S is f). I could get the subject wrong, or I could get the predicate wrong.
Say I see someone grab his head, and form the belief on that basis that Bob is in pain.
I may be right that the person is Bob, but wrong that it’s pain that Bob is in. Maybe Bob’s in shock, rather than pain.
Or I may be right that the person I saw is in pain, but wrong that the person is Bob. This is a misidentification. You’re wrong about which person or object is in question. [Pryor’s distinction need not be drawn for my purposes.]
Now, we have said that beliefs acquired through introspection are not infallible. They can be wrong. The point of 3 is that if they’re wrong, they’re always wrong about the predicate, not the subject. In introspection, you can’t misidentify yourself.
This the point made by Wittgenstein in the passage from The Blue Book.
Suppose you look down in a crowded dance hall, see a foot on fire and form the belief that your foot is on fire. But in fact it’s not your foot. You’ve misidentified the person whose foot it is, confusing him with yourself.
But suppose you introspect that you have a headache. Here there’s no possibility of misidentifying the person in question. In the previous example you could say, someone’s on fire, but is it me? But in this case it doesn’t make sense to think: someone is in pain, but is it me?
Again, the point is not that introspectively-derived beliefs are infallible. They’re not immune to being wrong; they’re immune to being wrong in a particular way.
Consider another case. Suppose you introspect that you love person X. Here, unlike the case of being in pain, we can make sense of the possibility that you’re wrong. But how might you go wrong? Well, maybe you just like X. It’s not love, but like, that you have for X. You’re wrong about the predicate. But does it make any sense in this context to think, someone loves X, but is it me? No.
Can point the point this way: in introspection, the question of who is undergoing the mental states is simply not in question. It is automatically you, the person introspecting.
Beliefs about one’s mental states acquired in other ways are not so immune. Suppose Motherwell looks at a painting and thinks, “My god, I really hate my mother”.
Here there is room for both kinds of error. Maybe it’s not hate that is really what the painter feels for his mother. But also, maybe Motherwell didn’t paint it. He just thinks he did. What the painting really evinces is that Jackson Pollock hates his mother.
Two ways in which Wittgenstein is wrong, according to Evans.
First, it is not beliefs as individuated by their contents that are sensibly said to be immune or not immune. This is a category mistake. Whether a belief (now individuated differently) counts as immune or not immune depends upon how it was acquired.
Second, Wittgenstein is wrong to restrict immunity to error through misidentification only to beliefs about one’s mental states. Beliefs about one’s bodily states also can exhibit it.
Take his own example, the belief that my hair is blowing. I may form this belief by looking in the mirror and seeing someone’s hair blowing. In this case, we can have a misidentification.
But I may also form that belief simply by feeling that my hair is blowing in the wind. And in that case, there is no sense to raising the question, someone’s hair is blowing, but is it mine?
Or consider the belief that my knee is bent. Again, I may form this belief by looking at a bent knee, and the bent knee might not be mine.
But I might just attend to my bodily position (proprioception, direct awareness of the position of one’s body.) Here, I can be wrong. Maybe my leg was amputated—it’s a ‘phantom’ feeling. But I can’t be wrong by misidentifying whose knee is in question.
This shows that the restriction of the term, “introspection” to beliefs about mental states is somewhat arbitrary. For we acquire beliefs about many bodily states in a quite analogous way:
“My knee is bent.” “I’m bleeding.” Etc.
Beliefs to these effects can be acquired through directly attending to one’s bodily condition, and when they are they are immune to error through misidentification.
They are also authoritative. But perhaps not as authoritative. There’s more room for imagining errors.
It is a tough question whether such beliefs are inferential in any sense.
We might call the ways of forming beliefs at issue ‘bodily introspection’.
2. Criteria of identity
Strawson uses Kantian terminology, which I will stay away from. He makes much of the ‘transcendental unity of apperception.’ See his  and . But I think this concept actually plays at best a subsidiary role in his criticism and diagnosis of Descartes. We can regard - and  as giving the main argument. Strawson interpolates the Kantian material to try to make plausible that the objection to Descartes that Strawson articulates is Kant’s own objection. We are going to ignore that interpretive question.
Begin with Strawson’s talk of “criteria of identity.” A criterion of identity for a kind of object f is a condition that must be satisfied by x1 and x2 if x1 is to be the same f as x2. It is a necessary condition for x1’s being the same f as x2 (a condition, perhaps, which can join with a finite number of other criteria to yield a sufficient condition). Strawson means more than this when he speaks of criteria of identity, but we can go on this understanding for our purposes.
An attempt to formulate criteria of identity for a given kind of thing is an attempt to fill in an instance of the schema:
x1 is the same f as x2 only if (maybe iff) __________
In the case of physical objects and events, the main question concerns identity over time:
x1 at t1 is the same f as x2 at t2 iff: __________
Let’s substitute “baseball game” for “f”. Suppose you’re watching the first game of a doubleheader. You leave for a while and come back. You’re not immediately sure if this is the same game as the one you were watching, or the second game of the doubleheader. But whether you know it or not, there is some fact of the matter. And we know the kind of facts that would help decide the question.
[Put criteria on board.]
As we see, it’s difficult to give complete and non-circular accounts of criteria of identity. We listed some non-circular criteria. But a complete account of criteria may have to be circular, employing the very concept of an f that is in question.
In fact, the problem is easiest to solve in the case of rule-governed, artificial things like games. In the case of kinds of physical object like dogs or tables, it may be just impossible to give non-circular accounts of criteria of identity.
That is, it’s very difficult, maybe impossible, to say what makes it the case that the cat in my house now is the same as the cat that was in my house last Tuesday without making use of the concept of a cat. It’s controversial whether this is true and whether it matters, whether accounts of criteria of identity need to be non-circular.
Nonetheless, one can at least give non-circular necessary conditions of identity.
What would have to have happened between now and last Tuesday for this to be the very chair that was in this position then?
One thing often thought necessary for identity over time of physical objects such as this chair is that there be a continuous spatial path over time from the chair then to the chair now. This is called spatiotemporal continuity.
One way this condition could have been realized is if the chair hasn’t moved since last Tuesday. Another way if someone took it out of the room and then brought it back in and put it in its original place. But if the chair that was here on Tuesday was taken out and put in the dumpster, and then the chair here now was brought in from a factory in Michigan, they must be different chairs.
The criterion of spatiotemporal continuity, unlike others we might formulate, packs an epistemic punch. It is extremely useful for deciding questions of identity from a position of incomplete information. I may ask myself whether it’s plausible that this cat in front of me could be spatiotemporally continuous with the cat that was in my house this morning. If I’ve just flown to Italy, leaving my cat at home, and am standing in the middle of Florence, it’s not plausible.
(Is the following a criterion of identity for chairs? If the chair that is here now is to be the same chair as the chair that was here last Tuesday, it now be composed of exactly the same molecules as it was then.
No, this is not a criterion.
Does the failure of this condition entail a violation of Leibniz’s law? Recall Leibniz’s law: if x1 and x2 are the same object, then x1 and x2 have all the same properties.
One might argue: x1 on Thursday was composed of different matter than x2 is now. Therefore, they have different properties. But they’re the very same chair. So Leibniz’s law isn’t true.
This argument is fallacious. If x1 and x2 are very same chair, and if x1 was composed of matter M on Tuesday, then x2 was composed of matter M on Tuesday. If the chair that was here on Tuesday was composed of matter M on Tuesday, and that is the very same chair that is here now, then the chair that is here now also was composed of matter M on Tuesday.
Both the chair here now and the chair that was here then have the very same property: they were composed of matter M on Thursday.
When Joe was nine, he was four feet tall. He’s not four feet tall now. That doesn’t mean he’s not the very same person as that nine-year-old who was four feet tall. Joe then and Joe now have the property of being four feet tall at age nine.
People aren’t certain heights, period. They are certain heights are certain times.
Chairs aren’t composed of particular chunks of matter, period. They’re composed of different chunks at different times.
So the properties in question are of being composed of particular matter-chunks at particular times.)
One final point about criteria of identity. Strawson’s argument involves a fairly uncontroversial principle, often encapsulated thusly: no entity without identity. That is, there is no such thing as an f unless there are answers to questions about whether this is the same f as that.
A corollary (or at least closely related claim): If you purport to talk of some kind of thing f but can give no criteria at all for deciding the question of whether x1 and x2 are the same f, then it can fairly be said that you have no idea what you’re talking about. You cannot possess the concept of an f unless you grasp criteria for the identity of f’s.
If I ask if this cat is my cat, and you say that not only do you not know, but you also have no idea what could even help decide such a question, then you don’t know what a cat is. You lack the concept of a cat. (You’re also pretty screwed vis-à-vis the material world in general.)
This Strawson applies to the particular case of subjects, that is, those entities that believe, feel, perceive, etc. It’s in his ; we’ll call it premise 1.
We can think of ‘empirically applicable’ in Strawson’s text simply as meaning: criteria we can imagine applying in particular cases. Criteria we understand and can bring to bear.
3. Introspection, criteria of identity and the concept of a human being
Now, Strawson’s :
Well, a ground for it is the fact that beliefs formed through introspection are immune to error through misidentification.
As we saw, it never makes sense to ask, of such a belief, “But is it me?” “Someone has a headache, but is it me?” “Someone hates his mother, but is it me who hates his mother?” And so forth. If such doubts were well-placed, we would have to answer them by appealing to criteria of identity. But such doubts are not well-placed, so we don’t.
Contrast the case where you go home, see someone watching TV, and form the belief that your roommate is watching TV. Here it’s possible to ask the question, “Someone is watching television, but is it my roommate?” If you know what your roommate looks like, you likely won’t raise this question. But your belief is still based on the assumption that the person you’re seeing is your roommate, and even if you don’t raise a doubt, we can imagine you questioning that assumption. And then criteria of identity would be needed to decide the matter. And we can imagine you being wrong about that assumption. Maybe it’s the long-lost twin of your roommate.
In introspection, no such mistake is possible.
As Strawson puts it (without recourse to the technical concept of immunity to error through misidentification), “When a man…” (p.165)
To put the point in its most general form, in introspection, there’s nothing I need to do, no application of criteria, no employment of knowledge, no exercise of any capacity at all, to be sure that it is myself that is the subject in question. In introspection, my self is automatically present for me, impossible to escape, impossible to misidentify.
To put it another way (as McDowell points out, this is actually a stronger thought), you don’t have to keep track of yourself. Imagine following a bird across the sky. You have to keep track of it, to follow it with your eye. If you look away for a bit and then look back, the question arises whether this is the same bird as before.
Not with yourself. It’s always there for you; you can’t miss it. There’s no inner eye that you have to keep fixed on yourself, such that if you close it, you’ll potentially lose track of yourself.
This what Strawson means by 2.
As Strawson sees it, 1 and 2 together entail:
Why? Well, if you have no concept of a self unless you have criteria of identity for the persistence and identity of a self, and if no such criteria are involved in introspection, than you must get them from somewhere else. You must, as it were, construct your concept of the self out of materials that are not found in introspection.
Where do you get these criteria? Strawson thinks there’s an obvious answer to this question: from your knowledge of human beings, those inimitable physical objects.
A human being is a physical object. It’s a mammal, an animal. Like other physical objects, its persistence and identity are a matter of conditions like spatiotemporal continuity. Strawson’s claim is that it’s our ordinary understanding of criteria of identity for such objects, which is evinced in our everyday commerce with other people and ourselves, that provides the needed additional material for the concept of a self.
And he wants to say that even in introspection, this is the kind of object that is in question. When you say, “I am in pain,” you are ascribing pain to a certain physical object, namely, the human being that you are. It’s the same object that other people talk of when they look at you and say, “She’s in pain.”
The fact that the beliefs you form through introspection, once you express them, can be taken up by other people from their third-person perspectives and used to predict your behavior or just to come to a better understanding of you, shows that these introspective beliefs are still about the human being that you are.
So even if in introspection you don’t actively employ criteria of identity for human beings, still, as Strawson puts it, “the links” to those criteria are not “severed”, for your introspective beliefs are integrated into a larger picture. These beliefs are one piece of a much larger puzzle.
Strawson’s mater thought, to put it sweepingly, is that we must understand introspection as just one perspective on the life of a human being, a life that can be observed and thought about in other ways.
To recap, Strawson’s argument, as I interpret is:
4. Descartes’ error
Strawson’s diagnosis of Descartes’ error ascribes to Descartes the following argument:
Why conclusion c)? Well, we make no use of criteria of identity in introspection. If the concept of the self is fully revealed in introspection, then it can’t involve any criteria of identity. Thus we arrive at the idea that the self, the subject of experiences, is somehow uniquely pure, simple. It, alone among all objects, requires no criteria of identity. Correlatively, I am aware of myself as a persisting thing through my experiences, but I do not need any grasp of any criteria whatsoever to be so aware of it. This is the ‘illusion’ Strawson recounts in .
Our idea of the self is the idea of a bare persistence, where that is to say an object whose persistence doesn’t consist in anything. This is the view of the self at issue for Nagel at pp.32-32 of The View from Nowhere.
Why the inference from c) to d)? Well such a pure, simple persistence can’t be material. We know that the persistence of material objects does involve criteria of identity, notably spatiotemporal continuity.
Indeed, a bare persistence can’t have any properties at all except thought, thought being the only thing (more or less) available to you in introspection and so attachable to the subject without the deployment of criteria.
The mistake, as Strawson sees it, is that c) has to be wrong, for Strawson. Why? Because of premise 1. There’s just no such thing as a kind of entity that has no criteria of identity associated with it. There’s no such thing as a bare persistence, of an object so “simple” there are no specifiable conditions for its persistence over time. The reason we had to gesture at this idea with vague words like “bare” and “simple” is because the idea doesn’t make sense.
Quote Nagel, pp. 34-5. We can “extract nothing”, as Nagel puts it, because there’s no criteria of identity there. And no entity without identity.
5. Wittgenstein’s error
By Strawson’s lights, Wittgenstein is more insightful than Descartes, although still wrong, when he denies that the “I” in introspective beliefs refers at all.
A Strawsonian diagnosis of Wittgenstein might go this way:
Wittgenstein moves from pointing to the immunity to error through misidentification of introspective beliefs to the conclusion that those beliefs aren’t really about selves at all. Unlike Descartes, he realizes there can’t be such a thing as a concept of a self without criteria of identity. The difference between Descartes and Wittgenstein is that Wittgenstein accepts both 1 and 2, whereas Descartes accepts only 2. Given that Wittgenstein also accepts a close cousin of Descartes’ b), he is led to his conclusion.
Strawson by contrast, thinks that the solution is not to deny the self that is the subject of experience, but to realize that the self that is the subject of experiences is the human being, the physical object, that thing we move to and comfort when it says it’s in pain.
We might understand (what I called) Strawson’s master thought as follows: the essential thing is to hold onto an integrated conception of the self, acknowledging both its subjective and objective aspects. In one way or another, this is something both Descartes and Wittgenstein fail to do.
As we will see next time, Evans offers some material that might help us to achieve the integrated conception.