The First-Person Perspective and Other Essays (Cambridge Studies in Philosophy)

by Sydney Shoemaker

The essays in this collection deal with the way in which we know our own minds. Professor Shoemaker opposes the "inner sense" conception of introspective self-knowledge. He defends the view that perceptual and sensory states have nonrepresentational features--"qualia"--that determine what it is like to have them. Among the other topics covered are the unity of consciousness, and the idea that the "first-person perspective" gives a privileged route to philosophical understanding of the nature of mind.

Genres: Philosophy

Paperback

First published September 13, 1996

About the author

Sydney Shoemaker is an American philosopher. Until his retirement, he was a Susan Linn Sage Professor of Philosophy at Cornell University. He holds a PhD from Cornell and BA from Reed. In 1971, he delivered the John Locke Lectures at Oxford University. He has worked primarily in the philosophy of mind and metaphysics, and has many classic papers in both of these areas (as well as their overlap). In "Functionalism and Qualia" (1975), he argues that functionalism about mental states can account for the qualitative character (or 'raw feel') of mental states. In "Self-Reference and Self-Awareness" (1968), he argues that the phenomenon of absolute 'Immunity to Error Through Misidentification' is what distinguishes self-attributions of mental states (such as "I see a canary") from self-attributions of physical states (such as "I weigh 200 pounds"). In metaphysics, he has defended the view that laws are metaphysically necessary, a position that follows from his view of properties as clusters of conditional causal powers. He has also applied his view of properties to the problem of mental causation. He also has distinguished contributions to the literature on self-knowledge and personal identity, where he defended a Lockean psychological continuity theory in his influential paper "Persons and their Pasts". In his recent work on the content of perception, he has argued for a distinctive version of internalist representationalism.

Reviews

[Alina · Rating 5 out of 5]

I read through the part of this collection of essays on first-person authority and self-knowledge (essays 1-4). I still need to digest them and reread parts of them. Nevertheless, here I’ll attempt to summarize the primary claims of these essays, then I’ll make some remarks on Shoemaker’s style, which I appreciate.

In essay 1 “Introspection and the self,” Shoemaker assesses Hume’s claim that there is no self which one could observe and ultimately concludes that while the self can’t be encountered through observation or perception, we can still have knowledge of the self or get acquainted with it through some other means. Shoemaker assesses a contemporaneous position which might stand in for the sort of picture Hume denies: one could think that reference to oneself via these quasi-perceptual means is a variety of demonstrative reference via perception at large (e.g., as in when one sees a cat, pointing to it and saying ‘that cat over there…’).

He argues that the two forms of reference are significantly different. Self-reference is immune to error through misidentification across time, whereas perceptual, demonstrative reference is immune in this way only for the present moment (when I remember whether there was a cat on the porch the other day, determining this is more like a matter of observation and inference, and I could get things wrong). Shoemaker uses this point of disanalogy to support his conclusion that the perceptual model of self-knowledge fails; this model assumes that the reference of “I” is just a case of demonstrative reference, since you’d be picking up on what “I” refers to through perceptual means as you do for the rest of the world when known through demonstrative reference.

Essay 2 “On knowing one’s own mind” was my favorite. Here, Shoemaker assesses two positions which one might call “Cartesian.” First, the subject has a unique, special way of accessing their own mind, which is different from both the way by which they access facts of the world and the way by which other people might learn about their own mind. Second, the judgments a subject is prone to make of their own mind when formed through these special first-person means are infallible. Third, this special first-person means of access is essential to the mind itself; it is impossible to have the mind with which we’re familiar without having this first-person sort of self-knowledge. Shoemaker will reject the second claim, but affirm the first and the third claims.

To argue for this, Shoemaker asks us to see what follows from the supposition that a person is “self-blind”; they can only come to know of their own mental states through the means available to any bystander, namely, through observing their behavior and drawing inferences about what mental states might be there on the basis of that. The position which Shoemaker opposes, that distinctively first-personal self-knowledge is not essential to or constitutive of the mind itself, would hold if it can be shown that a self-blind person can still have all the essential trappings of a normal mind (one with rationality, language use, and grasp of certain standard concepts, e.g., the concept of belief and desire). If being self-blind, on the contrary, implied that the person would no longer be rational or have another one of these essential markers of a human mind, then this supports that self-blindness is an impossibility for the mind, and the mind is rather essentially self-knowing in the distinctively first-person manner (which I’ll abbreviated as “self-knowledge” from here out).

At first, it may seem that a person can engage in all of the behaviors which exemplify rationality while being self-blind. When asked “Do you believe that P?” the self-blind person could simply look to the world, ascertain whether P is the case, and say yes or no accordingly. More generally, a self-blind person could find out certain rules by which they use starting materials of acts learnt through observational means to manufacture a behavioral response which is normally yielded through a normal person’s first-personal means of knowing themself. If one can exhibit all the behaviors typical of rationality, one counts as rational, for the purposes of this argument. So, it might seem that self-knowledge is inessential to having a mind.

Shoemaker challenges these initial appearances. He points out that self-knowledge is necessary for being able to use language in an assertive manner at all, which is presupposed in practices of reasoning and sharing information. In asserting a sentence, you need to follow a rule along the lines of “if I believe that P, then say that P” — and this presupposes knowledge of what you believe, or self-knowledge. Sometimes we think a system can follow rules without knowing what it is doing (e.g., a thermometer following rules which output display of the temperature), but taking this lens upon the current issue won’t do; a creature who follows any rule blindly like this would not be rational or recognizably have a mind anymore.

Shoemaker offers an independent argument against the possibility of self-blindness. Consider the domain of motivation, as opposed to belief. With belief, one can get away with much self-aware behavior by just looking at the world and seeing what’s the case. The same can’t hold for motivations and desires. When you’re asked “What do you desire?” you can’t just look at the world and see what’s desirable, as you can look at the world and see what’s the case when you’re asked “What do you believe?” To say that you can simply see that certain objects or states of affairs are desirable won’t do. It’s not clear what it means to see an object as desirable; does this mean it ought to be desired or it is possible for it to be desired, for example? It seems that knowledge of the desirability of objects in the world requires self-knowledge or some sort of knowledge independent of or in addition to knowledge of what’s going on in one’s environment.

Moreover, Shoemaker argues, self-knowledge is necessary for predicting one’s own behavior into the future and so undertaking projects which take time. To know of your intentions only through third-personal means isn’t enough. I’m forgetting the details of this argument, but it proceeds from the insight that self-knowledge is necessary for getting what David Kaplan called “cognitive dynamics” off the ground; this refers to the capacity of the mind to automatically transform a second-order belief about one’s doing something at the present moment to being past-tense, to the effect that one has done that thing at a certain past moment. If you’re self-blind, you could only update your second-order beliefs like this through voluntary effort and observation; it wouldn’t happen automatically.

As a whole, Shoemaker’s argumentative strategy involves showing that once it’s supposed that a person is self-blind, the only way of preserving that they exhibit essential behaviors and practices for having a rational mind is to assume that they in fact are not self-blind, so the supposition is contradictory. On top of that, Shoemaker suggests that we should think of second-order beliefs about certain first-order beliefs and desires as supervening upon those first-order mental phenomenon; the second-order or self-knowledge is not distinct or independent from the first-order (I’m more skeptical about whether he in fact establishes this positive account of through what means self-knowledge is achieved, but it seems right to me that some distinctively first-person self-knowledge is essential for having a mind).

In essay 3 “First-person access” Shoemaker argues against the Cartesian transparency thesis about the mind (that having any mental state entails that one has an infallibly correct judgment about the state one has), and instead argues in favor of his “self-intimation” thesis, that having a mental state normally comes with an implicit second-order belief about one’s having that state, which is normally true. Shoemaker proceeds by asking whether the claim that our beliefs are generally true as justified by appeal to externalism of mental content and a causal theory of mental content (e.g., Putnam’s view that mental contents are individuated by the environmental states which the subject is causally related to) can be understood as the more general case, of which the claim that our beliefs about our mental states should be generally true can be seen as a subclass.

Shoemaker argues self-knowledge isn’t a variety of knowledge of the environment. This has to do with the fact that to get externalism about mental content off the ground, we need to proceed by distinguishing between “wide” and “narrow” mental contents and by imagining that some aspect of mental content can be the same while the environmental conditions vary (e.g., water and twin-earth water). In the case of the mind, we’re not allowed to vary the relevant “environmental conditions” in the same way; under this case, those conditions would be some first-order mental phenomenon, but it is essential for that phenomenon to be of the same overall kind (as water and twin-earth water are both of some shared kind of ‘apparent water’) that it preserve a certain functional role in the mind.

For example, if the first-order phenomenon is the belief that it’s raining, that proposition would need to guide behavior and serve as a premise in reasoning, in order to still count as belief. I didn’t follow all the key steps in Shoemaker’s argument, but off this basis, he shows that the second-order belief about the first-order one is not going to be independent of the first-order, as a belief about water will be independent of water; rather, the first-order belief itself, defined partly by its functional role in the mind, essentially depends upon the second-order belief about one’s having that belief. Self-knowledge and perceptual belief understood under a causal or externalist theory of mental content are alike only in a minimal way; they both don’t depend upon a subject’s drawing inferences, and their contents need to be determined causally.

In essay 4 “Moore’s paradox and self-knowledge” Shoemaker argues that an account of the contradictory nature of Moorean sentences (e.g., “P but I don’t believe that P”) should not be made on the basis of conditions of assertion, or the linguistic expression of belief. Rather, their contradictory nature can be found in the nature of belief itself. Notice that logically, we couldn’t believe that we have a belief in a Moorean sentence, if we could feasibly have such belief. It’s a contradiction to say “It’s raining and I don’t believe that it’s raining, and that this is so is something I believe” - since ‘I believe,’ as in the final conjunct, will distribute over the former two conjuncts, getting you both belief that it’s raining and disbelief that it’s raining. Shoemaker takes this to motivate that it is impossible to be aware of a belief in a Moorean sentence, if it were possible to have this first-order belief. Is it possible to have such first-order belief? Shoemaker thinks it is, as long as the mind can be ‘compartmentalized.’ He suggests that it is possible that different parts of the mind each could believe in a conjunct of a Moorean sentence; so a Moorean sentence may accurately describe the total belief of a person at a time, even if that person could not possibly assent to the sentence when it’s presented to them, due to the fact that it’s impossible to be aware of having such a first-order belief.

In essay 11 “Self-knowledge and ‘inner sense,’ Lecture II: The broad perceptual model,” Shoemaker assesses a perceptual model of self-knowledge. He defines it as consisting in two conditions: (1) the transparency of the mind, such that judgments about one’s own mental states will be infallible, and (2) the existence of mental entities and facts is independent of our knowing about them. Against this, Shoemaker argues that conceptually, the first- and second-order mental phenomena constitutively related, such that we get something close to (1), but which isn’t full-out transparency: rather, certain mental states are “self-intimating,” which is a similar claim to transparency but weaker.

To argue for this, he starts off with a traditional case of a mental state, pain. Shoemaker asks what would follow if pain were not self-intimating, that is, if we could normally have pains without being aware of the fact that we do. If this were the case, we wouldn’t be able to be aware of any of the behaviors and activities which pain rationalizes, such as finding the cause of pain unpleasant and avoiding it. Ultimately, Shoemaker shows, the person with non-self-intimating pain would appear as if there were two independent selves housed in the same body, one which remains agnostic of whether they are in pain and is ignorant of what’s going on for the other self, and the other self which is in pain and goes through the behaviors and activities which pain causes. This is not acceptable, so we should reject the premise and rather think that pain is normally self-intimating.

As a whole, I appreciate Shoemaker’s philosophical style. I like how in most of his papers, he first motivates a problem and explicates a major position on it. Then he assesses such positions by revealing their conceptual consequences, or what else must be true of the mind if the position were the case. Most of his argument is negative like this, at least in the handful of papers I read, and often he is able to have a negative argument perform a simultaneous function of also defending a positive position; Shoemaker subtly sneaks in suggestions about what must positively be the case, given all of the problematic predictions/consequences which fall from a certain position which he rejects.

I’m convinced by his overall position that distinctively first-person self-knowledge is essential to having a mind and the nature of first-order mental phenomena of various mental types. I’m less convinced by that these types include all sorts of beliefs, desires, and intentions. It’s interesting when Shoemaker acknowledges the complications of desire and of cases of self-deception; he always puts them away, after examining them, by saying that these are exceptions to the rule, rather than the rule. I am doubtful that is the case. I’ve been thinking about how it seems to be the rule, rather than the exception, that for any piece of apparent self-knowledge of anything which goes beyond bare physical facts we believe in, it’s possible to imagine a counterfactual case where we tell our piece of purported self-knowledge to an audience, and that audience is of a different enough of a cultural background than our own that they can reasonably think we’re self-deceived. For example, think of how certain theocracies think that Americans are extremely selfish, when these Americans would think of themselves as liberated and exercising their autonomy. Or, think of how hedonists might think that anybody who claims to be generous or selfless is in fact motivated by the pleasure that their act gives them. Maybe one reason for this high applicability of the attribute of being self-deceived is that at the “first-order” level of the mind, there are many different beliefs and desires, or maybe there are complicated, ambiguous, or even indeterminate mental states which are open for being interpreted in diverse ways. This is just some musing for the time being, however.

If it turned out “self-deception” in this broad, revisionary sense were the rule rather than the exception, then this would point in favor for that there shouldn’t be a uniform account of self-knowledge given to all mental types at the first-order level, but different types of mental phenomenon can be known or fail to be known in different ways. Maybe, even, thinking about all these cases in terms of “knowledge” might be misleading, if for example there were a range of how indeterminate the first-order level can be, and/or a range of how susceptible the first-order level is to being changed by how a subject appears to know it.