Author Information: Sarah Sawyer, University of Sussex, s.a.sawyer@sussex.ac.uk

Sawyer, Sarah. “Contrastivism and Anti-Individualism: A Response to Aikin and Dabay.” Social Epistemology Review and Reply Collective 3, no. 9 (2014): 1-6.

The PDF of the article gives specific page numbers. Shortlink: http://wp.me/p1Bfg0-1zz

Please refer to:

• Sawyer, Sarah. “Contrastive Self-Knowledge.” Social Epistemology 28, no. 2 (2014): 139-152.
• Aikin, Scott F and Thomas Dabay. “Contrastivism and Individualism: A Reply to Sawyer.” Social Epistemology Review and Reply Collective 3, no. 6 (2014): 87-90.

Introduction

In “Contrastive Self-Knowledge” (2014), I develop an account of self-knowledge as contrastive. According to that account, apparently non-contrastive statements of the form:

S knows that she ψs that p

are elliptical for explicitly contrastive statements of the form:

S knows that [she ψs that p] rather than that S* φs that q.

In this formulation, S* marks out a contrast in the subject of the first-order attitude, φ marks out a contrast in the attitude taken towards the proposition, and q marks out a contrast in the proposition to which that attitude is taken. Self-knowledge, then, is contrastive along three dimensions. The square brackets are simply a device for indicating the element of the statement that represents the content of the self-knowledge ascribed. This serves to highlight the fact that the contrastivity of self-knowledge lies not in the content of self-knowledge but in the subject’s state of self-knowledge. When Jack knows that she believes that the negative representation of women in the media perpetuates sexism, the content of her self-knowledge is simply: I believe that the negative representation of women in the media perpetuates sexism. But Jack’s state of self-knowledge is determined not by her standing in relation to that proposition in isolation, but by her standing in relation to that proposition relative to a set of contrasting propositions. The contrast class that serves to individuate Jack’s state of self-knowledge need not itself be represented by her in thought—only the content of her thought will be represented by her. Nonetheless, the contrast class will be determined by her conceptual capacities. Since subjects can differ in their conceptual capacities, different subjects can each know that they believe that the negative representation of women in the media perpetuates sexism while nonetheless being in different states of self-knowledge. 

The claim that self-knowledge is contrastive is the main claim of my paper. However, the paper contains two further, related claims: first, that the contrastivity of self-knowledge derives ultimately from the contrastivity of the propositional attitudes—specifically from the contrastivity of belief; and second, that the contrastivity of self-knowledge entails anti-individualism about concepts. It is the second of these further claims that is the focus of Aiken and Dabay’s recent paper “Contrastivism and Individualism: A Reply to Sawyer” (2014). Aiken and Dabay frame their paper as a request for clarification concerning the argument from the contrastivity of self-knowledge to anti-individualism. They are right to point out that in my paper the argument from the contrastivity of self-knowledge to anti-individualism is not clear—indeed, my remarks in the paper are rather cursory. I aim to rectify matters in this response by providing an explicit statement of the argument I had in mind.

Contrastive Self-Knowledge

I begin by looking at contrastive knowledge. According to the contrastivist account of knowledge, knowledge is a three-place, contrastive relation holding between a subject, a proposition and a contrast class: S knows that p rather than that q. In order for knowledge to be genuinely contrastive, the contrast class denoted by ‘q’ in this schema must be understood as a set of propositions containing some but not all propositions that contrast with p. That is, it is crucial to the contrastivist account that (1) not be equivalent to (2):

(1) S knows that p rather than that q;
(2) S knows that p rather than that ~p.

If (1) and (2) were equivalent, then contrastive knowledge would reduce to non-contrastive knowledge, which would undermine the claim that knowledge is genuinely contrastive. It is, in fact, the non-equivalence of (1) and (2) that furnishes the contrastivist account with its solution to the closure paradox, since it paves the way for a subject to know that she has hands relative to one set of contrasts, but not know that she has hands relative to another set of contrasts, to know that there is a zebra in the enclosure relative to one set of contrasts, but not know that there is a zebra in the enclosure relative to another set of contrasts, and so on. S’s contrastive knowledge of a proposition p, then, requires what I will call a positive contrast class—the set of propositions in contrast to which S knows that p—and a negative contrast class—the set of propositions in contrast to which S does not know that p. The positive contrast class will map onto S’s ability at the time, given her evidence, to distinguish a range of objects, facts or events referred to by the elements of p from a range of possible objects, facts or events referred to by the corresponding elements of the propositions in the positive contrast class; the negative contrast class will map onto S’s inability at the time, given her evidence, to distinguish a range of objects, facts or events referred to by the elements of p from a range of possible objects, facts or events referred to by the corresponding elements of the propositions in the negative contrast class.

Since contrastive self-knowledge is a form of contrastive knowledge, analogous considerations naturally apply—modulo some differences that arise from the fact that self-knowledge is knowledge of one’s propositional attitudes rather than knowledge of objects, facts or events in the surrounding world.[1] In order for self-knowledge to be genuinely contrastive, then, (3) cannot be equivalent to (4): [2]

(3) S knows that she ψs that p rather than that S* φs that q;
(4) S knows that she ψs that p rather than that it’s not the case that she ψs that p.

If (3) and (4) were equivalent, then contrastive self-knowledge would reduce to non-contrastive self-knowledge, which would undermine the claim that self-knowledge is genuinely contrastive. In a manner that parallels the contrastivist solution to the closure paradox mentioned above, I have recently argued that the non-equivalence of (3) and (4) allows the contrastivist account of self-knowledge to solve the McKinsey Paradox [3]—a paradox which purports to undermine the compatibility of anti-individualism and self-knowledge. I will not rehearse the argument here. What interests us at present is the fact that S’s contrastive knowledge of the proposition that she ψs that p requires there to be a positive contrast class—the set of propositions in contrast to which S knows that she ψs that p—and a negative contrast class—the set of propositions in contrast to which S does not know that she ψs that p. But although the contrast classes will once again be marked out by S’s discriminatory capacities, in the case of self-knowledge it is not S’s ability and inability to distinguish amongst objects, facts or events that is relevant, but her ability and inability to distinguish amongst propositional attitudes, and hence concepts.

Suppose a subject S knows that she believes that p. As just stated, in order for S’s self-knowledge to be contrastive, there must be a positive contrast class containing propositions relative to which S knows that she believes that p, and a negative contrast class containing propositions relative to which S does not know that she believes that p. Let C be one of the concepts in the proposition S knows. Let PC1, PC2, PC3, and so on be concepts that are such that, when they replace C in p, the resulting propositions are ones that S knows that she believes that p in contrast to—the concepts that feature in the propositions that comprise the positive contrast class relative to which S knows that she believes that p. And let NC1, NC2, NC3, and so on be concepts that are such that, when they replace C in p, the resulting propositions are ones that S does not know that she believes that p in contrast to—the concepts that feature in the propositions that comprise the negative contrast class relative to which S knows that she believes that p. It follows trivially that S can distinguish (at the time, given her evidence) C from PC1, PC2, and PC3, but cannot distinguish (at the time, given her evidence) C from NC1, NC2, or NC3.

On Anti-Individualism

My argument from the contrastivity of self-knowledge to anti-individualism, stated briefly, is that on the assumption of individualism, there is no set of concepts that can fulfil the role of NC1, NC2, NC3, and so on. This means that individualism eliminates the possibility of a negative contrast class, which renders (3) and (4) equivalent, and hence renders self-knowledge non-contrastive. Consequently, contrastive self-knowledge entails anti-individualism. Let us look at this in more detail, starting with a closer look at individualism.

According to individualism, a subject’s concepts are determined by her individualistically specified discriminatory capacities and behavioural dispositions—broadly speaking, a subject’s conceptual capacities cannot outstrip the way things seem from her point of view.[4] According to anti-individualism, in contrast, a subject’s concepts are determined in part by facts that go beyond her individualistically specified discriminatory capacities and behavioural dispositions. To illustrate the difference between the positions, consider a range of different metals, M1, M2 and M3, and a subject S who cannot distinguish between them. Since, according to the individualist, S’s concepts are determined by her discriminatory capacities and behavioural dispositions, according to the individualist, S would possess a general concept that ranges over instances of M1, M2 and M3. According to the anti-individualist, in contrast, S might well possess a concept that has instances of M1 in its extension, but not instances of M2 or M3, despite the fact that she cannot distinguish between them. If we call to mind the standard descriptions of the Twin–Earth thought experiment, we can see that this would be so, according to the anti-individualist, if, for instance, S had had causal contact with instances of M1, but never with instances of M2 or M3.

Now let us consider the conjunction of individualism and contrastive self-knowledge. According to the individualist, as we have just seen, S’s concepts are determined by her discriminatory capacities and behavioural dispositions. According to the contrastivist, as we saw earlier, S’s knowing that she believes that p requires that she possess some concept C that occurs in the proposition she knows, that there be a set of positively contrasting concepts PC1, PC2, PC3, and so on, that she can distinguish from C, and that there be a set of negatively contrasting concepts NC1, NC2, NC3, and so on, that she cannot distinguish from C. This is where the incompatibility arises. On an individualist’s understanding of concepts, there could be no concepts which are both (i) distinct from S’s concept C and (ii) such that S cannot distinguish them from C. This is because an individualist individuates a subject’s concepts in terms of distinctions available from within her subjective perspective. But there must be such concepts if self-knowledge is contrastive, since this is how negatively contrasting concepts are defined.

The problem, at root, is that the claim that there is a concept C that S cannot distinguish from NC1, NC2, or NC3, in fact makes sense only if we identify the concept C independently of S’s discriminatory capacities and behavioural dispositions. But if we identify it in this way, it is not identified as a concept possessed by S at all. Indeed, since S by hypothesis cannot distinguish C from NC1, NC2, or NC3, the individualist account implies that S possesses instead a broader, more general concept that ranges over instances of C, NC1, NC2, NC3, and so on. But to say that S will have a broader, more general concept that ranges over instances of C, NC1, NC2, NC3, and so on is to say—given that the argument is general in form—that there can be, on the individualist account, no negatively contrasting concepts. And, of course, if there can be no negatively contrasting concepts, then there can be no negative contrast class for S’s self-knowledge. But if there is no negative contrast class for S’s self-knowledge, then S’s self-knowledge is not contrastive. It follows directly that if self-knowledge is contrastive, concepts are anti-individualistically individuated.

We tend to talk about concepts pre-reflectively as if they are shared—as if a single concept can be possessed by different people with different discriminatory capacities and behavioural dispositions. Thus we talk, for instance, of the concept water in abstraction from any particular individual thinker. There is in the main nothing wrong with this pre-reflective way of talking, and both the individualist and the anti-individualist can accommodate it. But this kind of talk only makes literal sense if concepts are anti-individualistically individuated. On an individualist understanding, what we think of as a single term in the language—‘water’, say—is used to express a variety of related but distinct concepts had by different individuals with different but overlapping discriminatory capacities and behavioural dispositions. As such, when it comes to assessing the implications of a position that depends on the individuation of concepts—such as, in this case, the implications of contrastive self-knowledge—we need to be careful not to let our tendency to talk as if concepts are shared lead us astray.

Contrastive self-knowledge depends upon the individuation of concepts because it depends on them being individuated in such a way as to allow for positively and negatively contrasting concepts. Pre-reflectively this makes sense. However, on reflection, it turns out to be inconsistent with individualism. ‘C’ is a term that we are using first to identify a concept that S has, and second to identify one of a set of concepts between which S cannot discriminate. But on an individualist understanding of concepts, there is no concept that can fulfil both roles. It is when we realise that the concept C is subject-relative on the individualist account that we realise that on the individualist account the very possibility of negatively contrasting concepts disappears.

Conclusion

Aiken and Dabay’s attempt to spell out the argument I had in mind errs in not distinguishing clearly enough between the positive contrast class and the negative contrast class required for self-knowledge to be genuinely contrastive. Much of what they say is right, but it doesn’t address the core issue, which is the availability of negatively contrasting concepts. On reflection, it strikes me that the argument from the contrastivity of self-knowledge to anti-individualism is more general than I had initially anticipated. First, the argument is not restricted to consideration of propositional attitude concepts, as I had previously suggested; rather, the argument, when laid out in abstract terms, clearly applies no matter what kind of concept ‘C’ stands for. Second, and somewhat more tentatively, the argument may not even be restricted to consideration of the contrastivity of self-knowledge; I suspect that there is an argument from the contrastivity of knowledge per se to anti-individualism. I do not have the space here to develop this line of thought further, but its future development promises to explain why individualism implies that individual subjects are confined to knowledge of their own internal worlds and precluded from knowledge of the world beyond—and this in turn promises to explain why the very possibility of representation in thought and language depends on an anti-individualistic understanding of representation.

References

Aikin, Scott F. and Thomas Dabay. “Contrastivism and Individualism: A Reply to Sawyer.” Social Epistemology Review and Reply Collective 3, no. 6 (2014): 87-90.

Sawyer, Sarah. “There is No Viable Notion of Narrow Content.” In Contemporary Debates in the Philosophy of Mind, edited Brian McLaughlin and Jonathan Cohen, 20-34. Oxford: Blackwell Publishers Ltd, 2007.

Sawyer, Sarah. “Contrastive Self-Knowledge.” Social Epistemology 28, no. 2 (2014): 139-152.

Sawyer, Sarah. Forthcoming. “Contrastive Self-Knowledge and the McKinsey Paradox.” In Externalism and Skepticism, edited by Sanford Goldberg. Cambridge: Cambridge University Press.

[1] Self-knowledge encompasses knowledge of more than one’s propositional attitudes, but I restrict my focus to knowledge of one’s propositional attitudes here for the sake of simplicity.

[2] (3) and (4) are, of course, just substitution instances of (1) and (2) respectively.

[3] See my (forthcoming).

[4] This is a crude characterisation in at least two respects. First, the expression ‘the way things seem from a subject’s point of view’ has an individualist and an anti-individualist reading. I leave this complication to one side here. Second, there are a number of different versions of individualism. I take the differences to be irrelevant for present purposes, as I think ultimately they are all committed to something like this crude characterisation. See my (2007) for further discussion of the differences and commitments of each.

Categories: Critical Replies

Tags: anti-individualism, contrastive self-knowledge, contrastivism, Sarah Sawyer, Scott F. Aikin, Thomas Dabay