We recently read Derek Brown’s “Indirect perceptual realism and demonstratives” (Philosophical Studies, forthcoming). I’m currently writing up a paper on perceptual demonstratives, and have some sympathies to views in the ballpark of Indirect Realism, so I was especially interested in Brown’s paper. Brown’s task in the paper is to defend Indirect Realism from two related arguments he finds in Snowdon (1992, Crane volume) and A.D. Smith (2002, The Problem of Perception). Smith argues that Indirect Realism is incoherent (there is no genuinely Realist view once we abandon the idea that we directly perceive the external world), and Snowdon argues that perceptual demonstrative judgments are fundamental to our belief-acquisition and knowledge of the external world but that Indirect Realism fails to render those judgments true, and thus undermines our knowledge of the external world. Both arguments rely on some claims about the status of demonstrative judgments about external objects on the Indirect Realist view. Brown defends Indirect Realism by arguing that there is nothing problematic about those judgments according to Indirect Realism.
Without going into too much detail on the particulars of Snowdon or Smith (although I will reconstruct Smith’s argument or something like it, below), I might simply note that Brown finds both arguments to rely on an ungrounded assumption that demonstrative judgments such as “That is a violin” would be false under Indirect Realism. Brown thinks that both authors might be confusing the epistemic dependence of such judgments (on the ability to perceive a perceptual intermediary) with a kind of semantic dependence. Brown’s notion of semantic dependence is that such judgments, in the case of indirect perception, are about propositions that include the dependence relation as a constituent. He suggests that both Smith and Snowdon might mistakenly be assuming that perceptual demonstrative judgments include the fact that we are non-dependently perceiving as part of their content. If they did, then such judgments would come out as false on the Indirect Realist view.
Now I certainly agree with Brown that it is implausible to suppose that perceptual dependence has a semantic status in his sense. But I also think it is unlikely that either Snowdon or Smith had such a view in mind. Perhaps Smith comes closest to such a view, since he puts weight on the claim that we take ourselves to be referring with a demonstrative to an item that is directly perceived (p. 16). But perhaps these claims are meant to motivate something like my premise 1 below, rather than a claim about the content of demonstrative judgments.
I find two problems with Brown’s response to these arguments. First, his positive support for the claim that a demonstrative can refer to an item that is only indirectly perceived is based on the use of demonstratives in a public language. His primary example is of a person viewing a violin on television and saying “That is a violin”. I share the intuition that this sentence can be true, and that the demonstrative here plausibly refers to a violin and not to something more directly perceived (such as a region of the television screen). But the opponent of Indirect Realism might concede that the use of demonstratives in a public language have this feature while denying that a certain sort of demonstrative thought is possible toward items that are only indirectly perceived.
This leads to my second worry. I think a more plausible version of the “semantic dependence” view would be precisely one that insists that items that are indirectly perceived can only be referred to in thought descriptively, whereas items that are directly perceived have a special status as items toward which we can directly refer via a demonstrative. Direct perception puts us into cognitive contact with externalia in a way that is not possible otherwise. This strikes me as a view that many Direct Realists appear to hold, and as part of the explanation for why they think defending Direct Realism is so important.
Right or wrong, the above suggestion allows for us to make sense of Smith’s argument in a way that makes it more plausible. Smith appears to argue as follows:
1. To be a Realist about the external world, one must hold that the items we can perceive and thereby demonstratively refer to in thought are “public objects of common reference”. The subject matter for which the question of Realism arises is precisely that which we demonstratively refer to in this way.
2. We can only demonstratively refer in thought to items that are directly perceived.
3. According to Indirect Realism, that which we directly perceive is never a “public object of common reference”.
4. If Indirect Realism is true, then we cannot demonstratively refer in thought to “public objects of common reference”.
5. Therefore, “Indirect Realism” is not a genuine form of Realism.
Obviously premises 1 and 2 deserve scrutiny. But they each have some plausibility. In defense of 2, it does seem that for items that we only perceive indirectly, they are in some sense thought about or known only via a description. They are in a sense theoretical entities, items posited on the basis of what we perceive directly. Smith’s defense of something like premise 1 is that the world for which the question of Realism arises is the world we take ourselves to directly perceive. We ask ourselves, “Is this book a public object?” But if on the Indirect Realist view genuine demonstrative reference (in thought) can only be toward sensory intermediaries, then the answer to such questions must always be “no”.
I find this point about the subject matter of Realism somewhat compelling. On the other hand, it is worth noting that this argument does not show that Indirect Realism is incompatible with the claim that there exists a mind-independent external world of “ordinary objects”. I think many would find that Smith’s notion of Realism is overly restrictive, and that the existence of a mind-independent external world that each of us inhabits and causally interacts with is sufficient for Realism.
First let me thank Dr. Thompson for taking the time to read and comment on my article. His discussion is helpful and informative although (not surprisingly) I do not find his critique compelling.
Background. The underlying issue concerns the relationship between demonstrative reference and the directness of perception. Snowdon and Smith believe that only directly perceived things can be demonstratively referred to; I disagree. My strategy begins by considering an example of indirect perception that is neutral, that does not on its own involve commitments to indirect or direct realism. That is the point of the television case: Chloe is looking at a television that is depicting a red violin. She is thus in a straightforward respect (endorsed by direct realists like Gibson) indirectly perceiving the violin. The trouble is that in such a case the idea that only directly perceived things can be demonstratively referred to is remarkably implausible. Chloe says “That is a violin”. [We can refine the scenario to make the violin, as opposed to some portion of the television screen, the only plausible referent of ‘that’.] Why deny that she is demonstratively referring to the violin depicted by the television, particularly when she seemingly does so with ease? If Snowdon and Smith are right then they need some additional material to explain why. I then search their work for that additional material and find all avenues wanting (i.e., including the semantic dependence avenue Thompson refers to).
Thompson’s reconstruction of Smith’s argument contains as its 2nd premise: “We can only demonstratively refer in thought to items that are directly perceived.” I agree that this is a premise in Smith’s argument, but fail to see why it is plausible. Thompson, by contrast, argues that “In defense of 2, it does seem that for items that we only perceive indirectly, they are in some sense thought about or known only via a description. They are in a sense theoretical entities, items posited on the basis of what we perceive directly.” To see why this is not compelling, apply it to the violin-television case: Chloe is looking at the violin depicted by the television and says “That is a violin,” yet according to Thompson the violin is “in some sense thought about or known only via a description.” We have straightforward evidence against Thompson’s claim, and thus, if that claim is to be endorsed, it must be endorsed on the basis of additional argumentation.
Thompson’s additional claim, that these indirectly perceived items are “in a sense theoretical entities, items posited on the basis of what we perceive directly” may be true on some or even various senses of the term ‘theoretical entity’, but I do not believe this is relevant. Consider other familiar sorts of perceptions that are routinely regarded as indirect (e.g., those involving telescopes, microscopes, etc.). We need never directly perceive some cell to be able to demonstratively refer to it through the aid of a microscope: one can look at a screen depicting the microscope’s output and speak of ‘this’ cell and ‘that’ one, and the fact that these cells are theoretical entities ‘posited on the basis of what we perceive directly’, does nothing to question this. The same holds for distant objects we perceive through telescopes. I thus do not find Thompson’s justification for premise 2 compelling. Consider now Thompson’s two more specific charges.
(1) Regarding my ‘That is a violin’ example Thompson states: “I share the intuition that this sentence can be true, and that the demonstrative here plausibly refers to a violin and not to something more directly perceived (such as a region of the television screen). But the opponent of Indirect Realism might concede that the use of demonstratives in a public language have this feature while denying that a certain sort of demonstrative thought is possible toward items that are only indirectly perceived.”
As I understand the worry, Thompson is concerned about the difference between demonstratives in language and demonstratives in thought. I believe the burden here is on the objector. If we again consider Chloe, she is demonstratively thinking about the violin, not about the surface of the television and its features (more generally, I rarely think about features of the television screen when watching it and routinely think about what it depicts). In this way her demonstrative language mirrors her thoughts, and if the objector wishes us to deny this then she must offer some reason, beyond mere skepticism, for us to do so.
(2) Thompson suggests that the direct realist simply “insists that items that are indirectly perceived can only be referred to in thought descriptively, whereas items that are directly perceived have a special status as items toward which we can directly refer via a demonstrative”. I agree that one can argue that Snowdon and Smith are insisting this, but I again question their justification for holding it. The claim is not in itself plausible when one reflects on various neutral cases involving indirect perception, and thus, without further argument, I do not see why it need worry the indirect realist.
General assessment. Smith, Snowdon and Thompson all have the intuition that indirect realists face a difficulty when faced with demonstratives. I agree. Where we disagree is on what the difficulty is. Whereas they are inclined to believe indirect realists have a hard time saying that we demonstratively refer to external items, I believe indirect realists have a hard time interpreting the referents of demonstratives in certain underspecified or ambiguous scenarios. In my view their belief is largely the result of thinking about indirect perception only within the context of indirect realism, and not in a broader context that includes party-neutral samples of indirect perception. The interpretive difficulties on the other hand can be dealt with, but do require further analysis (see Brown, 2008. Indirect perceptual realism and multiple reference. dialectica, 62, 3, pp. 323–334). Fortunately, the needed concession – multiple reference – is one indirect realists should have already made for a variety of linguistic tools outside of demonstratives.
Left by Derek Brown on December 2nd, 2008