Externalism, Visual Content, and Burge: Part 2

I claimed that whether Burge's strategy will succeed in the case of visual contents depends, remember, on whether experiences that P and the thoughts that might aspire to be knowledge of the contents of those experiences share contents as a matter of course.

I will argue that the Burge-style account cannot be extended to our knowledge of the content of our own experiences because (i) the representational resources utilized by experience and belief are different, (ii) the beliefs that accompany experience are not part of the experiences they accompany, and (iii) the beliefs that accompany experience are not about the experiences they accompany.

There is abundant reason to suspect that belief-forming mechanisms use representational resources that differ from those used by the visual system. Accordingly, if there is to be a justificatory account of our beliefs about our own experiences, it must be one that differs from Burge's.

A Red Herring: It is commonly claimed that experiences have non-conceptual contents. It is sometimes thought that this identifies a particular kind of content -- a kind of content that differs intrinsically from the contents of conceptual states. Perhaps one has in mind the view that propositional attitudes have contents that are composed of concepts -- whatever those are -- while experiences do not. If so, it follows that beliefs and experiences cannot have the same contents, since they take different sorts of things as objects. If so, it follows that Burge's maneuver cannot work. Since the experience that P cannot share contents with beliefs about it, these beliefs won't count as knowledge of the content of the experience since they do not share contents with the experience. Call this line of argument Red Herring. It turns out, however, that Red Herring is mistaken. Or, more carefully, the conception of the contents of experience Red Herring relies on is not justified by the standard arguments in favor of the existence of non-conceptual content.

The claim that experiences have non-conceptual contents should probably be understood as requiring only that a your experience can represent a feature even if you lack the conceptual resources to judge that things possesses that feature. It is possible for your experience to represent what you cannot. There is, in principle, no obstacle to someone's possession of the conceptual wherewithal to match his visual system's representational capacities. However, the fineness of grain and inexpressibility arguments used to establish the existence of non-conceptual content turn on there being actual differences between our conceptual resources and the representational capacity of our visual systems.

While Red Herring is no friend to someone who would seek to criticize the Burge strategies application to the case of visual content, we can now see how, if properly understood, the thesis that visual contents are non-conceptual poses a threat the Burge strategy.

In the best case for Burge's strategy, experiences and thoughts about them will share contents. But, even if experiences and thoughts about them share contents, their sharing those contents is not explained by the use of the same representational capacities. Thoughts deploy conceptual representational resources, while experiences do not. Since the representational capacities in virtue of which we believe and experience things differ, the relevant beliefs cannot be constituted as knowledge by the sharing of representational capacities. If they are constituted as knowledge, the explanation for their status must differ from the one Burge gives for judgments and our beliefs about our own judgments.

At any rate, remember, we are not the sort of creatures who can attribute in thought all of the properties our experience attributes to objects.

Let's set aside the challenge posed to a Burge-style account by non-conceptual contents for a moment. What if the contents of a subject's experience were such that they were thinkable by that subject? That is, what if there were a subject who could think what he experiences? What if, moreover, the relevant experiences and beliefs utilized the same representational resources. Even so there are functional differences between judgments and experiences that get in the way of a Burge-style account of our knowledge of the content of our own experience.

There are two problems for the Burge-style account here. First, the beliefs that accompany experience are not the right kind of beliefs to count as knowledge of the contents of the experiences they accompany. Second, the relation between the relevant experiences and beliefs is of the wrong sort.

I'll address this second point first. On Burge's account, we have certain kinds of knowledge of (at least some of) our own judgments because judgment-representing beliefs are part of these judgments. It is not simply that the judgments-representing beliefs share contents with and utilize the same representational resources as the judgments they represent. In addition to all that, the judgment-representing beliefs are part of the judgments they represent. And this fact explains their status as knowledge.

On Burge's account the relevant judgment that P always occasions the belief that I have judged that P. The reason for this is that the cases that Burge is giving an account of are cases of what he calls second-order reflexive judgments -- judgments that could be perspicuously expressed by sentences like "I am hereby thinking that water is wet." In such cases, the content of the judgment that P is "taken right up" by the second-order reflexive judgment.

(Such cases are probably the norm for judgment. The second-order belief needn't be tokened in a reflexive conscious judgment. You needn't assert inwardly "I hereby think that water is wet" in order for it to be true that the judgment that water is wet be accompanied by the belief that you have judged that P. Indeed, that a judgment counts as a judgment probably requires that it be accompanied by the belief. Judgments reflect a settling of the mind. In judgment, a temporary uncertainty is, at least for the time, resolved. Unless one believed that one had judged the matter, why would one feel settled?)

It is plausible that judgments occasion second-order beliefs and it is plausible that they do so in such a way that the second-order belief counts as knowledge. However, things are different with experience. The beliefs I form while I am experiencing are not part of the experience.

Consider now my worry that the beliefs that, in humans, accompany experience are of the wrong sort to constitute knowledge of the contents of those experiences. The beliefs that normally accompany judgment are judgment-representing beliefs. However, the beliefs that normally accompany experience are not experience-representing beliefs. Instead the beliefs that normally accompany experience are environment-representing beliefs. These beliefs are the ones of the sort we have when we "believe what we see". Thus the beliefs are formed as normal accompaniments to experience do not have the right subject-matter to count as knowledge of the contents of our own experiences.

The Burge-style account cannot be extended to our knowledge of the content of our own experiences because (i) the representational resources utilized by experience and belief are different, (ii) the beliefs that accompany experience are not part of the experiences they accompany, and (iii) the beliefs that accompany experience are not about the experiences they accompany.

Externalism, Visual Content, and Burge

In this post, I will argue that Burge's strategy for explaining our knowledge of our own thoughts cannot be extended to our knowledge of our own visual experiences.

Externalism about the contents of our propositional attitudes is often thought to be inconsistent with the claim that we know by introspection -- i.e. without the benefit of knowledge of the environment around us and our embedding in it -- what we're thinking.

The arguments generally go something like this: Suppose it is true that I don't (indeed cannot) know that I am not a brain in a vat. The justification for this claim rests on the idea that I could not distinguish between being a brain in a vat and not being a brain in a vat. Now suppose that whether I am thinking that P depends on me not being a brain in a vat. Now suppose that I am now judging that P. Since I cannot know that I am not a brain in a vat, I cannot know that I am judging that P, rather than something other than P.

Burge provides both an answer to this kind of argument and an independent account of why introspection might provide knowledge. I won't address his diagnosis of the skeptical argument, I would, however, like to comment on his account of self-knowledge.

Burge argues as follows: Suppose that your cognitive faculties are in working order. Suppose you judge that P. It is plausible that as you judge that P you believe that you judge that P. Does this belief -- that you are judging that P -- rise to the happy status of knowledge? Burge argues that it does. The reason, as far as I can tell is this: the judgment that P and the belief that you are judging that P are grounded in the same representational resources. Thus, whatever mechanisms determine the content of your judgement that P also fix the content of your belief that you are judging that P. So, since the content of your judgment that P and the content of your belief that you have judged that P must be the same, there is a non-accidental connection between the truth of your belief that you have judged that P and your believing that you have judged that P. There is, that is, a non-accidental connection between your judgment and your belief of the knowledge constituting sort.

I propose to accept what Burge says here. When you judge that P, you know that you are judging that P. You know this because the belief and the judgment rely on the same representational resources. My worry is this: this strategy will not succeed in justifying the knowledge we have of the content of our own perceptual experiences.

When I look around I am subject to visual experience. This visual experience has a content in virtue of which particular things sensuously seem this way or that (blue-right-there or curved-like-that, for example). The content of my experience, presumably, contributes to the phenomenal character of my experience. Perhaps I cannot put it into words, but it seems that I am intimately acquainted with how things seem to me.

But there is abundant reason to suspect that the contents of visual experiences -- like the contents of the propositional attitudes -- depend on how one is embedded in an environment. Suppose, then, that how things seem to you depends on what kind of environment you're embedded in. Thus, for example, whether the content of your experience is P or Q depends on whether you're a brain in a vat or not.

Now we get an argument that parallels the argument that, given externalism, we don't know the contents of our own thoughts. Since you cannot distinguish between being a brain in a vat from not being a brain in a vat, you do not know that your experience has content P, rather than content Q. Thus, since how things seem to you depends on what the contents of your experience are, you do not know how things seem to you. Aggh!

Probably there are places to object to this argument, but suppose we wished to explain how we have self-knowledge regarding the contents of our own experiences. Could we use Burge's strategy? We'll want to look at two things: (i) experiences that P and (ii) the thoughts that might aspire to be knowledge of the contents of those experiences. Whether Burge's strategy can be extended to this case depends on whether these experiences and thoughts utilize the same representational resources.

There is abundant reason to suspect that belief-forming mechanisms use representational resources that differ from those used by the visual system. Accordingly, if there is to be a justificatory account of our beliefs about our own experiences, it must be one that differs from Burge's.

In a follow-up post, I'll say something about why we should think that the representational resources grounding our belief-forming capacity differ from those grounding our visual experiences. Stay tuned!