1 Three plausible theses about perception

So far, I have suggested that we should adopt minimal intentionalism:

I have also suggested that it is plausible to think that objects and natural kinds can be among the things represented in our experience. If this is right, then it looks as though the contents of one’s experience are not always fixed by one’s intrinsic properties. For consider two experiences which are identical in every respect but that they are directed at two distinct identical twins. If objects can be among the contents of experience, then it looks like these two experiences will have different contents. But the subjects of the two perceptions might have been in the same intrinsic state -- they might have been molecule-for-molecule duplicates. This sort of case strongly supports externalism about perceptual content:

(I.e., content does not supervene on intrinsic properties.)

But it is plausible to think that phenomenal character is not like content in this respect; i.e., it is plausible to think that phenomenal character does supervene on intrinsic properties of perceivers. As Hawthorne puts it,

It may be overstating things to call this perfectly obvious. But there is an intuition here which seems to be difficult to reject. Let’s call the view here internalism about phenomenal character:

These three theses are not themselves contradictory. But many have thought that they were in tension. The tension can be expressed informally like this: intentionalism says that there is a necessary connection between content and phenomenal character; but if there’s this close of a link between them, how can one of these supervene on intrinsic properties, and the other not?

What we want to see is whether this tension can be turned into an argument against the conjunction of minimal intentionalism, externalism about content, and internalism about phenomenal character. The form this will take is the examination of various theses which have seemed plausible to many which can be added to this conjunction to make an inconsistent set of propositions.

2 From minimal intentionalism to biconditional intentionalism

Minimal intentionalism is a one direction supervenience claim: it says that sameness of content guarantees sameness of phenomenal character. Many people who endorse minimal intentionalism also go on to endorse the converse of this claim. This amounts to what I earlier called biconditional intentionalism:

Let’s suppose that we adopt biconditional intentionalism. This can be shown to be inconsistent with the conjunction of internalism about phenomenal character and externalism about content as follows:

(Here ‘phenomenal character’ means something like ‘total phenomenal character.’)

So at least one of (1), (2), and (4) must be false.

We could block this argument by giving up biconditional intentionalism. But many people seem to think that minimal intentionalism without biconditional intentionalism is a kind of bizarre stopping point. Isn’t the whole motivation of intentionalism the reduction of phenomenal character to content? And doesn’t this involve saying that having an experience with a certain phenomenal character just is having an experience with a certain content? And doesn’t this kind of property identity entail biconditional intentionalism?

The property identity does entail this biconditional. But it’s not obvious that the reduction of phenomenal character to content should be the motivation for minimal intentionalism, for a few reasons:

• One important motivation for intentionalism is the transparency of experience: the idea that any introspectible change is a change in how the world is represented as being. But transparency only counts in favor of minimal intentionalism.
• If we adopt biconditional intentionalism, then we have to restrict our view of what kinds of things can be parts of the contents of experience. But it is plausible that, e.g., external objects can be a part of the contents of experience; and it is plausible that two experiences can be of distinct but indistinguishable objects. So we have independent reasons for rejecting biconditional intentionalism.
• Sometimes biconditional intentionalism is motivated as follows: it seems extremely difficult to give a physicalistically acceptable account of phenomenal character; but it seems (comparatively) easy to give a physicalistically acceptable account of representational content; so, if phenomenal character is representational content, this will improve our chances of giving a physicalistically acceptable model of the mind. But to say this we have to endorse biconditional intentionalism. This line of argument does not seem at all convincing to me. If it is extremely difficult to give an account of phenomenal character, this task does not automatically get easier if we find that phenomenal character = representational content. Why not think that this kind of equivalence, if true, shows that giving a physicalist account of representational content is much harder than we thought?

So, to me, it seems plausible that the best way around this argument is be sticking with minimal intentionalism, but rejecting biconditional intentionalism.

3 From externalism to ‘strong externalism’

Above I suggested that there are good grounds for being an externalist about the contents of experiences. The formulation of this claim above is equivalent to the following:

The truth of this can only be explained by the fact that sometimes the content of a subject’s experience is shaped by facts external to him. This might suggest the following view: what it is for a subject to represent the world in a certain way just is for him to be connected to, or embedded in, his environment in a certain way. But then it might seem that we can advance a stronger claim than the bare externalist one above; perhaps it’s not just the case that some contents of experience fail to supervene on intrinsic properties, but that all such contents fail to supervene on intrinsic properties of perceivers. This is the move from what I have been calling ‘externalism’ to what Egan and John call ‘strong externalism.’ We can express it like this:

Let’s suppose that this is true. Then one might think that this makes plausible the idea that ‘content inversion’ is possible. This is the idea that two people can be inverted with respect to the contents of their experience; so when if when I am in intrinsic state I I am representing o as red, then when Invert is in state I, he is representing o as green. (It is important top see that this is different from the usual spectrum inversion cases, which are inversions in phenomenal character, not in content.)

Let’s suppose that this is possible. Then we argue as follows:

A key strength of this argument is that it does not rely on biconditional intentionalism; but deduces a contradiction from strong externalism and internalism about phenomenal character with the addition of only minimal intentionalism. One way to read this is as showing that the real culprit in these kinds of cases is minimal intentionalism, and that trying to solve the problem by switching from biconditional intentionalism to minimal intentionalism is only a temporary stopgap.

This seems wrong to me. Minimal intentionalism does not actually contribute all that much to the above argument. To run the above argument, you don’t even need anything as strong as the possibility of content inversion. The following scenario is already enough to generate a contradiction with internalism about phenomenal character:

This scenario leads to a contradiction with internalism about phenomenal character if it is metaphysically possible for two subjects A and B to satisfy this description. So, to block a contradiction, given internalism about phenomenal character + the possibility of content-inversion, we need something much stronger than the negation of intentionalism, which, as applied to the case of content inversion, says

Rather, we need the following claim:

This isn’t just the negation of intentionalism, which just says that sometimes sameness of content can go with difference of phenomenal character; this says that, in the case of content-inverts, sameness of content must always go with difference of phenomenal character. I do not see any plausibility in this claim.

So what should we say instead? It seems to me like this. Maybe content-inversion is possible, maybe not. We really have no way of knowing at this point. The standard externalist arguments do not support the possibility of content-inversion; if we had a well-supported externalist theory of mental representation which entailed its possibility, then this would be evidence; but we have no well-supported theory of mental representation, externalist or otherwise. But if it is possible, then it seems plausible to me that spectrum inversion among intrinsic duplicates is possible too. But then the problem lies with internalism about phenomenal character, not with intentionalism. So I suggest that the minimal intentionalist should be confident in the truth of the following disjunction:

but not worry too much about which disjunct makes it true.

4 Conclusions

I have suggested that there is a very plausible position which avoids these arguments. We should hold the following views:

• Minimal intentionalism; there can be no change in phenomenal character without a change in content. This is motivated by the intuition that experience is ‘transparent.’
• The items which can be represented in experience fall into two classes. There are those with which, in Johnston’s sense, we could become originally acquainted in hallucinatory experience, and those which we could not. Among the former class are the colors; among the latter are natural kinds and objects. Call the items with which we can have original acquaintance in hallucination hallucination-available.
• It is natural to think that having an experience which includes a hallucination-available property supervenes on one’s intrinsic properties. The fact that a property could be the content of a hallucinatory experience is good grounds for thinking that its presence in the content of a perception supervenes on intrinsic properties of the perceiver -- after all, hallucinations just are cases in which the normal external setting of a veridical perception is absent.
• We should reject the move from minimal intentionalism to biconditional intentionalism; we should allow that two experiences can have different contents, but the same phenomenal character. This is consistent with motivating intentionalism via transparency, since the transparency intuition only motivates the one-direction supervenience claim.
• We should say further that when two experiences have different contents but the same phenomenology, the difference in content is always a difference with respect to items which are not hallucination-available. Examination of cases makes this claim plausible.
• A further strength of this constellation of views is that it is consistent with qualia internalism: the view that phenomenal character of experience supervenes on intrinsic properties of the subject of the experience. This is because any variance in the contents of experience which is consistent with sameness of intrinsic properties will be variance in items which are non-hallucination-available; and this is the kind of variance which does not bring about a change in phenomenal character.