« Do perceptual experiences have A-contents?
Shoemaker’s “Realization and Mental Causation” »

Inversion and the ways things seem

I’ve been thinking about the following hypothesis:

(Hype) In terms of their colors, fire hydrants seem to Laverne the way grass seems to Shirley.

According to Hype, while things seem to Laverne the way they seem to us, the way things seem are for Shirley systematically inverted. According to Gibbons, this hypothesis entails a contradiction and isn’t a genuine possibility (Here).
The derivation works something like this. Suppose L and S see a fire hydrant and:
(1) L and S both say (sincerely) “That fire hydrant is red”.

Assuming that they mean what they say and that having been raised in the same linguistic community that that they mean the same thing we can say:
(2) L and S both believe that the fire hydrant is red.

Now, suppose that L and S are told that the lighting conditions are funny and that this leads them to revise the belief ascribed in (2):
(3) L and S both say “Oh, perhaps I don’t know the color of the hydrant but I know that it seems red”.

Assuming that they both mean what they say and were raised in the same linguistic community, we can say:
(4) L and S both believe that in terms of its color, the fire hydrant seems red.

Assuming that there’s no reason to assume that S cannot know how things seem to her, we can say that her belief constitutes knowledge in which case:
(5) In terms of its color, the fire hydrant seems red to S.

But S is the invert so (5) is incompatible with Hype.

I’m not certain the derivation goes through. Suppose, as we must, that Shirley is an invert and that when she sees something red (in normal conditions) it looks the way green things look to Laverne. Someone could (should?) say that she’d be initially disposed to believe that the red thing is green. She’ll be disposed to think that it looks green. However, as she learns the language, she’ll also be disposed to say that the things she believes to be green are “red” and that if pressed to make the same sort of qualification by saying “it seems red”. As we know from the literature on the content of thought and speech, she will come to believe what she says. That is to say, she’ll come to believe that the things she had been disposed to believe are green are red. When pressed, she’ll come to believe the things that she now classifies as red seem red when someone tells her stories about how her perceptual evidence has been undermined.

After she comes to use the language like others, she can be said to believe what they do. As she’s no less reliable in her judgments than the rest of us, she can be said to know what they do. To say that they know how things seem and how things are in terms of their colors is not to say that they are both in equally good position to know how to use public language to properly describe all aspects of such things. The most natural view I can think of that manages to avoid Gibbon’s attack is one on which the facts in virtue of which judgments that involve the phenomenal sense of “seems” are correct are distinct from those in virtue of which the judgments that involve an evidential sense of “seems” are correct. In ordinary observers, the distinction isn’t terribly important. Shirley has to overcome the tendency she has for making mistaken judgments about how things are or the look of things owing to her status as an invert and she can with a little help from her friends. The disposition to come to mean what they mean, the availability of the distinction between evidential and phenomenal seems, and the tendency to use the evidential sense in the sort of contexts in which someone would assert (3) having been made to revise the beliefs mentioned in (2).

The situation (to me) seems just like one which is pretty clearly possible. Suppose that a neuroscientist has a device that prevents you from seeing green. Apart from the interference from the device, your color version is perfect. When you see something that is green, the device messes with the rods and cones in your eyes so that the objects look similar to gray objects but owing to the effects of the device, there is some distortion so that the objects with green surfaces shimmer a bit. Knowing the set up, you know that what the scientist hasn’t done is diminished your power of determining whether something is green by looking. By hypothesis, however, things seem different to you. Suppose you’re shown a pear. You judge spontaneously that it is green. The scientist tells you that there might be something wrong with the machine. It might affect your experiences when confronted by red objects. You say that while you don’t know that what you saw was green, it seemed green to you just then. That these things are true, that you know them to be true, that all seems perfectly consistent with the fact that there is a difference between how things are with the device on or off. The things you seem don’t seem any different but things seem different to you. Introspection tells you whether the device is on or off, for example. And yet, it seems the very same considerations that would tell us that there couldn’t be anyone like Shirley would tell us that there couldn’t be this sort of device. Surely, this is mistaken.

Posted by Clayton Littlejohn on Monday, August 28th, 2006 at 2:38 pm.

9 Responses to “Inversion and the ways things seem”

Clayton, this is an intriguing post. I’m curious about how you are thinking of the situation with Shirley in your third to last paragraph. I’m especially curious as to how it is supposed to work given the distinction you mention in the second to last paragraph concerning phenomenal and evidential seeming. Prior to learning the language, are the way things seem to Shirley inverted both evidentially and phenomenally or only phenomenally? (I assume that you don’t intend them to be inverted only evidentially).

If she is inverted only phenomenally, then I don’t quite grasp how it is that she can be, prior to learning the language, disposed to believe that hydrants are green. I am assuming that if she is not an invert with respect to evidential seemings, then Shirley would have all the same color-related belief dispositions as Laverne.

If she is inverted both phenomenally and evidentially, what reason could there be for any one, including Shirley, to believe this (as opposed to believing that Shirley is inverted only evidentially)?

Left by Pete Mandik on September 1st, 2006

A comment!

Pete, I hope that I can offer a satisfactory answer. Let me say that the view I’m outlining above is not my view (I don’t have one). Rather, I’m trying to describe a view that makes sense of how there might really be no contradiction in saying that Hype is possible while admitting that L & S know what John says they know.

The basic idea is that we can say that our experiences have various phenomenal properties in virtue of which we are disposed to believe that the objects we perceive have “corresponding” sensible qualities. The evidential properties of such conscious states might change, however, owing to the way in which a subject’s language community will alter the disposition a subject has to form certain beliefs in response to certain experiences. The evidential sense of “seems” will be sensitive to those features of experience (i.e., tell us what sorts of beliefs the subject will form when taking experience at face value) and the phenomenal sense will pick out the intrinsic and unchanging features of an experience. (I take it that the coherence of the remark that things seem different but the things I see don’t seem to have changed in the neuroscientist example depends upon this phenomenal/evidential distinction).

Because the invert S’s experiences have the phenomenal qualities they do, she’s naturally disposed to believe red things are green (green things are red, blue things are yellow, etc…). However, because she is embedded in a language community in which “red” means red rather than green, she’ll come to say that x is red/believe that x is red when she assents to “x is red” (Were she not in such a community, she’d still be disposed to make systematically false judgments about the colors of objects). Her disposition to assent overrides another disposition to make systematically false judgments about the colors of objects she sees. However, this doesn’t involve a change in the phenomenal character of her experience. That remains constant. What changes as she develops into a mature adult is only the evidential significance or “doxastic impact” of such experiences.

So, she is a phenomenal invert from cradle to grave. The evidential significance of her phenomenal states alter thanks to her being situated in the language community she’s in. It’s influence explains how it comes to be that the phenomenal properties of her experiences play the same evidential role as the different phenomenal proeprties of her invert’s experiences. [In some respects my story about how to make sense of Hype is just stolen from John's story about how ways things seem can be inverted. I don't agree with some of what he says in describing the inverted Earth story, but there's a bit of theft over honest toil going on here].

Okay, that was quite lengthy but hopefully it makes more sense of what I said above.

Left by Clayton on September 1st, 2006

Thanks, Clayton. That is quite helpful. I do have some residual concerns, though, and hopefully the following can help bring them into clearer focus. I gather, then, it is ok to describe the view you are exploring as one in which pre-linguistic Shirley is both phenomenally inverted and evidentially inverted and post-linguistic Shirley is still phenomenally inverted but evidentially just like Laverne. Part of what I wonder concerns what being evidentially just like Laverne entails. I’ve been assuming that it entails having all of the same belief dispositions as Laverne (modulo indexicals). Is this assumption correct? If so, puzzles arise concerning how it is possible for Shirley to have (post-linguistic) beliefs like the belief “that the things she had been disposed to believe are green are red”. (This certainly is not a belief Laverne has, and not just because of indexcals.) If, on the other hand, my assumption is incorrect, then I’m not sure what the difference is between the suggestion that pre-linguistic Shirley is inverted only evidentially and the suggestion that she is inverted both phenomenally and evidentially.

Left by Pete Mandik on September 2nd, 2006

Pete,

I’m not sure what to say about Shirley’s beliefs after she adapts or assimilates in part because it is part of the story that for Shirley, she has no epistemic access to the change. I take it that SHE would never believe:

(1) The things she had been disposed to believe are green are red.

It is part of the view that this is true, but that (1) never would believe such a thing or have introspective grounds for such a belief. I’m modeling my story about Shirley on the stories we’re told about folks who are switched without their knowing and come to mean XYZ by their uses of “water”. Subjects like these would never be disposed to endorse the parallel:

(2) The things I’ve been disposed to believe are water (XYZ) are not what I used to believe was referred to by “water”.

If she believed (2), she’d be disposed to infer that there’s been a switch and she lacks that disposition. If she believed (1), she’d be disposed to infer that she’s been assimilated in a way that corrects her initial dispositions and she lacks that disposition.

So, I think if I understand the worry right it is that my hypothesis that S’s experiences come to have the evidential significance that L’s experiences would have all along, they’d be disposed to form the same beliefs when placed in the same surroundings. I take it that you’re hunting for that proposition that one will believe that the other won’t (ignoring indexicals) but I’m not sure that (1) is it or that there is some other sort of proposition such that one but not the other is disposed to believe it by their experiences or previously held beliefs. But that might just be because I’m not quite getting why you think (1) is the sort of thing that Shirley would be disposed to accept.

Left by Clayton on September 2nd, 2006

Thanks, Clayton. I got (1) from something you wrote in the post, namely:

“That is to say, she’ll come to believe that the things she had been disposed to believe are green are red.”

But I’m happy to grant that there are ways of reading that sentence that don’t entail (1).

Left by Pete Mandik on September 3rd, 2006

Aha,

I see what you’re saying. You must assume that I mean believe by “believe” :)

Anyway, yes, I didn’t mean to say that she’d believe that proposition but only be disposed to believe something’s red when shown a greeny.

Left by Clayton on September 3rd, 2006

Ok, that’s cool.

This helps clear up the questions I had concerning how Laverene and Shirley compare with respect to their belief dispositions. However, there remain some questions, then, about how, given that L and S are belief-identical, there can be any way of distinguishing phenomenal inversion from evidential inversion (and, of course, phenomenal seeming from evidential seeming).

I am actually quite skeptical about distinctions such as these and perhaps this can be brought out in consideration of your example of the neuroscientific intervention. I take it that your thought experiment is supposed to show that there is such a distinction (let’s call it the “P/E distinction”) but I think the experiment actually falls a bit short.

I am assuming that in order to establish the P/E distinction, we would need a case in which things vary only phenomenally, and thus without any variation with respect to, e.g., your dispositions to sort physical objects by their surface reflectance properties (aka, their colors). And this is something that the neural intervention thought experiment fails to provide. In the thought experiment, you are able to distinguish green from, say, red things by the former seeming shimmery-gray. At first glance it may look as though you and Laverne are evidentially identical because you would sort red and green objects the same way. But, if we dig a bit deeper, evidential differences arise. There are certain sorting mistakes that you would be prone to that Laverne would not. To wit, consider the challenge of sorting objects that are green from objects that really are shimmery-gray. Laverene’s sorting superiority would be evident here. I take it, then, that the neural intervention thought experiment is conceivable without there being any P/E distinction.

Of course, none of my above remarks suffice to show that there is no P/E distinction. My aim here is primarily to emphasize the question of how one is to make the distinction. One big worry I have with respect to this issue is that if phenomenal seemings are genuinely distinct from evidential seemings, then phenomenal seemings turn out to be something that not even their possessor has epistemic access to. Such a view is not absolutely incoherent, but I gather that John’s target concerns the folk notion of seeming and I would think that whatever the folk notion of seeming is, it’s a notion of something one has epistemic access to.

Left by Pete Mandik on September 4th, 2006

Pete,

I should apologize for not getting back to you earlier but it’s been rather hectic. Making matters worse, I’ve just returned from a lengthy expedition to retrieve my car that was towed (unjustly, I should add). Alright, here goes…

I think that you’re right about the belief dispositions of the relevant subjects. However, I didn’t want to offer the neuro case to illustrate the P/E distinction exactly, but to offer a very clear case in which it seems one could assert consistently:
(1) Things seem different but the things don’t seem different or to have changed.

The idea I had was that the evidence the subject has puts him in a position to notice that the phenomenal and evidential have “come apart”. It’s not as if the evidence is required for them to come apart. The case that I’d offer to illustrate the P/E distinction is just the original inversion case. I understand that that case doesn’t move some people and I’m not sure that it moves me but its incoherence I don’t believe is demonstrated in John’s paper.

It’s part of the view that I’m exploring that what it is for an experiential state to have the phenomenal qualities it does is for it to have intrinsic, introspectible qualities that dispose you to form beliefs about the sensible qualities of objects (e.g., that that thing is red, shiny, sweet, shrill, etc…). That disposition can be overridden by various things (e.g., additional evidence, the influence of the language community) but the intrinsic qualities of experience and that disposition remain constant through Shirley’s development. When that disposition is overridden, we have a state that comes to play an evidential role it didn’t initially. The phenomenal aspects of experience are the constant, unaltered intrinsic qualities and the evidential are those “relational” and altered qualities of experience that explain your belief dispositions.

You raised some sceptical concerns that I thought I should address. I wasn’t sure if it was some general scepticism of the sort that if we insist on a P/E distinction, we won’t know how things (phenomenally) seem to us or a worry that for inverts like Shirley she won’t know how things seem phenomenally to her because of the bifurcation.

As for the general sceptical worry, I’m not sure it is so serious. At least, it is no more serious than the worries that arise in the literature on the compatibility of content externalism and self-knowledge. In the usual sorts of switching cases, the evidential properties of the beliefs expressed by different utterances of “That’s water” vary without the subject’s having noticed but the subject knows how things seem on Earth and Twin Earth and the possibility of such switches don’t seem to undermine our claims to know our own minds.

As for the particular sceptical worry, I’ll have to think more about that. I don’t yet see why we have to deny that Shirley could have introspective access to how things seem phenomenally and have knowledge of how things seem evidentially based upon knowledge of her disposition to form beliefs about the colors of objects. What she won’t have, presumably, is the ability to use public language to properly describe her phenomenal seemings. That’s what distinguishes her from Laverne, but that difference seems compatible with Shirley possessing just the knowledge that John thinks we should ascribe her.

Left by Clayton on September 7th, 2006

Thanks for the detailed response, Clayton. (And sorry to hear about your car.) I too will have to think about this stuff some more. I probably will put something up on my own blog about it. I did, by the way, have both sceptical concerns in mind. I think the one you call the general concern is actually quite a bit more serious than you do, but I’ll have to think some more to develop the concern. I think the crucial question will come down to whether intrinsic properties can indeed play the epistemic roles of being simultaneously knowable to their posessors and inexpressible to anyone else or whether, instead, that is an essentially relational characterization of a mental state.

Left by Pete Mandik on September 8th, 2006

Something to say?