Comments on Jim Pryor’s “What is De Re Thought?”
APA Eastern Division, 2003 * Invited Symposium * Susanna Siegel
Jim thinks, and I agree, that there is at least one intuitive contrast between
(a) believing, on purely general grounds, that the oldest mammal is town is a squirrel
(b) believing some animal in particular—an animal that happens to be the oldest mammal in town—to be a squirrel.
He thinks this pair illustrates the contrast between mental states that are and mental states that are not object-directed. Jim says ‘thought’ where I’ve said ‘mental state’ – from now on I’ll talk like Jim, but ‘thoughts’ are supposed to cover desires, suppositions, intentions, hopes, and other attitudes besides belief. Things are simplest if we take it not to include perceptual experiences as well. Jim’s way of pinning down what it is to be object-directed is to contrast thoughts that are relevantly similar to (b), on the one hand, with (on the other hand) both thoughts reported by relational reports, and thoughts you can have when you’re in a position to demonstrate something. I won’t review those contrasts. I take it that once they’ve been drawn, the main point of Jim’s paper is to give an account (or at least to start giving one) of, as Jim says, “when should we say we have object-directed thoughts” – where (b) is supposed to be a paradigm object-directed thought. So that’s his basic question.
Jim’s basic question: When do you have an object-directed thought?
He makes a start on answering the basic question by focusing on cases where perceptual experience, as opposed to testimony or anything else, puts you in a position to have an object-directed thought. The result is a sufficiency claim:
Jim’s sufficiency claim: You’re in a position to have an object-directed thought about o if you have a perceptual experience that attributes a property to o.
(A perceptual experience is one you can have even if you’re hallucinating).
This is a start on answering the basic question about object-directed thought. But as it stands, it isn’t terribly informative. The sufficiency claim is a bit hard to evaluate if we’re in the dark about an analogous basic question about perceptual experience:
Basic question about perceptual experience: When does perceptual experience attribute a property to an object?
Much of the work of the sufficiency claim is going to be done by an answer to the basic question about perceptual experience. Jim’s sufficiency claim will have to be supplemented by theories in the philosophy of perception about when experience attributes a property to an object, and the virtues of the theory are going to depend on the supplement. So I’m going to focus on that.
Here’s a pretty natural answer to the basic question about perceptual experience: a perceptual experience attributes a property to an object o just in case you perceive o. Call it the Simple view.
Simple view: An experience attributes the property F to an object o, iff the subject of the experience is perceiving o.
The sufficiency part of the Simple view – that perceiving o suffices for a perceptual experience to attribute a property to it – is part of the starting point of Jim’s discussion. Jim is addressing the basic question about object-directed thought for the perceptual case only, and it was pretty clear at the outset that perceiving an object is one way to be in a position to think a thought that’s directed at it (as the examples Jim starts with suggest).
It’s not clear that Jim holds the necessity part of the Simple view. Here are three putative examples he gives of perceptual experiences attributing a property to an object that is not perceived by the person having the experience:
· Hacker: Your (visual) experience attributes the property of hacking into a computer to a hacker, when you see the real-time log of his activities.
· Jet: Your (visual) experience attributes the property of having a vapor trail to a jet, when you see only the vapor trail in the sky—without even so much as a grey speck at the end.
· Intruder: Your (auditory) experience attributes the property of entering the house to an intruder, when you hear the door open and shut.
More generally, Jim suggests that experience attributes a property to an object if there is an event of o’s phi-ing, and you “perceive this event as an event in which something is phi-ing” (25). (Note that this is just a sufficient condition. Jim wants to allow that a hallucinatory experience could attribute a property to an object—for instance, you could hallucinate your father. In this case there wouldn’t be a father-involving event that you perceive, assuming it’s a total hallucination, not a partial one).
Jim doesn’t say much about when you count as “seeing (or more generally, perceiving) an event of o’s phi-ing as an event in which something is phi-ing”. I’d like to explore some options for what more might plausibly be said.
In each of the three cases above, it is easy to imagine completely ignorant perceivers who hear the door open and shut but do not believe that the house has been entered, or see the log on the computer screen but do not believe that it has been hacked-into, or see the vapor trail in the sky but don’t believe that a jet has just flown by—and moreover they don’t take any attitude toward these propositions at all, a fortiori, they don’t represent them in perceptual experience. This is certainly sufficient for them not perceiving the phi-event as a phi-event. But (and here’s the first question for Jim), What’s the minimum you’d have to add to the cognitive situation of these ignorant perceivers, in order to get them to perceive the phi-event (intrusion, hacking) as a phi-event, or to perceive the property of being a jet’s vapor trail as such? Maybe there are various minimal conditions that could be added – if so, what are some of those?
One answer that seems plausible is that what you have to add is just beliefs. I’ll return to this option later.
A different option, and one I think Jim has in mind, is that what matters at bottom is having the right kind of phenomenology. On this option, if you need to add beliefs, this just because adding them produces the right kind of phenomenology. So what kind of phenomenology is the right kind?
Here are two situations that I’d say don’t have the right phenomenology. I’d like to see whether Jim agrees.
First situation: a variant on Jim’s case of the little girl in the balcony.
Balcony: You look into the darkness, and see nothing but a dark expanse. But you believe there’s a little girl in the balcony. And you’re right, there is.
The second situation:
Franco: You look into the clear blue sky through the window of a skyscraper. You seem to see an undifferentiated blue expanse. But you believe that there is a guy in the sky who is disguised in such a way that he blends in with the background perfectly. And you’re right: there is a guy there, call him Franco. Franco is camouflaged by a special costume that blends in with the sky.
It seems plausible to me that having the beliefs in question—that there is a little girl in the dark balcony, or a guy in the sky—could change the overall phenomenology in these cases. I’m less sure whether having these beliefs could change the visual phenomenology in these cases, but maybe it could. But I want to stress that as I’m thinking of these cases, you don’t come to see outlines of either a girl on the balcony or a guy in the sky, as a result of having the belief. If the visual phenomenology changes, it is not in this way.
I feel pretty sure that even if having such beliefs did change the visual phenomenology in some way compatible with your seeing undifferentiated darkness and blue expanse, it wouldn’t change it in a way that would make the visual experience attribute a property to an object.
Suppose you agree that the phenomenology in these cases isn’t the right kind for experience to attribute a property to an object. This at least suggests a positive moral for the visual case, that the right kind of phenomenology lets us differentiate the object from its surroundings. This may be extendable to the jet case. In the jet case, though you don’t see the jet, you know (I’m switching to knowledge from true belief) where it is, and Jim might say that this affects the look of the spot at the end of the vapor trail. More cautiously, if you agree that the phenomenology is these cases isn’t the right kind of experience to attribute a property to an object, then this suggests the negative claim that some kinds of phenomenology are ruled out.
In any case, if the answer to the Basic question for perceptual experience has the form, “You need to have the right phenomenology”, it seems important for the account of object-directed thought to say something about what kind of phenomenology this is. And either the positive suggestion or the negative claim is a start on giving an answer to that.
Background beliefs and seeing-as
Whatever we say about the Franco/Balcony cases, in many of the cases Jim wants to describe as ones where an experience attributes a property to an object that you don’t perceive, it seems plausible to suppose that you come to “see an event of phi-ing as an event in which something is phi-ing” in part by having certain beliefs. (One thing I wonder about is whether it has to be beliefs, or suppositions or imaginings would work as well. But I’ll focus on beliefs). So suppose that you come to hear the event of intrusion (in the intruder case) as an event in which someone is intruding, in part by having certain beliefs about what you’re hearing (I’m being purposefully unspecific about the content of this belief). Call this the background belief.
Now, one sort of link between the background belief and the phenomenology of seeing an event of phi-ing as an event in which something is phi-ing would be a constitutive link. Another would be a merely causal link. Either way, a question arises about the background belief, and this is the second question for Jim: Is the belief itself directed toward the intruder?
If the background belief is intruder-directed, then the experience doesn’t seem to be playing any crucial role in putting you in a position to have a belief that’s directed toward the intruder. This seems to hold, whether the link to belief is constitutive or causal. If the background belief is not only intruder-directed, but anchored by perceptual experience, then there seems to be some circularity in the account.
If the background belief isn’t itself directed toward the intruder, then presumably the account is going to go in stages: at the first stage, you have an experience with a non-object-directed belief. Then, this belief and the experience join forces to make it possible for you to have an object-directed belief about the intruder. It’s not very clear to me how the belief at the first stage is supposed to differ from the belief at the second stage. If this is the way you think seeing-as works, how are the background beliefs supposed to differ from the object-directed ones that the seeing-as experience puts you in a position to have? That’s the third question for Jim.
In the Franco and Balcony cases, if Jim is going to say that the experience does attribute a property to the guy in the sky or the girl in the balcony, the intuition seems even stronger that it is some other background state, such as a belief, that is giving rise (either causally or constitutively) to these (purported) cases of seeing-as.
One line that seems amenable to Jim’s account is this: you start out with a non-perceptually anchored object-directed thought in the background, and end up with the sort of experience that puts you in position to think an object-directed thought.
If this is what happens in the Franco/Balcony cases, then there isn’t any circularity worry. But it doesn’t seem to be a way of rescuing the idea that it’s the experience that is anchoring you to the object. What seems to be doing this is rather the non-perceptually anchored object-directed background belief. I’m inclined to agree that you can have some sort of object-directed thought about the guy in the sky or the girl in the balcony. But I’m inclined to deny that the visual experiences in these cases have anything much to do with them.
I’ve been saying that to get an informative theory of object-directed thought along Jim’s lines, what’s needed is an answer to the Basic question about perceptual experience. I want to end by raising an issue about perceptual reports.
Earlier I said it’s not clear that Jim holds the necessity part of the Simple view. He’d go farther and say he rejects it. But when a perceptual experience attributes a property to an object, there is some sort of perceptual link to it. Taking “perception” in a broad sense that includes any case of experientially attributing a property to an object, Jim accepts the Simple view.
There are lots of cases where intuitions aren’t clear about whether the relations you stand in to something suffice for you to perceive it. E.g., do you see the wine when it’s in a wineskin? Do you see the boat when it’s covered by a tarp? If you have the intuitions in these cases that wine- and boat-directed thoughts are available to be had, and you like the Simple view (even in the broad version), then you’d better be prepared to say that you do perceive them. But it’s not obvious that you see the wine or the boat. It’s pretty easy to imagine cases where we’d say we don’t see them (eg, “I don’t know what color the boat/wine is, I didn’t see it”). So, if you accept the Simple view, there’s a tension between intuition about object-directed thought on the one hand, and ordinary language reports on the other.
You might think Jim could get respond to this by saying “See, I told you that you don’t have to perceive an object in order to have a thought directed at it, these cases just support me in that.” But someone could reply to this by saying that we should take ordinary language reports not just as a guide to when we perceive objects in the restrictive sense Jim is talking about when he says he rejects the Simple view, but that they should also be a guide to when a perceptual experience attributes a property to an object. And then the fact that in some cases we’d say we don’t see the wine or the boat, would count against even the broad version of the Simple view, the version Jim accepts.
I don’t think ordinary language reports are a perfect guide to perceptual acquaintance (where this is just defined as a relation that puts you in a position to think an object-directed thought). But suppose you think they are a perfect guide, and that you accept the broad Simple view. Then it seems hard to avoid the conclusion that whether you’re in a position to have an object-directed thought is sensitive to whatever factors make perceptual reports true.
Moreover, those factors seem to include things other than phenomenology, external circumstances, and causal relations between the two. For instance, it’s also easy to imagine someone saying that she did see a certain boat in the harbor, even if it was covered in a tarp, and it’s prima facie it seems like she could say something true. If so, then conversational context (or its analog for solitary silent thought) seems to play a role in determining whether you see.
Putting together the assumptions that perceptual reports are a guide to perceptual acquaintance (premise (i)), with the assumption that the truth of the reports varies with conversational context (premise (ii)), seems to get you the conclusion that if the Simple view is true, then whether we’re in a position to have object-directed thoughts varies with conversational context.
(i) ordinary language reports are a very good guide to the kind of perception that’s needed for acquaintance.
(ii) whether we say we see the boat/wine varies with conversational context (or some analog for solitary silent thought). SO
(iii) if the Simple view is true, then whether we are in a position to have object-directed thoughts via perception varies with conversational context (or some analog for solitary silent thought) in the same way as described in (ii).
I don’t like the conclusion. I also don’t like premise (i). But if you deny premise (i), then you need some notion of object-seeing that more regimented than the notion reflected in ordinary language reports. So last question for Jim is, if you don’t like (i) either, then what’s the more regimented notion.