« Functional Laws and Determinable Properties
Inversion and the ways things seem »

Do perceptual experiences have A-contents?

As David Chalmers pointed out in his cool summary presentation at the end of the Time & Consciousness Conference in Sydney last month (see here and also Dave’s blog here), 3 issues need to be distinguished when thinking about the connection between temporal experiences and the metaphysics of time:

i) A-PROPERTIES OF WORLDLY EVENTS: do the events we perceive actually have A-properties (namely, can such events be present, past, or future? Is there such a thing as temporal passage?)?

ii) REPRESENTED A-PROPERTIES: do our experiences of worldly events represent them as having A-properties? (In particular, are events represented in experience as being present? Is temporal passage represented in experience?)

iii) A-PROPERTIES OF EXPERIENCES: do our experiences have A-properties? Or do they present themselves as having A-properties? (In particular, do experience feel to be present when we introspect them?).

Dave nicely spelt out some of the relations between these three issues. Here I’m interested in (ii) in connection with (i). Also, I’ll try to leave (iii) aside, if possible. There is a familiar argument going from how experience represents the world to the A-theory of time—Craig Callender calls it the “Argument from Experience”. Presumably, the argument (a fairly detailed version of it, that is) goes something like this:

(1) perceptual experiences of worldly events have A-phenomenology: when we introspect such experiences, they seem to present events in a way that we would naturally describe in A-terms (namely, as present).

(2) there is a tight connection between phenomenology and content (either phenomenology supervenes on content, or content supervenes on phenomenology).

(3) Hence, perceptual experiences represent worldly events in A-terms—as present.

(4) Perceptual experiences accurately represent worldly events as present.

(5) Therefore, worldly events have A-properties.

There is an obvious move B-theorists can make in response to this argument: deny (4).

In his paper at the conference, Craig Callender attempted to motivate that move by appealing to certain psychological results in the study of time perception. Like many participants, I failed to see how the empirical data he appealed to undermines (4) (for one thing, most of the data was about judgements of simultaneity and non-simultaneity, which are B-properties).

But intuitively, premise (4) seems question-begging (at least in the way I set up the argument). In order to show that a given experience E correctly represents some property P, one has to show at least that (i) E represents P and that (ii) P is instantiated in the environment where the subject has experience E. In the case of A-properties, however, clause (ii) is the conclusion of the argument. Hence, it seems, premise (4) is acceptable only if one already accepts the conclusion of the argument.

More importantly, it seems to me that there is a more interesting way to disarm the argument, one usually associated with Mellor’s work, and which focuses on the inference from premises (1) and (2) to (3).

Suppose we grant (1) and (2). Premise (1) is a claim about the phenomenology of experience to the effect that, pre-theoretically, it is most natural to describe perceptual experiences of worldly events using A-terms (I’m assuming here that it’s natural to describe some temporal dimension when trying to describe the phenomenology of such experiences). Premise (2) voices either a commitment to Intentionalism or to the converse view that perceptual content is determined by phenomenology.

Now, it seems to me, one could easily accept premises (1) and (2) and replace premise (3) by (3*):

(3*) perceptual experiences represent worldly events in B-terms only: for instance, experience E represents event W as being simultaneous with itself.

In other words, even if perceptual experiences of worldly events only represented these events as having B-properties, the phenomenology of such experiences would seem to be exactly the same.

If so, the phenomenology of such experiences strikes me as neutral between A-descriptions and B-descriptions of the content of such experiences. Presumably, then, A-theorists must do more to motivate the move from (1) and (2) to (3). One thing they ought to do is show that the phenomenology of the relevant experiences would be different if they represented worldly events only as having B-properties. But I don’t see how they could do that.

So far, I’ve over-simplified the issue by focusing only on A-properties such as being present. What about temporal passage? On his blog, Dave Chalmers writes:

“I think one can argue that at least at the level of representation, experience represents time as passing in a robust A-theoretic sense, so that Eden (the world where our experience is perfectly veridical) is a world of temporal passage, rather than a “block universe”. Of course this does not entail that our world is such a world: as with color experience, our temporal experience might be only imperfectly veridical.” (http://fragments.consc.net; July 27).

As Dave points out, B-theorists can always reject the counterpart of premise (4) concerning the veridicality of such experiences. And I’m happy to grant that he’s right about the phenomenology of such experiences. But does it follow that experiences represent “time as passing in a robust A-theoretic sense”?

I’m not so sure. Again, it seems that B-theorists can redescribe the content of such experiences in a way that’s compatible with the apparent A-character of their phenomenology.

What would be the content of an experience of temporal passage in A-terms? Presumably, something like this:

(A) event W is present now & (later, event W is past now & event W’ is present) & (even later, events W and W’ are past now, event W is later than W’ & event W” is present) & ….

Admittedly, this paraphrase of the A-content of an experience of temporal passage is less than ideal. Hopefully, it’ll do for the point to go through. Again, it seems as though B-theorists can redescribe the content of such an experience as follows:

(B) even W is co-simultaneous with this experience E & (later, event W and experience E are later than this experience E’ & event W’ is co-simultaneous with this experience E’) & (even later, event W and experience W are later than this experience E”, event W’ and experience E’ are later than this experience E”, W and E are later than W’ and E’, & event W” is co-simultaneous with this experience E”) & …

Or something along these lines. Again, it seems that if (B) captured the content of such an experience—rather than (A)—there is no reason to think that the phenomenology of that experience would differ. At least, I don’t really see how it could. Am I missing something?

If not, there is no entailment from the apparently A-phenomenology of such experiences to the claim that they have an A-content. In which case, it would seem as though the Argument from Experience for the A-theory is pretty hopeless.

(Apologies for the long post!)

Posted by Philippe Chuard on Sunday, August 27th, 2006 at 12:29 pm.

9 Responses to “Do perceptual experiences have A-contents?”

I’m inclined to agree, Philippe, but I’d like to get clarification on a thing or two. You question the move from 1&2 to 3 in the argument below:

(1) perceptual experiences of worldly events have A-phenomenology: when we introspect such experiences, they seem to present events in a way that we would naturally describe in A-terms (namely, as present).

(2) there is a tight connection between phenomenology and content (either phenomenology supervenes on content, or content supervenes on phenomenology).

(3) Hence, perceptual experiences represent worldly events in A-terms—as present.

Instead, you think 1&2 are consistent with:

(3*) perceptual experiences represent worldly events in B-terms only: for instance, experience E represents event W as being simultaneous with itself.

At times, you sound like you are denying 1, claiming that phenomenology really doesn’t commit to A-properties. That’s not what you intend, though–since that wouldn’t establish consistency, and you seem willing to grant Chalmers phenomenology. But when you talk about the B-theorist being able to redescribe the content in other terms, you presumably mean that he is able to redescribe it correctly. Doesn’t this mean that the A-theorist’s description is incorrect? If so, doesn’t that commit you to denying 1 or to denying that the correct description of content is determined by the correct description of phenomenology, which amounts–I think–to a denial of 2? I guess I’m just looking for a little more explanation as to where you are saying the initial argument goes wrong. It must be that you think it is invalid, but could you explain that a bit?

Left by rhowell on August 28th, 2006

For the record, I’m not sure what is to be made of the idea of an experience at a time representing temporal flow, other than that it causes certain beliefs about befores and afters. What sort of experiences are these? Isn’t an experience at a time limited to representing what is the case at that time? Doesn’t it seem that the representation of it’s relationship to other times (which, I am tempted to say, is the representation of its relationship to other experiences at different times) is a matter of something on a level higher than the experiences themselves? Such as beliefs?
I’ve been accused of being a sense-data theorist because I never know where to draw the line between what the experience contributes to content and what the judgments that come with them contribute. Perhaps an idiosyncratic confusion of mine.

Left by rhowell on August 28th, 2006

I agree with the above argument. It seems to me that it extends to more than presentness as well. Iconic memories of the past, in particular, seem to present their objects as past. I believe that’s uncontroversial among phenomenologists anyways.

Left by dbourget on August 28th, 2006

I have a similar question/comment as Robert. Whether 3* is compatible with 1 & 2 depends, I think, on how we are to understand the claim in 1 that “perceptual experiences of worldly events have A-phenomenology”. If the claim were merely that our *experiences* seem to have A-properties, then I would agree that this premise does not combine with 2 to entail 3. But 1 is stated in such a way that it seems to be claiming that the phenomenology of our experiences are such that the *perceived events* seem to have A-properties. I don’t know how to understand that claim except as the claim that our experiences represent the perceived events as having A-properties. But this amounts to 3, and is incompatible with 3*.

Left by Brad Thompson on August 29th, 2006

Hi Philippe,

I’ve posted a response here.

Left by djc on August 29th, 2006

Robert and Brad, good points.

I realise I was a bit careless in my presentation of the issues and committed the very mistake I wanted to warn against.

This is what I really had in mind: A-theorists claim (and Dave Chalmers seemed to agree with them) that we have experiences of succession of events, which represent events as flowing through time, where this flow irreducibly involves A-properties. (Brad, it’s about representation of A-properties of events, not the A-properties of experiences: I wanted to leave that 3rd issue aside). Usually, it seems as though the motivation for such a claim is the intuition that this is how the phenomenal character of such experiences strikes us. This is precisely what I want to question.

Call the phenomenal character of such experiences of succession of events P (assuming that we have experiences of succession of this kind: like Robert, I have some doubts).

I shouldn’t have said that P = A-phenomenology. What I meant to say is that P can be naturally (or intuitively) characterised in A-terms, at least at a pre-theoretical level. But nothing more than that.

It seems to me that granting this doesn’t (or at least ought not to) commit one to the claim that experiences with P must have A-contents of the kind described by Dave. Indeed, it seems plausible that P could supervene on complex B-contents (B-contents which resemble, but perhaps do not perfectly capture, the relevant A-contents) just as well as it could supervene on A-contents. In other words, the phenomenal character P of experiences of succession of events seems to me to be neutral between accounts of the contents of such experiences in B-terms and accounts in A-terms.

Perhaps, this is a better way to put the point. The following could quite plausibly happen: first, suppose you ask a subject to describe the phenomenology of the relevant experiences. It’s quite possible that she will use A-terms. Then ask her if she thinks the experience has an A-content and represents the flow of time. My guess is she’ll say ‘yes’. But then, and this is the crucial point, provide her with a complex description of the content of such an experience which only uses B-terms, and asks her whether that captures the content of the relevant experience. My guess is she’ll say ‘yes’ to that too.

Now, how should we characterise P in premise (1) of the experience argument? Is it possible to characterise it in a neutral way?

As I said, I’m happy to grant that, intuitively, we naturally use A-terms to describe P. Why doesn’t this entail that experiences with P have A-contents?

4 options.
a) although Intentionalism is true (suppose) and phenomenology supervenes on content, it ought to be possible to characterise the phenomenology more of less independently (even if usually, we don’t) of how we characterise the content. If so, even if we use A-terms in describing P, this by itself doesn’t entail that P supervenes on A-content. This is one assumption Robert seemed to reject in his post.

b) Perhaps, we should be more careful in using A-terms when describing P. What’s right is that in having experiences of succession of events, it seems AS IF there is temporal flow, and it seems AS IF such temporal flow irreducibly involves A-properties. But this is compatible with the claim that P in fact supervenes on B-contents.

c) Or it may be that, even though A-descriptions of P are natural and correct, they are correct only insofar as they amount to rough approximations of the nature of P. A more fine-grained and complete description of P would in fact involve B-terms (only).

d) And it’s quite possible too, as Robert suggests, that what really happens when we find such A-descriptions of P intuitive is that we impose our A-beliefs onto the phenomenology of such experiences, since most of us do pre-theoretically believe that there is temporal flow. This leads us to misrepresent P, because, in fact, P supervenes on B-contents.

I’m not sure which of these options I prefer.

But the important point is that B-theorists can resist the entailment from premises (1) and (2) to premise (3), when (1) is appropriately characterised, even if it can be naturally characterised in A-terms (at a pre-theoretical level, again).

What surprises me is that many people seem to claim that the phenomenal character of the relevant experiences obviously supervenes on A-contents. That’s what I took Dave to be claiming. And that’s what I don’t see. It seems to me there is room for resistance on the part of B-theorists here.

Left by Philippe Chuard on August 29th, 2006

For my part, I think I like a little combination of b and d, where the “as if q” comes as a result of the beliefs P tends to cause.
All in all, I think I’m on your side on this one, but I haven’t read Dave’s post yet…

Left by rhowell on August 31st, 2006

Ok, I’ve read Dave’s reply, which seems to be that even if one characterizes the content in B-terms–as you do in (B) of your original post–temporal passage would still be required for veridicality. I’m not sure why this is supposed to be the case, however. Here’s some possibilities, and i wonder if any of them are what Dave had in mind.
a) By characterizing things in terms of later/before, there is an implicit reference to a “now” which seems to be part of the content of experiences. The B-description must not contain indexicals, however, so any “now”-description is really an A-description.

b) What makes event x before event x’? Why not describe x’ as before x? Implicit is a stance on the direction of time, and if an experience commits on the direction of time, it implicitly presupposes a notion of temporal flow. A mere ordering has no direction.

I guess a) and b) question whether your B-reconstrual of experiences are really B-descriptions after all.

Is this what you had in mind, Dave? Even if not, what do you think Philippe?

Left by rhowell on August 31st, 2006

Hi, Robert. Definitely not (a), something closer to (b). I.e. the claim is not so much that the content involves temporal properties (presentness, pastness, etc) as opposed to temporal relations (later-than, before, etc). Rather it’s that the temporal properties/relations involved in the content are passage-involving properties/relations, as opposed to the non-passage-involving relations invoked by the B-theorist. So for (BC) to be phenomenally adequate, it would have to be read as invoking the relation Later-than, where ‘x is Later-than y’ entails that time passes between x and y, in the robust sense of passage such that the A-theorist but not the B-theorist accepts that there is passage in our world. There might also be an interpretation of (BC) as involving a B-theoretic later-than relation (small l) that does not entail passage. But the claim will be that the phenomenology determines the first, stronger version (in virtue of representing temporal passage in the world), rather than the second. Perhaps directional asymmetry in phenomenology is part of the case for this claim, though I’m not sure that this is the crux. As I said, I’m just stating the view, not trying to argue for it — I’m not even certain that it’s true, though I think it has some intuitive plausibility.

Left by djc on August 31st, 2006

Something to say?