Friday, October 19, 2007

Dennett on the "Cartesian Theater"

In his seminal 1991 book Consciousness Explained, Daniel Dennett famously criticizes what he calls the "Cartesian Theater" view of the mind. I find the criticism odd.

The central "Cartesian" claim Dennett targets is that there is a specific location in the brain "arrival at which is the necessary and sufficient condition for conscious experience" (p. 106). His argument consists mainly in denying that there's always a fact of the matter about when, exactly, an experience occurs, if one considers events at very small time scales (on the order of tenths of a second). He appears to draw from this argument what seems to be the fairly radical anti-"Cartesian" conclusion that there are, in general, no definitive facts of the matter about the flow of conscious experiences independent of the changing "narratives" we construct about them. (Elsewhere in the book, however, Dennett writes as though there are such facts. I criticize his apparent inconsistency about such matters here.)

The argument is odd in two ways:

First: Dennett does not want to deny the intuitive idea that there are "afferent" (inbound) brain processes that are not in themselves conscious, such as early visual processes in the retina and early visual cortex. Nor does he want to deny that there may be similarly non-conscious "efferent" neural processes, going out from the brain -- for example, motor impulses travelling from the supplementary motor area down the spinal cord (p. 108-109). So evidently there is a center in the brain where everything comes together, on his view. The only question is how large that center is. But how could that question of size be theoretically deep enough to drive the general conclusions Dennett wants and his characterization of the issue as one on which most previous philosophers have gone radically wrong?

Second: Ordinary external events may also be temporally indeterminate, if one looks at narrow enough time slices (even independently of issues of Einsteinian relativity). Consider an example from a real theater: An elephant and an acrobat charge onto stage right and stage left respectively, at about the same time. The elephant's trunk comes in at t + 0 milliseconds but his tail doesn't come in until t + 600 milliseconds. The acrobat's leading foot comes in at t + 200 milliseconds but his trailing foot doesn't come in until t + 350 milliseconds. Did the elephant or the acrobat enter first? Obviously, many variations of this scenario are possible. But does this support any radical, general conclusion about the temporal order of events? Does it show that the best way to think of the processing of events is in terms of multiple scripts and that there are no facts independent of our narratives? Of course not! In real theaters as in Cartesian theaters, there is blurriness at the edges. That's how the world works in general (except maybe at the quantum level). Nothing radical follows.


Erika said...

I’m not exactly sure what you are suggesting here- so forgive me if this is completely off. I am taking this to mean that there are inbound and outbound processes in the brain which are not conscious, yet still occur, so this implies that there is some “center” in the brain where these processes are being received or coming from? If this is right, I don’t really see that this creates a problem for Dennett’s denial of the “Cartesian” theater. It seems to me that even if Dennett is interpreted to require a center to the brain, that this “center” for him would be ephemeral, so to speak- not have a definite “size” or location. In other words, a conceptual center rather than a physically determinate center. So the “center” where unconscious input is processed in me now may be different from the “center” where a different input is processed in me a few seconds from now (same with output). In other words, not a center at all in the conventional sense, but rather differing places where processes take place. (And maybe these processes simply are not "famous" enough in the brain to be conscious at the moment?) At least that’s kind of what I gathered from it.

Nevertheless, this seems to create problems for Dennett considering that he has a “thin” view of consciousness. If he is allowing for such processes, then we need some reliable way of determining which processes are “allowed” to unconsciously occur in the brain and those that are excluded. I say this because if Dennett were to allow for a great number of unconscious processes to occur in the brain, it would suggest a richer view of access consciousness (I hope I am using the term right here) than he is willing to endorse, I think.

The problem that I really have with his denial of the Cartesian theater is that it seems (though I haven’t read the whole book yet obviously) he fails to adequately explain things such as a persisting personality, a sense of “self”, etc. It appears that there needs to be some cohesive and persisting element present in cognitive processes in order to account for that. Unless there were neurons constantly firing in the “personality” area of the brain, I don’t think he can have a good explanation of this. Even if he did say that, it would seem that the personality part of the brain is perhaps the part which is necessary for “conscious” experience- and thus he’s not really denying anything.

I hope that some of this is relevant and that I am not way off the mark here.

Richard said...

Hi Eric, a couple of quick thoughts:

(1) I take you to have suggested there's an area of the brain where all conscious experience occurs, i.e. a location arrival at which is necessary for conscious experience. But nothing you've said suggests that arrival at this location suffices for consciousness, right? All sorts of non-conscious processing goes on beside the conscious stuff in the large 'center' of our brain.

(2) I don't see any real blurriness in the ordinary theater case. There may be some indeterminacy in how we may choose to describe it, but the fundamental facts are perfectly precise, as your description itself illustrates. This seems importantly different from the kind of temporal indeterminacy Dennett ascribes to our experiences. It's not as though we can describe away the vagueness by specifying which parts of which experiences were had in what order, etc.

Anibal said...

The unity of consciousness (See, Bayne 2007)and the criteria to establish a relationary link between the conscious contents (prima facie Bayne reveals four unity relations: subjective, representational, phenomenal and acces unity) seems to reflect a universal common experience in conscious beings,say, that conciousness comes as whole experience or "unite" experience.

If we take the terminology from vision research this is summarize as the binding problem or how the different properties of a visual stimulus cohere to form the visual percept. It is suppose that the visual percept and their conjoint properties are bound in some neural locus, otherwise how we can experience a unite scene.

Though there are some skeptics voices such as Zeki and his microconsciousness hypothesis, that suggests that orientation is processes before motion, and motion before colour with a temporal lapse between them of some miliseconds.

Another influential voice in the neurobiology of conciousness, C. Koch in his latest book (The Quest For Concsiousness)shows no reluctance to accept the idea of an hommunculus (albeit unconscious) indicated by the functions of the frontal lobe (working memory)

From a philosophical standpoint Dennett´s view is right in stating that the "cartesian theater" goes down to a vicious circularity. If there is a little person who is concious of what we are concious how the conciousnes of that little person arise without and endless regression.
And even according to cognitive architecture thaories such as those celebrating the global workspace model favour by Dennett indicates a parallel processing of "demons" (tiny unconcious representations or scripts of the flow of concious information) but neverthless there is an ordinary concious experience, a unity conscious experience, to be eplained yet.

Eric Schwitzgebel said...

Thanks for the comment, Erika!

Regarding the "center", my thought is that if Dennett is saying that consciousness does not occur in certain areas of the brain such as early visual areas and late motor areas, then he is allowing that the brain does have some functional center where consciousness occurs (even if it may be large). Maybe the precise locus moves around (or maybe it's a distributed fact of some sort), but I don't see why that couldn't be true on a "Cartesian theater" view of the sort Dennett rebukes as well. I'm not inclined to think that the issues about precise time and space at very small intervals is qualitatively different depending on the size of the region -- whether it's, say, all but the most afferent and efferent extremes of the brain or whether it's a tiny nodule in the center.

I'm not clear on your comment about the rich/thin issue. It sounds a little like you're saying that unconscious processes would be "access conscious"...?

Dennett will talk later in the book about the "self". Whether you'll find his remarks on that topic satisfying or not, though, remains to be seen!

Eric Schwitzgebel said...

Hi, Richard! The sufficiency issue is tricky. If Dennett leans too hard on it, he turns the Cartesian materialist into a straw man, for few would say that *any* activity at all in some privileged region of the brain is sufficient for consciousness. (For example, neurons are almost always firing at some rate, even in dreamless sleep.) So presumably a certain kind of activity, in the relevant region, is sufficient for consciousness, on the Cartesian theater view -- but then we're back to the mere size issue again, since Dennett presumably also thinks that a certain kind of activity (e.g. "fame in the brain" in his later work) in the relevant regions of the brain is sufficient for consciousness.

On your second point, I think you can find blurriness everywhere. Suppose the audience on one side of the theater sees the acrobat enter slightly before the audience on the other side does. When, exactly *did* his leading foot enter, then? Or mess around with the metaphor a little: Make it street theater in the round, with shifting boundaries for the "stage". Or take different sorts of events -- say a soda spilling and a ref's blowing a whistle. What counts as the exact time at which the soda spills, down to hundredths of a second? Before or after the whistle? When exactly does the whistle start, for that matter? I can't see how facts about the blurriness and temporal indeterminacy of consciousness at very fine time intervals makes it any different than the blurriness and temporal indeterminacy of outward events.

Eric Schwitzgebel said...

Hello again, Anibal! I don't see why any of those types of approaches to the binding problem might work. I agree that we shouldn't assume that binding occurs at a specific, small region in the brain; but I also don't see any circularity in the hypothesis that it does, if that's where the evidence points....

As for the homunculus thing: I don't see why thinking that there's a region (large or small) in the brain that is functionally central to consciousness is tantamount to positing a homonculus in any vicious sense.

And just to be clear: It's not so much that I'm advocating a Cartesian theater view, as that Dennett's arguments against it seem to me peculiar.

Dave said...

Hi Eric,

I understand Dennett as arguing that certain phenomena (the phi-phenomenon and the others he considers in CE) bring out the fact that an 'atomistic' view of consciousness, upon which there is always a definite fact of the matter about the contents of consciousness at a particular time, faces some questions that we can't see how to resolve; what sort of evidence could tell against an Orwellian interpretation and in favour of a Stalinesque interpretation?

A multiple drafts model of consciousness lets us avoid these questions, since it rejects the idea that there are always determinate answers about the content of consciousness at a time. I guess I see this point about determinacy as the central anti-Cartesian claim Dennett is making. A consequence of this is that no physical or functional state of the brain at an instant suffices to fix the content of consciousness.

I take your first point to be that because Dennett identifies some inbound and some outbound processes as non-conscious, the locus of consciousness is whatever is 'in between' those processes. I think Dennett might respond that the multiple drafts model denies that any momentary state of the 'in between' suffices for consciousness - rather, being conscious is a matter of enjoying a certain amount of influence over other processing systems in the brain, something that has to happen over time, not at an instant.

I'm not sure I get the second point about the ubiquity of indeterminacy. We could say which of the elephant and the acrobat entered first if we fixed on a criterion for each of them that says what suffices for entering (e.g. the whole object needs to be in, so the acrobat enters first). But Dennett's examples are meant to make us skeptical about the idea that we could find such a criterion for consciousness-at-an-instant. To compare them to your other examples of soda and whistles - just as we might have to make an arbitrary choice when we decide to specify the instant the whistle started blowing, there's no non-arbitrary way of fixing a criterion for what's in a subject's consciousness at an instant. If you agree with that, I think you agree with the rejection of 'atomism' at the heart of Dennett's attack on the Cartesian theatre. The Multiple Drafts model is meant to be an alternative to that atomistic conception, and the radical conclusions about the dependency of the contents of consciousness on our narratives is a consequence of that model.

Great blog, by the way!

Jim said...

Hi Eric:

Just a few brief responses to some of the recent posts.

Erika: Maybe the personality part of the brain consists primarily of the private narratives within our individual streams of consciousness.

Anibal: Maybe all of our reasoning is basically circular (that each one of our mini-projects in life is a a movement largely within a charmed circle of our own making.)

Eric: what seems fascinating to me at the quantum level is that some of the famous theorists in this field (Bohr, Heisenberg, von Neuman) appear to choose a horizon of analysis similar to that supposedly experienced by Buddhist meditators (ie that what we are left with is the observing mind and the observable brain and the interaction between them).

At the macro level then the issue becomes how one cuts the horizon of analysis which in turn seems to imply that the language of problem indentification may be largely metaphoric--it does not suggest a situation that necessarily exists independently.

If all of our logic does end up being circular then I end up feeling quite dizzy.

Eric Schwitzgebel said...

Thanks for the thoughtful and helpful comment (as usual), Dave!

It seems to me that your first point is related to Richard's point about sufficiency. I agree I didn't handle this as well in the post as I might have; but as I said to Richard, I don't think Dennett can lean too hard on the sufficiency condition without making a straw man of his opponent. So, sure, he can deny that any process's being inward of the afferent and efferent extremes is sufficient for its being conscious, but can't a (reasonable, non-caricatured) "Cartesian materialist" too say that what matters is what is dominating the stage (or the pineal gland) or whatever? Then, again, it seems that it comes down to a matter of size, rather than a difference in principle.

On your second point: I do agree with Dennett's rejection of atomism. But I disagree with his suggestion that something radical or unintuitive follows. I think we should be non-atomists in the relevant sense about basically every macroscopic event -- at *some* level of spatiotemporal precision things get indeterminate. I don't see that anything radical follows about our ordinary claims, most of the time. If facts about consciousness are only as indeterminate as facts about when the football play is over or when John spilled his Coke or when the acrobat entered the stage, I don't think that should worry any reasonable hard-core phenomenal realist very much!

Eric Schwitzgebel said...

Jim, I'm not sure I understood your penultimate paragraph! In a sense it's trivial that our judgments about the world are not mind-independent. But whether the world itself depends (in some surprising, robust way) upon our minds... Well, I can't see my way through that question very clearly!

Wayne said...

Hello Eric,

A few years ago, I took an undergrad philosophy course in which we were to write some reply to Dennett's consciousness explained. As seems always to be the case we only had a few weeks to cobble something together, and even worse, our prof (who is actually an excellent teacher) had never taught this course before, and we only ended up reading 1/2 of the book!! Nevertheless my paper ended up claiming that Dennett had simply reframed the Cartesian theater as a "larger" part of the brain. In that sense, his constant railing against the term, and criticism of the usage of it, often seems, (in my admittedly extremely limited knowledge and experience) like wordplay....and borderline disengenuous! I think think that your criticism here is similar, and I take that as sweet validation of my thesis! haha. (I am not a phil major, though I have been told I should have been)

I did receive an A+ for that paper, but never got feedback since it was the end of the semester.

Anyways, thanks for the interesting post. I am currently trying to take Dennett to task for some flawed arguments I believe he made on another subject, and happened upon your site!

Norman Stone said...

The afferent/afferent problem can be mooted if you regard every neural transaction as being either predominantly efferent or afferent, depending on the amount of processing versus the amount of triggering from external sources. The degree of signal group alignment corresponds to the amount of processing (and facilitates response), while the degree of externality corresponds to amount of information (and the diversity of stimulation) in the signal group. Thus the turnaround from afferent to efferent is a continuum. The gradual turnaround model works best in the cortices, and becomes very one-sided at the periphery (sensorimotor activity), but there is no turnaround point, and there can be many regions of gradual turnaround.

Norman Stone said...

The afferent/afferent problem can be mooted if you regard every neural transaction as being either predominantly efferent or afferent, depending on the amount of processing versus the amount of triggering from external sources. The degree of signal group alignment corresponds to the amount of processing (and facilitates response), while the degree of externality corresponds to amount of information (and the diversity of stimulation) in the signal group. Thus the turnaround from afferent to efferent is a continuum. The gradual turnaround model works best in the cortices, and becomes very one-sided at the periphery (sensorimotor activity), but there is no turnaround point, and there can be many regions of gradual turnaround.