Friday, December 31, 2004

Multiple Drafts

In this post I want to provide a quick overview of Dennett's "Multiple Drafts" theory of consciousness, as presented in his book, Consciousness Explained. Several of the key ideas are ones I've discussed previously. Most central is Dennett's rejection of the Cartesian Theatre - a place in the brain where "it all comes together" and consciousness occurs.

Dennett maintains that cognitive discriminations need only be made once. The information does not then need to travel to any special area of the brain in order to become conscious. Without the 'theatre', there is no need for such a 'presentation' to take place. As I quoted before: "That would be a waste of time and (shall we say?) paint. The judgment is already in, so the brain can get on with other tasks!" (p.128)

Another important consequence of rejecting the Cartesian picture (with a focus now on the 'consciousness occurs' aspect of it), is that our brains can represent time using a medium other than time itself - so our subjective experience of temporal order need not match the objective order of discrimination (this allows us to explain many otherwise-puzzling phenomena; follow the link for details). The counter-intuitive upshot of this is that consciousness does not occur on a moment-by-moment basis. Because our brain's representations are spread out throught time as well as space, you cannot 'freeze' time and ask what is being consciously represented at any given instant.

But in fact Dennett goes even further than this. Even if we allow for temporally-extended representations, he suggests there simply is no objective fact of the matter as to which representations are conscious and which are not. His alternative proposal is the 'Multiple Drafts' model. It can be introduced via analogy with electronic publishing (pp.125-126):
In the world of publishing there is a traditional and usually quite hard-edged distinction between pre-publication editing, and postpublication correction of "errata." In the academic world today, however, things have been speeded up by electronic communication. [... Often] several different drafts of an article are simultaneously in circulation, with the author readily making revisions in response to comments received by electronic mail. Fixing a moment of publication, and thus calling one of the drafts of an article the canonical text [...] becomes a somewhat arbitrary matter. [... Most of] the important effects of writing a journal article are spread out over many drafts [...] if we feel we need the distinction at all, we will have to decide arbitrarily what is to count as publishing a text. There is no natural summit or turning point in the path from draft to archive.

To apply the analogy, there is no natural distinction between pre-experiential (Stalinesque) and post-experiential (Orwellian) revisions. Our brains construct (and continually edit) multiple 'narrative streams', and there is no objective fact of the matter as to which of these is the 'canonical text' of our conscious experience.

An intuitive example of this multiplicity is provided by (apparent) "unconscious perception and intelligent action." You've probably been lost in thought or conversation whilst driving, and discovered that you have no memory of your past few minutes' car-driving activities. But as Dennett asks:
But were you really unconscious of all those passing cars, stop lights, bends in the road at the time? You were probably paying attention to other things, but surely if you had been probed about what you had just seen at various moments on the drive, you would have had at least some sketchy details to report. The "unconscious driving" phenomenon is better seen as a case of rolling consciousness with swift memory loss.
An even more striking case is the phenomenon of being able to count, retrospectively in experience memory, the chimes of a clock which you only noticed was striking after four or five chimes. But how could you so clearly remember hearing something you hadn't been conscious of in the first place? The question betrays a commitment to the Cartesian model; there are no fixed facts about the stream of consciousness independent of particular probes. (pp.137-138, original emphasis)

I'll finish up with one last quote, which draws on ideas I've discussed above, and highlights some central aspects of the Multiple Drafts theory: (pp.135-136)
As soon as any [perceptual] discrimination has been accomplished, it becames available for eliciting some behaviour [...] or for modulating some internal informational state. [... T]his multitrack process occurs over hundreds of milliseconds, during which time various [changes...] in content can occur, in various orders. These yield, over time, something rather like a narrative stream or sequence, which can be thought of as subject to continual editing by many processes distributed around in the brain, and continuing indefinitely into the future. [...] This skein of contents is only rather like a narrative because of its multiplicity; at any point in time there are multiple drafts of narrative fragments at various stages of editing in various places in the brain. While some of the contents in these drafts will make their brief contributions and fade without further effect - and some will make no contribution at all - others will persist to play a variety of roles in the further modulation of internal state and behavior and a few will even persist to the point of making their presence known through press releases issued in the form of verbal behaviour.

Probing this stream at various intervals produces different effects, precipitating different narratives - and these are narratives: single versions of a portion of "the stream of consciousness." If one delays the probe too long, the result is apt to be no narrative left at all. If one probes "too early," one may gather data on how early a particular discrimination is achieved in the stream, but at the cost of disrupting the normal progression of the stream.

To illustrate this last sentence, consider the example of the cutaneous rabbit. Normally, subjects report feeling a series of equidistant taps along their arm. However, if you probe them too early (perhaps by asking them to push a button as soon as they feel the second tap), the illusion of 'hopping' never develops. They will instead (accurately) report that the second tap occurred in the same place as the first. This demonstrates that what we are conscious of is dependent upon how and when our stream(s) of consciousness is 'probed'.


  1. OK, so that's the overview. Now let's talk plausibility... What do you think, Richard? I've never been a fan. It's just about conceivable to me that multiple drafting is a necessary condition for consciousness. But it seems completely unbelievable that it's a sufficient condition. 

    Posted by uriah

  2. "no fixed facts about the stream of consciousness independent of particular probes"
    Er maybe i didnt understand it properly but doesn't all of this mean we CANT TELL about conciousness without probes
    the conciousness should exist without the probes it is just that they help a person to remember and to prove things about the conciousness.

    as to the rabit again he mostly said this but htee may be a point of disagreement - what may happen is that the brain recieves information it them tries to match it with previous experiences (or somthing similar) - this may actualy take some time maybe a fraction of a second maybe longer depending on all sorts of factors. But if you ask to early the brain has not had time to find that sort of an answer (as argued it seems above) so it gives the sensory answer instead or a half complete answer or even poentially a mangled one. later on it records the second answer AS WELL as the first one but since you can only recall option it may be as if it overwrote the sensory information but technically I doubt it would do exactly that though. This means that there may well be a narative but if somthing dominant is recorded that narative may be hard to find - and be subject to the normal decay of neuron connections.
    Then one has a problem of defining what conciousness is. is it the chain of instaintanious "you"s or is it the agregate apparent you as you seem to be suggesting here or for example can we take it to a whole new level and define it as some sort of chain of events not even bound to you physically? 

    Posted by GeniusNZ

  3. Uriah,
    Something like the Multiple Drafts model seems necessary to explain the empirical data. I think the theory provides a very good explanation of what we might call 'informational' or 'access' consciousness (I don't know the proper term, but I'm talking about the sense of 'consciousness' which simply means what info is [introspectively] 'accessible' to an agent.)

    I have a lot of trouble understanding how it is supposed to account for the qualitative aspects of consciousness, however. (Dennett seems to want to do away with qualia altogether!) But that's a topic that warrants a future post of its own.

    Genius: "doesn't all of this mean we CANT TELL about conciousness without probes"

    One could say that, I suppose, but as I quoted Dennett in my 'Cartesian Theatre' post:
    "Putative facts about consciousness that swim out of reach of both 'outside' and 'inside' observers are strange facts indeed."

    Further, it follows from the multiple drafts model that there really are no facts of consciousness independent of particular probes. This is because consciousness requires a single narrative stream, and this is only triggered by probes. Until probing occurs, we have multiple narratives ('drafts') in progress, so we cannot say which one of them is the objectively true narrative of our consciousness. There simply is no fact of the matter, according to this theory.

    So even though it's possible that we just "can't tell" the contents of unprobed consciousness, we would be justified in accepting the "no fact of the matter" interpretation, if that is what our best scientific theory tells us. 

    Posted by Richard

  4. Right, Dennett wants to quine qualia. He also holds that phenomenal and access consciousness cannot be separated (see his response to Block's BBS target article on this). So perhaps one way to organize Dennett's overall view of consciousness is as follows: (i) phenomenal consciousness has a qualitative and an access aspect; (ii) let's be eliminativists about the qualitative aspect; (iii) let's be multiple-draftists about the access aspect. 

    Posted by uriah


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)