On Philosophy

February 16, 2007

Two Readings Of The Self-Representational Thesis

Filed under: Essays,Mind — Peter @ 12:02 am

There are two ways of understanding the self-representational thesis about consciousness. The first is as an explanatory thesis, which attempts to explain why some systems are conscious by appealing to some of the facts about those systems, in this case their self-representational nature. The second is to understand the thesis as a phenomenal one, as an attempt to analyze and better understand the structure of experience as experienced (which may or may not move us closer to an understanding of consciousness in objective terms). However, the self-representational thesis has problems under both of these readings.

Certainly the first reading, that self-representation explains why a system is conscious in objective terms, seems to be favored by the self-representationalists themselves, as well as their opponents. For example R. Van Gulick says “transforming a nonconscious state into a conscious one is a process of recruiting it into a globally integrated complex whose organization [embodies] reflexive self awareness” (K&R 24), while Levine criticizes the thesis by arguing that it can’t explain the phenomenal, qualia, in non-phenomenal terms (thus implying that this was what the thesis was trying to do in the first place).

Let me begin then by posing a problem for the self-representational thesis that is an adaptation of Levine’s. Even given a self-representational system can’t we still meaningfully ask if that system is conscious? Specifically it still seems as if there is a gap between being consciousness, the subject of experience, and being self-representational. This is not to say that consciousness cannot simply be the property of being self-representational, but it does point out that we need to explain why self-representation should make a system conscious when being represented by another state, or representing something else, doesn’t. One way to argue that closing this gap can’t be done, and thus that there is more to consciousness than self-representation, is to provide examples of systems that are self-representational but not conscious, thus demonstrating that self-representation alone isn’t a complete explanation. One notable example of just such a system is the Gödel sentence, a sentence that states that it itself is not provable. And the Gödel sentence doesn’t achieve this self-representational quality through the use of a pronoun as we have here, which might be seen as only representing itself via our understanding of it, and not intrinsically; instead it contains a complete representation of itself (with the representation itself containing a representation of itself, ad infinitum). And certainly the Gödel sentence isn’t conscious. Thus if the self-representational thesis is to be understood as an explanation of consciousness it must be refined, in at least one of two possible ways, in order to exclude the Gödel sentence, and hopefully close the gap. The possibilities are either to claim that self-representation makes only certain kinds of things conscious, and the Gödel sentence isn’t one of them, or to define representation in a way that excludes the kind of self-representation that is contained in the Gödel sentence.

Let’s start with the first possibility, which in its most natural form would be to claim that it is only self-representational mental states that are conscious, instead of simply any self-representational system. But this seems to simply move the problem to a different situation, not eliminate it; if I hold the Gödel sentence in my mind it would seem that there is some mental state that represents the Gödel sentence, by virtue of the fact that I am holding it in my mind. And this mental state would be self-representational by virtue of the fact that the Gödel sentence is self-representational. But this mental state is not itself conscious to the best of my knowledge, since that would imply that while we were thinking of the Gödel sentence there are two consciousnesses present, one from my experiential mental state representing itself, and another from the sub-state that is my conception of the Gödel sentence representing itself. But maybe there is some way to avoid this, that somehow the mental state that contains my conception of the Gödel sentence isn’t of the right kind, or maybe it is only the complete mental state that has the potential to be conscious. But even if we can legitimately make this move there is another problem, which is that we now need to define what a mental state is. What separates a mental state from a non-mental one? Well clearly it can’t be self-representation, since we showed that some self-representing systems aren’t minds. The other obvious possibility is to define mental states as parts of systems that are conscious. But to endorse that definition would be to enter a vicious circle. In any case, limiting what self-representation can make conscious leaves a gap in the theory, an explanation is required as to why self-representation makes these things conscious and not others. Perhaps this gap can be filled, but the self-representational thesis as it stands is thus shown to be lacking as an explanation of consciousness.

The other possibility, to define representation in a way that excludes things such as the Gödel sentence, may therefore seem more attractive, since it doesn’t introduce a new gap into the theory. One popular definition of representation is as a certain kind of causal relation, and that would certainly exclude the Gödel sentence. However, it excludes the possibility of self-representation as well, assuming we accept that causation is unidirectional, and that we are working with one of the SOMT variations described by Kriegel. This follows because if M* was to represent M, M must have existed before M*, in order to be a cause of it. And thus M* cannot become a part of M, as if this were the case the resulting state of M* + M would be some new state, say M’, or if they were to become part of a complex with each other in more than name M and M* must change as a result of being put together. But M* is not causally connected to M’, nor with the changed M that results from being joined with M* in a complex, it is only causally connected to the earlier M, and thus it can represent only that M. So while the causal definition of representation is compatible with a higher order theory, where M remains distinct from M*, it is not compatible with a same order theory, where they become a single state. So defining representation as causation simply doesn’t seem to allow for self-representation at all. The other popular definition of representation is as intentionality. But this would introduce a gap into the theory, because we would then need a non-causal definition of intentionality in objective terms that permits self-directed intentional states, which the self-representational thesis doesn’t provide.

So as an explanation of consciousness the self-representational thesis falls short. Essentially it suffers from the same problems as higher order theories and the representational theory of consciousness; both of these theories explain consciousness in terms of representation, and both fail to exclude many cases that seem obviously non-conscious. Which doesn’t mean that the self-representational thesis is necessarily invalid, just that details remain to be filled in before it can either be endorsed or rejected. So let me turn then to the phenomenal reading of the self-representational thesis, which seems more promising. Under this reading the thesis is simply that every conscious state is experienced as self-representational, with different authors putting forward different theories about how this self-representation is possible. For example Williford’s treatment of the thesis is, in his own words, “largely descriptive” of conscious experience.

It is assumed by many that we are simply conscious of our experience, in fact Williford calls the thesis ubiquity. But are we really conscious of our experience in basically the same way as we are conscious of our sensations, or is it just a customary way of talking about something we find hard to understand? I think it is the latter. When I introspect on my own phenomenology it doesn’t feel like I am conscious of my experience, rather it feels like my experience is conscious, that my current consciousness and my current experience are identical. As I live through the experience I am conscious of past moments (retentions) and of my expectations about future moments (protentions). In both these cases I can bring them before “my mind’s eye” and reflect upon them, be conscious of them, explicitly, although I suppose they are always there somewhere. However, I am not able to bring my current experience before myself in this way. The closest I can get is to hold on to an experience as it passes and then have the experience of being conscious of what I have just experienced. Another way to put this would be to say that the content of the experience is not “this frog” but rather “I see this frog”, illustrating that the experiencer is part of the experience itself. However, some might be tempted to extend this content to “In this very experience …”, which I would agree is a valid way of talking, but not a reflection of the structure of my experience. To me these considerations imply that conscious mental states are not directed at themselves in any way, at most they contain, or are directed at, a conception of the self. And since I am conscious, to the best of my knowledge, I feel justified in concluding that our experience of being conscious is not necessarily self-representational.

Why do I think that self-referential talk about consciousness is just talk, and not a reflection of the structure of experience? Well, consider the property of being wet. We might define wetness in objects as being covered by some amount of water, which certainly agrees with our experience. But how then would we define the wetness of water itself? If it is wet only because it is covered by some amount of water then that water too must be wet because it is covered in water, ect. The way to avoid this regress is simply to accept that water is fundamentally wet, for reasons that aren’t revealed to us directly by our experience of it. I hope the analogy with consciousness is obvious. We are used to things being conscious because they are part of our current experience, making us conscious of them. But when it comes to our current experience itself it is possible that our usual way of defining what is conscious is misleading. Although it is normal to say that we have an experience of things, or that we are conscious of things, I think that it is a mistake to talk this way about experience itself, and that if we do talk about experience in this way it is not because we experience it as self-representational, but because we are simply used to talking about everything else in experience in this way.

Again, I can’t speak for the structure of your experience, but I feel safe in thinking about my own experience as representing my self, but not itself. Admittedly these phenomenological considerations don’t rule out the possibility that my experience is self-representational “under the hood” in an unconscious way. But the only way that claim could be motivated would be either to have a theory about the neural correlates of consciousness that implies that it is self-representational on an unconscious level, or to be working with a theory that required self-representation in systems that were conscious. The second of course would be the explanatory reading of the self-representational thesis. Of course I can’t rule out the first, since we don’t have a full theory about the neural correlates of consciousness, but the absence of this theory doesn’t justify positing unconscious self-representation either. So, with these considerations in mind, I conclude that the explanatory reading of the self-representational thesis is at best incomplete, while the phenomenal reading is misleading, at least in describing the conscious experience I have access to.

Advertisements

1 Comment

  1. I agree that self-representationality is not a defining property of consciousness, if only for the reason that consciousness isn’t perfectly available for introspection, but I think the Gödel sentence example misses the mark.

    The claim is not that consciousness is a self-representation, but that it is self-representing, as in an on-going process. A Gödel sentence by itself is not a process. It’s a “static thing” (for values of “thing” depending on what sort of ontology you’ll allow for mathematical objects like that…).

    A better example of something self-representing is a cell. A cell has DNA which represents the blueprints for how to build another cell (a self-representation), but a cell also interprets those blueprints when it replicates, meaning that the DNA is used for self-representing. So, the cell’s self-representation isn’t static, like a Gödel sentence or you previous example of a picture of itself, but a dynamic self-interpretation.

    Comment by Carl — February 16, 2007 @ 1:16 am


RSS feed for comments on this post.

Blog at WordPress.com.

%d bloggers like this: