Perusing the new issue of Philosophical Studies that came out I came across Chad Kidd’s paper Phenomenal Consciousness with Infallible Self-Representation, which happens to be freely available on Phil Studies home page. The paper is interesting and aims to respond to the challenge raised by Josh Weisberg’s paper Same Old, Same Old: The Same-Order Theory of Consciousness and the Division of Phenomenal Labor. Defenders of the Same-Order view often claim to have an advantage over higher-order theories when it comes to the problematic empty higher-order thought cases. Since the state that is being represented and the state that is doing the representing are parts of the very same sate there is thought to be no issue with the self-representing part occurring without the first-order state. Josh argued that the very same problems arise for any view that divides the phenomenal labor in the way that the higher-order theory does.
One way to avoid the problem is to have the first-order state a part of the higher-order state. On this kind of view a conscious mental state consists of the first-order state together with the higher-order self-representation. This does not allow for empty higher-order states and Kidd acknowledges this. He argues that this strategy incures costs, though, as it must say that there are some causal relationships that are necessary and this seems implausible. The reason for this, Kidd argues, is that according to this view it is necessary that consciousness can’t misrepresent, and if that is to be naturalizable it must be explainable in terms of natural relationships like causation. Instead Kidd wants to present a new version of the same order theory that incorporates insights from philosophical work on indexicals. Kidd wants to acknowledge, with Weisberg, that theories that invoke representation as part of the explanatory story must accept the possibility of mis-representation. This is why the quotational view defended by Block has problems, it cannot account for this. Kidd argues that if we adopt a Kaplan-esque semantics for the self-representational content we get a view that allows that there are possible cases of mis-representation but denies that in the actual world this is possible. In other words Kidd is arguing that it is only contingently the case that there can be no empty self-representational states in exactly the same way as that it is contingently necessary that every utterance of ‘I am here now’ must refer. Thus Kidd thinks that there is a same-order view of consciousness that does not have the empty HOT problem and also allows for the possibility of misrepresentation.
What Kidd seemingly fails to notice is that in principle one could have a higher-order view which employed the kind of semantics that he does. On this view one would have a separate HOT to the effect that one was in a red* state where ‘red*’ functions as an indexical like ‘here’ or ‘now’ does in ‘I am here now’. So whether one adopts a 2-dimensional view of the semantics of the mental states or not is independent of the question of whether one is a same or higher-order theorist. So even on the same-order view you have it being true that in some possible world there is an empty self-representation. The same problem then seems to arise. What is it like for the creature that has this empty state? Kidd suggests that if there is something that it is like then the theorist has given up on the explanatory power of the theory. Kidd says,
if it is possible to have an awareness of an experience with blue phenomenal qualities without actually having such an experience tokened in one’s mind, then it seems the production of the phenomenal blueness for the subject in such cases would be due to the higher-order mental state alone, and not a representation relation between two mental states.
Kidd is here assuming that the explanatory power of higher-order theories comes from their positing a relation between two states. If one gives up the relational structure of the theory then one gives up the explanatory power. This notion seems prevalent. I have heard Ned Block make similar remarks. I think that David Rosenthal is right that this idea really stems from thinking of higher-order views in quasi-perceptual terms. But at any rate, this i snot what the theory says. It is not the relationship between the states that explains one’s phenomenality. Rather it is that one is conscious of oneself as being in various states with mental qualities and that is just all there is to phenomenality. It is, if you will, the appearance of the relation that is doing the work.
It is here that the wine-tasting argument becomes important. We seem to have some kind of evidence that simply acquiring a new concept changes our phenomenal experience. What would explain that? One explanation is that the concept became available for deployment in higher-order thoughts. Another is that having the concept somehow changes the first-order states. Which is true seems like an empirical question. If we found that one’s first order taste states (whatever they turn out to be) are unchanged by learning the word ‘tanin’, for instance, that would be support for the higher-order theory. As of now this is an open empirical issue. And if this is right then the Kaplan-esque same-order view still has no advantages over the higher-order view.