5.3.15.3.1
5.3.1 ---- TheTheThe WAYTheWAYWAYWAY problemproblemproblemproblem
The suggestion that perception involves representation invites a broad threat to the Policy of Containment, namely the charge that perceptual consciousness, in particular the phenomenal qualities that characterise it, must be accounted for not by the law-like patterns of dependence
that govern the perceiver’s sensorimotor engagements, but by properties of the representations themselves. In dialogue with O’Regan, Block makes the following point:
In having a cognitive appreciation of a law involving inputs and outputs, one has to think of or represent those inputs and outputs in some WAY. A machine or a creature from outer space might be able to think of human inputs and outputs in WAYS that do not involve any conscious experience. Alternatively, the WAY might itself be phenomenal—say if our cognitive appreciation is coded in imagery. Given that cognition cannot grasp anything without grasping it in some WAY, the appeal to COGNIZING in explaining sensory qualities smuggles in the very notion that is supposed to be explained. (O’Regan and Block, 2012)
This is a fairly broad point, from which we might identify various specific problems. I will discuss two of them in the remainder of this passage, and a third in the section that follows.
One of Block’s complaints is that there are ways of representing sensorimotor contingencies that do not involve consciousness, meaning that an appeal to nothing more than the representation of sensorimotor contingencies cannot be sufficient to account for the existence of consciousness. This is a limitation that the sensorimotor theory can and should accept, given that certain other conditions must be in place for an agent to be conscious at all. O’Regan (2011) asserts that consciousness requires the functional realisation of a faculty described as the
self, while Noë (2004) expresses sympathy toward the view that consciousness
is connected somehow to life; and there may be distinct proposals about consciousness that are equally compatible with the sensorimotor approach. While offering some such background conditions for perceptual awareness is no small task, it is not obvious that an account of perception and the phenomenal character of perceptual consciousness is committed to offering a fully-developed answer to this broader question. With the assumption that appropriate background conditions are in place, the sensorimotor theorist can claim not just that sensorimotor contingencies explain differences and similarities in phenomenal quality, but, more fundamentally, that it is by virtue of representations of sensorimotor contingencies that there are, in general, such things as phenomenal qualities at all.More seriously, Block’s objection suggests that once you have accounted for the presence of conscious qualities by appeal to the representation of some feature, the
explanatory weight is borne by some property of the representation other than what it represents - making any proposal about the particular properties represented a dispensable feature of the account. To assume this suggestion is correct without further argument would be a prejudice. Developing the criticism, if it is not the representation’s extension that does the explanatory work, it might be an intrinsic property of the neural vehicle that could inhere even if that vehicle lacked other properties, in particular extrinsic ones, by virtue of which it bears representational content.
Block pursues this line of argument elsewhere in an argument against representationalism about phenomenal qualities. In a thought experiment similar to Davidson’s Swampman (1987), Block (1998) imagines that an atom-for-atom duplicate of himself appears, by pure chance, which lacks the causal history needed to possess content-bearing inner states. The duplicate nonetheless undergoes the same phenomenal states, suggesting it cannot be content that yields the phenomenal state. Transposing this objection to the sensorimotor theory, Block imagines a neural duplicate of himself that lacks the relevant ongoing or historical sensorimotor interactions but enjoys identical phenomenal states (O’Regan and Block, 2012). Assuming that this is possible, the scenario appears to suggest that ongoing or historical sensorimotor interactions themselves cannot explain phenomenal qualities, because the neural duplicate does not have them. Similarly, the neural duplicate’s brain cannot track or represent the appropriate sensorimotor laws, again because there are no sensorimotor interactions to track or represent.
The sensorimotor theory, on a dynamical systems approach, could respond that duplicating the necessary neural states depends on duplicating the appropriate sensorimotor interactions, as we saw in chapter 3. However, the present aim is to defend a more conservative version of the sensorimotor theory that cannot make use of this argument. Another possible response is to endorse disjunctivism, the claim that veridical perceptual experiences are a different kind of state to non-veridical experiences like hallucinations, even if they are phenomenologically indistinguishable. The disjunctivist response would claim that since the duplicates’ states differ in kind, with one undergoing a perceptual state and the other something else, we are entitled to provide a distinct explanation for each. This response may fail to convince, however, on the ground that the sensorimotor account of phenomenal
qualities is not meant to explain perception itself, but the phenomenal character common to both the perceptual and the non-perceptual states.
It is perhaps for this reason that O’Regan endorses a different kind of response. He replies that the neural duplicates share a pertinent feature, but, contra Block, that this is not some set of intrinsic neural properties. The pertinent feature is, instead, the fact that the duplicates’ brains are each in the state they would ordinarily be in