Consciousness, Computation, and Cognition

Much cognitive processing is unconscious, but some is conscious. By ‘consciousness,’ I mean phenomenal consciousness, conscious experience, what it’s like to be someone, or qualia. The relation between consciousness, computation, and cognition is a very difficult question.
A second point is that consciousness may not be entirely a matter of causal roles but also, at least in part, a matter of instantiating certain qualities (categorical properties). This is plausible given that, if anything appears to have a qualitative character, consciousness does. If consciousness is at least in part a matter of physical qualities, then the correct metaphysical account of consciousness must include qualities as well as causal powers. Although in the book I still use the term “functionalism,” strictly speaking an account of consciousness that includes qualities as well as causal powers goes beyond functionalism properly so called. (Unless, perhaps, qualities reduce to causal powers—an implausible view that I set aside here.)
Notice that an account of consciousness in terms of both qualities and causal powers is not a version of the type identity theory either. Type identity requires that conscious state types be identical to their lower-level realizers. Nothing in what I said requires such an identity. On the contrary, as I argued in my first post, higher-level properties—including qualities—are aspects of their realizers. Thus, according to this view, conscious states are aspects of their lower-level realizers that are partially (noncomputational) functional and partially qualitative.
A first point is that even if consciousness is a matter of causal roles and therefore of performing functions—as functionalism maintains—it need not be a matter of computational functions. Consciousness may be at least in part a matter of manipulating a specific physical medium in specific ways—that is, a matter of noncomputational functions. If this is correct, the result is a noncomputational version of functionalism.
The first distinction is between qualities and causal powers. Qualities (not to be confused with qualia) are categorical properties, i.e., ways that objects are on their own without depending on other objects. Paradigmatic examples include shapes and sizes. In contrast, causal powers are ways that objects are that are directed towards specific manifestations, which require partners in order to occur. Paradigmatic examples include solubility and fragility.
The two main traditional views are functionalism and the identity theory. According to functionalism, conscious mental states are states with certain causal roles, which are typically assumed to be computational. (Many people don’t see much of a difference between the causal and the computational.) According to the type identity theory, conscious mental state types are identical to lower-level physical state types. Both theories have serious drawbacks. Contra functionalism, consciousness does not feel like it’s just a matter of causal roles. Contra the type-identity theory, mental state types appear to be multiply realizable and hence not identical to physical state types.
To make progress, we need two distinctions.
(This concludes my guest blogging on my new book, Neurocognitive Mechanisms: Explaining Biological Cognition. Heartful thanks to the people—too many to list here—who helped me along the way.)
The second distinction is between computational functions and noncomputational functions. As I mentioned in a previous post, in my view, computational functions are functions defined in terms of manipulating multiply realizable vehicles in accordance with a rule. Paradigmatic examples include prime factorization or finding square roots. Noncomputational functions are functions defined in terms of producing specific physical effects on a specific physical medium. Paradigmatic examples include letting fuel into pistons (of an internal combustion engine) or lifting corks out of bottlenecks.
The relation between qualities and causal powers is debatable. Some metaphysicians have attempted to reduce qualities to causal powers, others have attempted to reduce causal powers to qualities plus laws of nature, and yet others have argued that qualities and causal powers are mutually irreducible. For present purposes, let’s assume that objects have both qualities and causal powers and leave it to others to sort out their relation.
Metaphysicians of mind have often ignored or underappreciated these distinctions and the philosophical work they can do. When these distinctions are properly deployed, I argue that they can help us make progress on the relation between consciousness, computation, and cognition.