Chapter 8 of The Given discusses the topic of cognitive phenomenology. My view of the matter is simple: either accept cognitive phenomenology or deny that there is such a thing as conscious thought. How can you deny the existence of conscious thought?! So, grant that cognitive phenomenology exists.
Cognitive phenomenology, as I define it, is a kind of phenomenology that is essentially over and above sensory phenomenology, and that is paradigmatically associated with conscious thought (although it is also present in cases of conscious perception and emotion.) On this view, there is something it is like to consciously think that the weather is depressing or to consciously think that humans are puzzling, something that is irreducible to any sensory phenomenology that may be associated with these thoughts.
The thesis that conscious thought is an essentially cognitive-phenomenological phenomenon involves two claims:
(1) Conscious thought is an essentially phenomenological or experiential phenomenon, just as perceptual experience and emotional experience are essentially phenomenological or experiential phenomena
and
(2) Conscious thought has its own distinctive irreducible kind of phenomenology: cognitive phenomenology.
Given these two claims, my overall argument proceeds in two stages. I argue that non-phenomenological accounts of conscious thought fail. I focus on accounts that rely on the idea that a thought’s being conscious somehow consists in the nature of its ‘informational connection’ to other mental states. I argue that no amount of informational connection to other mental states is sufficient for a thought’s being conscious. For a thought to be conscious it must also have phenomenological character. The second stage of the argument is then directed at those who accept that conscious thought must involve phenomenology, but who maintain, contrary to (2), that we need only appeal to sensory phenomenology in saying what a conscious thought consists in. I argue that no amount of sensory phenomenology that may be somehow tied up with or integral to the occurrence of a thought can account for that thought’s being conscious. Here I’ll just summarize the second stage of my argument.
It begins with the following simple principle, which I call the conscious content principle or ‘CC’ for short:
CC: If an occurrent thought T is to be a conscious thought, the representational content of that thought must in some manner be consciously entertained, the representational content must be in some manner consciously occurrent.
I argue that the prospects of satisfying CC while appealing only to sensory phenomenology are very bleak. I consider two sensory proposals: one based on inner speech and one based on a causal connection between unconscious content and sensory phenomenology.
The first sensory proposal for satisfying CC is to link the representational content of thought to inner speech or verbal imagery—to occurrent mental images of sentences, which could be acoustic. In ‘The Sensory Basis of Cognitive Phenomenology’, for example, Prinz suggests that
sentences do not merely stand in for thought, but actually constitute thoughts. When we produce sentences in silent speech, they issue forth from unconscious representations [emphasis added] that correspond to what those sentences mean (these are perceptual representations if empiricism is true). Sentences inherit their truth conditions from the unconscious ideas that generate them. So produced, these sentences aren’t arbitrary marks, but rather meaningful symbols. If we define athought as a mental state that represents a proposition, then mental sentences qualify as thoughts. (Prinz 2011, 187)
On one natural reading this proposal fails to satisfy CC. If Prinz’s proposal is committed to a thought’s content being nonconscious, it looks as if his claim is that the thought is really nonconscious and that some associated sensory phenomenology is conscious. On this proposal, conscious thoughts come out as nonconscious. In fact, according to this proposal, it looks as if there is no such thing as conscious thought at all.
Suppose we allow that the occurrence of some silent inner speech is literally part of the occurrence of a conscious thought. Can it be the occurrence of the inner speech that makes the representational content of the thought conscious? By claiming that the tokens of the inner speech sentences inherit truth conditions from their corresponding unconscious representations, Prinz may be surreptitiously counting representational contents as conscious. But then something more is conscious than just inner speech, given that inner speech is not itself intrinsically contentful. And now we need to hear more about what a content’s being conscious consists in. If this is not Prinz’s intended way of thinking about the mechanism of inner speech, and the representational content remains unconscious, then again it seems that all we really have is some associated sensory phenomenology that is conscious, while the thought itself—with the representational content that it has—remains unconscious.
The second sensory proposal is that CC can be satisfied for a particular thought when the nonconscious representational content causes some co-occurrent sensory phenomenology and that it is in virtue of this causal relation that the sensory phenomenology makes the representational content conscious. It seems clear, however, that appealing to a (the) causal connection between the representational content of the thought and the associated sensory phenomenology is just not enough to establish the right type of connection. One possibility is this: that the representational content of the thought should simply cause one to have some sensory phenomenology without making the thought conscious at all. Consider the conscious thought that grass is green. The representational content that grass is green might in such a case simply cause a (possibly co-occurring) green patch image or perhaps a thought about green patches. And rather than explaining the sense in which representational content is conscious, this proposal seems again like the claim that the representational content is nonconscious, although it does cause some conscious sensory phenomenology.