This week, I’m blogging about my new book, The Epistemic Role of Consciousness (Oxford University Press, September 2019). Today, I’ll discuss the epistemic role of consciousness in cognition.
Could there be a cognitive zombie – that is, an unconscious creature with the capacity for cognition? As I use the term, ‘cognition’ refers to the kinds of representational states and processes that are subject to epistemic norms of justification, rationality, or reasonableness. While zombies can have mental representations, their mental representations are not subject to epistemic norms. Hence, zombies cannot have beliefs and they cannot engage in reasoning or other forms of belief-revision. At best, zombies have z-liefs, rather than beliefs. Z-liefs are no substitute for beliefs, however, since they are not subject to epistemic norms. They cannot give you knowledge of the external world.
In Chapter 4: Cognition, I start by drawing a distinction between beliefs and subdoxastic states. These states are in many ways alike: they are unconscious representational states that figure in representational processes. Even so, there is a normative difference between them: beliefs are subject to norms of epistemic rationality, whereas subdoxastic states are not. Epistemic rationality requires that your beliefs are logically or probabilistically coherent, whereas there is no rational requirement that your subdoxastic states should be logically consistent with your beliefs (or with subdoxastic states in other subsystems). For example, there is nothing rationally defective about someone who subdoxastically represents the syntactic principle that p while also believing not-p.
What explains this normative difference between beliefs and subdoxastic states? Presumably, it is not just a brute fact. There must be some non-normative difference between beliefs and subdoxastic states that explains the normative difference between them. I argue that we cannot explain this normative difference without appealing to phenomenal consciousness.
More specifically, I argue that beliefs are subject to epistemic norms only because their contents are accessible to phenomenal consciousness as the contents of judgments. Beliefs are not phenomenally conscious experiences. They are standing dispositional states that persist through periods of dreamless sleep in which there is nothing in your stream of phenomenal consciousness. Even so, beliefs are disposed to cause phenomenally conscious experiences of judgment under the right conditions. When you believe that p, you are thereby disposed to judge that p when you consider whether p. Moreover, when you exercise this disposition, your judgment expresses what you believe. I argue that this reflects the nature of belief. Beliefs are individuated by their phenomenal dispositions to cause cognitive experiences of judgment. What it is to believe that p is to be disposed to judge that p when you consider whether p.
In contrast, subdoxastic states are not subject to epistemic norms because their contents are inaccessible to phenomenal consciousness in the same way. When you subdoxastically represent that p, you are not thereby disposed to judge that p. If you are disposed to judge that p for independent reasons, exercising the disposition doesn’t express what you subdoxastically represent, but merely coincides with it. Again, this reflects the nature of the representational state. Subdoxastic states are not individuated by their phenomenal dispositions, but rather by their dispositions to figure in unconscious representational processes.
Why not ground the normative difference between beliefs and subdoxastic states in some other non-normative difference in their psychological role? Stephen Stich (1978) proposes that beliefs are inferentially integrated with each other, whereas subdoxastic states are inferentially isolated from beliefs and from subdoxastic states in other subsystems. To assess this competing proposal, we need to consider cases in which conscious accessibility and inferential integration come apart.
Case 1: Suppose your mental representation of syntactic principle p is accessible to consciousness but inferentially isolated from your beliefs. You believe that if p, then Chomsky is mistaken, and you are disposed to judge that p, but you are not disposed to infer that Chomsky is mistaken. That seems irrational. If you believe a conditional, and you are disposed to judge its antecedent, then you are rationally committed to inferring its consequent. But if an unconscious mental representation grounds a rational commitment to make inferences, then it is a belief, rather than a subdoxastic state. Therefore, a disposition to judge that p is sufficient for believing that p even if it is not inferentially integrated with your other beliefs.
Case 2: Suppose your mental representation of syntactic principle p is inaccessible to consciousness but inferentially integrated with your beliefs. You believe that if p, then Chomsky is mistaken, and you are disposed to infer that Chomsky is mistaken, although you are not disposed to judge that p. Again, that seems irrational. If you believe a conditional, and you are not disposed to judge its antecedent, then all else being equal, you are rationally committed to refrain from inferring the consequent. But if an unconscious mental representation doesn’t ground a rational commitment to make inferences, then it is subdoxastic state, rather than a belief. Therefore, it’s not sufficient for believing that p that you have a mental representation that p that is inferentially integrated with your other beliefs. On the contrary, it’s necessary for believing that p that you’re disposed to judge that p.
Why must the contents of beliefs be accessible to phenomenal consciousness in order to be subject to epistemic norms? My answer is that belief, just like perceptual experience, plays a dual epistemic role. Beliefs play an epistemic role in justifying other beliefs. What you have epistemic justification to believe at any given time depends not just on what you experience at that time, but also on what you believe at that time. But your beliefs can justify other beliefs only because they also justify introspective beliefs about themselves. And your beliefs justify introspective beliefs about themselves only because their contents are accessible to phenomenal consciousness.
Subdoxastic states don’t justify introspective beliefs about themselves because their contents are inaccessible to phenomenal consciousness. If we suppose that subdoxastic states can justify beliefs about the world, without also justifying beliefs about themselves, then we must be prepared to countenance scenarios in which you have justification to believe abominable conjunctions of the following forms:
- p and I don’t have justification to believe that p; or
- p and it’s an open question whether I have justification to believe that p.
I argue, however, that you can never have justification to believe abominable conjunctions of these forms. This bolsters the argument in Chapter 4 that the contents of beliefs must be accessible to phenomenal consciousness in order to play an epistemic role in justifying beliefs about the external world.
Hey,
I really like the blog posts and the direction of focusing on the epistemic role of consciousness.
One thing that I just don’t get, probably because I’m a complete noob in epistemology , is why only mental states that are subject to epistemic norms can give you knowledge of the external world.
It seems that perception, attention and memory all give reliable information about the environment.
All these capacities have subdoxastic states (not completely sure what this means so I might be shooting blanks) and at least some connection to doxastic states.
Is it that only in virtue of the connection to beliefs that these capacities give us knowledge? that doesn’t seem right.
I can imagine a perceptual zombie that has no beliefs at all but seems to possess the same information that the non-zombie possess. If they hold the same information and act in the same ways why will they differ in the knowledge they have?
I, personally, don’t think that much hinges on such examples but it appears that they are part of a standard methodological move you make.
I suppose I could run the same example but with more simpler organisms than humans instead of zombies. Do birds have no knowledge of the external world because they are not subject to epistemic norms?
So what am I missing? I guess I should probably just read the book, but can you point me to where you discuss these or related issues?
Jonathan, I don’t think only mental states that are subject to epistemic norms can give you knowledge or epistemically justified belief. Perceptual experience is a good counterexample: it’s not subject to epistemic norms, but it can justify beliefs about the external world.
If we set aside conscious experiences, though, I do think these two epistemic properties go together: an unconscious mental representation justifies belief only if it stands in need of justification and is thereby subject to epistemic norms. Beliefs have both properties, whereas subdoxastic states have neither.
On perceptual zombies: check out my post on blindsight and super-blindsight for my arguments that unconscious perception cannot play the same epistemic role as conscious perception.
On animal perception: I do think think birds and other animals can acquire perceptual knowledge of the environment on the basis of conscious perception.
If you’re interested in reading more about my view on the epistemic role of consciousness in perception and cognition, I’d recommend looking at Chapters 3 & 4 in particular.
Hi Declan,
I was probably misinterpreting you but I thought that the connection between norm-governed mental states and external-world knowledge followed from your claim that –
“Z-liefs are no substitute for beliefs, however, since they are not subject to epistemic norms. They cannot give you knowledge of the external world.”
I read the blindsight post and I must say I’m still not fully convinced.
The argument just shows that blindsight patients need to learn that their unconscious perception is reliable. It doesn’t show that unconscious perception doesn’t provide evidence. It just shows it doesn’t provide evidence at a certain point in time. After they learn, it does provide them with evidence.
I always thought, but might be way-off here, that justification should be evaluated in light of your entire evidence. Perhaps there is more evidence supporting withholding over accepting P. So I’ll just tweak the epistemic principle you offer and would not have to accept the conclusion.
I do assume that learning about one’s reliability need not be conscious. It may require self-reference but I take it that this is not the same as phenomenal consciousness. This is not my field so again I might be just saying nonsense.
Anyway thanks for the reference, I’ll be sure to check it out!
Jonathan, I agree that the blindsighted patient can acquire knowledge about the blind field when he learns of his own reliability. But I don’t think this knowledge is based on evidence provided by unconscious perception. After all, it can be fully explained in terms of the following inference: (i) I’m inclined to guess that there’s an ‘X’ in my blind field, and (ii) I’m generally reliable, so (iii) there is probably an ‘X’ in my blind field. The blindsighted patient is justified in believing (i) on the basis of introspection, and (ii) on the basis of induction or testimony. Unconscious perception isn’t doing any work at all!