Contribution
A handbook chapter that compares the three frameworks neuroscientists take most seriously, and sets aside quantum, panpsychist and representationist theories. Its comparison uses one criterion, whether a theory recognises the explanatory gap, and argues that only the biological theory does. It also separates an “ambitious” from a “modest” higher-order theory and argues against the first.
Key insight
A theory that analyses consciousness in functional or cognitive terms cannot have the explanatory gap. It has instead its own question, why this function or this combination of thought and representation should be conscious, and that question would arise for any realisation of the function. The real gap, why this neural basis is the basis of this experience, presupposes only that experience has a brain basis. So a theory’s fit with the gap is a test of whether it is a theory of experience at all.
Assumptions
- Phenomenal consciousness is “what it is like” (Nagel), and the gap is real at least for current neuroscience.
- All scientific accounts accept that consciousness is based in the brain; that alone generates the gap.
- Representation is used in a less liberal sense than mere correlation: an MT+ activation is a representation only if strong enough to be harnessed in choice, at minimum in priming.
- The consciousness-of relation is granted for the sake of argument: a conscious state is one one is conscious of being in.
Key results
- Three distinctions (pp. 1112–1113). A non-representational MT+ activation, a genuine representation of motion, and a representation that is a key part of a phenomenally conscious one. In workspace terms: a minimal sensory activation, a dominant peripheral coalition, and a global activation.
- What consciousness does versus is (p. 1113). On the biological view, broadcasting and higher-order thought are functions of consciousness, not its nature.
- HOT and the gap (p. 1114). HOT’s ingredients (thought, aboutness, representation) each exist unconsciously, so neuroscience could explain each without finding “some magic glow” when they combine. The puzzle is HOT’s own, as a “prime order thought” theory would have its own puzzle about prime embeddings. HOT is “a kind of ‘no consciousness’ theory of consciousness”.
- Gap and dualism (pp. 1114–1115). As a pre-Socratic could not see how heat is motion, we lack the objective concept that would show how a subjective concept picks out the same thing. New Wave physicalism: subjectivity and objectivity are properties of concepts. Papineau’s attempt to dissolve the gap this way fails, since two concepts of orange produce no gap.
- Consciousness-of (pp. 1115–1116). The deflationary (Sosa) and same-order (Brentano, Burge, Kriegel) accounts explain why a conscious state is one we are conscious of, without HOT.
- Against ambitious HOT (pp. 1116–1118). Rosenthal’s conditions (non-inferential, mental states only, present tense) are added to meet counterexamples. The mismatch problem: what is experienced if a HOT about toe pain accompanies a visual red, or nothing? Infants: auditory synaptic density peaks around 3 months, frontal association cortex around 15 months; children to 3–4 have trouble reporting the source of their knowledge (Gopnik and Graf).
- The self (pp. 1118–1119). Visual-field relations make an experience a subject’s own even if he cannot know of it, as might be the case for extinction patient GK. Goldberg, Harel and Malach found an “intrinsic” self-related system active in introspection and inhibited in fast categorisation, which “disconfirms” Rosenthal’s view that a visual experience consists in a perception causing a thought that I have it.
- Machines (p. 1119). Workspace and HOT allow machine consciousness; the biological theory may require the electrical–chemical–electrical transfer of neural coding.
Claims
| id | claim | strength | support |
|---|---|---|---|
| C1 | Only the biological theory fully acknowledges the explanatory gap | moderate: conceptual argument (the “prime order thought” analogy) | pp. 1114 |
| C2 | Concept dualism reconciles the gap with physicalism | moderate: standard New Wave argument, debated in the cited Block 2006 and White 2006 | pp. 1114–1115 |
| C3 | Ambitious HOT relies on ad hoc stipulation and faces the mismatch problem | moderate | pp. 1116–1117 |
| C4 | Our justification for experience in infants and autistic children exceeds that for higher-order thought | moderate: developmental evidence, “far from overwhelming” by the author’s account | pp. 1117–1118 |
| C5 | Integrated information theory cannot distinguish consciousness from intelligence | weak to moderate: one example (seizure) against the 1998 version | pp. 1111–1112 |
| C6 | The biological theory is less friendly to machine consciousness | strong as a statement of the theory’s commitments; whether biology matters is not shown | p. 1119 |
Concepts
- biological theory: consciousness is a biological state of the brain (Place, Smart, Feigl; Block 1978; Lamme).
- ambitious vs modest HOT: an analysis of phenomenal consciousness in terms of higher-order thought, versus the claim that there is also a higher-order kind of consciousness.
- explanatory gap: we do not understand why the neural basis of an experience is the basis of that experience rather than another or none.
- New Wave physicalism: the gap is a dualism of concepts, not of properties.
- mismatch problem: a HOT whose content does not match the sensory state it is about, or that is about no sensory state.
Connections
- Block (2002), LIT-639. The concept/property move against Mary is stated there at length; here in a page.
- Block (2007), LIT-646. Cited for the claim that the biological account allows experience a subject does not know about.
- Chalmers (LIT-406, NOTE-336). Cited for the hard problem and as the definer of dualism; the chapter’s concept dualism is the rival to Chalmers’s view that experience is fundamental, declared on LIT-639.
- Integrated information (LIT-445, NOTE-365). The chapter’s criticism targets the 1998 version; IIT 3.0 identifies experience with a maximally irreducible conceptual structure, which this chapter does not address.
- Chalmers’s meta-problem paper (NOTE-354) cites this chapter as “Block 2009” for the biological theory and argues that it does not explain problem intuitions.
Bearing on the record
- THEORY-023. States the biological alternative to the computational functionalism that theory says architectural indicators presuppose, and says outright that the workspace is functionalist and realisable in silicon. Added as one of the theory’s sources (v2), for its account of what the dispute is about.
- No ML instruction. The machine section is about what theories allow, not about how to build systems.
Limitations
- A handbook overview. The empirical points are brief and second-hand.
- The comparison is by one criterion, the gap, which a deflationist rejects at the outset; the chapter grants that this makes HOT attractive to those who deny the gap.
- The biological theory is left unspecified beyond the MT+ example and the thalamic-switch background, which the chapter says “is in flux”.
Open questions
- What would make a biological property, rather than the function it realises, the basis of consciousness, and how could that be tested? Block (2002) argues we have no conception of how.
Corrections
- none to a seeded skim (there was no seed)