NOTE-1068: Lycan — Representational Theories of Consciousness (SEP)

Contribution

An encyclopedia entry on theories that explain consciousness by representation. It separates four things “conscious” can mean and treats two of them. One is the sensory quality, such as the green of an after-image. The other is “what it’s like” to experience that quality. It sends awareness of one’s states and introspection to the companion entry on higher-order theories (LIT-1362). The author is a party to the debate he surveys. He is one of the three named defenders of strong representationalism, and several of the replies the entry gives are his own (Lycan 1996, 1998, 2001). The entry argues for some things. It argues that pure representationalism fails and strong representationalism does not fail for the same reason, and that quality and “what it’s like” are distinct. On wide against narrow content it ends undecided.

Key insight

A representational theory of sensory qualities is a theory of what is experienced, not of what makes it experienced. The quality is “what the world is or seems like” (§5.1), a represented property of a represented object. It can be present in a state the subject is wholly unaware of. Whether there is anything it is like to have it is a further, higher-order matter that needs awareness. “As always, the representational theory addresses only sense (3)” (§4.5.1).

Assumptions

• Intentionality is representation (introduction). Each theory “attempts to explain its target phenomenon in terms of intentionality, and assumes that intentionality is representation”.
• Four senses of “conscious” (introduction): (1) states whose subjects are “directly, internally aware of them”; (2) introspection; (3) having a sensory state with “a distinctive qualitative property”; (4) “what it’s like” for the subject. The entry uses “qualitative” for (3) and “phenomenal” for (4), and notes that Block (1995) runs (3) and (4) together, and that Dretske and Tye use “what it’s like” for (3).
• Materialism is the usual motive (§1, §2.4, §3.1), though representationalism “need not be reductive” (§2.4).
• Nonactual objects are “everyone’s problem” (§4.1). The theory says only “that when you hallucinate a lemon, the yellowness you experience is that of the lemon”, and is neutral on the metaphysics of nonexistence.

Key results

• The problem (§1). Bertie sees a green after-image. There is no physical green thing outside his head and none inside it, so the green thing is not physical: “a valid deductive argument against materialism, and its premises are hard to deny”.
• The representationalist resolution (§1). “sensory qualities are actually intentional contents, represented properties of represented objects”. The after-image is an illusion of a green blob in the room. The “there is” in “there is a green thing that Bertie is experiencing” is the lenient one of “There is at least one mythical god that the Greeks worshipped”, so no actual green entity follows. A consequence the entry calls “slightly surprising but harmless”: qualities “are not themselves properties of the experiences that present them”. The experiences have the property of representing them. Most hold the representation to be nonconceptual. Dretske and Tye supply teleologized indicator semantics.
• Representation is not enough (§2). Saying “red”, semaphoring it or pointing to a chip represents redness without anything looking red. The representation must be “specifically a visual representation”. Dretske’s further factor is that it is sensory and “systemic”. Tye’s is that it is nonconceptual and “poised”. Lycan’s is functional role.
• Grades (§2.1). Pure representationalism, that representation alone suffices, is held by no one, “representation alone is cheap and ubiquitous”. Strong representationalism (Dretske, Tye, Lycan): “representation of a certain kind suffices for a sensory quality”, with the kind given in functionalist or other materialist terms. Block’s “quasi-representationism” counts as strong. Weak representationalism, that qualitative states have representational content, is “fairly uncontroversial”. It is now denied by Campbell, Travis, Noë, Brewer and Fish. Strong representationalism is “all the less defensible” for pains and itches. Dretske and Tye hold that qualities exhaust phenomenal character; Lycan (1998) denies it.
• Wide and narrow (§2.2). If qualities are represented environmental properties, they are wide, “and molecularly identical subjects could experience different qualities”. Dretske and Lycan defend this “phenomenal externalism”, which the entry says should be called “qualitative” externalism. Shoemaker, Horgan, Kriegel, Levine and Chalmers are narrow representationalists.
• Contents and reduction (§§2.3–2.4). Shoemaker takes a colour experience to represent a disposition to cause that type of experience. Chalmers argues that perceptual content is Fregean, which suits a narrow account. Chalmers’s nonreductive representationalism identifies qualities with contents that “cannot be characterized without reference to sensory qualities”.
• Five arguments for (§3).
  • Materialism (§3.1): the alternatives are sense-data, property dualism, type-identity (an error theory about greenness), adverbialism (“severely and tellingly criticized”) and eliminativism. Dretske (1996) holds that nothing intrinsic to the brain distinguishes a red quality from a green one.
  • Veridicality (§3.2): the experience’s veridicality condition “seems to exhaust not only its representational content but its qualitative content”.
  • Transparency (§3.3), after Harman (1990): “Look at a tree and try to turn your attention to intrinsic features of your visual experience.” This shows the weak thesis only. The strong one needs two premises more: that further mental properties would be introspectible, and that none is introspected. The counterexamples include mental paint, blur, moods, the sense modality and “what it’s like” itself, and they “take a lot of overcoming”. Representationalism as defined does not need transparency.
  • Seeming (§3.4), after Byrne (2001): no change in qualitative character without a change in how the world seems. The entry says the argument presupposes that Block’s two senses of “looks” do not come apart.
  • Hallucination (§3.5), after Pautz: a relational view, then representationalism as the surviving relational view.
• Awareness of qualities is not the theory’s problem (§3.6). The options are higher-order representation, self-representation, attention, acquaintance, or replication of content. “the problems of awareness of content are already with us, and do not afflict the representational theory of sensory qualities in particular”.
• The supervenience thesis (§4). The theory entails “that there can be no difference in sensory qualities without a representational difference”. Most objections are counterexamples to it.
• Unconscious representation (§4.2). Sturgeon, Kriegel and Chalmers: representation cannot suffice for a quality “because representation can occur unconsciously”. The reply turns on the senses of “conscious”. In senses (1) and (2) most representationalists already hold “that a sensory quality can occur without being noticed by its host”: the driver “on autopilot” who saw the red light and its redness but was unaware of it. In sense (3) the assumption is tautologous and the objection begs the question. In sense (4) “so far as has been shown, a (first-order) quality can occur without there being anything it is like for the subject to experience that quality on that occasion”. A state contains a quality “in virtue of representing that quality in a distinctive way”: the functional facts that fix the modality, and, where experience implies awareness, “whatever is called for by one’s account of awareness-of”.
• Phenomenal intentionality (§4.3). Horgan and Tienson: an internal, narrow intentionality constituted by phenomenology, “the fundamental kind of intentionality: the narrow, phenomenal kind that is a prerequisite for wide content and wide truth conditions”. It makes “what it’s like” prior to intentionality, contrary to the spirit of representationalism, and “certainly it poses a general threat of circularity”. Against it: unconscious intentional states. Kriegel (2011) replies that their intentionality is not “intrinsic” but “interpretivist,” merely charitable interpretation à la Davidson and Dennett. Mendelovici and Bourget (2014) argue that naturalistic considerations favour it. It blocks any direct argument from transparency, which Kriegel uses for phenomenal intentionality.
• Laws of appearance (§4.4). Pautz (2017): nothing can appear both round and square, and these impossibilities have no parallel in belief. Speaks (2017) doubts that they are metaphysical.
• Counterexamples (§4.5).
  • Moods and sensations (§4.5.1): Rey’s elation and anxiety “seem to be just overall states of oneself”. Replies: Brentano’s thesis that every mental state is intentional (Lycan, Tye); or, against transparency, Lycan’s (1998) split of a pain into its sensory quality and its affective aspect. “The quality is what remains under morphine; what morphine blocks is the affective aspect”. The affective part can then be treated as functional.
  • Same content, different quality (§4.5.2): Peacocke’s trees, the Necker cube, the duck-rabbit, Nickel’s squares seen “as prominent”, attention raising apparent contrast (Block 2010, Carrasco), blur. In each, representationalists look for a difference in what is represented. Macpherson rebuts the replies, and Green (2016) defends them.
  • Inversions (§4.5.3). Shoemaker’s inversion is only conceivable. In Levine’s fish-heads the two eyes differ functionally, so a psychosemantics may give them different representata. On Block’s Inverted Earth there are three replies: visual contents would not shift; qualities are a special narrow content (Rey, Chalmers’s Fregean content); or qualities are wide, so molecular duplicates can differ. Dretske, Tye and Lycan take the third. The entry offers “two arguments, though not very strong arguments, for the claim that the qualities are wide”. Biggs’s Scrambler: qualities cut off functionally and causally from the environment represent nothing.
• Against wide representationalism (§4.6). The arguments are introspection, narrow content, modelling a shift, modes of presentation, memory and structural mismatch, each with a reply.
  • Memory (Block 1996): long-term memory “gives you good reason to think that the phenomenal character of the experience has not changed”. The reply, after Lycan (1996) and Tye (1998): memory contents are wide too, so after the shift “your post-shift memories of good old Earth are false” and “memory is not after all a reliable check on the sensory qualities”.
  • Structural mismatch (Pautz, after Hardin): blue resembles purple more than green, while the reflectances go the other way. So qualities match neural substrates, not representata.
  • The verdict: “Hardly anyone will accept all of the foregoing replies. But no one should now find it uncontestable either that sensory qualities are narrow or that they are wide.”
• “What it’s like” is distinct (§5, §5.1). The higher-order sense is “the property of what it is like to experience that quality”. It cannot be identical with the quality, because it is “itself a property of the quality”. Two reasons are given. First, the quality is describable (“it looks green”), and what it is like to experience it “seems to be ineffable”. Second, qualities can occur unnoticed, and then “it would not be like anything for the subject to experience that quality”. So the property “requires awareness and so is something distinct from the quality itself”. Carruthers (2000) is credited with the distinction.
• Anti-materialist arguments (§5.2). The knowledge, conceivability and explanatory-gap arguments concern sense (4). The “perspectivalist” response is that Mary acquires “a new, introspective or first-person way of representing” a fact she already knew. It requires that the first-order state itself be represented, and so “encourages a representational theory of higher-order conscious awareness”. Phenomenal concepts come in many accounts (recognitional, indexical, demonstrative, quotational or constitutive, unimodal). Chalmers’s “Master Argument” is against all of them.
• Illusionism (§5.3). To deny WIL properties, with Dennett and Dretske, “is of course not to defend a representational theory or any other theory of them”. Some explain belief in them as misrepresentation in introspection, which Frankish (2016) calls illusionism. On this view a sensory quality may be mistaken for a WIL property. Rey: introspection “mistakes the lack of detail it delivers for the accurate representation of a simple ineffable property”. Kammerer (2018), sympathetic, argues that “no previously existing account can explain the strength of the illusion”.

Claims

id	claim	strength	support
C1	Representation alone does not suffice for a sensory quality; a further sensory or functional condition is needed	strong, and uncontested in the entry	the semaphore and colour-chip cases, §2; no one holds pure representationalism, §2.1
C2	Sensory qualities are represented properties of represented, possibly nonactual objects	the thesis surveyed; argued, not settled	five arguments, §3; counterexamples answered case by case, §4.5
C3	A sensory quality can occur without the subject’s awareness, and then there is nothing it is like to experience it	moderate as argument; contested by those who identify the quality with what it is like	the autopilot driver, §4.2; §5.1, citing Armstrong, Nelkin, Rosenthal, Lycan
C4	What it is like to experience a quality is a higher-order property distinct from the quality	moderate	ineffability and the unnoticed-quality argument, §5.1
C5	The objection from unconscious representation refutes only pure representationalism	moderate, on C3	§4.2
C6	Whether sensory qualities are wide or narrow is unsettled	the entry’s own verdict	the arguments and replies of §4.6
C7	Phenomenal intentionality reverses the representationalist order of explanation and threatens circularity when combined with it	weak here: stated, not argued at length	§4.3
C8	A pain’s sensory quality and its affective aspect come apart, as under morphine	weak here: reported from Lycan (1998)	§4.5.1

Concepts

• Sensory quality (the strict, “first-order” sense of “quale”, after C. I. Lewis and Goodman): “a qualitative property inhering in a sensory state”, introspectible, apparently monadic, describable in words like “green” (§1).
• Pure, strong, weak representationalism (§2.1): representation alone suffices; representation of a certain kind suffices; qualitative states have representational content.
• Wide and narrow (§2.2): whether qualities supervene on what is in the head.
• Transparency thesis (§3.3): “not even the most determined introspection ever reveals any such additional properties”, that is, properties beyond the represented ones.
• Mental paint (Harman, Block; §3.3): introspectible intrinsic features of a representation in virtue of which it represents what it does.
• WIL property (§5): “what it’s like” in the higher-order sense, a property of a quality or of the experience that has it.
• Phenomenal intentionality (§4.3): intentionality constituted by phenomenology and prior to wide content.
• Perspectivalist response (§5.2): Mary learns an old fact under a new, first-person mode of presentation. “Perspective” here is a way of representing a fact, not a standpoint.

Connections

The entry builds on Harman (1990), Dretske (1995, 1996), Tye (1995, 2002, 2003a), Lycan (1996, 1998, 2001), Byrne (2001) and Pautz, and engages Block (1990, 1995, 1996, 2010), Peacocke (1983), Shoemaker, Levine (2003), Chalmers (2004) and the phenomenal-intentionality programme (Horgan and Tienson 2002; Kriegel; Mendelovici 2018). Of these the record holds Block’s “Inverted Earth” (LIT-634, read for THEORY-095), “On a confusion about a function of consciousness” (LIT-643) and “Troubles with Functionalism” (LIT-407), Nagel (LIT-096), Frankish (2016, LIT-458) and Chalmers (2018, LIT-425). The record’s Block “Mental Paint” (LIT-635) is the 2003 paper; the entry cites “Mental Paint and Mental Latex” (1996), which the record does not hold. Kammerer (2018), “Can You Believe It?”, is not the Kammerer paper the record holds (LIT-409).

Three cross-references in the entry point to the wrong subsection: §2.3 sends inversion cases to “Section 4.4” (they are §4.5.3), §3.2 sends a further kind of content to “Section 4.5”, and §4.5.3 sends Rey to “Section 4.5” for a point made in §2.2. The opening of §3 promises “at least four direct arguments” and gives five. None changes the content.

Bearing on the record

• CLAIM-412: the exchange’s description needs correcting. The exchange took strong representationalism as “the standing position nearest the constitutive thesis”. On this entry that is true only for the character of experience, not for its occurrence.
  • What is near. Strong representationalism is an identity, not a production claim: sensory qualities “just are representational contents of a certain kind, but not that this is analytically or conceptually true” (§4.5.3, the representationalist’s claim as stated). It also needs a further organizational condition beyond representation (§2), as the claim needs “appropriately organized”.
  • What is not. The theory “addresses only sense (3)”, and it holds that a quality can occur with nothing it is like to have it (§4.2, §5.1). What makes a state experienced is handed to “whatever is called for by one’s account of awareness-of” (§4.2), that is, to the higher-order entry. The claim says that organization constitutes the experience itself, occurrence included. So representationalism is not a version of the claim. It is the claim’s part about content, with the occurrence question split off.
  • Wide content cuts against the claim. The defenders the exchange named (Dretske, Tye, Lycan) are wide representationalists. On their view molecular duplicates can differ in quality. The claim, as it reads THEORY-095, needs its interpretation to be internal. The nearer relatives are the narrow representationalists (Shoemaker, Kriegel, Levine, Chalmers’s Fregean content), and of these the nonreductive ones define content by reference to the qualities, which the claim’s nonphenomenal definition forbids.
  • The phenomenal-intentionality programme runs the other way, as the exchange said: phenomenology constitutes the fundamental kind of intentionality (§4.3). The entry adds that its proponents must treat the intentionality of unconscious states as derived or “interpretivist”.
• CLAIM-411: supports it, from the other side of the literature. The objection that “representation can occur unconsciously” (§4.2) is this claim’s objection, put to representationalism. The entry’s answer is the claim’s own diagnosis: representation is “cheap and ubiquitous” (§2.1), so the theory must name a further condition. For the quality, that condition is sensory and functional. For its being experienced, it is an account of awareness the entry does not give. So representationalism has the same open sufficiency condition the claim names, and it moves the condition to a different theory.
• QUESTION-254. Two points.
  • The autopilot driver. On the entry’s view this is a case of sensory quality without experience. Any answer to the question must classify the case, and the entry takes the side on which it is not experienced.
  • The further factors that make a representation sensory (“systemic”, “poised”, functional role; §2) are candidates for the question’s processing-against-perspective step only if they fix occurrence. The entry uses them to fix which quality is represented, not whether it is experienced.
• TERM-073. “Interpretation” in this entry is the “interpretivist” intentionality of Davidson and Dennett (§4.3): content ascribed by a charitable interpreter. That is the sense the TERM excludes, an “association supplied by an outside observer”. The entry’s own process word is “representation”, for environmental properties, which is narrower than the TERM’s determination of significance within an organization.
• TERM-072. The entry’s only “perspective” is the perspectivalist’s first-person mode of presentation (§5.2), a way of representing a fact. It is nearest the TERM’s sense 2 (represented), and the entry says it “encourages a representational theory of higher-order conscious awareness”. It is none of the three senses exactly, so a claim that cites this literature for “perspective” should say which it means.
• THEORY-095 (Inverted Earth). The entry gives the representationalist side the THEORY’s promote_when names and the record did not hold.
  • Lycan’s reply is (iii): qualities are wide, so the traveller’s qualities shift with content. On memory, his reply is the error theory the THEORY’s refutation clause describes: post-shift memories “are false” (§4.6).
  • It does not meet the refutation clause. The clause asks for a reason “not drawn from representationism itself”, and this reason is drawn from it: memory contents are wide by the same psychosemantics. Nor does it promote the THEORY. It denies phenomenal sameness rather than accepting the shift and finding sameness.
  • The entry’s verdict is that neither narrow nor wide is now “uncontestable” (§4.6). The THEORY stays Proposed.
  • Its other leads. Stalnaker’s (1996) version of the argument does not assume narrowness, and Pautz’s structural mismatch argues for narrowness on independent grounds. Neither is held.
• THEORY-633 (illusionism). §5.3 states illusionism as the denial of WIL properties (sense 4), keeping sensory qualities as represented properties. That matches the THEORY’s “quasi-phenomenal properties” if these are the qualities. The entry reports Kammerer’s (2018) verdict that no existing account explains “the strength of the illusion”. That is the gap the THEORY’s promote_when asks a worked theory to fill. It names critics (Strawson, Prinz, Balog, Nida-Rümelin, Schwitzgebel, Chalmers 2018) and none is reproduced.
• THEORY-103 and THEORY-091. Block’s overflow is experience beyond access. The entry’s unnoticed quality is a quality short of experience. The two disagree over the autopilot case, which Block (1995, LIT-643) uses for phenomenality without access. The entry notes that Block “fails to distinguish” senses (3) and (4) (§4.5.3, §5.1).
• QUESTION-255 and THEORY-063. Bourget (2017a) argues that only individuating the senses by content explains multimodal experiences “which unify distinct sub-experiences occurring in different modalities” (§3.3). That is a third candidate for unity beside the question’s two: unity as the content of one represented scene. The entry gives it one sentence.
• Filed from this reading: THEORY-879, that representing a sensory quality does not make it experienced.
• No instruction for machine-learning practice; nothing for the anthology.

Limitations

• A survey by a protagonist. Lycan’s own replies (wide qualities, the memory reply, the pain split, the layered “shapes”) are given more room than the opposition’s, and the entry says of two of its arguments for wideness that they are “not very strong”.
• Every argument reported is summarized, and none of the cited works was read here.
• It treats representationalism as a theory of sensory qualities only. It says almost nothing about affect beyond pain and mood (§4.5.1), and nothing about the neural or empirical literature.
• The 2023 revision leaves its newest material thin: Kriegel 2017 and Bourget 2017 on affect and vividness, Pautz 2017 and 2019, Shabasson 2022. It does not report the later phenomenal-intentionality literature in detail.
• Cross-reference slips, noted under Connections.

Open questions

• What is the account of awareness that turns an unnoticed quality into one there is something it is like to have? The entry defers it to the higher-order literature (LIT-1362).
• Is the absent-minded driver’s perception unexperienced, as this entry holds, or experienced and forgotten or unattended, as Block would hold? No evidence is given either way, and QUESTION-048’s difficulty applies.
• Does Pautz’s structural mismatch (§4.6) favour narrow qualities over wide ones, as he argues, or only show illusion about resemblance, as the reply says? Reading Pautz (2014, 2019) would test it.