Hedda Hassel Mørch: IIT, Russellian Monism and the Combination Problem

Transcript

1. Is IIT compatible with 
 dual-aspect (Russellian) panpsychism? Hedda Hassel Mørch NYU/Oslo
2. Dual-aspect panpsychism • Panpsychism: every physical thing is associated with consciousness. – Every physical thing is either (1) conscious, (2) made of parts that are all conscious, or (3) itself forms part of a greater conscious whole. • Dual-aspect (Russellian) panpsychism: – Every physical thing is associated with consciousness because consciousness is the intrinsic aspect of physical structure. • Avoids the main problems of physicalism and dualism (Chalmers, Strawson): – Physicalism: the epistemic gap – Dualism: mental causation – Dual-aspect panpsychism: • No (or reduced) epistemic gap between structure and realizers. • No problem of mental causation: Phenomenal properties are causally relevant in virtue of realizing physical structure.
3. Integrated Information Theory • Consciousness correlates with maximal integrated information (Φ) (Tononi 2008, 2013) – Information: how much does the system causally constrain its own immediate past and future state? – Integration: how much does the information depend on causal interconnections between the parts? – Maximality: does any overlapping system (part or whole) have higher Φ? • The Exclusion principle • Empiricial support: – Explains why consciousness is absent in epileptic seizures, deep sleep, anesthesia, cerebellum. • IIT entails panpsychism: – Every system is either (1) a maximum of Φ, (2) made of parts that are maxima of Φ (3) forms part of a maximum of Φ. • But IIT doesn’t entail dual-aspect panpsychism: – Consciousness can be correlated with every physical thing in other ways (emergence, identity, etc.)
4. Objections to IIT and dual-aspect panpsychism • If IIT and dual-aspect panpsychism can be combined, it would bring advantages for both views. • IIT gets response to philosophical objections: – Panpsychism is absurd (Searle) – But if dual-panpsychism avoid the problems of physicalism and dualism, compatibility with panpsychism is an advantage. • Dual-aspect panpsychism gets help with the combination problem – How does complex macroconsciousness arise from simple microconsciousness? – Results in problems analogous to the main problems of physicalism and dualism: • The epistemic gap: no entailment from microphenomenal properties in any physical relation to macrophenomenal properties. • Macro-mental causation: microphenomenal properties fill in all physical structure; no room for distinct macrophenomenal properties
5. IIT and the combination problem • IIT as a complete solution in and of itself – Exclusion principle: macroconsciousness (with max Φ) replaces microconsciousness • no problem of macro-mental causation. – Tononi: max Φ → combination follows a priori from phenomenological axioms • no epistemic gap. – Problem: many disagree about axioms or fail to see entailment.
6. IIT and the combination problem • IIT with the phenomenal bonding view (Goff forthcoming, 2009): – Some physical relations have an unknown intrinsic nature. – No epistemic gap between microphenomenal properties, their physical relations and the intrinsic nature of the relations to macrophenomenal properties – Problem: which physical relation? – IIT: causal relations that connect elements of systems with maximum Φ are (or become) phenomenal bonding relations. • IIT with the fusion view (Seager 2010, forthcoming; Mørch 2014): – Macrophenomenal properties are fusions of microphenomenal properties. – No problem of macro-mental causation: macrophenomenal fusions replace their micro-phenomenal causes. – Problem: where are the physical fusions? – IIT: maxima of Φ are physical fusions – they replace their parts (causally and phenomenally)
7. Coarse-graining in IIT • IIT selects a spatiotemporal grain where Φ is maximal. – The brain (or NCC) considered as a system of particles: low Φ – The brain considered as a system of neurons: high Φ – The brain considered on the nanosecond time scale: low Φ – The brain considered on the millisecond (or longer) time scale: high Φ • Information below the selected grain doesn’t matter to the experience. – If a brain has maximum Φ as a system of neurons, it doesn’t matter whether the neurons are organic or silicon. It will have the same experience regardless. • Multiple (microphysical) realization
8. The coarse-graining problem • Given coarse-graining, the microphysical structure of our brain has no experiential aspect/realizer. • As an argument: – Intrinsic-structural supervenience: If dual-aspect panpsychism is true, then physical structure nomologically supervenes on phenomenal properties. – No independent variation: If physical structure nomologically supervenes on phenomenal properties, there can’t be things with different physical structure, but identical phenomenal properties, given the same laws. – Multiple realization: Neuron brains and silicon brains would have different microphysical structure, but identical phenomenal properties, given the same laws. – Therefore, dual-aspect panpsychism is false.
9. The Exclusion postulate • Can’t overlapping microphenomenal properties realize microphysical structure? • Ruled out by IIT’s Exclusion postulate: consciousness never overlaps – only the system with maximal Φ is conscious. • Therefore, dual-aspect panpsychism is incompatible with IIT in view of Exclusion and coarse-graining.
10. Abandon Exclusion? • Philosophically problematic: – Societies, the internet, galaxies will be conscious. – Problem of the many: NCC minus 1 neuron, NCC minus 2 neurons, etc., will also be conscious. • Empirically problematic: – With Exclusion: a system loses consciousness when its Φ goes below level of parts. – Without Exclusion, IIT can’t explain (straightforwardly) loss of consciousness occurs before Φ reaches zero. – Alternative explanation without Exclusion: reportability or memory is correlated with Φ above a threshold only – but is this necessarily so?
11. Abandon coarse-graining? • Philosophically problematic: – The grain problem (Lockwood, Sellars) • The grain of our experience is coarser than the microphysical grain of the brain, therefore our experience should have a macrophysical correlate. • As it gets with coarse-graining, but loses without it. • Empirically problematic: – Without coarse-graining, Φ will peak in individual neurons/molecules/atoms, not in single extended brain area. – Vast number of consciousnesses in the brain with same level of consciousness.
12. Proposal: modify coarse-graining • Tension between two aspects of the combination problem, the grain problem and the palette problem: – Grain problem: macroconsciousness has too little structure – Palette problem: macroconsciousness has too many qualities • ~17 fundamental particles in physics; uncountably many macroqualities – This suggests: missing structure is encoded in extra qualities. • Modified coarse-graining principle: information below max Φ grain matters to the quality, but not the structure, of experience. – If your organic neurons were replaced by silicon neurons, your experience would have the same structure (contrastive, motivational), but different qualities (shifted spectrum). • Problem: – How is structure encoded or grounded in unstructured qualities?
13. Summary • If IIT is compatible with dual-aspect (Russellian) panpsychism, it can help solve the combination problem. • IIT is incompatible with dual-aspect panpsychism in view of Exclusion postulate and coarse-graining principle. – Entails that fine-grained microphysical structure of our brains has no experiential aspect/realizer; dissolves necessary connection between the mental and the physical. • Neither Exclusion nor coarse-graining can be abandoned without generating both philosophical and empirical problems for IIT. • But they can perhaps be modified: – Modified coarse-graining principle: microphysical realization matters to the quality but not the structure of experience. – (In the full paper, I also suggest a way of modifying Exclusion) • There might be further obstacles to unifying IIT with dual-aspect panpsychism and solving the combination problem – more research is needed!