6. Physical Theories of Consciousness

6.1 Physical C-Theories

The physical world contains elementary wave-particles (quarks, leptons, bosons). These are arranged into structures at different spatial scales. Quarks and electrons form atoms. Atoms are the constituent parts of molecules, which are the constituent parts of neurons, blood and bone. A structure at one level of the physical world will be referred to as a material:

D11. A material is an arrangement of elementary wave-particles at a particular spatial scale.

The constituent parts of materials are formed from other materials—carpets are made from nylon, which is made from molecules, and so on. Some of a material’s properties are not attributable to its constituent parts: water is wet; the electrons, protons and neutrons in water are not wet. Spatiotemporal patterns occur in materials (tartan, waves, etc.).

Physical c-theories are defined as follows:

D12. A physical c-theory links consciousness to spatiotemporal patterns in materials. Physical CC sets consist of one or more patterns and the materials in which these patterns occur.1

Physical c-theories map p-descriptions of patterns in materials onto c-descriptions of conscious states. They also map c-descriptions of conscious states onto p-descriptions of patterns in materials.

In a physical c-theory the materials are essential members of the CC sets: each pattern has to occur in a particular material. A physical c-theory that links a conscious state to an electromagnetic wave pattern would not attribute consciousness to a pile of beer cans that happened to instantiate the same pattern.

Physical c-theories fit in neatly with the standard sciences (physics, chemistry, biology, geology), which identify patterns in particular physical things (planets, molecules, proteins, glaciers). They have the same level of objectivity as these other sciences (C1).

6.2 Potential Physical CC Sets in the Brain

The normally functioning adult human brain is our only platinum standard system (A4). So the science of consciousness can only look for physical CC sets that contain one or more of the materials that are present in the human brain and one or more of the spatiotemporal patterns that occur in these materials.

Some of the brain’s properties depend on other parts of the physical world. For example, brains reflect electromagnetic waves and every neuron is a particular distance from the North Pole. These are not intrinsic properties, so they are not potential members of CC sets (C2).2

Physical CC sets can exchange physically conserved quantities, so they can e-cause c-reports during consciousness experiments (C4). It is not necessary that everything in a CC set e-causes c-reports, but the set as a whole must be capable of this. The changes in the balance of oxygenated and de-oxygenated blood that are measured by fMRI cannot e-cause c-reports because they peak several seconds after the c-report. Blood flow patterns that occur on an appropriate time scale are potential members of CC sets.3

The materials that could be members of CC sets include neurons, glia, blood, cerebrospinal fluid, electromagnetic waves, quantum states and novel materials.4 With the possible exception of novel materials (see Section 6.3), a physical CC set cannot solely consist of materials, which are typically present when the brain is unconscious (C3). Some of the materials in a physical CC set must contain patterns that only occur when the brain is conscious.

The patterns in physical CC sets could be computational structures, such as a global workspace,5 or patterns in the functional or effective connectivity between neurons.6 I have suggested elsewhere that the neural patterns caused by sensory input could be linked to conscious sensations, and that a combination of sensory and sensorimotor patterns might be linked to our conscious perception of a three-dimensional world.7 We could also use Tononi’s information integration algorithms to identify patterns in materials that might be linked to consciousness.8

Some examples of physical CC sets:

• {neuron firing pattern p3}
• {neuron firing pattern p3, electromagnetic wave pattern p4}
• {quantum pattern p5}
• {neuron firing pattern p3, haemoglobin}

In the last example the simple presence of haemoglobin is a member of the CC set. It does not matter which pattern occurs in the haemoglobin: a conscious state would only occur when p3 is present in neurons surrounded by blood.

It is essential that the members of a physical CC set can be precisely and unambiguously described. Mathematical c-theories work with formally structured p-descriptions—they cannot convert natural language descriptions of the physical world into c-descriptions. It is easy to construct p-descriptions of elementary wave-particles, atoms and molecules. It is much harder to p-describe biological materials, such as neurons (see Section 5.1).9

6.3 Novel Materials?

It has been suggested that consciousness could be linked to unknown materials, such as a novel wave-particle.10 The novel material could contain patterns that are linked to conscious states. Or it could be a passive member of CC sets that include patterns in other materials. It is conceivable that each conscious state is linked to a different novel material.

Novel materials must have e-causal powers. Novel materials without e-causal powers could not e-cause c-reports (C4) and they could not be detected with scientific instruments. So we could not verify their existence or use them to infer the presence of consciousness. This type of material should be sliced off with Ockham’s razor.

Up to this point it has not been necessary to posit novel materials to explain the brain’s operation.11 Most scientists believe that the known physical properties of the brain can account for the firing patterns that send spikes to the larynx and lead to c-reports. Novel materials might be needed if we observed brain events that did not have an identifiable cause—this might lead us to hypothesize new wave-particles. But no such cases have come to light.

The most plausible novel material is something with weak e-causal powers that plays a minor role in the e-causation of c-reports. Such a consciousness force or particle might be detectable by special instruments, but it would be invisible to our current technology. There is no pressing need for a consciousness force or particle, but we might believe in it if it was a necessary consequence of a c-theory that had been thoroughly tested in other ways.12

6.4 Simplifying Assumptions about Physical C-Theories

Experiments on physical c-theories have to demonstrate that some patterns in some materials are linked to consciousness and other patterns in other materials are not. For example, a physical c-theory might claim that some neuron activity patterns form CC sets. To prove this we would need to show that consciousness is correlated with the proposed neuron activity patterns independently of glia patterns, electromagnetic wave patterns, other neuron activity patterns, and so on.

The best way to prove that consciousness is linked to patterns in particular materials is to carry out studies that test all combinations of materials (see Section 5.3). However, the link between consciousness and particular materials cannot be fully tested in natural experiments, because the brain does not naturally change into different materials, such as silicon (see Section 5.4). So we are unlikely to be able to identify the minimal sets of spatiotemporal structures that form physical CC sets. For example, we will be unable to experimentally distinguish between these potential CC sets:

• {neuron firing pattern p3}
• {neuron firing pattern p3, haemoglobin}
• {neuron firing pattern p3, haemoglobin, cerebrospinal fluid}

Some potential CC sets can be eliminated by assuming that passive materials are not members of CC sets. For example, we can assume that the simple presence of haemoglobin is not linked to consciousness. This assumption should only be made when natural experiments cannot prove the link between consciousness and the simple presence of a material:

A7. CC sets do not contain passive materials. If the link between consciousness and the simple presence of a material cannot be demonstrated in a natural experiment, then this material can be excluded from potential CC sets.

Passive materials are only passive relative to consciousness—the materials could contain patterns that are not correlated with consciousness.

We can also assume that constant patterns are not members of CC sets:

A8. CC sets do not contain patterns that are present when the system is conscious and unconscious. If the link between consciousness and a constant pattern cannot be demonstrated in a natural experiment, then this pattern can be excluded from potential CC sets.

This assumption only applies to constant patterns that occur in the same materials as the patterns that are linked to consciousness. Constant patterns that occur in other materials can be excluded using assumption A7.

Suppose a conscious brain has neuron activity patterns p6 and p7, and the unconscious brain has neuron activity patterns p6 and p8. If a natural experiment cannot demonstrate that p6 is correlated with consciousness, then it can be excluded from the CC set using A8. The CC set would just consist of neuron activity pattern p7.

There are strong connections between the brain’s materials. When a neuron fires, there are chemical changes, fluctuations in electromagnetic fields and an altered balance between oxygenated and de-oxygenated blood. These changes are partially correlated with each other, but only one of them might be linked to consciousness.

Some of these partially correlated changes can be excluded from potential CC sets on the basis of their timing relationship with c-reports. It is more difficult to identify the correlation between consciousness and patterns that occur simultaneously. This could be done by replacing the brain’s materials, but there is little scope for this in natural experiments. To address this problem we can exclude weakly correlated patterns from potential CC sets:

A9. CC sets do not contain partially correlated patterns. When several different materials have the same spatiotemporal pattern, the material(s) in which the spatiotemporal pattern is strongest will be considered to be the potential member(s) of the CC set that is associated with the conscious state, unless the partially correlated patterns can be separated out in a natural experiment.13

Suppose a conscious brain has neuron activity pattern p9 with strength 7, and p9 occurs in the glia with strength 5 and in the electromagnetic waves with strength 8. p9 is completely absent from the unconscious brain. A natural experiment cannot demonstrate that p9 in neurons and glia should be excluded from the CC set, so the CC set could consist of p9 in any combination of the three materials. We can set this possibility aside by making assumption A9. The CC set would just consist of electromagnetic wave pattern p9.

A7-A9 enable us to develop more compact c-theories from ambiguous experimental data. They should not be rigidly adhered to because they are motivated by pragmatic considerations and go beyond the experimental evidence. CC sets might contain passive materials, constant patterns and partially correlated patterns.

6.5 Summary

A physical c-theory is a mathematical relationship between patterns in materials (captured in p-descriptions) and formal descriptions of conscious states (c-descriptions). Physical c-theories conform to the constraints and fit in well with mainstream scientific methodology.

A large number of materials and patterns are potential members of physical CC sets. We are unlikely to be able to completely separate them out in natural experiments, but we can reduce the number of potential CC sets by making assumptions A7-A9, which exclude passive materials, constant patterns and partially correlated patterns from CC sets.