I. Introduction

If one knows everything physical about the world, does one necessarily know everything about the world? In this essay, I will argue that the answer to this question is yes and that the view of mental phenomena that holds this to be true, namely physicalism, is to the best of our knowledge both necessary and correct, despite an appearance to the contrary.1 I shall first show that it is possible, and indeed quite plausible to construe mental events, states or properties — these being the most likely counterexample to the physicalist position — as having a purely physical basis. I shall then present an argument as to why this is necessarily so. I will begin by examining the strongest argument against physicalism, Jackson’s (1982) knowledge argument and a  credible physicalist response to it. I will then move on to describe the precise type of physicalism that I wish to defend, along with the epistemological difficulties involved in providing a definitive answer to this question in the absence of a completed science of consciousness.

Firstly, it is important to be clear about the meaning of the term ‘physical’. To avoid any circularity in its definition, I shall take this to refer to all spatiotemporal phenomena; i.e. those that take place in a particular region of space at a particular time. This includes all forms of matter, energy, fields, and other theoretical entities proposed by the natural sciences both as they exist today and in the future. For brevity I shall refer only to mental and physical events, but my argument should also be taken to apply to an account of mental states or properties, or any combination of the three, which I do not believe present any particular difficulties for the views discussed.

II. The Knowledge Argument

Jackson’s knowledge argument (ibid.) arises from an intuition generated by the following thought experiment and represents a powerful argument against physicalism. It involves Mary, a colour scientist who is an expert on all aspects of colour perception, and fully conversant in the completed sciences of physics, chemistry and human neurophysiology. (I will later argue that this latter premise is doing more work than would first appear, but for now let us assume that Mary is in possession of all the physical facts about how we perceive colour, and what happens in our brains when we do so.) Mary herself, however, has never seen any colours except for black, white and various intermediate shades of grey. She has lived her whole life in an entirely monochrome environment, experiencing the outside world only through black and white video monitors, and even her own body and clothing have been coloured black or white to complete the illusion. The point of the thought experiment is that upon seeing a colour — red, for example — for the first time we are inclined to think that Mary would learn something new. ‘Aha!’, we can imagine her saying, ‘that’s what it’s like to see red!’. In his 1982 paper, Jackson argues that if Mary was already in possession of all the physical facts before she saw red for the first time, then what she would learn as a result of her experience must therefore be a non-physical fact. If there are such things as non-physical facts, then it is not possible to know everything in virtue of just the physical facts, and therefore physicalism is false.

This is a persuasive and logically valid argument that is difficult to refute. However, it is unclear that what Mary learns (if indeed she learns anything) can properly be described as a fact at all. Unlike her existing knowledge of the physical facts, which is essentially propositional in nature, what she learns is not something that can be either true or false, as the usual definition of a fact requires.2 Rather, it would be more natural to say that she has a kind of experience that she has not had before. If this is true, then how do we explain the intuition that Mary acquires some new knowledge, namely what it is like to see red? Firstly, when Mary learns a physical fact about colour perception, she is employing her cognitive capacity for rational thought, the application of concepts, and so on. Her ability to reason requires the acquisition of propositional knowledge which, on the physicalist view, may be described in terms of particular modifications to the micro-structure of Mary’s brain corresponding to the particular propositions or beliefs in question. When Mary actually sees a colour, on the other hand, she is using an entirely different cognitive faculty, namely perception, which plausibly involves different parts of her brain. Such experiences are represented as different modifications to the micro-structure of Mary’s brain to those that occur when she learns a new concept. As such, they result in an entirely different sort of knowledge, which is essentially non-propositional in nature. As Papineau (1998: 382) argues, there is no reason why learning the precise neurological description of an experience should have the same effect upon us as actually having the experience itself, even though both experiences represent the same thing; e.g. seeing the colour red. Instead, learning about something and actually experiencing it generate different ‘representational states’ (Jackson 1999) within the conscious mind.

On this view, the fact that we call both first- and third-person modes of experience ‘learning’ or ‘knowing’ does not reflect the type of representation involved, but rather the acquisition of a new representation, whose precise form may differ depending upon the experience in question. In effect, the words ‘know’ and ‘learn’ contain an equivocation, as they cover both propositional and non-propositional states. This distinction mirrors the division between objective and subjective, or third-person and first-person, modes of experience, which is entirely unmysterious once expressed in terms of physical modifications to the subject’s brain. That physicalism is able to explain this distinction so clearly is itself a strong argument in favour of the theory. It also explains our intuition that the Mary thought experiment yields ‘knowledge’, whilst enabling the physicalist to deny that this kind of knowledge can be gained in any way except than through experience.

This line of argument may be further strengthened by considering a parallel thought experiment in which Marty, an equally brilliant colour scientist of the future who has also never seen any colours first-hand, has been equipped with the latest in brain simulation technology. He has been implanted with an electronic device that enables him to enter all the physical data from his colour vision experiments along with the relevant physical theories in order to simulate the effect upon a normal human subject of, for example, seeing the colour red. The device then makes the same modifications to his own neural pathways as would occur in the simulated subject as a result of having had this experience. In the modified thought experiment, it is plausible that Marty would know what it’s like to see red without actually having been exposed to any red objects.3 Furthermore, he would be able to do so solely upon the basis of the physical facts, thus refuting the intuition that perceptual experience must be non-physical in nature. Although certain aspects of this modified thought experiment are open to question, none of these present a particular problem for the physicalist. Rather, the experiment can be taken to show that (i) we have conflicting intuitions about whether experience can be gained solely on the basis of the physical facts, and (ii) our lack of cognitive ability to translate objective (or third-person) propositional knowledge into subjective (i.e. first-person) experience does not entail that physicalism is false. If we were able to supplant our natural abilities with the artificial means to do this, as in the Marty thought experiment, then the mystery of knowing ‘what it’s like’ on the basis of the physical facts would simply go away.

III. Psychophysical Reductionism

Having established that physicalism is defensible in the face of the knowledge argument, I will now address the question of its necessity, starting with an outline of the precise type of physicalism that I wish to defend. The central claim of physicalism that mental events, states or properties have a purely physical basis. In order to allow for the numerous physiological differences between species and individuals — as well as the fact that mental terms such as ‘pleasure’ can encompass a diverse range of physical phenomena — such relationships must be expressed in the form of broad generalisations or ‘patterns’ (Lewis 1986). Each pattern of physical events corresponds to a particular class of mental events, giving it the character of a functional specification. Furthermore, this relationship — sometimes characterised as ‘psychophysical supervenience’ (Kim 1982) — must involve a genuine reduction of mental events to their physical bases if it is to explain their causal efficacy in terms of physical laws (Kim 1987: 282). Such a reduction would in principle enable psychological laws to be derived from other, more fundamental physical laws via a series of ‘bridge principles’ (op. cit: 268), although whether this is possible in practice is debatable due to the complexity of such a task. Conversely, the reverse process of deriving physical laws from psychological ones is not possible due to the disjunctive nature of the bridge principles, which may contain a potentially infinite number of terms (op. cit: 272). This effectively rules out simple type-type reductions like:

Pain = Stimulation of C-fibres (Kripke 1980: 98)

in favour of token-token reductions, e.g:

This pain = Stimulation of those C-fibres

based on broad ‘species-specific’ (op. cit: 274) identity relationships. I will refer to this combination of type-supervenience and token-identities as psychophysical reductionism, and it is the form of physicalism that I will argue for throughout the remainder of this essay.

As Kripke (op. cit: 97–110) famously argued, all non-tautologous statements of identity between names are both necessary and a posteriori. In other words, they hold across all possible worlds and can only be discovered by empirical investigation as opposed to conceptual analysis. Identity statements about natural kinds, such as ‘Water = H2O’ only give the illusion of contingency because we can imagine a world that is qualitatively and epistemically identical to our own where this identity relation does not hold (ibid.). Kripke argues that such a world, which contains a clear colourless liquid that has all the characteristics of water but that is not H2O, cannot be said to contain water. Rather, it contains a counterpart substance — call it pseudo-water — and consequently the necessity of identity is preserved. This poses a problem for the kind of psychophysical reductions we are considering here because in order to explain the apparent contingency of, for example, a particular tickling sensation being identical to its physical description, T, there must be a qualitatively and epistemically identical world in which T is not felt as tickling, but as pseudo-tickling. However, this cannot be the case as pseudo-tickling is, by definition, not tickling and so the existence of such a possible world would contradict the necessity of the identity relationship that were trying to defend. Kripke takes this to show that type-type, and by extension token-token, psychophysical identity statements are incoherent, but as Stoljar (2000) points out, this argument only rules out the possibility of a posteriori psychophysical reductions. If the relevant identity relationships are a priori, then Kripke’s argument does not apply.

On the face of it, a priori reductionism is an unattractive route for the physicalist to take. Arguing that the relevant psychophysical identities can, in principle at least, be deduced a priori seems implausible due to the obvious difficulty of such an exercise. If this were the case then, for example, ‘how pleasure feels’ would have to be some kind of complex tautology involving a neurophysiological description of the phenomena in question, along with various physical facts about the brain, central nervous system, and so on. However, here we must distinguish between the way that we learn something in practice and whether it is knowable a priori. For example, we generally learn to count and perform simple arithmetic on our fingers long before we are able to deduce the truth of number statements by a priori logical analysis.4 However, this cannot be taken to show that such statements are necessarily a posteriori. Similarly, it is conceivable that the identity relationships equating psychological descriptions to physical ones may in principle be derivable a priori but that lacking the knowledge or cognitive ability to do so, we are forced to discover them empirically through scientific investigation. Moreover, our current physical sciences lack the vocabulary and theoretical constructs to even describe such relationships due to the relative immaturity and incompleteness of the relevant disciplines (neurophysiology, cognitive science, and so on). Consequently, it may not be possible for us to know whether such identities hold, a priori or otherwise, at this point in time. Only once the relevant theories and abstractions have been developed to a point of constituting a complete physical theory of the mind would it be possible to demonstrate the a priori nature of psychophysical reductions involved. Until this point is reached there remains an ‘explanatory gap’ (Levine op. cit.) between the incomplete theories of neurophysiology on the one hand, and the felt quality of experience on the other. This gap can only be bridged by further scientific study and theorising about the precise relationship between physical and mental phenomena, along with a greatly improved understanding of the workings of the human brain, without which it is impossible to correctly frame, let alone answer, the question.

According to the necessity of identity, if physicalism is true then it must be necessarily true. But this is not to say that physicalism is true. So far, the best evidence we have for this comes from empirical data about the relationship between mind and brain. Functional magnetic resonance imaging (fMRI) shows a strong correlation between physical events in the brain and activity in the conscious mind (Edelman & Tononi 2000: 60). Furthermore, experiments involving direct stimulation of a conscious subject’s brain with an electrode during surgery have been found to evoke vivid mental images, sounds, memories, and so on (op. cit: 68). This suggests at least a causal, if not a full blown identity relationship between the two. Such techniques are rapidly approaching the point where we may be able to identify the precise physical phenomena corresponding to each individual thought or experience in the conscious mind, paving the way for the development of a completed science of consciousness. At such time it would be implausible to deny some sort of identity relationship between mental and physical events due to the overwhelming weight of evidence in favour of this hypothesis, much of which already supports the physicalist position. Although it would be premature to conclude that physicalism is true, the currently available evidence would certainly seem to be in its favour, and it is perhaps the simplest and most obvious explanation of the mind-brain relation. This suggests that physicalism is the strongest basis that we have for a future, more fully developed theory of the mind of the kind envisaged by Jackson’s knowledge argument.

V. Conclusion

Physicalism, in the form of psychophysical type-supervenience and the token-token identity relationships between mental and physical events, states or properties that this implies, is both plausible and potentially the best systematic explanation of the available empirical evidence. It is able to overcome the knowledge argument by positing two distinct and incompatible methods of representation, namely propositional knowledge, which is factive, and experience, which is not. It follows from the necessity of identity and the proposed a priori nature of psychophysical reductions that because they are  identity relations, they must also be necessary truths (cf. Kripke, op. cit: 108). We can therefore conclude that if physicalism is true at all, then it is necessarily so. Furthermore, the apparent contingency and a posteriori nature of the proposed identity statements may be explained by appeal to the incompleteness and inadequacy of current scientific theory and vocabulary, which precludes an accurate statement of the theory at this time. Only as this may be achieved will it be possible to say whether physicalism is correct or not. At present, this fact is simply unknowable and we are left in the position of the medieval alchemist who, in the absence of an adequate science of chemistry, is unable to say whether Water = H2O, XYZ, or any other chemical formulae, be it a priori or otherwise.

——————

Bibliography

Edelman, Gerald M & Tononi, Giulio 2000: Consciousness: How Matter Becomes Imagination. London: Penguin Books

Frege, Gottlob 1884: The Foundations of Arithmetic. In The Frege Reader. Beaney (ed.) 1997, pp. 84–129.

Jackson, Frank 1982: ‘Epiphenomenal Qualia’. Philosophical Quarterly vol. 42,
pp. 127–36.

————— 1999: ‘Mind and Illusion’. In There’s Something About Mary. Ludlow, Nagasawa, Stoljar (eds.), pp. 421–42.

Kim, Jaegwon 1982: ‘Psychophysical Supervenience’. In Supervenience & Mind, 1993, pp. 175–193.

————— 1987: ‘The Myth of Non-Reductive Physicalism’. In Supervenience & Mind, 1993, pp. 265–284.

Kripke, Saul 1980: Naming and Necessity. Oxford: Blackwell.

Levine, Joseph 1983: ‘Materialism and Qualia: the explanatory gap’, Pacific Philosophical Quarterly vol. 64, pp. 354–61.

Lewis, David 1986: On the Plurality of Worlds, Oxford: Blackwell.

Stoljar, Daniel 2000: ‘Physicalism and the Necessary A Posteriori’. The Journal of Philosophy vol. 97, no. 1, pp. 33–54.

Papineau, David 1998: ‘Mind the Gap’. In Philosophical Perspectives 12: Language, Mind and Ontology. Tomberlin (ed.), pp. 373–88.