Remy Pereira – 20/01/2020
According to Putnam (1967), type-identity theory is incompatible with multiple realisability. In other words, if each type of mental state is identical with exactly one type of physical state, then that same mental state cannot be instantiated by different physical states in different contexts. In this essay I will argue, contra Putnam, that a restricted version of type- identity theory can be compatible with multiple realisability. I begin by explaining type- identity theory, before explaining multiple realisability and why they are taken to be incompatible. I proceed to introduce the restricted version of type-identity theory, which says that mental states are only identical with one physical state per system, e.g. in human beings. I cash this out in terms of realiser-states and role-state, which then allows me to overcome Block’s objection and also reveals precisely the kind of claim restricted type- identity theorists are making. I conclude by showing how this restricted type-identity theorist is compatible with multiple realisability.
To explain type-identity theory, I must first explain the type/token distinction. A ‘type’ is a general abstract category. There can be ‘tokens’ of a type, which are particular concrete instances of something belonging to that category. For example, ‘mothers’ refers to a general category of people, i.e. women with at least one child, and so mothers is a type (Iwill henceforth italicise types). ‘My mother’ refers to a particular individual, i.e. the woman who gave birth to me, who belongs to the category of mothers. Thus, my mother is a token of the type mothers. There can be different tokens of the same type. For example, your mother is also a woman with at least one child. Thus, my mother and your mother are distinct tokens of the same type: mothers.
Pain, hunger, belief, etc. are all types of mental states, of which there can be tokens. For example, belief is a type of mental state, and my belief that it is Tuesday would be a token of this type. The type-identity theorist makes ‘type-identity claims’ about these mental states. These are claims that identify a type of mental state with a type of physical state, e.g. the type of mental state pain is identical with the type of physical state C-fibres firing in the brain. This entails that all tokens of pain are tokens of C-fibres firing in the brain. On this standard view of type-identity theory then, a type-identity claim like “ ‘consciousness is a process in the brain’ is a general or universal proposition applying to all states of consciousness whatever” (Place, 1956, p. 46). Type-identity theory states that a type- identity claim can be made for every type of mental state.
It has been said that type-identity theory is incompatible with the possibility that mental states are multiply realisable (Putnam, 1967, pp. 164-5). To say that mental states are multiply realisable is to say that the same mental state can be instantiated or implemented by different physical states. This is analogous to how the same software can be run ondifferent hardware, e.g. I can run Microsoft Office on my laptop and on the PC’s in the university library. For example, functionalism is the view that mental states can be analysed in terms of what they do, rather than their material constitution. So a functionalist might analyse pain as any state which “tends to be caused by bodily injury, to produce the belief that something is wrong with the body and the desire to be out of that state, to produce anxiety, and, in the absence of any stronger, conflicting desires, to cause wincing or moaning” (Levin, 2018). Let’s call this set of causes and effects the ‘pain role’. Anything that played the pain role would be a realiser of pain.
Suppose scientists discover that although C-fibres firing plays the pain role in humans, due to neurological differences across species, different brain states play the pain role in different organisms. For example, in dogs, D-fibres firing plays the pain role. In this scenario, pain is multiply realisable; it is realised by both D-fibres firing and C-fibres firing. Recall that, according to type-identity theory, each type of mental state is identical with one type of physical state. It seems this cannot be the case for pain if pain is multiple realisable. If tokens of pain can be identical with tokens of either C-fibres firing or D-fibres firing, yet tokens C-fibres firing are not identical with tokens of D-fibres firing, then pain cannot be identical with either type of physical state; identity is transitive after all. It is for this reason that type-identity theory is taken to be incompatible with the multiple realisability of mental states.
Following Kim (1972) (1992) and Lewis (1969), I believe that a version of type-identity theory can be proposed which is compatible with multiple realisability. To accomplish this, I must present a theory which (1) makes type-identity claims between types of mental states and types of physical states, and (2) allows for mental states to be multiply realisable. I will present what I call ‘restricted type-identity theory’, which satisfies both conditions. ‘Restricted type-identity theory’ does not make universal type-identity claims of the form: ‘the type of mental state pain = the type of physical state C-fibres firing’. Rather, they make type-identity claims restricted to a particular domain, e.g. a particular species, organism, system etc.
For this essay I will focus on the example of species. Returning to the earlier example of dogs and humans, the relativised type-identity theorist can identify “human pains with human brain state A, canine pains with canine brain state B, and so on” (Kim, 1972, p. 190). Or in our example, one could say that ‘the type of mental state dog pain = the type of physical state D-fibres firing’, and ‘the type of mental state human pain = the type of physical state C-fibres firing’, without thereby making any universal claim about pain per se.
A common objection to this move comes from Block who says that species-specific identity theory “sidesteps the main metaphysical question: ‘What is common to the pains of dogs and people (and all other species) in virtue of which they are pains?’” (Block, 1980, pp. 178- 9). I want to introduce the distinction between role-states and realiser-states. Thisdistinction allows me to provide an answer to Block’s question, while also making the relationship between the restricted type human pain and the general type pain more explicit, as well as clarifying the kind of identity claim that restricted type-identity theorists are making. To understand the distinction, suppose John stubs his toe, and his C-fibres fire. A functionalist would say that John is in pain if and only if he has a state that plays the pain role, and thus in this example he is in pain.
Nonetheless, this account leaves open “a further question to be answered, namely, what is the property of pain itself? Is it the higher-level relational property of being in some state or other that plays the ‘pain role’ in the theory, or the C-fiber stimulation that actually plays this role?” (Levin, 2018). To take the first option would be to view pain as a ‘role-state’. Arole-state is a higher-order state of being a state that plays a certain role. C-fibres firing and D-fibres firing are different physical states. However, both occupy the same higher-order state, namely the state of being states that play the pain role. If pain refers to the role-state, then pain is the higher-order state of being a state that plays the pain role. To take the second option would be to say that pain is the realiser-state. A realiser-state is a state that actually plays a certain role. The realiser-state of John’s pain is C-fibres firing. If pain refers to the realiser-state, then pain is whatever state plays the pain role.
Equipped with the distinction between role-states and realiser-states, I will now explain the relation between the restricted type human pain and the general type pain. This will also answer Block’s question. I want to say that the general type of mental state, pain (or pain per se) is a role-state. Namely, it is the higher-order state of being a state that plays the pain role. The species-specific type of mental state human pain and dog pain, on the other hand, are realiser-states. More specifically, human pain and dog pain are definite descriptions, where ‘human pain’ is a shorthand for ‘whatever state realises the pain role in humans’, and ‘dog pain’ is a shorthand for ‘whatever state realises the pain role in dogs’. In the example I have used in this essay, these descriptions uniquely pick out C-fibres firing and D-fibres firing respectively.
To answer Block’s question, then: what human pain and dog pain have in common in virtue of which they are pains is that they both refer to realisers of pain. Cashing out restricted type-identity theory in this way, by relying on the distinction between role-states and realiser-states also sheds light on the kind of type-identity claims that are being made in these cases. Specifically, they are contingent identity statements, identifying a type of mental state as defined as definite descriptions of the form ‘whatever realises role X in species Y’ with the type of physical state which satisfies this description. This is analogous to more familiar examples of contingent identity statements identifying definite descriptions with their referents, e.g. ‘Benjamin Franklin is the inventor of bifocals’ or ‘the author of Sherlock Holmes = Arthur Conan Doyle’.
I will now show how restricted type-identity theory simultaneously allows for multiple realisability while still being a type-identity theory. Analysing the general type of mental state pain as a role state allows for multiple realisability because, as demonstrated earlier, the same role can be performed by different types of state. In the example I have used here, the pain role is played by both C-fibres firing and D-fibres firing. Nonetheless, type-identity claims can still be made because pain is being realised by different states in different species. This means that, relative to each species, type-identity claims still hold. Thus, restricted type-identity theory satisfies the two desiderata outlined before: it (1) makes type-identity claims between types of mental states and types of physical states, and (2) allows for mental states to be multiply realisable.
To conclude, what has this essay achieved? First, contra Putnam, it has shown that by restricting type-identity theory to specific domains, e.g. species, the theory can be made compatible with multiple realisability. Second, by explaining this theory with reference to the distinction between role-states and realiser-states, this essay how provided some further insight into the kind of identity claim that is being made when one says ‘human pain = C-fibres firing’. It is also worth noting what this essay has not achieved. I do not intend this to be an argument for the truth of type-identity theory, nor even of restricted type-identity theory. This theory may be refuted by other philosophical arguments. Moreover, the truth of restricted type-identity theory depends on empirical facts, such as whether all pain in humans is in fact realised by one type of brain state. Nonetheless, I have aimed to defend type-identity theory from the challenge of multiple realisability, which is often raised against it. In this respect I believe I have succeeded.
Bibliography
Block, N., 1980. Readings in Philosophy of Psychology. s.l.:Harvard University Press.
Kim, J., 1972. Phenomenal properties, psycho-physical laws, and Identity Theory. The Monist, 56(2), pp. 177-192.
Kim, J., 1992. Multiple Realization and the Metaphysics of Reduction. Philosophy and Phenomenological Research, 52(1), pp. 1-26.
Levin, J., 2018. Functionalism. [Online] Available at: https://plato.stanford.edu/archives/fall2018/entries/functionalism/ [Accessed 15 November 2019].
Lewis, D., 1969. Review of ‘Art, Mind, and Religion’. The Journal of Philosophy, 66(1), pp. 22- 27.
Place, U. T., 1956. Is Consciousness a Brain Process. British Journal of Psychology, 47(1), pp. 44-50.
Putnam, H., 1967. Psychological Predicates. In: W. Capitan & D. D. Merrill, eds. Art, Mind, and Religion. s.l.:University of Pittsburgh Press, pp. 158-167.