Hilary Putnam

Richard Hamming, an American mathematician, once held a speech in front of 200 scientists and asked: “What are the most important problems in your field? Are you working on one of them? Why not?” Nobody manages to work on important problems all the time, but if you catch yourself more often than not working on things that are not going to lead to anything important, make sure that this is what you really want. If your field is philosophy of mind, the central question to think about is how mind and body are related to each other.

Intuitively, mind is fundamentally different from matter, sensations are fundamentally different from brain processes. According to John Smart’s identity theory of mind, they are identical in the same sense that lightning is just an electric discharge. I am first going to discuss the objections to identity theory the American philosopher Hilary Putnam raises in his essay “The nature of mental states” and then present Putnam’s alternative, functionalism.

Objections to Identity Theory

Putnam does not dismiss Smart’s identity theory on a priori grounds. In fact, he talks about several possible objections to this theory, for instance the objection, that the reductive “P1 is P2″can only be true if P1 and P2 are within the same spatio-temporal region, and rejects them. Nonetheless he concludes that another hypothesis explains the relationship between mind and body in a way that is more plausible.

According to Putnam, identity theory requires us to accept a few assumptions about the real world that are not per se impossible, but quite unlikely.

The first assumption is that any organism that can be in pain possesses a brain that can be in a certain physical-chemical state and, when this organism is in pain, its brain is in this physical-chemical state. For identity theory to be true, the brain of any creature that cannot feel pain must never be in this state. The quantifier “any” includes not only mammals or currently-known organisms but also every physically possible creature.

Secondly, if the brain of any organism is in this physical-chemical state, the organism must necessarily be in pain.

These two assumptions required by identity theory need to be generalized if we consider that not only pain, but every psychological state is a brain state. Putnam states that we need to find only one psychological predicate that can be applied to two organisms that don’t have the same correlating brain state in order to falsify identity theory.

It appears to me that the second assumption is quite likely to be true since it is one of the basic assumptions of neuroscience and it does not look like it will be falsified any time soon.

In order to falsify identity theory on the grounds of the first assumption, one does not only need to find a generic psychological predicate that applies to two organisms with a different physical-chemical brain structure but one needs to show that these two organisms are in the very same mental state. I do not think that I ever felt exactly the same on two different occasions. How likely is it that two organisms with different brain structures ever feel the same?

How easy it is to see problems with identity theory depends on the level of abstraction that you use to compare two mental states. If you consider two mental states the same as soon as both can be named “hungry”, you run into problems with identity theory pretty soon.

Putnam’s proposal: Functionalism

Because Putnam sees these problems with identity theory and because of similar ones with behaviorism, he considers an alternative view of the body-mind problem: Sensations are not brain states in the physical-chemical sense of the word but functional states of whole organisms. The mind is the neural software running on the hardware of the brain, and, in theory, it can be implemented by different kinds of hardware.

In detail and applied to the concept of pain, machine-state functionalism makes four claims:

1. All organisms capable of feeling pain are probabilistic automata.

A Turing Machine is a very simple abstract model of a computer that can, in principle, implement every computation any other universal computers can implement. A Turing Machine is completely specified by a machine table that specifies how the machine behaves (e.g. “write a 1 on the current position of the tape and then move one element to the right”) for every possible combination of input symbol and current state.

A probabilistic automaton as described by Putnam is an indeterministic Turing Machine, that is, a Turing Machine with a Machine Table that does not specify exactly how to react on a combination of input symbol (”sensory input”) and current state (”motor output”) but gives transition probabilities.

It is certainly justified to view the universe and all the physical systems within it, including organisms, as behaving according to probability measures — in fact, no physical system has been shown not to behave according to a computable probability measure. This does not necessarily mean that no such systems exist or that mind is not in some way related to such a system, but it makes Putnam’s claim 1) more likely to be true than not.

2. Every organism capable of feeling pain possesses at least one Description of a certain kind.

A Description of a system is any true statement about the system that talks about the different states that the system possesses and how they are related to one another and to the Machine Table. The Machine Table itself is the Functional Organization of the system relative to that Description.

If an organism is a probabilistic automaton, it is self-evident that descriptions that talk about the different states the automaton possesses and their interrelations with motor inputs and sensory outputs are possible.

3. No organism capable of feeling pain possesses a decomposition into parts which separately possess Descriptions of the kind referred to in 2).

This assumption, introduced in order to rule out such systems as ant colonies or bee swarms as single pain-feelers seems quite ad-hoc. Without further evidence to one side or the other, I do not think that we can tell whether systems that have sensations may or may not be composed of simpler systems that themselves have sensations that are equivalent or comparable. Leaving this claim aside does not invalidate functionalism — all it does is making it even more vague than it is already.

4. For every such Description, there exists a subset of the sensory inputs such that an organism with that Description is in pain when and only when some of its sensory inputs are in the subset.

In other words, if an organism with a certain Functional Organization receives sensory inputs of a certain kind, usually from its ‘pain sensors’, it feels pain. If such an organism feels pain, it is because it receives the type of input required to make it feel pain. This claim may appear to be more strict than it actually is; all it demands is that whether an organism feels pain or not is not independent of the data the organism receives from its environment and that the relationship between ‘feeling pain’ and ‘having certain sensory inputs’ is not random but, depending on the Description of the organism, probabilistic. Although not provable, this claim is in accordance with both our intuitions and natural science and should thus be regarded as a sensible assertion.

The central claim of functionalism is that what characterizes mental states is not that they are physical in nature but their causal role — their relationship to other mental states and to bodily behavior. This makes functionalism incompatible with identity theory but still leaves a lot of room for other philosophies of mind. Both dualism and every form of physicalism that is weaker than identity theory can be seen as subtypes of functionalism. From this perspective, functionalism is quite a cautious theory of mind — one that does not say much about the actual relationship between mind and body.