This paper is one of the classic statements of functionalism, an the approach to the mind-body problem that Putnam presents as an alternative to both mind-brain identity and to behaviorism. It seems fair to see it as something like the approach via “semantic engines” that Haugeland describes.
§I. Putnam’s distinction between concept and property is intended to accomplish the same work as Place’s distinction between the ‘is’ of definition and the ‘is’ of composition. To see the connection, compare Place’s example that lightning is an electrical discharge with Putnam’s examples (such as “temperature is mean molecular kinetic energy,” p. 41 c. 1). Putnam’s comments on different philosophical approaches are quite similar to Armstrong’s comments in his first section (p. 31).
§II. Putnam next describes his alternative to the view that pains are brain states. You might think of the “machine table” he talks about as having the form shown in the example below. It has 9 cells, each containing two lists of probabilities. The cell specifying the machine’s behavior at a given moment is determined by the machine’s current state (the rows) and its sensory input (the columns). The list of probabilities then indicate, for each of a number of possible next states, the chance that the machine will enter that state next and, for each of a number of possible actions, the chance that the machine will now perform that action.
| sensory input | |||||||
| red | green | blue | |||||
|
current state | state 1 | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput |
| state 2 | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput | |
| state 3 | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput | probs of next state | probs of motor ouput | |
Putnam says this is a generalization of a Turing machine. If you want to compare it to Haugeland’s description of a Turing machine (p. 200), you can use the following translations:
| Putnam | Haugeland |
|---|---|
| state | execution unit |
| sensory input | token in the “in” bin |
| motor output | token in the “out” bin |
| current state | active execution unit |
| row of the table | rule of an execution unit |
A “Description” simply states that the system has a certain table. Putnam emphasizes that we know about the states only their function as described by the table; we do not know what they are “really.” That’s why they are “functional states.”
§III. Pay especial attention to the arguments in this section; they are the heart of the paper for our purposes. Putnam offers essentially one argument against the brain state hypothesis and one in favor of his own—what are they?
§IV. You can take the target of the arguments in this section to be the views of Gilbert Ryle (who both Place and Armstrong have in mind when they speak of behaviorism). The short last paragraph of the section is more important for Putnam’s argument than its length might indicate.
§V. Putnam here returns to the issues discussed in §I and argues that functionalism has the same value as reductions like “temperature is mean molecular kinetic energy” have in the physical sciences. Are the cases really the same methodologically? (The philosopher Ludwig Wittgenstein, who Place and Armstrong mention along with Ryle, held that a search for scientific explanation in the case of the mind was the result of philosophical confusion.)