Computers as models of the mind
In recent years, a new alternative to behaviorism has been suggested, which treats the mind neither as absurdly public, in the way behaviorism does, nor as completely private, in the way Cartesianism did.
It is, in other words, a halfway house between behaviorism and Cartesianism, and it is called functionalism. Its recent appeal derives from the development of the very computers with which we began. For one way of expressing what functionalism claims is to say that it is the view that having a mind, for a body, is like having a program, for a machine.A good way to start thinking about functionalist theories, however, is to look at similar theories of a simpler kind. Consider, then, what sort of theory you would need to give if you were trying to explain the workings not of something really complex, like a mind, but of something fairly simple and familiar, like a thermostat designed to keep the temperature above a certain level. What should a theory of such a thermostat say?
It should say, of course, that a thermostat is a device that turns a heater on and off in such a way as to keep the temperature above a certain level. Consider a thermostat that keeps the temperature above 60 degrees. An analysis of what something has to be like to do this job can be stated in a little theory of the thermostat.
A thermostat has to have three working parts. The first, which is the heat sensor, has to have two states: in one state the heat sensor is ON, in the other it is OFF. It should be ON when the external temperature is below 60 degrees and OFF when it is above. It doesn't matter how the heat sensor is made. If it is a bimetallic strip, then maybe whether it is ON or OFF will depend on how bent the strip is; if it is a balloon of gas that expands and contracts as the temperature changes, then ON will be below a certain volume, OFF will be above.
The second part is the switch, which needs to have two states also. It should go into the ON state if the heat sensor goes into its ON state and into its OFF state if the heat sensor goes OFF. Finally, we need the heat source, which should produce heat when the switch goes ON and stop producing heat when the switch goes OFF. (What I said about the heat sensor applies to the other parts too: it doesn't matter what they are made of as long as they do the job I have just described.)This explanation of the nature of a thermostat also shows what a functionalist theory is, for this little theory is a functionalist theory. And what makes it functionalist is that it has all of the following characteristics:
class=22 style='margin-left:0cm;text-indent:18.0pt'>It says how a thermostat functions by saying:a) what external events in the world produce changes inside the system—here, changes in temperature cause the sensor to go ON and OFF;
b) what internal events produce other internal events—here, changes from ON to OFF in the sensor produce changes from ON to OFF in the switch; and
c) what internal events lead to changes in the external world—here changes from OFF to ON in the switch lead to increased heat-output; changes from ON to OFF produced reduced heat-output.
Anything at all that meets these specifications functions as a thermostat, and anything that has parts that play these roles can be said to have a heat sensor, a switch, and a heat source of the appropriate kind. In other words, at the most general level, a functionalist theory says what the internal states of a system are by fixing how they interact with input, and with other internal states, to produce output.
What I mean by saying that the theory says what states are, can be explained by way of an example: our thermostat theory says what a heat sensor is by saying that ita) changes from ON to OFF (and back again) as the external temperature falls below (and rises above) 60 degrees, and
b) causes changes that lead to an increase in heat-output if it is ON, and to a decrease when it is OFF.
A heat sensor is thus characterized by its functional role, which is the way it functions in mediating between input and output in interaction with other internal states. And we can say, in general, that a functionalist theory says what a state is by saying how it functions in the internal working of a system.
We can apply this general model to computers. They have large numbers of internal, usually electronic, states. Programming a computer involves linking up these states to each other and to the outside of the machine so that when you put some input into the machine, the internal states change in certain predictable ways, and sometimes these changes lead it to produce some output. So, in a simple case, you put in a string of symbols like “2 + 2 =” at a terminal, and the machine's internal states change in such a way that it outputs “4” at a printer. We can now see why computer programs can be thought of as functionalist theories of the computer. For a computer program is just a way of specifying how the internal states of the computer will be changed by inputting signals from disk or tape or from a keyboard, and how those changes in internal state will lead to output from the computer.
From one point of view—the engineer's—all that is going on in a computer is a series of electronic changes.
From another—the pro- grammer's—the machine is adding 2 and 2 to make 4. People who are functionalists about the mind—which is what I shall mean by “functionalists” from now on—believe that there are similarly two ways of looking at the mind-brain. The neurophysiologist's way, which is like the engineer's, sees the brain in terms of electrical currents or biochemical reactions. The psychologist's way, which is like the programmer's, sees the mind in terms of beliefs, thoughts, desires, and other mental states and events. Yet just as there is only one computer, with two levels of description, so, the functionalist claims, there is only one mind-brain, with its two levels of description. In fact, just as we can say what electrical events in a computer correspond to its adding numbers, a functionalist can claim that we can find out which brain events correspond to which thoughts. Functionalism thus leads to monism. There is only one kind of thing, even though there are different levels of theory about it.Functionalism starts with an analogy between computers and minds. It doesn't say that computers have minds. But if we go carefully through the functionalist's arguments, we will see how you might end up holding that they could have minds, even if they don't yet.
1.5