How hard would it be to create a computer simulation of a washing machine washing clothes? To achieve a high level of accuracy, the simulation software would have to simulate the physical mechanics of the controls, the motor and other moving parts. The Navier-Stokes equation could be used to model the water, together with diffusion equations to model interaction between the water, the detergent and the dirt. And not forgetting the clothes. All of this might be achievable with a very powerful supercomputer.
Assume, for the sake of argument, that such a supercomputer is sitting on your desk, with some washing machine simulation software running on it. Could the software be made interactive? You would expect to be able to use your mouse to simulate the operation of the controls, putting the detergent in, and putting the clothes in. The simulation should show a display of the current state of the machine and its contents. At the end of the simulated washing cycle you could use your mouse to simulate pulling the washed clothes out of the machine and putting them into a simulated washing basket.
The question is, could you use this software to wash your clothes?
The obvious answer is: no you can't. The simulation is not really washing clothes, it is just simulating the washing of simulated clothes in a simulated washing machine using simulated detergent. If you have a pile of dirty clothes, there is no way to insert them into the computer and get them washed by the washing simulation software.
The boundary between real and simulated seems obvious. The boundary plays an important role in many philosophical discussions about the human mind. For example, if you simulate the conscious human mind on a computer, is that simulation itself a conscious human mind? By analogy with the washing machine example, the obvious answer is: no it isn't.
But, what if we had picked a different example?
Suppose that, instead of considering the simulation of a washing machine, we considered the simulation of an electronic calculator. To be really specific, let us suppose that it is a basic four-function calculator that does plus, minus, times, divides. The calculator has a button for each digit, one for each operation, as well as an "equals" button, a "clear" button and maybe an "on/off" button or switch. There is a display which shows the entered values and the answers as decimal numbers to some predetermined level of accuracy.
And now let us suppose that we wish to simulate this electronic calculator on a computer. We write a computer program which simulates the entry of values through the calculator keyboard. Just as was the case with the washing machine simulator, we wish the simulation to be interactive, so the simulation software displays a picture of the current state of the calculator, and allows us to simulate pressing buttons by pressing the simulated buttons with the mouse.
And then we ask the same question: can we use our simulation of an electronic calculator to do calculations? For example, if we can use the real calculator to work out 35 times 59, can we use the simulated calculator to work out 35 times 59? This time, the answer seems to be: yes we can. We can use the mouse to simulate pressing the keys "3", "5", "times", "5", "9" and "equals", and then we can look at the simulated display to see what the answer is. In fact our simulated calculator is very similar in its functionality to actual "Calculator" programs, such as those bundled with most modern personal computer operating systems.
So we have to conclude that a simulation of a calculator actually is a calculator. What is it about a calculator that makes it so different from a washing machine? The best answer seems to be that it is an information processing system, and that a simulation of any information processing system must necessarily process exactly the same information and get the same result if it is to be considered an accurate simulation.
Does this discussion resolve deep questions about the philosophy of the mind? Is a simulation of a conscious mind itself a conscious mind, or isn't it? Unfortunately, we have not resolved anything. If a "conscious mind" is a type of information processing system, then, yes, a simulation of a conscious mind is a conscious mind. But if a "conscious mind" is actually something else, then a simulation of a conscious mind might not be the same thing as a real conscious mind.
What the discussion does show is that arguments that use the difference between simulation and reality to prove that the human mind is not a mechanical information processing system are fallacious. They assume the thing they set out to prove, because the proposition that a simulation of something is different from the real something depends on the assumption that the "something" is not an information processing system. The standard name for this type of fallacy, where someone assumes the very thing they are claiming to prove, is "begging the question".
