What is the Turing Test, anyway?

The Turing Test, right?

Comes up in discussion fairly often, most recently for me in an OUP blog post. I have thought that I understood it, and it turns out that I am not sure I do.

I’m going to quote from Computing Machinery and Intelligence (Mind 49: 433-460), which I believe is the article that the test comes from.

[…] the “imitation game” […] is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart from the other two. The object of the game for the interrogator is to determine which of the other two is the man and which is the woman. […] It is A’s object in the game to try and cause C to make the wrong identification.”

There are a lot of details, but that’s the basic game: A is trying to convince C that he is a woman and that B is a man pretending to be a woman.

We now ask the question, “What will happen when a machine takes the part of A in this game?”

So. What I had thought, for many years, is that C is still attempting to determine which of the other two is the man and which is the woman, and is unaware that either might be a machine. And in that circumstance—C’s focus is A’s sex (or gender), not its machine nature—if it is not obvious to C that A is a machine, then A passes the test.

Reading it again, though, it reads as if we were to replace the word ‘man’ with ‘machine’—C’s goal is to determine which of the two is the machine and which is the woman. That’s significantly different, isn’t it?

In the rest of the paper, Mr. Turing did not develop the idea that C was trying to learn something about A that was not its machine nature. Instead, he talks more generally about A (or A’s constructor) trying to fool C, and the possibility of that happening.

So: have I been interpreting the test wrong all these years? Is Alan Turing’s writing sloppy on this point? Does it matter?

Tolerabimus quod tolerare debemus,
-Vardibidian.