On Wed, Jul 3, 2013 at 8:41 PM, James Bowery <jabow...@gmail.com> wrote:
They are necessary so you can perform experiments. If you don't like an > operational definition then you need to say why. > It seems like it is possible to make progress on a question like this without requiring a formal definition. Perhaps a similar question to whether artificial intelligence is possible is whether computers can create art. A well-conceived experiment might involve a panel of judges who use their experience and intuition, perhaps along with some guidelines, to judge submissions of "art," who then try to decide whether the submissions were from from a person or from a computer. A formal definition might seek to spell out exactly what art is so that we can tell with great assurance whether a computer has produced it. But art is something that is hard to define, and many people produce very poor art. I remember reading about a contest where they had a person who served as a judge on one side of a terminal and either a computer or a person on the other, and the judge had to decide whether he or she was interacting with a computer. This seems like a test and one that can sort out whether artificial intelligence has been achieved to a certain extent (the computer fools most of the judges over a period of trials), without weighing down the challenge with the need to spell out what intelligence is. Eric