why does turing say that the question "can machines think?" is too meaningless to deserve discussion

Answer 1

I am not aware that he said that, but I could believe it.

I think the fundamental issue is that we still define a machine as a deterministic mechanism, whereas we think of "thinking" beings as being essentially non-deterministic and having "free will". However, I would argue that it is easy to imagine a non-deterministic machine by allowing loops, recursion, etc., to the point where you cannot tell if the machine will stop, or what state it will be in. Especially if you allow pseudo-random functions, etc., to enter into the picture.

As for "free will", it is an essentially self-referential concept which takes as its premise the notion that we cannot determine the internal state of a thinking being. But the truth is, all of OUR mechanisms are at the lowest level either deterministic or stochastic (acting in a statistical or probabilistic fashion).

The question becomes one of pattern and structure. If I can duplicate the patterns produced by a "thinking mind" in a machine, can it be said that the machine thinks? To date, all arguments against it have been slowly taken down -- among them, the massive parallel computing power of the brain, its pattenr recognition capability, its ability to judge based on fuzzy sets, etc., etc. -- and duplicated this non-programmatic behavior on machines.

While it is still arrogant of us to think we have "cracked the code" on how thinking beings work, neither is it impossible to conceive of a thinking machine, if we are capable of making fair judgments and comparisons of the two.

Answer 2

Almost everyone has heard of the Turing Test (see http://www.thenewatlantis.com/archive/11/halpern.htm ) Basically, if you can't tell the difference between the responses of the computer and the responses of a human and you're willing to say the human thinks, then you should be willing to say the same about the computer.

But from the mathematical/logical point of view, what passes for human thinking is not that well defined, and so asking if a machine can do "whatever it is" would not be well-enough defined to make for meaningful discussion.

Answer 3

This is called "the Chinese Room", it's a thought experiment:

Imagine you are in a room by yourself, with a hole in the wall to your right and a hole in the wall to your left. You have a basket full of cards with Chinese symbols on them. Someone is handing you more of these cards through the wall on your left, and you have an instruction book that tells you which symbol from your basket you're supposed to hand back out through the right wall depending on which one you get handed through the left wall. If you didn't speak Chinese, how would you ever learn what those symbols meant, just by having them handed to you, and following your instruction book to choose which ones to hand back out the other side? The answer is that you couldn't, because if you don't know what either the inputs or the outputs mean, you can't deduce what other things mean. They will remain nothing more than symbols to you, no matter how long you do this.

This is what a computer does. It gets inputs, and based on how it is programmed (the instruction book), hands back the correct outputs. It doesn't "know" what any of the data means, because all the data is just a bunch of meaningless symbols to it, so to speak. It can't think for itself. However, unlike the human, this will be the case for the computer no matter what language the inputs and outputs are in. Going back to the Chinese room, if you were handed cards with things written in your native language on them, you'd be able to follow what was going on because you'd know the meaning of the cards. A computer doesn't have a native language- all languages are just meaningless symbols to a computer.

Answer 4

I think the root of the question is really a question of freewill and you could compare humans to machines in that we are only products of the materials we were made of and the information that we were given.

A machine can compute just like a human mind, using certain information to come to a conclusion.

The difference/or question is can machines feel?

Answer 5

Turing is thinking of the multi-function computer not of a device that is capable of thought. That requires totally different architecture. One which may or may not involve computation.

Answer 6

The crux of the question is what you think "thinking" is or entails. The scientific/philosophical community in general has established no consensus on what "thinking" basically is. Having no set definition for what constitutes thought, the question "can machines think?" is necessarily impossible to answer satisfactorily.

This link should help:
http://en.wikipedia.org/wiki/Philosophy_of_the_mind

Answer 7

Nope, machine can only do what they have been programmed to do, they do not learn things on their own, they cannot tell the difference between object unless it was programmed to do so

im talking about a computer machine lol not robots lol

Answer 8

Because they compute!

Thinking and computing are different functions. Computing in some terms can be Superior to thinking. Thinking invokes a different function though you can compute when you think.

Answer 9

They are automatons.