My major linguistic interest lies in the problem of getting a computer to interact with us humans using language in exactly the way that we do. It is just not acceptable to any rational human being that they should be forced to adopt a jargon or baby-talk just to get a computer to do what it was invented for: making life easier. Under no circumstances should we accept a computer operating system or program that has the audacity to issue commands to a human.
I believe that in about fifty years' time it will be possible, to programme computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 per cent chance of making the right identification after five minutes of questioning. The original question, "Can machines think?" I believe to be too meaningless to deserve discussion. Nevertheless I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted. I believe further that no useful purpose is served by concealing these beliefs. The popular view that scientists proceed inexorably from well-established fact to well-established fact, never being influenced by any improved conjecture, is quite mistaken. Provided it is made clear which are proved facts and which are conjectures, no harm can result. Conjectures are of great importance since they suggest useful lines of research.Turing, A.M. (1950). Computing machinery and intelligence. Mind, 59, 433-460.
Could a machine think?
In a recent article, Thinking Machines and The Semantic Quagmire, I attempt to show that, at least in theory a computer could think. However, in practice their are many problems remaining to be solved. Many problems of ambiguity and strangeness in language remain as problems for any computer program. But not for the internet's smartest readers.
Ambiguity and strangeness in speech and writing, some examples.
There's a long weight by the weighbridge.
Which way does the weighbridge in Weybridge weigh whey?
I walked round the square to The Round Table and paid for a round.
A musical error was caused by the clash of symbols.
Shed like a garage.
The man who owned the green fell.
Judge: It is impossible to sentence with one bound to free him.
Defence counsel: Your Honour is bound to free him.
Prosecutor: I'm bound he is not bound.
Defendant: I'm bound to accept being bound over.
Attractor roles combine. [1]
The argument was about time.
Must is not nice.
Charlie shot up the mall in his car because it was Sunday.
The electric car goes a long way towards the green target.
Dealing with these language uses and their disambiguation, and reacting in a manner appropriate to the entire context in which they arise requires that a computer must analyse the syntactic, semantic and pragmatic components together. This is a tall order for any single computer program.
It is over fifty years since Alan Turing's famous conjectures about computer intelligence. I am confident that the creation of an intelligent robot is almost within our grasp. But if we fail to give it good laws2 to operate by, it is not the computer, but the human who will have failed the intelligence test.
[1] See e.g. The Roles Played by External Input and Synaptic Modulations in The Dynamics of Neural Systems. Free pdf.
[2] Isaac Asimov's Three Laws of Robotics, as a minimum.
Further reading:
For discussions and links relating to robotics, artificial intelligence, I recommend 3 Laws Unsafe.
Copyright notice:
Original tractor picture by elcajonfarms, Wikipedia used here under Creative Commons licence.
Comments