The Turing test

How can we tell whether something is capable of thinking?

Readings

Alan Turing (1912-1954) : famous British mathematician / logician, founder of computer science. He also helped cracked the German U-boat Enigma code in WWII. He was a homosexual and because of this was arrested in 1952 and put on trial. He resisted prison by submitting to humiliating treatment. He killed himself by eating an apple dipped in cyanide.

Turing proposed to consider the question "Can machines think?", and suggested that it should be replaced by the question of how machines will perform in an "imitation game".
Original version of the imitation game : judge talks to a man and woman through teletype and has to decide which is which. Turing asks : what if a machine takes the man’s place?
Passing the Turing test = the judge cannot do better than guessing. Sometimes the Turing test is conceived as a simple test where a computer tries to deceive a judge into thinking that it is human.
Other versions - A jury asks questions of a computer, and the computer passes the test when the majority believes that it is a human. Similar to the Loebner competition (but the most human-like machine wins).

Does the Turing test aim to provide a definition of thinking or intelligence?
It is sometimes suggested that the Turing test provides a (behavioural) definition of what thinking is, i.e. provides necessary and sufficient conditions for thinking.
But Turing said explicitly that he is not trying to give such a definition. Turing thought that the question of whether machines can think is "too meaningless to deserve discussion".
Even if we disregard what he said the definition would not be acceptable.

Necessary? What about machines or creatures (e.g. babies) that can think but are not able to communicate with a language, or perhaps too shy or paranoid to do so? What if the judge is a computer expert who can detect subtle hints?
Turing actually recognized an objection similar to the last point. He said that "we need not be troubled" as long as there are machines that can pass the test. This suggested that passing the test is not necessary for being able to think.
Also, it is not sufficient. Who can be a judge? What if a computer passes the test because of a stupid judge? (Can we specify the criteria without bringing mentioning thinking or other mental concepts?) Or what about the unintelligent Blockhead?

Proposal A - It is a practical test that avoids controversies about the nature of thinking.
- Compare : determine how old a person is by his or her face.
- Is it really practically useful? The first batch of thinking machines would probably fail the Turing test. If there are machines clever enough to pass the test, we’d probably do not need the test to be convinced that they can think anyway.
Proposal B - The test specifies one of the central aims of AI. Researchers should try to build machines that can pass the test.
- But isn't it better to study the human mind instead of building a program to trick people?