The New Ecology 25
It was in his 1950 paper on “Computing Machinery and Intelligence”
that Turing introduced his famous test (called by him “the imitation
game” and only later named after him) for determining whether or not
it was possible for an interrogator to mistake a machine for a person, offering a setup which originally involved one person asking questions and
two other participants, one human, another a machine, providing typed
answers to questions.
56
Turing himself considered a number of objections to the consideration of whether we could mistake a machine for a
human, some of which (such as religious objections) he dealt with fairly
cursorily, while others – notably the argument from the informality of
behaviour, later codified by Hubert Dreyfus as the assertion that human
reason and behaviour cannot be fully captured in rules
57
– were not so
quickly glossed over. While there were some half-hearted attempts to use
the test in the 1960s, most famously via Joseph Weizenbaum’s ELIZA
programme which appeared to fool some humans, it was not fully implemented until 1991 when Hugh Loebner offered a $100,000 pledge to the
first entrant to pass the test. Before this, John Searle had offered one of
the trenchant critiques of the Turing Test as a means of measuring true
intelligence: in his famous “Chinese Room” analogy, presented in his
1980 paper “Minds, Brains, and Programs”, Searle offered the analogy
of a human operator in a sealed room, which he later summarised more
concisely in 1999:
Imagine a native English speaker who knows no Chinese locked in
a room full of boxes of Chinese symbols (a data base) together with
a book of instructions for manipulating the symbols (the program).
Imagine that people outside the room send in other Chinese symbols which, unknown to the person in the room, are questions in
Chinese (the input). And imagine that by following the instructions
in the program the man in the room is able to pass out Chinese
symbols which are correct answers to the questions (the output).
The program enables the person in the room to pass the Turing Test
for understanding Chinese but he does not understand a word of
Chinese.
58
There have been responses to Searle’s analogy – whether along the
lines that just because the man does not understand Chinese does
not mean that comprehension has not taken place, to conceding that
this is an argument against strong AI but that the process can still
be useful for dealing with natural language processing and applied
AI. In any case, with regard to Turing’s own conception of machine
intelligence, the earlier mathematician would have probably not been
phased the slightest by Searle’s objections: his opening statement is
that he will consider the question “Can machines think?” and, within
a paragraph, he has dismissed that question as fatuous, replacing it
It was in his 1950 paper on “Computing Machinery and Intelligence”
that Turing introduced his famous test (called by him “the imitation
game” and only later named after him) for determining whether or not
it was possible for an interrogator to mistake a machine for a person, offering a setup which originally involved one person asking questions and
two other participants, one human, another a machine, providing typed
answers to questions.
56
Turing himself considered a number of objections to the consideration of whether we could mistake a machine for a
human, some of which (such as religious objections) he dealt with fairly
cursorily, while others – notably the argument from the informality of
behaviour, later codified by Hubert Dreyfus as the assertion that human
reason and behaviour cannot be fully captured in rules
57
– were not so
quickly glossed over. While there were some half-hearted attempts to use
the test in the 1960s, most famously via Joseph Weizenbaum’s ELIZA
programme which appeared to fool some humans, it was not fully implemented until 1991 when Hugh Loebner offered a $100,000 pledge to the
first entrant to pass the test. Before this, John Searle had offered one of
the trenchant critiques of the Turing Test as a means of measuring true
intelligence: in his famous “Chinese Room” analogy, presented in his
1980 paper “Minds, Brains, and Programs”, Searle offered the analogy
of a human operator in a sealed room, which he later summarised more
concisely in 1999:
Imagine a native English speaker who knows no Chinese locked in
a room full of boxes of Chinese symbols (a data base) together with
a book of instructions for manipulating the symbols (the program).
Imagine that people outside the room send in other Chinese symbols which, unknown to the person in the room, are questions in
Chinese (the input). And imagine that by following the instructions
in the program the man in the room is able to pass out Chinese
symbols which are correct answers to the questions (the output).
The program enables the person in the room to pass the Turing Test
for understanding Chinese but he does not understand a word of
Chinese.
58
There have been responses to Searle’s analogy – whether along the
lines that just because the man does not understand Chinese does
not mean that comprehension has not taken place, to conceding that
this is an argument against strong AI but that the process can still
be useful for dealing with natural language processing and applied
AI. In any case, with regard to Turing’s own conception of machine
intelligence, the earlier mathematician would have probably not been
phased the slightest by Searle’s objections: his opening statement is
that he will consider the question “Can machines think?” and, within
a paragraph, he has dismissed that question as fatuous, replacing it