The Turing Test was designed by a man named Alan Turing in 1950. It was initially called the “imitation game.” Originally, the test was designed to differentiate between males and females. It was played with three peoplea man, a woman, and an interrogator. The interrogator would go into a separate room and try to determine who was the man and who was a woman by asking various questions such as “How long is your hair?” or “Do you have an Adam’s apple?” Based on the answers to the participants’ replies, the interrogator would decide who was the man and who was the woman. Often times this wasn’t easy since the participants would be allowed to lie in order to try to throw the interrogator off.
Turing went a step further with the “imitation game” idea by incorporating computers into it. He believed that in approximately fifty years (today’s time) computers would be programmed to acquire abilities rivaling those of human intelligence. As part of his argument, Turing put forth the proposal in which a human being and a computer would be interrogated through textual messages by an interrogator who didn’t know which was which. Ideally, if the interrogator were unable to distinguish them by questioning, then it would be unfair not to call the computer “intelligent.” Passing this test was considered regularly and reliably fooling an interrogator at least 50% of the time.
Turing and Godwin both believed that anything that could pass the Turing Test was genuinely a thinking, intelligent being. In particular, they felt that passing the test illustrated that the computer had the ability to interact with humans by sensibly “talking” about topics that humans talked about. Also, passing the test according to Godwin reflected that the computer was able to understand how humans thought and interacted.
Despite Turing and Godwin’s obstinate belief that computers could think, many believed that this was not the case. In the book Can Animals and Machines Be Persons?, Goodman set out an objection called the “Chinese-box” argument. Essentially, a man (who had no knowledge of Chinese) would be placed in a box and textual messages similar to those found in the Turing Test would be displayed on the screen in either English or Chinese. Then, man inside the machine would give the appropriate responses in Chinese. Despite his lack of knowledge of Chinese, the man would be able to give responses by using a large “Chinese Turing Test Crib Book.” Ideally, the person inputting the questions would be unable to distinguish that man’s Chinese from a native speaker’s. That argument was extremely damaging.
By describing the Chinese-box argument, Goodman was pointing out that externally it would seem that the man in the box understood both English and Chinese when in reality he wasn’t “thinking in Chinese” the way he did in English he was really just translating the symbols he saw into different symbols. Fundamentally, computers did the same thing. They would translate their binary code into symbols which we could understand. To do so, they would use rules analagous to those found in the “Chinese Turing Test Crib Book.” Overall, the Chinese-box argument supported the idea that a computer could cleverly imitate thinking and understanding but could never be a real, literal “thinker” or “person.”