In his 1950 article for Mind, Alan Turing (p.433) explores whether machines can think by using an imitation game as an analogy for his explorations. The imitation game is played by three people: a man, a woman, and an interrogator of either sex. The interrogator must establish, by asking a series of questions, which is the man and which is the woman. The interrogator cannot see the respondents or hear their replies as the questions and answers are in written form. In Turing’s game the man or the woman is replaced by a computer. Turing (p.434) then asks the question whether the interrogator would be fooled as often in the game if one of the respondents was a computer, and infers that the computer would be able to respond as succinctly and comprehensively as any human.
The game Turing envisages will use only digital computers (p.436) as it must follow a set of rules just as a human computer should and, like Turing’s understanding of human computers, the digital computer has a store which is the memory, an executive unit which carries out various individual operations, and a control which is a ‘book of rules’ (p.437). The digital computer will also be what Turing terms a ‘discrete state machine’ which only has a finite number of possibilities so that the principle of uncertainty cannot occur (p.440). This allows all future states of the machine to be predicted by the input signals of the machines initial state. Turing states that such a machine with an adequate, suitably increasing storage capacity and appropriately connected programming should be able to convince the interrogator that it is human (p.442).
One of the objections to Turing’s assertion that computers can think is a mathematical objection (p.444). Gödel’s theorem infers that such a powerful logical system must have an infinite capacity for it to be successful otherwise there will be certain things that a finite capacity system such as the discrete state machine will not be able to accomplish. In the imitation game there may be questions that the finite machine may not be able to answer even given adequate time. Although Turing answers this objection by stating that it is not proven that humans have such an infinite intellectual capacity either, and that, just as there could be humans that are cleverer than machines, there might also be machines that can be cleverer than humans (p.445). A further response to this reply could be that even infant humans have a capacity for such things as humour, which might be considered open-ended possibilities to the human intellect (Hurley et al. 2011, p.5) such that Turing’s closed state machine could not replicate. If such a contention is found to be the case then Turing’s assertion that machine’s can think may turn out to be false, in that an interrogator will be able to identify the machine by it lacking the capacity to answer questions laced with implied humour.
- 1. Hurley M., Dennett D. C., Adams R. B., (2011) Inside Jokes: Using Humour to Reverse-Engineer the Mind, Massachusetts Institute of Technology
- 2. Turing, A.M. (1950). Computing machinery and intelligence. Mind, 59, 433-460.