Is passing the Turing Test a good test for whether something can think
Is passing the Turing Test a good test for whether something can think? Alan Turing proposed a test to in assist answering the question if machines can think. If the system passes the test, it is an indication of the system is an intelligent one. Intelligent not in terms of being smart but more so being a thinker. The machine is rational in a sense of having a capacity to form expressions, needs, and desires of mental states. Moreover, the machine is able to make sense of what is going on in conversations. I will first give a clarification on what the Turing Test consist of and then explain why one might believe that if something passes the test, it is considered a thinking system. Using John Searle’s thought experiment; I can suggest reasons as to why the Turing Test is not a good test. However, I strongly believe that the Chinese Room thought experiment is not suffice to have the same level of language capacity as the actual Turing machine. Therefore, being able to pass the Turing Test is adequate to declare something as a thinking system. Instead of answering if machines can think, Turing took into consideration if a programmed computer can play a certain game and at the same time be indistinguishable from a human being playing it. Turing proposes an imitation game where it involves a man, a woman, and an interrogator who may be of either sex in separated rooms from the man and the woman (Turing). The interrogator has to determine which of the two are the
man and the woman by communicating through teletype. It is then asked if the interrogator would decide incorrectly as often as it would have if played with a machine rather than a human being (Turing). The machines which are concerned in this game are logical computing machines that can do anything mechanical. The idea is to start with something empirical by going through different steps; each step is determined which in turn gives rise to a definite result. The machine has an infinite tape which is divided into squares that the machine writes and erase depending on certain types of ...
This is a preview of the whole essay
man and the woman by communicating through teletype. It is then asked if the interrogator would decide incorrectly as often as it would have if played with a machine rather than a human being (Turing). The machines which are concerned in this game are logical computing machines that can do anything mechanical. The idea is to start with something empirical by going through different steps; each step is determined which in turn gives rise to a definite result. The machine has an infinite tape which is divided into squares that the machine writes and erase depending on certain types of rules or instructions given (Clark 11). This illustrates that a machine can produce an adequate answer to any problem solving through simple computation. Turing proved that these machines can do anything which is universal. Turing proposed that a machine is considered an intelligent agent if the interrogator cannot tell if the machine is not a human in an open-ended conversation (Clark 21). Intelligent, for Turing, means having the capacity to perform linguistically in a way equivalent to a human being. A functional kind is multiple realizable, it can be realized or embodied in different material. Here, intelligence is investigated as a functional kind using functional analysis. Just like how mental representation is required for thinking, the mind manipulates its representations. Thinking is an inner activity that manipulates symbols basically in the way a digital computer dictate in its programming language. A computer is a formal system, it is made up of elements that operate according to formal rules and it is a system that has different level of meaning. The Turing Machine does not know what the symbols mean, it treats them as manipulations because of their shapes and the output can be semantically interpretable. However, some might believe that one cannot simply observe the behaviour on the outside of the machine without knowing what is going on in the inside. To further illustrate, I will now use Searle’s thought experiment, the Chinese Room, to suggest that there is a possibility of no thinking occurring inside a system that is able to pass the Turing Test. In the Chinese Room, Searle supposed that he was given a batch of Chinese writing in a room locked up and he has no understanding of the language Chinese at all (Searle). He then was given a second batch of Chinese writing along with rules in English, which he understands, to correlate the second batch with the original batch. This allows him to “correlate one set of formal symbols with another set of formal symbols, and all that ‘formal’ means here is that I can identify the symbols entirely by their shapes. (Searle).” He is then given a third batch with instructions, in English again, to correlate this batch with the first two batches. The rules allowed him to match certain Chinese symbols with certain shapes. Eventually he becomes so good at manipulating the Chinese symbols that the people external from the room believes that he is a native Chinese speaker. This shows that computer programs can be realized in multiple physical forms and not intrinsic. He instantiate the computer for manipulating the symbols but do not understand a word of Chinese. There is no understanding present anyway even if it looks intelligent or rational but on the inside contains no understanding. Searle believed that the manipulation of the rules will never give you an understanding, even if given another room with different symbols. One would still not understand the language or symbol. In addition, this thought experiment is supposed counter the Turing Test since seems like it can past it but has no understanding content. One might think that the Turing Test is misguided and is not a good test to pass for whether something can think. The system is not intelligent, he might be intelligent in following the rules but not in any Chinese thinking. Computation is defined syntactically using symbol manipulations. Searle believes that syntax alone is not sufficed for semantics because syntax has no casual power (Searle). One has these physical patterns in the system and these physical patterns, for instances can be changes in voltages, are talked about as ‘on or off’ but are not intrinsically ‘on or off’. One will treat them in syntax code so that one can interpret semantics, intrinsically there is nothing semantic. Rather than working with the idea that the computer itself is sufficient as a mind, the mind is intrinsically a computer. Nonetheless, to successfully pass the Turing Test, the machine must be able to make sense of what is going on to have the capacity to carryon the conversation. The Chinese Room experiment it is unable to have the exact language capacity, to keep up in a conversation, as the Turing Machine would have. If the man in the Chinese Room is asked to do a certain movement or to make a sound, he would be unable to because he lacks the knowledge of the Chinese language. Relying on syntactic alone does not make you a native Chinese speaker. The man inside must not only has an understanding of the words, but indeed the concepts. Hence, we cannot declare the man in the room a genuine Chinese speaker since he does not understand any of the content given to him. The Turing Machine would have such contact with the outside world for him to respond sufficiently in an ordinary conversation. This way, one can say that the machine does have something external, to us, which it maps onto. The Chinese Room objection is not sufficed to indicate that there is no thinking when the machine is manipulating the symbols. The man in the room does not understand the concepts in the Chinese Room; hence, not being able to be a real native speaker and would be unable to perform the same language capacity a Turing Machine would. Therefore, the Turing Test can be constituted as to whether a system is thinking. Bibliography Clark, Andy. Mindware An Introduction to the Philosophy of Cognitive Science. New York: Oxford University Press, 2001. Searle, John R. MINDS, BRAINS, AND PROGRAMS. 1980. <http://www.bbsonline.org/documents/a/00/00/04/84/index.html>. Turing, Alan. Computing machinery and intelligence. 1936. <http://www.abelard.org/turpap/turpap.php>.