Turing vs. Searle: What Makes Meaning?

From the time it was first presented, the Turing Test has been a significant part of the artificial intelligence debate. Whether it is used in a literal discussion of artificial intelligence or a more abstract discussion of what defines human intelligence, it is popularly cited by both proponents and opponents thanks the many supporting and disputing statements that have been made since. One such disputation, made by John R. Searle, is popularly known as the Chinese Room. Searle’s thought experiment aimed to dispute Turing’s test of intelligence by suggesting that no matter how realistic a computer’s answers are, that computer is incapable of understanding the meaning of those answers or the questions that elicit them. While Searle makes a compelling point, his disputation is very limited in scope and is not entirely convincing. While the Chinese Room does present a solid argument that a computer processor might be highly proficient at comprehending object and symbol interpretation without being intelligent, it fails to consider the computer as a whole and so fails to dispute Turing’s Test.

In Searle’s Chinese Room scenario, two Chinese women encounter a tent that appears to answer questions of any kind as realistically as a human might. They are effectively applying the Turing Test in that they consider a response that a human should be able to give and pose the situation to the tent in the form of written Chinese. The tent then responds in written Chinese, apparently with a full understanding of the question because its answer is relevant. To the two women, the tent appears to have human intelligence. Inside the tent, however, is where the disputation happens. There is an operator in the tent who takes the questions written by the women and, through the use of a book, essentially a program, identifies appropriate responses to send out. The operator does not speak Chinese and never knows the meaning of the symbols he is receiving or sending.

Because the operator does not understand his responses and is depending on a book to tell him what received symbols correspond to which response symbols, effectively acting out a program, he is not aware of his answers. Searle’s argument is that no matter how accurately symbols can be translated into other symbols if the thing doing the translating doesn’t attribute any meaning to them, it is not intelligent. In this way, if the operator alone were the subject of the artificial intelligence question (often raised in STEM classes), then Searle would be very difficult to disagree with. An operator is a simple machine, putting pieces together according to a program.

Searle suggests that the operator represents the limitation of a digital computer, basic symbol recognition, but his thought experiment presents a total situation that is much more complex than that. Assuming the tent is the machine rather than just the operator, the tent is quite compellingly intelligent. The operator represents only a part of the machine. The book, or program, is doing the actual thinking and if its program is sufficient to respond to any conceivable question in a real human way, then the book is the part that attributes meaning to symbols, not the operator.

In the Chinese Room, the operator is roughly as important to the thinking process as the ears and vocal cords are to human thought. His purpose is input and output, not actual processing. The focus of the Chinese Room should be to discredit the book as an intelligent processor if it is intended to dispute artificial intelligence. In order to accurately deliver responses, the book must have some kind of understanding of the meaning of the symbols it receives. In the case of the 2 + 2 question, there is no real meaning to be had, math is straightforward. But the women acknowledge that and move on to abstracts. In the case of the joke, the book understands what a joke is. While it might just have a list of “jokes” ready when that input comes up, is that so different from how a human responds when asked to tell a joke? If a person was not considered intelligent because they are unable to create an original joke on the spot, there aren’t many intelligent humans. As far as an observer would be concerned, the tent is intelligent because it understands questions and gives relevant answers.

The only question that remains is whether or not the machine genuinely understands the questions or answers. As far as the operator is concerned, there is no meaning, but since the book is the real keeper and giver of meaning in the tent, it is where the question of meaning must be applied. While the way that the book attributes and interprets meaning may seem contrived, how it came by its understanding is not relevant. The fact remains that it understands questions and gives meaningful answers. Beyond that, it is only an abstract, human-centric standard that could deny it meaning because it lacks an organic brain.

As far as the Turing Test is concerned, the Chinese Room thought the experiment is not convincing in its efforts to disprove artificial intelligence. It does disprove the intelligence of the operator, but the book is actually the center of meaning in the Chinese Room because it is what receives the symbols and responds to them in a meaningful way. While there might be other arguable limitations of this kind of artificial intelligence, it seems that the Chinese Room suggests a brilliantly programmed computer that resembles human intelligence so closely that there is no clear division. Only the concept of “meaning” that is being applied could come into question, but since the book understands the statements well enough to respond to them, then there is no clear lack of meaning. The form of meaning is alien to a human observer, but that does not discount it as intelligent.