Johnny Forgot to Cover All His Bases…

The issue we’re picking at this week is whether under certain circumstances a digital computer may perfectly replicate human behavior to the degree that it should necessarily be considered as having a mind or not. Generally speaking, there are two differing perspectives on the subject. The first of the two is that of ‘Weak AI’, which “is defined as the view that the computer plays the same role in studying cognition as it does in any other discipline. It is a useful device for simulating and therefore studying mental processes, but the programmed computer does not automatically guarantee the presence of mental states in the computer (MIT Encyclopedia, p 115).” The other view is that of ‘Strong AI’, which “is defined as the view that an appropriately programmed digital computer with the right inputs and outputs, one that satisfies the Turing test, would necessarily have a mind (Ibid).”

            In response to the Strong AI vs. Weak AI debate, American philosopher John Searle developed the Chinese Room Argument, which essentially “shows that passing the Turing Test is compatible with the computer having no mental sates at all (Rachels, p 95),” thereby refuting the notion that an implemented program itself constitutes having a mind. The thought experiment of the argument is structured as follows:

a.      A man who is strictly an English speaker is found locked in a room in which there are boxes with Chinese symbols/characters on them, which he clearly does not understand.

b.     The man is provided with an instruction manual for manipulating the Chinese symbols (the program), and his ‘job’ is to send out the input boxes but with different markings on them that he does not understand either, although his manual tells him what marks to make on which boxes before they’re sent out.

c.      Imagine that the external person who sends the input boxes into the man’s room is really having an extended conversation in Chinese where the inputs are questions and the outputs are answers, and because of the program, each response that the man in the room gives seems to make perfect sense to the person outside the room asking the questions.

d.     The external person would naturally think that he/she is having a meaningful conversation with another person who understands Chinese, yet clearly this thought experiment shows that this is not the case. The conclusion here is: “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 (MIT Encyclopedia, p 115.)”

In the Chinese Room Argument, Searle highlights the fact that the man in the room,

Manipulates symbols according to the instructions he has been given, but he has no way of attaching meaning to those symbols. This seems to show that the Turing Test is unsound. After all, the man in the room passes the Turing Test for understanding Chinese, yet he does not understand Chinese. (Rachels, p96)

Therefore, being that the man in the Chinese Room does not understand or speak Chinese (still passing the Turing test), this implies that even if a sufficiently capable digital computer replaced the human in the Chinese room, it could not possibly understand or speak Chinese since the mental faculties of the human greatly outweigh those of a digital computer (at least in our day in age). As Rachels puts it, “It is possible for a zombie to behave like a person but be mindless. Machines that pass the Turing Test may be such zombies (Rachels, p 95),” and as such, it is not appropriate to say that a machine has a mind simply because it can accurately simulate real meaningful conversation in Chinese.

            And I say, “Not bad Johnny, but you forgot about a few bits …” There are several objections to Searle’s argument. First, the man in the room represents only a piece of the whole. The man represents a central processing unit, but does not accurately represent the whole, which may be something or other that knows, understands, and speaks Chinese. Though this is a good objection, I think the best objection to this argument is as follows: the Chinese Room thought experiment is egocentric and defines intelligence, loosely speaking, in a very rigid manner, essentially taking for granted that the human manners of intelligence are the only yardstick by which we can assess any type of intelligence. In a sense, Searle makes a great point in comparison to human intellect; yet, it remains closed to any debate regarding alien intelligence, for instance, where were it to be found, it would ruffle the foundations of the argument which is solely based on a comparison of human vs. ‘possible’ machine or ‘other’ intelligence. It seems Johnny S. forgot to cover all his bases…

 

References

Rachels, James. Problems from Philosophy. Boston: McGraw-Hill, 2005.

“Chinese Room Argument” MIT Encyclopedia of Cognitive Science, Cambridge, MA: MIT Press, 1999.