Searle's Argument Analysis

Introduction

The object of this essay is to depict as to whether or not artificial intelligence (A.I.) is possible from the use of arguments by Alan Turing, John Searle, and Jerry Fodor. To accomplish the task at hand; I shall firstly, describe the Turing Test and explain how it works, secondly, describe Functionalism and to detail on how it allows for future A.I. Thirdly, I will describe and explain Searle’s argument and example of the “Chinese room”, and finally I shall describe and explain a few replies to Searle’s “Chinese room” argument. However, due to the time constraint I will be unable to fully analyze Searle’s reply to all of his critiques, rather I will now state Searle’s counter to the objections with a simple point; they all are inadequate because they fail to come to

grips with the actual Chinese Room argument. In the end, this essay will demonstrate as to why not only A.I. is possible, but also that it is on its way and will arrive in the near future.
(i)
The Turing Test was created by Alan Turing in an attempt to answer his proposed question as to if a machine could pass as a thinking machine. To begin with attempting to solve the question, Turing proposes a game which he addresses as the “Imitation Game”. As stated by Turing, the game is played with a man (A), a woman (B), and an interrogator (C) who may be of either gender. However, the one who is the interrogator must stay in a separate room apart from the other two. According to Turing, the objective of the game is for the interrogator to determine which one of the two players is the man and which is the woman. Furthermore, Turing points out that the interrogator knows the other players as X and Y, and that the interrogator must make a decision at the end of the game by stating that either “X is A and Y is B” or “X is B and Y is A”. Moreover, the interrogator in the game is allowed to ask questions (which should be responded to by either written or typewritten answers ensuring that the tone of the voices do not assist the interrogator) to the man and woman in order to help identify the gender. In addition, the objective of A is to try to cause C to make the wrong identification, while B’s objective is to help C. For instance, Turing asserts that the interrogator could impose the question of X’s hair, which X could reply to as, “My hair is shingled, and the longest strands are about nine inches long”. As depicted by the example, the interrogator poses a question to assist him/herself in identifying the gender, however, the man makes in attempt in leading the interrogator to make the wrong identification by supplying an answer based upon general female appearances.
Since the task is of identifying the gender of the two players, it will be a difficult task for the interrogator, one that they may not succeed in. Therefore, Turing puts forward the question of “What will happen when a machine takes the part of A in this game?”. Additionally, Turing wonders if the interrogator will decide wrongly as often when the game is played with a machine than when the game is played with a man and woman. Turing also addresses the objection of whether a machine can be constructed to play the imitation game satisfactorily, which he believes we should not be troubled by. In other words, Turing is stating that being able to fool the interrogator often enough is a sufficient condition for deciding that the computer can think, even though perhaps not a necessary condition: there might be computers that do not succeed in the game, but are still thinking in some other way.
While the goal of the Turing Test is clearly defined and depended on Turing’s original question of whether machines can think, it should be noted that the game may have another goal in breaking gender inequality. As discussed in class, Turing was a homosexual that faced enormous amounts of discrimination, all after his vital contributions to computer science and his large part in breaking the German code in World War II. Furthermore, he was prosecuted in 1953 on the basis of his sexuality and given “medicine” from which he could have been cyanide poisoned which lead to his death. In class it was mentioned that since Turing faced this discrimination, he would understand the amount of pain and suffering those who are also discriminated against would face. Therefore, with the Turing Test there was a strong possibility in demonstrating that a man and woman are not different by depicting how one cannot even notice the difference between the genders if placed in a separate room where they cannot physically identify the gender.
(ii)
Functionalism has been a vital component to the idea of A.I., and as discussed later on, it was the key target of Searle’s objection to A.I. Thus, Functionalism is a theory about the nature of mental states, and how mental states are identified by what they do rather than what they are made up of. In other words, functionalism is the arrangement of the components involved that is essential to the overall structure (in this case the brain). In addition, according to functionalism mental states are functional states. In the case of computers they themselves are machines that implement functions. Thus, if functionalism is true, mental states can be similar to program states of computers, which would allow for a machine running the right computer program to have mental states. The machine could possess beliefs, memories, and pain, and could have a mind that can be seen as intelligent, and allowing for A.I. to exist.
Moreover, it is important that we note that computationalism states that roles or activities of the mind are computational, for instance, the mind manipulates symbols like flipping images in the brain (which requires calculations in mathematics), and therefore, the mind is sort of a computer. The relevance of computationalism lies upon the basis that according to functionalism and computationalism “the program is to the hardware as the mind is to the brain”. Hence, if minds can perform calculations (which are considered as a mental activity) and what the mind is made up of not mattering, then it is possible for machines to think.
(iii)
In John Searle’s “Is the Brain’s Mind a Computer Program”, he constructs his argument on the premise of functionalism to state that A.I.

is false. To accomplish this, Searle uses the example of “the Chinese Room” to challenge strong AI, and to object to Turing’s test. Searle begins by stating to imagine himself in a room with a box of Chinese characters which he could not understand, but in the room he had a book of instructions in English which he could understand. Searle then states that if there was a group Chinese speakers’ outside of the room passing him messages in Chinese, he would not understand, but could reply with the symbols in with the use of the instructions to form an appropriate response. Furthermore, Searle states that the Chinese speakers would think that the speakers were speaking to a Chinese speaker; however, realistically they were talking to a confused John Searle. Therefore, as Searle states, if a computer were to be placed in Searle’s position, the rule would become the “computer program”, and the basket of symbols the “data base”, it would prove that the machine would not understand Chinese, but only simulate that knowledge, which is not truly

intelligence.
Furthermore, Searle’s Chinese Room argument is particularly directed functionalism’s concept of strong AI, and not weak AI since weak AI makes no claim that machines can think. The claim from Strong A.I. according to Searle is that it claims that thinking is merely the manipulation of formal symbols. Additionally, as stated by Searle, this view can be summarized to “the mind is to the brain as the program is to the hardware”. Searle demonstrates his argument against strong A.I. in the Chinese room, however, there is a much deeper argument that Searle makes and is derived from the Chinese room. Searle makes a point in demonstrating that programs by themselves are not constitutive of nor sufficient for the mind, and to demonstrate, Searle depicts the disparity between programs and the mind. Searle argues that programs implemented by computers are just syntactical, computer operations are formal in that they respond only to the physical form of the strings of symbols, not the meaning of the symbols, whereas, minds have states with meanings unlike programs. Therefore, Searle constructs a four premise argument; the first of which, states that computer programs are formal or syntactic (syntax). The second premise states that human minds have mental contents (semantics). However, syntax is neither constitutive of, nor sufficient for semantic content (the third premise). Hence, the fourth premise asserts (and as stated previously) that programs by themselves are not constitutive of, nor sufficient for minds, or strong A.I is false.
(iv)
The Chinese room argument from Searle has seen its fair share of replies, two of which have rose to prominence over the others. The first is the Systems Reply and the reply points out that although Searle’s argument is logical against AI, it can be seen as sort of unfair due to the critique of one part of the Chinese Room system. In other words, Searle sets out three parts to the argument; the man, the box, and the rule book, however, Searle’s critique is concentrated on the fact that only the man does not understand, therefore, existing to be unfair since the man is one part of the whole. For instance, the functionalist would see it as illogical to state that the man does not comprehend the language for the reason that it would be painfully obvious that he would not understand for he is one part. Furthermore, the functionalist would reply by stating that while the man does not understand, the whole room does understand due to parts making the product according to functionalism.
The second famous respond to Searle is known as the “Robot Reply” and it first attacks semantics for the reason it is not clear. For instance, the definition of semantics is that it is the meaning of a word, however, how does one give a word a meaning? For illustration, consider that I do not know what the word cat means and am completely unfamiliar with the word. However, one day as I walk I notice a cat and some person near me points at the cat and tells me that it is a cat. Therefore, with this in mind, I will associate the word “cat” with the physical image of the cat I have just seen to form a meaning of the word cat. As this illustration depicts, a meaning of the word cat was able to come from the association of the word with the physical image of the cat. Hence, given how this may be one way one becomes to know what cats are, the Robot Reply proposes that if we were to put a computer into a robot’s body, attached with television cameras onto the robot head, and installed transducers connecting the TV messages to the computer; it could learn by seeing and doing likewise to a human. Furthermore, this type of computer in a robot body which is freed from the room could attach meanings to symbols and actually understand natural language, allowing for A.I. to exist. Thus, according to the Robot Reply if we were to give the symbols in the program a meaning, then Searle’s argument against A.I. would fail.
Furthermore, Searle lists a few other common objections to his illustration of the Chinese Room, which also happen to amaze him due to the number and vehemence of the defenders. The first of which is stating that the man in the Chinese room does indeed understand, albeit Searle explicitly stating that the man does not indeed understand. The explanation for this is asserted to be that despite the fact the man in the room does not understand Chinese, if it is possible after all, to understand something without knowing that one understands it. Similarly, the second objection states that while Searle does not understand Chinese, however, the unconscious subsystem in Searle does, since it is possible to have unconscious mental states. Therefore, there is no reason that Searle understanding Chinese is not “wholly” unconscious. Thirdly, the objection puts forward that the semantics Searle alludes to, do not exist anyway; there is only syntax. For instance, the objection states that it is a kind of prescientific illusion to suppose that there exist in the brain some mysterious “mental contents”, hence, only some sort of synaptic symbol manipulation exists in the brain (similarly on computers). The fourth objection asserts that Searle is not truly running the computer program, but only thinks he is. Since once the conscious agent of his goes through the steps of the program, it ceases to be a case of implementing a program at all. The last objection to Searle’s argument states that if the program were to simulate the operation of the brain of a Chinese speaker, then it would understand Chinese due to all the neuron level being identical to the brain of a Chinese individual. Thus, this type of system would most likely understand Chinese as well as a Chinese speaker.
Problem with Functionalism & Searle’s argument
Due to functionalism’s concept of organization of parts being essential to itself, it can led to absurd logics when this concept is applied. For instance, we know that according to functionalism it does not matter what the brain is made of, if they are organized properly and if it does the job, hence why A.I. can exist in terms of functionalism. However, imagine if we were able to gather all of the citizens of China, and specifically assigned them to accomplish tasks as a brain would, and this would lead to the remembrance of; for say one’s “grandmother”. Since the parts are organized properly, functionalists would have to accept this, therefore, stating that China is a brain, thus depicting the absurdness of functionalism.
Conclusion
In conclusion, it is clear that while Searle’s argument seems logical in a sense, it proves false as exemplified by the Robot Reply, and likewise from the Systems Reply (a reply which even Searle himself has agreed with). Despite the fact that the concept of functionalism can be proven to be absurd, and the fact that there has been no machine which has come close to passing the Turing Test as of yet, nonetheless it still should be concluded with confidence that not only A.I. is possible, but also that it is on its way, and will arrive in the near future.