John Searle’s Argument on Strong Artificial Intelligence

John Searle in his paper “Minds, Brain and Programs” presented the strong critics of the strong intelligence. First of all in the paper Searle differentiates between different types of artificial intelligence: weak AI, which is just a helping tool in study of the mind, and strong AI, which is considered to be appropriately designed computer able to perform cognitive operations itself. Searle conducted Chinese room experiment, the primary goal of which is to prove that machines cannot posses the states of conscious awareness, like perceiving, understanding or knowing (Searle, 1980).

According to Stevan Harnad, in his research Searle actually does not argue about artificial intelligence at all but in fact he attacked the main positions of computationalism, “a position (unlike ‘Strong AI’) that is actually held by many thinkers, and hence one worth refuting” (Harnad, 2001). Shortly, Chinese Room thought experiment can be described the following way. Searle places himself on the place of the computer. He supposed that he had to process a batch of Chinese characters with the help of the definite program and produce the output.

Searle is completely unfamiliar with Chinese; he can’t even differentiate Chinese characters from Japanese ones. He could only distinguish them by their shapes. Searle was able to process those symbols due to the fact that the rules were given in English. That enabled him to operate with the Chinese characters. In some time as the person learns the rules better, processing the Chinese words becomes easier and the answers are more correct.

So, people who ask questions in Chinese and receive answers are sure that the person knows the language as the answers are just undistinguishable from the answers of native speakers of Chinese. The conclusion is that obeying definite rules Searle can process Chinese questions and give correct answers to them without even knowing a word in Chinese. “Nobody just looking at my answers can tell that I don’t speak a word of Chinese”, Searle writes (Searle, 1980). The same thing is with computers. They are in fact in the same position as Searle.

Computers don’t have mind, they don’t think in Chinese, but they are manipulating with symbols just as Searle did. That’s why people might have the impression that computers can possess intelligence. However, this work of Searle was not a complete explanation on the problem of artificial intelligence. It was just a beginning and it raised the wave of critics and argument. On the one hand almost all researchers couldn’t but agree with the Searle’s statement that he was able to give correct answers to the questions without knowing the word in Chinese.

But still there were a number of people who considered that Searle’s experiment couldn’t be judged as a valid critics of the artificial intelligence. All the replies can be roughly divided into the following main groups (Cole, 2004). The first group argued the Searle’s experiment by identifying, who it is who speaks Chinese. The second group of critics researches the way how meaningless symbols can become meaningful. The third group of scholars believes that there is a need to redesign the Chinese room along the lines of a brain.

Finally the last group of scholars considers that there are numerous points which testify to the fact that Searle’s argument is completely misleading. So, as it was already mentioned the first argument was concerned with the mind source. This group of researchers was interested in the question where the mind was since the person in the room wasn’t speaking Chinese. The main issues under research were main ontological controversies of mind and body and simulation and reality (Cole, 2004; Hauser, 2005; Hearn, 2007). The group of the researchers attempting to answer this question fell into several categories.

The first category proposed systems reply (Searle, 1980; Cole, 2004, Hauser, 2005; Russel & Norvig, 2003; Dennett, 1991; Hearn, 2007, Crevier, 1993), which believes that since the person is not the one who possesses the knowledge of Chinese but the answers are still correct, it is the system, comprising the man, batch of words and rules for processing the words, which comprehends Chinese. The person in the room is just a part of this “understanding” system, which implies that the fact that the person does not understand and does not know the Chinese language is completely irrelevant.

However, Searle was able to answer this critical response saying that the man can be the whole system in case he memorizes all the rules for processing the Chinese words and will keep them in his mind. However, this won’t change the fact that he does not understand Chinese (Searle, 1980). The other point on which Searle argued this response was that critics are in fact missing the point as they on the one hand were trying to find the mind, but on the other hand point that it belongs to some “system”, which is a room.

But this doesn’t make sense as the room itself has nothing to do with the mind. It can be true only on the point when the critics explain this from the metaphysical point of view, which means that the mind is something that appears or “emerges” in the room and continues to exist there (Harnad, 2005; Searle, 1980; Crevier, 1993). The other response, which belongs to the group of “mind finders”, is virtual mind reply (Cole, 2004). This seems to be a more correct reply, which sticks to the idea that there is some Chinese-speaking mind in the room but it is virtual.

It was argued that computing machinery possesses the ability to “implement” another computer, which implies that any computer can simulate other machines step-by-step, performing the functions of both. Cole even argues that a program can be created, which in fact is able to implement two minds at once. So, despite the fact that there exists only one man in the room and one system, the number of “virtual minds” can be unlimited (Cole, 2004).

However, Searle’s response was that such a mind is nothing but a simulation by itself: “No one supposes that computer simulations of a five-alarm fire will burn the neighborhood down or that a computer simulation of a rainstorm will leave us all drenched”(Searle, 1980). This statement was argued by the supporter of the virtual mind idea, Nicholas Fearn, in the following way: “When we call up the pocket calculator function on a desktop computer, the image of a pocket calculator appears on the screen. We don’t complain that “it isn’t really a calculator”, because the physical attributes of the device do not matter” (Fearn, 2007).

Anyway, the following conclusion can be made: on the one hand these scholars were able to argue the Searle’s statement that “strong artificial intelligence” is false due to the fact that the man in the room doesn’t understand Chinese, which implies that nothing in the room understands Chinese (Cole, 2004). On the other hand the scholars still failed to prove the existence of the strong AI as they couldn’t prove that the system or virtual mind understands Chinese. Searle maintains that “the systems reply simply begs the question by insisting that system must understand Chinese” (Searle, 1980).

The other groups of scholars, who argue Searle’s work, were concerned with finding the meaning. Their replies are generally referred to as robot and semantics replies. The main concern of these scholars is to argue the Searle’s work at the point of intentionality and syntax-semantics controversy. For the person in the room Chinese characters are just meaningless “squiggles”, however, if the Chinese room can really comprehend Chinese words, there should be the source of the meaning. Thus, this group of scholars was trying to find the connection between the symbols and the items they symbolize.

According to the proposed replies to these questions, several categories could be differentiated. First one is robot reply (Searle, 1980; Cole, 2004; Hauser, 2006; Hearn, 2007), which states that if the program is placed in the robot instead of the room nobody would doubt that he understands what he’s doing due to the establishment of the “causal connection” between the symbols and things, which are represented by them. According to Hans Moravec “If we could graft a robot to a reasoning program, we wouldn’t need a person to provide the meaning anymore: it would come from the physical world” (in Crevier, 1993).

However, Searle argued this idea by stating that there is no difference who operates the words, as the person in the room is just following the rules without understanding what the words actually mean. Searle further says that “he doesn’t see what comes into the robots eyes” (Searle, 1980). The second group proposed derived meaning theory (Hauser, 2006; Cole, 2004), which there is a connection between the room and the world through Chinese speakers and programmers, which implies that the symbols the person works with are already meaningful in general, which does not necessarily mean that they should be meaningful to him.

However, Searle argues that symbols can only possess derived meaning, which depends on the conscious comprehension of Chinese speakers and programmers outside the room, which does not at all mean that the room by itself possesses the ability to understand by itself (Cole, 2004). The other semantic replies were concerned with the commonsense knowledge idea (Dennett, 2007), which states that the meaning of symbols could be derived from the background of the commonsense knowledge, which serves as a “context” providing meaning for the symbols.

Searle argument was based on the idea that although the background does exist, still it can’t be built in programs. So, it is obvious that Searle supports the viewpoint that there is no difference in the amount of knowledge written into the program and the connection of the later with the world. Still the person is the only one, who operates in the room and his actions is purely syntactic, which do not provide him with the meaning of the words, thus, the main Searle’s statement is that “syntax is insufficient for semantics”(Searle, 1984; Searle, 1989).

However, it should be admitted that there is some sense in the virtual mind theory, saying that even though the symbols mean nothing to Searle, they acquire their meaning from the virtual mind, which is connected with the outside worlds through Chinese speakers and programmers, which implies that it is irrelevant whether these symbols mean anything to Searle. The third group of scholars argued Searle’s work on the point that it the system needs to be redefined.

Thus, according to brain simulator reply (Searle, 1980; Cole, 2004; Hauser, 2006; Churchland & Churchland, 1990.) the program is sure to understand Chinese in case it is a simulation of the interaction of the neurons in the brain of a speaker of the Chinese language. Searle argues this reply saying that this type of simulation is unable to reproduce such basic features of the brain as its causal and intentional states, saying that “human mental phenomena are dependent on actual physical-chemical properties of actual human brains” (Searle, 1980). He further states that only brains can cause mind (Hauser, 2006).

According to the brain replacement scenario (Russell Norvig, 2003; Cole, 2004; Moravec, 1988; Kurzweil, 2005; Crevier, 1993,) the scholars maintain that in case one small computer is able to simulate the work of one individual neuron, this won’t cause that much difference to the system in general, however, in case all the neurons are replaced, we would create digital computer stimulating the brain. This means that if we support Searle’s point of view this will lead to the disappearance of the whole conscious awareness (Searle, 1992; Russell & Norvig, 2003).

Combination reply (Searle, 1980; Hauser, 2006) supported the idea that in case there is a robot created on the basis of brain simulation, which is linked to the world in the way that it has the causal power of the real brain, it is able to think. Connectionist reply (Cole, 2004 Hauser, 2006) has much in common with the brain simulator reply and believes that the real comprehension is possible in case there is a massively parallel connectionist architecture. So, basically these arguments can be divided into two main groups.

The first one believes that Searle is true in this Chinese room experiment, however, in case some changes are made in the room or the program, it can acquire mind and consciousness (Cole, 2004). The second group considers that redesigning should be made in order to see at which point Searle is wrong. Searle argues that machines still are unable to understand anything even if they are redesigned. The other argument is that in case there is a need of a robot body or a connectionist architecture are necessary, this would mean that we can’t speak any longer of strong AI (Searle, 1980; Harnad, 2001).

According to Searle “I thought the whole idea of strong AI was that we don’t need to know how the brain works to know how the mind works” (Searle, 1980) So, as far as we can see Searle’s argument of the strong artificial intelligence has its grounds. It is thoroughly based and well-considered. There was a lot of argument on his Chinese room experiment, however, hardly any critic was able to prove that Searle was completely wrong at some point.

References:

1. Churchland, Paul and Churchland, Patricia. (January 1990). Could a machine think?. Scientific American 262: 32-39.

2. Cole, David. (Fall 2004). The Chinese Room Argument, in Zalta, Edward N. , The Stanford Encyclopedia of Philosophy.

3. Crevier, Daniel. (1993), AI: The Tumultuous Search for Artificial Intelligence. NY: BasicBooks.

4. Dennett, Daniel. (1991). Consciousness Explained. The Penguin Press.

5. Fearn, Nicholas. (2007). The Latest Answers to the Oldest Questions: A Philosophical Adventure with the World’s Greatest Thinkers. New York: Grove Press.

6. Harnad, Stevan. (2001). What’s Wrong and Right About Searle’s Chinese Room Argument. in M. & Preston, J., Essays on Searle’s Chinese Room Argument, Oxford University Press.

7. Harnad, Stevan. (2005). Searle’s Chinese Room Argument, Encyclopedia of Philosophy, Macmillan.

8. Hauser, Larry. (1997). Searle’s Chinese Box: Debunking the Chinese Room Argument. Minds and Machines, 7: 199-226.

9. Hauser, Larry. (2006). Searle’s Chinese Room, Internet Encyclopedia of Philosophy.

10. Kurzweil, Ray. (2005). The Singularity is Near. Viking Press.

11. Moravec, Hans. (1988). Mind Children. Harvard University Press.

12. Russell, Stuart J. and Norvig, Peter. (2003). Artificial Intelligence: A Modern Approach (2nd ed.). Upper Saddle River. NJ: Prentice Hall.

13. Searle, John. (1980). Minds, Brains and Programs. Behavioral and Brain Sciences 3 (3): 417-457.

14. Searle, John. (1983). Can Computers Think? , in Chalmers, David, Philosophy of Mind: Classical and Contemporary Readings, Oxford.

15. Searle, John. (1984). Minds, Brains and Science: The 1984 Reith Lectures. Harvard University Press.

16. Searle, John. (January 1990). Is the Brain’s Mind a Computer Program? Scientific American 262: 26-31.

17. Searle, John. (1992). The Rediscovery of the Mind. Massachusetts: M. I. T. Press.