sign in sign up
<<
Home , Intelligence

Can machines think?

Even though Descartes’ argument that the is distinct from the body fails, the general point that Descartes wants to make remains a challenge. It is that if the mind has essential properties that are not shared by the body, then it must be distinct from the body. The he uses is:

If X has at least one which Y cannot have, then X and Y are not identical.

In fact, if X has a property which nothing else has, then that property is the essential property of X. That is it is the property which makes X what it is and makes X distinct from all the other things.

Thus when Socrates asks Euthyphro to give a single definite characteristic that makes piety what it is, he is asking Euthyphro to tell him the essential property of piety. An example is ’s of man ( ) which is: Man is a rational animal. What he means is that are a kind of animals, but they share a characteristic which marks them off from all the other animals which is the characteristic of rational or having . No other animals have this property, thus being rational is the essential property of human beings. (Aristotle has been proved wrong, but what really matters is that the of an essential property is illuminated by his definition).

Descartes wants to point out that the property of indubitability is the essential property of the mind. In other places, he lists other essential properties that the mind has but the body does not have. The issue is whether these so-called essential properties of the mind can be explained in terms of physicality. If they can, then they are not essential properties of the mind at all, but are properties that some physical things also have; and thus the mind is not distinct from the body.

In this lecture, I will focus upon the claim that thinking is an essential property of the mind and upon an attempt to invalidate the claim. The reason why I will do this is because Descartes gives prominence to it in his argument and this leads people who argue against his thesis to do the same. But thinking is only one among many, the question thus must be asked why Descartes should focus upon it.

It is clear that Descartes that thinking is something unique to the mind, and that the thinking mind is unique to humans. He believes that animals and machines built to be like humans do not have reason. They therefore do not have a mind. His assumptions can be put in the following way: For a thing to have a mind it must have reason. For a thing to have reason it must be able to think.

His argument is that there are two tests whether these non-humans have reason or not and they fail the test. The first test is that anything that has reason must be able to use . Both animals and machines fail the test. Animals can only imitate words and sentences uttered by us; machines can only put words into particular sequences as programed by us. Both cannot put those words into various sentences to express different things. To be able to do that requires the use of reason and thus failure to do so shows that animals and machines do not have reason. The second test is a generalization of the first. Animals and machines might be able to perform certain tasks far better than us, but they cannot invent new tasks apart from those determined by or by human artifice. This shows that they do not have reason. In the case of machines, to make a machine perform a different task than what it is designed to do requires putting in another of parts to perform another task. To keep on adding parts so they can perform like humans is improbable. Humans can use reason to think of ways to do various tasks. In short, Descartes assumes that reason is inventive; in other words, reason is free to go beyond instinct. To prove that machines can think, one has to prove that a machine can pass the tests. (Can you think of other criterions different from the ones Descartes uses? Tell me what you think.)

Not only Descartes, but we also often think that human beings are privileged creatures in the universe. But why do we think that thinking is unique to humans? Human beings are unique because they have . To have a free will means to have a of action; and to be able to choose requires thinking. Animals only act according to their instincts. Humans also have instincts but we often act against them; and that proves that we do have free will. On the other hand, religious beliefs lead us to think that we have a mind distinct from the body, something that might survive death, something which we call ‘the ’ which will go either to heaven or hell after death; and since we believe that reasoning and thinking are what only we, among all animals, are capable of, we tend to link thinking and the mind together in such a way that the of the mind is thinking.

Morever, our body is composed purely of i.e. it is a physical thing. When we think of the mind as distinct from the body, we are taking it to be an immaterial thing, a soul which survives the death of the body. When we link thinking with the mind, we then link thinking with an immaterial thing. We do say that the brain does the thinking, but when we regard the mind as distinct from the body we are assuming that in fact the mind does the thinking and uses the brain as a vehicle. We thus believe that purely material things cannot think.

Scientists nowadays have proofs that some animals can think. Therefore dualists (people who believe the mind is distinct from the body) have to admit that some animals have and that human beings are not unique in that regard. Nevertheless, dualists can consistently continue to think that the mind is distinct from the body.

The above picture the dualists have will be shattered if somebody can prove that some material things can think. For if some material things can think, then it is a simpler explanation to say that thinking is done by the brain and there is no immaterial mind behind the thinking. To conceive of humans as composing of two entirely different elements provides a very complicated explanation which faces a serious problem. Material things can be causes and effects of one another because they have mass, but how can an immaterial thing affect material things when it doesn’t have a mass. If we have a reason to believe that some material things can think, then this solves the problem. We don’t have to have two completely different things the connection between which is mysterious. The problem with dualism is that it believes material things can’t think, thus it has to posit another entity to explain thinking and in the process creates a very difficult problem.

The idea that some material things can think was suggested by Alan Turing (1912- 1954), who was a famous British scientist. His work with computers made him believe that since computers were getting smarter and were doing intelligent work as good as or even better than humans, one day computers would be made to function at the level of . How do we know when that day will come? Turing answered by first restating the question “Can machines think?” because he this question was not precise and might generate prejudices. He suggested the idea of ‘the game’ as a way to rephrase the question and at the same as a model for the test that would decide whether machines can think or not. This test is nowadays called “the ”. I will let Turing explain in his own words what the imitation game is and how the question should be rephrased.

I propose to consider the question, "Can machines think?" This should begin with of the meaning of the terms "machine" and "think." The definitions might be framed so as to reflect so far as possible the normal use of the words, but this attitude is dangerous, If the meaning of the words "machine" and "think" are to be found by examining how they are commonly used it is difficult to escape the conclusion that the meaning and the answer to the question, "Can machines think?" is to be sought in a statistical survey such as a Gallup poll. But this is absurd. Instead of attempting such a definition I shall replace the question by another, which is closely related to it and is expressed in relatively unambiguous words.

The new form of the problem can be described in terms of a game which we call the 'imitation game." It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart front the other two. The of the game for the interrogator is to determine which of the other two is the man and which is the woman. He knows them by labels X and Y, and at the end of the game he says either "X is A and Y is B" or "X is B and Y is A." The interrogator is allowed to put questions to A and B thus:

C: Will X please tell me the length of his or her hair?

Now suppose X is actually A, then A must answer. It is A's object in the game to try and cause C to make the wrong identification. His answer might therefore be:

"My hair is shingled, and the longest strands are about nine inches long."

In order that tones of voice may not help the interrogator the answers should be written, or better still, typewritten. The arrangement is to have a teleprinter communicating between the two rooms. Alternatively the question and answers can be repeated by an intermediary. The object of the game for the third player (B) is to help the interrogator. The best strategy for her is probably to give truthful answers. She can add such things as "I am the woman, don't listen to him!" to her answers, but it will avail nothing as the man can make similar remarks.

We now ask the question, "What will happen when a machine takes the part of A in this game?" Will the interrogator decide wrongly as often when the game is played like this as he does when the game is played between a man and a woman? These questions replace our original, "Can machines think?"

He then went on to describe the test. Since in his days there were no personal computers, we will adapt his test by using personal computers as the machines in the test. Two computers are put in two closed rooms and are connected to the third computer outside. One computer is controlled by a person, the other is not controlled by anybody but has a software that can communicate like a human being. The third computer is operated by a person whose task is to talk with the two computers in the rooms. He of course doesn’t know which of the two computers in the rooms is operated by a person. He can have a conversation with the two computers on any matters he likes, he can ask questions to try to determine which computer is not controlled by human beings. He can have as many sessions as he wishes. In the end he is asked which one is controlled by a person and which one is not. If he gives the wrong answer most of the or just says he can’t tell, then the uncontrolled computer can think. In other words, that computer has .

One might question whether this test is a valid method. Think about this: when we want to determine whether a person is smart or not, we have a conversation with him and we judge from what he says how intelligent he is. That is, we don’t try to look inside his head whether he’s smart or not. We judge from his behavior and from his conversations with people.

It is also futile to object to the test on the ground that the software in the computer is created by human beings. The point is it can think, it does not matter how this capacity comes about. People who believe in God also believe that God created everything including humans. So God made us intelligent. Would these people say that we are not really intelligent because our capacity to think was created by God?

We have to be clear about one thing. Nowadays some computers can play with world champions and win. Computers can calculate faster and more accurate than humans. But these computers will fail the Turing test because when we talk to them about the weather or about the Great Flood of 2011, they will not be able to answer us. To build a computer that passes the Turing test, scientists have to program it so it has multiple dimensions of thinking and has in various subjects in such a way that it can communicate with people as if it were a human being. We might think this is a gargantuan task, but Turing suggested that it could be done by creating a child-like computer which can learn the way a human child learns and which finally develops into a computer with an mind.

When will that day come? Alan Turing predicted that by the end of the 20th century we would have a computer that passes the test. We are now in the 21st century and no computers have passed the test. Nevertheless, scientists think we are getting closer and closer since computers are more and more powerful, and many of them think that day is near. When that day comes, dualism will be refuted. We will have to admit that humans are purely physical things and death is the ultimate end. Too bad we won’t get to heaven, but it’s a relief to know we won’t go to hell either.

In June 2014, it was claimed that a computer has finally passed the Turing test. Read the news from the following site: http://www.theguardian.com/technology/2014/jun/08/super-computer-simulates-13- year-old-boy-passes-turing-test

If Turing’s expectation is fulfilled some day soon, then we can claim that there is a machine that passes Descartes’ two tests. Even though Turing does not mention Descartes in his article, his thought points out that a machine can pass these tests. If a machine can pass the Turing test which consists in linguistic , then it is obvious that that machine can use language. Again we have to throw away our prejudice and think about how we test whether a person can use a language or not. We talk to him in that language and if he can respond to us in the same language, then we judge him to be able to use that language. The same criterion must be applied to the machine, otherwise we will be guilty of using a double standard. Therefore a machine that passes the Turing test must be judged as capable of using a language and thus it passes Descartes’ first criterion.

As for Descartes’ second criterion which specifies that a machine must be able to do a variety of tasks, Turing has an answer to this in his idea of the machine. He proposes that with the development of the digital computer towards more storage (more ) and faster processor, we can design one machine that can perform a variety of tasks without having to build numerous machines, each doing a specific task. For example, if we want to build one machine that can perform the many tasks that a human being can, then we build it in the form of a robot that resembles a human body, that is, with arms and legs that can be ordered to do things that humans do. If we want it to perform one task, we write a program ordering it to perform that task. If we want it to perform another task, we just write another program. Finally there will be numerous programs including a huge amount of data and all these require a very large storage. For the robot to perform like humans it has to process these programs quickly. In the future when a computer is equipped with a very fast processor and massive storage capacity, that robot then can be programmed to perform as many tasks as an ordinary human being can. Turing calls this kind of machine the universal machine and we can see that this machine can pass Descartes’ second criterion. In conclusion, given Descartes’ benchmarks, if Alan Turing’s vision is correct, then machines can think.

In fact, for a computer to pass the Turing test, it has to be able to use language like an ordinary human being does; that requires also a large memory to store a lot of data equivalent to ordinary human and , and a fast processor in order to respond as quickly as an average human does. In other words, only a universal machine can pass the Turing test. Since a universal machine can perform a variety of tasks, the computer that passes the Turing test passes both Descartes’ criterion.

The child machine

For a computer to be able to pass the Turing test, it has to communicate with an adult person. Therefore for it to convincingly fool the interrogator into thinking it is a human, it has to have the and intelligence that an adult has. The task of the scientist then is to program the computer in such a way that it has the average adult intelligence. We know that for an average adult, to have such intelligence requires years of of gathering data and of learning from his interaction with other people and his environment. It is then a gargantuan task for the scientist to program the computer to imitate such intelligence. Turing suggests an idea to overcome this difficulty. Instead of building a computer with the average adult intelligence, the scientist should build a child machine. That is he should build a computer with the child intelligence and design it in such a way that it can learn and develop its intelligence to be on a par with the average adult intelligence. Since in his article, Turing presents this idea very clearly. Therefore I will let him speak for himself:

In the process of trying to imitate an adult human mind we are bound to think a good deal about the process which has brought it to the state that it is in. We may notice three components.

(a) The initial state of the mind, say at birth,

(b) The education to which it has been subjected,

(c) Other experience, not to be described as education, to which it has been subjected.

Instead of trying to produce a programme to simulate the adult mind, why not rather try to produce one which simulates the child's? If this were then subjected to an appropriate course of education one would obtain the adult brain. Presumably the child brain is something like a notebook as one buys it from the stationer's. Rather little mechanism, and lots of blank sheets. (Mechanism and writing are from our point of view almost synonymous.) Our hope is that there is so little mechanism in the child brain that something like it can be easily programmed. The amount of work in the education we can assume, as a first approximation, to be much the same as for the human child.

We have thus divided our problem into two parts: the child programme and the education process. These two remain very closely connected. We cannot expect to find a good child machine at the first attempt. One must experiment with teaching one such machine and see how well it learns. One can then try another and see if it is better or worse. There is an obvious connection between this process and evolution, by the identifications

Structure of the child machine = hereditary material Changes of the child machine = mutation, Natural selection = judgment of the experimenter

One may hope, however, that this process will be more expeditious than evolution. The survival of the fittest is a slow method for measuring advantages. The experimenter, by the exercise of intelligence, should he able to speed it up. Equally important is the fact that he is not restricted to random mutations. If he can trace a cause for some weakness he can probably think of the kind of mutation which will improve it.

The

Turing’s idea has a lot of followers, but it also faces a formidable objection. It is mentioned above that some scientists believe that future developments will prove that a machine can pass the Turing test. But a philosopher thinks we don’t have to wait for the day when a computer passes the test. Let’s suppose that a computer has already passed the test. That computer is still unable to think. Why?

A philosopher named objects to the Turing test, not on the point that we can never build a computer that can pass the Turing test, but on the point that a computer that passes this test still cannot think. This is how Searle puts it:

From , Brains and Science by John Searle, 1984

…It is essential to our conception of a digital computer that its operations can be specified purely formally, that is ,we specify the steps in the operation of the computer in terms of abstract symbols – sequence of zeroes and ones…

…imagine that you are locked in a room, and in this room are several baskets full of Chinese symbols. Imagine that you (like me) do not understand a word of Chinese, but that you are given a rule book in English for manipulating these Chinese symbols. The rules specify the manipulations of the symbols purely formally, in terms of their syntax, not their semantics. So the rule might say: ‘Take a squiggle-squiggle sign ot of basket number one and put it next to a squiggle-squiggle sign from basket number two.’ Now suppose that some other Chinese symbols are passed into the room, and that you are given further rules for passing back Chinese symbols out of the room. Suppose that unknown to you the symbols passed into the room are called ‘questions’ by the people outside the room, and the symbols you pass back out of the room are called ‘answers to the questions’. Suppose, furthermore, that the programmers are so good at designing the programs and that you are so good at manipulating the symbols, that very soon your answers are indistinguishable from those of a native Chinese speaker. There you are locked in your room shuffling your Chinese symbols and passing out Chinese symbols in response to incoming Chinese symbols. On the basis of the situation as I have described it, there is no way you could learn any Chinese simply by manipulating these formal symbols. Now the point of the story is simply this: by virtue of implementing a formal computer program from the point of view of an outside observer, you behave exactly as if you understood Chinese, but all the same you don’t understand a word of Chinese. But if going through the appropriate computer program for Chinese is not enough to give you an understanding of Chinese, then it is not enough to give any other digital computer an understanding of Chinese. And again, the reason for this can be stated quite simply. If you don’t understand Chinese, then no other computer could understand Chinese because no digital computer, just by virtue of running a program, has anything that you don’t have. All that the computer has, as you have, is a formal program for manipulating uninterpreted Chinese symbols. To repeat, a computer has a syntax, but no semantics.

What Searle means is that a computer that passes the Turing test can communicate successfully with people, yet it does not understand a word of what it and people say. All that it can do is manipulate symbols according to the rules embedded in the software. To be intelligent requires that one understands the meanings of the symbols one uses. Therefore, a computer that passes the Turing test still cannot think.

The above argument is an analogical argument (see Acharn Pratoom’s lesson on analogy). That is to say, it has the following form: The person in the CR has the properties A B C…and he doesn’t understand Chinese A digital computer has the properties A B C …like the person in the CR Therefore, the computer doesn’t understand Chinese.

Can you fill in the properties A B C…?

The controversy comes down to this. What does it mean to understand symbols? If I ask a person to bring me a book, and he goes into the library and brings me a book, isn’t that enough to say that he has understood me? If that is not enough, what else do I have to do to test whether he understands me or not. You can ask yourself this: I understand the language I’m using, what else do I have apart from the ability to use symbols to communicate with other people that makes me understand the language.

There is one thing we have to be clear about. If Searle were right, it wouldn’t mean dualism is correct. Searle’s point is that human intelligence cannot be duplicated by machines. It does not follow that humans have minds distinct from their bodies. It is compatible with the fact that all thinking comes from the brain and the is such that its ability cannot be duplicated by a computer software. In fact, Searle himself believes that the mind is just the brain. But if Searle were right, then one attempt to refute dualism has failed.

So you think that if humans are purely physical things, then death is the end and immortality is out of the question. You might be wrong. A person who believes that the brain is just a computer argues that in 2045 humans will become immortal. You should be interested in this because by that time you will still be alive and so you will become immortal (if that person is correct). By that time I’ll be dead and gone so count yourselves lucky. What I’ve just said about this possibility appears in an issue of the Time magazine. I have put this article on my web page (look for “Immortality”). Reading this article is not required (it won’t be in the exam). The article might be long but it’s very easy to read. It will cheer you up if you think immortality is desirable (it could be boring too!), but only if you believe the brain is just a computer.