The following was my first foray into the subject, followed up by much longer pieces which can be looked at later. It’s a response to John Searle’s (hideous) thoughts on Can Computers Think?.
Can Computers Think?
Firstly, it should be stated that the interpretation of the question should not be a significant obstacle. Whether or not it is implied in a present or future tense is irrelevant, since the word ‘computer’ refers to a concept rather than an existing machine - I concur with John Searle, in that ‘the nature of the refutation is completely independent of any state of technology. It has to do with the very definition of a digital computer, with what a digital computer is.’ Few would dispute the simplistic but accurate three-stage model of a computer: input, process, and output. Having discounted the relevance of the first two words, we contemplate thought. I postulate that, if we are to accept the simplistic model of a computer (as compared to the incredible complexity of the actual circuitry and components), then we should also be prepared to accept a simple model of the human mind. I propose that there is no difference: the mind essentially operates in exactly the same way as a computer.
The layout is thus: the input devices (which to a computer are keyboard and mouse, amongst others) of a human are the sensory inputs. Without the senses of touch, sight, hearing, smell, and taste, there is nothing for the mind to consider. Note that I refer always to the entire lifespan of each entity - a human who exists with all senses and then suddenly without is an entirely different case from one who has existed forever without sense: the former has the extra ‘sense’, of memory, or experience. From these inputs, information is supplied to the central processing unit, or the mind. The mind analyses this information, typically in a comparative way with information already gathered and stored in its databanks (memory), and arrives at a conclusion. This conclusion is then output - possibly through a physical means, utilising the powers of the body, or in an internal sense, storing the conclusion in memory for future reference. This is the power of thought.
In broad terms, this is sufficient to answer the question raised. The difficulty occurs with the concept of thought, and the meaning attached to it. Some who tackle this issue tend to use the notion of thought as a universal description of the superiority of the human mind over other substances - the advantages of the mind can be summed up as the ability to think. I wish to isolate further than this: firstly, there is constantly a cycle of input-output in a being. Any sensory input produces a relevant output, which is typically in the form of memory storage - for the majority of inputs, there is never a noticeable output: if one were to listen to music and upon hearing each chord, say its name, life would be incredibly loud. This leads us to assume that the mind is not totally reactionary, as described, but has a rational program, or system of processing information that determines the importance or relevance of the input information, and whether any subsequent course of action is required. Upon the sight and smell of chocolate, one must determine whether or not one can actually eat chocolate at that moment in time (there are a multitude of reasons for and against, based on knowledge already known - i.e. an allergy, or having eaten too much an hour previously). In short, there is obviously a middle, processing, stage involved with the function of the human mind - thought.
This is not sufficient to describe the activity of a human, however. There are other activities which do not at first glance fall under the same banner, the most obvious of which is emotion, or feeling. If one is to feed this concept through the ‘machine’ outlined above, it does not seem to fit, but the problem is easily solve - emotion is a mental thing, but all that this means is that the input comes not from the senses, but from the output of a previous ‘operation’ or thought process. One occurrence, either in the physical or mental world, leads to an emotion or feeling - but this feeling is no different to any other thought process - it has an input, thought the context is different from sensory input, it has a process, where the field of knowledge compared to is also entirely different from that of sensory consideration, and it has an output - which as always, is dependent solely on the other two stages and their interaction. Emotion is not a human feature, created by a superior mind or extra series of genes. It is the result of an acquired experience, the knowledge of the individual, and it’s role in analysing, classifying, and justifying new information. Fundamentally, emotions are a progression from thought.
This seems a bold statement. Take, however, the image of a mother and her child, of a post-infantile age. If the mother should die, and subsequently the child feels sad, the logical explanation is as follows:
-
The information possessed by the child articulates that the mother was the one who gave birth to the child, and who nurtured him or her.
-
Therefore, given the significance of birth and nurturing with relation to existence (a significance typically learnt by growing up, revealed and probably amplified in the long term), it follows from 1 that the existence of the child is because of, and was maintained by, the mother.
-
Given 2, and various others factors developed during childhood and beyond, it is therefore logical that the mother is and was a significant part of that child’s life. This is not a feeling, but a knowing, even though an emotion will likely evolve from the contemplation of it.
-
If then, the mother is significant in the child’s life, the loss of this person contradicts the knowledge of the child, and disturbs the related information in the child’s mind - it has to adapt, by altering and replacing all the information relevant to the mother that no longer exists. It is this process of contradiction, or more generally, the presentation of a mental problem, that results in the output of an emotion.
Does a cat feel sad when it’s mother dies? We assume not, because we believe that they as a species are not as evolved as humans. Why does we think this, though? One of the primary explanations is the notion of language. Try to imagine a world totally void of language, as speech and written word. Communication reverts to its original form: a primitive recognition system involving sight and sound. Here there is no established language or mode of communication, one must try to understand what is happening all by themselves. The most obvious example of such an environment is that of the new-born baby. Can such a child think? Certainly not in the way that you and I do, otherwise there would be no need for nurturing, education, or university! A new-born needs nurturing and feeding by another, more ‘evolved’ being, if it is to survive. It may feed the physical pang of hunger, and respond to it through tears and screams, but knows not what it means, nor how to cure the problem. It is only through repeated feeding by the parent that it comes to realise what is necessary for survival. Through continued exposal to the outside world and the interaction with it, the baby begins to pick up key aspects of life, and starts to evolve the concept of thought: the input (hunger pangs), the process (through experience, remedy of hunger pangs comes through food), and the output (the physical ‘hunt’ for food). This is the construction of the program that allows thought to emerge from birth - it is a routine, a procedure, a computer program. Before the child learns why it is that hunger pangs occur, they are like error messages on a computer - any digital computer does not posses the knowledge to explain what these messages are and how to resolve them (the beginnings of AI), but this is merely because it has not been exposed to them, and furthermore, does not posses the ability to identify them by itself. A computer does not posses the 5 senses - all it has is a keyboard and mouse, both of which are reliant on human intervention. It is no wonder that a computer cannot learn, if it has no body to go with the brain!
In his essay, Can Computers Think?, Searle proves nothing other than to enforce that a) if one has no input, one can have no output, and b) a simple recognition game (the Chinese Rooms) is possible by all. This asserts little other than what I have stated - recognition is not thought, it is comparison. When a human sees an apple, it isolates the shape and colouration of the object, compares these attributes to its memory or database, and arrives at the conclusion that it is most probably an apple, with certain characteristics (i.e. that it can be eaten). If a human is shown that the Chinese symbol for apple corresponds directly with the English, or even physical, equivalent of an apple, then the human will of course recognise this in the future. A computer can do both of these things, and can do it much quicker. In fact, the whole of a computer’s life is spent doing this - receiving information, processing it, and outputting it. I again argue that this is exactly the same system that is involved in human thought, and that the reason why it appears inferior is because a computer has a restricted array of inputs (restricted by the human engineers who created it). There is nothing more to it than this. Searle repeatedly states his belief in semantics, but what is semantics without syntax? What is meaning without something to apply meaning to? Is semantics not merely an extension of syntax, a meaning/conclusion derived by comparing inputted information to already accumulated data? Syntax can exists without semantics (the example of a baby or cat), but semantics does not exist without syntax. Which, then, is most important?
That, is for the reader to decide, but my position should by now be clear. Computers are not the same as minds, or brains - that is obvious. But in order to try and compare the possibility of thought in either, we must look beneath the surface, to the simpler inner workings of sentience. Human thought as a concept is an incredible thing, but that does not make it any more complicated or elaborate than we may think. Given the body, and the experience, a computer can be intelligent, and think for itself. In the same way as we are nurtured and given new ‘programs’ or procedures for carrying out different tasks or taking care of ourselves, so to would a computer given the same array of inputs and outputs as a human. The need for a distinction between man and machine arises primarily from fear - do we really want computers to be able to think for themselves, knowing the power they possess when we harness them? Are we really prepared to grant computers with the ability to be intelligent and self-sufficient? Whether fortunate or not, in the world of technology, man plays God, and so it will be until we give something away - the key to the cell in which AI is contained.
footnotes
J. Searle, Can Computers Think, in J. White’s Introduction To Philosophy, p.198
