Machine Thought

   
Recent changes
Table of contents
Links to this page
FRONT PAGE / INDEX

(I had dreadful trouble starting this essay, so I used this "device" of pretending it was an excerpt from a larger work to avoid the problem if getting started. I had intended to go back and fix it, then ran out of time. I'll leave it stand - it is here as it was when handed in for marking.)

An Excerpt from "The Complete and Final Solution to the Ultimate Question",

Translated by C. Douglas.

.... and as a consequence of this, consideration falls upon the question of whether or not there is a nonphysical part to the existence of a person, i.e. a mind, which is distinct from the brain, and which is not material in substance. Clearly this question cannot be treated in full in this paper, but it suffices to show that the argument presented in favor of the existence of the mind fails, for it is the failure of this particular argument that has a bearing on the question of extraterrestrial intelligent life, as mentioned above. (The argument, of course, being that by Descartes set out in Discourse 5, as alluded to earlier.)

This argument of Descartes falls in two parts. The first of these says that the logical possibility of a machine that imitates thoughts naturally follows from the assumption that people do not have a non-physical "organ," i.e. a mind. This part of his argument is quite sound and irrefutable, for if the brain were all there is to thinking then it would be possible (in principle at least) to build hundreds of millions of simulated neurons and link them together to form a "brain." Thus, if people do not have a mind, it would be possible to build a machine to imitate thinking.

The second part of Descartes' argument is designed to show that it is impossible for machines to imitate the behavior of rational beings; and hence, thought. Should Descartes succeed in showing this, he will have shown that people must indeed have a non-physical part to them, i.e. a dualist-style mind, for if people did not have a dualist-style mind, then, by the earlier argument, a machine to imitate thought would be possible; a contradictory result.

Descartes gives two reasons for the impossibility of a machine imitating thought.

"Of these the first is, that they could never use words or other signs, composing them as we do to declare our thoughts to others/ For one can well conceive that a machine may be so made as to emit words, and even that it may emit some in relation to bodily actions which cause a change in its organs, .......; but not that it may arrange words in various ways to reply to the sense of everything that is said in its presence."

Discourse 5, p74

The program is called SHRDLU and is described in detail in Winograd's thesis, published under the title of "UNDERSTANDING NATURAL LANGUAGE."
This argument can be seen to have failed when one considers a computer program written by T.Winograd for his 1971 Ph.D. thesis. The heart (mind?) of SHRDLU is a computer sub-program which can accept, interpret, act upon, and reply to English sentences which talk about or refer to the limited domain of a world inhabited by blocks, a box, a table, and SHRDLU's own "desires," "motivations," and actions. SHRDLU can refer to past events and reasons for doing things. She can accept and act upon ownership of the objects in her world, and expand her vocabulary with new words given to her and defined by the user. SHRDLU replies in English to anything said to her in English, even if it is to say that she doesn't understand, and thus is replying to the sense of everything that is "said" in her presence. Even if one does not accept that the competence shown by SHRDLU qualifies because it is on such a limited domain, if equipped with sufficient memory, a suitable learning method (as given by PURR-PUSS, described in "THINKING WITH THE TEACHABLE MACHINE by J.H.Andreae) and sufficient time (4 to 5 years, the time taken for a child to show the same degree of "comprehension" and "understanding") one can see that Descartes' first objection to machine thought is suitably rebuffed.

In 1981 as part of my undergraduate degree at Monash University in Melbourne, Australia, I took first year philosophy, a 25% module, and classed it as a second year minor, which would normally be a 20% module.

This essay was one of those required in the "Problems of Philosophy" sub-module. I found it recently during a tidy-up, and thought it vaguely amusing, as well as a little embarrassing.

Why not share.

This is one of the Essays In Philosophy.




Thus attention turns to the second reason;

"And the second is that, although they might do many things as well as, or perhaps better than, many of us, they would fail, without doubt, in others, whereby one would discover that they did not act through knowledge, but simply through the disposition of their organs: for, whereas reason is a universal instrument which can serve on any kind of occasion, there organs need a particular disposition for each particular action; whence it is that it is morally impossible to have enough different organs in a machine to make it act in all the occurrences of life in the same way as our reason makes us act."

Discourse 5, p 74

Here, Descartes says that in order for a machine to act in a given situation it needs to have the appropriate action explicitly built into it, along with the situation in which it is appropriate. If this were indeed the case then Descartes surely is correct when he says that it would be impossible to enumerate every situation along with the appropriate action. However, modern computers are programmed to give an appropriate response to a wide variety of situations without these situations being explicitly stored in the computer in a way that Descartes thought would be necessary. A good example of this is the chess program SARGON 2.5 written by Kathe a Dan Spraklen (explained in the book entitled "SARGON," by the same authors). In chess there are an estimated 10120 possible positions (10120 is a 1 followed by 120 zeroes!) (10100 " ... is a number so large that it exceeds the number of raindrops that would fall on New York, Los Angeles, and Chicago, in more than a century." -- The How and Why Wonder Book of Mathematics)

Allowing that only one thousandth of a percent of these are likely to occur, this still leaves 10115 positions that the computer would have to "remember" in order to play chess in the way Descartes envisaged; yet this program runs on a machine in which it is physically impossible to store and more than 820 positions and it still plays an exceptionally good game. (It scored 3 out of 6 in the Monash Chess Club qualifying tournament.)

It does this by taking the position on the board and applying deduction and evaluation techniques to produce a score for each move in a manner similar in kind if not degree to the manner in which people select a move. Thus computers can apply techniques which can serve on any kind of occasion. The program GPS (described in "ARTIFICIAL INTELLIGENCE" by P Winston) uses techniques which apply in every situation, indeed, it is called GPS (General Problem Solver) because it can solve problems in every situation. Hence it is clear that Descartes' second objection to machine thought is no longer valid.

It has therefore been shown that the reasons Descartes gives for the impossibility of machine thought are no longer valid, but this does not mean that people do not have a non-physically part to their existence. All that has been shown is that Descartes' proof does not work, which does not show that there is no other proof which is correct. The possibility of the existence of a non-physical mind has thus neither been proved nor disproved.

It is, however, as mentioned earlier, the reason for the failure of Descartes' argument that is of particular interest. It can be seen that the ...


Contents

 

Links on this page

 
Site hosted by Colin and Rachel Wright:
  • Maths, Design, Juggling, Computing,
  • Embroidery, Proof-reading,
  • and other clever stuff.

Suggest a change ( <-- What does this mean?) / Send me email
Front Page / All pages by date / Site overview / Top of page

Universally Browser Friendly     Quotation from
Tim Berners-Lee
    Valid HTML 3.2!