(Artificial intelligence, cognitive science, Hungarian literature)


Ways of Thinking: The Limits of Rational Thought and Artificial Intelligence

László Mérő



Contents

0. Chapter Zero 1

I. The diversity of thinking 9
1. Logical thinking 11
2. Common sense 26
3. Puzzles and science 36
4. Ways of thinking in different cultures 45
5. Levels of thinking 62

II. The building blocks of thinking 81
6. Cognitive schemata 83
7. The magic number seven 89
8. Some tens of thousands of schemata 97
9. Some tens of thousands of what? 104
10. A challenge for programmers 110
11. From beginners to grandmasters 115
12. Profession — language — way of thinking 126
13. Artificial intelligence at candidate master level 143

III. The strength of diversity 163
14. The limits of rationality 165
15. High-level cognitive schemata 175
16. Mystical thinking 191
17. The trick of evolution 200
18. Alternating the reference systems 215

Bibliography 236
Sources 243
Index 246



0. Chapter Zero

There are as many ways of thinking as persons. There is no thought, kindness or insult two persons would react the same way to. "You who read me, are you sure of understanding my language?" — asks Jorge Luis Borges in one of his short stories, stepping out from within the strict frame of his fable for a moment. The well-defined frame and its deliberate breakdown are both means of expression.

Some people will nod in agreement with the last sentence, others will argue about it. Still others will just shrug at best. All those reactions are very human. Those who agree probably react to something I was not even thinking of when I was writing that sentence. The same is true for those who disagree. Those who shrug are left unaffected by this sentence: they watch the issue from the outside. They may be of the scientific breed that would first want to see a well-built system of concepts around a statement like this, in which they can decide the validity of the statement. Or they want it to be proven to them by the usual scientific methods.

The phenomena of thinking and intellect are studied by several branches of science — all from different points of view. Biologists, philosophers, psychologists and even mathematicians interested in logic, or engineers wishing to build artificial intelligence, all pry into the essential nature of reason. Nevertheless, they have hardly anything to say to each other, for their frames of investigation are so different. Which should then be the branch of science whose methods and system we will accept as competent? I would like to start by admitting that you are not holding a so-called interdisciplinary book in your hands. Such books perhaps do not even exist, as will be expounded in greater detail later, in Chapter 12. If you wish you may read this book as an interdisciplinary study till then, for we are going to use the results of several branches of science indeed. I would also like to avoid mislabeling: this is not a popular scientific book either. We will talk about science, the book may possibly propagate public knowledge, even unintentionally, but not within the frame of popular science. This time I am not against the genre. It is a justified and

1

nice undertaking for someone to try and familiarize outsiders with his profession's way of thinking. This book, however, goes right into the causes and reasons of the diversity of ways of thinking.

I have written this piece of reading — or if you wish to switch to Latin: this legend — to provoke readers' minds. This is a legend or a tale about the tricks of how our thinking works and about the efforts and failures of artificial intelligence. Rather than any special previous education, I hope some mental effort will be necessary during reading. In order to facilitate continuous reading, I do not interrupt the text with exact citations or notes. But if the reader wants to dig deeper into the subject, look up some of the scientific results or simply check what the author means, it is necessary to make the sources available to him even in such a book. The detailed Index helps him in that, with the references also indicated. Instead of footnotes, the sources of quotations can be found under Sources at the end of the book.

Let us start with the statement: there are a lot of ways of thinking. However, we do know some guiding principles along which the ways of thinking can be classified into types. It can be seen day by day how differently specialists and laymen think within any profession. We can talk about common and scientific ways of thinking, and we feel that the artists' way of seeing things radically differs from both. Abstract and concrete ways of thinking can also be differentiated. We also talk about rational and irrational thinking. What do these expressions actually mean? In fact, we can now even talk about artificial intelligence that is created with the help of computers. Is it of the same form of reason as natural thinking, or is it basically different? Can this distinction also be listed among the above, not necessarily exact but commonly used, categories?

Meeting a young man

Artificial intelligence still owes us the fulfillment of its greatest promises. Sometimes one cannot help confronting "his young self" who impertinently asks the questions: What about the translator program? What about the dictaphone that types a text from a tape recorder at least as correctly as an average secretary? What about the World Champion chessprogram? What about the visual form perception, poetry and music?

2

The answers are similar to those Frigyes Karinthy, the father of satire in Hungary, gave to a young man who — as his own former self in a short story — sprang the questions on him: "What about your flying machine?" "What about your great symphony? Your awe-inspiring play about the grey horizon and the vibrations and convolutions of the proud gods beyond the horizon?" "What about Hungary, proud and independent?" Karinthy answers in embarrassment: "We are working towards that goal, I and other people. Still, it isn't something you can achieve overnight. After all, one's got to make a living, too." Today the commercially successful practical products of artificial intelligence give a wide berth to the unexpectedly difficult basic questions, and embark on solving only simple partial problems.

Nobody has been able so far to solve the above great problems through artificial intelligence, albeit they are surely not theoretically unsolvable, since man's natural intelligence can solve them. That is why the investigators of artificial intelligence have become increasingly interested in the results of psychology. If man's thinking mechanisms were understood better, then somehow it would also become possible to model and simulate them by artificial means.

I have also met the above young man and repeated Karinthy's excuses almost word by word: "Well, you see... It was no go... The Grey Horizon — that isn't something an actor can play... But I have written a pretty good sonnet on the theme... It appeared in print in a distinguished review... People liked it... And I've been a better-paid writer ever since..." Today I am not quite sure any more that the great promises of artificial intelligence can be realized at all (although I am not convinced of its opposite either), but I am certain that we are on a much longer and more difficult road than was originally thought by the enthusiastic founders of artificial intelligence some twenty or thirty years ago, who were encouraged by the initial spectacular successes.

When the library comes to life

There are libraries where the elaborately and spaciously arranged shelves invite one for a kind of browsing instead of absorbed reading. Having borrowed the book one wanted, one may spend even hours there, rambling frivolously among the books, dipping into some of them, reading a couple of pages in the theory of relativity, then five pages about the habits

3

of primitive people. There is something interesting in each of them, and one feels it would be nice to read them all thoroughly. Not as thoroughly, though, as to spend too much energy on it: they are just pleasant passing adventures.

Then there comes a time when the visitor to the library becomes deeply interested in something: e.g. he receives eye-glasses, she gets pregnant or becomes possessed by a scientific problem. He or she goes to the same library and starts browsing. He/she is astonished to find that every book is meant for him/her: every book has something to say about seeing, birth, or the scientific problem he/she is interested in. Even the radio programmes are talking about the topic, i.e. it is all over town. The library comes to life.

As time passes, one gets accustomed to wearing spectacles, bringing up a child, or investigating a scientific problem. The picture becomes complete and one forgets its components. Or at least one sees the problems within a quite different system. One reorganizes those incidences that were not forgotten: probably there is a reason for keeping them in memory. And, of course — as it will be discussed abundantly — the whole is more than the sum of its parts.

Logic is doubtless unshakable, but it cannot withstand a man who wants to go on living.

(Franz Kafka)

A happiness that often madness hits on, which reason and sanity could not so prosperously be delivered of.

(William Shakespeare)

Reason is a harmonizing, controlling force rather than a creative one. Even in the most purely logical realm, it is insight that first arrives at what is new.

(Bertrand Russell)

Reason is a good tool, in fact, it is an essential tool, yet, it is a subordinate tool. We must not believe in it, we ought to believe only in description and reality — but it is our duty to be sceptical by the aid of reason.

(Géza Ottlik)

Our washing up is just like our language. We have dirty water and dirty dishcloths, and yet we manage to get the plates and the glasses clean. In language, too, we have to work with unclear

4

concepts and form a logic whose scope is restricted in an unknown way, and yet we use it to bring some clarity into our understanding of nature.

(Niels Bohr)

He had two characteristic features, two passions: an unusual power of clear and logical reasoning, and a great moral purity and sense of justice: he was ardent and honorable.

But he would not have made a scientist of the sort who breaks new ground. His intelligence lacked the capacity for bold leaps into the unknown, the sudden flashes of insight that transcend barren, logical deductions.

And if he were really to do good, he would have needed, in addition to his principles, a heart capable of violating them — a heart which knows only of particular, not of general, cases, and which achieves greatness in little actions.

(Boris Pasternak)

The heart has its reasons which reason does not understand.

(Blaise Pascal)

My smartest sayings are those that even I had not expected.

(Jules Renard)

In the beginner's mind there are many possibilities, but in the expert's there are few.

(Shunryu Suzuki)

5



A friendly conversation could be important for us, too: we can get accommodation in a foreign country, or we can study the gestures and mimic of our partner, etc. It is an exciting game, like the conversation of Kornél Esti with the Bulgarian conductor in the Hungarian poet Kosztolányi's novel. But there is always something at stake in human affairs. Unwittingly, Kornél Esti first outraged the Bulgarian conductor, then conciliated him, while he deeply felt and understood the affair, although at a level that was quite different from where the conductor's story took place. Kornél Esti would certainly have had a very bad feeling if he could not have conciliated the Bulgarian conductor. There is no such understanding and feeling in ELIZA. It cannot be found in any of the present artificial intelligence programs either. Not only is this level unattainable, but even direct understanding, which is much simpler, leaves much to be desired, even if the program (or rather its author) aims at mastering only a very narrow field.

30


[On holism and reductionism]

The author should probably take sides with one of these trends. I am not going to do this now, or later. In my everyday scientific (i.e. puzzle solving) work I naturally adhere strictly to the norms of science — primarily not because otherwise the scientific community would expel me from among them, but because it is worth making science and solving puzzles only within the strictly given frames. This, however, does not prevent us from trying to look at our daily work from the outside of

51

science occasionally — like at the moments of meeting a young man, as in Chapter Zero. On these occasions a more holistic approach may be helpful. This cannot be expressed by the methods of science, although there are generally accepted methods by which it can be expressed, such as the arts, for instance. Karinthy's work is a piece of literary art rather than a scientific work or case history about the unfulfilled hopes of a genius. I do not know whether science will ever be able to grasp the essence of literature, but at present it is far from it.

An atheist could also build a good cathedral. It is enough to understand the religion whose cathedral he is entrusted to build. It is the employer who needs the belief to concentrate his resouces just on this goal. The situation is the same in science. All the scientific community requires of its members is to work according to the appropriate paradigm; in the meanwhile, the members may think of the paradigm as they like. I can see no more trustworthy method than science to study thinking, either. My native language is rationality; my everyday logic cannot accept conclusions that contradict scientific results. Yet at the same time I clearly feel that there are many fields that slip out of the present range of science — and I do not deem them unworthy of reflection. I cannot and do not find it important to decide where — if they exist — are the boundaries of the fields that can ever be possessed by science. This is why I do not take sides on the question of reductionism versus holism.

On the other hand, I feel that boundaries have been outlined by now that perhaps can be reached through the foreseeable methods of artificial intelligence, but that can hardly be exceeded. I would like to draft these boundaries, and also to show the areas where science may be competent (whether you like it or not), as well as those areas where we justly look for (and find) other means of expression and presentation.

52



Thus, the number of ways of choosing seven schemata from the LTM [Long Term Memory] of the subject must be greater than the number of all possible chess positions. "This is the reason of every self-torture and song," as the professor of mathematics, Pál Turán, liked to cite the Hungarian poet Ady when he was over the most difficult part of a complicated train of thought, reached a resting point and only had to score the points.

99


[On mysticism]

If everything is one, then that one thing must be very complex, as it is manifested in so many forms. Today's science agrees: the world is complex. In fact, science also hopes that the world is not as complex as that; it hopes that the world can be described by not too many well-chosen general concepts and rules. Einstein also said that "The fairest thing we

can experience is the mysterious. It is the fundamental emotion which stands at the cradle of true art and true science." But mystics do not want to learn about or describe the unity of the universe; they want to experience it. Complexity is meaningless in this sphere of thinking; it is replaced by an extreme kind of trance logic. Who could be the person who wants to learn everything if everything were one? Or as the Hungarian poet Sándor Weöres says,

If the wide world were a dipper,
it would not fit in my apron;
but how could I get an apron,
if the wide world were a dipper?

193


SOURCE: Mérő, László. Ways of Thinking: The Limits of Rational Thought and Artificial Intelligence, translation by Anna C. Gősi-Greguss; English version edited by Viktor Mészáros. Singapore; Teaneck, N.J.: World Scientific, 1990. Extracts, pp. 1-5, 30, 51-52, 99, 193. References to Frigyes Karinthy, Géza Ottlik, Dezső Kosztolányi, Endre Ady, Sándor Weöres, Jorge Luis Borges, Franz Kafka, et al.


Frigyes Karinthy: philosophical fragments / filozofiaj fragmentoj

Futurology, Science Fiction, Utopia, and Alienation
in the Work of Imre Madách, György Lukács, and Other Hungarian Writers:
Select Bibliography

Frigyes & Ferenc Karinthy in English

Géza OTTLIK (1912 - 1990) Study Guide & Bibliography

Robert Zend (Hungarian-Canadian writer, 1929-1985): Dedications, Works, Links

Jorge Luis Borges: Selected Study Materials on the Web

Science Fiction & Utopia Research Resources: A Selective Work in Progress


Home Page | Site Map | What's New | Coming Attractions | Book News
Bibliography | Mini-Bibliographies | Study Guides | Special Sections
My Writings | Other Authors' Texts | Philosophical Quotations
Blogs | Images & Sounds | External Links

CONTACT Ralph Dumain

Uploaded 7 February 2016

Site ©1999-2016 Ralph Dumain