Abstract
It is hardly possible to talk about thinking by computers without saying something first about thinking by people. There are two reasons why this is so. First, the only definitions of thinking that are of any use at all are ostensive ones. We can point to a person in a certain state of activity and say, ‘Thinking is a set of processes like those now taking place in the central nervous system of that person.’ Alternatively, we can point to the statement of a problem and to its solution and say, ‘Thinking is a set of processes like those that enabled a person to produce this problem solution from this problem statement.’ I do not mean that these two definitions are necessarily equivalent, but they might serve equally well as a basis for delimiting the set of phenomena we wish to understand when we investigate thinking.
That is the aim of that great seience which I am used to calling Characteristic, of which what we call Algebra, or Analysis, is only a very small branch, since it is this Characteristic which gives words to languages, letters to words, numbers to Arithmetic, notes to Music. It teaches us how to fix our reasoning, and to require it to leave, as it were, visible traces on the paper of a notebook for inspection at leisure. Finally, it enables us to reason with economy, by substituting characters in the place of things in order to relieve the imagination. . . .
Leibniz, On the Method of Universality (1674)
[R. Colodny (ed.), Mind and Cosmos, Pittsburgh: Pittsburgh University Press, 1966, pp. 3–21].
The work on which this chapter is based was supported in part by a grant from the Carnegie Corporation and in part by Research Grant MH-07722-01 from the National Institutes of Health. Several of my examples are drawn from joint work with K. Kotovsky, P. A. Simon, and L. W. Gregg. My debts to Allen Newell are too numerous to acknowledge in detail. To all of these, I give thanks and offer absolution for the particular conclusions reached here, which are my own.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
A typical example is Edna Heidbreder’s concluding comment in her article on ‘Thinking’ in the 1960 Encyclopedia Brittanica: “Thinking remains one of the unsolved problems of psychology.”
It might be mentioned that although there was much resistance to the atomic hypothesis through the first two-thirds of the nineteenth century, the grounds for this resistance were not what a radical operationalist might suppose. There were few objections to atoms because of their hypothetical character — only a few philosophers of science like Mach, Poincaré, and Russell, anachronistically stressed this toward the end of the century. The main objection was to the neglect of the ‘qualities’, like color, in a theory that took mass as the significant atomic property. The sceptics were humanists, not operationalists. See Stephen Toulmin and June Goodfield, The Architecture of Matter (New York: Harper & Row, 1962), pp. 234–37, 263–68; or Harvard Case Histories in Experimental Science, ed. by James B. Conant and Leonard K. Nash, Vol. I, 215–321 (1950).
Toulmin and Goodfield, pp. 365–68.
Compare Kekulé’s prescient observation, “Should the progress of science lead to a theory of the constitution of chemical atoms, it would make but little alteration in chemistry itself. The chemical atoms will always remain the chemical unit..” (Quoted by Toulmin and Goodfield, p. 265.) ‘Little alteration’ sounds too strong in the light of modern physical chemistry, but the import of the statement, that there is a distinct ‘chemical’ level, is still substantially correct.
Symbolic of this progress was the award of the 1963 Nobel Prize in Physiology and Medicine to Eccles, to Hodgkin, and to Huxley for their work on transmission of neural signals. See the brief appreciation of this work, by M. G. F. Fuortes, in Science 142, 468–70 (1963).
See Note 3 above.
Of course ‘next’ must be put in quotation marks since the differential equations describe the changes in the limit as the time interval is taken shorter and shorter.
Allen Newell et al., IPL-V Programmers’ Reference Manual (New York: Prentice-Hall, 2d ed.. 1964).
The analysis here is based on H. A. Simon and K. Kotovsky, ‘Human Acquistion of Concepts for Serial Patterns’, Psychological Review 70, 534–46 (1963).
For similar theories applied to closely related tasks, see J. Feldman, F. Tonge, and H. Kanter, ‘Empirical Explorations of a Hypothesis-Testing Model of Binary Choice Behavior’, in Symposium on Simulation Models, ed. by Hoggatt and Balderston (Cincinnati: South-Western Publishing, 1963), pp. 55–100;
and K. R. Laughery and L. W. Gregg, ‘Simulation of Human Problem-Solving Behavior,’ Psychometrika 27, 265–82 (1962).
In the partial reinforcement experiment, the subject is asked to predict whether the next stimulus in a series will be a ‘plus’ or ‘minus’. The sequence is in fact random, each symbol having a certain probability of occurring. Subjects, however, typically search for patterns: ‘a run of plusses’, ‘an alternation of plus and minus’, or the like. See the chapter by J. Feldman in Computers and Thought, Feigenbaum and Feldman (eds.), (New York: McGraw-Hill, 1964).
For a survey of these theories see A. Newell and H. A. Simon, ‘Computers in Psychology’, Handbook of Mathematical Psychology, ed. by Luce, Bush, and Galanter (New York: Wiley, 1963), I, and the references therein.
See references in ‘Computers in Psychology’, to the work of Hiller and Isaacson on musical composition, Clarkson on investiment decisions, and Feigenbaum and Simon on memorizing.
See H. A. Simon, ‘The Architecture of Complexity’, Proceedings of the American Philosophical Society 106, 467–82 (1962), Noam Chomsky, Syntactic Structures (The Hague: Mouton, 1957); and Toulmin and Goodfield, pp. 301–02. See also Section 4 of this volume.
For reasons both of economics and organizational simplicity, a typical computer has only a few ‘active’ memory locations (sometimes called accumulators) where processing can be carried out. Information is brought in from ‘passive’ storage locations, processed, then returned to storage. Thus, the steps involved in adding the number in storage location A to the number in storage B and storing the sum in C might be the following: (1) copy contents of A into accumulator, (2) add contents of B to contents of accumulator, (3) store contents of accumulator in C. With only one or a few active accumulators, the action of such a system is necessarily serial rather than parallel. Increasing the number of accumulators is expensive; it also creates an extremely difficult problem of co-ordinating their activity.
For some quantitative analysis, see A. Newell, J. C. Shaw, and H. A. Simon, The Processes of Creative Thinking’, Contemporary Approaches to Creative Thinking, ed. by Gruber, Terrell, Wertheimer (New York: Atherton Press, 1962),
Chapter 3; and H. A. Simon, and P. A. Simon, ‘Trial and Error Search in Solving Difficult Problems: Evidence from the Game of Chess’, Behavioral Science 7, 425–29 (1962).
In general, see A. Newell and H. A. Simon, Human Problem Solving, (Englewood Cliffs, N.J.: Prentice-Hall, 1972).
The planning heuristic is described briefly in ‘The Processes of Creative Thinking’, pp. 91–96.
Data on the behavior of subjects performing the Donald-Gerald task will be found in Sir Frederic Bartlett, Thinking (New York: Basic Books, 1958), Chapter 4.
A discussion of satisficing heuristics and aspiration levels will be found in H. A. Simon, Models of Man (New York: Wiley, 1957), Introduction to Part IV and Chapters 14 and 15.
D. W. Taylor, ‘Toward an Information-Processing Theory of Motivation’, Nebraska Symposium on Motivation, ed. by Jones (Lincoln: Univ. of Nebraska Press, 1960); Walter R. Reitman, ‘Personality as a Problem-Solving Coalition’, and Silvan S. Tomkins, ‘Simulation of Personality’, Computer Simulation of Personality, ed. by Tomkins and Messick (New York: Wiley, 1963).
Ulric Neisser, ‘The Imitation of Man by Machine’, Science 139, 193–97 (1963).
A.M. Turing, ‘Computing Machinery and Intelligence’, Mind 59, 433–60. (1950), reprinted in The World of Mathematics, ed. by James R. Newman (New York: Simon & Schuster, 1956), IV, and in Feigenbaum and Feldman, op. cit.: J. C. C. Smart, ‘Gödel’s Theorem, Church’s Theorem, and Mechanism’, Synthèse 13, 105–10 (June 1961).
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1977 D. Reidel Publishing Company, Dordrecht, Holland
About this chapter
Cite this chapter
Simon, H.A. (1977). Thinking by Computers. In: Models of Discovery. Boston Studies in the Philosophy of Science, vol 54. Springer, Dordrecht. https://doi.org/10.1007/978-94-010-9521-1_15
Download citation
DOI: https://doi.org/10.1007/978-94-010-9521-1_15
Publisher Name: Springer, Dordrecht
Print ISBN: 978-90-277-0970-7
Online ISBN: 978-94-010-9521-1
eBook Packages: Springer Book Archive