Causality, Complexity and Computers
23
is but one part of what we, as scientists, are trying to do. The problem of
complexity should not be relegated to a transparent box which we trust but
do not understand.
We are therefore led to consider complexity as a problem in its own
right apart from the simulation of a particular system under review.
There are three aspects of a system which can be usefully considered as
influencing its behaviour: (1) the number of elements, (2) the structure,
(3) the values of the parameters. By constructing computer models and
varying any of these three classes, we can ascertain their influence on
systemic properties. But since the computer gives only particular solutions,
we are using it as a source of empirical data from which empirical generalisation must be made. By starting with relatively simple systems we might
hope to build up an understanding of systemic behaviour which we are,
as yet, lacking. We must begin to learn what our evolutionary or educational
past has apparently not forced upon us: namely, to think in non-linear and
quantitative terms. I do not think that our inability to comprehend complex
situations is due to the 'structure of the mind' or other such limitation but
essentially due to unfamiliarity with the behaviour of such systems. A good
chess player, e.g. can understand situations which completely baffle most
of us. By making a large number of observations on computer results and
relating these to structure and quantities, it should be possible to obtain a
direct and intuitive understanding of 'standard' types of complex systems.
We then might build up a 'dictionary of complexity' which could become
as familiar as the rules of logic are alleged to be.
Computer technology, far from solving the problems of complexity,
has high-lighted the necessity to approach it in a different way. The computer is not only a powerful tool for simulation and prediction, but if
properly used can also aid our understanding. But we must deliberately
set out to use it in this way-as a 20th century abacus-not only for calculation but for the training of our unsophisticated minds.
References
1. BUCHER, T., U. W. RUSSMANN: Angew. Chern. 75, 881 (1963).
2. BURNS, J. A.: 3rd Symposium on Quant. Biol. of Metabolism, p. 75 (1968).
3. BURTON, A. c.: J. Cell Compo Physiol. 14, 327 (1939).
4. GARFINKEL, D.: Biochem. Soc. Symp. 26, 81 (1966).
5. KAeSER, H.: in: Biological Organisation at the cellular and supercellular
Level, p. 25. Academic Press 1963.
6. KREBS, H. A., u. H. L. KORNBERG: Ergebn. Physiol. biol. Chern. expo
Pharmakol. 49, 212 (1957).
7. ROSEN, R.: Int. Revs. Cytology, 23, 25 (1968).
23
is but one part of what we, as scientists, are trying to do. The problem of
complexity should not be relegated to a transparent box which we trust but
do not understand.
We are therefore led to consider complexity as a problem in its own
right apart from the simulation of a particular system under review.
There are three aspects of a system which can be usefully considered as
influencing its behaviour: (1) the number of elements, (2) the structure,
(3) the values of the parameters. By constructing computer models and
varying any of these three classes, we can ascertain their influence on
systemic properties. But since the computer gives only particular solutions,
we are using it as a source of empirical data from which empirical generalisation must be made. By starting with relatively simple systems we might
hope to build up an understanding of systemic behaviour which we are,
as yet, lacking. We must begin to learn what our evolutionary or educational
past has apparently not forced upon us: namely, to think in non-linear and
quantitative terms. I do not think that our inability to comprehend complex
situations is due to the 'structure of the mind' or other such limitation but
essentially due to unfamiliarity with the behaviour of such systems. A good
chess player, e.g. can understand situations which completely baffle most
of us. By making a large number of observations on computer results and
relating these to structure and quantities, it should be possible to obtain a
direct and intuitive understanding of 'standard' types of complex systems.
We then might build up a 'dictionary of complexity' which could become
as familiar as the rules of logic are alleged to be.
Computer technology, far from solving the problems of complexity,
has high-lighted the necessity to approach it in a different way. The computer is not only a powerful tool for simulation and prediction, but if
properly used can also aid our understanding. But we must deliberately
set out to use it in this way-as a 20th century abacus-not only for calculation but for the training of our unsophisticated minds.
References
1. BUCHER, T., U. W. RUSSMANN: Angew. Chern. 75, 881 (1963).
2. BURNS, J. A.: 3rd Symposium on Quant. Biol. of Metabolism, p. 75 (1968).
3. BURTON, A. c.: J. Cell Compo Physiol. 14, 327 (1939).
4. GARFINKEL, D.: Biochem. Soc. Symp. 26, 81 (1966).
5. KAeSER, H.: in: Biological Organisation at the cellular and supercellular
Level, p. 25. Academic Press 1963.
6. KREBS, H. A., u. H. L. KORNBERG: Ergebn. Physiol. biol. Chern. expo
Pharmakol. 49, 212 (1957).
7. ROSEN, R.: Int. Revs. Cytology, 23, 25 (1968).