Popper’s World

John Maynard Smith

  • The Open Universe: An Argument for Indeterminism by Karl Popper, edited by W.W. Bartley
    Hutchinson, 185 pp, £15.00, July 1982, ISBN 0 09 146180 4

Karl Popper is perhaps the only living philosopher of science who has had a substantial influence on the way scientists do what they do. I say ‘perhaps’ because the same claim might be made for Thomas Kuhn. However, Kuhn seems to me a perceptive sociologist of science, but a poor philosopher. Also, in so far as he has had an effect on the way scientists behave, it has been pernicious: to be a great scientist, according to Kuhn, you must do revolutionary science, and the best evidence that you are doing it is that you are so obscure and inconsistent in your statements as to be wholly incomprehensible to others. This does not make for good science. In contrast, Popper has encouraged us to speculate boldly, but to be fiercely critical; it is true that we usually manage the latter only for other people’s ideas, and not for our own, but since science is a social activity, that suffices.

Popper’s main contribution to the philosophy of science was contained in his Logic of Scientific Discovery, first published in 1934, and in English translation in 1959. It was during the process of translation that Popper wrote a mass of additional material, which was at first intended to appear as an appendix to the Logic. However, the material accumulated until it was longer than the original book. Most of it is now published for the first time, in the form which it had reached by 1962. It appears in three volumes: Realism and the Aim of Science; The Open Universe: An Argument for Indeterminism; and Quantum Theory and the Schism in Physics.[*] This review treats only the second of these volumes, because that is as far as my competence will stretch. Indeed, many would doubt my competence to review even this volume, since I lack any training in philosophy. My excuse is that I have spent much of my life thinking about a branch of science – evolutionary biology – to which questions of chance and necessity are central. Also, although I regret not tackling the other two volumes, I think this is the one of most interest to the general reader.

As is made clear in his autobiography, Unended Quest, Popper’s life has been dominated by two problems, both presented to him during his youth in pre-Nazi Vienna. The first was to diagnose the difference between what seemed to him to be genuine knowledge, as represented by the work of Einstein, and what he saw as mere rhetoric, as represented by Freud, Marx, Adler and Jung. It was this problem which led him to the famous demarcation criterion of falsifiability, and to The Logic of Scientific Discovery. The second problem was to find a philosophical justification for human freedom; it is this which underlies the Postscript.

That this is Popper’s preoccupation is made clear in the Preface to this volume, in which he writes: ‘This book is then a kind of prolegomenon to the question of human freedom and creativity.’ He points out that there is an apparent contradiction between the widely held view that every event has a cause, and the common-sense conviction that, at least sometimes, men are free to choose what they will do – that men are sometimes masters of their fate. It is this apparent contradiction which he sets out to resolve.

The strongest version of the determinist thesis is due to Laplace; the first part of this review will therefore be concerned with Popper’s refutation of Laplace’s argument. I then digress briefly to discuss Popper’s ‘propensity’ theory of probability, which, although not essential to the main theme of the book, is of real interest to scientists. Finally, I discuss Popper’s more recent views on the freedom of human action. At this point, reluctantly, I find myself in total disagreement with him. I suspect that this is because my view of the world is conditioned primarily by biology, and his by physics.

Laplace, the great formaliser of Newtonian physics, argues as follows. If some universal intelligence could know all the laws of physics, and the present state of the universe (in Newtonian terms, the positions and velocities of all particles in the universe), then it could calculate the future with complete certainty. In a sense, therefore, the future already exists. To use a modern analogy, the universe – past, present and future – resembles a cine film: you and I are restricted to observing only the contemporary frames, but the future frames exist, with just the same reality as the past. Nothing we can do will alter that future: free will is an illusion.

I remember being totally persuaded by this argument when I was about sixteen, although I do not think it made me any less wilful. More recently, I have rejected it on the grounds that the knowledge, and the calculations, are in fact impossible. No calculator smaller than the universe itself could contain the necessary information. I suppose that it is true that the universe knows what it is going to do, in the sense that something is going to happen, but nothing smaller than the universe can possibly know what.

Popper offers similar but more coherent grounds for rejecting Laplace’s argument. He says that he does not regard philosophical determinism – the cine-film view of the universe – as ultimately refutable. He is concerned only to show that it does not follow from our knowledge of science. Further, it would not follow if we took an entirely Newtonian view of physics: Popper’s rejection of determinacy does not rest on the indeterminacy of quantum physics, which he is in any case unwilling to accept at face value – but that is an argument I do not want to get into. He argues that it is quite possible to hold that every event has a cause, and yet to reject Laplace’s argument.

The accuracy of one’s predictions, Popper points out, depends on the accuracy with which one knows the present. Since one can never know the present with absolute precision, Laplacean determinism requires that one answer the following question: ‘Given that I want to know the future with some specified degree of accuracy, how accurately must I know the present?’ Quite rightly, Popper argues that this question is unanswerable, because a vanishingly small difference in initial conditions can give rise to an indefinitely large difference in end-results. He quotes a mathematical result of Hadamard’s (of which I was unaware) to prove this. Today he would probably refer to the mathematics of ‘chaos’, which explains such apparently random and unpredictable phenomena as turbulence in terms of a fully deterministic model. Popper gives other reasons for rejecting the Laplacean argument, including the reasonable one that no calculator can predict its own future with certainty.

I now digress to discuss Popper’s propensity theory of probability. When I first met this (in one of the actually published appendices to The Logic of Scientific Discovery), I found it hard to believe that a man of such clarity of mind could hold such an odd view. Rereading it after twenty years, I found myself accepting it like an old friend. I can best explain Popper’s theory by first describing what might be called the old-fashioned scientist’s view of the matter. If you toss a penny, it comes down heads about half of the time (unless you happen to be Rosencrantz, or was it Guildenstern?), but any particular throw is unpredictable. However, if you measured exactly the initial position, velocity, and spin of the penny, you could calculate exactly what would happen; the initial conditions determine the outcome, in a thoroughly Laplacean manner. To this, Popper would reply: how do you account for the fact that the initial conditions, on different tosses, are such as to generate a random sequence, with probability one-half? Indeed, he might trap the old-fashioned scientist in an infinite regress: the initial conditions were as they were because their initial conditions were as they were, and so on ad infinitum. Is it not better simply to say that a penny has a propensity of one-half to give a head? This propensity is a property of the penny, just like its mass or its colour. As I say, this seemed idiotic when I first met it. Now it seems to fit rather well with the attitude I have come to accept (originally acquired, I think, from J.B.S. Haldane): that a random event is an event into whose causes it is not at present efficient to inquire. Popper would perhaps replace ‘at present’ by ‘ever’.

So far, so good. But now I come to the parting of the ways. What is the relation between philosophical determinism and the concept of human freedom? For Popper, the whole motive for wishing to dethrone determinism is to underwrite freedom. Now of course there is free will. I am writing this review of my own free will: no one made me do it. But does this mean that my action was uncaused? A psychologist might say that it was caused by my respect for Popper, by my wish to dissociate myself from his views on free will, or even by my knowledge of the state of my bank balance. But my ‘respect’, my ‘wish’ and my ‘knowledge’ are, or so I would hold, represented by physical states of my brain. If not by physical states of my brain, then how do they come to influence the movements of my hand as I write?

I think that Popper would agree that a free act, such as the writing of this review, can be caused. However, in collaboration with Eccles, he has developed a way of talking about causation which seems to me mistaken. I think it is relevant to discuss these ideas here, because they are treated in addenda to the present volume, written since 1962, and also because they form a continuation of Popper’s search for a philosophical justification for freedom. He proposes that we should recognise three ‘Worlds’. World 1 is the world of things – of pens, paper, neurones and muscles. World 2 is the world of consciousness. World 3 is the world of cultural artifacts, in their conceptual rather than their physical form. The concept of falsification is part of World 3, but the many copies of the book in which it was proposed are each part of World 1. This is fine. Popper then goes on to say that World 1 is ‘causally open’ to World 2, and that World 2 is causally open to World 3. This is also acceptable, provided we are careful. The concept of falsification (World 3) may influence the way I think about an experiment (World 2), and this in turn may influence which instruments are connected in what ways (World 1). But in that last sentence, the ‘concept of falsification’ which influences what I do is not an immaterial concept, but the physical representation of that concept in my brain, and so is ‘the way I think about an experiment’.

Popper and Eccles appear to think that concepts and feelings are causes of actions, independent of the events in the brain which embody those concepts and feelings. Now just as Popper admits that philosophical determinism (the cine-film view of the universe) is irrefutable, I admit that I cannot refute the idea that a feeling can be an immaterial cause of a material event. We must be clear what this means. As I write this review, my hand moves, because the muscles in my hand and arm contract, because neurones descending from my brain and spinal cord carry impulses which cause them to contract. The Popper-Eccles view, if I understand it (and if I do not, the fault is not entirely mine), is that at some point this chain of material causation is broken, and something immaterial causes a neurone to fire.

There is no way of showing that this is not so. Why, then, do I object to supposing that it is? Fundamentally, because I would like to know how brains work, and if one once admits that the interesting things that brains do are not caused by material events, the question becomes unanswerable. I object because I think Popper’s views are defeatist. I suggested earlier that my disagreement with Popper arises because I am a biologist. I am not altered in this opinion by the fact that Eccles is a distinguished neurobiologist, because it is an occupational risk of biologists to claim, towards the end of their careers, that problems which they have not solved are insoluble. But I think that Popper is sometimes too ready to treat as insoluble problems I would like to see solved. This is most vividly illustrated by his remark: ‘It seems reasonable to regard the emergence of consciousness and previously that of life as two comparatively recent events in the evolution of the universe; as events which, like the beginning of the universe, are at present, and perhaps for ever, beyond our scientific understanding.’ I agree that we are not making much headway with consciousness right now, but I hope to live to see the solution to the problem of the origin of life.

[*] The Postscript to the Logic of Scientific Discovery by Karl Popper, edited by W.W. Bartley. Vol. I: Realism and the Aim of Science (Hutchinson, 420 pp., £20, 21 March, 0 09 151450 9); Vol. II: The Open Universe: An Argument for Indeterminism; Vol. III: Quantum Theory and the Schism in Physics (Hutchinson, 229 pp., £15, 26 July 1982, 0 09 146170 7).