Read an Excerpt

Science, Faith And Society

The University of Chicago Press

SCIENCE AND REALITY

I

What is the nature of science? Given any amount of experience, can scientific propositions be derived from it by the application of some explicit rules of procedure? Let us limit ourselves for the sake of simplicity to the exact sciences and conveniently assume that all relevant experience is given us in the form of numerical measurements; so that we are presented with a list of figures representing positions, masses, times, velocities wavelengths, etc, from which we have to derive some mathematical law of nature. Could we do that by the application of definite operations? Certainly not. Granted for the sake of argument that we could discover somehow which of the figures can be connected so that one group determines the other; there would be an infinite number of mathematical functions available for the representation of the former in terms of the latter. There are many forms of mathematical series—such as power series, harmonic series, etc.—each of which can be used in an infinite variety of fashions to approximate the existing relationship between any given set of numerical data to any desired degree. Never yet has a definite rule been laid down by which any particular mathematical function can be recognized, among the infinite number of those offering themselves for choice, as the one which expresses a natural law. It is true that each of the infinite number of available functions will, in general, lead to a different prediction when applied to new observations, but this does not provide the requisite test for making a selection among them. If we pick out those which predict rightly, we still have an infinite number on our hands. The situation is in fact only changed by the addition of a few more data—namely, the 'predicted' data—to those from which we had originally started. We are not brought appreciably nearer towards definitely selecting any particular function from the infinite number of those available.

Now, I am not suggesting that it is impossible to find natural laws; but only that this is not done, and cannot be done, by applying some explicitly known operation to the given evidence of measurements. And to bring my argument a little closer to the actual experience of science, I shall now restate it as follows. We ask: Could a mathematical function connecting observable instrument readings ever constitute what we are accustomed to regard as a natural law in science? For example, if we were to state our knowledge concerning the path of a planet in these terms: 'That setting certain telescopes at certain angles at certain times a luminous disc of a certain size will be observed'—does that properly express a natural law of planetary motion? No: it is obvious that such a prediction is not equivalent to a proposition concerning planetary motion. Firstly, because we will in general be claiming too much and our prediction will prove often false even though the underlying proposition on planetary motion was correct: for a cloud may make the planet invisible to the eye, or else the soil may give way under the observatory, or some other of a hundred and one possible errors or obstacles may falsify observation or make it unworkable. Secondly, we would be claiming too little, since the presence of a planet at certain points of space—as postulated by its law of motion—may manifest itself in an indefinite variety of ways, the majority of which could not, on account of their sheer multitude, ever be explicitly predicted; and many of which may even be unthinkable to-day as they may be due to arise from yet unknown properties of matter or a host of other factors unknown at present, though inherent in our system.

There is, in fact, an essential feature lacking in both of the foregoing representations of science, which can be perhaps best pointed out by using yet a third picture of science. Suppose we wake up at night to the sound of a noise as of rummaging in a neighbouring unoccupied room. Is it the wind? A burglar? A rat? ... We try to guess. Was that a footfall? That means a burglar! Convinced, we pluck up courage, rise, and proceed to verify our assumption.

Here are some of the features of a scientific discovery that we had missed before. The theory of the burglar—which represents our discovery—does not involve any definite relation of observational data from which further new observations can be definitely predicted. It is consistent with an infinite number of possible future observations. Yet the theory of the burglar is substantial and definite enough; it may even be capable of proof beyond any reasonable doubt in a court of law. In the light of common sense there is nothing curious in this: it merely makes it clear that the burglar is being assumed to be a real entity; a real burglar. So that we may even reverse this by saying that science is assuming something real whenever its propositions resemble the theory of the burglar. In this sense an assertion concerning the path of a planet may be said to be a proposition concerning something real, it being open to verification not only by some definite but also by many as yet quite undefined observations. We often hear of scientific theories gaining confirmation by later observations in a manner described as most surprising and audacious. The feat of Max v. Laue (1912) jointly confirming by the diffraction of X-rays in crystals both the wave nature of the X-rays and the lattice structure of crystals, is often praised as a striking feat of genius. It appears of the essence of scientific propositions that they are capable of bearing such distant and unexpected fruit; and we may conclude, therefore, that it is also of their essence to be concerned with reality.

A second significant feature of the discovery of the burglar, closely connected with what has just been said, is the way in which it is made. Curious noises are noticed; speculations about wind, rats, burglars, follow, and finally one more clue being noticed and taken to be decisive, the burglar theory is established. We see here a consistent effort at guessing—and at guessing right. The process starts with the very moment when, certain impressions being felt to be unusual and suggestive, a 'problem' is presenting itself to the mind; it continues with the collection of clues with an eye to a definite line of solving the problem; and it culminates in the guess of a definite solution.

But there is a difference between the solution offered by the burglar theory and that offered by a new scientific proposition. The first selects for its solution a known element of reality—namely burglars—the second often postulates an entirely new one. The vast growth of science in the last 300 years proves massively that new aspects of reality are constantly being added to those known before. Whence can we guess the presence of a real relationship between observed data, if its existence has never before been known?

We must go back to the process by which we usually first establish the reality of certain things around us. Our principal clue to the reality of an object is its possession of a coherent outline. It was the merit of Gestalt psychology to make us aware of the remarkable performance involved in perceiving shapes. Take, for example, a ball or an egg: we can see their shapes at a glance. Yet suppose that instead of the impression made on our eye by an aggregate of white points forming the surface of an egg, we were presented with another, logically equivalent, presentation of these points as given by a list of their spacial co-ordinate values. It would take years of labour to discover the shape inherent in this aggregate of figures—provided it could be guessed at all. The perception of the egg from the list of co-ordinate values would, in fact, be a feat rather similar in nature and measure of intellectual achievement to the discovery of the Copernican system. We can say, therefore, that the capacity of scientists to guess the presence of shapes as tokens of reality differs from the capacity of our ordinary perception, only by the fact that it can integrate shapes presented to it in terms which the perception of ordinary people cannot readily handle. The scientist's intuition can integrate widely dispersed data, camouflaged by sundry irrelevant connexions, and indeed seek out such data by experiments guided by a dim foreknowledge of the possibilities which lie ahead. These perceptions may be erroneous; just as the shape of a camouflaged body may be erroneously perceived in everyday life. I am concerned here only with showing that some of the characteristic features of the propositions of science exclude the possibility of deriving these by definite operations applied to primary observations; and to demonstrate that the process of their discovery must involve an intuitive perception of the real structure of natural phenomena. In the rest of this lecture I shall examine this position further and also point out (in section v) the necessity of amplifying it in some important respects.

II

However, would it not seem that our daily experience compels us with the force of logical necessity to accept certain natural laws as true? Generalizations such as 'all men must die' or 'the sun sheds daylight' seem to follow from experience without any intervention of an intuitive faculty on the part of ourselves as observers. But this only shows that we incline to regard our own particular convictions as inescapable. For these generalizations are quite commonly denied by primitive peoples. Such people believe that no man ever dies, except as a victim of evil magic, and some of them also believe that the sun crosses back by night to the east without shedding any light in its course. Their denial of natural death is part of their general belief that events which are harmful to man are never natural, but always the outcome of magic wrought by some malevolent person. In this magical interpretation of experience we see some causes which to us are massive and plain (such as a stone smashing a man's skull) regarded as incidental or even irrelevant to the event, while certain remote incidents (like the passing overhead of a rare bird) which to us appear to have no conceivable bearing on it are seized upon as its effective causes.

The primitive peoples holding these magical views are of normal intelligence. Yet they not only find their views wholly consistent with everyday experience, but will uphold them firmly in the face of any attempt on the part of Europeans to refute them by reference to such experience. For the terms of interpretation which we derive from our intuition of the fundamental nature of external reality cannot be readily proved inadequate by pointing at any particular new element of experience.

We are thus, it would seem, in danger of the opposite extreme: namely, of losing sight of any difference between the rival claims of the magical and the naturalistic interpretations of events. Now, it is true that there is a poetic truth expressed in primitive magical theory which is commonly found in our works of fiction. If a man in a novel is killed by accident, the event must have some human justification; the question of the Bridge of San Luis Rey can never be disregarded in a work of art. The naturalistic view of a man's death, say by a rail accident, robs human fate of some of its proper meaning; tending to reduce it to 'a tale told by an idiot, signifying nothing'. But at the same time the naturalistic view opens such a noble vista of the natural order of things which are inaccessible to the magical view, and establishes so much more decent and responsible relationships between human beings, that we must not hesitate to accept it as the truer of the two.

A similar competitive conflict comes into view in contrasting the medieval and the scientific outlook. It is usually overlooked that medieval catholic philosophy was first established in a world imbued with scientific rationalism. St. Augustine, who above all laid the foundations of catholic philosophy, testifies amply in his Confessions to his profound interest in science before his conversion. But as he approached conversion he came to regard all scientific knowledge as barren and its pursuit as spiritually misleading. The battle which round the year 380 was fought in Augustine's mind was won by his fervent desire for a certainty of God which he felt to be endangered by the intellectual pride of men pursuing the chain of second causes. 'Nor doest Thou draw near', he wrote, 'but to the contrite in heart, nor art found by the proud, no, not though by curious skill they could number the stars and the sand, and measure the starry heavens, and track the course of the planets' (Conf., bk. v, p. 3).

Eleven hundred years later we see St. Augustine's spell broken in its turn by a gradual change in the balance of mental desires. The secular spirit, critical, extrovert, rationalist, spread into many other fields before it revived the scientific study of nature. Science was a late child of the Renaissance; in fact by the time of Copernicus' and Vesalius' discoveries, the Renaissance had passed its peak and was falling under the shadow of the Counter-Reformation. Both Copernicus and Vesalius discovered new facts because they abandoned established authority—and not the other way round. Copernicus was affected by the new spirit while studying canon law at Italian universities around the year 1500. He returned home from Italy where so-called Pythagorean doctrines were then freely discussed, in strong and irrevocable possession of the heliocentric view. When Vesalius first examined the human heart and did not find the channel through the septum postulated by Galen, he assumed that it was invisible to the eye; but some years later, with his faith in authority shaken, he declared dramatically that it did not exist.

And I think that to-day we can feel the balance of mental needs tilting back once again. Science is not so emphatic any more in disregarding how far its generalizations make sense when extended to the world as a whole. It is doubtful whether to-day scientists would accept without murmur, as they still did at the end of the nineteenth century, a view like that of Laplace and Poincaré about the nature of the universe. Poincaré had shown that from Laplace's mechanical theory there followed that every phase of atomic configuration must go on recurring cyclically to infinity and that every conceivable configuration (of the same total energy) keeps recurring likewise—so that on revisiting our universe one day we may have a chance of finding ourselves going through life once more, but this time in the reverse direction starting with a revival of our dead bodies and ending our lives as babies, eventually to be absorbed by the maternal womb. To-day, I believe such manifestly absurd conclusions would be seriously held against a scientific system which ventured to put them forward. In fact the modern study of cosmogony has involved—as Sir Edmund Whittaker has pointed out in his Riddell Lectures of 1944—a renewal of interest in the universe as one comprehensive whole. Moreover, since the advent of relativity, scientists have become increasingly confident that natural laws can be discovered by a systematic elimination of unwarranted assumptions implied in our way of thinking and this has strengthened our sense of rationality in the universe.

We conclude that objective experience cannot compel a decision either between the magical and the naturalist interpretation of daily life or between the scientific and the theological interpretation of nature; it may favour one or the other, but the decision can be found only by a process of arbitration in which alternative forms of mental satisfaction will be weighed in the balance. The foundations of such decisions will be ascertained in my third lecture. Now I return to the analysis of science.

III

The part played by new observations and experiment in the process of discovery in science is usually over-estimated. The popular conception of the scientist patiently collecting observations, unprejudiced by any theory, until finally he succeeds in establishing a great new generalization, is quite false. 'science advances in two ways,' remarks Jeans, 'by the discovery of new facts, and by the discovery of mechanisms or systems which account for the facts already known. The outstanding landmarks in the progress of science have all been of the second kind.' As examples he quotes the work of Copernicus, Newton, Darwin, and Einstein. We could add Dalton's atomic theory of chemical combination, de Broglie's wave theory of matter, Heisenberg's and Schrödinger's quantum-mechanics, Dirac's theory of the electron and positron. In a number of these discoveries predictions of the highest importance were involved which often came to light only years after the discovery was made. All this new knowledge of nature was acquired merely by the reconsideration of known phenomena in a new context which was felt to be more rational and more real.

---

Excerpted from Science, Faith And Society by Michael Polanyi. Copyright © 1946 Michael Polanyi. Excerpted by permission of The University of Chicago Press.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

---