(a) What Scientists Do
(b) Philosophical Problems Arising out of Natural Science
(c) Scientific Laws
(d) Scientific Objects
(e) Probability
(f) Determinism and Indeterminacy
(g) The Value and Danger of Science
(a) What Scientists Do — Having formed some idea as to the nature of reason, we must now consider the scope and limitations of actual human reasoning. In our day reason's most spectacular achievement is natural science. How does the scientist go about his work? What sort of truth can he tell us?
It may be objected that these questions concern science more than philosophy. But philosophy is concerned with every subject, or a special aspect of every subject. Certainly it has much concern with science. Some modern philosophers go so far as to define philosophy as "the logic of the sciences" Without agreeing with this limitation of philosophy, we must agree that philosophy at any rate involves a study of the logical basis and structure of science.
What, then, does the scientist do? All human activity springs from complicated motives. The guiding motive of any particular scientific worker probably includes, along with sheer intellectual curiosity, such ulterior motives as the will to shine in his profession, the will to serve the community, and (in capitalistic societies) the urgent need to secure a livelihood by selling his skilled labour as dearly as possible.
For one reason or another a scientist's attention is directed to a particular science, such as physics or biology, and to some highly specialised field of study within his chosen science, such as the breaking-point of metals, or the inheritance of characters in cereals. Most scientific work to-day is very highly specialised. All the more obvious fields of research have already been at least roughly and often minutely mapped, and a subtle technique, appropriate to a special field, equips the worker for enterprises which formerly would have been quite impossible.
Let us consider the form of that technique so far as it is common to all sciences. Let us take as an example the formulation of the law of gravity. When things are let go; they fall. How fast do they fall in varying circumstances? Does their weight make any difference to their speed? Pioneering, the scientific mind made a vast number of observations of falling bodies, and devised a mathematical formula which would enable the behaviour of future failings to be predicted. It was found that they moved, and might be expected to move, with an acceleration of thirty-two feet per second every second. The colour, temperature, odour, etc., of the falling bodies were found to be irrelevant. Their weight and shape were relevant only in relation to air-resistance, and irrelevant to gravitation itself.
We may summarise the nature of all scientific enquiry as follows. Whatever his ulterior motives, the scientist's immediate aim is to describe how things happen in his particular field of enquiry. He wants his description to be as simple and handy as possible, and as coherent as possible with other scientific descriptions. He seeks some principle, or preferably some precise mathematical formula, in terms of which he can explain his problem, or rather describe his data. But first he must procure clear and significant data. He therefore analyses the crude facts, distinguishing between those that seem relevant and those which seem irrelevant. He discovers how to make crucial observations and, if possible, experiments, to help him to get a clear view of what actually happens. Whenever possible he measures the significant factors in his data. Factors which seem not to be significant for his purpose he simply ignores. He imagines hypothetical descriptions, or hypothetical laws, and tries these out; until at last he discovers one which compendiously describes the whole mass of data, and enables him to predict the future course of events.
(b) Philosophical Problems Arising out of Natural Science — This procedure confronts the philosopher with a number of problems. What precisely is a scientific law? Clearly, as' we have already seen, it is not a law with binding force. There is no "must" about it. At most it describes how events are observed to happen. But if this is so, by what right do we use the law for prediction of future events? Thus we raise the problems of the validity of inductive reasoning, the nature of causation, of probability, and the issue between determinism and indeterminism.
Another very difficult problem is raised. How far is the method of analysis reliable? How far can we discover the truth about natural events by analysing them, and ignoring all those aspects which seem irrelevant? Thus we come once more upon the question of the scope and danger of abstraction. We also encounter the issue between pluralism and monism. Which is the more significant and useful view, that the world consists of many independent things in relation, or that it consists of one thing, and is a seamless whole, such that nothing can be truly said about its parts without reference to the whole? Some, but not all, of these problems we shall discuss in this chapter.
(c) Scientific Laws — We have already seen that scientific laws are not binding laws. There is no necessity in them. For all we know, they may be violated at any moment. They are at best descriptive. Some philosophers hesitate even to allow that they are descriptions, in the ordinary sense of the word, for the following reasons.
The observations from which a law is derived are, of course, erratic. Instruments that measure time and space are never perfectly accurate. The manipulating and observing experimenter himself introduces further complications. Strictly, the law derived from the observations does not describe the actual data but a simplified principle to which the data, taken as a whole, approximate. The law, in fact, is a sort of graph, near which all the past data fall, and all future data may be expected to fall.
The Logical Positivists, bearing this in mind, insist that a scientific law is not really a proposition about a set of data; for it is not a proposition at all but only a formula by means of which propositions about actual events may be constructed. They say this because they are anxious not to attribute semi-mystical "principles" to nature. Rightly they seek to avoid thinking of abstractions, such as gravity, as mysterious "things" or "spirits" controlling nature. Rightly they insist that a scientific law is more like a rule of grammar than a sentence. It is a human dodge for simplifying description. Other dodges might work equally well.
But surely there is an important difference between a mere formula and a formula that is a scientific law. The law, after all, is derived from events, and is predictive of events. As such, it is descriptive of nature, in the sense that it describes not particular events but a set of relations between certain kinds of events. In fact, it describes a complex universal character. Of course, if universals are nothing but the names we "give" them, then a scientific law is nothing but a complicated word. But if, as I have maintained, universals have real being as "distributive unities," then a scientific law is actually a description of a universal character inherent in a large class of events.
The fact that scientific laws can be true or false, that they can be tested in sense-experience, shows that they really are, in some sense, descriptive of nature. The fact that there may be different and equally good, or even better, ways of formulating laws raises no more difficulty than the fact that " It rains " and "Il pleut" are equally good descriptions of a certain kind of natural event. These statements are no less true, though less precise, than the statement that H2O, in drops of a certain size and frequency, is descending on the earth.
When Newton, in a flash of creative imagination, guessed that there was a connection between the laws descriptive of falling bodies and of the movements of the planets, he set about testing this hypothesis by further observations and calculations; and discovered that his original formula did, in fact, describe the principle common to both sets of events. When Einstein, intrigued by certain minute discrepancies between prediction and observation, devised a much more subtle formula to comprise much more than gravitation, he did not overthrow Newton's law. He merely invented a more exact "language" by which to describe more precisely what Newton's language had described less precisely. Both laws, however, are descriptive of nature. But Einstein's is the more precise and comprehensive description.
(d) Scientific Objects — So much for scientific "laws." What of scientific "objects," such as electrons, protons, neutrons, positrons? Are they to be regarded as real factors in nature or as mere formulae, useful for scientific prediction? Obviously very little is known about them. They are mere calculable potentialities for affecting our instruments. An electron, for instance, is (in this view) a very abstruse formula descriptive of a very subtle "permanent possibility of sensation." It is a mere system of probability. We can assign to it no quality known to us. The little that we do know of it is often self-contradictory. An electron is apparently to be conceived as at one and the same time a particle and a system of waves. Nevertheless, the logical status of scientific objects is at bottom the same as that of ordinary unperceived physical objects, such as the earth's metal core, or the stony centre of Cleopatra's Needle, or a man's own brain. If these are real factors in nature, so are electrons. The only difference is that our knowledge of scientific objects is reached by a much more indirect method, is far less detailed, and cannot be accurately conceived in terms of familiar sensory characters.
On the other hand, if scientific objects are mere. formulae, useful for prediction of perceptible events, but not to be regarded as objective entities, then ordinary unperceived physical objects must be regarded in the same way. Not only so, but perceived physical objects, too, though of course not pure sense data, must be regarded as mere formulae, useful in action, but no more.
This view we have rejected. In doing so we pledge ourselves also to a realistic view of scientific objects.
(e) Probability — It is fairly clear that scientific laws are compendious descriptions of past sensory experience, or at the very least formulae from which such descriptions can be derived; but by what right do we use them also for prediction of future sensory experiences?
It used to be said that the first assumption of all science was the "uniformity of nature," the conviction that, wherever and whenever events occur, the same fundamental physical laws must hold good of them. To-day it would rather be said that though the scientist hopes for and seeks regularity, he makes no assumption that it must exist. An immense amount of regularity has been discovered and is found to hold good from day to day. But we know no reason, inherent in the nature of things, why this regularity should continue. At any moment gravitation might cease, or the sky might roll back and reveal the Celestial City, or sheer chaos might supervene.
We have a strong expectation that none of these things will happen. The "probability" of their happening, we say, is infinitely small. What is this " probability"? Is it simply the degree of the intensity of our sense of expectation, or rather of the strength of our mental habit of expectation, which becomes more and more insistent the more often a familiar sequence of events is experienced? Or is probability in some manner objective in nature?
Sometimes probability can actually be calculated and assigned a percentage. In dice-throwing we can easily calculate the probability that the six will turn up so I many times out of so many throws. Put to the test of experience, the prediction proves the more accurate the greater the number of throws. If it were to fail completely, if the six were to turn up much more often than we expected, we should at once infer that some special influence was at work. Perhaps the dice might be biased, or the throw itself nicely calculated to turn up sixes.
If we knew all the relevant data for any particular throw (the centre of gravity of each of the dice, the initial position of both, the strength and direction of the movement, and so on) we could predict the result of that throw without leaving any more to probability than is left in all statements about the world of fact. As it is, we know only (let us say) that the dice are not appreciably biased, and that the throws are genuinely haphazard. Each side of the die, we say, has as good a chance of turning up as any other. Since there are six sides, each side has one chance in six for any particular throw. The probability is that, out of every six throws of one die, one throw will produce the coveted side with six dots on it. This statement is clearly not simply a statement about our expectation. Whatever anyone expects, the statement is in some sense true. Yet it is not in the ordinary sense a statement of fact. Actually the six might turn up six times in succession, or not at all in a score of throws. Then what is the statement about?
It is a logical statement about the implications of a hypothesis or definition. If the die is unbiased, and the throw is random, and if the accepted principles of dynamics still hold good, then no side has the advantage. The reasoning is "necessary" in this hypothetical sense. But there is no observable necessity in its application to any particular group of throws. Indeed, strictly it does not by itself apply to particular throws at all, since it is incomplete. In every particular case the issue is determined strictly by the dynamics of the situation. But the formula is useful, because over a large number of throws the idiosyncrasies cancel out. So long as the conditions hold good, the formula is a true description of a universal principle which has had instances and may have others.
On the face of it there is a great difference between the probability that a six will turn up in a particular throw and the probability that the laws of dynamics themselves, or any natural laws, will hold good. The one probability can be calculated, the other not. And in the one case possible interferences can at least be conceived and studied; in the other not. But the underlying principle is identical in both cases.:fn each, certain factors are known, others are not known. In the case of the die, what is demanded is prediction of a particular result, and for this prediction the known factors are insufficient. Only a general principle can be established. In the case of a natural law, a general principle is all that is demanded; and for this the knowledge that we have has proved adequate. But, of course, in both cases the great unknown makes certainty impossible.
(f) Determinism and Indeterminacy — In the nineteenth century the growth of rationalism combined with the success of science to suggest that all physical events were connected together in one great causal system. Every physical event was regarded as a necessary effect of preceding events and a cause of succeeding events. Mental events in human minds were thought either to be links of a non-physical kind in the causal chains or to be mere consequences of purely physical causation. They were supposed not to be themselves causally efficient.
Although in the physical sciences determinism was generally accepted, in the biological sciences a long- drawn-out war raged over it. The usefulness of organs and of modes of behaviour strongly suggested that in some way purpose was a controlling factor in biological causation. The determinists clung to the concept of mechanism, and declared that natural selection was enough to explain the process of evolution. The vitalists insisted that natural selection was negative, and that some positive and teleological or purposeful drive, some "entelechy," or "élan vital," was obviously at work.
Into this controversy we need not now enter. All that we need do is to try to see clearly what is at stake. The issue can be stated in terms of purely descriptive law, without any reference to underlying forces) whether physical or teleological. Are there, or are there not, some sequences of biological events which cannot even in theory, even if we had all the relevant data, be fully described by the formulae of physical mechanism, and which in fact involve a teleological infringement of the purely mechanical course? To use an analogy, are there points at which the stream, instead of taking the line of least physical resistance, actually gathers itself together and leaps over barriers? And are we justified in holding that these leapings can be described only by reference to a goal?
The issue of the controversy must be left to the scientists. Perhaps, like so many controversies, it will be decided not by the victory of one side but by the discovery that the alternatives have been wrongly conceived, so that neither is true and neither is false.
Let us note, however, that even if the teleological view is correct, determinism (though not, of course, mechanism) might still hold good. Particular events, though not determined solely by preceding physical events, might still be determined. They might still occur systematically in relation to determining factors. They might show a teleological bias that was constant and regular; and in relation to this bias prediction of future events might still be possible, in the manner in which a man's act may, up to a point, be predicted from knowledge of his purpose.
On the whole it is probably fair to say that though mechanical descriptive laws have proved increasingly useful in biological research, the issue between teleology and mechanism is not yet decided. The steady advance of biochemistry strongly suggests that in time all biological phenomena will be accounted for in terms of physical mechanism. On the other hand, it may also turn out that thoroughgoing mechanism in the abstract field of the biological sciences is not, after all, incompatible with teleology in more concrete studies.
Strangely, while the biological sciences have tended to provide increasing evidence of determinism and even of mechanical determinism, physics itself has been shaken by a serious attack of "indeterminacy." It would be folly on my part to pretend that I have more than a superficial understanding of this scientific controversy. Consequently the reader must take my comments on its philosophical aspect as .merely a starting-point for further study. If he wishes to pursue the matter he should read, not only the popular works of Eddington and Jeans, but the penetrating criticism of them in Professor Susan Stebbing's Philosophy and the Physicists.
The trouble seems to have had two sources. One, we are told, was the complete failure to find any reason why an electron should change its orbit at one time rather than another; the second source of difficulty lay in the discovery that in principle there was no possibility of knowing both the position and the velocity of an electron in its orbit. If one was known, the other was in principle unknowable.
The common-sense reaction to these troubles was simply to attribute them to our ignorance. If we knew enough, it was said, we should be able to predict the electron's leap; and we should be able to correlate its speed and its position.
But the eminent physicists pointed out that this was sheer assumption. We were so accustomed to discover system in nature that we irrationally took it as certain that system must hold throughout. When at last we stumbled upon a fundamental arbitrariness in physical events, we could not recognise it, but regarded it merely as a case of veiled determinism. Instead, we should have recognised that, after all, at bottom nature was not systematic. The ultra-microscopic events within the atom contained a factor of sheer arbitrariness. No doubt in the mass, in "macroscopic" physics, these arbitrary events average out and yield the systematic, predictable events from which the theory of determinism was derived. But when we look more minutely into the matter, determinism (they said) turns out to be illusory.
To enforce their argument the opponents of determinism cited the analogy of life-insurance. The actuary is able to predict that so many people of a given age will die each year, though the death of any individual is unpredictable. From a host of accidents a statistical law of probability emerges, by means of which prediction is possible.
Some have found in this supposed arbitrariness of physical nature an argument for free will in human beings. The bogey of physical determinism, they say, is destroyed. If physical events themselves are at bottom arbitrary, they cannot impose determinism on the mind. This, however, is a very unconvincing argument. A man's behaviour consists of physical events of the "macroscopic," not the microscopic, order; and therefore, even according to this theory, should be subject to the determinacy of "macroscopic" physics. Putting the matter very crudely, we may say that what the champions of free will must establish is not that the individual electrons in a brain have "free will" but that the single mind of the man has "free will."
But quite apart from the question of free will, what bearing have these arguments on the problem of determinism in physical nature? From the point of view of common sense the fact that there is system on the "macroscopic" physical plane seems to imply system. also on the ultra-microscopic plane, even if we cannot yet discover the laws of that system. The analogy of the actuary was misinterpreted. His generalisations would not hold good unless the individual deaths, though unpredictable, were as a matter of fact systematic. Generalisations about deaths from road accidents, diseases, and suicide would be impossible if the individual deaths were not in fact determinate instances of general principles — physical, biological, psychological, social. Similarly, if the behaviour of electrons was really indeterminate in detail it would prove indeterminate also in the mass. And whatever is the truth about the behaviour of individual electrons, it is certainly true that in the mass, or on the "macroscopic" scale, their behaviour is determinate in the only sense in which any matter-of-fact is ever determinate, namely, that in very many cases it can be predicted and subsequently verified with great precision. Of course, there is no discoverable logical necessity in their behaviour, or in any actual events. But science has established a huge system of exact statistical laws about their behaviour; and, though these laws are not necessary, they have an almost infinite degree of probability.
Much confusion arises from the ambiguity of the words "determined" and "determinism." If determinism involves logical necessity, then clearly we have no right to say that physical events are determined, since we know of no logical necessity in the sequence of events. Even if determinism involves merely causal necessity, we have now, according to Professor Stebbing, no right to attribute determinism to physical events, since in the microscopic foundations of physics causal laws have given place to statistical laws, necessity to probability. (But surely this is nothing new.) If, on the other hand, determinism involves merely determinate or systematic or regular behaviour, then the new developments of physics do not disprove determinism, since on the macroscopic level and even on the sub-atomic level there is an immense amount of regularity and predictability. It is important to emphasise this point since the works of Eddington and Jeans tend to give a different impression. As Professor Stebbing has pointed out, the new concepts of physical science do not show that there is anything indeterminate or arbitrary in physical nature. There is nothing lawless in the basic phenomena of physics.
The upshot seems to be that recent developments of physics have no special bearing on the philosophical problem of determinism. Independently of these developments it is recognised that all scientific laws are descriptive laws, not necessary laws. They describe observed regularities in the spontaneous course of events. At most they can only suggest a determinism which can never be proved. Sub-atomic physics does nothing to diminish the suggestion.
(g) The Value and Danger of Science — It is obvious that natural science has given man extensive knowledge and great powers. It is equally obvious that those powers have been used unwisely; and that the knowledge which science has given has in some important respects led not to wisdom but to blindness, folly, destruction, and grave peril to civilisation.
The method by which science went to work was that of attending to those aspects of the world which could most easily be observed with accuracy, and ignoring the rest. Roughly, it studied the movement of material things, and whatever was clearly related with movement. It ignored "secondary qualities," such as colour and sound, save as symptoms of movement. It also ignored mental facts, such as desiring.
Thus in time was built up the amazingly complex system of the physical sciences; and, along with this, industrial power. Meanwhile, with high confidence in his new explorative technique, man applied the concepts which had proved so useful in the study of lifeless matter to the study of living matter and of mind. By observation and analysis he strove to single out the determining factors of vital and of mental behaviour, with the expectation that these could be explained in terms of the laws of matter in motion. He succeeded at least to the extent of discovering many important and unexpected ways in which behaviour depended on obscure physical factors in the body or in the environment. It seemed clear that in time the dream of the materialist would be fulfilled, and everything would be thus explained.
I shall consider Materialism in more detail in a later chapter. Meanwhile, we must note that the theoretical and practical triumphs of physical science led to an unjustified confidence in it as a key to the metaphysical understanding of the universe.