THE COMMON SENSE OF SCIENCE

(1)

We have now crossed the tangled uneven fields of science at several places. More than this, we have at critical points mined below the surface for the solid strata on which that rests. I am using the images of exploration and of search, and they are just, because this halting step-by-step into the sense of science is a voyage of discovery. And the dimension for this voyage is time. Like the voyages of the Spaniards into the fabulous West, science even at its boldest does the will of history, and in turn helps to determine its movement. Like civilization and like our societies, it exists in the larger setting of history: it does not exist, it grows. Civilization is less than ten thousand years old: in this moment of time man has created the world we know, from Ur to Radio City, and from Confucius and Pythagoras to Rabelais and Einstein. And in that short and spirited adventure science fills a still smaller moment.

Science as we know it indeed is a creation of the last three hundred years. It has been made in and by the world that took its settled shape about I66o, when Europe at last shook off the long nightmare of religious wars and settled into a life of inquisitive trade and industry. Science is embodied in those new societies; it has been made by them and has helped to make them.

The medieval world was passive and symbolic; it saw in the forms of nature the signatures of the Creator. From the first stirrings of science among the Italian merchant adventurers of the Renaissance, the modern world has been an active machine. That world became the everyday world of trade in the seventeenth century, and the interests of science were appropriately astronomy and the instruments of voyage, among them the magnet. A hundred years later, at the Industrial Revolution, the interest shifted to the creation and use of power. This drive to extend the strength of man and what he can do in a day's work has remained our interest since. In the last century it moved from steam to electricity. Then in 1905, in that wonderful year when at the age of twenty-six he published papers which made outstanding advances in three different branches of physics, Einstein first wrote down the equations which suggested that matter and energy are interchangeable states. Less than fifty years later, we command a reservoir of power in matter almost as large as the sun, which we now realize manufactures its heat for us in just this way, by the annihilation of its matter.

These great historical movements must underlie everything that can be said about science. We should be proud of their share in science, and of the share which science has had in them.. And in them the actual influence, the interpenetration of all our actions goes deeper than the mere surface of society, the radar screen, the indirect heating and the vitamin pill of our century, or the white bread, the leather shoes, the cotton dress, and the iron bedstead of the Industrial Revolution. Science has entered into the life and structure of society, so that the man who makes a living in a kitchen garden

in Kent and the man who draws strips about blonde heroines in space-ships can be seen equally to owe their market to our technical society. And if the one is not allowed to employ boys of ten, and the other must spice his cartoons with glib and sexy tortures, that sensibility, good and bad, is largely the creation of science. Human life is social life, and there is no science which is not in some part a social science.

For this reason, I have looked at the ideas of science always in the setting of their times. From year to year they grow larger until at last the outline is quite changed. And the growth does not go on in empty space, it does not even go on in an abstract space where there is nothing but ideas. It goes on in the world, the rational and empirical world. The mastery and the greatness of science rests in the end on this, that here the rational and the empirical are knotted together. Science is fact and thought giving strength to one another.

(2)

But we ought at last to make a map of the land we have explored; and here it is time to put aside history and the other aids to navigation. Even in following the stages by which science has grown, we have come in the last chapters more and more to ask, What is the basis of scientific method today? And it is now time to sum up what we have found when our journey has brought us to the present.

The map we are making is as it were a geological map; it plots the strata on which our technical skill rests. For the skill of head and hand go together. As the instrument makers and engine builders of the eighteenth century showed, our understanding of nature can only be as accurate as the machine parts with which we explore and control her. And equally, as the whole progress-of quantum physics has shown from the first equations of Max Planck in 1899 to the atomic piles of today, our technical success rests on skill and boldness of mind in thinking through the implications of experiment with no regard to our habits of philosophy-whether these habits are skeptical or materialist.

Whether they are skeptical or materialist, these habits are founded deeply in the way we have long come to think that science must conceive the real world. We are all aware, although we rarely think about it, that all human forethought depends on our recognizing or putting some kind of order into the world. As much as book-keeping, government and doing the week-end shopping, science is an activity of putting order into our experience. So much was true even of the science of Aquinas. To this was added in the sixteenth and seventeenth centuries a new assumption about the kind of order which science sets out to find or make. Roughly, the assumption amounts to this, that science is to get rid of angels, blue fairies with red noses, and other agents whose intervention would reduce the explanation of physical events to other than physical terms. The world is regular in itself; the world is a machine.

In order to simulate the workings of this machine, we usually describe a model made of simple units and obeying simple laws whose motions are then shown to take it to just those points in time and space where experiment can check it against the physical world. It does not matter whether this model is made with pulleys and springs and cathode tubes whose behavior has become familiar to us, or whether it is simply an array of equations to be solved. Either is a model. The true essence of the model is that it is an axiomatic construction like that of Euclid. It postulates the world built out of repeated units, atoms or cells or reflexes, which obey defined laws and whose behavior is then simply the action of these laws through time.

Finally we have come to take it for granted that these laws must have very much the shape of Euclid's axioms. Euclid's axioms determine what happens when you draw a given configuration of lines, and they determine it precisely and once for all. If three lines are drawn which meet by pairs in three different places, then they enclose a triangle and its angles add up to one hundred and eighty degrees. They do not enclose a triangle half the time, and something else the other half. The angle is not one hundred and eighty degrees seven times out of ten, and something else three times. And the angle is not nearly a hundred and eighty degrees, within some area of uncertainty. In Euclid's world, everything happens as has been foretold. Or so mathematicians thought, until the recent flutter caused by the existence of theorems which cannot be proved to be either true or false. Of course, Euclid's world happens to contain no time; and this is a difference of far-reaching meaning. Nevertheless, we have grown accustomed for three hundred years to think all laws like his: precise, determinate, and invariable. In a world with time in it, they are causal laws. And these are the laws which we have thought to be of the essence of science.

That background I have marked in detail in this book. And I have added to it in detail again another kind of law which can be put into a working world. This world will still be orderly, a machine, and it can have a model, although it need not. It differs essentially in being moved by laws which have a different form: the form of chance in place of cause and effect. But in the map which we are now making, we should look deeper than this. We should look below the differences in method, to their origin in the nature of science as we now see its. What is the nature of science? That is our question in this chapter. From its answer the new methods of science must be seen to grow directly. And here our speculation must be most searching and original.

(3)

If we are to begin at the beginning, we must grasp that we are all part of the world we observe. We cannot divide the world into ourselves on one side of the screen as spectators, and everything else as a spectacle on the other side, which we remotely observe. This may seem merely a philosophical point. And it is of course possible to put together a good approximate working science on a false philosophy: to make steam engines and to fix the nitrogen in the air and solve several differential equations. But there comes a point of fineness when these rough and ready habits fail; and then it is not possible to get the right answers until we have the right notion of what it is that we are doing. At this point our philosophy must be right, if philosophy is the word for this critical attitude to our own habits of thought. We must look not at some abstract view of science, but at the actual processes which we carry through when we practice science.

I have already recalled the most remarkable practical example of this. Physicists since Newton have been describing the world as a network of events. But physics does not consist of events; it consists of observations, and between the event and us who observe it there must pass a signal-a ray of light perhaps, a wave or an impulse-which simply cannot be taken out of the observation. This is the insight which Einstein showed in 1905. It came to him when, looking at the discrepancies within physics then, he asked himself how in fact one would set about doing what Newton took for granted, namely comparing the time in two places far apart. Once the question is put, everyone can answer it: you cannot make any comparison at two different places without sending a signal and observing its arrival. The insight is not in answering the question: it was in asking it. Event, signal and observer: that is the relationship which Einstein saw as the fundamental unit in physics. Relativity is the understanding of the world not as events but as relations.

Something like this had been said by philosophers for some time: that science must get rid of abstractions, and make its system only out of what is in fact observed. But Einstein was the first to take the philosophy seriously. He put it into equations; and physicists were astonished to find that it explained the erratic behavior of Mercury, and predicted the bending of light near the sun.

I stress this example from large-scale physics for this reason. Examples are often quoted from quantum physics to show that the act of observing itself effects the particles we are looking at, much as a rabbit scurries away from our headlights at night. In the same way, it is hard in the social sciences to take a poll of opinion and frame the question so that it does not bias the replies. And in psychology the method of asking oneself questions has now been shown to be most fallible: you cannot watch your own mind and pretend to yourself that you are not looking. But none of these difficulties is as fundamental as that which Einstein revealed. In these examples, observation merely intrudes into the experiment. But Relativity went deeper and showed that the observations are the bricks and stuff of science.

By habit, this is not an easy point to grasp. We accept it during the experiment, and when that is over we slip back into making some model whose pieces are not observations but idealized things. Why not? we ask, It is only a model. And indeed, it will work well enough as an approximate model of large events, such as eclipses and hydro-electric dams and the action of penicillin in arresting the multiplication of bacteria. But when we come to finer effects we must be more modest and more realistic. For then we must use science as it is, and that is an assembly of observations so ordered that they tell us what we may expect to observe in the future.

(4)

In using the word observation, I am conscious still of having drawn too passive a picture of the process of science. We may be tempted still to think of the world as going its mighty way and merely impressing on the scientist in passing a glimpse from time to time of its imperturbable motion. This would be a grave misunderstanding. Indeed it would perpetuate the breach between the world and the experimenter which I have been trying to close. Science is not only rational; it is also empirical. Science is experiment, that is orderly and reasoned activity. The essence of experiment and of all science is, that it is active. It does not watch the world, it tackles it.

This of course is not peculiar to science. All living is action, and human living is thoughtful action. If this is plain enough as a statement about living, it still needs to be underlined about science: that science is a characteristic activity of human life. The characteristic of human action is that it is a choice at each step between what are conceived to be several alternative courses open to us. Men can visualize these alternatives and animals probably cannot; but in both, action means choice-and this, whether we suppose the choice to be free or circumscribed. In both, action is directed towards the future. Men are conscious of this direction, and choose one action rather than another in the conscious hope that it will lead to one rather than another kind of future. I add that this statement describes what they do correctly, whether we think that their choice is free or determined.

This seems to me, the most important point which I can make; and oddly enough, it has had least attention in the past. The characteristic of living things is that their actions are directed towards the future. We could put this more bluntly, and say that it is simply the characteristic of action; but this seems to me a needless abstraction, since action and living are in effect interchangeable notions. Living things change; they are different tomorrow from what they were today; and their actions today are directed towards tomorrow. The enzymes in the cell are unaware that what they do will make the cell divide in twenty minutes from now; but if they fail to do it, neither they nor the cell has a future; both die. We do not know what sets in motion the life cycle of the thread worm or the liver-fluke or the oak; but we know that each stage of that cycle is a getting ready for the next; and if the organism misses one cue, it dies. The mechanism of getting ready is odd and elaborate: we see the shadow and close our eyes, we hear a noise and our glands squirt adrenalin into our blood, so that the pulses quicken, the muscles tense and the nerves are alert. But in every case our actions are directed towards some obscurely foreseen future. And this is true of the most primitive cell, and of Gibbon mining mountains of scholarship for the pleasure at last of minting one ringing footnote.

All this is hidden in the process of life; but it becomes plain and explicit when we look for scientific laws. For of course a scientific law is a rule by which we guide our conduct and try to ensure that it shall lead to a known

future. The law formulates our anticipation of the future in a systematic way, as a kind of shorthand. And the wider the conditions in which the law applies, and the more compact as it were its shorthand, the more powerful and remarkable we think the law. But a scientific law differs from our own habitual way of pointing our actions towards the future only in being more systematic and explicit. We are all forward-looking creatures. Life is a process of looking forward. It turns towards the future as mites shift towards the light. Indeed, living things alone go through processes, such as age and decay, by which the future can plainly be told from the past. They make time visible; whereas there is nothing in the dead world at all easily accessible to tell the past from the future. In classical mechanics time has no recognizable direction, and the universe would work just as well if every atom in it were rolling backwards.

(5)

The key to the action of living things then is this, that it is directed towards the future. They have a way of knowing what is going to happen next, or more precisely, how to act in anticipation of what is going to happen next. Most of this knowledge is unconscious. We need not be astonished about this foresight, or at any rate we need not find it more astonishing than we find the rest of the world. For plainly it has always b fen the condition for the survival of living things, individually and in species. Unless they could adapt themselves to the future, and interpret its signals in advance, they were bound to perish. Whatever the rhythm and the uniformities of nature, what has survived of life has necessarily had to be in tune throughout; this chiming has been the condition for survival. Galileo is said to have discovered that a pendulum keeps nearly steady time by timing a swinging lamp by his pulse. The story makes my point in a neat symbolic manner; for of course all that Galileo or my doctor or anyone else has discovered is, not that either the pendulum or the pulse keeps steady time, but that they both keep the same time. Whatever their rhythm, they keep the same rhythm. We find the world regular as we find it beautiful, because we are in step with it.

I have said that in using foresight, whether unconsciously in instinct and habit or consciously in inference, living things have had to adapt themselves or die out. And we could put this more trenchantly: the act of foresight is itself the adaptation to the future. By this act individuals adapt themselves, and so do societies and all living assemblies. So the adaptation- of a species is a slow action directed towards the future, in which the whole society interprets the signals, whether of the coming ice-age or of the erosion of a continent, and unconsciously changes its structure to meet the change.

I repeat the word "unconsciously" because of course there is nothing in this which need be rationally understood or consciously willed. For the species as a whole, the mechanism of adaptation may be quite impersonal, and may even be at odds with the survival of the individual, as is the bee's dying sting. Selection inevitably acts now, and yet the species inevitably is adapting itself to the future: the generations as it were prepare one for the next. We need see no master mind behind this and no driving purpose. It is I repeat the condition of life for the individual and the species. The present is not like the future, but it is not unlike it either; it is a signal of the future; and living things, singly or in species, are predictors which interpret the signal so that they make ready for the future.

The idea of a machine which is a predictor is altogether new. But it is of outstanding importance, and we must become used to it. It covers all the basic actions of living things, from the search for food in the lowest cell to the boldest creations of the human imagination. To my mind, it gives us an insight into the function and processes of the human mind which has been missed by older philosophies. And this is not to be wondered at, because it is hard to grasp the full scope of a predictor until you have tried to build one.

A predictor is a machine which uses information about the past and the present in order to make ready for the future. In the nature of things, neither its information nor its forecasts can be complete. But they do not seek to be: the predictor is not trying to be a pocket version of the hypothetical angel of Laplace, a sort of scientific Tiresias who knows everything and has foreseen everything. A predictor takes its information in the form of signals, and its mechanism interprets these signals so that it acts in anticipation of the future. This action is a continuous process. The predictor goes on accepting signals even while it is adjusting itself towards the future, and it feeds these back into the mechanism so that, as it were, it keeps on tracking the future from moment to moment. This picture will do equally well for a predictor which is in fact tracking an aircraft so that the guns may fire on it at the right instant, or for the bat sending out its shortwave note to detect obstacles, or for the mechanisms which keep our body temperature constant or send blood to the brain when we are thinking. What I have called the interpretation of these signals is itself a fascinating business, because in every mechanical system, living or constructed, it implies a sorting out of the meaning of the message from the meaningless oscillations which are carried with it. But I want rather to single out the essential relation: that the present provides a set of signals, which are made continually to yield a meaning from which the future is anticipated. At each moment the machine must integrate its signals as a whole; the function of the process is a synthesis, not an analysis.

(6)

We are interested in science, where the process of prediction is conscious and rational. Even in human beings this is not the only kind of prediction. Men have sound intuitions which have certainly not been analyzed into rational steps, and some of which may never be. It may be for example, as is sometimes claimed, that moss people are a little better at guessing an unseen card, and some people much better, than would be a machine which merely picks its answers by chance. This would not be altogether surprising, for whatever the human mind is, it is certainly not a machine making only chance guesses, like a table of random numbers. Certainly evolution has selected us rapidly because we do possess gifts of foresight much above those of other animals. The rational intelligence is one such gift, and is at bottom as remarkable and as unexplained. And where the rational intelligence turns to the future, and makes inferences from past experiences to an unknown tomorrow, its success is quite as great a mystery as the very modest successes of even the most gifted guessers-off the music hall stage-who have yet been found.

There are two points here which we need to see very clearly. The first is an old confusion. For two hundred years now, philosophers have distinguished between reasoning by pure deductive processes, such as is found in Euclid, and inductive reasoning which extends the experience of the past into the future. But this distinction is much overrated. All that can be said about deduction is that we can state its processes, and give rules for deciding what is acceptable, in a precise form. But the sanction for believing that its conclusions will be true tomorrow because they were true yesterday are no different from those which apply to any other theory which claims to reach into the future. If a triangle has three equal sides, then its three angles will be equal, we say. But what we mean is still that the three angles are equal;

we have deduced that they are so by steps of logic which have always yielded sound results. If we say that the three angles will be equal, then we claim that these steps will continue to be allowable and will yield true results in the future. And this claim is typically an induction from the past to the future.

The second point goes even deeper. There is an unspoken assumption in all our speculations, that the ideal of science is to make predictions which shall always be fulfilled. We are hankering after the Laplace predictor, which shall be a perfect machine to get all the answers right. This is equivalent to saying that we want a model which shall be indistinguishable from the real world at every observation. This is not the aim of any predictor. Here bluntly we have the difference between the model and the predictor; and this is why I brought in the word "predictor": because it is not a machine which claims to act out the future in advance. It tries to forecast it, by its own process. And its forecasts` are not always right. It does not assume the future already to exist, to be dredged or conjured up in advance at our bidding. It makes no larger claim than that the future can in general be predicted, within defined limits of uncertainty. And since there are uncertainties, the predictor will sometimes be wrong.

We must face this fact, that in the nature of things predictions sometimes turn out wrong. Naturally, our aim is to have them right as often as possible and at least more often than not. But forecasts can be useful even if they are quite often wrong. We joke about the weather forecasts, but it was necessary to keep them secret during the war. And in major biological processes like evolution, the mistaken prediction has an important function. Genetic factors which remain in a species even though their effect is to make it less well fitted to its environment are a kind of mistake in prediction, and as it were a residual error. Yet without them, the species cannot adapt itself to new changes. Some heavily armored monsters have probably died out for lack of these means of future adaption, just as the pure strains of white mice would die outside the laboratory for which they have been bred to excessive perfection. Fitness for use must retain an element of unfitness and elasticity in order that it may also be fitness for change. When Bolingbroke and Paley argued that man is designed like a watch, which fits its use perfectly, they had no thought of further evolution. Characteristically, the eighteenth century was for them the peak and the point of rest of nature's history.

By contrast, we have learnt to see the world in motion and in change. We are clearer about our own shortcomings, but we have also learnt not to stop at them smugly. For what is true of the species when it faces the future is true of the individual. Both adapt themselves to, the future by continued correction, as a predictor does. The process is one of trial and error. This is the process we call learning, and the errors are as essentially part of it as the successes. If you put a mouse into a maze and it gets it right first time, it has not learnt to run the maze. It does not learn until it makes some mistakes and learns to avoid them. One mouse may learn from its mistakes quicker than another, but not even the ideal mouse of the psychology laboratory can learn otherwise than by making some mistakes.

The process of learning is essential to our lives. All higher animals seek it deliberately. They are inquisitive and they experiment. An experiment is a sort of harmless trial run of some action which we shall have to make in the real world; and this, whether it is made in the laboratory by scientists or by fox-cubs outside their earth. The scientist experiments and the cub plays; both are learning to correct their errors of judgment in a setting in which errors are not fatal. Perhaps this is what gives them both their air of happiness and freedom in these activities.

We must therefore understand that it is the nature of predictions to be sometimes mistaken. It is only so that we learn as individuals and as species. And science learns in the same way. This is precisely the step which Galileo and Francis Bacon took more than three hundred years ago, which was the beginning of our science. For until they set the Scientific Revolution going, men held to the medieval belief that the workings of nature could be understood by intellectual insight alone. Galileo and Bacon coupled with this appeal to reason a new appeal to fact. Since then, the test of scientific explanation has in the last place always been empirical: does it match the facts? Science itself has therefore been conceived, though unconsciously, as a process of learning; for to appeal to fact in speculation is to grant the possibility of error. Science itself is a predictor mechanism in process of continual self-correction. The steps from Ptolemy's astronomy to Newton's and thence to Relativity are precisely stages of learning; each step corrects the small but demonstrable error which has opened between prediction and fact. Let us not be contemptuous of mistakes; they are the fulcrum on which the process of life moves. At the very time that Paley was tracing God's design in the watch-like perfection of man, William Blake said more modestly, but with deeper insight, "To be an error and to be cast out is a part of God's design."

(7)

The fundamental ideas which I have been putting forward are these. Every living action is an act of choice. It is directed towards the future. The machine which we conceive within it is a predictor, which interprets past and present information as signals to accommodate itself to an expected future. And interpretation and accommodation cannot be made altogether free from error, for error is essential to the process of learning which directs them.

There is in all this a bold analogy between the way in which individuals learn, the way in which species adapt themselves, and the way in which science works. But, of course, it is my point that this is not merely an analogy: it is a true and close relation. For science is not a special activity. It is a type of all human activity. An Italian who goes to New York soon learns to adapt his habits to eating a factory-made cereal for breakfast. There is some evidence that the cereal eaters, as a species, are adapting their jaws to their diet by the slow workings of natural selection. But between these extremes there lies the equally human activity of scientific development. The invention and popularization of the breakfast cereal is itself a scientific solution to a complex of problems, which range all the way from cutting down the time between getting out of bed and catching the train, to the full use of the most readily won foods of North America.

What marks out science as a system of prediction and adaptation from those of the individual and of the species is at bottom this, that it is a method which is shared by the whole society consciously and at one time. This at once implies that science must be communicable and systematic. Both the signals and the predictions must be of a kind which everyone can have in common. To my mind, philosophers put the cart before the horse when they say that science constructs a world by sorting out what the experiences of different people have in common. On the contrary, the practice of science supposes the existence of a real and a common world, and assumes that its impact on each individual who is part of it is modified by him in a way which constitutes his personal experience. We do not construct the world from our experiences; we are aware of the world in our experiences. Science is a language for talking not about experience but about the world.

But what is most striking about the predictions of science is that they are not an assembly of piecemeal guesses. Science is a way of ordering events: its search is for laws on which to base the single predictions. This is the stroke which rounds our picture: that science is systematic in method because it seeks a system of prediction. The aim of science is to order the particular example by articulating it on a skeleton of general law.

Once again, what I have said about science is not peculiar to it. All human conduct is shaped by what the individuals believe to be general laws. The human predictor interprets the signal by an act of recognition which puts it into some general category. We then assume that the future will have some general likeness with futures we have met before which followed this kind of signal, and this is the kind of future we prepare for. We recognize a pair of dumb-bells and brace ourselves to lift them; when they turn out to be made of cardboard, the shock is unpleasant because unexpected. What is odd about the generalizations of science is not even that they are far wider, and cover a range of facts beyond the habits of any one individual. This is a real difference, but it is not the essential difference. The essential difference is that the generalizations of science are explicit. And this derives at once from the fact that science is communicated. The individual need never make a list of his habits, that is his generalizations, because he does not need to pass them on to anyone else. He will form habits of anticipating the future from present signals even if he never expects to meet another person. Robinson Crusoe did so; and Defoe shows striking psychological insight when he describes the disorder into which Crusoe was thrown when he saw the footprint, not because Crusoe feared the presence of other people, but because their presence had ceased to be part of his conceptual world. Although we cannot be sure, it is likely that some animals lack any form of communication; yet it is certain that they still form habits.

It is the explicit character of its laws which makes science a different activity; and this character derives from communication. Science is the activity of learning by a whole society, even though that society may so divide its labor that it passes the responsibility for this activity to a few men. And the laws of science are those principles of prediction and adaptation to the future which apply to the whole society, and can be learnt by all its members in explicit form. This need to meet two requirements at once, universal usefulness and explicit statement, is precisely what makes a world pictured by science seem strange to our personal experience. As persons we will not find ourselves analyzing the world into cells and co-enzymes and mesons and genes and curved space, because that is not an individual's analysis of his own experience. And precisely the individual's analysis of his own experience is the subject of discourse in Berkeley and Hume and McTaggart and Moore, whose philosophies all start from a point inside the head of one person. It is not to be wondered at that science and philosophy have more and more been losing touch when they have been talking about such different things. The nucleus and energy and the central nervous system are entities which are reached as we seek for the common world under the random fluctuations of individual experience. And the odd properties which they have are part of the price we pay for making them explicit. The world impinges on our experience in ways which we can recognize implicitly for ourselves as design, meaning, and cause and effect, and these will do admirably as approximations to the experience of all of us. But when we try to refine this language to describe in detail the real world which underlies our experiences, we meet the difficulties of all language. No explicit statement, no communicable language can formulate generalizations which are more precise than the common agreements between those who use them. So we cannot make scientific laws which have a greater finality than the measurements and rules which we can share. Our laws of prediction are limited by our human and necessary errors. There is nothing pathetic about this; it is no more tragic a shortcoming than others which make us men and not something else, than hunger or ambition. These are the nature and the driving forces of human societies; and I have shown that error in the laws of science also partakes of both.

(8)

The basis of the map which I have drawn in this chapter has been the relation between present and future. It is as if the future were our North and Pole-star: this gives their direction and their structure to the act and the thought, in life and in science. I am therefore not much troubled by the difficulties which philosophers have in trying to rationalize the process of intellectual inference or induction. Philosophers have wanted to give to induction about the future the same status that deduction holds in a timeless science like geometry. And I have already remarked that as soon as deduction is used in a science which takes account of the passing of time, it has no higher status than has induction.

But more generally the philosopher and the man in the street begin their speculations by thinking about the past and the present, as a solid basis of knowledge. For two reasons, this is not useful. First, it is only the past and the present of our own experience which are known. The real world which we share with others is just as mysterious in the past and the present as in the future. And second, it is a mistake to suppose that the basic process in thought is looking back at what is known; and that looking forward to the future is to be justified from this. This is a reversal of the process of life. Anticipating the future is the fundamental activity; babies do it before they are born. Analyzing the past and the present is a subsidiary process, whose purpose is still that we shall learn to recognize and interpret signals for the future. It is absurd to ask why the future should turn out to chime with our knowledge of the past. This puts the question upside down, and makes nonsense of it. What we have learnt from the past is knowledge only because the future proves it to be true.

The only question which can sensibly be asked about the method of induction into the future is this. What are our grounds for preferring one prediction to another? Why do we choose this rather than that course of action, in circumstances where the future which we foresee remains uncertain whichever we follow? It is not enough to answer that one prediction has a smaller calculated area of uncertainty than the other; for like every scientific law, this calculation already assumes a preference, if not between these forecasts then between more fundamental ones. And it certainly will not do to say that one prediction `has proved to be right more often than the other; because the next event is not the same as the last, and there is in fact no way of comparing events as such. No, our preference is not between forecasts but between ways of forecasting. We are not preferring one prediction, but one scientific law to another. And laws of course, unlike events, can be weighed by past evidence-although we ought to beware of the word "past": what we really mean is by other occasions on which we predicted the future on the basis of these laws.

One of the difficulties which have troubled philosophers and men in the street on all such issues, is that they have had so static a picture of the future. They have thought of the future as like the past or present, simply a moment along an endless red carpet of time, unrolling before us and rolling up behind. The future is just like the present, they have said; it just happens to beat another time. This error derives from Newton's picture of time, which had no direction at all, and might as well have run backwards. But since the middle of the last century, there has been one physical property which has given direction to time. It is this. If you look at a stream of gas which has come out of a nozzle, you can tell which part of the gas is further from the nozzle, that is which has come out earlier, without seeing the nozzle. The part which flowed out earlier is by now more disorderly, and its molecules are drifting about more at random. They have lost the direction imposed by the flow through the nozzle. So the passage of time in the universe at large is marked by an increasing state of physical disorder or randomness. It is remarkable that this is itself a chance effect, yet only this gives time (and with it cause and effect) its direction.

But the essential point is that this distinguishes the future from the past: it is the one general law about the future to which we are all attuned. We do not know how we sense this; but we certainly do. Indeed, the essential property of life is that it is opposed to this current: life imposes greater order from moment to moment, while the physical universe is drifting into greater disorder. Even the guessing of an unseen card is not beyond the boundaries of the intelligible, once we understand that the future has distinctive properties which make it recognizably different from the present. Recognizably different, in the statistical sense: for the future differs from the present by being statistically more random. The guess would be inexplicable only if it were always right.

(From J. Bronowski, The Common Sense of Science, Harvard University Press, Cambridge, 1978. ISBN 0-674-14651-4. Chapter VII).