The Nature of Science

There is a popular account of the scientific enterprise which presents its method as surefire and its achievement as the inexorable establishment of certain truth. Experimental testing verifies or falsifies the proposals offered by theory. Matters are thus settled to lasting satisfaction; laws which never shall be broken are displayed for all to see. In actual fact, as we shall find out, the matter is a good deal subtler than that. Nevertheless, the great enhancement that the twentieth century has seen in our understanding of the world in which we live, even encompassing an account of its earliest moments fourteen thousand million years ago and including the beginnings of a comprehension of how life could have evolved from inanimate matter, together with the remarkable technological developments stemming from scientific advance, lends a certain credibility to this triumphalist point of view. Such splendid successes suggest that here is the key to real knowledge. In the bright light of science's achievements, other forms of discourse are in danger of appearing mere expressions of opinion. The widespread thought that science has somehow "disproved religion" is based on psychological effect rather than logical analysis. It is a continuation of the Enlightenment distrust of all knowledge which is not patterned according to the paradigm of scientific method.

It is ironic that at the same time that there is this widespread popular attitude there is also, in circles more austerely intellectual, a critical review of the nature of the scientific method and of its actual achievement. The practices of science have been reassessed and its procedures found to be more complex and questionable than the simple popular account acknowledges. The picture of the professor in his laboratory watching the pointer move across the scale to the expected reading, and thereby establishing his theory beyond the possibility of doubt, bears about as much relation to reality as does the simplicity of the comic-strip detective to the complexities of actual police investigation. If the method of science is open to revaluation, so, of course, will be the nature of the conclusions resulting from it. It is to these matters that we must now turn.

Certainly science seems to be successful in settling issues to the satisfaction of those concerned. At the beginning of this century, there were still a few physicists like Ernst Mach who did not believe in atoms. They thought the idea just a useful figment of the chemists' imaginations, but they did not accept the existence of a real granularity in nature. Nowadays, you would not find a scientist who would espouse such an anti-atomic opinion. We all believe in atoms, even if the elementary particle physicists think of them as "large" composite systems, and it is to the quarks and gluons that we now look for the basic constituents of matter. Such achieved agreement is impressive. It contrasts with many other forms of knowledge where debates continue without prospect of universal settlement. Karl Popper said, "But science is one of the very few human activities—perhaps the only one—in which errors are systematically criticized and fairly often, in time, corrected... in other fields there is change but rarely progress."1 This apparent scientific progress is pretty clearly connected with the exploitation of the experimental method. Is it not science's power to manipulate and interrogate the material at its disposal which enables it to provide agreed answers to the questions raised? I write as a theoretical elementary particle physicist who for more than twenty-five years worked in that discipline. For the greater part of that period, the subject was experimentally led. It was the discoveries of our experimental colleagues which largely set the theoretical agenda.2 Latterly theory has regained the initiative, with its development of gauge theories of fundamental interactions, but even so it is still to experiment that it must look for the confirmation of its ideas. In 1967 Abdus Salam and Steven Weinberg independently proposed an attractive theory which combined electromagnetism and the weak nuclear force (responsible for such effects as the beta decay of nuclei) into a single unified account. It was an idea comparable with the unification of the apparently dissimilar forces of electricity and magnetism which Clerk Maxwell had achieved in the nineteenth century. Nevertheless, Salam and Weinberg's work gained comparatively little attention until the confirmation in the second half of the 1970s of the existence of an effect that their theory had postulated, the so-called neutral current. Now the theory seems to us to be completely established because of the discovery in 1983 by physicists at CERN of the very heavy W and Z particles which are the cornerstone of its construction.

Told like that, it all sounds like a textbook example of the simple view that theory plus experimental verification equals established truth. Yet the story is a little bit more complicated in its details. The experimentalists could have discovered the neutral current in the 1960s, before Salam and Weinberg had formulated their ideas. They actually saw events which we would now understand as due to its effects. However, it is difficult to sift out this elusive phenomenon from the experiments, because many other things are also going on. Amongst this background (as the physicists call it) were events induced by neutrons which could look very much like those caused by a neutral current. To interpret the results, therefore, it was necessary to estimate how great these neutron effects would be. In the 1960s it was believed that these spurious background events would be sufficiently numerous to explain away altogether the apparent neutral current interactions. In the 1970s improved calculations showed that this was not the case. I do not doubt that the new calculations really were better than the old ones, but we have to recognize that people were motivated to do them partly by the theoretical expectation, by then existing, that there might well be an actual neutral current to be observed. This is just one of many possible examples of how difficult it is for an experimentalist to see what he is not looking for.

Furthermore, I have said that in 1983 physicists at CERN discovered the Ws and the Zs. Indeed, Rubbia and van der Meer were given a well-justified Nobel Prize in 1984 for doing so. What they and their colleagues actually saw, however, was a complicated pattern of readings in a very large and expensive array of electronic counters. An extensive chain of interpretation is necessary to translate those patterns into "Here we have a W," or "There is a Z."

The trouble with the simple view of scientific method is that it does not take into account the sophisticated web of interpretation and judgment involved in any experimental result of interest. To be told that the needle of the galvanometer moved to 7.6 on the scale, or that there are certain marks on a photographic plate, is not in itself a source of insight or excitement. It can only become so if, through careful assessment of possible competing and obscuring effects (the background calculations of which we have spoken) and from an interpretative point of view (which sets the result in a matrix of theoretical understanding and expectation), we are led to the conclusion that something of significance has actually happened. Experiments are always theory-laden. The dialogue between observation and comprehension is more subtle and mutually interactive than is represented by the simple confrontation of prediction and result.

In order scientifically to interrogate the world, we have to do so from a point of view. It is precisely this need for an (admittedly corrigible) theoretical expectation which distinguishes science from its precursor, natural history, which is simply content to take in the flux of apparent experience as it happens. In a famous phrase, Russell Hanson referred3 to this theory-laden character of our observation as "the spectacles behind the eyes." Our scientific seeing is always "seeing as."

To recognize this is to raise the question of the character of our experimental knowledge. The role of observation as the stern and impartial arbiter of scientific theory is somewhat compromised if in fact the image of nature we receive is always refracted by those spectacles behind the eyes. Might there not be a variety of possible perspectives on the world of which the received scientific view at any time is just one option?

In books on the philosophy of science, this possible dilemma is often illustrated by the notorious duck/rabbit, a sketch which, looked at one way, can be seen as a duck and which looked at another way, can be seen as a rabbit, the open bill of one becoming the ears of the other. Actually, this particular ambiguity is rather readily resolved by acknowledging that what is before us is a rather exiguous line drawing. Physics itself provides a much more striking example of such ambivalence.

The conventional view of quantum theory,4 accepted by the vast majority of physicists, states, for example, that there is no assignable cause for the decay of a radioactively unstable nucleus at any particular moment. All that can be asserted is that there is a calculable probability for such a decay taking place within a specified period of time. The quantum physicist is in the same practical position as the actuary of a large insurance company who is unable to say whether any particular client will die in the coming year, but who can be toler ably sure that a calculable number of clients in a particular age group will die within that period. However, there is an important difference between the physicist and the actuary, according to conventional quantum theory. There are causes why the actuary's clients die, even if they are not known to him. There are asserted to be no causes for individual events in the quantum world.

To this conventional quantum interpretation, there is an alternative point of view, first worked out successfully by David Bohm. It asserts that all events are causally determined, but some of these causes (called in the trade "hidden variables") are inaccessible to us. That is the reason, in Bohm's view, why our actual knowledge has to be statistical. It is a matter, not of principle, but of ignorance. This point of view is, of course, identical with that of the actuary, whose clients die of causes, to him unknown.

In the realm of non-relativistic quantum theory (that is, concerning the behavior of very small and slowly moving systems), the conventional theory and Bohm's theory give exactly the same experimental results. Yet the understandings they offer are radically different. Here is a duck/rabbit with a vengeance! Why then do the majority of physicists believe the one in preference to the other? It is clearly not a matter of observational decision.

I think there are two reasons for the majority preference for conventional quantum theory (which I share). The first is that Bohm's theory, though very clever and instructive, has a contrived air about it. It is significant that this is enough to put off most professionals despite the theory's "common sense" determinism, which might seem an overwhelmingly attractive feature to a layman. Matters of taste, judgments of elegance and economy, play an important part in the development of science. By these canons conventional quantum theory seems to most of us more elegant, and so more compelling, than Bohm's ingenious ideas. But why should the more elegant prove scientifically the more compelling, other things being experimentally equal? Here we see the coming into play of a factor, the search for simplicity, which goes beyond the impersonality of the popular account of the scientific enterprise. After all, is not one man's simplicity another man's complication? Does it not all depend on those spectacles behind the eyes? To Copernicus as much as to Ptolemy, the circle was the perfection of simplicity. It was only natural, in their view, that heavenly motion should be explained in circular terms. Kepler's introduction of ellipses must have seemed to many of his contemporaries a most ugly and unwelcome development. Simplicity only returned to celestial mechanics with the totally different beauty of the inverse square law inserted into Newtonian dynamics. Today we retain a belief in the elegance and economy of gravitational physics, though its current expression would be in terms of the geometrical curvature of space-time described by Einstein's general relativity (if one uses the language of classical physics) or in the gauge theory of massless gravitons (if one uses the language of quantum theory). Beauty is indeed in (or behind) the eye of the beholder. Its influence on scientific thought is undeniable, but that very statement raises the question of the true nature of that thought. A second reason for preferring conventional quantum theory to that of Bohm is that the former is much more readily combined with special relativity to give an account of small physical systems whose velocities approach that of light. Although many of its predictions are consistent with relativity, Bohm's theory requires a specific reference frame for its formulation. The contrasting success of conventional quantum theory in being readily open to this extension illustrates another, and reassuring, feature of "good" science—its fruitfulness, the way ideas can continue to be applied in circumstances going far beyond those for which they were originally invented. A curious twist to this part of the story is that no one has yet found a perfect reconciliation of conventional quantum theory with general relativity. In other words, despite great efforts, the quantum theory of gravity is not yet on a firm foundation, and the two modern points of view about gravitation sketched at the end of the precedi ng paragraph are not perfectly at one with each other. There is hope that the speculative theory of superstrings may provide the resolution of this dilemma. For the present it is interesting to note that physics can manage to survive with two of its fundamental theories, quantum theory and general relativity, imperfectly reconciled. Even pure theory is never exhaustively rational.

The simple account of science sees its activity as the operation of a methodological threshing machine in which the flail of experiment separates the grain of truth from the chaff of error. You turn the theoretic-experimental handle and out comes certain knowledge. The consideration of actual scientific practice reveals a more subtle activity in which the judgments of the participants are critically involved. If you wish to give an experimental physicist an uneasy moment, look him straight in the eyes and say, "Are you sure you have got the background right in your latest experiment?" (In other words, "Are you sure you have eliminated all possible sources of spurious effects and are actually measuring what you claim to measure?") If you wish to give a theoretical physicist an uneasy moment, look him straight in the eyes and say, "That latest theory of yours looks a little contrived to me." (In other words, "I do not see in it that look of elegant inevitability which time and again has proved the hallmark of true theoretical insight.") Their answers will not depend upon simple ineluctable prediction confronting indisputable fact. Rather, they will involve a reasoned discussion of how those concerned evaluate and interpret the situation.

This role of personal judgment in scientific work was emphasized by Michael Polanyi.5 He called it tacit skill. Acts of discrimination are called for in concocting a successful scientific theory which are no more exhaustively specifiable than are the skills of a wine-taster in blending a good sherry. But just as the sherry blender has to submit the result of his labors to the judgment of the discerning public, so the scientist has to persuade his colleagues of the soundness of his judgment. This necessity saves personal knowledge from degenerating into mere idiosyncrasy.

Once one has acknowledged the part that personal discrimination has to play in scientific endeavor, the whole enterprise may seem to have become dangerously creaky, its rationality diminished or even destroyed, by the importation of acts of individual judgment, even if they are claimed to be validated by the eventual assent of the scientific community. Has not the austere search for truth degenerated into the proclamation of an ideology, even if democratically endorsed by its adherents? There have certainly been philosophers of science who have taken such a view, and it is from them that the scientific method has received its most severe criticism.

Thomas Kuhn studied those rare moments in the history of science when a major change occurs in the scientific worldview. Most of the time, scientists are engaged in problem solving, applying an agreed overall understanding to the attempt to explain particular phenomena. Just occasionally, however, it is the overall understanding itself which is subject to radical revision. An example of such a paradigm shift, as Kuhn calls it, would be the transition from classical to relativistic dynamics. For Newton there is a universal uniformly flowing time; for Einstein each observer experiences his own time so that two observers in relative motion will not agree about which events are simultaneous with each other. For Newton a particle's mass is an unchanging quantity; for Einstein it varies with the motion of the particle. Clearly there is a striking difference between these two systems of mechanics. We can all agree on that. But Kuhn proclaims a divorce between the two so absolute that he can say that "In a sense that I am unable to explicate further, the proponents of two competing paradigms practice their trades in different worlds."6 This is his celebrated claim that two competing paradigms, such as Newtonian and Einsteinian mechanics, are incommensurable; that is, there is no point of contact and comparison between them. If this were really so, it would imply that there were also no rational grounds for preferring one to the other, since such grounds would depend on the possibility of making just such a critical comparison. Kuhn does not flinch from drawing that conclusion:

As in a political revolution, so in paradigm choice—there is no standard higher than the consent of the relevant community. To discover how scientific revolutions are effected, we shall therefore have to examine not only the impact of nature and of logic, but also the techniques of persuasive argumentation effective within the quite special groups that constitute the community of scientists.7

Thus Kuhn's study of scientific revolutions has led him to accentuate the role of the personal factor to the extraordinary extent of proclaiming the efficacy of scientific mob rule.

All this is really very curious and greatly overdone. Did special relativity really come to be adopted because Einstein had a propaganda machine superior to that of Lorentz? Experimental evidence (such as the eventual confirmation of the slowing of moving clocks via observation of the lifetimes of rapidly moving particles) presents perfectly adequate nonideological reasons for accepting the theory. While Newton's and Einstein's understandings of mass are very different, is there not sufficient residual common ground for us to be able to say that they are offering alternative, and so comparable, accounts of inertia? Kuhn dismisses as an irrelevancy the well-known fact that Newtonian mechanics is the slow-moving limit of Einstein's mechanics. Yet to physicists this relationship would seem to be important, for it explains why classical mechanics was so long an adequate theory and why it remains so for systems whose velocities are small compared with the velocity of light.

Of course, study of persuasive techniques can help us understand why a new scientific viewpoint gains quick or slow acceptance, but to suppose that this provides the major part of the story of how new ideas are embraced is surely preposterous. Indeed, in later writings

Kuhn himself seems to have withdrawn from so extreme a position.

Kuhn's revolutionary incommensurability, if true, would undermine the idea that science can claim our rational, as opposed to rhetorical, assent. An even stronger threat to that idea is posed by the writings of Paul Feyerabend. He is a philosophical enfant terrible who does not hesitate to proclaim that the scientific emperor has no methodological clothes. Our discussion of skill has made it clear that there is no totally specifiable set of rules for scientific theory choice. There is no algorithmic machine, the turning of whose handle is guaranteed to lead to the Nobel Prize. At best, there are only guiding principles, exercised with discrimination by experts whose conclusions are subject to the collective judgment of the scientific community. Feyerabend seizes on this tacit, unspecifiable element and blows it up into a dominating principle of scientific laissez-faire. He claims that in science "the only principle that does not inhibit progress is anything goes."8 He is a self-proclaimed scientific anarchist. What in Kuhn was simply preposterous becomes in Feyerabend the Theatre of the Absurd.

If science is an intellectual free-for-all, then there is no reason for preferring astronomy to astrology, the oxygen theory of combustion to the phlogiston theory. Feyerabend honestly recalls that "having listened to one of my anarchic sermons, Professor Wigner [a distinguished theoretical physicist] replied 'But surely you do not read all the manuscripts that people send you, but you throw most of them into the wastepaper basket.'" He acknowledges that he does so, but "partly because I can't be bothered to read what does not interest me... partly because I am convinced that Mankind, and even science, will profit from everyone doing his own thing."9 His impishness is irrepressible.

Yet another assault on the rationality of science is mounted by adherents of what is called the "strong program" in the history of science. They assign to social forces a prime causative role in scientific change. For example, Andrew Pickering wrote of the recent sequence of investigations in high energy physics which have led physicists to believe that matter is composed of quarks and gluons, "The world of HEP [High Energy Physics] was socially produced."10 The claim is that the largely unconscious adoption of certain conventions of experimental interpretation, together with a collective expectation framed in particular theoretical terms, has so molded the thought of the invisible college of high energy physicists that a quark model of matter was imposed on the supposedly plastic mass of available data. The assertion is there in the title of his book; it is "constructing" quarks, not "discovering" them. Weighty grounds would be required for so startling a conclusion. On investigation we find that all that is offered is an analysis of such incidents as the differing background calculations, which in the 1960s seemed to exclude neutral currents, but which in their revised form were the basis for confirming the by then theoretically acceptable neutral current in the 1970s (see p. 11). We can readily agree that this is an excellent example of how social forces can retard or accelerate the pace of scientific discovery, but there are no grounds at all for going on to assert that they actually control the nature of that discovery. After all, the dust does settle. No one would now claim that the neutral current is an artifact of background calculations. The cumulative weight of evidence for its existence, made clearer by increased understanding of how to perform the calculations of neutron-induced background, has simply settled that issue.

The wider question of the quark structure of matter is a more complex story, but it contains an incident which illustrates how it is that significant signals from nature provide the stimulus for the development of understanding. In the late 1960s it was discovered at Stanford that hard scattering took place from protons. That is to say, when projectiles such as electrons were hurled at protons, sometimes they bounced back. This was the direct analogue at the subnuclear level of Rutherford's famous work in 1911 when he found that alpha particles bounced back off atoms. He said later that the result was as surprising as if a 15-inch shell had recoiled on impact with a sheet of tissue paper. Rutherford interpreted his experiment as showing the existence of a pointlike concentration of positive electric charge within the atom. He had discovered the nucleus. In an exactly similar way, it was natural to interpret the Stanford hard scattering experiments as indicating the presence of pointlike constituents within the proton. From that time onward many of us felt sure that some form of quark model was an inevitable next development in particle physics. Of course, that conviction arose within an interpretative framework, as all scientific understanding has to do, but it arose in response to phenomena, not as an imposition upon them. If hard scattering had not taken place, no amount of social pressure would have succeeded in constructing a quark model.

The recognition of a role for judgment in the scientific enterprise, a tacit element not wholly reducible to the application of rules specifiable a priori, gives it a kindred character to aesthetic, ethical, and religious thinking. Many have asserted these latter modes of thought to be of a different and inferior kind, matters of mere opinion. We now see that what is involved in the comparison is a question of degree rather than an absolute distinction. To say so is not, as Kuhn and Fey-erabend and others have suggested, to open the door to irrationality but simply to recognize that reason has a broader base than corresponds to a totally specifiable method of verification. (The qualification total is vital here; we are not saying that anything goes.) The mind has its reasons that computers know not of. The justification of this rational claim depends, I believe, on an assessment of the actual nature of the scientific achievement. By their fruits ye shall know them. We must consider what those fruits actually are.

At first sight the prospect might seem discouraging. Paradoxically, the advancing success of science appears subversive of its attainment of truth. Do not all theories in the end prove inadequate and have to be replaced? We once thought that the basic constituents of matter were atoms; then nuclei; then protons and neutrons; then quarks and glu-ons; next—maybe strings? Bigger fleas have lesser fleas, and so ad infinitum. Isn't ultimate scientific truth a will-o'-the-wisp? Newton-Smith calls this the pessimistic induction "any theory will be discovered to be false within, say, two hundred years of being propounded."11 There is excellent evidence for adopting this maxim. However, to do so is only fatal if we thought that certain truth is our necessary goal. In fact we shall have to be content with the more modest aim of verisimilitude. Our understanding of the physical world will never be total, but it can become progressively more accurate. The analogy of a sequence of maps of increasingly larger scale may be helpful. None will ever tell us all there is to be told about that particular piece of terrain. Each is a kind of coarse-grained isomorphism, representing accurately features from a certain size upwards but ignoring or smoothing out those which are smaller. For different purposes different maps are adequate. As a motorist I do not need the detail I would demand as a hiker. In the same way established scientific theories do not disappear; they simply have their domain of applicability circumscribed. Newtonian mechanics is satisfactory for largish objects moving at ten miles an hour, unsatisfactory for the same objects moving at a hundred thousand miles a second. Scientific theories are corrigible, but the result is a tightening grasp of a never completely comprehended reality.

So I would wish to say. But there are many who would deny it. First, there is the problem of how we know, even within a prescribed domain, that we have arrived at an adequate map of its physical behavior. Newtonian mechanics has so far proved excellent for describing the collisions of billiard balls, but how can I be sure that the next time I approach the table they will not be found to be behaving differently? Anyone attempting to make a general statement faces the problem of induction, of how to produce universal laws from the study of specific instances. One cannot examine every electron in the universe before saying anything about electrons in general, nor can one survey every billiard ball collision that ever has been, or ever will be, before pronouncing on how such objects behave. That resolute skeptic, David Hume, was the first to emphasize the logical difficulty this presents. How then can science proceed?

One possible response is to moderate the claim. This is the attitude of Karl Popper. He exhibits a maximal distrust of induction. However many "for instances" there may be in favor of a theory, there is always an infinity of untried cases in which it might prove wrong. The odds are thus permanently stacked against its validity. In Popper's view, therefore, we abandon all hope of verification. The best that can be done is to settle for falsifiability. While any number of successes will never count in a theory's favor, one failure will prove fatal. "Only the falsity of the theory can be inferred from empirical evidence, and this inference is a purely deductive one."12 This chilling message is conveyed in the original with the emphasis of italics. Clearly if that is all that can be said, the nature of the scientific enterprise is precarious indeed. "The empirical basis of objective science has then nothing 'absolute' about it. Science does not rest upon rock-bottom. The bold structure of its theories rises, as it were, above a swamp."13 Popper is driven to this gloomy assessment because he exalts logic above intuition. (It is only the "deductive" which is safe.) Since the ratio of the number of successful answers to the number of potential questions is inevitably, for any theory, a finite number over infinity, those who rely on such wary calculation of odds will always condemn themselves to a state of intellectual pessimism. It is imposed by their timidity. Newton-Smith says "one cannot over-stress the counterintuitive character of [Popper's] position."14 Is the Newtonian mechanics of billiard balls really in a state of permanent jeopardy? I think not.

Those who balk at induction do so because there is no exhaustively specifiable set of rules which enable one to lay down a priori when its application is justifiable. Its employment involves an act of judgment, even though in the case of a theory well tried in a defi nite domain, such as Newtonian mechanics, one cannot feel that great powers of discrimination are required for its successful exercise. We have already recognized that such acts of judgment enter other aspects of the scientific enterprise. That being so, an answer to the Humean criticism which is preferable to the partial surrender of Popper is simply to assert that we shall rely upon inductive method exercised with an appropriate degree of skill. Undoubtedly that attitude corresponds to the actual practice of science, and it seems to have stood the subject in good stead.

Science certainly appears successful. It has the air of progress about it. But what exactly is the nature of its achievement? Here we come to the second set of objections to any claim that science results in a tightening grasp of a never completely comprehended reality. It is asserted that the use of that last word is a naive misapprehension of what science is actually about. We have reached the parting of the ways between the positivists, the idealists, and the realists.

Those of a positivist persuasion lay stress on perceptions which can be intersubjectively agreed; the scientific task is the harmonization of such experience. Entities not directly accessible to experience, such as electromagnetic fields and quarks, which form the staple of the discourse of fundamental physics, are said to be just manners of speaking which are useful simply as means to that reconciling end. They do not represent actually existing realities. The scientific world is populated by pointers moving across scales and marks on photographic plates, rather than potentials or electrons; theories are just convenient summaries of data. There are, in fact, very considerable difficulties in drawing that clear distinction between the facts of data and the devices of theory which my simple summary of positivism has assumed.15 The objections to positivism, however, go beyond that. Its arid account seems totally inadequate to explain the actual practice of science. After all, if all that happens is the reconciliation of various bits of experience, much of it recondite, why is it worth all the painful labor involved? Bernard d'Espagnat, speaking of the activity of elementary particle physicists, wrote:

Whereas the activity appears essential as long as we believe in the independent existence of fundamental laws which we can still hope to know better, it loses practically its whole motivation as soon as we believe that the sole object of the scientists is to make their impressions mutually consistent. These impressions are not of a kind that occur in our daily life. They are extremely special, are produced at great cost, and it is doubtful that the mere pleasure their harmony gives to a selected happy few is worth such large public expenditure.16

Or, I would add, the dedication and toil of those involved. The philosophical problems of positivism, together with its impoverished account of scientific motivation, mean that it has few adherents in its pure form. However, there are accounts of the nature of scientific achievement, less inadequate than positivism but substantially influenced by it, which make claims that fall short of the realism I am wishing to defend.

Science is concerned with the power to predict or the power to manipulate phenomena, we are told. These two abilities are closely connected, for to foresee is to be forewarned and so at least to some extent to be in a position to take action to obtain a desired outcome. Science does indeed manifest such instrumental capacity, but should we be content with that and not go on to claim that its final goal is understanding? An instrumentalist would maintain that the only question to ask about a theory is, "Does it work?" If it does, then we are not to bother whether it is true or not. The suggestion urged on Galileo by Cardinal Bellarmine, that the Copernican system was just a means of "saving the appearances" (of getting the answers right) but did not describe how things actually were, would be endorsed enthusiastically by someone of this persuasion. However, it will not do.

Suppose that the Meteorological Office was given a sealed machine which had the property that if you fed in details of today's weather, then the machine would correctly predict the weather for any day ahead in the following year. The predictive role of the Met Office in weather forecasting would be perfectly fulfilled. Would that mean that all its meteorologists would simply pack up and go home? Not at all! They are also interested in understanding the way in which the earth's atmosphere and the sea and the landmasses interact as a giant heat engine to produce our climate. Before long some of them would be tampering with the seals on the machine in the hope of finding out how it worked, expecting that that would lead to an improved comprehension of the weather system that it modeled so accurately. No account of science is adequate which does not take seriously this search for understanding, together with the experience of discovery which vividly conveys to the participants the impression that understanding is what they are actually attaining. I have never known anyone working in fundamental physics who was not motivated by the desire to comprehend better the way the world is. It is because they yield understanding, though often having low or zero predictive power, that theories of origins, such as cosmology or evolution, are rightly classed as parts of science.

To claim that understanding is the true goal of science and the nature of its actual achievement is not of itself to have reached the realist position I wish to defend. We have to ask the further question of where this understanding comes from. Is it imposed by us, or is it dictated by the nature of the world with which we interact?

The former account would be given by those who take an idealist position. The modern grandfather of this point of view is Immanuel Kant, who believed that space and time are necessary mental categories which we impose on the flux of experience in order to be able to cope with it at all. This kind of view has not been without its supporters in the scientific community. Sir Arthur Eddington, in a famous parable, compared physicists to fishermen using nets with a certain width of mesh, who concluded that there were no fish in the sea smaller than that particular size. In other words, the apparent ordered reality that we think we perceive is alleged to be the product of our observational procedures. The American physicist Henry Margenau was bold enough to admit the consequences of such ideas and said, "I am perfectly willing to admit that reality does change as discovery proceeds."17 In his view the neutron did not exist prior to its "discovery" in 1932. One's feeling that such a statement is, to say the least, highly unsatisfactory is reinforced by a consideration of the track record of idealist claims. Kant believed that he had demonstrated that space had to be three-dimensional Euclidean in structure. With our knowledge of non-Euclidean curved spaces, actually realized in general relativity, we can see that all that he succeeded in doing was to produce a specious rationalization of what at the time was thought to be the only physical possibility. Eddington spent the last years of his life developing the tortuous ideas published posthumously in his book Fundamental Theory. 18Its supposedly rationally established conclusions have signally failed to correspond to the structure of the physical world revealed to subsequent investigation. If fruitfulness for the future is a good test of scientific creativity, idealist notions have proved a dismal failure.

We need not be surprised. The world, though ordered, is strange and subtle. Our powers of rational prevision are pretty myopic and limited by the contingency of the way things are, existing independently of how we think they ought to be. The natural convincing explanation of the success of science is that it is gaining a tightening grasp of an actual reality. The true goal of scientific endeavor is understanding of the structure of the physical world, an understanding which is never complete but ever capable of further improvement. The terms of that understanding are dictated by the way things are.

That is the realist position. It certainly corresponds to the way scientists themselves see their activity and are encouraged to persevere with it. Of course, most of them are philosophically unreflective people, and it might be that this is just a shared naive misapprehension.

Yet the way devotees of a subject view their practice must surely count for something in its evaluation. Many philosophers of science have been unwilling to give this due recognition, feeling that they knew best, without paying sufficient attention to what the honest toilers had to say. The realist view, it seems to me, is the only one adequate to scientific experience, carefully considered.

If realism is to prove defensible it has to be a critical, rather than a naive, realism. First, it has to recognize that at any particular moment verisimilitude is all that can be claimed as science's achievement—an adequate account of a circumscribed physical regime, a map good enough for some, but not for all, purposes. Once one moves outside regimes already explored, to hitherto unattained high energies for example, then there is every prospect that modification of our theories will be required to take account of unforeseeable phenomena. These modifications may, at times, be drastic (as when Einstein takes over from Newton), but there is sufficient residual continuity to discount the Kuhnian claim that we have lurched from one world to another, disjoint from it.

Second, our everyday notions of objectivity may prove insufficient as we move into regimes ever more remote from familiar experience. Quantum theory presents us with exactly this happening. According to Heisenberg's uncertainty principle, for entities like electrons we cannot know both where they are and what they are doing. This abolishes picturability in the quantum world. I shall discuss later (p. 53) in what sense we can still assert that an electron is "real," but it is clearly not that of naive objectivity. Realism is not tied to such simple notions derived from everyday experience alone.

Third, a critical realism is not blind to the role of judgment in the pursuit of science. It acknowledges that the simple picture of definite theoretical prediction confronting unquestionable experimental fact and leading to certain truth is too unsubtle an account of what science is about. As Newton-Smith says, "The story of SM [Scientific Method] will not produce a methodologist's stone capable of turning the dross of the laboratory into the gold of theoretical truth."19 There are always unspecifiable discretionary elements involved.

We cannot take off our spectacles behind the eyes, but if experiments are theory-laden, it is also true, as Carnes points out, that theories are fact-laden.20 They are responses to what is perceived to be there and in need of explanation. Perhaps the most troublesome question for the critical realist arises from the fact that for any finite set of data, there will always be a variety of possible theories which could fit it. (One could call this the duck/rabbit problem.) A rational criterion of choice is provided by demanding that an acceptable theory should prove its fruitfulness. It can do so in two ways: by a capacity to continue to cope with data as their range and accuracy expands, and by the theory being shown to have correct conclusions unforeseen at the time of its devising.

As an example of the former, consider the Newtonian account of the solar system. For about two centuries every new result coming from increased observational accuracy could be explained by a natural refinement of calculational technique. These theoretical responses represented fine-tuning in accuracy (for example, by taking into account hitherto neglected interplanetary effects) which was wholly in accord with the spirit of the theory and in no sense imposed upon it. (In contrast, a stubborn adherent of the Ptolemaic theory would have had to introduce ad hoc a new set of epicycles every time better observations were available.) The most striking illustration of such natural development of Newtonian ideas was provided by the work of Adams and Leverrier. They explained perturbations in the orbit of Uranus by supposing them to be due to a further, and till then unknown, planet. Their suggestions were triumphantly confirmed by the discovery of Neptune. The power of a theory to respond to progressive experimental probing without arbitrary manipulation is strong evidence of its verisimilitude. We cannot go on to say its truth, because its fruitfulness is not unlimited. An unresolved small discrepancy in the advance of the perihelion of Mercury eventually showed that even the Newtonian theory of gravitation had its limited domain of applicability. The explanation of this phenomenon required Einstein's general theory of relativity.

As an example of the second type of fruitfulness, we can consider Dirac's theory of the electron. In 1928 he devised an equation which successfully combined quantum mechanics with special relativity. Such a nontrivial synthesis was necessary to describe particles which are small and fast moving. It was an unexpected bonus when it was found that the same equation also explained the fact, till then mysterious, that the electron's magnetic properties were twice as strong as one would naively have expected. When this sort of thing happens, it is very convincing evidence for the verisimilitude of the theory. Again, it was no more than that, for it was eventually found that there are small corrections to the electron's magnetic behavior which require for their explanation the much more elaborate theory called quantum electrodynamics.

I believe that, after a certain time of testing, theories which gain wide acceptance in the scientific community have exhibited their reasonableness by demonstrating just such fruitfulness. Such rational staying power conveys an impression of naturalness and lack of contrivance which is convincing. Thus the underdetermination of theory by data does not pose a fatal difficulty for realism, since the theories which survive have been selected by the rational criterion of sustained success. Nor do I think that the lack of effective competing theories is to be attributed to a slothful acquiescence in a socially induced consensus. Scientists are active in a continual attempt to devise alternatives to received opinion, impelled not only by the search for truth but also by the desire to establish personal reputation.

I have attempted to defend a view of science which asserts its achievement to be a tightening grasp of an actual reality. In the course of the discussion, we have acknowledged the role that personal judgment, presented for the approval of the community and pursued along lines which are rational but not wholly specifiable, has to play in the enterprise. In my view this means that science is not different in kind from other kinds of human understanding involving evaluation by the knower, but only different in degree. It is clear that the personal element is less significant in science than in, say, judging the beauty of a painting, but it is not absent. We are to take what science tells us with great seriousness, but we are not to assign it an absolute superiority over other forms of knowledge so that they are neglected, relegated to the status of mere opinion. Our discussion has taken science off the pedestal of rational invulnerability and placed it in the arena of human discourse. It is not the only subject with something worth saying. If differing disciplines, such as science and theology, both have insights to offer concerning a question (the nature of humanity, for example), then each is to be listened to with respect at its appropriate level of discourse.

Finally, mathematics itself, the natural language of physical science, has not proved exempt from critical reassessment. This surprising development springs from G5del's theorem.21 This result asserts that in any mathematical system sufficiently complex to include arithmetic (that is, containing the whole numbers), there are propositions which are capable of being stated but not capable of being either proved or disproved. These undecidable propositions are, moreover, known not to be just pathological oddities but to include results of manifest significance. Instead of the completeness we would have expected from mathematics, it appears that every interesting mathematical system is open and incomplete. Because that is so, the consistency of mathematical systems becomes an incalculable question. Thus, even the exercise of mathematics involves an act of faith!

Hofstadter commented on this curious state of affairs, "G5del showed that provability is a weaker notion than truth whatever axiomatic system is involved."22 In other words, truth transcends theo-remhood. Even in the austere discipline of mathematics, there is more than meets the calculating eye.

The Health Zen

The Health Zen

There's no magic bullet that will make you slim down without trying. No particular diet that lets you eat a big amount of food and drop pounds quickly. No ab-machine or exercise bike that you see at three fifteen in the morning on an infomercial is truly going to make that much difference to you.

Get My Free Ebook


Post a comment