|
PHILOSOPHY Pierre Duhem’s ‘Aim & Structure of Physical Theory’Victor Suchar Member2 September 2003
A potted history If there is a problem in the philosophy of science, which can be said to be central, it is that of the nature and structure of scientific theories. A philosophy of science, is in effect, little more than an analysis of theories and their roles in the scientific enterprise (Why's that? How well are these theories constructed? Are they well founded? What are their pre-suppositions and are they coherent? Whom should we believe?). My claim in this talk is that Duhem's The Aim and Structure of Scientific Theory (originally published in 1906) was the most profound of these analysis, (related, of course to classical physics), and the precursor of all others. But in only 20 years, Duhem's great work became neglected, even forgotten (outside France) - it did not reflect the then recent and fundamental developments in physics. By late 1920s, it became normal for philosophers of science to construe scientific theories as axiomatic calculi which are given a partial observational interpretation by means of correspondence rules (between statement and experiment), as described in Rudolf Carnap's The Logical Structure of the World, published in 1928. This was an attempt to translate all knowledge claims of both natural and social sciences into a meta-(or higher symbolic) language, which would distinguish observational terms - those which could be verified by experiment - from theoretical terms, and then to use calculative techniques to establish logical connections and judge those claims. It can be described schematically, and in a highly simplified manner, as a black box containing axioms, theorems and mathematical deductions, into which observations would be introduced as inputs, and discernments would result as outputs. This bold experimental concept, commonly referred to as ‘The Received View’ (a term coined by Hilary Putnam in 1962), has been used explicitly or implicitly in most results obtained in the philosophy of science from the late 1920s to 1950s. In 1934, Carnap's work came under attack by Popper in his Logic of Scientific Discovery, which proposed falsification-ism as replacement of inductionism as a methodology of science –which was considered by many as a revision of the first draft, not a change in kind. The Received View remained in place. By 1951 with the publication of Quine’s ‘Two Dogmas of Empiricism’, a seminal work in analytical philosophy, and specifically in the philosophy of science, with ‘Observation and Interpretation’, a Symposium of Philosophers and Physicists, held in Bristol (Popper, Ayer, Feyerband, Bohm, participating), this analysis became subject to fundamental critical attack, which continued to the 1960s. The critical attacks were two pronged: firstly, attacks on specific features of the ‘Received View’, such as the notion of observational-theoretical distinction, show that they are defective; secondly, alternative philosophies of science advanced by Kuhn, Feyerbend and others, rejected the ‘Received View’ and argued for a different conception of theories and scientific knowledge. By late 1960s the consensus among philosophers of science was that the ‘Received View’ was inadequate, but the proposed alternatives also received significant criticism. The 1969 Illinois ‘Symposium on Structure of Scientific Theories’ was convened to bring together the leading philosophers of science (including Bohm, Putnam and Kuhn) to sort out the chaos and look for new directions (which will be the subject of my next year's talk on Kuhn’s ‘Structure, and the Road after Structure’. But my claim in this talk is that the problems inherited from the ‘Received View’ to be sorted out in 1969, were addressed much earlier by Duhem, and that Quine’s reading of his work and the publication of ‘Two Dogmas of Empiricism’ in 1951 which used his ideas, was the first step in the passage from positivism (or logico-empiricism), which underpinned the ‘Received View’ to a pragmatic form of empiricism (the new empiricism). Aim and Structure was finally translated and published by Princeton in 1954, almost fifty years after its original appearance in print. Biographical note and general outline Duhem was born in Paris in 1861 and died in Bordeaux in 1916. He entered the Ecole Normale Superieure, first in the yearly competition, and proved his early promise by completing his dissertation in thermodynamic physics in only three years. However, with no evident fault in the work, the dissertation was rejected. Two years later, Duhem success-fully submitted another dissertation in thermodynamics and earned his doctorate in mathematics. His first dissertation was subsequently published to much acclaim from the scientific community. Apparently, the reason for Duhem's first rejection was that he opposed the then leading chemist at the time, Marcellin Berthelot, by repudiating his thermo-dynamical views on minimal work (evidently not a good strategy even at the Normale at that time). According to Duhem's biographers, Berthellot managed to confine him to teaching at provincial universities. In 1887 Duhem became maitre de conferences (reader) at Lille, where he taught analytical mechanics. Following a pedagogic dispute, however, he moved at Rennes in 1893, and took the chair as Professor of Physics at Bordeaux in 1895, which he occupied until his death. Duhem believed in the vital importance of history to scientific progress and made considerable contributions to the history of science: The Origins of Statics traced the development of the concept of equilibrium; ‘The Evolution of Mechanics’ extended the framework of classical mechanics to include thermodynamics and prepared the ground for the use of variational methods; Saving the Phenomena - An Essay on the Notion of Physical Theory from Plato to Galileo; and, most of all, Le Systeme du Monde, a 10 volume work of which only the first 5 were published during his life (abridged and translated in English as the one volume Mediaeval Cosmology), is recognised as a fundamental work for the study of mediaeval physics and mechanics. His corpus of published research papers and books on thermodynamics, electrodynamics, hydrodynamics, theory of elasticity, acoustics is considerable. Duhem’s work was from the beginning controversial. His staunch Catholic faith was an obstacle to the secular tone of modern science, and Catholics were uneasy with his totally modern opposition to neo-scholasticism (he was particularly unimpressed by the neo-Thomism of the day). He constructed arguments purporting to show both the impotence of science in the service of anti-religious cause, and the uselessness of science as a model for ‘scientific’ intellectual standards supposedly superior to those upheld by the Catholic faith (i.e. uselessness of science for apologetic purposes). In effect he was an iconoclast, considered as a Catholic apologist by many non-Catholics, and as a ‘modernist’ by the French Catholic intellectual elite. In addition, he made some major mistakes: along with Oswald and Mach, and against Boltzmann, he was an energeticist who opposed the concept of atomic structure, and took energy and thermodynamics as fundamental. There were also many errors in his studies of Mediaeval Scholastics. According to his, not unsympathetic biographer, R N D Martin ‘Duhem was a brilliant maverick who continually got things frustratingly wrong: producing brilliant arguments against atomic explanations in physics and chemistry, a muddled instrumentalist in the philosophy of science, and a voluminous collection of misreadings of mediaeval Scholastics’. Like others who pioneer major new territory, he took many chances - big men, in the process of discovery, make big mistakes - but his work remains seminal. Martin’s observation that ‘Duhem seems to have fallen between every available stool’ seems very apt today. Let’s consider for a moment his ‘muddled instrument-alism’, since his definition of physical theory put forward early in the Aim and Structure is the quintessence of instrumentalism: ‘A physical theory is not an explanation. It is a system of mathematical propositions, deduced from a small number of principles, which aims to represent as simply, as completely, and as exactly as possible, a set of experimental laws.’ For Duhem, an explanation is a metaphysical entity and science should be independent of metaphysics. But this is not intended, as with the positivists, to dispose of metaphysics as irrational or meaningless; it is rather an assertion of the autonomy and dignity of metaphysics as alone capable of expressing the truth of how things are in the world. Metaphysics, according to Duhem is not independent of experience; but its methods are not those of science, and its conclusions stand independent of the changing fashions in science. There is a complete separation between the spheres of influence of science and of metaphysics. Understandably, he infuriated both the positivists and the religious apologists. Furthermore, it is useful to know about his influences. We should note when reading his work that he is not a neo-Kantian, as he appears to many, but a Pascalian. According to the great French mathematician Emile Picard, who knew Duhem well (and who wrote his biography in 1922): ‘He was once labelled a Kantian, but no, Duhem depends not on Kant, but on Pascal, whose Pensees he knows by heart.’ So what did Duhem take from Pascal? First of all is the distinction between the spirit of ‘geometrie’ and that of ‘finesse’, between geometrical and intuitive minds that Pascal sets out in the Pensees using the term ‘geometrie’ for pure deductive reason, and ‘finesse’ for intuitive judgement, those mental abilities that escape deductive reason. In effect, Duhem maintains that deductive reason or ‘geometrie’ alone cannot establish any science, whether rational or experimental, and that any sound work needs the assistance of intuitive factors, such as common sense (bon sens) and finesse - principally the later. According to Martin, ‘Duhem establishes the insufficiency of deductive reason by a dialectical argument in which successive attempts of deductive reason to impose its empire on physics are defeated, leaving finesse or intuitive argument in charge. The dialectical structure means that points established early in the work do not stand for Duhem’s final position, like the reductio ad absurdum proof in geometry, they may only be there for subsequent modification or refutation. In the name of bon sens, Duhem rejects any position that would leave physical theory completely determined by logic, and at the same time in the name of common sense he rejects any theory that excludes logical ordering. He argues for balance between deductive reason and intuitive judgement in physical theory, and Aim and Structure exemplifies that concept in its own structure.’ This, in my view, is an apt statement about this book, and shows, I hope, that Duhem’s instrumentalism was not muddled - on the contrary, it is important both for the contemporary practice of physics and of the philosophy of physics. I will make here a large claim for Duhem - in his work he made six fundamental statements of great importance today: [1] There is no fundamental distinction between the theoretical and observational aspects of physics. This notion was used by Quine in his seminal ‘Two Dogmas of Empiricism’ to attack the second dogma: verificationism and reductionism (the paper was originally read at the American Philosophical Association in 1950, and published in the Philosophical Review (1951). Quine states in ‘Two Dogmas’ that ‘the dogma of reductionism survives in the supposition that each statement, taken in isolation from its fellows, can admit of confirmation or information at all. My counter-suggestion, issuing essentially from Carnap’s doctrine of the physical world in The Aufbau, is that our statements about the external world face the tribunal of sense experience not individually but only as a corporate body. This doctrine was well argued by Duhem’. [2] The under-determination of any theory by empirical facts - in essence the notion that any body of evidence can be explained by any number of mutually incompatible theories. Duhem believed that under-determination arose because of the laboratory conditions in which experiments are conducted. The notion was later universalised by Quine and became the famous Duhem-Quine thesis. [3] An experiment in physics cannot condemn an isolated hypothesis. [4] A crucial experiment is im-possible in physics. [5] The provisional and approximate nature of physical law. [6] The role of intuition in the choice of hypotheses. These statements taken together form the basis of the attack and subsequent demise of the ‘Received View’ both for the inductionist/verificationist and the falsificationist, making him not only the precursor, whose work we neglect at our peril, but also (once we rediscover him) the founder and leading figure of the philosophy of science in the 20th century - whose work is still valid as the basis for the new pragmatic empiricism which began the 60s. In many ways, the rediscovered Duhem is the midwife of the transition from positivism to the new pragmatism of Quine and later, Kuhn. Quine states in ‘The Two Dogmas’ that ‘We can improve our conceptual scheme, our philosophy bit by bit while continuing to depend on it for support; but we cannot detach ourselves from it and compare it objectively with an unconceptualised reality. Hence it is meaningless, I suggest, to inquire into the absolute correctness of a conceptual scheme as a mirror of reality. Our standard for appraising basic changes in the conceptual scheme must be not a realistic standard of correspondence to reality, but a pragmatic standard. Concepts are language, and the purpose of concepts and of language is the efficacy in communication and prediction. On this theme see Duhem.’ To Duhem, the construction of physical theory is more reflective of the contingencies of nature than of the necessities imposed by logic. As a practising scientist, he understood that science is not a pure endeavour. But there is more about this question of importance to contemporary practice, as I hope to say in the Conclusion, after we look at some significant statements from this book. Duhem is often paraphrased, even caricatured but rarely heard directly. Some significant statements made by Duhem in Aim and Structure: On experiments in physics [1] An experiment in physics is not simply the observation of a phenomenon; it is the theoretical interpretation of this phenomenon. (NB: this is the Duhem notion used by Quine in his seminal paper ‘Two Dogmas of Empiricism’.) The aim of all physical theory is the representation of experimental laws. The words ‘truth’ and ‘certainty’ have only one significance with respect to such theory; they express the concordance between the conclusions of the theory and the rules established by the observers. We cannot, therefore, push our critical examination of physical theory further if we do not analyse the exact nature of the laws stated by experimenters, and if we do not note precisely what sort of certainty they yield. Moreover, a law of physics is but the summary of a very large number of experiments that have been ask the questions: What exactly is an experiment in physics? Is there a need to raise it, and is the answer not evident? What more does ‘doing an experiment in physics’ mean to anybody producing a physical phenomenon under conditions such that it can be observed exactly and minutely by means of appropriate instruments? An experiment in physics is the precise observation of phenomena accompanied by an interpretation of these phenomena. This interpretation substitutes for the concrete data really gathered by observation, abstract and symbolic representations which correspond to them by virtue of the theories submitted by the observer. [2] The result of an experiment in physics is an abstract and symbolic judgement. The characteristics, which so clearly distinguish the experiment in physics from common experience, by introducing into the former, as an essential element, a theoretical interpretation, excluded from the latter, mark the results arrived at by these two sorts of experiences. The result of common experience is the perception of a relation between diverse concrete facts. Having been artificially produced, some other fact has resulted from it. The result of the operations, in which an experimental physicist is engaged, is by no means the perception of a group of concrete facts - it is the formulation of a judgement interrelating certain abstract and symbolic ideas, which theories alone correlate with the facts already observed. Open any report of an experiment in physics and read its conclusions - in no way are they simply an exposition of certain phenomena. They are abstract propositions to which no meaning can be attached if the physical theories admitted by the author are not known. When we read, for example, that the electromotive force of a certain gas battery increases by so many volts when pressure increases by so many atmospheres, what does this proposition mean? We cannot attribute any meaning to it without recourse to the most varied and advanced theories of physics. Pressure is a quantitative symbol introduced by theoretical mechanics, and one of the subtlest notions which science has to deal with. In order to understand ‘electromotive force’ we must appeal to the electro-kinetic theory founded by Ohm and Kirchoff. The volt is the unit of electromagnetic force in the practical electromagnetic system of units - the definition of this unit is drawn from the equations of electromagnetism introduced by Ampere and Maxwell. Not one of the words serving to state the result of such an experiment directly represents a visible and tangible object - each of them has an abstract and symbolic meaning related to concrete realities only by long and complicated theoretical intermediaries. Isn’t this simply a language which the experimenter uses and the general public does not - a kind of French for an English speaker who does not have a dictionary? In other words, a technical language in which a physicist expresses the results of his experiments, and which the initiate can translate into facts? It resembles it; but differs in that a given sentence in a technical language expresses a specific operation performed on very specific objects, whereas a sentence in the physicist’s language may be translated into facts in a variety of different ways. Duhem, a physicist of great experience, was very much opposed to the conventionalism of Poincaré, the greatest mathematician of the era, for whom physical theory should be simply a vocabulary allowing one to translate concrete facts into a simple, convenient conventional language. According to Poincaré, ‘a scientific fact is nothing but a brute fact translated in a convenient language’, and again: ‘All that the scientist creates in a fact is the language in which he states it.’ When I observe a galvanometer (states Duhem), and ask a lay visitor ‘Is the current passing through?’ he goes and looks at the wire in order to see if something is passing through it. But if I put the same question to my assistant he will realise that that means, ‘Is the spot displaced’ and will look at the scale. Is it then correct to say that the words ‘the current is on’ are simply a conventional manner of expressing the fact that the magnetised little bar of the galvanometer has deviated? But the assistant may well answer: ‘The current is on, and yet the magnet has not deviated - the galvanometer shows some defect.’ Why does he say that the current is on despite the absence of the galvanometer reading? Because he has observed that in a voltameter, placed in the same circuit as a galvanometer, bubbles of gas were released; or that an incandescent light inserted on the same wire was glowing; or that a coil which this wire is wrapped is becoming warm; or that a break in the conductor was accompanied by sparks. In virtue of accepted theories, each of these facts, as well as the deviation of the galvanometer, may be translated by the words ‘the current is on’. This group of words does not express in a technical and conventional language a certain concrete fact. As a symbolic formula it has no meaning for one ignorant of physical theories, but meaning for one who knows these theories; it can be translated into concrete facts in an infinity of different way, because all these disparate facts admit the same theoretical interpretation. The role of the physicist is not limited to creating a clear and precise language in which to express concrete facts; rather, it is to create a clear and precise language that presupposes the creation of a physical theory. Between an abstract symbol and a concrete fact there may be a correspondence, but not complete parity. The abstract symbol cannot be an adequate representation of the concrete fact. The concrete fact cannot be the exact realisation of the abstract symbol. The abstract and symbolic formula, by which a physicist expresses the concrete facts he has observed in the course of an experiment, cannot be the exact equivalent or the faithful story of these observations. The disparity between the practical fact really observed and the theoretical fact, the symbolic, abstract formula stated by the physicist, is revealed to us when different concrete facts interpreted by a theory fuse into one another to constitute but one and the same experiment, and they are expressed by a single symbolic proposition: The same theoretical fact may correspond to an infinity of distinct practical facts [3] The theoretical interpretation of phenomena alone makes possible the use of instruments. On physical law [1] The laws of physics are symbolic relations. Just as the laws of common sense are based on observation of facts by means natural to man, so the laws of physics are based on experiment. Let’s consider one of the simplest and most certain of common sense laws: All men are mortal. The law relates two abstract concepts - the abstract idea of man in general rather than the concrete idea of man in particular, and the abstract idea of this or that form of death. Indeed it is only on this condition (i.e. that the concepts related are abstract) that the law can be general. But these abstractions are in no way theoretical symbols, they merely extract what are universal in each of the particular cases to which the law applies. Thus, in each of the particular cases where we apply the law, we shall find concrete objects in which these abstract ideas are realised. Let’s take another law, one about an object belonging to physics, but when this particular branch of knowledge existed only as a dependency of common sense, without reaching the stage of a rational science. We see a flash of lightning before we hear thunder. The ideas of lightning and thunder, which tie this statement together, are abstract and general ideas drawn naturally from particular data that with each bolt of lightning we perceive a glare and a rumbling which we recognise as the concrete form of our ideas of lightning and thunder. This is not true of the laws of physics. Take for example, Boyle’s law: at a constant temperature, the volumes occupied by a constant mass of gas are in inverse ratio to the pressure they support. The terms it introduces: the ideas of mass, temperature, pressure, are still abstract ideas. But these ideas are not only abstract, they are also symbolic: the symbols assume meaning only by grace of physical theories. Let’s put ourselves in front of a real gas to apply Boyle’s law. We are not dealing with a certain concrete temperature embodying the general idea of temperature, but with some more or less warm gas - we are not facing a certain pressure embodying the general idea of pressure, but a certain pump on which a weight is brought to bear in a certain manner. No doubt, a certain temperature corresponds to this more or less warm gas, and a certain pressure corresponds to this effort exerted on the pump, but this correspondence is by no means immediately given - it is established with the aid of instruments and measurements, and is a complicated process. In order to assign a definite temperature to this more or less warm gas, we must have recourse to a thermometer; and in order to evaluate in the form of pressure the effort exerted by the pump, we must use a manometer. To use a thermometer and a manometer implies the use of physical theories. The abstract terms referred to in a common sense law, being no more than whatever is general in the concretely observed objects, the transition from the concrete to the abstract is a necessary and spontaneous operation that remains unconscious. This instinctive and unreflective operation yields unanalysed general ideas - abstractions taken grossly, so to speak. No doubt, the thinker may analyse these general ideas; he may wonder what man is, what death is, and seek to penetrate the deep and full sense of these words. This inquiry will lead him to a better understanding of the reasons for the law, but not necessarily to understand the law - it is sufficient to take the terms related in their obvious sense in order to understand this law. The symbolic terms connected by a law of physics are not the obvious abstractions that emerge from concrete reality - they are abstractions produced by slow, complicated and conscious work - the labour that has elaborated the physical theories. If we do not do this work or if we do not know physical theories, we cannot understand the law or apply it. According to whether we adopt one theory or another, the very words which figure in a physical law change their meaning, so that the law may be accepted by one physicist who admits a certain theory, and rejected by another who admits some other theory. Consider for example the following physical law: ‘All gases contract and expand in the same manner’, and let’s ask different physicists if this law is, or is not, violated by iodine vapour. The first physicist professes theories according to which iodine vapour is a single gas, and draws from the foregoing law the consequence that the density of iodine vapour relative to the air is a constant. Experiment shows that this density depends on the temperature and pressure. Our physicist therefore concludes that iodine vapour is not subject to the stated law. A second physicist will state that iodine vapour is not a single gas, but a mixture of two gases, which are polymers of each other and capable of being transformed into each other. Consequently, the law mentioned does not require the iodine vapour density relative to the air to be constant, but claims this density varies with temperature and pressure according to a formula established by Willard Gibbs. This formula represents the results of experimental determinations. Our second physicist concludes that iodine vapour is not an exception to the rule that all gases contract or expand in the same manner. Thus the two physicists have different opinions concerning a law which both enunciate in the same form: one finds fault because of a certain fact, the other finds it is confirmed by that very fact. The different theories they hold do not determine uniquely the meaning suited to the words ‘a single gas’ so that though they both pronounce the same sentence, they mean two different propositions, and in order to compare his proposition with reality each makes different calculations, so that is possible to for one to verify this law which the other’s finds contradicted by the same facts. This is plain proof of the following truth: A physical law is a symbolic relation whose application to concrete reality requires that a whole group of laws be known and accepted. [2] A law of physics is, properly speaking, neither True nor False but Approximate. A common sense law is merely a general judgement - this judgement is either true or false. This is not the case with the laws of physical science at maturity, which are stated in the form of mathematical propositions - such laws are symbolic. But a symbol is neither true nor false, it is selected to stand for the reality it represents, and pictures that reality in a more or less precise or detailed manner. To apply to a symbol the words ‘truth’ or ‘error’ no longer has any meaning, so the logician concerned with the strict meaning of words will have to answer anyone who asks whether physics is true or false, as ‘I do not understand the question’. This answer seems paradoxical. The experimental method, as practised in physics, does not make a given fact correspond to only one, but to a variety of symbolic judgements - the degree of symbolic indetermination is the degree of approximation of the experiment in question. Finding a law for a sequence of analogous facts means the physicist must find a formula which contains the symbolic representation of each of these facts. The symbolic indetermination corresponding to each fact consequently entails the indetermination of the formula, which must unite these symbols - we can make a variety of different formulas or distinct physical laws correspond to the same group of facts. In order for these laws to be accepted, each fact should correspond not the symbol of this fact, but to one of the symbols, infinite in number, which can represent the fact. This is what is meant by the statement ‘the laws of physics are approximate’. (This notion is the basis of the famous Duhem-Quine ‘Principle of Under-Determination’.) The physicist will prefer one law to another, the first to follow theories he admits. For example, he will ask for the theory of universal attraction to decide which formulas he should prefer among all those which could represent the motion of the sun. But physical theories are only a means of classifying and bringing together the approximate laws to which experiments are subject. Theories, therefore, cannot modify the nature of these experimental laws and cannot confer absolute truth on them. Thus, every physical law is approximate. Consequently, it cannot be, for the strict logician, either true or false - any other law, representing the same experiments with the same approximation, may lay claim as the first to the title of a true law, or more precisely, of an acceptable law. [3] Every law of physics is provisional and relative because it is approximate. What characterises a law that is fixed and absolute? Is it not contradictory to say that the law is provisional and acceptable to one person, but not to another? No, if by laws we mean the laws that physics states in mathematical form. Such laws are provisional, not because a law is true for a certain time, then false, but because at no time is it either true or false. It is provisional because it represents the facts to which it applies with an approximation that physicists today judge to be sufficient, but will some day cease to judge satisfactory. Such a law is always relative, not because it is true for one physicist and false for another, but its approximation is sufficient for the use the first physicist, but not sufficient for the second. The degree of approx-imation is not fixed - it increases gradually as instruments are perfected. As experimental methods improve, we lessen the indetermination of the abstract symbol brought into correspondence with the concrete facts of physical experiment. As the indetermination of experimental results becomes narrower, the indetermination of the formulas used to condense these results become more restricted. The laws of physics are therefore provisional: the symbols too simple to represent reality completely. The task of continual modification by which the laws of physics avoid more and more adequately the refutations provided by experiment, plays an essential role in the development of science. Scientific laws, based on the experiments of physics, are symbolic relations whose meaning remains unintelligible to those who do not know physical theories. Since they are symbolic, they are never true or false - like the experiments on which they rest - they are approximate. The degree of approximation of a law, sufficient today, will become insufficient in the future as experimental methods progress sufficient for the needs of the physicist, so that a law of physics is provisional and relative - provisional also because it does not connect realities but symbols. And there will always be cases where the symbol no longer corresponds to reality. The laws of physics must be continually retouched and modified. A law of physics possesses a less immediate and more difficult certainty than a law of common sense, but it surpasses the latter by the minute and detailed precision of its predictions.
On physical theory and experiment [1] An experiment in physics can never condemn an isolated hypothesis, but only a whole theoretical group. A physicist, deciding to demonstrate the inaccuracy of a proposition in order to deduce from this proposition the prediction of a phenomenon, institutes an experiment to show whether this phenomenon is or not produced. In order to interpret the results of this experiment and establish that the predicted phenomenon is not produced, he does not confine himself to the use of the proposition in question, but makes use of a whole group of acceptable, undisputable theories (auxiliary theories). The prediction of the phenomenon whose non-production is to cut off debate, does not derive from the proposition challenged if taken by itself, but from the proposition at issue joined to that group of theories (the auxiliary theories). If the predicted phenomenon is not produced, not only is the proposition questioned at fault, so is the physicist’s whole theoretical scaffolding. This experiment teaches us that among the propositions used to predict the phenomenon and to establish whether it would be produced, there is at least one error - but it does not tell us where the error lies. The physicist might declare that this error is contained in exactly the proposition he wishes to refute, but is he sure that it is not in another proposition? If he is, he accepts implicitly the accuracy of all the other propositions he has used, and the validity of his conclusion is as great as the validity of his confidence. (NB the frequent example of looking at the stars and offering astronomical theories which willy-nilly include theories of optics, on how telescopes work, and general electro-magnetic theory, etc., shows that we are not testing, refuting, or giving probable evidence for any one hypothesis, say about planetary motions, but a hypothesis and auxiliaries. There is not much to argue about that). [2] A crucial experiment is impossible in physics. Between two contradictory theorems of geometry there is no room for a third judgement. If one is false, the other is necessarily true. Do two hypotheses in physics ever constitute such a strict dilemma? Unlike the reduction to absurdity employed by the geometers, experimental contradiction does not have the power to transform a physical hypothesis into an indisputable truth. In order to confer this power to it, it is necessary to enumerate completely the various hypotheses that may cover a determinate group of phenomena. But the physicist is never sure he has exhausted all the imaginable assumptions. The truth of a physical theory is not decided heads or tails. (NB: This is called the strong Duhem thesis. Taken together with his earlier theses about the entanglement of interpretation and observation, and about under-determination, they are used to deny both verificationism and falsificationism - the concepts that underpin the ‘Received View’.)[3] Certain physical theories rest on hypotheses, which do not by themselves have any physical meaning (theoretical terms). For example, the principle of equality of action and reaction may be stated as ‘The centre of gravity of an isolated system can have only an uniform rectilinear motion.’ Can we verify this principle? For that it is necessary for isolated systems to exist. But these systems do not exist - the only isolated system is the whole universe. We can observe only relative motions. The absolute motion at the centre of the universe, therefore, remains unknown. We can never know if it is rectilinear and uniform. The question has no meaning. Whatever facts we may observe, we shall henceforth be free to assume our principle is true. In truth, hypotheses, which by themselves have no physical meaning, undergo experimental testing in the same manner as other hypotheses. Whatever the nature of a hypothesis is, it is never contradicted in isolation by an experiment - experimental contradiction always bears as a whole, on the entire group constituting a theory without any possibility of designating which proposition in this group should be rejected. How then do we decide which hypotheses should be abandoned? [4] Good sense is the judge of hypotheses that ought to be abandoned. When certain consequences of a theory are struck by experimental contradiction, we learn that this theory should be modified, but are not told by the experiment what must be changed. It is left to the physicist to find the weak spot that impairs the whole system. (But what if we change auxiliaries in order to save a fundamental hypothesis? Take for example, a different kind of hypothesis, one that had a different career in the history of science, starting well after Duhem. The hypothesis of the Neutrino was offered as a shift in an auxiliary theory ‘A’. By shifting ‘A’, one was able to preserve a certain hypothesis ‘H’ about the conservation of energy and momentum - fundamental in physics. It seemed that one could preserve what appeared to be refuted by data by shifting auxiliary assumptions. The auxiliaries are changed by including a new assumption about the existence of an otherwise unverifiable particle, and therefore cast doubt on the status of verifiability in scientific theories. It all turned happily in the end, but at the time, physicists were unable to give independent, empirical confirmation of the existence of such a particle. This new auxiliary was a Duhem way of saving a hypothesis - if one wanted to pay the price of postulating an almost mass less and charge-less particle, then one could save the fundamental conservation law. It appeared that one could save what one wanted by paying the price of assuming the existence of something else. But then, the process of saving is itself based on intuition and on the consensus in the scientific community of the importance of what is to be saved). Since logic does not determine with strict precision the time when an inadequate hypothesis should give way to a more fruitful assumption; and since recognising that the moment belongs to good sense, physicists may hasten this judgement and increase the rapidity of scientific progress by trying consciously to make good sense more lucid and vigilant. But nothing impairs good sense and disturbs insight more than passions and interests. In order to estimate correctly the agreement of a physical theory with the facts, it is not enough to be a good mathematician and skilful experimenter, one must be also an impartial and faithful judge. [5] The importance in physics of the historical method. In geometry the clear knowledge obtained by deductive logic and the certainty stemming from common sense are so exactly juxtaposed that we cannot discern the mixed zone where all our knowledge operates simultaneously and in rivalry - that is why the mathematician wants to construct in physics a perfect science. In its pursuit he creates what Ernst Mach calls ‘a false rigour’. How do we protect against the dangers of such methods? How can we survey the enormous territory separating the domain of ordinary experience from the theoretical domain of clear principles? The legitimate and fruitful method of preparing a student to receive a physical hypothesis is the historical method - to retrace the transformations through which empirical matter accrued when the theoretical form was first sketched. Besides, the history of science can keep the physicist from the mad ambitions of dogmatism as well as the despair of scepticism - every time the physicist is on the point of going to an extreme, the history can rectify him with an appropriate correction. On metaphysics Physics is independent of metaphysics. I have constantly aimed to prove that physics proceeds an autonomous method independent of any metaphysical opinion. Having carefully analysed the method in order to exhibit through this analysis the proper character and exact scope of the theories that summarise and classify its discoveries, I have denied that these theories have any ability to penetrate beyond the teaching of experiment or any capacity to surmise realities hidden under data observable by the senses; and have denied these theories the power to draw the plan of any metaphysical system, as I have denied metaphysical doctrines the right to testify for or against any physical theory. Our reflections on the meaning and scope of physical theories were induced by preoccupations that have nothing to do with metaphysical doctrines and nothing to do with religious dogmas.
Tendency toward an ideal form Physical Theory has as its limiting form a natural classification. No scientific method carries within itself its full and entire justification. It cannot through its principles alone, explain all these principles. We should not be astonished that theoretical physics rests on postulates that can be authorised only by reasons foreign to physics (for example, the Principle of Least Action - teleological or organisational). The right of a physicist to develop a logically incoherent theory is arrived at by those who analyse the method of physics without recourse to any principle foreign to the method. For them the representations of theory are convenient summaries and artificial devices aimed to facilitate the work of discovery. Why should we forbid workers the successive use of disparate instruments to those adapted to a certain task and not so well adapted to others? This doctrine with its sceptical overtones shocks most of those working to advance physics. Although the merely logical study of the procedures they employ does not provide convincing arguments in support of their viewpoint, they feel that this way is the right one; they have an intuition that logical unity is imposed on physical theory as an ideal to which it tends constantly. And this conviction is shared even by those who defend the right of a theory to logical incoherence - there is no hesitation among them for the preference of a rigorously coordinated theory to a junk heap of irreconcilable theories. Therefore it is not with a whole heart that they proclaim the right to logical incoherence - like all physicists they regard the physical theory which would represent all experimental laws by means of a single, logically coordinated system as the ideal theory, but they regard this as an unrealisable ideal. The physicist is then lead to justify the tendency toward logical unity by the following metaphysical assertion- ‘The ideal form of physical theory is a natural classification of experimental laws’.On the value of physical theory Is theory an artificial construction which makes the truths of empirical knowledge easier to handle, enabling us to make prompter and more advantageous use of it in our acting on the external world, but teaching us nothing concerning the world which is not already taught by experiment alone? Or, does theory teach us something concerning reality which experiment has not and cannot possibly teach us, something transcending merely empirical knowledge? If we answer the last question affirmatively, we are saying that physical theory has value as knowledge, but if we answer the first question affirmatively (is theory a construct), we are saying that physical theory is merely convenient, that it has no value as knowledge, but only practical value. According to Duhem: ‘Physical theory confirms to us a certain knowledge of the external world which is not simple reducible to empirical knowledge; this knowledge comes neither through experi-ment nor from the mathematical procedures employed by theory, so that the merely logical dissection of theory cannot discover the fissure through which this knowledge is introduced into the structure of physics; through an avenue whose reality the physicist cannot deny, any more than he can describe its course, this knowledge derives from a truth other than the truths apt to be possessed by our instruments; the order in which theory arranges the results of observation does not find its adequate and complete justification in its practical or aesthetic characteristics. We surmise, in addition that there is or tends to be a natural classification. Through an analogy whose nature escapes the confines of physics but whose existence is imposed on the mind of the physicist, we surmise that it corresponds to a certain supremely eminent order. In a word, the physicist is compelled to recognise that it would be unreasonable to work for the progress of physical theory if this theory were not the increasingly better defined and more precise reflection of a metaphysics - the belief in an order transcending physics is the sole justification of physical theory.’ (The alternating hostile or favourable attitude that a physicist may take to this affirmation is summarised in Pascal's words: ‘We have the impotence to prove invincible by any dogmatism, and we have an idea of truth invincible by any scepticism.’) Conclusion The principles of the new Duhem-Quine empiricism, based on the statements made in Duhem’s ‘Aim and Structure’ and can be summarised as follows: [1] There is no need to make a fundamental epistemological distinction between the theoretical and observational aspects of science either in regard to decidability of truth-value or transparency of empirical meaning. [2] Empirical applications of observation statements are not incorrigible, and the empirical laws are not infallible. [3] The meaning of a physical law is to be determined by the context of scientific practice and the network of related laws involved in determining the meaning of that law. This principle provides the basis for the current under-determinist perspective on the relationship between experimental evidence and theory and the network of related theories in the work of Quine and Davidson. We may note here that under-determination gives rise to the problem of theory choice - how do we choose a theory among all those supported by the same body of evidence - a very complex problem. Duhem, as we have seen from his statements, believed that inadequate theories could be rejected by using intuition. Quine replaced intuition with an ongoing inquiry into the track record - the theory with the best track record is the theory of choice - a kind of evolutionary naturalism. We could say that it is a testimony to the seminal character and influence of Duhem's work, that today it is thought necessary to return to his philosophic understanding of scientific practice and of the role of history in science (Feyerabend, Kuhn). These insights unfortunately failed to be transmitted into subsequent debates until the 1950s. The main conceptual difference between his work and that of his successors, particularly Carnap, is that he attempted to construct a rough guide using the context and the language of scientific practice, not a meta-language expressed axiom-atically in which to translate all claims of knowledge in order to expose logical relations. It is this contextual approach which takes into account both scientific practice and the requirement for logical coordination to which we return. Victor Suchar References Duhem, Pierre: The Aim & Structure of Physical Theory, Princeton Univ., 1991 (first ed. 1906; first English ed. Princeton 1954). Koerner, S. (ed): Observation & Interpretation. Symposium of Philosophers & Physicists. Proceedings of the Ninth Symposium of the Colston Research Society, Univ. of Bristol, 1957 (London: Butterworth Scientific Publications, 1957) Martin, R N D: Pierre Duhem; Philosophy & History in the Work of a Believing Physicist (La Salle, Illinois, Open Court, 1991) Picard, Emile: La Vie et l'Oeuvre de Pierre Duhem (Paris: Gauthier-Villars, 1922) Quine, Willard V O: ‘Two Dogmas of Empiricism’ in From a Logical Point of View (NY: Harper & Row, 1963 [first ed. Harvard 1953]) Suppe, Frederick (ed): The Structure of Scientific Theories. Proceedings of the Symposium held in Urbana Illinois, 1969. ([first ed.1973] Univ. of Illinois, enlarged ed. with Afterword by Suppe, 1977
|
