The Nature of Things

Presidential Address

given to the British Society for the Philosophy of Science


J.R. Lucas

on June 7th, 1993

It would be improper for a President to play safe. After two years of curbing my tongue and not making all sorts of observations that have sprung to my mind, in order to let you have an opportunity of having your say, I am now off the leash. And whereas mostly in academic life it is appropriate to adopt a prudential strategy, and not say anything that might be wrong, I owe it to you on this occasion to play a maximax strategy, to speak out and say what I really think, being willing to run the risk of being wrong in order not to forgo the chance of actually being right in an area of the philosophy of science which must, I think for ever, be largely a matter of metaphysical speculation.

I stand before you a failed materialist. Like Lucretius, from whom I have borrowed my title, I should have liked to be able to explain the nature of things in terms of ultimate thing-like entities of which everything was made, and in terms of which all phenomena could be explained. I have always been a would-be materialist. I remember, when I was six, telling my brother, who was only two, in the corner of a garden in Guildford, that everything was made of electricity and believing that electrons were hard knobbly sparks, and later pondering whether position was a sort of quality, and deciding that it was, absolutely considered, but that relative positions, that is to say patterns, as seen in the constellations in the sky, were only in the eye of the beholder. I am still impelled to a very thing-like view of reality, and would like to explain electricity in terms of whirring wheels, and subatomic entities as absolutely indivisible point-particles, each always remaining its individual self, and possessed of all the qualities that really signified. I find it painful to be dragged into the twentieth century, and though my rational self is forced to acknowledge that things aren't what they used to be, I find it hard to come to terms with their not being what I instinctively feel they have got to be, and am still liable to scream that that the world-view we are being forced to adopt cannot be true, and that somehow it must be fitted back into the Procrustean bed of our inherited prejudices.

But I am not going to ask you to listen to my screams. Rather, I shall share with you my attempts to overcome them, and work out new categories for thinking about the nature of the world, and a correspondingly less rigid paradigm of possible explanation. It has taken me in two different directions. On the one hand reality is much softer and squodgier than I used to think. It is not only that the knobbliness is less impenetrable, as quantum tunnelling takes over, nor that it is fuzzier, without the sharp outlines of yestercentury, but, more difficult to comprehend, the very concept of haecceitas, as Duns Scotus called it, this-i-ness, or transcendental individuality, in Michael Redhead's terminology, 1 has disappeared from the categories of ultimate reality. On the other hand, reason has become much wider and more hospitable to new insights from various disciplines. The two changes are connected. Our concept of a thing, in order to be more truly a thing, has been developed into that of a substance, and substances have come to need to have more and more perfections, and we have therefore come to identify as substances more sophisticated combinations of more recherché features; and with this change in what we regard as a thing has come also a corresponding change in our canons of explanation. It will be my chief aim this evening to show how our changed apprehension of reality has opened up new vistas of rationality, and how the wider concept of rationality we have been led to adopt has in turn altered our view of what constitute real substances.

The corpuscularian philosophy posited the ultimate constituents of the universe as qualitatively identical but numerically distinct, possessing only the properties of spatial position and its time-derivatives, and developing according to some deterministic law. In the beginning, on this view, God created atoms and the void. The atoms, or corpuscles, or point-particles, were thing-like entities persisting over time, each for ever distinct from every other one, each always remaining the same, each capable of changing its position in space while retaining its individual identity. Spatial position constituted the changeable parameter which explained change without altering the corpuscle's own identity. Space was the realm of mostly unactualised possibility, of changes that might, but mostly did not, occur. But space also performed the logical function of both distinguishing between qualitatively identical corpuscles---two thing-like entities cannot be in the same place at the same time---and providing, in spatio-temporal continuity, a criterion of identity over time. It was thus possible for each point-particle to be like every other one, but to be a different particular individual, and this particularity affected the corpuscularians' ideal of explanation, articulated by Laplace, and much refined in our own time by Hempel. Scientists seek generality, and eschew the contingent and the coincidental. In the Hempelian paradigm, the focus of interest is on the covering law, which is general, and not on the initial conditions, which just happen to be what they are, and can only themselves be explained by the way earlier states of the universe happened to be. Boundary conditions, being the particular positions and velocities of particular point-particles, are too particular to constitute the sort of causes that scientists, in their search for generality, are willing to take seriously as genuinely explanatory.

The corpuscularian philosophy had many merits. It reflected our experience of things: stable objects that persist over time, clearly individuated by exclusive space-occupancy, capable of change without losing their identity. As a metaphysical system it had great economy and power. All macroscopic things, all events and phenomena, were to be explained in terms of the positions and movements of these ultimate entities. There was a clear ontology, a clear canon of explanation, and a clear demarcation between physically necessary laws and purely contingent initial conditions. Of course, there were also grave demerits. From my own point of view---though I have failed to persuade Robin Le Poidevin of this 2---time is essentially tensed, and it counts against the corpuscularian scheme that it did not account for the direction of time or the uniqueness of the present: more influential in the history of science was the account of space, and the difficulty in formulating a plausible account of how corpuscles could interact with one another, which in due course led us to replace corpuscularian by field theories, as being better able to account for the propagation of causal influence. The vacuum, though adequate for giving things room to exist and move in, was too thin to let them interact with one another, and Voltaire has had to return from London to Paris.

But it was not only space that proved too thin to do its job. The ultimate thing-like entities not only failed to accommodate the things of our actual experience, but have turned out not to be thing-like at all. Although the atoms of modern chemistry and physics are moderately thing-like, subatomic entities are not. We do not obtain predictions borne out by observation if we count as different the case of this electron being here and that there and the case of that being here and this there. Instead of thinking of the word `electron' being a substantive referring to a substantial, identifiable thing, we do better to think of it as an adjective, with some sense of `negatively charged hereabouts'. We do not feel tempted to distinguish two pictures, one of which is red here and red there, and the other of which is red there and red here; the qualities referred to by adjectives lack haecceitas, this-i-ness, and are real only in so far as they are instantiated.

We are forced to deny this-i-ness to electrons and other sub-atomic entities in order to accommodate empirical observations, but it is not just a brute fact, but rather the reflection of the probabilistic structure of quantum mechanics. The loss of determinateness in our ultimate ontology is the concomitant of our abandoning determinism in our basic scheme of explanation. Probabilities attach naturally not to specific singular propositions, but to general propositional functions, or, as Colin Howson puts it, 3 generic events, or, in Donald Gillies' terminology, 4 repeatable situations. Although you can intelligibly ask what the probability is of my dying in the next twelve months, the answer is nearly always only an estimate, extrapolated from the probabilities of Englishmen, males, oldie academics, non-smokers, non-diabetics, and other relevant general types, not dying within the subsequent year. Calculations of probabilities depend on the law of large numbers, assumptions of equiprobability, or Bayes' Theorem, which all ascribe probabilities to propositional functions dealing with general properties rather than to singular propositions asserting ultimate particularities. If we accept the probabilistic view of the world, we can no longer picture the universe as made up of particular thing-like entities that Newton could have asked Adam to name, but as a featured something, whose underlying propensities could be characterized in quantum-mechanical terms, and whose features calculated up to a point, and found to be borne out in experience.

The loss of particularity legitimises a paradigm shift in our canon of explanation. In his Presidential Address, Professor Redhead noted the shift from a Nineteenth Century ideal, in which we could deduce the occurrence of events granted a unified theory together with certain boundary conditions, to a Twentieth Century schema, which, although less demanding, in as much as it is not deterministic, is more demanding, in that it seeks to explain the boundary conditions too. 5 Outside physics that has always been the case---and often within physics too. It is one of the chief objections to the Hempelian canon, an objection expressed by many of those present here tonight---Nancy Cartwright, John Worrall, Peter Lipton---that it fails to accommodate the types of explanation scientists actually put forward. 6 It depends on the science concerned what patterns of law-like association, to use a phrase of David Papineau's, count as causes. 7 Different sciences count different patterns of law-like associations as causes because they ask different questions and therefore need to have different answers explaining differently with different becauses. The fact that different sciences ask different questions is of crucial importance. Once we distinguish questions from answers, we can resolve ancient quarrels between different disciplines. 8 The biologists have long felt threatened by reductionism, and felt that there was something amiss with the claim that it was all in the Schrödinger equation, or as Francis Crick put it, ``the ultimate aim of the modern movement in biology is in fact to explain all biology in terms of physics and chemistry''. 9 But their claim that there was something else, not in the realm understood by physicists, smacked of vitalism, and was rejected out of hand by all practising physicists. Vitalism made out that answers were in principle unavailable, whereas what is really at issue is not a shortage of answers but an abundance of questions. It was not a case of biologists asking straightforward physicists' questions and claiming to get non-physicists' answers, but of their asking non-physicists' questions, to which the physicists' answers were germane, but could not, in the nature of the enquiry, constitute an exhaustive answer to what was being asked. Biologists differ from physicists in what they are interested in---no hint of vitalism in pointing out that the life sciences investigate the phenomenon of life---and in pursuing their enquiries pick on features which are significant according to their canons of interest, not the physicists'. What is at issue is not whether there is some physical causal process of which the physicists know nothing, but whether there are principles of classification outside the purview of physics. It is a question of concepts rather than causality.

My favourite, excessively simpliste example is that of the series of bagatelle balls running down through a set of evenly spaced pins and being collected in separate slots at the bottom: we cannot predict into which slot any particular ball will go, but we can say that after a fair number have run down through the pins, the number of balls in each slot will approximate to a Gaussian distribution. There is nothing vitalist about a Gaussian distribution, but it is a probabilistic concept, unknown to Newtonian mechanics. In order to recognise it, we have to move from strict corpuscularian individualism to a set, an ensemble, or a Kollectiv of similar instances, and consider the properties of the whole lot. More professionally, all the insights of thermodynamics depend on not following through the position and momentum of each molecule, but viewing the ensemble in a more coarse-grained way, and considering only the mean momentum of those molecules impinging on a wall, or the mean kinetic energy of all the molecules in the vessel. Equally the chemist and the biologist are not concerned with the life histories of any particular atoms or molecules, and reckon one hydrogen ion as good as another, and one molecule of oxygen absorbed in the lungs of a blackbird as good as another. 10 The chemist is concerned with the reaction as a whole, the biologist with the organism in relation to its environment and other members of its species. A biologist is not interested in the precise accounting for the exact position and momentum of every atom, even if that were feasible. Such a wealth of information would only be noise, drowning the signal he was anxious to discern, namely the activities and functioning of organisms, and their interactions with one another and with their ecological environment. It is the song of Mr Blackbird as he tries to attract the attention of Mrs Blackbird that concerns the ethologist. He is not concerned with exactly which oxygen molecules are in the blackbird's lungs or blood stream, but in the notes that he trills as dawn breaks, and their significance for his potential mate. If he were presented with a complete Laplacian picture, his first task would be to try and discern the relevant patterns of interacting carbon, oxygen, hydrogen and nitrogen atoms that constituted continuing organisms, and to pick out the wood from the trees. In this change of focus the precise detail becomes irrelevant. He is not, in Professor Watkins' terminology, a methodological individualist. What interests him is not the life history of particular molecules of oxygen, but the metabolic state of the organism, which will be the same in either case. Different disciplines, because they concentrate on different questions, abstract from irrelevant detail, in order to adduce the information that is relevant to their concerns.

In practice scientists have long recognised that in order to see the wood they must often turn their attention away from the trees. But whereas that shift was to be defended simply as a matter of choice on their part, now it is legitimised by our new understanding of logical status of the boundary conditions we are interested in. If our ultimate theory of everything can talk only in general terms, and cannot assign positions and velocities to particular atoms, it follows that it is no criticism of other theories that they can talk only in general terms too. Hitherto there has been a sense of information being thrown away, information which was there and ultimately important, so that we were, in some profound way, being given less than the whole truth. There was a Laplacian Theory of Everything which was in principle knowable and in principle held the key to all ologies. Every other discipline was only a partial apprehension of ultimate truth, useful perhaps because more accessible for our imperfect minds, but conveying only imperfect information none the less. Just as we rely on journalists to reduce the welter of information about the Balkans or South America to manageable size, so chemists and biologists seemed to select and distil from total truth to tell us things in a form we were capable of taking in. Compared with the high priests of total truth, they were mere popularisers. I may discern Gaussian patterns in long runs of bagatelle balls, but they are patterns only in the eye of an ill-informed beholder: better informed, I should see why each ball went into the slot that it did, and be aware of the occasions when a non-Gaussian distribution emerged. My Gaussian discernment would seem a rough and ready approximation, like describing France as hexagonal, which is fair enough for some purposes, but falls far short of being fully true. Even though the things we pick on as worthy of note and in need of explanation---the shape of the Gaussian curve, the significance of bird-song---lie outside the compass of the limited concepts and explanation of a Theory of Everything, the possession of perfect information trumps curiosity.

The case is altered if there is no fully particularised ultimate reality, and no complete theory of it. We cannot claim that ultimately there are trees which exist in their own right, whereas the woods are only convenient shorthand indications of there being trees there: we cannot trump the different, admittedly partial, explanations put forward by different disciplines by a paradigm one that claims to be complete, nor can we suppose that there is some bottom line that establishes a final reckoning to which all other explanations must be held accountable. All natural sciences concern themselves with general features of the universe, and there is no reason to discountenance any science because it selects some general features rather than others. Questions about boundary conditions cannot, then, be faulted on grounds of their being general, and not ultimately particular. The answers, too, are to be assessed differently, once the mirage of a complete Laplacian explanation is dispelled. Not only is it irrelevant to the ethologist's purposes, which particular mate the blackbird seeks to attract, or which oxygen molecules are in the blackbird's lungs or blood stream, it is, in its precise detail, causally irrelevant too. The blackbird's song is not addressed to a particular Mrs Blackbird in all her individuality, but to potential Mrs Blackbirds in general, and if one mate proves hard to win, another will do. Much more so at lower levels of existence: if one worm escapes the early bird, another will be equally succulent; if one molecule of oxygen is not absorbed by his haemoglobin, another will. Explanations are inherently universalisable, and if the physical universe is one of qualitatively identical features that cannot, even in principle, be numerically distinguished, then the explanations offered by other disciplines are ones that cannot, even in principle, be improved upon by a fuller physical explanation. Indistinguishability and indeterminism imply a looseness of fit on the part of physical explanation which take away its Procrustean character. The new world-view makes room for there being different sciences which are autonomous without invoking any mysterious causal powers beyond the reach of physical investigation.

The autonomy I am arguing for is, in the words of Bechner, 11 theory autonomy rather than a process autonomy: we use new concepts to ask new questions, rather than find that old questions have suddenly acquired surprising new answers. But this distinction between questions and answers offers a solution to the problem of reductionism only if there is some further fundamental difference between the concepts involved in framing the questions asked by different sciences. Otherwise, they might still be vulnerable to a take-over bid on the part of physics. A reductionist programme whereby every concept of chemistry and physics is exhaustively defined in terms of physical concepts alone might still be mounted. Thus far I have only cited examples---Gaussian curves, temperature, blackbird song---where reductive analysis seems out of the question. But the unavailablity of reductive analyses is much wider than that. Tony Dale bowled me out recently, when I had overlooked the fact that the concept of a finite number cannot be expressed in first-order logic. The very concept of a set, and more generally of a relational structure, is a holistic one. But rather than multiply examples, let me cite an in-principle argument. Tarski's theorem shows that the concept of truth cannot be defined within a logistic calculus: roughly, although we can teach a computer many tricks, we cannot program it to use the term `true' in just the way we do. It therefore seems reasonable to hold that other concepts, too, are irreducible, and the failure of the reductionist programme is due not to some mysterious forms of causality but to our endless capacity to form new concepts and in terms of them to ask new questions and seek new types of explanation.

The new world-view we are being forced to adopt not only permits us to concern ourselves, qua scientists, with general features, but impels us to do so. Even the corpuscularian philosophy gave somewhat short shrift to the things of ordinary experience. Most configurations of atoms were transitory. Even rocks were subject to the attrition of time, and the mountains, far from being eternal, were being eroded by the wind and the rain. Processes could in principle withstand the ravages of time, and at first glance Liouville's theorem seemed to suggest that point-particles whose initial conditions were close to one another would end up close still. But although, indeed, there was a one-one correlation between initial and final conditions, the correlation was much less stable than at first sight appeared. True, the volume in phase-space remains constant, but its shape does not, and may become spreadeagled with the elapse of time, so that the very smallest difference in initial conditions can lead to a wide difference in outcome. Poincaré pointed out the logic of the roulette wheel, 12 and we now regularly hear of the damage done by irresponsible butterflies on the other side of the universe destroying the reliability of met office forecasts. No longer can Newton number the ultimate things among the (kumaton anerithmon gelasma), the innumerable laughter of quantum waves, but if he wants atoms, must raise his sights to those stable solutions of the Schrödinger time-independent equation, which one way or another, will be realised. And although some solid objects are likely to remain substantially the same over time, most collocations of atoms are evanescent. If we seek stability amid the flux of molecular movement, we are likely to find it at a higher level of generality where chaos theory can indicate the recurrence of relatively stable patterns. In the Heraclitean swirl eddies may last long enough to be identified. Flames are processes, but possess the thingly property of subsisting and sometimes of being identified and individuated. So if we want permanence, we shall be led to focus on certain general features, certain types of boundary condition, which can persist over reasonable stretches of time. Just as chemists look to the time-independent Schrödinger equation to show them what stable atomic configurations there are, and would like to be able to work out in detail what molecules are stable too, so at a much higher level, biologists take note of organisms and species of organisms, which are the basic things of their discipline. Organisms are homeostatic, self-perpetuating and self-replicating. They are processes, like flames, but longer lasting and with greater adaptability in the face of adventitious change. They react to adverse changes in the environment so as to keep some variables the same, which together constitute the same organism that survives over time in the face of alterations in the environment. There is thus an essential difference between organism and environment which differentiates all the life sciences from the physical ones. Thinghood has become modal as well as diachronic. It is not enough to continue to be the same over time: organisms need to be able to change in some respects in order to remain the same in others, more important. Even if I were to alter the environment by watering the garden, moving the bird table, replacing the coconut with peanuts, the flora and fauna, though responding in various ways to the altered situation, would mostly persist as the self-same organisms as if I make no alterations. This invariance under a limited range of altered circumstance is more like the invariance of operation of natural laws than the continuance of atomic matter, but goes further; laws of nature would operate even if initial conditions were different, but do not characteristically alter their mode of operation so as to restore some antecedent condition, whereas biological organisms typically do, provided the alteration of initial conditions is not too drastic.

Homeostasis is a familiar concept in science---but logically a treacherous one. A homeostatic system tends to maintain the same state, and sameness can easily shift without our noticing it. The simple negative feedback of a flame or an eddy or a thunderstorm results in the process not being interrupted by every adventitious alteration of circumstance, but the persistence is short-lived none the less. Living organisms last longer, and are better able to withstand the attrition of time, because they react to counter the effect of a wider variety of circumstances. The requirement of persistence alters what we count as the substance that persists, and per contra as the concept of substance develops, so also does our idea of what counts as survival, and more generally what goals the substance seeks to secure and maintain. We begin to recognise as important explanatory schemata not only the survival of the organism, but the survival of the species, and now, even, the survival of the biosphere. And we begin to see not only the individual's maximising its own advantage as a rational goal, but the value of co-operative action, if we are to escape from the Prisoners' Dilemma and not be driven by individual selfishness into collective sub-optimality. Beyond that, I find it difficult to peer, but still hope dimly to discern the lineaments of what, if I may borrow a suggestive phrase from Nicholas Maxwell, 13 we might describe as an aim-oriented rationality.

The concept of homeostasis is borrowed from control engineering. It leads on naturally into information theory, and information theory provides the key concepts for understanding genetics. As self-perpetuation gives rise to self-replication, there is a greater need for the exact specification of the self, and the chromosome needs to be understood not only biochemically as a complicated molecule of DNA, but as a genetic code specifying what the new organism is to be like. Once again, the change of emphasis from the particular physical configuration to the general boundary condition, and the looseness of fit between the probabilistic explanations of the underlying physics and the quite different explanations of the emergent discipline allow us to accommodate the new insights without falling into obscurantist obfuscation. 14 Homeostasis also implies sensitivity. If an organism is to be independent of its environment, it must respond to it so as to counteract the changes which the changes in the environment would otherwise bring about within the organism itself: if I am to maintain a constant body temperature, I must sweat when it is hot outside and shiver when it is cold. Even plants must respond to light and to the earth's gravitational field. The greater the independence and the more marked the distinction between the self and the non-self, the greater the awareness the self needs to have of the non-self, and the more it needs to register, so as to be able to offset, untoward changes in the world around it. We are still in the dark as to what exactly consciousness is or how it evolved, but can see in outline why it is needed. A windowless monad cannot survive the changes and chances of this fleeting life---sensitivity to clouds on the horizon no bigger than a man's hand is the price of not being destroyed by unanticipated storms.

My interest lies in the end of this line of development. We can give a general characterization of what it is for a system to be able to represent within itself some other system, and so can think of organisms in terms not of biochemistry or evolutionary biology but of information theory and formal logic. And from this point of view we can consider not only consciousness but self-consciousness, and a system that can represent within itself not just some other system but itself as well. There are a whole series of self-reflexive arguments. Popper, a former President of our Society, has devoted much energy to arguing from them to an open universe; in particular, he argues from the impossibility of self-prediction. MacKay argues similarly---other people may predict what I am going to do, but I cannot. 15 Many people, Haldane, Joseph, Malcolm, Mascall, Popper, Price, Wick and others, have been concerned about rationality, and have argued that if determinism or materialism were true, we could not be rationally convinced of it. 16 Reductive metaphysics, which reduces rationality to something else---the movement of physical particles, for example---cannot leave room for the rational arguments which alone could establish its truth. I myself found these arguments intriguing, and indeed, compelling, but extraordinarily difficult to formulate in a cast-iron way. Eventually I came up with an argument based on Gödel's theorem, which is indeed a version of these arguments, and is intended to show in one swoop the failure of any reductionist programme as regards reason. I have received much stick for using Gödel's theorem to show that the mind is not a Turing machine, but I am quite impenitent on that score, and believe that the argument goes much further, and shows not only the impossibility of reducing reason to the mere following of rules, but the essential creativity of reason. We can never formalise reason completely or tie it down to any set of canonical forms, for we can always step outside and view all that has been thus far settled from a fresh standpoint. In particular we can find fresh features that seem significant, and seek fresh sorts of explanation of them. It does, I believe, establish the essential openness of the universe, granted only that there is at least one rational agent. If there be rational agents, since we are rational agents, it follows that the course of events in the universe cannot be reduced to a system of things evolving according to a determinate algorithm, but that there are always new opportunities and further possible exercises of rationality.

The interplay between things and explanations is illuminating. Instead of starting with things, we are able to identify things only at higher levels of organization, and the higher we go the more thingly properties we find. Atoms have stability (usually), but are qualitatively identical with many others. Organisms have more individuality, and are less commonly clones, but still view their environment if not in terms of chemical similarity nevertheless in terms of fungibility, readily replacing one food supply by another. Nor is it only the environment that organisms regard fungibly: although some birds are faithfully monogamous, many are not, and if one Mrs Blackbird fails to respond to the musical blandishments of her would-be mate, another will serve his reproductive purposes just as well. Human love likewise is not uniformly faithful to the individual ideal, but with human beings we can see this as a derogation from humanity, and can construct a coherent concept of unique individuality, according to which this person is irreducibly himself, and essentially different from anybody else. 17 Our idea of thinghood leads us from the utterly simple and essentially similar atoms of the corpuscularians to infinitely complex and unique persons, each necessarily different from every other.

The different ideals of thinghood support different paradigms of explanation. Since different sorts of feature characterize things at different levels, and the features that characterize at the higher levels cannot be completely defined in terms of those that play a part in lower-level explanations, the higher-level explanations cannot be reduced to lower-level ones. As we have seen, a Gaussian curve cannot be defined in terms of a Laplacian explanation, for it essentially involves the notion of an ensemble or Kollectiv. Higher-level systems are not derivable from some fundamental system, but are, instead, autonomous. We cannot predict the exact position or velocity of a sub-atomic entity, but by means of the time-independent Schrödinger equation we can say what properties a hydrogen atom would have if it existed, and we can have good reason for supposing that many such atoms will exist, since they are stable configurations of quantum-mechanical systems. The explanations sought by a chemist are in terms of energy levels and the valency bonds they generate: those sought by the biologist are in terms of the maintenance of life and the continuation of the species. And as these explanations differ, so also do the things they are explanations about. Explanations influence what is to count as a thing, and ideas of what it truly is to be a thing influence what questions we ask, and what explanations we seek to discover. 18 We can see this, if we like, as a form of emergent evolutionary development: new levels of being evolve from lower, chemical elements from the flux after the Big Bang, molecules, organisms, consciousness, and self-consciousness, in the fullness of time; but we can also see it in terms of a hierarchy of Platonic forms and explanations, each going beyond the limits of its predecessors, and at the higher levels reaching out to ever new kinds of creative rationality.

To summarise, then. The new scientific world-view differs from traditional corpuscularianism in not postulating some ultimate thing-like entities whose motions determine completely the state of the world not only at that time but at all subsequent ones too. Instead of there being particular point-particles, there are only general features, and instead of a rigid determinist law, there are only probabilities, which are, indeed, enough to enable us to make reliable predictions about many aspects of the world, but do not foreclose the possibility of other types of explanation being the best available. Other types of explanation are answers to other types of questions, and it is because we ask different questions that the different sciences are different. These different questions pick on different general features, often different types of boundary condition; and once we acknowledge that there is no metaphysical reason to reduce the generic characterization of boundary conditions typical of other sciences to the paradigm physical terms of Laplacian corpuscularianism, we can accept these other sciences as sciences in their own right, since, metaphysics apart, we have good reason to resist reductionism as applied to questions rather than answers.

The abolition of ultimate things thus opens the way to our acknowledging the autonomy of the various sciences. At the same time, the notion of a thing leads us to pick out various types of boundary condition as instantiating, to a greater and greater degree, certain characteristic features of being a thing---permanence, stability, ability to survive adventitious alterations in the environment, and the like. As we follow these through, we find a natural hierarchy of the sciences in which we ask questions about more and more complicated entities, possessing more and more thing-like perfections. Things have gone up market. By an almost Hegelian dialectic our notion of a thing becomes transmuted into that of a substance, and in so far as we remain pluralists at all, we move from the minimal qualitatively identical, though numerically distinct, atoms of the corpuscularians to the infinitely complex, though windowed, monads of a latter-day Leibniz. Whether Lucretius would have been pleased at this outcome of the complex interplay between ontological intimations of existence and rationalist requirements of explicability, I do not know. But he could hardly complain at my taking this as my theme, here at an address to the British Society for the Philosophy of Science taking place in the London School of Economics, whose motto is taken from Virgil's description of him, and also expresses the common sentiment of all our members,

Felix qui potuit rerum cognoscere causas

Happy he who understands the explanations of things

To return from footnote to text, click on footnote number

1. Michael Redhead, ``A Philosopher Looks at Quantum Field Theory'', in Harvey Brown and Rom Harré, eds., Philosophical Foundations of Quantum Mechanics, Oxford, 1988, p.10.
2. Robin Le Poidevin, Change, Cause and Contradiction, London, 1991, esp. ch.8.
3. C.A.Howson and P.Urbach, Scientific Reasoning: the Bayesian Approach, La Salle, Illinois, USA, 1989, p.19.
4. D.A.Gillies, Objective Probability, Cambridge, 1973, esp. ch.5.
5. Reprinted in S.French and H.Kamminga, eds., Correspondence, Invariance and Heuristics, (Kluwer Academic Publishers, Holland), 1993, p. 329.
6. Nancy Cartwright, How the Laws of Physics Lie, Oxford, 1983, ch.2, esp. pp.44-46. Peter Lipton, Inference to the Best Explanation, London, 1991, esp. ch.3; John Worrall, ``The Value of a Fixed Methodology'', British Journal for the Philosophy of Science, 39, 1988.
7. David Papineau, British Journal for the Philosophy of Science, 47, 1991, p.399.
8. I owe this point to H.C.Longuet-Higgins, The Nature of Mind, Edinburgh, 1972, ch.2, pp.16-21, esp. p.19; reprinted in H.C.Longuet-Higgins, Mental Processes, Cambridge, Mass., 1987, ch.2, pp.13-18, esp.p.16. I am also particularly indebted to C.F.A.Pantin, The Relations between the Sciences, Cambridge, 1968; and to A.R.Peacocke, God and the New Biology, London, 1986, and Theology for a Scientific Age, Oxford, 1990. Michael Polanyi emphasized the importance of boundary conditions and their relevance to the different sorts of explanation sought by different disciplines. In his ``Tacit Knowing'', Reviews of Modern Physics, October, 1962, pp.257-259, he cites the example of a steam engine, which although entirely subject to the laws of chemistry and physics, cannot be explained in terms of those disciplines alone, but must be explained in terms of the function it is capable, in view of its construction, perform. What is interesting about the steam engine is not the laws of chemistry and physics, but the boundary conditions, which in view of those laws, make it capable of transforming heat into mechanical energy; it is the province of engineering science, not physics. The example of the steam engine is illuminating in that no question of vitalism arises. See also, Michael Polanyi, ``Life Transcending Physics and Chemistry'', Chemical and Engineering News, August 21, 1067, pp.54-66; and ``Life's Irreducible Structure'', Science, 160, 1968, pp.1308-1312.
9. F.H.C.Crick, Of Molecules and Man, University of Washington Press, Seattle and London, 1966, p.10.
10. That the biologist is primarily concerned with boundary conditions of a special type is pointed out by Bernd-Olaf Küppers, Information and the Origin of Life, M.I.T. Press, Cambridge, Mass, U.S.A., 1990, p.163.
11. Compare the distinction drawn by M.Bechner between theory autonomy and process autonomy in his ``Reduction, Hierarchies and Organicism'' in F.J.Ayala and T.Dozbhanski, Studies in the Philosophy of Biology: Reduction and Related Problems, London, 1974, p.170; cited by A.R.Peacocke, God and the New Biology, London, 1986, p.9.
12. Henri Poincaré, Science and Method, tr. F.Maitland, London, 1914, p.68.
13. Nicholas Maxwell, From Knowledge to Wisdom, Oxford, 1984, esp. ch.4.
14. Nicholas Maxwell, From Knowledge to Wisdom, Oxford, 1984, esp. ch.4.
15. D.M.MacKay, ``On the Logical Indeterminacy of a Free Choice'', Mind, LXIX, 1960, pp.31-40.
16. See K.R.Popper, The Open Universe, ed. W.W.Bartley, III, London, 1982, ch.III, $$ 23,24. Popper traces the argument back to Descartes and St Augustine. A further list is given in J.R.Lucas, The Freedom of the Will, Oxford, 1970, p.174. Further arguments and fuller references may be found in Behavioral Sciences, 1990, 13, 4.
17. I argue this in my ``A Mind of One's Own'', Philosophy, October, 1993.
18. Compare A.R. Peacocke, Theology for a Scientific Age, Oxford, 1990, p.41: Because of widely pervasive reductionist presuppositions, there has been a tendency to regard the level of atoms and molecules as alone `real'. However, there are good grounds for not affirming any special priority to the physical and chemical levels of description and for believing that what is real is what the various levels of description actually refer to. There is no sense in which subatomic particles are to be graded as `more real' than, say, a bacterial cell a human person or a social fact. Each level has to be regarded as a slice thorough the totality of reality, in the sense that we have to take account of its mode of operation at that level.
Return to bibliography

Return to home page