The Unity of Science without Reductionism
A Lecture given in Dubrovnik
on April 11, 1995
reprinted in Acta Analytica, 15, 1996; pp.89-95
some typos corrected July 26th, 2006
The Unity of Science is often thought to be reductionist, but this is because we fail to distinguish questions from answers. The questions asked by different sciences are different---the biologist is interested in different topics from the physicist, and seeks different explanations---but the answers are not peculiar to each particular science, and can range over the whole of scientific knowledge. The biologist is interested in organisms--- concept unknown to physics---but explains physiological processes in terms of chemistry, not a mysterious vital force. The replacement of Laplacian determinism by quantum mechanics further erodes the tendency towards reductionism. The answers given in different explanations are not subsumed under one complete theory; and quantum mechanics does not have a concept of haecceitas, "this-i-ness" which would make its entities the fundamental constituents of everything.
The Unity of Science has often been seen as a reductionist slogan. Science is to be unified, it is thought, by reducing it all to physics. Many physicists are already half-persuaded that they have the key to all knowledge, at least to all scientific knowledge. "It is all in the Schrödinger Equation," they feel, and reckon that in the fullness of time all other sciences will be worked out from first physical principles. Some non-physicists agree: ``the ultimate aim of the modern movement in biology is in fact to explain all biology in terms of physics and chemistry''. 1 But many biologists are not happy with the take-over bid of their science by the physicists: in time past they have appealed to mysterious vital forces or  entelechies to defend their subject as being in principle not subsumable under physical laws or principles, and have been correspondingly despised for mystery-mongering. The 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.
But although their arguments were bad, and their objections to physicalism obscure, they had a point. There is something wrong with reductionism. I want to do justice to their objections without invoking obscure non-physical physical causes or processes, and to preserve the unity of science while giving proper rein to the autonomy of different sciences. I shall do this partly by distinguishing questions from answers, and partly by reviewing what is now the fundamental physical theory, quantum mechanics, and pointing out that it is not only not a determinist theory but one which denies "haecceitas", this-i-ness to its fundamental particles.
Different sciences 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. 2 Whereas vitalism made out that answers were in principle unavailable, 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. 3 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 with 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. It is like the change of attention being urged on us yesterday by Jean-Pierre Marquis, when he commended category theory, with its emphasis on similarity of structure in mathematical entities, rather than insisting on an exact ontological account of what their constituent components were. Different disciplines, because they concentrate on different questions, abstract from irrelevant detail, in order to adduce the information that is relevant to their concerns.
Thus far, thus good. But hitherto, it has seemed that the practitioners of other sciences were simply throwing away information they did not need, information offered by the fundamental physical theory, and always available in principle. On the traditional Laplacian view, there was a fundamental theory about fundamental entities which offered a complete account of everything. We might choose to make use of only part of the information available, much as classical thermodynamics makes use of only part of the information about the motions of each molecule of a gas which is in principle available; but the information was, in principle, there, and gave us the last word about the fundamental constituents of reality. But Laplace was wrong. The fundamental theory, quantum mechanics, is, as we all know, indeterminist, and, what is not so widely recognised, quantum-mechanical entities are not really entities at all, and cannot claim to be the fundamental substances, the ultimate constituents of reality. They cannot claim to be the fundamental substances, because they lack what Duns Scotus called "haecceitas", this-i-ness. If electrons and photons were genuine substances in the way the corpuscles of the atomists were supposed to be, they would obey Maxwell-Boltzmannn statistics: that is, we should count as separate cases this quantum-mechanical entity being in one box, that in another, and that quantum-mechanical entity being in the one box and this in the other. But, in fact, they do not obey Maxwell-Boltzmann statistics, but rather Bose-Einstein or Fermi-Dirac statistics. Bose-Einstein statistics count the two cases described above as the same; quantum-mechanical entities that obey Bose-Einstein statistics are called generically "bosons", and the statistics can be understood if we regard bosons not as genuine substances to be referred to by substantives, but as adjectives describing the state of the universe in a particular region. If I have a still-life picture of a flower with yellow petals, and describe it as having yellow here and yellow there, I have not described a different picture if I describe it as having yellow there and yellow here. It is a distinction without a difference, and should be seen as one single case, not two different ones. This is how we should see the quantum-mechanical world. It is not made up of fundamental things, each with its own individuality, but is a many-coloured whole, coloured differently in different parts, these colours being fundamental qualities perhaps, but not fundamental things. Whereas Newton thought of God as having created the void and in it atoms, like sands on the sea shore for multitude, with God able none the less to give each atom its own name and know it for the atom it is, we now think of the world as a kumaton anerithmon gelasma, an innumerable laughter of waves, with no individuality at all, but only shimmering with merely evanescent qualities. There is therefore no ontological reason to think that somehow the physicists are giving us an account of the fundamental substances which is the truth about what is really real.
Nor is the account that they can give complete. Quantum mechanics gives us only probabilistic predictions, not certain ones. There is no complete account in terms of quantum mechanics of how a physical system will evolve. We therefore cannot extract from the quantum-mechanical account some cruder account suitable for some subordinate discipline by throwing away the information we do not need so as to concentrate on the details that are relevant. Instead we have to rely on the other discipline to provide us with the concepts which, one way or another, will be instantiated. We do not know how the time-dependent Schrödinger equation will collapse into an eigen-state, but granted certain boundary conditions we can use the time-independent Schrödinger equation to tell us what stable atomic, or molecular, configurations to expect. Much as chaos theory cannot tell us precisely what will happen but can indicate what patterns are likely to recur or to persist, so fundamental physics cannot tell us precisely what will happen, but granted certain further boundary conditions and constraints on stability, can give us confidence that one way or another certain patterns will be found.
The epistemological and the ontological argument are connected. What prevents the various different explanations offered by different sciences being trumped by an in-principle omnicompetent Laplacian explanation capable, in principle, of explaining everything, is the indeterminism of quantum mechanics, itself based on the probabilistic account of the collapse of the wave-packet. But probabilities are to be most properly ascribed to generic propositional functions rather than particular single propositions; in Popper's terminology, they are propensities, a sort of quality, not an individual thing. This is what was right in the old frequency theory. We cannot in quantum mechanics be always dealing only with ensembles, because sometimes we have to deal with single cases: but single cases are to be considered as typical instances of a type, a type that could be exemplified by a whole ensemble. Determinism could deal with determinate individuals, but probabilistic indeterminism deals naturally with un-individualised instantiations of generic types. In moving away from Laplacian atomism, we abandon both the epistemological completeness of determinism and the ultimate substances of atomism.
But substances are not easily dispensed with. And I want to conclude by showing how we seek to re-introduce substance into our scientific thinking, and to note the conceptual pressure the various marks of substantiality exercise upon our thought. We need permanence, we need stability, if we are to be able to communicate, and allow that while things are not what they used to be, they have changed without thereby ceasing to be the things they were before they changed. And we can, as I have indicated, pick out stable configurations in the swirl of events, chemical atoms and molecules, rocks and sticks and stones, and regard them as substances. But though stable, they are liable to be altered by adventitious circumstances, and if altered, to remain altered. Molecules may be involved in chemical reactions, sticks may be burnt, rocks split. Greater permanence can be accorded to those substances that are homoeostatic, and in the face of adventitious alteration will maintain their original condition. Biological organisms maintain themselves---provided the alteration of their environment is not too drastic---and therefore have a better claim to be substances than mere material objects. But with homoeostasis come new possibilities and new questions. Homoeostasis means maintaining the same state, and there are many different ways of being the same. I maintain the same temperature in my body, roughly the same mass, the same habits, the same goals, the same ambitions. The organism as an agent pursues a number of goals, in particular survival and self-replication, which can together be  summed up as its self-interest. But just as the concept of the self begins to emerge, we are led in cases where many organisms interact---especially in situations like the Prisoners' Dilemma---to make a rational move from individual self-interest to the collective interest of a wider self. Biologists need to think not just of organisms but of populations and groups, and the intellectual possibilities of altruism revealed by the theory of games are realised in the course of biological evolution.
Homoeostasis raises questions not only of goals but of sensitivity. I maintain a constant body temperature by shivering when it is cold and sweating when it is hot. I maintain my independence of the environment in one respect by responding to it in some other way. 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. I used one yesterday when Jim was putting forward, not on his own behalf, but in order that it could be refuted, a version of naturalist epistemology which explained away truth and rationality, and explained why he held the views he did, but could explain why anyone else should. This is a common metaphysical argument---we reject those reductive metaphysical systems which saw off the branch on which they are sitting so to speak. Once we admit self-conscious, self-reflexive substances, which we are led to do because they have more marks of substantiality that any others, we are led to tangle with the most difficult issues of logic and metaphysics.
NOTESTo return from footnote to text, click on footnote number 1. F.H.C.Crick, Of Molecules and Men, University of Washington Press, Seattle and London, 1966, p.10.
Return to home page