Janus

by Arthur Koestler

  University of Geneva), Seymour S. Kety (Professor of Psychiatry,

  Harvard University), Arthur Koestler (Writer, London), Paul

  D. MacLean (Head of the Laboratory of Brain Evolution and

  Behaviour, NIMH, Bethesda, Maryland), David McNeill (Professor

  of Psychology, University of Chicago), Jean Piaget (Professor of

  Experimental Psychology, University of Geneva) J. R. Smythies

  (Reader in Psychiatry, University of Edinburgh), W. H. Thorpe

  (Director, Sub-Department of Animal Behaviour, Department of

  Zoology, University of Cambridge), C. H. Waddington (Profrssor and

  Chairman, Department of Genetics, University of Edinburgh), Paul

  A. Weiss (Emeritus Member and Professor, Rockefeller University,

  New York).

  Now it is quite possible that some sexual (or even scatological) motivation may enter into an artist's work; yet it is absurd to proclaim that art is 'nothing but' goal-inhibited sexuality, because it begs the question of what makes Goethe's art a work of genius, quite unlike other premature ejaculators'. The reductionist attempt to explain artistic creation by the action of sex-hormones is futile, because that action, though biologically vital, does not give us an inkling of the aesthetic criteria which apply to a work of art. Those criteria pertain to the level of conscious mental processes, which cannot be reduced to the level of biological processes without losing their specifically mental attributes in the course of the operation. Reductionist psychiatry is a Procrustean host to the weary traveller.

  It is easy to make fun of those latter-day orthodox Freudians who have reduced the master's teaching to a caricature. In other fields, however, the reductionist fallacy is more discreetly implied, less obvious and therefore more insidious. Pavlov's dogs, Skinner's rats, Lorenz's geese, each served for a while as fashionable paradigms of the human condition. Desmond Morris's bestselling book The Naked Ape opened with the statement that man is a hairless ape 'self-named homo sapiens . . . I am a zoologist and the naked ape is an animal. He is, therefore, fair game for my pen.' To what extremes this zoomorphic approach may lead is illustrated by a further quotation from Morris:

  The insides of houses or flats can be decorated and filled with

  ornaments, bric-à-brac and personal belongings in

  profusion. This is usually explained as being done to make the

  place 'look nice'. In fact, it is the exact equivalent to another

  territorial species depositing its personal scent on a landmark near

  its den. When you put a name on a door, or hang a painting on a wall,

  you are, in dog or wolf terms, for example, simply cocking your leg

  on them and leaving your personal mark there. [2]

  On a more serious level (though the passage quoted was obviously meant to be taken in all seriousness) we are faced with two impressive strongholds of reductionist orthodoxy. One is the neo-Darwinian (or 'Synthetic') theory which holds that evolution is the outcome of 'nothing but' chance mutations retained by natural selection -- a doctrine recently exposed to growing criticism* which nevertheless is still taught as gospel truth. The other is the behaviourist psychology of the Watson-Skinner school which holds that all human behaviour can be 'explained, predicted and controlled' by methods exemplified in the conditioning of rats and pigeons. 'Values and meanings are nothing but defence mechanisms and reaction formations' is another of Frankl's telling quotes from a behaviourist textbook.

  * See below, Part Three.

  By its persistent denial of a place for values, meaning and purpose in the interplay of blind forces, the reductionist attitude has cast its shadow beyond the confines of science, affecting our whole cultural and even political climate. Its philosophy may be epitomized by a last quote from a recent college textbook, in which man is defined as 'nothing but a complex biochemical mechanism, powered by a combustion system which energises computers with prodigious storage facilities for retaining encoded information'. [3]

  Now the reductionist fallacy lies not in comparing man to a 'mechanism powered by a combustion system' but in declaring that he is 'nothing but' such a mechanism and that his activities consist of 'nothing but' a chain of conditioned responses which are also found in rats. For it is of course perfectly legitimate, and in fact indispensable, for the scientist to try to analyse complex phenomena into their constituent elements -- provided he remains conscious of the fact that in the course of the analyses something essential is always lost, because the whole is more than the sum of its parts, and its attributes as a whole are more complex than the attributes of its parts. Thus the analysis of complex phenomena elucidates only a certain segment or aspect of the picture and does not entitle us to say that it is 'nothing but' this or that. Yet such 'nothing-but-ism' as it has been called, is still the -- explicit or implied -- world-view of reductionist orthodoxy. If it were to be taken literally, man could be ultimately defined as consisting of nothing but 90 per cent water and 10 per cent minerals -- a statement which is no doubt true, but not very helpful.

  2

  Nevertheless, reductionism proved an eminently successful method within its limited range of applicability in the exact sciences, while its antithesis, holism, never really got off the ground. Holism may be defined by the statement that the whole is more than the sum of its parts. The term was coined by Jan Smuts in the 1920s in a remarkable book [4] which for a while enjoyed great popularity. But holism never got a grip on academic science* -- partly because it went against the Zeitgeist, partly perhaps because it represented more of a philosophical than an empirical approach and did not lend itself to laboratory tests.

  * Except indirectly through Gestalt psychology.

  In fact both reductionism and holism, if taken as sole guides, lead into a cul-de-sac. 'A rose is a rose is a rose' may be regarded as a holistic statement, but it tells us no more about the rose than the formulae of its chemical constituents. For our inquiry we need a third approach, beyond reductionism and holism, which incorporates the valid aspects of both. It must start with the seemingly abstract yet fundamental problem of the relations between the whole and its parts -- any 'whole', whether the universe or human society, and any 'part', whether an atom or a human being. This may seem an odd, not to say perverse, way to get at a diagnosis of man's condition, but the reader will eventually realize, I hope, that the apparent detour though the theoretical considerations in the present chapter may be the shortest way out of the labyrinth.

  3

  To start with a deceptively simple question: what exactly do we mean by the familiar words 'part' and 'whole'? 'Part' conveys the meaning of something fragmentary and incomplete, which by itself has no claim to autonomous existence. On the other hand, a 'whole' is considered as something complete in itself which needs no further explanation. However, contrary to these deeply ingrained habits of thought and their reflection in some philosophical schools, 'parts' and 'wholes' in an absolute sense do not exist anywhere, either in the domain of living organisms, or in social organizations, or in the universe at large.

  A living organism is not an aggregation of elementary parts, and its activities cannot be reduced to elementary 'atoms of behaviour' forming a chain of conditioned responses. In its bodily aspects, the organism is a whole consisting of 'sub-wholes', such as the circulatory system, digestive system, etc., which in turn branch into sub-wholes of a lower order, such as organs and tissues -- and so down to individual cells, and to the organelles inside the cells. In other words, the structure and behaviour of an organism cannot be explained by, or 'reduced to', elementary physico-chemical processes; it is a multi-levelled, stratified hierarchy of sub-wholes, which can be conveniently diagrammed as a pyramid or an inverted tree, where the sub-wholes form the nodes, and the branching lines symbolize channels of communication and control: see diagram .

  The point first to be emphasized is that each member of this hierarchy, on whatever level, is a sub-who
le or 'holon' in its own right -- a stable, integrated structure, equipped with self-regulatory devices and enjoying a considerable degree of autonomy or self-government. Cells, muscles, nerves, organs, all have their intrinsic rhythms and patterns of activity, often manifested spontaneously without external stimulation; they are subordinated as parts to the higher centres in the hierarchy, but at the same time function as quasi-autonomous wholes. They are Janus-faced. The face turned upward, toward the higher levels, is that of a dependent part; the face turned downward, towards its own constituents, is that of a whole of remarkable self-sufficiency.

  The heart, for instance, has its own pacemakers -- actually several pacemakers, capable of taking over from each other when the need arises. Other major organs are equipped with different types of coordinating devices and feedback controls. Their autonomy is convincingly demonstrated by transplant surgery. At the beginning of our century, Alexis Carrell showed that a minute strip of tissue taken from the heart of a chicken embryo and put into a nutrient solution will go on pulsating for years. Since then, whole organs were shown to be capable of functioning as quasi-independent wholes when taken out of the body and kept in vitro, or transplanted into another body. And as we descend the steps of the hierarchy to the lowest level observable through the electron microscope, we come upon sub-cellular structures -- organelles -- which are neither 'simple' nor 'elementary', but systems of staggering complexity. Each of these minuscule parts of a cell functions as a self-governing whole in its own right, each apparently obeying a built-in code of rules. One type, or tribe, of organelles looks after the cell's growth, others after its energy supply, reproduction, communication, and so on. The mitochondria, for instance, are power plants which extract energy from nutrients by a chain of chemical reactions involving some fifty different steps; and a single cell may have up to five thousand such power plants. The activities of a mitochondrion can be switched on or off by controls on higher levels; but once triggered into action it will follow its own code of rules. It cooperates with other organdies in keeping the cell happy, but at the same time each mitochondrion is a law unto itself, an autonomous unit which will assert its individuality even if the cell around it is dying.

  4

  Science is only just beginning to rid itself of the mechanistic preconceptions of the nineteenth century -- the world as a billiard table of colliding atoms -- and to realize that hierarchic organization is a fundamental principle of living nature; that it is 'the essential and distinguishing characteristic of life' (Pattee) [5]; and that it is 'a real phenomenon, presented to us by the biological object, and not the fiction of a speculative mind' (P. Weiss). [6] It is at the same time a conceptual tool which on some occasions acts as an Open Sesame. All complex structures and processes of a relatively stable character display hierarchic organization, regardless whether we consider galactic systems, living organisms and their activities, or social organizations. The tree diagram with its series of levels can be used to represent the evolutionary branching of species into the 'tree of life'; or the stepwise differentiation of tissues and integration of functions in the development of the embryo. Anatomists use the tree diagram to demonstrate the locomotor hierarchy of limbs, joints, individual muscles, and so down to fibres, fibrils and filaments of contractile proteins. Ethologists use it to illustrate the various sub-routines and action-patterns involved in such complex instinctive activities as a bird building a nest; but it is also an indispensable tool to the new school of psycholinguistics started by Chomsky. It is equally indispensable for an understanding of the processes by which the chaotic stimuli impinging on our sense organs are filtered and classified in their ascent though the nervous system into consciousness. Lastly, the branching tree illustrates the hierarchic ordering of knowledge in the subject-index of library catalogues -- and the personal memory stores inside our skulls.

  The universal applicability of the hierarchic model may arouse the suspicion that it is logically empty. I hope to show that this is not the case, and that the search for the fundamental properties, or laws, which all these varied hierarchies have in common amounts to more than a play on superficial analogies -- or to riding a hobby horse. It should rather be called an exercise in General Systems Theory -- that relatively recent inter-disciplinary school, founded by von Bertalanffy, whose purpose is to construct theoretical models and discover general principles which are universally applicable to biological, social and symbolic systems of any kind -- in other words, a search for common denominators in the flux of phenomena, for unity-in-diversity.

  As early as 1936, Joseph Needham wrote:

  The hierarchy of relations, from the molecular structure of carbon

  compounds to the equilibrium of species and ecological wholes,

  will perhaps be the leading idea of the future. [7]

  Even earlier Lloyd Morgan, C. D. Broad, and J. Woodger among others emphasized the importance of recognizing hierarchically ordered 'levels of organization', and the emergence on each higher level of new 'organizing relations' between (sub) wholes of greater complexity, whose properties cannot be reduced to, nor predicted from, the lower level. To quote Needham again:

  Once we adopt the general picture of the universe as a series of

  levels of organisation and complexity, each level having unique

  properties of structure and behaviour, which, though depending on the

  properties of the constituent elements, appear only when these are

  combined into the higher whole, we see that there are qualitatively

  different laws holding good at each levels. [8]

  But such a multi-levelled view went against the materialist Zeitgeist, because it implied that the biological laws which govern life are qualitatively different from the laws of physics which govern inanimate matter, and that accordingly life cannot be 'reduced' to the blind dance of atoms; and similarly, that the mentality of man is qualitatively different from the conditioned responses of Pavlov's dogs or Skinner's rats, which the dominant school in psychology considered as the paradigms of human behaviour. Harmless as the word 'hierarchy' sounded, it turned out to be subversive. It did not even appear in the index of most modern textbooks of psychology or biology.

  Yet there have always been voices in the wilderness, insisting that the concept of hierarchic organization was an indispensable prerequisite -- a conditio sine qua non -- of any methodical attempt to bring unity into the diversity of science, and might eventually lead to a coherent philosophy of nature -- which at present is conspicuous by its absence.

  To this minority chorus there was also added the small voice of the author, expressed in several books in which 'the ubiquitous hierarchy' [9] played a major, and often dominant part. Taken together, the relevant passages would add up to a fairly comprehensive textbook on hierarchic order (which may see the light some day). But this is not the purpose of the present volume. As already said, the hierarchic approach is a conceptual tool -- not an end in itself, but a key capable of opening some of nature's combination-locks which stubbornly resist other methods.*

  * Cf. also Jevons: 'The organisation hierarchy, forming as it does a

  bridge between parts and whole, is one of the really vital, central

  concepts of biology.' [11]

  However, before attempting to use the key, it is necessary to gain some insight into the way it works. The present chapter is meant to convey some of the basic principles of hierarchic thought in order to provide a platform or runway for the more speculative flights that follow.

  5

  To say it once more: if we look at any form of stable social organization, from the insect state to the Pentagon, we shall find that it is hierarchically structured; the same applies to the individual organism, and, less obviously, to its innate and acquired skills. However, to prove the validity and significance of the model, it must be shown that there exist specific principles and laws which apply (a) to all levels of a given hierarchy, and (b) to hierarchies in different f
ields -- in other words, which define the term 'hierarchic order'. Some of these principles might appear self-evident, others rather abstract; taken together, they form the stepping stones for a new approach to some old problems.

  'A good terminology', someone has said, 'is half the game.' To get away from the traditional misuse of the words 'whole' and 'part', one is compelled to operate with such awkward terms as 'sub-whole', or 'part-whole', 'sub-structures', 'sub-skills', 'sub-assemblies', and so forth. To avoid these jarring expressions, I proposed, some years ago [10], a new term to designate those Janus-faced entities on the intermediate levels of any hierarchy, which can be described either as wholes or as parts, depending on the way you look at them from 'below' or from 'above'. The term I proposed was the 'holon', from the Greek holos = whole, with the suffix on, which, as in proton or neutron, suggests a particle or part.

  The holon seems to have filled a genuine need, for it is gradually finding its way into the terminology of various branches of science, from biology to communication theory. It was particularly gratifying to discover that it has also insinuated itself into French: in Professor Raymond Ruyer's much discussed book La Gnose de Princeton [12] there is a chapter entitled: 'Les accolades domaniales et les holons' -- with a footnote which says: 'If I am not mistaken, the word originated with Koestler.' New words are like parvenus: once their origin is forgotten, they have made it.