The conceptual foundations of decision-making in a democracy

[p. 249]

J&A to Part Two:
Life and Information

Nurture Versus Nature (2a.4)

Confronting the draft of this book with the writings of contemporary biologists and sociologists yielded a surprising dispute between scientists about the general priority of nature (genes) versus nurture (environment) in the explanation of features and actions of men. Possibly the subject will have been settled as a non-issue by the time you read this. If not, a good summary of the discussions and problems involved is given by Susan Oyama's ‘The ontogeny of information’. The discord often finds its origin in an ‘imperialist’ inclination, a tendency towards hubris of many great scientist, leading them to extend the explanatory scope of their theories beyond their possibilities. The pressures to which professionals are subjected by their peers and the market are not conductive to a careful and unprejudiced evaluation and search for consensus. The outsider can approach the subject with an open mind and without those pressures.

Obviously these clarifications and additions of a layman cannot concern the theories themselves. The purpose of the chapters of this J&A is to avoid being drawn into controversies, such as ‘nature versus nurture’. Already Monod's thirty-year-old essay, addressed to the general public, shows why the systems approach advocated by Oyama (and me) is mandatory for an adequate apprehension of life. Once we have decided that life is a complex, interdependent, dynamic and usually stochastic process of elements which can learn, the discussion about nurture versus nature becomes a non-issue, a conclusion which any economist versed in systems theory would have drawn immediately.

Any generally valid explanation of a living being must always consider both aspects, and usually we cannot assign a general priority to either of them. Of course in a specific case and for a specific purpose, one may be more important than the other. No one can object to the statement that sickle-cell anaemia is caused by a genetic defect. It is equally justifiable to attribute the appearance of a cold in a person having just returned from a fall into ice-cold water to that mishap or to a virus without any reference to genes. What would not be justified is the generalisation that catching a cold is always attributable only to circumstances: a difference in genes may very well explain why certain persons catch a cold in circumstances where others do not; but that propensity could also follow from ‘nurtural’ circumstances such as malnutrition, exhaustion or overprotection during youth.

The controversy about nurture versus nature could, and should, have been settled by now, and a generally accepted overview of the present state of the art in sciences dealing with life, written in a language accessible to the layman should have been available as a source to which I could have referred in writing this book. The fact that they are not illustrates the necessity of the institutional improvements which were the conclusion of the first volume.

[p. 250]

The reason why the genes are more prevalent in my book, and particularly in its examples, than other aspects of life, is that they are particularly well suited to illustrate a theory involving information, and to explain why any empirical or normative epistemology must be based on some implicit or explicit concept of the process of life and of the role of information in that process. For the genes are the most ubiquitous and probably very first means of recording information. It does not imply the dominance of genetic information over other aspects of the process of life. In fact, I think that any attempt to explain everything by one aspect, one theory, is doomed to failure and bound to end in a sterile dogmatism, as evidenced by the more extreme forms of sociobiology which simply ignore the basic feature of man that, however important inherited and genetically determined tendencies of behaviour may be, once we are conscious of them we can attempt to counteract them.

Back to the controversy between nurture and nature. There exists no ‘natural’ non-tautological criterion by which we can a priori assign any event or object to one of these two categories. ‘Nature’, meaning the innate predispositions through which a being reacts to his environment, cannot be imputed solely to the genes even at its moment of conception. For the parent bestows on its offspring not only the genes, but also some environment without which these genes cannot develop into a being, would not be alive, like cytoplasm and mitochondria of the egg cell. All these ‘inherited’ features will influence the way in which it develops. Trying to classify any of these inherited features as either nurture or nature on basis of some generally valid criterion will convince the reader of the limited scope and of the artificiality of any such classification. It is very useful, even indispensable, as a heuristic device. But used in any other way, it will be in contradiction with the nature of life as a process, and the living being as an interdependent, complex, dynamic and stochastic system.

Yet often specific cases in personal and social hygiëne and policy require a determination of the character and relative importance of the two factors. Left to its own devices, the discussion will be drawn into ideological and political struggles and degenerate into a dialogue of the deaf, a situation which only can be remedied by a dispassionate investigation of a disinterested and unprejudiced group taking account of both the de factii and de jurii under the constraints of democratie argumentation acknowledging the limitations imposed on the precision and reliability of any conclusion by the nature of the subject. In accordance with the conclusion of Volume One, the contribution of this group will mainly be the invalidation of evidently biased misconceptions and the prevention of mishaps such as - in my country - the sabotage of a perfectly legitimate investigation into the genetic component of aggressive criminal behaviour by Prof. Buikhuizen. Truth would have sanctioned his research and justice would have imposed on it the conditions for preventing abuse of its findings.

Evolution as an Explanation of the Origin of Species. (2a.6)

The essence of evolution is that life is in constant change, that this change is mostly evolutionary, that it is gradual, step by step, while every new situation has its roots in the previous one. Through this process of evolution, we can explain about all changes in the living world which we experience today, and most of the changes which seem to have occurred in the history of life.

[p. 251]

‘Most’, not ‘all’. Really fundamental changes, like the appearance of totally new features, often seem to defy explanation by the Darwinistic process of mutation and selection. Sometimes we have found no fossils of transitory creatures. More puzzling is that certain new features, like the feathers of the archaeopteryx, seem to have no adaptive advantage until they have developed to a level where they permit flight, which requires that they have evolved simultaneously with the bone and muscle structures required to form some sort of wing. That kind of change induced some scientists to develop alternatives to Darwinism. Marjorie Greene mentions a theory by somebody called Schindelwolff, but I have been unable to find his work in any Dutch library. Neo-Darwinists, notably G.G. Simpson, explain these problems away by attributing this lack of fossils to the relatively short period in which these transitory forms lived, and their relatively small number. In the absence of any empirical support for these statements, Darwinists defend their point of view by two arguments:

1) ‘An adaptive advantage must have existed for the intermediary phases, otherwise the feature would not have evolved’. That is either a petitio principii or a tautology; Marjorie Greene gives a decisive refutation of this argument (p 144-149). The poet Christian Morgenstern must have had that kind of argumentation in mind when he wrote the poem about Palmström who, run over by an automobilist driving on the wrong side of the road, resolutely refuses to die, stands up and walks away ‘weil nicht sein kann was nicht sein darf’.

2) ‘Even if explanations based entirely on Darwinistic principles have not yet been found, nobody can prove that they may not be found in the future’. Until the necessity to abandon the logically simple and totally mechanistic neo-Darwinistic theory is established, we should follow Ockham and stick to the most simple and well established theory until we can prove that another is more adequate. In her work quoted above, Marjorie Greene also takes up arms against this position. My argument against the Darwinistic claim of monopoly differs somewhat from hers. It is based on the position defended later on in the chapters about science that such a proof can never be given; that not only - following Popper - a theory can never be proved right in any objective way, but contrary to Popper it often also cannot be proved to be wrong.

There is a route between the Charybdis of accepting a theory showing clear deficiencies and the Scylla of rejecting a useful theory in its entirety because it has failed tests. That is the ‘scientific’ position which tells us to accept that Darwinism explains much of what we know about life, but that there may be other mechanisms at work which are not yet known. Since the time of my first draft, biologists have for instance found that the cooperative exchange of inheritable material such as mitochondria, has played a role in the emergence of new species. That however is complementary to neo-Darwinism, not a refutation of it, it poses no challenge to the process of mutation of inheritable material and selection on its viability. Darwinists are justified in claiming that nobody can teil whether the remaining gaps in their history of evolution are due to an error in their theory or that fossils of the intermediary stages have been destroyed or just not yet unearthed and that their theory has not yet been contradicted by any of the many documented changes of species such as bacteria and certain fast reproducing insects, while it has explained those changes and has often been successfully used to generate them in laboratories. The only scientific challenge is another scientific theory with an equal or higher degree of verisimilitude.

[p. 252]

The Quantification of Information. (2a.9)

Quantitative theories of information address the following two subjects:

a)	ITS POTENTIAL, THE THEORETICAL MAXIMUM CONTRIBUTION which the information can make to the objective sought.
b)	THE THEORETICAL MAXIMUM EFFICIENCY THAT CAN BE ACHIEVED in processing information, given the kind of information needed and the means available to process it.

But what do they measure? Not information content. Quantitative information theory can help us to relate information to magnitudes of physics and chemistry such as energy and entropy, but only by analogy. It cannot otherwise contribute to understanding its nature.

We will not succumb to the temptation to present here the theories of Shannon, Fisher and consorts, however fascinating these may be. For the purpose of this book it is enough to have the reader get a feel of what these concepts really represent by applying the underlying concept of these theories (not their formulae) to our example of synthesis of milk sugar.

a) THE POTENTIAL OF INFORMATION. The significance, the information content, in the milk sugar process is contained in the representation of the availability of milk sugar in exploitable quantity. To simplify matters, we will assume that all advantages in terms of the objective of the bacteria can be expressed as gain (or prevention of loss) of energy required to enable the bacteria to survive and propagate. To determine the potential of information, in this case to establish the potential in energy saving, we determine:

1:	the lowest possible expense of energy for surviving without the information process
2:	the lowest expense of energy we can achieve with the information process and then deduct the second from the first.

If the bacteria had no switch to turn the production of enzymes on or off and no information about how to set it, it would have to choose either for trying to do without the enzyme galacto-side-permease or for a continuous production of them. This is tantamount to choosing between the assumptions: ‘there-never-is-any-sugar’ or ‘there-always-is-sugar’. This illustrates that even without knowledge of the information process concerned, we can often decide which alternative is better. Clearly, in the absence of information about the availability of sugar, the best choice is continuous production of enzymes, as without the digestion of milk sugar the bacteria would always starve, whether there is any sugar or not, while it will survive as long as there is.

Information processes do break down. In most information processes there is a default value, a value which leads to the action to take if no information is available. The choice of the default value can have important consequences. The best default value is the one which minimises the resulting damage, the well-known rule of minimising the probability of the maximal disaster. In the above case the default value should be ‘there is sugar’, and that is indeed what Escherria coli has ‘chosen’.

[p. 253]

The ‘best’ default value can serve as referent against which to measure the potential of information process. In the case of the milk sugar example the advantage of the information process is the savings of that energy which an uninformed bacteria would waste on the production of enzymes for which there is no milk sugar to process. The less likely the occurrence of milk sugar in the environment of the bacteria is, the greater the quantity of enzymes produced in vain by a bacteria which lacks that information system, the greater the ‘gross’ energy saving potential of the information process. The net energy saving is obtained by subtracting from the gross savings the consumption of energy required to run the information system. If milk sugar always would be available, the net potential savings are therefore negative. Processing information is efficient because it usually consumes very little energy in comparison to the purely physical or chemical processes required to sustain life.

Conclusion: the potential of this information process is proportional to the improbability of the occurrence of what (in this case the presence of milk sugar) we would have assumed in the absence of information. It is a verbal rendering of the well-known formula for expressing the power of information: Popper's concept of information content.

We cannot however, from this potential alone, deduce any value of the information process itself. That would require the determination of the values which the different outcomes of the process (which is directed by information) have for the information processing subject. It is well known that (always subjective) values can at best been ranked, never measured cardinally. We can attempt to find an equivalent magnitude which we can measure, for instance energy in the above case, but even then we cannot measure the value which energy has for the bacterium. Most human information processes are too complex to allow such a measure. The concept of information potential is valuable for a comparison of various information processes, especially scientific theories, provided that we remember that it never measures information itself.

b) THE EFFICIENCY OF INFORMATION PROCESSING TOOLS. A conventional Standard for measuring the efficiency of a process for obtaining information is the number of steps required in the process-to-be-evaluated to alway obtain the desired information, as compared to the minimum number of steps it would have required in the most simple process, namely one which uses exclusively queries put in a binary code such as 0/1, right/wrong, sugar/no-sugar etc. One step in such a binary-code process is called a ‘bit’. Suppose the information process is one of asking questions; if we find a specific type of question by which we can obtain the desired information by asking just once, while it would minimally take three questions of the yes-no type, we say that this specific question has an information value of three bits.

Another valid justification for using the binary code as a point of reference is that decision-making in the end must always result in choosing for one course of action and rejecting at least one other alternative; otherwise there would be no decision to take. Even if we could not find any objective reason for preferring one alternative above another, we would have to decide for one, even if we have to do so by flipping coins. All judgements which must result in a decision can always be reduced to a series of steps of the yes/no type, a universality exclusive to the binary code.

[p. 254]

Let us take an example. We are faced with four cups, each containing three identical coins. That is, the coins within a cup are identical. Three of the cups contain legitimate coins, each weighting three grams, and one contains three counterfeit coins which can be distinguished from the real article only by their weight, which is two grams. We need to know which cup has the counterfeits.

The ‘binary’ system of yes/no requires comparing the weight of the cups two by two. We can feel the difference of one third by hand. Sometimes we might get our answer in comparing the first cup to the second one; if not, we need two steps to get our answer. However, there is one way to always find out in just one step. We take all three coins from the first cup, two from the second, one from the third (for a total of six coins), and none from the fourth, and measure the total weight of the coins thus collected, which is just one information gathering step. If the fourth cup contains the counterfeit coins, my six coins would weight 18 grams, for they would all be good ones. If the third cup contains the counterfeit, one of my six coins would weight only two grams, for a total of 17. If the second cup contains it, the scales would read 16, and if the counterfeit is in the first cup, half of the coins would be counterfeit, and the weight of the six of them would be 15 grams. Thus the information content of this method is two.

Note that we did not measure the total efficiency of the process, for it ignores the extra cost of the most efficient system such as the thinking required to devise it, the cost of a sufficiently sensitive balance and its operating costs. The extra fixed costs of the one-step system will make it more costly for a first trial; it will become economical only if we have to repeat this process many times. Herein lies the economie justification of this kind of measure of the efficiency of information. For life and information usually are stochastic processes repeated over and over again, and their theory - like most theories of such processes - is not concerned with single occurrences. When explaining biological and sociological phenomena by using this measure of information, the cost of developing and operating the information process whose efficiency we attempt to measure must always be considered.

Clearly what is thus measured is the efficiency of the means used to obtain information; it bears no relation to either the quality or quantity of that information. For whatever the way in which we ask the question, the correct answer - and thus the information obtained - is always the same, and in the above example consists of the identification of the cup which contains the counterfeit coins.

Measuring information is not the subject of this book, so please note that the possible shortcomings of this paragraph should not affect the validity of the rest of it. It had to be written because these quantitative theories are often presented as explaining and measuring information itself, while in fact they do not quantify the information content itself. It also is needed to show that such objective standards of measurement in no way conflict with the fundamentally subjective nature of any information. It is not intended to belittle the importance of these theories.

[p. 255]

More Than Physics and Chemistry. 2a.9)

In the introduction I have already pointed out the contrast between the success of physics and chemistry in dealing with the inert world and the failure of social science to give us tools for organising our own society. A similar failure bedevils our dealings with living creatures, including our own body. Medicine finds it difficult to deal with ‘functional’ diseases. The problems we have generated in attempting to combat other forms of life like pests or forcing them to cooperate - for instance through fertilisers - in excess of their ‘natural’ ability are front page news. It is not so much the failure in reaching the original objective which is in cause; rather, it seems to be impossible to reach only that objective. As soon as we deal with living systems, manipulations of such systems have many and very often unpredicted effects.

Intuitively we sense that the laws of physics and chemistry which man has developed to explain the inert world are insufficiënt to deal with the living one, especially if we include man in it. Many have deduced from this that there must exist some spiritual world that is not apprehensible by the methods and laws developed in physics and chemistry. Yes and no.

YES, we need more than physics and chemistry to explain man. This ‘more’ however is also required for the explanation of all other living systems; the major division runs between science dealing with the inert world and science dealing with the living one.

NO, this ‘more’ does not imply any spiritualistic consideration, as we explain in more detail in the chapter Body and Mind, especially about Feigl, p. 362. All it takes to develop the required additional methods is the application of scientific method with a really scientific - that is open - mind.

If we think about what has been said of life in Part Two A and about the holistic nature of living systems and the limitations it imposes on their predictability and control (vol.1, 2b.5) we find one word which recurs in nealy every paragraph: ‘information’. ‘Information’ is seldom encountered in a theory about the inert world.

Physics and chemistry deal with their objects in terms of mass, energy, acceleration, specific weight, atomic number, oxidation or reduction, acid or basic, radicals etc. The specific place in time and space of the objects we investigate does not enter our considerations. And rightly so. For inert objects are largely invariant in their properties with respect to time and place. Water boils at 100 degrees at 1000 millibar irrespective of other aspects of its surroundings and - most important - of the number of times it has been boiled before.

When testing the second law of thermodynamics, we can repeat the same experiment over and over again, using the same gas artificially returned to its original state. Provided we took all required precautions in the experimental setup, we will get the same result in the last trial as in the first one.

The way living beings react is very much influenced by the context in which the reaction takes place and by their previous experience of similar situations. The place of objects in time and space is a basic element in the information process which forms life. ‘Life’ sciences need an

[p. 256]

explanation of this information process. That clearly does not mean that we can explain any living being without recourse to physics or chemistry: ‘something’ having mass and energy must be there to generate, carry and receive information.

Information implies order, a concept which we also encounter in physics. But the role order plays in physics is incomparable to the one it plays in the living world. It is the nature and dimension of that role which we have to redefine and for that we need to develop new mathematical tools and new types of models. Incidentally, we will argue in the chapter ‘Order, disorder, chaos and information’, p. 419, that the very concept of order seems, on basis of its use, as yet poorly defined. Specifically, it does not refer to an ex ante definable property of whatever entity is called ordered or chaotic, but derives its meaning from the specific process under consideration: the same arrangement of elements which is called ‘perfect order’ in one type of system may be called ‘total chaos’ in another.

In the inert world, order is totally passive. In a living being the order has a self-centred ‘function’: the arrangements of molecules etc. is one which is ‘advantageous’ to the being. This advantage can in principle be totally defined in terms of physics and chemistry. The new element is the point of view, the ‘I’, implied in ‘advantage’. While the notion of ‘statistical bias’ is familiar to the physicist and chemist, its role as function, as advantage, in living systems is a concept totally alien to physics and chemistry, as are the notions and phenomena of subjectivity, value and meaning. These, and the conditions under which their use in empirical science is justified, have been explained at various places in Volume One, notably in the chapter ‘Some remarks about functionality’, p. 44.

It has been explained why the methods of physics and chemistry, which rely entirely on today's analytical method for their explications, are insufficiënt to deal with holistic processes like learning, why we need additional methods enabling us to deal with the self-centred, telenomic (goal-directed) and holistic information process which includes learning. And why this process does not require for its explanation any metaphysical concepts such as spirit, soul or God. Why the insufficiency of physics and chemistry does no concern the nature of the entities involved, but the manageability and adequacy of any model of their interrelations in space and time.

If the basic cleft between types of science runs between science dealing with the inert world and science dealing with the living one, then the time-honoured distinction (P.C. Snow's two worlds) between hard or exact science dealing with nature, and soft science dealing with man and especially his mind, is confusing rather than revealing. For biology is only in degrees more exact than social science. The actual or perceived difference between the two disciplines in terms of exactness owes more to the attitudes, predilections and mystifications of their respective practitioners than to the nature of the phenomena investigated.

Biologists look mainly for theories which can be formulated in - or reduced to - the ‘exact’ language of physics and chemistry. Many social scientists on the other hand seem mainly attracted to their discipline because - in its present status - it appears to provide the licence to dispense with the precise formulations required by the ‘natural’ sciences, and with the discipline imposed by generally accepted criteria for evaluation.

[p. 257]

As asserted in the part about the ‘truth’ of scientific theories, all scientific theories are generalisations created by man for a purpose, an assertion which will be further substantiated (in J&A 3, ‘Against the autonomous existence of world three objects’, p. 296). Man imposes these generalisations on his self-created inner world of representations in an effort to organise his mental dealings with the real world outside. Because of their general applicability and apparent stability over time (until a new theory appears) we often fail to realise that the laws of physics and chemistry do not represent any ‘object’ of this outside world: except in our mind, there is no such thing as a law of gravity, only the consistent recurrence of falling objects. These ‘laws’ are a representation of a stability which we assume to exist in the relations between the objects of the outside world, a stability which is corroborated by our experience and for which we attempt to find an explanation. The justification for the attempt to find an explanation for this stability rather than just recording it was given in the paragraph about induction in Part Three, p. 99).

The essence of the set of laws of the inert world, the basic feature which differentiates them from any set of laws we could conceive for the living world, is that these laws, and the nature of inert objects, are considered to be beyond the influence of the objects themselves. Insofar as shape and place which inert objects take can be determined at all, we consider their shape and place to be totally determined by the laws of physics and chemistry. If we encounter a phenomenon in the inert world which does not seem to be determinable by these laws, such as the exact place of particles in quantum theory, we assume it to be intrinsically indeterminate, to be random.

By introducing in the living world the notion of a self-directed, and thus subjective, process of information, we face a field of a fundamentally different nature. The very fact that the significance of a message is independent of the nature of the medium in which it is encoded has freed its significance from the constraints imposed by the laws of physics and chemistry. Contrary to physical events on a human scale, the information process, and thus the development of life, is open ended. The central elements of the information process (for instance the expected contribution to the objective of the receiver of this information and the feedback of the actual contribution as perceived by the being concerned, on future expectations) cannot be caught in laws having the same kind of universality and specificity as those which rule the inert world.

This process is always subject to some random disturbances which may not make it totally unpredictable but which are sufficiënt to make every step of the process a stochastic one. Given its interdependent and dynamic nature, such a process becomes unpredictable in the long run and often even in just a few steps. Any long range stability it may exhibit is not provided by any ‘law’ but by the built-in bias of the process towards survival and propagation, by the very subjectivity that is so suspect to many scientists. An analysis of this bias can only provide an explanation ex post of why a being survived. It can never provide any law-like rule for the emergence of the bias itself, nor for a reliable prediction of future behaviour. Provided we hark to that caveat, such a description of the information process does not involve any concepts which are incompatible with the scientific method.

Beings endowed with a central nervous system learn at the level of the individual. The speed at which they learn, what they learn and how they translate this knowledge into behaviour,

[p. 258]

depends among others on the expected consequences of changing, respectively maintaining, behaviour. One single disturbance on the habitual trail of a deer is enough to make it change its daily route. On the other hand, if instead of shooting at it, we place salt licks along this trail without disturbing the deer, we will reinforce its habit. Using salt licks to lure it to another route can make it change its habits, but much more slowly. For while salt is desirable, safety is vital and comes first.

The above consideration underlines the importance of ensuring that the motivations of individuals which are entered as external variables into social models are realistic and are put at the correct level in their hierarchy of motivations: some objectives are derived from more fundamental ones and valid only in today's situation. Others - like the maximisation of utility - are so abstract and general as to be irrelevant at the level of actual decision-making unless we give utility a more specific content, which must be done explicitly if we want to avoid all kinds of sins against the rules for democratie argumentation.

This introduces a very interesting and well known notion which is however hardly ever encountered in social science. With a higher animal, changes in its environment may increase the stability of the model we had made of its behaviour, or they may speed up its change. Which way the influence goes depends on whether the changes confirm the adequacy of the behaviour described in the model or on the contrary exposé its shortcomings. The usefulness of the model of a system as a means to control the behaviour of the system therefore is in direct relation to the purpose to which we put the model: if, when applying the model to take a decision, our objectives are congruent with those of the elements of the system described, the model may work. If the objectives are conflicting, our very use of the model will result in an experience of inadequacy of its reactions, its behaviour, by the very elements we want to influence and induce to change their behaviour. The model then loses its usefulness and the system has become less predictable for all its members and therefore less stable. Applying that concept to a human society, we can expect that a really democratie system, that is one which respects, and aims to fulfil, the objectives of its members, will in the long run be the most stable and successful one.

In modem man, the learning function has reached its contemporary maximum in terms of generality and of the speed of creation and spreading of the knowledge it generates. Most of the information available to the builders of a model of human society is also part of the experience-based knowledge of the elements of that model (the members of that society).

The livelier the society, the less useful analytical models are as a guide to policies aiming to change it. The obvious solution would be to add to our model a representation of how we learn and how we develop objectives for decision-making. The attempt to do so will exposé the intrinsic inadequacy of the traditional analytical method and provide a better feeling for the word ‘holistic’. Because information is the main cause for the need of new methods, the flowchart type of models used by system analysts look promising.

The conclusions of this chapter have been given in Volume One, chapter 2a.9, p. 48. As food for thought, I might add that, given the fact that each of us is part of this living world, the limitation

[p. 259]

of science in providing control over the living world is no cause for sorrow as it is the best guarantee for our freedom. It also provides a general and scientific justification for what has become more evident every day, namely that a cooperative attitude towards the rest of the living world is a better basis for achieving our own objectives than the attitude of confrontation.

Meaning: Michael Polanyi. (2b.1)

He is the ‘father’ of meaning, alerting us to the essential subjectivity of meaning and the holistic nature of all knowledge. His best-known work is ‘Personal Knowledge’, but I will refer to a later book, ‘Meaning’, in which he summarises his ideas and which is closest to the subject of this paragraph. A major reason why we do not come to the same conclusions is that, like all other authors on the subject, he starts from the most sophisticated information process which the living world knows, the self-reflective human mind. He thus overlooks those basic elements of the information process which are most evident at the lower levels.

He correctly sees meaning as a construction - by the subject - that is neither inherent in the medium nor in the object of the information process. He concurs that - in the case of verbal communication - the medium, the word, is by itself a totally meaningless assembly of sounds or signs which could just as well have been different, as evidenced by the multitude of languages. He too sees the physical indeterminacy of the DNA molecule as the reason why it is so suitable as a medium for the genetic code (Polanyi, p 172). As he says ‘... every living organism is a meaningful organisation of meaningless matter... .’ He also acknowledges teleology (I call it telenomy or functionality) and sees a direction - he calls it a gradiënt - to evolution, analogous but inverse to the gradiënt of inert matter which follows the second law of thermodynamics. The major difference between his work and mine is that:

-	he presents a mechanism which generates this gradient
-	he sees this teleology not only ex post but also ex ante and asserts that it cannot be explained in physical and chemical terms alone

In Part Two I have explained how the telenomy evidenced by a living creature could have emerged as result of totally random mutations selected on the basis of their proven functionality in terms of survival and reproduction of the subject. That functionality however only becomes apparent ex post and in this it differs from classic teleological thinking. This telenomy is not itself part of the laws of physics and chemistry, yet is an essential feature of living creatures and cannot be ignored by any explanation pretending to be comprehensive. An investigation of meaning should start with this telenomy and not, as is the case with Polanyi, end with it.

The basic instrument for reversing the direction of the second law, for developing this telenomy of all living beings, is the information process. Meaning is part of this process, and should therefore be explained in terms of this subject-oriented telenomy. Polanyi's explanation should then be rephrased accordingly, an interesting project for a doctoral thesis. Here, I just want to acknowledge the stimulus his work has provided, and explain why I cannot use his concepts. I will examine the fundamental proposition of Polanyi and show how I would interpret the same phenomenon to arrive at different conclusions by starting at the lower end of the living world: a bacteria.

[p. 260]

His theory of perception (gaining knowledge) is based on dividing awareness in subsidiary and focal awareness. When gathering knowledge, we ‘look’ at a particular object about which we want information, we direct our attention to that object. As a result we will achieve a certain awareness, knowledge. This knowledge is however the end-result of a complex process of gaining intermediary knowledge. This intermediate knowledge then is integrated by focussing our attention on the focal target. We are unaware of this intermediary knowledge, which Polanyi calls subsidiary knowledge. The focal target is called by Polanyi the ‘meaning’ of this subsidiary knowledge.

Polanyi gives various examples to illustrate what he means, of which we will take one: the use of a hand-held probe to explore the interior of a cavity. We transpose the shocks transmitted to our hand and to the muscles holding the probe into an awareness of the things touched by the point of our probe. We attend subsidiarily (and without being conscious of it) to the feeling of holding the probe in the hand, while the focus of our attention is fixed on the far end of the probe, where it touches an object on its path. What we are aware of are all these feelings in the hand, all these subsidiaries, yet what we ‘know’ is their joint image integrated through our focusing on a target, which then is their joint meaning. The subsidiaries stand in a functional relation to meaning. The process also involves a very interesting phenomenal transformation: we have lost the sensation of the probe or stick pressing on the fingers and palm, and on the muscles guiding the probe. Instead, we ‘feel’ the point of the probe as it touches an object. Giving sense to the resulting image is the semantic aspect of this process, which is never explicit: as soon as we attempt to describe how we arrived at the meaning we focus our attention away from the original target towards one of the subsidiaries. Therefore Polanyi calls this method of perception ‘tacit knowledge’ in contrast to knowledge we arrived at though explicit deduction.

There are three centres of tacit knowledge, says Polanyi, (p 38): the subsidiary-particulars, the focal target and the ‘knower’ who links the subsidiaries to the target. They form a triad, controlled by the knowing entity who causes the subsidiaries to bear on the focus of attention. The ‘knower’ integrates the subsidiaries into a focal target, or in other words, the subsidiaries have for the ‘knower’ a meaning which fills the centre of his focal attention. If the knower shifts his focus of attention, for instance to the subsidiaries, the triad dissolves.

If we do shift our attention to the subsidiaries, we find that they too have a meaning; but if we focus on subsidiaries, the joint meaning we had gathered when focusing on the previous target disappears. Polanyi thus deduces that the subsidiaries have in themselves a meaning which we do not see when looking at the focal target, but which appears if we shift our attention to them. When focusing on the joint target we - as it were - look ‘through’ the meaning of subsidiaries. In a nutshell and using as far as possible his words, that is Polanyi's theory of meaning. All knowledge of living creatures except deductive knowledge of mankind then is tacit knowledge.

Turning, as in Volume One, our attention to the information process of very simple creatures we see that the components and relations of tacit knowledge do not quite conform to Polanyi's analysis. First, we see that the triad is different: we have a subject (an observer) who has an objective, a purpose. We have the world outside the observer. Finally we have the process by which the observer acts on this outside world. In the case of an information process, the observer's

[p. 261]

objective is to make a representation of this outside world, usually to bias a decision for action towards his objectives, such as survival and reproduction. The ‘meaning’ of the information gathered is largely or totally determined as soon as the observer has focused on a target, and the subsidiary elements have no meaning in themselves at all except as derived from this focal meaning. What Polanyi calls subsidiaries are simply steps in the process to arrive at a representation of the focal target. These steps are sub-processes which can involve the generation of knowledge, but the process of ‘giving meaning’ is directed by the overall process aimed at the focal target.

It is tempting to consider the intermediary steps of a complex information process as separate mini-processes in their own right, and thus as having a meaning of their own. That is Polanyi's fallacy. The subsidiary processes do have a meaning, but that is derived from their contribution to the generation of meaning of the basic process investigated and follows mainly from their natural adequacy for serving as a medium or as clues in the total process. The information generated in the subsidiary process is a selection amongst the many types of information which the subsidiary process could provide, the criterion being its bearing on the focal target. When focusing subsidiarily on an object, we may be aware of colour, pattern or movement. What we will actually be aware of depends on the kind of object we look at and the purpose which directs the information process. They determine the properties of the subsidiary object which we will attempt to establish.

This fact will be easily conceded if you think of the different meaning the same physical intermediary process will generate when applied to different focal targets. The blind man swinging a cane and a drummer testing the response of a drum to various kinds of strokes both use the same hand and arm muscles and nerves, yet give a completely different interpretation to the sensations derived from it. Only if the process itself becomes the focal target does the process become similar for both cases; but then we are engaged in a different information process.

That is why the focal meaning disappears when we shift towards the subsidiaries: we have abandoned the original process, and with it the meaning it was intended to generate, and engaged in a new one. Contrary to Polanyi I hold that any information process only generates one (set of) meaning of which we can be aware. What he calls secondary awareness or subsidiary meaning is nothing of the sort. They are intermediary processes which derive their meaning from the primary process, from the focal target. As the focal target changes, so will the ‘meaning’ generated by the intermediary processes. We cannot within the same process become aware of both the focal and the intermediary ‘meanings’; these appear only if from intermediate they were promoted to focal meaning. We may be able to become aware of some bits of the knowledge generated in subsidiary processes, but only to the extent that it has been recorded in our medium or long term memory; that is an extremely small selection of all intermediate meanings which we generate.

We are of course capable of introspection. But that requires engaging in a quite difficult process specifically aimed at such introspection, and usually involves repeating the activities we remember from a previous process, and focusing on them. The blind man would start by probing the cavity, and then stop that activity at various stages and focus on what he did. We can obtain knowledge about our intermediate information processes only by building up a record of

[p. 262]

what we do. Except for rare cases, we have to build such a record specifically for the purpose of introspection.

Obviously, the above applies to sub-processes involving mainly tacit knowledge. One of the rare cases where we can in principle retrace a whole and complex information process is deduction, especially if it is explicit. But even then, deduction also involves tacit, totally unconscious information sub-processes in our brain, many of which we will never become aware of at all.

Polanyi distinguishes between ‘from’ and ‘focal’ awareness. Focal awareness (becoming aware of what we focus on) needs no explanation. The ‘from’ awareness is the tacit knowledge embodied in what the subsidiaries teil us: the ‘feelings’ of each muscle etc, which he calls the ‘meaning’ of the subsidiaries. The two are - according to Polanyi - mutually exclusive: the focal awareness, which we indubitably have, disappears as soon as we focus our attention on the ‘meaning’ of the subsidiary, the ‘from’ awareness, and vice versa. Just as with meaning, I hold that there are no two states of awareness in the same information process, but only the focal awareness, and that Polanyi in fact deals with not one, but two or more separate and consecutive processes in which the first process itself becomes the focal target of the second etc.

We also differ in another dimension. Polanyi identifies explicit knowledge with deduction. He calls ‘tacit’ any knowledge of which we are aware, but which does not involve deduction, for instance, all perception. And there is to him no knowledge outside conscious knowledge: the tacit integration of subsidiary knowledge towards the focus of attention can be carried out ‘only by a conscious act of the mind’. That makes ‘conscious’ a synonym of ‘aware’.

Polanyi's view is applicable only to human beings. What would we call the state of the central nervous system of a dog who sees a rabbit? Is he aware of the rabbit or not? If we say he is aware of it, then animal and human consciousness are equal. If we say that the dog is not aware of it, because his tacit perception is not integrated by a conscious act of the mind, then what is it? I would say that both dog and man are aware of the rabbit, but that only man can perceive that he is aware. Man can make this awareness the subject of another information process, while the dog cannot. But to do so, he must leave the natural, parallel, flow of information processing. Being conscious of awareness is the essence of reflective thinking. It is the major difference between humans and animals, possibly with some intermediate cases like certain primates and sea-mammals. It is the result of an information process which I call conscious or analytical, as opposed to unconscious or holistic. The difference in meaning between conscious and aware thus defined seems to me conform to the common usage of these words. As any hunter knows, a dog clearly shows, by wagging his tail in a specific way, that he ‘knows’ that there is a rabbit; so he must be aware of its presence.

The above illustrates the fruitfulness of my approach of looking at information processes in general and start with the simplest ones; human information processing then becomes a very particular case specific to human beings within the general field of information processes. Below is another example of why my approach has more ‘empirical content’.

[p. 263]

Polanyi illustrates the difference between tacit and explicit deductive knowledge by stating that we cannot ‘act’, for instance play the piano or ride a bicycle, by explicit knowledge and that we also cannot provide an explanation for that inability. If we follow the view of information presented in this book, the explanation is evident: to create conscious, explicit knowledge we must temporarily move outside the unconscious, holistic, information process (which is the only one which can run the extremely complex information system required by the above physical actions within a reasonable time span). Conscious thought can thus only direct the flow of that unconscious process, it can never be part of it. In accordance with my view of reason as controlling the holistic process, not replacing it, it is clear that while we agree that we cannot ride a bicycle by using only explicit knowledge, such explicit knowledge can help us to learn to ride a bicycle by telling us not to try to go straight ahead, but to follow a slight wave pattern around that straight line. We will then learn it faster than a chimpanzee who has to learn exclusively by trial and error. At first we will be conscious of what we are doing and the riding will be awkward. We will be proficient in cycling only after we have integrated that explicit knowledge into our holistic, unconscious information process.

It is remarkable to what extent people can look at one and the same phenomenon and yet arrive at an opposite explanation, depending on the preconceived point of view from which they look at it. Explicit knowledge, says Polanyi, ‘is totally ineffectual unless it is known tacitly, that is, unless it is known subsidiarily, unless it is lived in’. He mentions Konrad Lorenz's view that science is based on Gestalt-like integration of particulars, and that the speed and complexity of tacit integration far exceeds any explicit selection of supporting evidence. And, ‘An integration... will often override single items of contrary evidence. It can be damaged by new contradictory facts only if these items can be absorbed in an alternative integration...’ (Polanyi, p. 41/42)

This statement is in accordance with Part Two of Volume One. But then comes interpretation. ‘... that it is our urge to understand and control our experience which causes us to rely on some parts of it subsidiarily in order to attend to our main objective focally.’ In the light of what has been said before, that conclusion seems to be, as we say in Holland, ‘drawn by its hairs’. The simple statement that the holistic, tacit, process is much more efficiënt would have been enough, and would apply to any generation of information, whether it is for action or for understanding. My view of course is that - as we can focus only on one information process at a time - we must always relegate whatever we do not focus on to subsidiary knowledge. Only that element on which we focus can become part of explicit knowledge in that process.

Conscious thinking is the exception, the bulk of our information processing is unconscious. Why such focus on understanding? Because according to Polanyi that is needed to introduce another ‘important feature of personal knowledge. It will appear that all personal knowing is intrinsically guided by impersonal standards of valuation set by a self for itself’. Such standards ‘...are generally applicable rather than personally idiosyncratic.’ or: ‘...to which he attributes universality.’ His examples could however all be explained by the social dimension of man, fact which he ignores because his real objective is to prove wrong the notion that standards of evaluation are instrumental, objective-directed, and thus conventional.

[p. 264]

He implicitly attacks the notion that a demarcation criterion must be conventional by arguing that to obtain knowledge through ‘indwelling’, by personal participation and thus looking at the actions of others from their point of view is not at all ‘unscientific’. Obviously that is totally congruent with my contention that all knowledge involves a subjective element. His statement that ‘ALL knowing is personal knowing - participation through indwelling’ supports my view that there exists no totally objective knowledge. But those like Popper or myself who attempt to discriminate between scientific and other knowledge never put in cause the means to acquire knowledge. We only deal with the evaluation of knowledge which has been presented to us as being scientific and as meriting the qualification that it is the best available. I reject knowledge as unscientific only if it is presented in a form which prevents us from evaluating it in this respect, or if the evaluation shows that it cannot improve our decision-making about facts pertaining to our world of empirical experience.

Contrary to what Polanyi implies, neither science nor a demarcation criterion demands that we study man and society in a detached manner, and thus refrain from ‘indwelling’ as a method for creation of knowledge. My view of information clearly accepts indwelling both as a method for developing knowledge, and for evaluation of knowledge by the individual. Indwelling is rejected only as a method of evaluation of scientific knowledge as to its suitability for social decision-making. As the evaluation of a scientific proposition almost never achieves unanimity, we need something else to decide as a society about which scientific theory to use in establishing facts.

‘Now at last we are in a position to show how philosophic thought and the methodological principles of science have been misguided by not having achieved a clear knowledge of tacit knowing.’ (Polanyi, p 46). (I would say the same thing, but replace ‘tacit knowing’ by the information process, and include Polanyi in those who have been misguided.) Polanyi then presents his view on the mind-body problem, on behaviourism, on universal terms, on principles of explanation, on empirical generalisations, on scientific discovery, on determinacy versus indeterminacy, and on the objectivity of scientific theories. The interested reader will be easily able to rephrase these views in terms of the concepts presented in Part Two and three of volume one. While I disagree on the specifics, I agree with Polanyi that all meaningful integrations (including those of science) ‘are inescapably personal.’ But - as stated above, his theory targets the view that science is factually objective knowledge; that view is held by no modern epistemologist, not even Popper. Polanyi does not address the real problem with which these epistemologists attempt to deal: the evaluation of scientific theories.

He states that the distinction between facts and values does not rest on any existential difference. Agreed. But that provides no justification for denying that we can and do differentiate between facts and values... unless one rejects any instrumental view of knowledge, and does not acknowledge the necessity in social decision-making to distinguish the de factii from the de jurii (i.e. value).

‘... meanings created in the sciences stand in no more favoured relation to reality than do meanings created in the arts, in moral judgements, and in religion’. (Polanyi, p 65) Quite so. But in practical decision making, I would rather rely on physics than on religion for making my car

[p. 265]

go, and on economics rather than on art to deal with unemployment. That choice is justified by Polanyi's own standard of general custom. But as previously argued, a personal and subjective preference need not be arbitrary, but can be perfectly rational if directed by the objectives I have in applying that knowledge to decision-making. This instrumental view is anathema to Polanyi for exactly the same reason it is to Popper: it robs any moral theory of its universal and absolute status above human authority.

Polanyi disagrees with Popper's statement that any free society must be an open one. Yet both are alike in one respect: their shared abhorrence of totalitarian societies. (Polanyi, p 3) Both develop a theory which - if carried to its logical conclusion - would point to the inevitable subjectivity of all knowledge, and thus to the rejection of any absolute authority. Yet both shrink away from this conclusion. ‘Here the inconsistency of a liberalism based on philosophic doubt becomes apparent: freedom of thought is destroyed by the extension of doubt to the field of the traditional ideals, which includes the basis for freedom of thought. (Polanyi, p. 10).’

Polanyi is not a professional philosopher, yet one might expect him to be aware of the refutation of this kind of argument. Russell - as I do - considers freedom of thought to be a meta-ideal applicable to all kinds of specific thoughts and ideals, including itself, putting them all on a par and rejecting only any a priori authority of any of them over the others. How to reconcile freedom of thought with the justification of imposing it on those who do not subscribe to it is precisely one of the major subjects of my book (PART ONE and FOUR A).

I have shown that the extension of doubt to all fields of thought need not have the consequence Polanyi and his consorts fear. On the contrary, it protects us from the deadly temptation to shift responsibility for the defence of freedom to some abstract principle.

Polanyi also points to the negative nature of any liberalism based uniquely on a rejection of moral authority and unfettered individualism. I share that objection, which is precisely the reason for Part Two on life which introduces the notion of a purpose beyond the satisfaction of material comforts of the individual: to live in a society and to safeguard and develop the information, genetic and cultural, which the individual and his society represent.

The will of man to live in a healthy society which respects his ultimate authority over himself, and the consequent rejection of the a priori authority of any other moral principle, is itself a moral principle of the highest rank. It is not an abdication of thought; it is thought permeating all human endeavour. In short, while Polanyi's work did provide many ideas for the present book, his general philosophy is not relevant for my subject.

‘Subjective’ and ‘Functional’ Have No Connection With ‘Selfish’. (2b.3.3)

A professor who had a glance at my book wrote: ‘I agree with much of what you say [follows a list of subjects] but disagree violently with your comments on evolutionary biology and the writings of Richard Dawkins’. I did not base my conclusions on Dawkins' work. On basis of a

[p. 266]

review of Dawkins' ‘The Selfish Gene’ I decided that it did not seem to add to Monod anything I needed. What it did add seemed to suffer from some hubris, to overstretch the scope of its biological conclusions. Whatever the merits of his book, its title does not at all reflect my view of subjectivity, functionality and genes. The factor in the genes which is responsible for the bias towards self-realisation and propagation is the fact that the sequence of nucleotides has in the past managed to survive and propagate. That sequence does not by itself express or contain any specific element to which we can point and say: that is the base of that bias, that ‘selfishness’. The bias arises only because the process - of which that sequence of nucleotides is a part - has managed to keep going, and its significance becomes apparent only within that process. We could properly call ‘selfish’ only the tendency towards survival and reproduction which is an ex post feature of the whole process of living. If we equate this tendency with selfish, the word ‘selfish’ loses all discriminatory power, because all living beings which are not in that sense selfish in all their aspects cease to be part of the living world, will not propagate.

To obtain discriminatory power, its opposite, ‘unselfish’, must be a meaningful and possible alternative concept. If we give ‘selfish’ the above general meaning, any unselfish motivation is conceivable only in a creature capable of choosing between objectives. To be distinct from gratuitous, unselfish must be willed, have an objective. That objective can only be the well-being of other beings and therefore is conditional on their existence and presupposes the real opportunity to choose in any other than a selfish way.

The choice to act in a selfish or unselfish way therefore can arise only within a creature capable of reflection and imagination and conscious of having to live in a society, but also conscious of its own individuality, i.e. man. In short, ‘selfish’ either is seen as a moral category, and then is applicable only to man, or it is intended to refer to the general tendency, the decision-making bias, of all living beings towards survival and reproduction. In that last case it is an ex post and somewhat tautological concept not particular to genes. Applying ‘selfish’ to genes may make a catching title, but is it good science?

I write this addendum in the hope of preventing any misunderstanding about my use of subjectivity and functionality. In my book they have no direct connection with selfishness. Also, Dawkins’ ‘selfishness’ promotes a distorted view of nature by putting the accent - like the concept of survival of the fittest and the struggle of all against all - on fight and egoism at the expense of cooperation, symbiosis, and - in the case of humans - morals, which have a prominent function in their survival and propagation as social beings.

The systems approach to life leads to the same conclusion. If we use ‘selfish’ in the above, generalised - and to my mind tautological - sense, then we cannot ascribe it to specific genes as there are no genes whose specific function is the generation of a bias towards survival and reproduction. Neither can we ascribe that bias exclusively to the genes, for that bias requires the contribution of all those elements of the nascent beings which are indispensable for its survival, such as cytoplasm and mitochondria. The genes do stand out from all other elements because they are the only one which is necessary for all living beings, even retro-viruses, to keep the process of life going and because they are almost pure information. That does not however justify the application of ‘selfish’, whatever that may mean, exclusively to the genes.

[p. 267]

Sense and Referent: Frege. (2b.4)

Many philosophers see an information process, especially communication, as containing three elements:

-	the object of the process = the referent
-	the subjective representations generated in the message
-	something in between, the sense, to which they impute an existence which is objective, that is independent from the sending or receiving subject.

The problems and paradoxes bothering these philosophers originate in this autonomous ‘sense’. The analysis of the nature of information (PART TWO B) teaches us that such an objective sense is an illusion, and that only the object and the representations in the minds of the subjects exist. This illusion is fostered by the extensive, though never complete, ‘overlap’ of these subjective representations, fruit of the enormous effort we invest in its creation, and the similarity of the information processing apparatus of men. The godfather to the problem is Gottlob Frege, and we will take his ‘Funktion, Begriff, Bedeutung’ as our point of departure. (Frege was studiously ignored by the academic establishment and killed himself twenty-three years after publication of his work, two years before its first acknowledgment.)

Using his simplest version, let us take two statements of equality:

a = a and b = a

Such statements are of any consequence only if the symbols are used to represent something, such as objects, in which case they are signs, names, for these objects.

If ‘a’ and ‘b’ refer to objects, the second statement will be true only if both ‘a’ and ‘b’ refer to the same object. If we understand the equality sign in these statements as meaning ‘referring to the same object’, both statements would tell us the same thing, and are totally interchangeable.

Yet the two differ profoundly in what they can tell us. For a = a is a tautology and tells us nothing we did not already know, while we may be ignorant of the fact that ‘b’ refers to the same object as ‘a’. Knowing that the morning star and the evening star both refer to Venus certainly is not trivial.

Frege gives another example which is especially convenient for my purpose. If we connect each of the three apexes of a triangle with the centre of their opposite side, the lines starting at the apexes A and B will intersect at the same point as the lines starting at C and B. The commonality of the intersection of all three lines is by no means a foregone conclusion. If we call the point of intersection of the first set of lines x, and of the second set y, then their equality as expressed by x = y cannot be the whole story even though both refer to the same geometrical point.

For the equality does not really concern the names themselves, the knowledge that they both refer to the same object is trivial; it only tells us something about the language, like the fact that ‘lit’ in French points to the same kind of object as ‘bed’ in English. But the complete statement tells us more than that the names refer to the same object. That something more, according to

[p. 268]

Frege, is the ‘sense’ of the statement. The morning star is in some way different from the evening star although they refer to the same object. Names therefore may convey something which is different from just the referent. Unless we are concerned primarily with semantics, statements become interesting only if they do more than just point to the same referent, namely if they contain an assertion, if they are a ‘Behauptungssatz’. Thus Frege distinguishes sense from referent as being that which is added by this assertion to the object, the referent. After two pages of introduction, Frege's remaining pages are mainly devoted to properties of this ‘sense’; yet he never tackles the question: what is it?

Just once, and summarily, does Frege touch on the basic nature of ‘sense’, namely when he introduces the notion of a representation generated in the mind of the writer or reader of a statement (my translation): ‘The referent of a name is the object itself which we designate by it; the representation which we thus have is totally subjective; in between lies the sense, which is not any more subjective like the representation, yet is not the object itself either.’ (Frege, p.44)

Frege rejects the conflation of sense and representation because the representation - being subjective - can and does vary from individual to individual, while to him it is evident that the sense does not. He therefore believes that the sense of statements has an autonomous existence, independent of individuals who hold it. In this he is the precursor of Popper. This subject will dealt with extensively in the chapter ‘Against the autonomous existence of world three objects.’, p. 296. Here only the Frege's arguments will be scrutinized.

As an analogy, Frege takes a group of individuals looking at the moon through a telescope and the representation of the moon which each makes. There is a real picture, real rays of light, created by the telescope through its lenses, and there is the image on the retina of the observers. The picture created by the telescope is objective in the sense that it can in principle serve as a common object to diverse astronomers. The images formed by the retina would however vary from one astronomer to another because the configuration of the eyes differs per individual. Let us extend, Frege says, the analogy by imagining an observer of the astronomers who through some kind of apparatus can see the images projected by the eye on the retina of the astronomers; these images would be objects to the common observer. Thus, says Frege, we can imagine that the subjective image of the moon produced on the retina of the astronomers can at the same time be an ‘object’, have an objective existence, to an observer of the astronomers.

He implies that sense is in some way similar to such an image. Unfortunately, he does not pursue this illustration further as to its relation to ‘sense’, because ‘...dies zu vervolgen, wuerde wohl zu weit abfuehren’. That is a pity; for had he done so, he would have noticed that he is not dealing with one, but with two information processes:

-	astronomers gathering information about the moon,
-	an observer gathering information about astronomers.

He did notice that the eyes of astronomers apparently are constructed in a sufficiently similar way to produce a reasonably similar image of the same object on the retina, but that there remain some individual differences which - coupled to differences of positions in space and time - will result in some variation between the representations. In his time it was not known that these ‘images’ on the retina are not what the astronomers ‘see’. The representation which

[p. 269]

we can integrate into our information processing, about which we can talk, is not the image on the retina, but the interpretation by our brain of the electro-chemical impulses generated in it by the image on the retina. That interpretation is another element of variation between subjects, and this element is quite different in its nature from the images recorded by the retina in that it is generated not only by the inherited configuration of our brain. From the first day we first see light, the interpretations of the signals transmitted by the optic nerve are also determined by complementary experience (touching the object). Seeing has to be learned.

He would also have noticed that the first analogy - the enlarged picture created by the telescope - is totally unsuitable as an analogy for an information process as it is a purely mechanical manipulation of rays of light. It simply states that there is more than one step between the referent (the moon) and the external impulses (rays of light) as registered by the astronomer. It is an analogy for any multi-step information process.

There is no evidence of any element totally independent of the astronomers in between their object (the rays of light coming from the telescope aimed at the moon) and the subjective representation they make of it. What Frege sees as the ‘sense’ does not pertain to the information process of the astronomers, but pertains to the representation of this process made by the observer of the astronomers which again contains the inevitable subjective component. The impression of commonality and consequent ‘objectivity’ of this sense, is the observer's interpretation of what he sees when looking at the configurations on the retina of the astronomers and of subsequent conversations about the experience.

Let us look at Frege's problem in the light of what we have said about information and use our example of the milk sugar system (Volume One, Part Two, 2b.3, p. 61). The referent here is the presence or absence of milk sugar in the immediate vicinity of the bacteria. The medium is the inhibitor, either ‘naked’ or ‘clad’ by a sugar molecule. A free inhibitor is the message, the name, the sign for ‘no sugar’. An inhibitor bound to a molecule of sugar is not a name for sugar: it is not recognised at all by the operators as a name for anything and thus contains no specific message; it is interpreted as ‘no information’.

What else do we find in that information process? We find two elements: the producer of an inhibitor and the operator who has an affinity with the inhibitor in its original, free, state. That must be the elements which produce the representation, the information ‘there is no sugar’. The representation ‘no sugar’ is created by an event, the encounter of an operator with a free inhibitor.

The production of the inhibitor must precede the making of the representation. But that production ‘makes sense’, leads to an information process, only if there is an operator who can ‘understand’ it. All elements of the process, including the sense, must be in place before the information process can produce the representation.

We see now the real nature of the phenomenon which give rise to the notion of ‘sense’ and which is responsible for the impression of objectivity which it gives to the superficial observer. ‘Sense’ is the set of potential representations which the subject can make in a specific

[p. 270]

information process. It is chronologically, not logically, a priori in relation to that process, having arisen in the evolution of the subject up to the moment of the actual information process under consideration, which evolution includes his inherited apparatus plus what he has learned since his conception.

The unifying element generating the impression of objectivity of the set of possible representations is the function of the information process which defines the part of its environment about which the subject seeks information and the requirement that the representation be correct, coupled to the objective existence of this environment. The first two result from evolution through learning, either or both at the level of the sort and at the level of the individual. In communication there also must be an a priori and thus learned commonality of language.

Frege was not concerned with bacteria, but with human language. As I have said, we can safely assume that no two individuals ever make exactly the same representation upon hearing the same word and even within the same individual the representation of a word will vary over time. Yet there seems to be a sense to a word that is independent of the individual which utters it: if we say ‘one table’, nobody thinks about two chairs. The details of the representation we make of one table may vary, but there seems to be an essence which is the same for everybody. That is why Frege, and in his wake many others, for instance Russell and Popper, distinguish between the representation made by an individual when confronted with a statement, and something which is different in nature from this representation and which they call ‘sense’. As this ‘sense’ seems to be independent of the individual, it is tempting to attribute an autonomous existence to it.

The view of the information process in this book provides another explanation for this independency which logically is just as sound, namely to consider this ‘common’ sense of a word as the ‘overlap’ (not to be taken graphically, but I could find no better word) common to all representations which individuals make of a message, an overlap which is the result of a strong motivation to learn from each other, a necessity for a species who is so dependent on social interaction. The overlap view can explain all facts and thought experiments put forward in support of the autonomy hypothesis, while it also explains something for which the proponents of autonomy have to introduce purely ad hoc hypotheses, namely how a sense that is considered autonomous can change according to the state (in time and space) of the societies concerned.

We could call that overlap the social sense of a word. All the problems which bothered Frege then disappear, at least as problems of logic. They reappear as practical ones. For the social sense does not arise out of the blue but requires a constant effort of our cultural institutions to create an overlap sufficient to ensure the unity of ‘sense’ necessary for the coexistence and cooperation of the members of the living system, of the society, concerned. Much of this book is devoted to that problem in the field of philosophy and social science.

The overlap will always be imperfect and contain ambiguities and unwarranted connotations. In another work (‘Begriffsschrift’), Frege attempts to improve overlap so as to make it perfect. That effort has been both a success and a failure. It has succeeded in drawing attention to the ambiguity of language and to formal languages as a means to avoid such ambiguity. In Part

[p. 271]

Five I have recommended the formal language of logic as a control procedure for the evaluation of statements designed to serve as a basis for social decision-making. The ‘Begriffsschrift’ is a failure in that Frege hoped that all statements of science could and would be made in his language. That objective clearly has not been met, nor is it likely ever to be.

Sense-giving is part of a holistic process, is always done in a context, which endows our common information processing and communication ability with a richness compared to which a formal language is practically barren. Formal languages are unsuitable for any purpose except to control that holistic process. The richness of the holistic process follows from its contextuality. But this contextuality also generates ambiguities and connotations which are often unwarranted but in most cases unavoidable. How bad are these imperfections? In decision-making our objective is limited to achieving adequate communication through an effort commensurate with the importance of the decision to be taken, given a - finite - set of available alternatives. Any serious attempt to meet Frege's standards will show that eliminating all imprecisions will take up an amount of our limited information processing capacity which is totally out of proportion to the contribution which this precision can make to our objectives.

Unaware of the nature of life as a constant fight against chaos and of the role of information in that fight, the notion of functionality and efficiency was simply not part of the field investigated by Frege and his successors. They could not appreciate the efficiency of holistic processes which - however imperfect - are the only way to achieve a timely and energy-efficient communication about the reality around us for decision-making. These imperfections become counterproductive only when leading to misunderstandings of such a dimension that the wrong alternative might be chosen.

Whenever our objective is to construct tools for universal application, such as scientific laws, such imperfections are bound to have very serious consequences. In such circumstances the use of formal languages comes into its own and - as explained in Part Five - are an indispensable means for achieving agreement about the de factii in a democratic society.

Frege marked a step in philosophy, making us conscious of these problems and helping us to ask the right questions. A century later it has become clear that solutions should not be sought exclusively in logic, but also engage empirical science (psychology and social science) coupled to a clear understanding of the kind of society we strive for.