The conceptual foundations of decision-making in a democracy

[p. 401]

Knowledge and Human Understanding as a Social Venture.

At various places I have pointed out that to understand man, we must acknowledge both his individuality and the fact that he is a social creature. The reader may feel that in the Part Three about knowledge and truth the social aspect did not receive the attention it merits. He would be right... if my objective had been to present a scientific and empirical theory of human knowledge and understanding. As clearly stated, my objective is norms in decision-making. Epistemology enters the picture only to the extent that it is relevant to that objective. All decisions, also those of society, are in the end taken by individuals. Criteria and norms are developed from the point of view of an individual having to take a decision or to evaluate the decision of somebody else, for instance of a politician.

That may justify abstraction from the social context in which knowledge is actually developed and applied. But as soon as we want to translate the very general criteria proposed in my book into more specific ones and attempt to apply them, we will have to take account of that social contex. Specifically, we will have to explain how the current theories about the social aspects of knowledge are congruent with my Part Three. This chapter is a first modest step in that direction, an attempt to point out some avenues for reconciling the social theory of knowledge with my text.

From the literature dealing with social context I distilled three subjects which seem relevant to my work:

1)	the problem of fitting it into a modern view of human understanding such as Toulmin's
2)	the problem of incommensurability of research programs (Collingwood and Kuhn)
3)	the problems which the social context might generate in the evaluation and selection of scientific theories for social decision-making (Ravetz).

As usual, the major differences between the above authors and my book derive from the fact that:

-	they start right at the apex of information processing: conceptual systems and scientific theories.
-	partly for that reason they do not do justice to the function of knowledge in the process of life.

1) Stephen Toulmin: ‘Human Understanding’.

(Abbreviated as HU)

The primary concern of Toulmin was not with the practice of decision-making in science, but the process by which science and its theories came into being. It is gratifying to find that all major differences with my work can be traced to that difference in objectives and in points of view from which we approach the subject.

As other coryphees of epistemology, Toulmin did not base his theory on a general and explicit concept of information. He took it for granted and started from the most complex type of

[p. 402]

knowledge, science. He came to the same conclusion as I did: the inevitable subjectivity of all knowledge. But this conclusion was not deduced from his view of what understanding is, but only by default: he noted that we have never found criteria for total objectivity which could survive the scrutiny of reason and that, any time we believed we had found one, that belief landed us in totalitarian societies. Had he started at the bottom of the hierarchy of knowledge, he would have found the obvious explanation that subjectivity is an essential element of any information process, that it is precisely that subjective element which distinguishes the information process from other processes.

We also differ in our objectives. Toulmin mainly describes how we develop (scientific) knowledge, while I take the point of view of a user who needs to evaluate knowledge in its function of a means for deciding on the de factii component of social decision making in a democracy. Yet Toulmin also arrived at the conclusion that decisions on scientific theories by society are in the end a matter of authority.

1a) SCIENCE AS A SOCIAL VENTURE. The fundamental contribution (in my view) of Toulmin is the notion that to understand scientific knowledge, we must see science as a social venture, an enterprise of scientists engaged in the activity of producing scientific knowledge, and that - if we are to understand either that venture or its products - we must always take account of both, for they are part of one interdependent system: science.

Specifically, Toulmin notes that laws and theories never ‘explain’ phenomena; only scientists do. Such laws and theories may be, and usually are, invoked whenever a scientist attempts to explain a phenomenon; ‘...they “serve as” an explanation, when produced in the appropriate context and applied correctly.’ (Toulmin, p 157). ‘In the practical business of giving explanations, scientists - after all - rely not only on the presentation of explicit deductive arguments, but on such alternative activities as the drawing of graphs, the construction of intellectual models, or the programming of computers’. I would add that the use of metaphors plays an important part in most attempts to get a new theory understood and accepted, a fact of which Galileo already was well aware. Toulmin also notes that formal arguments do not in such presentations enjoy some clear priority by those who make the presentation, but are used on par with all other forms of rhetoric. The selection criterion which the orator uses for choosing between them is their expected contribution to his objective, which is to get his point (for instance a scientific theory) accepted by the audience (Toulmin, p 158). I agree: the primary objective of the scientist is to get his view accepted, to persuade if he cannot convince.

But in a book about democracy, the appropriate point of view is not that of the scientist making the presentation, but that of his audience. Cabaret for instance primarily intends to entertain the audience. In a presentation of a scientific statement, the people in the audience want to gather knowledge which can help them in their own ventures (for instance to develop their own theories). If they value their autonomy, they want to remain as free as possible of the interests of the person making the presentation. They want to be able to evaluate scientific knowledge on basis of their own objectives, which - as noted - requires a maximum of objectivity, and thus of compliance to the rules of democratic argumentation.

[p. 403]

It is not possible to devise any objective and analytic criterion of evaluation for science itself - which includes scientists - because it is a living system. But its theories are not living systems, they are the product of one. Analysing a theory and its products will, if at all effective, produce new information, and thus will change the whole field of inquiry because of its interdependency. As the theories themselves are not a living system, they need not change by the event of being analysed; only their evaluation and interpretation will change. Except in rare cases where the very nature of the field of investigation prohibits it, the product of scientific enterprise - laws and theories - can be described in the form of logical, deductive, systems of concepts, and can be accorded truth values, i.e. be axiomatised. That offers the most objective (in our sense of inter-subjectively valid) evaluation of the theory, and the audience in a democracy is entitled to require such axiomatisation to the extent that it is possible.

Given the immense opportunities for fraud and the limited possibilities of the lay audience to evaluate the presentation of a scientist, we must to some extent rely on his good will. Credibility thus becomes a valuable asset for a scientist, and that gives his audience a weapon: it can police, or have policed, these presentations and disqualify scientists who obviously frustrate the democratie venture of being as objective as possible in the establishing of facts.

In his introduction (p. 36) Toulmin emphasises the distinction between private versus social knowledge which I presented in Part Three (p. 96). He points out (p. 156) that ‘the content of science is, thus, transmitted from one generation of scientists to the next by a process of enculturation’. His description of that process is quite similar to what I wrote: we learn the meanings of words and concepts by using them and by interaction with our teachers and colleagues, thus creating the minimum congruence of individual meanings necessary for effective communication. Specifically, if we want to decide whether an apprentice has or has not grasped the significance of the concepts of a particular discipline, Toulmin notes that we do not investigate his ‘private’ mental grasp, his (hypothetical) inner life, but that we test his ability to explain them to others and to apply them. His private knowledge is irrelevant and not directly accessible to society.

Toulmin investigates the concept of rationality and gives it a content similar to mine: he states that the concept should not be applied to the products of human activity, but to the activities from which such products arose. The reader will note that this is exactly the way I explained away Russell's problem about judging the rationality of a person claiming to be a poached egg. Applied to scientific theories, it tells us that we should not judge the rationality of specific scientific theories, but of the procedures we use in evaluating them. Unfortunately, Toulmin does not make explicit the logical structure of his concept of rationality, and thus fails to notice that such a judgement can only be made in relation to the objective, the purpose, of the activity concerned, which in the case of the evaluation of scientific theories would be their application by others, i.e. in social decision-making.

Toulmin opposes the knowledge of nature to the - reflective - knowledge of ourselves. The general problem of human understanding is, he says, to draw an epistemic self-portrait which is both well-founded and trustworthy; a portrait which is effective because its theoretical basis is realistic and which is realistic because its practical applications are effective. Considered that

[p. 404]

way, our practical standards of rational judgement are one face of the coin whose other face is philosophical. The practical (‘outward-looking’) task is to apply these standards of judgement to decide which ideas, concepts or points of view have the strongest current claim to intellectual authority over our thoughts and actions. The theoretical (‘inward-looking’) task is to give an analytical account of the considerations by which the authority of these very standards is to be judged. He writes (p. 3/4): ‘...any excessive separation between the theory of knowledge and the practice of knowers should prompt us to ask: “Are scientists becoming unreflective? Or are philosophers losing their sense of relevance?”.’ To the student of the philosophy of justice that should sound familiar: it is the application to epistemology of Rawls' ‘reflective equilibrium’.

TOULMIN AND KANT'S ‘VORSTELLUNG VERSUS DARSTELLUNG’. Because Toulmin does not conceive human understanding and the concepts it uses as a special case of a general information process, he runs into a problem when dealing with Kant's notion that all our experience is concerned with ‘representations’ (Vorstellungen) of objects of the ‘outside world’ rather than with the objects themselves.

‘Let us assume’, says Toulmin on page 194, ‘that Kant's assertion is to apply to all experience and knowledge including the mechanisms generating knowledge from sense-perception. Then the logical necessity of imputing a transcendental nature to these a priori disappears as soon as we have a non-transcendental explanation for the a priori which is inevitably involved in the making of a representation.’ In that Toulmin is perfectly correct.

He concludes that if we have such a non-transcendental explanation, Kant's attempt to escape from subjective idealism would then be frustrated, and that therefore Kant cannot have meant that his view of understanding is to include the mechanisms of sense-representation, but must have intended his theory to be restricted to the way we talk about such direct sense-knowledge, not to the sense-knowledge itself. We should then apply Kant's concept of understanding only to the way we talk or think about knowledge, not about objects. Kant could have avoided - in Toulmin's view - such a ‘misunderstanding’ by using the word ‘Darstellung’ instead of ‘Vorstellung’. However much I reread Kant, I can find in his writings no substantiation of Toulmin's assertion. And for a good reason: Toulmin's expectation that it might be possible to find an explanation for sense-representations which does not involve some a priori is simply wrong.

The analysis of the information process presented in Volume One totally supports Kant's view of the inevitability of an a priori to any ‘Vorstellung’. We do differ with Kant about the nature of this ‘a priori’ which today we know to be generated by a previous process and thus chronologically a priori (evolutionary) information process. Such a non-transcendental a priori opens a passage between the Charybdis of subjective idealism and the Scylla of the idealism of the universal and transcendental.

Right at the beginning, Toulmin relegates Kant to the camp of the idealists because he has recourse to absolute, universal principles of rationality, of pure Reason. We should however notice that Kant is a very special kind of idealist, namely a reluctant one. His transcendental ideas are introduced only at those points where a transcendent explanation becomes factually

[p. 405]

impossible within the knowledge available in his time. He also had the integrity to draw the consequences of the presumed transcendental nature of these ideas by very explicitly refraining from giving them any ‘empirical’ content and saying that we cannot make any assertion about them in terms of ‘reality’.

But if applied to ‘thinking and talking about thought’, i.e. to reflection and communication by symbolic language, Toulmin's distinction between Vorstellung and Darstellung is perfectly justified and relevant. In these activities we never deal with objects but exclusively with representations: we make representations of a representation. That always involves two different information processes, and thus two different subjects: the one generating the ‘object’-representation and the other generating a representation of the object-representation. If we have a non-transcendental explanation of the generation of these representations such as proposed in my book, then we do not need any idealistic element such as an autonomous existence of abstract concepts.

TOULMIN'S FUTILE SEARCH FOR AN ‘IMPARTIAL’ POINT OF VIEW. Toulmin mainly deals with intellectual information processes and communication. He is not really interested in the way an individual makes sense-representations. One may then wonder why he bothered at all with such a highly personal subject as ‘a point of view’. The explanation is that his real objective is not an explanation of understanding, but to find a passage between the Scylla of absolute principles and the Charybdis of total relativism.

(Toulmin, p 478) ‘... how, despite the diversity of concepts and rational standards in different human enterprises, situations and milieus, we can nevertheless - in appropriate cases, at least - define for ourselves an impartial standpoint of rationality, and so escape from the threats or temptations of relativism.’ As with Popper and many other philosophers, we find that the elements to which I object are not directly derivable from the ‘hard core’ of their theory. Accepting my objections does not therefore invalidate their theory. Their objectionable conclusions are deduced from auxiliary and rather ad hoc hypotheses which are introduced precisely because they fear certain logical consequences of that hard core. That fear always concerns the abuse of their theory to justify totalitarian aspirations. That effort is doomed because, as explained in Volume One, no theory can ever provide a defence against tyrants; the only defence is the concerted use of power by those who oppose any tyranny.

The artifice which Toulmin uses to find his impartial point of view is his differentiation between theories and the historically developing enterprise, science, of which they are a formal artifact or abstraction. The rationality of the scientific enterprise lies primarily in its procedures for conceptual change. According to Toulmin, applying comparative logicality - as the only measure of rationality - to successive theories based on different sets of concepts leads only to confusion. He notes that efforts by Popper, Lakatos cum suis to extend an analysis of rationality which is still formal (i.e. a-historic) to scientific procedures has induced them to view conceptual change as an anomaly.

Instead of demonstrating that rational procedures of scientific enquiry have a kind of ‘logic’, Toulmin says we should investigate how the formal structures and relations of propositional

[p. 406]

logic are put to work at the service of rational enterprises. For such logic is no more than an instrument of scientific investigation and explanation. We should resist, he says, any attempt to derive standards of judgments of collective enterprises from an absolute authority such as logic. Neither can they be found in some external, a priori (demarcation) criterion or imposed by the relative and transitory authority of a local reference group.

I certainly share that view. But it may be possible to define, in a chronological a priori, an objective of the enterprise which is shared by all participants. That would provide a factual and rational standard for collective decision-making. I have argued that consensus about an objective and the acceptance of the standards deduced from it is at least conceivable and can often be achieved, provided we accept the premiss that no-one can claim any a priori authority. This premiss implies the rejection of the claim of total objective validity or total impartiality for any standpoint or rational judgement. Toulmin argues that we can define such an objective, impartial standard, namely by a ‘...sufficiently full and detailed understanding of the features that make alternative strategies and procedural innovations “adaptive” to particular problem-situations: his “ecological approach”’. We can judge, can make rational comparisons, of concepts, procedures [and theories, auth.], of collective enterprises, only in the light of the problems they were supposed to solve, ‘...how far, and in what respects, the alternative strategies employed in each milieu have actually fulfilled the historically developing purposes of the relevant enterprises.’ (Toulmin, p 498). That judgement, by his own words, cannot be made a priori. Yet, if we can make it only a posteriori, we would be thrown back into total relativism in all decisions, as these always concern future actions, and would have reverted to Hume's scepticism.

Toulmin asserts that we can escape the choice between a priori and a posteriori by ‘... investigating what there is to be meant by “rationality” and its associated concepts, on what conditions rational questions become operative, and how the practical force of those questions reflects the conditions of their applicability’ (Toulmin, p 499). ‘...there is one basis, and one alone, on which our judgements of “rationality” and conceptual “merit” can truly be impartial. This is one which takes into account the experience which men have accumulated when dealing with the relevant aspects of human life - explanatory or judicial, medical or technological - in ALL cultures and historical periods.’ (Toulmin, p 500) I certainly will not contest that assertion.

His next conclusions also are totally in line with my own propositions: ‘...our procedures are never to be considered final, we are forever compelled to lay bets about our future, and we can ground the strategic estimates on which rational changes of policy are based only on a well-digested appreciation of earlier achievements...’. I agree with Toulmin that our past experience is all we have in terms of objective data to guide us, and that we must not a priori exclude any of it if we are to be impartial. But even if it were humanly possible to achieve that feat - which I think is totally beyond our capabilities - that still would not be enough in any concrete situation of social decision-making. What are we to do if two competing policies both claim to be based on a ‘well digested appreciation of earlier achievements’?

In fact, Toulmin remains primarily concerned with the explanation of the venture of science as a product of human activity which we would like to explain on the basis of the criterion of

[p. 407]

rationality. As such I have great respect for his achievement, although I am not qualified to evaluate it on the basis of its pretensions in that direction. But, as explained, it cannot provide us with any totally objective basis for deciding for or against a certain theory as soon as that decision is contested. It does lend support to my view about scientific knowledge as a social venture. And if we accept my theory of democratic decision-making, it can help us in putting it into practice.

Summing up, Toulmin's research program could be useful in improving and substantiating what I said about information, and my assertions about the nature of scientific knowledge. Toulmin correctly poses the problem of authority in the evaluation of scientific theories and is relevant for the normative part in that he supports the rejection of any absolute (as opposed to conditional and functional) authority of any a priori concept. However, he relegates that authority to history, which may be correct but is of little help in everyday business which is to take decisions whose consequences lie in the future.

THE HOLISTIC NATURE OF SCIENCE. In Part Six I pointed to the ‘wholeness’ of science and the ravages of its compartmentalisation. Toulmin provides additional arguments. ‘So, if we are seriously to revive this central philosophic aim [the rational appraisal of human understanding, author], we must ignore contemporary attempts to divide off various epistemic disciplines by academic frontiers with professional check-points.’ (p. 6). ‘Questions about the processes, procedures, and mechanisms by which our concepts are developed, acquired, used, and/or improved may be topics for particular sciences or disciplines... The philosophical question in epistemics is, by contrast, the question from what sources our concepts ultimately derive their intellectual authority.’ (p. 10)

ABSOLUTE TRUTH VERSUS RELATIVISM. As I do, Toulmin sees in a procedural rationality a way out between an absolutist and a relativistic (and thus positivistic) view, (p 85). Agreed. But where does Toulmin present his criteria for procedural rationality? In any case not his Volume I.

In volume I Toulmin promised a volume III which is to deal with ‘Rational Adequacy and Appraisal of Concepts’. Unfortunately, it is totally unavailable in Holland. That is surprising, considering Toulmin's status and the fact that his volume I is present in all academic libraries. I therefore doubt if it has ever been published. Fortunately, a pupil of Toulmin, Ravetz, has published a book dealing with those subjects, and I will rely on him to represent at least Toulmin's school of thought on these subjects, if not his text.

2) Incommensurability of Conceptual Systems and Scientific Research Programs: Collingwood/Kuhn.

My knowledge of Collingwood is limited to the summary of his ‘An Essay on Metaphysics’ (Oxford, 1940) in Toulmin's above book (Toulmin, p. 66/78). From that summary it seems that Collingwood's view about conceptual systems comes close to mine, except that I extend it to purely deductive systems like mathematics. A conceptual system, he says, is

[p. 408]

hierarchical and its most comprehensive principles are located at the apex of the structure. Conceptual systems are not axiomatic systems, which are dependent on the truth of these most comprehensive principles; they are ‘systems of presuppositions, where the logical relations between propositions on different levels of generality are not truth relations but meaning relations.’ Collingwood's standards are not truth, but logical efficacy, relevancy and applicability. They include the tacit assumptions which so often underlie both our interpretations of phenomena as well as our choice of the phenomena we look at.

At the head of any self-supporting conceptual system lies a set of absolute presuppositions, presuppositions which do not, in turn, depend for their meaning on other, more general, presuppositions but whose meaning is taken for granted. Abandoning a set of absolute presuppositions means abandoning the whole conceptual system because the meaning of its propositions is hierarchically dependent on those absolute presuppositions. Toulmin has drawn attention to the similarity of Collingwood's absolute presuppositions and Kuhn's paradigms and indeed that is the reason for mentioning both authors in the same chapter.

Collingwood did not address the question of where these absolute presuppositions come from, nor how and why they change. His failure in this respect originates in the same omission which landed him irrevocably in historical relativism: the failure to acknowledge the functional character of knowledge. As with so many others, by starting at the highest-level information process he did not conceive clearly the subjective and (ex post) goal-directed character of any ‘meaning’.

The view of information presented in my book supports his view that in their conception such systems are systems of meaning. But my view also entails that - in their application to decision-making - we transform conceptual systems into axiomatic ones by adjoining a truth value to meaning, and evaluate such axiomatic systems in terms of expected contribution to the de factii aspect of decision-making.

I also hold that there is one presupposition common to all scientific conceptual and axiomatic systems, and that all others, at least in empirical science as well as in most formalised systems, are dependent on it: the assumption of ‘no miracles’, and thus the concept of causality: no event is by its very nature beyond possible explanation in terms of causality, at least at the level of normal human decision-making. Problems of causality generated for instance by the indeterminacy of elementary particles are seen as not (yet) solved, but not intrinsically insoluble provided we do not look for absolutely true solutions. (See my chapter on quantum theory, p. 423)

Collingwood's notion of absolute presuppositions gets him into trouble when explaining changes in these absolute presuppositions. Clearly his error is the use of the word ‘absolute’ instead of ‘uncontested’ or ‘incontestable within the system’. Change becomes explainable as soon as we acknowledge that the acceptance of these presuppositions is conventional, and that their semblance of absoluteness derived from the perceived ‘wholeness’ of the system.

Collingwood's view of conceptual systems highlights their uniqueness, their incommensurability, in short their holistic character. There are two reasons why I do not from this conclude to

[p. 409]

Feyerabend's intrinsic incommensurability of different conceptual systems, and even less of the theories which they spawn.

If two conceptual systems are different in their ‘absolute presuppositions’ and their meaning, it may indeed be impossible to compare them in their original totality and form. But there is no law of man or nature which prevents us from making the presuppositions explicit and translating the meaning of those of one system into concepts understandable by the other; in fact, that should be a major task of epistemologists. And if the systems have any pretense to deal with our empirical world, then such a translation must be feasible. As Toulmin has documented, it is also a job to which authors of a new scientific system devote quite a lot of energy, often with success. It is true that success depends on the willingness of the audience to cooperate, but that is not a property of the systems. If successful, we can then evaluate the two sets of presuppositions in a rational way.

Toulmin correctly criticises Kuhn's view of scientific revolutions which implies total incommensurability of the old and the new system, and the absence of any basis for a rational discussion about their relative merits. Kuhn's view cannot in that respect withstand a confrontation with history: the transitions took a long time and were subject to rational argumentation all the way, even though rational argumentation was not always the most conspicuous element of the process of transition.

The holistic and interactive nature of learning entails that a new concept will become acceptable and accepted only to the extent that it can be rooted in concepts which are already familiar. Such ‘new’ concepts do not arise out of nothing. Copernican and Einsteinian revolutions are extensions of already existing concepts to new combinations and to new aspects of reality which form a new (logical and conceptual) system; their new concepts get their meaning from an evolutionary and interactive learning process by applying them to problems, actual and new, and old, real and imagined. The conclusions of the new are then confronted with those obtained when applying the old concepts. In that same process, these old concepts may acquire a new meaning. The main tool in this process is (the pretension of) rational argumentation.

If we see a conceptual system as hierarchical, with at its apex a set of ‘most general’ principles (or presuppositions) essential to all subsystems, then that set functions as Kuhn's paradigm. And if that set is altered in any substantive way, the resulting system will be sufficiently different from the original one to merit the qualification of new (as opposed to ‘updated’). Its structure may be so radically different that it spawns a new cosmology which will be prima facie incommensurable with the old one. If such a change occurs within a short period, say one or two decades, then it will present itself as revolutionary both in its appearance and in its consequences, however evolutionary the process may in fact have been.

With Toulmin, we should therefore reject both Kuhn's concept of revolution as an adequate epistemological explanation of change as well as the total incommensurability of the new and old system. But we should retain the notion of hierarchical conceptual systems and of paradigms (as the set of most general principles).

[p. 410]

Yet neither Collingwood nor Kuhn has drawn attention to what I hold to be most relevant in the notion of ‘most general principles’ or paradigms, namely the extent to which various disciplines share fundamental presuppositions. Direct sharing of the fundamental presuppositions may be the exception. But with the possible exception of physics, the knowledge developed by one discipline will often serve as part of a ‘most general principle’ or axiom in another. For instance a main tenet of my book is that the findings of biology (from molecular chemistry to neuro-physiology) should, and certainly will in the future, serve as the most fundamental presuppositions in all humanities, and that the main task of today's philosophers is to expedite that process and prevent its abuse by false prophets.

Such expediting is urgently needed. While no such ‘most general’ principle can claim absolute validity, it is a fact of life that a large part of the scientific community is taking their principles for granted up to the point of becoming totally unaware of them and of their implications. And they would be unwilling to question them, the more so because the work of many scientists will be invalidated by a change of principles, as Kuhn has pointed out. If the principle is indeed a ‘most general’ one, then all members of a scientific discipline will be affected, and resistance to change maximal.

As Toulmin noted, once a certain alternative has become public knowledge scientists are quite willing to rationally discuss it. The problem is to achieve public awareness. The fact that these same scientists whose work will be affected are predominant in the procedures which today determine the selection of scientific papers for publication does generate a problem which Toulmin did not address, but which we will discuss in the next chapter (about Ravetz). It is a problem of social decision making, not of epistemology.

The main objection to Collingwood's and Kuhn's theories is their implication of change in scientific systems as some historically driven inevitability. There is no scientific law which requires us to see it as such, and the fact that a theory emerges ‘when the time seems right for it’ only provides an explanation of the date at which it emerged, not of its conception. Current discussions about the evaluation of scientific theories and about the best way to organise and manage scientific institutions testify to the fact that most of us hold that this process can be influenced; standards for accepting or rejecting theories into the body of science are acknowledged and discussed. Even the evolutionary, ecological, view of science should, as we have seen, never lead us to confound ex post explanations and causality with determinism, with the notion that it could not have been different. Our efforts to direct the process of selection of scientific theories towards our objectives need not a priori be irrational or immoral, need not be doomed... provided we do not extend it to the conception of theories which contains a large stochastic and personal element.

3) The Social Responsibility of Science, Ravetz.

Ravetz worked with Toulmin and presumably has digested and included in his book the relevant ideas of his mentor. His ‘Scientific knowledge and its social problems’ is an excellent review of the practice of science in its social context and is warmly recommended to those readers who are interested in the application of the concepts proposed in my book. I cannot do

[p. 411]

justice to his work within the scope of a chapter, so I will not attempt to describe its content. It is relevant to this chapter because it deals with what I presume I would have found in the third volume of Toulmin, namely how society evaluates and selects scientific theories and the authority which that selection implies.

It is tempting to show how my concepts of ‘holistic’, of the subjective element of information and of the two-faced, self-assertive and integrative, nature of man can be illustrated by examples gleaned from Ravetz's book. I will resist that temptation and limit this chapter to the major differences between Ravetz's work and mine, which all result from a different point of view (the scholar versus the decision-maker), a different direction of approach (producer versus user-oriented) which lead to different conclusions about today's basic problem in science and what has to be done about it.

A good part of Ravetz's book is devoted to the description and explanation of the social aspects within the practice of science, and its consequences for its product: knowledge. The connection of the scientific community and its products with the outside world receives less attention. That connection, the use of scientific knowledge and the compensation of scientists is my focal point.

The scientist develops knowledge. The technologist uses that knowledge to create tools for performing certain functions. The rest of society uses these tools to solve practical problems. Ultimately, how the rest of society perceives the contribution of scientists determines the income and prestige of scientists, and the most prominent contribution is the increase of our capability to solve practical problems.

Ravetz did not write his book just to show his erudition. He focused on a real problem: how to ensure the quality of scientific knowledge, and to what extent scientists should and can be held responsible for the consequences of the application of their discoveries by others. He finds two factors who threaten the quality of scientific knowledge. One of them operates within the scientific community. The other concerns the relationship between that community and the rest of society in a modern economically developed country.

Ravetz agrees - and offers additional arguments - that there are no totally objective and conclusive standards and methods for evaluating the truth of scientific knowledge and safeguarding the quality of a product which pretends to improve our grasp of reality. There is no general and conclusive definition of ‘better’. Evaluation then becomes a craft which is performed by peer review and generates three problems:

-	to fully appreciate the portent and quality of a work of science requires a craftsmanshi at least equal to that of its author
-	at the highest level these peers are few in number
-	there is no instance reviewing how they do their job, no courts of appeal for their verdicts.

That is how the problem presents itself if looked at from the point of view of the producer of science, if we go in the direction from science to technology to application. But if we take the point of view of the consumer of science, of the decision-maker, if we start from a problem and look for the best technology to apply, we have to rely on what science has to offer, and this

[p. 412]

situation is quite different. To users of the products of science, any theory which improves their decision-making is welcome, and the concrete problem situation enables us to give a clear and operational content to the word ‘better’.

We, the ‘consumers’ of science, need some methods and procedures which, although far from solving all problems, can help us to bias our decision-making in the right direction. For instance, by applying my single purpose demarcation criterion we can eliminate at least those theories which show little promise of helping us to deal with reality. By enforcing the rules of democratic argumentation and - in the case of scientific theories - by requiring axiomatisation, we can eliminate from the evaluation those forms of subjectivity, prejudice and fraud which are avoidable.

While I agree with Ravetz that total and perfect axiomatisation is not practicable, we can often in decision-making do without such perfection. For we need not decide upon a theory once and for all. We use the iterative method (remember Achilles and the tortoise) and proceed step by step. All we need is to determine the level of precision which we require. At each step we can then define a question whose answer will determine whether we have reached that level. If so, we can stop right there and take a decision. If not, we have to move to the next step and define a new question. As shown in the first volume, the general form of that question is whether a further increase in precision could lead to answers which would make us change the decision taken on the basis of the latest results. If the decision-maker is competent, he will in most cases be able to justify a decision after just a few steps. Decision-makers can thus usually arrive at a decision about a scientific theory even if their craftsmanship and competence in the field concerned is inferior to that of the author.... provided the scientist will collaborate and answer their questions to the best of his ability and as truthfully as possible! If we accept my proposals for democratic argumentation, we can disqualify a scientist and his theory on the evidence of an attempt to sabotage the process of evaluation. That makes such sabotage costly to its perpetrator in terms of credibility of himself and of his theory. It may thus be quite realistic to assume that it is possible, by establishing adequate institutions, to improve the social evaluation of scientific knowledge, a function which today is fulfilled almost exclusively by those in power in the organisation of which the scientist is a part and by the referees they appoint.

Ravetz provides an overview of the problems which arise if a scientist is dependent on his own organisation for acceptance and accreditation, usually by referees and peer-review, of an important innovation and notes that if that innovation goes against the body of theories etc. acknowledged by his organisation, his chance of success is practically nil.

I think that Ravetz understates the problems inherent to the evaluation by referees, for he does not seem to be aware of the extent to which the referees (have to) rely on the outer signs and trappings of craftsmanship and the academic status of its author rather than on an assessment of the content of the work submitted to their judgement. If the author is unknown, referees will often take as the hallmark of craftsmanship the number and complexity of formulae, the maximum use of the idiom particular to the craft and the number of other authors cited in the work. That might explain why so many works of science and philosophy make a maximum use of professional jargon and references to other authors, and thus become unintelligible to any

[p. 413]

layman and difficult to assess on basis of the criteria proposed in my book. Sometimes that may be the inevitable consequence of the nature of the subject. But it also can and does serve to hide a vacuity of content. The book of Polten to which I referred in the chapter about body and mind (p. 362) is a good example.

Relying on Ravetz's signs of craftsmanship may not do much harm if the work is targeted only at professionals in the same field as that of the author or if the subject is an elaboration and refinement within a well-developed research program. In short, if it is part of what Kuhn calls ‘normal’ science. Yet even in ‘normal’ science, the system of referees and patronage by the own organisation of the scientist often fails in exposing vacuity of content, and thus in its function of quality control. It is almost bound to fail in recognising and acknowledging the value of a true and profound innovation. Really cross-disciplinary work may not even reach the stage of review.

The above assessment of the situation is not intended to question the integrity and professional competence of the referees. Rather, it follows from the conditions under which they have to work. Assessing the content of a piece of science, and even more of philosophy, is both far more labour-intensive and risky than assessing its craftsmanship.

If the subject is really innovative, the referee ideally should himself have the creative set of mind which leads the innovator to discern a new ‘fact’ in what others just considered an anomaly. If it is cross-disciplinary, he must have access to the money, time and modesty required for obtaining the cooperation of like-minded colleagues in those fields in which he is less competent. To judge content and not only form, he must in both cases invest the time required for analysing the paper in terms of logic, method and facts. Given the current deluge of papers resulting from the pressure on scientists to publish, it is patently impossible to satisfy the above conditions in today's practice of evaluation by referees paid by the publisher or the university.

It would be naïve in the extreme to expect that we will ever be able to solve the problem of selection of papers. As also argued by Ravetz, the very nature of scientific innovation is an open-ended process which precludes any rigid method of selection. In tune with our functional, evolutionary and procedural approach, that impossibility should not deter us from doing all we can to improve the selection procedures so as to further our goals, in this case to assess not so much the craftsmanship of the author, but rather the value of his product in terms of improving our knowledge of the world in which we live.

The rejection by publishers usually is polite: ‘Very good, but your book does not fit our list, we have no capacity etc.’ As I come from a business environment, it is easy for me to guess how the letter would have read if the truth be told: ‘Dear Sir, we have never heard of you, nor have we found any reference to a publication in your name. The subject you treat is not currently popular. Considering the supply of high quality material submitted to us by well-known authors, we regret........’ I completely respect and understand the editor's decision. ‘Try to get part of your book published as an article in a well-known professional journal’ is an ubiquitous suggestion by well-meaning friends of a desperate author. Alas, the same problem applies: articles are easier to get published if you already have a book on the subject to your credit. Referring to the book lends authority and helps to keep the text within the format of an article. And

[p. 414]

editors are necessarily biased. The damage of an erroneous rejection is a lost opportunity which will become an embarrassment to the editor only in the rare cases where the rejected work becomes somebody else's successful publication during his tenure. Considering the time it takes for many innovations to achieve recognition, this is a not a big risk. The costs of a flop are immediately evident!

All these problems just illustrate that everything of value comes at a cost. The most aggravating aspect of today's selection procedure is that the verdict of the publisher or professor is final and there is no point in contesting it. Discussions with frustrated authors would blow up the publisher's overhead to astronomical proportions. And in the rare cases where a professor whose judgement carries weight will accept to evaluate a work which did not originate in a current program of his discipline, he will do so only if that does not entail any further engagement. That is where an improvement of today's procedure can be achieved. In a healthy democratic society there should be no such thing as a final decision, certainly not in the first round. It should always provide for a court of appeals to limit the damage from the inevitable errors to which any human judgement is prone. The history of science greatly understates the damage done by the current situation because it only documents the problems of those inventions which in the end (sometimes after the suicide of the inventor) did achieve acknowledgement. The others have vanished without a trace.

If publishers and the academy of science cannot be expected to provide such a service, who can? For reasons which will be clear to anyone having read Ravetz's book, we cannot call on the other institutions of society either; in fact, it is difficult to conceive how any fully formalised and thus bureaucratic institution could improve today's situation. But it is not impossible. As my proposal to that effect is the same as the one I propose for the second problem put forward by Ravetz, we will first present his second problem.

Ravetz felt that the most urgent problem of science did not lie within the scientific community, but in its relations with the outside world. Private industry and the state have displaced the private Maecenas and the independent universities as promoters and financiers of scientific research. Consequently, the efforts of scientists are now directed and exploited to the benefit of those in power in industry and state, at the expense of other legitimate and often vital interests, such as future generations (healthy environment and natural resources) and of other, often ethical, problems with which we are confronted today. The obvious failure of western democracies to adequately deal with that kind of problem was the very motive which led to my investigations. And as shown in Part Six, I agree with Ravetz that scientists cannot escape responsibility for what happens in our societies. But I come to different conclusions both in diagnosis and in cure.

According to Ravetz, the mistaken notion that science is value-free is at the core of the problem. That notion provided scientists with an excuse for not taking a professional stand on moral issues and for not generating within science a countervailing power looking after those interests of society which are neglected by the powers of industry and state. Ravetz finds his justification for rejecting the notion of a value free science in the classic stance embraced, in a different form, by Marxists. New findings in science, he says, are not value free because their application

[p. 415]

in technology and problem solving is selective, and thus favours certain functions and certain types of technology, and as a result certain end-uses and interests. The scientist who publishes such findings then cannot escape responsibility for the fact that by publishing, he favours these technologies and purposes. I agree with this assertion, but I think that it is not the core of the problem which we are facing, for two reasons.

First, the emergence of new insights into the way of nature is only partly accidental: the general situation of the field of knowledge must be ‘ripe’ for this event. If it is, if a discovery can be made, it is only a matter of time before it will be made; the role of chance errors in our information processing ensures that even deep ‘ruts’ of the current reserach programmes cannot prevent an innovation for ever. Retranslating from a quotation from J. Robert Oppenheimer (Atom and void, p. 87) which I found in a Dutch article: ‘Still, it is a profound and necessary truth that the deep things in science are not found because they are useful; they are found because it was possible to find them.’. Scientists are aware of that. We cannot expect them to refrain from publishing a discovery because of its possible or even likely abuse, for they know that their sacrifice will at best result in a short delay.

Secondly, it is an error to assume that any man, even a scientist, can ever assess the full impact of an innovation. We simply cannot predict the future of scientific knowledge as a whole. A discovery in a certain field may very well lead to another discovery in a different field whose benefit may outweigh the damage done in the first one. We cannot even predict all the practical applications to which a discovery may be put. There is one thing which we do know: in today's constellation, all discoveries will be used by the powers that be if they find any use for them. But that alone is insufficient for judging their end-impact.

Ravetz finds his hope in the emergence of the ‘critical science’ first publicised by Barry Commoner (which I have not read). I have great respect for scientists forgoing comfortable jobs in a traditional scientific environment to further their ideals, and for what they have accomplished. But it is clear today that these accomplishments fall far short of what is necessary for the survival of humanity and that their results are totally out of proportion to the effort invested in them. An honest assessment of today's situation must conclude that the problem does not lie with a lack of knowledge and technology, but with the way we apply them. In short, the problem is not one of scientific discovery, but of decision- making in society as a whole.

The social context of ‘critical science’ has been that of contestation against the powers that be, especially industry. Many ‘critical’ scientists fell for the temptations of Marxism and the consequent accent on the ‘negation’ of what ‘is’ and turned against capitalism and industry. It is very satisfying and unifying to fight a common enemy. While defining what we stand for, developing solutions and attempting to implement them is hard and often heartbreaking work, for we will have to find a common denominator for all diverging individual interests and ideals, all this in face of an amorphous and paralysing indifference of the rest of society. What a contrast with a juicy exposure of the evils of the powers that be by aggressive ‘action groups’!

The prevailing contestational approach confronts the public with two (or more) incompatible views - which all claim to be ‘scientific’ - of groups hawking their interests or their own

[p. 416]

anti-establishment agenda aand where the more radical elements gain most of the publicity. In both camps there are scientists who must be aware that the claims put forward by their own camp are biased and inflated as to their precision and reliability. Anyone who has actively participated in political decision-making will understand how loyalty and the fear of disavowal of their own group prevents such scientists from taking a stand for truth in sufficient numbers to provide the public with some guidance as to what the relevant facts really are.

Opposition to ‘establishment’ science is necessary. But it will become efficient only after some kind of process has been put in motion which - out of conflicting views - generates the minimum of consensus required for democratic decision-making. Until then decision-making will be done exclusively on the basis of power plays in which the establishment will dominate, followed by extremists and demagogues; the values which science claims to stand for will come in a distant third.

The point of view of Ravetz is inevitable if one starts, as he does, by looking at science and introduces the rest of society only to the extent that his inquiry into science directly requires it. As with any inquiry into a living system, we miss relevant information if we do not have a picture of the system as a whole and the place of science in it. As shown in Part Two of volume one, adequate apprehension of a hiearchic living system requires that we make two investigations: one starting from the subsystem, and one starting from the whole system.

Ravetz himself noted that the ‘material’ of science, both its problems and its products, are mental constructs whose main use lies in decision-making: at the level of science itself, at the level of technology, and sometimes at the level of practical problems. The appropriate field of human activity for an investigation into the social aspects of science, and on how to integrate science into the rest of society, is therefore decision-making, which is the point of departure of my book.

The road that Ravetz explored so well can be travelled by a single very competent and talented professional, for he does not have to venture outside his field - epistemology - except for some exploratory forays. I have travelled the other road and looked at what science can tell us about decision-making of living systems in general and of a human democratic society in particular, and what it takes for science to adequately fulfil its role in decision-making by the system as a whole. Such a venture has to deal with all aspects of social decision-making which - as stated in the introduction - is beyond the capability of any single human being. But that job has to be done. The main ambition of my book is to define that job and stimulate others who are better qualified in various fields to come together and rewrite my book. Whatever their conclusions will be, the start must be the formation of a loose association as explained in Part Six. Can such a group meet the problems raised by Ravetz?

That is totally unpredictable, but not impossible. Its effectiveness and authority will depend on the quality of its judgements both in terms of objectivity and of correctness. Such quality can be achieved, provided the group is modest in its objectives and claims. It should not attempt to tell politicians what to do, nor should it venture into judgements which it cannot totally justify. In many instances it will have to confess its inability to arrive at a decisive conclusion.

[p. 417]

Even with such limitations, the impact of such a group on social decision-making can be enormous. It will be open to both ‘critical’ and ‘establishment’ scientists, provided they put democracy (and thus objectivity) above the interests and beliefs of their background. In fact, it is necessary that the group contains both kinds of scientists.

This same group can serve as a court of appeal to scientists whose work is rejected. On the basis of a synopsis aimed at showing the benefit of a work of science or philosophy, the group could decide that a certain paper is worth further investigation. It could then superficially read the papers thus selected and submit those that survive this scrutiny to a real evaluation of their content on basis of the criteria proposed in my book. If the group comes to the conclusion that a paper rejected by publishers or peers can make a real contribution to our knowledge, the author can then be helped along, for instance by a letter of recommendation.

The very existence of such a group plus a few successes will increase the risk to publishers of missing out on what could soon prove to be a success. In the course of rescuing unjustly rejected work, the group might also achieve the expertise and authority necessary to expose the serious inadequacies of publications evidently not meriting publication, thus providing a countervailing power to pressure which the authority of fame, hierarchy and money might put on referees to refrain from disqualifying works which enjoy such support. The group could provide the necessary support for a referee or a critic who has incurred the ire of a famous scientist for exposing the vacuity of content of the publication of a protégé. While the objective of discouraging the writing and publication of a vacuous work may not today be the most urgent one, we inevitably must tackle this problem in the near future because of the exponential growth of publications (Internet!).