SCIENCE AND SOCIETY

When a philosopher develops his theories of diverse topics these may be unrelated. When they are related, the question, how well they fit together to form a system, may intrigue a reader.

The conventions of science which we have discussed, particularly the convention to examine and test scientific hypotheses, are social institu­tions. Popper attacks inductivism as naturalistic and thus psychological and thus question begging. In his Logic of Scientific Discovery, the view of science as an institution is merely implied - by contrasts. He stresses the contrast between psychologism and the view of science as intersubjectivity, as well as the contrast between the naturalistic view of science (be it psychologism or apriorism) and the view of science as based on agree­ment. “Agreement... puts... theory to the test” he says (106). “Coming to an agreement... is... to perform a purposeful action...” And, finally, he contrasts “a justification with a decision - a decision reached in accor­dance with a procedure governed by rules...” (109) - which contrast he clarifies by an analogy with a verdict by a jury. Perhaps the nearest he comes to institutionalism proper is where he stresses that language is a given for scientific inquiry (104).

Popper speaks explicitly of science as a social institution, or of the institutional aspect of science, especially the conventions of testing as an institution, in his Poverty of Historicism, section 32. In a similar vein he speaks in his The Open Society and Its Enemies (ii, 218), of the “public character of scientific method”, and of “the working of the various social institutions which have been designed to further scientific objectivity and criticism”.

(I wish here to express my great indebtedness to Sir Karl Popper’s un­published Postcript) to which I tend to ascribe a sociological view of science more distinct than in any of his published writings. But I check myself: though I have studied that work very closely, in manuscripts and in gallies, it is nearly two decades since I saw it last, an I do not trust my memory to such an extent on a matter of interpretation; in memory what is kept is usually interpretation, not the verbatim text, and this is liable to surreptitious change unless checked.)

Popper goes further. He attacks both psychologistic individualism and holistic or organicist collectivism - and both are versions of naturalism ([Open Society] i, 60; ii, 88 if., 206 if.).²

He contrasts this with what I shall call conventionalism, though he calls it legal or moral positivism, namely the theory according to which social rules and codes are merely arbitrary conventions (i, 68, 71; ii, 206). He offers instead his own view which I shall call modified conventiona­lism, though he calls it conventionalism or critical conventionalism (i, 60 if; ii, 178 ff). According to this view, first, social institutions are conventions, including codes, rules, coordinations, etc.; second, they contain arbitrary elements as well as built-in errors; and third, conventions can be improved with the aid of criticism. On behalf of the third point Popper likens institu­tions to scientific hypotheses (i, 163; ii, 218). To conclude this picture, Popper does not claim that all social factors are comparable to scientific hypotheses. In his ‘Towards a Rational Theory of Tradition’ ([Conjee- turesj) he describes traditions as myths; and in his ‘Back to the Preso- cratics’ (in the same volume) he shows how mythology may transmute into science (or traditions into institutions) with the addition of a dash of defi­niteness and a dash of the critical attitude.

How far can we draw the comparison between science and society, between hypotheses and institutions? Surely, there is the (irreducible) dualism of facts and decisions: hypothesis is merely factual, whereas an institution has both a factual element and, more significantly, a con­ventional element.

This may be a restriction on the length to which Popper’s analogy will go; and so I understand him to view matters. For my part, I would like to push the analogy all the way, for reasons to be outlined below. In my opinion a hypothesis which may be viewed as scientific is not merely one which answers certain desiderata and thus qualifies for the honorific status of science. On the whole, the idea of the honorific status of science, the implicit view that the task of the philosopher of science is to offer medals to scientists, appeals neither to Popper nor to his former students; but as for Bartley and myself at least, we feel that Popper still follows the tradi­tion and demarcates science by a set of desiderata, which makes it difficult to dissociate Popper from those who give medals to scientists. In contra­distinction, I prefer to follow Derek J. De Solla Price in considering hypotheses as scientific - for better and for worse - only if they are taken up by the world of science ([Babylon], 116, 124). Popper’s view of science as the better, never the worse, hypotheses should at least be modified to say, the (better) hypothesis which answers certain desiderata is scientific only if, and to the extent that, the interpersonal conventions of critical ex­aminations and testing apply to it. There are many crucial remarks in Popper’s own works supporting this view. He does say in his Logic of Scien­tific Discovery (82-4) and elsewhere that scientific character is not a quality inherent in a proposition or a theory, but largely our attitude towards it, in particular our agreement not to protect it from refutations: when we de­cide to rescue a hypothesis from refutation it ceases to be scientific.

There is another limitation on the extent to which we may draw the analogy between science and society. Science, according to Popper’s views as expressed in his later period, aims at the truth as an ideal goal. Social institutions, inasmuch as they embed factual contentions and expec­tations, may share this characteristic. But whether social conventions qua decisions may aim at ideal goal analogous to the truth is an open question. This open question can, and perhaps should, be divided as to moral decisions and as to social conventions. Is there such a thing as the ideal moral code, the moral truth, final and unattainable? Is there such a thing as the ideal society, the kingdom of freedom, final and unattainable? Are these two ideals one? These questions about ideal or final ends are tanta­lizing. Some serious reviewers of Popper’s The Open Society and Its Ene­mies have felt that it is most regrettable that Popper does not address himself to these questions of ideal or final ends. This may mean the desire to have these questions answered, coupled with the compliment to Popper as the man who is more likely than others to be able to satisfy this desire.

Personally, I do not think Popper meant to answer these questions. At the same time we may want to extrapolate possible answers to them in order to better be able to criticize Popper’s theory of science. Is Popper’s theory of science criticizable? If not, should we endorse it, should we apply it and, if so, where; or should we strengthen it so as to render it criticiz­able? If it is criticizable, in what manner? This may hinge on the unan­swered problems about ideal ends or final ends.

The dualism of facts and decisions is reflected in conventions or insti­tutions: these are agreements of dual characters. We may agree about ends, and we may agree about means to execute them; in these cases the agreements are of decisions, different people come together to make similar decisions. But the question may be, what are the available means to choose from, or even what is the problem of choice. Here we have to form opinions; and again we have to do so in agreement in order to form a convention. It is obvious that we can criticize the factual element of a convention much in the way we can criticize any factual contention. Pop­per states ([Logic], 38, 83, 107) that there is a factual element in the con­ventions of science, namely the claim that following these may be fruitful in some definite sense of fruitfulness.

Somehow, to a reader familiar with Popper’s exciting Logic of Scientific Discovery, the very mundane quality of my example may sound very strange and awkward. Philosophers of science seldom speak of a lingua franca, even when discussing the metaphysical tradition on the topic (from Leibniz to Frege, Russell, and Wittgenstein); because the topic sounds either too metaphysical for ordinary condiserations, or all too ordinary for even rather ordinary philosophers. Where, between the bluntly metaphysical and the bluntly empirical does the philosopher stand? What kinds of conventions does Popper discuss? Ordinary? Meta­physical? Say, ordinary; which institutions in Britain, in France, in the U.S., guarantee scientific criticisms? Which of these is more effective? Can they be improved? Are there direct incentives in these countries for criticisms? Does the Nobel Prize Committee consult a Popper-like hand­book? For many years I puzzled as to whether Popper’s statements con­cerning conventions are descriptive, in the way in which reports made by an anthropologist describe conventions in a given society (his or alien), or whether they are proposals, put in the way a member of a commonwealth (of learning) presents a proposed reform. Examinations of the relevant texts indicate to me that Popper is systematically ambiguous; I cannot say whether or not his view is descriptive or prescriptive. And so I cannot say whether or not the factual element in his view of the conventions of science is refutable, say, by historical examples.

Yet the problem becomes more difficult when we ask, should the aim of science be the search for the truth? And then, even when we agree about the ends, and the circumstances, and the most effective means, we can still raise further questions and ask whether the sole criterion forjudging the organizations or the conventions of science is the alleged maximal efficiency with which they allow us to approach these ends: we may have - we do have - conflicting ends and competing institutions.

I think it is clear that Popper does assume that there are ideal ends, but that he balks at the suggestion that there are ideal conventions to further them. Perhaps the very intertwining, then, of science and society, or rather of thought and action, absolves him from answering this question - if it does not destroy this question itself: the science which will outline the ideal society will be ideal knowledge itself (and the individual may be ideal too), and thus will make the very need for institutions to safeguard criticism and to test hypotheses (and individuals) quite unnecessary.

One could put it this way. Proper fallibilism is one which includes human fallibility as an essential ingredient of any picture of mankind. Therefore, any ideal which conflicts with this must be rejected: a properly ideal humanity is not humanity. The ideal of final truths or of final ends, or even of both, is not in conflict with fallibilism. The perfect method cannot exist even as an ideal, because the ideal method ideally is the road to perfection and therefore impossible.³

Thus, however vague and abstract our description of the conventions of science, it cannot be the ideal or even part of it. Hence, if the conventions are described too vaguely to be criticizable, there is no reason to assume that we have captured the core of all reasonable descriptions of the con­ventions of science, the core which must be true. Rather, we may wish to render the description more informative and thus more criticizable. Yet we have to see to it that some intuitive demarcation is retained (after all the increased informativeness), between the conventions of science in general and, say, the rules and regulations of the Royal Society at its foundation. This seems to me to be a strong desideratum. Those who distrust intuition and want a reason for this desideratum, may find it in the fact that the rules of the Royal Society were designed by conscientious inductivists who hardly believed in rules and regulations in science (since inductivism is naturalistic, of course), and in the following way. The con­ventions a philosopher of science describes, be he Poincare, Duhem, or Popper, are definitely not of the same type as the laws of the Royal Society, or else the description is empirically refuted by historical evidence. The conventions described by a philosopher, then, belong to a higher type - like constitutions and canons.

I do not wish to decide here the criticizability and modes of criticiza­bility of constitutions or canons. I only wish to draw attention to the fact that some constitutions and canons do allow for legal reform and even constitutional amendments, while some do not. And here, we can say, the constitution of science which Popper describes - or rather outlines - is the same as the constitution of the open society which he describes - or rather outlines - in the sense that both coincide with his doctrine of modified conventionalism. We cannot, therefore, deny that we observe the existence of a modified conventionalism in operation, much as Kant could not deny the fact of the operation in human beings of a moral sense and of moral enthusiasm. This is, indeed, how even ethics shifts from the domain of one’s conscience to the domain of conventions ([open Society], Ch. 5): its canons of necessity pertain to society.

We see here the operation of modified conventionalism already noted in the previous section. The full force and novelty of this idea seems to me to stand out against the background of an almost universal rejection of the doctrine in the midst of a regular and daily adherence to it practice. In theory we have the old Greek tradition of polarizing nature and conven­tion with its enormous impact on the whole Western tradition in social and legal philosophy, and in the philosophy of science, which still rests on polarizing nature and convention and on viewing nature as truth and convention as entirely arbitrary. This dichotomy, nature versus convention, still occurs both in social philosophy and in the philosophy of science, in the face of all the reforms which regularly occur in both society and science. This may be explained thus: attempts at a scientific philosophy start with canons of justification which impose the dichotomy. Attempts at a scientific social philosophy usually started from either a naturalist or a conventionalist philosophy of science, thus further imposing the same dichotomy on social philosophy (see my [‘Individualism⁹]). It seems to me fairly obvious that this last observation, though partly an interpretation of history and partly an application of Popper’s ideas, is the transforma­tion of an important philosophy into a new ideology, the idealogy of the New Renaissance. Its author is Popper’s former student, W. W. Bartley, III. If my memory serves me right, there was a lecture, broadcast over the University of Illinois radio station in Urbana, Illinois, in 1963 and repeat­ed there in 1964, in which Sir Karl expressed a similar appreciation of Bartley’s broader view of non-justificationism. Unless Popper has re­tained his transcript or his tape, the lecture would seem to be lost.

To repeat, the broad idea offered here is both that justificationism is the moving spirit in the Western tradition (contrary to Popper’s emphatic and repeated claim that it is the critical or skeptical approach), and that traditionally the dichotomy between nature and convention was the ex­pression of justificationism. To repeat, both these components of Bart­ley’s novel - and I do mean novel - philosophy are to be found in a way in Popper’s writings, and even put with emphasis (but not sufficiently consistently). In particular, it was Popper who has shown that certain justificationist doctrines (though the term ‘justificationist’ is Bartley’s) impose the dichotomy on us; that the dichotomy, nature versus conven­tion, is the same as the true versus the false and arbitrary ([Open Society], i, 61; [Conjectures], 18); that for conventionalists the main thrust of their philosophy was their dissociation from naturalism; although they them­selves often balked at the total arbitrariness they found themselves en­dorsing.

In moral philosophy, one of the constant and heroic aspects of Russell’s studies, throughout his long career, was his confession that he could not jettison the dichotomy, but found each of its components highly objec­tionable, the one as too authoritarian, the other as too nihilist. Extreme conventionalism is advocated by quite a few legal philosophers, especially Hans Kelsen and his followers, all of whom find it necessary to qualify it somehow so as to escape total arbitrariness. Yet in legal philosophy it is H. L. A. Hart who first clearly broke away from the dichotomy, playing ethics against convention, and speaking of moralizing our laws in stages (thus employing a higher level convention to improve conventions). In the latest edition of the Encyclopedia of the Social Sciences, in the article on legislation, Benjamin Akzin still wonders at the degree of self-decep­tion which legal philosophy engages in when justifying the law either on the basis of nature or on the basis of convention, but in the face of ever present reforms.

So much for the persistence of the dichotomy in social philosophy. In the philosophy of science the same phenomenon occurs. In his The Value of Science, Henri Poincare finds in simplicity an anchor in non-arbitrari- ness. This has been criticized by Popper ([Logic], Section 46) who argues that for the conventionalist the preference for simplicity is itself merely a convention. “It is curious that conventionalists themselves have over­looked the conventional character of their own fundamental concept”, he adds. “That they must have overlooked it is clear, for otherwise they would have noticed that their appeal to simplicity could never save them from arbitrariness, once they have chosen the way of arbitrary conven­tion.”

This criticism which Popper launches in his Logic of Scientific Discovery clinches the charge that, contrary to accepted desiderata, conventiona­lism presents science as utterly arbitrary. This criticism is echoed in Popper’s Open Society (i, 237), where he draws the parallel between the philosophy of society and of science, and where, in the name of the dualism of facts and decisions, he dismisses conventionalism in science as monism, rather than dualism (since, if theoretical science is utterly arbi­trary, it reflects only decisions and not views on facts). Popper is not very easy to follow on this, and his terminology is not too helpful. We may, perhaps, summarize his view thus. Naturalism, both in science and in society, is the monism of reducing all convention to truth or of disposing of them as falsehoods; conventionalism, both in science and in society, is seeing nothing but arbitrariness with no relation to truth or co facts or to nature. Popper recommends dualism both in science and in society. He accepts both a natural component and a conventional component as irreducible; he recommends criticism to improve the natural element; he recommends that this recommendation of criticism be incorporated in the conventional element (as a higher level convention), and he recom­mends that even the (higher level) conventions concerning criticism be open to criticism and improvement. But no improvement is ever the achie­vement of the ideal of final truth or of final ends. Convention, finally, has no anchor in the truth, even when the concept of truth is broadened to include moral truths; we cannot speak of conventions as false, merely as more or less adequate.

This, then, is as tight an amalgamation of The Open Society and The Logic of Scientific Discovery as I could achieve in my reading which I would expect Popper to endorse. Also, all this points at Popper’s later work, his ‘Three Views Concerning Human Knowledge’, in that it consi­ders institutions as mere instruments, but not of theories, not even of the factual elements embodied in institutions.

The question raised early in this essay is largely answered, then. The conventionalist, disregarding any factual elements in theoretical science, cannot but view theoretical science as purely an instrument. Consequently, for conventionalism the desideratum of a criterion of progress can only be satisfied by a pragmatic criterion. Both Poincare and Duhem expressed their hope that the criterion of simplicity and of usefulness would coin­cide. This is either a pious hope or a theory. And so is Mach’s psychological theory of simplicity as mental economy and an (implicit but quite wide­spread) economic theory of economy as the most useful mental commo­dity. Mach’s psychological theory is patently false, and the economic theory could be refuted even before the day of the computer by showing that novelty, not economy, is the most highly prized mental commodity. Conventionalism, then, leads to instrumentalism, and to an obviously erroneous one at that.

Modified conventionalism, on the other hand, assuming the existence of a factual element (in science and in society), includes the assumption of an unattainable ideal of truth, which latter assumption is what has been labelled ‘modified essentialism’. Modified conventionalism, then, is broader than, and includes, modified essentialism.

But, though I read all this in Popper, and though I would expect him to endorse it, I happen to know of his outright rejection of my reading; nor, on the other hand, can I overlook passages in Popper which conflict with my reading and which I find no reason to accept. Anyway, my concern is not with acceptance or rejection: I follow Faraday here, not to mention the gentle Rabbi IJelbo, in considering proselytizing as not the worthiest of occupations.

III.