POSITIVE EVIDENCE AS A SOCIAL INSTITUTION

The current literature in methodology and epistemology is almost exclu­sively devoted to one topic: positive evidence, or favourable evidence, or empirical support, or harmony between theory and experience, con­firmation, etc.

Once we take institutionalized theories as if a priori true, then many theories can be verified as if conclusively. Philosophers who observe veri­fications are considered naive, and naive they are; but they do observe verifications - in applied science and in technology; and even in historical research. When our theoretical framework entails a limited number of empirical possibilities we can perform eliminative induction and verify an hypothesis by eliminating all alternatives to it. When the verified possibili­ty is then refuted as well, so does its a priori framework: this is how ap­plied science can aid pure science. Pure science aids applied science by allowing its tentative ideas, when sufficiently corroborated by some social­ly determined standards of sufficiency, to become the a priori framework to be taken for granted. The standard of corroboration is not Popper’s, since the corroborated theory remains accepted even after its refutation, whereas for Popper the degree of corroboration of a refuted theory is negative; nor is the standard of corroboration logical probability, need I say, since logical probability precludes the Popperian idea of positive evidence as being nothing short of failed negative evidence. Rather, a corroborated theory is one which is highly probable relative to recognized institutionalized theory and in the light of evidence procured under test conditions. Hence, though the test is scientific, corroboration is of mere technological value, but of no theoretical value. For pure science some theory of verisimilitude looks, at first blush at least, more promising.

I

The current literature seems to echo the following ideas which are tradi­tional since the founding of the Royal Society.

(1) Enlightenment is rational belief, belief which can be justified ob­jectively.

(2) Science alone is rational. In opposition to science we have over­belief and under-belief - credulity and undue skepticism.

(3) Scientists practice what they preach. Rational technology is the applied science.

Since point (2) declares science and only science to be rational the sense of point (1), point (3), concerning rational technology being scientific, is a corollary to points (1) and (2). All three points may be viewed as defini­tions, but then the traditional view may be reduced to the proposition that the definitions are not empty: the traditional theory will read, there exist known instances or paradigms of rationality in science and in tech­nology as so defined, and no cases of rationality besides the ones which fit the definition.

Of all the criticisms of the traditional views I prefer the one from the broad - and shallow - outline of intellectual history. The argument (which I have discussed in Chapter 9) rests on the observation that general intel­lectual frameworks contain general metaphysical presuppositions shared by scientists and other reasonable men, yet which alter in time. Some in- ductivist philosophers have noticed this and have modified inductivism accordingly to be the claim that induction is valid only with respect to such frameworks. Though this view goes back to Descartes and to Kant, it seems that Frederick L. Will is the originator of this view in our century. The great advantage of this view, of course, is that it solves the problem of induction within deductive logic. The slight disadvantage of this view is that it reinforces the adherence to an existing framework up to the very last moment of its existence, and the switch to the new framework im­mediately after it replaces the old one. But who should demolish the old view and who should erect the new one? If what Will recommends is the rational thing to do, then all our big medals have gone to the brave irra- tionalists.

ii

Popper has replaced the received standard of rationality by an alternative standard:

(1') Learning from experience is learning from facts to locate our mis­takes. There is no rational justification of belief, but only rational justifica­tion of ^belief.

(2') Science is rational in the sense that it is critical, in the sense that it consists of refutable conjectures plus the empirical refutations of some of them.

(3') Technology often employes rejected doctrines (Newton’s, Max­well’s) and sheer rules of thumb (concerning safety margins and procedures not amenable as yet to any scientific explanation).

To this I wish to add the following:

(4) Empirical support to a theory is a necessary condition for its ap­plicability, legally imposed by bureaux of standards, or by food and drug offices, or by trade and industry offices and their like. Empirical support is never a sufficient reason for application. Empirical support never forces us to apply a theory; it only permits us to apply a theory if we are ready to take the further risk. The theories we are normally forced to accept are those which are accepted by leading reasonable people in our society - as the law specifies in many cases of bona fide mistakes.

Popper has devoted much of his time and energy to the criticism of the current doctrines. Again and again he tries to appeal to the rationality of the multitude of philosophers by devising newer and easier refutations of their views - but to little or no avail. I consider it rather unreasonable to act so instead of reason out an obvious set of problems: Need one bother with the current doctrine? If so, why? Need one bother with it merely because it is current? Why? What are the best means available to one decided to demolish a philosophical fashion? It seems to me that these problems are not particularly interesting, though they are on occasion practically important; intellectually it seems preferable to pursue, in re­lative solitude if need be, certain problems raised by the best available ideas, but social and educational considerations may put on us some demands.

As a sort of compromise between the intellectual and the practical aims, when these compete for our time, we may choose the middle ground. We may persue certain problems of some interest and not little usefulness. In the present case I suggest the following. What, if any, is the role of positive evidence in pure science? How can one explain the prevalence of the fashionable predileliction for positive evidence? In particular, can the traditional theory of rationality (point (1) above) be construed as a first approximation to a better theory? Why is there a semblance of universal agreement in science? If science thrives on dissent and criticism, why is dissent played down and dissenters disparaged rather then encouraged? To what extent is research handicapped by dogma, superstition, and prejudice, be it Aristotelianism, Marxism, inductivism, instrumentalism, or Popperianism? Finally, what is the role of negative and of positive evidence in technology, in applied science, and in fundamental research?

There are two peculiar aspects to these Popperian problems: they have an unmistaken philosophical flavour, yet they lend themselves to both empirical investigation and (social) technological application. The empirical study of the actual role of evidence is almost entirely a virgin field, and as far as the role of evidence in the commonwealth of learning is concerned, in- ductivist preconceived notions have precluded all empirical investigation of it. Even well-known facts have been constantly disregarded. Boyle, Duhem, Popper, and others, have assumed that in legal practice evidence is free of all philosophical problems. This seems to be obviously false, as Clarence Darrow knew and as many expert-witnesses can report. Also, the problem of the permissibility of using background information when dis­cussing positive evidence is common to both philosophy and jurisprudence, but the twain have not yet met. The application of probabilities to linking testimonies to specific cases is extremely rare: the law often demands of witnesses to bridge gaps precariously rather than have their testimony and use experts to make positive identifications.

in

Let us begin with the role of positive evidence in pure science. Is positive evidence in pure science of any significance? Why is it considered extreme­ly important? According to Popper, science consists of series of bold con­jectures and the empirical refutations of some of them. If so, science could, in principle, be gloriously successful with no failure of any test, i.e., with no positive evidence. Somehow this does not ring true. Popper himself has modified his own view, declaring it desirable to have some of our conjectures corroborated before they be refuted. Empirical corrobora­tion, he says, is a sign of progress, of having come an iota nearer to the truth. This claim cannot be taken seriously. First, if corroboration plays an indispensable role, more than such a passing remark about this role is necessary. Second, Popper’s demarcation of science must be altered so as to make room for any substantial role corroboration is alleged to play in the development of science. Thus, Poincare’s refuted but never corro­borated theory of radioactivity may be unscientific according to Popper’s new view of science as conjectures and corroborations and refutations.

Popper’s characterization of the empirical character of science as em­pirical refutability is very satisfying first as the explanation of what we view as pseudo-scientific, as pseudo-empirical, namely as empirically irre­futable; second, it is part of a modern version of critical realism (“I err, therefore the objective world exists”). His restriction of all empirical character to science, however, seems objectionable, since empirical charac­ter may occasionally accompany theories traditionally viewed as para­digms of nonscience, this including superstitions (e.g., spiritualistic claims empirically refuted by Faraday), and prescientific empirical technology, and folk-medicine, metaphysical doctrines (e.g. Thales’ theory that all is water, refuted by Cavendish and Watt), and epistemological doctrines (e.g., the theory - once universally believed - of the finality or utter veri­fication of scientific doctrines, especially Newton’s, refuted by the Ein- steinian overthrow of Newtonianism), large chunks of value theory (such as that desirability and undesirability always resolve into final verdicts making each item either desirable or not) and much common sense (es­pecially concerning sex and concerning race, past and present). Popper himself has sometimes characterized science as the set of all refutable con­jectures, sometimes as the set of all explanatory refutable conjectures. He seems to consider these two characterizations as coextensive, but they are not. The latter is preferable not only as being the narrower but also as characterizing science by its aim, namely, comprehension or explanation (preferably true). Viewing science chiefly as purposive or as goal-directed activity is much more in accord with the sociological approach favoured by Popper, namely the rationality principle, or situational logic, or the principle of rational reconstruction. The word ‘rationality’ in the sense of purposiveness should not be confused with ‘rationality’ in the sense of enlightenment. Assuming, then, the purpose of science to be enlighten­ment, i.e., true explanation, we may try to explain the behaviour of men of science as the search for the true explanation through the testing of the best explanations we have.

Now, assuming that scientific theory is refutable explanation, our esteem of a given theory may be high if it is highly refutable and/or if it is highly explanatory. Popper speaks mainly of a high degree of refutability. Re­futability, however, is rather a practical requirement, and practically the ability to construct one test to a given theory suffices to keep us busy for a while - provided the theory is valuable; the more a theory explains the more it fits our purpose, and if we are ambitious we try to test it first. (Einstein was encouraged when he discovered the high explanatory power of his own theory of gravitation.) If the test fails, then the failed refuta­tion increases the value of the tested theory in that it increases by one the number of observed facts which it explains. This is why the empirical support of an admittedly false theory which, however, explains a sur­prising number of facts, may still be quite significant; e.g., the empirical support of Bohr’s model of the atom. By Popper’s theory of corrobora­tion, as Bohr’s theory was refuted it could never be corroborated. Hence, Popper’s theory of corroboration is empirically refuted. A better theory of corroboration should take account of the desiderata in Bohr’s case - not to refute a theory but to use it as a stepping stone for the construction of a better theory. Thus, specific cases may have specific ends on top of the general ones.

Ideally, then, in pure science we would like to invent wide explanations, repeatedly support the true ones and very quickly refute the false ones. But we can find empirical support of false ones, so that support is no cri­terion of truth; we may still try to eliminate false explanations, in the hope of progressing towards the truth, and the wider an explanation the more seriously we shall take it and hence the more anxious we shall be to examine it. We may, in the interim, try to find out as many positive and as many negative instances to a given theory - until we discover a more in­teresting and challenging theory.

IV

And now to applied science. To a very large extent applied science receives its goals ready-made from technology (though here operational research is an exception and not such a simple case). In technology, our aims are to devise the best ways of doing things. We wish to devise new ways, and verify our hopes and refute our fears concerning them. This cannot be done. Alternatively, we would like to verify our realizable hopes, and refute the vain ones; we would like to verify our fears which have grounds in fact, and refute the others. Even this is too ideal. Yet, it is already suf­ficient to show the difference between pure science and technology - or even between pure and applied science: in pure science positive evidence in favour of an error is the failure to eliminate it; for technology the same may be success - if the evidence in question happens to be useful. In pure science, being right in prediction for the wrong reasons is not very desir­able: in applied science being right for any reason may do very well, if the case is useful for technology; hence, in applied science things are not as clear-cut as in pure science.

Existing techniques of research are embedded within given theoretical frameworks. Within such frameworks conclusive verification is often pos­sible. If such a verification turns out to be inconclusive after all, the frame­work within which it occurs has thereby been transcended. In such a case, applied science or technology gives up the problem as temporarily in­soluble, perhaps to be relegated to pure science or to fundamental re­search. Pure scientists may be dissappointed when a phenomenon they are studying is shown to be explicable within the existing framework which they wish to transcend (e.g., Faraday’s explanation of Chladni’s figures within classical mechanics, contrary to Oersted’s hopes). By contrast, ap­plied scientists do not wish to transcend but to employ more and more the existing hitherto successful framework. And so they may be very happy to explain within the framework a hitherto unexplained phenom­enon, (e.g., colloidal chemistry, much biochemistry, even genetics); the same holds for the technologists, especially when they explain success­fully an undesired feature of a process, (e.g, borings in internal combustion engines due to rapid combustion, leading to the use of delaying additives), or a side-effect (in pharmacology often due to impurity), which they can thus hope to eliminate. If the explanation is verified, the undesired feature is often on the way to being removed. But such research is confined to the existing theoretical framework. When the limit of the existing framework is reached the pure scientist may become very interested since the im­provement of the existing theoretical framework is his concern. Examples of this kind are rare, but they exemplify the possible contribution of ap­plied science to pure science; the failure to pump water above eleven yards was an interesting scientific discovery; similarly the failure to create controlled nuclear fusion was an interesting discovery in plasma physics. Examples of locating technical failures for the purely technological pur­pose of eliminating them may occasionally be more technologically im­portant then scientifically interesting; e.g., the discovery of a certain cause of the standard failure to grow bacterial culture (every now and then cultures were known to perish) which quickly was instrumental in creating penicillin. But whereas the successful elimination of a failure may be scientifically interesting, the failure to eliminate some failure should be of interest for pure science if and when it is a scientific refutation. That is, when we have scientific reason to eliminate a failure, yet fail to eliminate it, then the technologist calls in the scientist.

So far reference was made to existing techniques and to their improve­ments within the existing theoretical frameworks or even on the limits of such frameworks. Yet the existing theoretical framework is something not sufficiently clear and definite. Moreover, the existing framework is irrelevant for too many technically important questions which technolog­ical research may pursue by trial and error or rules of thumb. Once a solu­tion to such a question is backed by facts it may be incorporated within the existing framework and thus vastly improved within a surprisingly- short time. But even then the incorporation is done very vaguely and into a rather vaguely defined framework. Consequently, the risk involved in such implementation of an innovation is highly problematic. The prob­lems are varied; they may be specific or general, social or epistemological. These problems are unstudied because of the prevalent mythology which views anyone alert to them as a Luddite or at least a suspect.

V

I now come to the problem of the fashion favouring positive evidence. I shall link it first to a much deeper problem. Primitive people personally depend on society but economically are quite autonomous. In the West, economic autonomy has been ousted by division of labour, but personal autonomy is increasing. Here are conflicting tendencies, already discussed by Russell and others long ago: economic interdependence calls for more conventional coordination, personal independence for less. How are these reconciled? Even though technology reduces our working-time it increases our need to coordinate. How is this need gratified? I suggest that the coordination is best achieved by an accepted common theoretical frame­work which we all agree to follow socially while considering ourselves utterly free to disbelieve it personally: we call this conventionalims, frank and open hypocracy, etc.; but we cannot shake it off. The framework is institutionalized and adhered to not because it is true but in order to coordinate wherever coordination is needed, whether in the theatre, in railroad time-tables, or in industry. We may try to reform the institu­tionalized theoretical framework, especially when this framework permits too much stupidity and immorality. But the western idea of personal free­dom does not go so far as to permit people to act exclusively on their own best judgment or on the best scientific opinion of the day. The claim for an individually determined rationality, for subjective probability of the de Finetti and Savage school, is an epistemological fiction: it was invented in the eighteenth century as an ideal of a totally rational society - or, to be precise, not society at all but a bunch of rational individuals with clear- cut interests. In real circumstances, institutionalized public opinion must be considered as well, even when deemed mistaken. It is no accident that current epistemology shies away from its own social implications: these are rather ludicrous. In only one point is there deviation from all this: somewhere the leading philosopher Carnap agrees with Popper the heretic that we conventionally accept corroborations and stop testing a theory: hence, induction is much a matter of convention!

Let us examine the risks taken in the stopping of a test prematurely - logically all stopping of a test is premature - and the implementation of an innovation based on the corroborated theory. These risks are socially, not personally, determined; they differ in different Western countries. Again, remember thalidomide. Psychologically, a test-pilot may feel confident enough to fly a newly designed airplane, but the law and the underwriters are there to prevent him from so doing until certain ground-tests are accom­plished. If the ground-tests are not successful, there will be no test-flights; if the ground-tests are successful they still do not insure safe flights, and therefore test-flights are still necessary prior to determining a plane’s airwor­thiness, and that is why test-pilots take risks when they fly on test-missions.

Thus, the high degree of permitted risk is regulated institutionally. So must be the low degree of permitted risk, or else the spirit of adventure may decline. An underconfident test-pilot, or even an underconfident aircraft company, may decline taking risks; social institutions exist to insure that the underconfident will be replaced by the more adventurous. One method of securing this is institutionalized competition, whether be­tween various military powers, various entrepreneurs, technologists, medical research-workers, or test-pilots, who covet prizes, medals, etc. Failure to undertake a risk may involve military inferiority, losing markets or new investments, etc. A competitor may consider his loss or gain in making the decision to pursue a venture or not, depending on whether the venture is going to be successful or not. Another method is that of in­surance, and insurance often depends on such considerations and esti­mates too. These considerations and estimates are made within a theo­retical framework and the detailed theory of them is due to Abraham Wald. The framework is normally a composite set of assumptions, party institutionalized, partly not. It is usually very reasonable to work on the institutionalized theory because failure in such cases enables one to fall back on one's society as a victim of uncontrollable circumstances. (A simple example: relying on a false railway timetable is more excusable an error than relying on one’s own false judgment on the same matter even if one’s own judgment is usually better than the railways’ own timetable!) This makes the adherence to institutionalized theory a form of insurance. (One who does only what is expected from one in one’s station is safer than one who shows initiative - even where initiative is recommended!) Those not able to draw this kind of insurance benefit, be they derelicts, outcasts, or spies, or even simply highly independent people (free lance or academics), are better off when they get rid of as much of their theoretical background as soon as they can. (This is what Popper ignores when he invests in efforts to convince his opponents.)

Hence, the following of institutionalized theoretical frameworks is al­ways a form of insurance (particularly when the framework includes in­surance methods of all sorts). Thus the framework operates as pooling risks. This hangs one person’s and one institution’s safety and stability on the safety and stability of the whole society. This is particularly useful for reform: we make a new and weak institution depend on old established ones. The question is, doubtlessly, does not this process of minimizing individual risks lead to the increased risk of destruction of the whole society?

Since the calculated risk of any venture is considered only within a theoretical framework, the risk of holding on to that framework is in­calculable. Nor is the risk calculable of changing it for a better one rather than patching it up ad hoc. Those who dreamt that scientific advancement on the whole decreases the risks of mankind as a whole were forced to reconsider their views after science created the new possibilities of global destruction (by population explosion, radioactive fallout, etc.) These risks are known to us, but there may be risks not known to us, as there were for our forefathers. Global risks, known or unknown, are not insurable, since insurance is merely the pooling or sharing of risks and the diversi­fication of resources. In a catastrophy there are not enough resources to compensate everyone. Progress permits further diversifications of re­sources, but it may lead to new global risks. The question, which of these two will win in the end (e.g., will space colonization precede nuclear de­struction?) is handled by science fiction. It belongs to science fiction since it essentially transcends our knowledge.

Our ignorance of some possible risks does not permit us to take known risks lightly, and, at least in this respect, it is not sufficient for us to criticize the existing institutionalized theory or even to offer an alternative to it. The existing institutionalized theory at least works, and the new one has to be carefully examined and verified on a small scale before it is institutionally accepted.

Here we see that testing in pure science is rooted in the purpose of science, the scientific search for truth, but testing in applied science and in technology in general is rooted in the institution fostering competition: whatever is not prone to competition is taken on faith. Why does this ostrich policy succeed? This is a problem for pure science of a Darwinian stock.

All this may easily be tested as well as applied to cases of retention and replacement of frameworks which are intellectual and institutional at one and the same time. I shall offer only one case - which is not much interest­ing but which is of enormous practical possibilities.

Popper’s philosophy comes to replace a philosophy which was institu­tional for three centuries, and very successful it was. Popper’s theory can improve matters by making controversy more openly the order of the day, by cutting out dead wood and accretions. But it will be unreasonable to expect it to become the institutional theory prior to its verification on a small scale in some pilot plants, such as workshops of science and tech­nology. This is why philosophers are so hesitant to consider Popper’s opinions freely: they fear to deviate too much from the institutionalized opinion. If and when scientists and technologists institutionally endorse Popperism, the philosophical fashion will follow suit all too quickly. Hence, Popper is mistaken in his efforts to dissuade people from their old fashioned theories merely by an intellectual appeal. It is more useful to try and create pilot plants, to work out wind-tunnels for ground tests prior to testflights leading to institutionalized flights of the imagination. But there are different ways of institutionalizing a philosophy. In par­ticular, it is easy to declare a philosophy acceptable by the public but only after incarcerating and sterilizing it and destroying all its possible social implications. I can hear the strains of the approaching bandwagon, and can only express my hope that when it passes by, not all philosophers of science will be adding their weight to the procession.

appendix: the logic of technological development

There is psychological evidence against the suggestion that we can assess the credibility of a theory and then adjust our credence in it so as to make its credence equal its credibility. There is also psychological evidence against the suggestion that credence is an additive measure in any way. There is also historical evidence that much of scientific progress heavily depended on some metaphysical faith in is some new incredible ideas. And so the psychologists theories of scientific assurance must be rejected. In many cases where tests are obviously too risky, even permission to test is conditioned by the prior procuring of positive evidence. The most obvious instance of this is the use of laboratory animals in medicine. This, however, is not to justify any of these procedures, rather these procedures are legal (stipulated) codes of justification, themselves unjustified (and unjustifiable), and perhaps even harmful.

This involvement plays a crucial role also in ventures into the unknown, and of insurance in technological innovation. The specific rules governing ventures are related both to involvement and to the use of positive evidence in insurance practice (e.g., the insurance of a test-flight is void if certain procedures of ground tests are waived). The rules, thus, insure stability by localizing and minimizing risks of new ventures. To proceed further one would have to fall back on a theory of social stability and planned social change, and, more specifically, of the balance between stability and planned change (including desirable and permissible rates of change, etc.).

Popper has developed a theory of social progress akin to his theory of scientific progress - it is the outcome of negative evidence. This does not yet explain social stability. For this we need add his theory of social in­stitutions as interpersonal means of coordination and as accepted by con­vention. And here positive evidence may be the conventional, defective as it may be, means of checking institutional change: the higher the standards of required positive evidence leading to change, the more ‘static’ the society in question. But a ‘static’ society is not necessarily stable in the sense of durable. (Bellarmine applied the highest standards to Galileo’s proposals, thus keeping the Church ‘static’, but leading to its loss of hege­mony in the world of learning.)

A theory of social stability and of assurance in general can in this man­ner be incorporated into Popper’s philosophical framework, then, with little or no alteration. Within this theory a theory of the role of positive evidence in implementing technological (physical, biological, social) in­novations or reforms can be developed on lines adumbrated here (and in some of my other papers on the topic). Briefly, the guiding line is this: whereas assurance is no guarantee of success, it is a prerequisite, a social condition for permission for trial. The current inductive confirmation theory presents the acceptance of confirmation or scientific assurance - and hence also the actions based on confirmed theories - as obligatory for all rational beings. By distinction, the theory adumbrated here is that positive evidence should be viewed as a permit, in some cases a permit necessary by stipulation: for many types of novelty, positive evidence prior to im­plementation is required by laws; hence the case is open both to empirical study and to criticism and improvement. Whereas the obligation which the current confirmation theory posits is psychological (pertaining to the individual’s rationality) and somewhat elusive (to say the least), the license of the present theory is conventional and hence can be examined by being applied to social cases, e.g., with food and drug regulations, public health regulations, binding testing procedures (in medicine and certain industries) or patent laws. This, then, opens up the possibility of comparative soci­ology of assurance or of positive evidence. Assurance in the psychological sense, then, need not abide by rules of propriety; but socially, propriety of assurance depends on severe test not resulting with negative evidence. Propriety, then, is no guarantee of success, but a mark of accepted degress of responsibility; this is in accord with common sense.

The principle, “there is no insurance against epidemics”, when applied to the insurance provided to social institutions (rather than to an individ­ual or even to a subsidiary institution, as usual) amounts to the rule “don’t put all yours eggs in one basket”, which is the idea of interdependence of social institutions, which is a version of generalized functionalism. The guarantees which institutions - insurance companies or governments alike - can offer, are valid only within the social framework: they do not guarantee the viability of the social framework itself. The social frame­work, thus, plays a similar role in technology as the current scientific framework in scientific research. For all we know both may collapse; but, also, both are susceptible to improvement through criticism. (This is a combination of Kantian and Popperian principles.)

The distinct role which positive evidence plays, then, is in the field of social implementation of technological innovations and alterations (phy­sical, biological, and social). A particular case of this is that of social changes relating to the implementation and improvement of the institu­tions designed to aid the advancement of learning; and these include both current research techniques and current scientific ideas. If so, Popper’s proposals for the social reforms of research conventions will not be widely implemented prior to tests leading to positive evidence sufficient by current standards. This may be achieved by instituting pilot-plants and their likes.

The usphot of all this is that a sharp dichotomy is introduced between science whose chief role is to provide testable explanations and eliminate the worst of these, and technology whose chief role is to introduce prac­tical proposals, eliminate the worst of these, and volunteer implementa-

tion of the uneliminated. Whereas standards of criticism in science may be raised as much as it is within our reach, one can easily overdo the stan­dards of assurance necessary prior to implementation. This may lead to stagnation or sluggishness which, in a competitive world both military and commercial, may be suicidal. It is the competitiveness, built into the social system, which both guarantees that positive evidence should be in accord with the Whewell-Popper theory and that not ‘too’ much positive evidence be required prior to implementation. These two guarantees are evidently in conflict which may be studied, regulated, perhaps even mini­mized, but only after critical procedures are institutionally reformed, as mentioned above.