I.C. Jarvie

York University, Toronto

All rights reserved by the author. This file may be downloaded and/or printed only for the purposes of individual private study. The making of multiple copies, either directly or indirectly, is forbidden.

Distinguished professor of philosophy
Faculty of Arts,
York University
4700 Keele St. North York
Ontario, M3J1P3, Canada

It might indeed be said that the majority of the problems of theoretical philosophy, and the most interesting ones, can be re-interpreted...as problems of method.
- Popper (LScD, p. 56)

Insoluble difficulties in the 'pure' logic of science led Popper to propose, in 1935, that we treat science as a social institution constituted by a set of rules, or methodology, that furthers certain aims. Popper's subsequent work is consistent with this 'social' reading, but he nowhere makes it explicit. Virtually all of his critics (and some of his followers) overlook this decisive shift from the logical to the social, and as a result simply fail to appreciate this most original feature of his thought.
Although I find a 'social turn' at the centre of Popper's thought, it needs stressing that his concerns were first and always philosophical. The epigraph of this paper makes a bold and ambitious philosophical claim. So far as I know, Popper never portrayed himself as having put forward a social view of science. This is fair enough. The social turn we shall find in his work was not followed up by empirical work. Popper never carried out empirical study of the workings of the institutions of science, nor did he write a scientific biography. Indeed, it would be fair to say that he has not even done any historical studies of science, at least ones which involved the study of primary documents. He reframes his problem as methodological, and methodology is a set of social conventions. Thus it also is to be stressed that Popper does not reduce science to being just another social institution. It is a social institution, but a very special one in both its aims (transcendent truth) and its results (universal scientific knowledge). There must be no confusion between Popper's 'sociologism' and that of "the Strong Programme in the Sociology of Knowledge". Nonetheless, the thesis of this chapter is that Popper is a major theorist of the sociology of science: The Logic of Scientific Discovery contains many profound and original ideas on the social character and constitution of science, and, more to our purpose, those social ideas on science contain the germ of his later, very influential ideas on society and politics. In other words, I shall argue that Popper did not philosophise about science and then turn to apply that philosophy to social and political thought. He was thinking socially from the beginning.

1. The Received View of Popper on Science

To begin with, it may be helpful to outline the main ideas of The Logic of Scientific Discovery as they are usually described. The work is in two parts. In Part I the ideas are sketched, in Part II they are worked out and defended in detail as a series of answers to objections.
The book focuses on two problems which are labelled 'the problem of induction' and 'the problem of demarcation'.
The problem of induction is that of how to gain general theoretical knowledge from experience. Experience always comes in particulars: we observe something about this, that or the other. Knowledge, by contrast, is general, even universal. Logic allows that no number of particular statements describing the experience of observing white swans shall permit the deduction of a general statement of scientific knowledge such as 'all swans are white'. How, then, do we get from particular experience to universal knowledge?
The problem of demarcation is that of how sharply to differentiate science from non-science, including pre-science, pseudo-science, folk wisdom, and metaphysics, as well as logic and mathematics.
Since, for Popper, knowledge presupposes language and must be formulated in statements, he is able to reformulate the two problems more crisply: the problem of induction becomes, how can we reach the general statements of scientific law when the statements reporting our experience are necessarily limited and particular; the problem of demarcation becomes, how can we demarcate scientific statements from other statements?
If we accept his autobiographical account, Popper had been engaged in thinking about these issues since 1919, when he was 17.1 The first problem he says he solved fairly quickly: the relationship of scientific statements to statements of experience was deductive, not inductive. Logic allowed particular statements of experience to refute proposed universal scientific statements, and that, it turned out, was sufficient for the task at hand. Provided one did not expect science to consist of certain, or proved, knowledge, it was enough to have general statements that had survived attempts at refutation: they were not supported or recommended by anything other than the fact that they had so far survived whatever challenges experience had thrown at them. This made them the most qualified candidates for consideration as scientific knowledge.
Popper relates that it took him some years to realise that the two problems of induction and demarcation were connected, and that the demarcation problem was the deeper of the two: induction was one solution to the demarcation problem, deduction another. The solution to the demarcation problem that Popper poses and elaborates is that scientific statements be characterised as general statements which are falsifiable. Those already falsified belong to the history of science. Those not yet falsified are current science. Statements that are not falsifiable come in more than one type: there are the statements of logic and mathematics, which are tautologous truths; there are general statements that sound empirical but which nothing could falsify, these are metaphysical.
The reason science is best characterised as consisting of falsifiable general statements has to do with the manner in which science permits us to learn from experience. Puzzled by some logical conflict between our received ideas and statements capturing the evidence of experience, we seek a resolution. Three choices confront us. Either the report of experience is faulty;if not that, then the logical claim to detect a contradiction is faulty; if neither of those, then the received ideas are faulty. Thus is science the critical engagement of ideas with experience. It is the process of culling from among the ideas we have those that are worth criticising by the check2 of experience. The check of experience is not easy to meet. We need to guard against formulating our ideas in a vague, weak or evasive manner, for these make empirical criticism difficult, even impossible. To foster the confrontation of ideas with experience Popper makes a revolutionary proposal.
Sir Francis Bacon, founder of the view that induction is what characterises scientific method, saw clearly that adoption of the correct method was not, in itself, sufficient. The impulse to find what we want, to observe what is not there, is overwhelming. The mind is filled with what he termed 'prejudices'. Bacon therefore proposed that scientists must first work on emptying their minds of all preconceptions and curbing the almost irrepressible desire to jump to conclusions. Thus, for Bacon, being a scientist involves a good deal of inner struggle to achieve a certain psychological condition, a condition where the mind is free from prejudice and thus open and receptive to the experience of Nature as it really is. Only in that purified state, Bacon thinks, will the methodological canons of inductive reasoning, patiently pursued, keep one clear of error.
There are two features of Bacon's account that are worth noting here for their contrast with what Popper proposes. One is that Bacon's account is individualistic - the budding scientist struggles alone against his prejudices. The other is that it is psychologistic: bad habits of mind are the trouble; science can only be attained after one has developed good habits of mind. Elsewhere, Bacon did envisage scientific institutions,3 but not in his account of method.

2. Methodology and Social Practice

Popper's revolutionary move is not just to shift the methodological emphasis from induction to deduction - although that is a standard way to put it. Popper goes further: he also rejects Bacon's individualistic and psychologistic approach to readying oneself for science. Prejudices are not to be purged by some sixteenth century version of psychotherapy, but by confrontation with experience in a social setting. Science is a creative endeavour, it is the search for new knowledge, so Bacon's emphasis on the old - psychological state of the discoverer, on how ideas are obtained - is misplaced. What counts is how those ideas stand up to various checks we carry out on them, most importantly, the check of experience (LScD,q7 2). Instead of working on the state of our minds, Popper argues for an institutional approach. As he sees it, the cooperative and social character of scientific research checks our Baconians impulses to see what we want to see, to jump to premature conclusions, and so on.4 He refers to this as the friendly-hostile cooperation of scientists, and, more abstractly, as the intersubjective nature of the scientific enterprise (LScD, p. 44-48). Scientific ideas are formulated in publicly examinable ways, open to being checked by others as well as their proposers. Those others may be motivated in various ways: to find the truth, or by envy and hatred of the authors of the ideas. But the checking takes place within institutionalised methodological rules that are the best we can devise in an imperfect world. These rules do not ensure there will be no error, but they do create incentives to discover and expose error, rather than to evade it and cover it up.
How does Popper get from an inquiry firmly centred in the logic of science to an insight into aspects of scientific method that go beyond the logical into the social? Popper early in his discussion considers three objections to his view (LScD,q7 6). The first objection is that we expect science to deliver positive information, so it is wrong-headed to characterise it in a negative way, by refutability. This objection Popper parries by asserting that he will later show that positive information has a logical connection to refutability (the promise is cashed in LScD,q7q731-46 ).The second objection is to the effect that, in the same way that inductions cannot be verified, no more can falsifications. In rebuttal Popper maintains that there is here an asymmetry in the logical relations. One contradictory evidence statement is sufficient to refute a generalisation; no number of noncontradicting evidence statements is sufficient to establish a generalisation. The third objection is said to be more serious: refutations can be handily evaded by conventionalist stratagems, such as the ad hoc introduction of auxiliary hypotheses or the redefinition of terms, or simply by refusing to look at the contrary evidence. Popper finds this objection logically insuperable as it stands. He writes:

it is impossible to decide, by analysing its logical form, whether a system of statements is a conventional system of irrefutable implicit definitions, or whether it is a system which is empirical in my sense; that is, a refutable system... Only with reference to the method applied to a theoretical system is it at all possible to ask whether we are dealing with a conventionalist or an empirical theory. The only way to avoid conventionalism is by taking a decision: the decision not to apply its methods. (LScD, p. 82.)

Thus Popper accepts a decisive criticism of the very position often attributed to him - what I have termed the received interpretation. This is important because it is recognition of the problem created by the third objection that moves Popper into proposing what I call his social view of scientific method.
The conventionalist objection to the refutability criterion of science cannot be overcome on its substance. If someone chooses to indulge in ad hoc manoeuvres, or to ignore refuting evidence, they commit no logical or factual error. This does not mean their position is unobjectionable. What they do, Popper maintains, is to impoverish science by refusing to submit its claims to the tribunal of experience. A system sheltered in this way can easily and unnoticingly degenerate into metaphysics, that is, untestable statements. Popper thus shifts the discussion from logical matters to questions about choices, aims, policies and their consequences. For those concerned to keep science anchored in experience, Popper suggests adopting a supreme or meta-methodological rule not to avoid falsification (LScD, p.54).
Here, in the shift from purely logical criteria for science to methodology, we find the beginnings of Popper's social view of science. A methodology consists of methodological rules; each rule represents a decision, a choice to act in a certain way; we make these choices, in turn, in order to foster certain aims. They are thus open to discussion. Both the rationale of the choices and whether the choices will in fact foster the desired aims are matters on which there can be reasoned dispute.
Let me spell all this out in a little more detail. If we look closely at the opening pages of The Logic of Scientific Discovery we find that what is proposed is a theory of method. A scientist is stipulated to be someone who puts forward statements or systems of statements and tests them step by step. The task of a logic of scientific discovery, or a logic of knowledge, is to give a logical analysis of this procedure, "that is, to analyse the method of the empirical sciences" (LScD, p. 27). Popper proposes a method of deductive testing of theories ("deductivism") by deriving their consequences and checking these by various means: against each other (are they self-consistent?); by their logical form (are they empirical or tautological?); against other theories (are they consistent with them?); and finally against empirical experience (are they consistent with known and with newly discovered experimental facts?).
Discussing the problem of demarcation, Popper challenges the idea that science is a way of thinking (psychologism), or the method of induction, which is caught in a vicious infinite regress, or indeed that there exists any "natural" boundary to it. Instead, he suggests treating the refutability criterion as a proposal for a suitable agreement or convention to help us to "be able to say of a given system of statements whether or not its closer study is the concern of empirical science" (LScD, p.37). Reasonable discussion of the suitability of a convention is "only possible between parties having some purpose in common" (LScD, p. 37). Parties with some purpose in common refers, of course, to a social group of some sort.
The social group in question seeks knowledge of the world of our experience, the real world. How is its theoretical system to be distinguished from others? By the fact that it is submitted to tests and has stood up to tests. This is a methodological view of what constitutes experience. The two alternatives Popper is discussing are the inductive method and the deductive method. Deductivism, he proposes, admits to empirical science only those systems of statements that can be refuted by experience. However, this idea of falsifiability is insufficient (LScD, pp. 42, 50 and 54), because it is always possible to evade or deny a threatening falsification.
Does this mean that science is simply a subjective choice to proceed in a certain way? Not at all. But Bacon mistakenly identified the problem: objectivity has nothing to do with a psychological state of disinterestedness or of being free of prejudice. Instead, Popper draws on Kant's idea that "objective" refers to a statement that can in principle be understood and tested by anybody. This Popper calls the "inter-subjective testability" of scientific claims. In an important footnote added to the English translation he says "inter-subjective testing is merely a very important aspect of the more general idea of inter-subjective criticism, or in other words, of the idea of mutual rational control by critical discussion" (LScD, p. 44), making explicit reference to his own later works, The Poverty of Historicism, The Open Society and Its Enemies and Postscript to the Logic of Scientific Discovery.
This point about objectivity made, Popper resumes his only slightly veiled sociology. The theory of method, he announces, is concerned with the purely logical relations between scientific statements and with the choice of methods -

decisions about the way in which scientific statements are to be dealt with. These decisions will of course depend in their turn upon the aim which we choose from among a number of possible aims. The decision here proposed for laying down suitable rules for what I call the 'empirical method' is closely connected with my criterion of demarcation: I propose to adopt such rules as will ensure the testability of scientific statements; which is to say, their falsifiability (LScD, p. 49).

What are these rules, why do we need them, can there be a theory of such rules? A methodology turns, he writes, on one's attitude to science. If one is interested in the advancement of science, in its constant revisions and corrections, one will be led to a very different answer than if one takes a narrowly logical and naturalistic view of science. Purely logical means offer no defence against metaphysics or anything else because of the possibility that refutation can be evaded. Instead we have to set up conventions for which we take responsibility: "by what we do with them and what we do to them. Thus I shall try to establish the rules, or if you will the norms, by which the scientist is guided" (LScD, p. 50).
That Popper is thinking institutionally could not, I think, be clearer; all the more puzzling, then, that it has been overlooked. Popper is here proposing that science is to be seen as an interested group that shares an aim and then legislates conventions for itself in order the better to pursue that aim. He does not explicitly say that his view is social, but he offers some analogies with the social institutions of games and of trial by jury, going so far as to refer to "the game of empirical science" (LScD, p. 53) and comparing its rules to the rules of chess. Certainly he seems to be arguing that science is constituted by its rules, as is chess. He also seems to be allowing that the rules of science can be debated, hence they are not immutable. Much the same goes for chess. The rules of chess have evolved and might evolve more. A rule revision would not necessarily make for a new game, especially if the rule was adopted by the International Federation. The fact that, in baseball, the American League permits the designated hitter to substitute for the pitcher and the National League does not, hardly raises serious questions about which league really plays baseball.
But the analogy to games has a flaw: games are relatively frivolous activities, engaged in for recreation and play. Science, by contrast, is the search for knowledge, a rather more weighty matter. Early scientists in the Age of Reason saw their activity as directed towards blowing away the cobwebs of error and superstition from the past, clearing the corridors that eventually led to the enlightenment of mankind. The very success of this project has led to mankind becoming seriously dependent on science and its applications for its economic livelihood and for preserving and extending its longevity. Thus the building of institutional norms or rules for science might better, I suggest, be compared not to a game like chess, but to the creation of dedicated social institutions - universities, for example, or learned societies. This might be especially apt as universities and some learned societies are subordinate institutions within the overall institutional creation we call science.
Popper does not pursue matters in this direction. Although methodology as he conceives of it is clearly institutional, he does not examine any actual scientific institutions and their workings, including the methodological and other norms prescribed. One of the very few other places where Popper brings to the surface the social nature of the view he is developing is the passage where he compares science to the institution of trial by jury (LScD, pp. 109-110).
Popper uses trial by jury to make a number of very important points about the theoretical context of inquiry and the way it directs our approach to the facts. Juries, he notes, decide issues of fact. The questions they are asked to decide will depend on the actual laws in force, and the procedures followed. He gives no examples. He has mind, I conjecture, the decisions built into a legal system that distinguish it from others: for example, 'minor child' may be defined differently in different places, restricting the charges that may be levelled at someone in one jurisdiction rather than the other; English-speaking readers will likely be familiar with a trial system that is adversarial, leading questions being permitted only in certain parts of the examination, and a jury that must sit mute, and not disclose its deliberations. What the jury does is come to a decision about a matter of fact. This Popper compares to scientists deciding to accept a basic statement. From this statement, together with statements about the law, consequences can be deduced (for example, that an accused has or has not committed an offence). Although the trial, and the conduct of the jury, are governed by rules, Popper is at pains to stress that the jury verdict never justifies or gives grounds for the truth of what it asserts it finds. By contrast, he notes, the judge's judgment is expected to be 'reasoned'. If the reasoning is unsound, that is a ground for challenge; no comparable challenge can be made to the substance of the verdict of the jury. Popper here seems to recommend a scientific parallel: the experimentalist experts decide the finding of fact, the basic statements; the presiding scientific community then tries to judge the implications of that finding. The finding of fact is far less frequently disputed than are the implications.
The analogies drawn with chess and jury trial serve to highlight on the one hand the constitutive nature of rules and on the other the embedding of crucial decisions in institutional procedures. The equivalent institution in science to the International Chess Federation, the jury and the judge, are not spelled out. Popper's focus remains on the logic of science as he endeavours to show that falsifiability is a viable criterion of scientific character once embodied in a methodology. In the course of defending the view in LScD he offers many further suggestions for methodological rules. What he does not engage in, as we shall see, is any discussion of the general picture: do the rules come as a set, or can we pick and choose and yet stay with the game of science?; how are decisions made when new rules are offered or modifications to old ones suggested?; is not submission to the rules constitutive of the institution of science broadly conceived?

3. Popper's Suggested Methodological Rules for Science

In the final section of Part I of The Logic of Scientific Discovery Popper discusses "methodological rules as conventions". He proposes a supreme or meta-rule which governs the later rules to be proposed. This reads:

(SR) "the other rules of scientific procedure must be designed in such a way that they do not protect any statement in science against falsification" (LScD, p. 54).

This meta-rule enshrines the falsifiability criterion of demarcation as a control for the whole system of rules. It takes care of the objection that falsification can always be avoided by mandating that it will not be avoided. Thus we see why I said that the received view of Popper is that he characterises science by falsifiability. In fact he finds falsifiability insufficient in itself. It is not self-justifying like a tautological truth. It has to be adopted, by decision, a decision governed by aims. We then impose it upon ourselves as a procedure, a methodology. Popper is no more a "naive falsificationist" than Bacon is a naive inductivist. Each was fully aware of the logical deficiencies of their alleged positions and did their level best to remedy them.
Although he writes of "proceeding systematically", Popper does not go on to write up a complete list of the rules which govern scientific method. He does not say so, but there are good reasons for not doing so. To expect this would be like expecting a legal commentator to specify the complete list of laws. No such list can exist. Law-making and law-reforming are on-going endeavours. Attempts are made from time to time to codify areas of the law, but never the system as a whole. All such codifications need constant maintenance and up-dating. In this respect, a methodology is not like the closed set of rules that constitute chess. Popper begins by proposing two examples, and we shall be able to extract others from later in the book. The first example is:

(R1) "The game of science is, in principle, without end. He who decides one day that scientific statements do not call for any further test, and that they can be regarded as finally verified, retires from the game" (LScD, p. 53).

This rule directly attacks those philosophers who, like Bacon, envisage science as eventually resulting in a body of finally verified truths. Such spokespeople are a rare breed today, sixty years into the Popperian age, but they are not yet extinct. It also raises a fundamental question against all those physicists, the most recent of whom is Steven Weinberg, who think science might culminate on one grand, final, Theory of Everything.

The second example is:

(R2) "Once a hypothesis has been proposed and tested, and has proved its mettle, it may not be allowed to drop out without 'good reason'" (LScD, p. 53-54).

Among the good reasons for a hypothesis being allowed to drop out of science are its replacement by one better testable, or the falsification of one of its consequences. This rule clearly addresses what is sometimes known as the 'stability of science'. The question is raised, if science is simply the set of unrefuted hypotheses, are we free to pick and choose among them? This rule suggests that no, once a hypotheses has achieved a certain status it can only lose that status in an orderly procedure. Although the content of science changes, sometimes quite frequently, science comes as a package; the institution treats picking and choosing as unscientific. At least, in principle it does; in practice things are less clear cut.
These examples of rules are nothing at all like the rules of inductive logic that other methodologists have sought for, and they can be debated and revised by the community of scientists, other interested parties, and the community at large, without special technical preparation. Consider the question of the plausibility of (R2). The analogy of law-making in the British Parliament may help. When a Bill passes all the stages and receives Royal Assent it gains a special status - namely Law. It gets printed in statute books, applied in the courts and cited in precedents. However, it is a fact that some laws simply fade away. They may stay on the statute books for centuries and yet not be enforced or even attended to. A good example is the laws against suicide. Does anything comparable happen in science? I suspect it does. That is, I suspect that hypotheses which once proved their mettle sometimes just fade out of sight, out of mind. Not because of the fallibility of memory (Morgan 1985), but because of an entirely new way of looking at things has arisen, and it is not worth bothering to refute each established result one by one - those in the know ignore them, those not in the know go on thinking that these are still current results. A good example from anthropology is the many, many stories from around the globe that Sir James Frazer collected in The Golden Bough. This best seller is still in print, but it would be hard to find anthropologists who could cite chapter and verse of where and when this or that report in it has been falsified. Almost all of them could be falsified in principle, but they may not have been in fact.
More problematic are cases advanced by Agassi and others where a scientist in good standing resists some scientific idea, perhaps even a dominant scientific idea, that appears to have proved its mettle. Sometimes the objection is described as metaphysical, sometimes aesthetic, sometimes it is inarticulate. Is such resistance unscientific? In so far as it motivates attempts to criticise and refute the idea, or to develop an alternative, the answer is no: motivation is irrelevant. In so far as it involves refusal to face facts, to suppress, or fail to acknowledge, then, yes.
The supreme rule (SR) and the two simple rules (R1 and R2) are offered by Popper early in the book as examples of the sorts of rules that can be used to govern science and ensure that its theories are connected to experience. The whole exercise Popper conceives of as his theory of experience. Without denying that characterisation, it is here being represented as a set of putative rules for the constitution and work of the Republic of Science. Nothing like a complete set of such rules exists in Popper's text. On the analogy with the law, mentioned above, this makes sense. But if the rules are to be the Constitution of Science, then they deserve to be spelled out so that prospective members know the dimensions of the regime they are agreeing to be governed by. But it is also possible to view the rules as a kind of common law - unwritten customs long established and only articulated and discussed at problematic moments. But Popper supplies no clear decision procedure in the case of dispute or infraction. Some of the epistemological and other sins Popper criticises and erects rules against are perpetrated by scientists in good standing, not just by philosophers. So how are complaints to be dealt with, how are infractions to be dealt with? It should be remembered that Popper is implicitly embedding the institution of science inside two larger entities. First in an open civil society governed by the rule of law. Thus general matters of criminal conduct and of dishonesty are already sanctioned. The second is that science is an activity for those with time and inclination to want rerum cognoscere causas.5 Putting this another way round, Popper is taking it for granted that there are barriers to entry other than those erected by the necessity of subscribing to the rules he is articulating. Science is an institution or set of institutions embedded amongst other institutions, to some of which it is subordinate.
Popper concentrates mainly on rules of procedure - rules that guide us in how to proceed. But there are a few rules he mentions, like the SR, which in effect define the enterprise altogether. In the course of what follows we will occasionally find Popper offering meta-rules that are more like defining than procedural rules. But before continuing it is good to reemphasise that Popper's book is anything but a systematic treatment. The rules themselves are not always fully formulated, and are often couched as subjunctives. Nevertheless, let us proceed to extract the rest of the rules and thus to build them into a set that we can examine as a whole.
In discussing the metaphysical difficulties involved in ideas about causality, and the dubious standing of principles of causality (such as every event has a cause, no effect without a cause) Popper proposes to cut through them with a rule. The rule does not dogmatise about cause, but instead rather enjoins us.

(R3) "[We] are not to abandon the search for universal laws and for a coherent theoretical system, nor ever give up our attempts to explain causally any kind of event we can describe" (LScD, p. 61).

The resemblance of this rule to (SR) should be apparent. The implication is that those who do give up the search for universal laws or cease their attempts to explain events we can describe are opting to leave the republic of science; those who merely wish to do so are wishing to leave science. The normative value of this rule is considerable: it sets an overall aim for science (the search for universal laws, a coherent theoretical system and causal explanations), an aim that ties the group together. Science can then be seen as a special form of voluntary society. No-one insists you dedicate yourself to the search for universal laws and causal explanations, but if that is your concern then science is probably what you are doing (depending on your rules of procedure), and if that is not your concern you can aid in the fight against confusion by not calling yourself a scientist.6
Proceeding through the book, the next time Popper brings up a methodological rule is when he is considering the problem of how to maintain the empirical character of an axiom system, to prevent it from becoming conventional. There is, he argues, no natural solution, only a methodological decision will do the job. Accordingly:

(R4) "I shall...adopt a rule not to use undefined concepts as if they were implicitly defined" (LScD, p. 75).

This rule is a corollary, no more, of (SR) which enjoined citizens of the republic not to permit the evasion of falsification. Abruptly using empirical concepts as though they were conventions is a technique of protecting a system from empirical difficulty which Popper named the "conventionalist twist". (R4) functions specifically to help us avoid unnoticingly giving matters a conventionalist twist when dealing with formalised areas of science.
Generalising, Popper argues that
it is impossible to decide, by analysing its logical form, whether a system of statements is a conventional system of irrefutable implicit definitions, or whether it is a system which is empirical in my sense; that is, a refutable system...my criterion of demarcation cannot be applied immediately to a system of statements...Only with reference to the method applied to a theoretical system is it at all possible to ask whether we are dealing with a conventionalist or an empirical theory. The only way to avoid conventionalism is by taking a decision: the decision not to apply its methods (LScD, p. 82).

Implicitly here is a general rule, not formulated, but which might be put as, "avoid conventionalist stratagems". Instead of such a general formulation, Popper went at it piecemeal. At the end of q719 of The Logic of Scientific Discovery he pinpointed four conventionalist stratagems, and for each of them he devised a methodological rule. The four are: the introduction of auxiliary hypotheses ad hoc; the modification of the ostensive definitions; the raising of doubts about the reliability of the experimenter [and/or his apparatus]7; the raising of doubts about the theoretician. In one of his few remarks about the social sciences in this book Popper comments that, compared to the physicist, the sociologist and the psychologist need constantly to guard against the temptation of these stratagems, and he singles out psycho-analysts as particularly prone (LScD, p.82).
The four rules Popper devises to meet the specific challenge of these four stratagems are as follows:

(R5) "[O]nly those [auxiliary hypotheses] are acceptable whose introduction does not diminish the degree of falsifiability or testability of the system in question but, on the contrary, increases it" (LScD, p. 83).

(R6) "We shall forbid surreptitious alterations of usage" (LScD, p. 84.)

(R7) "Inter-subjectively testable experiments are either to be accepted, or to be rejected in the light of counter-experiments" (LScD, p. 84).

(R8) "The bare appeal to logical derivations to be discovered in future can be disregarded" (LScD, p. 84).

The wording of (R7) is not wholly satisfactory - read carelessly, it seems to say experiments can either be accepted or rejected. What is plainly intended is a presumption that inter-subjectively testable experimental work be accepted. It should only be rejected in the face of counter-experiment. Thus in the debates about Pons and Fleischman's work on "cold fusion" these rules would prescribe attention to their inter-subjectivity, i.e. their repeatability. And, indeed, the scientific community did in the first instance try to repeat their work. Reiterated lack of success, as well as counter-experiment (thought experiments included), was what permitted challenges to come forward to the experimenters, their apparatus and their theoretical deficiencies.
Plainly, close adherence to these rules would be fairly devastating on much of the social sciences, where verbal sleight-of-hand, the introduction of ad hoc excuses and ad hominem indictments of the (class or other "interests" of the) investigators are almost standard.
Moving on, we come to what I call (R9), which is anomalous in the present discussion because it is not in the original text of Popper's book of 1935 but in a starred footnote, the star signalling its introduction in the translation of 1959. My reason for not excluding it is that while not stated in 1935 it was implicit and actually employed, as a careful reader will plainly see. It is a very general rule that has application well beyond science. Indeed, it might be termed a general methodological rule for the conduct of critical inquiry.

(R9) "[A]fter having produced some criticism of a rival theory, we should always make a serious attempt to apply this criticism to our own theory" (LScD, p. 85n).

An immediate consequence of this rule is that it should apply to the present endeavour. My theory that Popper is offering the skeleton of a constitution for the republic of science must itself be subjected to vigorous criticism, which I will undertake once the expository section of this chapter is laid out. An example of Popper utilising this rule is to be found in his discussion of what he terms the asymmetry between verifiability and falsifiability. The obvious objection to the demand that scientific statements be verifiable by experience is that it is impossible to satisfy. Science consists of universal theories and universal theories make assertions about infinitely large classes of objects. There is neither time nor opportunity to examine all those objects and verify the scientific claim. Also, many scientific statements refer to invisible (quarks, black holes), long disappeared (big bang), or entirely abstract objects (relations) and it is utterly unclear how they are to be verified at all.
Under the belief that Popper was proposing a simple substitution of verifiability for falsifiability, many philosophers made the criticism that falsification was in principle no more final or conclusive than verifiability was. A falsifying instance (this swan is black) presupposed verification (it is black and always will be black) if it was to falsify "all swans are white".
But Popper anticipated this objection and showed it can be answered. At issue was the logical asymmetry between verification and falsification. A single contradictory falsifies, a single verifier does not verify, except in the trivial sense that it does not falsify.8 Conclusiveness was a red herring as far as the logical situation was concerned. But Popper recognised that formal demarcation criteria like verifiability and falsifiability were inadequate. Falsifiability is a logical property, but falsification is not. A statement is declared falsified, it is a matter for a decision. The necessity of such a decision is what sponsors their articulation into a set of rules, a methodology.
Discussing the difference between falsifiability, which is the logical property we wish a system of statements to have if it is to encounter experience and thus be classed as science, and falsification, Popper stresses that a decision accepting a basic statement which contradicts a theory is a necessary but not a sufficient condition of falsification, because non-reproducible single occurrences are of no significance to science. This is a rule-like assertion, but it is not formulated as one. Although on p. 86 Popper says, "Thus a few stray basic statements contradicting a theory will hardly induce us to reject it as falsified", he does not get round to a rule for this situation until p. 106.

(R10) "[W]e should not accept stray basic statements - i.e. logically disconnected ones - but ... we should accept basic statements in the course of testing theories; or raising searching questions about these theories, to be answered by the acceptance of basic statements" (LScD, p. 106).

A science, Popper holds, needs a point of view and theoretical problems. It is only in the context of investigations into those matters that interesting as opposed to stray basic statements are developed around reproducible effects.
Four more rules can be isolated from the text of Popper's book. They involve a reiteration, the exclusion of accidents, the specification of random samples and the equivalent of materialism. They have much to do with what sorts of theories and what sorts of estimates of theories are welcome in science. On p. 121 Popper brings in a rule not previously formulated about empirical content and shows its equivalence to another about severe tests. Arguing that he regards the comparison of the empirical content of two statements as equivalent to the comparison of their degrees of falsifiability:

(R11) "This makes our methodological rule that those theories should be given preference which can be most severely tested...equivalent to a rule favouring theories with the highest possible empirical content" (LScD, p. 121).

The dots replace a cross-reference to the place where rules (R5)- (R8) are formulated, (R5) clearly implying that severity of tests is a positive value.
Again as a corollary, this time to (R1), Popper wants to estop one of the commonest evasion devices in common practice, the appeal to accident:

(R12) "I propose that we take the methodological decision never to explain physical effects, i.e. reproducible regularities, as accumulations of accidents" (LScD, p. 199).

This rule is introduced during a technical discussion of probability in physics, where the issue is how to prevent probabilistic hypotheses from rendering the system of statements unfalsifiable. Yet it has broad implications, suggesting that in science we should refuse to be satisfied with explanations from accumulation of accidents. This sharply distinguishes science from technology, where an air crash, for example, will as a rule be explained by an accumulation of accidents.
In the same context of discussing the falsifiability of probability statements, a further rule is considered necessary:

(R13) "a rule...which might demand that the agreement between basic statements and the probability estimate should conform to some minimum standard. Thus the rule might draw some arbitrary line and decree that only reasonably representative segments (or reasonably 'fair samples') are 'permitted', while a-typical or non-representative segments are 'forbidden'" (LScD, p. 204).

A final example from this context of probability statements and their confinement is:

(R14) "[T]he rule that we should see whether we can simplify or generalise or unify our theories by employing explanatory hypotheses of the type mentioned (that is to say, hypotheses explaining observable effects as summations or integrations of micro events)" (LScD, p. 207).

Here the challenge was the claim that all observable events should be explained by micro events. Noting that this doctrine is similar to certain forms of materialism, Popper calls that a "metaphysical hypostatization of a methodological rule which is in itself quite unobjectionable".

4. Assessment

What then is to be made of this list of 15 rules ((SR) + (R1)-(R14))? First and foremost to reiterate: it is incomplete. A Constitution for the Republic of Science would need many more rules, and some specification of their institutional embodiment, including rules for dispute settlement. Popper makes no effort to organise the rules systematically and lay them out in a table so that they can be checked against one another and debated in relation to one another and the aims. This may explain why they are an aspect of his philosophy that is seldom discussed. The occasional challenge to one or another rule is seldom framed in a manner that suggests appreciation of the innovative brilliance of the idea. What seems to be at work is that both Popper and his critics suffer from a blindness to the institutional turn. Let us re-frame the rules by spelling this out.
What Popper's rules amount to is this. The demarcation between science and non-science cannot be stated in an abstract way, only in a practical way. Science is an activity carried on in a select community and this community is obedient to a set of rules which guide its activities. The community is both real and partial. Real because it consists of actual human beings, partial because their participation in this community is only one of the roles they play as social actors, and by and large their other roles are played in distinct, but overlapping communities. All scientists are, for example, citizens of nation states; virtually all scientists belong to families of orientation (their natal family), and the majority also to families of procreation. Most scientists belong to communities and groups in civil society (political, religious, voluntary, recreational). The vast majority of scientists belong to some one or other large institution that provides them with a livelihood: government departments, universities, laboratories, museums, business corporations, etc.
In none of these "outside" involvements is it necessarily the case that they conduct themselves by the same set of rules of procedure that Popper has been sketching out. Notoriously, there are critical and open-minded scientists who cleave to dogmatic religious beliefs and raise their children with a much stronger expectation they will also cleave to the religion than that they might in turn become scientists. The social institution of science and its constitutive rules are, then, a special niche in the overall social life of its members.
What kind of a social institution is it? Popper nowhere says anything about its internal organisation. The rules give us no guidance to the manner in which it is governed, or to whether there is established leadership. Indeed, although many rules are put forward for discussion, the issues of how these discussion are to be conducted and how decisions about amending the rules are to be made are not entered. Thus it is a stretch to see these rules as a Constitution for the Republic of Science or even as a proto-constitution. They are more like a proposed set of procedural rules for discussion by a body already in place.
Now of course science was at the time Popper wrote well and truly established in its place in society. It was in many ways unique among social institutions. Its acitivities tended to centre around the subinstitutions of the learned society, associated journals, universities, laboratories, and international conferences. Physics was understood by physicists to be a far-flung invisible college uniting colleagues from around the globe in common endeavours. But to point this out is also immediately to point out that there was a leadership. To be the senior professor at certain institutions, or the head of a learned society, was a concrete kind of leadership for which there was fierce competition. There was also intellectual leadership, independent of such posts, which turned solely on the esteem directed at the work of a particular person, regardless of formal post. And, without a doubt, there were some elements of charismatic leadership - scientists who could energise others with a vision, regardless of whether their own work was important.
Concrete leadership resided mostly in prestige institutions that themselves had constitutional means of selecting those position-holders. Election to professorships, elections in learned societies, the selection of journal editors, were all conducted by sets of rules bearing no relation to Popper's methodological or procedural rules. It is not at all clear that the weight of these offices and rules was brought to bear to move science in the direction Popper intended (especially if Kuhn is to be believed). Institutions and office-holders develop vested interests, some of an intellectual character. Thus, for example, the injunction to prefer the most severely tested theory, or always to apply a new criticism of the theory one opposes to one's own theory might not at all conform to institutional pressure and so be ignored.
Because his set of methodological rules is free-floating, not attached to any institutional framework, the task of defending them, and of debating them in a responsible way, is left up in the air. Popper has not provided a constitution for science, not even a declaration or charter of rights (and responsibilities) which the constitution is trying to embody in a system of governance. Perhaps that is as well. For the methodological rules Popper proposed are not intended for this or that university, this or that physics laboratory, nor are they intended for some kind of United Nations Organisation of physics. They are directed to science conceived of as a general and abstract republic. They enjoin: here, if you want to respect and advance these aims, is a set of proposed procedures. But "you" is unspecified, and what "you" are to do if you have what you think is a better idea is also unspecified. These proposals being published by a philosopher of science you thereby elect a kind of specialism of that name to debate these proposals. Yet since they are proposals to guide the practice of research in its most general aspect their testing and debate might seem to be most appropriately carried out by the community of scientists proper.
Another way to think of science is not as a series of concrete institutions but as an invisible college, an abstract institution rather like language. We might see Popper's rules as addressed to this wider community of science, one that has to do with self-identification and not with institutional gatekeepers. Popper's arguments and rules are then seen as directed to men and women of good will who want to advance the project of science. This could explain why individual scientists, in their autonomy, often pledge allegiance to Popper's vision of science, while the institutions to which they belong seldom do so. And parasitic institutions like those of the philosophy of science are even harder pressed to find any role for appreciative incorporation of Popper's ideas. Popper's lack of attention to the concrete institutional embodiments of science, and his efforts to formulate rules outside such constraints, betrays, I believe, a fundamental mistrust of existing institutions and practices. They were and are hierarchical and authoritarian and no longer welcoming to the man or woman of good will, as well as being subject to the corruptions of power and wealth. Later in his career, commenting on Thomas Kuhn's notion of "normal science", i.e. its concrete institutional embodiment, Popper declared that:

'Normal' science, in Kuhn's sense, exists. It is the activity of the non-revolutionary, or more precisely, the not-too-critical professional: of the science student who accepts the ruling dogma of the day; who does not wish to challenge it; and who accepts a revolutionary theory only if almost everybody else is ready to accept it -- if it becomes fashionable by a kind of bandwagon effect. To resist a new fashion needs perhaps as much courage as was needed to bring it about...
I admit that this kind of attitude exists; and it exists not only among engineers, but among people trained as scientists. I can only say that I see a very great danger in it and in the possibility of its becoming normal (just as I see a great danger in the increase of specialisation, which also is an undeniable historical fact): a danger to science and, indeed, to our civilization (Popper 1968, pp. 52-53).

In utilising this quotation I am leaping well out of my self-imposed restriction of period. But nothing in my present line of argument relies on this quotation. It simply is an explicit affirmation of what I think we will find, when we discuss Popper's ideas on society, history and politics, in the classical period that is my main concern: that his philosophy of science portended a radical critique of concrete institutions and a desire to hold them to normative standards that men and women of good will could agree upon, not those laid down by the entrenched experts.

1. According to the account that appears in C&R, p. 34 (a chapter first published in 1957) and elaborated in the autobiography of 1974/76.
2. The use of this term for the present purpose is first found in W. W. Bartley, III's The Retreat to Commitment, New York:Knopf 1962.
3. Cite New Atlantis.
4. Bacon identified four 'idols' as common dangers.
5. The second half of Virgil's line "Felix qui potuit rerum cognoscere causas" (Geogics, II, l. 490) is the motto of the London School of Economics, where Popper spent the latter half of his academic life.
6. The aim of science built into (R3) is not quite the same as that outlined in 'The Aim of Science' (1957) - 'satisfactory explanation'. Whether the difference is significant or represents a strengthening and deepening of it is an exegetical point I leave to one side. Agassi claims the 1957 paper portends a new philosophy of science.
7. The phrase in parentheses is mine but is clearly implicit in the text.
8. As to the objection that every falsifying basic statement verifies the contradictory of the statement being tested (a basic statement about a black swan could contradict "all swans are white" and verify "Some swans are not white"). this involves the same equivocation on the word 'verify'. Two contradictory statements cannot be true together. Any statement that does not contradict another can be said to verify it, even if it is false. The positivist demand for verification involved a stronger sense of 'verification' as a demand that statements be 'verified as true or probable', which required lots of verifying statements. A single one was of no interest. Pace Kuhn, a single falsifying statement is of great interest.

* The text of this talk is taken from a book in progress. It was prepared for this meeting of the Japan Popper Society during a visit to Japan and, as a result, some of the references are incomplete.