POPPER’S PROBLEMS OF DEMARCATION

The plural in the subtitle, ‘Popper’s problems of demarcation’, is novel, and a serious contribution to the understanding of Popper’s philosophy, made by his former student, W. W.

I think, again with Bartley, that there is more to it than that. Let us lump all the various problems of demarcation mentioned above into one, and ignore, for example, the fact that changing (institutional) attitudes may render some or even all scientific theories unscientific. Still, the title ‘the problem of demarcation’, is somewhat misleading, considering that the expression is Popper’s own and that he uses it from the start in two rather distinct senses, one historical and the other modern - his own­sense. Usually this does not matter much, except when contrasting his view with those of his predecessors, and when attempting to decide whether Popper’s criterion of demarcation legislates what one should view as scientific or whether it characterizes science in its social and histo­rical setting.

(a) what makes science empirical!

(b) what makes a theoretical system scientific!

Now, classically, the people Popper mentions, Hume, Kant, Russell, and Whitehead, answer question (b) by the answer:

(J): scientific=certain or near-certain.

Now, (J) (for justificationism) is objectionable, as Popper explains; he replaces it by his new

(E): scientific=empirical.

Of course (E) (for empiricism) of necessity reduces (b) to (a); but, for the tradition, (J) makes (b) and (a) possibly distinct! In other words from (b) and (E), we can say, (a) in a sense follows; but not from (b) and (J): (J) answers (b) with no reference to (a).

Traditionally, particularly for the philosophers Popper mentions, a special status was allotted to question (b), what makes a theory scientific?, in view of the alleged obviousness of the answer to it. Indeed, this was said not to be a topic for serious consideration. (J) regularly was taken to be the answer to (b) as a matter of course, in spite of clear know­ledge of the classical objections to (J) - which are unsurmountable. The philosophers Popper mentions have, each in his own way, tackled this situation, expressing the desire to be critically minded, while accepting (J) as a matter of course, in spite of unanswerable criticism of it. This, no doubt, is a tall order by any standard. Traditionally, the situation is even more complicated. Most philosophers of science, particularly at present, endorse two further assumptions:

(S) a theory is certain or near-certain when and only when backed by evidence;

(C) a theory is empirical when and only when backed by evidence.

(R) a theory is empirical only when it can be refuted by evidence. In a compromising mood an embracing meaning was offered: ‘empirical’ should now mean, some measure of empirical decidability one way or another:

(D) a theory is empirical if it can be backed or undermined by evidence.

After this change, (J) and (S), which most contemporary philosophers of science endorse, still easily entail (E). In other words, due to Popper’s influence, philosophers of science who still stick to their views ((J) and (S)) now answer consistently two or three - indeed ten - problems of demarca­tion together:

(OMN) scientific=(near) certain=empirical=confirmed=backed by experience.

The situation can easily be further complicated, and as it was compli­cated already in two ways - probability and meaning - when Popper appeared on the scene. They have persisted, and Popper found it advisable to attack them repeatedly.

The probability complication is the addition of one of the two follow­ing formulas, perhaps or both:

(Pi) empirical backing=high probability

(P₂) empirical backing=increased probability

where ‘empirical backing’ is backing by evidence as in (S) and where probability is some (unspecified) additive function. (Similar formulas can be written for undermining evidence rather than backing evidence.) These two rules, (P_x) and (P₂), are clearly auxiliary hypotheses. Criticizing them really amount to attacking only the periphery of the problems, and I cannot imagine that Popper would have fallen for this digression but for the fact that Carnap has endorsed (PJ or (P₂).

How is one to choose between (P_x) and (P₂)? Let us accept (J) and (S), and let us try to retain (E) as a corollary if we can.

(R) empirical = refutable,

since, clearly, refutability is high initial improbability.

We are running in circles now, and are playing on different meanings given to ‘empirical’, sliding back and forth from the classical (C) ‘support­able by evidence’ to Popper’s (R) ‘refutable by evidence’ and back - via (D) which is the inclusive disjunction of the two. To make this rather trite exercise somewhat more sophisticated, let us add the other confusion which was there before Popper entered the scene.

Let us consider the problem of empirical meaning and observe how easily it effects a serious change in the problem of demarcation, and yet does so quite incidentally. Whereas the negation of a (certainly) true statement is not true, the negation of a meaningful statement is yet meaningful. Hence, whereas for classical thinkers the negation of a scien­tific theory was decidedly not scientific, for the followers of Wittgenstein the negation of a scientific theory was decidedly scientific. Yet they refused to view refuted statements as scientific. (It is difficult to blame them for this, since old traditions die hard.) They therefore had to work with two concepts simultaneously, potentially-scientific and actually-scientific, to wit, verifiable and verified. This does not help: the negation of a verified existential statement is a universal statement that never was even veri­fiable.

The confusions mentioned here, between testability and meaning, between probability and increased probability, between actual and potential increased probability, are all easy to cross-fertilize. The question, what attitude one takes towards a literature relating to one’s interest when this is not up to one’s standard is very tricky; concerning it there exists practically no literature. I consider Boyle’s ruling, in which he proclaimed that he would ignore the publications of those chemists who would not describe their experiments in a sufficiently clear manner to render them repeatable, very wise, indeed. Even William Whewell’s harsher ruling that he could not take the German philosophers seriously because they took Hegel seriously (meaning Hegel’s Naturphilosophie, of course), I consider not unfounded.

This is not to belittle Popper’s original work. It is something of a miracle - in my opinion for what it is worth - that Popper has emerged from the morass of the current philosophy of science so-called with a classical work like his Logic of Scientific Discovery. But I do not think he was left completely untouched by the morass. When Popper says that a refutable theory is scientific he may mean, first, that it is actually scienti­fic: initially and in the light of any experience, corroborating or refuting; he may mean, second, that it is only potentially scientific until it be corro­borated. Popper (but not his colleagues who engage in studying scienti­fic meaning) can choose either reading; because, for Popper, as for the classical philosophers, the negation of a scientific theory is not scientific. As for myself, I had understood Popper to have meant my first reading rather than my second reading. I know now that that was mistaken: he really means to say, a theory is potentially-scientific if and only if it is refutable, and actually-scientific if and only as long as it is well-corroborated in the light of all existing evidence. In Popper’s early works both my present readings are allowed; in 1960 he wrote a note at my suggestion ([Conject­ures], 248 n), admitting that much, and further endorsing something akin to my present second reading.

Let me take up Popper’s problem of demarcation following his own presentation ([Conjectures], Ch. 1). He had been puzzled, he tells us, by the status of certain theories which were topical at the time: Einstein’s, Freud’s, Marx’s. Why is it, he asked, that Einstein’s theory is scientific, but not Freud’s? What makes the one but not the other scientific? What was fishy in Freud’s claim to empirical backing?

In retrospect I find this problem somewhat embarrassing. Perhaps I should say I find that formulation embarrassing. It may mean - and I say this after having checked and rechecked the relevant texts much more closely than the normal standards require-any of the following problems. Why should we honor theory E but not theory Fl Why should we study and examine E but not Fl Why should we accept the claim that E was backed by evidence but not Fl Why should we accept the claim that E exhibits empirical nature but not Fl Why should we allow ourselves to believe E but not Fl Why must we believe E but we need not believe Fl Why must we believe E and we must refrain from believing F (but either disbelieve it or suspend judgment on it)? Why should we act on the as­sumption that E is true but not on the assumption that Fis true?

These few questions (the list is not exhausted; see above), each of which fits Popper’s text well enough, may look like variants of each other - but only to the untutored eye. Let me sketch only two differences, those which happen to be substantial. First, belief and grounds for action are not the same: as Popper stresses in his discussion of instrumentalism in his ‘Three Views’, we knowingly employ false theories in technology, but we do not knowingly believe a false theory. The second difference is between empirical theory and empirical backing. The confusion between the demarcations of these two is the one which eluded me and troubled me for a few years. In his early autobiographical sketch ([Conjectures], Chapter 1) Popper discusses both the nature of empirical backing and the empirical nature of a theory. He may have suggested that an empirical theory is a refutable theory and an empirical theory may be refuted or alternatively empirically backed. He may have suggested that a theory is potentially empirical if it were refutable and actually empirical if it has been properly empirically backed. To repeat, I think Popper has acknow­ledged that earlier he had been unclear on the topic, and that he now considers a scientific theory proper only one which is properly empirically backed before it is empirically refuted. On top of this past confusion of mine, for a long time I found it difficult to notice, what now seems clear, that in his autobiographical sketch Popper says that Freud is in error both because Freud’s theory is unempirical and because this theory cannot have been properly backed by evidence. First, says Popper, a theory is empirical only if it is refutable; second, an empirical support proper can result only from a proper test (i.e., the attempt at a refutation); third, as a corollary, only an empirical theory can be properly empirically supported; fourth, and finally, a potentially scientific theory needs some empirical support to become actually scientific.

It looks obvious to me now that this is what Popper has been saying all along. I was a devoted student of his and worked closely with him for over seven years, yet I really understood him only a few years later. I am not clear as to whether hermeneutics is ever a worthwhile activity: I liked Popper’s ideas better when I understood them less. In addition to the various criticisms already alluded to, I wish to say that proper refutations and proper corroborations of both scientific and unscientific theories are quite possible. It is easy to illustrate cases of proper and improper backing and refutations of scientific and of unscientific theories, thus having (at least) 4 kinds of backing and 4 kinds of refutations.

The cases mentioned in this table are those of tests which ended in empirical backing and in empirical refutations. Both Pare and Jenner were scientifically testing known superstitions: Jenner found empirical backing to one, and Pare refuted many. Contemporary aerodynamics is classical and so, qua scientific theory, is plainly refuted; but, qua rational techno­logy its forecasts are properly tested - with supervision of well known government agencies.

I do not wish to decry Popper’s study of the problem of demarcation of science. I join Bartley in viewing it as an important problem which Popper’s studies have rendered unimportant (see Appendix). Also, I see more value in Popper’s equation of the (actually) empirical with the (potentially) refutable; though I think he disowns this very equation. What I view as the more important and most original contribution of Popper’s to the philosophy of science is his solution to the problem of induction: we gain theoretical knowledge from experience by refuting some of our theories. (But, I equate neither refutation nor empirical character with science. See my [Novelty] and my [Nature].) Unfortunately, Popper himself, far from seeing that the problem of demarcation of science is already a conglomerate of problems, has further collapsed it with the problem of induction into one. Thus, he says, ([Conjectures], 52), “ it took me a few years to notice that the two problems - of demarcation and of induction - were in a sense one”. No doubt, in some sense any two prob­lems are one; and so, in a sense, this statement is true. But Popper himself paraphrases this same statement (two pages later) thus: “the problem of induction is only an instance or facet of the problem of demarcation”, which I consider patently false (see Bartley’s [‘Demarcation*], 64). I do not know how literally to take this. Elsewhere ([Logic], 42), he says, “the proposed criterion of demarcation also leads us to a solution of Hume’s problem” - which, taken literally, may well be true. Similarly, he is cautious when he says (ibid,, 313), “the solution suggested here has the advantage of preparing the way also for a solution of the second and more fundamental of the two problems of the theory of knowledge (or of the theory of the empirical method).” Nonetheless, in my view, the first problem, of demarcation, is concerning knowledge, and the second, of induction, is concerning method. Popper often seems to use the terms ‘methodology’ and ‘epistemology’ almost interchangeably. This is re­grettable. So long as Popper indentifies the two problems, of insuction and of demarcation, and thus, the two fields, methodology and epistemology (see Wellmer, [Methodologies, the theory of learning and the theory of knowledge, he does not add to clarity. In particular, he forces us thus to identify conventionalism and instrumentalism and thus come to the very incongruity with which we started: modified instrumentalism is trite and modified conventionalism intriguing; yet they modify the same doctrine in the same way!

There are a number of questions concerning Popper’s early theory of demarcation and its application to the gradualist or approximationist approach as endorsed in his later ‘Three Views’ and as even adumbrated at the very end of his Logic of Scientific Discovery, Popper assumes the following: if we take the most highly testable theory available, corrobo­rate it, refute it; if we repeat this process of conjectures and corroborations and refutations, then we shall find a series of theories approximating truth. This sounds like a miracle, but is partially explained by Popper when he declares that explanatory power, degree of generality, and content, all increase with degrees of refutability. I have refuted these claims ([‘Nature⁹]); and on this score I have suggested that harmony can be achieved better by narrowing Popper’s problem of demarcation to that of empirical charac­ter, not to scientific one; and by demarcating science itself by its empirical nature as well as by its explanatory power, generality, abstractness, depth, etc.

Meanwhile I have gone further. I now find it hardly possible to charac­terize science even by a large set of characteristics. I can only view science as a tradition where such and such characteristics are manifest to this or that degree and even in conflict (see Appendix to Chapter 8). And, inasmuch as I endorse Popper’s view of traditions as akin to myths, I cannot avoid seeing science as largely a myth. Scientism, which Popper rightly views as a myth akin to some religious myth, is rather obviously part and parcel of the Western scientific tradition, and Popper is mistaken in denying this. Consequently, it is with some surprise that I find that, though I have gone further away from Popper’s teaching towards modified conventionalism as broader than his modified essentialism, my modified view is in spirit though not to the letter, more Popperian than Popper’s declared view. Or should I say Bartley an? This is a rather ironic point of departure. Natu­rally, it seems inevitable that at a sufficiently large distance of time wide differences of opinions vanish and historians who notice them at all wonder what the vehement disputes of the past were all about; but it is a strange sensation when it occurs to one’s own colleagues and associates. But let those historians, if any, who pay any attention to details smaller than Popper’s broad ideas, let them not conclude that the Popper-Bartley vanishing controversy did not feel any the less real, that by-standers did not feel caught up in its heat.

IV.