The Popperian Image of Science Gerard RADNITZKY Ruhr-Universitat Bochum, West-Germany 1. Skeleton outline of the methodology 10. According to Popper, a research enterprise starts from a problem (Pi in fig. 1). Wheredoes the problemcome from? Popper focusseson the type of research situation in which progress of a dramatic order may occur rather than on the "piecemeal" mode production of new knowledge characteristic of research enter prises approximating the Kuhnian ideal type of Normal Science. In the typical situation of this type a "veteran" theory has run into diffculty, "clashed" with some experimental results. The "clash" may be more or less manifest, and more or less serious, with logical contradiction as a limit case. It may be latent, in particular in a type of research situation where pieces of knowledgeproduced in an "empirical" manner cannot be explained by the pertinent veteran theory. The residual may grow but resist explanation, as e.g., spectroscopicknowledge did before 1913. Such a residual-knowledgewhich cannot be explained by the veteran theory is practically always in evidence; however, when it accumulates or manifestly con flicts with the veteran theory, it constitutes a challenge which can no longer be ignored. The afore-mentioned type of situation of manifest conflict typically arises when attempts are being made to apply a theory to a new territory. Indeed it must occur, sooner or later, since the kernel of the theory is a conceptual model of a real (physical, biological,...) system, and because it is in the natur eof a model that it also has negative analogy. Thus a theory, once created, will sooneror later give rise to an objectiveproblem situation of that sort: a problem which confronts us or which we literally have to dis-cover. In particular, in case of a "clash", it is not just a question, but a problem in the etymological sense of the word 'problema': situation in which one cannot remain. The case under discussionis just one example of the phenomenon that objective problems arise whether the researcher wishes it or not, which phenomenon serves as part of Popper's "argument from independence" for giving ontological status to World-3.1 11. Where does the theory came from? Figure 1 sketches only two possibilities: the theory being one of the theories which, at the time, are accepted 1. E .g. OK pp. 160, 118 f., 147. -3- 4 G. RADNITZKY Vol. 5 in the discipline concerned; and the situation where an Ansatz towards a theory has just been made on the basis of the background knowledge, in particular on the basis of preconceptions about the general nature of the new "territory" (new sort of real system) which are derived from "world-picture" hypotheses. In the creation of theories about a hitherto unexplored realm "metaphysics" is structurally indispensible, and not merely indispensible as a part of psychological heuristics (world-2).2 normative (praxiological) model of the "game of science" legend: _??_•« " mapping" _??_ steering _??_ processes T theories/hypotheses X subject matter/territory •› results/store P problems Y other territories •Þ selector HG hypothesis generation ISF "internal steering factors" _??_ decision process HC hypothesis control Fig. 1 (artwork, text, legend) Preconceptions about the general nature of the subject matter at the level of world-picture hypotheses ("cosmological"hypotheses) also play an important role in the first-mentioned type of case, even if they remain implicit. They constitute the kernel of a system of factors -preconceptions, evaluations,... -which may 2. Cf . LSCD pp. 38, 277 f. -4- No. 1 The Popperian Image of Science 5 be called 'Internal Steering Factors' (ISF in fig. 1).3 Partially based upon the preconceptions is a programmatic conception about the discipline to which the theory belongs: evaluative statments about what the discipline (the "X-ology" in question, 9 in figs. 1 and 2) should be like, which may include an "ideal of science." According to the methodology under discussion a research enterprise that is properly conducted is governed by the rules4 derived from Critical Rationalism (right column of fig. 1); and thus the models presented by figs. 1 and 2 consist of the gambits and moves resulting from the application of these general rules to the various moments of research. The rules also specify the concept of "scientific progress" or the comparative concept of scientificquality, and the methodology of appraisal is spelled out in figure 2 as the operational characteristics of the "Hypotheses-Checking/Control Station" HC. These methodological gambits and moves apply to all research enterprises worthy that name. In "Normal Science" a certain substantive system of ISF, in particular a certain set of pre concepstions (cosmological hypotheses), steers the concrete procedures of a certain ensemble of concrete research enterprises in more detail; it provides sub stantive criteria for the appraisal of products and gives these research enterprises a certain unity and direction: gathers them together in a research "direction", "tradition" , "school", "style",...5 3. Preconceptions at the level of world-picture hypotheses do not belong to the theories themselves. E.g. among the preconceptions characteristic for the ISF of "Newtonian" type of theory would be the ontological posit that the world (that world with which physics is concerned) consists of matter spread out in empty space; in such a world, for two bodies not in contact with each other (not even through intermediation of other bodies travelling from one to the other) to interact causally, the simplest thing to assume is "simultaneous action at a distance" as a causal connector, (cf., e.g., (Radnitzky, 1974b) p. 33) etc. For a discussion of the concept of "preconcep tions" cf., e.g., (Radnitzky, 1974a) •˜ 212, Radnitzky, 1972) pp. 141-147; on the role of preconceptions in research cf. (Radnitzky, (1974b) with that title, esp. pp. 27 ff. and for examples from the "distilled history of science" pp. 32-43. Imre Lakatos confirmed to me in personal conversation that the "hard core" of an RP basically consists of such preconceptions; but there are passages in his writing which suggest that the hard core is made up mainly to the central components of the theory, which, over its history, constitutes the RP. Of course, the concept of the ISF is close to one of the many meanings of the Kuhnian "paradigm". Such as the "master-rule" mentioned in fig. l. 5. Cf., e.g., (Radnitzky, 1974b) pp. 7f. The claim that the rules apply to all research should not be taken to imply any form of "unity of science" thesis. We only wish to imply that the sciences humaines a very heterogenous group too contain certain moments (which Apel calls 'quasi-naturalistic' or 'explanatory' moments) to which the model of figs. I and 2 may be adapted. Again, to claim this is not to deny that the typical sciences humaines, the Geisteswissenschaftenhave additional methodological problems of their own, which this model does not touch upon. Some of them very likely cannot be tackeld without taking into consideration "knowledge-constitutive interests": there is a great difference between the task of "giving an explanation of understanding" (Erklaren des Verstehens) and the task of "understanding explanation" (Verstehen des Erkldrens). -5- 4 6 G. RADNITZKY Vol. 5 12. The proposed solution to the problem (Pi in fig. 1) is a tentative theory (Ti+1 in fig. 1). Where does Ti+1 comefrom? In figure 1 the box "Hypothesis Generation", HG, marks the "station" producing the tentative theory. Following Popper's usage, in figure 1 'theory' and 'hypothesis' are used interchangeably. Among the general rules (applying to all research) the one particularly important for the operation of HG is that labelled "preference rule": "Prefer theories (hypotheses)with high informationcontent (and hence more "falsifie-able" i.e. involv ing greater risks of being wrong, allowingmore severetests, theories whose absolute logical probability is as low as possible) to hypotheses with a lesser amount of information!" (Roughly: Nothing risked, nothing gained: the more "daring" the theory, the greater its potential for corroboration by tests.) (This rule of course follows from striving for severe tests.) However this rule gives only a very broad "policy" direction. In a concrete research situation the operation characteristics of HG are governed, in more detail, by the ISF, especially the (substantive) preconceptions ("cosmological" hypotheses). The ISF also steer the operation of the selector (marked in the way customary in cybernetic charts, by a "standing" triangle), i.e. they guide the decision concerning which part of the background knowledgeavailable at t is made use of. Of course all the input into HG are world-3 entities. Within the scope of the operation characteristics of this type of "Hypotheses - Generator", a logical move may be accomodated which, if used as a move of appraisal, would be highly objectionable (because then appraisal and generation, HG and HC, would have been conflated): a logical move which is often called "inductive" and which might more properly be labelled "observationist inductivist": from a set of data sentences to a generalizationof the same kind. If the data are taken as basic only in the sense of not being problematized in the particular research context at hand, then, although the logical move is not governed by precise formal rules, the procedure may nonetheless lead to a probabilistic conjecture. This move is certainly made use of in experimental work. But it must not be forgotten that not even this procedure is purely "empirical" (as one objectionable branch of "inductivism" would have it) since it is guided by the ISF all the way, even if this guidance may not be prima facie recognizable. In the type of research situation Popper focusseson (situation with chances for a "big step forward", other than "normal science"), in situations where the essential output of the HG is a theory rather than a law hypothesis, the the essential input will come from the preconceptions, from the "cosmological" hypotheses.