Printer Friendly

Philosophy of Science in the Twentieth Century: Four Central Themes.

In this book Donald Gillies set out to provide both an introduction to and 'a history of philosophy of science in the twentieth century' by focusing on four themes that 'all would agree . . . are indeed of the greatest importance' (p. xi). These are: (I) inductivism and its critics, (II) conventionalism and the Duhem-Quine thesis, (III) the nature of observation, and (IV) the demarcation between science and metaphysics. Gillies does not, however, simply present a survey of past opinions. Rather, the discussions of the different topics are meant to be cumulative and make an original contribution to the literature by issuing in a new answer to the problems faced by the methodological doctrine of falsificationism. Here a pointed over-view must suffice to motivate some brief remarks on the book's success as an introductory text.

Part I lays out the distinctions basic to and the problems associated with inductivism. (Along the way, it also introduces the main historical actors dealt with in the book as a whole: Russell, the Vienna Circle, Popper, and Poincare and Duhem - only Wittgenstein appears later.) The story told is a familiar one: the Baconian view that science starts with observations and moves on from them to generalizations and predictions fails, first, for want of observations without a theoretical background and, second, because science can make do without any inductive inferences. Thus Popper 'rejected both inductivism as a theory of discovery and Bayesianism as a theory of justification, (p. 32). Dissenting from Popper's further claim that scientific conjectures can never be positively justified, Gillies confronts the general method of conjectures and refutations with examples from the history of science, lending his qualified support. A shorter account of Duhem's critique of Newton's version of inductivism follows. With 'creative theorizing' taking the place of traditional inductivism, 'conjectural induction' (an explanatory generalization which, being based on observational data, may nevertheless contradict them in part, pp. 47, 62) emerges as Gillies' candidate for the 'kind of synthesis between inductivism and falsificationism' he had announced earlier (p. 36). (The reader will have to wait until the end of Part IV to learn whether there is anything distinctly Popperian left in the synthesis provided and what the positive justification of hypotheses might consist in.)

Part II begins with Poincare's conventionalism: the basic principles of geometry and laws of physics are viewed as definitions of the terms involved which could never be invalidated by experiment. Noting the limits of Poincare's conventionalism, Gillies turns to a discussion of the various strands of the so-called 'Duhem-Quine' thesis (often held to support even extreme forms of conventionalism). The 'Duhem thesis' (also called the 'holistic thesis') says that 'an experiment in physics can never condemn an isolated hypothesis but only a whole theoretical group'. For Gillies, Duhem was not a conventionalist, because he held that no conventions were in principle immune from experimentally generated doubt. His theory rather was a 'modified falsificationism' (p. 104), for he distinguished between physical laws falsifiable in isolation (called 'level-1 hypotheses') and those that are not ('level-2 hypotheses') and assigned to 'bon sens' the decision when to regard experiments as 'in practice crucial' for level-2 hypotheses. The 'Quine thesis', Gillies then notes, is stronger than Duhem's, because it concerns not only physics, but all of science, and because the unit of empirical significance is now the whole of science, not just a certain group of hypotheses. Part II closes with Gillies' own version of the Duhem-Quine thesis: (i) the holistic thesis applies to level-2 theoretical hypotheses in all sciences (pace Duhem), (ii) its unit of significance is smaller than the whole of science (pace Quine), and (iii) 'scientific good sense' would in many situations overrule the logical possibility of making adjustments elsewhere (p. 115).

Part III begins with a brief chapter on the Vienna Circle's and Popper's views on evidence statements, before turning to the question of the theory-ladenness of observation. Notable about the former are the swift parallels Gillies draws between Carnap's and Neurath's reasonings in favour of 'physicalist' construals of the empirical basis of science and the later Wittgenstein's rejection of private languages (points I have independently and extensively discussed elsewhere, e.g. Overcoming Logical Positivism From Within, Rodopi, Amsterdam [1992]), as well as Gillies' ecumenicism about Popper's and the Vienna Circle's proposals: science uses both personal reports of sense experience and impersonal observation statements. For the thesis that observation statements are theory-laden we are directed to Duhem's view of scientific experiments as 'interpretations' of phenomena and 'Neurath's principle' (the test of a scientific statement involves assuming some others which, in due course, could also be called into question). Gillies recommends that Duhem's view should be extended to all observation statements whatsoever. An observation statement is to be defined as 'a statement which is the result of some sensory input interpreted, whether consciously or unconsciously, using a set of theories . . . theories of the instrument are part of the set of theories used to interpret the sensory input' (p. 146).

Part IV addresses the demarcation problem between science and non-science. Since Wittgenstein's Tractatus, in the view of the Circle, correctly held that all meaningful statements are truth-functions of simple observation statements, they concluded that all those statements that were not so reducible were meaningless, for instance those of metaphysics. Gillies favours Popper's replacement of verifiability as a general meaning criterion by falsifiability as a mere demarcation criterion. In support of Popper's 'almost wholly convincing' rehabilitation of metaphysics as meaningful and productive Gillies introduces the later Wittgenstein's theory of meaning as use - metaphysics could be viewed as a rule-guided discourse, another language game - and cites the metaphysical theory of atomism as a useful 'heuristic' for science. Nevertheless Gillies recommends that 'restraint should be put on enthusiasm for metaphysics' (p. 191). How this is to be done remains unclear until the final chapter (when the problem left open from Part I is also resolved). Dealing with criticisms of falsifiability Gillies focuses particularly on the Duhem-Quine problem and rejects Popper's recent answer (that it is 'systems of theories' to which falsifiability applies) as falling victim to similar arguments as the verifiability criterion did. Gillies' own candidate is 'confirmability or corroborability', supplemented - so as to avoid the old arguments - by the 'principle of explanatory surplus' (PES) which 'denies that if e follows logically from h, this automatically means that e supports h' (p. 215). Accordingly, we must count the number of theoretical assumptions and that of facts explained; when the number of the former exceeds that of the latter, an explanatory deficit is obtained and no support of the former is said to be given by the latter; 'only a rough qualitative estimate' of the surplus (or deficit) is required for the principle to work. But what then, finally, of falsificationism? Despite his switching to confirmability, Gillies argues that nearly all of the prescriptions of falsificationism can be retained, given that we can now let level-2 hypotheses be confirmed modulo PES via level-1 hypotheses derived from them. Thus we arrive at Gillies' own 'modified falsificationism', a 'systematic development' of Duhem's position. Popper has been left behind, after all: Gillies' modification avows that the aim of science is to find theories that work, not only to refute candidates (p. 223)! And how is metaphysics demarcated? By dropping the truth requirement on 'knowledge' and insisting on the empirical nature of justification: unlike scientific ones, metaphysical beliefs cannot attain the status of knowledge (p. 230)!

As will be evident already from this broad overview, Gillies's book covers standard territory in an original and stimulating fashion. Given its deliberately introductory nature, it may not be fair to criticize the lack of conclusive argument for many of its more controversial readings (several quibbles have been suppressed above and must be here) and proposed solutions (which no doubt will attract independent attention), especially since this is compensated for by a clear and incisive style and a rich store of examples from the history of science. Viewed as a history of twentieth-century philosophy of science, however, Gillies' book is open to several objections. Did the present century end with Popper's post-Second World War elaboration of his opposition to the Vienna Circle? Surely the absence of any sustained discussion of the work of Kuhn and his opponents is remarkable in a book with the present main title. Perhaps its author feels that whatever advance there has been in philosophy of science since Popper's day, has occurred in probability and/or confirmation theory which, due to their formal complexity, are not much discussed here. Given the complete neglect thus entailed of the various contemporary approaches to the material culture of science, however, Gillies at least fails to live up to the catholicism vis-a-vis the 'logical' and the 'historical approach to philosophy of science' he himself attributes to 'the ideal philosopher of science'. A second complaint concerns the treatment of the half of the century Gillies does discuss. While learned and enlightening about most of its protagonists, the book is disappointing in its discussion of the philosophies of the Logical Positivists. No account has been taken of the new scholarship on the matter (for instance, Coffa, Friedman, or Haller, to name but three of the more senior writers involved). Too often still the Vienna Circle serves as a somewhat simplistic prelude for Sir Karl - even in instances where the latter's claims are less than original (e.g. that 'something had gone wrong' when the Tractatus declared itself meaningless on its last pages was not left for Popper to discover, but had long troubled Neurath and Carnap). More importantly perhaps, the very possibility of questioning the (not only) Popperian tendency to attribute philosophically foundationalist ambitions to the Circle even in its early stages is not recognized at all. This complaint connects with the first one: after all, the project of a philosophy of science has been contested territory not only since Kuhn, but throughout this century. Some may find the utter neglect of this metaphilosophical dimension a rather high price to pay for Gillies' admirable introduction to that project as traditionally understood.
COPYRIGHT 1994 Oxford University Press
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1994 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Uebel, Thomas E.
Publication:The British Journal for the Philosophy of Science
Article Type:Book Review
Date:Dec 1, 1994
Previous Article:Scrutinizing Science: Empirical Studies of Scientific Change.
Next Article:Philosophical Naturalism.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |