What is the status of scientific truth-claims? Can they purport to hold good for all time across vastly differing contexts of language, culture, and society? That is to say: is science in the business of providing valid explanations of physical objects and events whose nature remains constant despite such deep-laid shifts of cultural perspective? Or is it not rather the case—as currently argued by relativists, pragmatists, and ‘strong’ sociologists of knowledge—that those contexts provide the only means of understanding why science has taken such diverse forms (and come up with such a range of competing ‘truths’) throughout its history to date?footnote1

These questions are of interest not only to philosophers and historians of science but also, increasingly, to cultural and critical theorists influenced by the widespread ‘linguistic turn’ across various disciplines of thought.footnote2 They are often linked with the issue of ontological relativity, that is, the argument—deriving principally from W.V. Quine’s famous essay ‘Two Dogmas of Empiricism’—that there exist as many ways of describing or explaining some given phenomenon as there exist ontological schemes or systems for redistributing predicates over the entire range of sentences held true at any particular time.footnote3 On this holistic account there is no means of drawing a firm, categorical line between synthetic and analytic propositions, or matters of factual (contingent) truth which might always be subject to revision in the light of further evidence, and on the other hand those so-called logical ‘laws of thought’ whose truth is assumed to be a matter of a priori necessity and hence—by definition valid—for all possible contexts of enquiry.footnote4 And with the collapse of this distinction, so Quine argues, we must also relinquish the idea that philosophy of science might yet come up with an adequate method for linking observation-sentences to theories (or vice versa) through a clear-cut set of logical procedures. For in a holist perspective those sentences can possess meaning—that is to say, be assigned determinate truth-values—only as a function of their role within the entire existing ‘fabric’ or ‘web’ of beliefs, or the entire set of truth-claims (‘empirical’ and ‘logical’ alike) that currently happen to command widespread assent. Which is really to say that there are no such determinate truth-values, since theories are always at some point ‘underdetermined’ by the best evidence to hand, while that evidence is always ‘theory-laden’—or committed to some prior ontological scheme—right down to the level of its basic data as given in first-hand observation-sentences. Thus for Quine it follows that one must apply a principle of strict ontological parity as between (for instance) Homer’s gods, centaurs, numbers, set-theoretical classes, and brick houses on Elm Street. Any preference in the matter—and Quine admits readily that he has a whole range of such preferences—must in the end come down to one’s particular choice of ontological scheme.footnote5

There are many other sources of this relativist trend in contemporary philosophy of science. They include Thomas Kuhn’s highly influential account of the way that science alternates between periods of ‘normal’ and ‘revolutionary’ activity, the former characterized by broad agreement on what counts as a proper (constructive and disciplined) approach to certain well-defined problems, the latter by a sense of impending crisis—and an absence of agreement on even the most basic principles—which heralds the transition to a new epoch.footnote6 Here as with Quine it is taken for granted that all the components of a given scientific ‘paradigm’—from observation-sentences to high-level theories—are intelligible only in terms of the prevailing consensus, or according to the overall framework of beliefs that provides its own (strictly immanent) criteria of truth, progress, theoretical consistency, evidential warrant and so forth. But it then becomes difficult—if not impossible—to explain how we could ever gain insight into scientific world-views other than our own, or again, how historians of science could ever claim to understand the reasons (i.e. the scientific grounds) for some decisive paradigm-shift, as distinct from the various short-term cultural, social or historical factors that may have played some part in bringing them about. Hence Quine’s recourse to the idea of ‘radical translation’ as a means of (purportedly) bridging this otherwise insuperable gulf between different observation-languages or ontological schemes.footnote7 Hence also the difficulties that Kuhn confronted in his 1969 Postscript to The Structure of Scientific Revolutions when responding to his critics on the issue of relativism and its self-disabling consequences. For it is far from clear that these difficulties are in any way resolved by his Quinean (radical-empiricist) line in the face of such strong counter-arguments.footnote8

The problem is yet more acute with those kinds of ultra-relativist position adopted by proponents of the present-day ‘linguistic turn’ in its full-fledged (postmodern) guise. Thus it is sometimes claimed—as for instance by Richard Rorty—that our best model for interpreting the process of scientific paradigm-change is what happens when poets and novelists come up with striking new ‘metaphors we can live by’, or again, when strong-revisionist literary critics interpret such metaphors after their own fashion.footnote9 Then again there are those—the late Paul Feyerabend chief among them—who espouse an anarchistic philosophy of science which rejects all appeals to truth, logic, reason, consistency, experimental proof, and the like.footnote10 On this view the idea of scientific ‘progress’ is nothing more than a piece of bogus mythology, one that takes hold through our myopically equating ‘truth’ with what currently counts as such according to this or that (self-authorized) ‘expert’ community. Much better, Feyerabend thinks, to have done with this misplaced reverence for science and instead take account of the various factors—social, political, psychological, careerist and so forth—which have always played a decisive role in the history of scientific thought. For we can then see how mixed were the motives (and often how random or opportunist the methods) which gave rise to some so-called ‘discovery’ or ‘advance’ that is nowadays treated as a text-book example of its kind. And this will bring two great benefits, as Feyerabend sees it. First, it will help to demythologize science—to remove some of its false prestige—and thereby open it up to criticism from other (i.e. non-‘expert’ but socially and ethically more responsible) quarters. Second, it will encourage scientists to become more adventurous in framing risky conjectures or in pursuing novel and heterodox lines of thought.

There are various explanations that might be adduced for the current appeal of such ideas. One is the widely-held view that philosophy of science can no longer have recourse to any version of the logical positivist (or logical empiricist) distinction between truths-of-observation on the one hand and self-evident (tautologous) truths of reason on the other.footnote11 There are similar problems—so it is argued—with the appeal to deductive-nomological (or covering-law) theories, those that would seek to account for observational data by bringing them under some higher-level (metalinguistic) order of logical entailment-relations.footnote12 For here again the way is open for sceptics like Quine to argue that any such distinction will always be drawn according to some preferred ontological scheme, some languageor culture-specific set of descriptive or explanatory priorities. One alternative that has enjoyed wide favour, not least among practising scientists, is Karl Popper’s hypothetico-deductive account whereby the measure of a theory’s claim to genuine scientific status is not so much its truth as established by the best current methods of experimental testing but its openness to falsification by those same methods.footnote13 This account has the signal advantage of explaining how a great many scientific theories that once enjoyed widespread credence should eventually have turned out mistaken, or—as with Newton’s conceptions of absolute space and time—‘true’ only relative to a certain restricted domain. It thus meets the criticism of those, like Feyerabend, who would exploit such evidence to the point of denying that notions of truth have any role to play in the history and philosophy of science.