Science, philosophy of

DOI: 10.4324/9780415249126-Q120-1
Version: v1,  Published online: 1998
Retrieved July 17, 2024, from

2. Contemporary philosophy of science: the theory of scientific knowledge

Scientists propose theories and assess those theories in the light of observational and experimental evidence; what distinguishes science is the careful and systematic way in which its claims are based on evidence (see Scientific method). These simple claims, which I suppose would win fairly universal agreement, hide any number of complex issues.

First, concerning theories: how exactly are these best represented? Is Newton’s theory of gravitation, or the neo-Darwinian theory of evolution, or the general theory of relativity, best represented – as logical empiricists such as Carnap supposed – as sets of (at least potentially) formally axiomatized sentences, linked to their observational bases by some sort of correspondence rules? Or are they best represented, as various recent ‘semantic theorists’ have argued, as sets of models (see Models; Theories, scientific)? Is this simply a representational matter or does the difference between the two sorts of approach matter scientifically and philosophically? This issue ties in with the increasingly recognized role of idealizations in science and of the role of models as intermediates between fundamental theory and empirical laws (see Campbell, N.R.; Idealizations). It also relates to an important issue about how best to think of the state of a scientific field at a given time: is a scientist best thought of as accepting (in some sense or other) a single theory or set of such theories or rather as accepting some sort of more general and hierarchically-organized set of assumptions and techniques in the manner of Kuhnian paradigms or Lakatosian research programmes? It seems likely that arriving at the correct account of scientific development and in particular of theory-change in science will depend on identifying the ‘right’ account of theories.

Next concerning the evidence: it has long been recognized that many of the statements that scientists are happy to regard as ‘observation sentences’ in fact presuppose a certain amount of theory, and that all observation sentences, short perhaps of purely subjective reports of current introspection, depend on some sort of minimal theory (even ‘the needle points to around 5 on the scale’ presupposes that the needle and the scale exist independently of the observer and that the observer’s perception of them is not systematically deluded by a Cartesian demon). Does this mean that there is no real epistemic distinction between observational and theoretical claims? Does it mean that there is no secure basis or foundation for science in the form of observational and experimental results (see Observation)? If so, what becomes of the whole empiricist idea of basing scientific theories on the evidence? It can be argued that those who have drawn dire consequences from these considerations have confused fallibility with (serious) corrigibility: that there are observation statements, such as reports of meter readings and the like, of a sufficiently low level as to be, once independently and intersubjectively verified, not seriously corrigible despite being trivially strictly fallible (see Measurement, theory of). Aside from this issue, experiment was for a long time regarded as raising barely any independent, philosophical or methodological concern – experiments being thought of as very largely simply means for testing theories (see Experiment). More recently, there has been better appreciation of the extent to which experimental science has a life of its own, independent of fundamental theory, and of the extent to which philosophical issues concerning testing, realism, underdetermination and so on can be illuminated by studying experiments.

Suppose that we have characterized scientific theories and drawn a line between theoretical and observational statements, what exactly is involved in ‘basing’ theoretical claims ‘systematically and carefully’ on the evidence? This question has of course been perhaps the central question of general philosophy of science in this century. We have known at least since David Hume that the answer cannot be that the correct theories are deducible from observation results. Indeed not only do our theories universally generalize the (inevitably finite) data as Hume pointed out, they also generally ‘transcend’ the data by explaining that data in terms of underlying, but non-observable, theoretical entities. This means that there must always in principle be (indefinitely) many theories that clash with one another at the theoretical level but yet entail all the same observational results (see Underdetermination). What extra factors then are involved over and beyond simply having the right observational consequences? What roles do such factors as simplicity (see Simplicity (in scientific theories)), and explanatory power (see Explanation), play in accrediting theories on the basis of evidence? Moreover what status do these factors have – are they purely pragmatic (the sorts of features we like theories to have) or are they truth-indicating, and if so why? Some have argued that the whole process can be codified in probabilistic terms – the theories that we see as accredited by the evidence being the ones that are at any rate more probable in the light of that evidence than any of their rivals (see Confirmation theory; Inductive inference; Probability theory and epistemology).

Finally, suppose we have characterized the correct scientific way of reasoning to theories from evidence, what exactly does this tell us about the theories that have been thus ‘accredited’ by the evidence? And what does it tell us about the entities – such as electrons, quarks, and the rest – apparently postulated by such theories? Is it reasonable to believe that these accredited theories are true descriptions of an underlying reality, that their theoretical terms refer to real, though unobservable entities? (Or at least to believe that they are probably true? or approximately true? or perhaps probably approximately true?) More strongly still, is any one of these beliefs the uniquely rational one? Or is it instead more, or at least equally, reasonable – at least equally explanatory of the way that science operates – to hold that these ‘accredited’ theories are no more than empirically adequate, even that they are simply instruments for prediction, the theoretical ‘entities’ they involve being no more than convenient fictions (see Conventionalism; Fictionalism; Incommensurability; Putnam, H.; Scientific realism and antirealism)? One major problem faced by realists is to develop a plausible response to once accepted theories that are now rejected either by arguing that they were in some sense immature – not ‘fully scientific’ – or that, despite having been rejected, they nonetheless somehow live on as ‘limiting cases’ of current theories (see Alchemy; Chemistry, philosophical aspects of §2; Field theory, classical; Mechanics, Aristotelian; Mechanics, classical; Optics; Vitalism).

Clearly an antirealist view of theories would be indicated if it could convincingly be argued that the accreditation of theories in science is not simply a function of evidential and other truth-related factors or even of epistemic pragmatic factors, but also of broader cultural and social matters. Although such arguments are heard increasingly often, many remain unconvinced – seeing those arguments as based either on confusion of discovery with validational issues or on fairly naïve views of evidential support (see Constructivism; Discovery, logic of; Gender and science; Marxist philosophy of science).

Citing this article:
Worrall, John. Contemporary philosophy of science: the theory of scientific knowledge. Science, philosophy of, 1998, doi:10.4324/9780415249126-Q120-1. Routledge Encyclopedia of Philosophy, Taylor and Francis,
Copyright © 1998-2024 Routledge.

Related Searches


Related Articles