MEANING VARIANCE AND THE COMPARABILITY OF THEORIES

The central problem to emerge from the controversy between Positivism and the 'new empiricism' of Thomas Kuhn and Paul Feyerabend is the problem of the comparability of different scientific theories. Its solution promises to elucidate the sense in which theories may be said to compete and the justification for preferring one theory to another, and thus provide a satisfactory analysis of the notion of scientific progress. But the interest which this problem commands is actually quite recent and results at least in part from a peculiarity of the controversy. Both of the opposing philosophical positions are concerned primarily with the relations between theory and observation, and neither sees comparability as a genuine problem. Nevertheless, the two positions lead to radically different views of comparability, and in fact the debate between them can be fairly well expressed in terms of these views. Thus for Positivism there is no question of analysing the relations between competing theories. It is required only to invent a procedure for comparing theories. And this invention is an immediate consequence of a solution to the problem of confirmation. For Kuhn and Feyerabend the emphasis is on the autonomy and complete empirical adequacy of individual theories. Different theories are held incomparable. And although Feyerabend stresses the importance of theoretical competition, his interest is in competition between equally adequate theories, and in his major writings he treats such competition as completely unproblematical. To the criticism that theories cannot be at once incomparable and competing, Feyerabend has responded by upholding incomparability. One or the other of these positions on comparability is inescapable if one follows Positivism or Kuhn and Feyerabend in upholding a rigid view of the relation between observation and theory. The independence of observation from theory guarantees that theories using the same observation terms be comparable. The dependence of observation upon theory precludes the comparability of theories. For the positivist there are terms which necessarily have the same meaning in different theories. For Feyerabend no term can possibly have the same meaning in different theories. Thus there can be no problem of