ABSTRACT
The most widely accepted positivist conception of theories is known as the 'Hypothetico-Deductive' conception; 'deductive' because, like the positivist account of explanation, it represents theories as deductive systems, and 'hypothetico' for reasons which can be clear later. Although these arguments against the instrumentalist interpretation of scientific theories have considerable force against the most radical forms of positivism and empiricism, they are far from decisive as objections to the Hypothetico-Deductive account of theories when it is combined with a much less stringent conception of empirical testability. The duck-rabbit diagram can be argued, in the case of the perceptual gestalt analogue, that some description of, say, the ambiguous figure which is independent of both the duck and the rabbit 'interpretations' can be given. Instrumentalism carries with it a distinctive account of the status of bridge-principles. They must be regarded not as statements of empirical relationships between micro-level processes and macro-level regularities, but as partial definitions or rules of inference.
