Fishery science is usually perceived by its practitioners as being a critical and quantitative activity, deeply dependent on mathematical analysis; indeed, the introduction to a well-known text on fisheries science suggests that only those who are comfortable writing computer programs or playing with numbers should become involved in fisheries management. This blunt statement demonstrates what went wrong with the science: it forgot that it is heavily dependent on two other disciplines – biology and ecology – in which numerical predictions are quite often unsatisfactory. Consequently, there is a fundamental contradiction between the potential capability of fishery science and its stated task of making routine and quantitative predictions concerning the effects of specified levels of fishing on a stock of fish.

Biology is notorious for its lack of predictive theory, and for its high content of inductive and a-priori generalisations that are based on simple observation of nature. As Murray noted, ‘the fact that biology lacks . . . universal laws and predictive theory . . . poses a serious problem for both biologists and philosophers’. Nevertheless, it is possible to deduce simple biological laws and to verify them by the prediction of ecological observables: I shall discuss one such example in Chapter 2. Such laws may be used to falsify theories that have been arrived at by inductive methods.