ing expediency can pose a pitfall, leading us beyond objective evidence evaluation and obscuring a broader question: how much do I want the idea to be confirmed, for other reasons?
***
Like all values, these seven are “effective guidance in the presence of conflict and equivocation” [Kuhn, 1977], not rigid criteria that dictate an unambiguous conclusion. “The criteria of choice . . . function not as rules, which determine choice, but as values, which influence it.” Like all values, these differ among individuals. Thus the disagreements between scientists about a hypothesis do not imply that one has misinterpreted data or made an error. More likely, they employ different subjective weightings of conflicting evidence. In Chapter 6, I argued that such disagreements are actually scientifically healthy and that they are an efficient means for advancing science; group objectivity grows from individuals’ subjectivity.
Scientific values differ between fields, and they may evolve within a field. For example, engineers and applied scientists emphasize the value of social utility as a key evaluation criterion, and they differ with physicists concerning the relative value of fruitfulness and simplicity. Kuhn [1977] notes that quantitative accuracy has become an evaluation criterion for different sciences at different times: it was achievable and valued by astronomy many centuries ago; it reached mechanics three centuries ago, chemistry two centuries ago, and biology in this century.
Like all human values and unlike rules, the scientific values are implicitly imprecise and often contradictory. For example, more complex hypotheses are usually more accurate than simple ones, and hypotheses with a narrow scope tend to be more accurate than those with a broad scope. Even a single value such as accuracy may have contradictory implications: a hypothesis may be more accurate than a competing idea in one respect and less accurate in another, and the scientist must decide which is more diagnostic.
An extreme example of the conflict between values is the quantum mechanics concept of complementarity, which achieves utility and expediency by abandoning consistency. According to complementarity, no single theory can account for all aspects of quantum mechanics. Concepts such as light as waves and light as particles are complementary. Similarly, the concepts of determining position precisely and determining momentum precisely are complementary. Furthermore, the concept of determining location in space-time is complementary to the concept of determinacy. In each case the pair of concepts is apparently contradictory; assuming one seems to exclude the other in an individual experiment. But full ‘explanation’ requires both concepts to be embraced, each in different situations. Complementary concepts only seem to be contradictory, because our perceptions are unable to reconcile the contradictions. The actual physical universe, independent of our observing process, has no such contradictions.
***
Evaluation Aids
Scientific progress depends on proper appraisal of evidence, on successful rejection of incorrect hypotheses and adoption of correct (or at least useful) hypotheses. Yet the evaluation techniques employed most often are incredibly haphazard, leading to conclusions such as ‘sounds reasonable’ or ‘seems rather dubious’.
Evaluation of evidence is a scientific skill, perhaps the most important ability of a successful scientist. Like any skill, its techniques must be practiced deliberately and systematically, before one
