Characterizing scientific kinds. The role of generic generalizations in science.

Generics are generalizations that are not explicitly quantified, like ‘Dogs bark’ for example. Although they appear to be very simple generalizations, their semantics are very complex. To see this, consider that a generic can be true even though only a small minority of the kind instantiates the property (e.g. ‘Ticks carry Lyme disease’) and can be false even though the majority does instantiate the property (e.g. ‘Humans are right-handed’). To account for this complexity, the best semantic theories for generics hold that they can be true in several different ways.
Scientists also use generic generalizations when they characterize the kinds in their domain, stating for example that ‘Metals are excellent conductors of heat’. Philosophers of science, however, have always analyzed such sentences as 'ceteris paribus' laws, even when this clause is not stated. This research project starts from the observation that scientists do regularly use generalizations with generic meaning to characterize kinds, and aims to describe and evaluate the scientific role of such generalizations. For they immediately pose a puzzle. If generics can be true in different ways, how can such statements play any theoretical or communicative role in science? I will investigate the hypothesis that this can be explained based on an innate cognitive heuristic that we all use when reasoning about the characterizing properties of kinds