It is often presumed that empirical considerations provide epistemic objectivity for claims about the boundaries and classification of scientific categories. This has seemed especially plausible in chemistry. Focusing on the category chemical element, we describe two 20^th century developments that undermine epistemic objectivism about it. But our second thesis is that, in practice, this shortfall is bridged by relying on a little-recognized species of pragmatic norm: classificatory norms. We contend this precludes the objectivity, yet ironically affords the rationality, of related category and classification claims.

There is an ongoing debate about what chemical bonds are and whether they exist. Based on Daniel Dennett's seminal paper Real Patterns, I argue that chemical bonds are patterns of sub-molecular interactions. This proposal resolves the problems that have been raised in the context of existing understandings of the chemical bond. Moreover, it provides a novel perspective through which one can defend the reality of chemical bonds, but also through which one can reexamine structural realism.

The nature of ammonia occupied Humphry Davy for five years. Despite the lack of a clear outcome, Davy's work on ammonia is interesting because it illustrates his use of analogy. The aim of this paper is to show that Davy's analogical reasoning on ammonia revolved around the idea that similarities in chemical properties indicated a similar composition. Davy used the analogies between ammonia and metallic compounds, on the one hand to speculate on the metallic nature of ammonium (a possible metallic component of ammonia), and on the other hand to speculate on the internal composition of all metals.