Vanessa Seifert (University of Bristol), Geoffrey Blumenthal (University of Bristol), James Ladyman (University of Bristol) - Reference to chemical kinds is often considered in the case of terms such as ‘gold’ and ’water.’ This paper considers the history of reference to types of colourless air, which raise particular issues for the various theories of reference. Many examples of such reference are given from the primary literature. It is shown that in the cases of most of the examples, reference to a gaseous substance started using descriptions of a method of production of the gas and of at least one related distinctive observable property, but without the gas being named. In cases in which initial descriptions were deemed to be vague, more descriptions of distinctive observable properties were added as desired. In most of the examples, such descriptions of a gaseous substance were stated prior to theories about the inner constitution of a gas, or about the dispositions of a gas concerning reactions with other substances, being stated. When names were added, they were generally less definite than the descriptions, which in several cases caused problems. In each of the examples, more than one researcher gave a name to the same gas. While some researchers routinely used terms and substance names belonging to a single type of theory, the examples show other cases in which a researcher adhering to one theory wrote to or concerning the work of a researcher adhering to another theory, using more than one name for a specific substance interchangeably, even though terms within those names formed part of the terminology of different theories. That is for example, the substance name ‘dephlogisticated air’ translated virtually exactly with other names for ‘oxygen’, although ‘dephlogisticated’ did not translate into an antiphlogistic term. This paper argues that in the history of chemistry, reference to a gas by one chemist was intended to give others epistemological, methodological and practical access to it. This paper argues that, generalising from the examples, there were four aspects of the reference that took place in practice: it included or was generally intended to include definite descriptions, which were chosen to refer uniquely or as uniquely as was practicable, to enable a user of the descriptions to know whether they were referring uniquely, and to enable a user of the descriptions to know whether a specific substance was what the descriptions described. Reference to the gas was generally subsequently continuous, and this was knowable in all these ways by researchers due to the initial descriptions used in reference. Such reference underlaid the interchangeable use of names for substances using terms from different theories which is seen in the examples. This paper argues that there are implications of the cases discussed for theories of reference, incommensurability and realism, which deserve further exploration.
