Anders Landig

Theories of partial reference have been developed in order to retrospectively interpret rather stubborn past scientific theories like Newtonian dynamics and the phlogiston theory in a realist way, i.e., as approximately true. This is done by allowing for a term to refer to more than one entity at the same time and by providing semantic structures that determine the truth values of sentences containing partially referring terms. Two versions of theories of partial reference will be presented, a conjunctive (by Hartry Field, 1973) and a disjunctive one (by Christina McLeish, 2006). In this paper, I will analyze them with regard to modal and epistemic aspects of those theories. It will be argued that a) theories of partial reference are (surprisingly) compatible with the rigidity of natural kind terms but face a weaker form of the so called “no-failures-of-reference-problem” and b) that the disjunctive account of partial reference suffers from a serious weakness: the impossibility of discriminating between descriptions that fix the reference of a term and those merely associated with it leads to the unacceptable result that past scientific theories containing such partially referring terms will come out as epistemically necessary, i.e., as a priori true.