Identity Criteria and Sortal Concepts

In this paper we focus on a specific aspect of the notion of conceptualisation, i.e. on the issue of the specification of a certain kind of concept: sortal concept. Our starting point is the intuitive idea that a sortal concept cannot be specified in isolation from a general notion of entity. We think that this idea has some bearing on the way in which identity criteria should be conceived, since they are usually taken as a fundamental tool for the specification of a sortal concept. The first goal of our paper is to discuss and point out some difficulties concerning the relation between sortals and identity criteria. In general, we think that the specification of a sortal presupposes - in Russellian terms - a range of significance on which the concept is defined. It follows that the sortal cannot be stated without a specification of its range of significance. This means that identity criteria for a sortal K providing conditions of identity only for objects falling under K are not enough to specify K. Independently of this point, we are quite skeptical about the possibility of achieving a formal satisfactory definition of sortal. We will try to show that even Guarino and Welty's last proposal [4] does not succeed. Our second goal concerns which concepts are to be taken as sortals. By thinking that identity criteria are necessary and sufficient conditions for identity, Guarino and Welty cannot accept that, say, a is S1 and a is S2, in the case that S1 and S2 are associated with incompatible criteria. So they are lead to postulate, for example, that not a but b is S2 and a is constituted by b. We defend the thesis that it is possible to take some concepts - in our case S2 - as endowed only with necessary conditions for identity. We will argue for some specific choices.