Abstract

In [Laurence, Margolis 2003] the authors try - within their polemics against F.Jackson’s views in [Jackson 1998] - to decide the question whether concepts are a priori (in their formulation “to be defined a priori”). Their discussion suffers - as a number of similar articles - from a typical drawback: some problem whose solution requires an exact notion of concept is handled as if the latter were quite clear. The consequence of this ‘conceptual laxity’ is that a) the topic of the discussion is not very clear (what does the phrase ‘concepts must be defined a priori’ mean?); b) the relevance of the Quinean criticism of the “second dogma of empiricism”, i.e., of Quine’s claim that “science sometimes overturns our most cherished beliefs” and therefore there is no sharp boundary between analytic and synthetic is uncritically accepted; c) no distinction is made between the question whether the relation between an expression and its meaning is a priori and the question whether the relation between a concept and the object identified by the concept is a priori. The present article intends to elucidate and then to answer the questions that can be asked when we say something like “concepts are a priori ”