Abstract
Bernard Bolzano's most fruitful invention was his method of variation. He used it in defining such fundamental logical concepts as logical consequence, analyticity and probability. The following three puzzles concerning this method of variation seem particularly worth considering, (i) How can we define the range of variation of an idea or the categorial conformity of two ideas without already using the concept of variation? This question was raised by Mark Siebel in his M. A. thesis, (ii) Why must we define analyticity by means of (simultaneous or successive) variation of several ideas rather than by means of replacing a single idea? This problem is suggested by an example due to W.V.O. Quine, John R. Myhill and Benson Mates, (iii) Must every 'there is...' sentence be synthetic for Bolzano, as his pupil Franz Příhonský claims in his booklet Neuer Anti-Kant, or can a 'there is...' sentence be logically analytic?