A somewhat simplified version of Jerrold J. Katz’s theory of the analytic/ synthetic distinction for natural languages is presented. Katz’s account is criticized on the following grounds. (1) the ‘antonymy operator’ is not well defined; it leaves certain sentences without readings. (2) The account of negation is defective; it has the consequence that certain nonsynonymous sentences are marked as synonymous. (3) The account of entailment is defective; it has the consequence that analytic sentences entail synthetic ones. (4) Katz’s account of “indeterminable sentences” is criticized; it has the consequence that certain logical truths are not marked as analytic. (5) Katz’s semantics provides no account of truth, so that he is unable to show that analytic sentences are true and that ‘indeterminable’ sentences are not.