My statement of the well-known result which does follow from Godel's theorem was rather redundant: (correction follows) There is a sentence which is not provable in PA but which is true in the standard model of PA as defined in ZFC (because provable in ZFC of that structure).