This thesis explores and makes explicit various aspects of Richard Rorty's rhetorical program for shifting our traditional conceptions of reference and truth. Rorty wants to persuade us to adopt verification (coping) semantics in place of correspondence seeking semantics. I argue against his intuition pumps by considering Keith Donnellan's remarks on description and reference and argue for a view of correspondence truth that is based on what the object, whatever the object, permits us to say. Making this point allows us to see a purposeful conflation in Rorty's work. If beliefs are true because they are justified, Rorty's fallibilistic remark that any of our beliefs may not be true (in the cautionary sense) would follow. But truths may pay because they follow (as "attributive representations") from 'unblocked' objects, or they may just pay. Thus, I suggest that Donnellan preserves William James' remark that we desire correspondence truth, an everyday explanatory notion.