Emphasis on autonomy of texts presupposes that there are perennial concepts. But researchers' expectations may turn history into mythology of ideas; researchers forget that an agent cannot be described as doing something he could not understand as a description, and that thinking may be inconsistent. They will never uncover voluntary oblique strategies and by treating ideas as units will confuse sentences with statements. On the other hand, a contextual approach to the meaning of texts dismisses ideas as unimportant effects. Neither (...) method shows how what was said was meant is crucial for understanding. There are no perennial problems; philosophers of different times do not speak directly to us. But history of ideas helps us recognize the contingency of many of our beliefs. (shrink)

Arthur Lovejoy’s history of unit-ideas and the history of concepts are often criticized for being historically insensitive forms of history-writing. Critics claim that one cannot find invariable ideas or concepts in several contexts or times in history without resorting to some distortion. One popular reaction is to reject the history of ideas and concepts altogether, and take linguistic entities as the main theoretical units. Another reaction is to try to make ideas or concepts context-sensitive and to see their histories as (...) dynamic processes of transformation. The main argument in this paper is that we cannot abandon ideas or concepts as theoretical notions if we want to write an intelligible history of thought. They are needed for the categorization and classification of thinking, and in communication with contemporaries. Further, the criterion needed to subsume historical concepts under a general concept cannot be determined merely on the basis of their family resemblances, which allows variation without an end, since talk of the same concepts implies that they share something in common. I suggest that a concept in history should be seen to be composed of two components: the core of a concept and the margin of a concept. On the basis of this, we can develop a vocabulary for talking about conceptual changes. The main idea is that conceptual continuity requires the stability of the core of the concept, but not necessarily that of the margin, which is something that enables a description of context-specific features. If the core changes, we ought to see it as a conceptual replacement. (shrink)