A COMPARATIVE STUDY of sources written in different languages encounters problems reminiscent of those described by Elizabeth Brown and Susan Reynolds in their critique of the classic theory of feudalism. The two catchphrases capturing the essence of this critique are “the tyranny of a construct,” and “the confusion of words, concepts, and phenomena.” The confusion, according to Reynolds, results from the way historians typically proceed: first, they employ one word, “fief,” to translate many different medieval terms; then they define fief as a concept of “dependent noble or military tenure.” Finally, they assume the existence of the phenomenon corresponding to this concept every time they see words conventionally translated as “fief.” This habit of “starting our investigation of phenomena by focusing on particular words” leads to circular argumentation that distorts the realities of the past.

Shedding the habit, however, is not simply a matter of scholarly integrity. Historians behave in this way not necessarily out of prejudice or sloppiness, but because it is often hard to come up with an alternative. Reynolds admits that “historians who work from written sources have to begin with words: they are all we have”; she just calls for thinking hard about what is being discussed (the phenomena) before coming up with generalizations. To avoid generalizations entirely and to adhere strictly to concepts and notions found in medieval texts, one must also avoid the use of any modern language and write in Latin or whichever languages are used in the sources. Indeed, recent works on medieval social and political history contain almost as many words in italics as not: they are peppered with milites, homines, fideles, suis, benefitia, feva, casamenta, castra and the like, whereas earlier scholarship would have used “vassals,” “fiefs,” or “castles.” Nowadays, historians are careful not to to distort the actual content of their sources by bringing in all the theoretical baggage carried by terms that had been used to describe “feudal society.” They know that a vassalus was not necessarily a vassal in the textbook sense of the word, and that there is no reason to think that every homo did homage and every fidelis swore an oath of fealty.

Sticking with the original terminology of the sources may work well for research that concentrates on one linguistic area. Scholars using medieval Latin texts can productively discuss accounts of homines receiving casamenta and performing, or not performing, homagium.