All scholars who work with manuscripts must confront the question of their dating. But in most cases, this is a preliminary operation that is tackled when the stage is being set. Once dealt with, the problem is forgotten as one gets on with the main subject of the research, whether it is philological, historical, artistic, or something else. By virtue of its use as an auxiliary activity, the operation ends up seeming minor or marginal, while in reality it is vital. In order to simplify matters, I shall treat only the question of dating in a temporal framework and from a paleographer's point of view. But my considerations also apply, mutatis mutandis, to the localization of manuscripts in space and to the analysis of their decoration.
Why must we date manuscripts? This is a naive, even trivial question, but it is here that we must start if we are to proceed methodically. The answer is just as elementary. Manuscripts are historical objects, and as such they cannot be correctly interpreted unless they have been placed in their proper chronological and geographic context. But it is equally possible to reverse the relation and consider that each manuscript is an archaeological object that carries information about the milieu that produced it, a milieu that needs to be defined in part by its position in space and time. The difference between the two aspects of that relationship, which one can refer to as a passive relation and an active relation respectively, is essential but subtle. It seems to escape most scholars, namely those for whom the history of the book and writing is not a chief concern.
These scholars tend to consider only the first aspect of the relation, and do so in a very restrictive manner. I mean that they are usually interested in dating the manuscripts they are working with only if that dating has a direct impact on the research at hand; they do not try to go beyond the degree of precision immediately useful to this framework. The approach is shortsighted and extremely self-centered. For we lose the possibility of exploiting the manuscript for other purposes or simply using it in another context, and we lose all sorts of potential information conveyed by the manuscript in question.
I would like to illustrate the consequences of this unjustified negligence by examining the almost farcical example of MS 444 at the Bibliotheque municipale in Laon. (1) The book contains an enormous Greco-Latin glossary, a major witness to the study of Greek in the ninth century, and is thus of extraordinary importance to the history of Carolingian culture. Part of the manuscript was published in 1887 in the Monumenta Germaniae historica. The edition is preceded, as is usual, by a very brief and erudite introduction. In an early note we learn that the manuscript was written in part by Martin of Ireland, who died in 875 according to the annals. On the following page, another note indicates that the book contains a funeral oration for the wife of Charles the Bald, the empress Ermentrude, who died in 869. In other words, for more than a century anyone could deduce by comparing the two notes that the manuscript had to have been copied during a six-year span between 869 and 875, a degree of precision that is rare for the period. Strange as it may seem, the comparison of these two notes has never been made, even though the manuscript has been cited in countless articles. All the authors refer simply to a manuscript of the "ninth century," or at best "of the second half of the ninth century." And the consequences of this unpardonable negligence have sometimes been rather regrettable, because a certain number of hypotheses that have been formulated about this book are simply untenable if one takes its exact dating into account. Moreover, within the framework of such restrictiveness, we are the victims of preconceptions and habits that have maintained precision in dating at a very low level, at least for general use. …