Faced with information overload, we have no alternative but pattern recognition

An examination of the manuscript tradition of what is often called Robert de Boron's 'Arthurian trilogy" (Joseph d'Arimathie, Merlin and Didot-Perceval) helps us see how modern criticism constructs, from medieval data, meaningful structures that run the risk of being anachronistic and not reflecting the reality of medieval reception. In its manuscripts, the so-called trilogy is presented much more often as a binomial Joseph-Merlin, with Perceval appearing in only two very unusual witnessings. Moreover, this third text, presented as an autonomous romance by modern Criticism, functions much more like a suite of Merlin, perhaps a forerunner of the other suites this romance received between 1230 and 1250. The difference between the tripartite description usually given and its manuscript realizations is not just a matter of quantification: it has an important impact on our interpretation of the emergence of prose around 1200, of the development of the cyclical form and of the literary posterity of these works. The aim of this article is not so much to "rectify" the traditional representation of these three texts in favour of a more "correct" version, but to highlight the vagaries of pattern recognition, which is the means by which the philologist constructs his object, sifting through a surfeit of data: the modern reader, faced with an ancient object, is in a constant process of reconstruction and negotiation.