Textual traditions and segmentation
For each MS, this database captures what episodes of the core text it contains, in what order and in what version. The key tool articulating this information, and the invention of the KCL DDH team, is the Folio Sequence. Each of our core texts posed different challenges, and the specific solutions adopted are described and explained in the Textual Traditions and Segments sections relating to each text. We recommend that you consult these for any core text before exploring what insights the database can generate.
Segmentation
In order to capture what episodes appear in a specific MS and in what order, we have divided the narratives into sections or 'segments'. Not all MSS contain every episode found in a modern edition of the text. Although some MSS are fragmentary by accident, others select only certain episodes for inclusion, often apparently by design. Conversely, some MSS add supplementary episodes. For pragmatic reasons our segmentation of core texts refers to modern critical editions, but it is important to bear in mind that the texts these editions posit do not always map perfectly on to the texts actually found in extant MSS. For instance, the full range of episodes considered nowadays to belong to the canonical version of Guiron, as established by Lathuillère (1966), is found in no single surviving MS. On the other hand, although the work published as the Old French Roman d'Alexandre (Armstrong 1937-76) did appear in medieval MSS in that form, it was also often embroidered with minor elaborations or had its episodes reordered. Some MSS also interpolate what modern readers consider distinct texts, such as (for the Alexandre tradition) the Venjance Alixandre or the Voeux du paon, inserting them at an appropriate place within the diegetic chronology of the host text. A text's influential status during the Middle Ages may be measured by observing how prone it was both to generate related works and to pull existing ones into its orbit so that they commonly circulated with it.
Our core texts thus belong to what we have called 'textual traditions': clusters of related works, varying internally as well as in their external circumstances. What may be broadly termed cyclicity accounts for much of this phenomenon: extending the chronological scope of a narrative into past or future, linking together semi-independent works, filling in (sometimes retrospectively identified) blanks in the narrative sequence, or adapting, abridging, or amplifying existing episodes. Here again, the distinctive textual tradition that we have identified for each of our core texts reflects modern textual scholarship as well as medieval circulation. For instance, it proved impossible to discuss the prose Lancelot without including the other Vulgate prose Arthurian romances which often accompany it in MSS. From a different perspective, Guiron and the Roman d'Alexandre may be considered not so much as individual texts as already textual traditions, for they present themselves as compiling existing materials into new wholes - a process then continued by various MSS. In contrast, the Roman de Troie's textual tradition consists less of amplifications grafted onto the original stem than of reworkings into prose, sometimes for inclusion in other textual frameworks. Medieval evidence dictates that some textual traditions are rhizome-like, spreading out 'horizontally' through association with one segment or another; others are more 'vertical' and temporal, undergoing modification in time through translation, addition, and embedding into compilations. Both phenomena have a bearing on the question of what version of a text we are looking at (on which more below).
In most cases what we call a 'segment' approximates to a long 'narrative chunk'. Segments are numbered according to an ideal narrative chronology, often that established in modern editions; even episodes which appear only in a few MSS are included in the sequence (see the Segmentation of Tristan). However, some cases needed special treatment. The most distinctive concerns the Roman de Troie: instead of splitting the narrative into chunks, we have recorded the treatment of five short passages which scholarship regards as crucial for determining the specific version of the text at issue; our data advances the state of knowledge on this topic. For the other traditions, we have introduced 'interstitial segments': these represent shorter portions of text that are sometimes (but not always) included within the longer narrative segments, and which may (or may not) have been part of the original. Their particular meaning and function is described in the Segmentation section for each key text. The Histoire Ancienne, for example, is generally considered to be a prose text, however some MSS include verse 'moralizations' which are thought to be original to the work. Some, all, or none of these verse passages may appear in a MS. In order to represent this state of affairs, we have divided the main prose account into 11 sections, and treated each moralization as an 'interstitial segment' separate from the main run of chronological segmentation. We have then noted those cases where each interstitial segment is present (our research shows that they are much more common than previously thought, although often embedded in the surrounding text, or rendered in long lines or as prose). Analogously, in the Lancelot 'interstitial segments' represent interpolated episodes found in only some MSS. In certain cases, what appears in one of our traditions as a text related to the core text (such as the Queste del saint Graal in the prose Lancelot tradition, or Prose 5 in the Roman de Troie tradition) appears in another as an interstitial segment (the Queste is incorporated into many MSS of the Tristan, while Prose 5 replaces the Troy section of the Histoire Ancienne in the latter's second redaction).
Textual Tradition
It is also through segmentation that we address the question of what version of a text is found in each MS. Many segments exist in several versions; these are designated as different 'sources' in our various modes of searching the database.
No two medieval MSS of a single text, of course, are ever quite identical, but not all dissimilarities are equal. In some cases, variations are sufficiently substantial in content, form, or tone that they amount to distinct redactions of the text. We recognize, for instance, two redactions of the Histoire Ancienne and three of Lancelot (our different approach to the Roman de Troie acknowledges the significant critical disagreement over its textual tradition). Nevertheless, a MS that contains a 'pure' redaction is a rarity; most contain versions that scholarship describes as 'mixed', 'contaminated', or 'vacillating'. We have recorded as 'source shifts' those cases where a MS appears to change redaction, whether this occurs at the end of a segment or at some point during it (for example: Grenoble 865). Our findings in several instances nuance or correct previous scholarship.
MSS of the same redaction also disagree among themselves, and where those disagreements allow, sub-groups called textual 'families' are identified. To put this another way: where 2 MSS share what philologists term 'monogenetic innovations' (i.e. the innovative reading is of a sort unlikely to have been arrived at independently by different scribes), then these MSS are taken to be related - or more properly, the relevant sections of these MSS are taken to be related. That is, one may have been copied directly from the other, or both may have been copied from the same 'parent'. On the basis of such shared innovations, scholars construct stemmata, family trees which group extant MSS and conjectured lost intermediaries relative to their descent from an 'archetype' (i.e. the - usually hypothetical - model that presented all the monogenetic innovations shared by the extant witnesses; the archetype should not be confused with the author's original, autograph text). The stemmata for our texts feed into the scholarship which has informed our more narratological, broader approach to segmentation and to textual tradition, however except in the case of the Troie, we have not recorded the different readings (variae lectiones) on which stemmata typically rest.
Searching
Identifying similar patterns of variation in MSS locatable to different places and periods allows us to track the movement of texts in sometimes considerable detail. Alternatively – to start from the other end – sometimes a strictly textual variation allows us to postulate MS travels for which no other evidence may survive. The clearest example here is how the verse moralizations illuminate the paths taken by the Histoire Ancienne, but this database promises further discoveries. Searching by Manuscript will show you not only codicological and paleographical information about the material object, but also what texts it contains, what segments, in what order, and in what version(s). Searching by Textual Tradition enables you to see whether particular configurations of episodes and versions recur in different MSS. Consult the 'Show Similar Manuscripts' feature (at the top of the Folio Sequence): it will display MSS that share the same segments in the same order (Longest Seg. Match) and also MSS that share the same version of those ordered segments (Longest Src./Seg. Match). A high figure in both columns invites further investigation. The connection may be weak, for instance where two MSS contain what is in fact a widespread and common combination of source and segment (such as the interpolation of the Queste into Tristan). Stronger evidence for a relationship is worth probing, since the tool has its limitations. For example, BnF f.fr. 356 does not return the top results for similarity with Arsenal 3477, their longest source-segment sequences in common being only 4 segments long. However, closer inspection of the Folio Sequence for each MS shows that the two share a second source-segment sequence that is nearly as long as the first. Delving into the particular MS records will provide further details of origin and provenance that can be used to map the conjectured circulation of both MSS and text. In this case, the similarities are so extensive that it has been suggested that both MSS came from the same Paris atelier (Delcourt 2009, 205). If, however, as is also possible, one was in fact produced in Arras and the other in the Veneto, then their similarities provide evidence of commerce between these places. In fact, the importance of the traffic between northern France/the southern Low Countries and Italy that the database brings to light has surprised us. Map search offers you visual representations of MSS' connections to places at different periods in their careers.