Gregory of Tours, Historiae

Displayed to the right are two networks drawn from our data for Gregory of Tours' Ten Books of History (Decem Libri Historiarum), completed before 594. The top graph shows just the hostile connections Gregory describes in his work; the bottom graph shows just the kinship connections (excluding the category of marriage).

Blue nodes = male characters; magenta nodes = female characters, green nodes = groups of mixed gender; circle nodes = named characters, square nodes = nameless characters. The network is based on the number of connections characters had (= Degree centrality).

For our definition of hostile connections and kinship connections see here.

In the network of hostile connections the character in the middle of the central dense cluster is Guntram, Merovingian king of Orléans from 561 to 592. The two named characters are Fredegund (d. 597), the queen consort of Chilperic, king of Neustria and Guntram's brother, and Brunhild (d. 613), queen consort of Sigebert, king of Austrasia and also Guntram's brother. Fredegund is the third most connected character in this network, and Brunhild is the tenth most connected. Normally women being so prominent on centrality measures, especially those calculating hostility, is unusual in the texts we are studying in this project. It is however not so surprising in this case, as our results confirm the dominant role Fredegund and Brunhild play in Gregory's Histories, as is apparent from analogue reading. We are planning to investigate further their different rankings by degree centrality and compare this with their respective performance on other centrality measures.

The network of kinship connections that Gregory describes is unusually fragmented, also compared with the hostility network. The densely connected clusters are kinship groups for whom ancestry was significant for family identity. The cluster in the centre represents the Merovingian royal families, while the cluster above it are the patriarchs of the Old Testament, beginning with Adam. The network does not include marriage connections, however. Inserting these into the calculation may make it less disparate. This is something we would like to investigate further too.