Ingredients Network of Medieval Medicines

In this paper, we incorporate a network-based approach to explore the impact of ingredients in medieval medical recipes, as identified in the most ambitious work of Bernard of Gordon, the Lilium Medicinae. Bernard of Gordon was a medical doctor and lecturer in Montpellier, who completed the Lilium Medicinae, in 1305 AD. The Lilium is an extensive treatise on disease aetiology, diagnostics, personal case studies, and treatment recipes. The text was widely disseminated during the medieval period; the Latin text of the Lilium is extant in several dozen manuscripts and multiple printed editions. Furthermore, it was translated into a range of vernaculars, including Hebrew, French, Spanish, and Irish. There is only one extant translation in English, the Lylye of Medicynes, which survives in a fifteenth century manuscript. There are 359 clearly indicated and formally presented recipes in the extant Middle English manuscript, which in several cases include weights, measures, and clear preparation instructions. Based on the dataset of medieval recipes, as extracted for the needs of our study from the Lylye of Medicynes, we built a bipartite network (Newman, 2003; Albert & Barabasi, 2002; Dorogovtsev et al., 2008) consisting of two types of nodes and one type of ties: i) the 359 medieval recipes, and ii) the 656 ingredients identified in the recipes, while the ties indicated the relationship between ingredients and medicines. Whilst ingredient networks are usually represented as bipartite networks, network analysis is rarely applied to them directly, and most often, such networks are first projected to one-mode ones (Ahn et al., 2011; Teng et al., 2012). Projection is a process that transforms a bipartite network to a one-mode one, which in the case of our dataset occurs by linking together the ingredients that are co-present in the same recipe. Thus, for the analysis of the bipartite network of medicines-ingredients we projected the bipartite network of medicines-ingredients, to a one-mode network of ingredients. In our study we explore the importance of the ingredients of medieval medical recipes through their structural properties in the projected ingredients network. The findings of our study reveal the most significant ingredients from a network perspective, and highlight the need for further research on medieval medicines, and more specifically the systemic and systematic analysis of their ingredients networks. To the best of our knowledge this is the first study exploring the ingredients of medieval medicines through a network-based approach.