In glossed books, several practices of knowledge selection can be seen in action. In fact, selection is one of the key functions of marginal annotations according to the four s’s of information management mentioned by Anne Blair in her monograph Too Much to Know, on early modern book practices. That these practices are not an invention of the early modern period but a continuation of a whole set of older techniques can be easily illustrated with early medieval manuscripts, with their nota signs, cut-and-paste signs, marginal keywords and summaries, etc. But in the annotated books of the early Middle Ages, the more hidden kind of selection at which the introductory text for this workshop hints is also abundantly present: in each copy with annotations, a selection is made from previous authorities, from existing sets of glosses and from texts that were used to create new glosses. The rich dynamic that results has boggled the minds of traditional philologists, but is exciting material for intellectual historians. It shows the “entangled intellectual networks” that this workshop seeks to analyse. Yet it is not easy to interpret the material evidence in such a way that these narratives of entanglement can be laid bare. I intend to show some examples to illustrate both possibilities and obstacles of the approach.