The Book of Llandaf, to use the Welsh spelling (Llandaff in English), or Liber Landavensis (Aberystwyth, National Library of Wales ms 17110E, digitised on the NLW website), was compiled in the first half of the twelfth century in the diocese of Llandaff in Glamorgan, south Wales. Described by the historian Wendy Davies (Reference Davies1979), who published some of the most influential modern studies of the manuscript, as ‘part-cartulary, part-history, part-register’, the manuscript is among the earliest surviving from Wales and is often invoked as a supplement to Domesday Book by scholars investigating Norman settlements in south-east Wales in the decades following the Conquest.

Yet the manuscript has also been controversial among modern historians, not all of whom have been convinced by the historicity of its contents. Written over a period of time, from about 1120 to 1134, in a number of hands, mainly in Latin but with some important Old Welsh and Norman French forms, the primary purpose of the book was to establish the boundaries of the diocese of Llandaff since its supposed foundation in the fifth century. The man who wanted thus to set Llandaff up as an ancient and extensive episcopal see in order to fend off competition from St Davids in the west and Hereford to the east was Urban, bishop of Glamorgan from 1107 until his death in 1134. The Book of Llandaf was created by Urban to lend support to his ambitions, and, like a number of other monastic histories devised for similar propaganda purposes, was based on an essentially spurious claim to ancient land rights. This is what led some twentieth-century historians, including Christopher Brooke, to classify the manuscript as an example of a twelfth-century forgery.

Patrick Sims-Williams, a leading Celticist and one of the foremost scholars of early Welsh literature and history, has taken up the task, begun by Wendy Davies, of recuperating the Book of Llandaf as an important and valid source of early Welsh history. His main focus is that section of the manuscript which contains 159 charters purporting to date from the fifth to the eleventh century; charters that record grants made to the supposed founders of the see of Llandaff and its subsequent bishops. The book is divided into fourteen short chapters, each dealing with a different aspect of the charters, including their chronology, the witness lists, the use of diplomatic, genealogical references, and the larger structure of the manuscript. Together, the chapters build an impressive body of evidence that the manuscript is, in total, a carefully curated compilation drawn for the most part from existing sources.

Through a painstaking textual archaeology, Sims-Williams retrieves a chronological order for the charters which is at odds with the manuscript’s page order, indicating that compilers added in charters of different dates as they made the manuscript. He also provides a new and persuasive interpretation of the process of compilation, namely that the charters were taken from single sheets containing one or more charters of different dates that were then copied out continuously in the manuscript, a practice that explains the wayward chronology.

Further evidence comes from the formulaic style of the charters, which is almost entirely consistent with twelfth-century diplomatic even though the charters claim to be of different dates. Using frequency diagrams for key formulae, Sims-Williams is able to conclude that the style of the charters represents editorial decisions made by the compilers at Llandaff rather than the language of their earlier exemplars. Though the diplomatic may not be authentic to a period earlier than the twelfth century, Sims-Williams agrees with Davies that the charters were not all calculated forgeries but edited versions of earlier charters. Based on his collection of fine-grained data from the witness lists, doublets and personal names, displayed in tables and appendices, Sims-Williams goes much further than Davies in establishing the status of the charters as largely a compilation of earlier material.

Challengingly complex, this is the most detailed study of the Book of Llandaf yet published and now, surely, the standard work of reference on the manuscript for historians, codicologists, linguists and armchair scholars who enjoy linguistic and codicological puzzles. Though Sims-Williams shows that there are few easy solutions, and that Urban’s claims were unconvincing even at the time, he establishes beyond any doubt that the Book of Llandaf is indeed a ‘genuine archive’ and thus a vital historical source for a study of Wales and its border with England in the eleventh and twelfth centuries.