The Robert ap Huw Manuscript and the Dilemma of Transcription PEKKA TOIVANEN 1 Introduction There is little written information from the first millennium con- cerning instrumental performance, context and practice, and scarcely any music actually intended for instruments before c.1450 survives. Most instrumental pieces were improvised, often on structural or har- monic patterns, and since music was usually learned and transmitted orally, it was seldom written down. Western musicology tends to rely too exclusively on written evidence, underestimating or even ignor- ing the processes so essential within any musical culture which pave the way for notated sources and musical treatises. This is equally true for the repertory under discussion: of the numerous studies relating to the Robert ap Huw manuscript, only four attempt to address the oral character and transmission of harp music in Wales. Existing transcriptions reflect the notion that early music cannot be properly discussed unless it is presented in a modern notated form. Yet tran- scriptions can only represent a literal version at second remove from a musical tradition originally transmitted orally, where the first layer of literal translation was most likely the work of Robert himself. Inevitably, such attempts to transmit the essentially oral tradition of cerdd dant through modern notation may be misleading, if not er- roneous. This article re-examines the transmission of oral tradition from an ethnomusicological perspective, considering some of the musical and cultural features of Robert ap Huw's environment. 2 Tradition and transmission A salient characteristic of ethnomusicological scholarship has always been its interest in process: cultural change, whether encouraged or