In 1979, in his edition of Medieval English Songs, Eric Dobson castigated the established practice of the then chief lyric collections where:

Why this lamentable failure of editors to edit? The answer lay partly in the chaotically corrupt state in which Middle English lyrics have survived. However incredible it may seem, where these lyrics are found in multiple copies, in not a single instance are any two versions of any one lyric found to be identical. Textual variations in different copies of the same lyric frequently manifest differences in wording, in word order, in order of lines, and even in order of stanzas, differences sometimes so bizarre as to suggest that scribes must not only often have copied carelessly but may sometimes have done so from memory, and often from bad memory at that. Indeed, as in the instances where lyrics survive in shorter and longer versions, it is evident that rewriting sometimes entered into the transmission process. This state of affairs gave rise to the assumption that Middle English lyrics were, by and large, corrupt beyond repair, and that the traditional procedures of textual criticism revered by editors of previous generations were defeated from the start. In particular, it seemed impossible to analyse surviving manuscripts in terms of the stemmata of classical editorial procedure whereby manuscript superiority leading to the privileging of more authoritative readings could be established.

Yet another important requirement of traditional editorial practice was lacking. As E. T. Donaldson observed concerning Chaucer, ‘it is impossible to edit at all without having in mind some fairly strong preconception concerning the metre’. Whether one thinks of classical Latin poetry or, indeed, of Beowulf, metre has played a vital role in arriving at editorial decisions. In the case of Middle English lyrics, however, one will search the older standard editions in vain for any sustained discussion of metre. A notion commonly, albeit often tacitly, accepted was that lyric scansion was appropriately to be analysed in terms of the number of stresses per line.