Second thoughts on "Sumer is icumen in" (John, February 2019)

Those who have sung "Sumer is icumen in" with me will know that I have had serious doubts as to its dating. I found it difficult to believe that it had suddenly come into being as early as the thirteenth century with no evidence of the simpler two-part and three-part canonical essays which must surely have preceded it, and even more difficult to believe that it had inspired no successors for perhaps a hundred and fifty years. Recent correspondence with the British Library has convinced me that these doubts were misguided.

"Sumer is icumen in" is on folio 11v (verso, back, left-hand page, "recto" being the front) of British Library manuscript Harley 978, a digitisation of which can be found at

It is a singularly lovely thing. The consensus of scholarly opinion seems to be that its Gothic script is typical of the second half of the thirteenth or very early fourteenth century; on the grounds of its content, the more specific date "between 1261 and 1265" is put forward by Andrew Taylor and Alan E. Coates in The dates of the Reading Calendar and the Summer Canon, Notes and Queries 45.1 (1998), 22-24. There is an extensive bibliography in the online record cited above, and my attention has been drawn in particular to Nino Pirrotta, On the Problem of "Sumer is icumen in", Musica Disciplina 2 (1948), 20516, and Ross W. Duffin, The Summer Canon: a New Revision, Speculum 63 (1988), 122.

If we accept this date for the manuscript, how can we answer my doubts regarding the music? As regards the apparently sudden emergence of this piece of canonical writing, the musical section of the manuscript comprises folios 2-15, and includes three pieces in two parts (folios 8v-9r) and one in three (folios 9v-10r). There is a convenient modern edition of this last (it is in "Five Anglo-Norman Motets", Antico edition AE24) and at first sight its regular cadences on bare fifths and even octaves are in a style wholly alien to that of "Sumer". However, having actually sung it, together with some of the broadly contemporary Robin and Marion motets which comprise Antico editions AE22/25, I find a link more believable. The motets all seem to be built on tenors in long notes, just as "Sumer" seems to be built on the long notes of its two-part bass; true, the piece on folios 9v-10r is not in canonical form, but the writing of canons is not difficult (I have put together a couple of simple three-part canons myself, and I am no more than an amateur madrigal singer with no formal training in composition whatever) and I find it possible to believe that whoever produced this could also have written "Sumer". It is by no means unknown for a composer suddenly to produce something quite outside his normal style, and indeed outside his normal quality.

The apparent absence of any immediate successor is less easy to explain, but the most obvious possibility is that there were indeed successors but they have been lost or I am otherwise unaware of them. I have tried to find a modern edition containing all known canons written before the fifteenth century, but have been unsuccessful; if anybody reading this knows of one, please will he or she tell me. Wikipedia, in article "Canon (music)" accessed 14 February 2019, mentions (a) "Tosto che l'alba" (two parts in canon above a bass) by Gherardello da Firenze who apparently died in 1362 or 1363, (b) "Si je chant mains" (three parts all in canon) with a link to a YouTube performance in which it is possibly attributed to Denis le Grant who died in 1352, and (c) some three-part movements by Guillaume de Machaut in his "Le Lai de la Fontaine" of 1361. The Duffin paper cited above mentions "Talent m'est pris de chanter" from the broadly contemporary manuscript of Ivrea, Biblioteca Capitolare. These do not completely fill the gap following "Sumer", particularly if the Taylor-Coates date of 1261-5 is accepted, but they do at least reduce it.

It is possible that future advances in scientific knowledge will allow the material of manuscripts like this to be directly and accurately dated without destructive testing, but as long as we are reliant on indirect dating by style and content I am happy to accept the British Library's dating. I still find it difficult to believe that this delightful little canon inspired no immediate successors, but the possibility that they existed but are now lost cannot be excluded. Or was "Sumer" so far ahead of its time that composers in its immediate aftermath were overawed into not attempting to compete, just as Pearsall's magnificent eight-part madrigals "Lay a Garland" and "Great God of Love" seem to have overawed composers since?