A consensus has long existed that the Parson's Tale and the Retraction were intended by Chaucer to bring The Canterbury Tales to an end. A superficial reading of the manuscripts would tend to confirm this consensus, and only one modern edition, the Manly--Rickert Text, has raised the issue of its validity. The statistics, as we shall see, are impressive. A closer look, however, will show an uneasiness with the Parson's Tale, expressed mainly at its juncture with the Retraction, but also in its intentional omission from some collections of the Tales. This uneasiness begins with what is probably the second extant manuscript, London, British Library, MS Harley 7334, which has as heading for the Retraction Preces de Chaucer. It finds expression in a number of ways. It never successfully distinguishes where the Parson stops speaking and Chaucer 'the maker' begins; nor does it make sense of some of the language employed in the Retraction -- the word 'treatise', the thanks given to God for what is pleasing to the reader in a work being largely retracted, the way the Canterbury Tales themselves are mentioned in a list of Chaucer's works. All the responses observable in the manuscripts have continued to our own day. Those troubled by the Parson's Tale have directed their efforts towards assimilating it into a reading of the work as a whole. They have failed to consider the possibility that Chaucer intended it as an independent work, the Treatise on Penitence, with the Retraction as a fitting conclusion.
Let us look first at manuscripts that come to us in undamaged condition.(1) Ten of these, and the two Caxtons as well, contain both the Parson's Tale and the Retraction. Only one manuscript, Jean d'Angouleme's BN f. angl. 39, breaks this pattern. The personal idiosyncrasies of its patron--owner so clearly played a part in the choice of tales included in the Paris MS that the failure to mention the Parson's Tale might mean simply the patron's disapproval rather than any questioning of its validity at the end of the work. Jean's taste played an interfering and intrusive role in the very production of the manuscript.(2)
A similar picture emerges from the slightly damaged manuscripts. Of the sixteen with folios but no complete quires missing, only one, London, British Library, MS Sloane 1686, leaves out the Parson's Tale and the Retraction.(3) Of the three with a single quire missing, MS Barlow 20 is the exception.(4) Balow 20 and Sloane 1686 should weigh more heavily than the Paris MS, however. Neither is selective in its inclusions until the end of the Manciple's Tale. Here each clearly intended to bring The Canterbury Tales to an end.
A number of the more heavily damaged manuscripts should also be considered. Eighteen of these include part of the Parson's Tale, and no doubt once included the whole of it and the Retraction as well.(5) Oxford, Bodleian Library, MS Bodley 686, on the other hand, basing its ordering of the fragments on the a-Ellesmere model, dwindles to a close after Sir Thopas by including only the 'Tale of a Crow', which the scribe attributes in running heads to Lydgate, and the St Cecilia, which he never assigns to a pilgrim. This ending is the more surprising since the manuscript includes an elaborate if spurious conclusion for the Cook's Tale.(6)
Another manuscript, Harley 7333, a great anthology of secular literature, written by Austin canons at St Mary de Pratis near Leicester, breaks off after 253 lines of the Parson's Tale, at the foot of a recto. Not only the blank versopage but a blank leaf follow, plus four small booklets mainly consisting of poems by Chaucer, Gower, Lydgate and Hoccleve. That the break happens at the foot of a page suggests that the copy-text was not responsible. Apparently the Austin canons found reason to stop copying the Parson's Tale.(7)
Of the forty-nine manuscripts so far considered four fail to include the …