Pretty obviously, at the centre of any edition stands a reproduction of the text in question. This represents that primary contact with an audience that any editor seeks. It allows that audience to read and puzzle over the document, and it provides them with a handy reference system, should they wish to quote the work in the course of some discussion. But where should this representation of ancient documents for modern use come from?

For the various types of editions predicated upon single manuscripts and discussed above, the answer to this question is easy. One has selected a manuscript as being of special interest or as representative, and one simply copies it out and presents it in print or code. But a ‘critical edition’, predicated upon a range of copies, clearly demands more complex procedures. How would one decide what version of the text to place before readers?

At least initially, the answer to the question is relatively straightforward. Just as in a single-manuscript edition, one wants to provide one's readers with a single continuous textual source. But, faced with a range of copies, fifteen of them in the case of Rolle's ‘Super Canticum’, section 4, one would seem to have a plethora of choices. How would one adjudicate between claims for one manuscript or another in these circumstances?

An initial response would return to one specific type of single-manuscript edition discussed above, that predicated upon ‘the most complete copy’. If one seeks a single continuous source for presenting the text, it is plainly unintelligent to seize upon something manifestly incomplete as the basis for operations. To do so means that, at a minimum, one will have to supplement what one has written from another source – and would immediately raise the question of why one had not followed that more complete version in the first instance. Thus, initially, one looks for a copy that, on general inspection, appears to be a relatively complete version of the text to be presented.