Sunt Lacrimae Rerum

A few days ago, I became aware of Jerome McGann's A Critique of Modern Textual Criticism, which I had previously been unaware of. I used to be interested in the field, and in fact did a graduate course in the theory of mediaeval textual criticism with a teacher with whom McGann had been discussing these issues at the time (Lee Paterson, credited at the beginning of the book). I gather that it has had some considerable impact.

Regardless of the fact that the only two reviews on LibraryThing are dismissive (neither of which seems to have been written by someone with prior knowledge of post-18th century bibliography - one is at least written by someone with an exposure to classical stemmatics - and who are not aware of the context into which it fits) it is a well-written and cogent piece of work. The problem it addresses - how one conceives of the task of the editor as conditioned by how one views authorial composition - is refracted through various scholarly disagreements on editions of specific texts. At the time, considerable debate centred around the Kane and Donaldson Piers Plowman (referenced by McGann, despite the fact that it's not in the period he is mainly dealing with) and the Hans Gabler Ulysses (not referenced, but very relevant, and about which McGann has written elsewhere). The question of whether one is trying to recover a text reflecting "authorial intention" or something different, conditioned by either/both of a socially mediated agreement regarding a standard text (some versions become iconic, as with Auden's September 1 1939), or by a model of composition as collaborative, will not go away with modern texts. In addition, the issues are closely related to the perspective shifts associated with the New Historicism and other forms of theory which stress the social context of and contraints on the author at a different level.

I have seen it suggested that all of this is becoming less relevant now that we have entered a world in which composition, in the old sense, is dead, having been replaced by reformatting an author's electronic text so that at no time is there a process of transcription.

If anything, this is diametrically wrong. Consider the following, not entirely atypical, life of a novel by a modern author (with each stage preserved in complete texts for the future critic):

1. The first complete version is generated by the author after partial drafts have been circulated to beta readers. It now being trivial to generate and share copies, the author shares it with some critical friends, who respond in more or less detail. (This would have been much rarer and clumsier when it involved handing around and retyping a paper MS.) Following the input of these friends, the author then prepares



2. A new version revised both in detail and in structural ways, which is then sent to the publishing house (possibly by way of an agent, if the book has not been written pursuant to an existing contract). If the work has been written under a contract, the editor may have already have had input into the book.



3. The publishing house uses the author's MS and adds editorial queries and emendations, using Word's revision-tracking capabilities. The author makes some substantive changes and some minor changes as a result of the editorial feedback. At some point a copyeditor corrects the spelling and punctuation to match house style, with no minor allowances for places where the author wants specific effects. The resulting manuscript is frequently shorter than the author's initial submission.



4. At some point the corrected Word document is poured into an XML markup which can be used to generate both an e-book and typesetter's markup. This becomes the text of the first printing.



5. During the life of the first edition minor errata are reported. These are integrated into subsequent printings (new plates being easy to generate as long as they do not change pagination) and into the e-book edition. (In the old days, these would have been tipped in to later printings, and then integrated into later editions - if the publisher were conscientious about them.)



6. After several years the publisher decides to reissue the book in a new trade edition. To encourage the public to buy the new edition, the author, who has has time to rethink done things and is more generally experienced, does a moderate revision to the text creating a slightly longer work from which a few episodes have been deleted, to which a few have been added, and with noticeable, though not overwhelming, stylistic changes.

Under the model of original authorial intention, instance (1) would be seen as the best substantive text. (The parallel here is with the treatment of the different endings to Great Expectations , where the second ending is a result of feedback from a friend after publication.) This would also correspond, roughly, to Gabler's preferencing of the Rosenbach MS as a substantive source over the printed versions of Ulysses.

The Greg model of copy-text would suggest the adoption of (2) as a classic copy-text, i.e. as the source for punctuation and indifferent variants. It is also a good candidate for representing the author's final intention on a view which accepts a moderate social/collaborative model of composition but would draw a line before editorial input.

The substantively edited copy in (3) corresponds to the text taken as copy-text in McGann's approach. It recognizes the collaborative cooperation between author and editor as a positive part of the process, but there are two issues which need to have an eye kept on them: first, the typical blanket imposition of house style has no particular authority, even when the author has passed them, but can be hard to distinguish from punctuation adjustments which reflect actual thought; secondly, some (but by no means all) editorial changes are driven by pure publishing concerns (e.g. meeting general length guidelines driven by market/cost concerns as opposed to simply "tightening up" a text for better effect) and are likely to represent something the author agreed to reluctantly rather than embraced. It is also the version which most readers will see as "standard".

The text in (4) should be textually the same as (3), but formatting and layout (which may be important) will be different (consider the typeface issues in Heydt's The Interior Life). Some forms of transformation of the author's text is probably necessary, but in some cases the author's express preference may be different from what was actually achieved.

Emendations in (5) will typically be included in any edition as preferred readings, unless they emend readings which themselves are secondary. If the text in (3) were to be taken as copy-text, though, this text has an even better claim to authority.

With (6) we are into the territory occupied by the New York Edition of Henry James, the A, B, and C texts of (the first seven passus of) Piers Plowman, and the current view of Q2 and F Hamlet. (It is also the case of Stross' world-walker series.) Especially in the case where (1) or (2) is taken as the copy-text, does one integrate changes to create an eclectic text which never existed? Does one publish in parallel? Does it merely turn on the volume of changes (given that even the addition or suppression of a minor scene might transform the interpretation of a work)?

If (2) is published with the author's cooperation during the author's lifetime, as with Feist's riftwar books, it becomes a variant of (3), but raising the same kinds of questions as (5) does.

There are no easy answers here, and certainly the modern composition and submission process has not simplified the work of a later textual editor significantly.