Dear Hans Walter,

Thank you very much for your reaction.

I am coming to this from a very different angle than what you seem to be assuming. I am a book historian and for me a stemma is interesting because it shows/can show which version or versions of the text were most popular/numerous, where and when the reception of the work happened, which manuscripts are textually related, what the manuscripts that were ancestors to important textual families looked like (with or without rubrics, with or without illustrations, with or without certain interpolations), etc.

So while I agree that the stemmatic method was developed by editors to try to reconstruct archetypes of existing manuscripts, that is not the sort of scholarly activity I am mainly engaged in. I am reconstructing stemmata, not to reconstruct a 'clean' text or to produce an edition but in order to study the relationships between surviving and lost manuscripts. I am certainly not trying to use the stemmatic method in order to identify authors' originals and I am sorry if I gave that impression.

The tradition I am working on is the Chronicles of Jean Froissart. For the first three books virtually all scholars (including all previous editors) agree that each of them exists in one or more authorial versions. One can disagree with this (and a few scholars have expressed doubts about the authorial status of one or two of these versions), but I think that arguments in support of the hypothesis cannot be based on stemmatic reasoning, although one can of course disprove such a hypothesis using stemmatic methods (if a descendent of one supposedly authorial version is in fact shown to be derived from a descendent of another authorial version).

So I see the different authorial versions of Froissart's Books I-III as a given, or at least as a hypothesis I am happy to accept until proven wrong. The particular situation with many of these versions of Froissart's Chronicles, is that the differences are very localised. So the differences between the A and B versions are all in the first twelve chapters, while the rest of the text seems to be identical in these two versions. A similar situation is found in Book III, where a number of chapters across Book III have apparently been rewritten by the author. I am trying to use this situation to help me with establishing a stemma as I have set out in my earlier e-mail. I'll try to reformulate what I am proposing to do using as an example the A and B versions of Book I referred to above. 

The text of chapters 0-11 in these two versions of Book I is different, but for the rest of Book I the text seems to be identical. There are of course textual variants across the extant manuscripts, but if we assume that between version A and version B Froissart decided to rewrite only the first 12 chapters, than we must assume that the text of the other chapters, as contained in the surviving MSS, ultimately goes back to Froissart's own autograph MS of this text. 

It may be that Froissart had two autograph copies, one of version A and one of version B, in which case we must assume that the B autograph, for the chapters after chapter 11, was (directly or indirectly) copied from the A autograph. It may also be that he had only a single autograph and that after finishing the A version he physically added his rewritten chapters 0-11 to his codex, either at the end or by removing his earlier text and thus effectively substituting the earlier version of these chapters.

If we now try to establish the stemma of the A and B manuscripts, we can first group them under these two (A and B) families on the basis of their text for chapters 0-11. Each of these families will ultimately go back to a different autograph (the A and B autographs, whatever they looked like). When we try to establish the stemma of each family we will look at variant readings in the surviving manuscripts. Now here comes the interesting bit. Say we are looking at chapter 15, which is in principle the same text in the A and B versions. At a certain place in the text we find three variant readings which are simply synonyms, but one of these is found in witnesses of both the A and B families. In a 'normal' situation, simple synonyms would not be considered as relevant variants because none of them is 'wrong' and therefore one cannot work each which of them are original and which of them are not. In our situation, however, we have very strong grounds to assume that the variant reading which appears in witnesses of the two families is there simply because it was in both the A and B autographs and is therefore the original authorial reading.

This seems to me a perfectly legitimate and logical way of reasoning. One could of course assume that in some cases one can end up with variants found in both families/versions that are simply the result of chance, but if there is a pattern emerging, then this sort of argument would be very useful to establish a stemma. Indeed, it means we can use many more variants, because we don't have to work out which are 'significant' variants, and which readings or correct and which are corrupted.

I am sorry I have been rambling on for a bit, but it seems that my original posting was confusing so I have tried to explain the problem a bit better.

Best,

Godfried Croenen
 





> -----Original Message-----
> From: The list of the European Society for Textual Scholarship and the
> Society for Textual Scholarship
> [mailto:[log in to unmask]] On Behalf Of Hans Walter
> Gabler
> Sent: 29 October 2009 08:16
> To: [log in to unmask]
> Subject: Re: stemmatology and authorial versions
> 
> I have only now found Godfried Croenen's query of 17 September on my
> computer -- I was on holiday at the time.
> 
> I am puzzled, because I do not readily see a link between stemmatic
> reasoning and authors' originals -- or rather, only exceptionally so
> if one by way of analyzing derivative witnesses begins to suspect
> that the archetype (i.e., the logical point of confluence-in-reverse
> of surviving witnesses) might actually coincide with an assumed real
> author's original. The idea, however, of perhaps even being able to
> distinguish more than one author's original by hypothesis from
> surviving variation seems to me exceedingly daring. Hypotheses do not
> win, as we know, by proliferation; only by reduction: for they do not
> (in conceptual terms, at least) mirror realities -- they are logical
> constructs. Moreover, I believe it is important to realise that the
> whole stemmatic method was never out after author's texts -- it was
> devised, rather, in pursuit of the 'purest' text, i.e., the text
> least overlaid by contaminations of transmission. The concern with
> the author and the author's authentic (authorized, or whatever) text
> came in at a later stage in the history of ideas, as well as in the
> history of textual criticism.
> 
> What would be intriguing to learn is which specific case of
> transmission, if any, sparked Godfried Croenen's query.
> 
> Hans Walter Gabler.
> 
> 
> At 16:55 17.09.2009, you wrote:
> >Dear all,
> >
> >I am trying to get some relevant methodological literature on a
> >situation which I am dealing with and which must occur quite
> >regularly in textual studies, although in the few handbooks I have
> >looked at I found no discussion of this specific case.
> >
> >A lot of criticism of the Lachmannian method has focused on the
> >problematic status of the "common error", i.e. how do scholars know
> >that a variant is indeed an error, why do scholars think the
> >author's text did not contain errors, how can we be certain that a
> >correct reading has not been corrected by the scribe and is
> >therefore younger then the 'error', why should we not use variants
> >that are not errors, etc.
> >
> >In a tradition where we have two or more authorial versions of an
> >author's text, there may to an extent be a methodological way out of
> >this impasse, which would allow scholars to use variants without
> >worrying whether or not the variants are proper 'errors' and even
> >whether or not they are 'significant' (i.e. whether or not they are
> >of such a nature that it would be difficult for scribes to make the
> >same error independently, or change the error back to the original
> >reading without having access to a witness of the original reading).
> >
> >The argument goes as follows. If at a point in a section of the text
> >which is contained in both redactions/versions of the text, the
> >textual tradition has more than one reading, then the variant shared
> >by witnesses of the two redactions is likely to be the one that is
> >original. Indeed, each of these witnesses goes back to an archetype
> >which in turn is derived from the author's (two or more) originals.
> >Since it is likely that the author's second (third) original, or the
> >later stages of his original manuscript were based on his first
> >original or on the earlier stages of his original, then the variant
> >transmitted to the witnesses of both (or more than one) version of
> >the text is likely to go back to the author's original(s).
> >Therefore, the fact that the variant appears in witnesses of both
> >redactions is in itself significant and in many cases enough to
> >accept that this is the original reading (and therefore 'correct',
> >although not in the Lachmannian sense). This means that, as long as
> >one can find variants which occur in witnesses of both versions, one
> >does not need to (and maybe should not) worry about which is the
> >correct and which is the faulty reading because one can with a fair
> >amount of certainty establish which reading is original and which is
> >a variation on the original.
> >
> >The argument is of course not entirely unproblematic (what with the
> >changes in the author's successive versions, what the tendency of
> >scribes to simplify readings, etc.), but I hope you get the gist. I
> >would be grateful for any pointers towards either theoretical
> >discussions of this issue, or to examples of where
> >editors/stemmatologists have (successfully) used this sort of
> >argument in their reconstruction of textual filiations.
> >
> >Thanks,
> >
> >Godfried Croenen