while we are on this very useful subject of figuring out how to represent variation —

something which current TEI does not deal with very well, I think, is ‘block level’ information in a collation.  It is very usual in the collation we do to divide the text into named segments: e.g. Book 4, chapter 1, paragraph 1. As well as show the variation within each segment we commonly want to represent information about the instances of ‘each block’ in each witness, as follows:

1.  A list of the witnesses which have this block (including: the witnesses which contain multiple texts of the block, both because the block appears more than once in the document, or because there is scribal variation within blocks)
2.  A list of the witnesses which do NOT have this block, because it is simply ‘not present’
3.  A list of the witnesses who do NOT have thsi block, because of apparent lacunae 
4.  Lists of the variant orders in which this block appears: i.e., list of the different blocks which PRECEDE this block in the witnesses, and the witnesses which have those orders; and the same for blocks which FOLLOW in the witnesses.

This task keeps growing. Again, CollateX (in its Birmingham incarnation) is ahead on most of these, with strong mechanisms in place for dealing with lacunose etc. witnesses.


On Nov 30, 2015, at 1:04 PM, Hugh Cayless <[log in to unmask]> wrote:

I'd been toying with the idea of doing something like:

<witness xml:id="λ" n="λ">Florence, Biblioteca Medicea Laurenziana Aed. 203, XV. 
    <witness xml:id="λ1" n="λ¹">first hand</witness>
    <witness xml:id="λ2" n="λ²">second hand</witness>

That is, treating the different hands as sort of "sub-witnesses". Can you give me (or point me at) an example of what the extended CollateX JSON format looks like? It would be nice to have them be compatible.


On Mon, Nov 30, 2015 at 1:12 PM, Robinson, Peter <[log in to unmask]> wrote:
This hits right on a problem which has nagging at me for some time.
We have, very frequently, the situation where corrector x changes a word or two, or much more, in Ms A. We cope with this very well within transcription by something like, for the case where original “les” is overwritten by “los” by hand c1:

  <rdg type="lit">
    <seg type="overwritten">
      <seg type="1">les</seg><seg type=“2">los</seg>
  <rdg type="orig" resp="c1">les</rdg>
  <rdg type="mod" resp="c2">los</rdg> 

The first rdg element presents a ‘topographical’ transcription, which we use to represent just how this appears in the ms.  The second two rdg elements represent what the transcriber deduces as the original and modified state of the readings.  We can collate the two readings, to end up with an apparatus entry which looks like:

les A-c1] los A-c2

One could indeed create separate entries for each witness in the witlist, no problem.  But what this does NOT do is represent the critical information, in the apparatus, that A-c1 and A-c2 are actually ‘sub-manuscripts’, or some such expression, of Ms A.  In fact, Cat Smith’s collation editor (see https://github.com/catsmith/collation_editor) does this rather better, by representing witnesses A-c1 and A-c2 within the structure representing A: this is done as an extension of the CollateX JSON format for representing witnesses for collation input. 

Obviously, it would be nice to have routines for roundtripping data from TEI encoded collations to and from something like CollateX form. Which means, at least, the TEI encoding should be able to do all that CollateX can do. Here is one example of something which can be represented in CollateX but not in TEI, at present. Sub variation and overlapping variants, both which the collation_editor/collateX can do, are others.


On Nov 30, 2015, at 10:57 AM, Hugh Cayless <[log in to unmask]> wrote:

Dear All,

I'd like to know how producers of critical editions in TEI usually handle the documentation of hands in their documents. I'm working with an edition that has this kind of thing:

genista] genestra λ¹δ, genesta εγμλ²χ²

so MS. λ has "genestra" in the first hand, corrected to "genesta" in the second (the edition doesn't actually tell us how this works in the MS). I want to associate λ1 and λ2 with λ.

The situation with describing hands seems rather confusing. In msDesc, we have a handDesc, which contains handNotes, but you can also have handNotes inside the profileDesc. I'd really like to do neither of these, but rather list the hands inside each witness. I could do this by embedding msDesc inside witness, but that's rather overkill, since msDesc requires msIdentifier, which, while it does contain the same info as will go in the witness, also forces me to encode it in a structured way rather than in prose.

So my questions are: how do you all do this? Does handNote need some cleaning up? Should I not use handNote at all for this sort of thing?