[tei-council] Fwd: multiple lem in TEI

[Martin Holmes]

Would a fourth option be to move the constraint from <app> to <rdgGrp>? 
Even if there's a tortuous argument to be made for this particular 
example, presumably no-one is suggesting that a single <rdgGrp> can have 
multiple <lem>s?

Cheers,
Martin

On 13-06-12 05:01 AM, Lou Burnard wrote:
> I don't think Elena had anything to do with this particular example :
> the text licensing this usage of multiply nested <rdgGrp>s as an
> alternative to multiply nested <app>s goes back to TEI P3 if not before.
>
> This example is from section 19.1.3 ("Indicating subvariation in
> Apparatus Entries") of P3, reappears more or less without change in P4,
> and eventually becomes section  12.1.3 of P5. The  comment at the end of
> the section " Some encoders may find the use of nested apparatus entries
> less intuitive than the use of reading groups, however, so both methods
> of classifying the readings of a variation are allowed." seems somewhat
> humorous in retrospect.
>
> Amongst possible solutions that occur to me
>
> a) modify the schematron rule to distinguish rdgGrp/rdgGrp/lem from
> rdgGrp/lem
>
> b) disallow the nesting of rdgGrps (if you want to do that, you have to
> use nested <app>s unintuitive as that may seem)
>
> c) remove the schematron rule completely
>
>
>
>
> On 12/06/13 12:26, Martin Holmes wrote:
>> I went through the svn history, and it looks like this example was added
>> in this commit:
>>
>> ------------------------------------------------------------------------
>> r4552 | louburnard | 2008-04-29 16:11:49 -0700 (Tue, 29 Apr 2008) | 2 lines
>>
>> add examples from Elena
>>
>> ------------------------------------------------------------------------
>>
>> So we should definitely check with Elena about this. I'll do that and
>> copy the group.
>>
>> Cheers,
>> Martin
>>
>>
>> On 13-06-12 02:36 AM, Gabriel Bodard wrote:
>>> lemGrp is not a terrible idea, as unintuitive as it may sound. On the
>>> other hand, if Marjorie is right about what the example in the GLs
>>> means, then shouldn't it perhaps be something like:
>>>
>>> <app>
>>>       <rdgGrp>
>>>         <lem/>
>>>         <rdg/>
>>>       </rdgGrp>
>>>       <rdgGrp>
>>>         <rdg/>
>>>         <rdg/>
>>>       </rdgGrp>
>>>       <rdgGrp>
>>>         <rdg/>
>>>         <rdg/>
>>>       </rdgGrp>
>>> </app>
>>>
>>> so that there is one "preferred" reading for someone who wants to
>>> generate a "good text" from this, and all of the variants are still
>>> available to someone wanting to generate a dynamic view of all the
>>> witnesses? The schematron rule would therefore stand...
>>>
>>> G
>>>
>>> On 2013-06-12 09:20, Sebastian Rahtz wrote:
>>>> Marjorie has sent a helpful discussion of the invalid example,
>>>> but it confuses me even further about which way we should go with this.
>>>>
>>>> More input needed: is the example or the constraint wrong?
>>>>
>>>> Begin forwarded message:
>>>>
>>>> From: Marjorie Burghart <marjorie.burghart at ehess.fr<mailto:marjorie.burghart at ehess.fr>>
>>>>
>>>> It's indeed an awkward example... If I understand correctly, its problem is not so much that it has several <lem> as that it does NOT really have one.
>>>> If you look at the example above the last one, you see that "Experience" is the lemma, and "Experiment" and "Eryment" are rejected readings - which is very fine: one and only one lemma, 1 or more readings.
>>>> The last example, as I understand it, expands on the previous one: it notes orthographic variants in <rdg>s, grouped with the main word they are subvariants in <rdgGrp>s. As a side note, in the area of philology with which I am familiar (Latin language, "literary" religious texts) minor orthographic variants are not considered significant in the history of the transmission of the text, and most philologists do not bother with them (they just announce in the introduction a list of words that have been given a standard spelling, usually). It is quite different though with people editing vernacular texts, so it all depends. But noting the orthographic variants can of course be useful if you are preparing an edition with links to digital facsimile of the witnesses.
>>>> In this case, <rdgGrp> comes in handy, but it seems that <lem> has a different meaning within a rdgGrp: apparently it does not mean that this is the reading of choice to retain in the main text, but the canonical spelling of the word. Therefore, in the last example, I do NOT see any real lemma, and I would be really bothered if I had to build a "main text". There should be something somewhere to note that the first <rdgGrp> containing "Experience" and its minor orthographic variants IS the actual lemma of this <app>.
>>>>
>>>> Should there be a <lemGrp> then? :)
>>>>
>>>> --
>>>> Sebastian Rahtz
>>>> Director (Research) of Academic IT
>>>> University of Oxford IT Services
>>>> 13 Banbury Road, Oxford OX2 6NN. Phone +44 1865 283431
>>>>
>