Add att.canonical to <monogr> (and allow it to contain nothing but a text node or mixed content)

[awagner-mainz]

This takes up and is somewhat similar to the argument in #2392 ("Add att.canonical to <bibl>"). I have often seen bibliographic entries - even structured ones as in bibliographies/lists of references - to report the <monogr> part of an "analytic" publication by way of a shorthand. In lists of references this can be author-date (where the <imprint> that currently is mandatory would have its use), but it can also be author-shorttitle, and in footnotes it often is a short title or even just a "id." or "op.cit.". IMHO, it would be convenient to allow encoding these phenomena as instances of <biblStruct> and not force editors to either fall back on <bibl> or feign a full, redundant <monogr> (including the mandatory <imprint>).

I guess one could argue that <biblStruct> is meant to facilitate automated processing and thus it makes some sense to, yes, indeed, force editors to enter the information another time even though it is not present in the source (in this place at least). But with the @ref attribute pointing to a proper, more elaborate <monogr> entry, the processability could still be ensured and the whole encoding would be more true to the original source text.

Example:

• Fraser, N. and L. Gordon,: "Civil Citizenship against Social Citizenship?’", in: van Steenbergen (ed.), Condition of Citizenship.
• van Steenbergen, B. (ed.): The Condition of Citizenship. London, Thousand Oaks, New Delhi: Sage, 1994.

could be:

<listBibl>
  ...
  <biblStruct>
      <analytic>
          <title level="a">Civil Citizenship against Social Citizenship?</title>
          <author>
              <persName>
                  <forename full="init">N.</forename>
                  <surname>Fraser</surname>
              </persName>
          </author>
          <author>
              <persName>
                  <forename full="init">L.</forename>
                  <surname>Gordon</surname>
              </persName>
          </author>
      </analytic>
      <monogr ref="#monogr-resolved-biblstruct-8">in: <editor>van Steenbergen</editor> (ed.), <title>Condition of Citizenship</title></monogr>
  </biblStruct>
...
  <biblStruct>
      <monogr xml:id="monogr-resolved-biblstruct-8">
          <title level="m">The Condition of Citizenship</title>
          <editor>
              <persName>
                  <forename full="init">B.</forename>
                  <surname>van Steenbergen</surname>
              </persName>
          </editor>
          <imprint>
              <pubPlace>London, Thousand Oaks, New Delhi</pubPlace>
              <publisher>Sage</publisher
              <date when="1994">1994</date>
          </imprint>
      </monogr>
  </biblStruct>
</listBibl>

[awagner-mainz]

Hmm, it just occurred to me that the placement of <biblScope> poses another problem with this approach: It is a child/descendant of <monogr>, but it cannot be in the referred-to monograph, instead it needs to be in the present <biblStruct> (the Fraser/Gordon chapter in the example). If there were a <biblScope> in the referred-to <monogr> (van Steenbergen (ed.), Condition of Citizenship), it would certainly not be the one applicable to the chapter containing the reference to it. Thus, it seems that in order to accommodate the idea about referrals I had in mind, it would be necessary to not only do all the changes mentioned above, but also additionally move the <biblScope> element from <monogr> to <analytic>.

I could imagine that, on an ontological level, this would even make some sense. And I understand the wording of the guidelines to actually specify this precise behaviour: "The biblScope element [...] specifies where to find the component in which it appears [...]." (3.12.2.1 Analytic, Monographic, and Series Levels) Or does it? The accompanying examples and current practice of course use it differently. And doing it in the way suggested here presents other problems, when the <analytic> component has been published in several different <monograph> components (then the <analytic> component probably would have to contain mutliple <biblScope> entries and a way to convey which of them applies to which <monogr>).

On the other hand, however, I also get the impression that the changes I had in mind turn out to be quite (a bit too?) far-reaching. And possibly there have been arguments about these issues that I am not even aware of. I leave the issue here anyway, maybe someone feels inspired to write their thoughts or report on earlier discussions... What do you think?

PS. Yet another alternative approach would be to allow <ref> to altogether replace <monogr> in <biblStruct> (perhaps the same would apply to <ref> replacing <analytic>), but this would not solve the biblScope-needs-to-go-with-analytic-not-monogr problem either...