> Hi Christian,
> I see Lou's point that you can't tell from that @corresp="#s2 #s3 #s4"
> <s xml:id="s1" xml:lang="en" corresp="#s2 #s3 #s4">The monkey is in the
> <s xml:id="s2" xml:lang="fr">Le singe</s>
> <s xml:id="s3" xml:lang="fr">est dans l'arbre</s>
> <s xml:id="s4" xml:lang="de">Der affe in baum ist</s>
> that s2 and s3 as a group correspond to s4 and s1. I'd be tempted to
> link those together with a @corresp on s2 pointing to s3 and vice versa.
> ;-) (Or use something like <join> elsewhere)
That's true. Actually, one of uses of this data is as an input to a
workflow that semiautomatically tries to distinguish these cases, and also
to narrow down the type of relationship involved. I would need a more
expressive formalism for this anyway (hence, SAWS), and then represent
that along these lines. But for the representation of input data it would
probably suffice, especially if the ambiguity is still unresolved ;)
What we do have in the case of bibles are the Eusebian sections that lump
together corresponding parts of the gospels. "Correspondence" here means
that an exhaustive understanding of the respective event requires the
consultation of the other gospel verses in the same Eusebian section. They
refer to the same event and possibly provide additional detail. This may
mean that corresponding passages from different gospels describe different
aspects of the same event (something like concatenation), but this may
also mean that they refer to the same event (something like paraphrase).
It still is problematic, because the unambiguous concatenation relation
between different verses from the same gospel can only be inferred through
their provenance (as we might for the two French <s> elements in Lou's
Applied Computational Linguistics
Johann Wolfgang Goethe Universitšt Frankfurt a. M.
60054 Frankfurt am Main, Germany
office: Robert-Mayer-Str. 10, #401b
mail: [log in to unmask]