That seems like quite a bit of hardwork for your processor. [You could
also use @domains I suppose, which supposedly indicates a parent element
type for the target]
Another way of doing this, since the goal is to distinguish amongst
interpretations (provided by the encoder) concerning the function of the
gloss might be to use the @ana attribute:
<gloss target"#xxx" ana="#definitional">....
Then you'd need to predefine the categories somewhere:
<interp xml:id="definitional">Used for glosses that supply a
<interp xml:id="translational">Used for glosses that translate into
However, it would obviously be a lot less work to classify your <gloss>
elements by means of a @type attribute. Since <gloss> as defined
clearly has multiple functions, I think there's a good case to be made
for it having one, and will add one unless I hear dissenting voices
within the next 24 hours!
Martin Holmes wrote:
> Hi there,
> If you link some <gloss> elements to <term> tags, and some to a
> different type of tag (whatever you're using for examples), then that in
> itself is a method of distinguishing them (in XPath/XSLT, at any rate).
> It may seem a bit indirect, but it's very practical, and saves using any
> attribute at all.
> Stephen SHIMANEK wrote:
>> Hi all,
>> First, apologies for my last message in which the first paragraph
>> was garbled and unclear due to rapid typing. I sent this correction
>> through last night, but it appears that if the listserve is in CC: it
>> does not go through.
>> My problem is simple: it seems that gloss does not allow for type
>> attributes (unless I am mistaken).
>> But we need to code both _definitions of terms_ and _translations of
>> examples_ in a corpus of linguistic texts. Martin suggested that I use
>> gloss for the latter, citing the definition of <gloss> given in the
>> <gloss> identifies a phrase or word used to provide a gloss or
>> definition for some other word or phrase.
>> My problem is that we will need to dissociate these two (very different)
>> types of gloss. Someone searching for examples of particular verbs
>> (like the ergative "break") in a corpus of linguistic texts might very
>> well be interested in glosses of related examples in other languages.
>> Giving them the option of searching through all of the elements typed as
>> "examples" (whether they be <cit>, <q>, <mention>, or <gloss>) would be
>> beneficial. On the other hand, it would be nice to be able to search
>> for linguistic terminology both in that which has been marked up with a
>> tag <term>, but also in the definitions of these terms, which I gather
>> should also be marked with <gloss>.
>> Has any thought been given to dissociating these two cases covered by
>> the tag?
>> Best wishes,
>> Steve Shimanek