In my corpus of signed language data, I am using a multilingual ECV on the sign gloss tiers (via Global Signbank). I notice that when I search these tiers (structured search multiple EAF) across the corpus, language matters. So for example, if I am trying to find all tokens of a particular sign that means “mother”–it seems I need to search both MOTHER (english) and MOR (norwegian) to get all the tokens.
(Conducting this search leads to: 12 hits for MOR and 10 hits for MOTHER (but they are the same sign, linked in Global Signbank)
ELAN seems to only search according to the exact letter string, which will not match if different files have different languages set locally (in that I can choose to view the tiers in English or Norwegian in specific files). Do you know if there is a way for ELAN to search both languages simultaneously–according to the linked glosses in the ECV?
Yes, the ELAN search functions search the annotation strings as they are stored in the .eaf files. I believe there is currently no exception to that in the current search functions. The structured search function is a more or less separate search engine which used to be the base of the search function in the online Language Archive and was not only used for .eaf files.
So, I don’t think there is a friendly way to search for both languages simultaneously, or even to search for a single language independent of what happened to be stored in the individual eaf files.
I don’t know if this works and, if so, if it is a viable ‘workaround’ workflow or not, but you could maybe try this:
use File->Multiple File Processing->Update Transcriptions for ECV’s. Make sure to select a different folder as the output folder, select a language (in this case English or Norwegian), make sure the “Don’t change annotation value…” checkbox is unticked and click Start. If all goes well all annotations linked to a ECV entry should then be in the same language and searchable. Not convenient, but I can’t think of anything better at this moment.
