On Sun, 17 Sep 2006, Peter Constable wrote:
Moreover, it is possible to use IPA to describe words that appear the same (in pronunciation) in different languages, in which case we would probably use "mul" if we had to tag the text.I suppose one could use IPA notation to cite an abstracted utterance. Abstracted usage -- i.e. independent of any language -- is often done when discussing individual phonemes, but I don't know that I've ever seen a single instance of a transcription of an abstracted utterance.
Just to clarify: I did not mean an abstract utterance but something like"The word [kala] 'fish' has probably remained the same from the Uralian protolanguage to Finnish and Estonian but often changed considerably
in other Uralian languages." (I'm cheating a bit here, in order to be able to use ASCII only, since [a] is not the most adequate vowel symbol here.)So it would not be an abstract utterance but a word that is allegedly the same in three languages, one of which is an uncertain reconstruction without a specific code.
The same situation may, of course, also appear when words are written using some normal script instead of IPA.
But I'd suggest that, if a document contained such a thing, it would make most sense either to leave it untagged, or to tag it "zxx" as it is not content in any language, let alone multiple languages; it is just a abstract phonetic sound.
Perhaps. It depends on the context and requirements. I think the actual _utilization_ of language tagging in software is far behind the specification of codes for it, and it is therefore difficult to estimate the practical impact of tagging decisions in borderline cases. Most texts in the world have no language tagging, still less tagging at the level of individual words (or characters) even in cases where the tagging decision would be straightforward (e.g., in tagging personal names that clearly belong to a language other than the surrounding text).
-- Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/ _______________________________________________ Ltru mailing list Ltru at ietf.org https://www1.ietf.org/mailman/listinfo/ltru
Note Well: Messages sent to this mailing list are the opinions of the senders and do not imply endorsement by the IETF.