Translation in language documentation and revitalisation: LIP discussion

Alan Ray recaps June’s Linguistics in the Pub.

The June Melbourne LIP discussed the vexed topic of translation, particularly in the context of endangered languages. The context for the discussion was provided by Evans and Sasse (2007) and Hellwig (2010). Present were linguists from Monash, Melbourne and La Trobe universities.

The first observation, supported by personal experience and the above references, was that the longer a linguist works with a language and its speakers, the greater appreciation there is for the complexities and subtleties of that language. The challenge is how to show that complexity. In a standard three line example of text, gloss and free translation, the last is where idiom and other complexity can be shown. Of course the free translation can also mislead as it does not directly reflect the exact text and there is frequently no transparency as to how the free translation was arrived at. There was support in the group that the process should be more transparent. For example, at times a fourth line should be added before the free translation; a literal translation which most accurately reflected the base text.

There was considerable discussion on the question of context; how to show it and how important it was. Various aspects of context such as discourse information, cultural knowledge, gesture and physical landscape could all be important in establishing meaning.

Following from this it was agreed that translation / transcription recordings should capture as much as possible of context and especially discussions, clarifications and amendments by consultants. A benefit of recording everything is that the researcher may miss something that can come to light later when reviewing the recordings. Also consultants may want to add words such as pronouns which were not originally recorded because they believe the outcome to be more ‘correct’, itself an insight into the language in question.

A particular item of language that several participants had experienced difficulty with is that of documenting TAM (tense, aspect mood) in the language being analysed. The cultural and linguistic biases of the researcher can interfere greatly with understanding the language internal specifics of TAM, particularly that of aspect.

Finally, various challenges with dealing with glosses were raised. Should a gloss be invariant and meaningful and what does that say about your analysis? For example, Australian languages such as Wambaya have homophonous case endings for ergative, locative and instrumental. Glossing as separate case endings clarifies roles but reflects that the analysis is that they are different while an invariant gloss says that the analysis is that they are the same. A suggested treatment was to have a non-meaningful gloss such as the numbering used for Murrinpatha verb slots. Such non-meaningful glosses force the reader to read elsewhere to fully understand the language specifics. Cross-linguistic concepts such as dative case are useful but the label can hide subtle cross-linguistic differences. Lexical items also pose challenges, as an item may embody concepts not expressed easily in English, a simple example being dual pronouns.

References

Follow