UniversalDependencies / UD_Irish-IDT

Irish data
Other
6 stars 7 forks source link

how to annotate ní hamháin (not only) #10

Open tlynn747 opened 4 years ago

tlynn747 commented 4 years ago

Currently the phrase ní hamháin 'not only' is inconsistently annotated throughout the Irish treebank.

The difficulty with comparing the phrase to other languages for inspiration/advice on how to annotate is the syntactic structure of this phrase in Irish: ní (neg form of copula) hamháin (adjective). This differs for e.g. to not only in English and non seulement in French where they make use of amod and advmod labels.

Another problem is that sometimes ní hamháin is proceeded by a comma (which would present the argument for a parataxis interrupting a string of PPs, for example),

sentence 1481 (training file) bhí naíonraí rathúla lánEabhraise faoi lánseol, ní hamháin ar na gabháltais, ach i mbailte agus i gcathracha chomh maith 'successful all-Hebrew crèches were in full-swing, not only in the holdings, but in towns and cities as well'

but quite often there is no comma:

sentence 3518 (v2.7 to be released) Cuidíonn an Ghaeilge mar theanga pobail sa Ghaeltacht ní hamháin leis na sainghnéithe luachmhara de thraidisiún... 'As a community language in the Gaeltacht Irish not only helps valuable traits of tradition...

sentence 138 (test file) amhrán a chan Gaeilgeoirí le fuinneamh, le scléip agus le gliondar ní hamháin ar fuaid na hÉireann, ach go deimhin ar fuaid an domhain. 'a song that Irish speakers sing with energy, with fun, with joy not only across all of Ireland, but of course across the world.'

One option is to follow the English/French approach and attach hamháin 'only' as amod or advmod to the head of the subsequent phrase depending on whether it's a noun or verb, and just connect as a copula, ignoring the fact that normally when the copula is attached to an adjective, it's considered part of a copular phrase.

Another option is to treat ní hamháin like a copular phrase interrupting a string of NPs or PPs etc. And the question that arises then is how to attach this "inserted" copular phrase to the matrix sentence - to the root or to the following NP or PP.

Note that if the phrase is ní hamháin sin 'not only that' - should probably be treated differently to the examples above, ie as a phrase... discourse? advcl?. (e.g. "Not only that, they also believe that XYZ")

Similarly if it introduces a complement phrase, in which case it seems the adjective would be the root of the sentence (with left-adjoining cop attachment) followed by a complement phrase in the same way we analyse sentences such as Is dóigh go bhfillfidh sé 'It is likely that he will return' ... i.e. a copular phrase with clausal subject:

sentence 96 Ní hamháin nach bhfaca sé aon rogha eile áfach (lit not-be only that he did not see any other choice however') 'Not only did he not see any other choice however' ((nach is a complementiser, negative form of 'that'))

colinbatchelor commented 4 years ago

There is the parallel construction cha mhòr 'almost' in ScG. Again cha is the negative form of the copula and mhòr is an adjective.

I realise I have been inconsistent in how I have annotated these in Scottish Gaelic so far.

s08_082 (dev): tha mi a' smaoineachadh cha mhòr nach deach ainm Aitken dhan a’ leabhar ann an siud 'I think the name of Aitken almost went into the book there'

f03_049 (test): ... ach cha mhòr gu robh sad air fhàgail innte... 'but there was almost no dust left on it'

f04_027 (train): ... shaoil mi gu robh e agam, cha mhòr 'I wonder whether I had it, almost'

The last case I annotated as an advcl but the other two I made the head of deach and robh respectively, but I wonder if that was the wrong way round and they should both be advcls?

tlynn747 commented 3 years ago

Coming back to this - I think advcl is a good solution @colinbatchelor !

any thoughts @laurenCassidy ?

e.g. sent_id: 1417 Ní hamháin mar shaoránach ach mar ealaíontóir fosta goilleann tosca an tsaoil laethúil go dtí na scairteacha ar an fhile. "Not only as a citizen but also as an artist, the circumstances of everyday life touch the poet's (?)."

sent_id: 3169 Cuidíonn an Ghaeilge mar theanga pobail sa Ghaeltacht ní hamháin leis na sainghnéithe luachmhara de thraidisiún ...., ach tá sí mar chuid dílis agus inspioráideach... "Irish helps as a community language in the Gaeltacht, not only with the invaluable traditional characteristics but it's an integral and inspiring part...."