tibetan-nlp / annotation-docs

Tibetan annotation docs
2 stars 0 forks source link

Auxiliaries should not depend on auxiliaries #13

Closed heacu closed 7 years ago

heacu commented 7 years ago

Under "The Status of Function Words" in the UD syntax guidelines, it says:

The primacy of content words implies that function words normally do not have dependents of their own. In particular, it means that multiple function words related to the same content word always appear as siblings, never in a nested structure, regardless of their interpretation. A typical case is that of auxiliary verbs, which never depend on each other.

In the following example from page 11a, auxiliaries are chained, with one depending on another:

screen shot 2017-11-10 at 2 46 16 pm

I will correct this example in light of the above advice, and see if there are other sentences to fix from the other annotated pages.

heacu commented 7 years ago

Here's a similar example from Mila 13a which an auxiliary depends on a copula.

screen shot 2017-11-10 at 3 25 57 pm

heacu commented 7 years ago

Another example from Mila 13a:

screen shot 2017-11-10 at 3 29 13 pm

heacu commented 7 years ago

Examples from Mila 13b:

screen shot 2017-11-10 at 3 31 13 pm

heacu commented 7 years ago

From Mila 14b:

screen shot 2017-11-10 at 3 34 08 pm

And from Mila 15b:

screen shot 2017-11-10 at 3 36 30 pm

screen shot 2017-11-10 at 3 37 28 pm

heacu commented 7 years ago

I have now modified all of the above examples in accordance with the advice that auxiliaries should not depend on auxiliaries, etc.

heacu commented 7 years ago

Here is one other example from Mila 13b:

screen shot 2017-11-10 at 9 01 47 pm

Fixing this example required loosening the linking conditions for auxiliaries. Now the aux relation is permitted to point back to all manner of POS categories, due to the possibility of an intervening copula. This might seem surprising, but it does follow the UD guidelines for cop, where it states:

If the copula is accompanied by other verbal auxiliaries for tense, aspect, etc., then they are also given a flat structure, and taken as dependents of the lexical predicate.

heacu commented 7 years ago

The documentation now mentions the requirement that stacked auxiliaries should all depend on the main verb rather than hierarchically depending on each other: https://tibetan-nlp.github.io/lim-annodoc/#auxiliaries