search

amrisi / amr-guidelines

240 stars 86 forks source link

have-degree-91 disagreements #215

Open timjogorman opened 6 years ago

timjogorman commented 6 years ago

I wanted to write some quick notes about the have-degree-91 disagreements; hopefully Claire can look and help clarify things too. There's one core thing that I think was unclear when we did this: when to put have-degree-91 on top:

githubhavedeg1 and githubhavedeg2

cbonial commented 6 years ago

The guidance in the proposed guidelines is this: Annotators are encouraged to use have-degree-91 as the root concept (as opposed to the adjective with a particular degree, or the entity characterized by that adjective) when a comparison seems to be the main focus of the sentence. I think that this leaves open the same type of disagreement that we see in AMR already concerning what the root concept should be, but I'm not sure what the best way to alleviate that is. I do think that if the matrix verb is a copula, and what follows is some kind of comparison of properties or quantities, then certainly have-degree-91 should be the root concept (e.g., (30) above). This explicit mention of copular structures could be mentioned in the guidelines. Beyond that, I'm not sure that the disagreement I see in the challenge pilot set is much greater than what we see in AMR generally given that this is a somewhat debatable choice?