Open uhermjakob opened 6 years ago
I like request-response-91
. The other pragmatic situations seem more difficult to circumscribe....
Thanks Ulf! My temptation would be to handle this with some specific MWP frames; we could just do something like:
A few of the "that's what it is" instances do seem to me to just call for (t/ that :domain (i / it)) -- "that" and "it" being both actually anaphoric -- as in:
I think there can be details (what happened when, who tried to do what, etc.) without spreading his own propaganda. Because that's what it is.
There are a few idiomatic phrases in English with pragmatic meanings that are not easily captured using traditional frames for AMR annotation.
Here are a few examples, along with some original and current annotations that I think are too shallow or that are just placeholders for a better pragmatic roleset.
(i / it :domain (t / that))
, currently annotated as:(t / that-is-it-00)
# That's the main_point/difficulty/solution/all/the end/enough.(a / all :domain (t / that))
(e / enough :domain (t / that))
(e / enough :domain (e2 / enough))
CU's frame enough.01 (" sufficient, meeting a minimum requirement") is not really suitable, nor is
(have-quant-91 ... :ARG3 enough)
. Maybe we need an additional roleset for enough in the sense of "unwilling to tolerate any more of something undesirable".Person A: Let me tell you about this crazy thing that happened to me yesterday. long silence Person B: Yes?
For cases such as above or below, I propose to use request-response-91, in analogy to request-confirmation-91.
"Well?" said the geographer expectantly.
(r / request-response-91 :ARG0 (g / geographer) :manner ...)
Thoughts?