Mass destruction about amr-guidelines

[cbonial]

Another common phrase is "nuclear reactor," which is not being verbalized to "react," but "nuclear" is being stemmed to "nucleus". These sorts of pertainym stemmings are still difficult for me to judge -- does this seem appropriate to everyone?

[uhermjakob]

I concur with Claire with respect to :degree for mass:

(w / weapon
  :ARG2-of (d / destroy-01
             :degree (m / mass)))

I'd be happy to add more such sentences and phrases to the consensus guidelines workset, which gets priority listing in searches with the AMR Editor search tool.

Regarding nuclear/nucleus: Yes, it can sometimes be a bit hard to judge which way to map, which is why we've been working on reference lists such as http://www.isi.edu/~ulf/amr/tmp/new-pert.txt

(r / reactor
  :mod (n / nucleus))

I can't remember how to format these with the correct indentation in Github.

I usually just leave a blank line before and after an AMR and indent at least 4 spaces at the top level.

[nschneid]

My two cents: I dislike nuclear -> nucleus in the sense of nuclear weapons, nuclear energy, etc. (Yes, there is a technical reason they are called "nuclear", but these are cultural concepts that probably do not bring nuclei to mind whenever we talk about them.)

If it is a scientific discussion of atomic structure, then sure, "nucleus" should apply.

I like always normalizing atomic -> atom, but only because we can say (informally) "atom bomb" as well as "atomic bomb".

[cbonial]

I agree with Nathan; sometimes the semantic distances between the pertainym and the original concept has grown, as we see with other etymologically related parts of speech. It's a tough call, but I think as long as we can do it consistently with the sorts of tools Ulf is developing, that's the most important thing.

[cbonial]

Closed: See Ulf's AMRs above with "degree" and "nucleus" for correct annotation.