nschneid
commented
8 years ago Never mind—the AMR has apparently been updated to remove the cycle, by introducing another person entity. Here is the original one with the cycle involving p2:
(h2 / have-concession-91
:ARG1 (a2 / and
:op1 (p2 / person
:mod (o3 / ordinary)
:domain p2
:mod (a3 / after-all))
:op2 (p3 / possible
:domain (l / look-01
:ARG0 p2
:manner (f / fury
:mod (s / silent))
:mod (o4 / only)
:manner (h / helpless))))
:ARG2 (g / gnash-01
:ARG0 (p / person
:mod (o / ordinary)
:quant (a / all))
:ARG1 (t / tooth
:part-of p)
:time (b / bring-02
:ARG1 (c2 / corrupt-01
:ARG1 (p4 / person
:ARG0-of (h3 / have-org-role-91
:ARG1 (g2 / government-organization
:ARG0-of (g3 / govern-01))
:ARG2 (o2 / official)))))))
@mdtux89 found at least one example of a cyclic AMR due to a tautological statement. Should we introduce a frame or other policy for tautologies to avoid this?
cjconsensus bolt12_3991_0111.12 (snt. 383 in workset bolt12-tune-master-dt, last updated on Sat Jan 10, 2015)
When corruption among government officials is brought up, the ordinary people all gnash their teeth, but the ordinary people are, after all, the ordinary people, and can only look on helplessly in silent fury.