You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Until recently, corpora like EWT used the relations {nmod:tmod, nmod:npmod, obl:tmod, obl:npmod} for temporal and non-temporal NP adjuncts. This was recently revised to :unmarked for all NP adjuncts regardless of temporality. UniversalDependencies/docs/issues/1028
In Atis, however, :npmod is not used, and :tmod is applied on what appears to be a purely semantic basis, including to PPs (though it is not 100% consistent, as can be seen from the attachments of "tomorrow").
To bring this in line with the new English standard,
Current :tmod nominals (including PPs) should move the temporality information to a MISC feature, Temporal=Yes
nmod and obl nominals lacking a case dependent should (at least in most cases) add the :unmarked subtype
The text was updated successfully, but these errors were encountered:
Until recently, corpora like EWT used the relations {
nmod:tmod
,nmod:npmod
,obl:tmod
,obl:npmod
} for temporal and non-temporal NP adjuncts. This was recently revised to:unmarked
for all NP adjuncts regardless of temporality. UniversalDependencies/docs/issues/1028In Atis, however,
:npmod
is not used, and:tmod
is applied on what appears to be a purely semantic basis, including to PPs (though it is not 100% consistent, as can be seen from the attachments of "tomorrow").To bring this in line with the new English standard,
:tmod
nominals (including PPs) should move the temporality information to a MISC feature,Temporal=Yes
nmod
andobl
nominals lacking acase
dependent should (at least in most cases) add the:unmarked
subtypeThe text was updated successfully, but these errors were encountered: