streusle
streusle copied to clipboard
STREUSLE: a corpus with comprehensive lexical semantic annotation (multiword expressions, supersenses)
(based on efforts to clean up Little Prince annotations) - sit\_down, stand\_up, lie\_down, sit\_up - far\_away, close\_by - as\_for (DISC) - belong\~to - out\_of - because\_of
Per carmls/snacs-guidelines#116
"out-of-place" is treated as an MWE. Even though syntactically "of" has "out" as its governor and "place" as its object, the `heuristic_relation` info is stored on strong expressions, so "of"...
(in sentence mentioning Wildwood) For other Approximators the modified NP is treated as the obj, not the gov. Here it seems both "for" and "over" are syntactically analyzed as `case`,...
"Buy your kids somethin nice >>with
❌ marks ones with the idiomatic meaning of 'slightly': ``` $ egrep '^[#] mwe =' streusle.conllulex | fgrep 'a little' # mwe = Anyone else find it a❌little suspicious that...
Omitting spatial in_front_of and on_top_of: ``` ./tquery.py streusle.go.json +lc=P +ll!="on top of|in front of" +ss +ss2 upos="ADP.*NOUN.*ADP'[)]$" # b'a38b99b' ~ ./tquery.py streusle.go.json +lc=P '+ll!=on top of|in front of' +ss +ss2...
"they seemed more interested~in helping me find the right car rather_then just make_ a _sale" (245160.4) "rather then" has gov "make" and obj null, but in fact "make" should be...
We have been using `DET` for most quantity MWEs starting with "a", but this seems dubious: om "a bit tired" or "a bit of cake", single-word determiners cannot be substituted...
be a big baby be a joke be a no brainer be a pain be a plus be first call be happy camper be in for a treat be in...