Predicate: be
be: Frames file for 'be' based on sentences in wsj. No Verbnet entry.
Roleset id: be.01 , copula, vncls: , framnet: Performers_and_roles
Roles:
Arg2 will probably be a PRD node, but NEVER a VP node!.
Arg1: topic
Arg2: comment
Example: poor John
person: third, tense: present, aspect: ns, voice: active, form: full
John is an idiot.
Arg1: John
Rel: is
Arg2: an idiot
Roleset id: be.02 , existential, vncls: , framnet:
Roles:
Arg1: thing that is
Example: what it is, what it was, what it shall be
person: third, tense: present, aspect: ns, voice: active, form: full
John doesn't think, therefore he isn't.
Arg1: he
Rel: is
Argm-neg: n't
Example: with support
person: third, tense: present, aspect: ns, voice: active, form: full
There's a foul odor in the air whenever John's around.
Arg1: a foul odor in the air
Rel: 's
Argm-tmp: whenever John's around
Roleset id: be.03 , auxiliary, vncls: , framnet:
Roles:
do not tag--no arguments! Will be followed by a VP node.
Example: one of many
Do not tag!
person: ns, tense: ns, aspect: ns, voice: ns, form: ns
John is eating.
Predicate: be_like
be_like: This usage is common in data that contains spoken English. Examples taken
from EBC.
Roleset id: be.04 , be_like: multiword expression akin to "say", vncls: , framnet:
Roles:
Arg0: speaker
Arg1: utterance
Example: definitely spoken
When it comes to ABC *T*-1 they be like A B.
Argm-tmp: When it comes to ABC *T*-1
Arg0: they
Rel: [ be] [ like]
Arg1: A B
Example: say
And I 'm like why 's the door open *T*?
Argm-dis: And
Arg0: I
Rel: [ 'm] [ like]
Arg1: why 's the door open *T*?
Example: definitely spoken
When it comes to ABC *T*-1 they be like A B.
Argm-tmp: When it comes to ABC *T*-1
Arg0: they
Rel: [ be] [ like]
Arg1: A B
Example: one more example
I know 0 my producers are probably like ah uh.
Arg0: my producers
Rel: [ are] [ like]
Arg1: ah uh