Roleset help.01 Verbnet Class: NONE "help, aid":

Roles:

Arg0:helper


Arg1:project


Arg2:benefactive

Examples:

sentential argument (-)
Japanese money will help *trace* turn SE Asia into a more cohesive economic region.
Arg0:	Japanese money
Argm-MOD:	will
REL:	help
Arg1:	*trace* turn SE Asia into a more cohesive economic region.




nominal argument (-)
She was out *trace* to help the poor underpriviledged child.
Arg0:	*trace*=she
REL:	help
Arg1:	the poor underpriviledged child




passivized (-)
The stocks of banking concerns weren't helped much by the announcement.
Arg1:	The stocks of banking concerns
Argm-NEG:	n't
REL:	helped
Argm-MNR:	much
Arg0-by:	the announcement

benefactive, only for when the benefactor and the project are split under two nodes (no node contains both to the exclusion of other material):


benefactive (-)
Mary helped John with his homework.
Arg0:	Mary
REL:	helped
Arg2:	John
Arg1-with:	his homework




another transitive (-)
You'll be helping a neighborhood S&L in areas crucial to its survival.
Arg0:	You
Argm-MOD:	'll
REL:	helping
Arg1:	a neighborhood S&L
Argm-LOC:	in areas crucial to its survival

Sometimes one logical argument is split across a zillion nodes: Manufacturers' backlogs rose 0.5% in September to $4.36 billion, *trace* helped *trace* by defense spending. Logically, the Arg1 of help is "Manufacturers' backlogs rose 0.5% in September to $4.36 billion", but that material is spread across at least three nodes. The lower trace of 'help' is linked to "Manufacturers' backlogs", which is close enough to the truth for us to accept.