Predicate: total
total: Frames file for 'total' based on survey of initial sentences of big
corpus
Roleset id: total.01 , sums to, vncls: 54.1 22.2, framnet: Amounting_to , Adding_up
total.01: Many, many adjectives to be avoided (eg "total panic"). Member of VNcls register-54.1, amalgamate-22.2 (although some of the examples in the VNcls are questionable).
Roles:
        Arg1: thing being counted (vnrole: 54.1-theme, 22.2-patient)
        Arg2: count (vnrole: 54.1-attribute)
Example: ergative
        Newsweek's ad pages totaled 1620, a drop of 3.2% from last year.
        Arg1: Newsweek's ad pages
        Rel: totaled
        Arg2: 1620, a drop of 3.2% from last year
Roleset id: total.02 , slang: wreck completely, vncls: , framnet: Render_nonfunctional
total.02: I doubt this will appear in WSJ.
Roles:
        Arg0: wrecker
        Arg1: wreck
Example: oh too true
        John totaled his car.
        Arg0: John
        Rel: totaled
        Arg1: his car
Predicate: total_up
total_up: I can say this; not sure about WSJ
Roleset id: total.03 , count, vncls: 54.1, framnet:
total.03: Corresponds to Vncls register-54.1 (although VPC not member)
Roles:
        Possibly takes arg2 "EXT" as well.
        Arg0: counter (vnrole: 54.1-agent)
        Arg1: thing counted (vnrole: 54.1-theme)
Example: bills
        John totaled up the restaurant bill.
        Arg0: John
        Rel: [totaled] [up]
        Arg1: the restaurant bill