Semantic Role Labeling

22.1 Semantic roles

1. What do the subjects of “Sasha broke the window” and “Pat opened the door” have in common?
   Both are volitional, often animate actors with direct causal responsibility for their events: both are AGENTs. The objects, prototypically inanimate things affected by the action, are THEMEs.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 22, §22.1, p. 500
2. What are thematic roles?
   Roles capturing semantic commonalities between arguments of different predicates (AGENT, THEME…); modern formulations are due to Fillmore (1968) and Gruber (1965). There is no agreed universal set; most have about a dozen.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 22, §22.1, p. 500
3. Define the common thematic roles.

   role	definition
   AGENT	volitional causer of an event
   EXPERIENCER	experiencer of an event
   FORCE	non-volitional causer
   THEME	participant most directly affected
   RESULT	end product of an event
   CONTENT	proposition or content of a propositional event
   INSTRUMENT	instrument used in an event
   BENEFICIARY	beneficiary of an event
   SOURCE	origin of the object of a transfer
   GOAL	destination of the object of a transfer

   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 22, §22.1, p. 500
4. What thematic roles appear in: “The wind blows debris from the mall into our yards”?
   The wind is a FORCE (non-volitional causer), debris the THEME, from the mall the SOURCE, into our yards the GOAL.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 22, §22.1, p. 500
5. Why do computational systems use semantic roles?
   As a shallow meaning representation that supports simple inferences impossible from surface strings or parse trees, e.g. that “Company A acquired Company B” answers “Was Company B acquired?”; also as an intermediate language, e.g. for MT.
   • Jurafsky & Martin, Speech and Language Processing, 3rd ed. draft, ch. 22, §22.1, p. 500