FrameNet

FrameNet is a hand-built, corpus-based lexical resource for English that describes word meaning through the situations words evoke. It is developed at the International Computer Science Institute (ICSI) in Berkeley, California, drawing on the frame semantics of the linguist Charles J. Fillmore. Its starting point is simple but powerful: many words cannot be understood in isolation, only against a frame, a schematic scene with participants. You cannot grasp «to sell» without a seller, a buyer, some goods and a price.

In 1998, Baker, Fillmore and Lowe presented the project as a lexical database based on corpus annotations. That publication establishes a documented origin; a current and changeable project role need not become part of the definition.

Frames, frame elements and lexical units

FrameNet rests on three building blocks. Frames are the schematic situations: «Commerce_buy» (a purchase), «Apply_heat» (cooking with heat) or «Being_born». Each frame defines its frame elements, the roles involved: in «Commerce_buy» these are the buyer, the seller, the goods and the money. And each frame gathers the lexical units, the words that evoke it: «buy» and «purchase» trigger «Commerce_buy»; «fry», «bake» and «boil» trigger «Apply_heat». A single scene thus groups verbs, nouns and adjectives that share a background, and every sense is illustrated with real annotated sentences.

What it is for: semantic roles

One concrete task built on FrameNet is frame-semantic parsing: identifying the target, classifying the frame and labeling its semantic roles. In «Ana bought a car from Luis for 9,000 euros», those roles can represent Ana as buyer, Luis as seller, the car as goods and the figure as price. A 2016 paper also documents an application of frame semantics to question interpretation; that is a demonstrated use, not a promise about language understanding in general.

Against WordNet and in the age of LLMs

FrameNet should not be confused with WordNet. WordNet organises words into synonym sets and hierarchies —what is a synonym of what, what is a kind of what—; FrameNet organises situations and the roles they contain. One maps the vocabulary; the other maps the world the vocabulary refers to. Its limits should be stated through evidence: a FrameNet-based question-answering system published in 2007 described coverage problems in earlier approaches, and a 2025 study specifically evaluated whether LLMs can extract frame-semantic arguments. These are dated results on defined tasks, not a general guarantee of accuracy or transparency.