Tree Adjoining Grammar

from Wikipedia, the free encyclopedia

Tree-adjoining grammar s (TAG), also tree -adjoining grammars , are formal grammars that were introduced by Aravind Joshi and are used in computational linguistics to describe natural languages .

TAGs are similar to context-free grammars , but use trees instead of rules as the smallest elements.

TAGs are described as mildly context-sensitive ; so they are stronger than context-free grammars, but weaker than context-sensitive grammars in the Chomsky hierarchy . Hence, they are probably strong enough to generate natural languages, but also weak enough to be efficiently parsable .

Web links