model,1-3-H89-1036,ak constraints </term> can be stated . These <term> constraints </term> either hold within the <term> elementary
other,11-10-H89-1036,ak <term> parser </term> selects a set of <term> elementary structures </term> associated with the <term> lexical
other,20-3-H89-1036,ak structures can be composed with a given <term> elementary structure </term> . The <term> ' grammar ' </term> consists
model,4-5-H89-1036,ak head </term> . There are no separate <term> grammar rules </term> . There are , of course , <term> '
lr,1-7-H89-1036,ak these structures are composed . A <term> grammar </term> of this form will be said to be '
other,4-18-H89-1036,ak TAGs </term> . We assign them regular <term> syntactic structures </term> while representing them semantically
tech,9-12-H89-1036,ak be adapted to take advantage of the <term> two steps parsing strategy </term> . The <term> system </term> parses <term>
other,6-3-H89-1036,ak constraints </term> either hold within the <term> elementary structure </term> itself or specify what other structures
other,10-4-H89-1036,ak of a <term> lexicon </term> where each <term> lexical item </term> is associated with a finite number
tech,10-19-H89-1036,ak finally show how they can be parsed by a <term> parsing strategy </term> as mentioned above . This paper proposes
other,8-13-H89-1036,ak unification formalisms </term> that have a <term> CFG skeleton </term> and that have a <term> TAG skeleton
lr,3-16-H89-1036,ak being developed . Subsets of these <term> lexicons </term> are being incrementally interfaced
model,6-6-H89-1036,ak rules </term> . There are , of course , <term> ' rules ' </term> which tell us how these structures
other,3-15-H89-1036,ak development . A <term> lexicons </term> for <term> French </term> is also being developed . Subsets
other,12-2-H89-1036,ak locality </term> ( as compared to a <term> context-free grammar </term> ) over which <term> constraints </term>
other,9-8-H89-1036,ak grammar </term> naturally follows from the <term> extended domain of locality </term> of <term> TAGs </term> . A general <term>
other,3-13-H89-1036,ak </term> . The <term> system </term> parses <term> unification formalisms </term> that have a <term> CFG skeleton </term>
other,12-18-H89-1036,ak representing them semantically as one <term> entry </term> . We finally show how they can be
lr,7-4-H89-1036,ak <term> ' grammar ' </term> consists of a <term> lexicon </term> where each <term> lexical item </term>
other,16-10-H89-1036,ak structures </term> associated with the <term> lexical items </term> in the <term> input sentence </term>
hide detail