File Information
File: 05-lr/acl_arc_1_sum/cleansed_text/xml_by_section/intro/91/w91-0103_intro.xml
Size: 1,850 bytes
Last Modified: 2025-10-06 14:05:07
<?xml version="1.0" standalone="yes"?> <Paper uid="W91-0103"> <Title>Towards Uniform Processing of Constraint-based Categorial Grammars</Title> <Section position="3" start_page="12" end_page="12" type="intro"> <SectionTitle> L </SectionTitle> <Paragraph position="0"> mantics and string representation ensures termination. This condition on lexical entries can be seen as a lexicalized !and computationally motivated version of GB's projection principle.</Paragraph> <Paragraph position="1"> Word-order domains. The string associated with a linguistic object (sign) is defined in terms of its word-order domain (Reape, 1989; Reape, 1990a). I take a word=order domain as a sequence of signs. Each of theSSe signs is associated with a word-order domain recursively, or with a sequence of words. A word-order domain is thus a tree.</Paragraph> <Paragraph position="2"> Linear precedence rules are defined that constrain possible orderings of signs in such a word-order domain. Surface strings are a direct function of word-order domains.' In the lexicon, the word-order domain of a lexical entry is defined by sharing parts of this domain with the arguments it subcategorizes for. Word-order domains are percolated upward. Hence word-order domains are constructed in a derivation by gradual instantiations (hence strings are constructued in a derivation by gradual instantiation as well). Note that this implies that an unsaturated sign is not associated with one string, but merely with a set of possible strings (this is similar to the semantic interpretation of unsaturated signs (Moore, 1989)).</Paragraph> <Paragraph position="3"> In lexical entries, word order domains are defined using Reape's sequence union operation (R.eape, 1990a). Hence the grammars are not only based on context-free string concatenation.</Paragraph> </Section> class="xml-element"></Paper>