Abstract
After introducing the notion of lexicalization, this chapter reviews the state of the art in natural language generation with respect to this topic: Sections 2.2 to 2.5 look at the problem from four different viewpoints. Then, Section 2.6 summarizes several of the more important ideas and identifies taks that urgently need progress. The specific goals for this book are set forth, which were, in general terms, already sketched at the end of the previous chapter.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Notes
The theoretical feasibility of separating these tasks has often been questioned (e.g., by Danlos [1987]), and some practical generators that employ a truly integrated architecture have been proposed (e.g., [Ward 1991; Reithinger 1992; Kantrowitz and Bates 1992]). Still, the major argument in favor of a two-step, modular design is that it keeps control flow simple and separates the different knowledge sources involved.
For example, even in the seemingly innocent choice of prepositions, we notice stylistic differences like the one between on and upon. More importantly, connectives can play a significant role in conveying aspects of meaning, as investigated for instance by Elhadad and McKeown [1990], and by [Grote et al. 1997]. Also, see the distinction between discourse-oriented and proposition-oriented closed-class items, made by Pustejovsky and Nirenburg [1987].
While theoretical, or formal, linguistics often prefers to ignore idioms and other “messy” aspects of linguistic reality, there is a fair amount of literature on idioms stemming from applied linguistics, where they are of more interest because of, inter alia, their role in language teaching.
For example, Becker lists the oldest profession and to blow up in the same category ‘polywords’, but the first is a completely frozen expression, whereas the second is a fully inflectable verb phrase that can also be passivized.
Possibly as a sign of change, there is a collection of recent work in [Everaert et al. 1995].
This INGEST does not correspond to the Schankian CD primitive mentioned earlier.
See, for instance, [Appelt 1985; Novak 1988; Dale 1992] and, focusing on the notion of text cohesion and avoiding the repetition of identical noun groups, [Granville 1984; Buch-berger and Horacek 1988]. A broader survey of ‘discursive constraints’ on lexicalization, including pronominalization decisions, can be found in [Robin 1990].
Cumming [1986, p. 11] concludes this in her survey as well as McDonald [1991, p. 229] and Matthiessen [1991, p. 277] in their analyses of the role of lexicalization.
For a detailed discussion of the interaction between lexical and syntactic decisions with specific English/German examples, see Mehl [1995].
This problem is, so to speak, the sentence-planning version of the “generation gap” that Meteer [1992] has dealt with on the level of text planning.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 1999 Springer Science+Business Media New York
About this chapter
Cite this chapter
Stede, M. (1999). Lexicalization in NLG. In: Lexical Semantics and Knowledge Representation in Multilingual Text Generation. The Springer International Series in Engineering and Computer Science, vol 492. Springer, Boston, MA. https://doi.org/10.1007/978-1-4615-5179-9_2
Download citation
DOI: https://doi.org/10.1007/978-1-4615-5179-9_2
Publisher Name: Springer, Boston, MA
Print ISBN: 978-1-4613-7359-9
Online ISBN: 978-1-4615-5179-9
eBook Packages: Springer Book Archive