Representing and Rendering Linguistic Paradigms

Representing and Rendering Linguistic Paradigms David Penton and Steven Bird Department of Computer Science and Software Engineering The University of Melbourne, Victoria 3010, Australia {djpenton,sb}@csse.unimelb.edu.au Abstract Linguistic forms are inherently multi-dimensional. They exhibit a variety of phonological, orthographic, morphosyntactic, semantic and pragmatic properties. Accordingly, linguistic analysis involves multi- dimensional exploration, a process in which the same collection of forms is laid out in many ways until clear patterns emerge. Equally, language documentation usually contains tabulations of linguistic forms to illustrate systematic patterns and variations. In all such cases, multi-dimensional data is projected onto a two-dimensional table known as a linguistic paradigm, the most widespread format for linguistic data presentation. In this paper we develop an XML data model for linguistic paradigms, and show how XSL transforms can render them. We describe a high-level interface which gives linguists flexible, high-level control of paradigm layout. The work provides a simple, general, and extensible model for the preservation and access of linguistic data. 1 Introduction A linguistic paradigm is any kind of rational tabu- lation of linguistic forms, such as phrases, words, or phonemes, intended to illustrate contrasts and systematic variation (Bird, 1999). A characteristic property of paradigms is that interchanging entire rows or columns does not change the interpretation of the information. We view paradigms as a two- dimensional arrangement of elements and attributes, with optional row and column labels. An exam- ple of a paradigm for the German definite article is shown in Figure 1, with the labelling of number and gender at the top, and case on the left. This paper describes a relational data model for linguistic paradigms, together with an XML based approach for representing and rendering them. An XSLT implementation provides proof of concept.1 This work presents a simple and general model for 1The implementation is available from //www.csse.unimelb.edu.au/research/lt/http: projects/paradigms/ Figure 1: Paradigm for German definite article (Finegan, 1999, 60) the preservation and access of linguistic paradigms, and can generate an extensive range of useful visu- alisations. This paper is organised as follows. In §2 we dis- cuss the existing computational models in linguistic paradigms and the lack of an existing formalism. In §3 we discuss the data model, while §4 and §5 we provide an example of how to generate and visualise a linguistic paradigm. In §6 we discuss the query engine implementation and motivate each opera- tion. In §7 we describe the transformation of the paradigm data into an XHTML presentation form. Finally, §8 discusses the significance of the work and outlines several areas for further investigation. 2 Background Traditionally, there have been two sources for com- putational representations for linguistic paradigms, descriptive linguistic tools and technologies for lan- guages having complex morphology. There has been little formal work in either area concerning the best form for this linguistic data type. Here we examine some of the more widely-used models and their drawbacks. Of the descriptive linguistic tools, perhaps the foremost are Shoebox2 and CHILDES.3 Shoebox is an interlinear text editor popular among field linguists …

Content truncated to 500 words