Dialect Corpora Taken Further: The DynaSAND Corpus and Its Application in Newer Tools

Dialect Corpora Taken Further: The DynaSAND corpus and its application in newer tools Jan Pieter Kunst a and Franca Wesseling b aMeertens Institute, Royal Netherlands Academy of Arts and Sciences, Joan Muyskenweg 25, 1090 GG Amsterdam, The Netherlands janpieter.kunst@meertens.knaw.nl bMeertens Institute, Royal Netherlands Academy of Arts and Sciences, Joan Muyskenweg 25, 1090 GG Amsterdam, The Netherlands franca.wesseling@meertens.knaw.nl Abstract. In this paper we will expand on the creation and structure of the DynaSAND database as a case study of a corpus tool. Furthermore we will focus on its implementation in other search engines, thereby illustrating how the underlying data is decoupled from its original interface and used in new ways. Keywords: Dutch dialects, corpus, user interfaces, multiple corpora search 1 Introduction The Syntactic Atlas of the Dutch Dialects (SAND) corpus is one of the results of a large-scale dialect syntax project conducted between 2000 and 2003 in the Netherlands and the Dutch- speaking parts of Belgium and France (cf. Barbiers, Cornips & Kunst 2007, Barbiers & Bennis 2007). The goal of the SAND project was to gain insight into the (morpho)syntactic variation within and between Dutch dialects, and to make this variation visible in a clear manner. The information uncovered is of great interest to linguists, in particular syntacticians and dialect researchers, because it provides them with a huge amount of dialect data, structured in a well- organized manner. The output of the SAND project, i.e. the dialect data, has been made accessible in various ways. First, there are two printed atlases that bring into view the dialectal variation concerning various morphosyntactic features such as the use of reflexive pronouns and the construction of relative clauses. Second, there is a freely accessible web-based dynamic atlas (DynaSand, available at www.meertens.knaw.nl/sand/), which displays the dialect data in an orderly fashion. The DynaSAND web application is meant to make the data that was collected in the SAND project available to linguists who do not have advanced computational skills. The linguistic data is stored in a relational database, a search engine is included, and the data is linked to the audio files of the original fieldwork recordings. Also, the geographic coordinates of the locations of the fieldwork are added to the database, which are used for drawing dynamic maps on the basis of search results. A new development is the incorporation of individual language resources like DynaSAND in larger entities that are used to access and search different language resources simultaneously. This is accomplished by the addition of a web service interface to the DynaSAND corpus, i.e. a search interface that is not meant for humans but for other computer programs, so that the data from the corpus can be used in other applications besides DynaSAND itself. Two examples of projects that make use of this extra interface to the DynaSAND corpus are Edisyn (European Dialect Syntax) and MIMORE (Microcomparative Morphosyntax Research Tool). 2 The SAND corpus and DynaSAND web application PACLIC 24 Proceedings 759 The SAND corpus contains data from 267 dialects collected in …

Content truncated to 500 words