A Computational Model of Finnish Sentence Structure
Esa Nelimarkka, Harri Jäppinen, Aarno Lehtola · DSpace repository (University of Tartu) · 1983
Introdjcbion The present paper propounds an outline of a computational model of Finnish sentence structures.Although we focus on Finnish we feel that the ideas behind the model might be applicable to other languages as well, in particular to other inflectional free word order languages.A parser based on this model is being implemented as a component of a larger system, namely a natural language data base interface.There it will follow a component of morphological analysis (see JSppinen et al C83); hence, throughout the present paper it is assumed that all relevant morpho logical and lexical information is computationally available for all words in a sentence.Even though we have a data base application in mind, sen tence analysis will be based on general linguistic knowledge.All applicatio-.dependent inferences are left to subsequent modules which are not discussed here.The linguistic foundations We shall freely borrow ideas of Anderson [1], [2], Tarvainen [11] and Pajunen [9] , and Fillmore C33, m and Siro ClOj concerning dependency and case grammars.We shall use the latter grammar to introduce semantics into syntacs and the former to give a basis for a functional syntax where the subordinate dependency relations are specified with formal binary rela tions.The structure which these grammars will impose on sentences is a dependency-constituency hybrid structure with labelled dependants, similar to the sister-dependency structure of Hudson [7].We shall briefly expli cate and reason our choices.