Determination of Syntactic Functions in Estonian Constraint Grammar

Proceedings of EACL '99 Determination of Syntactic Functions in Estonian Constraint Grammar Kaili Mfifirisep Institute of Computer Science University of Tartu Liivi 2, 50409 Tartu ESTONIA kaili~ut.ee Abstract This article describes the current state of syntactic analysis of Estonian using Constraint Grammar. Constraint Gram- mar framework divides parsing into two different modules: morphological disam- biguation and determination of syntac- tic functions. This article focuses on the last module in detail. If the morphologi- cal disambiguator achieves the precision more than 85% and error rate is smaller than 2% then 80-88% of words becomes syntactically unambiguous. The error rate of parser is 1-4% depending on the ambiguity rate of input. The main goal of this work is to elaborate an efficient parser for Estonian and annotate the Corpus of Estonian Written Texts syn- tactically. It is the first attempt to write a parser for Estonian. 1 Introduction The main idea of the Constraint Grammar (Karls- son, 1990) is that it determines the surface-level syntactic analysis of the text which has gone through prior morphological analysis. The process of syntactic analysis consists of three stages: mor- phological disambiguation, identification of clause boundaries, and identification of syntactic func- tions of words. This article focuses on the last module in detail. Grammatical features of words are presented in the forms of tags which are at- tached to words. The tags indicate the inflectional and derivational properties of the word and the word class membership, the tags attached during the last stage of the analysis indicate its syntactic functions. The underlying principle in determin- ing both the morphological interpretation and the syntactic functions is the same: first all the pos- sible labels are attached to words and then the ones that do not fit the context are removed by applying special rules or constraints. Constraint Grammar consists of hand written rules which by checking the context decide whether an interpre- tation is correct or has to be removed. Constraint Grammar seemed to suit best for the analysis of Estonian texts because its mechanism is simple and easily implementable, it can be well adapted for the Estonian language, it is at the same time sufficiently reliable (robust) and the re- sulting syntactic analysis that the Grammar gives suits various practical applications. 2 Syntactic Analysis of Estonian The Estonian language is a Finno-Ugric language and has got a rich structure of declensional and conjugational forms. The order of sentence con- stituents in Estonian is relatively free and influ- enced more by semantic and pragmatic factors. For morphological analysis of Estonian, we use the morphological analyser ESTMORF (Kaalep, 1997) that assigns adequate morphological de- scriptions to about 98% of tokens in a text. Mor- phologically analysed text is disambiguated by Constraint Grammar disambiguator of Estonian. The development of disambiguator is in process but 85-90% of words become morphologically un- ambiguous and the error rate of this disambigua- tot is less than 2% (Puolakainen, 1998). All the syntactic information is given by syntac- tic tags in constraint grammar framework. …

Content truncated to 500 words