The ǂconstruction of reference corpus of contemporary Serbian

[M. Utvić]

PDF/A (417 listova)

Korpusna lingvistika / Corpus linguistics

U prvom delu rada se razmatraju opˇsta pitanja koja se odnose na definiciju, istorijat, parametre i klasifikaciju korpusa, kao i na korpusnu lingvistiku kao metodologiju istraˇzivanja jezika. Posebna paˇznja je posve´cena pitanjima reprezentativnosti i balansiranosti korpusa kao uzorka jezika. Takode je detaljno razmotren i uticaj Interneta, odnosno veba, na kritiˇcko preispitivanje definicije korpusa. Kao parametri korpusa, posebno su analizirani nosaˇc, domen i namena, obim (veliˇcina), period, izvor/medijum, anotacija i viˇsejeziˇcnost. Na osnovu tih parametara su opisane mogu´ce klasifikacije korpusa i posebno su izdvojeni nacionalni korpusi kao opˇsti, referentni korpusi koji pretenduju da reprezentuju jezik jedne zemlje. Detaljno su analizirani nacionalni korpusi slovenskih jezika. Poseban odeljak je posve´cen istorijatu srpske korpusne lingvistike. Na kraju prvog dela rada su navedeni ciljevi rada: razmatranje mogu´cnosti izgradnje opˇsteg korpusa srpskog jezika koji bi bio elektronski, dinamiˇcki, sinhroni, balansiran, anotiran (morfoloˇski, strukturno, bibliografski), kao i mogu´cnosti izgradnje prate´cih viˇsejeziˇcnih paralelnih korpusa u kojima je srpski izvorni ili ciljni jezik...

The problem regarding the methods and tools to construct a corpus of contemporary Serbian as a reference language resource is considered in this thesis. The thesis consists of three parts. General questions related to definition, history, parameters and classification of corpora, as well as to corpus linguistics as a methodology in language research, are considered in the first part of the thesis. The special attention is paid to questions regarding representativeness and balance of corpus as a language sample. The affect of Internet/Web on critical review of corpus definition is considered in detail, too. Corpus parameters (storage medium, domain/purpose, size, time span, mode of communication, annotation and multilinguality) are particularly analysed. Possible classifications of corpora, based on these parameters, are described with emphasis on national corpora as general reference corpora which are supposed to represent the national language of a country. National corpora of Slavic languages are analysed exhaustively. A special section is dedicated to the history of Serbian corpus linguistics. The goals of thesis are listed in the end of the first part of the thesis: considering possibilities for construction of general, electronic, dynamic, synchronous, balanced, morphosyntactically and bibliographically-annotated corpus, as well as the possibilities for construction of multilingual parallel corpora with Serbian as source or target language...

korpusna lingvistika, srpski jezik, raˇcunarska lingvistika, obrada prirodnog jezika, morfosintaksiˇcka anotacija, referentni korpus

corpus linguistics, Serbian, computational linguistics, natural language processing, morphosyntactic annotation, reference corpus

811.163.41'322.2(043.3)

Srpski jezik

Tekst