Lietuvių kalbos denominatyviniai veiksmažodžiai ir greta jų vartojami pamatiniai žodžiai (tekstyno tyrimas)

LTReikšminiai žodžiai: Veiksmažodžiai; Pamatiniai žodžiai; Verbs; Base words.

ENThis paper aims to study cases when denominal and deadjectival (further referred to as denominal) verbs co-occur with their base words, cf. atstov-au-ti 'to represent' ^ atstov-as 'representative' both where the base word precedes the derivative and where the base word is used after the derivative: [PIRMININKAS]. Aliu. Komiteto atstovas ponas K. Dirgėla, kuris komiteto pavedimu atstovauja šiam dokumentui [...] '[CHAIR-MAN] Thank you. The representative of the committee Mr. K. Dirgėla who represents this document'. Bet vis dėlto mes, patinka kam ar nepatinka, atstovaujam tautai, esam tautos atstovai ir formuojam tam tikrą politiką [...] 'But nevertheless, whether someone likes it or not, we represent the nation, we are the representatives of the nation and we form certain policies'. In these excerpts, the speaker uses the same stem both as a noun and as a verb (with the corresponding derivational suffix) and, as Lipka (1987: 64) argues, the use of the constituents of complex lexemes in the preceding or the following co-text creates lexical cohesion and coherence. We started with a 1 million-word corpus of modern Lithuanian to test the methods of our study and then moved to an 82 million-word corpus. The lists of denominal verbs of both corpora were compiled and a text search application was developed to locate the cases when denominal verbs co-occur with their base words. To determine the size of the search window, we used a test list of denominal verbs with the suffix -au-ti. The window size determines the maximum distance the base word may be found before or after the denominal verb. When the window was gradually increased, the number of cases of co-occurrences also increased until the window of 170 words in one direction (= 340 words in both directions) words was reached. At this point, the number of cases found somewhat stabilized.The window of 170 words (= 340 words in both directions) was used to retrieve all co-occurrences of denominal verbs with their base words from the 1 million-word corpus, and these excerpts were manually reviewed to mark the cases when the derivative and the base word had a certain textual relationship. The relationship was recognized when the denominal verb and the base word belonged to the same episode and they were related to the same referent or one of them had a generic reference while another one was specific. To make the analysis of the 82 million-word corpus less time-consuming, the search window was reduced to 50 words (= 100 words in both directions) and only 5% of the excerpts for each suffix (4 800 excerpts in total) were manually reviewed to mark the cases of textual relationship. In the case of smaller search window, more cases of textually related co-occurrences were found (48% vs. 27% in the case of the larger search window). The relation between the total number of all co-occurrences of denominal verbs with their base words and the total number of denominal verbs was used to determine the productivity of each suffix. The denominal suffixes were also ranked according to the percentage of textually related co-occurrences and the ranking was virtually the same in both corpora. It can be concluded that the main division lies between the verbs typically based on adjectives (suffixes -ė-ti and -in-ti) and having lower productivity and less textually related co-occurrences and the verbs which are mostly based on nouns (suffixes -y-ti, -o-ti, -au-ti, -uo-ti) and typically have higher productivity and more textually related co-occurrences. The differences among the denominal suffixes were more evident in the case of the 1 million-word corpus and the window of 170 (340) words, while the data of the 82 million-word corpus and the window of 50 (100) words show no or less significant differences. [From the publication]