ASV-Label
Login

16px-feed-icon Veröffentlichungen View this page in English

CRVH

Abstract:

This paper presents corpus based evidence from Russian for an ‘aspect coding asymmetry’ of verbs, the research question being whether and if then how differing lengths of verb forms ac- cording to aspect can be predicted. We assumed that each verb has a default aspectual value and that this value can be determined based on fre- quency. The most frequent aspectual value of a verb is taken as its default value. Whereas Zipf’s law says that frequency correlates with word length, our results disclose that lengths of verb forms in Russian correlate more strongly with their aspectual default value (default vs. non- default) than with frequency. We observed in addition a positive, but weaker, impact of infor- mation content (Cohen Priva 2008; Piantadosi et al. 2011) which was estimated from verb de- pendents in the syntactic contexts of verbs. A final result is the tendency that predictors mutu- ally exclude each other: The stronger the impact of frequency the weaker the impact of infor- mation content and vice versa.

Type: Inproceedings

Author: Giuseppe Celano and Michael Richter and Rebecca Voll and Gerhard Heyer
Title: Aspect coding asymmetries of verbs: The case of Russian
Booktitle: KONVENS 2018. PROCEEDINGS of the 14th Conference on Natural Language Processing
Year: 2018
Editor:Adrien Barbaresi, Hanno Biber, Friedrich Neubarth, Rainer Osswald
Pages:34 -- 39
Address:
@INPROCEEDINGS{CRVH,
AUTHOR = {Giuseppe Celano and Michael Richter and Rebecca Voll and Gerhard Heyer},
TITLE = {Aspect coding asymmetries of verbs: The case of Russian},
BOOKTITLE = {KONVENS 2018. PROCEEDINGS of the 14th Conference on Natural Language Processing},
YEAR = {2018},
EDITOR = {Adrien Barbaresi, Hanno Biber, Friedrich Neubarth, Rainer Osswald},
PAGES = {34 -- 39}
}