Cross-linguistic variations in syntactic complexity: insights from a multilingual parallel corpus

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216208%3A11210%2F24%3A10489058" target="_blank" >RIV/00216208:11210/24:10489058 - isvavai.cz</a>

Výsledek na webu

<a href="https://jakobson.korpus.cz/~rosen/public/2024_Singapur.pdf" target="_blank" >https://jakobson.korpus.cz/~rosen/public/2024_Singapur.pdf</a>

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Cross-linguistic variations in syntactic complexity: insights from a multilingual parallel corpus

Popis výsledku v původním jazyce

The study examines cross-linguistic variation in syntactic complexity (SC). Using the InterCorp multilingual corpus, with 49 languages annotated by Universal Dependencies (UD) and 6 syntactic complexity measures (SCMs), we analyse parallel texts both cross-linguistically and intra-linguistically (across text types), using InterCorp, a multilingual parallel corpus with Universal Dependencies (UD) annotation and syntactic complexity measures (SCMs). The pilot study of 12 languages has shown that: (i) Both language and text type influence the syntactic complexity of texts and sentences, but text type shows a larger effect size than language (in our sample). This highlights the importance of distinguishing text types in cross-linguistic analyses. (ii) Within fiction, languages cluster differently according to different SCMs (sLength and subRatio). The differences are given by structural variations between languages, but also - in Japanese - partly by the specifics of the UD annotation (tokenization). In fiction, languages show high degree of variation, due to the stylistic diversity of the analyzed texts.(iii) A strong correlation was observed between the two NP measures (maxNPDepth and maxNPLength) and between the two clausal measures (maxTreeDepth and subRatio). Most of the languages in the sample show a similar PCA pattern, but in other languages, patterns vary both cross-linguistically and intra-linguistically (text type variation).

Název v anglickém jazyce

Cross-linguistic variations in syntactic complexity: insights from a multilingual parallel corpus

Popis výsledku anglicky

The study examines cross-linguistic variation in syntactic complexity (SC). Using the InterCorp multilingual corpus, with 49 languages annotated by Universal Dependencies (UD) and 6 syntactic complexity measures (SCMs), we analyse parallel texts both cross-linguistically and intra-linguistically (across text types), using InterCorp, a multilingual parallel corpus with Universal Dependencies (UD) annotation and syntactic complexity measures (SCMs). The pilot study of 12 languages has shown that: (i) Both language and text type influence the syntactic complexity of texts and sentences, but text type shows a larger effect size than language (in our sample). This highlights the importance of distinguishing text types in cross-linguistic analyses. (ii) Within fiction, languages cluster differently according to different SCMs (sLength and subRatio). The differences are given by structural variations between languages, but also - in Japanese - partly by the specifics of the UD annotation (tokenization). In fiction, languages show high degree of variation, due to the stylistic diversity of the analyzed texts.(iii) A strong correlation was observed between the two NP measures (maxNPDepth and maxNPLength) and between the two clausal measures (maxTreeDepth and subRatio). Most of the languages in the sample show a similar PCA pattern, but in other languages, patterns vary both cross-linguistically and intra-linguistically (text type variation).

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

60203 - Linguistics

Projekt

<a href="/cs/project/LM2023044" target="_blank" >LM2023044: Český národní korpus</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2024

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů