Probability distribution of dependency distance and dependency type in translational language

Bibliographic Data

ID22236256
AuthorsLu Fan (0000-0002-0896-6294, Shaanxi Provincial People's Hospital, corresponding author), Yue Jiang (0000-0002-0310-2657, Xi'an Jiaotong University)
Year2023
Volume10
Issue1
Publication date2023-12-06
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueHumanities and Social Sciences Communications (JOURNAL)
Journal identifiersISSN: 2662-9992 • E-ISSN: 2662-9992
PublisherSpringer Science and Business Media LLC (PUBLISHER)
DOI10.1057/s41599-023-02427-x
OpenAlexW4389399755
LanguageEN
Citations received4
References cited26

As a “third code”, translational language attracts considerable attention in linguistics research due to its distinctive features. Adopting the quantitative linguistic approach, the current study examines its features by investigating the mean dependency distance (MDD), as well as the probability distribution of the individual dependency distances (DDs) and distribution of a high-frequency dependency type in translational language. The MDD and the distributions were tested in a self-built corpus which contains parallel and comparable language materials in both Chinese-English and English-Chinese translations. The results show that: (1) compared with source texts and native texts, translated texts in both translation directions yield an MDD in between; (2) both the distribution of DDs and that of the dependency type nsubj follow the Zipf-Alekseev distribution in translated texts, as in source texts and native texts; (3) the in-between feature is further confirmed by parameters a and b in Chinese-English translation materials when fitting the distribution of DDs to Zipf-Alekseev distribution; (4) translational texts in both directions show higher a and lower b than their source and native texts when fitting the DD Distribution of dependency type nsubj to Zipf-Alekseev distribution. These findings suggest that, on the one hand, dependency distance minimization (DDM) occurs in translational language, which is consistent with native language and reflects a general tendency of natural languages to reduce cognitive load; on the other hand, translational language presents distinctive feature in nsubj type, but in most cases, it is subject to the gravitational pull of both source and target language systems, exhibiting a “compromise” feature in between. The current study highlights the contribution of syntactic quantitative methods to deeper understanding of the complexity of translational language and its cognitive underpinnings

Linguistics · Natural language processing · Statistics · Zipf's law · Authorship Attribution and Profiling · Chemistry · Computer Science · Mathematics · Natural Language Processing Techniques · Text Readability and Simplification · Artificial Intelligence