Results of the summarizing process for the arXiv paper: 1709.04359v1

In their paper titled "Linguistic Features of Genre and Method Variation in Translation: A Computational Perspective," authors Ekaterina Lapshninova-Koltunski and Marcos Zampieri explore the use of text classification methods to investigate genre and method variation in an English-German translation corpus. The authors utilize linguistically motivated features, which represent texts using a combination of part-of-speech tags arranged in bigrams, trigrams, and 4-grams. They employ a Bayesian classifier with Laplace smoothing as the classification method. The output of these classifiers is then used to carry out an extensive feature analysis on the main differences between genres and methods of translation. The study's findings reveal that certain linguistic features are more prevalent in specific genres or translation methods. For instance, translations produced by human translators tend to have more complex sentence structures than those generated by machine translation systems. Additionally, literary translations exhibit more variation in vocabulary compared to technical translations. The authors argue that their approach can be used for further research into genre and method variation across different language pairs and can inform the development of better machine translation systems. This paper is set to appear as a book chapter in Grammar of Genres and Styles published by De Gruyter.

• The paper explores genre and method variation in an English-German translation corpus
• Linguistically motivated features are used to represent texts using part-of-speech tags arranged in bigrams, trigrams, and 4-grams
• A Bayesian classifier with Laplace smoothing is employed as the classification method
• Certain linguistic features are more prevalent in specific genres or translation methods
• Human translators tend to produce translations with more complex sentence structures than machine translation systems
• Literary translations exhibit more variation in vocabulary compared to technical translations
• The approach can be used for further research into genre and method variation across different language pairs
• The study's findings can inform the development of better machine translation systems.

Exploring Genre and Method Variation in Translation: A Computational Perspective

In their paper titled "Linguistic Features of Genre and Method Variation in Translation: A Computational Perspective," authors Ekaterina Lapshninova-Koltunski and Marcos Zampieri explore the use of text classification methods to investigate genre and method variation in an English-German translation corpus. The paper is set to appear as a book chapter in Grammar of Genres and Styles published by De Gruyter.

Background

Translation has been studied extensively from a linguistic perspective, but there has been limited research into the differences between genres or methods of translation. To address this gap, Lapshninova-Koltunski and Zampieri propose a computational approach that utilizes linguistically motivated features to represent texts using part-of-speech tags arranged in bigrams, trigrams, and 4-grams. They employ a Bayesian classifier with Laplace smoothing as the classification method. The output of these classifiers is then used to carry out an extensive feature analysis on the main differences between genres or methods of translation.

Findings

The study's findings reveal that certain linguistic features are more prevalent in specific genres or translation methods. For instance, translations produced by human translators tend to have more complex sentence structures than those generated by machine translation systems. Additionally, literary translations exhibit more variation in vocabulary compared to technical translations. This suggests that different types of texts require different approaches when it comes to translating them accurately from one language into another.

Conclusion

The authors argue that their approach can be used for further research into genre and method variation across different language pairs and can inform the development of better machine translation systems. By understanding how certain linguistic features vary across genres or methods, researchers can gain valuable insight into how best to translate texts accurately while preserving their original meaning as much as possible.