Using Computers to Analyze and Compare Text in Translation

What are some ways to identify minute differences between multiple translations of the same text? How might DH tools aid in analysis of literature in translation? This workshop explores some of the thought processes and methodologies involved in using computational text analysis to study multiple translations of the same text, and demonstrates using Voyant Tools and AntConc to analyze and compare five translations of Anton Chekhov's story "Rothschild's Fiddle." Slides are available here.

Correction regarding the Voyant Tools Readability Index: Voyant Tools uses the Coleman-Liau index formula to calculate the Readability Index. The Coleman-Liau formula produces a metric that corresponds to a U.S. grade level. The Readability Index for my corpus' five translations of "Rothschild's Fiddle" indicate that the Long and Pevear/Volokhonsky translations have a higher metric (8.321 and 8.047 respectively), meaning they require a higher U.S. grade level to comprehend and, by that standard, are less readable. The Garnett and Fell translations have lower metrics, meaning they require a lower U.S. grade level to comprehend and are more readable. Contrary to my workshop statement, these results challenge my initial impressions from analyzing the translations' opening sentences. This discrepancy in readability demonstrates that there are various ways of defining readability. A comprehensive analysis of readability probably requires exploring multiple methods for calculating readability metrics.