Subject
DMITRI NABOKOV on "mathematical analysis of ADA translation
samples
samples
From
Date
Body
EDNOTE. Some days ago Lyuba Tarvi offered a statistical analysis of the four posted ADA samples. Below DN comments.
I would add that one of the difficulties of statistical analysis of literary works is that it often (more or less accurately) measures features that constitute (in the case at hand) only some of the features that make a good translation. To put it an another way, IQ tests tend to measure whatever they measure but it is far from clear that they provide a global quantification of that mysterious thing called Intelligence. Dr. Tarvi's analysis is presumably an accurate measure of some features of the translations but not all--as I am sure she would be the first to admit.
--------------------------------
FROM: DMITRI NABOKOV
"Both the Ilin & Skyarenko versions are pretty good. ... With all due respect, the mathematical analysis proposed by Lyuba Tarvi is not my idea of comparison. This is borne out by the fact that Kirichenko's choices of individual words may coincide with mine, while her sentences and paragraphs are sometimes an incomprehensible jumble, which is not reflected in the "Token Equivalence Method," even if it does approximate a correct assessment of the four translations limited to its specific criteria."
I would add that one of the difficulties of statistical analysis of literary works is that it often (more or less accurately) measures features that constitute (in the case at hand) only some of the features that make a good translation. To put it an another way, IQ tests tend to measure whatever they measure but it is far from clear that they provide a global quantification of that mysterious thing called Intelligence. Dr. Tarvi's analysis is presumably an accurate measure of some features of the translations but not all--as I am sure she would be the first to admit.
--------------------------------
FROM: DMITRI NABOKOV
"Both the Ilin & Skyarenko versions are pretty good. ... With all due respect, the mathematical analysis proposed by Lyuba Tarvi is not my idea of comparison. This is borne out by the fact that Kirichenko's choices of individual words may coincide with mine, while her sentences and paragraphs are sometimes an incomprehensible jumble, which is not reflected in the "Token Equivalence Method," even if it does approximate a correct assessment of the four translations limited to its specific criteria."