Cognitively Difficult Phrases in Post-editing from English to Finnish: A Comparison of Two Analysis Methods
Nylund, Tommy (2023-11-13)
Cognitively Difficult Phrases in Post-editing from English to Finnish: A Comparison of Two Analysis Methods
(13.11.2023)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
avoin
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe20231211153103
https://urn.fi/URN:NBN:fi-fe20231211153103
Tiivistelmä
Post-editing machine translations have become more common in the translation field in recent years. Studies in this field have examined the types errors post-editors have to correct, what sort of errors machine translations produce, and how useful machine translations are, but there is far less research on post-editing time and cognitive strain it produces.
In this paper, an analysis of cognitive strain in post-editing is presented where the method Choice network analysis is used together with the data processing tool Translog-II to explore, how cognitive difficulty manifests with five professional translators who post-edit the same English to Finnish text.
The different post-edits made by each participant of the same word or phrase in various sentences are examined with the Choice network analysis and the Translog-II tool measuring the time it took to post-edit these words or phrases, and then comparing these results with each other.
The results were mixed, where in 6 cases the two analysis methods showed similar cognitive difficulty and in 7 cases, they showed differing results of each other. These findings suggest that these two analysis methods might not be best suited to work by themselves but could yield better results when used together.
With Choice network analysis it is possible to analyse multiple possible post-edits with possible cognitive difficulty and with Translog-II it is possible to explore the time spent on the task and see the various alterations the post-editors make.
In this paper, an analysis of cognitive strain in post-editing is presented where the method Choice network analysis is used together with the data processing tool Translog-II to explore, how cognitive difficulty manifests with five professional translators who post-edit the same English to Finnish text.
The different post-edits made by each participant of the same word or phrase in various sentences are examined with the Choice network analysis and the Translog-II tool measuring the time it took to post-edit these words or phrases, and then comparing these results with each other.
The results were mixed, where in 6 cases the two analysis methods showed similar cognitive difficulty and in 7 cases, they showed differing results of each other. These findings suggest that these two analysis methods might not be best suited to work by themselves but could yield better results when used together.
With Choice network analysis it is possible to analyse multiple possible post-edits with possible cognitive difficulty and with Translog-II it is possible to explore the time spent on the task and see the various alterations the post-editors make.