The present work reports on an empirical study aimed at investigating translation problems across multiple language pairs. In particular, the analysis is aimed at developing a methodological approach to study concordance search logs taken as manifestations of translation problems and, in a wider perspective, information needs. As search logs are a relatively unexplored data type within translation process research, a controlled environment was needed in order to carry out this exploratory analysis without incurring in additional problems caused by an excessive amount of variables. The logs were collected at the European Commission and contain a large volume of searches from English into 20 EU languages that staff translators working for the EU translation services submitted to an internally available multilingual concordancer. The study attempts to (i) identify differences in the searches (i.e. problems) based on the language pairs; and (ii) group problems into types. Furthermore, the interactions between concordance users and the tool itself have been examined to provide a translation-oriented perspective on the domain of Human-Computer Interaction. The study draws on the literature on translation problems, Information Retrieval and Web search log analysis, moving from the assumption that in the perspective of concordance searching, translation problems are best interpreted as information needs for which the concordancer is chosen as a form of external support. The structure of a concordance search is examined in all its parts and is eventually broken down into two main components: the 'Search Strategy' component and the 'Problem Unit' component. The former was further analyzed using a mainly quantitative approach, whereas the latter was addressed from a more qualitative perspective. The analysis of the Problem Unit takes into account the length of the search strings as well as their content and linguistic form, each addressed with a different methodological approach. Based on the understanding of concordance searches as manifestations of translation problems, a user- centered classification of translation-oriented information needs is developed to account for as many "problem" scenarios as possible. According to the initial expectations, different languages should experience different problems. This assumption could not be verified: the 20 different language pairs considered in this study behaved consistently on many levels and, due to the specific research environment, no definite conclusions could be reached as regards the role of the language family criterion for problem identification. The analysis of the 'Problem Unit' component has highlighted automatized support for translating Named Entities as a possible area for further research in translation technology and the development of computer-based translation support tools. Finally, the study indicates (concordance) search logs as an additional data type to be used in experiments on the translation process and for triangulation purposes, while drawing attention on the concordancer as a type of translation aid to be further fine-tuned for the needs of professional translators. ***

Concordancing Software in Practice: An investigation of searches and translation problems across EU official languages

-
2013

Abstract

The present work reports on an empirical study aimed at investigating translation problems across multiple language pairs. In particular, the analysis is aimed at developing a methodological approach to study concordance search logs taken as manifestations of translation problems and, in a wider perspective, information needs. As search logs are a relatively unexplored data type within translation process research, a controlled environment was needed in order to carry out this exploratory analysis without incurring in additional problems caused by an excessive amount of variables. The logs were collected at the European Commission and contain a large volume of searches from English into 20 EU languages that staff translators working for the EU translation services submitted to an internally available multilingual concordancer. The study attempts to (i) identify differences in the searches (i.e. problems) based on the language pairs; and (ii) group problems into types. Furthermore, the interactions between concordance users and the tool itself have been examined to provide a translation-oriented perspective on the domain of Human-Computer Interaction. The study draws on the literature on translation problems, Information Retrieval and Web search log analysis, moving from the assumption that in the perspective of concordance searching, translation problems are best interpreted as information needs for which the concordancer is chosen as a form of external support. The structure of a concordance search is examined in all its parts and is eventually broken down into two main components: the 'Search Strategy' component and the 'Problem Unit' component. The former was further analyzed using a mainly quantitative approach, whereas the latter was addressed from a more qualitative perspective. The analysis of the Problem Unit takes into account the length of the search strings as well as their content and linguistic form, each addressed with a different methodological approach. Based on the understanding of concordance searches as manifestations of translation problems, a user- centered classification of translation-oriented information needs is developed to account for as many "problem" scenarios as possible. According to the initial expectations, different languages should experience different problems. This assumption could not be verified: the 20 different language pairs considered in this study behaved consistently on many levels and, due to the specific research environment, no definite conclusions could be reached as regards the role of the language family criterion for problem identification. The analysis of the 'Problem Unit' component has highlighted automatized support for translating Named Entities as a possible area for further research in translation technology and the development of computer-based translation support tools. Finally, the study indicates (concordance) search logs as an additional data type to be used in experiments on the translation process and for triangulation purposes, while drawing attention on the concordancer as a type of translation aid to be further fine-tuned for the needs of professional translators. ***
2013
en
Concordancing Tools
European Union
Information Need
SCIENZE DELL'INTERPRETAZIONE E DELLA TRADUZIONE
Search Log
Translation Problems
Translation Process
Translation Studies
Università degli Studi di Trieste
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.14242/269642
Il codice NBN di questa tesi è URN:NBN:IT:UNITS-269642