Conceptual definitions and operational definitions

Daniel Gile

Both empirical research (ERP) and LAP research of all types use terms to denote concepts. In many cases, researchers do not feel the need to define them. I do not remember having encountered in TS discussions on the definition of the concept ‘language’ in spite of the very central position of language in Translation and the existence of extensive discussions of its nature in other disciplines.

Some definitions are only naming conventions. Such is the case of ‘source language’ and ‘target language’. Other definitions are more complex and are processed differently depending on the research approach of the relevant scholar.

Before taking up these differences, I should like to add to the Empirical Research Pole (ERP) and the Liberal Arts Pole (LAP) already discussed in this Research Issues section another approach to investigation into Translation, namely the Professional Pole (PP), which is basically the Translation practitioner’s approach – and to no small extent the Translator trainer’s approach. (Readers are asked to keep in mind that these classifications are tools for the exploration of the TS scene, and that as such, they are necessarily tentative).

Take the concept of Translation quality. Its nature is by no means evident, as it can be seen and has been analyzed and described under many angles, relative, ‘absolute’, Text-dependent, client-dependent, receiver-dependent, etc., and it seems difficult to find a single consensual definition which would satisfy demanding thinkers. In contrast, in empirical research, individual studies tend to have a narrow focus and researchers may be happy with an operational and even implicit definition which will allow them to carry out their investigation, especially if quality is the dependent variable used to assess the effect of an independent variable (such as working conditions, a training method etc.) and is seen through an indicator (assessment by a panel of evaluators, the number of errors of a certain type etc.). Empirical researchers are well aware that the number of errors or proportions of source Text propositions rendered correctly in the target Text are only an indicator of part of something they think of as ‘Translation quality’, but the absence of a strict definition of ‘quality’ does not bother them. Neither will it bother other ERP critics who may point out that the indicators chosen for one particular study may have left out another meaningful facet of ‘quality’ and will seek to complement it with additional studies which will take on board other facets – without attempting to find an accurate and comprehensive definition of the nature of ‘quality’ which might satisfy philosophers interested in the same issue. Professionals with no particular academic interest in TS (the Professional Pole) may be interested in quality as reflected in their clients’ reactions, or as measured intersubjectively by assessors in professional accreditation tests, while Translation trainers experimenting with training methods may focus on quality as an indicator of progress in the acquisition of skills, again without worrying about its deep nature.

ESIT’s ‘deverbalization’ concept is another interesting example. Philosophers may discuss at length the possibility of having language-independent ‘sense’, but empirical researchers interested in the concept tend to seek an operational definition which will allow them to test its existence and/or circumstances. As to Translation trainers, they have no problem with its use in the classroom without a strict conceptual definition: field observation shows that it is generally quite clear to them when a student has translated with or without ‘deverbalization’.

Both professionals and empirical researchers may even be puzzled about the usefulness of a comprehensive definition of ‘Translation’, or about statements such as “TS cannot proceed without a thorough understanding of what Translation is”. They need operational definitions, enough to do their work. In this sense, one could argue that LAP is conceptually more rigorous than ERP or PP.