ABSTRACT
Choice network analysis (CNA) is presented here as a method for constructing models of the mental processing underlying translation. 1 CNA is seen as both an alternative and a complement to other research methods, such as think-aloud protocols and word-based psychological experiments. The method operates on the principle that models of mental processes can be inferred from the analysis of translations of the same source texts by multiple subjects. Examples of networks are presented covering translation between Arabic and English and Spanish and English in such areas as lexis, complex noun phrases, ellipsis and passives. CNA is seen to be theory-free in that the analyst can use the method to test hypotheses from a variety of theoretical standpoints. The examples discussed here draw on insights from theories of working memory, language production , metaphor, syntax and cognitive style. Principles behind the construction of networks are discussed, including the issue of linearity and the possible incorporation of connectionist modelling. Some applications of CNA are proposed, such as researching text difficulty and expert-novice behaviour, and the generation of hypotheses.
