Translation and understanding: mental models as an interface in the process of translation
This thesis discusses two characteristics of language which affect translation, using English and Japanese examples. However, the general points made in the thesis are not specific to these two languages. One characteristic of language is that it encodes particular perceptions of experience by its users. Word meaning is defined in this thesis in terms of the typical experience the language user associates with a word. Concepts for which there are no single lexical items are encoded by putting together words which the speaker thinks best characterise the concept. This particular characterisation of a concept may become established in the language community. If the members of a language community form a habit of characterising a concept in a particular way, it may become difficult to perceive the concept in any other way. In translation, this may lead the translator to impose characterizations established in her own language on the other language. However, such difficulties can be overcome because of the creative capacity of people everywhere to learn new ways to perceive the world. And language provides the mechanism to encode such novel perception. This is the other characteristic of language discussed in this thesis. We can use an existing word to encode a new kind of experience which we perceive as having some connection with the kind of experience associated with the word. Such novel application of a word can be understood because upon hearing the word, the typical experience associated with the word is evoked in the hearer's mind, and using her knowledge, the hearer constructs a mental model which she thinks best accounts for the combination of experiences evoked in her mind by the linguistic forms. Defining word meaning and sentence meaning in terms of mental images allows us to understand the process of translation: Upon hearing/reading the source language text, the translator constructs a mental model based on the text. She then bases her translation on this mental model, which becomes a rich source of information. Because the translator is not moving directly from one language to the other, no direct correspondences between the linguistic forms of the two languages need to be sought. This also explains why it is relatively easy to translate between two languages whose users share similar experiences and therefore can build similar mental models, even if the languages are typologically very different from each other.