Robin Setton proposes a procedure for discovering processes in simultaneous interpretation through detailed corpus analysis, followed by a brief discussion of the fundamental insights it offers, and the prospects for fuller formalisation and more ambitious applications such as the evaluation of translations. The methodology proposed applies a composite of tools of linguistic analysis at different levels with the ultimate aim of generating viable representations of the meanings available from the input and output speech streams. Transcription conventions are described and special attention is devoted to context modelling. Conventionally, contexts are recognized as vital quality determining factors, and assumed to be individual and unconstrained. The author points out that if translation research is to be connected with cognitive science, context – as a knowledge base - needs to be constrained to input into a model of interpreting.