Abstract
This article focuses on the nature of non-linguistic mental representations constructed by
simultaneous interpreters drawing on the conceptual complex (CC) model originally introduced
by Funayama (e.g. 2005, 2007). While processing source text, two types of mental
representations are considered relevant: linguistic and conceptual representations. Conceptual
representations remain longer in the interpreter’s mind than linguistic ones, facilitating the
processing of source text and the production of target text. Although conceptual representations
develop incrementally, they keep their identity during discourse processing. Consequently, some
of their features remain unaltered, while others change. The conceptual skeleton is identified as
the solid part of non-linguistic representations. Analysing linguistic differences found between
source and target text as recorded during simultaneous interpreting, this study elucidates the
nature of the skeleton, describes the consistency and liquidity of conceptual representations as a
whole, and explores how they develop based on the skeleton and their role in the interpreting
performance.
