In this paper I present a METONYMIC THEORY OF TRANSLATION. I argue that this cognitive approach
allows us to develop a better understanding of what is actually involved in the daily work of the translator. Translation is concerned with the processing of metonymic equivalents between source
and target language more than with the literal one-
to-one substitution of linguistic items. Categories
rarely correspond exactly between languages and the relationship between source and target text is
rarely metaphorical; instead, it is the middle ground of near approximations and partial overlaps which best describes translation. Work from cognitive linguistics on metonymy helps provide us with these insights. I discuss two areas in translation studies: ‘shift theory’ (Catford 1965, Vinay & Darbelnet 1995) and ‘metaphor translation’ theory (Baker 2011, Broeck 1981, Dagut 1976, Newmark 1982). Work in these areas makes a useful contribution but does not go far enough, I feel. I reinterpret the notion of ‘shift’ in terms of metonymic relations and re-characterize non-literalness as a universal enabler, rather than an occasional obstacle met when translating idioms. I argue that metonymic differences are the basis of both ‘interlingual translation’ (original to first draft) and ‘intralingual translation’ (first draft to final version, through revision). The METONYMIC THEORY OF TRANSLATION presented here is grounded in a more general Metonymic Theory of Language and Communication, the cognitive ability to recognize part-whole relations between signs and parts of signs playing a vital role across the entire spectrum of linguistic phenomena.