Conference interpreters routinely perform bilingual processing under extreme temporal constraints and extreme cognitive load. Investigating how they process lexis might inform our understanding of lexical processing by a bilingual mind in general. In this talk, I will present bilingual control mechanisms, such as facilitation and inhibition, and will show some data from both experimental and corpus-based studies that use accuracy, ear-voice span, fluency and eye-tracking data to shed more light on how interpreters process cognates, homonyms, high-frequency and low-frequency words, as well as words with one and many translation equivalents when working in a booth.