The articles in this special issue focus on the interactions between cognitive and language processes in conference interpreters. The extreme language demands of their task provide us, as exceptional groups and exceptional performance often do, a useful tool to determine the ways the brain engages cognition to process and produce language.Consider the language tasks that conference interpreters perform virtually simultaneously: listening to speech in a given language, comprehending it, then speaking the message in a different language. Sometimes the input is fast, mumbled, not loud enough, or, if the speaker is reading a text aloud, the natural intonation variability may be reduced. Often the interpreter must start speaking before a sentence or even clause has ended, in order to keep up with the input. No wonder the International Association of Conference Interpreters (AIIC), when negotiating working conditions with some of the major employers of interpreters (e.g., the European Union and the United Nations) stipulated not only that conference interpreters need working-space with good ventilation and reasonable temperature and allowing direct visual contact with the speakers, but also that they work no more than six hours a day, alternating approximately 30 minutes at the microphone with 30-minute breaks during which a colleague takes over (Agreement 2007-2011 as ratified by AIIC and the Organizations, n.d.). Conference interpreters generally listen in a highly proficient second language and speak in their first language, so their comprehension in the non-first language must be excellent, and their production will be native.1 A good number of conference interpreters work in both directions, which at first glance might seem to be not much harder than working only in one. In fact, however, interpreters develop strategies at macro and micro levels to deal with the challenges in interpreting from a given language to another, and these strategies are not necessarily commutable. As a result, conference interpreters' proficiency levels in each language they work in must be quite high, and extensive practice with interpretation between the two languages in the direction or directions they will employ professionally is crucial. Indeed, interpreters report that only after approximately 10 years of post- training professional work do they feel they were at the peak of their performance.2A part of what those years of developing expertise involves is setting up and automating a set of strategies for particular instances of challenging translation moments. On the one hand, there is rarely one-to-one correspondence of lexical items or idioms, or structures, between a given pair of languages; a lexical item or idiomatic usage in a given language may have more than one translation in another one, so the correct translation must be selected based on context. Structural and/or word-order differences between languages can also pose particular difficulty, as Seeber & Kerzel (this issue; see also Seeber, forthcoming) points out in his experiment with conference interpreters working between German with its clause-final verbs in embedded clauses and English with its verbs positioned much earlier in the clause in comparable clauses.Lack of one-to-one structural and lexical correspondence and not-quite-native proficiency are challenges to the cognitive system underpinning the language abilities stressed in simultaneous interpretation. It is these cognitive underpinnings of interpretation that form the foci of the articles in this issue. Most studied has been working memory (WM, e.g., Moser, 1978), covered here by Kopke and Signorelli, Tzou, Eslame, Chen, and Vaid, and by Signorelli, Haarmann, and Obler. Although different scholars use somewhat different operational definitions of working memory, it is generally understood to be that ability to hold material in mind 'verbatim' in order to 'work' with it as one processes sentences for comprehension and production. …