We examined the role of general inhibitory control and general set shifting processes in bilingual language production in 51 native Finnish speakers with English as L2, mainly learnt after the age of 7. We tested the hypothesis that inhibitory control, measured with the Simon and Flanker tasks, is central when switching into L1 (Green, 1998) and, more generally, that general set shifting processes, measured with the Number-Letter task, underlie language switching and mixing (Meuter & Allport, 1999). The results were inconsistent. The basic language switch cost effects were in line with the inhibitory control model, but the interactions with the executive tasks did not support the model and were partly contrary to it. The general set shifting hypothesis received some support. Alternative explanations of the sources of the switching and mixing cost asymmetries in bilingual language production are discussed.