For several decades, context effects in picture naming and word reading have been extensively investigated. However, researchers have found no agreement on the explanation of the effects. Whereas it has long been assumed that several types of effect reflect competition in word selection (e.g., Levelt, Roelofs, & Meyer, 1999), recently it has been argued that these effects reflect the exclusion of articulatory responses from an output buffer (Finkbeiner & Caramazza, 2006). Here, we first critically evaluate the findings on context effects in picture naming that have been taken as evidence against the competition account, and we argue that the findings are, in fact, compatible with the competition account. Moreover, some of the findings appear to challenge rather than support the response exclusion account. Next, we compare the response exclusion and competition accounts with respect to their ability to explain data on word reading. It appears that response exclusion does not account well for context effects on word reading times, whereas computer simulations reveal that a competition model like WEAVER++ (Levelt et al., 1999) accounts for the findings.