“…As well, formal models of the Stroop effect typically rely on inhibitory connections (e.g., Botvinick, Braver, Barch, Carter, & Cohen, 2001;Zhang, Zhang, & Kornblum, 1999). As a slight variant, some accounts suggest that a top-down inhibitory mechanism dampens activation of concepts that are closely related to the target (e.g., Dagenbach & Carr, 1994). Although atypical, not only is a solution based on facilitation more efficient computationally (only related concepts need to be connected, whereas in an inhibition-based model, every single semantic concept needs to be connected to every other semantic concept, which would require approximately 200 million connections in a vocabulary of 20,000 words), but it also seems more reasonable (and biologically plausible) to assume that related concepts should be connected in a facilitative manner and that unrelated concepts should not be connected at all.…”