Abstract
Lexical selection (i.e., selection of relevant words from the lexicon) is one of the essential linguistic processes associated with speaking. A prominent theory of language production developed by Levelt, Roelofs, and Meyer (1999) posits that lexical selection is a competitive process: Every produced word is a result of comparison of activation levels of multiple alternatives, which “compete” to be selected. In this dissertation, new evidence supporting this view was provided. In the first study, I demonstrated that the distractor effects in the picture-word interference task (PWI) and in the Stroop task are both undermined by competitive lexical selection. This is in line with the competitive account of Levelt et al. and in contrast with some recently advanced findings that linked the distractor effect in PWI to perceptual and conceptual processing. In the second study, I explored whether and how speakers can modify their top-down control in PWI. It was observed that the distractor effect size is dependent on whether the speaker is exerting more or less control. Importantly, under both high and low control the distractor effect was linked to the lexical selection time-window of the processing stream. A series of experiments performed in the third study showed that such control adjustments in PWI were driven by relaxation of control on post-congruent trials, rather than enhancement of control on post-incongruent trials, as a currently dominant conflict-monitoring account suggests. The fourth study demonstrated that speakers can also adjust allocation of processing capacity based on available information about the complexity of an upcoming trial.
