How do speakers retrieve words from their mental lexicon? There is a long-standing debate on the question of whether only words selected for articulation are phonologically activated or whether this is also true for their semantic competitors. Past behavioral and electrophysiological research has addressed this issue by testing whether retrieval of a target word (e.g., cat) affects – or is affected by – the processing of a word that is phonologically related to a semantic coordinate of the target (e.g., log, related to dog). This research has consistently failed to demonstrate such mediated priming effects. Our experiments tested whether this pattern also extends to lexical retrieval in a speaker's second language. Using an ERP paradigm and experimental materials that had previously been employed to study lexical retrieval in speakers' first language (Jescheniak et al., 2002, JoCN, 2003, Cog Brain Res), we now obtained such effects surfacing as a widely distributed negativity starting around 600 ms after stimulus onset. The overall pattern is compatible with the view that increased language experience leads to a fine-tuning of lexical activation during speech planning.