To what degree are talker processing costs driven by the fact that items serving as targets on some trials can serve as distractors on others? Similarly, to what degree are multi-talker processing costs driven by having to respond to multiple target stimuli on mixed trials but not on blocked trials?