We show that retrieval from semantic memory is vulnerable even to the mere presence of speech. Irrelevant speech impairs semantic fluency-namely lexical retrieval cued by a semantic category name-more than meaningless speech (reversed speech or nonwords). Moreover, speech related semantically to the retrieval category is more disruptive than unrelated speech. That phonemic fluency-in which participants are cued with the first letter of words they are to report-was not disrupted by the mere presence of speech, only by speech in a related phonemic category, suggests that distraction is not mediated by executive processing load. The pattern of sensitivity to different properties of sound as a function of the type of retrieval cue is in line with an interference-by-process approach to auditory distraction.Keywords: Auditory Distraction, Lexical Retrieval, Semantic Fluency, Phonemic Fluency Retrieval and Distraction 3 Lexical retrieval is a fundamental capacity of language that underpins oral and written language production, involving the access and selection of context-appropriate lexical items to express an intended meaning. One of the most frequently-used tests of lexical retrieval is semantic fluency, in which a semantic category cue is given (e.g., 'Animals') and the task is to retrieve from long-term semantic memory as many examples as possible from that category (Bousfield & Sedgewick, 1944;Newcombe, 1969). We ask whether this process of search by semantic criterion is vulnerable to the presence of concurrent to-be-ignored sound, given that there is a body of work suggesting that retrieval processes generally are largely inviolable to all but the most attentionally-demanding concurrent tasks (Craik, Govoni, Naveh-Benjamin, & N. D. Anderson, 1996). We also attempt to discover how the lexical and semantic status of the sound's contents determine the degree to which semantic fluency is impaired, with a view to revealing the level of abstraction at which distraction occurs. Phonemic fluency (e.g., Benton, 1968), in which the criterion for production is phonemic, not semantic, is also studied in order to provide baseline conditions that share some processes-including key executive processes-with semantic fluency, but not those related to semantic retrieval. This allows examination of whether semantic auditory distraction has a process-specific effect on semantic retrieval, namely whether it only occurs when the focal task also requires semantic (but not phonemic) retrieval, or whether common (executive) processes are vulnerable.Fluency tasks have not been used to study auditory distraction and the work reported here is the first of its kind. Their attraction as a research tool for students of distraction is that they are relatively process-pure in terms of retrieval. Evidence collected so far relating to disruption by sound of lexical-semantic processing has used, almost exclusively, list-based tasks, ones that are not to the same degree process-pure as the fluency task. Because they are list-based, perform...