In 7 experiments the authors investigated the locus of word frequency effects in speech production. Experiment 1 demonstrated a frequency effect in picture naming that was robust over repetitions. Experiments 2, 3, and 7 excluded contributions from object identification and initiation of articulation. Experiments 4 and 5 investigated whether the effect arises in accessing the syntactic word (lemma) by using a grammatical gender decision task. Although a frequency effect was found, it dissipated under repeated access to a word's gender. Experiment 6 tested whether the robust frequency effect arises in accessing the phonological form (lexeme) by having Ss translate words that produced homophones. Low-frequent homophones behaved like high-frequent controls, inheriting the accessing speed of their high-frequent homophone twins. Because homophones share the lexeme, not the lemma, this suggests a lexeme-level origin of the robust effect.The word frequency effect in speech production was discovered by Oldfield and Wingficld (1965). In a picture-naming task, they found that pictures with low-frequency (LF) names (such as syringe) took longer to name than pictures with high-frequency (HF) names (such as basket). Wingfield (1968) established this effect as a genuinely lexical one. The effect was not due to differential speeds of object recognition but to naming itself.In this research, we consider the word frequency effect in light of recent models of lexical access in speech production. In particular, we consider whether the frequency effect arises early, in selecting the semantically appropriate lexical item, or late, in the retrieval of the item's form information.The article is organized as follows. We begin by sketching a model of the production lexicon. Against the background of this model, we then derive hypotheses about the potential loci of frequency effects from data in the literature on picture naming, hesitations in spontaneous speech, and speech errors. Finally, we turn to the core of this article, an experimental investigation of these hypotheses.In seven experiments, we analyze whether the frequency effect is due to either conceptualization or articulation (it is neither), whether it arises at the early, so-called lemma level (there is a frequency effect there, but it turns out not to be the frequency effect), or whether it originates at the late word form, or lexeme level (it does). Jorg D. Jescheniak and Willem J. M. Levelt, Max-Planck-Institut fiir Psycholinguistik, Nijmegen, The Netherlands.We wish to thank Ger Desserjer, Hans Franssen, and Johan Weustink for invaluable technical assistance and Monika Baumann, Jean Fox Tree, Antje Meyer, Paul Meyer, Thomas Pechmann, Ardi Roelofs, Herbert Schriefers, Linda Wheeldon and four reviewers for helpful comments and discussions.Correspondence concerning this article should be addressed to Jorg D. Jescheniak, who is now at Institut fiir Psychologic FB12, Freie Universitat Berlin, Habelschwerdter Allee 45, D-14195 Berlin, Germany. Electronic mail may be sent to jo...