Three experiments investigated the processing of pitch contour during spoken word recognition. In Experi- ment 1 event-related brain potentials (ERPs) were recorded while subjects made decisions to artificially pitch manipulated bisyllabic words. ERPs revealed that pitch contours are discriminated already within the first syllable of a word. In Experiment 2 subjects heard spoken word-initial syllables with re- synthesized pitch contours and were asked to name a word starting with that syllable. They answered more likely with an initially stressed word if the syllable carried a stressed pitch contour. However, if the same syllable carried an unstressed pitch contour, subjects more frequently responded with an initially unstressed word. In a cross-modal priming paradigm we pre- sented pitch modulated initial syllables as auditory primes followed by visual targets (Experiment 3). Reaction times as well as ERPs were found to be sen- sitive to pitch information. Taken together the results indicate that pitch is automatically extracted during spoken word recognition and that this prosodic pa- rameter guides lexical access.