This research examines the representation and access of lexical form during spoken language comprehension. Previous research using a phonological priming methodology has shown contrasting results depending on whether overlap is initial or final. When overlap occurs in onset position, inhibitory effects are found that are mediated by relatedness proportion, ISI, and task demands. Explanations for such effects point to the involvement of lexical competition. With final overlap, results are often facilitatory in nature although this facilitation seems to require rime overlap. These results have been attributed to the operation of automatic perceptual processes. A series of segmental phonetic priming experiments were conducted in Dutch to address the automatic and perceptual nature of the processes involved. A single phonetic segment (rather than a word or nonword) was used to prime a target containing that segment. The phonetic priming paradigm offers several advantages. As with previous methodologies, it allows variation in overlap (e.g., initial or final) between prime and target. Importantly, however, using a single phonetic segment strips the prime of lexical status. Onset and offset priming can be directly compared. Presentation of a single phonetic segment also allows for a reduction of the temporal interval between prime and target, enhancing sensitivity to early processes in auditory word recognition.

In the onset phonetic priming experiments, fricative segments [f,s,x] were used as primes for target items (e.g., faam 'fame', soep 'soup', and gek 'weird'). Matching pairs ([f]-faam, [s]-soep, [x]-gek) were contrasted to non-matching items ([s]-faam, [x]-faam, [f]-soep, [x]-soep, [f]-gek, [s]-gek). Both primes and targets were presented auditorily in a shadowing task. The offset phonetic priming experiments were identical to the onset experiments except that overlap occurred at the end of target items (e.g., [f]-dief, [s]-dief, [x]-dief 'thief'). Compared to the phonological priming results, the present preliminary data (60 subjects) show a different pattern of results. For onset overlap, the data show a small but consistent facilitation, with matching items facilitated compared with non-matching controls. Since primes are phonetic segments, lexical level inhibition does not contribute. In contrast to the onset results, the offset fricative experiments show an inhibitory effect, with naming latencies to matching items slowed compared to non-matching controls. Two additional experiments investigated CV-onset and VC-rime contrasts to evaluate the supplementary role of the nucleus in mediating these facilitatory and inhibitory effects (e.g., onset: [bu]-boek 'book'; offset: [it]-riet 'reed'). Results will be discussed in terms of recent models of auditory word recognition.