Experimental priming of phonological variant identification

In conversational speech, it has been observed that speakers tend to reuse the variant that they have recently used. This recency effect, also called 'persistence' or 'repetitiveness', has often been attributed to 'priming' in the psycholinguistic terms. However, it is still not clear whether one can really be primed to choose one variant over another with discrete phonological variation in an experimental set-up. This study empirically tests the hypothesis that having most recently perceived one variant increases the probability of reusing the same variant in speech perception on a subsequent trial, controlling for overall variant rates. Using the well-studied variable ING, i.e., the alternation between -ing and -in' as a test case, our results reveal that phonological variant identification can be primed, and that this variant priming decays rapidly, only after one intervening word. The difference between variant persistence and variant priming regarding their decay profiles ultimately calls into question whether repetitiveness in conversational speech is really driven by priming.