We report the results of three experiments designed to test priming percolation (alignment boost effects) from one grammatical level to another. In the first two experiments, we set off to replicate in Dutch the results of Branigan, Pickering & Cleland (2000) for lexical boosts of syntactic alignment, adding a baseline control condition without priming. In the third experiment, we tested direct syntactic boosts of phonological alignment, using invented verbs. The direct link between syntax and phonology (without any interference from the lexicon) has been postulated in the past, but so far no empirical evidence has been offered in its favor. Our experimental results so far largely confirm the predictions of the Alignment Model (Pickering & Garrod, 2004), including the relation between syntax and phonology. Speakers, who were instructed to use the same syntactic structure as their dialogue partner did, also invented a verb that resembled more their partners invented verb.