Articulatory factors influencing regressive place assimilation across word boundaries in German

This EMA study examined articulatory properties of nasal–stop and stop–stop sequences across word boundaries in German. The immediate aim was to investigate whether the greater tendency of nasals as compared to stops to exhibit regressive place assimilation results from differences in tongue-tip reduction due, in turn, to acoustic–perceptual properties of nasals. In parallel with this aim, the relevance of word frequency for assimilatory processes was investigated. Analysis of data from four speakers showed greater tongue-tip reduction in high-frequency words with a nasal, indicating a combination of factors that causes tongue-tip reduction.A further route to capturing assimilatory processes is the amount of overlap between C1 and C2. Analysis of this was complicated precisely by the fact that particularly for the high-frequency nasals, movement reduction was sometimes so strong that no kinematic analysis could be performed. A tentative conclusion was nonetheless that overlap tended to be greater in nasal–stop than stop–stop sequences.The discussion points out that high-frequency words in German mostly end in a nasal and concludes that it is word frequency and speakers' knowledge of acoustic–perceptual properties of nasality that allows them to simplify articulation more in nasals than stops. This account presupposes that the nasality itself remains robustly present even when the lingual gesture of the nasal consonant is strongly reduced (or overlapped by the following consonant). Velum movement data available for some of the speakers and inspection of the acoustics indicated that this was indeed the case.

► Electromagnetic articulography was used to study coronal assimilation.
► Movements of tongue, lip and velum were analyzed.
► More tongue tip reduction and overlap was found for nasals than stops.
► Word-frequency or perhaps co-occurrence frequency also affected reduction processes.
► Acoustic properties of nasals (vs. stops) are assumed to make reductions less salient.