Door Gertjan Postma
Nijen Twilhaar & Van Oostendorp (2000) wonder why the following assimilation contrast exists in Helledoorn Dutch. While the personal ending morpheme /-n/ shows place assimilation to the root consonant in the present tense, it lacks assimilation in the past tense. For instance [pɑkn̩] ‘took’ contrasts with present tense cases such as [pɑkŋ̩] ‘take’. This distinction in assimilation functions as a distinction in tense in this case. How can assimilation be a marker of tense? T&O assume deletion of an erstwhile tense marker –de which is now without segmental expression. So they analyze /pak+n/ for the present tense and /pak+ø+n/ for the past tense. The k-n sequence gives rise to an assimilated [pɑkŋ̩], while in the latter structure, the past morpheme blocks place spreading, giving rise to [pɑkn̩], both with a syllabic nasal. T&O assume that this past tense morpheme is not completely deleted because of recoverability (captured as their constraint Trace): a coronal feature is retained to which the suffixal /n/ assimilates. Roos (2009) criticizes this hypothetical past morpheme and take –en as the past morpheme in this dialect and all the similar Low Saxon dialects. It cumulates in his map of the weak past tense, where some dialects (nr 21 and 22) have –en as their weak preterit morpheme, some have -Ø, some the more usual –de.
Roos even doubts on T&O’s empirical data: “there is little evidence corroborating the claim (…) that the past tense nasal does not assimilate to the stem-final consonant of weak verbs”. Roos calls upon the data from RND and GTRP. However, Roos should have looked more seriously. Consider the following data from Kuinre, location F052 of the GTRP in Salland. In this case, the assimilation facts are explicitly IPA-encoded in the GTRP (columns 2 and 4). The analysis in columns 3 and 5 is mine.
These data are relevant for T&O’s assimilation blocking, even more so because the past -ø morpheme and the –de morpheme alternate over the paradigm in function of number in this dialect. The conclusion is therefore justified that T&O’s assimilation facts are trustworthy, and that their hypothetical zero past morpheme is a plausible cause of the difference in assimilation. Whether blocking is phonological (by their [cor]), or morphological (past tense de/Ø functions as “first head” in the sense of Creemers, Don & Fenger (2015) creating a phase barrier) is a further point of study.
Creemers, Don & Fengers (2015). Stress and categorial flexibility as a consequence of morphological structure. Proceedings of NELS 45 • Nijen Twilhaar, J. & M. van Oostendorp (2000). Feature domains in Hellendoorn Dutch. Ms. Meertens Institute Amsterdam. • Roos, Niels (2009). The weak past tense in Dutch and Low German. PhD Dissertation Nijmegen University.