29th Manchester Phonology Meeting, University of Manchester, Manchester, UK
The file is open to annotations via Hypothes.is. Comments are always welcome!
(with Yonatan Goldshtein)
In this paper we demonstrate that the patterning of stød in Standard Danish submits to a complete, non-stipulative account within an approach that embraces metrical structure as the mechanism behind accentual contrasts (Köhnlein 2019), and the main tenets of Stratal Phonology (Bermúdez- Otero 2018). Our approach maintains the empirical coverage of the ‘Non-Stød Model’ developed by Basbøll (2005 et passim), but builds on widely accepted principles of morphology-phonology interaction instead of a proliferation of bespoke domains, in addition to incorporating insights from other approaches (notably Itô & Mester 2015).
Stød is a laryngealization prosody that, in Standard Danish, is restricted to stressed syllables with heavy, high-sonority rhymes (with a long vowel or a sonorant coda). Subject to this pho- notactic constraint, we formulate the following basic generalization: by default, stød is assigned to a stressed syllable at the word level unless some disyllabic domain was constructed at the stem level. Stem-level disyllabic domains arise when either the root is underlyingly disyllabic, or a suffix is attached at the stem level. We demonstrate that this simple generalization interacts with regular phonological phenomena (e. g. lexical word-final extrametricality and epenthesis in rising-sonority clusters) and morphologically specific patterns to derive precisely the attested paradigms of stød assignment.
The analysis relies on widely accepted principles of morphology-phonology interaction, and is consistent with relevant morphological generalizations: for instance, the unpro- ductive plural suffix -e always shows root attachment, triggering stem-level phonology (Giegerich 1999, Bermúdez-Otero 2018), whilst for the productive -er suffix this pattern is exceptional. In the paper, we show that the basic analysis extends to disyllabic and longer nouns, nominal compounds, and simple and prefixed verbs, as well as to exceptional patterns.
Overall, we show that a metrical approach to accentual distinctions (Morén-Duolljá 2013, Köhnlein 2016, 2019) extends readily beyond ‘tonal’ accents to other distinctions such as stød (Iosad 2016, Morrison 2019). Furthermore, integrating the metrical, domain-based approach with the principles of Stratal Phonology yields a consistent and empirically adequate account of this phenomenon that builds on a small number of first principles. This, we suggest, reinforces the nature of both Stratal Phonology and the metrical view of accentuation as productive research programmes.