This projects explores a novel cyclic optimization theory that explains recurring problems in the morpho-phonology of tone.

The goal of this  project is to develop a new take on tonal morphophonology combining stratal phonology (Kiparsky 2015, Bermudez-Otero 2018, Trommer 2011) with  the recently developed theory of Gradient Symbolic Representations (Smolensky & Goldrick 2016, Zimmermann 2019). In this novel „Harmonic Layer Theory“, tones can get incrementally stronger or weaker by every cycle of optimization. The same tone can hence react differently to identical tonotactic problems in larger morphological domains since it has a different activation at this level and constraints are hence violated differently. Our hypothesis is that Harmonic Layer Theory allows for principled solutions to four persistent empirical problems tone poses for morphological and phonological theories: 1) global rules and strata straddling (Hyman 1993), 2) inter-stratal conspiracies (Myers 1991, 1997), 3) competition of overwriting patterns (Hyman 2013), and 4) tonal attraction phenomena (Trommer 2019). 
Data for the project will be drawn from African languages which exhibit complex word-internal morphologies, and intricate systems of tonal phonology which are both crucial to the project. Our focus is on three groups of African languages — Bantu, West-African and Nilo-Saharan —  which show a particular richness in the relevant phenomena. 


Red onions cut in half
Tone strata (Symbolic picture); Photo: Colourbox