Morphological Strata of Tone

点击翻译按钮获取中文摘要

The goal of this project is to develop a new take on tonal morphophonology combining stratal morphophonology in a Harmonic Grammar version with the recently developed theory of Gradient Symbolic Representations. Under this approach, phonological representations are phonologically (but not necessarily phonetically) gradient, i.e., have continuous abstract activation. Due to the stratal organization of grammar, these activations change iteratively in tandem with the cyclic construction of increasingly bigger morphological units. In contrast to most previous work in stratal phonology, this allows for eliminating different grammars (i.e., different constraint rankings/weightings) in different strata. This new overall approach makes highly specific predictions on possible morphophonological systems which we will systematically test in the project: (a) Monotonicity across Strata: Changes of activation should be unidirectional (monotonically increasing or monotonically decreasing) since these changes are triggered by the same constraint ranking in all strata (b) Consistency of Strength in a Given Stratum: Specific phonological units should behave consistently ’strong’ or ‘weak’ with respect to all phonological processes at a stratum (pace approaches in construction phonology which do not make this prediction), (c) Pervasiveness of Cooperation: Cooperation (the fact that idiosyncratic phonological effects are triggered cooperatively by multiple morphemes) falls out naturally in Gradient Symbolic Representation Theory by the fusion of several phonological units with different activation levels. This is in stark contrast with most previous theories to lexical exceptions, where morphophonological idiosyncrasies have a single morphological trigger. Concretely, we will investigate these general predictions in four empirical areas where tone poses unsolved problems for morphological and phonological theories: (1) global rules and strata straddling (phonological processes which require simultaneous access to information specific to single strata), (2) inter-stratal conspiracies (the same phonological well-formedness constraint is satisfied by different repair strategies in different strata ), (3) competition of overwriting patterns (different morphologically enforced tone patterns compete for realization on a given base word), and (4) tonal attraction phenomena, where underlying base tones are attracted to different linear positions. Empirical data for the project will be taken exclusively from African languages, which exhibit especially rich interactions between morphosyntax and tonal phonology, especially from Bantu, Nilo-Saharan, and West-African languages of different genetic affiliations. The majority of these data will be drawn from written, published sources, but there will be supplementary fieldwork on Nilo-Saharan in cooperation with Maria Kouneli’s co-project on Eastern Sudanic in the Research Unit.