It is well known that prosody can be used to express meaning in language, evidenced through word-level contrastive tone, accent, and stress, as well as sentence-level intonation. This talk will explore the use of prosody in expressing grammatical categories, which I call “grammatical prosody”. Grammatical prosody can stretch to the limits of morphological exponence, having the potential for very wide domains of realization while also subject to competition with lexical or other grammatical prosodies within smaller domains, and are therefore excellent testing grounds for model comparison (e.g. item-and-arrangement vs. item-and-process, phase theories of linguistic interfaces, prosodic hierarchy debates, a.o.). At the same time, however, grammatical prosody is severely underdocumented, thereby rendering the theoretician without a stable empirical target.
In this talk, I further the efforts to describe and analyze grammatical prosody by examining prosodic patterns in the verb system of the Bolivian language Ese Ejja. In this language, stress in the verbal domain exclusively expones inflectional categories tense/mood/agreement and transitivity, and is subject to complex but highly regular rules. These surface stress patterns can be placed profitably within a morphological paradigm, allowing us the ask the following question: what relationships do the paradigmatically-related cells have, e.g. in enforcing uniformity and/or distinctiveness?
I will make three claims in this talk which fall out of the Ese Ejja data. The first is a typological claim: Ese Ejja grammatical prosody demonstrates a rare type of paradigmatic relationship resulting in prosodic uniformity between inflected verbs with derivational morphology and inflected verbs without derivational morphology. Schematically (where R=root, I=infl, D=deriv), the position of stress in an inflected verb [αR-αI] is maintained in related words [αR-D-αI], regardless of other rules~constraints.
From here, I make a theoretical claim: Ese Ejja requires us to modify the formal model Output-Output Correspondence (OO-Corr), designed to capture such unexpected uniformity effects. I call this modification Transparadigmatic Output-Output Correspondence (Tr-OO-C, Rolle 2018), and formalize it using Agreement-By-Correspondence. In short, Tr-OO-C is unique in that it requires correspondence between forms sharing both the same verb root and the same exact verbal inflection; other more familiar types of OO-Corr require only the same root/stem.
I will conclude this talk by discussing a third, functionalist claim: this paradigmatic relationship emerges as a response to difficulties in the ‘on-line’ computation of stress position. The inflectional controller appears towards the end of the word but grammatical prosody appears at the beginning, a difficulty which is only increased when medial derivational affixes are present. In response, in these contexts the grammar simply refers to a related and ‘simpler’ paradigm, and copies its position of stress.