Abstract:
This dissertation develops a theory of lexical accent where the central role is given to the notion
of accent competition as the defining property of lexical accent systems. Languages with
complex morphology (traditionally known as ‘polysynthetic’) are the empirical basis for this
study as they provide a particularly fruitful ground for investigating the effects of both
phonological and morphological factors in the assignment of lexical accent. Novel in-depth
analyses are developed for Arapaho (Plains Algonquian), Nez Perce (Sahaptian), Ichishkiin
Sɨnwit (Sahaptian), and Choguita Rarámuri (Uto-Aztecan). I argue that accent competition
across languages is resolved in formally similar ways and that no idiosyncratic, languagespecific
analyses for individual lexical accent systems are warranted. It is thus proposed that
the idiosyncrasy in lexical accent systems is found in the distribution of underlying accents,
but not in the rules of the systems. I propose a typology of lexical accent systems and argue
that they fall into one of two types based on the mechanism of accent competition resolution:
I. Cyclic: In an accent competition, accent in the outermost derivational layer within the domain
wins, or II. Directional: In an accent competition, either the right-most or the left-most accent
within the domain wins.
The second group of proposals made in this dissertation concerns the status of
fundamental properties of stress and prosody – Culminativity and Obligatoriness of stress
(Trubezkoy 1939/1960; Hyman 2006, 2009), and primary versus non-primary stress. It has
previously been claimed that Culminativity of stress can be breached in highly synthetic
languages (e.g. Blackfoot, Stacy 2004; Arapaho, Bogomolets 2014a,b; Mapudungun,
Molineaux 2018; Yupik, Woodbury 1987).
I argue that stress is in fact always culminative, but Culminativity should be regarded as a
macroparameter allowing for a set of language-specific ways to implement it, including a
mechanism of clash avoidance and an enforcement of ‘one and only one’ stress within domains
smaller than a morphological word. Finally, this dissertation addresses the general structure of
the word-level prosodic system. I propose that the word-level prosodic system is not bipartite:
primary stress vs. rhythm, but tripartite: primary stress vs. secondary stress vs. rhythm.