Event
Below is the abstract for Karlos Arregi: Paradigmatic order in Nuer
Nuer (Nilotic, South Sudan & Ethiopia) has a particularly complex
system of nominal inflection that includes (at least) 25 separate
classes of nouns (Frank 1999). In well-studied Indo-European
inflectional systems (e.g. Latin, Russian) word classes are determined
on the basis of inflectional allomorphy: different classes trigger
different allomorphs of the same inflectional morpheme (e.g. in Latin,
nominative singular is -a in class 1, but -us in class 2 masculines).
Although allomorphy does play a role in Nuer, this is not the main
factor that contributes to the large number of noun classes in this
language. In fact, the number of allomorphs available for each
inflectional morpheme is comparatively small, considering the high
number of noun classes (e.g. the nominative singular is always zero,
and the locative plural can only be -ni or zero), and so is the number
of cells in each paradigm (only six: two numbers and three cases). The
main factor distinguishing the classes is patterns of syncretism. For
instance, while in some classes nominative singular is zero and both
genitive and locative singular are realized as syncretic -kä, in
others the syncretic pair is either nominative and genitive (both
zero, with locative as -kä) or nominative and locative (both zero,
with genitive as -kä), yet in others the whole singular paradigm is
syncretic (zero).
Contra Baerman 2012, I argue that, despite this apparent complexity,
any framework that incorporates some of the basic tools of
realizational theories of morphology can account for Nuer nominal
inflection, and I illustrate this claim with an analysis in the
framework of Distributed Morphology (DM). These tools are: (i)
diacritic features (used in defining the different classes), and (ii)
rules of referral (more specifically, rules of impoverishment, which
account for certain kinds of syncretism). The key to understanding the
Nuer paradigm is the simple assumption (common at least in DM) that
diacritic features are features, and as such can both trigger rules
and be themselves subject to rules. To the extent that the analysis is
successful in describing the Nuer facts, it provides evidence against
extending the toolbox of realizational theories beyond what was
proposed in Zwicky 1985 and related work.