John Kupchik

<kupchik@hawaii.edu>

Department of Linguistics
University of Hawaii at Manoa

On the variation of morpheme order in Mari declension: variable morphotactics or morpheme scrambling?

The ordering of inflectional affixes is largely held to be universally fixed in the sense that with a particular polyaffixal construction the morpheme order is unchangeable, or if there is an alternate order it co-occurs with a different meaning and thus has a different underlying representation. Exceptions to this rule are so rare as to be widely considered non-existent, and indeed linguists have often stated the fixedness of morpheme order as a universal (cf. Perlmutter 1970).

Mari (old name: Cheremis), an agglutinative Finno-Ugric language in the Uralic family, rather remarkably goes against the idea morpheme ordering is universally fixed. What occurs in Mari declension is a truly fascinating and unique phenomenon of debatable origin, purpose, and usage. But simply put, the ordering of bisuffixal and trisuffixal strings of case, possessive, and plural markers is variable, with almost always at least two possible (=grammatical) orders for any particular suffix string, and in some cases three possible orders.


The main study on this phenomenon is Luutonen's dissertation (1997), in which he terms it "variable morphotactics", his conclusion being it is essentially the result of rule-ordering conflicts among the Mari dialects along with plural markers in the process of being grammaticalised. In this presentation I will give an overview of Luutonen's study, which focuses on the Eastern and Meadow dialects. I will also present original data I have gathered from a native speaker of Mari (Western, Hill dialect). I will then argue that what occurs in Mari declension should not be termed variable morphotactics, but rather "morpheme scrambling". I will show in detail how this morpheme scrambling operates in the Hill dialect, and contrast it against the Eastern dialect. I will also compare morpheme scrambling in Mari to syntactic scrambling. In doing so, I will show that the scrambling in Mari is simply within a different domain from typical scrambling, that being within the domain of the word as opposed to the sentential, or phrasal domain. This phenomenon is unattested in any other language, and is thus rife with theoretical implications. 

Luutonen, Jorma. 1997. The variation of morpheme order in Mari declension. Mémoires de la Société Finno-Ougrienne vol. 226. Helsinki: Suomalais-Ugrilainen Seura.

Perlmutter, David M. 1970. Surface structure constraints in syntax. Linguistic inquiry 1:2.