Special Focus on Umlaut in Germanic
Optimality Theory and Morphology: The Role of Umlaut in MSG
Paul Houseman UW-Madison
ABSTRACT
This paper analyzes the highly complex system of Standard German plural formation from within the framework of Optimality Theory (Prince and Smolensky 1993), where a grammar is conceived of as a set of ranked, violable constraints on well-formedness of the output. The system has two components: the function Gen, which generates a set of possible candidates from the input, and Eval, a set of ranked constraints that winnow the candidate set in order to single-out the most optimal output. Both will play important roles in this paper. The project in general has two goals: to explain the productive plural system of Standard German without recourse to lexical exception or vague terms like analogy and to explain why new language learners (German children, second language learners, etc.) typically make the mistakes in acquisition that they do. That is to say, I break with standard Optimality Theory (OT) in positing a default constraint ranking in children (at least as it relates to plural production) as well as constraint rankings that have been lexically conditioned in adult speakers. This applies in particular to the thorny issue of the use of umlaut, or stem vowel change, to mark plurality and I will show that OT offers at least a partial solution for handling this phenomenon by using both standard phonological constraints and purely morphological constriants adapted from current work being done in Natural Morphology. Specifically, I develop Natural Morphological ideas of markedness, iconicity, and salience into a OT constraint framework.
Previous accounts of German plural formation, whether in standard generative theory or natural morphological accounts, have failed to deal with umlaut data‹ precisely the data that needs to be explained before a true picture of the system can emerge. OT accounts, on the other hand, while looking at umlaut data in German have so far either not attempted to cover plural formation‹ one writer going so far as to claim that all use of umlaut in plural formation is lexicalized (and hence memorized by speakers)‹or dealt with only a tiny fraction of plurals in the language. In section 3, I briefly discuss these various attempts within other frameworks and some of the serious problems associated with them. Finally, one particular advantage of an OT framework is that it can be used for both synchronic and diachronic explanations of plural formation in Germanic languages. That is, there is no need to create new rules to handle data at each successive temporal stage in the language, but rather a single set of universal constraints is posited where solely differences in ranking of constraints can account for changes in the data. I will then show how the OT framework I set up for Standard German could thus be applied diachronically for older stages of the language and for dialect variation within a language family.