I am an Assistant Professor of Slavic Linguistics at The Ohio State University.
My theoretical research interests fall primarily in the area of synchronic morphology: paradigms and word-based models, information theory as it relates to inflectional structure, inflectional defectiveness (a.k.a. paradigmatic gaps) and other form-meaning mismatches, the structure of the lexicon, and cross-linguistic differences in morphological processing.
My areal research interests include Croatian (Serbian, Bosnian), Russian, and Modern Greek, roughly in that order. I also have a side interest in the other South Slavic languages, and Albanian.
At the graduate level I have most recently taught morphological theory, the structure of Bosnian/Croatian/Serbian and quantitative research methods in linguistics (corpus-based linguistic analysis). My Winter 2010 seminar will be about frequency effects in grammatical structure. At the undergraduate level I teach various identity issues, esp. the immigrant experience. A new undergrad course on language and politics in SE Europe is in development.
Current Projects
Learning Defectiveness: Given that inflectional morphology is relevant to syntax, we implicitly expect all lexemes to have forms expressing all relevant inflectional property sets. In this light, a persistent mystery is how (and if) paradigmatic gaps can be generationally stable in the absence of ongoing structural causation. Why are such gaps not 'automatically' filled in during the learning process? Janet Pierrehumbert (Northwestern U.), Robert Daland (UCLA) and I are applying a statistical learning approach. Contrary to claims in much of the literature on defectiveness, we argue that gaps can be learned and persist under such conditions. Using a multi-agent network model with a simple Bayesian learning mechanism, we show that a sensitivity to and balance between neighborhood similarity effects and tracking of relative usage within the inflectional paradigm is sufficient to model the persistence of gaps in Russian (e.g. *pobežu 'I will be victorious) and Modern Greek (e.g. *koton 'chicken.gpl'). So far, this work has resulted in a 2007 ACL paper and a few conference presentations.
Probabilistic Paradigmatics: The idea of intraparadigmatic dependencies is an old one in morphology. Traditional Word and Paradigm (WP) models included the idea of a principal part, i.e. the word-form(s) in the paradigm from which all others could be predicted. But while modern WP models have argued for a variety of paradigm-based dependencies, principal parts have played little role. In my dissertation (The Ohio State University, 2006), I argued that the fundamental idea of a principal part is still needed in paradigmatic morphology. Specifically, I argued that the distribution of inflectional defectiveness in Modern Greek can only be understood in these terms (gaps occur exactly in the cells and inflection classes which are neither predictive of other paradigm cells, nor predictable from them). In ongoing work, I am using the information theoretic concept of entropy to quantify this notion, and exploring other areas in which entropy seems to have explanatory force in inflectional morphology.
Cross-Linguistic Differences in Morphological Parsing: Race models of morphological processing posit that multiple factors influence the likelihood of a complex word being decomposed (parsed) vs. accessed 'whole' (stored) in the lexicon: productivity, junctural phonotactics (probability of a morpheme boundary based on phoneme bigrams), relative frequency, etc. Since these factors are language-specific, the balance between parsing and storage is predicted to vary cross-linguistically (but most work has been done on English or typologically similar languages). Members of my lab are testing aspects of this prediction for Russian. Lauren Ressue is working on the parsability of lexical vs. superlexical aspectual prefixes, and semantic consequences thereof. Jeff Parker is working on implications for derivational affix ordering.
Case Syncretism and Structural Attraction: In morphological theory, it is often unclear whether a given instance of syncretism is systematic or accidental. Decisive distributional and/or historical evidence is not always available. This project seeks to develop a behavioral test that will identify systematic syncretism, based on structural attraction effects (false agreement). So far this work has resulted in a conference presentation at the 16th Balkan and South Slavic Linguistics Conference (2008).