Maryland Mayfest 1999

My talk will begin with a brief review of the logic of Baker's 1985 Mirror Principle, and why the issue of whether morphological derivations reflect syntactic derivations still splits the field of Morphology 15 years later. In short, this is arguably the main empirical issue for deciding what the relationship of morphology to syntax is.

Next, I will survey some recent advances in the field, which suggest that morphology reflects syntax even more than previously thought. In particular, the Mirror Principle logic seems now to apply to inflectional morphology as well as to derivational and complex-predicate forming morphology, given (especially) the Macro-study of Cinque 1999 and the micro-studies of Mohawk (Baker and Travis 1999) and Athapaskan (Rice, in progress). There is also some indication that syntax can explain absolute morpheme order as well as relative morpheme order in a substantial range of cases.

Finally, I will propose a (modest) new contribution to this body of work, involving preposition incorporation/applicatives. Post-1988 descriptive work has brought to light a new kind of preposition incorporation, not forseen by Baker 1988, but found in diverse languages such as Slave (Rice 1989) and Abaza (O'Herin 1995). This phenomenon differs from normal applicatives in a cluster of syntactic ways, as discussed clearly by O'Herin: there is no transitivity restriction on applicative, multiple applicatives are possible, applicative does not change binding domains, etc. Strikingly, this phenomenon also differs from "normal" applicatives in a cluster of morphological ways: the adpositional morpheme is adjacent to agreement with the applied object, agreement with the applied object cooccurs with agreement with the base object, the incorporated adposition appears outside other inflectional morphology, etc. Crucially, this is another case of differences in word structure covarying with differences in syntactic properties, showing that the two are not independent. The two clusters of properties can be explained in an incorporation framework by simply saying that the "applied affix" in "normal" applicatives is a lexical category, whereas in Abaza and Slave it is a functional category. Because of this, a pronominal head (D) can incorporate into the adpositional element prior to P-incorporation in these languages, but then P-incorporation must also target the verb in Infl, not the verb in VP by Li's Generalization. As a result, the P shows up outside inflectional material in these languages.

Back to the previous page