xxix, 918 p. : ill. (some col.), maps / This dissertation is a comprehensive grammar of Kokama-Kokamilla (KK), as spoken by about 1000 elders in the Peruvian Amazon. It presents detailed documentation of the structures of the language and the functions they serve, with rich exemplification. This study is based on significant fieldwork since 1997, allowing the analysis to be grounded in text data. One of KK's most salient typological features is a morphological distinction between male and female speech in several grammatical categories. Major grammatical categories like person, number, tense, and modality are conveyed by positionally fixed clitics. Five tense clitics encode three degrees of distance into the past and two into the future. There are also six epistemic modal clitics, which interact to create a four-way modal system. None of the twelve suffixes is obligatory, but, in language use, as many as four can occur together on a single verb, followed by up to two clitics. Syntactically, KK has intransitive and transitive clauses, but semantically three-place predicates are syntactically encoded by means of transitive clauses. There are six directive constructions that distinguish degrees of pragmatic force. Another noteworthy point is the multiple types of purpose clause which differ in terms of coreference, controlled by the matrix clause absolutive argument rather than the subject. Clause nominalization is a central subordination strategy, particularly in relativization, which is largely achieved via an absolutive nominalizer. Pragmatically, KK has constructions that explicitly distinguish subtypes of focus according to scope (narrow/broad) and pragmatic information (contrastive/noncontrastive). This bears on theories about whether contrast simply emerges from conversational implicature, versus can be explicitly coded by dedicated grammar. Information structure also explains the distribution of alternating pronominal forms and constituent orders. Though long classified as Tupí-Guaraní, recent research claims that KK is the product of a contact language situation and hence has a mixed grammar. Although this claim seems accurate, this study shows the possible Tupí-Guaraní origin of a number of grammatical morphemes. What languages (and families) contributed to the rest of the mix remains to be determined. Thus, this comprehensive description is an important step in advancing comparative studies among the languages of the region. / Adviser: Spike Gildea, Chairperson; Committee members: Doris Payne; Scott DeLancey; Francisco Queixalos; Lawrence Sugiyama