This dissertation develops a theoretical and empirical framework for the analysis of the ideological and interactional constitution of language. It discusses the process of "making language," namely, how language emerges as an object of speakers' attention, the historical processes leading to this type of language consciousness, and the interactional means through which it is achieved and becomes recognizable and analyzable. Integrating work on language ideologies, phenomenology, language socialization, practice theory, conversation analysis, and the ethnographic description of ontologies, this work offers insights into the underlying mechanism of how language becomes a meaningful entity in the lifeworld of its speakers.

Focusing on the constitution of language opens up new avenues for the investigation into its ontological status. Language is here understood as an equivocation that might index potential referential alterity. Individual languages need not always be tokens of the same type and thus arbitrary and translatable. Language and languages are specific objects that result from the socialization of speakers into conceiving of and attending to particular communicative practices as languages.

To analyze the constitution of language, the dissertation introduces the concept of metalinguistic repair, understood as the deliberate replacement of a term from one code with a semantically equivalent term from another in ongoing interaction. Together with other metalinguistic strategies in language play and language teaching, metalinguistic repairs are theorized as phenomenological modifications by which the code is highlighted and language is constituted as an object that is distinct from the speaker, the meaning, and the context of the utterance. The consequence of these modifications is what is called here enlanguagement, a term from studies of pidgin and creole genesis that is redefined to designate the process through which speakers are oriented to notice particular pragmatically salient linguistic features as belonging to different languages, thereby constituting these as distinct entities.

This work is based on ethnographic research in an indigenous Aché community in Eastern Paraguay. It draws on five years (2008-2013) of language documentation work with the Aché, as well as one year (2013-2014) of in-depth language socialization research in one Aché community through video-recordings of children's everyday interactions, interviews, and participant observation. The Aché are a recently settled hunter-gatherer collective, currently experiencing language shift from their heritage language, Aché, to a Paraguayan national language, Guaraní. The presently dominant medium of communication in the communities is a mixed code, using elements from Aché and Guaraní.

The context in which the Aché children grow up is unique and ideal for this study, because despite the fact that language differences are not relevant in everyday interaction since language mixing is the default mode of communication, the children do attend to them in everyday conversation and play. Through spontaneous repairs and corrections, the deliberate use of specific forms, and discussions about language, they demonstrate an awareness of the linguistic code as a distinct aspect of language use. Such situations are analyzed in detail as key moments in which "language" and "languages" are created. The Aché children do not merely use different languages that are somehow already constituted as given entities in their lifeworld. Rather, by employing a multiplicity of linguistic resources in their everyday interactions they end up making language and languages and making them over.

This dissertation bridges the domains of ideology and interaction in order to provide an integrated account of how language emerges as a cultural and historical product on the one hand, and as an interactional achievement on the other.