Atilà Vital, Federal University of Minas Gerais, Brazil, Carlos Silva, PhD, University of Porto, Portugal, Fabio Granja, PhD

Despite being a rare phenomenon in Indo-European languages, reduplication has been extensively studied in the field of descriptive linguistics. For the purposes of this study, it can be defined as a non-concatenative process that does not function through the addition of specific and regular morphemes to a base. Instead, the repetition of a root is what gives the reduplicated word its final meaning, challenging the claim that language is completely arbitrary. This work presents a critical comparison of the grammatical description of reduplication in two languages: Modern Tupi (a.k.a. Nheengatu) and Tsonga (a.k.a. Xichangana), spoken in the Amazon Rainforest and in Southeast Africa, respectively. We have compared grammars and papers about these languages in terms of their reduplicative systems within the framework of descriptive linguistics and linguistic typology. Our analysis considered the characterization of the process in dynamic and stative verbs and in transitive and intransitive verbs. The main issues found in the materials were: i) the traditional classification of certain words as adjectives in Nheengatu, considered as outdated; ii) the semantic analysis of verbs that can be partially reduplicated in Tsonga. The former can be better explained as in Tsonga, which instead of adjectives has stative verbs which act as modifiers. The latter can be best understood as single iterative verbal actions that affect the same entity, as is the case in Nheengatu, and not as micro-repetitions of these same actions such as proposed by Langa (2013). The conclusions we have drawn from this research project point to the notion that comparing languages as different from each other as Nheengatu and Tsonga helps develop a typology of reduplication. Moreover, the problems found in the materials analyzed seem to point to the fact that Eurocentric grammatical descriptions lead to misunderstandings in the way some languages work. The limitations of the study are mainly related to the scarcity of descriptive materials of both languages and the constraint of our findings to only two languages. Further research could benefit from expanding the proposals presented here to languages whose grammatical descriptions present similar problems.