Prosody as a means to express Tense in the Kaingang language Márcia Nascimento 1 , Marcus Maia 1 , Leticia Rebollo-Couto 1 1 Federal University of Rio de Janeiro indiaedai@yahoo.com.br, maiamarcus@gmail.com, rebollocouto@yahoo.fr Abstract This study describes a prosodic process used in the Brazilian Indigenous language Kaingang (Je family, Macro-Je stock), to distinguish tense in a specific verb class ended by consonantal segments. Pitch distinguishes past and future tenses in that verb class, in both declarative and in interrogative sentences. A psycholinguistic experiment and acoustic analyses showed that Kaingang, considered an accentual language in the existing literature, displays a grammatical tone functioning as a morphological feature, in addition to a set of particles and morphemes. A rising contour of fundamental frequency or a high tone on the domain of the word verb indicates future tense whereas a falling contour or a low tone indicates past tense. Index Terms: Tense, Kaingang language, Tone languages, Speech perception. 1. Introduction The Kaingang language is one of the more than 150 indigenous languages, spoken by the 239 indigenous peoples. of Brazil. Kaingang is spoken by a population of about 30,000 divided among 32 indigenous lands, spread in the three southern states of Brazil as well as in the São Paulo state. This language presents a dialectal variation that could be classified between four to five variants. Following [1], within the Jê languages family, Kaingang forms a well-distinguished branch, with similarities with Xokleng. Together, they form the southern branch of the family. We can distinguish four regional languages in Kaingang [2]. This research is centered on the language spoken among the Nonoai community, located in the north of “Rio Grande do Sul” state, and amounting to about 2.500 speakers. Typologically, Kaingang is an isolating language, and its pattern word order is SOV – Subject, Object, Verb. The subject is morphologically marked by different particles (subject “tỹ” for declaratives or “mỹ” for interrogatives, gender “fi” for female, etc.). Kaingang is described in the literature as an accentual language, but presents a case of prosodic or melodic inflections to mark time in verbs which end by consonants, like: (1) “han” [do] (2) “krãn” [plant] (3) “nũr” [sleep] (4) “kusũg” [paint] (5) “kym” [cut] (6) “kãtĩg” [come] (7) “jẽn” [feed] (8) “jẽnkrig” [clean] This process is accomplished by means of a fundamental frequency (F0) inflection that does not follow Kaingang's standard description of lexical stress, nor utterance-level intonation. Contrasting to these two levels, this F0 inflection is implemented at the level of verb morphology, as proposed in our analysis, which is elaborated in [3]. A perception test was designed at the Experimental Psycholinguistics lab of the Federal University of Rio de Janeiro (LAPEX/UFRJ), and was taken by members of the Kaingang Nonoai village. The experiment tested if a melodic inflection could distinguish tense, opposing past and future in non-morphologically inflected verbs, ending in consonants and which do not receive suffixes. To support the perceptual results, an acoustic analysis is proposed, making explicit and confirming the hypothesis that melodic height is the distinctive feature opposing past and future tenses for such type of verbs. 2. Corpus and Methods 2.1. Corpus The data is based on two sets. All were read aloud by an L1 Kaingang female speaker in a sound proof room for creating the recording material. The first set is used for the perceptual experiment and the first acoustic analysis; it consists of 64 utterances, divided in 32 interrogative utterances and 32 declarative utterances. In order to verify if the melodic inflection was due to the domain of the verb word or to the domain of the sentence, a second set was recorded, adding an adverb at the end of the sentences. This set contains 32 utterances: 16 assertive utterances presenting at the end of the sentence the rẽkétá adverb ( meaning ‘yesterday’) and 16 sentences presenting the adverb vajkỹ (meaning ‘tomorrow’), also at the end of the sentence. The interrogative utterances were also distributed in the same way, using the same time adverbs in the end of the sentences. The second set was used for the second acoustic analysis. Example with verb “jẽnkrig” [clean]: (1) Fẽrá fi tỹ ĩn jẽnkrig rẽkétá. Fẽrá Fem. DECL. house clean yesterday ‘Fẽrá cleaned the house yesterday’ (2) Fẽrá fi tỹ ĩn jẽnkrig vajkỹ. Fẽrá Fem. DECL. House clean tomorrow ‘Fẽrá will clean the house tomorrow’ (3) Fẽrá fi mỹ ĩn jẽnkrig rẽkétá? Fẽrá Fem. INT house clean yesterday? ‘Fẽrá cleaned the house yesterday’ ? Speech Prosody 2016 31 May - 3 Jun 2106, Boston, USA 326 doi: 10.21437/SpeechProsody.2016-67