The Brazilian Portuguese-Russian Corpus (BraPoRus) of older heritage Russian speakers

Abstract

The Brazilian Portuguese-Russian Corpus (BraPoRus) is the first corpus of spontaneous speech by cognitively healthy older heritage speakers of Russian (ages 60–100) who are dominant in Brazilian Portuguese yet retaining functional proficiency in heritage language (HL) Russian. BraPoRus includes over 200 hours of naturalistic speech from 43 participants and includes a subcorpus, BraPoRus-1.0, already pre-processed and available for research via BilingBank of TalkBank. BraPoRus-1.0 comprises 15-min. extracts from different interview sessions that contained descriptions of various life events from 16 participants (women n=10, men n=6; M_age = 80.5, age range 65-98). All recordings were transcribed into Cyrillic automatically by Sonix.ai and cross-checked manually; they were also processed with a BatchAlign-2 pipeline for automatic processing of media files. In addition to linguistic documentation, all BraPoRus participants underwent cognitive testing, enabling multimodal analyses. BraPoRus thus provides a critical resource for investigating HL maintenance, attrition, and linguistic aging. Its integration with psycholinguistic experiments enables cross-methodological studies, while its open-access design supports replication, comparative analysis, and tool development for HL pedagogy. As such, BraPoRus contributes a much-needed lifespan perspective to corpus-based bilingualism research.