The current presentation focuses on the functions of multilingualism in interaction between the researcher and participants in a small corpus of videotaped materials, produced on the Taimyr Peninsula, Russia, in the 1990s (SKNA Helimski). The tapes, filmed by the late linguist Eugene Helimski, contain interaction in multiple languages, mainly Nganasan (Uralic) and Russian. Despite its relatively small size (approx. 7 hours), the corpus features a diverse array of situations. The Nganasan language community is in an advanced stage of language shift to Russian, and many speakers demonstrate linguistic insecurity typical for language-shift situations (cf. Abtahian & Quinn 2017). On the other hand, some speakers have difficulty fully expressing themselves in Russian, which they have learned as a second language.

Using the methods of interactional linguistics, I observe how participants employ different linguistic resources, such as repair and code-switching, to establish their roles in each situation, both in relation to each other and the codes used (Raymond 2018: 169–170). The speakers, possessing epistemic rights to the language, take the role of the ’teacher’ in relation to the linguist (cf. Sandman & Grzech 2022: 96–98), who occupies an ’outsider’ position compared to the other participants. However, Helimski’s mode of conversation is often casual, and he is an active participant rather than a passive observer. By using Nganasan in, e.g., other-repairs, the researcher takes on a domain expected to be outside of his epistemic rights by default, which is socially risky (Drew 2018). However, I argue that by taking a linguistic risk, the researcher can also demonstrate willingness to learn and make mistakes, which lowers the threshold of the other participants to speak freely, leading to more naturalistic data and better researcher–community rapport.