Informal, discourse or spontaneous data are probably the most important types of data to collect documenting a language,
since it tells us how the people actually use the language on a daily basis and gives us access to many different registers and speech styles,
than narratives and elicited datas alone.
Per say, how people actually speak in quotidian situations is usually quite different from how they translate elicited sentences or tell a story.
In addition, studying other people’s conversations is extremely useful to learn how to speak fluently and naturalistically,
and is essential for language revitalisation programmes.
In the case of highly endangered languages, such datas might be quite difficult to obtain.
Occasions to record group of people using naturally the language might be rare, and people might feel uncomfortable speaking spontaneously while being recorded.
In order to get the best naturalistic data possible, speakers need to feel comfortable being recorded
and a strong relationship of trust and respect with the field researches should be established in the first place.