The pitfalls of working on endangered languages #3

When conducting an experiment designed to elicit language-in-use tokens of demonstratives, it is best to ensure in advance that the speakers know the language names of the items you expect them to refer to in the endangered language - no matter how 'basic' the vocabulary might seem to the non-native user of the endangered language.

While the situation may generate perfectly natural language-in-use tokens of speakers searching in vain to for the (incidental to the researcher in this context at least) lexeme, and plenty of tokens of the 'whatsit' demonstrative, straightforward tokens of 'this tooth' may, in the end, be few and far between.