When linguistic research takes place in the natural setting where the language under investigation is spoken rather than at a desk in an air-conditioned office at one's home university, this has consequences for the endeavor. In the field, one becomes part of a social network in the speech community under investigation, and thus this type of research necessarily involves as much personal and social effort as it does linguistic “brain work.”

Descriptive linguistics appears to have more standard ways of working with informants than, say, ethnography does. In current anthropology, skepticism about the ethnographer's ability to understand and convey (what are assumed to be) subjective experiences encountered in the field seems to dominate the discussion. In empirical linguistic research, one may also run into epistemological problems, some of which the linguistic fieldworker – and the theoretical linguist consulting descriptive sources – can ill afford to ignore. The interpretation of most linguistic signs requires a context in time and space. Researchers have to try and find out whether variation in speech between informants they consult are the result of elicitation techniques or whether they truly reflect linguistically interesting variables in the data. Formalizing linguistic field methods is possible only to a certain extent. The rest depends on the serendipity of the individual linguist and elusive insights that no clearly defined eliciting procedure seems to be able to insure. Below, I present some practical guidelines for the basic investigation of (relatively) undescribed languages that is to be carried out in conjunction with native-speaker informants.