Legacy materials

[Rudy Troike]

To add my two cents to what Susan Penfield said, sometimes one may be in a
"salvage" mode, to collect as much language as possible before it is gone
forever, as has happened in the past (Samuel Gatschet's experience with
Karankawa is an example -- much of what we have would not exist if he had
waited six months to interview a person who remembered hearing the language
as a child). But apart from recognizing -- as Bill Poser noted -- the often
radical differences among genres (unfortunately ignored by most linguists in
the past), and recording samples of various general types, if there is time
there should be concurrent analysis, since some aspects may turn out to be
uninterpretable later on without a living speaker to check them with. Bill's
example of traditional texts' containing archaic words is one example, since
the meaning may still be known to the person telling the stories, but might
not be known to younger speakers or might not be determinable from mere later
analysis. Or a rare grammatical form might occur only once in certain types
of texts for pragmatic reasons, and because of its uniqueness might not be
interpretable on later inspection or might not be known to younger speakers,
if there are any. Constraints on the possible extension of grammatical forms
(such as the influence of an animacy hierarchy) might be undiscoverable
without concurrent analysis and exploration with competent speakers.

    Rudy Troike