230
plants are now rare or have disappeared completely, and Solega TEK is now heavy
with the memories of the sights and smells that were once taken for granted.
Medicinal plant s that are still not completely extinct have retreated to lower altitudes
where Lantana has yet to establish its grip, and the Solega have had to update their
knowledge of these newer habitats.
Increased contact with the outside world is also leading to an infl ux of new con-
cepts and methodologies, which could either be incorporated unchanged into a pre-
existing traditional framework, or modifi ed to suit indigenous needs and beliefs.
Outside infl uences include new crop plants, agricultural techniques, medical knowl-
edge, religious infl uences and naturally, all the vocabulary associated with these and
other innovations. I noted at least two instances of parts of the Lantana plant being
used by Solega nowadays—the crushed leaves are applied to the skin following a
beesting to reduce infl ammation, and the charcoal of Lantana branches, obtained
from a camp fi re , is used to clean one’s teeth in the morning (Table 3.4 ). In the pro-
cess of destroying much of Solega TEK, then, Lantana is also able to make a small
contribution towards increasing it.
8.3 Fact or Fiction: TEK from the Analyst’s Point of View
I have argued repeatedly in this book that two things are essential when investigat-
ing the potentially complex mental representations of natural phenomena. The fi rst
is the elicitation of data in contexts that are as natural as possible, in order to ensure
that the categories and the relationships between categories so obtained are native to
the language under investigation, and not artefacts of the data collection method.
Failing this, any conclusions drawn from data obtained through artifi cial contexts
(i.e. interviews, goal-oriented experimental tasks) should refl ect the shortcomings
of the data collection method, and the biases that may be present in such data should
be acknowledged explicitly. In the words of William Labov [ 73 ], founder of the
fi eld of variationist sociolinguistics:
Either our theories are about the language that ordinary people use on the street, arguing
with friends, or at home blaming their children , or they are about very little indeed. (p. 109)
The second point follows directly from the fi rst: naturalistic speech data should
be given priority over elicited data, as it is only through studying the patterns of
naturalistic speech that we can gauge the accuracy of our analytical models of peo-
ple’s mental representations. The study of such data also reduce the chances of the
analyst applying his/her own ontological prejudices to the analysis of native catego-
ries in another language.
The study of categorisation in an ethnographic context is one fi eld of enquiry that
is especially prone to context effects and over-analysis, as discussed in Sect. 1.3.
The existence of ‘black’ and ‘white’ as basic colour categories is often cited as a
cross-linguistic universal, but as Wierzbicka [ 109 ] has pointed out, not even ‘colour’
can be assumed to be a concept shared by all languages. Careful examination of the
8 Conclusions