232
The objections raised by Baird et al. and Tversky against a neat, simplistic and
unifi ed view of cognitive map s echo my own doubts about the status of elaborate folk
classifi cations (one for ‘Animals’ and one for ‘Plants’) as real, static mental representa-
tions shared by members of a speech community. As I have argued in Chaps. 1 – 4 , and
as pointed out by other authors [ 13 , 14 , 71 , 151 ], folk classifi cations are inseparable
from the social contexts in which they are routinely used, and any effort to elicit a
complex taxonomy in an interview or experimental situation will result in the docu-
mentation of categories and relationships that are alien to the speakers’ usual way of
thinking. Far too many studies on classifi cation systems concern themselves with only
the lexemes that label biological categories, and the hierarchical relationships between
them. Covert categories of natural kinds are frequently posited in classifi cation sys-
tems, but these rely heavily on perceptual sorting and grouping tasks, with linguistic
data being rarely used. The results of the Boster and Johnson [ 56 ] experiment men-
tioned in Chap. 2 suggest that people will use purely perceptual cues as a last resort
when no other information—linguistic or encyclopaedic—is available to them. The
case for ‘perception’ in the construction of ethnobiological classifi cation s is probably
highly overstated, and as some authors have argued, the classifi cations derived from
perception-based sorting and grouping tasks are probably not accurate representations
of native conceptions of the relationships between natural kinds. Ellen [ 232 ] notes that
card-sorting tasks suffer from serious methodological fl aws (even though the task itself
may be a useful way to elicit novel information from respondents), and that any hierar-
chies generated by this method should be regarded with caution because:
the human mind is suffi ciently fl exible to ‘rewrite’ classifi cations in a taxonomic form, to
rely entirely on binary discriminations, or upon morphological criteria, depending on
implicit prompting. (p. 28)
Additionally, the fact that these tasks are carried out by individuals in isolation
(i.e. without consultation with other speakers) exposes an assumption that there is a
single accepted classifi cation evenly distributed across the speech community, and
that uncovering that classifi cation is simply a matter of averaging the responses of
an adequate number of members of that community. The Solega bird naming study
presented in Chap. 4 refutes such an idea, showing instead that variation is indeed
the norm even at the level of lexical labels for common bird species. This variation
was uncovered in spite of the fact that in my elicitation sessions, consultants were
presented with visual and auditory stimuli as a group, and were allowed to freely
consult with one another to reach a consensus.
Similarly, it has to be remembered that cognitive map s of the type described in
Chap. 5 (especially Fig. 5.5 ) are probably not permanently stored mental represen-
tations analogous to a cartographic map—it is most likely not a three-dimensional
snapshot of the external environment, with herds of elephants and swarms of bees
moving across the surface in concert with the seasonal cycle , and accompanied by
bursts of fl owering fi rst in one habitat, and then in another. Instead, elements of the
cognitive map, such as the phenological signature of a particular forest/landscape
type, are likely to be reconstructed from the diverse pieces of encyclopaedic
information that each Solega speaker possesses, and with a clear goal, such as trav-
elling to a particular location for a specifi c purpose, in mind.
8 Conclusions