2338.4 Culture and Classifi cations
An ethnoclassifi cation scheme is merely the skeleton onto which the fl esh of a
community’s TEK is draped. As such, we should be wary of giving too much
importance to this abstract construct, given that the true richness of a people’s
knowledge of the natural world lies elsewhere—in the way they talk about the
relationships between groups of living entities, between living entities and the
physical world, and between living entities and humans. If an ethnoclassifi cation
scheme is to be elicited and analysed, it should be borne in mind that the scheme
needs to be considered in its social context , for there are few linguistic phenom-
ena that remain unchanged either in different contexts, or in a decontextualised
form. Roy Ellen , in particular, has long been suspicious of folk classifi cations
presented as static crystallisations of people’s conceptions of the relations between
natural kinds, and has urged fi eldworkers to pay greater attention to the amor-
phous mass of ‘culture’ that is often rejected in studies of classifi cation (see [ 233 ]
for a collection of essays that attempt to present “Classifi cations in their Social
Context ”). Ralph Bulmer , too, felt that Berlin ’s predictions (I would prefer to use
the label ‘prescriptions’) were far too restrictive to account for the cross-linguistic
variation commonly seen in classifi cation systems [ 151 ]—variation that could
uncontroversially be attributed to cultural differences. However, when the search
for universals turns into a prescriptivist framework into which all data are to be
squeezed, variation and anomalies are often ignored, and the resulting classifi ca-
tions start to resemble one another closely.
Berlin is right to say that a Kingdom level taxon can be present even in languages
that do not have lexifi ed labels for ‘Plant’ and ‘Animal’. After all, languages such as
Tzeltal ([ 9 ], p. 193) have classifi er morphemes that indicate that this is a very real
cognitive division for speakers of those languages. Even in the absence of such
morphosyntactic elements that pervade a language, however, Berlin insists that the
Plant/Animal Kingdom level taxa are cross-linguistic universals. This is where I
disagree, because the fact that respondents can categorise all natural kinds known to
them into groups that approximate Animalia and Plantae cannot be taken as evi-
dence that this is how they usually perceive the natural world. Moreover, some
languages have explicit category labels that speakers will volunteer as indicating the
highest-level taxonomic division in their language—a good example is the
Australian language Kayardild which distinguishes, with primary lexemes, ‘big
game’, which includes marine mammals and turtles, ‘game’, which includes non-
marine creatures, ‘bony fi sh’ and ‘sharks and rays’ [ 234 ]. Surely in the face of
concrete linguistic evidence to the contrary, as well as native speaker insistence that
“ this is how we divide living things ” it makes no sense to keep arguing that there
must necessarily be covert Plant/Animal categories in Kayardild (recall Wierzbicka ’s
earlier comment on ‘sadness’ and ‘marmalade’). If, as Hallpike [ 190 ] suggests,
categories such as ‘things of the forest’ and ‘things of the sea’ are explicitly named
in a language (p. 203), is it valid to disregard such groupings in favour of ‘plant’ and
‘animal’, which may not be named?
8.4 Culture and Classifi cations