39scientifi c taxonomy, and that this observation indicates a universal human propen-
sity to recognise real divisions in nature, Berlin assumes that there is something
absolute about the relationships between the binomial Latin names that have domi-
nated western science for the last two centuries. Berlin also points to a phenomenon
called the “ hollow curve ” in scientifi c taxonomy, where most plant and animal
higher taxa contain large numbers of very small (often monotypic) genera, and a
few very large genera. This, he claims, is identical to the pattern found in folk tax-
onomies, and lends support to his hypothesis that all humans perceive and classify
natural divisions in similar ways. This view has been challenged in recent years, not
only by supporters of the phylogenetic species concept mentioned above, but also
by more traditional taxonomists such as S. M. Walters, former Director of the
Cambridge University Botanic Garden. Walters [ 121 ] is remarkably forthright about
his attitude towards traditional Linnaean classifi cation, saying that the classifi cation
of fl owering plants, at least, suffers from a “European bias” that is largely ignored
by other taxonomists.
Drawing from a number of historical sources, and from the writings of contem-
porary taxonomists, Walters argues that the hollow curve is actually an artefact of
the attachment of conservative taxonomists to the categories fi rst proposed by
Linnaeus. These categories were, after all, heavily infl uenced or even based directly
on pre-existing European folk taxa. All of the 20 largest fl owering plant families in
existence today were fi rst recognised by Linnaeus or by his predecessor Tournefort,
and were based on plants that were either common or culturally important in
Europe—these include plants such as the rose, daisy, nightshade, carrot, lily, mint,
and so on. Walters’ argument is that as specimens of new species poured into Europe
from around the world, taxonomists loyally tried to group these species into familiar
Linnaean (and thus, European) genera, leading to an artifi cial infl ation of species
number in these old genera.
Any group of fl owering plants possessing some obvious ‘natural resemblance’, which con-
tains a number of common, widespread European species of importance in the life of the
community, is likely to contain an ‘excess’ of genera and ipso facto constitutes a relatively
large modern family. In addition to the carrot family, Umbelliferae, familiar examples are
the mint family, the Labiatae, and the cabbage family, Cruciferae. It has long been familiar
to both professional taxonomists and the users of Floras written by those professionals that
these three families are peculiar in that it is usually possible to say that any given plant does
or does not belong to one of these families but it is often very diffi cult to decide in which
genus within the particular family it should be placed. Putting it crudely, the three families
are ‘good’ or ‘natural’, the genera within them are ‘bad’ or ‘artifi cial’. Such families were
established in classical and Medieval European folk taxonomy, and the names of their con-
stituents in the classical languages became the modern genera (p. 6).
Recent examples of widespread revision of generic boundaries in these venera-
ble families are not hard to fi nd, especially outside the Old World. Mitchell [ 122 ],
for instance, presents a heavily revised taxonomy of New Zealand Brassicaceae (the
current name for Cruciferae), noting that the taxonomic history of this group of
plants refl ects the infl uence of Northern Hemisphere taxonomists. In particular, he
mentions that “ most of the New Zealand species were initially placed in well-known
or common Northern Hemisphere genera, but as better quality and more plant
2.4 Problems with Berlin’s Ethnobiological Classifi cation