languages must share,but which are not necessarily found among artifi-
cial languages or languagelike artifacts.For those whose major concern
is evolutionary biology,a weakness of the former approach is suggested
by the frequency with which that approach draws attention to what it
calls implicational universals (generalizations of the type,“if language A
has feature x,it will also have feature y”) or partial or statistical univer-
sals,in which languages share a given feature with a frequency far greater
than that of chance.Clearly in such cases there exists the possibility that
the universals concerned could be merely contingent and in no way
directly derive from the biological makeup of the species.Surely we
would be on firmer ground if we confined ourselves to properties that
are exceptionless and specific to natural language.Similarly,in biomusi-
cology one should not expect to find universals lying conveniently on the
surface ready for any untrained investigator to pick up.However,
attempts to discover these universals,if properly conducted,should be
crucial in determining whether music is a species-specific adaptation,like
language,or something that may be shared,at least in part,by members
of other species.
It should be borne in mind,too,that music may not be a natural kind
(Molino,this volume) and may accordingly be decomposable into dis-
tinct modules.Chomsky (1980) suggested that language may be similarly
decomposable into two components,conceptual and computational.
This distinction was developed into the one between a meaningful but
unstructured protolanguage,potentially sharable with other species,and
a syntactic mechanism that imposed a complex hierarchical,parsable
structure on this protolanguage to yield contemporary human language
(Bickerton 1990,1995).It is equally possible that music may turn out to
contain elements specific to our species mingled with other elements that
may be much more widely shared.
A further problem for an evolutionary study of human behavior con-
cerns misplacing emphasis on one of the two basic ingredients that make
up an evolutionary process.For any trait to emerge in the course of
evolution,some kind of selective pressure must exist that is a set of cir-
cumstances that renders the trait adaptive in terms of increased progeny,
and a degree of genetic variability must be present from which the trait
can be selected.However,a number of recent studies have concentrated
exclusively on possible candidates for the selective pressure that affected
language,and ignored the variation that must have existed for the pres-
sure to work.Thus a number of studies sought to attribute the emergence
of language to the growing complexities of life in hominid groups,
ignoring the fact that social life already achieved near-human complex-
ity among many primate species (de Waal 1996) and that any additional
hominid complexity was more likely a resultof language (e.g.,complex-156 Derek Bickerton