Theravada Buddhism and Nazism satisfy the defining conditions proposed by Geertz but
not those proposed by Spiro, if we assume that neither
end p.408
of them postulates superhuman beings. This disagreement could be accounted for on the
supposition that the two definitions serve to analyze the concepts of religion the two
anthropologists bring to their research. For they might well have slightly different
concepts of religion. Of course, we would expect their concepts to overlap to a large
extent, differing only in their applicability to a minority of cases. But it is possible that
each definition satisfies a criterion of intuitiveness with respect to the intuitions of its
framer. The two definitions also disagree about the kind of entity a religion is. For
Geertz, a religion consists of symbols; for Spiro, a religion is an institution. It seems that
a perspicuous social ontology would treat symbols and institutions as belonging to rather
different metaphysical kinds. The difference in kind between the American flag (a
symbol) and the American League (an institution) is at least as large as the difference
between chalk and cheese. This disagreement could be accounted for by the supposition
that the two anthropologists favor somewhat different approaches to their discipline. If
one has been trained in cultural anthropology and it provides useful methods for studying
symbols, then defining religions in terms of symbols will portray them us proper objects
of investigation for cultural anthropology. Similarly, if one has been trained in social
anthropology and it provides useful methods for studying institutions, then defining
religions as institutions will portray them as proper objects of investigation for social
anthropology. So this disagreement could be rooted in pragmatic considerations.
However, these two disagreements are not trivial even if they turn out to have rather
simple explanations.
It should not be thought that the problem Theravada Buddhism raises for certain attempts
to define religion is without parallels to other cases. Chad Hansen (1997) has argued that,
though we usually classify Confucianism and Daoism among the world's major religions,
we will doubt that they are religions if we compare them with our conceptual stereotype
religions. This is because “classical Chinese philosophy shows signs neither of creation
myths, of attempts to explain `why we are here,' of a mind/body (or spirit/body)
dichotomy, nor of supernaturalism” (25). So we might expect otherwise attractive
definitions of religion to disagree about whether classical Confucianism and Daoism
count as religions.
Such disagreements have given rise to skepticism about the possibility of analyzing or
defining the concept of religion in terms of conceptually necessary and sufficient
conditions. An alternative view, derived from Wittgenstein, who uses the example of
games to illustrate his point, is that the concept of religion is a family-resemblance
concept. John Hick, an advocate of this view, argues that it is “illuminating to see the
different traditions, movements and ideologies whose religious character is either
generally agreed or responsibly debated, not as exemplifying a common essence, but as
forming a complex continuum of resemblances and differences analogous to those found
within a family” (1989, 4). Thus, for example, “the bloodthirsty worship of Moloch in the
ancient Near East had nothing directly in common with Theravada Buddhism; but on the
other hand,