46 Peter Beyer
to do with boundaries, and one with the valuation of these concepts in their social
contexts. Thus, we have disagreement about what does and does not belong in a cate-
gory like religion or culture; we debate the boundaries between members of a category,
such as where one religion ends and another begins; but we can also contest the status
of the categories themselves. We do this, for instance, when we discuss the legitimacy
of what can be claimed by appealing to categories like religion, culture, sport, art, or a
number of other social forms.
In one sense, problems of this nature are as old as human history. Boundaries are
the very stuff of social structures and human knowledge: We make distinctions and
thereby create ordered worlds. Yet, although social order would be impossible without
them, these social forms also always seem in one way or another to be problematic, to
not quite “work” (e.g., Berger and Luckmann 1966; Douglas 1966). That said, however,
the specific ways that this general feature works itself out in contemporary society has
its particular and somewhat unique characteristics when compared to societies of the
past. It is to the contemporary situation with respect to the idea of religion that this
chapter addresses itself.
Sociological discussions about defining religion have almost always come to the
conclusion that this is a difficult exercise about which there is little agreement. Gen-
erally, these debates hover around the central organizing distinction between substan-
tive and functional definitions or restrictive and expansive ones (e.g., O’Toole 1984;
Hervieu-Leger 2000). More often than not, substantive definitions, which focus on ́
what religionis, tend to be restrictive; and functional definitions, which center on what
religiondoes, lean toward being more expansive in what they include. Accordingly, the
most typical criticism of substantive/restrictive definitions is that they include too lit-
tle, perhaps on the basis of an implicit theological bias that wishes to exclude “false”
religion. By contrast, a frequently cited weakness of functional/expansive definitions
is that they exclude too little, thus rendering the term meaningless and perhaps even
betraying an antireligious bias: What “religion” does can be done (better) by many other
things, like the state, art, sport, medicine, or science. Thus, from the nature of the func-
tional/substantive difference and the criticisms of either side, it becomes evident that
all three of the axes of dispute I mentioned above are at work. Sociologists have disputed
the boundary between religion and nonreligion, what counts and doesn’t count. They
have disagreed on how valuable or important religion is, whether it is necessary or not.
And behind both issues is that of internal variety: They assume that there are many
religions, irrespective of whether the favored approach is substantive or functional.
In both sociological and nonsociological realms, therefore, the term religion re-
mains somewhat elusive. And this along similar lines of dispute. One reason for this
parallelism is undoubtedly that the two domains exist in the same social and historical
context. That fact leads to this hypothesis: The definitional or conceptual difficulties
with respect to religion point to a social context that encourages and perhaps even
requires “religion” to be multivalent. In other words, the problem is not in the am-
biguous or variable nature of “religion itself,” whatever that may be but, rather, in a
social context that makes such ambiguity sensible. It is this variability of religions in
that social context which is the specific focus of this chapter. The sections that follow
explore various aspects of this overall question, and they do so by translating it into
two interrelated matters: The social context of contemporary global society and the
social forms that religion and religions typically seem to take in this context. The main