end p.409
although in most other ways in startling contrast to Christianity, the cult of Moloch
overlaps with it in involving the worship of a personal deity; and Christianity in turn
overlaps with the Theravada in the quite different respect that it offers a comprehensive
interpretation of life” (5). If the concept of religion is a family-resemblance concept, such
resemblances allow us to classify the cult of Moloch, Christianity, and Theravada
Buddhism as religions without supposing that all three of them satisfy a single set of
necessary and sufficient conditions for being a religion. According to Hick, once we
understand the concept of religion in this way, we have resolved or, perhaps, dissolved
the problem of defining religion.
Cognitive psychologists have developed more refined versions of the family-resemblance
view. According to one version, some concepts are organized around an example that
serves as a prototype (see Rosch and Mervis ). As a result of complex patterns of
resemblances to and differences from the prototype, other cases lie at various distances
from the prototype in a similarity space. Cases near enough to the prototype clearly fall
under the concept; cases far enough away from the prototype clearly do not fall under the
concept. And there may be a gray area in between where borderline or contested cases are
to be found. Applying this general idea to my own concept of religion, we may imagine
that the prototype is the religion I was brought up in, which is Christianity. Judaism and
Islam are near enough to my prototype to be clear cases of religion, but soccer is far
enough away that it clearly does not fall under my concept of religion, though I
understand what is meant when it is said that some people make a religion of soccer.
Confucianism, Daoism, and Thervada Buddhism are near enough to my prototype to
count as religions, and Nazism and Soviet Marxism are far enough away to lie in the gray
area for me.
It seems to me this refinement of the family-resemblance view does some real work in
explaining the classificatory practices I engage in using my concept of religion. So I do
not find it surprising that the family-resemblance view is currently the dominant or
received view of the concept of religion. However, it does not command unanimous
agreement. Jim Stone () has recently proposed and defended against putative
counterexamples a new theory of religion that provides its own definition of religion.
According to Stone, “A religion is a system of practices meant to place us in a relation-
of-value to a supermundane reality so grand that it can figure centrally in the satisfaction
of substantial human needs” (188). He advertises his achievement with the bold claim
that “religion has an essence, I will maintain, which the new theory reveals” (177). It may
be that Stone's definition will prove to be immune from clear counterexamples. But even
if this turns out to be the case, I doubt that he will convince many philosophers that it
reveals the essence of religion. Structural theories in the natural sciences provide
definitions of natural kinds that are plausibly construed as revealing their essences.
Familiar examples are the claim that water is H 2 O and the claim that gold is the element
with atomic number 79. However, there is no good reason to suppose
that religion is a natural kind, for religions are social products rather than things we
discover in nature. So even if Stone's definition cannot be shown by means of
counterexamples to be mistaken about the extension of the concept of religion, there is no