ceive a difference, naturally in their group's favor, in terms of attrac-
tiveness, honesty or intelligence. They are far more willing to cheat
or indeed inflict violence on members of the other group. Even when
all participants are fully aware that the division is arbitrary, even when
that is demonstrated to them, it seems difficult for them not to
develop such feelings, together with the notion that there is some
essential feature underlying group membership.^13
These well-known results demonstrate the extraordinary strength
of the human propensity toward group solidarity, toward what Matt
Ridley called "groupishness." Humans seem desperate to join some
[288] group and to demonstrate loyalty to it. But the same findings also
show how easy it is to trigger essentialist understandings of groups,
and even to get those understandings to translate into actual behavior
in the context of a coalition. That is, the groups in question are invited to
collaborate more with members of their group than with outsiders.
They quickly develop intuitions and emotions about trust and reliabil-
ity connected to group membership, intuitions and emotions that are
necessary to coalition building.^14
I think artificial laboratory conditions and actual social behavior
converge here in a way that suggests why essence-based understand-
ings become salient and emotionally important: They are the concepts
we spontaneously use to describe intuitions that are in fact not about categories
but about coalitions.
This idea may be abstruse in formulation, but the point is really
simple. Our naive view of social interaction around us is that we are
often dealing with people with whom we share some essential fea-
tures—lineage, tribe, religious practices and so on. But I think we can
get a better sense of how such interaction is actually built if we realize
that many of these groups are in fact coalitional arrangements in
which a calculation of cost and benefit makes membership more desir-
able than defection, and which are therefore stable. In previous chap-
ters, I mentioned the flexible, ad hoc coalitions found in many social
contexts, from the informal alliances and friendship networks between
coworkers to the hunting parties organized by nomadic hunter-gath-
erers. Coalitions require very sophisticated computations of other
people's reliability, based on signals that are often ambiguous and
sometimes could be faked, as well as estimates of the costs and benefits
that result from joining a coalition. Yet people generally do not need
to make these computations explicit. Rather, they justify their behav-
ior by thinking and saying that some people are intrinsically reliable
RELIGION EXPLAINED