Attributing human or animal agency to events is an explanatory strategy which, while it sometimes fails, is in general extremely effective. Since other humans and, after them, animals are the most important things in our environment, it is vitally important to take them into account when they are there — important enough that erring on the side of caution means accepting regular anthropomorphic and animistic "errors". The result is that our conceptual and perceptual schemata incline us to animism and anthropomorphism. This is the central argument of chapters two and four of Faces in the Clouds, in which Guthrie pulls together ideas from all over — from Piaget and Gombrich as well as from modern cognitive science and psychology.
This forms the core of Guthrie's explanation of the generality and ubiquity of anthropomorphism. He goes on, in chapters five and six, to illustrate that ubiquity, with surveys of anthropomorphism in literature and the arts and in philosophy and science. Illustrated with thirty two pages of black and white photographs, this is, once again, an excellent summary, which should appeal even to students of art, literature, and the history of ideas without a special interest in theory of religion.
The final chapter pulls these strands together to argue that religion is anthropomorphism. Superficially non-anthropomorphic religions either turn out to be anthropomorphic on closer inspection or are better described as philosophical systems, while attempts by some theologians to strip religion of anthropomorphism produce results which are barely recognizable as religion. Of course not all anthropomorphism is religious — but then there is no clear-cut divide between religion and other forms of thought which exhibit anthropomorphism, so this should be expected. Religion is, then, an explanatory system, but one whose basic operations are mostly unconscious. As explanatory systems, religions lend themselves to intellectual systematisation, but religious belief may at the same time be "felt rather than thought".
There has been relatively little new work on theory of religion in recent decades: there seems to have been a general acceptance that the subject is too hard, that the existing theoretical morass is undrainable. While I don't think Faces is likely to change this, I do wonder whether it shouldn't. Allowing that definitions of religion have a certain amount of arbitrariness about them and that some degree of eclecticism in explanation is likely to be unavoidable, Faces is the most convincing attempt at a unitary model I have seen. It is, in any event, a significant contribution to the field: many have linked religion and anthropomorphism before, but none have done it so convincingly and few have explored the causal psychological mechanisms underlying the connection.