Specifically, we want to verify Frazer's argument by encoding all of the ethnographic literature that he piles up and then seeing
- if the model selection algorithm that he uses can find the argument that he makes from the cases that he presents (internal validity)
- if the model selection algorithm that he uses finds other arguments as well, using the same case material (internal decisiveness)
- in the case of model ambiguity, whether Frazer selected the best model, most likely in the minimal-message reformulated sense of Ockham's Razor (internal conciseness)
More on what we mean by model selection in a future post.
This is distinct from showing that Frazer's argument is consistent with current or even past research agendas (communal validity). In fact, few today will share his assumptions about the uniformity of religious belief of an Aryan Race, which entails
... every inquiry into the primitive religion of the Aryans should either start from the superstitious beliefs and observances of the peasantry, or should at least be constantly checked and controlled by reference to them. (Preface, p.viii, ed. 1894)Frazer himself uses this stance to downplay the reports in the old literature vis-a-vis the cultural context in which the non-literate peasantry has been reared.
... and so it has come about that in Europe at the present day the superstitious beliefs and practices which have been handed | down by word of mouth are generally of a far more archaic type than the religion depicted in the most ancient literature of the Aryan race. (p.viiif)The idea for the title so far has been "The Logic of Sympathetic Magic: Taking a Leaf from Frazer's Golden Bough". I think it's funny, too.