Isn’t that the essence of the issue, that those models are loaded with biases, that might or might not overlap with dominant ones in inscrutable ways, hence producing new levels of confusion and indirection?