and kindly, in Lowland Scotland and in France, for example. Meanwhile, the very nature of the incidents—a bestial mother (totemism, or worse?), a helpful beast (Manitou), a magical tree, a talking bird—are of that kind which the savage fancy undeniably and universally evolves. These things, as Sainte-Beuve says, would not be introduced now, could not be invented now, without the old examples, inherited, as I suggest, from a period of barbarism. "But", it may be urged, "if you allow that polygamous might be altered into monogamous details, why should men have retained beast-mothers, talking birds, helpful animals, revivified bones?" Well, first, even polygamous peoples have romantic love affairs. The polygamy need never have been conspicuous in the story, and, at most, a jealous co-wife could easily become a jealous stepmother. Secondly, without the talking birds, helpful animals, revivified bones, talking trees, you no longer have the story. You have to do what Perrault did, and to introduce a new "machinery", a fairy godmother (new, here), transformed rats (even that, in essence, is as old as Circe), and though Monsieur Perrault could do all this, it was a task rather beyond peasant grandmothers. To drop polygamy, if ever there was a trace of it in the tale, was very much more easy. But, even in a polygamous country, the institution need not have been introduced into Cinderella.
Thus I see no proof that a tale full of savage fancy, most manifest in the forms which seem oldest, and are rudest, did not arise in a savage state of society. I admit that the tale has been diffused, the tale as it stands in most versions, shoe and all, but, as Mr. Jacobs allows, this present version may not be the original. He suggests "a later and inartistic junction of the sea maiden formula" in the conclusion of some Celtic versions, and an ingenious dovetailing in of elements from another and more archaic tale, in "the earlier part". How much then is left of the original? What is the original? In truth, any tale may
