For my folk song class, I'm writing an analysis of the old Scottish ballad Kemp Owyne (specifically the version called Kempion; here are some related ballads*). The most interesting part, I find, is looking at the relationships between the different versions and between Kemp Owyne and other stories.
Discovery number one: no version of Kemp Owyne is 'complete' in the sense of containing all the plot elements found in any version of Kemp Owyne. Furthermore, the different versions ('subspecies'?) seem to differ not randomly but systematically in the elements they omit. The Kempion version leaves out the vast majority of the beginning, so for example, you never find out why the evil stepmother curses the princess (in fact, you don't even find out it was the evil stepmother's fault until the very end), whereas other versions start out with a thread where the stepmother is angry about not being called "the fairest of them all". It's hard for me to even imagine what it's like to only know one version of the story, since I'm in the privileged position of being able to read them all and cross-reference.
Discovery number two: there's a lot more swapping of story elements than I thought -- a lot more random ligation of characters into new situations. I suppose most folktales are fanfiction of other folktales, for some sense of 'fanfiction'. For example, Kemp Owyne is supposed to be identified with Ywain/Yvain from the Arthurian legends, aka the historical Owain mab Urien (and "Kemp" means 'hero' or 'champion'). Why? No reason. He's just a convenient fictional knightly hero type. But interestingly enough, there's an invocation to St. Mungo at the end of Kempion... and yet, Yvain is also supposedly St. Mungo's father! How did that happen?? Either there is massive storycest going on, or someone chose to invoke a saint who wasn't even born yet. (You have to admit that time-traveling Scots heroes and saints would be pretty awesome.)
It really is like evolution, complete with horizontal gene transfer. But also with revival and retelling and all sorts of processes that don't have obvious biological analogues. Fascinating!
[*] Those texts have some typos in them, but I wasn't about to type all of Kemp Owyne again, and I'm really fond of that whole website because it makes up in quantity what little it lacks in quality. (The typos are few enough that I feel like only scholars really ought to care about them.)