Yeah I genuinely do not buy the argument into the value of mythology as a valid form of comparanda.
Again, I wrote up some stuff about it in how pop feminism is a victim of our pattern seeking brains and I think that a lot of comparative Indo-European mythological scholarship is also a victim of this predisposed nature of how we can like... attribute patterns where there are none to be had.
Like we can establish sound changes and their correspondences by a set of constraints (or rules) that govern them. What do we have that can establish formal relationships between myths? What are there to make sure that they aren't coincidences?
Sure, you can say that gold and geld are not through the reconstruction of sound changes that go back to PIE—but how can one even be sure in the regard that people want to see patterns in things with archeological material? Or the stories that people tell?
A sky deity that can be reconstructed in the mythology of IE-speaking people? Yep, must be a shared ancestor and not some kind of pattern that we independently invented like 30 different times.
Again I don't know how people can take this seriously except for works that project one's ideologies into the past.
Yeah I'm of the opinion that the most we can do with linguistic data is infer the mays and may nots of the culture but we cannot really say for sure what they were like.
From linguistic evidence we can infer that the shared ancestor of the Evenki and Even had reindeer pastoralism because they have words that trace back to Proto-Ewenic; however, we can't really say for sure for shared immaterial cultural traditions and such.
Like, of course the word for 'headhunting' can go back to Proto-Malayo-Polynesian, and can infer the existence of headhunting at least. but like… hypothesizing in which how they headhunted through comparative cultural data seems to be very odd to say the least; as shared innovations can just be parallel or just… independent formed from one another. It's also to note that culture is not driven by unconscious thought and that the drift itself is not… quantifiable by any means that I've read.
Like this is what gets me when people cite stuff like "Oh the Pleiades myth is so old that Aboriginal Australians and Greeks share this story" but like—personification of celestial bodies is so common, and the deification of the Pleiades is shared between other cultures like Japan for example (hence, Subaru)
Like, it's hard to say for sure what is what—like you can make connections for sure, but it's kinda hard to infer whether these are substratal influences, coincidence, or shared ancestors.
I've never been particularly interested in this aspect of IE studies myself, but from what I can tell of the studies in this area that try to establish a common origin with some rigor, they rely not on the general similarity of the ideas or elements of the story, but how the elements of the story fit within the culture generally, if anything stands out, and particular details and/or specific actions (e.g. not that there is an underworld guarded by animal guard(s), but instead that the animal is a dog with many eyes; not that there are three hunters pursuing an animal, but that the second hunter carries a cooking pot) – even more though, they (should) try establishing shared linguistic components, which can mean names of the same origin, archaic morphology or syntactic structures, similar (metaphorical) use of specific words (especially as they deviate from their expected use), identical stylistic features and so on (in appx. order of significance). Once you go beyond a single language family, the thing that remains are syntactic structures (depending on the characteristics of the language families involved, of course), metaphorical use of words, stylistic features and story details (the latter not really linguistically relevant, as mentioned), so there's automatically fewer things you can operate with and the conclusions must necessarily be more tentative.
But I think there's one other thing that should be pointed out: the default position with regards to whether certain myths are related is "we don't know." Both shared origin and independent innovation are non-trivial claims, and there's always the possibility of borrowing. Of these the shared origin one definitely requires a higher threshold of confidence, but if it cannot be demonstrated, that doesn't automatically make the myths independent innovations – that has to be substantiated as well, plus there's always the possibility of borrowing as mentioned.
And in particular in a situation where there is evidence that speakers of X, Y and Z were once part of a single community of speakers, then if they share a story that has similar elements, both the independent innovation and borrowing options seem to be even more non-trivial, i.e. require even more substantiating for one or the other option – an independent innovation of X vs. something demonstrably shared by Y and Z for example requires a break in tradition with subsequent unrelated introduction of the same idea; and it seems to me that something that is trivially introduced is similarly trivially maintained.
Attempts at a reconstruction will always be an imperfect approximation of a reality we have no access to and things can happen in such a way that by their very nature we will never be able to tell how exactly they happened. This is why when e.g. you cannot demonstrate that two myths share a common origin that means just that – it cannot be demonstrated. In reality they might well share an origin, yet all the relevant details and linguistic elements have been replaced, stripped away etc. to the point where all we can say is we don't know. To take independent innovation as the automatic default option would be wrong too, then. We might of course say that given all the available information it is equally or even more likely to be an independent innovation or a borrowing and that would be methodologically sound and even true! But note crucially "given all the available information," "likely," "or."
official linguistics post