This is your occasional reminder that the ‘bear taboo’ is not actually a Proto-Indo-European understanding.
The archaic Indo-Europeans of the Urheimat do not appear to have had a problem referring to a Bear quite directly – as *h₂ŕ̥tḱos.
We can tell this because there’s a suite of directly cognate terms for ‘Bear’ in not only Greek and Latin (ᾰ̓́ρκτος – ‘Arktos’; and ‘Ursus’, respectively), (or, for that matter, the Celtic languages – *Artos, Arzh, Arth, and Art, for Proto-Celtic, Breton, Welsh, and Old Irish, respectively) but also the most archaic IE languages that are directly attested to us. In Sanskrit – Riksha ( ऋक्ष – also anglicized ‘Rksa’); and in Hittite, 𒄯𒁖𒂵𒀸 (anglicized as Hartaggas).
Why do we mention those last two? Well, it is often thought that Anatolian IE languages ‘branched off’ pretty early from the PIE sphere proper – in fact, so early that some features which developed later and are pretty pervasive across much of the rest of the IE sphere hadn’t formed yet. (The usual ones highlighted are Hittite’s approach to verb tenses; and the fact that its primary ‘genders’ are ‘animate’ and ‘inanimate’ rather than ‘male’, ‘female’, ‘neuter’)
So archaic Hittite etc. can be a pretty interesting window in on what may have been going on in PIE. Although subject to a few caveats – like the interesting question as to whether some developments Hittite ‘missed out on’, were in fact independent and parallel developments post-PIE in these other IE spheres; and, of course, once we step away a bit from the linguistics into the culture etc. – how much of what we’re witnessing has resulted from contact, influence, or outright incorporation from non-IE peoples they encountered on the way to where they eventually ended up or at their borders, through trade, under their dominion, etc.
Anyway, we can tell that *h₂ŕ̥tḱos means ‘Bear’, not simply because the word’s descendants occur in those languages aforementioned … but because it’s *directly applied* to something pretty enduring: the Stars.
More specifically, the constellation of Ursa Major / Arktos Megale [oh hey, look, cognate! PIE *méǵh₂s … ] – which, in the Vedic sphere, we find hailed as the Rksa. [Of course, there’s also *other* names for this Constellation – and most prominently, we have the ‘SaptaRsi’ labelling, wherein the Seven Stars are said to be the Seven Vedic Rsis [‘Seers’]; but more upon all of that some other time. I detailed a few thoughts in my earlier – ‘Arktos, Ursa, Rksa SaptaRsi – The Seven Bear-Seers Amidst The Stars And The Foundational Act Of Piety Of The Maidens Of Milk And Fate’]
Meanwhile, it’s in the Germanic and Slavic spheres that we see PIE *h₂ŕ̥tḱos being rendered conspicuous via absence – with an array of other more ‘circuitous’ expressions being employed instead. ‘Honey-Eater’ [which gets into a Four Lions reference], ‘Brown One’ (Bjorn, Bear, etc.), those sorts of things. It may be important to note that there’s not really good coterminity between these two different styles and linguistic spheres here – showing that it’s likely a *significantly* post-PIE and independent development in either ‘side’ (a position somewhat strengthened by what seems to be a persistence for ‘proper’ PIE Bear terminology even as late as Proto-Balto-Slavic, given Lithuanian has ‘irštvà’, with the ‘irs’ referring to ‘Bear’, as in iršas .. which, well, you can see what it means (think ‘Ursus’). It’s rare, but it is there)
(As a brief point of minor interest – the Slavic ‘Medvedev’ [from Proto-Slavic *medvě̀dь; and meaning ‘Honey/Sweet Eater’] finds a reasonably direct cognate in Sanskrit मधुअदः [‘Madhvadah’ – that is to say, ‘Madhu-Adah’] … although in RV I 164 22, it is the ‘Suparna’ (potentially interprable either as Birds / Eagles, or as the rays of illumination from a heavenly body such as the Sun or Moon … or both) that is eating of the Sweetness, and rather not a bear)
The supposition is that for some reason, the PIE term fell out of use – because, it is presumed, speaking the name of the creature was likely to draw its attention and summon it forth.
Why didn’t this happen for those other PIE-descended peoples?
Well, perhaps it is that as they were in areas rather removed from the forests of Northern Europe, they were less overtly likely to run into large and aggressive Bears in the first place. And therefore did not develop the same aversion to the more archaic term and its accompanying totemic value. Which runs into the rather difficult-to-reconcile situation of the Celtic IE seemingly not having this issue.
There’s another suggestion that maybe the ‘taboo’ manifested in other ways for some cultures – instead of wholesale replacement of the term, the supposition that some part of the word was twisted so you weren’t quite pronouncing the ‘forbidden’ name. This underpins a theory about why Armenian արջ (‘Arj’) sounds the way it does; although there are other interpretations.
A third possibility is that the whole thing’s taboos – all the way down. And that the *h₂ŕ̥tḱos is, itself, a shift from a more archaic term likely meaning ‘Destroyer’ (PIE *h₂rétḱ-os – whence ‘Rakshasa’, etc.) – or, for the sake of argument, an adjectival term for destruction that gets turned into a noun itself.
That … seems rather peculiar; and basically just like a case of really wanting to keep the taboo thing going as a PIE concept, even though the way the taboo was supposedly observed ‘in the wild’ was through people not using that actual PIE root and its immediate descendants.
Meanwhile, the concept makes no sense. If you can’t say ‘Bear’ because you might attract the Bear … then how can you possibly say ‘Destructive’, ‘Destroyer’, ‘Destruction’ without running the similar risk of attracting the … well … that. We just simply wouldn’t have words in PIE for a great many things except via ellipsis, then.
Further, as a certain philological crocodile has observed – it must be one very odd taboo if there’s no apparent issue with directly and overtly naming an entire classification of Demons with the Sanskrit descendant from the ‘Destroyer’ root … which are, themselves, not subject to any such ‘taboo’ of a linguistic nature to mention. And this is before we get into the other potential PIE explanations for ‘Rakshasa’ that are rather more removed from *h₂ŕ̥tḱos and which surface from time to time.
[Before somebody suggests it, I strongly reject the suggestion that the Rakshasa is, in essence, a ‘Bear’ – because we can demonstrate that there’s a rather pervasive and more positive (or, at least, divinely-linked) mythic situation for the Bear in the archaic IE spheres]
In any case … it IS plausible to contemplate the notion that ‘words have power’ in a PIE context. Certainly, many post-PIE Indo-European ‘magical’ (in fact, more ‘ontological’ – as it really is a metaphysics of language, speech, nature, and being that is not so ‘removed’ or ‘relegated’ as ‘magic’ might perhaps unfairly imply) perspectives enthusiastically run on such a notion.
But I’m not Quite sure that they uh .. work like this. At least, on a general enough level to explicate the near-wholesale disappearance of a perfectly viable PIE root from two rather major IE language families.
The ‘taboo’ of the Bear non-mentioning has, in some ways, become quite the ‘totemic’ shibboleth itself.