I had read some papers about these Altaic language hypothesis, but there are a few theories with more ore less differences. These are all hypotheses, no final proof. One group didn't accept the arguments of the other and vice versa. It's like kindergarten.
For me there is enough proof for an Altaic language family. You could also split the IE language family into Romance, Iranian, German, etc language families if you look at the differences. It's just the same with Turkic, Mongolian and Tungusic.
Sent from my KFAUWI using
Eupedia Forum mobile app
Well, of course it's not hard science, and it hasn't been definitely established, but on one side you have the majority opinion of mainstream scholars in linguistics, on the other you have proponentes of a hypothesis discredited by most reliable and not too speculative methods of comparative linguistics and debunken by the majority of linguists. Basically, the very hypothesis of an Altaic language family started on weak grounds, because it assumed that Uralic was linked to Turkic and other language families of Northeast Asia, but it was soon discarded that Uralic was part of that hypothetical language family.
The main problem with the Altaic hypothesis is that it simply doesn't resist when you apply to it the same requirements that are applied to other well attested language families like Sino-Tibetan, Indo-European or even the most elusive of all (because it's probably the oldest split of all), Afro-Asiatic. Comparative linguistics do not work in an uncontroversial way for the hypothetical Altaic languages, especially if you include the most "unlikely" members of it, like Korean and Japanese (let alone Uralic, which is rightfully often excluded even by the modern proponents of Altaic).
Unlike virtually any language family, in which the languages become more similar the further back you go in the evidences of their evolution, the Altaic languages become
more different from each other in their ancient stages than they are nowadays. That's a classic clue that these are not phylogenetically related languages, but distinct language families that formed a strong Sprachbund and mutually shared areal features and loanwords. The very "basis" of the Altaic hypothesis is pretty weak because it can be explained away easily as a Sprachbund effect much like many common features developed after the Iron Age in different European language groups, namely: Altaic languages are supposed to have a common origin because they're all very aggluttinative (not exactly a rare feature crosslinguistically), they are mostly SOV and most of them have vowel harmony. Those are very generic similarities, not really specific features that are very unlikely to be developed independently "by random chance". Even vocabulary-wise the extremely generous (some would say excessively speculative) pro-Altaic analysis by Starostin estimates "only" 16-22% of lexical correspondence in the 110 most common words.
Take a look at the reconstruction efforts by Starostin and Blazek (and not just in Altaic, but in many other hypotheses their work is often considered a bit too self-confident and bold, too devoid of scientifically healthy caution, for example with reconstructions of "macro-families" supposed to have been spoken dozens of thousands of years ago out of very thin and unreliable evidence). The reconstruction has been widely criticized by linguists because of many mistakes: incorrect or distorted meanings to "force" similar-sounding words to be hypothetical cognnates; incorrect words in some languages, period, using words that sound similar to those of other language families, but actually never meant that; assuming as cognates similar-sounding morphological particles (e.g. case affixes) even though they have completely different functions. The are also few regular sound correspondences that work consistently and that have been accepted extensively by other scholars.
Comparisons with Romance, Germanic and so on are untenable. Those language families are clearly related even by having a simple glance at the most basic parts of the speech: pronouns, numerals, basic verbs and nouns, etc. The assumed relationship between the Altaic subfamilies would have to be much, much older (not the 4k-5.5k assumed for most IE branches) to account for how different they are and how hard it is to find regular sound correspondences and a numerous enough quantity of consistent morphological and syntactic correspondences between them. And that would still not explain why on Earth for instance Turkic and Mongolic are much more similar now than 1000 or 2000 years ago (using either written attestation or linguistic reconstructions), exactly the contrary of Romance vs. Germanic or Iranian vs. Greek.
As for aDNA, I'm pretty sure that you understand that it is much more probable to make a link between a language expansion and specific, "recent" subclades of Y-DNA haplogroups than between languages . Using the uncontroversially accepted methods of historical linguistics, the furhest ago you can go tracing back the origins of languages is about 10-15 kya, if that, so anything further back will be just unrecoverable even linguistically, let alone any relationship with changes in genetics or culture. But here we're not talking about very basal haplogroups like R, Q or N here, but about things like R1a-Z93, with a much more recent, chronologically and geographically defined origin and expansion in the aDNA record and associated with much more recent and therefore much more traceable historic events. We're talking of movements and changes that happened a mere 1500-2000 years ago, in historic times, with written documentation about some facts, and not about genetic and linguistic processes 30,000 or 40,000 years ago. The further back you go the more likely that genetics and linguistics become disjointed, but if you have a lot of contemporaneous (i.e ancient) genetic and linguistic evidence and the facts are still reasonably close to us historically to be plausibly linked to some sequence of known historic facts, then the likelihood that those dynamics between language and genetics became totally unrecoverable is lower.
Additionally, I did not correlate languages with Y-DNA haplogroups alone, but with all the genetic evidences: autosomal DNA, Y-DNA, Mt-DNA... and that all fits together with the historic (documented), archaeological and linguistic evidences, too. I would never make any conclusion about the most probable linguistic affinity of a population based only on their Y-DNA haplogroups, even if really specific and nearly contemporaneous (in origin) subclades. It's the combination of many lines of evidence (and not just many genetic evidences via different data, but also evidences provided by other sciences) that substantiates that claim, and it's of course not certain, but it is more plausible and reasonable, because that hypothesis (Scythians being mostly - not necessarily exclusively - IE speakers who then got gradually Turkified beginning some 2200 years ago and especially in the Late Antiquity i.e. 1400-1800 years ago) is the one that best fits the many data that have been collected.
If you want to know why I concluded that, and based on several lines of evidence combined in the way that seemed most reasonable way, I explain it extensively in my Quora answer about this very same issue:
https://qr.ae/TWXm4Q