Talk:Indo-European migrations

Seconding need for update
The first sections of the article are pretty outdated in light of works like Lazaridis et al 2022 as well as Kroonen et al 2022 and other linguistic works that refute the argument that Balto-Slavic is related to Indo-Iranian. The Kurgan hypothesis has also changed: Anthony has proposed a revised version which is not a Kurgan hypothesis. The article is like an incomplete mashup right now and would need a rewrite, perhaps even summing up older arguments and information and focusing on more recent ones, which are not as thoroughly explained. To put it plainly, the narrative it weaves is one mostly stuck in the past, ignoring modern research. This is due to a lack of detailed analysis of recent research and a disproportionate focus on older analysis and hypotheses. 2A02:85F:E03B:3E00:2946:D607:82A3:9EBA (talk) 22:44, 9 March 2023 (UTC)
 * Balto-Slavic is obviously related to Indo-Iranian. We just don't know how closely. There are obvious similarities between the two branches, but there isn't a consensus in Indo-European linguistics on what that means exactly, and there hasn't been for a long time. It's by no means a new idea that they might not be particularly closely related (no closer even than any of the other "core IE" braches, perhaps), nor is it particularly relevant to the homeland debate (and this article isn't even primarily about the homeland). That early forms of Balto-Slavic and Indo-Iranian were spoken in Eastern Europe in the third millennium BC is still highly likely, regardless of how closely related they might have been.
 * Determining the relationships between the major branches (Indo-Iranian, Balto-Slavic, Albanian, Armenian, Graeco-Phrygian, Anatolian, Tocharian, Italic, Celtic, Germanic) is notoriously difficult anyway, so it's not such a big deal. Linguists have long tended to effectively treat the differentiation of IE (especially "core IE") as an "explosion" into a number of distinct branches without any significant interrelationships (with the possible exception of Italo-Celtic), only later areal contact that has obscured the picture.
 * Anthony (2007) has only dropped the "Kurgan culture" moniker, which has long been controversial anyway, but the steppe hypothesis hasn't been essentially changed by him, so the relevance is unclear. Even if the term "Kurgan culture" is not favoured anymore, the term "Kurgan hypothesis" is still a valid synonym of "steppe hypothesis". --Florian Blaschke (talk) 19:47, 30 July 2023 (UTC)

Heggarty
There is an article in the Independent of 28 July 2023 about the paper Haagerty at al. (2023), Language trees with sampled ancestors support a hybrid model for the origin of Indo-European languages, published in Science  on the same day. This includes ‘The latest research points to a new hybrid hypothesis for the origin of the Indo-European languages with a homeland south of the Caucasus and a subsequent branch northwards onto the Steppe, as a secondary homeland for some branches of Indo-European entering Europe with the later Yamnaya and Corded Ware-associated expansions.’and ‘“Recent ancient DNA data suggest that the Anatolian branch of Indo-European did not emerge from the Steppe, but from further south, in or near the northern arc of the Fertile Crescent—as the earliest source of the Indo-European family,” Paul Heggarty, another author of the study, said. “Our language family tree topology, and our lineage split dates, point to other early branches that may also have spread directly from there, not through the Steppe,” Dr Heggarty said. The summary of the Science paper (I do not have access to the full paper) also includes: ‘Indo-Iranic has no close relationship with Balto-Slavic, weakening the case for it having spread via the steppe.’

There is a map in the Independent, which is not very clear, but it seems to show Greek and Albanian as having spread directly from Anatolia, and leaves the origin of Celtic as unspecified. There are various arrows and question marks for the spread of the Indo-Iranian languages.

I request that someone who has more technical knowledge of this subject than I have should add information about this latest hypothesis to this article, and to the article on the Proto-Indo-European homeland. Sweet6970 (talk) 10:35, 30 July 2023 (UTC)
 * I don't think that this source should be added because it represents a very WP:FRINGE point which won't find any acceptance in the academic community. It directly contradicts all studies which have been published in 2023 and all studies which will be published after September and all studies which are scheduled to be published in early 2024. Many glottochronological studies which have been published over the years have proposed various alternative dispersion routes for IE languages, but they're not included in relevant articles because most times these alternative scenarios are highly improbable.--Maleschreiber (talk) 11:17, 30 July 2023 (UTC)
 * According to WP:FRINGE a Wikipedia article should not make a fringe theory appear more notable or more widely accepted than it is.. It does not say that it should not be mentioned at all. And according to Wikipdedia, Science is one of the world's top academic journals so I don’t think we should ignore this. Sweet6970 (talk) 11:40, 30 July 2023 (UTC) |


 * I think it is rather flippant to dismiss this new study from Science out of hand as "very fringe," and a mischaracterization. The Near Eastern model for the origin of Indo-European is not an outlier theory, and has been gaining rigorous scholarly attention in recent years. Further, your criticism isn't based on anything concrete, such as methodology. I think the fact that this new study is more than just a glottochronological study, but also an interdisciplinary work that draws from insights in archeology, anthropology, and genetics, warrants that it be given serious attention. Jpd50616 (talk) 11:46, 30 July 2023 (UTC)


 * Max Planck Institute... I suggest we wait for some scholarly responses, before we add this. Joshua Jonathan  -  Let's talk!  11:49, 30 July 2023 (UTC)


 * I concur with User:Maleschreiber and User:Joshua Jonathan, although not out of a priori rejection. It's a new, uncited paper for which the jury is still out, so presenting it here and now violates WP:DUE. We may include novel research results from subject matter experts published in subject matter-related journals with due weight, but not from sources that partially use WP:FRINGE-methodology published in journals that are not dedicated to the field. Science is specialized in science, but not in historical-comparative linguistics. If the linguistic part of this interdisciplinary project was based on uncontroversial mainstream historical-comparative methods, I would consider a preliminary mention of the paper much less problematic.
 * If such sources gain major attention in secondary sources (beyond news reports), we may include some mention of them with due weight. But let's all have a look and thorough read first, maybe things aren't as bad as the Independent makes them look. Keep in mind that the Independent was capable of shitposts like calling the Tarim mummies "China's celtic mummies". –Austronesier (talk) 11:51, 30 July 2023 (UTC)
 * I agree with the statement that we shouldn't make  and that we shouldn't just ignore the study. As this was just published, we can only compare it to other high profile publications which have been published in the last 5 years and they don't support such an opinion. I believe that throughout the year there will be several reviews of this study and then we can decide how to engage with it. There's no need to rush for its inclusion as we can wait for academic reviews to be published and then we can discuss how to depict them in the article.  Claims of interdisciplinarity in such studies often mask a complete lack of interdisciplinarity, but I agree with  that we should we wait for responses from the academic community.--Maleschreiber (talk) 11:55, 30 July 2023 (UTC)


 * Interestingly, I recall R1a, Part of the South Asian genetic ancestry derives from west Eurasian populations, and some researchers have implied that Z93 may have come to India via Iran[36] and expanded there during the Indus Valley civilization.[2][37]. That always seemed weird, but noteworthy, and in this context, quite relevant. Joshua Jonathan  -  Let's talk!  12:01, 30 July 2023 (UTC)


 * I agree that it is a very recent publication, and we are here to report established consensus. It is also not fringe, it is a new analysis which uses newish evidence to unite the current leading hypothesis with lagging, but longstanding and respectable, hypotheses in this area of study into a coherent, albeit complicated, synthesis. I'm not competent to criticize the methods, though they look respectable to my eye, and whether it achieves general acceptance only time and much analysis will tell. But I do suggest that it's worth giving a very brief outline of its main points. Richard Keatinge (talk) 12:47, 30 July 2023 (UTC)


 * The claim of interdisciplinarity might well be technically correct (no doubt genetics, archaeology and anthropology are distinct disciplines), but a study on languages without linguistics included in the interdisciplinary mix does little to inspire confidence. --Florian Blaschke (talk) 20:15, 30 July 2023 (UTC)


 * I concur that I it seems best to wait until the article has had some response from the relevant expert community/communities. And it does seem to make some extraordinary claims, e.g. the idea that Indo-Iranian doesn't derive from the steppe, which as far as I know, conflicts with the genetic evidence of steppe DNA in Indo-Iranian populations (e.g. in Iran and India), as well as the scholarly opinion that Indo-Iranian derived from the Corded Ware culture (through the Sintashta culture), which in turn derived from the Yamnaya or something related. Skllagyook (talk) 13:39, 30 July 2023 (UTC)


 * The study puts PIE at c. 6000 BC (We find a median root age for Indo-European of ~8120 yr B.P. (95% highest posterior density: 6740 to 9610 yr B.P.).), well outside the 4500–2500 BC range derived from linguistic evidence, so this looks like yet another rehash of Gray/Atkinson: trying to do historical linguistics without doing historical linguistics, building trees with methods derived from genetics but without consulting actual historical linguists or having sufficient competence in the field. Not noteworthy. Ringe has already shown how to execute the same idea competently. --Florian Blaschke (talk) 20:08, 30 July 2023 (UTC)
 * While I don't subscribe to the theories Haagerty et al. propose, and I do think there's a big problem with "historical linguistics" studies not consulting historical linguists (has been for years, and the media loves it. "Mapping the Origins" people!), I don't think this disqualifies Haagerty et al. from being noteworthy. If it gets enough attention (again, like "Mapping the Origins"), then I think it should warrant a mention, with a line detailing criticisms as well.
 * Wikipedia is an encyclopedia, I think we should document ALL theories about IE origins/expansion, if they get enough attention. JungleEntity (talk) 01:23, 3 August 2023 (UTC)
 * There are so many of those papers now – fringy, attention-grabbing papers that make a big splash among non-linguists (i. e. laypeople) but are severely criticised by linguists (i. e. relevant experts), especially their methods and conclusions – that we cannot document all of them (and definitely not as soon as they are released). It just gets tedious, and we shouldn't give bad science more attention than it deserves. --Florian Blaschke (talk) 01:41, 3 August 2023 (UTC)
 * That is sadly the case with a lot of anthropological academia nowadays. I agree, we shouldn’t give these people a podium, but if it gets enough attention, I think it should be included in Wikipedia. It’s better to include it (once again, if it has enough attention), while also pointing out that most experts in the field disagree with the data or methodology of the project. The alternative is not including it at all, which I think brings more harm. I’ve been lucky enough to study IE linguistics in an academic setting, and I can see how saturated the field is right now. I can’t imagine what it might be like for a layperson, with the IE journals left are dying or are off the wall, and fringe theories seeping there way more and more into the top of YouTube and Google results. In recent years, the only big publication I’ve seen addressing this problem is “The Indo-European Controversy” by Pereltsvaig and Lewis, and that is still paywalled (I think? I can’t tell with university access, although I remember not being able to find a copy when away from uni).
 * Wikipedia might be the only place where people interested in IE studies can easily see “Yes, this research project has gotten a bunch of attention in the press recently, but historical linguists have criticized it for x and y.” JungleEntity (talk) 15:25, 3 August 2023 (UTC)

After a cursory reading of the paper and parts of the Supplementary Information, I want to clarify on three points:

1. "It's an interdisciplinary paper" – True, the list of authors includes scholars from various disciplines, and the final conclusions of the paper concerning linguistic archeology (= speculations about the linguistic identity of archeological cultures and genomically defined populations) are certainly a collective effort. But the main part of the paper on which all subsequent conclusions hinge is the computational phylogeny of the IE language complete with split date estimates. No new archeological and genomic are presented to complete the picture; the latest phylogeny of Gray's team is just grafted onto existing models of the demic spread of genes and cultures. I.e., it is primarily a linguistic paper with an interdisciplinary appendix. And again, trying to sell research result from one's discipline in a non-specialist journal is a big red flag.

2. "It is not fringe" - If the conclusions of this paper are at odds with a long-grown consensus about the linguistic archeology of IE languages, that certainly doesn't make it a fringe paper. Linguistic archeology is essentially speculative and rests on the plausibilty of inherently unprovable assertions (such as the linguistic identity of pre-literary ancient peoples). BUT: the methodology employed to arrive at the proposed phylogeny IS fringe. Quantitative computational methods in linguistics are increasingly accepted in the field as long as they are not promoted as supplanting well-established qualitative methods. Quantitative methods remain controversial in the field, and among computational linguists, Gray's methods are not widely accepted. As Ringe has nicely put it, Gray's methods have been destroyed under scrutiny from experts with knowledge in both "conventional" historical linguistics and computational linguists.

3. "Historical linguists were not consulted" – Historical linguists were involved, but just as "cognacy deciders". Consider the implications: a big computational apparatus is set into motion in order find the objectively best fit of the data (NB the entire paper itself is data-free), but at the bottom, the tree rest on heuristic subjective judgements that are directly linked to a preconceived notion of the phylogeny. We cannot reconstruct proto-forms at the highest level without a subgrouping model, otherwise we cannot distinguish retentions from innovations. So unlike in genomics, where we have unambiguous objective matches between A, C, G, and T, in linguistics it is the tree that implicitly determines cognacy decisions, which in turn serve as input to build a tree. Historical linguists were certainly consulted, but not for their expert capability to produce results through reasoning. Their role is reduced to serve as data feeders. –Austronesier (talk) 21:35, 30 July 2023 (UTC)


 * Here are some more critical comments. I can't access the Science-article (yet), but the map seems to suggest that (Indo-)Iranian arrived in India from Iran, and that the IVC was IE-speaking. That's bizarre. The IVC gene pole was partly derived from Iranian hunter-gatherers (the same sort of people who contributed CHG to the steppe, I suppose), but that's another migration; are they mixing-up different migrations? And what about Sintashta, and the relation between Vedic practices and Sintashta? Some sort of Out-of-India? Looking forward to Davinsky's Stalin-orgel going loose on this study... Joshua Jonathan  -  Let's talk!  21:49, 30 July 2023 (UTC)
 * Their map is a combination of the steppe-model and the farming-model; see here. Hocus pocus. Joshua Jonathan  -  Let's talk!  04:34, 31 July 2023 (UTC)
 * The Imgur link gives a 404 error for me. --Florian Blaschke (talk) 17:26, 2 August 2023 (UTC)
 * Davidski's commentary: We're dealing with a bunch of [insert preferred insult here]. Joshua Jonathan  -  Let's talk!  17:47, 12 August 2023 (UTC)
 * Davidski is an absolute idiot who doesn't believe in any scientific research. 204.18.231.97 (talk) 03:46, 13 August 2023 (UTC)
 * @IP (or @MojtabaShahmiri, it's you, no?): I rarely feel the urge to agree with you, but this time I concur that certain amateur voices simply don't need even to be mentioned here in a talk page when it comes to the assessement of a linguistic phylogeny. Whatever comes out from linguistic research needs to be evaluated as such and not from a dogmatic POV that only can handle linguistic data when it provides a one-to-one match with population genomics. –Austronesier (talk) 18:22, 15 August 2023 (UTC)

Criticism
This article contains scholarly criticism of Heggarty et al. (2023). Joshua Jonathan -  Let's talk!  20:57, 5 March 2024 (UTC)

"Hypothesized" in opener
Given the state of the evidence right now, does it make sense to call the Indo-European migrations "hypothesized"? It seems like a strong consensus around major migrations has developed since we started to get lots of autosomal DNA evidence a little over a decade ago. Obviously there are still many details, some major, to be worked out, but are there any real competing hypotheses still out there?

Even if the term isn't wrong per se in this context, to the average person "hypothesis" means something like "educated guess." Just think about how much of a field day Creationists have had with the ridiculous "evolution is just a theory" argument.

I just made an edit to a similar effect on the Bantu expansion page. DuxEgregius (talk) 22:24, 29 March 2024 (UTC)


 * I don't think it's necessary in either case. Even with DNA evidence, reconstructions of historical events are always hypothetical, that's how history and archaeology work. Remsense  诉  02:34, 30 March 2024 (UTC)
 * Good point. I think in both cases it's a relic from when the evidence was less conclusive. DuxEgregius (talk) 08:00, 30 March 2024 (UTC)
 * As an archaeologist, I'd push back rather strongly against the idea that we (or historian and geneticist colleagues) cannot offer anything more certain than 'hypothetical' reconstructions of events. I don't think that's in line with mainstream thinking on the philosophy of archaeology and other palaeosciences, at all.
 * On the specific point, I think it's still appropriate to describe these as "hypothesized" migrations. aDNA has proven that there was substantial gene flow from the Eurasian steppe outwards c. 4500 years ago. Whether that gene flow is the result of the specific form of human movement implied by migration, as well as to what extent it was associated with Indo-European languages, is still very much up for debate. –&#8239;Joe (talk) 08:11, 30 March 2024 (UTC)
 * I think we believe the same thing but our public use of words have different boundaries. I'm equally happy retracting my point and just saying "model" or "theory" instead of "hypothesis" at any rate. Remsense  诉  08:49, 30 March 2024 (UTC)
 * As a historical linguist, I don't think that "hypothesized" puts our readers on the wrong track, even when understood in the "popular" sense" of the word. We could safely remove "hypothesized" if this article was entitled "Bronze-Age pastoralist migrations", but unfortunately, it carries the linguistic term "Indo-European" in its title. Archaeogenetics has uncovered a lot of rapid gene flow in certain parts of Eurasia during the Bronze Age paired with the spread of technology, subsistence methods and cultural practices which can well be labelled migrations (especially in the steppe, but less so in Central and SE Europe, where intensity and speed of steppe-related gene flow is compatible with less dramatic scenarios of population shifts).
 * But the association of these spatiotemporally manifest events with the expansion of the Indo-European languages is by its very nature hypothetical, and most likely will remain so until the invention of time machines or devices that allow to reconstruct sound waves at any time and place in history. The only records of Indo-European languages in the Bronze Age come from Ancient Greece and Anatolia, which means that for the most part, the Indo-European migrations happened behind the veil of the literary record. Yes, linguists have developed very sophisitcated methods to probe into the past way beyond literacy (I myself work in an area with minimal literary records and therefore heavily rely on such methods), but matching the findings of this methods with manifest archaeological and biological (archaeogenetic, palaeobotantical etc.) data is always hypothetical—ranging from "speculative" to "highly plausible". –Austronesier (talk) 12:06, 30 March 2024 (UTC)

Massive new paper out on the origins of Indo-Europeans
By Harvard https://www.biorxiv.org/content/10.1101/2024.04.17.589597v1 David Anthony himself is co-author, is this pointing to abandoning the Kurgan model as mainstream? This seems to be the mainstream now, endorsed by the major genetic labs and Anthony himself, at least regarding the very first expansion of IE 2A02:85F:E0D4:3F00:A0BA:B4E2:FF3E:2B0 (talk) 12:20, 26 April 2024 (UTC)


 * No, it's definitely not an abandonment of the Kurgan-model, more a modification. Joshua Jonathan  -  Let's talk!  14:32, 26 April 2024 (UTC)

Armenian - edits 17 May 2024
Do you have a source for the statement The Hayasa-Azzi confederation is considered by some to have spoken Proto-Armenian. And for the comment that the ‘most prominent’ view is the third? Sweet6970 (talk) 12:08, 19 May 2024 (UTC)


 * Yeah,
 * Petrosyan, Armen (2007). "The Problem Of Identification Of The Proto-Armenians: A Critical Review". Journal of the Society for Armenian Studies. p. 55.
 * "The Hayasa hypothesis has been criticized for it's proponent Kapantsyan's unacceptable linguistic approaches. In later (post-Kapantsyan) versions, it is in fact the only hypothesis widely accepted by competent scholars."
 * Criticism of the first view can be found on Armeno-Phrygians and Armeno-Phrygian languages. Sakaiberian (talk) 13:20, 19 May 2024 (UTC)


 * Thank you for providing the source. Sweet6970 (talk) 11:04, 20 May 2024 (UTC)