I continue to find thought-provoking nuggets in How to Read the Bible
As with the story of Cain and Abel, the Tower of Babel narrative has, for modern scholars, certain clearly etiological elements: not only its explanation for the name Babel, but also its accounting for the dispersion of peoples across the Near East and the replacement of an originally single, common language by an array of different, mutually incomprehensible idioms. Behind this latter element, too, modern scholars see a message not about the world as a whole, but something rather more local and specific. Semitic languages all appeared to be related: any native speaker could tell that Babylonian and Assyrian and Aramaic and Hebrew all had common roots and expressions, but a speaker of one tongue would not necessarily understand much of what was being said in the others. It is this reality, rather than the existence of different languages per se, that the story seems out to explain: all the peoples of the ancient Near East did, it says, originally speak the same language, but that unity was destroyed quite intentionally by God.
Of course, once Sumerian had fallen into desuetude, all languages spoken in the area would have been clearly related, and the story makes much more sense.
I’m curious if anyone here has an opinion on Gordon Whittaker’s Euphratic theory. He argues for an Indo-European substratum in Sumerian. One of his lines of reasoning is that the few polysyllabic words in Sumerian are probably borrowings. He sees IE roots in many of those polysyllabic words.
The people who wrote the OT would have been keenly aware of Persians, who spoke an IE language. Their ancestors warred with the Philistines, another Indo-European people. They must have been conscious of Greeks and Hittites as well.
I’m curious if anyone here has an opinion on Gordon Whittaker’s Euphratic theory.
Huh. I’d never heard of it; here‘s his 2008 paper (pdf) if anyone wants to check it out.
I wonder if Egyptian seemed obviously related to speakers of Semitic languages.
I’ve read a couple of Whittaker’s papers on Euphratic. I didn’t see any obvious signs of crackpottery in them, but as an amateur I’m not really qualified to judge how strong his arguments are.
Thinking geographically, I don’t have a problem with an Indo-European substratum to Sumerian. After all, the topography of the region suggests that the route of the Indo-Europeans as they traveled to India passed fairly close by. They would also have been prevented from moving westward from within that region by the desert. The Semites encroached from the Arabian desert and the Fertile Crescent. Which still begs the question: whence the Sumerians?
There are obviously different theories as to how old different bits of the OT are (and older bits may on some accounts have been repurposed into newer contexts by mysterious hypothesized redactors), and one presumably needs to take sides on various open controversies in order to be confident what other languages the authors/editors/redactors/etc of the particular passage would have necessarily been familiar with. Certainly as time went on and OT-readers came into closer contact with more and more non-Semitic languages (Greek, for example), I don’t think that made the Babel account seem any *less* plausible. But are there classical sources who actually seem to have figured out that e.g. Persian and/or Gothic is related to Greek and/or Latin in a way that the Semitic languages or Egyptian (which I think used to be “Hamitic” until not too many generations ago) are not?
Wherever the homeland was located, the Indo-European route to India was, by all likelihood and archaeological evidence, north and east of the Caspian sea. And an Indo-European substratum (or ad- or superstratum) in Sumerian would have been much earlier anyway, at a level comparable to Anatolian.
It’s not an issue I know anything about, but the wiki-discussion of the non-Semitic origin of the original Philistine language (and the specific claim it is IE) is very cautious/hedged in its wording: http://en.wikipedia.org/wiki/Philistine_language. My summary – we really don’t know, although the speculative theory of IE origin has yet not been affirmatively shown to be BS. And whatever the Philistines originally spoke got progressively semiticized over time, so you’re again back to the issue of what century B.C. you think is relevant for the composition/editing/redaction of the Babel account. (If at the other extreme you took the old-fashioned position that the text of Genesis was fixed no later than the death of Moses, the Philistines haven’t even really come onstage yet.)
Anatolian
Not too hard to imagine Indo-Europeans of the Anatolian persuasion drifting down the Euphrates. How does that work chronologically with respect to the Sumerians?
John Cowan is quite right (naturally) about Egyptian.
Although Egyptian is Afroasiatic it’s far from obviously related to the Semitic languages, and it seems unlikely that the relationship would have been apparent to any Canaanites interacting with Egyptians in any plausible timeframe that any of the Bible came from. We’re talking about something like the difference between English and Urdu, at least.
On the other hand it’s true enough that the Semitic languages at pretty well any stage BC were all close enough that anybody curious about languages doubtless would have noticed something. But Canaanites must have been in contact with speakers of non-Semitic languages at pretty much any time you care to pick, whether Egyptian, or Hittite or Persian or whatever.
And I don’t really buy the thesis that this has much bearing on the Babel. The whole point of it is surely about languages ending up different, not recognisably similar. It’s pretty much the weirdest thing about human language as opposed to animal signalling, after all, and something which is much more obvious than any similarities – we all talk, but we use completely different words for the same things. If you’re going to mythologise, that would be what you would want to explain with your myth. Only a linguist or a modern Bible scholar would imagine that people would naturally imagine the similarities of languages needed an explanation.
Classical ideas about relatedness of languages: Varro noticed similarities between Latin and Greek, and drew the culturally natural but wrong conclusion that they showed that Latin was in fact derived from Greek, or at least that the obviously related forms were. I doubt whether the cultural preconceptions of the time would really have allowed any serious thought about whether barbarian languages might be related to Latin and Greek; but in any case the whole notion of “related languages” in the modern sense was absent, like modern notions of language change. Even Dante supposed that Latin was a sort of interlingua and that his own Tuscan had always been the vulgar language of Tuscany, in a diglossic relationship with Latin.
As a further illustration of just how far the ancient Greeks were from our understanding of language change and language relatedness, there’s the familiar story from Herodotus of the Pharaoh Psammetichus trying to find out which was the original human language by raising two children in linguistic isolation to see what language they would speak. In the story He concluded that the oldest language was Phrygian because the children called bread “bekos”, which was Phrygian for bread. Apparently.
Their understanding, in a nutshell, was no more advanced than that of a modern journalist.
Still, some Greeks may have figured it out in Hellenistic times, even if the fact didn’t stick in the world’s memory. After Alexander there must have been plenty of L2 Persian speakers among the Greek soldiery, and it wasn’t beyond the power of a Greek philosopher (if he didn’t happen to share Aristotle’s views on the complete worthlessness of all barbarian culture) to see a few similarities between the Greek and Persian languages of his day.
In de Camp’s historical novel An Elephant for Aristotle the (historical) Greek philosopher Pyrron uses such resemblances to argue that Hellenes and Persians were once a single people, and therefore marriages between them are properly endogamous. The story’s young lovers find this convincing, but their parents do not, until all is saved by Alexander’s edict of -323 legitimizing marriages between Hellene soldiers and non-Hellene women.
The complexity of Greek intellectual life kept increasing for several centuries after Herodotus, peaking in late Hellenistic times. 99.9% of pagan Greco-Roman literature is lost. I don’t think it’s silly to imagine that some Greek author could have stumbled upon sound correspondences and discovered the IE family, and that all record of this was subsequently lost. I think that the ancient discovery of heliocentrism, for example, only came down to us through a few off-hand remarks.
A critique of Whittaker, and others, here.
A good find, thanks. I’d say Rubio disposes convincingly of the IE substrate hypothesis, along with some other poorly grounded ideas. I love the end of his first footnote:
LH, I think Christian’s life’s work deserves to be presented in fuller form. The previous sentence said:
.
I am only surprised not to see Dravidian in this mix.
“and older bits may on some accounts have been repurposed into newer contexts by mysterious hypothesised redactors”: well, that certainly raised a smile here.
“Thinking geographically, I don’t have a problem with an Indo-European substratum to Sumerian. After all, the topography of the region suggests that the route of the Indo-Europeans as they traveled to India passed fairly close by. ”
There’s a problem with the timing of this that makes it irrelvant to Sumerian.
“I am only surprised not to see Dravidian in this mix.”
They were still up at the North Pole.
Dravidians at the North Pole: yes, of course! How could I forget.
Dravidians at the North Pole: well at least they got there, which is more than Admiral Byrd did, apparently.
Of course, when the Dravidians went there the North Pole was hot. Then there was global cooling and they had to go south.
English and Urdu are so different that a phrase “My name is David” is translated as “Mera naam David hai…”
You’ll have to be a linguistics genius to recognize similarity 😉
SFR, Seeing a short sentence written, or spoken slowly, using known cognate words and keeping them carefully separated, is one thing. Another is to hear them in a conversation between native speakers, where the cognate words, especially spoken without a break, might be unrecognizable. For instance, it is not obvious that “meranaam” should be analyzed as “mera naam” rather than “me ranaam”. Occasional resemblances of vocabulary between any two languages is a common occurrence even if they are unlikely to be related.
Nevetheless, many people quite removed from modern linguistics noticed similarity of languages and correctly speculated about their relatedness.
I recall reading about Turkish officer during siege of Vienna in 1683 who noted in his diary that Germans speak language close to Persian. His conclusion was probably based on something very similar
Mein name ist David.
Name man David ast.
From the Wikipedia:
“In 1583, Thomas Stephens, an English Jesuit missionary in Goa, in a letter to his brother that was not published until the 20th century,[4] noted similarities between Indian languages, specifically Sanskrit, and Greek and Latin.
Another account to mention the ancient language Sanskrit came from Filippo Sassetti (born in Florence in 1540), a merchant who travelled to the Indian subcontinent. Writing in 1585, he noted some word similarities between Sanskrit and Italian (these included devaḥ/dio “God”, sarpaḥ/serpe “serpent”, sapta/sette “seven”, aṣṭa/otto “eight”, nava/nove “nine”).[4] However, neither Stephens’s nor Sassetti’s observations led to further scholarly inquiry.[4]”
Noting similarities between languages (especially in individual words) is one thing. Formulating a hypothesis about the origin of such similarities is another. Communicating the result to others with the background necessary to evaluate the hypothesis and pursue its implications is yet another.
Indeed, in the past, many people, especially travellers likely to have acquired at least a functional knowledge of several languages, noted similarities between their own language and others spoken far away (as with Italian and Sanskrit), but did not reach the conclusion that they had a common origin (at the time, any common origin was sought either in Hebrew or in another known language). The Turkish officer who found German “close to Persian” (certainly not close to Turkish) did not necessarily think that the two languages were ‘related’ in the sense that we use today. In any case, single words as found in isolation are not enough to reach this conclusion, since the occurrence of similar words often results from borrowing rather than genetic relationship (as between English and French, for instance). Sir William Jones, considered the founder of historical linguistics, was a scholar familiar with several languages, and his claim to fame does not rest merely on his observations of the resemblances in individual words and in the forms of these words: his critical contribution was the hypothesis that Sanskrit, Greek and Latin must have had a common ancestor “which perhaps no longer exists”. The next step was for scholars to try to figure out what this hypothesized common ancestor, later dubbed “Proto-Indo-European”, was like, based on the comparison of existing languages likely to be related.
People untrained in comparative linguistics but who have heard of languages being ‘related’ tend to speculate about specific languages on very flimsy evidence. When working in a native community I was told by some people that their language was “familiar with French and German”, and a teacher of Dutch origin, having heard the sound [X] (as in Dutch g), asked me if the local language was related to Dutch.
M-L,
“Noting similarities between languages (especially in individual words) is one thing. Formulating a hypothesis about the origin of such similarities is another. Communicating the result to others with the background necessary to evaluate the hypothesis and pursue its implications is yet another.”
Indeed they are. Case in point is l’affaire Penutian. I wonder if you are the same Marie-Lucie, and I bet you are, who is mentioned repeatedly on the talk page of the Wiki on Penutian. If you are you in a position to have an opinion on something I have been wondering.
This is what I have been wondering is the disparity between they way Africanists, not just Greenberg, are such willing lumpers they they accept the reality of Niger-Congo, whereas Americanists are so persnickety that they hesitate to accept Penutian. Is this really just the case of not having a standard metric to access these relationships that this appears to be?
I get that WRT to Penutian the history is deep and the case is pretty cold; the evidence has perhaps passed beyond the “data horizon’ into the black hole of time. And I understand that after so many centuries of living cheek by jowl with each other, and with other clearly unrelated languages, the sub-groups are going to have so many adstrate effects as to compromise the data hopelessly. But how is the situation in Niger-Congo any different?
Jim, yes, I am that person. Once in a great while I look at the Penutian page on Wiki, but I find joining discussions very frustrating. If they are talking so much about me, I should take another look!
I can’t comment on what Africanists do or don’t do, as I know very little about African languages, but it looks like they and Americanists have quite different backgrounds and types of training. In my opinion, the Penutian question is far from being dead, but few people have addressed it productively. I am also at fault in not seeking more avenues of publication. The Wiki page shoulde include a reference to my 1997 article, most of which not only is still valid but keeps being reinforced by my further research. The appendix does need revision. If you are interested in further information and discussion, send me a private message on Facebook.
Will do, M-L, if I can figure out how.
The whole question interests me as a Californian and as a Westerner in general. And all the related questions interest me: the question of possible Penutian-Uto-Aztecan affiliations, the historical question of Uto-Aztecan displacement of Penutian in the Great Basin, Penutian interactions with Hokan groups (presuming there is any reality to that grouping.)
That answer about the disparity of criteria for grouping is about what i suspected was the case. Resolving that is a task for another generation, if one comes along that is not pied-pipered by theoretical fantasies.
Well, I am certainly not “pied-pipered by theoretical fantasies” and I am pretty sure of having answers to some of the questions (at least linguistically; I leave historical interpretations to the historians). As for Hokan, it has no relation to Penutian except that some of the languages are neighbours and there may be borrowings or other areal influence. The term “Hokan-Penutian” does not refer to any kind of relatedness, it only refers to a conference series which included papers about both of these groups (separately).
p.s. I suggested facebook because so many people use it. If you don’t, you probably have friends who do. In any case, you can sign up (free) without having to reveal much about yourself, and then you can search for my name and leave me a message which I will answer.
p.p.s. If you include your email address in the message, I will reply to that.
Some of Marie-Lucie’s papers are available online:
“Ergative and accusative: A single representation of grammatical relations with evidence from Nisgha” (Working Papers of the Linguistic Circle 1982, freely available PDF). I found two generative-based responses to this by Robert S. Belvin: “Nisgha syntax and the ergativity hypothesis” (M.A. thesis 1984, freely available PDF), “Ergativity and Accusativity in Nisgha Syntax” (Berkeley Linguistics Society 1990, freely available PDF).
“Morphophonemics of Nisgha plural formation: A step towards Proto-Tsimshian reconstruction” (Kansas University Working Papers 1983, freely available PDF).
“Morphophonemics of Nisgha plural formation: A step towards Proto-Tsimshian reconstruction” (Working Papers of the Linguistic Circle 1983, freely available PDF).
“From the Nisgha speaker’s point of view: the evidential postclitics” (WPLC 1984, freely available PDF).
“The Evolution of the Nisgha Counting System: A Window on Cultural Change” (1983, WPLC, freely available PDF).
“Tsimshianic and Penutian: Problems, methods, results, and implications” (IJAL 1997, can be put on personal JSTOR shelf)
“On the eve of a new paradigm: The current challenges to comparative linguistics in a Kuhnian perspective” (1999: Historical Linguistics, Google Books partial view)
“From Erosion to renovation: The emergence of the Coast Tsimshian preposition da“ (2008, New reflections on Grammaticalization 4, freely available PDF).
I did not expect to see (or post) this list here! These papers present parts of my work. Some of them are very old and no longer represent all my views on the language(s). For those tho might want to read them, I have a few remarks.
Information contained in papers on the Nisgha/Nisga’a language was incorporated or updated in my grammar of the language (1989). For the two 1983 papers on morphophonemics, the WPLC one was a course paper, which I revised and expanded for the Kansas article. Most of the first half of the paper still stands, but the second part, dealing with some irregular forms, would need updating in the light of comparison with other members of the family. The one on the counting system would also need the same type of updating.
After I wrote the 1982 paper on ergativity, Belvin wrote an MA thesis (later revised as the Berkeley paper) disputing my results because they did not agree with Marantz’s generative description of ergativity. I still do not agree with Marantz (or Belvin, whose work also contains some errors of data). (I had not seen the 2nd paper until now). When I researched works on ergativity, it was obvious which authors had actually studied languages with this feature and which of them had only done some picking and choosing in grammars of such languages, never getting a “feel” for the feature. Marantz was in the second group.
The Penutian connection, first proposed by Sapir, at first seemed to me very unlikely but ended up becoming my major area of research in 1995, along with my ongoing reconstruction of Proto-Tsimshianic (based on the four varieties in the family). The 1997 article is still valid except for a couple of minor revisions (the entire journal issue deals with Penutian languages). My contribution to comparative Penutian research consists mostly of morphological comparison of Proto-Tsimshianic with as many as possible of the languages and families usually included under the name Penutian. Since the article was written I have found a few additional morphemes, but have not had to remove any from the 1997 list. On the other hand, the appendix (which gives a number of apparent lexical correspondences) would need revision in the light of my more recent research.
Another paper that is very important for the Penutian connection is not mentioned here but should be available online: “Tsimshianic l-initial plurals: a relic of an ancient Penutian pattern” (2002, Berkeley Ling society).
The 1999 paper deals with my views of the situation of historical linguistics. Things have not changed much since I wrote it, although some generativists have started to admit that historical phenomena might be relevant.
The last mentioned document is a one-page abstract for a paper that I was scheduled to present at a conference that I was unable to attend. I presented a new version in 2012.
Here’s the l-initial plurals paper (freely downloadable PDF). I ran Google Scholar searches to find the papers and their links. If you don’t want the list to be here, I’m sure Steve will take it down.
The second Belvin paper apparently exists in both 1990 and 2012 versions, but the Google Scholar links are identical. I found yet another paper of his on Nisga’a: “The causation hierarchy, semantic control and eventivity in Nisgha” (ASJU 1997, freely downloadable PDF).
Hat: If m-l is okay with keeping the above list here, could you fix up the typos and markup errors? Thanks. (It’s things like this that really make me feel the lack of preview.)
It’s very easy for unsophisticated people to come up with ideas of the relatedness of languages. I recently met an Inner Mongolian who spoke good Japanese and was fascinated with English because of the supposed similarities between English and Mongolian. One of the examples cited was ‘ideh’ being similar to ‘eat’. It’s nice when people take an interest in other languages, but really, if your basis for doing so is based on a couple of random observations on vocabulary similarities, surely it would be more intelligent to try and find out what has already been researched about the topic than just hazard guesses in the dark.
Hat: If m-l is okay with keeping the above list here, could you fix up the typos and markup errors? Thanks.
Done; let me know if I missed anything.
surely it would be more intelligent to try and find out what has already been researched about the topic than just hazard guesses in the dark.
Well, of course it would be. It would be more intelligent to look before one leaps, too, and not marry until one was ready. We humans are capable of acting intelligently but usually prefer to hazard guesses in the dark. What are ya gonna do? As Stalin allegedly said about authors, I have no other humans to offer you.
@Jim:
I think the reason why Africanists have been much readier to adopt Greenbergian lumpism than Americanists have is due to a number of things:
1. The level of documentation of American languages, including by some of the greatest descriptive linguists ever, is much greater than for all but a handful of African languages.
2. Greenberg’s classification was a real advance at least on some of its predecessors inasmuch as it is at least based on entirely linguistic criteria, rather than vague racial or cultural principles. Mind you, its less controversial parts tend to be the less original too.
3. A lot more African languages than American probably really are related in a potentially scientifically demonstrable way.
Cynically, I think (1) may be the most significant. As far as I can make out, few experts in the actual “Khoi-San” languages buy into Greenberg’s notion that they really are a genetic unity; “Nilosaharan” is pretty much on a level with Altaic in terms of rigour.
That really leaves Afroasiatic and Niger-Khordofanian.
Afroasiatic is a sort of special case as probably the deepest (in terms of probable time depth) language family accepted by non-fringe historical linguists, due to the fact that it includes most of the earliest languages ever to be written, and that the languages often show extremely unusual features like the ablaut-on-steroids of the Semitic branch which seem to go way back and mean even in the absence of much in the way of real reconstruction it seems pretty certain we’re looking at a genetic group.
Niger-Khordofanian …
I know one Gur language, Kusaal, fairly well. It is obviously closely related to some of the neighbouring languages, like Moore, Dagbani, Mampruli and Dagaari, to a degree which is apparent to any inquisitive native speaker. (Interestingly, these resemblances are accounted for in terms of supposed common descent of the speakers, much as seems to be going on in Genesis with some though not all of the “Shemitic” groups.)
If the documentation of these languages were better, it would be straightforward to reconstruct a protolanguage. As far as I know this has never been done – what work I have seen makes elementary errors, for example in reconstructing an entire additional contrast of alveolars vs dentals because the conditioning factors for the development of the reflexes in the actual languages have been overlooked.
This is the sort of thing which ought to be the *beginning* of setting up Gur as a linguistic genetic group and it hasn’t been done. When you start comparing these languages with remoter “Gur” languages like Kasem it soon becomes apparent that the Gur group is on a level with Indoeuropean in terms of diversity, not Germanic or Slavonic. Senoufo, generally called Gur, is remoter still.
So even individual subgroups of “Niger-Khodofanian” are on a level with Indoeuropean; in other words, the groups as a whole is a pretty long-range hypothesis. Parts of it are nevertheless quite plausible – for example the Gur noun class system does look very much as if it is related to the Bantu. Others very much less so – the evidence that Mande belongs would never convince a typical sceptical Americanist, for example.
LH: I had not intended to splash my work over the blog, but since those articles have already been mentioned, and my remarks too, they might as well stay.
Americanists vs Africanists: Among North American languages, several families are very well established, the major one being Algic (comprising the huge Algonkian family which includes dozens of languages, plus two isolated distant relatives on the California coast: Yurok and Wiyot). Compared to Algonkian, spoken over an extensive, largely flat territory in the US and Canada), Penutian is still considered a hypothetical group of languages spoken in varied territory between the Pacific Coast and the Rockies, squeezed between several other language families, and mostly quite distinct from each other. In terms of comparative and historical study, the Algonkian work started with a very competently trained linguist (Bloomfield), while the Penutian work has been pursued mosly by linguists well trained in descriptive methods but minimally in historical methods.
As for the reception of Greenberg’s work on American languages, he largely accepted other linguists’ proposed classifications, except that he lumped together all the languages which did not fit into the two Northernmost families (Eskimo-Aleut and Na-Dene, which are each very distinctive) and called the rest Amerind. While Greenberg had some valid criticisms to make about some of the current historical work (especially – reading between the lines – on California languages), his own methodology was VERY questionable and his conclusions inacceptable. My feeling is that the currently very large number of American families will eventually be reduced through regrouping, and instead there will be several groups of intermediate size (on the order of Indo-Euopean, which is a family of language families). Penutian should be one of those groups (probably with some redefining of its overall membership and internal subgroups).
M-L,
“As for Hokan, it has no relation to Penutian except that some of the languages are neighbours and there may be borrowings or other areal influence.”
It even feels different. When you are looking at Atsuge and Klamath-Modoc, you know you are in very, very different langugaes, especially structurally. In fact given the time depth of the contacts, it’s surprising how distinct they stayed.
John and David, thank you very much for all of that, especially M-L’s papers!
David,
” for example the Gur noun class system does look very much as if it is related to the Bantu”
That sounds convincing just on the geography, not much chance for contact. But yet.
All these N-K groups have been in contact for God kknows how long. And God only knows what their settlement and migration hisotry is. The expansion of the sahara is bound to have crammed languages into contact, but it’s speculation to assume all the movement was only southward. And really this was a main argument against retaining Pentutian, as I understood it.
And those first two reasons sound like arguments from silence to me. And the third one, it seems to me, fits for Penutian as much as for N-K. It hangs on overt evidence, but that evidence can’t disprove contact influence.
Afroasiatic – yeah, that system of derivation you see in all of them is probabaly unique in the world.
I guess it really comes down to what constitutes a proto-language and how that fits with creolization and other contatc effects – and what bearing that really has on how this or that actual language came to be – what is the definition of relatedness, looking at a language as a system rather than at the history – common ancestry – of the community speaking it.
One final thing – I wonder – if Penutian does end up being demonstrated, that may provide tools for loking at all these long range proposals rather than just dsmissing them as unprovable.
I had not intended to splash my work over the blog, but since those articles have already been mentioned, and my remarks too, they might as well stay.
I’m glad you feel that way; I appreciate your desire not to cause a splash, but I think it’s good to have your work available in this convenient fashion, and there’s such a thing as excessive modesty!
I’d like to thank everyone involved in the discussion of North American and African language families; I’m learning a lot (though I probably won’t retain much of it).
“LH: I had not intended to splash my work over the blog, but since those articles have already been mentioned, and my remarks too, they might as well stay.”
I am very grateful that John went and found this for me. And I am grateful for your work. I understand good manners and reticence, but this work represents a real contribution and it should be shared as widely as possible. I think Mary Haas would be pleased.
“Americanists vs Africanists: Among North American languages, several families are very well established, the major one being Algic (comprising the huge Algonkian family which includes dozens of languages, plus two isolated distant relatives on the California coast: Yurok and Wiyot). Compared to Algonkian, spoken over an extensive, largely flat territory in the US and Canada), Penutian is still considered a hypothetical group of languages spoken in varied territory between the Pacific Coast and the Rockies, squeezed between several other language families, and mostly quite distinct from each other.”
If somehow Indo-iranian had all gone extinct we would be saying the same thing about IE versus Turkic.
“In terms of comparative and historical study, the Algonkian work started with a very competently trained linguist (Bloomfield), while the Penutian work has been pursued mosly by linguists well trained in descriptive methods but minimally in historical methods. ”
Ain’t that the truth. a grounding in IE or Chinese research would help train people. Chinese in aprticualr, wiht the way consonants get treated and the way the writing system is so ambiguous wrt to phonetics, would be like Ranger School.
Just wikipedia’d for Afroasiatic and found this gem:
There are two etymological dictionaries of Afroasiatic, one by Christopher Ehret, and one by Vladimir Orel and Olga Stolbova. The two dictionaries disagree on almost everything.
That’s absolutely wonderful.
In a sense, it’s not surprising. Proto-Semitic must be dated as far back as Proto-Indoeuropean, after all, and that’s just one branch of the tree. Old Kingdom Egyptian from the 3rd millenium BC is less like Protosemitic than modern Aramaic dialects are. So any putative proto-Afroasiatic would be likely to date back so far you wouldn’t realistically expect to be able to identify much shared vocabulary any more. Closer to the last Ice Age than the present.
As I said, Afroasiatic generally is thought to be a real thing even so, on the basis of some shared, often fairly unusual morphology which seems very unlikely to be due to borrowing, like the prefix-conjugation in y+ 3sgm t+ 3sgf n+ 3pl etc, feminines in +t, 2sg possessives with suffixed +k, abstract nouns with prefixed m+, plural forms with infixed -a- and so forth. All this turns up in a language as remote from Hebrew and Arabic as Hausa.
What’s depressing about the dictionaries in question is the way they happily accept vast semantic and phonological latitude in the comparisons made. They are like the various Nostratic compilations that so much effort has gone into, and it says it all that they differ so much from one another.
With Niger-Khordofanian there isn’t even anything as compelling, it seems to me. Gur certainly has noun classes, like Bantu; but to start with, they are marked by suffixes instead of prefixes. I’ve seen it sweepingly asserted that Gur originally had prefixes and the suffixes are a secondary development from fused, agreeing, pronominal elements, but there are AFAIK no Gur languages with only prefixes, most show no trace of any noun class prefixes, and in the minority which have both, like Gurmanche, the synchronic position is transparently the other way round – the prefixes are more loosely attached and retain a deictic meaning. Presumably people have been simply looking at dictionaries and not bothering with grammar. No way to do historical linguistics.
In Essentialist Explanations, I find this: “Gur languages are essentially typical Niger-Congo languages, only with the nouns spoken backwards.” It’s attributed to me, which means I saw it in some source decades ago and don’t remember where.
I assume this is about suffixes rather than prefixes, but I can’t be sure.
I’ve seen it posited that Proto-Afroasiatic was the language of the Natufian culture of Palestine, the first farmers of ca. 10,000 BCE, and that it spread with agriculture and herding. I think this is plausible, afaict, in spite of the fact that Colin Renfrew is one of the people that believes it.
Afro-Asiatic: A friend of mine took courses in Ancient Egyptian, became interested in Afro-Asiatic, went to Nigeria to learn Hausa and do fieldwork on Chadic, and published a comparative word list of some Chadic languages, before the situation in Nigeria became too dangerous for further fieldwork. He does not have a high opinion of Ehret’s work.
It’s been truly said that Greenberg-style mass comparison can only be at best a way of suggesting relationships to be confirmed (or not) by later proper comparative work. Unfortunately in Africa his conclusions seem often to have been accepted as givens and (with some sterling exceptions) the proper bottom-up comparative work has yet to get properly off the ground.
The possibility of resemblances being due to large-scale borrowing between originally distinct language families doesn’t seem to have been much considered, which is a pity given that a lot of contemporary West Africa is in many ways one big Sprachbund, with things like: aspect being more prominent than tense in the verb, serial verb constructions, no case inflection for nouns, terracing tone systems and a huge amount of semantic congruence turning up again and again in languages with little or no close genetic connexion – unsurprisingly when a great many people are multilingual. There seems no reason to suppose that things were much different in the past.
I don’t think it’s inconceivable that a general feature of noun classes, with a close association of nouns in noun phrases with a preceding or following class-marking pronominal element, could have been an areal feature spreading among proto-Gur and other groups. The actual markers are all short, V or CV, there aren’t that many potentially different consonants involved, and there are so many different sets in the different Niger-Khordofanian languages that have noun classes that by pure chance you’d expect lookalikes. Human-plural “ba” turns up all over but apart from that the similarity is not so much in the details as in the basic concept of lots of noun classes with dissimilar singular and plural markers and some correlation of class and meaning.
I haven’t read Whittaker 1998, which is the subject of Rubio’s critique, a rather mild one by the standards of the other critiques in his paper. But I have read Whittaker 2008 now, and it doesn’t seem so unreasonable to me. I note that he draws the mysterious PE.TA.SI, discussed in the recent Sumerian thread, to late PIE *potis: see the section on Euphratean culture. That might explain why it was written so but pronounced differently.
Interestingly, Whittaker (2008) finds lots of feminines in his material. That would mean that “Euphratic” belonged to the non-Anatolian branch.
PA.TE.SI would be a perfect fit for *potis: lack of /o/ in Sumerian-as-understood-by-Akkadians, neutralization of /e/-/i/ distinction in closed syllables in Sumerian.
Huh. I like it!
Cross-linking the other Sumerian thread.
And of course it is PA.TE.SI; there’s something about my mind or fingers that wants to transpose the first two vowels. Ah well, I’ve been programming for forty years, and I still find myself typing reutnr for return on occasion.
I remember reading and discussing it a few years ago, probably in sci.lang. I seem to remember that it was well-argued and that the evidence was good enough to be interesting but not extraordinary enough to turn IE upside down quite yet. But it has been in the back of my head when thinking about Yamnaya, Maykop and cultures further south.
reutnr for return
blockquote
bvlockquote
blcokquote
blcokqtuoe
blockquote
blocvkqutoe
blockquote
blcokqutoe
blcokquote
blcokqtuoe
blcokqutoe
blcokqutoe
blcokqutoe
blcokqutoe
blockquote
blcokqutoe
blcokweutoe
blockquote
blcokqutoe
blockqutoe
🙂
From Steven Shapin’s LRB review of Scientific Babel: The Language of Science from the Fall of Latin to the Rise of English by Michael Gordin:
I used Google Translate on the Maori and got “Do you talk nonsense, I am tired of the tower.”
Just a note about PA.TE.SI:
There’s no evidence it was ever pronounced anything like [patesi]. It’s used to spell the word ensi(k) (no one’s sure how or if Sumerian /k/ was pronounced in a syllable coda). The first sign PA is probably a semantic element for an authority figure. On its own it can mean ‘overseer’, and it’s found in compound signs for ‘shepherd’ and certain administrative positions. The SI is a phonetic element that gives the last syllable of the word. I don’t remember what the story for TE is off the top of my head. This kind of mish-mash of phonetic and semantic elements is pretty common in compound signs.
Interestingly enough, as important and ubiquitous as the title is in later documents the word “ensi” doesn’t show up until the texts are unambiguously written in Sumerian. The sign list “Lu A” (which is first attested during the Uruk IV period (~3100 BC) and contains a variety of official titles that were already obsolete when scribes started noting Sumerian grammatical elements in their writing ~2800 BC) doesn’t list “ensi”. So if PA.TE.SI did come from a substrate loan, it was never used in the very earliest times when it’s conceivable that cuneiform was used to write something other than Sumerian (or Akkadian), and only shows up much later as the spelling of an unambiguously Sumerian word “ensi”.
“Shepherd” of all things?
You mean *pah₂-? 🙂
I get “Do not you talk nonsense”, which makes more sense, though whether the exact nuance is “Don’t talk nonsense” or “Stop talking nonsense” or just “You are talking nonsense” I don’t know.
Of course ‘shepherd’ = ‘overseer (of the flock)’ is a world-old metaphor: see English pastor.
I think David was remarking on the remarkable similarity of IE pa– to the first syllable of the alleged patesi.
when it’s conceivable that cuneiform was used to write something other than Sumerian (or Akkadian)
I wonder what the current state of thinking on this is. Totally a Sumerological dilettante here, but I thought the identification with Sumerian was held to be fairly secure, given sound correspondences between the phonetic and logographic values of the same signs. The only one coming to mind is the sign meaning “life” (TIL) also being used for the phonetic TI, but I know there were more examples. (Not that this is necessarily a clincher: Japan is a good example of a pseudo-phonetic script being repurposed to new phonetic uses in an entirely different language community from its inventors. And rebus is exactly the sort of thing that gets reinvented.)
Of course, signs or the idea of signs could have been taken over by the Sumerians before their development into a serviceable system of phonetic representation. Alternatively such passes of the baton could easily have happened a large number of times before cuneiform came to Sumerian speakers from its distant ur-inventors. And given the time depth, surely nothing should be ruled out. But if your comment is reflecting some more specific examples suggesting that there was a pre-Sumerian stage to the script’s use as a script, I’d be fascinated to know.
Both: the coincidence of that sound with that meaning.
a pre-Sumerian script
The invention of writing did not occur in a vacuum but is linked to a particular type of society necessitating the keeping of accurate records for purposes of large-scale receiving, storage, redistribution, taxation, commerce, and more, before writing was also used for recording speech as such (see for instance the uses of the “Minoan” script). What society prior to Sumer’s could have fulfilled these sorts of conditions?
The idea is that the society was more or less continuous, but its language changed – as it later did again (to Akkadian… and then to Aramaic, and then to Arabic).
But then where did Sumerian come from? Smacks of obscurum per obscurius.
Sure, but with Sumerian you can’t get rid of that question anyway.
“Where do we come from?
And why do some of us
speak Basque?”
– headline in Scientific American long ago, original line breaks
Anyway, most of the paper looks quite convincing to me, and not just to me.
It’s very interesting. My main problem is how to reconcile it with what we believe to know about PIE. For one thing, it’s very old. Old to be established well outside of the homeland (whichever you prefer), and old to show such late features. Whittaker’s contention that Euphratic patterns with northwestern IE could be an indication of its place near the bottom of the tree, for so does Tocharian. Anyway, there would certainly be implications for the reconstruction of PIE.
As far as I can tell, it (or most of it) would fit quite well somewhere around the base of the non-Anatolian branch. Where the homeland of that branch was is another question.
Yeah, but how did that have rime to establish itself as a substratum, no the substratum, in Sumerian?
Could have been just the preceding wave of barbarians coming down from the mountains.
Of course everything falls into place if PIE was spoken where Gamq’relidze & Ivanov hypothesized (eastern Anatolia/south of the Caucasus). But I haven’t read their book, so I can’t tell how much is left of their decidedly non-mainstream arguments after the most obvious problems are accounted for. Rumor even has it that the Caucasian Hunter-Gatherer contribution to the Yamnaya people was exclusively female…
One solution is to put PIE in the Golden Crescent, and let it move north through Majkop. That’s essentially G&I. Another is to move PIE back in time as the language of the first Trans-Eurasian trade network. If that originated in (or established) BMAC, it’s Johanna Nichols.
“Substratal” traces of enigmatic Indo-European languages sometimes pop up in unexpected places, cf, the Bangani substrate controversy (which turned into a vitriol-throwing contest between Claus Peter Zoller and his critics, George van Driem and Suhnu Ram Sharma). Before the discovery of Tocharian, who would have expected to find an outlying non-satem IE language in the Taklamakan Desert, of all places?
If the Euphratic substratum turns out to be real, the Tocharians might be less outlying, and one should start looking for lost Kentum languages all over the place. Maybe even reassess Bangani. I’d suggest an early 4th millennium horse-and-wheel-powered Kentum dispersal followed by a 2nd millennium Satem explosion based on the chariot. In this scenario, Yamnaya is not the origin of IE but the geographical hub of a long distance trade network using IE as its common language, taking advantage of technological innovations from across the continent.
Of course, this may well be the case even without Euphratic. Last year I stumbled upon this and this (never mind the date) on Michael Frachetti’s work on early pastoral nomadism.
(Hoi! Out of the moderation queue!)
Then again, IE might be bronze age Yamnaya taking over the established intercontinental network. If Burushaski actually is related to Yeniseian, they could be the Botai.
A millennium is a long time. Many innovations turn up all over the place almost instantaneously, The wheel is in the Balkans and in Mesopotamia in the mid-4th millennium. The horse is in Central Europe long before Corded Ware. Antedating the taming of the horse, or the wagon, or the potter’s wheel, or bronze metallurgy, or the ard, a few centuries could turn the dynamics of inner Eurasia upside down more than once.
Maybe even reassess Bangani.
I don’t think the Bangani substrate has been debunked. The latest (though by now quite old) news was that Professor Anvita Abbi went to the places Zoller had visited, collected more data from the local informants, and basically replicated Zoller’s results. I’m not aware of any further significant developments, so I tentatively accept Zoller’s data (as do several mainstream Indologists, to my knowledge)
Anvita Abbi on Bangani.
Exactly. See also Hans Heinrich Hock’s statement on the Bangani war.
Botai?
Moving it back in time seems increasingly difficult, now that molecular-style dating agrees with the mainstream. And the other hypothesis about the BMAC is that it was violently destroyed by Indo-Iranian invaders and could have been Dravidian-speaking.
I know. I’m playing with the implications of moving PIE back before Yamnaya, which follows from accepting Euphratic. If we also accept the Kentum substrate in Bangani, the natural conclusion to me is that IE expanded from the core at least twice, maybe several times. What sort of vector might have carried this first expansion? My suggestion is the spread of long distance trade slash pastoral nomadism.
The Botai is a second order hypothetical. If the continental trade revolution happened very fast after the taming of the horse, I think it could well have included the Botai and other steppe HG cultures before being taken over by Kentum IE.
As for Europe, the genetic evidence for a steppe origin is pretty convincing, but Is the genetic data fine-grained enough to discern between one or more expansions? Between fourth and third millennium?
Also, in this scenario, BMAC may be Kentum. Proto-Tocharian?
who would have expected to find an outlying non-satem IE language in the Taklamakan Desert, of all places?
Indo-European was traditionally divided into Western (kentum) and Eastern (satem) languages, but such a binary division is not the only possible one. In fact, later sociolinguistic studies showed that the spread (and spitting) of a language proceeds from the centre to the periphery of its territory. It is now well-known that the variety spoken in a capital city or other major cultural centre (which receives internal migrants from all over the country) is subject to many more and faster changes than remote regions (which lose inhabitants but rarely receive migrants and therefore are not in contact with speakers of other varieties). So indeed an early split taking some PIE speakers towards the East before significant changes occurred in the centre could have carried kentum unchanged to the East, while the Central PIE variety, that of speakers who remained in place, later developed the changes which eventually resulted in satem (in an area from which some speakers later migrated East).
The trick is that both kentum and satəm are innovations. Satəm is the combination of a merely phonetic innovation – /kʲ gʲ gʲʰ/ becoming various affricates or fricatives – with the phonological merger of /kʷ gʷ gʷʰ/ into /k g gʰ/. Kentum is the phonological merger of /kʲ gʲ gʲʰ/ into /k g gʰ/.
The Luwian branch of Anatolian is thus neither kentum nor satəm, and neither are Albanian and maybe Armenian: they have the phonetic innovation of satəm, but not the phonological one (or only in some environments).
What this all really shows is the danger of doing phylogenetics with one or two characters. You need more. Mooooore. I’ll soon write a grant proposal based on the fact that way too few characters (only ever less than 350 at once) have been used in my field!
And finally, cultural centers should not be confused with geographic ones. Remote, isolated regions often develop innovations, and sometimes these innovations spread later; when they do, they’re simply less likely to reach the entire area in question than an innovation from the geographical center is. An example of this may be the High German consonant shift, which came from the southern edge.
In fact, later sociolinguistic studies showed that the spread (and spitting) of a language proceeds from the centre to the periphery of its territory. It is now well-known that the variety spoken in a capital city or other major cultural centre (which receives internal migrants from all over the country) is subject to many more and faster changes than remote regions (which lose inhabitants but rarely receive migrants and therefore are not in contact with speakers of other varieties).
This also works only in some special circumstances. For example, the capital of the UK is not located anywhere near the centre of Britain. The “wave model” has its own limitations just like the “family-tree model”. Splits are more often due to migrations and communication barriers than to local innovations which can radiate to the surrounding dialects. “Indo-Europia” did not have a political capital, and if it had one or more “major cultural centres”, there is no evidence that the satem innovation was associated with one of them. Of course we know with hindsight that the satem/centum division was incorrect, but it was by no means clear to linguists in the first decade of the 20th century, and the discovery of Tocharian on the family’s easternmost periphery came as a great surprise to everybody.
The Luwian branch of Anatolian is thus neither kentum nor satəm, and neither are Albanian and maybe Armenian: they have the phonetic innovation of satəm, but not the phonological one (or only in some environments).
They are all “satemoid” in their own ways, but with important differences. For example, in Luwian and Albanian (and possibly in Armenian) the velar and labiovelar series show different reflexes at least in some positions. There is no evidence, in those branches, of the RUKI treatment of *s, which is another non-trivial innovation shared by Balto-Slavic and Indo-Iranian (actually more characteristic than the “satem” shift itself).
Yes, satemization has happened several times, with different outcomes, and it’s a good idea to limit the designation to II and BS.
I don’t think the kentum expansion hypothesis needs a political center. I certainly didn’t mean to suggest more than a geographical meaning to the word. Well, maybe economic or cultural, but then as a secondary effect of being the geographical hub.
Thank you for your various comments on “centre” and periphery. I meant “centre” mostly in a geographical sense, and even then, not as an absolute requirement: there can be regional centres as well, becoming local cultural poles for various reasons.
An exemple is the current regionalization of English in North America. It was thought some decades ago that existing regional differences would be ironed out, especially by the sharing of movies and TV in the US, but instead the continent is linguistically fragmenting into several major areas and there is no common centre from which innovations might be radiating outward throughout the continent or even just the US.
I wonder if the “kentum” words in Bangāṇī have a Tocharian connection.
existing regional differences would be ironed out
Many of them have been. The old dialect boundaries (mostly in vocabulary) have been reinterpreted or re-expressed as accent boundaries, and mostly phonetic ones at that (only a few new splits and mergers here and there). The various short-vowel chain shifts, though phonetically substantial, are literally superficial. So while there are more varieties than there used to be, they are not nearly so different from one another.
And I don’t think we should paint a uniform picture of accent divergence either. For example, here in central Massachusetts my peers are much more likely to be rhotic near-General American speakers compared with their parents or grandparents who lived in the same places, and I think the same holds true at least for some social segments in the NYC area. And in the South there’s been a leveling trend, with traditional local accents losing ground to a generalized Southern pattern. That isn’t to discount things like the Northern Cities and California shifts, but overall I think it’s a bit of a wash.
Just so. The Eastern Mass accent is retreating eastward toward its core, and the NYC accent, which never had much of a periphery, is turning toward more national forms and mixing with AAVE.
I was going by what I had read, you Americans know best!
That’s what we keep telling the world!
John Cowan, Lazar, Marie-Lucie…William Labov argued back in 1987 that Black and White English in Philadelphia were diverging, not only phonologically but also grammatically. Young black speakers, he found, were using such things as habitual BE far more than their grandparents, and their speech also exhibited some remarkable innovations, with the verb ending /s/ turning into a marker of indirect speech, for instance. He saw this state of affairs as a consequence of growing DE FACTO segregation and expressed the fear that the teaching of standard English to Black speakers would become more of a challenge as a result of these innovations.
If we consider socio-economic trends in the United States over the past thirty years, with a sharp decrease in social mobility and increasingly rigid racial and class lines, I strongly suspect that as a whole the story of American English is now one of divergence more than one of convergence. Unfortunately this reality is unlikely to be explored by linguists in academia too closely, if the past is any guide. The Labov study I refer to above triggered a firestorm of controversy, and my impression is that if a less famous linguist had submitted the same article, with the same data and conclusions, it would never have been published.
I can only report what I hear as a NYC resident, which is that young white New Yorkers are phonologically, lexically, and syntactically more and more AAVE, beginning with my daughter (who is not white, but certainly not AA either).
One thing that I find striking about language variation in the NYC area is how many adult black residents speak in a totally non-AAVEish way because they are pretty obviously (to judge from their speech) immigrants from the West Indies (with Haiti obviously being different from the Anglophone islands and so on) or Africa (I assume also considerable variation in accent in English depending on country of origin). Their US-born kids don’t tend to retain those accents, but it’s not clear to me that they’re converging on “traditional” AAVE rather than some new/shifting synthesis which *might* (as John Cowan suggests) be less distant from the current speech of non-black New Yorkers of similar age/SES.
Indeed. I dub it “New York Multicultural English”.
How about “New York Estuary English”? Hands across the ocean and all that.
Hudson Estuary English, as opposed to e.g. Thames, Parramatta, and Liffey,
Estuary Englishes.
I was already reaching hands across with NYCME as compared with London ME, which is much more like what I’m talking about than Estuary English is.
“The closer you get to humans, the worse the science gets.”
As Stalin allegedly said about authors, I have no other humans to offer you.
What did he say and in what circumstances? I have had no luck finding this.
http://www.bibliotekar.ru/encSlov/5/91.htm
On the Botai horses (via Dmitri Pruss on Facebook). I don’t have access to the full article.
(I meant “the full paper”)
A better article, from Science.
Interesting! So Przewalski’s horse, like the North American wild turkey, is a descendant of a domesticate.
Seems so. And the Botai domesticate did not contribute significantly to the modern horse. One interpretation is that the modern horse is descended from a more western domesticate. Dmitry is reading it differently. Seeing that the genetic diversity of modern mares was present already in the Botai stock, he suggests that the domesticated stallion colts were castrated or slaughtered and the genetic replacement happened partly by catching of wild mares, partly as the stock of mares breeded with wild stallions across the continent. The two interpretations aren’t mutually exclusive.
Neither of the two main genetic conclusions, that the Przewalski’s horses are descended from the Botai and that the Botai lineage is apparently extinct among domestic horses, seems particularly startling on its own. It’s that both of them are true simultaneously that is most peculiar.
To me too — though it took me a while to get there. In the first article Sandra Olsen speculates that the Botai people migrated east into Mongolia, based on close similarities in the burial practice for slaughtered horses. One story might be that the Botai kept their domestic mares apart from wild stallions, while the mares that were brought (traded, stolen) westwards were bred with wild stallions along the way. When the Indo-Europeans took over the central steppe, the Botai retracted to the Mongolian highlands with their divine breed of horses. Eventually, after horseback hunting finished off the wild horse, some Botai horses escaped and multiplied in the wild. Maybe horsekeepers could be more lax in their horsekeeping habits. Maybe the domesticated horse lost importance when horseback horse hunting ceased. Maybe the Botai were overrun and their horses let loose. Or maybe time went by and divine horses weren’t as divine anymore, and they simply got some of those new ones from the Saka.
But I’m not sure that the Botai were the first riders. There’s a big increase in horse remains on archaeological sites on the western steppe in the first half of the 4th millenium, and on the Lower Volga already in the 5th, along with sheep and cattle.
3500 BCE is also a little late for the start of the IE expansion. Intrusions into the Danube valley by people carrying horsehead amulets had been going on for a while already. The great trek to Afanasevo happened in 3300 BCE at the latest.
Now it’s the time to check zebras.
Maybe they were once domesticated too.
https://upload.wikimedia.org/wikipedia/commons/2/2e/WalterRothschildWithZebras.jpg
Zebras are said to be too aggressive to be reasonably domesticable.
Horses, yes. Horses pulling chariots, yes. But riding?
Me: There’s a big increase in horse remains on archaeological sites on the western steppe in the first half of the 4th millenium, and on the Lower Volga already in the 5th, along with sheep and cattle
Admittedly, this means domestication and not necessarily riding. But keeping large stocks of grazing animals only makes sense as economic specialization. You need somebody who will give you something in exchange for your surplus. As soon as you go somewhere to exchange surplus, you have invented pastoral nomadism. When you go somewhere else too, and you exchange some of the stuff you got for your surplus at the other place, you have invented the trading caravan. Per my links somewhere upthread, this started developing across the whole Eurasian steppe in the early fourth millennium. Horses as pack animals would have made it easier to bring goods. Horseback riding would have made it easier to contain, guard, and move, herds.
Horses, yes. Horses pulling chariots, yes. But riding?
Dating is a dangerous game. You often end up being antedated. It’s unlikely that the oldest we know is the first that ever was. I don’t mean to be certain about anything. I’m just saying that there’s archaeological evidence for a special significance of the horse on the western steppe well before 3500 BCE, and that the incentives for inventing riding would have been in place there for a time.
It’s unlikely that the oldest we know is the first that ever was.
And yet our brains are so constructed that we almost automatically leap to that conclusion. This word was first used by Shakespeare! The first human lived right here! …oops, I mean right there! ….oops, I mean over there! It’s taken me decades to make serious progress in internalizing the intellectual awareness that finding something just means finding that thing, and it says nothing about what else might exist or have existed.
I’ve read that there is some weak evidence of Magdalenian people of France having tamed horses back in Upper Paleolithic (circa 15,000 BC)
In addition, claims that modern population X is not descended from ancient population Y are approximations at best; it may merely mean that none of the ancient chromosomes have survived repeated meiosis. As I’ve pointed out before, the chances that you or I have a particular ancestral chromosome from a particular ancestor even 400 years ago (16 generations) is only 1 in 65,536, and for 5000 years ago (200 generations) it is 1 in 1,606,938,044,258,990,275,541,962,092,341,162,602,522,202,993,782,792,835,301,376.
Also, taming is not domestication. Elephants, for example, are tamed but not domesticated: they are part of the wild elephant breeding pool.
it may merely mean that none of the ancient chromosomes have survived repeated meiosis
Yes. Add to that that there sample size is some 20 genomes from Botai horses and a similar number from later horses from across time and space. But the bigger the breeding pool, the better the chances, and there have been very many horses for a very long time. Natural selection or selective breeding could do things, though. Human preferences in colour, or size, or shape of the hooves.
Also, taming is not domestication.
True, and possibly relevant in the case of horses for a long time. One scenario here involves mares being kept but breeding in the wild. It’s a middle position with a one-way stream from the wild horse breeding pool and some level of managed reproduction and selection for desirable properties on genes transmitted through the maternal line. I wonder if they could tweak that distinction out of their old genomes. It would probably take a lot more than 20 horses.
And horses as technology is different from horses as biology. It’s likely that they interact and co-evolve, but in a complex way.
These numbers shrink, at least slightly, when there’s been any selection at all.
for 5000 years ago (200 generations) it is 1 in 1,606,938,044,258,990,275,541,962,092,341,162,602,522,202,99
this sort of statistical perversity relies on unrealistic assumption that there is no interbreeding whatsoever between ALL your ancestors in 200 generations.
but actually in most cultures people start marrying their first, second and third cousins and so on, wrecking these calculations from the very beginning.
in practical terms, you don’t have 65,536 ancestors 16 generations ago and you certainly don’t have 1,606,938,044,258,990,275,541,962,092,341,162,602,522,202,99 ancestors 200 generations ago.
Actual number of your ancestors 200 generations back is more like 10,000,000,000,000,000,000,000,000,000,000,000 times less than that.
Of course, and that’s why all we Europeans are descended from Charlemagne, though by varying numbers of descent lines.
As for the 5000 years, I was just basically showing off the Scheme language’s facility of (and with) bignums.
1,6 e+26 are still far too many ancestors. Even 1,6 e+10 are too many, conceding the cutoff after ,99.
And even if you don’t carry the chromosome of a certain ancestor, some other descendant might, so the effect is evened out on a population level. The chance of transmission would rather be a function of population size. The smaller the gene pool, the more is lost due to random processes (drift) per generation.
Parpola sez (p. 251): “The Rigvedic hymns were composed in Old Indo-Aryan around 1300–1100 BCE in the Indo-Iranian borderlands and the northern Indus Valley. The horse-drawn chariot is mentioned hundreds of times in them, while there are just two clear references to horse-riding.”
Yes, but the invention of the chariot was much later than the Botai corrals. The chariot seems quite firmly connected to the Indo-Iranian expansion out of the Sintashta culture.
I say “seems”. There could be older chariots waiting to be found.
Chariot in Sintashta, riding in Botai, and later the twain shall meet?
Unrelatedly, Parpola should have stayed away from the technobabble. “With 19 radiocarbon dates, the duration of the Sejma-Turbino network was fixed to 2150–1600 BCE (with sigma 1 calibration)” – reverse hull polarity and set the shields to an oscillating frequency! Rather, σ is the standard deviation, and all dates from 2150 to 1600 BCE are within 1 σ of the mean estimate. In other words, the probability that the true date lies outside that range is close to a third.
(Except it’s not anymore, because now – as the rest of the sentence makes clear – there are 38 dates, so that the range – presumably 1 σ again – has narrowed down to “c. 2200–1900 BCE”.)
The development of equine transport technology was a process in several stages. It’s far from completely understood yet, not at least because the archaeological evidence is mostly circumstancial, with few material remains. Even the interpratation of bits and tooth deformations from the Botai is contested. But the connection of chariots to Sintashta seems clear enough.
Thanks for the Parpola paper. I’ll read it eagerly. With those dates Sejma-Turbino is exactly contemporary with Sintashta. Probably no coincidence at all.
From later in the paper (p. 276):
Not through yet, but a few surprise points I’ve noted:
1. Parpola’s view is very much a work in progress, developing as a conversation with other experts in the field. It’s very interesting to see those processes at work — and very welcoming of him to invite us all to see them.
2. He sees the Sejma-Turbino phenomenon as a joint venture between Andronovo Indo-Aryans and Proto-Uralics from the Middle-Volga. They developed together in the forest zone, probably with Indo-Aryans as a ruling superstrate. The combination of Indo-Aryan metallurgy, horsekeeping, and warfare and Proto-Uralic knowledge of the forests made them able to branch out alle the way to the Altai.
3. Sejma-Turbino daggers are decorated with horses. One is decorated with a horse pulling a skier. It’s likely that the huge distances — from the Altai to the Baltic — were covered in the winter by horse-drawn sleds. So maybe the sled was developed in the Sejma-Turbino by adding a transport platform to the pair of skis and putting the skier on top.
4. The Shintashta arose at the very north of the steppe, close to the forest zone. This may be significant. Parpola doesn’t say so, but he makes me think that the chariot was invented as a summer sled. It didn’t get anywhere in the swampy forest region, but it worked well on the dry steppe. The Indo_Aryans never picked up riding. (Earlier today I was close to suggesting in jest that the Indo-Iranians learned their horsemanship, not from the Botai, but from proto-Uralics with reindeer.)
5. He’s as fascinated with the Mälar type axes as I am. There’s even a hillfort with East Baltic pottery showing that this was a real intrusion of a people from the east. He suggests — warily — that these became a forming element in South Saami. I’ll repeat my half-serious suggestion that they were the linguistic ancestors of the south Scandinavian Finns of the sagas.
6. Meanwhile, the Iranians-to-be lived on the Don-Volga steppe and got their copper from the Caucasus. The Maikop culture was formed by this intrusion into the Caucasus. The Iranians took up horseriding around 1600 BCE and spread explosively. This means that there weren’t two waves of Indo-Iranian but three, the middle one being Indo-Aryans from Shintashta. It also means that technologically the Botai were a dead end — unless they used their bits not for riding but for skijoring*.
* A strange adaptation of No. skikjøring. I would call it snørekjøring “line driving”.
As it happens, I stopped exactly at page 276 to write this up!
(Which was quite stupid, being two pages before the start of the reference list.)
the Sejma-Turbino phenomenon
In Russian this is the Се́йминско-ту́рбинский феномен, if anyone else is curious.
And Синташта [Sintashta] is stressed on the last syllable, according to this YouTube video; contrariwise, the Wikipedia article indicates penultimate stress, but I’m guessing that’s a guess.
Sintashta. Argh!
OK, Sintashtá.
And it’s actually at least four waves. First Andronovo Indo-Aryan. Then Iranian. Then wheeled Sintashta Indo-Aryan. Then Iranian again.
I’ll also take up another speculation from upthread: Whatever came before the first wave of Indo-Aryan could have been Proto-Tocharian. Reading about Sejma-Turbino I discovered the Qijia culture in Qinghai.
A point I noted and expected to reemerge towards the end, which it didn’t, is the unknown origin of the more easterly Fëdorovo branch of the Andronovo culture. This is also a candidate for ancestor of Tocharian.
proto-Uralics with reindeer
Obligatory reminder: reindeer herding among the Sami only dates from the Middle Ages, while the northern branches of Komi, Ob-Ugrians and Selkups all picked it up still more recently from the Nenets. I don’t know of a good terminus post quem for the Nenets and their relatives getting started with it, but it cannot be older than the Samoyedic expansion into the high Arctic, seeing how reindeer are not herdable in the taiga zone.
Reindeer herding may be relatively recent among other practitioners as well. The Uralic term for ‘male animal’ (*kojəra, > Finnish koiras etc.) developed the meaning ‘reindeer bull’ in Samoyedic (*korå), and this seems to have then taken off as a Wanderwort that has spread to at least Chukotkan (as in the ethnonym “Koryak”) and probably also Tungusic. It could have happened independently, but more reasonably this would be connected to some kind of a new technology, such as the idea to herd reindeer in the first place instead of merely following wild herds.
It is very conceivable that some small-scale reindeer taming went on much earlier, though, seeing how in much of Uralic, the word for ‘(tame) reindeer’ is an old loanword deriving from IE *peḱu-.
Obligatory reminder: […] Reindeer herding may be relatively recent
Thanks. I know, but I should have made that more explicit. But as you say, small scale reindeer taming is probably earlier. How old is the transport pulk? Just up there I stopped short of suggesting the pulk as an alternative to the sled for transport in the Sejma-Turbino. But it would be a good vehicle for e.g. the Sami takeover of the North Scandinavian trade routes.
Oh, and very interesting about *kojəra!
That’s in footnote 2, from which I forgot to quote:
And some of the earliest domesticated horses got bronze antlers put on them, right?
So guess I. Having watched the video, where the name occurs 3 times in 3 different cases, all end-stressed, I went ahead and fixed it.
Молодец!
^_^
@JC it may merely mean that none of the ancient chromosomes have survived repeated meiosis … the chances that you or I have a particular ancestral chromosome from a particular ancestor even 400 years ago (16 generations) is only 1 in 65,536
I’m surprised that no one has yet jumped in to correct chromosome. As Dmitry has explained here before, we don’t inherit entire chromosomes from our parents; instead they give us mixtures of their own homologous pairs. The word you want is probably genetic marker.
@j. The Uralic term for ‘male animal’ (*kojəra, > Finnish koiras etc.)
Aha, so Finnish koira (dog) must be a specialization of koiras, then? This explains why Google Translate Finnish-to-English keeps turning male birds into dogs, e.g. the male’s plumage comes out as “the dog’s suit” or “the dog’s costume”. I had guessed that Finnish used koira analogously to English “dog-fox” for a male fox, but now I see they’re two different words, which Google Translate can’t distinguish.
so Finnish koira (dog) must be a specialization of koiras, then?
Likely so. Komi кыр /kɨr/ ‘male dog’ seems to preserve an intermediate stage. Also, now that I check, Votic has koira ‘dog’ versus peni ‘female dog’, the latter from the western Uralic generic term *penä(-j) ‘dog’; the semantic shift here seems likely to have happened before the bleaching of *koira to just ‘dog’.
The paper Dmitry linked to in the other thread refers to another recent paper: Fages et al: Tracking Five Millennia of Horse Management with Extensive Ancient Genome Time Series (Cell, 2019-05-02)
So:
• Domestic horses are a sister clade to Botai/Przewalski. The “Siberian” and “Iberian” horses are farther removed.
• Diversity has decreased through time. Some of the earliest domestic horses found, including one from a Bell Beaker context in Hungary, are in sister or cousin clades to all modern horses.
• The Hungarian horse shows admixture from Iberian horses, and it’s suggested that this was a result of exchange of horses within the Bell Beaker network. There were probably one or more distinct but not yet identified populations of horses that also contributed to the early European horses,
• Until sometime around the 7th century there was great diversification within the domestic lineage. Steppe, Middle East and European horses diverged early.
• There were also distinct Northern and Southern European lines. The modern Icelandic and Shetlandic horses belong to the former.
• Iron Age European domestic horses were replaced by horses from the Middle East after the Arabic expansion. So were Steppe horses — the horses of the Mongolian hordes were more Arabian than Xiongnu or Scythian. The only surviving old breeds identified in the study are Shetland and Iceland horses.
But do the authors ever take the opportunity to use the word “S/Iberian”?
Sadly no. Somewhere in there I was thinking of making a rhymed summary, but forgot before I got to the end.
Hat: Could you remove the last paragraph starting with “So:” from the blockquote? That’s my summary,.
Done!
Me, Dec. 15, 2015:
Similarly, PIE had an amazing three-way distinction of the clusters /kʲw/ and /kw/ from the single phoneme /kʷ/. In the satəm languages all three stay distinct (e.g. the Sanskrit outcomes are ś, kv, k). Germanic, Celtic, Italic and Greek are all kentum, and all have or used to have phonemic consonant length (unlike PIE!), but the first three merged the clusters into /kʷ/, while the Greek outcome of the clusters between vowels is the new phoneme /kːʷ/.
https://xkcd.com/2421/
Trond: is there any relationship between Icelandic and Mongol horses/ponies? Both are northern and smallish with stiff manes.
I read a paper almost two years ago, and now I’m the house expert on horses. No, as I understand it, the insular breeds are the only ones continuing the genetic profile of the ancient Northern European domesticated horse. All others in the study, including the Mongolian horse, are mainly descended from the Early Medieval new horses. I guess the local mounted warfare business was rather uncompetitive.
Me, March 1st, 2020:
Of course not: they’re śv, kv, k.
What’s depressing about the dictionaries in question is the way they happily accept vast semantic and phonological latitude in the comparisons made. They are like the various Nostratic compilations that so much effort has gone into, and it says it all that they differ so much from one another.
Robert Ratcliffe did a detailed and quantitative analysis of Ehret vs. Orelbov(a), and there are a great many things going on in there. Most of Ehret’s comparisons are Cushitic vs. the rest, whereas for Orelbov(a) it’s Chadic: Teeters’s law in operation. Then there’s the question of triliteral roots: for Ehret the third consonant is always a suffix, but for Orelbov(a) some but not all roots are primitively triliteral. The two agree on some 21 segments, of which only 12 have substantially the same correspondences, and then 15 segments in Ehret alone and 12 in Orelbov(a) alone. Lastly there are just too many possibilities around, particularly because Arabic dictionaries contain a random mixture of MSA and Old Arabic and everything between, so that all of Arabic qarn ‘horn’, qaar-at ‘hill, qarra ‘be cold’, and qar ‘shout’ are reconstructed by Orelbov(a) as *qar in all four senses, with the distinguishing phonetic material lost in the process. The more of this kind of polysemy there is, the more randomness.
Ratcliff winds up with “There is no way to determine how many chance matches exist or occur between two languages. One can only ask how many chance matches will be found given a particular method of conducting a comparison.” Hear, hear.
Thanks, JC! Very interesting find.
I must say, for all that I am always happy to read papers that confirm my prejudices, it seems over the top to conclude the outright inapplicability of the comparative method over great time depths from the deficiencies of two studies where the method has been extremely sloppily applied, with numerous basic methodological errors.
(Looking over this thread to refresh my memory about it, I notice that in 2014 I was still not convinced that “Gur” was provably related to Benue-Congo, though I didn’t come out and say so explicitly. I was certainly wrong about that. At this rate of progress, I shall probably be a full-fledged Nostraticist by 2028.)