Jaison Jeevan Sequeira, Swathy Krishna, George van Driem, Mohammed Shafiul Mustak, and Ranajit Das have an article (in preprint, open access) called “Novel 4,400-year-old ancestral component in a tribe speaking a Dravidian language“:
Abstract
Research has shown that the present-day population on the Indian subcontinent derives its ancestry from at least three components identified with pre-Indo-Iranian agriculturalists once inhabiting the Iranian plateau, pastoralists originating from the Pontic-Caspian steppe and ancient hunter-gatherer related to the Andamanese Islanders. The present-day Indian gene pool represents a gradient of mixtures from these three sources. However, with more sequences of ancient and modern genomes and fine structure analyses, we can expect a more complex picture of ancestry to emerge. In this study, we focus on Dravidian linguistic groups to propose a fourth putative source which may have branched out from the basal Middle Eastern component that gave rise to the Iranian plateau farmer related ancestry. The Elamo-Dravidian theory and the linguistic phylogeny of the Dravidian family tree provide chronological fits for the genetic findings presented here. Our findings show a correlation between the linguistic and genetic lineages in language communities speaking Dravidian languages when they are modelled together. We suggest that this source, which we shall call ‘Proto-Dravidian’ ancestry, emerged around the dawn of the Indus Valley civilisation. This ancestry is distinct from all other sources described so far, and its plausible origin not later than 4,400 years ago on the region between the Iranian plateau and the Indus valley supports a Dravidian heartland before the arrival of Indo-European languages on the Indian subcontinent. Admixture analysis shows that this Proto-Dravidian ancestry is still carried by most modern inhabitants of the Indian subcontinent other than the tribal populations. This momentous finding underscores the importance of population-specific fine structure studies. We also recommend informed sampling strategies for biobanks and to avoid oversimplification of ancestral reconstruction. Achieving this requires interdisciplinary collaboration.
I’ll be interested to see what knowledgeable Hatters think about this. Thanks, Dinesh!
Interesting, I just posted on fb a discussion of a strange Dravidian-related ancestry in the Coorgs
https://www.nature.com/articles/s42003-025-08073-0
(I strongly suspect that the Coorgs’ unusual DNA composition isn’t very ancient, though, but stemmed from many generations of reproductive isolation in a small community; my interest to the paper came more from a surprise at the Coorgs’ traditions and beliefs)
So either the subcontinent was mysteriously lightly-populated as of four or five millennia ago, or there was a huge massacre and/or cultural subjugation of the Andamanese-like “ancient hunter-gatherer” population at the hands of invaders, or?
Or germs. Domesticated animals brought germs, and the herders were better adapted to that than the local hunter-gatherers.
I noticed Dmitry’s Facebook post earlier today, but hadn’t had time to look at it, so I just assumed this was the same paper. It’s even more interesting when they’re not, and they differ.
Not necessarily a completely violent scenario. A higher density of agricultural populations should have contributed to gradual displacement of the hunter-gatherer DNA in all but the most favorable conditions for the latter (just like in Europe, hunter-gatherer ancestry lingered in the North due to very rich aquatic resources).
The Coraga paper in this post, BTW, also deals with an extremely bottlenecked population following long reproductive isolation and low numbers, and I am just a skeptical that it truly represents something ancient
Or you just amend my “mysteriously lightly-populated” to “no more densely populated than you would expect for societies that weren’t yet doing the agriculture thing,” although I don’t have a very good sense of what that means in terms of the humans:acre ratio.
I like it, but the linguistic evidence for Elamo-Dravidian is quite tenuous – Elamite (as understood today, which isn’t terribly much) shares about as much or as little with some other language families of the wider region.
My academia.edu page presented me just now with this paper by Juha Janhunen as its top suggestion (for once getting it right that I would find it interesting):
https://www.academia.edu/45035334/The_differential_diversification_of_Mongolic
(Can’t find any other open-access link, sorry.)
I plonk it in here as an excellent case of genetics in the DNA sense not correlating with “genetic” in the language sense, but everybody knows what I think about all that so I’ll leave well alone, and just say that it’s a very interesting paper anyway.
What actually appealed to me about it was that it talks about what a remarkably diverse family Mongolic is, given its very shallow time depth – something that would also be the case for Western Oti-Volta if its current spread is all down to the Mossi-Dagomba kingdoms. Very similar timescale, in fact …
On the other hand, the distances involved in Mongolic are vastly greater, and the exact origins of the chiefly and peasant strata in WOV-speaking societies are not known. But the point about the rate of language change not being a constant on any metric is very much to the point. My point of comparison with WOV diversity has tended to be Romance, but Mongolic might be nearer the mark …
may i be the rather uninformed voice of total skepticism?
based on the abstract, this sounds like (1) we found a possible new lineage cluster in our sample, (2) our reconstruction of a hypothetical timeline for that lineage can be interpreted to line up with the hypothetical timeline of someone else’s linguistic reconstruction, (3) we can pick a hypothetical geography that sutures those hypothetical timelines perfectly to reflect the racial ideology of the ruling party in the country where we’re taking samples!
all i can see it demonstrating is that BJP/SS rule continues to ensure a steady stream of race science.
rozele’s comment made me realize that I have no idea how the more hardcore Hindutva / “Out of India” folks account for or explain away the seemingly dramatic differences between Indic languages and Dravidian languages if they do not attribute the former to “invaders from out of town.” A few minutes of quick googling left me no more enlightened and I feel that I may lack the attention span I had back when I was seventeen and being handed Hare Krishna agitprop in a dubious part of Philadelphia full of barbed wire and broken glass.
I’m not sure the idea that “the region between the Iranian plateau and the Indus valley” spoke Dravidian is in line with the ideology of the party. (because of what JWB mentioned: its incompatibility with the Out of India scenario)
I came across a line “A slightly more sophisticated attempt at getting around the linguistic differences between Dravidian and Indo-Aryan languages has been made by Subhash Kak” (here) which I found promicing.
Even though the paper that contains it did not answer the question of what is the stance of Modi’s government (and couldn’t answer it because it predates this government), I was much amused by how the quoted line is placed between
“According to Rajaram, no well-informed scholar today takes either the Aryan invasion or the notion of the foreign origin of the Vedas and the Vedic civilization seriously” (same page, with disapproval)
and
“According to Dixon, there is no reputable historical linguist anywhere in the world who accepts the claims of the Nostraticists.[101]” (page below, with approval)
I think I’ll stay away from people who use the adjective “reputable”:-)
Funnily, it is not the usual effect of the fact that Western scientists were poorly informed of the work of Russian-speaking scientists (you may not subscribe* to Nostraticism but many Russian linguists who either accept the claims or work on it are “reputable” here). Dixon footnote makes me think he knew about its greater acceptance in USSR – he simply did not think Soviet linguists are “reputable”.
But he also says “claims of Greenberg and Nostraticists”.
*I’m indifferent to it and don’t understand why feel so strongly about it as some do. A hypothesis, promicing enough for some to work on it, not quite promicing for others, but without serious consequences for anything until proven because it’s a hypothesis. Yes, sadly some treat hypotheses as the truth. That is stupid and unscientific, but bullying, as Dixon does, enthusiasts of whatever hypothesis by referring to reputation is exactly same brand of stupid.
“Russian linguists… ‘reputable’ here” – I mean (as is known) people engaged in conventional historical linguistics and who don’t practice sloppy methods, at least (for those of them who work on it) outside of their “Nostratic” work which I’m poorly informed about. Because there is a plenty of scholars reputable in India who believe that historical linguistics is long-discredited racist bullshit from 19th century. One more reason to avoid referring to reputation “anywhere in the world” only to specify in the footnote that the whole world was an exaggeration.
PS au village, sans prétention, j’ai mauvaise réputation…
No reputable scholar believes in action at a distance – it’s mystical nonsense.
No reputable scholar believes in plate tectonics – there simply isn’t any plausible physical mechanism for it.
(But they also laughed at Bozo the Clown…)
All very no true Scotsman. I can use my own judgement to find a linguistic theory unconvincing or even absurd; why invoke a poorly specified bunch of other people’s?
@rozele: I haven’t followed Indian politics closely, but Hindutva ideology as I used to encounter it on the ‘Net would be fiercely opposed to the idea that anything could have come from outside and (implied) before Sanskrit. There’s a different Tamil nationalism that also could come into play, bur I don’t recognize that either. If anything, I detect a sympathy towards the peoples that were marginalized by the Brahmin purity prescriptions, but that could be wishful thinking on my part.
That said, I share the skepticism of your points 1 or 2. There could be something there, but without independent confirmation, there’s no way to know.
Greenberg was most certainly a well-informed and reputable scholar. It’s opinions that are reputable or disreputable. You can be a reputable scholar in one area and a total crackpot in another – even a different subfield of a discipline in which you are actually a highly competent expert in your own.
Many an excellent linguist has no idea at all of how to go about comparative linguistics, for example, in which they are so flaky that they’re don’t even know how ignorant they are.
(Greenberg himself was actually in a somewhat different category; I think he’d more painted himself into an intellectual corner. Perhaps more of a risk for a truly eminent scholar than for a merely competent one, even. Like being an author so successful that editors dare not point out your infelicities any more.)
I seem to recall that George van Driem has some pretty odd ideas in certain domains. But then, don’t we all?
Greenberg lucked out that, in spite of the flaws of his work, his posthumous reputation is nevertheless that of a great scholar, while his acolyte Merritt Ruhlen got a damning description by Larry Trask that probably still rings true: “Ruhlen is not recognized by anybody in linguistics as a member of the profession. Every single linguist who is acquainted with his work regards him as a crackpot and a charlatan.”
(I have remembered Trask’s first sentence verbatim for decades, but I was able to fill in the second because, a Google search reveals, I had commented about this here on LH before back in 2006.)
Back when I was young and naive I assumed that Ruhlen was not a crackpot or charlatan because obviously a respectable establishment/gatekeeper publisher like Stanford University Press wouldn’t have published his books if that were the case. I no longer subscribe to the premises underlying that reasoning.
It seems to be an actual trope in the more excoriating bring-popcorn type of academic reviews of monographs, to add a envoi reprimanding the “respectable academic publisher” for their involvement.
I think I first came across this in a review of the – truly, very bad – Transforming the Images, by Elke Novak, which deals with Inuit syntax. The reviewer, more in sorrow than in anger, pointed out that the guilty publisher had recently reissued Kleinschmidt’s classic Greenlandic grammar – a painful coincidence, they felt …
It has recently been determined by statistical analysis of a wide range of human genetic samples that all human beings are genetically related, at a approximate time depth of [N].
Excitingly, this is consistent with the view held by linguists* that all human languages descend from an ancestral “proto-World”, at an approximate time depth of [N].
This momentous finding …
[I meant to be satirical, but I think we actually did that very paper not long ago …]
*Merritt Ruhlen and Noam Chomsky! What more proof do you need?
You can read more about the tribe whose genes got analyzed in wikipedia, including the bad treatment by their neighbors that eventually prompted the state legislature to enact the Karnataka Koragas (Prohibition of Ajalu Practice) Act, 2000. https://en.wikipedia.org/wiki/Koraga_people
Reminds me of the last infuriating book I read, Latin Alive: The Survival of Latin in English and the Romance Languages by Joseph Solodow, published by Cambridge University Press. Half of the book is decent pop-sci showing how Latin, and specifically Vulgar Latin through French, had a major impact on the development of English. In the other half of the book, the author boldly attempts to “reconstruct Proto-Romance”, yet he completely ignores Balkan Latin (Romanian gets a single, vague mention in the book, other Balkan Romance none). I looked up the author and he was a classicist who, though presumably highly competent in Classical Latin, was a mere dilettante in matters Romance. What’s really disappointing here is that his CUP editor didn’t realize that Romance linguistics is a whole field of its own, with its own important literature to cite, and the manuscript needed competent review from that community.
@Christopher Culver: Lots of discussions of Proto-Romance just blow past the many questions related to Romanian.
rozele is on to something, but that something is not, I think, Hindutva ideology: of the five authors, the primary one has a Portuguese (presumably Christian name), two have Hindu names, one has a Muslim name, and one is European.
The guiding light here is, I think, the idea that Dravidian came from the northwest, and specifically that Harappan civilization was Dravidian, and that its script was indeed a script, representing a Dravidian language, and therefore Yay Dravidians! It’s less jingoistic and deluded than Tamil The Primordial Language and such, but it still rests on too-thin evidence.
The inspiration for these ideas come from several sources. I think those are, first, McAlpin’s Elamo-Dravidian, which also places the Northern Dravidian language Brahui (spoken in Pakistan) closer to Elamite than to Dravidian. Elamo-Dravidian is no better than any other wild long-range theory. Second, Parpola, Mahadevan, Zvelebil, and maybe others claimed the Indus/Harappan script is Dravidian, based on identifying supposed rebuses that only work in Dravidian. The supposed decipherment never got very far, but the idea stuck. Thirdly, Koraga, first described in the late 1960s, was claimed to be North Dravidian, despite its location in Karnataka. That was first suggested in Bhat’s 1971 description of the language (p. 3). Of the six similarities to Northern Dravidian in his list, only one is universal to North Dravidian and to Koraga, namely the past verbal suffix –k~g~gg, which Emeneau apparently figured to be a NDr innovation. On the other hand, Krishnamurti’s 2003 The Dravidian Languages (p. 300) says, “In my opinion this is an innovation in Koraga of giving a past meaning to a non-past suffix just like Kota, which uses -p- (<∗-pp), a non-past marker, as a past marker.”
The current paper was preceded by another in 2024, which argues similar same points, and also mentions the -k past suffix as its only evidence to consider Koraga NDr, then states that affiliation as a proven fact.
All this, along with the social, genetic, and geographical isolation of the Koraga, must make them a tempting subject for genetic studies. I know very little about genetic phylogeny, but a couple of things make me uneasy, and I hope the experts here will help me out. If I understand it right, the supposed uniqueness of the Koraga is based on the admixture analysis (fig. 3 of the most recent paper). The admixture diagram is based on choosing K=10 (that is, the number of components in the admixture.) As I understand it, choosing K is a bit of an art: too few components, and significant signal is lost: too many, and you see artifacts. That is why in papers of this sort you often see versions of the diagram for multiple K’s. Here all they have, in the supplementary information, is a cross-validation error vs. K diagram (fig. 2). At a glance, I would say they might have chosen too many components.
Other populations in the admixture diagram (“Other tribal populations across India” and “Mainstream Indian populations”) both look unrealistically uniform. The rest (Toda, Jarawa/Onge, “British”, etc.) each look even more uniform. Does that diagram make sense?
The PCA diagram (fig. 1) puzzles me too. I am not sure what the principal components are based on, perhaps on Asia in general, or the world. That makes discerning South Asian significant PCs less distinct. The Koraga are off the main “Dravidian cline” in the diagram. The authors say they belong to “clusters that have drifted away from the ANI-ASI cline”, without a clear explanation. Am I right to be puzzled?
Finally, one of the coauthors, George van Driem, has done a great deal of important and pioneering descriptive fieldwork on Himalayan languages. Since then, he has dabbled in long-range theorizing (Burushaski-Yeniseian in particular, which Vajda has very graciously criticized), and some grand hand-wavy theories of language evolution and population spreads. He would be the linguist in charge here, and yet he signed off on this paper, which takes Elamo-Dravidian and NDr Koraga as established facts.
@Y, thanks! I mean, for the information about Koraga.
But I’m confused now. The book you’re referring to is accessible for free via google scholar, I checked p 300 and other mentions of Koraga – there is not much.
“Koraga (Bhat 1971) is almost like Tuḷu in most respects and is tentatively shown as an off-shoot of Pre-Tuḷu.” ; “The location of Tuḷu in the family tree is doubtful and Koraga needs to be appropriately located in the subgrouping scheme.” ; “Koraga past -k-/-g- may appear to be like that of North Dravidian, but there are no other features that it shares with North Dravidian.”
So he simply treats it “as an off-shoot of Pre-Tuḷu“.
Contrast this to Sequeira et al: “has been influenced for centuries by surrounding by Tuḷu speakers, and many Koraga are bilingual in Tuḷu“.
Is that all?
Only one suffix which some scholars (without further comments) treat as divergence from Tuḷu and some (also without argumentation?) treat as a link to North Dravidian on the background of convergence with Tuḷu?
Or is there more to it?
The work of Zvelebil referred to in the article is of the same nature, Dravidian Linguistics: An Introduction and its treatment of Koraga is even less detailed (link):
“If we add to this the possibility that Koraga, a small Dravidian language of coastal Karnataka, may belong to this “Northern” branch of Dravidian (as its linguistic features seem to strongly indicate), then a location of Kuṛux-Malto on or near the coast is not unlikely.”
“…an isolate, as is the recently discovered Koraga which is at the moment a true enigma.”
“takes Elamo-Dravidian and NDr Koraga as established facts” – The paper intorduces Koraga with references to books that treat it as either NDr or an independent branch within D and does not mention the opinion that it is SDr. I don’t know whether it then speak about the hypothesis as a fact or not.
References to Elamo-Dravidian are not numerous, it does not take Elamo-Dravidian as a fact. Unless characteristics “hypothetical” and “a theory” amount to such treatment:) Also when you place them “between the Iranian plateau and the Indus valley” – if of course such a placement was not motivated by the E-D hypothesis, you can’t avoid mentioning it (as a theory and a hypothetical group).
“dabbled in long-range theorizing” – and mastrubates, I bet. Sloppy science is sloppy science and should be called so, but why speak disrespectfully about a scientist’s interest in “long-range theorising”? An interest is an interest, and nothing good will come out of censoring interests.
I do understand that historical linguistics lacks a solid foundation. In a sense.
I don’t mean that its methods are not good or its results are not reliable. They are.
What is lacking is an exact (as in exact sciences) procedure which would allow us calculate precisely (and quantitatively) how reliable (or likely) a theory or reconstruction is, and where it is circular and thus delusional. If such a procedur were available, there would not be conflicts over reconstructions.
Reliabilily is maintained by the use of excessive data. Or intuition. And there is the fear of less reliable (and hypothetical) connections – that alongside with teaching of hypothecal and unlikely results as facts and also individual scholars who waste years of their time on working on false reconstrcutions with circular motivation. Yet fear (in science) is always irrational and per se indicates that missing foundation.
Of course there are people who want to reach beyond the limits of reliable.
That is natural for curious people (and researchers who are not curious are not researchers but robots).
I admit, I’m not attracted by [time] depths and [enormous] scales but I understand people turned on by these. It is reasonable to still demand from them self-critical attitude.
I too would rather walk on a shakier ground in uncharted land, but if I were a linguist I’d be drawn to exploring things horisontal rather than vertical (language areas that is). Those are exactly shakier grounds.
P.S. my comment about NDr Koraga and E-D was meant as a partial confirmation of what Y said, in the NDr Koraga part. As I said – I’m confused:/ Brief mentions in introductory textbooks the first of which calls it an enygma and isolate, while the second treats it as a part of Tuḷu-Koraga don’t help with this confusion:( (I disagree with Y’s characteristic of the paper in the E-D part)
Lots of discussions of Proto-Romance just blow past the many questions related to Romanian.
Well, as every Dacianist knows, Latin is in fact descended from Romanian, so Romanian wouldn’t enter into a discussion of Proto-Romance.
In light of Christopher Culver’s post about a crackpot work with the imprimatur of Cambridge Univ. Pr., I certainly cannot be confident w/o reading it that another CUP-published work bearing on “Proto-Romance” is crackpottery-free, but _The Regional Diversification of Latin 200 BC – AD 600_, by the late J. N. Adams (published in 2007) certainly sounds interesting. By pushing the thesis that “Latin” at least once it had expanded pretty far away from Latium “was never geographically uniform” it strongly implies that there never was any sort of not-directly-attested “Proto-Romance” that can be distinguished from Latin itself. A glance at the table of contents suggests that Eastern/Balkan Romance may (yet again!) get short shrift, but that may just be due to the paucity of contemporaneous textual evidence, since it sounds like Adams is focused on indications of variation that can be found in that sort of evidence rather than in conjecturally reverse-engineering from daughter-language evidence many centuries later.
I got an 88 page preview of the book J.B. just mentioned. This includes the first half of the introductions, followed by sequences of 3-4 pages of intro and 2 pages missing. It continued that way into the chapter on Republican inscriptions.
I’m not clear whether everyone who clicks such things gets the same preview, so if anyone is interested, I hope you get a big chunk of it.
Ryan, that book by Adams (and in fact the whole series of monographs by Adams of which it is part) has been on internet shadow libraries for over a decade now. Why are you limiting yourself to 88-page previews?
With regard to Mr. Brewer’s comment:
“Crackpot” is a strong word and I wouldn’t use it about Sokolow’s book. It isn’t all that different from a number of pop-sci books on Romance from the twentieth century that, as a rule, ignored Balkan Romance. What’s disappointing is that scholarship had moved on and CUP put out a book that didn’t reflect that. To give one example, those old treatments of the Romance languages that Sokolow regurgitates describe monophthongization of *au to *o as pan-Romance, but this did not happen in Romanian, Aromanian, Megleno-Romanian, and Istro-Romanian. I’d like to see about Dalmatian, but not sure what reference to consult.
I think “crackpot” was in my earlier post that was the jumping-off point for Mr. Culver’s discussion of Sokolow, but I should not have assumed that he was suggesting a perfect parallel (in that regard) with his example. And I was sort of taking others’ criticisms of Ruhlen at face value without going back and devoting time to considering whether crackpot/charlatan were excessively strong labels.
That fits this paper from last year, which found there’s no difference between Proto-Romance and Classical Latin in basic vocabulary. From the conclusions:
It sounds like Proto-Romance was mainstream Latin as spoken during the Crisis of the Third Century.
>Ryan, that book by Adams (and in fact the whole series of monographs by Adams of which it is part) has been on internet shadow libraries
Cool. JWB didn’t provide a link. I looked at 5 or 6 sites to see if there was one, and the best I found was what I linked, I thought helpfully. I’d note that in disparaging my efforts, you didn’t bother to offer a better link.
Sigh. Hat, I must have made a typo in the html code when I tried to link above at 10:50 am. Any chance you can check and maybe correct it?
If you can reply with the actual error, I’d be interested. It’s odd that I made a typo that eliminated the link, yet still put the word in red letters, which is why I thought at a glance that the comment came out right.
I can’t correct it — it just says <a>preview</a>, with no URL. (I just sent you a link to the Adams book.)
Christopher Culver: As I recall Vegliote (I recently explained here at Casa hat why the term “Dalmatian” should be avoided) also preserved Latin /au/ (The same is true of Old Provençal and a great many modern Occitan varieties, and it must have persisted quite late in pre-Old French, as the evidence of “chose” from CAUSA shows: /k/ was palatalized before /a/ but not /o/, indicating that the change must have taken place before the reduction of /au/ to /o/).
The best summary of the sound changes from Latin to Vegliote which I know of is found in volume II of Pierre Bec’s MANUEL PRATIQUE DE PHILOLOGIE ROMANE (Basic summary: the consonants are amazingly conservative, in some ways even more than those of Sardinian, but the vowels have changed even more radically than those of French. The next time I teach a historical linguistics course this is the example I will be using to highlight that languages can be amazingly conservative in some aspects and amazingly innovatory in others).
(Some of it may now be out of date -I know that recent research has shown that in Vegliote, Latin /kt/ became /jt/, not /pt/ despite the evidence of “guapt” as a reflex of Latin OCTO. Because other instances of Latin /kt/ -notably irregular past participles such as FACTUM and DICTUM- show /jt/, it is now accepted that “guapt” owes its /pt/ cluster to the analogical influence of the Vegliote reflex of SEPTEM).
Only because I didn’t know how Hat feels about linking to or mentioning specific names of shadow libraries. But it’s getting rare now to meet someone interested in linguistics who doesn’t use them. Even hoary, starchy-collared old professors might pull a book off one in the middle of a lecture in order to show some example.
shadow libraries
Didn’t know the expression, but immediately found shadowlibraries on github.
Mr. Culver’s discussion of Sokolow
Confused me for a moment — the author of Latin Alive is Solodow (as in Christopher’s first mention), not Sokolow (as in his second).
Yes, sorry. I wish the edit button here worked for longer.
Ryan, red means the tag <a></a> worked, but the href part didn’t.
I wish the edit button here worked for longer.
Present help in time of need.
realizing i never spelled out how i saw the paper working within hindutva: it provides a Noble and Antient World-Historical genealogy, crafted specifically for the substrate on which The Fathers of the Nation rose, and whose latter-day heirs are thus Our Brothers in Both Faith and Blood. it’s hard to get classier than the elamites for that kind of thing – after all, they invented Civilization.
i noted the very precisely calibrated, quite unambiguous spread of the names, and said “o! that’s interesting – it’s the same demographics-forward move the u.s. far right formations are making these days.” but i know nothing, in fact, about these authors, so i have no idea whether the match i’m describing is an engineered result, or good-faith data presented through the dominant metanarrative, or what. i’m agnostic about all of that – the congruence is meaningful whether it’s accidental or not.
I see that one of the authors, Ranajit Das, has coauthored a couple of papers with Paul Wexler et al. on the origins of Ashkenazi Jews. Ooooo-kaaaay…
@JWB, you’ll find it comforting that Trask says this in his textbook:
“The best history of the attempts at classifying languages into families is Ruhlen (1991); be warned, however, that the book contains a few errors, and that Ruhlen accepts as valid some recent very large-scale groupings that are rejected as unsubstantiated by the vast majority of historical linguists.”
(in Wikipedia this reference supports words about criticism of a certain tree which is not Ruhlen’s)
@drasvi: Good to know although now I’m wondering if “Ruhlen (1991)” is just a different edition/printing of what I think of as “Ruhlen (1987)” or something else entirely I can’t immediately figure out from googling.
Trask is online* for free, sunlit not shadow, Ruhlen 1991 is what you think:
A guide to the world’s languages, vol I. Classification
* I mean, if anyone is curious.
Hat wrote:
>I just sent you a link to the Adams book.
To my email? I do use my real email to sign my posts, but I haven’t received anything.
I’d be interested. I did just get my copy of The Making of Europe. Thanks for the recommendation!
Shadow libraries: I think the lagest (but not most convenient) is Anna’s Archive.
That’s the one I use. I just re-sent the link, Ryan — let me know if you get it.
Got it. Thanks.
WP says, most of criticism of Ruhlen is caused by his support of mass comparison.
Is this a serious understatement?
_____
Also, Nichols (about Africa): “Greenberg 1963b is a paradigm case of scholarly success, its analysis still largely intact in received view and its method generally accepted in mainstream thinking.”
Is this true, about the method?
Fair enough about Ruhlen, since most of his original work has been in mass comparison and the defense thereof. The classification book was useful for lower-order families, as an imperfect but OK compilation of the then-current state of the art, but it’s outdated now.
As it turns out, Nichols’ enthusiam for Greenberg was misplaced. She was one of many who didn’t accept Amerind, Indo-Pacific, or Eurasiatic, but were willing to accept the African classification. By now, three of his African phyla (all, if Omotic is not AA) and many of the sub-phyla have now been broken up, for the same methodological faults that invalidate Amerind and many of its subdivisions.
@drasvi:
Sure. It’s largely accepted among people who know nothing about African comparative linguistics (who, after all, greatly outnumber those who do.)
Is that Johanna Nichols? She belongs in the aforesaid majority group.
(What has actually endured in G’s classification is what he adopted – largely without attribution – from previous studies. His own innovations have fared poorly. As you’d expect from his methods.)
Anyone who says that mass comparison is a “method generally accepted in mainstream thinking” in comparative work reveals that they should not be taken seriously at all when talking about comparative work – whatever real contributions they may have made elsewhere. (As both Nichols and Greenberg have done.)
“Niger-Congo” has lasted longer than “Amerind” (a) because it contains a very large core (Volta-Congo) of languages that really are genetically related, unlike “Amerind”, and (b) because there weren’t at that stage nearly so many highly capable scholars working on African languages as on American languages, who could point out the multiple vitiating errors in the data and the analyses.
As far as I understand, what has endured of Greenberg’s methods is the attempt to classify the languages of Africa – as opposed to classifying their speakers and imagining this gets you at least halfway to classifying the languages.
I do hope he lifted that from earlier work…
He did get that from predecessors – although the prestige of promoters of “race”-based classification, especially Meinhof, had caused better linguists, like Westermann, to either resile from previous correct views or at least downplay them. Greenberg does deserve credit for killing all that off, even though he didn’t himself originate the ideas he popularised, it was he who made them mainstream (again.)
Even Guthrie (unlike Meinhof, certainly no Nazi) refused to accept that Bantu was related to the “Western Sudanic” languages. So someone like Greenberg was certainly needed to undo the damage, whatever his own mistakes.
As you imply, that is itself a lasting contribution worth honouring.
[I just (not before time) got hold of a nice dead-tree copy of Bargery’s Hausa dictionary. It has a long intro by Westermann which is still worth reading.]
I was just looking at a good recent account of Dompo, a Ghanaian language of which the only previous notice was a very brief study by Roger Blench, who thought it was an isolate. Looking at it myself, I’m pretty sure it’s actually Volta-Congo, though I applaud Blench’s caution.
Seems to me that Blench has the makings of a true Splitter, if only he could rid himself of the idea that what Greenberg hath said, stands forever (or until someone can prove that the languages G lumped together are actually unrelated – i.e. forever.) Where G has not spoken, Blench seems to be wholly sensible. He’s sensible on subclassification within genuine genetic groups, too, very often.
On both of these points I recommend his article “Declassifying Arunachal”. It’s about regarding a lot of languages of Arunachal Pradesh as isolates or two-member families; they had all been referred to Sino-Tibetan on almost no evidence.
(Those recent papers purporting to support speculative long-distance language relationships by appeals to genetics-as-in-DNA are, of course, attempting to recreate the very mistake that Greenberg so successfully opposed.)
Crimes are not typically committed with clues just subtle enough that exactly one clever Poirot will see through them. Usually they get solved soon, or never (or until some new technology appears). Likewise language family relationships are typically either transparent, or hopelessly impenetrable, with a few in between. Prehistory is not obliged to give us any clues, archaeological, genetic, or linguistic, though I am grateful for the ones it does.
Indeed. The important thing is to have the ability to say “alas, we just don’t know.”
Tom Stoppard remarked somewhere “all hope must be built on a foundation of solid despair”; I think that knowledge, likewise, can only spring from frank confessions of ignorance.
[IIRC, Stoppard was either discussing or actually quoting from the remarkable play Next Time I’ll Sing to You, by James Saunders.]
@Y, thanks!
“Mass comparison” is a method of processing data. I think it’s untrue – and can’t be argued – that its output is total noise and tells nothing about the world. I think, like any such method it has a certain place in science, though many methods won’t ever be practiced by anyone because there is a better method who does the very same job but better. And of course it shouldn’t be used out of this place. Thus I think supporting it can’t make one a crackpot.
There are always questions of what exactly its output means (this is what people are arguing about), where and why its wide practice will lead us, and, like any method (e.g. comparative method) – and I believe this is what makes one a “crackpot” – unless implemented as a computer program it can be done well or it can be done bad.
Or happily and carelessly bad, leading to happy and careless delusions, that is crackpottery.
So what I want to understand is whether Ruhlen does it, crackpot support and implementation of the method. When Trask says “crackpot” I understand it this way.
I followed some references for “objections” to Ruhlen in WP and was disapponted: say, the 5 objections in the Kusunda section (and it is not even a “Ruhlen’s” work, it is four people) are supported by a reference to Poser’s mail with 2 objections to…McWhorter.
In the context of the artcile but without reading it.
The proper use of mass comparison is in suggesting hypotheses for testing by comparative methods. Doing that is entirely valid and not evidence of crackpottery.
Deciding that the testing by comparative methods is unnecessary because the issue has already been settled by mass comparison is either stupidity or crackpottery.
Presenting the results of mass comparison which have not been tested by comparative methods as being established facts is scientific fraud; the culpability of this may be mitigated by how stupid you are or how much of a crackpot you are. (M’Naghton rules.)
(Greenberg himself did not – quite – proceed to this Stage Three. Ruhlen did.)
It can be legitimate to present unverified mass-comparison results as worthy of interest in cases where the comparative work has not been done at all, is at a primitive stage, or is proving very difficult. It is not legitimate to cross the line and to suggest to your readers that they are being told established facts.
Mass comparison cannot demonstrate that languages are genetically related. Ever. It can only ever suggest that the possibility is worth investigating.
@DM, oh, that’s one the first artciles (the second one, if I haven’t forgot anything) by Blench I ever read.
It made me hm… No, not a “fan”, but very determined to follow his work. I realised that he’s extremely itneresting.
DE, I agree of course with your first paragraph – and that’s why I asked if what WP says is understatement. Because it is not a serious problem, that someone supports something.
I slighly disagree with the second: I’d call that idiocy. Not stupidity (too mild) and not crackpottery either.
As for the third… Er. I don’t know what is the “fact” of classification. Tell me what exact historical process is claimed to be a fact here, and I’ll know what to think.
But it is normal for scientists to have contradictory very firm beliefs that this or that etymology is true or untrue, and I’m not ready to call them ALL “stupid crackpots”.
P.S. Ruhlen maintains that such classification is the first step in the comparative method and that the other operations of historical linguistics, in particular the formulation of sound correspondences and the reconstruction of a protolanguage, can only be carried out after a hypothesis of classification has been established. – WP, again, without references.
“…and not crackpottery either…” – because there is a plenty of people who happily keep doing incredibly sloppy and delusional work without slightest intent to refine their methods or otherwise be self-critical and sometimes even make extra-terrestrials part of their theories:) We need to call them somehow.
To clarify: by “comparative methods”, above, I mean, looking for regular phonological and morphological correspondences, working toward reconstructing protolanguages, and developing criteria for recognising loanwords. All that neogrammarian stuff. I’m specifically not including mass comparison in the term.
Mass comparison is just the warm-up to the real event. Warm-ups have their place. But I’ll demand my ticket money back if all I get to see is a warm-up.
I’d call that idiocy. Not stupidity
I was subsuming ignorance under stupidity. That was a bad move: many of the most egregious perpetrators are not stupid at all, they’ve just never quite got the hang of comparative linguistics, for some reason. Trouble is, they don’t realise that they haven’t got the hang of it …
Their productions then go on to pollute the noosphere. There seem to be entire self-reinforcing and self-referencing subcultures of Crap Historical Linguistics.
I expect that many are in between in that they become reasonably transparent after a lot of hard work.
I think we’re seeing this in the fact that new sound laws are still being discovered within IE, indeed within Germanic for instance; they just have fewer examples and more counterexamples-at-first-and-second-but-not-third-glance than the classics do.
Sure, there is endless work to be done on subclassification. I meant discovering new relationships between top-level families. Even not so transparent language families were recognized early on, like Sino-Tibetan.
There are a few outliers-yet-not-true-crackpots who deny that Chinese is related to to Tibeto-Burman genetically, attributing the many clear resemblances to thousands of years of contact instead. (Not so much Splitters as Antilumpers, I guess. Atheists rather than mere agnostics. The true Orthodox Splitter only declines to be persuaded, rather than actively denying the very possibility of a genetic relationship.)
So do I, because I don’t see a reason to view this as qualitatively different – having read attempts in both.
Are there? I thought Laurent Sagart was the last one, and he gave up around the turn of the millennium (though he still maintains, or did as of 10 years ago, that the closest relative of Sino-Tibetan is Austronesian).
the 5 objections in the Kusunda section
I’ve met the first two authors of that paper; at the time, they were energetic but untrained amateurs, trusting rather too much in Ruhlen’s judgment. One gave up on historical linguistics, and the other, under the influence of Sergei Starostin, eventually became a serious historical linguist and realised that his first paper was basically worthless.
David Eddyshaw: Tom Stoppard remarked somewhere “all hope must be built on a foundation of solid despair”
Can you be more specific? The passage from Saunders that Stoppard notably quoted was this:
What you quoted sounds more like this, from Bertrand Russell:
Also, the internet informs me that Tom Stoppard did the film script adaptation of a novel by Nabokov called Despair, but that doesn’t seem to be relevant to the quote.
Can you be more specific?
Unfortunately, no.
I think I remember this from a television interview, from the distant days when I watched television (so early 1990s, or before.)
I’m pretty sure of the wording (which is memorable, as you’d expect from Stoppard); he may very well have had Russell in mind, of course, or indeed have been directly quoting him (more or less.)
I’m certain that my immediate source was Stoppard, not Russell himself. I admire Russell (mostly) in an abstract sort of way, but have never been able to get on with his pop-philosophy writings.
The passage you cite from Saunders is the one I recall Stoppard citing. (It, too, is very memorable. I actually bought a copy of the play after hearing it.)
@Y, @DE, thanks!
Yes, Johanna, from the introduction to her book.
No, I don’t think she means “all people think it is good, be with all people!” (neither an opinion nor a fact and hardly good for a scientific work)
I think she means “folks think it’s good”. But I think she too thinks it’s good.
What made me wonder:
If you’re building a tower and have sewn and pitched a tent, waitign for workers to arrive, and it is a good tent (and workers sleeping in your tent agree fully) it would be strange to say: bricklayers and stonecutters “accept” your sewing method.
Your tent is to be replaced with stone tower. The sewing method with stonecutting and laying methods.
Africa is Very big.
Greenberg’s work was done by one man in a few years.
IF it is warm and dry, then it is a “tent”, and the method is “sewing”: a method which is good for one researcher working with data from all over Africa for a few years.
If the method is “masonry”, then rain wets you and sun burns you and snow or sand pile inside: a “tower” is too much for one man and few years.
This, I think, true for any talented researcher and any wonderful method.
I don’t mean that Nichols’swording are “strange”, but I wonder if people think of his method that it is (good or bad) masonry.
@Lameen, thanks!*
I haven’t yet read the Kusunda paper:
I hoped to find serious criticism, but with zero knowlege of the area it will be difficult for me to evaluate the work. I honestly don’t know what to think of correspondences listed in WP (in terms of “the chance of coincidence”: I’m much more accustomed to either longer lists or, conversely, individual words).
One of the objections listed in WP is “borrowing”. This actually refers to the exchange between McWhorter and Poser: McWhorter said pronouns are rarely borrowed (“languages do not exchange pronouns much. Usually, a language’s pronouns are original stock, not the result of later bartering“), Poser objected that it is not true that they are never borrowed, both claims are true:-/ I think some people do this hedging when speaking but not when thinking and have such strange argumenets.
What made me curious is that if there are arealisms (pronouns rather than cultural wanderwords) extending from Nepal to the Andamanese islands – to say nothing of New Guinea (with less perfect matches) – that would be no less interesting for me than shared origin of the languages. Unless of course, the area in question is already too well-known to have such arealisms. The historical processes underlying exchange of whole langauges and of their fragmetns are similar: human contact.
Do you rememeber if the paper is bad becuase
– bad processing of data (say, the forms listed are in reality different)
– the match is though to be chance coincidence
– the match is though to be a result of borrowing (which would have made it uninteresting for Power and maybe Ruhlen, if growing trees is of some sportive interest, but not me)?
* Also I’m of course pleased to hear that Starostin’s influence was towards more ratehr than less serious approach:) As I said, I’m not too interested in long-range work. But he is highly esteemed here (he and Zaliznyak are our celebrity linguists) and I know some people from his family.
@Lameen, and as I began asking question here and if you will have time to answer, I have a more practical (for me) one: What do you think of Stolbova’s work on Chadic? Reconstruction, the lexical database, the etymological dictionary*.
It is not talked about in English too often. In Russia she is simply a (or the?) leading specialist in Chadic and her work won’t be questioned. And it is not talked about often in Russian anyways because not many Russians work with Chadic.
I understand that it is more reliable than Orel and Stolbova’s HS dictionary but nothing comparable to what we have for IE or even Semitic. There is an enormous range in between. I don’t understand how sceptically I must take everything. I can understand something on my own, but what I can’t control is cherry-picking forms to support false reconstuction. Chadic languages are many, and I merely want to learn more about them, I don’t know them.
* Russian title, English contents as I rememebr. The DB is in English.
@DE: The Comparative Method is wonderful, but I think you are missing the point of where Greenberg (and Ruhlen) went astray.
Mass comparison, of the crudest Greenbergian sort, works as advertised — crude data, superficial resemblances, reliance of sheer numbers — within narrow circumstances. An English speaker can tell that speakers from London, New York, and Mississippi are speaking related languages, with justified certainty, while never working out any sound correspondences. Likewise a linguist looking for the first time at poorly recorded early wordlists of Polynesian, or California Athabascan, would be right to assert their genetic relationship by inspection.
Greenberg himself gives a toy example, of a short wordlist in several IE languages, where the overall relationship and the broad division into subfamilies is evident by inspection. That is fine as far as it goes. But then he goes too far.
If you compare wordlists of several languages, and 100% of the items agree perfectly in form and meaning, you need go no further. If for each gloss 90% of the languages have a near match, you’re good. The problem comes when one further drops that percentage, and loosens what is meant by “near match”. Greenberg took it to the extreme, where in a large set of languages any match between any two languages for a given gloss would count as support for group membership. If you look for matches between all or almost all languages in your set, having more languages lessens the probability of chance matching, even if you use impressionistic comparisons. In Greenberg’s method, having more languages increases the probability of chance matches. In a sense, megalocomparison baits you with one set of assumptions and then switches to where they do not apply.
This is quite a general issue, and the logic here is the same even in systems which don’t have the equivalent of the Comparative Method, e.g. anthropological classification by cultural traits. Linguistics is lucky to have such a thing as regularity in sound change, which enables the CM, and dramatically reduces the probability of chance matches. It is unlucky to not have the equivalent for semantic change, and unlikely semantic matches are still prevalent in the literature everywhere.
There is no exact definition of “crackpot”, but to me it carries a whiff of the aggrieved fanatic. By all accounts Greenberg was a pleasant and low-key person. I do think he was deluding himself at the same time that he was deluding others.
@Y,
One good thing about mass comparision is that now, when we have computers, it can be improved (and statistical measures of significancance applied). it is easier to algorithmise.
The other is good for me, and horrible for others:) It is sensitive to arealisms:))))
But I find those interesting and I do not find classifications terribly interesting: they’re but the first step that helps us in chosing what to reconstruct and what other work to do.
@Y:
I certainly wouldn’t characterise Greenberg as a crackpot. However, crackpots can be perfectly nice people, and quiet crackpottery is, in anything, probably commoner than the loud obnoxious kind.
https://poetrywithmathematics.blogspot.com/2010/03/w-h-audens-kingdom-of-number.html
The problem comes when one further drops that percentage, and loosens what is meant by “near match”.
This is a far more general problem; Normalization of Deviance gives a good description (and explains how it causes actual disasters):
It goes on to describe what happened with Challenger:
@David E. An association between crackpottery and loud obnoxiousness, in some circles, can be seen here.
At first reading I thought that was Feynman’s voice. It’s very compelling.
@Jerry: Baez says nothing at all about loudness or obnoxiousness. It sounds as if he has been annoyed by not a few physics crackpots in the past, but he is admirably restrained about it.
Of course,
– LOUD crackpots are the ones you notice.
– OBNOXIOUS crackpots are the ones you want to insult (by telling others that they’re crackpots)
This creates bias. (I know the name of the loudness bias, but not sure how to call the obnoxiouness bias:)).
What do you think of Stolbova’s work on Chadic? Reconstruction, the lexical database, the etymological dictionary*.
I may be biased against it by having encountered HSED first, but I haven’t gotten anything useful out of it; it strikes me as overreach to try to reconstruct that many lexical items without first reconstructing the intermediate languages, and without a clearer understanding of the rather complex morphology. My usual starting reference is still Newman’s proto-Chadic. From my own work, I would say a big issue in Chadic is that you really need to pay a lot of attention to contact to be able to do reconstruction without getting swamped by false cognates.
@Stu: All-caps is usually considered loudness, and writing to physicists one doesn’t know is at least not keeping one’s physics theories quiet.
Obnoxiousness is a matter of taste. I’d say it’s obnoxious to name something after oneself, compare oneself to Newton or Einstein, or say one deserves a Nobel Prize, and more obnoxious to say that those who rebut one’s theories are hidebound reactionaries, self-appointed defenders of the orthodoxy, or a conspiracy; to compare them to Nazis, stormtroopers, brownshirts, or the Inquisition; and to fantasize about show trials in which they’re forced to recant. Also to tell Baez that his index suppresses original thinkers.
I am not familiar with Greenberg, his theories, etc., but from what I’ve encountered on LH this doesn’t seem like crackpottery at all. He took a plausible scientific idea and extended it probably well beyond its known area of applicability. But that’s how science works. (In part, science works in many ways) If he didn’t hit a jackpot, that’s just what happens with most ideas.
For anyone unfamiliar with the reference to Feynman, he (chosen as probably the most prominent physical scientist in America) was on the commission that investigated the Challenger shuttle disaster. He famously demonstrated, on live television, what happened when an o-ring (of the real materials and transverse dimensions, but only a few centimeters across) was soaked in ice water and how rigid it became. His appendix to the commission’s report is also fascinating reading.
As aggressive crackpots go, few have ever topped the International Flat Earth Research Society, which existed in the latter half of the 20th century. Favorite quote, at the conclusion of a flyer of theirs I once picked up (text here, accurately transcribed):
which existed in the latter half of the 20th century.
Not that I’m supporting their claim “The International Flat Earth Society is the oldest continuous Society existing on the world today.”, but its origins are more like 1838 [see the second para of that WiPe].
It wasn’t until after the invention of the steam locomotive that humans got stupid enough to believe in a flat earth. Nowadays we have LLMs.
If you like reading about old-style crackpots, Fads and Falacies is probably still be the best book.
But is not so-called “roundness” (or “flatness” if you prefer) is a mere question of terminology?
In reality (or as some would say, locally) it is flat.
Of course at certain distances its metric is not Euclidean, which leads to some interesting effects. For example, the so-called horizon, when our perception is fooled by the metric.
But that’s some complicated mathematical abstraction. Professionals resort to this convenient abstraction when calculating routes, but it is not a real thing, it is math and nothing more.
I came across “what is doing” in a text from the 19th century. It did not occur to me that not only houses are building, but things are doing.
I wonder what is the semantical role of “you” in “how do you do?”, is it the agent or…?
Was tut sich…
Perhaps the most intriguing sentence in the wiki description of Feyman’s role in the so-called “Rogers* Commission” investigating the Challenger disaster is “Feynman later reported that, although he had believed he was making discoveries about the problems at NASA on his own, he eventually realized that either NASA or contractor personnel, in an apparent effort to anonymously focus attention on these problem areas, had carefully led him to the evidence which would support the conclusions on which he would later report.” In other words, inside whistleblowers who knew what had gone wrong but didn’t want to get in trouble for having been the source, had arranged to carefully set out a trail of breadcrumbs for Feynman, having astutely assessed that a “wow, famously brilliant guy apparently figured it out on his own!” narrative would minimize the odds of bad actors within NASA/contractor management trying to identify and punish the whistleblowers. General Donald Kutyna, USAF (apparently still alive at age 91), was the key link in the chain (as a technically knowledgeable guy who also understood the bureaucratic politics and bad incentive structures) in deliberately putting the clues under Feynman’s nose in a way that let Feynman initially think he had figured it out by himself.
*Bill Rogers was a very well-pedigreed Establishment guy – certainly not a technical rocket-design guy but he had served as both U.S. Attorney General (under Eisenhower) and Secretary of State (under Nixon, until he got marginalized by Kissinger) and was thus a natural choice for this sort of Important Blue-Ribbon Commission.
It speaks well of Feynman that he figured that out. Being able to make the assessment that his own (very high) capabilities were, in fact, not high enough to explain how he had really solved the problem. Admirable.
Henry Petroski talked about the normalization of deviance, though not by that name, in his essay on bridge collapses. For more than a century we have had a major collapse about every 20 years. I don’t remember most of them, but1940 was Galloping Gertie,1963 was the Silver Brldge. 1982 was the Mianus Bridge, and 2000 was the Millennium Bridge (which was fixed before it collapsed).
The recent Philadelphia and Baltimore bridge collapses don’t count because they resulted from traumatic stresses. But should the I-35 collapse of 2007 be in the list?
The Tacoma Narrows bridge (Galloping Gertie) is a fascinating example of normalization of deviance. As soon as it was built, it was observed to be susceptible to resonant interactions with the wind, but the resulting vibrations were analyzed and not deemed dangerous. However, over time the stresses weakened the structure, and when the fatigue had weakened things enough, cross coupling excited genuinely dangerous torsional oscillation that tore the decking apart.
For more than a century we have had a major [bridge] collapse about every 20 years.
“we” means USA? Or does that include Canada? And do dam collapses count in the toll?
AFAICR, there’s been no such collapses in Europe over that timescale. (Power outages seem to be their gig.) Nor in Aus/NZ. (Devastating floods. Earthquakes in NZ, happening places which are supposed not to be geologically active.) China, yes; but ‘tofu dregs’ construction means stuff collapses/needs to be demolished often before it’s even put into operation; it’s not wear and tear.
AFAICR, there’s been no such collapses in Europe over that timescale.
If only… (those are just the two that came to my mind immediately; there probably were more.)
I do a lot of bicycle travel around Europe and have had to re-route due to a collapsed bridge on more than one occasion.
Ponte Morandi is the first thing that comes to mind when I hear “bridge collapse”, that was quite the scandal.
Wikipedia has lists of “bridge collapses” both in German https://de.wikipedia.org/wiki/Liste_von_Br%C3%BCckeneinst%C3%BCrzen#21._Jahrhundert and English. Interestingly the lists are not identical, so you would need to combine for the full list.
Looking at the list Vanya linked to, I will say that I do not remember any English-language media coverage (which isn’t to say there wasn’t any, of course) of e.g. the Ponte Hintze Ribeiro disaster of 2001 in Portugal, despite the quite substantial number of fatalities. Not enough dead foreigners to be newsworthy if they don’t match up with some broader narrative about “terrorism” or political strife?
There really doesn’t seem to have been much coverage of the Ribeiro disaster in the US at the time. As far as I can tell the NYT never mentioned it, there were some items on CNN and CBS. To your point, the IRA bombing of the BCC in London was probably considered more newsworthy, as was the crushing to death of 35 pilgrims in Mecca during the Hajj.
Somewhat odd, because it really is quite the story, a 19th century landmark collapses on the same day two protesters are being hauled into court for previously blocking traffic because they believed the bridge to be unsafe.
I can’t find a mention either, and I too am surprised — you’d think there would at least have been a “Bus Plunge”–style paragraph.
Well, the point of the classic “bus plunge” story was to fill up a one-and-a-half-column-inch hole that was sitting there conspicuously unplugged very late in the layout process for the next day’s edition. I think changes in the news business have made that sort of niche less relevant.
When I was 19 my friend told me that apart of Die schöne Müllerin by Müller (lyrics) and Schubert (music)* there once was competing Die schöne Müllerin by Müller (lyrics) and Müller (music), with music by a cerain then known Müller.
I’m not sure this is true, but I learn that https://en.wikipedia.org/wiki/Max_Müller is a son of Wilhelm Müller, the poet from my friend’s joke.
* Thanks to the friend, some songs from their Winterreise even occasionaly play in my mind, in Russian…
(The context is that DM’s comment about Greenberg and classifiyng speakers made me read the classification of Cust – and of Friedrich Müller, an Austrian Müller, whose large gropings are used by Cust and in Müller’s books are made a part of even larger racial groupings. When I was googling for him, I found the WP article about [Friedrich] Max Müller)
@DM, about Greenberg and “classifying speakes”. I think in the 19th century external criteria – apart of proximity – and “mixed” argumentation that takes them into account played an auxiliary role. It is not what I read in books. And why a geek would do that?
However it played some role and there are books I won’t read.
There was a decline in it, I don’t know if it’s because of Greenberg or not.
But I think 19th-century classifications are good. In 1800 Europeans simply did not know the interior. In 1900 they didn’t know parts of the interior. Of course they were careful. Classification was a way to structure the arriving data somehow. The structure:
(1) a mess to the south of Sahara, in the “Negro” land. The “Negro” group.
(2) several groups based on some principle. Hamitic, Bantu. Nuba.
There is a book from 1883 by Cust, in English (it contains an account of classifications). He thinks that given the very fragmentary knowlege, classification of fragments won’t be scientific. His task is learning where the fragments come from. His large groups are taken from Müller 1873 and 1876 (“group” means: an areal group. Hamitic is a “group”). He refuses to take into account known small groups, his groups are lands (from the river Senegál to the Kamerún mountains).
Yes, external criteria. Yes, good.
Müller makes his groups parts of the racial classification. Each of his groups (apart of the “Negro group” and Fula) is contained in one of Greenberg’s groups. Greenberg extends groups known in the 19th century into the Negro land, he classifies the unclassified group.
(“@DM, …” – I don’t mean I’m speaking to specifically DM, but DM wrote about classifying speakers and Greenberg).
Cust’s definition of Hamitic is interesting.
It is (a) spoken in a continuous area (b) has gender (c)
they have flat nailsis not Semitic.(a) too is here because Müller and Cüst with him do not count Khoi-San as “Hamitic”. If the reason for this is race of speakers or certain exotic features of Khoi-San languages like clicks – then (a) too is “flat nails” criterion. If the reason is their location – it is a real one.
And no motivation for (c), he simply says it is “of course” (or something to that effect) a bad idea to mix them up in “Sub-Semitic”, “half-Semitic” (our para-) or “Hebraeo-African”.
If you mean Cust’s A Sketch of the Modern Languages of Africa, I would recommend his extremely sensible remarks about language versus “race” near the beginning of Chapter Four. (Also some very sensible remarks about classifications based on inadequate poorly-transcribed word-lists.)
Despite this, he does eventually adopt a “race”-based classification, notably separating “Negro” (viz Guinea-zone West African) languages from “Bantu”, but gives quite a good survey of then-contemporary views. In the process, he mentions that two superb Africanists, Bleek and Christaller (who he actually corresponded with) were firmly of the opinion that the “Negro” and “Bantu” language groups were in fact related. This is in 1883. The views of people like Meinhof were actually a regression, motivated by racism, from previous views based solidly on the linguistics.
Where Greenberg was right, he was following prior linguistically-based work. Where he “extended” groups on the basis of faulty methodology, he achieved nothing of value in terms of the classification of African languages.
[I shall magnanimously ignore Cust’s gushing dedication to the criminal Leopold II. We all have our blind spots …]
Cust was an interesting man:
https://en.wikipedia.org/wiki/Robert_Needham_Cust
Says WP: “He was one of the few Victorian intellectuals to oppose the racist theories popular at the time.”
All honor to his name, which I will try to remember. Such people should not be consigned to the dustheap.
@DE, yes, this book.
The problem is: this book is NOT a “genetical” classification. The author explains why he believes
– that there is not enough data for a genetical classification of a large part of Africa, and where there is such data, this classification has not been properly done for anything but Bantu.
– that not knowing where data comes from – where langauges are spoken) is a major and pressing issue for linguistics.
And for this reason his task is mapping languages. No more and no less.
He also unambiguously explains that “group” and “family” are two technical terms, where the former means an areal grouping and the latter means a genetical grouping. And his “families” are Semitic and Bantu – the rest are “groups”.
And nevertheless, Müller’s groups which he borrows are very exactly Greenberg’s.
The “Negro” group is but a pile of unclassified languages, and he understands that.
After this thinking that people in the 19th century instead of genetical classification practiced something different “because they were silly and unscientific and their methods were racialist” would be … silly.
@DE, about Negro…
But he says it himself, that “Negro” group is heterogeneous and conveniently but illogially lumping together (this these very words “lump” and “illogical”).
And about Bantu he cites numerous scholars who think that some of those Negro languages are para-Bantu.
And he speaks about determining if any of Negro languages are para-Bantu and which exact once as an improtant sceintific problem.
Does not this mean that for him “Negro” means precisely Unclassified? He even says that the name is convenient when the area and its langauges are largely unknown – but will be changed to a “sceintific” name.
Yes, there are some – importantly Müller – who believe that
– the area of these unclassified langauges is also the area where the “Negro” race is found (but sorry, HOW do you decide who in Africa a “Negro” and who is not? Are we sure that this “race” is not a mere projection of an area?)
– that languages of “Negroes” are as much genetically isolated from others as their speakers.
In the context of Bantu Müller is the only scholar cited by him, who believes in this second thing.
But Müller also believes the area contains numerous independent families.
So thinking that Fulfulde is “Hamitic” is not “silly”?
And nevertheless, Müller’s groups which he borrows are very exactly Greenberg’s.
No, they’re just not. Müller separates “Negro” (supposed to be a language group!) from “Bantu” and lumps together “Fula-Nuba.” Serer, the actual closest relation to Fulfulde, is “Negro.”
He separates “Hamitic” and “Semitic” (which Greenberg did not); like Greenberg, he has a spurious “Hottentot-Bushman/Khoisan” group, but even in Greenberg’s day there were maybe not yet enough actual data to show the spuriosity, so a venial sin.
Müller’s “Hamitic” does not include Chadic, which is “Negro”; Greenberg, on the other hand got this right (though he did not originate the idea.) Again, given the time depth of Afro-Asiatic, one can hardly fault Müller too severely; though frankly, anyone who looks at Hausa with eyes unclouded by “race”-based preconceptions would surely be struck by its structural resemblances to Semitic.
Cust was an interesting man:
Seems to have been an interesting family. His uncle ” was a member of the Canterbury [NZ] Association which organised European settlement of the area around 1850.” And another relative was Dean of York Minster.
(I’ve lived in both York and Canterbury, so recognised the name.)
Cust was an interesting man
His defense of opium is, um, original.
I see that Gulimancema (“Gurma”), Dagbani (“Dagwamba”), Farefare (“Guresha”) and Mooré (“Mose”) get a passing mention on pp209-210 of Cust’s book. Courtesy of the remarkable Koelle, I think.
All “darkest Africa” in those days, I suppose. (The Brits only got to Bawku in the twentieth century.)
@DE, aha!
I intended top ask you who’s “Guresha”:)
It hasn’t been shown to be spurious in the sense of showing that any of its parts belong elsewhere (as has been done with, say, “Negro”). Evidence that they belong together is certainly scarce even today, but apparently nonzero.
:-S
@drasvi:
The language names as given seem to be actually all ethnonyms, so that would do for Gure(n)se “Farefare people”, and it seems to be a reasonable conjecture given the name’s appearance along with others from that area.
That may in fact be at the back of the cryptic complaint on p51 about an unhelpful dictionary entry “Guresa or Gursea: African, a dialect of the Kouri Class … See Guren”; on p209 he says that “Latham” calls the the Farefare-et-al group “Kaure”; this looks like a variant of Koelle’s “Gur.” The reference to “Guren” would then be traceable to the actual glottonym used by the speakers themselves, Gurenne.
Having actually lived in the area, it still disconcerts me a bit to see it treated as a real terra incognita in even quite recent sources. Like Rattray’s title for his (actually very interesting and informative) Tribes of the Ashanti Hinterland …
One forgets how recent the European penetration into most of Africa really was. Félix Houphouët-Boigny, first president of Côte d’Ivoire, and French cabinet minister, was three years old when the French first came to his village.
@DE, DM, we sure keep in mind this:
When criticising their GROUPS?
And this
and this
>One forgets how recent the European penetration into most of Africa really was.
Or one’s mind is actively misled by the modern liberal belief that the slave trade, colonialism and effective European political control of Africa is a single four- to five-century dynamic of inescapable domination and cruelty, rather than a faltering effort often secondary to local power networks.
Not to underplay the degree of cruelty nor absolve European participants. But it’s important to recognize that the footprint was shallow for centuries, and the deeper impact lasted less than a lifetime.
@drasvi:
Is that from Müller?
The bit about “Negersprachen” strikes me as not so much a praiseworthy confession of ignorance as an attempt to run down the various languages as not even being capable of being compared historically, unlike Bantu: essentially, beneath the notice of comparativists: they must have all originated independently, what with being so formless and chaotic and all: they can’t possibly be related to any respectable languages.
Actual experts (like Christaller) knew better even then. As did Westermann later. This writer doesn’t want to know. Negersprachen? Pfui!
The passage about Khoisan certainly bears out my point that nobody at that time knew enough about the languages in question to be able to express a sensible opinion about their relatedness or otherwise.
Müller (if it be he) was no Africanist:
https://de.m.wikipedia.org/wiki/Friedrich_M%C3%BCller_(Sprachwissenschaftler)
I note that Vol III of his Grundriss der Sprachwissenschaft is entitled Die Sprachen der lockenhaarigen Rassen. Do you really want to maintain that all this is geography and praiseworthy confessions of ignorance rather than based on “race”?
@Ryan:
No. The effects of the slave trade were profoundly damaging over a much greater area than was directly controlled by Europeans at that time, and over several centuries. Exploitation need not involve troops on the ground. It’s much more cost-effective when done by remote control.
The Scramble for Africa made little sense economically – which is ultimately why direct occupation was a quite transient thing in the end.
Decolonisation has not ended the exploitation. This is not some wet liberal mea culpa on my part: it’s the way things actually are. There are still African power elites personally benefiting from it, just as there were during the slave trade.That in no way means that Africans are not being exploited by outsiders.
How long do you think these elites would last without foreign economic (and military) support?
The use of the terms Family and Branch, or Group and Sub-Group, has its special application. When the Languages are assumed to have a common origin, like the Semitic and Bantu, the first term is applicable, but where there is no certainty of such common origin, and the subdivision is only one of convenience, then the latter terms are used.
I had a disagreement once with someone about this point. Can’t one say “The Romance language family” instead of the unwieldier “The Romance subgroup of the Indo-European language family”?
“Phylum” seems to be the current hip locution for “set of languages I want to imply are genetically related, without actually demonstrating as much.”
“Branch” in this sense seems to be an idiosyncrasy of Cust’s, and in fact he doesn’t actually use it consistently.
On the face of it, it’s a bizarre usage: branch of what, exactly?
I think what’s going on is that he’s trying to have his cake and eat it too: working with Müller’s “race”-based overall scheme, which he is rightly uneasy about (for reasons he actually explains pretty well) but giving this odd technical meaning to “branches” within Müller’s classification so that he can still claim (again, reasonably) to be agnostic about whether these subdivisions are actual genetic groups.
It was unfortunate for Cust (who would probably have enjoyed the Hattery) that despite the work of real African-language experts like Bleek and Christaller, by this time the accepted wisdom on African languages was already being affected by racist conceptions of African cultural history, as ultimately seen in Meinhof and his ilk. I imagine that Cust himself would have been delighted to discover that it was all bollocks, but he was a well-informed amateur in this area, not a professional, and can be forgiven for not having the intellectual courage (or arrogance, maybe) to stand against the Zeitgeist in this domain.
African leaders seem able to change patron nations. European countries don’t today seem to have a great capacity for changing African leaders.
@DE, I think the word “classification” has a meaning – and this meaning is “placing in classes according to some [specified in the context] criteria or principle”.
And I think you use it very differently. You mean a hypothesis about trees. Which is merely one way of classification.
You use it differently because everyone does. And the fact that everyone is doing so follows from two things:
– one branch of linguistics is much better developed than many others.
– this branch works with a certain real thing that has (in reality) the structure of classes (namely: trees).
Typology and areal linguistics work with different structures which (in reality) are not trees – and thus a classification of languages won’t help typology or areal linguistics that much.
However, when all you have is Semitic and Bantu – I mean, these are only trees around you can speak about with confidence – this principle of classification does not help much.
And when you need to structure SOMEHOW either your book (that is when all you need is to list languages, without meaning much by this classification) or your theories, different principles are used.
And this does not mean people who use them are idiots.
Or racists*. A racist is not someone who studies people or peoples or uses the word “races” instead of the word “peoples”. Not even someone who studies wooly hair. A racist is someone who treats people differently based on whether the race they belong to has wooly or fleecy or tufty hair.
*which is not to say that racism was not the norm and that Müller is not a racist. I only mean: it does not follow from his classification of hair.
Müller compiled this classification for a book about peoples, not languages. from 1873.
He re-used it in his book about languages in 1876.
He very precisely classifies peoples based on their hair. This is not “bad linguistics”. It is NOT linguistics.
But this gives him only three groups of races in Africa, one of them Khoi-San. He further introduces “Neger” and “Kaffern”, “Nubas” and “Mittelländer”. 5 “races”
Of course he intends to make this is a racial classification.
I have doubts.
Kaf(f)irs speak Bantu. Bantu-speakers are Kaf(f)irs. For example, did Europeans first noticed the difference between bodies of Kaffirs, Negroes and Nuba and only then – by lucky coincidence – it turned out that all Kaffirs speak Bantu?
Are Oromo and Tuaregs so similar?
Oh, interesting – that’s a clear statement against “Khoisan”.
That’s not in the quote. It just says “no data, no evidence, nope”.
Does Russia count as “European”? La Françafrique hasn’t so much ended as been replaced by Wagner.
Does Russia count as “European”? La Françafrique hasn’t so much ended as been replaced by Wagner.
Sure, but I assume that counts as clients changing their patrons, not as patrons changing the leader of a country.
No doubt the freedom of military juntas in the Sahel to go to Russian mercenaries rather than French soldiers to support their rule is a true emancipation from oppressive colonialism for ordinary Africans. Nkrumah would have been thrilled.
Well, I’m definitely NOT sayig that Müller is (a) good linguist (b) not an idiot (c) not a racist.
I think Cust‘s book is good. I have read only a few selected places (and only linguistics, so I have no idea what he thinks or feels about people as such) in Müller…. and he does not say much in those places.
All I understand, he is an UBER splitter.
21 (!!!) small group in “Negro” – and he thinks many are historically independent. (a stronger claim than “no data” which is rather Cust’s point – but a speculative one).
But Cust borrowed his groups from Müller, and my claim is that if we treat “Negro” as “Unclassified” – and it does follow from Cust’s words that it IS “Unclassified”. Besides “Negro” refers all which is spoken in a region largely unknown to Europeans – groups “Bantu”, “Hottentot”, “Hamitic”, “Nuba-[without Fula]” are very precisely Greenberg’s groups.
That is racial or not (I don’t think “Bantu” and “Hamitic” are racial, I think “Khoisan” is both racial and linguistic (and also naturally arises as “not Bantu”) but have no idea about “Nuba”) – but somehow they came up with same groups.
To what @Y said about “mass comparison”.
That is simply because we DO use, constantly, one external criterion. Proximity.
But we don’t speak about it very often (which is not good).
Interpret any statistical measure as conditional probability (Bayesian reasoning).
You first form a priori hypotheses based on distance. To every link between any two languages you assign a likelyhood (based on distance, known history of human movements, natural obstacles like mountain chains).
Then modify it based on the degree of similarity, using the expression for conditional probability.
I don’t think it is anyhow important whetehr you obtain this “degree” statistically or manually. What you do manually – I mean normal methods – is still statistics, but more complex and… vague, intuitive.
Your normal reasoning does not give you measurable likelihoods – and this is the Big Problem of historical linguistics. It is not “good”. It sucks.
You degree of similarity will be a parameter.
And if this is confusing (why linguistics gives you a mere parameter?) – hey! A hypothesis about a language tree is always a hypothesis about SPEAKERS.
Normally you solve this by working with Very Large degrees of similarity.
You don’t care if the a priori likelihood is small: your measure of similarity compensates for that.
But I’m not sure this works for, say, Afro-Asiatic.
Your normal reasoning does not give you measurable likelihoods – and this is the Big Problem of historical linguistics. It is not “good”. It sucks.
Why pick on historical linguistics?
What is the “measurable likelihood” that the earth is round and not flat? The arguments in favour of the former are not in the least statistical, neither intuitive or “manual” nor otherwise.
There are vast areas of perfectly sound science where statistical methods have no place at all.
Case in point from comparative linguistics:
Mooré tʋ̀me ‘send, work’; Buli tom ‘send, work’; Yom tə̄m- ‘send’; Mbelime tōmū ‘send, work’; Miyobe (imperfective) tùm ̀ ‘send’, Samba Leko tùm ‘send’, Gbeya tomá ‘send’, tom ‘work’, Twi soma ‘send’, Proto-Bantu *tʊ́m- ‘send/work’ … (cf Swahili mtume ‘apostle’) …
OK: what is the “measurable” probability that these resemblances are due to chance?
There is no sensible way to answer this question: you can come up with virtually any number you like by fiddling with what parameters you count as mattering, how big a pool of languages you pick to look at …
Happily, there is also no actual need even to ask the question. The way you go about studying this is not by statistics but by verifying that, for example, the initial s of the Twi form is not isolated but a regular correspondence (cf Twi ɛsã ‘three’, Mooré atã) … looking for evidence that this ‘send/work’ word might in fact be some sort of Wanderwort … all the usual comparative historical stuff.
Mass comparison, too, is different in kind from this, not just a messier or alternative version of the same thing.
@DE, because you have data from numerous languages. Also you have theories about processes in languages.
You can build a model based on those and calculate likelihoods in the model.
@DE, Hausa: Müller thinks it is “Hamiticised”.
Cust provides a linguistic argument for this:
1. The Most Important thing about Hamitic is gender
2. Hausa has gender
3. Similar languages of the region are genderless.
Müller says nothing.
>the freedom of military juntas in the Sahel to go to Russian mercenaries rather than French soldiers to support their rule… Nkrumah would have been thrilled.
Oddly, you seem to be sarcastic. Yet surely Nkrumah would smile to see them following his model – indispensable rulers who become frustrated that people fail to see the benevolence and wisdom of their plans, and find it necessary to suppress dissent, build a cult of personality, turn to autocratic countries for support and name themselves Presidents for Life. He was truly influential — his model now reigns in much of the continent.
There are vast areas of perfectly sound science where statistical methods have no place at all.
My rule of thumb is that, if you need to resort to statistics to see whether a result is significant or not, you should assume it isn’t. Very 19th century of me, I suppose; but it would spare us a lot of unreliable papers.
I think Cust‘s book is good
I don’t think a modern reader would be well advised to rely on it …
I agree he’s an entertaining writer, though, and it’s hard not to like his evident enthusiasm for the subject. I was too dismissive in calling him an “amateur”, too: he’s assembled an enormous amount from his sources, and mentions corresponding with several actual Africanists. Much of what he reports is pretty solid stuff.
He makes some sensible points on the linguistic side, as I said. As you say, he’s a splitter by temperament, too, which was even more sensible in the context of African languages in his day than it is now.
The thing about Hausa was more or less the received idea in his day (and much later: Meinhof and his like, too.) Notice that, relatively enlightened though Cust is, he can’t get round to the idea that Hausa shared these features with “Hamitic” due to shared inheritance: it has to be Hamitic influence. I notice, incidentally, that he says “F. Müller groups it among the Negro Languages, and the question still being an open one, I follow him.” Should have followed his own instincts instead. (He turns out to be quite a fan of Hausa qua language.)
To be fair, it still surprises some moderns that black Africans might speak languages related to Egyptian and Semitic, and even now provokes genetic studies to “explain” the supposed anomaly.
Cust is largely free of genetic-determinist-type racism, though, which is not bad for the era. He’s well into the ranking of cultures on the contemporarily approved scale, but that was the norm at the time. Unrealistic to expect too much.
Lameen would no doubt appreciate
Myself, I feel a bit miffed by this, as I’ve always supposed Welsh to be the rudest and most degraded of Languages (owing to the admixture of tribes, and the predominant influence of the Latin.)
I’d forgotten the name “Quorra” for the Niger river in Nigeria, from the days before Europeans discovered it was all one river. Turns up a lot in Cust.
@Lameen, thanks! (about Stolbova) I delayed my responce because I was trying to understand what of her reconstructions is hers and what is based on what. (what bothers me is reconstructed historical changes).
The opium essay, at a glance, shows Cust as considerably less enlightened about the Chinese than about the Africans, presumably because he hadn’t spent time with the former as he had with the latter. His argument for continuing opium cultivation and trade in China goes along the lines of: the Chinese Empire is doing just fine, even with opium; we British are ones to talk, what with our liquor drinking (with which we have ruined the Africans); and, why abandon a valuable and profitable crop?
His enlightenment displays some pretty startling limits even in the African languages book. Not (alas) atypical is
(Note that this is, in fact, him opposing genetically-based racism. One just feels …)
I don’t have a great sense of how much that was just the cultural norm for his time and place; most of the stuff from that period dealing with Africa that I’ve actually read is by people like Christaller who were probably far from contemporary norms in this. I don’t think that that is unrelated to their frequently better insights into language relationships, incidentally: racism impairs one’s understanding.
Cust seems to have had a difficult relationship with his daughter
https://en.wikipedia.org/wiki/Albinia_Wherry
in a way that suggests that he may have been something of an ornery customer, though the sources seem rather one-sided. She had (understandably) something of the mishkid thing going on, by the look of it.
@DE, there is a thin book by a German author where he is (mostly) retelling the story of Barth’s travels. I read it, in Russian. It is exceptionally boring.
Was a funny feeling when I read the author’s explanation of why Barth’s books did not become popular – or influential. He says, that’s because Barth’s books are VERY boring.
And no, of course Cust is not an amateur.
Or else you define science (that thing one can be professional or amateur in) very differently from me. For me it is in the noble sense satisfying your curiousity, in a less noble sense expanding the humanity’s knowlege.
He simply does not do historical linguistics. He refuses to do it. He thinks doing it with languages of the Sudan when much of this Sudan is not known to Europeans won’t be very scientific. (Which does not mean there was not enough material for some reconstructions, but even today such work is done with subbranches of subgroups, not with “languages of Africa”). And he finds the fact that Europeans don’t know where many languages are spoken a pressing issue, so his work solves this problem. This problem is important.
Moreover the criterion of “proximity”, while external IS very important for historical linguistics.
I find this (a) professional (b) sceince in the less noble sense.
Barth was a great man for details, apparently. Not everyone shares this taste …
Thinking about the Niger river led me to look at the WP page for
https://en.wikipedia.org/wiki/Mungo_Park_(explorer)
whom I was duly taught about as a kid in Glasgow.
He apparently was the first to refer to the somewhat-immaterial
https://en.wikipedia.org/wiki/Mountains_of_Kong
which actually get a mention in the Cust book when he’s discussing Mooré etc. It did perplex me …
That last Wikipedia article is a delight. I’m perversely tempted, though, to stick [citation needed] after “None of these mountains actually exist.”
Better yet:
I just checked my New Reference Atlas of the World (C.S. Hammond, 1930), but alas, no mountains. It does, however, show a Boboland in that general vicinity (just west of Upper Volta).
because Barth’s books are VERY boring.
I object! A keen observer risking his life and health in places no German had ever been before, and noting down everything he hears or sees – how can anyone possibly read this and be bored?
(goes to Project Gutenberg and rereads some…)
… Ok, maybe you do have to be unusually interested in the Sahara and Sahel to enjoy this. But there are so many fascinating details!
Here’s how he gathered his Emghedesie vocabulary – not quite best practice, but without it the language would be entirely lost now…
@Lameen, I am interested (not “unusually” maybe, but yet) in both. My freind was disposing of a part of her library, I took this one and also a book about adventures of Czech herpetologists and conservationists in Algeria in more recent times, 60s and 70s (the time of collaboration between your country and the Soviet Block). Adventures in the company of snakes and lizards but also people and feneks. It was a couple thousands times more interesting (despite my general lack of interest in herpetology):)
Details were the only thing that kept me from falling asleep, so when reading about boring Barth overloaded with them I was not sure if I should believe the author. And I can’t remember a single travelogue that I found boring. One would think that this also apples to any attempt to retell a (believed to be boring) travelogue in a shorter, more accessible and interesting form….
However the author succeeded in making me Barth’s admirer (and would be pleased to know he succeeded in that) and generally interested in such explorers. I envied you when several years ago I was passing by your blog and noticed that you worked with papers of… I think it was Richardson.
Emghedesie previously on LH (…actually the first Google result for “Emghedesie”, which was surprising).
Françafrique with a cedilla? I’d have guessed at a /k/ there because my brain starts from Latin, but of course the word was constructed long after it became /s/ in France.
Francafrique might be the currency union.
The French government halved the value of my bank account deposit in CFA francs in Togo overnight without warning in 1994.
I can understand a certain bitterness regarding the high-handedness of the French on the part of West Africans.
Though the arguments regarding the CFA franc are not straightforward, by any means:
https://www.brookings.edu/articles/how-the-france-backed-african-cfa-franc-works-as-an-enabler-and-barrier-to-development/
“I was informed at Oran, in Algeria, by a person, whose opinion I value, that in that Port a Patois is forming of Spanish, French, and Hebrew, with a French basis.” Cust.
It seems that recent work (by McAlpin and Kobayashi , plus upcoming work by Basu) no longer supports a North Dravidian clade. Rather, Brahui and Kurux+Malto are seen as separate early offshoots.
(tattling) George van Driem says Greenberg applies racial notions to langauge groupings uncritically!
Greenberg can be accused of uncriticalness in quite a number of areas, but that one seems a rather bizarre one. Where does GvD get that from?
“Amerind” seems a better bet than Greenberg’s African classification, though.
Or maybe Indo-Pacific?
Mind you, baseless though those are, I was under the impression that Greenberg arrived at his uberlumpery by honest linguistic malpractice rather than racial notions.
[I have a book on my shelves that used to belong to van Driem. Small world …]
The exact words are
“Racial notions have continued to be uncritically applied to language groupings.”
(Languages of the Himalayas, volume 1 p. 139)
From a reference to the WP article about Ruhlen.
Without further context, that doesn’t explain your reference to Greenberg.
I know, but the quote is googleable and I think it is better to read it on the page in Google Books than torn out of context (as in the letter where I read it* – and as any larger chuch I’ll quote is goign to be).
* when reading the letter I though that “race” here is only to create a bad association, which is a ratehr dirty trick. But on the page he is talking about race, mostly.
About typological criteria, Cust: “I agree with those Scholars, who place but slight value upon resemblance of words, the lists of which have been made by different persons, often unskilled, and not using sufficient care to secure accurate Transliteration and reduction of the word to its simplest form.”
Also: “The time may come, when we shall know something of the Grammatical Structure of these Languages, and we have a warning from Lepsius, that in Africa the Vocabulary of an African Language is undergoing constant modification.”
Indeed, I remember Lepsius saying something about uselessness of vocabulary in estabishing links. This is funny. Accents are slightly different for Russians and for most people here in our respective understandings of the comparative method – as we already know from work of long-rangers of the Moscow school.
English-speaking linguists place somewhat greater emphasis on morphology.
And then somethign similar in the 19th century.
Of course no one will say that vocabulary is unimportant for reconstruction, because how you obtain phonetical changes without diverse lexicon? And I personally not sure what claims made without reconstruction mean.
Yet for this stage, when you don’t have a reconstruction, these two find vocabulary almost useless – and rely on what we now call “typology” (and within typology – on gender) and on morphological markers too when available.
Yeah, shared inherited morphology is the clincher for demonstrating a genetic relationship, but you can’t actually demonstrate that morphological resemblances are due to shared inheritance unless you can relate the form of the morphological elements by regular sound changes, and you can’t do that without vocabulary comparisons.
A lot of the supposed evidence for the more ambitiously inclusive forms of “Niger-Congo” relies on supposedly shared morphology, but with the regular formal correspondences not actually demonstrated, so that in reality all that’s being compared is lookalikes. That’s especially true with verbal derivational suffixes (usually called “verbal extensions” in this context, after Bantuist practice.) That matters, because along with the noun-class systems, resemblances in verbal derivation are usually regarded as the clinching evidence for the true genetic unity of the group.
As lots of “Niger-Congo” subgroups have quite a range of morphological affixes among their many and diverse individual languages to pick and choose among, finding chance lookalikes is pretty much as easy as with lexicon.*
It doesn’t help that the data used are often of very poor quality. When “Gur” appears at all, it’s usually via Gaston Canu’s astonishingly unreliable Mooré grammar, which goes into much more detail about derivational morphology than most. Unfortunately, Canu seems to have simply made up a lot of the forms he cites, and to have misinterpreted most of those which he didn’t simply invent.
The idea that comparative typology can substitute for comparative historical linguistics is one of those zombie ideas that no amount of clearcut refuting evidence seems able to kill. Johanna Nichols’ work is full of it. It plays a large part in the more maximalist “Niger-Congo” constructions, too, particularly with the inclusion of Kordofanian, but also with Atlantic.
* Kusaal Kʋsaa “Kusaasi person”, plural Kʋsaas. Baa “dog”, plural baas. Clear proof that Kusaal is closely related to English … you may think that’s silly, but I’ve seen equations every bit as stupid as “Kusaal plural -s = English plural -s” in serious publications purporting to reconstruct the noun classes in “proto-Niger-Kordofanian.”
@DE, Greenberg 1949 is largely lexicon.
” It should be borne in mind that for most of the languages all that is available is a short vocabulary of some hundred words and a few phrases, in some cases vocabulary only. This also accounts for the lack of prominence of verb forms. For many of the languages, the equivalents of even such common concepts as ‘eat,’ ‘drink,’ ‘give’ are not known.”
I haven’t read Westermann but it seems this is a serious change in method compared to this belief of unreliability and even uselessness of vocabulary expressed by Cust and Lepsius. Also Westermann 1911 studies only selected tokens: 5 well-known langauges from the west of Sudan, 3 from the east.
“The idea that comparative typology can substitute for comparative historical linguistics”
No, I don’t think THIS is the belief of the 19th century. What is “reconstruction” was well known, and that it is NOT typology was well understood.
Please remember how Cust defines a group, and that he states very clearly that Hamitic (and Semito-Hamitic) is NOTHING more than a [largely typological] “group”. This confusion can’t be ascribed to him, and I’m not sure if it can be ascribed to anyone. And an argument over classifications without reconstructions is a bit silly I think. Without reconstructions all you have is guesswork. While reconstructions are impossible to confuse with typology.
@DE, why stupid? I mean, can’t we use English -s as the translation in interlinear glosses?
He’s not talking about translation, he’s talking about equating them in terms of historical linguistics (common descent).
Quite so.
can’t we use English -s as the translation in interlinear glosses?
OK:
Ba anɛ biis.
they be.FOCUS childs
Ba anɛ pɛ’ɛs.
they be.FOCUS sheeps
Ba anɛ kuus.
they be.FOCUS mouses
@LH, he did not say it directly. His words only hint at this mistake. And he is speaking about “serious publications” with reconstructions. I still hope he meant something else. Something compared to it, but not this mistake.
(and I don’t know who of us two you intended to help but neither your responce not DE’s vague “quite so” do not help:))
I’m not sure why they don’t help. Perhaps you could refine your question, or make it clearer.
the quote is googleable
Ah, WP. On “Indo-Pacific Languages.” Indeed one of Greenberg’s wilder ideas. However, WP also says
“Greenberg’s proposal was based on rough estimation of lexical similarity and typological similarity.”
Are you privy to information showing that this is not actually so, and that Greenberg was really basing his classification on “race”? I have no access to Greenberg’s actual 1971 paper. If he was not, any implication by van Driem that that was what he was really up to would be an unwarranted slur, but I think you’re actually reading too much into that excerpt, and that no such implication is intended.
Rather, van Driem (and Roger Blench) seem to be invoking the completely irrelevant point that Greenberg’s proposal in fact coincides (so they say) in scope with previous “race”-based classifications. That has no bearing on its validity either way: the (only) proper way to critique the proposal is to point out its weakness on purely linguistic grounds.
If Greenberg’s methodology had been sound (which it was not), the fact that it coincided with a previous classification based on “race” would not have invalidated it in the least. In fact, the coincidence might even have been interesting – if you could find some independent way of attaching any useful meaning to the term “race” in this context, at least.
he did not say it directly.
OK: here I am, saying it directly. In that footnote (indeed, in the whole comment) I was referring, specifically, to historical comparative linguistics, and the “serious publications” I had in mind are specifically about the reconstruction of proto-Niger-Congo.
I am not attributing the wrong turn that African linguistics took, under the influence of racism, to Cust, who says many sensible things about the linguistics, and was apparently remarkably free of genetic-determinist racism for that period. He should have followed his own common-sense more that accepting what was becoming the mistaken consensus, but he does several times express his own misgivings about it. Good for him.
It does not follow from Cust’s individual intellectual virtues that there never was any such wrong turn. This wrong turn is abundantly documented.
Greenberg’s African-language classification was more successful than (perhaps) it should have been, precisely because it was at least a welcome return to the practice of deciding linguistic questions by linguistic methods (flawed though the methods were) rather than by appeals to (DNA-based) genetics which would have been irrelevant even if they hadn’t been based on pseudoscience and racist ideology.
@DE, then could you please tell me, WHO is that in van Driem’s opinion who CONTINUES to uncritically apply racial notions to langauges?
Sorry, mere “coincides” is impossible.
@DE,
Müller’s Negro is a list of languages spoken by what he thinks is a race (“Negro”). But there is no confusion here!
Lepsius is a caricature of a 19th-century scholar:
Die Schnalze, dieser charakteristische Ausdruck sprachlicher Indolenz und Verkommenheit, sollen bei den Buschmännern noch mehr vorwalten als bei den Hottentotten.
Da der Mann die Sprache bildet, so geht die Unterscheidung der Geschlechter von der Aussonderung des Femininums aus, daher wir dieses vorzugsweise ausgebildet finden.
Alles was vor und neben dieser neuen dreigetheilten Völkerquelle von andern Stämmen und Rassen lag und noch liegt, gehört zu dem vorgeschichtlichen Völker-Humus, der ohne eignes Licht und Wärme, nur von diesen allein geschichtlichen Völkern bestrahlt, erwärmt und verwendet wird.
But his groupings are based on his Theory, namely that Bantu is African, Hamitic is invaders and everything which is neither Bantu or Hamitic is the outcome of contact between Bantu and Hamitic.
He groups Hausa with Berber in “Libyan”.
His Khoisan is Hamitic and his Maasai is not based on this
im Oigob Mask..l..Fem..nim Aegypt..”...p,f.”...t,s
im Beya....”...b...”...t,s
im Hottent.”...b...”...t,s
Lepsius is a caricature of a 19th-century scholar
In fact, Lepsius was a major and (rightly) respected scholar.
https://en.wikipedia.org/wiki/Karl_Richard_Lepsius
What you seem to be missing is that this kind of thing was mainstream at the time. It wasn’t a few marginal crackpots coming out with this sort of stuff: it was the consensus. It was the dissenters who were marginalised. And most people just went with the flow.
Schnalze
Apropos: read what Cust says about clicks. Cust was too sensible too rank people by “bloodlines”, but he is perfectly happy to rank both cultures and languages. He comes out with some quite bloodcurdling stuff.
I don’t think Nichols suggests typological connections are indicative of a genetic relationship, just of something interesting to investigate, which can indicate some past mixture of genetic and areal connections. I like the conclusion to her and Peterson’s rejoinder to Campbell regarding the distribution of n/m pronouns in the Americas (Language 74(3), 605–614, 1998, p. 613, here): “Something happened. We do not and cannot know just what happened, but this does not preclude establishing when and where.”
@DE, why do I “seem” to be missing it?
Lepsius ENJOYS saying such things. He perceives a vertical structure in everything: men, women, peoples. And loves speaking about it. (Likely he sincerely believes that the main role of the woman (and one assigned by God to her) is to be the man’s helper).
Even his Fulah are above both hamites and bantu because that’s what often happens to mixed people (I wonder if he would also think that children of a Gemran woman and a Negro are going to be more talented than he)
It is simply not what I usually read in grammar books (the preface to Nuba grammar).
Remember your words about “mainsteam” when and if you find something dubious in a Meinhof’s book, but I’m afraid it is going to be a book aimed at German administrators and “something dubious” will be that different races have different talents (but perhaps with power of reason ascribed to his race, I don’t remember. Still WITHIN the European tradition of ascribing “talents” to peoples of Europe).
“read what Cust says about clicks”
if you mean “belong to the brute” it is not him, it is the voice crying in the wilderness:)))))))))))))
@DE, I know that snobbery IS mainstream. Still not all scholars express it so often and not all Love this vertical structure. I laugh at him because he’s funny.
If someone THINKS it but does not speak of it, or even does not love the idea – I don’t laugh.
It is not simply easier to be progressive when “everyone” is and when your progressive ideas are not simply “crazy” for everyone. MUCH, MUCH easier.
Of course most people here would support our war with Ukraine if you were Russians (to the extent such “if” can make sense).
People don’t only fear being “crazy”, they don’t want to be lonely in their thoughts, and adopt and analyse everyone’s model of the world even when they know it does not match the reality.
And don’t tell me that Lepsius is not unusual. He IS.
And yet, grouping Hausa with Berber as he and Meinhof do goes against the flow of racial thought. The criteria used for Maasai and Khoisan are linguistical (and how on earth Khoisan and Berber speakers same? The MOST exotic peoples of Africa in European perception – and the EUROPEAN people of Africa in European perception)
“Hamitic” and “Bantu” are two entities defined linguistically.
Seriously, read Lepsius and then read Müller or Meinhof or Cust and count how many times you react emotionally [with something else than “agreement” for I sometimes agree with Cust] and what is this reaction. It is going to be more frequent with Lepsius and he is so shameless that the reaction is going to be ” urge to giggle”.
@Y:
Nichols teeters on the brink, and sometimes falls over …
Not relevant to the point particularly, but I was struck (inevitably) by this bit in her (and Peterson’s) apologia: “Afroasiatic, mentioned by Campbell, is older than a stock; it is a proven family but not securely reconstructable.”
This is different from Linguistic Diversity in Space and Time, from which she says the classification is mostly taken: in that work, Afroasiatic is too a “stock.” So too is Niger-Congo, including Mande (and Atlantic, of course): of interest, because one of the many reasons for doubting that Mande is really related to Volta-Congo is that the personal pronouns really don’t show any connexion at all … (whereas Afroasiatic pronouns do show pretty clear relationships with each other, even at that breathtaking time depth …)
In the apologia, Nichols says “worldwide genetic classification into stocks, which is the level of relationship relevant to our sampling technique, is well advanced and eminently reliable” outside New Guinea and South America.
It really isn’t.
And her vaguely-formulated and over-adjustable stock/family terminology in itself blurs a distinction which ought not to be blurred.
(In saying that Afroasiatic is a “family” but not a “stock”, Nichols is in fact inverting the usage of Linguistic Diversity in Space and Time, but perhaps that can be passed over as a mere lapsus calami. It seems that she herself has some difficulty with keeping these terms straight. I sympathise. It did make me go back and check that I hadn’t completely misremembered her definitions in LDTS. Nope.)
I like the specification of “outside … South America” for the implication that by contrast we’ve got North America all cleanly sorted out. Maybe for a sufficiently idiosyncratic definition of stock-which-is-not-the-same-as-family we do, I suppose.
Well, OK. Nichols definitely has a weakness for grand statements (and what is a “stock” vs. a “phylum”? “Emperor” vs. “King”?) But the sentence I quoted is, in principle, thoughtful and right.
(As to the specifics of American n/m pronouns, it’s more complicated than that, but what isn’t.)
From the very little I’ve read about Nichols’s work (I’m not sure I’ve read anything by her other than a few pages about Chechen), it seems she takes for granted that the Comparative Method has a hard time limit of 6000 years or so, at which it suddenly stops working, so we have to use another method if we want to reconstruct through that barrier; and then she offers certain typological traits as possibly super-stable and therefore useful for this purpose. In short, The Wrongering.
Ah, I didn’t know daher had ever meant “because”. Another example of how much Standard German has changed.
The notion that there is a meaningful distinction between a “family” where you’ve reconstructed the protolanguage, and a “stock”, where you have (in some underspecified way) demonstrated a genetic relationship but not been able to reconstruct the protolanguage, is incoherent and unworkable.
Firstly, you can’t actually demonstrate a genetic relationship between languages at all without some reconstruction, even if it’s as unsophisticated as saying “all these words for ‘tree’ have roots beginning with a alveolar consonant – surely that is likely to be more than sheer coincidence?”
That implies a reconstructed protoform beginning with an alveolar, even if you haven’t formulated an actual hypothesis to that effect in so many words. Even mass comparison implicitly assumes some features of a hypothetical protolanguage in this way.
If you have no such possible correspondences, you have no business talking about the languages being related genetically at all. You’re just misusing the term.
Next: how much of the protolanguage do you need to construct before your language group qualifies as a “family”? How accurately? In how much detail?
Even with PIE, intensively worked on by geniuses for centuries, with vast amounts of data, some of it closer in time to the protolanguage than to the present, nobody could really speak the language comprehensibly or well enough to chat up a Yamnaya girl at a party or dissuade her brothers from beating them up afterwards. Whatever Schleicher thought. (She would be unimpressed by the fact that your entire vocabuly consisted of only 1600 words, for a start, even if she could actually understand any of the ones you were trying to pronounce.)
My reconstruction of proto-Oti-Volta is the best that there has ever been. It’s not very good. Does that mean that Oti-Volta is only a “stock”? Perhaps it will become a family if I work a bit harder …
Nobody has ever reconstructed proto-Volta-Congo (let alone proto-Niger-Congo) by anything even approaching adequate comparative methods. We don’t even know how to say “I ate it.” I think it is highly likely that we never will. But the evidence that Oti-Volta (say) is related to Bantu is, in my ultrasplitterist view, absolutely conclusive.
So is Volta-Congo a “family” or a “stock”?
Who cares?
My (possibly unworthy) suspicion is that the real point of introducing the term “stock” in this way is to enable assertions of genetic relatedness to be made even in cases where the real evidence is very much weaker than with Volta-Congo. To blur the line, in fact. Or rather, to move the goalposts.
Volta-Congo is a broth not a stock.
It’s a Gallerte.
Science, including historical linguistics, is a human and social practice. Therefore, I suggest:
— “Family”: Something a responsible, conscientious adviser would allow their student to work on for their dissertation (e.g. Semitic, IE).
— “Stock” / “Macrofamily”: Something that a tenured professor could conceivably work on without loss of reputation (e.g. Afro-Asiatic, Macro-Ge).
— “Phylum”: Something said professor would work on in their spare time and publish only after they are retired, not caring about insinuations of mental decline (Amerind, Nilo-Saharan).
The boundaries are of course fuzzy.
At last, a workable definition!
A social practice that makes you care about careers more than about truth?
Hostile to what I understand as “science”.
@DE, what you said about reconstruction is similar to what I say once in a while. Except I say it as my objection to everyone:)
“Family”: Something a responsible, conscientious adviser would allow their student to work on for their dissertation (e.g. Semitic, IE).
This creates the opportunity for a meta-study: By counting dissertation topics, one could determine what fraction of professors are responsible and conscientious.
Ah, but then one would need a separate definition of “family.” Which is Absyrd.
Y has cracked the problem. This is as momentous as Popper solving the Problem of Induction. Momentouser, even.
The linguistic sense of “stock” was apparently introduced in English by a translation (by Elizabeth Sabine) of Humboldt’s Cosmos:
The original German:
Doesn’t sound to me like “family”/“stock” were supposed to mean anything different there, it just sounds like elegant variation.
Elizabeth Sabine translated several scientific works “under the superintendence” (it says on the title page) of her husband Edward, a scientist and Arctic explorer. Other terms that she introduced into English include sastruga (snow ridge formed by wind on polar plains, from Russian) and wind rose (meteorological diagram).
“clincher”
@DE, as I understand you:
0. you find (somehow) promicing candidates – pairs of languages, pairs of their elements and pairs of segments – for comparison.
1. you reconstruct the sound changes based on vocabulary
2. you find that your reconstructed changes work as planned in… er, various subsystems with rigid structure.
3. you say “2 is the most important step!”
I’m not sure I understand Step 3 here.
Even if you are certain (and this needs some empirical basis, no?) that those subsystems* are less borrowable than “basic vocabulary” or whatever, Step 1 is still very important and moreover, without it your reconstruction is much less rigorous….
*of course they are highly useful for the Step 0 which is the least rigorous of all (and doesn’t even have to be rigorous)
it just sounds like elegant variation
Definitely. In fact, for the rhetorical figure* to work, I think it has to be interpreted like that in context: same with “race” and “nation(s).”
* I wanted to say “chiasmus”, but it’s not that. It’s a sort of semantically crossed parallelism. I’m sure it must have a proper name.
“abundantly documented”
@DE, are you advising me not to read Lepsius, Cust, Müller, Meinhof and others and not even read those who read them – but rely instead on 7th-hand sources?
Yes, I read the aricle about Meinhof in some linguistic encyclopedia, by Sara Pugach, a historian.
She says his classification is “part linguistic and part racial” and was criticised and gives two references. One to Sapir (not online) the other to some German who criticises him for not recognising that Nuba is Hamitic despite their outstanding civilsation**. This is your “abundant documentation”. Thanks and [….*]
*I don’t mean to criticise Sara. She is not a linguist and she trust linguists.
But some of them are liars:(
** Yes, THIS German guy DOES indeed support your point.
@drasvi:
It’s not the rigidity of the subsystem: it’s the observed rarity of borrowing of the kind of subsystem you’re dealing with. Even morphology can be borrowed (particularly derivational morphology, which is less systematic and regular than flexion), so whether a particular instance of regular correspondence of form and function will really be a “clincher” will depend on other factors too, such as whether its a suppletive or otherwise surprising set, and whether it is indeed part of a whole interlinked system (and so cannot have borrowed in isolation.)
A concrete example: one of many things that convinces me that Volta-Congo is really real is the class affix sg/pl pair corresponding to Bleek-Meinhof Bantu 5/6. This had the noun prefixes *dɪ/*ma in proto-Bantu, but the corresponding verb agreement prefixes were *dɪ/*ga (the additional nasal consonant on the noun forms is a particularly Bantu thing, which turns up in other classes too.)
This is usually the commonest non-human-countable noun-class pairing (“gender”) and doesn’t have any other very specific semantic associations. The commonest non-human-countable pairing in proto-Oti-Volta had the suffix pair *dɪ́/*ɦá, and similar pairings occur pretty much everywhere in Volta-Congo where the class affixes are preserved as a system at all: e.g. Lelemi (a Togo Mountain language) lɛ̀nyi “tooth”, plural ànyi; cf Mooré yēnde “tooth”, plural yēna (suffixes, not prefixes here.)
Outside Volta-Congo, no such pairing appears anywhere, even in (say) Fulfulde, which has the “human gender” pair o/ɓe, which look temptingly similar to (say) proto-Oti-Volta *wá/*bá (but where the excellent John Merrill has shown that the Fulfulde singular suffix is really from a old root meaning “head”, so that the resemblance is actually quite accidental …)
Now the variation in position of Volta-Congo class affixes (suffixed in “Gur-Adamawa” and the parts of “Ubangian” that are actually Volta-Congo at all, mostly prefixed elsewhere) suggests to me that these affixes were really clitic pronouns in proto-Volta-Congo, so this is in fact not, properly speaking, reconstructing morphology at all: but it’s still a highly suppletive pairing found in core vocabulary of a kind which is rarely borrowed, which is nevertheless found throughout the whole of Volta-Congo.
The fact that the 5/6 “gender” is fairly content-free semantically makes this very common pairing more arbitrary and hence less likely to have been borrowed or diffused than the semantically contentful human-plural *ba “they” and “liquid” *mV, which turn up in (some of) Atlantic too (though the relationship of Fulfulde ɓe to this *ba would still involve assuming two sound changes happened Just Because.)
[Apologies for the didactic tone, drasvi: I’m not trying to lecture you but to clarify the issues for myself a bit.]
A social practice that makes you care about careers more than about truth?
Hostile to what I understand as “science”.
I am not talking about winning popularity contests against the mean kids in school.
Scientific truth is subjective. You write a paper, have friendly colleagues review it. If they are encouraging, you submit it, and anonymous reviewers read it and recommend it, perhaps with corrections, or they recommend against it. This is done with math, linguistics, botany, or philosophy. In the end, the system works, not perfectly, but it does. Historical linguistics has advanced wonderfully in the past 10–20–50 years.
My funny-sounding but absolutely reasonable scheme is based on the assumption that established scholars are generally thoughtful people who like new things to be discovered but are aware of how exciting-sounding ideas can lead to dead ends. Every scientist has had exciting ideas that they had to abandon in the end.
As with everything, one needs to guess in advance whether an idea is worth pursuing. That initial exploration takes time. Getting from a “good hunch” to a “promising start” takes more time, and luck. It is unconscionable to have a grad student spend their limited time on something that may not pan out.
In Europe, at least, PhDs are now supposed to take 3 years. That’s not long enough to even work on a family, unless it happens to be an already well-studied one. And, in practice, if something is widely recognized as a phylum at all, it’s because there are professors who have worked on it before retirement.
I didn’t mean identify a new family necessarily, just work on some aspect of it, be it Romance or Afro-Asiatic.
BTW Macro-Ge (Jê) wasn’t a good example. Andrey Nikulin did his dissertation on it, and it is far clearer than Afro-Asiatic.
“Scientific truth is subjective.”
@Y, of course you can study a love song without asking “does the singer actualy love anyone?” and “is her neck like a tower?” and you can study anything this way.
But if a social practice needs this claim as its motivation, that’s no good. (I even wanted to mention in this thread – in connection to Christpher’s quote from Trask – someone whose disseration in Russian history makes this claim in its first paragraph. History is subjective and must serve interests of the country)
“it’s the observed rarity of borrowing of the kind of subsystem you’re dealing with.”
@DE, if such rarity of borrowing is actually observed, then this step is “imprortant”.
But I don’t understand why claim it is more important than the lexicon if lexicon is where the regular changes come from.
As with everything, one needs to guess in advance whether an idea is worth pursuing. That initial exploration takes time. Getting from a “good hunch” to a “promising start” takes more time, and luck. It is unconscionable to have a grad student spend their limited time on something that may not pan out.
Obviously dissertations in linguistics are very different from what I imagined. Why wouldn’t it be a good use of a grad student’s time to study a suitably limited part of the evidence for a proposed high-level taxon and conclude that it does or doesn’t support the proposal?
ETA: I have no, repeat, no plans to get any more degrees in anything.
I think part of the trouble is that the very reason that “high-level taxa” are controversial is that the solid data are too sparse and/or too equivocal for it to be really possible to definitively answer the relationship questions (meaning that purported answers are likely to entail some self-deception, or at least wishful thinking, on the part of the answerer.)
Genuine progress towards a yea-or-nay answer will only really be possible if a significant amount of new data turns up. (Which does happen.) If that hasn’t happened, you’d be sending your grad student out to die in the swamp.
And may Bopp, Brugmann and Bleek defend any young friend of mine from devoting their precious years of youth to questions of subgrouping of languages …
I agree.
As in “head of cattle”, because of course.
Sorry, my bad, the linguistic sense of “stock” was not introduced by that translation of Humboldt. I should have read more carefully: that was just the earliest instance of the *phrase* “linguistic stock” found by the OED when they revised linguistic. But that’s unimportant, since the application of “stock” to languages was already long-established. The OED’s entry for stock (unrevised from 1917) combines biological and language stocks in the same subsense; these quotations referring to language are earlier than the Humboldt translation:
So the different implications of “stock” vs. “family” didn’t develop until much later. In 1923, Kroeber’s Anthropology still said that “linguistic family” and “linguistic stock” are synonyms. The ranking of family, stock, phylum seems (as far as I can tell from a few minutes of looking) to have been proposed by Whorf and Trager, and supported by Swadesh.
@Y, what triggered me is “reputation loss”. The idea that an adult scientist will work not on what she believes to be a worthy problem, but on one optimal for her reputation (and particularly won’t make her lose her, which means reputation here is not a function of her “sportive” achievements) is scary for me.
Yes, not all scientific problems are suitable for qualificaton work of a student for a variety of uncsientific reasons.
But you have adults doing same qualification works for same or other unscientific reasons!
Yes, people both want to work on something others find interesting and want others to find interesting what they find so. But it is not about carrots and sticks. Yes, people who’re funding science want to know what to fund – which is mostly an issue for expensive experimental science – and use cittations and publication for this.
Yes, many researchers are only doing science because their friends are doing it (they want to play together), or even for prestige or money (as hired workers, no more).
If you choose not to marry someone afraid of “loss of reputation” you may lose my respect, but doing something per se unexciting because your friends are doing it is normal. If I ever publish some lingustic result, perhaps my motivation will be this. Or maybe not – a problem must be sexy enough to make me seriously work on it. However such people are not Newtons, they merely are keeping Newtons company.
I have been wondering about the reasons for decline of Muslim science since long before my interest in Arabic (and “Muslim” topics associated with it). I don’t know why. Maybe I like stars. Their names come from Arabic and sound very romantic. For whatever reasons, this historical mystery bothers me more than others.
But once I read something that frightened me. Namely, I realised that in the 19th century Bukhara was an exceptionally learned city. Handreds schools. Dozens of madaris, where sciences (and Persian poetry) are studied…from almost medieval books. A maddening concentration of learning.
And NO knowledge of achievements of europeans.
Institutionally or culturally a very uncurious environment.
It frightened me, because we – scientists especially – often think of spread of learning as the chief means of making the world better. But what follows is that learning is not THE important ingredient. And what if our period of institutionalised curiousity is but a brief (and bright) period, akin to Muslim Middle Ages? What if we will have much more learning than today, but without progress?
‘It’s not the rigidity of the subsystem:’
@DE, yes, ‘rigid’ is a bad word, interlinked is better. Perhaps I was thinkng of sticks, conneting forms:)
5th class reminded me another piece of sexism (or at least of sexism expected from Maasai) by Lepsius.
https://archive.org/details/nubischegrammati00leps_0/page/n79/
Not quite sure what he’s trying to say there, but it actually is true that in a number of Nilotic languages with a two-gender masculine/feminine system, things which aren’t actually male or female can get assigned “gender” based on shape or size; interestingly, the criteria don’t always work the same way round in different languages.
I don’t think that there actually are many Bantu languages which have collapsed the noun-class system into a binary animate/inanimate or rational/irrational system: I think he’s mixing up morphology with agreement and semantics.
Lots of Bantu languages are like Swahili, in which nouns for people and higher animals take “human” class agreement, regardless of what morphological noun class they belong to.
There are some Bantu languages (like Lingala) where all agreement has been reduced to that, but not many. (Kusaal has basically done that too, whereas most Oti-Volta languages still do the whole multi-class grammatical agreement thing.)
I think this is all part of the then-popular notion that whether languages mark “gender” constitutes a fundamental linguistic divide (which got tied up with batshit theories about prehistoric cultures in Africa.) The fact that Bantu “grammatical gender” works in a way which is about as remote from a two-gender “masculine/feminine” system as can well be imagined undermines this fantasy pretty fatally – so scholars like Lepsius (and many others) duly managed to see all these quite different systems as being all alike – spiritually, or something.
Bantu, accordingly, had to have acquired gender somehow from “Hamitic.” Because those scholars knew so pitifully little of the full range of West African languages, they blithely proclaimed that they all lacked “gender” – with exceptions like Hausa (which was therefore declared to be either Hamitic-influenced or a kind of degenerate Hamitic) and Fulfulde, which was simply declared to be Hamitic (as befitting lighter-skinned Herrenvolk.)
It’s a pity they didn’t know a single Oti-Volta or Grusi language – though even if they had, no doubt some Hamitic epicycles would have been conjured into being to “explain” them too.
The moral is that deeply embedded erroneous theories can make even learned and perceptive scholars quite unable to see what’s under their noses.
@DE, Lepsius’s theory is that the two linguistically definable entities, namely Bantu and Semito-Hamitic were, historically, the only ones around.
Bantu – and possibly some languages similar to them – are African languages.
Semito-Hamitic are invading languages.
Everything else (which is a mess, langauges which can’t be easily assigned to some linguistically well-defined group) is the outcome of the contact between Hamitic and Bantu (or Bantoid African langauges).
His argument here is that Maasai is not Hamitic despite havign the Most Important Hamitic characteristic, that is gender. The argument is formal:
[masc – fem] Maasai [l – n] Egyptian [p,f – t,s] Beǵa [b – t,s] Hottentot [b – t,s]
Semantically he theorises that Maasai gender is not as much gender as “strong, coarse” and “weak, small, gentle” and assignment of males and females in these classes is only secondary – and tries to find similarity with Bantu classes. Classes 5 [l or r] and 9 [n] are the only ones fomally similar.
Semantically Class 5 has things that have something Ausgezeichnetes, Auffallendes, Vollkommenes about them. Von der 9. Klasse sagt Hahn, daſs sie die meisten Thierklassen, wie auch die Früchte und eſsbaren Pflanzen, nicht die Bäume, die einer andern Klasse angehören, enthalte , was vielleicht von dem Begriffe des Beherrschten, dienst- und nutzbar Gemachten ausgehen mochte.
_____
I understand that “something Ausgezeichnetes, Auffallendes, Vollkommenes” is similar to “strong and rude” (not men) and “vielleicht von dem Begriffe des Beherrschten, dienst- und nutzbar Gemachten ausgehen mochte” is similar to “weak, small, gentle” (not women) but yet said with men and women in mind.:/
I have to say that, even accounting for the absence of much sound data in his day, L is there indulging in a shocking bit of special pleading to try to force recalcitrant facts to fit his pretty hypothesis:
If you can’t find a Bantu class prefix that looks rather like whatever you happen to be wanting to compare with Bantu, you just haven’t looked hard enough yet.
And most Bantu classes contain several distinct semantic groups (along with numerous exceptions), so there’s a lot of scope for cherry-picking there too.
The reality …
From Meeussen’s Bantu Grammatical Reconstructions (basically, proto-Bantu reconstructions):
Bleek/Meinhof Class 5:
ten cloud cheek breast buttocks hearth spear sun stone branch eye tooth shoulder belly twins embryo pregnancy coal spear word, voice name
Ausgezeichnetes, Auffallendes, Vollkommenes is “distinguished, striking, perfect”; in fact, as I said above, this class has really no particular semantic associations apart from non-human.
But I’m sure we can all agree that pregnancy is a male thing.
Bleek/Meinhof Class 9:
land, ground snail locust crocodile tortoise guinea fowl fish antelope buffalo hippopotamus leopard animal, meat fly crab snake eagle rat buffalo elephant wildcat neck breast, chest belly, bowels kidney hunger bag dog pig path drum gall penis shame cooking pot groundnut pepper goat sheep bovine oar shouting grave
It is right that this class contains animals a lot, though as you can see, that’s by no means all of it. It’s probably cognate with an Oti-Volta class that likewise contains animals, and also seeds (but not fruits, which go in 5.)
Again, “penis” is obviously associated with femininity.
Trees do indeed go in a different class, 3: but this is nothing to do with “usefulness”; a clue is the cognate class in Oti-Volta, which does not contain tree names, but is quite strongly associated with “long thin things”: arrows, fingers, roads, rivers, necks, tails …
For Class 3, Meeussen lists one or two trees species and
tree torch body head finger life… limit year root handle heart mouth lip spirit village bellows fire back forearm tail burden cultivated ground stage
Incidentally, an awkward-but-true fact about Volta-Congo noun classes is that the semantic associations with the classes are not always stable over time. There is a particularly clear example of this in Oti-Volta: proto-Oti-Volta had a distinct “tree” class, and nearly all the modern languages simply continue this: but all the Western languages have shifted the whole semantic group over to the “child” class, except for Nõotre, away in Benin. So has Moba, the Gurma language which borders on WOV; but not the rest of Gurma. It’s evidently an areal thing.
Actually, two other Gurma languages border on WOV too: Gulimancema (with Mooré) and Konkomba (with Dagbani.)
But the current wide extent of both Mooré and Gulimancema probably reflects relatively recent and rapid expansion, and relations between the Dagomba and Konkomba are mostly pretty hostile (sometimes extremely so.)
Whereas Moba (especially its Bimoba dialect, in Ghana) is spoken by people who often know Kusaal, and the language has quite a few WOV loanwords. It’s also (unlike every other Gurma language) lost noun-class/”gender” agreement in pronouns, which looks very much like the bad influence of Kusaal. Several phonological similarities too (partly concealed by Moba’s unhelpful orthography.)
@DE, I think I have all reasons to giggle both at his sexism, racism and meta-whatever-ism (because I think we’re dealing with a man who enjoys ranking humans) and at grand theories.
But
“If you can’t find a Bantu class prefix that looks rather like whatever…” – I don’t think he’s trying to support his grand theory with these parallels between Maasai and Bantu. Rather, the idea is that if Maasai is not Hamitic and if tis gender is not “feminine-masculine” and if one accepts his Grand Theory and this gender has to do with Bantu classes, then classes 5 and 9 are the best matches.
Also the idea is not that penis is feminine. Rather that Maasai genders are primarily not as much {masculine – feminine} genders as they are primarily {strong, large, rude – small, gentle} classes. (someone with background in math would have said: “define ‘primarily‘, without a definition I have no idea what you’re talking about”). Unfemininity of class 9 does not contradict this idea.
Penis thus is “made serviceable” because class 9, and in this is similar to Maasai “small, gentle”.
And same for everything: How doth the little crocodile…
Lepsius’ views on such things are obviously only of historical interest now, but in reality, I don’t think that gender systems in which human males and females are consistently* allocated to different grammatical genders based on something other than biological sex actually exist. Lepsius has invented an imaginary system to justify his gimcrack theory.
* There are systems with invidual lexical exceptions, as with e.g. French vedette or Lavukaleve ruima “old man” (feminine), and semantically motivated nonce-exceptions, like using masculine gender honorifically for women (no comment), but nothing of this kind actually instantiates Lepsius’ fantasy.
No, why. I don’t know anything about Oigob, but is there a problem a hypothetical langauge where each woman goes to Class A because she is a woman and women as a group are there because* A is the class of smaller things? In at least some (well-known) Afroasiatic languages feminine gender is also employed for diminution, but *all those “because” are highly speculative here.
___
I will only expect semantical stability of classes if classes (known to me) are semantically well-defined.
But are they so in African langauges?
(if the formal tree class was disbanded, then it is not semantical instability but some other instability)
Lepsius characterisation of Maasai gender is wrong.
What he’s doing is taking semantically-motivated exceptions (e.g. using “boy” as “feminine” to denote smallness, which is indeed a thing in Maasai) and treating these exceptions as representing the semantic core – the prototypical types – of the grammatical gender.
I very much doubt whether he would be doing anything so silly (he was neither ignorant nor stupid) if he did not have an Important Theory to support.
I will only expect semantical stability of classes if classes (known to me) are semantically well-defined.
But are they so in African langauges?
That varies tremendously by language. In Volta-Congo, it also varies a lot by noun class: some classes have much stronger semantic associations than others: notably the “human” and “liquid” classes. But even those can’t be relied on: the “human” class often ends up as the preferred home for loanwords, for example, which thoroughly screws up any semantic link. In general, the associations are weak.
And my WOV “tree” example involves a very semantically well-defined group, which didn’t stop every one of the group being transferred to a different formal noun class.
Not even well-defined “natural gender” is invulnerable: trees are consistently treated as animate in the 1976 Kusaal Bible version: as of 2016, they’re inanimate …
@DE, about the role of gender both in classification and in theorising about either cultural development or race:
Despite your words about “abundant documentation” of the confusion between race and language (and I think you’re referring here to “documentation” by people who have never read works from 19th-century:-() I believe the works I read by Cust, Müller and Lepsius are precisely the influential ones.
And thus are the significan ones when we are speaking about the consensus.
As I said, I do find in some of them racism – but not confusion.
Nevertheless works of Cust, Müller and Lepsius are not the only influential ones. They are works that contain full classifications.
One influential linguist who I haven’t read is Bleek. And I see that some ascribe these ideas about gender – BOTH its importance AND its racial, even racist connections – to him.
E.g. https://www.jstor.org/stable/715866 page 123
Is this true?
WP does not even recognise Alice Werner as an Africanist. A teacher of Bantu!
I understand why Greenberg calls her classification “Meinhof’s classification” – it IS a sythesis of Meinhof’s and Westermann’s work (Westermann is everything which is neither “Hamitic” nor Khoisan nor Bantu which is a lot) and it is more convenient to attack the German King of Africanists than “some lady” if you want to be crowned as the King of Africanists.
But I don’t understand what WP is doing here and if this is sexism* or what. I only can say: in reality teachers are important people, MUCH more important than people think, and some are much more important than almost all scientists, so perhaps I shouldn’t take it as an attempt to diminish her role even if others do take it so.
*How someone who’s not an (a) German (b) man can be an Africanist?
As so often, drasvi, I think we’re in substantive agreement: nineteenth-century Africanists were neither ignorant nor stupid, but suffered from a good many unexamined prejudices which misled their analyses quite a bit. And were trying to make sense of often very inadequate primary data, to a much greater degree than we are.
The moral (so far as there is one) is, I suppose, that we enlightened moderns should be wary that we don’t have a different set of unexamined prejudices liable to mislead our analyses … and should keep on collecting data before getting sidetracked into premature Grand Theories of Everything.
How someone who’s not an (a) German (b) man can be an Africanist?
It is true that, strictly speaking, only Germans can be linguists, but the term is conventionally extended to non-Germans as a courtesy title (just as UK physicians are conventionally called “doctor” despite not having actual doctorates.)
The truly daring even extend this courtesy to non-German women. In this case, the person in question was evidently worthy, having actually adopted a German surname. A clear case for becoming a naturalised linguist.
Aren’t all German ladies teaching Bantu to children in church kitchens?
Yes, now; but historically a significant proportion were Bundeskanzlerinnen, which restricted the time available for traditionally female pursuits such as teaching children Bantu in church kitchens.
History and Tradition go hand in hand.
You can’t put History before Tradition.
Happens all the time in the UK. Many of our age-old traditions were created in the nineteenth century, much later than the history they invoke.
Not to be outdone, here in Wales we have the ancient traditions of the pre-Christian Druids, which were invented by Iolo Morganwg right back in the eighteenth century.
There was a debate here back in the 80s about an endangered species in the Pacific Northwest, and the rhetoric against protecting it was about the destruction of livelihoods that went back generations, and I thought this territory wasn’t even a state, and was largely unsettled when my great grandfather was a kid, there’s literally no generation other than the living ones that could have had this job.
You could still sympathize. But that heritagizing bugged me.
It’s claimed that much British (royal) pomp was a back-reaction to the pomp of Mughal India (and that Japanese tea rituals were a reaction to the Christian eucharist)…
@DE, don’t question Scandi-KONGO. In silk or in hides women always teach Bantu.
That’s what makes them women.
Men teach Welsh; the secret, of course, is that it’s a distinction without a difference.
@jm:
A lot of classic Japanese culture seems to have been invented in the Tokugawa period. (Not a criticism at all, BTW. Quite the opposite.)
I recall reading an introduction to The Tale of Genji which pointed out at some length that very much of what one thinks of as traditional Japanese culture didn’t actually exist yet in Murasaki’s day. Not even Noh …
No wonder the Heian court spent so much time conducting elaborate love affairs and composing poetry. They didn’t even have Sudokus …
Men teach Welsh; the secret, of course, is that it’s a distinction without a difference
Very true; but who has been carelessly revealing our Druidic secrets? We evidently need a bit more security discipline in the office. I feel a human sacrifice coming on (pour encourager les autres.)
Unexpectedly, Bleek – known also for the term “Bantu” – speaks about Barth’s term “Ba-languages”.
But I was only able to find a mention of “Ba-langauges” in Barth (also comparison of ba with Hausa prefix ba with a mention of “a link with the Batī-language”, whatever Batī is and Kanuri Kānem-bū).
Bleek also uses a wonderful term Sexuelle Sprachen.
(or also like here: Allerdings wenn die durch die Sexuelle Form der Sprache hervorgerufene Personifikation von Naturerscheinungen den Himmel mit Göttern anfüllte, …)
At least in the English text of A Comparative Grammar of South African Languages Bleek does not say what Alice Werner ascribes to him, namely that genderless langauge denotes “mental deficiency”.
He instead says gendered langauge influences speakers and provokes a certain “higher poetical conception”.
____
Either he says it elsewhere, or there is not much difference for Alice, or she’s accustomed to racism, or the racist reading belongs to others, not Bleek – or it’s another specimen of ascribing racism to others.
Bleek seems to have been a pretty good sort.
https://en.wikipedia.org/wiki/Wilhelm_Bleek
He probably does say some things somewhere to get a modern liberal annoyed, but he was clearly not racist in any meaningful sense.
He wasn’t at all like Meinhof.
All through this period, you get this steady stream of people who clearly respected Africans, and (not coincidentally) did some of the very best linguistic work. My hero Christaller is another. His Twi grammar is astonishingly good – even now, probably the best available. These people were far from perfect, but they were still admirable – as people.
I assiduously call the Bantu noun classes “Bleek-Meinhof” classes (as you may have noticed); I do toy with just calling them “Bleek” classes, but Meinhof deserves his mention. He was an unequivocal racist and a literal Nazi, but he was nevertheless a great Bantuist. In that domain, he was no pseudoscientist. He is admirable as a Bantuist. Only.
Meinhof is a random German guy.
That is all.
“Alice Werner’s classification” did not sound good (and her classification is mostly a synthesis of Meinhof’s and Westermann’s ideas) and English speakers invented “Meinhof’s classification”.
When they needed to dethrone a German king and enthrone Greenberg they called this fully hypothetical (!!!FUCK!!!) classification “race-based”, without reading Meinhof. Greenberg at least says it is typological.
This all is disgusting. Maybe not racism but hardly much different. I don’t give a shit WHO you are treating this way: Germans (which I think has a lot to do with this story), blacks or women.
If you need a racist, Lepsius is shameless enough.
@DE, sorry, I’m reacting so fiercely (and with disgust) not at you, but at the system with enthroning and dethroning kings. If you happen to dislike Meinhof (a random German guy), I of course have no problem with that. But I think you dislike him based on words of people who don’t read in German. And Hamitic Fula – that’s Westermann. “Die Fulbe gehören ihrer Körperbildung und Hautfarbe, ebenso ihrer Kultur und Sprache nach nicht zu den Negern, sondern sind verwandt mit den sogenannten Hamitenvölkern Nord- und Ostafrikas..” – Westermann, Handbuch der Ful-Sprache, 1909.
@DE: Thanks, that clears things up about those high-level stocks and phyla, though it’s disappointing about the prospects for breakthroughs.
I note that the Wikiparticle on Afro-Asiatic doesn’t mention any doubts that it’s a valid grouping, though it does say that there aren’t any sound correspondences. An example of people who have made up their minds based on the current evidence, I guess.
Yes, even respectable splitters accept Afro-Asiatic (though proper splitters don’t think Omotic belongs with it.)
There are sound correspondences (cf Hausa ka, suffix for masculine singular “you”, Classical Arabic ka, Egyptian k.) However, nobody has managed to set up a coherent overall system for proto-AA: the time depth is just too great and there aren’t enough secure cognates to work with. (There are two elaborate and extensive AA etymological dictionaries, both of them very bad: full of fundamental methodological errors.)
So it’s a case like proto-Volta-Congo (but more so): it will probably never be possible to reconstruct more than scattered fragments of the protolanguage, but there are enough of those for the validity of the actual genetic grouping to be convincing.
It does help the AA case that AA languages tend to be typologically weird: despite what I said about typology being no guide to genetic affinity, AA languages tend to be typologically weird in a strikingly similar way, viz treating vowels rather as if they were all suprasegmental. But then, PIE was rather that way too …
“fundamental methodological errors” – means either that there are more professional linguists who can make a much better dictionary – or that the attempt as such is an error.
I’m not sure about this second option. The subtitle of HSED is “materials for reconstruction”.
“fundamental methodological errors” – means either that there are more professional linguists who can make a much better dictionary – or that the attempt as such is an error.
No, of course it doesn’t imply just those two possibilities. It means exactly what I said: that both works contain fundamental (and pervasive) methodological errors. (Search for Roger Blench’s paper on this: this is not just some quirky personal opinion confined to me.)
There are few people with the expertise and energy to attempt such a thing; that in no way means that anyone who actually has attempted it has done it well, nor does it mean that there would be no value in trying to do it properly, with sound methodology. My statement does not imply either of the conclusions you have foisted on it, let alone only those and no alternatives, as you are claiming.
It is not the case that anything at all is better than nothing in scientific endeavour. Badly done work, especially badly done work by prestigious scholars, actively retards real progress by making it appear that problems have been solved already when they haven’t, discouraging further enquiry, and by setting people on pursuing dead-end false leads.
What’s definitely going to help with AA is work within it. Proto-Central-Chadic was only reconstructed a few years ago, AFAIK, and Proto-Chadic seems to be a work in progress. And while the languages classified as Cushitic are all clearly AA, whether they form a single branch of it seems less clear, and there’s no reconstruction of its last common ancestor.
Omotic (…which was considered part of Cushitic not that many decades ago…) may actually be similar: the latest idea is that some of its constituents are AA after all, while others are not.
Apart from Omotic, however, AA is just plain obvious.
*e/o ablaut, yes. But the zero grade obviously started out as simple vowel reduction in unstressed syllables and was still less morphologized & lexicalized in Proto-Indo-Anatolian than in Proto-Indo-Actually-European.
Badly done work, especially badly done work by prestigious scholars, actively retards real progress by making it appear that problems have been solved already when they haven’t
Amen. Couldn’t have put it better.
I was just reflecting on personal pronouns, and their being so stable over time and all:
English: I, you, she, we, you, they
Latin: ego, tu, ea, nos, vos, eae
Well. one out of six ain’t bad …
(Yes, I know that the cognate of tu is “thou”, and that “us” is related to nos etc etc. But that’s only because I know how the story ends, because I looked at the back of the book.)
I was thinking about this originally in the context of Volta-Congo.
Farefare to “we”, Swahili (verb agreement prefix) tu- “we.” I was just reading that Starostin reckons this word is particularly stable, and this close agreement between distantly related languages certainly confirms his view. Further support comes from Lingala bísó “we”, Mbembe ɲí “we”, Yoruba a “we”, Twi yé “we”, Gbeya ɛ́rɛ́ “we” …
(All these languages really are related, too …)
I was reflecting on why WALS (https://wals.info/chapter/31) mixes up what I understand as “gender” and gendered pronouns:)
Could use pink for those and there wouldn’t be a problem….
In Berber, “we” (the independent pronoun) is somewhat unstable – it often gets restructured by analogy with “I”. “We” as in the 1pl subject prefix, however, is segmentally identical across all Berber languages, and still matches most of Semitic and parts of Cushitic into the bargain.
A form *tʊ really does look like proto-Volta-Congo “we”, despite my list above. But forms plausibly derivable from something like that seem to be oddly thin on the ground outside Central Gur and Bantu. There seem to be quite a lot beginning with /b/; it makes you wonder if there might perhaps have been an inclusive/exclusive distinction way-back-when in the protolanguage.
There’s also the fact that personal pronouns are highly subject to phonological attrition and remodelling: for all I know, some of those forms I cited might be derived from *tʊ after all. (Yoruba, in particular, seems to delight in reducing pretty much any proclitic syllable to a simple short vowel.)
Of course in colloquial French the “stable” Romance pronoun “nous”, as well as the verb forms, have largely been replaced in non-oblique cases by a third-person pronoun. This also happens to some degree in Italian colloquial speech, where a reflexive construction replaces the second person plural (e.g. “si mangia” for “mangiamo”) in some contexts.
Then there are many languages, such as Bahasa, where there is a distinction between exclusive (“kami”) and inclusive (“kita”) “we”, and over time that distinction can blur, such as in Mandarin where 我們 “women” is still exclusive in some Northern dialects but generally both inclusive and exclusive in standard Mandarin.
Just based on that evidence it seems like an odd pronoun for Starostin to build a case around.
Of course prolonged exposure to any East Asian language tends to make one dubious that there is any deep linguistic reason pronouns would be unusually stable over time. They are clearly culturally contingent just like any noun.
Well, in East Asia you get the phenomenon of politeness systems that, in the Japanese extreme, replace the entire concept of “personal pronoun” by titles – whether boku means “you” or “I” depends on the situation just as much as whether you can use it at all. And titles are cultural vocabulary, the least stable part of language. Plus, when there’s no person marking, are you really talking about yourself in the third person?
A less elaborate politeness system has eliminated thou.
I thought it was the other way around – the distinction is missing from the rest of Sinitic and was a northern borrowing from neighboring languages? (I was also taught both ambiguous wǒmen and inclusive-only zánmen as standard, and that this was a compromise between dialects with exclusive-only wǒmen & inclusive-only zánmen and dialects with wǒmen only.)
Somewhere in Germany the usual way to say “see you” is man sieht sich: “we’ll see each other” > “there will be seeing of each other”…
This has actually been proposed for the IE *w- vs. *n- forms, on the tenuous grounds that “exclusive” makes sense as a negation. Independently, some Nostraticist has found an exclusive comparandum to *n- without mentioning the proposal for pre-IE.
@Lameen, DE,
our realities are different:)
In yours – if I trust your words – everyone thinks that we have two good Afroasiatic dictionaries. Or not “everyone”: you’re speaking about those specialists in comparative AA who are able to make a better dictionary but won’t do it because they trust HSED and Ehret. Both at once.
You can’t mean that what you say will result from publication of such a dictionary – the very result which makes this publication a bad idea – simply did not happen despite publication of several AA dictionaries, that all those problems are hypothetical and the hypothesis in question already contradicts the reality.
In mine Stolbova taken alone took part in 3 projects in 90s (two AA “dictionaries” and one “database”) and says her Chadic dictionary (2016) could be a part “of a new Afrasian dictionary”.
And I don’t know anyone who trusts either HSED or Ehret.
Sorry for sarcasm, but I think the problem here is:
An owner of a restaurant from a city where raw ingredients are never sold comes to what she believes is an eatery elsewhere. She is served a bag of dirty potatoes from the field and is angered.
Which is not to say methodology is not important. But what you’re talking about is not methodology.
“we’ll see each other”
Увидимся.
(I do not of course understand English “see you” as a claim that an agent me will observe the patient you – rather as a verb unusually for itself denoting a mutual action, where the formal object is semantically the addressee and the implied subject is not the agent either. Have no idea to what extent native speakers understand it same way)
In East/Southeast Asia (with the rather major exception of Chinese), pronouns basically are nouns: an open word class with room for infinite gradations of politeness or whim. But this is really quite unusual; grammaticalization from noun to pronoun is reasonably well-attested, but I don’t think I could name a single language from any other part of the world where personal pronouns are an open class the way they are in Malay or Japanese.
‘‘replace the entire concept of “personal pronoun” by titles”
I thought about this when I read in WALS sex gender that for most scholars studying agreement anaphoric pronouns “agree” with antecedents.
So what about this replacement by titles?
And even without it, what about titles? Something like “Masha and I argued about it. I think [….] but my daugher thinks [….]” where Masha is one of several daughters. Is this agreement and is “my daughter” a gender?
Greenberg’s Afroasiatic comparison sets, or Bender’s for Nilo-Saharan, might fairly be compared to a bag of dirty potatoes, potentially useful if carefully washed. Ehret’s are more like a gratin painstakingly prepared from a mix of unwashed potatoes and rocks, and served up on a gilt-edged plate. Read his introduction; he absolutely claimed the authority of the Comparative Method, and the contents absolutely do not justify any such claim.
Or, to change metaphors: Ehret’s PAA (let alone his PNS) is basically the equivalent of Percival Lowell’s Martian canals – it’s what happens when you try to push your precision much further than the resolution of your observing equipment allows.
So what about this replacement by titles?
In the original text of The Tale of Genji, almost no characters are named at all (that would be rude.) The conventional names for the characters used in translations (including modern Japanese ones) are all actually official titles or allusions to some important event or circumstance surrounding that character. This also goes both for Prince Genji himself and his beloved Murasaki (after whom the author is herself conventionally named. We don’t know her name either.* One does not go round referring to high-born ladies by their names. What are we, barbarians?)
* She was a Fujiwara. But then, so was everybody. Apart from peasants and foreigners. So, everybody.
@Lameen, I never read Ehret’s dictionary (I read criticism of it).
HSED is very messy. Definitely raw.
Perhaps there are also methodological flaws that make it less usable or unusable for “cooks” – I didn’t think much about it.
I don’t mean it is good.
But there must be some place in science for publication of very raw results, so others (“cooks” who work with potatoes rather than delicious meals) could either work on gradual improvement of, say, a dictionary, or re-use some of those results.
(Think of how people work on Wiktionary)
Both the way science (particularly Western science) is funded and the culture and understanding of what is a “publication” or a “project” and why people publish books (linked to how sceince is funded) create certain obstacles for this.
(There is also a difference between Russian and Western understandings of “what is a “publication” or a “project” and why people publish books”.)
(There is also a difference between Russian and Western understandings ‘what is a “publication” or a “project” and why people publish books’.)
This does not mean Russian science is “good” in this specific sense. It’s somewhat different.
There is an important distinction to be drawn between “very raw” and “badly overcooked.”
Manessy’s Contribution à la Classification Généalogique des Langues Voltaïques is a worthy example of the former: his actual reconstructions leave much to be desired, even making allowances for the poor quality and scantiness of the data available to him compared with the present day. But he is clear about his methods, and provides masses of handily-tabulated presumed-cognate sets for the benefit of later scholars. A genuinely valuable source even if you reject most of his actual reconstructions.
Admittedly, “Gur”/Voltaic is a much less challenging proposition than Afro-Asiatic. But in a way, that’s the actual point: Manessy limited the scope of his study to a domain where proper comparative methods can realistically be applied, and is careful not to push his methods beyond the breaking point. He also frankly admits cases (like verb flexional morphology even in proto-Oti-Volta, let alone proto-Gur) where the data just don’t seem up to providing any of the concrete answers we would like to have.
I don’t think he always got it right, but he had the right attitude.
There is also a “vocabulary” by Bomhard of which I know nothing.
And a dictionary by Diakonoff’s team. In the 70s a team gathered to work on a project named, in translation (literally) “Comparative-Historical Dictionary of Afrasian”, or (in Diakonoff’s translation) “Comparative Historical Vocabulary of Afrasian” (one word for both in Russian, though I understand it as “dictionary” here)
They published 3 parts in the 80s (same title) then there was a break and during this break the team split. Then they published 5 issues with an English title “Historical and Comparative Vocabulary of Afrasian” in the 90s. In the fouth (1995) and fifth issues Stolbova is not listed among the co-authors.
I have never seen it and it is not online (Greenberg about differences from HSED: https://www.jstor.org/stable/30028602).
Weirdly, HSED is a part of a famous European series – while HCVA is something even I have never seen. I think Diakonoff was displeased:) Especially HSED is messy.
There was a pressing need of making work of Russian scholars visible to foreign scholars. No one understood how Western science works and what forms collaboration will take but informing people of our (or Russian, not mine of course) work was of course important – speaking of “why publish books”.
D’s team failed remarkably (though the few people working on similar projects saw HCVA… I think).
HSED in turn was “served” as a “meal”.
Drasvi,
For me “see you” is a normal sentence that blends and alternates in usage with the longer, uncontracted variants see you later, I’ll see you later and we’ll see you later.
Seth Myers’ late-night talk show runs a Web-only segment called “Corrections,” filmed after the last show of the week on Thursday afternoon. It started during the pandemic, when Myers, having no audience to give him real-time feedback on how the show was going, starting reading the YouTube comments on the clips that NBC posted from his show. It developed into the “Corrections” segment, in which he responds to the show’s mistakes, after the jackals from the Internet point them out. My sons and I find “Corrections” to be far better than the main show—no celebrity interviews, and virtually all the content written by Myers himself. The show’s staff are the only audience while “Corrections” is being taped, and lots of the jokes are clearly for the benefit of the employees at least as much as the YouTube viewers. It you watch one of the recent videos and really enjoy it, it’s probably worth going back to the beginning and watching all 130+ videos in order, so you can follow all the running gags.
Anyway, Myers ended the first bunch of videos with, “I’ll see you next week,” or a variation on that. Eventually, several jackals complained that Seth actually doesn’t see the viewers; but rather, they see him. So one of the running gags is that he now ends most episodes: “See me next week.”
There are also some papers with Afro-Asiatic etymologies by Militarev, in JOLR.
Not to mention Takacs’ various works.
For Militarev, I’ve given my thoughts previously here.
@Ryan, thanks! The problem is that when you say “see you” to someone, the fact that you will see her is of little interest to your addressee. What is more interesting is that you two will meet.
If we’re getting at meta-meaning, I suspect neither is right. What’s interesting is that you care enough that the idea of seeing her again is something you express fondly. Regardless of whether you’re likely to see each orher in any foreseeable timeline.
But whatever meta-meaning, the sentence is still straightforward and transparent for most English speakers. And often there is a fairly concrete expectation. I would be very likely to say see you if we were all getting in different cars to meet up somewhere. Or to do a few errands and then reconvene. It is not merely a conventional expression.
By contrast emphatic See ya! Is not transparent. It often means Glad to be rid of you.
‘There is an important distinction to be drawn between “very raw” and “badly overcooked.”‘
@DE, I agree.
“you care enough that the idea of seeing her again is something you express fondly”
@Ryan, thanks!
As I understand from this, for you this phrase is NOT “a lexicalised expression which means ‘we will meet soon’ which literally means something else but normally you’re not aware of this literal meaning and won’t think of it” but a phrase understood very literally.
Something similar to what you said occured to me when I was thinking of its etymology. My theory (an indirect way to refer to a meeting) works poorly here. When it was used for the first time, it was understood literally. Not only literally, but literally too. And it made sense literally, else the speaker would have chosen different words.
Meanwhile my theory is based on that it does not make sense. So why would I want to tell someone that I will (literally) see her? Perhaps what I’m speaking about is whole experience of seeing her, the fact but also how I feel about this fact, thought I.
for me, “see you” – which i use a lot less than “see you soon” — is quite concrete: i wouldn’t use it with someone who i don’t expect to see again (though i might say it to someone and end up never seeing them again). “see ya”, however, extends a bit further into abstraction, partly because it’s often said (especially sarcastically) more to other people present, as commentary on the situation, than to the ostensible adressee, but partly because for me its non-sarcastic version can be appropriate to more casual social connections than “see you”, including ones where there’s no certainty of a later re-encounter (someone working counter, say).
but i agree both with Ryan about the affective layer and with Drasvi’s original characterization of the weirdness of the construction: denoting a mutual action, where the formal object is semantically the addressee and the implied subject is not the agent either.
i agree both with Ryan … and with Drasvi’s…
I believe the joke I’m thinking about at the moment (where the rabbi agrees both with incompatible claims and with their incompatibility) is international* even if I only heard it in Russian:-)
*and in Russia some jokes are variously told about a rabbi OR a mullah.
“Not to mention Takacs’ various works.”
Takacs is working on an Afroasiatic dictionary of his own, since 1999. I’m afraid HSED and Ehret – both the fact of their publication and their imperfections – did not weaken his motivation but strengthened it instead:)
A snippet from Google:
The subject of “see” is not an agent; it’s an “experiencer.” English verb subjects are by no means always agents.
https://en.wikipedia.org/wiki/Thematic_relation
There’s nothing especially strange about “see” in this regard. Or about English: I know of no language at all in which verb subjects are always agents.
On the supposed weirdness of “[I’ll] see you [later]” as a parting formula, I take it that you find до свидания weird too, drasvi?
Ellipsis of subjects etc in formulae for greetings and the like is perfectly usual. Again, I can’t think of a language where it never happens.
@DE, yes, not an agent.
The weirdness is that “tomorrow I will receive some more visual data pertaining to you” does not sound as information* useful to the listener, unless she’s a dancer.
“Tomorrow we will meet” does.
*Especially if what you’re going to do tomorrow is watching a recording of your addresse – that will be uninteresting even for a dancer.
Consider that when ending a phone conversation, the more common ending is “talk to you later” even though “talk with you later” is available.
Or that a way of describing a coming job interview could be “the supervisor said she’ll see me on Tuesday!.”
@drasvi:
If you wanted to say “goodbye” to a work colleague at the end of the day in Kusaal, you would say (unless it was Saturday)
Bɛogʋ la.
“That’s tomorrow.”
This would not be expected to elicit the reply
Bɛog an bɔ?
“What is tomorrow?”
Greetings do not “provide information” at all, and providing information is very far indeed from the only common use of lsnguage.
https://en.wikipedia.org/wiki/Phatic_expression
(This is not the only kind of speech that does not convey information, but it’s a clear example.)
Seeing is asymmetrical.
– It changes your state but not the state of what you seeing.
– when you say “I saw her”, you don’t convey any information about whether she also saw you.
Meeting is symmetrical. It changes everyone’s state and when you say “I met her” I know she met you.
The source of weirdness is that understood literally this formula conveys information about YOU and only about you. And not very useful information about you.
https://hitchhikers.fandom.com/wiki/Ravenous_Bugblatter_Beast_of_Traal
If I want to emphasize that I really do expect to see soneone again, at a not too distant but undetermined point in thr future, I often use the mid-twentieth-century formula, “Catch you on the flip side.”
In Kusaal, verbs expressing direct physical perception, like “see” and “hear”, have different syntax from e.g. “think” when it comes to taking a following complement clause. You can say
M nyɛ nidibi kenna.
I see people.NOMINALISER come.IMPERFECTIVE.hither
“I see that people are arriving.” (literally, “I see people’s arriving”)
as pretty much synonymous with
M nyɛ ka nidib kenna.
I see and people. come.IMPERFECTIVE.hither
but you can only say
M tɛn’ɛs ka nidib kenna.
“I think that people are arriving.”
and not
*M tɛn’ɛs nidibi kenna.
You can see people’s arriving, but not think people’s arriving.
>The source of weirdness is that understood literally this formula conveys information about YOU and only about you. And not very useful information about you.
This isn’t true, as shown best by the example I gave of someone saying that the person making the hiring decisions “says she’ll see me”. It’s so informational that the person who is the object passes it on.
And also, as in the diss:
– See you later
– Much
(Meaning – I hope that we don’t see each other till much later.)
The meanings are all transparent, even if some of it isn’t wildly informative. You, drasvi, may assume someone will see you later, but they’re saying it carries meaning nonetheless.
Seeing is asymmetrical.
– It changes your state but not the state of what you seeing.
– when you say “I saw her”, you don’t convey any information about whether she also saw you.
Hence the witticism
“See you.”
“Not if I see you first.”
Which I remember hearing only as teasing, not as a candid statement that the speaker will avoid the listener.
Surely drasvi realizes that it doesn’t make sense to try to interpret idioms literally.
Seeing is asymmetrical.
– It changes your state but not the state of what you seeing.
“Kill” is asymmetrical. It describes a change of state of the object but not the subject. All this has absolutely nothing to do with whether a statement conveys useful information or not.
A verb like “die” implies a change of state of its subject. “See” does not. Once more, the thematic role of the subject of “see” is “experiencer.” It’s no more a “patient” (i.e. a state-changer) than it is an agent.
In fact, verbs which are “symmetrical”, in the sense that the subject and complement do have the same role, are a small minority (though an important one: cf “Venus is the second planet from the Sun.” )
(I should say, “constructions” which are symmetrical, rather than “verbs”: even the copula is not usually so. “Women are all doctors” does not mean the same as “All doctors are women.”)
Agent, patient, experiencer.
I think it was late at night.
Did anyone read Ryan’s and rozele’s comments?
“see you” is not same as “hi” or “salam” (a salam by an English speaker).
Not something an L2 speaker understands natively IFF she knows its contexts even when it is her only English word. And of course more things can be said about what it means to speakers apart of “it is idiomatic, you say it when…” (and even for salam there are rich connotations which are not contextual)
Also Takács is “weawer” in Hungarian and Slavic.
I wonder why the first a.
Is it what Hngarian does to either (a) Slavic [tk] (b) Slavic /tŭk/ or is it (c) from a Slavic dialect where *ŭ first gave /ă/ and then was dropped or (d) from a Slavic dialect that does not drop *ŭ’s?
@drasvi
Would an epenthetic schwa be turned in to a by vowel harmony?
I don’t know much about Hungarian language history, but here’s my conjecture:
1. We know that the Slavic substrate in Hungarian shared features with Bulgarian; e.g., the Pest in “Budapest” means “oven” and shows the typical Bulgarian development of the Pslav. cluster /kt/ before front vowels.
2. While all other surrounding Slavic languages treated the yer in *tъka- as weak and have words for “weaver” starting in tka-, Bulgarian treated the yer as strong and has тъкач.
3. Bulgarian “ъ” is normally described as /ɤ/, a mid-back and somewhat centralized vowel; whether it was the same in Pannonian Slavic or maybe nearer to [a], I have no idea. In any case, my bet would be that the actual scenario was closest to drasvi’s (d) – takács is a loan from a Slavic dialect similar to Bulgarian that didn’t drop the yer in this specific word (IIRC, there are other cases where Bulgarian doesn’t drop yer in the position /#CъCV-/, but it’s bedtime and I am too tired to look them up).
By Morphic Resonance, I was just reading a very interesting paper by Benjamin Suchard and Jorik Groen, which posits an epenthetic /a/ in proto-Semitic:
https://www.academia.edu/47400828/_Northwest_Semitic_sg_CVCC_pl_CVCaC_%C5%AB_Broken_plural_or_regular_reflex
I think I’ve lost track of what drasvi and others are actually disagreeing on, but the use of “see” to refer to a meeting rather than the specifics of visual perception is a lot wider than those particular idioms, sometimes implying more asymmetry than others…
“The doctor will see you now.”
“I’m going to see her on Thursday.”
I can’t say I disagree on something.
I simply find the Engish phrase rather peculiar, and I’m not sure if native speakers understand it in the same way as I do.
From Ryan’s explanation I understand that the literal asymmetrical meaning of “see” does affect its interpretation: what native speakers feel when saying or hearing it and how they understand the “meaning”.
Note, that “meaning” is a complex thing. Even when meaning is not “literal” it can be
1. I can say A implying B to make you aware of B.
OR
2. the meaning of A used too often to inform about B can shift towards B.
This second mean, when you hear “A” its “literal” meaning does not influence your feelings and thoughts.
Both times we also can ask, WHY people use A if B is what they want to say.
DE simplifies this all to a claim that it is some idiom. Does he mean: idioms aren’t interesting?
Surprisingly many puzzled English language learners are found on the web, inquiring about the meaning of “smell ya later.”
I was thinking why in Russian there is a well-defined verb for smelling like listening and looking but not for smelling like hearing and seeing.
Two verbs used for this in speech mean “to feel” and “to hear” and are used with the noun “smell [of]”.
The 3d is somewhat dated, can be used with smelling direct object (without “smell of”) if smell informs you about it rather than being of any interest per se, but it is also a somewhat dated word for “to feel”*
The 4th is bookish.
And no verbs for tasting.
* Its semantical develpments are interesting. It also can refer to intuition or foresight, also (uncommonly) to hearing and there is an interjection ču, I think linked to it.
It is not used in my Russian, but in books from 19th century you often read instead of “somewhere an elephant trumpeted” something like “Ču! Isn’t it an elephant trumpeting deep in the jungle amidst the steamy African night?”
God knows what is its NORMAL meaning.
We did that one before. Russian is by no means alone in that: Welsh, for example, has clywed for “hear” and “smell.”
Hausa ji covers all sensory modaliities except sight: hear, smell. taste, feel. It is even used for “feel” in the sense of “feel an emotion”: ji tsoro “feel fear.” And also for “understand” (something spoken): Kana jin Hausa ko? “Do you speak Hausa?”
Kusaal wʋm has all the senses of Hausa ji, except “feel an emotion”, for which it somewhat oddly uses zɔ “run, fear”, even for positive emotions: zɔ nimbaanlig “feel compassion.” Kusaal uses other verbs there too, e.g. di nyan “feel ashamed” (“eat shame”) versus Hausa ji kunya.
Kusaal nyɛ “see” is also the usual word for “find”, even in metaphorical senses like nyɛ laafi “get well” (“find health.”)
“eat shame”
“taste the whip in love not given lightly“!
By the way:
Perhaps Jerry will like Ratcliffe 2012 On calculating the reliability of the comparative method at long and medium distances.
The problem with those dictionaries is not a problem with the comparative method itself. The works in question do not begin to apply the comparative method properly. The Ratcliffe paper is pretty good on pointing out their really quite extraordinary failures.
Have a look at p264 on “bird” words: you don’t even need to know anything about linguistics to recognise that this is mere junk science, simply on common-sense grounds.
But that (admitting a grotesque degree of semantic latitude in comparanda) is only one of the fundamental methodological errors.
The fact is, that if you end up with hundreds of supposed cognates in a language family as old as Afro-Asiatic, you’re doing something very wrong.
We were just hearing the other day that the total number of reconstructed lexemes for PIE, vastly younger than AA, and intensively studied by experts for centuries, is about 1600.
Volta-Congo is pretty certainly significantly younger than Afro-Asiatic. I have managed to identify less than fifty lexical cognates between proto-Oti-Volta and proto-Bantu. Admittedly (a) I am not a professional and (b) I haven’t devoted years of my life to trying to reconstruct proto-Volta-Congo. On the other hand, this particular comparison is a relatively easy one, not least because proto-Bantu has been worked on for well over a century by professionals.
Neither I nor Ratcliffe say it is…
(What do you mean by “professional”?)
Paid to do it as your day job.
I was reading something of John McWhorter’s a while back (can’t remember exactly what) in which he was taking issue with the claim that you can only properly claim that languages are genetically related if you can reconstruct a substantial amount of vocabulary in the common protolanguage.
He said that this was disproved by Niger-Congo, in which only a handful of words had been reconstructed. As with much of what he says, he was simultaneously right and wrong: he believes in the Greenberg maximalist Niger-Congo, and the words which he assumes to have been reconstructed to proto-NC are pretty much all mama/papa words or phonaesthetic words (“blow, breathe” is a favourite of NC maximalists, for example.) But he was quite right in principle, nonetheless.
If you mean you aren’t paid for your work, that’s silly (also that means Russians science is done by amateurs:)).
Then there is the method itself, but nothing about this method that you don’t know is taught to students.
Then there is linguistics. It is LARGE. Of course there is a plenty of things useful to a historical linguist that you don’t know – but this is also true for any historical linguist, and a plenty of them are less knowlegeable.
And then there is something that I believe actually can make someone more professional than you are: amout of experience with the languages of interest. But I think you stuffed this in (b).
PS aha. Thanks. But that does not make you a worse or better specialist.
Yes, in linguistics in particular, there have been some quite famous amateurs who have done work of lasting value. (Though the example who immediately springs to mind in fact had a lot of formal links with academia.)
Despite my slightly flip answer, though, I think there is a (sometimes blurry, but significant) distinction to be made between scholars who have “come up through the ranks”, as it were, of a formal higher education in the relevant field, before going on to further research (not always paid, or conducted from the “security” of an actual academic position) and someone like me, who has bypassed the main entrance and climbed in through an upstairs window.
I know more about comparative Oti-Volta than anybody else does. In that sense, I am the world expert*. But this is an unsatisfactory state of affairs. Because I have never studied linguistics in a formal academic setting, there are huge gaps in my background knowledge, not all of which I am even aware of. I also lack the kind of socialisation in an appropriate academic network that leads to new ideas by cross-fertilisation.
It would greatly conduce to the future health of comparative “Gur” study if someone with a proper linguistic training (and – in a better world – a secure academic position) were the world expert, and not me. Meanwhile …
That’s what I really meant by “not professional.”
* Buligin ziŋ zi’ kɔligin yɛlaa.
“The fish in the pool knows nothing about how it is in the river.”
“Paid to do it as your day job.” – I met some people on the internet who seriously believe that the money they earn very precisely reflect their “usefulness” (I also heard about something called “Protestant ethics” which is not the same, but can I think be somehow combined with the above).
Actually it would not be difficult for you to find a job as a linguist in Russia (and I mean: among our leading Africanists). Not in the sense you would apply for it and immediately accepted. You would first meet people with similar interests (who all or almost all happen to be linguists) then someone would ask you if you would like to teach a course, then you would have been told that it will be more convenient for students if they hire you, then someone would have told that maybe if you expand what you alreayd have to a dissertation it will be even more convenient, and then you would work in some research institution.*
But as everyone you would need a second job, one where you’re paid well:))))))
* I describe here Anna Mouradova’s career. The first step “meeting people with similar interests and asked if she would like to teach a course in MSU” took some 20 minutes maybe.
Talks about hiring her… two month maybe. But I don’t know when she was invited to work in the Institute of Linguistics.
No, I don’t like this distinction. I like science.
Careers are not what I mean by science. They are careers.
I have in fact been accused of having a “Protestant work ethic”, but the accuser in question was a Catholic, so I forgave his misunderstanding. It is easy to misinterpret cultures from the outside …
(My Protestant wife has never told me that I have a Protestant work ethic, or, indeed, any kind of work ethic. For some reason.)
Oh. Something has changed.
When I googled her my first thought was “wait, why does she looks like a Syrian?”
And the second thought: “why not, she is an Assyrian”.
Perhaps Jerry will like Ratcliffe 2012 On calculating the reliability of the comparative method at long and medium distances
Thanks, drasvi. I certainly like the idea. Is it possible I said more than I remember here about how quantitative methods are needed, and the argument that the Indo-European languages are genetically related is a quantitative argument in disguise? That is, the probability that all those correspondences happened by chance is just too low.
Has anyone studied how many spurious sound correspondences occur between unrelated languages? For instance, how far could someone get in reconstructing Proto-Scandi-Congo by rigorous comparative methods? Of course, since languages are messy, applying quantitative methods would be much harder and less reliable than that sort of thing by itself would achieve, but I think it would be interesting.
Has anyone ever said, “We estimate that there’s about a 2/3 chance that these two languages have a common ancestor”? Or even “…it’s more likely than not…”?
Jerry, no, that is not why I think you may like it… Or well, yes.
I think both a linguist and someone with background in exact sciences can find it entertaining.
But the reason why “someone with background in exact sciences can find it entertaining” has a lot to do with what you said now:-)
However, human mind
1. performs exceptionally poorly with probabilities (say: the problem about three doors and goats. It is extremely simple, in the sense of “not complicated”. Yet it will confuse at first even a mathematician:))
2. will never enjoy processing millions of digits, whether those represent probabilities or not.
Computers do such things, not humans.
And of course no one wants to transfer linguists’ job or a part of it to computers. I don’t “want” it.
But it will make historical linguistics more scientific, solve a number of problems, make some rather heated arguments impossible and tell us more about the world, whether I like it or not.
Has anyone studied how many spurious sound correspondences occur between unrelated languages?
Yes. This is a well-known study:
https://languagelog.ldc.upenn.edu/myl/Ringe1992.pdf
There are others, too. This is a well-known problem, but it gets blithely ignored nevertheless.
The Ratcliffe paper that drasvi appositely linked to does a bit of that kind of analysis too.
The essential problem is the same one that vitiates many published medical papers: underestimating the actual degrees of freedom when you do your sums. These AA dictionaries greatly increase the chance of finding unrelated lookalikes and mistaking them for cognates by phonological and semantic latitude, cherry-picking variant forms within subgroups to get the best-looking match, and segmenting words in ways which have no basis in the morphology of the languages involved in order to produce lookalike word fragments. Estimating how likely it is that the resemblances are due to chance, while ignoring the fact that you have significantly altered the odds by such practices, produces entirely meaningless figures.
Counting loans, mama/papa words and phonaesthetic words as cognates degrades the reliability further.
Again, these are all well-known pitfalls. It makes you wonder why people ignore them. Sometimes they simply assert that they have not fallen into them, as if just listing the kinds of error protects you from committing them.
“It makes you wonder why…”
2-consontant root (which is what Afro-Asiatic root is believed to be by many) combined with uncertainty about vowels* (to the extent they are not grammatical) and morphology is an objective problem…
*Say, Diakonoff’s team reconstructs TWO of them. /a/ and schwa:)
https://languagelog.ldc.upenn.edu/myl/Ringe1992.pdf
That’s what I’m talking about! And yes, it’s a shame for anybody to ignore that kind of thing.
There are two somewhat different questions involved in this.
One is the reconstruction of spurious protoforms, which is what we’ve been mostly talking about.
The other is estimating the chances of two languages being genetically related. I don’t think these are the same question at all, though they are obviously linked.
I don’t think that there is some fixed lower limit of cognates you need to conclude that languages are related (McWhorter is right.) It reminds me of Abraham haggling with God about how many righteous people there have to be in Sodom and Gomorrah for God not to rain down the fire and brimstone …
In Niger-Congo work, the criterion for degrees of genetic relateness has tended to be the number of lookalikes on Swadesh lists (and the like.) It’s easy enough to do that with pairs of languages nobody but Merritt Ruhlen has ever considered to be related, and the figures you get are often not very different from those you get comparing pairs of “Niger-Congo” languages.
Ways are found to finesse this awkward fact, and to be fair, it’s hardly a good technique. It’s hard to apply consistently even within groups you know are related: does Byali yia “die” count as a match with Ditammari ku “die”, for example? By what consistent principle? (The words are in fact exact cognates* … and Byali and Ditammari even belong to the same subgroup of Oti-Volta.)
* Also cognate with Swahili kufa “die.” Unsurprising? Sure … but the ku- is actually a noun class prefix (it’s got secondarily incorporated into the stem) and the -a is a generic verbal suffix. The part that is actually cognate is f.
DE, perhpas we are still mixing up overcooked meals with dirty potatoes.
When we mix up needs of a cook with needs of an eater, we take away from cooks a medium for international collaboration.
My point is exactly this:
Paper publications are such a medium and cooks need this medium.
(a Wiktionary for historical linguists would be better,
but they are Professionals which means: apart of collaboration they also compete for money and need prestige and results they can boast with)Meanwhile the issues you are talking about are problematic for eaters rather than cooks.
It is enough to supply with each form with a field “confidence” (a digit 0 to 9 in this field) to make what you think is a BAD dictionary a GOOD one.
Or simply to make it short.
The problem is interdependence of reconstructions. Dependence of reconstructed sound changes on forms – and dependence of reconstructed forms on sound changes.
Actually IF Ehret is as bad as everyone says, I have all the reasons to be grateful to him.
And no reasons not to be so.
And that’s exactly how I feel.
Of course we won’t agree:)
“Fundamental methodological problems” (to quote Takacs) are by no means a reason to dislike his effort and the fact of its publication. Especially when it provides – as I think – motivation for Takacs and others.
WHY on Earth should I dislike the fact that someone informed me about her work?
Gratitude in turn is by no means a reason not to criticise the methodology.
There is one thing that can make it “bad”: if researchers working in other fields will use it. Or if it will be taught in universities.
Or what DE and Lameen said (once one dictionary is published, no one will work on another) except that what we have is (annoying?) proliferation of dictionaries. “
once one dictionary is published, no one will work on another
This is definitely not what I said, nor what I take David to be saying. It retards progress by obscuring the very existence of the problem and reducing the perceived value of the work that remains to be done. Archeologists think proto-Afroasiatic has already been reconstructed; ambitious Afroasiaticists think they have to write a replacement; and no one ends up directing their attention to the intermediate levels that would be necessary to actually reach a reliable PAA reconstruction. (Due credit to Stolbova there for at least attempting proto-Chadic first, but Chadic is already too high to reach without scaffolding.)
BTW, that also happens in my field. What I’m working on right now is a complex of problems that were thought solved in the late 90s; the “solution” became textbook wisdom and stayed so till 2017 or so, when suddenly the surprises began to pile up (and they’re not done yet).
@Lameen, sorry, I misunderstood you. (and if I misunderstood DE too, sorry too)
But we are still in different universes and absolutely nothing has changed.
You say “no one”.
If you need to work from the bottom up to reconstruct it, I also need to work from the bottom up to understand it. And in my universe a lot more work is being done today than in 80s. While I’m not going to say “that’s because of Ehret”, there is also I think growing understanding of importance of such work. And I belive Ehret’s dictionary – its criticism and its flaws – plays its part in it, small or large.
Also it is an axiom for me: I won’t think bad about a scientist because her work is silly or she is silly.
I beieve it is so for you, for DE, DM and me and also for people like you who potentially will work on such reconstructions and people like me who potentially will read and understand your works.
I think your aquitance with Ehret’s dictioanry increased value of such work as perceived by you.
I don’t know what archaeologists think about the value of reconstructions of branches of Chadic.
As for misunderstanding – yes, we all think that the natural way of erecting a building is from the bottom up, and you expect this understanding from me and I of course expect it from you and DE. I thought you two were thinking about something else (for DE: Greenberg’s classification). But I though you were speaking about dictionaries and thinking about something else, while in reality you simply were speaking about somthing else:) Sorry. DE specifically said (or that is how I understand him) that a good AA dictionary is possible, but if I misunderstood him too, as I said above, it changes little. I still think that the dynamics is different from the predicted one. The dynamics that you predict happens too, of course.
Speaking of Afro-Asiatic, a brand-new collection by Kuznetsova et al. Rarities in phonetics and phonology (here), has a paper by Wolff on the history of the ridiculously small vowel systems of Central Chadic. He suggests the possibility that Proto-Afro-Asiatic itself had a one- or zero-vowel system.
The rest of the volume is chock-full of phonetic and phonological giraffes.
Being familiar with a language with contrastive vowel glottalisation, I naturally turned to the chapter on the excitingly named “ballistic syllables.” (Seems like the feature is pretty much the opposite of vowel glottalisation – “hypermodal syllables”, if I’m interpreting it right. I’m so tempted to start calling Kusaal modal vowels “ballistic vowels”, but the term might be labelled as inaccurate by killjoy nitpickers …)
JenInE would be pleased to see the chapter on the rarity of preaspirated consonants, no doubt …
Juliette Blevins’ chapter about peculiar (or not) historical sound changes is very interesting, too. I’m not sure that she always avoids proving Blust’s point, though …
@Y, in the reconstruction of Diakonoff’s team, there is
– something that gives Semitic a
– something that gives Semitic u and i
– labialised velars (make the second something u)
This is obtained for Semitic and then for AA.
Not too different from what you say. When you have only one vowel and all syllables are CV….. you always can say that In Reality those syllables are C, CC, CCC… and the silly vocoid is “surface”.
Here
“for Semitic” – *From, not “for” Semitic, of course (based on distribution and alternations within nominal roots). And about a and ə: “It is not impossible that the difference … one of pitch”
SOme 20 years ago * was used this way in Russian chats.
A:[smoe txet]
A:*[some text]
Where does it come from?
“It is not impossible that the difference … one of pitch”
Sure, not impossible. Not very likely, though …
Anyone know of a real language in which [a] and [ə] are allophones according to tone? But perhaps by “pitch”, they mean “stress” (and, indeed, the boundary is not always clear-cut, by any means.)
Heath’s grammar of Tamasheq calls vowel patterns “melodies” (which he says are different from the vowel template patterns of Semitic.) To be quite sure of confusing poor Niger-Congo scholars further, in describing these “melodies”, he also adopts H for “high” vowel (meaning “close”) and L for “low” vowel (meaning “open.”)
Heath never shies away from language complexity. (I do have the occasional moment of doubt in which I wonder whether, if there isn’t enough complexity, he feels that he needs to introduce some …)
It is quite a nice illustration of the AA treating-vowels-as-suprasegmentals thing, though.
Rounding and fronting of vowels are treated as word-level suprasegmental features in quite a lot of Chadic grammars (and in Foley’s Yimas grammar, another language which can be argued to have – maybe – only one unequivocal vowel phoneme.) And Tucanoan languages treat nasalisation as a word-level prosody.
I can appreciate the theoretical beauty of analyses like these, but in terms of descriptive adequacy they sometimes don’t seem to be much different from traditional descriptions in terms of vowel harmony. Different when these prosodies actually work like flexions or devational affixes, though, as they can in AA.
The essential thing that AA vowel-sceptics are up to is stripping out as many of the distinctive features of vowels (e.g. front-back, open-closed, rounded-unrounded) into suprasegmental prosodies until, ideally, you leave the poor vowels with nothing but their undifferentiated vowelness, which you can then call “syllabicity.” The prosodies, having served their purpose, can then be quietly sidelined.*
It’s vowel height that is most resistant to this feature-stripping, which is why /ə/ (or /ɨ/, whatevs) tends to be the Last Great Hope of the Vowels. The strippers try to explain all its occurrences away as epenthetic, but it never quite works completely. Epic stuff. Pixar should make a movie.
* There’s an excellent study of Central Chadic (not Wolff’s one) which actually does also try to reconstruct the prosodies themselves historically, but I’ve unfortunately forgotten the author. Lameen will certainly know …
“Anyone know of a real language in which [a] and [ə] are allophones according to tone?”
Not [a] and [ə] , but “something that gives /a/, designated as *a” and “something that gives i or u, designated as *ə”. If that helps of course:-)
After all, reconstructed vowels are not phonemes. You can’t know if your vowel is an allophone or say more than one segment…
Wolff has 1-2 vowels and mentions that AA can be vowelless. But as I said, when our syllables are all CV[….], you can rename your V to “syllable nucleus” attached to C₁.
“One vowel” is only phonetically different from (or more specific than) “vowelless”…
But I wonder, how exactly a reconstruction without vowels is different from a claim that vowels can’t be reconstructed. (I don’t mean they are NOT different)
The feature-stripping thing has demonstrably happened in Ge’ez: velars became labiovelars before /u/; subsequently, /i/ and /u/ fell together in all positions, which left labiovelars contrasting with plain velars. The labiovelars were then reanalysed as root consonants in their own right, and consequently appear in vowel-template patterns even before /a/.
(Militarev apparently holds that such Ge’ez contrastive labiovelars were simply inherited from proto-Semitic, a view so daft that it gives one a headache even thinking about it.)
This is, however, the opposite process chronologically to that implied by the AA vowel-minimiser caucus.
If that helps of course
I don’t think it does: however you slice it, you’re proposing that a tone difference alone produced a vowel height difference, somewhere along the line. No doubt it’s possible, but I can’t think of any real-world examples. Not a problem if they mean “stress”, but I suspect that interpretation might lead to a different set of implausibilities.
And unless you have some other reason to ascribe lexical tone to AA (and can actually reconstruct it in some lexemes, based on tone in daughter languages), attributing the one remaining distinctive vowel feature in your reconstruction to “tone” rather than just calling it by its True Name of “vowel height” is just some sort of rebranding exercise. It doesn’t mean anything.
But I wonder, how exactly a reconstruction without vowels is different from a claim that vowels can’t be reconstructed. (I don’t mean they are NOT different)
I’ve often wondered about this (and similar issues) myself. It’s abundantly clear that we can only ever reconstruct parts of protolanguages. This is obvious with lexicon (and even more so with syntax) but it seems to me that this is true in principle of morphology and phonology too.
I suppose the difference is that with morphology, and even more so with phonology, you’re actually trying to reconstruct a system, not just a whole lot of random affixes or phonemes. The internal coherence of the whole system can reassure you that you can’t have missed an entire part of the jigsaw.
Not certainty, though. If every single daughter of PIE had independently merged (say) *ɛ into *e, without any other changes dependent on which vowel was originally there, we’d never know. Loans in other language families, or Nostratic-turned-rigorous might help, I suppose.
Even when we do have good reasons for being confident about our reconstruction, as with PIE, we should maybe be wary of assuming that we really know enough about the protolanguage to make grand typological claims about it. (How exactly did PIE speakers realise all those funky consonant clusters?)
There’s a nice example of the interplay between phonology and morphology (and hence typology) in Linn’s grammar of Yuchi (thus also providing a nice example of how All Threads are One.)
If you analyse the stress system properly, it turns out that quite a lot of “affixes” are really clitics, so the language is much less “agglutinative” than previously thought. (Still pretty agglutinative, though. But not nearly so “polysynthetic.”)
“True Name”
Imāla:))))
There are indeed too few Arabic technical terms in general linguistics …
My understanding of the one-vowel Chadic languages in question is that you can have Vowel between consonants, whose quality is determined fully by its environment; or no vowel, in which case an epenthetic phonetic schwa may break the consonant cluster.
@DE are there great many AA vowel-sceptics?
As for Ehtiopian labiovelars, as I said, there is a Russian dictionary barely known in the West (published in English in 90s), by a team of Diakonoff. Militarev did the Berber part.
And in this reconstruction AA has labiovelars.
The Chadicists seem to be the ringleaders. Don’t know how far they’ve corrupted their brethren. I’m not sure the idea works all that well outside Central Chadic, even.
But Lameen is the genuine expert.
Yimas is one of the more plausible cases, though even there it can’t be made to work completely, synchronically. Unfortunately there doesn’t seem to be anything useful online about this aspect. I got all this from Foley’s grammar. There are other Papuan families with very small vowel systems, too.
Northwest Caucasian is the other famous set of cases. Given their record-breaking consonant systems, the idea that vowels have been progressively robbed of features by neighbouring consonants there looks like a no-brainer to me. Like Ge’ez, but much more so.
Apropos, the Chadic language Margi is often cited as an example of NW-Caucasian-style consonantal exuberance with only two vowels. Hoffmann’s grammar has lots of vowels, but perhaps he was just Unenlightened. Or practical.
Bella Coola (Nuxálk) has three vowels, with /a/ somewhat commoner than the others. I can see how a three-vowel language with one vowel much rarer than the others might shed it altogether. Unfortunately there is no phoneme inventory database with relative frequencies that I know of.
I’m currently trying to explain away two out of the nine basic short vowels of Kusaal. Baby steps … only six more to go after that.
But, really, Niger-Congo has no truck with all this Chadic vowel-hatred. Even five short vowels is a bit on the stingy side for Niger-Congo. How can you have a proper ATR harmony system with only five vowels?
@DE: for Central Chadic reconstruction, Gravina’s thesis? I don’t know whether it was *good* reconstruction or not, it made my head spin.
Yeah, thst’s the one I was thinking of. Thanks!
I’ve actually got a copy lying about on the hard drive of one of my gadgets. I was being too lazy.
Richard Gravina, Central Chadic Reconstructions — that one?
This one, I think.
@Hat:
No, it’s his PhD thesis that I’ve actually seen. (I was more interested in his methodology than the actual reconstructions, as I really don’t know much about Chadic from a comparative point of view.)
That looks interesting too, though.
@Y:
Yup.
“an excellent study of Central Chadic (not Wolff’s one) which actually does also try to reconstruct the prosodies themselves historically,”
@DE, is this one Gravina’s?
My CCh to-do list since recently has been “read Gravina, learn W(M)andal(r)a”
Wandala is certainly very interesting. Frajzyngier’s grammar is pretty good. But I expect you’ve already seen it.
An unexpected two-vowel language is northern/Standard Mandarin. This analysis suddenly makes sense of why so few of the syllables that would otherwise be expected to exist actually do exist – and of the otherwise insane phonetic fact (insane enough not to be mentioned there) that the syllables yi, wu, yu are neither [i u y] nor [ji wu ɥy], but syllabic approximants [j̩ w̩ ɥ̩].
It marks a corrective footnote. That’s widespread throughout the innertubes.
There’s work on that. I recommend Andrew Byrd’s thesis from 2015, which concludes that PIE was actually much less funky than Polish in this respect; first, though, read this short book chapter by Martin Kümmel on why the assumption that PIE placed syllable boundaries through consonant clusters the same way Sanskrit and Classical Greek did is unjustified – in the thesis you can actually see the places where that assumption drove Byrd to despair.
In Ket, the difference between [e o] and [ɛ ɔ] is that the former occur in the level tone and the others in the other three.
…and in the article on Standard Mandarin phonology I just linked to, it says: “Some native Mandarin speakers may pronounce [wei̯], [jou̯], and [wən] as [ui], [iu], and [un] respectively in the first or second tone.[13]: 69 ”
Thanks, DM.
Syllabification in general is one of those things which you approach naively expecting it to be simple, if not trivial, that turns out to be more and more intractable the more you look into it. A tarpit problem.
It may be one of those things that we were never meant to know.
[On your second comment]
Good thing I hedged my bets in mealy-mouthed way I did. Nice examples.
(I still stand by my remark about the chutzpah of just blithely proposing to make “pitch” a stand-in for “vowel height.” What’s wrong with just saying that vowel height is itself a prosody, as Heath does for Tamasheq? With the added bonus of perplexing Niger-Congo specialists. I mean, fuck those guys …)
In Mandarin (epicycle alert!) the tones arose from segmental features, and also have correlates like glottalisation. Different from Niger-Congo tones …
@DE, yes.
9 Muslim Wandals (in pants) and a Chinese official brought this langauge to Cameroon from Carthage soon after the Arab conquest. There they converted back to paganism.
If you wonder why pants, find the riddle about pants on p. 106 [122] of Frajzyngier’s grammar.
(You can’t imagine how happy I was when, after a few seconds I remembered that I read the answer to the riddle, on p 319)
“Carthage” is novel. Most West Africans’ ancestors come from Egypt nowadays (creeping Americanisation, I calls it: the bon ton had solid Arabian origins in the Good Old Days.) But the Kusaasi were still stoutly maintaining that they are autochthonous, last I heard. Good for them!
Sorry. Autochthonous Wandals, of course. An autochthonous Mandarin with his Muslim daughter in pants*.
*The daughter is not autochthonous**.
**Badr al-Buduur!!!!
“saying that vowel height is itself a prosody, as Heath does for Tamasheq”
It would be “it is not impossible that vowel height is itself a prosody”. I think the idea is that the reconstruction of two vowels in reality means “two or one”. The quote is from 1970. I’m not sure if “a prosody” was an idea so easy to come up with.
A:[smoe txet] / A:*[some text] / Where does it come from?
in english texting too! (at least in my circles) to some extent, i think it evolved here into:
A:[smoe txet]
A:/[some text]
i dunno about the origins; all i’ve got is wondering whether it’s stylizing the correction as a philological footnote. the slash is, i think, to be more straightforwardly read as “or”, or maybe “or, better”.
The daughter is not autochthonous
patrilocality will do that to people.
I’m not sure if “a prosody” was an idea so easy to come up with
Indeed not. It’s perversely ingenious. It really does work well for some languages, though.
I first encountered the idea in Foley’s Yimas grammar*, where it’s really well explained. For all that, it seems to work rather better in various Chadic languages than it does in Yimas.
* It really is a pity that there is no pdf version of it that people could pirate. I mean. buy. It’s a work of art.
xkcd on asterisk corrections.
spent a while deciphering ʒam (was still waking up).
The three Russian issues of the dictionary I spoke about are here. 3 issues. It’s incomplete, the English version is 5 issues.
(I don’t think anyone will find it useful, but.
I also don’t think it is “good” – have no reason to think anything – though Takács thinks about himself as a continuator of this effort).
I should have mentioned that the Ket tone in question also comes with a bit of extra length. But in Mandarin, the first (high, level) and the third (dipping) are the sometimes-somewhat-long ones; the second (high rising) and the fourth (falling) are shorter if all else is equal.
I haven’t noticed glottalization in what Mandarin I’ve heard, but the third tone does come from a segmental glottal stop, and glottalized register-tones are all over the region (Burmese, Vietnamese…).
I came across this paper on tone and glottalisation in Mandarin when poking into the issue of glottalised vowels in general:
https://pubs.aip.org/asa/jasa/article/142/3/1693/613278/Covariation-between-voice-quality-and-pitch
Glottalised vowels actually seem to be a pretty heterogeneous group phonologically across languages; in fact, there seems to be often quite a lot of variation even within some single languages.
https://www.researchgate.net/publication/281119746_Acoustic_properties_of_different_kinds_of_creaky_voice
In Kusaal, and the other WOV languages that preserve it, vowel glottalisation is completely orthogonal to tone. That’s not especially unusual cross-linguistically (at least, no more than contrastive vowel glottalisation itself is unusual) though it would be odd for SE Asia. It probably ties in with how Niger-Congo tone systems don’t arise from segmental features in the first place.
Although I was conjuring up a mere epicycle to save the phenomena, the synchronic systems of tone in Niger-Congo really do seem to behave significantly differently from those of SE Asia. I suspect rather different phenomena have got lumped together by European scholars who found them all equally exotic (as with glottal vowels.)
“haven’t noticed glottalization in what Mandarin I’ve heard,”
A friend of mine substitutes the third tone for ʿayn [ʕ] ع and her Arabic teacher is quite happy with the result:)
Well, pharyngealisation can work like a prosody too …
Western Oti-Volta vowel glottalisation certainly arose from lost consonants after the vowels (in fact, one instance of it in Toende Kusaal has happened since Prost wrote his grammar in 1979.) There was definitely more than one source. You could nearly represent glottalised vowels phonemically as V + epenthetic vowel (as opposed to long V:) but it won’t work in Kusaal, because short glottal vowels are also common word-finally. The Kusaal habit of deleting final vowel morae spoils so many neat phonological solutions …
So “prosody” is that thing when a European only learns about a(n?) ع from what she perceives as colouring of the vowel?
Well, I meant it quite literally: pharyngealisation can spread across syllables and even words.
But yes, the vowel quality works as a cue for Europeans (and not just Europeans.)
The case of ع is striking to us because we don’t have that consonant, but actually all kinds of consonants blend into following vowels (and vice versa) in various ways. The neat segments we (seem to) hear are helpfully generated for us by preprocessing that we are unaware of. (DM is our local expert on this …)
DE. not only this: its articulation can be weak. But I don’t know what goes on in terms of articulation when it sounds as a “coloured vowel”.
As for its “spread” you’re perhaps* thinking of the “spread of emphasis” (quite distinct from what ع does).
*Sorry, not “perhaps”. That’s what you’re thinking about.
What you’re calling “emphasis” is pharyngealisation.
https://en.wikipedia.org/wiki/Pharyngeal_consonant
https://en.wikipedia.org/wiki/Pharyngealization
IS called so by some.
I’m not a phonetician and there is some disagreement as to what it should be called (and it is more than one thing anyway… unless everything else follows from pharyngealisation, and not being a phonetician I don’t know what “follows” from it)
One of the 19th-century Russian names for ṭ etc is дебелые. I totally like it. The word is dated now, and usually refers to people, so it sounds funny. I understand it as robust, stout – but Wiktionary marks these meanings as “dated”. The meaning which is not “dated” (“plump”) is unknown to me:) And the word itself is dated. (though I sort of understand how people lookign for words for plump people can euphemistically use д.)
Arabic “emphasis” is often uvularization, and ع is actually epiglottal in many places, pulling vowels in the opposite direction from what a pharyngeal or uvular consonant does.
Babylonian Akkadian fronted /a/ > /e/ next to (original) /ʕ/ and its voiceless counterpart; also, to confuse things further, often between original /ʔ/ and /r/ or /n/. When the language lost these “guttural” consonants, it thereby acquired a new short vowel phoneme.
Akkadian “emphatic consonants” were glottalised rather than pharyngealised/uvularised, and had no effect on vowel quality.
That’s why I think they were epiglottal there.
@DM, do you remember in what dialect(s) ʕ causes raising? (especially if you are speaking of ʕ making it more likely than between say /m/ and /k/)
In Tunisia (and not only) imāla is super variable. Intrapersonal variation, across dialects in same small region, sometimes lexicalised (not often). And affected by anything you want it to be affected to:)
or mṛā can have a variant with /r/ and imāla.
Though some of variation is interdialectal contact.
Ancient Egyptian seems to have managed to change *d > ʕ, which strikes me as a great candidate for the Kuznetsova-et-al zoo.
https://en.wikipedia.org/wiki/Egyptian_language#Classification
(Disputed, though; another sign of the in-reality-primitive state of AA reconstruction.)
“*d > ʕ” – wow.
“in-reality”
Redundant.
To you, maybe. Alas, not to all.
DE, not to whom, specifically?
Ehret, Orel, Stolbova …
Gábor Takács, “Semitic-Egyptian Relations”, in Weninger, The Semitic Languages (2011), pp. 9, 13–14:
Rössler (1971) is: Das Ägyptische als semitische Sprache. In: F. Altheim & R. Stiehl (edd.). Christentum am Roten Meer. Band I (Berlin, New York: Walter de Gruyter) 263–325.
I’m not talking about imāla. Pharyngeals (and dorso-uvulars but not [ʀ]) pull vowels towards [ɑ]; epiglottals pull vowels towards [æ]. Epiglottals are an areal feature of the Maghreb or part thereof (Korandjé has them, too), and I’ve heard them from elsewhere… Lebanon maybe?
There’s at least one language in Dagestan that has both (apparently as allophones conditioned by the large vowel system).
Egypt probably.
(Test to see if the edit clock reappears.)
Watters, Notes on Kusunda Grammar (2006), p. 26: “Pharyngealization in Kusunda produces sounds very different from the sounds of breathiness in Tibeto-Burman. Pharyngealization is manifested by a ‘ratchety’ quality on the vowel, almost as dramatic as the cawing of a crow, and on a spectogram displays a wave form with a series of rapidly occurring striations.” [Figure follows.]
I read this as a pharyngealized epiglottal trill, but that is just a guess.
Takács, A New Idea on the Origin of πῡραμίς, link (not sure it is interesting reading, but the title sounds ike something for LH)
Das Ägyptische als semitische Sprache
All together now: “Oh, no it isn’t!”
(Also: “Look behind you!”)
`@DE, I don’t like the fact that I can understand your comment.
Me to Lameen: ““Fundamental methodological problems” (to quote Takacs) are by no means a reason to dislike his effort and the fact of its publication. Especially when it provides – as I think – motivation for Takacs and others.”
Gábor Takács, Preliminary report on the new comparative-historical phonology and etymological dictionary of Southern Cushitic – here
“…stimulated my research on the Southern Cushitic comparative-historical phonology and lexicon”
Takács idea is that Ehret’s SC reconstruction, “the highly precious lexical treasure accumulated by Ehret (1980) has to be almost completely re-arranged” because it is “not being at all void of serious drawbacks and imperfections“.
Also that its use as a basis for strange Cushitic and AA reconstructions makes the task urging.
This is quite the extraneous footnote from Takacs.
> This timing is due to the happy fact that the obligation of an enormously time-consuming and intellectually destructive commuting to the remote Hungarian capital has ceased in April 2019 and, henceforth, I have been able to exclusively devote myself to a desired full-time research in my private Afro-Asiatic library, established at the turn of 1991/2 in Székesfehérvár and re-built in spring 2015 at Balatonederics.
I’ve never really seen anyone quite so bitter about having to go to the office. Where was he commuting from? The South Cushitic area?
Wow, “intellectually destructive commuting” — what a phrase!
Ryan, pay also attention what he says about the library. “Office” for many (not all) Soviet* scholars is the place where you can’t normally work. People distract you and there is no sofa. As for books you need for work, it depends. But his are at home.
(But when the road takes a hour and a half (3 hours a day) that sucks enormously as well)
It can be different when you job involves lots of communication with people, when you don’t have the habit of working at home – and thus haven’t turned your home into (or think of it as) a much better office or when their library is better than your library (which depends on the effort invested in making your home a better office.
Note also that here books are cheap, and while books he is using are not necessarily “cheap” while scans are cheap everywhere – having very large libraries strongly affects people habits and perception of home.
*Of course he is not Soviet, not in the narrow sense. But.
He should have spent that time profitably listening to South Cushitic podcasts and audio books.
JK
But living an hour and a half from your office is a choice.
It’s interesting to see him criticizing Ehret’s profligate cognate sets with shameless semantic stretches and then find the Phoenix blogger accusing him of same.
But I guess those are the coins of the realm of comparative linguistics.
If he lived in Székesfehérvár and commuted to and from Budapest daily, that’s at the upper end of what Europeans are willing to put up with. How bad it was I don’t know; ideally, he could’ve worked on a laptop on the train; worst case is he had to drive (and routinely get stuck in traffic jams).
I just tried to make an epiglottal trill (which has been reported from a few other languages that have epiglottals). It does sound strikingly crow-like. Pharyngealizing it seems to make trilling impossible; on the one hand, that may be inevitable because the epiglottis is in the pharynx*; on the other, maybe I’m just doing it wrong, being thoroughly unfamiliar with any languages that have anything pharyngeal** or epiglottal.
* That seems to be the only real objection to recognizing “epiglottal” as a separate place of articulation from “pharyngeal”.
** …aside from the Meidlinger L of Vienna’s 12th district and surroundings. Start with the Czech velarized one, and (sometimes) wwway overdo it, complete with spreading in all directions.
Or a financial constraint.
The problem is that there’s nothing approaching an overarching theory of semantic change. We do have a general idea of which changes are more likely than others, but the details are all unclear, and quite flabbergasting changes are well known in the documented history of, say, English and German, with more that have to be added for cognates between these two, so it’s hard to say a change is flat-out impossible.
Ryan, what blogger?
“1.5 hours” is my friend in Tehran, I don’t know abotu Takács. It’d call it “long, but not uncommon” for Russia.
“1 hour” is what it took for me when I went to school. (but THAT wasn’t a probelm)
If you mean this post, it makes the accusation (and many others!) about Berber, not South Cushitic.
South Cushitic might suffer from another problem: critical existence failure. Takács isn’t mentioned in that post, though.
Needless to say, when I quote Takács, I do NOT mean that Takács is a “good, serious scholar”. I simply can’t tell.
The point is different (simply that what I felt about him is true).
My objections to Ehret’s PAA and PNS do not apply with anything like equal force to his PSC; I haven’t looked at it closely, but it deals with a much more tractable problem
Lameen, objection to his reconstructions are, of course a good thing.
We are arguing about you and DE’s objections to the fact of their (reconstructions’) publication. Of course the point of publication of one’s results is also giving everyone the opportunity to criticise them!
But the question is whether this publication has slowed down the “progress”. And how it slowed it down.
And, yes. Takács is speaking about Ehret’s South Cushitic.
Yet he refers to its use in Ehret’s AA and says Ehret’s AA and its flaws make work on CS “triple urging”. Maybe rhetorically.
In Bucharest having a 90 minute commute is unfortunately not at all uncommon.
Lameen, this publication bothers you and perhaps for a good reason. But I think, instead of analysing the issue you – based on, perhaps, ethics – predict something, I believe, unrealistic. If this publication increased the value of work from the bottom for us, why think it will be different for everyone in the field? It is not different for Takács and i think it is not different for other comparative linguists. And why speak about it when we have so many exiting works?
As I understand, it not commuting from a place where he can’t work to a place where he can.
It is commuting from a place where he can work to aplace where he has some other business (and perhaps can work as well). This is more annoying and makes you dream about getting fired rather than about “moving to Budapesht”
@DM: I just tried to make an epiglottal trill (which has been reported from a few other languages that have epiglottals). It does sound strikingly crow-like. Pharyngealizing it seems to make trilling impossible; on the one hand, that may be inevitable because the epiglottis is in the pharynx; on the other, maybe I’m just doing it wrong, being thoroughly unfamiliar with any languages that have anything pharyngeal** or epiglottal.
The pharynx can still expand above the epiglottis, by its own muscles or by the root tongue moving forward, lowering the perceived pitch from crow toward raven.
At least some of the death-metal growls are epiglottal trills. There’s science and videos galore. I confess that I don’t know all of the guttural apparatus as well as I would like, particularly my own.
The WP article on Dutch phonology describes the “dark” /l/ in the north as pharyngealized [lˤ] vs. velarized in the south, and the word final one as [ɤˤ]. /ɔ/ is described as pharyngealized as well.
The /l/’s in the Arabic name of God are pharyngealised/velarised/whatever.
I gather that in most forms of Arabic the sound occurs exclusively in this word, which raises some interesting theoretical questions.
Not too unusual for a Russian who has a weird sound in Господи ‘Lord!’ and /x/ in бог ‘God’.
“I don’t know all of the guttural apparatus as well as I would like”
Made me think about that thread about “the guttural”.
Not too unusual for a Russian who has a weird sound in Господи ‘Lord!’ and /x/ in бог ‘God’
A similar but less extreme phenomenon appears in standard Scottish English (not the same thing as Scots at all), in which /x/ is a phoneme, but restricted to a pretty small group of what are actually loans from Scots. (The sociolinguistics of this is interesting and rather counterintuitive: working-class Scottish English does not have /x/ …)
Kusaal /h/ is confined to loans too, but nevertheless really has to be considered an echt Kusaal phoneme, because one such loanword (hali) is thoroughly established in the language (it’s the usual word for “very”, even. Also for “even”, even.)
There’s quite a bit of that sort of thing in West Africa – several languages only have /p/ in loanwords, for example.
You could probably say the same about the Russian words in question: loans from Church Slavonic. Does that work?
The weird thing about الله is that it isn’t a loanword. As far as I know … It’s also far from clear (at least to me) why it should have this unique phoneme in the first place. But I wouldn’t be totally surprised to hear that there is an extensive literature on this very question. If so, some Hatter will know …
Yes, almost this.
It’s a Slavic isogloss, not unlike /j/ in Arabic. Within East Slavic the letter Г can be either [g] or, in the south, voiced fricative sounds (more than one).
The difference is that the Bible was not revealed to its translators, does not come with descriptions of articutions by grammarians and schools of reading. The spread of the southern pronunciation is a mere fact of historical sociolinguistics (not only ‘sociolinguistics’: switching from /g/ to a fricative is a decision with a specific historical reason rather than something intuitive).
*/alʔilaːhu/ “the god” somehow losing its /i/, and the impossible cluster /lʔl/ contracting to the one-word phoneme /lˁː/ around the same time as the other “emphatics” stopped being ejective?
On Greenberg’s African classification, there’s a chapter by Harald Hammarström in the new Festschrift for Tom Güldemann (here), analyzing Greenberg’s methodology and the scholarly acceptance of his four-family model. An appendix (linked from the article) includes a list of subnodes in Greenberg’s classification with comments on problematic ones.
Could be most straightforwardly (a), since Hungarian also epenthesizes -a- ~ -e- (depending on, you guessed it, vowel harmony) for the majority of its Slavic loans with CR- clusters. (But they do come from many more than one dialect for over a millennium now, and few universal generalizations can be made.)
An evocative turn of phrase, I must say
…made all the more so by that, going by this analogy, I would claim even his Proto-Cushitic, Proto-South-Cushitic and Proto-East Cushitic reconstructions (the last less known since he has not been the first on the scene) do follow the same recipe. The proportion of rocks seems to merely happen to be smaller and some washed potatos seem to have now made their way in by the quirks of procurement (“a much more tractable problem”); regardless the product is unadvisable for eating and will have to be carefully dissected and re-cooked if we wish to retrieve edible nuggets from it.
I have so far not found much time to dig into Takács’ extensive work, but the way his ’21 paper linked above refers to “rich variety and such a full system” of sibilants and gutturals might not be indicative of the correct track happening here either. E.g. very few of the allegedly numerous sibilants have actually good correspondence sets to them (most surface inventories are modest) and it seems to me many could result, once outright rocks have been discarded, from internal or external borrowing. Given e.g. West Rift + Kʼwadza *tsʼ corresponding in Asa to both š and ʄ, the first might be native and the second might be mostly borrowings from relatives (plus a recent implosivization), instead of setting them up as something like *ts versus *tsʼ.
(Other reasons to suspect this kind of a history seem to exist also. I have been doing review of Ehret’s collected Kʼwadza data the last few days as it happens; it seems worrisome how many of the supposed South Cushitic material is reflected just in Kʼwadza and Dahalo specifically, and nowhere among the much better documented West Rift languages; and I may be seeing some opportunities to clean up overlapping sound correspondences here too.)
there’s a chapter by Harald Hammarström in the new Festschrift for Tom Güldemann, analyzing Greenberg’s methodology and the scholarly acceptance of his four-family model
Ouch.
There are some speculations on the sociological reasons why this deeply flawed edifice did not suffer the same fate as Greenberg’s other ill-founded long-range proposals; partly the awfulness of the Hamitic theory (though the relevant actual experts knew it was bollocks well before Greenberg) and partly Americans preferentially citing other Americans.
I look forward to the day when the consensus will finally get round to acknowledging that no part of “Atlantic” has so far actually been shown to be genetically related to Volta-Congo by acceptable comparative methods.
Hamitic was explicitly based on race. Greenberg’s Indo-Pacific implicitly so: Andamanese, Papuan, and Tasmanian, all spoken by the dark-skineed people of SE Asia and the neighboring islands (Blench called it “the crinkly hair hypothesis.”) I cruelly speculate Australia has been left out because by then its languages were too well documented, synchronically and diachronically, to allow for etymological wiggling.
‘Hamitic was explicitly based on race.’
Y, I don’t know what DE means by Hamitic theory. I wash I did, regardless of our disagreement.
But you, I think, mean the terminology: the two names, Semitic and Hamitic. Of course it is based not on isoglosses, but you can’t seriously claim that Berbers are ‘Hamites’ based on race.
I don’t, but others did. An old but good article by Sanders tells the story of how “Hamitic” was first created to absolve ancient Egypt of Africanness, and how it accreted other ethnicities over time, with increasing difficulties. Seligman, for one, managed to shoe-horn some dark-skinned groups into Hamitic by appealing to admixture.
Paul Theroux, in The Happy Isles of Oceania, where he meets some Mormons in Aitutaki, in the Cook Islands:
Y, sorry, I misunderstood you: I thought you meant the idea that Hamites are not Semites because they’re black (and objected that this doesn’t work for northern Berbers).
But if you mean, ‘both are unlike other Africans’, I’m confused. Semito-Hamitic is still in use. We use it! It’s 1.called differently in English literature, and 2. it’s not clear where they came from, Africa or not.
Of course drawing lines between us and them, while not racist per se (people are obsessed with classifying things and try hard to fit chunks of continua in boxes) is convenient for racists, and the simplicity (when lines between cultures (called civilisations), languages and appearances are drawn in same place) is convenient to them – and not only – too.
But this linguistic line to which H. refers is still drawn, and have always been based, I think, on grammar…
.
Sure. A Russian linguist using “семито-хамитские” in 2025 means and thinks something different than a British one using “Hamitic” in 1925.
@Y. thank you for the link. I understand Sanders, but nether you nor DE:(
DE is speakiong about linguistics and you, I think, are speaking in the context of it. Sanders is speaking about something that has nothing to do with linguistics:
“The Hamitic hypothesis is well-known to students of Africa. It states that everything of value ever found in Africa was brought there by the Hamites…“. (I thought this is NOT the “Hamitic theory” DE is speaking about, but if it is, what on Earth does it have to do with Greenberg?)
Y, no, why?
The exact shape of the group changes (still does), but already in 1840s it is more or less Berber, Egyptian, Cushitic – and Semitic.
Ueber das verhältniss der ägyptischen sprache zum semitischen sprachstamm, https://archive.org/details/ueberdasverhltni00benf/mode/2up (without the name “Hamitic”, but does that mean anything?)
Or wait. I haven’t read the whole book:) I read about it that Coptic “et toutes les langues de l’Afrique septentrionale jusqu’à l’Atlantique” are named there as a branch of Semitic. But I don’t know whether Beja or Somali (or Oromo or whatever) are counted as languages of the North Africa.
However I think it won’t be easy at all to find ‘explicit race-based’ argumentation in that book.
@Y: you are more optimistic than i am. there’s plenty of the theological race science that the terms are based on out there – i don’t know many youtube recommendation jumps it takes to get from zero to “hamites”, but it won’t be a large number – and neither academics nor scientists are immune from it. i mean, gimbutas and pearson – different flavor of the same dish, but one that’s much much more explicitly socially deprecated – have not exactly been wholeheartedly repudiated by their professions, even though more of the anthropologists and archaeologists who agree with their Aryanist work today than previously have to make a more active choice between the “legitimate” academy* and the equally viable career path in the openly fashy sphere, and tune their rhetoric accordingly.
but, really: both terms should be taken out of use in any linguistic context. “hamitic” because it’s always been an explicitly racist term in both theory and practice. “semitic” because even if the group of lects that’s been placed under it is perfectly legitimate, using a term for them that has several specific popular meanings that don’t match the intended referent, and that comes just as definitively from a racist classification system, is a bad habit worth discarding and will continue to actively cause problems** if it’s kept. as it has ever since before people started criticizing it because of those problems – especially the “consistently legitimizing racist ideologies” one.***
.
* which in the u.s. as i type, is, at the level not just of institutions but the industry as a whole, illustrating my main point about the problems caused by the continued use in linguistics of “semitic”.
** see note above, and also all the rest of the trashfire (from “anti-arab racism is also anti-semitic, because arabic is a semitic language” to “there’s no such thing as anti-semitism, because jews aren’t semites, because yiddish isn’t a semitic language”).
*** this is also why the use of “gender” for “noun-class” needs to be abandoned – it similarly serves no function but to exceptionalize SAE languages and to give aid and comfort to those promoting exterminationist anti-trans politics. the purpose of a terminology is what it does.
As for Russian linguists – yes, if you claim that
– the group called “Semito-Hamitic” is race-based
– once renamed it is not race-based
then technically (and only so!) you’re ascribing somehing to numerous linguists whose native language is not English and if you don’t intend anything like this, perhaps (logically and only so) you should find a different name for “explicitly race-based” groupings. But I personally don’t care at all. I know you don’t mean anything like this – and that’s enough. I never heard of a linguist who thinks that Russians are racists based on this.
@rozele
How about PHUAAA! languages (Punic/Hebrew/Ugaritic/Akkadian/Arabic/Aramaic)?
Jordano-Tigridic?
I like the Africanist hydronym-based system (or, in the case of Nilo-Saharan, dehydronym.) Sadly, my proposal of “Tarim-Liffey”, in place of the unsatisfactory “Indo-European”, has received surprisingly little support.
That’s because that’s just Indo-Tocharian!
Kızılırmak-Tarim-Liffey…
On reflection, “Awash-Tigridic” is better than “Jordano-Tigridic.”
there’s plenty of the theological race science that the terms are based on out there
The vile Curtis Yarvin, influential on the technofascist Thiel and the moral vacuum Vance, opined that US blacks had “recent hunter-gatherer” origins (engagement with fact is not a core fascist value) and combined this with adumbrations about the debased character of “clicks.”
These fuckers have never really gone away.
debased character of “clicks.” – oh, like Lepsius!
(he knows clicks!)
And even Cust, alas. It seems to have been the accepted wisdom at the time that clicks were basically Bad.
Idle to speculate on the “intellectual” sources of a creature like Yarvin, but given that he probably has never even heard a click used as a phoneme and evidently neither knows nor cares anything about the actual real Africa, one imagines that this “subhuman” characterisation of clicks has become a trope somewhere in the “alt-right” sewer system, where nineteenth-century and early twentieth-century “scientific” racism remains the canonical wisdom.
A sad aside, incidentally, in Steeman’s grammar of Sandawe, is that the presence of clicks in Sandawe makes the language readily recognisable to other Tanzanians, and that this causes Sandawe people to converse in Swahili even among themselves when outside their home area.
@DE, they’re sticky.
As is known, L1 English L2 Khoisan speakers have them in their English.
And everything truly GOOD must be DIFFICULT. Like, you know, being a good person:)
And of course, southern African Bantu languages have acquired them from “Khoisan.” I gather that even now it’s far from clear quite how this happened, exactly. Handy tabu-replacement strategies? Women speaking those languages, especially, have to cope with a lot of language tabus:
https://en.wikipedia.org/wiki/Avoidance_speech#Africa
That page links to this:
https://en.wikipedia.org/wiki/Noa-name
which was a new term for me. I shall have to work it into my conversation as soon as possible.
Words for “lion” quite often get noa-ised in Oti-Volta. Besides various paraphrases along the lines of “savanna lord”, even the ordinary Kusaal word gbigim and its cognates can probably be analysed etymologically as “the powerful one.”
DE, any reason to think usual areal convergence doesn’t work with them?
Work with who? Southern African Bantu-speakers?
I have only vague recollections of papers I’ve read on this, but I think that the problem is that the patterns of occurrence of clicks in these languages don’t seem to be explicable by them having been adopted along with loanwords which contain them (as with e.g. the /h/ of Kusaal hali “as far as”, where /h/ is not found in non-onomatopoeic words of native origin.)
They appear in words of undoubted proto-Bantu provenance, for example, but not in a way reducible to nice Neogrammarian sound laws. And in other words with no “Khoisan” cognates.
Obviously this is an areal thing, inasmuch as this is (very nearly) the only part of the world where you get click phonemes; but that doesn’t explain how the transfer of these sounds actually happened.
With clicks:)
DE, aha, I was thinking about it (loans or native words? are they predictable?) and even thought “native, unpredictable”.
I think, when something surprises us, it’s a good idea to ask why is it surprising. There is a plenty examples of phonological convergence which are not suprising. But of course we don’t expect convergence to take the form of “influenced by aliens, these Russians pronounce hard Russian /t/ as [t] and soft /t’/ as an epiglottal [anything]”. For speakers, t’ and epiglottal anything are different planets, but as I understand, clicks are not a different planet.
I’m resisting the temptation to look for the “taboo replacement and the bear” thread.
Taboo deformation is explicitly mentioned there as actually occurring, so the idea that that’s where clicks came in (which I’ve encountered before) is attractive.
Here you go.
Shurely in old-school racial-hierarchy pseudo-science, clicks are diagnostic of groups even more quote unquote primitive than yer generic Bantu or speaker of West-African-languages-Greenbergianly-lumped-with-Bantu. Click-users are quite rare in the ancestry of the African-descended populace of the New World (although so are Swahili-speakers). Is the takeaway supposed to therefore be congratulations to the slave-traders for drawing forced labor from not-quite-so-primitive sources? All of which is to say that I don’t understand the internal coherence of the line of argument being attributed by David E. to Yarvin. (I have not put in the time to try to locate actual running prose generated by Yarvin where this occurs, so there may be context I’m missing, of course.)
On the contrary: I think Yarvin was merely being abusive, drawing on the shared repertoire of “alt-right” cartoon race tropes. Pointless to look for any actual underlying factual basis, no matter how distorted. Coherence is not the objective here.
Yarvin seemed to be under the misapprehension that African slaves were drawn from hunter-gatherer groups (as opposed to the reality, that they came from societies that had been agricultural for millennia), but it’s a category mistake to ascribe error to a bullshitter. It would only be an error if he gave a fuck about facts.
It’s always useful to be reminded that people like Thiel who treat the Yarvins of this world as coherent thinkers are not geniuses. The question to ask of them is: If you’re so rich, why ain’t you smart?
OTOH, having click-using ancestors is positively correlated statistically with presently being an L1 speaker of Afrikaans, which as a West Germanic language is indicative or at least suggestive of sophistication and Kultur. Or something like that.
Well, obviously on a lower level than Celtic … but we can’t all be Brythonic Übermenschen. We need some Germanic types to be our hewers of wood and drawers of water while we superior beings devote ourselves to elevated pursuits like composing cywyddau and conducting human sacrifices. If the servants wish to entertain themselves with their primitive skaldic verse and suchlike, no harm in that, so long as they don’t shirk their duties.
The lack of semi-creolized Celtic tongues resulting from wholesale miscegenation (I think “admixture events” is what they’re called more politely) is I guess confirmatory of such high-minded devotion to supposedly elevated pursuits. Even in Patagonia, no creolization.
Purity of essence!
My Argentine grandfather used to claim to know a Patagonian war dance … (I was, however, quite young at the time, and with hindsight, the fact that the words ran “Om mane padme hum” does suggest he may not have intended his account of the matter to be interpreted quite as literally as I then supposed.)
“From the earliest times the Welsh have been looked upon as an unclean people. It is thus that they have preserved their racial integrity. Their sons and daughters rarely mate with human-kind except their own blood relations. In Wales there was no need for legislation to prevent the conquering people intermarrying with the conquered. In Ireland that was necessary, for there intermarriage was a political matter. In Wales it was moral. I hope, by the way, you have no Welsh blood?”
“None whatever,” said Paul.
I like that “with human kind”, allowing the possibility of friendly relations with the livestock, an old Saxon slur still occasionally believed by the unenlightened.
“the idea that that’s where clicks came in (which I’ve encountered before) is attractive.”
As an explanation of the distribution of clicks across Bantu words or as an explanation of the fact that due to contact and intermarriage a certain articulation from language A found its way in B?
What are we explaining?
“None whatever,” said Paul.
To discover the source of that quote, I first searched for “In Wales there was no need for legislation”. The Google KI response (always the top one on the page) was a bit stroppy:
#
The statement “In Wales there was no need for legislation” is incorrect. While the evolution of law-making in Wales has seen shifts, particularly with the establishment of the Senedd (Welsh Parliament) and the granting of primary legislative powers, legislation has always been a key part of the process. Early on, the National Assembly for Wales …
#
I wonder what “the process” refers to here. At any rate KI appear to have gotten over the abject fawning they used to deploy when they wanted to contradict you. Diddums is all grown up ! It seems like only yesterday that …
Hilde Gunnink has been doing much work recently on the history of clicks in Bantu languages (bibliography here). Specifically, here (“The adoption and proliferation of clicks in Bantu languages: the role of hlonipha revisited”) she concludes, “the adoption and proliferation of clicks in Bantu languages is unlikely to have resulted from hlonipha alone, and other factors, such as identity marking and sound symbolism, should also be considered.”
Curiouser and curiouser! Sound symbolism! Identity marking! Hlonifa! Are clicks so weird that no one will borrow them* without a properly resounding reason to name?
If you read h as Russian [x], it will sounds really like Russian slonixa, she-elephant.
hl here is [ɬ]. And with [ɬ]… it still sounds like children’s slonixa. More or less.
*Unless, again, we are talking about the distribution. There is something to explain there, but everyone formulates the question differently.
@JWB, “Khoisan” is not only hunter-gatherers, but also Hottentots. (I always thought that people of the desert are less prone to assimilation because until governments invented natural reserves [and some say, to profit form selling safari tours there] no one needed desert and that they are normally hunter-gatherers)
I don’t know if any hunters established a relationship with people who came from Europe to be counted as “Hottentots”, but I think some Khoisan-speaking herders did.
And I think, their role, distinct from “slaves” or “we don’t know how to deal with them at all” or “warfare” can be interpreted as a higher status than that of some other local groups.
(There also was a scene in a 19th-century travelogue where a Hottentot lady on the street looks at women wearing bustles, laughs, points at her buttocks and says something to the effect “these are real”.)
Thanks for the links, Y.
The occurrence of clicks in words of Bantu origin, where clicks don’t result from any known regular sound change.
I have the dead-tree book this chapter is from; just discovered this pdf of it:
https://pages.sandpoints.org/dotawo/library/Tom%20Guldemann/Historical%20Linguistics%20and%20Genealogical%20Language%20Classification%20in%20Africa%20(186)/Historical%20Linguistics%20and%20Genealogical%20La%20-%20Tom%20Guldemann.pdf
At least, I think that I just discovered it. It’s at this point that I’ll discover that I linked to it before. Whatevs. Good stuff, anyhow.
@DM, aha, thanx!
L2 users of Khoisan are said to click in their native European language.
That is unusual.
Of course rather then believing or not this account we should study it, but what follows is
– inter and intrapersonal variation in a society where a large portion of speakers are bilingual
– how do we model behaviour of those clicks? This doesn’t normally happen to normal consonants, I’m not tempted to replace my Russian phonemes with entirely different Arabic or English or French ones.