This thread developed into a discussion of the parallel between the development of evolution theory and historical linguistics. Now Mark Liberman has a post at the Log about “how close we should expect linguistic and biological descent to be, in general. There are too many ways, both wholesale and retail, for people to end up speaking a language different from the language of their ancestors, and similarly many ways for genes to flow from one speech community to another.” He links to and discusses the abstract of Hafid Laayouni et al., “A genome-wide survey does not show the genetic distinctiveness of Basques”, Human Genetics (published online 1/16/2010), and I urge anyone interested in the topic to check out his post.
Off-topic: It has been lamented in the comment threads here that there’s no OED-equivalent for German – the Grimms’ dictionary was one, but it’s way out of date. Turns out an update is under way; volumes A to F, the ones actually done by the Grimms, are supposed to be completed in 2012. This update process results from a decision made in 1957…
My angle on this is that nineteenth century historians were trying desperately to make history scientific, and historical linguistics and social Darwinism were two of their main tools. They wanted to find essences like atoms to which the chaos of human behavior could be reduced and made intelligible, the way atomic theory was reducing the chaos of chemical and mineral behavior (and some biology) to order and intelligibility.
So they ended up constructing an aspirational scientific concept called “The People” or “The Nation”, which was ideally defined as one race (gene pool), one language, one homeland, one origin, one culture, one state. It was a real mess, a jerry-build construction joining evolutionary biology, historical linguistics, nationalist ideology and political theory, and romantic culturalism. It also made little sense, because the homeland of origin was often different than the actual homeland, and because a group’s descent did not match its actual language.
And because of the very concept of origin. Granted the depth of history, wherever you put the origins of a people, the question arises “What was before then?”, just like the question “Who made God?” God and the origins of peoples were constructed asymmetrically as as unexplained and unexplainable explanation principles.
This idea was extremely robust, probably because of its very messiness — science, literature, the state, and popular movements all jumped in to defend the general idea whenever it was threatened by one of the others.
And it worked. Even though in many ways this was all was crap, it produced the nation-state, which has proven to be extraordinarily powerful. And that’s the world we’ve inherited.
One of my favorite quotations ever, anywhere, from a historical philologist studying Egypt:
W.B. Kristensen once remarked that the supposition that the origin of a phenomenon is simpler and more easily understood than that which proceeds from it, is untenable. Every origin is in itself already a complex phenomenon, sometimes of an even more mysterious nature than that which it is supposed to explain.
“Seth, God of Confusion”, H. Te Velde, Brill, 1977.
Mark Liberman’s post starts with Charles Darwin saw the history of languages as a model for “descent with modification” in biological evolution.
As I wrote in a comment there, those studies seem to assume that the Basques remained genetically isolated, in which case their DNA, etc profile would be quite different from their neighbours’. But as the once larger Basque-speaking territory shrank (already in pre-Roman times), through a shift to more dominant languages, the former Basque-speaking community must have intermarried with the newcomers, so that part of their genetic material still exists in the geographical territories which were once Basque-speaking, and there is not a sharp demarcation between them and their neighbours. And it is unrealistic to think that no intermarriages ever took place between the current Basque-speaking community and its neighbours, although it would have been less common for non-Basques to marry into the more remote Basque communities than for intermarrying Basques to join the larger, non-Basque-speaking community.
I’m now very skeptical of the possibilities of either linguistics or genetics to tell us much about history very far back, because changes can both obscure old differences and create new ones.
I’ve already speculated that with a couple cycles of creolization and complexification, plus borrowing and areal effects, the relationship between a language and its ancestor might be imperceptable at the distance of maybe 2,000 years.
In genetics, Cavalli-Sforza found four genetically distinct populations in Europe: Saami (Lapps), Basques, Sardinians, and Icelanders. the Saami are a survival or an invader from Siberia, the Basques’ uniqueness is being disputed, and the Sardinians and Icelanders are “island effects”.
All the genetic difference in Iceland was created in the last 1,000 years. A small known population of known origin expanded and contracted for that long, with several sever dieoffs, producing completely novel generic uniqueness which, however, has nothing at all with strangeness of origin, but only with a unique recent history.
As a Austronesianist who has studied tiny language communities along the coast of Papua New Guinea, where virtually everyone marries outside their village (pop. 300, typically), I’ve always been suspicious that current language affiliations tell us much about the more distant past. One of my own favorite research papers on the topic was entitled The Population Kaleidoscope (published in 1997 in the Journal of the Polynesian Society), in which I assembled historical, anthropological, and linguistic evidence for very different distributions of language communities around the Huon Gulf (in Morobe Province, PNG), just in the past 300 years. Volcanic eruptions, earthquakes, floods, droughts, and constant warfare ensured that most small communities include descendants of groups of refugees from other language communities and that linguistic isoglosses run right through every family.
Do you think some of the drift, vaguely intimated disappearances, undisentanglable pollinations, and so on in the history/ies of languages are the result of the equally long history of paper-writing-service advertising?
I had given a reference (Schmoll 1959) on the thread of the December 7 2009 posting “The most interesting language”. Schmoll argued that, in the Spanish Basque country, Basque was the result of language spread from the North after the fall of the Western Roman Empire. The absence of “Basque genes” does seem to point in the same direction.
It seems to me that historical linguists tend to confuse the micro- and the macro-social when it comes to language spread: the fact that, globally, over the past two millenia, Basque has been losing ground to Romance does not preclude that it expanded at the expense of Romance at some times/in some places (Incidentally, much of Basque historical phonology seems more compatible with a Romance substratum than adstratum/superstratum).
The same can be found in other instances of language spread: the fact that over the past few centuries Russian expanded at the expense of various Turkic and Uralic languages is quite compatible with the documented fact that, especially in the early stages of Russian colonization, some ethnic Russians shifted to various (locally prestigious) languages (I have at least one reference on this, should anyone be interested).
Here’s a historical look at the Basque question.
Unfortunately the article seems to toy with the idea that the Caucasian Iberians had something to do with the pre-IE Iberians in Spain, though it doesn’t put much stress on it.
As I understand, the Romans were pretty casual about naming their subject people, with Albania, Iberia, Galicia being slapped here and there like Alpha Kappa Gamma etc., and sometimes they forgot and used one name twice.
I have at least one reference on this, should anyone be interested
*raises hand*
Etienne: It seems to me that historical linguists tend to confuse the micro- and the macro-social when it comes to language spread: …
Don’t you consider yourself a historical linguist? Please avoid sweeping generalizations. You could add “many” or even “most”, which would leave room for “some” historical linguists to diverge from the main pack.
But I too raise my hand to ask for the reference you mentioned.
JE: thank you for the link. A succinct statement of the current state of research into Basque(s).
You don’t have to go as far as Basque. You could be excused for thinking that Bobby Brown might be (for instance) a white dairy farmer from East Anglia, since there’s nothing in his name that would tell you otherwise.
M-L, when Etienne said “tend” I think that means that not all historical linguists do that. When you say X (singular) tends to do something, it means they do it some of the time, but if you say that Xs (a category) tend to something, it can mean that many of them do it, but not all and not necesarily even half.
It’s a deliberately vague word.
Genetic analysis can be useful in historical linguistics if its limitations are properly understood. Cavalli-Sforza apparently overestimated the overlap between genes and language. Genetics is just another point on the graph.
Thanks, JE, I guess I read too fast and missed “tend”. Etienne, mes excuses. As a historical linguist myself, I tend to be sensitive to what may appear to me (rightly or wrongly) to be unjustified generalizations.
Marc, I don’t think that that was Cavalli-Sforza’s problem; he wasn’t trying to match genes and language, he was trying to trace human groups (described by history and archeology) back to prehistory, and basically decide who was related to whom.
For example, a recent study in his tradition ended up concluding, for better or worse, that the British population is primarily descended from the pre-Celtic inhabitants, whoever they were, and then in descending order from Celts arriving after 400 BC, and then in descending order from Franks, Norsemen, and “Romans” (who were soldiers and bureaucrats from varying backgrounds, including Sarmatians). The languages spoken by these groups weren’t a factor.
But geneticists say that his very sophisticated mathematical analysis of the genetics had some flaws.
Cavalli-Sforza apparently overestimated the overlap between genes and language. Genetics is just another point on the graph.
I have a question for all the linguistics professionals: do you think that the low repute that Cavalli-Sforza has in your circles (much lower than his reputation among professional geneticists) is related to the fact that he chose to collaborate with Merritt Ruhlen?
If C-S did cooperate with Ruhlen I’m sure that that brought his reputation down with linguists.
I remember now that that stuff was in his book, but I just skipped it. His accomplishment rises and falls with the genetics.
The criticism of the genetics may be the same as the criticism of Ruhlen, cherrypicking the variables you choose to analyze in order to get the answer you want.
Cavalli-Sforza, Greenberg and Ruhlen (Greenberg’s assistant) were all at Stanford, so it is not surprising that Cavalli-Sforza and the other team should get together. Not having expertise in linguistics, C-S would naturally believe G and R’s claim that they were engaged in cutting-edge research and that their many critics were just mindlessly following outdated traditional opinion (distorted by G and R’s presentation, in spite of some valid criticisms).
Cavalli-Sforza found that for the Americas, genetic evidence corroborated G and R’s tripartite linguistic classification (Eskimo-Aleut, Na-Dene, and “Amerind”), but the first two families had long been established on linguistic principles, and have always been considered as representing the most recently arrived populations. No new techniques were needed to confirm the general consensus among linguists, and it was no surprise that the genetic evidence corroborated it.
“Amerind” is a completely different case, as Greenberg lumped together a huge number of independently established families, using a demonstrably unreliable method of comparison and accumulating a very large percentage of errors in the data. His “subclassification” within his new group largely followed other proposals, many of which had been discarded. As far as I know, Cavalli-Sforza showed a genetic difference between the “Amerind” speakers and speakers of the other two groups, but his research did not seek to confirm the subclassifications within “Amerind”. If speakers of the languages subsumed under “Amerind” had been in the Americas millennia before the arrival of the two Northern groups, it would not be surprising that genetically they had developed separately from those people remaining longer in Asia, as well as mingled together. But genetic difference or mixing does not mean that the languages of the group had a common origin.
As a linguist working in the area of language classification, my feeling is that the huge number of currently accepted language families lumped together under “Amerind” will eventually be reduced to a much smaller number of independent groups.
do you think that the low repute that Cavalli-Sforza has in your circles is related to the fact that he chose to collaborate with Merritt Ruhlen?
Well, associating with Ruhlen demonstrates that he didn’t try very hard to understand anything about linguistics before finding someone to collaborate with.
But for me the main reason to put him in the crackpot category is that he didn’t pay much attention to the structure of tree he and Ruhlen created “showing” how biological and linguistic human family trees “align”.
The tree (in Italian, but quite comprehensible) is here:
http://www.sucuncordu.net/sci-genilingue.jpg
I wrote about it on languagehat before:
http://languagehat.com/archives/003074.php
The short version is that the linear order of the two trees almost aligns, but only by manipulating the internal structures of the trees (which have no inherent linear order of the leaf nodes). Categories that are as close as possible on one side are unrelated on the other side, but they are conveniently “next to” each other in the linear order. Casual observers will see an amazing correspondence, but it is completely cosmetic.
Anyone who lets their name be associated with such a ridiculous analysis is not reliable.
The page referred to by Trey:
http://languagehat.com/archives/003074.php
is well worth rereading in light of the present thread (it is mostly about Indo-European but is relevant to other cases of language classification).
I wrote about it on languagehat before:
http://languagehat.com/archives/003074.php
The short version is that the linear order of the two trees almost aligns, but only by manipulating the internal structures of the trees (which have no inherent linear order of the leaf nodes). Categories that are as close as possible on one side are unrelated on the other side, but they are conveniently “next to” each other in the linear order.
You make a very important point here (so I’m quoting it in a longish form), one that I have long thought many people who construct genetic trees (and doubtless the same is true of linguistic trees) are quite unconscious of. So, for example (a simpler example than yours), if your data allow a clustering ((A,B),(C,D)) you can present it as ((B,A),(D,C)) if you want to suggest (without saying so in so many words) that A and D are “similar” to one another. Have I understood your argument correctly, or are you say something different?
This distorts nothing in the data, but it suggests that the data provide more information than they really do. I don’t think people who do this are being consciously dishonest (though some probably are), because I find that few people (again in biology) seem to realize how much of a subjective component there is in the way a tree is drawn.
Having said that, I would be surprised if Luigi Cavalli-Sforza were guilty of deliberate distortion (he would have too much to lose from being exposed), but he may well have had the wool pulled over his eyes.
C-S’s accomplishment is in genetics. The extension to linguistics was a mistake on many levels and, as I said, I ignored it to the point of forgetting that it was even there. Trey is basically wrong. It’s like saying that Newton’s absurd later studies invalidated his physics. Many past thinkers have combined good stuff and bad stuff.
C-S’s genetic work can be criticized too, but it’s not like Ruhlen. As I’ve said recently, it now seems to me unlikely that either genetics or linguistics will tell what we want to know about the distant past, but I have concluded that partly on the basis of C-S’s European results.
Larry Trask, a historical linguist and Basque specialist, posted a bit about Ruhlen to the sci.lang list a year or two before he died. It makes for amusing reading.
Here’s one example. http://groups.google.com/group/sci.lang/browse_thread/thread/6db0fed4c11015e8/888412968e2c7b5b
“You could be excused for thinking that Bobby Brown might be (for instance) a white dairy farmer from East Anglia, since there’s nothing in his name that would tell you otherwise.”
However you could not be excused for assuming that Bobby Brown is not descended from a white dairy farmer, although more likely from Somerset or somewhere else in that area. The farmer’s name may have been Jefferson.
This distorts nothing in the data, but it suggests that the data provide more information than they really do. I don’t think people who do this are being consciously dishonest (though some probably are), because I find that few people (again in biology) seem to realize how much of a subjective component there is in the way a tree is drawn.
I just looked at the “tree” diagram to which Trey linked. I have seen such diagrams from time to time, but not paid any attention to them. It had never occurred to me that what I regard as a mere mathematical object – a graph, here in particular a tree – could be used in such a primitive fashion to visually suggest relationships – and among inner nodes in different branches, for pete’s sake ! Are there really people who argue by neatly arranged diagram ?!
Graphs don’t have any canonical, static representation. Instead, they have canonical “traversal algorithms”. The essence of graphs to be walked on, not looked at. There are various layout algorithms for displaying graphs in 2 or 3 dimensions, but these have precious little to do with the traversals. I would go even further than Athel – the subjective component accounts for 100% of the way a tree is laid out.
I really don’t get the point here. Both in biology and linguistics people draw trees. In both areas most people admit that there’s an element of arbitrariness (I don’t say subjectivity or politics) in the way a tree is drawn.
Subspecies are turned into species, species are turned into subspecies, species are moved into different families, families are merged, new families are created, and so on.
Tree diagrams are a pretty good way to present your conclusions: Spanish and Portuguese are closer to one another than either is to French, Sardinian and Romanian are each off by themselves, all five of them are closer to one another than any of them is to German, and so on indefinitely. What’s the problem with that? There are tough calls here and there, but it’s hardly “subjective.”
The C-S / Ruhlen tree was overambitious, but there’s not problem with trees. (Actually, there’s the cladistics argument, but that’s a sort of tree too.)
1) Marie-Lucie: apology accepted. John Emerson explained better than I could my choice of the word “tend”.
2) The reference [drumroll, please!] (sorry, can’t type in Cyrillic): Ubrjatova, E.I. 1985. JAZYK NORIL’SKIKH DOLGAN. Nauka, Novosibirsk.
Early in the book mention is made of one family which, over three generations, underwent language shift from Russian to Dolgan (a Turkic language closely related to Yakut).
3) One thing I have wondered about Ruhlen: his early work was on the history of the Romanian language, and is quite sound. Now Romanian, in terms of diachronic phonology, is a *very* conservative Romance language. I have often wondered whether Ruhlen’s belief (that surface similarities between languages/language families separated over millenia will be visible through surface inspection) isn’t due to his unconsciously taking Latin to Romanian as “typical”.
Perhaps if his early work had been on the history of a phonologically less conservative language (Irish? French?), he would be a sounder scholar today.
Hello.
A brief comment on John Emerson’s comment right up there at the top.
I’m interested by your theory that nineteenth century historians created the notion of ‘nation’ in their efforts to make history more scientific. I would have thought that the notion of ‘nation’ is something that comes from a natural psychological tendency to essentialise differences between peoples, arising simply because individuals do not witness history but instead experience a snapshot of it.
I would add that in reality nation-states are few and far between, in part because of the doctrine of uti possidetis which governed the process of decolonisation. Rather than a redrawing of borders on ethnic lines, independent states were deliminated according to the borders forged by imperial powers, regardless of the composition of the population. Secession is strongly discouraged by international law; every state fears the creation of a precedent.
(Deliminated isn’t a word. Oops. I meant delimited, demarcated.)
Etienne: Merci.
I have often wondered whether Ruhlen’s belief (that surface similarities between languages/language families separated over millenia will be visible through surface inspection) isn’t due to his unconsciously taking Latin to Romanian as “typical”.
A while ago I mentioned the difference in the type of work done by historical linguists working within the solid framework of a large, obvious, well-known group with plenty of historical documents and a long tradition of study by a large number of scholars (such as the Romance family), and linguists faced with little-known language groups with scant documentation, little or no historical depth, and little or no scholarly tradition, as with many American Indian (and Australian, etc) languages. There are many cases of linguists whose historical training, like Ruhlen’s, started within the context of Indo-European or one of its component families and who may have done competent work in that field, who later became involved in the comparative and historical study of American Indian languages. Unfortunately, many features both of language and of historical techniques that can be taken for granted in the first case are not necessarily present or immediately usable in the second case.
A comparison that comes to my mind is with an art historian specializing in, say, church architecture of the Middle Ages, or even early Greek temples, later starting a new career in the archeology of Ancient Peru. Of course, such a person may eventually become an important figure in the second field, but there will be a lot to figure out, to learn and to unlearn, before they can make credible contributions to that field. The asset of their previous training and experience can become a hindrance if they try to apply it blindly to the new context, forgetting the difference in circumstances. Instead, a consciousness of the general principles and the applicability of the methods learned in the old context will allow the new archeologist to adapt them to the new circumstances.
Tree diagrams are a pretty good way to present your conclusions: Spanish and Portuguese are closer to one another than either is to French, Sardinian and Romanian are each off by themselves, all five of them are closer to one another than any of them is to German, and so on indefinitely. What’s the problem with that? There are tough calls here and there, but it’s hardly “subjective.”
JE, how do you imagine that “closeness” can be adequately represented by a tree ? To put it abstractly, a tree is a set of parent-child relationships (a set of sets, here parent-child pairs). As Athel put it supra:
Here is a graph showing this node-rearrangement – the same kind of think done in the diagram at Trey’s link.
I have no idea what you’re talking about, Stu. The tree diagram I’m talking about is not a mathematical object, it’s lines on a piece of paper. Both biologists and linguists use them to show the kind of relationships we’re talking about, and as far as I’m concerned they work fine.
As far as I know, no one uses the method you link to to show that A and D are closer to one another than B and C. What they do show is that A and B are closer to one another than they are either to C or D, and similarly for C and D compared to A and B. But A, B, C, and D are all close to one another compared to anything not descended from their common root.
if your data allow a clustering ((A,B),(C,D)) you can present it as ((B,A),(D,C)) if you want to suggest (without saying so in so many words) that A and D are “similar” to one another.
Actually, this is something I have been doing myself for precisely this reason (listing four languages belonging to two groups within the same family). But perhaps such a diagram or list needs to be accompanied by an explanation in order to prevent misunderstandings.
Rachel: sentimental nationalism, romanticism, historical linguistics, politics, and historical biology all collaborated to produce the concept of nation or people (Volk).
A lot of scholars, historians, and quasi-scientists tried to use it as a scientific concept, according to which “nations” have identities, like atoms or or molecules or species*, and in my opinion this is bogus and has been harmful, though it was powerful because it reinforced certain political transformations which were already in the works.
*The essentialist understanding of species is also bogus; species change and their boundaries are not absolutely clear. (Fundamentalists and racists tend to be species essentialists).
There’s a lot of argument about the earlier history of the sorts of identification with “nations” that developed during the 19th century, but before then these identifications, though extant, tended to be subordinated to other identifications (as they still are within Islam, were religion tends to trump nationality, and ethnic identity tends to be more geographical than “national”.)
The scarcity of true nation-states is part a consequence of what I said — that “nation” was a constructed and defective concept — and is in part a social problem. States often define themselves as nation-states at the expense of their minority peoples; for example, France has been pretty aggressive in repressing the Bretons, Alsatians, Basques, and Provencals over the centuries, and in some new states this tendency has become murderous.
The reference [drumroll, please!] (sorry, can’t type in Cyrillic): Ubrjatova, E.I. 1985. JAZYK NORIL’SKIKH DOLGAN. Nauka, Novosibirsk.
Thanks! As for typing in Cyrillic, I use translit.ru. The reference in Cyrillic is Убрятова Е. И. Язык норильских долган; it’s at Google Books, but of course only in “snippet view.”
Genetics also effect the voice box, larynx, shape and size, the mouth, the ear drum and all the other mechanical pieces.
The ability to convert audio inputs and outputs to and from the brain are modified too.
just athought.
John Emerson: actually, trees in historical linguistics are treacherous. Obviously, when we draw a tree of the Latin-to-Romance type what we are describing is a historical reality: Latin did, over time, become the various Romance languages. But when we draw more detailed trees within a single language family we rapidly run into a problem: mixedness, which is much more of a problem within a language family than between languages belonging to different families.
Thus, the fact that English has a huge Romance element has no brearing on its place on a linguistic family tree: it remains firmly in the Germanic branch of Indo-European. This is because we can distinguish native from borrowed elements: English is a form of Germanic with Romance loans. Well and good.
However, consider the position of Catalan within Romance: is it closer to Spanish or to Provencal? Well, it depends. It has a number of features in common with both. But which of these are native, and which are borrowed?
The answer is: there is no way to know. For that reason any decision as to where to place Catalan is arbitrary (the political implications of the choice ought to be obvious).
A family tree diagram is thus treacherous: on the one hand it will show undoubted historical realities (Romance languages stemming from Latin, Germanic languages from Proto-Germanic…), and on the other it will show relationships that may correspond to present-day preferences (dictated by non-linguistic factors) more than to historical reality.
Sure, but the Catalan-Provencal branching is at the absolute lowest level of the branching. And the harder it is to decide where to put a language, the less important the question is. As I remember (from literature), you have Catalan-flavored Provencal and Provencal-flavored Catalan, and Spanish-flavored Catalan and Catalan-flavored Spanish, and so on, and probably even more complicated mixtures. But this really isn’t about trees (which describe hierarchies of division) but about dividing up the continuum into parts, which is somewhat arbitrary,
That kind of problem is inherent in pretty much any historical science. People seemed to be saying here that trees don’t convey useful information at all.
John Emerson: I think we are pretty much in agreement: I agree with you that a tree conveys information. My point is that the objective reliability of a tree differs. The fact that English and German have a common ancestor, which is not the ancestor of French or Russian, is something no serious linguist would dispute.
On the other hand, whether Catalan and Provencal had a common ancestor which is not also the ancestor of Spanish…now THAT is something serious linguists can disagree on (and have disagreed on!)
Yeah, I’ve seen trees where Catalan is paired with Spanish and Provencal with French, and I’ve also seen criticisms of that analysis (which might have been influenced by state boundaries). Even so, this is not a criticism of trees, just a way of saying that trees are hard to get right, and that the final distinction is hard to agree on.
The problem is that there’s no way to draw two trees (a language tree and a bloodgroup tree) on a single page so that you can see how similar or dissimilar they might be, which is what’s wanted here, and so people tend to assume that if the first-order connections look similar, the trees are similar.
Emerson said:
“All the genetic difference in Iceland was created in the last 1,000 years. A small known population of known origin expanded and contracted for that long, with several sever dieoffs, producing completely novel generic uniqueness which, however, has nothing at all with strangeness of origin, but only with a unique recent history.”
Off course, but Iceland was a truely isolated and almost closed population. Basques are not. They are a continental population. They have been invaded, (almost as often as British or any other European, for instance), they have undertaken commercial contact with other people, they have made war, they have moved and they have received foreign influence, not only cultural, but even genetic too. Just as many other European people. But they show some strange features. That remains unexplained.
My point about Iceland was that if you find a genetically unique population without knowing its history, you can’t reconstruct its history from the genetics. It might me a remnant population, and immigrant population, or the remnants of an island effect. It didn’t say anything about the Basques, whose history is known 1500 years back or more.
I think that a lot of C-S’s genetic work is being redone, which is normal for pioneering science.
Also, it’s possible that Icelandic and Sardinian islanders might be relatively unique within the European group just by an unusual mix of normal European genes, whereas the Basques and the Lapps might be unique in having genes not found in any other European group.
Re: Basque and Georgian. It would be interesting to know if any competent linguists actually speak both languages.
Learning either is said to drive a person mad, so I’m afraid learning both would be inadvisable.
Re knowing both Basque and Georgian, you could adapt the old anecdote about Bismarck being asked about the origin of (I think?) the Schleswig-Holstein controversy and him responding “There were only three men in Europe who ever understood that: one has since died, another has gone mad, and I have forgotten.”
History of Schleswig, Holstein, and Schleswig-Holstein.
Not Bismarck, you say? Well, slap my ass and call me Lord Palmerston.
Marc: actually, knowledge of Modern spoken Georgian would probably be of limited help in proving it to be related to Basque. First, Georgian has a written history dating back fifteen centuries, and the earliest stages of the language are what you’d want to look at. Second, Georgian is a member of a broader language family, Kartvelian, on which solid scholarship has been produced. *IF* Basque and Georgian are indeed related, I’d expect such a connection to be found by comparing Basque to Proto-Kartvelian, not to Modern Georgian.
I came across Schleswig-Holstein-Gottorp when reading about Karl XII of Sweden. The Swedish royal house married Danes and Holstein-Gottorpians in alternate generations, so if the Queen was a Dane the Queen Mother would probably be a Holstein-Gottorpian. The Swedes married Danes on the gamble that the accidents of death and birth would make a Swede the ruler of Denmark. They married Holstein-Gottorpians to use as leverage against the Swedes if they ever needed to.
At one time Karl was betrothed to Hedwig Sophia of Denmark, while his sister Sophia Hedwig was betrothed to the Duke of Holstein-Gottorp…. or something like that.
Etienne:
“I have often wondered whether Ruhlen’s belief (that surface similarities between languages/language families separated over millenia will be visible through surface inspection) isn’t due to his unconsciously taking Latin to Romanian as “typical”.”
I’ve seen a TV program in which Ruhlen quotes Romanian nepot as an example of a word unchanged since Indo-European. I didn’t know he was a Romanianist, but that’s circumstantial evidence for what you’re saying.
Ruhlen quotes Romanian nepot as an example of a word unchanged since Indo-European
Ruhlen and a few other people seem to be under the impression that somehow some meanings cause the words which contain them to be impervious to phonetic change. But the reason that the Romanian word nepot (‘nephew’) is close to its PIE ancestor is that in that language, the sounds which compose it have not changed much in the words where they occurred in the same position as in this one: this lack of change is not just found in few random words. Initial *n is preserved intact in all IE languages (witness the many words for “no” or “night”, all starting with n), middle *p remained unchanged in Latin and did not change in Romanian or Italian either (even though it changed to b or v in most of the other descendants of Latin), and final *t (which had a vowel after it in most of the Latin forms of the word) is also unchanged in Romanian and Italian (where the word is nipote), even though it also changed or disappeared in most of the other Romance languages. The vowel o has also remained the same in the same two languages, and the Italian change from e to i has happened in hundreds of words of that language.
These phonetic changes or lack thereof have nothing to do with the fact that the word means “nephew” rather than something completely different, and everything to do with the fact that phonetic change has affected some sounds in some positions in some languages and not in others. This particular word happens to contain sounds which have remained the same across the board in the evolution from PIE to Latin to Italian and Romanian, while the same sounds changed across the board in many of the other languages descended from Latin.
The fact that phonetic change occurs across the board in related languages is one of the foundational observations of comparative and historical linguistics. Nepot is not an exceptional case at all, it fits in with all the other words which have evolved at the same time in the same languages.
“Initial *n is preserved intact in all IE languages (witness the many words for “no” or “night”, all starting with n), middle *p remained unchanged in Latin and did not change in Romanian or Italian either (even though it changed to b or v in most of the other descendants of Latin), and final *t (which had a vowel after it in most of the Latin forms of the word) is also unchanged in Romanian and Italian (where the word is nipote), even though it also changed or disappeared in most of the other Romance languages. ….”
Others say that nephews are especially important in Romanian culture.
From the classical Chinese of about 800 BC until contemporary Mandarin, the syllable “fang” is the one least likely to change (sometimes it becomes “bang”. The actual pronunciation probably is slightly different, but all of the “fang” words from then are pronounced the same now and transcribed “fang”. (This conclusion is based on flipping through several historical dictionaries and probably isn’t 100% correct).
By contrast, contemporary Mandarin words pronounced “shi” or “si” had dozens of very dissimiliar pronunciations 2800 years ago. It’s like half the Chinese language fell into a black hole of total entropy.
contemporary Mandarin words pronounced “shi” or “si” had dozens of very dissimiliar pronunciations 2800 years ago
From what is known from the evidence of long-written languages, there are indeed many ways that other consonants or sound combinations can end up as “sh” or “s”. Comparing words in different languages should never start with those sounds, since there are so many sounds or combinations that they can have evolved from.
I would thank m-l again for explaining things better than I could, but I don’t want her to get a swelled head.
I understand that comparisons would be based on (I assume) Mitxelena’s reconstruction of proto-Basque (at least as far back as that goes) and proto-Kartvelian, but still, it would be nice to know both. I’m pretty far along in Basque (can read Berria (berria.info) articles with a dictionary, for example) and my sanity – knock wood – is still intact. 😉 It helps that the grammar (especially word order) is similar to Japanese, which I already spoke when I started Basque.
But I’m not going to learn Georgian any time soon. Plus, I’m not a linguist, just a lowly translator.
So is the answer that no one knows if there are any linguists who speak both languages? I’d expect it to be a Georgian who learned Basque, for some reason.
To clarify: members belonging to the “fang” group in 800 BC are still in the “fang” group now, but “fang” might be pronounced slightly differently now than it was then.
Others say that nephews are especially important in Romanian culture.
The sister’s son was especially important in Viking culture.
Marc, they’re all locked up.
Many had to die before the Matterhorn could be successfully climbed, and many have sacrificed their sanity in the attempt to learn both Georgian and Basque. The last aspirant was completely sane right up until the night when he started in on the 458,683th Basque noun inflection. So near, and yet so far.
But we are confident that you can be the one to finally triumph.
Basques are not. They are a continental population. They have been invaded, (almost as often as British or any other European, for instance),
Since when? Even the Romans made treaty with them and in return the Roman armies were guided over their impenetrable mountains.
I think it’s the verbs that drive you crazy. The nouns are Prozac compared to the verbs.
Lasciate ogne speranza, voi ch’intrate
That’s not even really the hard part, because you can just memorize the different pieces and put them together like Legos. It’s the allocutive forms (“You have me as an American” = “I am American”) that are hard, although they too have something vaguely analogous in Japanese.
Apparently the guy I referenced was merely approaching a false summit. Had he reached the top, as he almost did, he would merely have seen the true summit towering above him, dotted with the bodies of brave but foolhardy linguists.
I understand, LH, my head is quite big enough as it is.
I don’t know Basque, but I have a copy of Hamlet translated into Basque. I looked for “to be or not to be” and it was not hard to find (my copy of the book is somewhere where I can’t find it, so I can’t report what the translation is). I bought this book in the famous Shakespeare & Co. bookstore in Paris – as a linguist, I could not resist. I regret to say I don’t know Georgian either.
Learning either [Basque or Georgian] is said to drive a person mad, so I’m afraid learning both would be inadvisable.
I have feeling Nikolai Marr might have known both and he was a real oddball.
John E.: Middle Chinese had no /f/; even now, Min languages don’t.
I doubt very much that m-l could ever get spoiled; more likely that it would be too hard to come up with something adequate and even if you could, too embarrassing to say it. Not to mention her tireless efforts on the Goat Blog and the Martian Blog too.
“Izan edo ez izan” is probably the translation of “to be or not to be.”
You can find two versions of Hamlet in Basque here.
The first one is in pre-Batua orthography and therefore in some dialect (not sure which). The other looks like it’s in Batua.
Ah, I stand corrected. It’s “izan ala ez izan.” Here is Act Three. The soliloquy starts with “Izan ala ez izan.” Note that “Ophelia” becomes “Opele.”
For those who understand the impulse to buy a Basque edition of Hamlet w/o knowing or even necessarily planning to learn Basque, I will mention that the Strand in NYC currently has some copies of the first (and I think only) Hani-English / English-Hani dictionary ever published, at an incredibly deep (90%+) discount from the original $200+ list price. The website says they have 7 copies left in stock, which may not have been updated to net out the one I bought yesterday. I mean, it caught my eye while I was looking for something else entirely, but how could I resist?
I used to look with mild longing at that dictionary when I was still living in NYC and frequenting the Strand, but in those days they were selling it for $30 or $40, as I recall, and I wasn’t that interested in it. Even if it’s down to $20 or less now, I seem to be much less interested in oddball languages I’ll never study than I used to be. Sigh.
If they’ve still got 7 copies maybe I’ll wait ten years. By then, it’ll be down to $10.
So is the answer that no one knows if there are any linguists who speak both languages?
I know one. Thomas Widmann who used to be in sci.lang and dk.kultur.sprog some years ago.
Oh, sorry. I screwed up the HTML again.
You forgot the main point: is he OK?
On my last trip to Half Price Books, I somehow ended up with a Tibetan phrasebook. You never know when you’re going to want to know how to say, “I want to rent a yak.” Anyhow the script is so pretty.
You can always use the excuse “It was on sale.”
I was going to buy the Rasulid Hexaglot even though I only knew a little Mongolian out of the six languages in it. Armenian Arabic Greek Turkish Persian and Mongolian, from around 1300.
Actually the Amazon review makes it seem to be interesting for those without the languages.
I fixed Trond’s HTML, but I’m afraid there’s nothing I can do for poor Thomas Widmann.
I have a number of phrasebooks, especially from Lonely Planet. Most of them are for single languages, but the Central Asian one has about ten or twelve languages (more or less fully documented depending on their regional importance). I am not planning a trip to Central Asia, but with most of the languages being Indo-iranian or Turkic, I found the phrasebook a great source for teaching comparative linguistics.
I’ve read that the definition of different national languages in Central Asia was the Russian way of warding off pan-Turkish movements. (My source was a pan-Turkish spokesman, I think). So such a book would give you a dialect series Uzbek, Kirgiz, Kazakh, Turkmenian, Uighur. But as I was told, of these only Uighur is very distinct from the others.
Indeed, most of them are very similar, but the book, being for travellers, is organized by nation rather than simply by language.
I have that Central Asian phrasebook too; it’s quite well done.
If you are to buy only one Asian phrasebook, this should be it. It’s the book that all other Asian phrasebooks revolves around.
Yes, the Tibetan one is Lonely Planet. I was compelled to get one for Farsi too. So now I can say and write
The system of capitalization escapes me, especially since the Arabic alphabet doesn’t have capitals.
I notice they don’t use teh Rubiyat’s “A Jug of Wine, A Loaf of Bread and Thou” — probably not the best phrase to choose if you find yourself in Iran these days.
I was a little worried that a phrasebook might be too rinky-dink a thing to have on my bookshelf, but since all the most famous linguists have them, I feel much better.
Quite a hefty price for Rasulid Hexaglot, and I see my local library doesn’t have it.
The Rasulid family, maybe from rasul, رسول “prophet”, as in The Prophet?
I believe that “Rasulid” refers to a ruling dynasty, probably of Yemen. The answer is out there but I’m too lazy.
Nijma: I was a little worried that a phrasebook might be too rinky-dink a thing to have on my bookshelf,
There is (decorating) safety in numbers: one lonely little phrasebook next to big hardcover dictionaries looks small, but when you have a bunch of them next to each other they look fine. It is the same with CDs. And your visitors will be intrigued by those conversation pieces.
For a linguist, a phrasebook gives some idea of the language at a reasonable price, and if you are interested in further study you can invest in a grammar and dictionary later. As a bonus, a good phrasebook will give you some useful and current things to say in the relevant language, unlike most grammars.
The system of capitalization escapes me, especially since the Arabic alphabet doesn’t have capitals.
You’re aware that that’s Edward Fitzgerald’s translation, right? It has nothing to do with Arabic, and they used capitals differently in 1859 (though he used them a lot even for the Victorian era).
I believe that “Rasulid” refers to a ruling dynasty, probably of Yemen.
Give that man five silver dirhams. The Rasulids ruled southern Yemen from 1228 to 1454, and Bosworth (in his magisterial The New Islamic Dynasties, which I decided to get for myself as a splurge in December 1998 and called bookstore after bookstore till I found one that had it—I may have gotten the last copy in NYC) writes that it is “probable that they came from the Menjik clan of the Oghuz Turks, who had participated in the Turkish invasions of the Middle East under the Saljuqs, and that the original Rasūl [al-Malik al-Mansur Umar I b. Ali b. Rasul, Nur al-Din al-Ghassani] had been employed as an envoy (rasūl) by the ‘Abbāsid caliphs.”
And yes, that’s the same as in The Prophet; a rasūl is someone who is sent, whether as messenger, emissary, or envoy.
As for ‘poor Thomas Widmann’ – Imagine being married to someone who speaks Basque and Georgian!
Signed Thomas’s wife!
Oh, hi! I was going to assure that your husband seems fine – although disturbingly diverted also outside linguistics – but now I’ll leave that to someone in a better position to judge.
(I also meant to link to his blog, or maybe the linguistics part of it, in stead of your professional website, but couldn’t find a bold enough statement of mastered languages anywhere. He really ought to do something about that.)
My god, she is too. Small world. He also speaks dansk. Or English. Or both, probably.
Mrs Phyllis, Are there any advantages to being married to someone who speaks Basque and Georgian? (Apart from the obvious one.)
My god, she is too. Small world. He also speaks dansk. Or English. Or both, probably.
Mrs Phyllis, Are there any advantages to being married to someone who speaks Basque and Georgian? (Apart from the obvious one.)
If this man is in IT and computers, maybe he will produce the universal Danish-Georgian-Basque culturally-neutral Chomskyian computer language! And then our computers won’t crash all the time and get viruses, and spam screens will finally work right.
Isn’t the internet a wonderful place? Whole discussions on blogs unknown to you can be about your (in)sanity! 🙂
Anyway, to get this back on a linguistic track:
I learned Basque first. I was 20 or 21, and I did a summer course at Euskal Herriko Unibertsitatea in Donostia. I actually found Basque relatively easy, although I started mixing it up with Japanese at times. That said, I must admit that I never got close to learning all the inflections, not even on really dull winter days back in Denmark. However, I definitely would recommend learning Basque to everybody, it does wonders to your mind. Much better than Latin.
I didn’t approach Georgian till I was 24. I wanted to try and learn a hard language, after Japanese, Sanskrit, Basque and Russian hadn’t proven sufficiently maddening.
Fortunately, there was an exchange programme between my university and Tbilisis Saxelmc’ipo Universit’e’t’i, so I went to Tbilisi for a year.
However, that was too much for my poor brain. The first three months or so were OK, but one day, learning yet another irregular inflection, I suddenly felt my brain overflowing, and I never got as fluent as I wanted to.
Sadly, my brain never really recovered, and my ability to store new languages in long-term memory has never returned to pre-Georgian levels.
Surely that’s not madness, though.
Thomas and Phyllis, I’m delighted you dropped by and are such good sports about my cheeky speculation! Believe me, I’m in awe of anyone who could learn both those languages, even if not completely fluently. I’ve made two attacks on Georgian and at one point was starting to be able to read it without too much pain, but I doubt I’d be able to speak it without actually going to Georgia, which at my advanced age isn’t going to happen. As for Basque, I’ve got a dictionary and a (French) grammar, but I’m probably never even going to get to beginner level.
I’d respond to Trey at http://languagehat.com/archives/003074.php but comments are closed there. The places where language families cross gaps in the genetic tree are exactly the exceptions where language shift and population mixture are thought to have happened: Cushitic and then Semitic speakers colonizing where Khoisan was formerly dominant, Uralic crossing the Caucasoid-Mongoloid interface in North Asia, Austronesian movement across Papuan territory and into the Pacific. I know some here have low opinions of Cavalli-Sforza and Ruhlen, but it is silly to accuse them of naive ignorance of what they illustrated in the diagram.
It’s a cautionary tale I’ll never need to worry about. But I wonder if there are any native Basque speakers who have learnt Georgian successfully, or vice versa.
I never met anybody in the Basque country who had any significant knowledge of Georgian, whereas the Georgians seemed to be quite well informed about Basque. I never met anybody who had actually learnt it, though, but I would be surprised if nobody had attempted it.
(This might have something to do with the Georgians’ deep desire to be considered European, whereas there is no such prestige in being Caucasian.)
For what it’s worth, I’m not personally convinced Basque is related to South Caucasian. There are a couple of similarities, such as the affix for 1pl (Georgian gv-, Basque g-, -gu), but I think people get carried away because of ergativity and such stuff.
My personal guess (but it’s just conjecture) is that South Caucasian is related to Sino-Tibetan, but I don’t know enough about the latter to say anything intelligent about it. I’m just fascinated by the Georgian numerals: ert-, or-, sam-.
Well, there you go. This Thomas character sounds mad as a hatter! 😉
Serendipitously, one of the regular members of my Russian conversation group is an ex-pat Georgian/Russian (father/mother) (ex-patted by his parents at around 5, so he didn’t learn Georgian, just Russian) and it turns out he has recently started studying Georgian, so I think I’m going to try my hand at Georgian now.
We’ll see if I go crazy. Maybe I should blog it so people can warn me if I start sounding off-kilter. 😉
I’m just fascinated by the Georgian numerals: ert-, or-, sam-.
Oh, is it something like Danish? I just LOVE your Danish vigesimal counting and that you don’t have to be a math major to use a word like syvoghalvtredsindstyvende! Fifty-seventh just seems so… prosaic, in comparison.
I think I’m going to try my hand at Georgian now.
I highly recommend it. All joking aside, it’s a fascinating language, and there are a number of good textbooks available. (Dictionaries are a problem, though, unless you read German.)
I picture the ancient Danes living in a fluid world with few discrete units, a world where enumerating or counting things was seldom necessary. They lived in bogs, didn’t they?
They died in bogs (some of them).
It’s really amazing the Danes have lasted this long, considering they can’t understand each other.
Does the name Bogdanovich originally refer to bogged Danes?
Vich vones?
Vich bogs? Peat bogs.
I can get by in German. What dictionary do you recommend?
(I mean, listen, I should start out all my comments here with “My name is Marc, and I’m a dictionary-addict.” It’s almost like my love of learning languages is simply an excuse to buy dictionaries…)
But wait! Только что вспомнил, что знаю русский язык гораздо лучше чем немецкий. 🙂 Самые большие словори грузинского наверника русские, нет?
Found this: pronunciation examples for all Georgian letters, for the curious
Нет. The classic dictionary is Kita Tschenkéli [Chkhenkeli], Georgisch-Deutsches Wörterbuch (Zurich: Amirani-Verlag, 1965-1974, 3 vol., 2,470 pp.), one of the great works of obsessive lexicography. It’s quirky (all verb forms are entered under the root, so that the entry for svla ‘go’ is 22 pages long), but once you get used to it it’s indispensable. Needless to say, it’s expensive and hard to find; I xeroxed mine from a library copy. If your Georgian gets up to speed, you can make use of Chikobava’s Kartuli enis ganmartebiti leksikoni (Tbilisi, 1950-1964, 8 vol., 12,302 pp.), but that’s probably even harder to find.
Apparently Tschenkéli has been reissued. Here‘s the publisher’s page; isn’t that a pretty book? And if I’m reading aright, it’s only 145 euros, a comparative steal.
Ah, kamelåså!
Found this: pronunciation examples for all Georgian letters, for the curious
I was curious, but the letters all look the same to me: just little squares.
Marie-Lucie — I’ve got Windows Vista and have my computer set up for the complex scripts (since I need Japanese for work), so I don’t think you have to download or install Georgian fonts.
m-l,
Here are some screenshots of the Georgian Transkription page : |1| |2| |3|
They’re probably too wide to view in blog format — they loose too much resolution when I make them that small — but it should be easy to save them to disk.
(I can see them fine in both Vista and XP, but m-l has a Mac.)
There is (decorating) safety in numbers: one lonely little phrasebook next to big hardcover dictionaries looks small, but when you have a bunch of them next to each other they look fine.
I never thought of the decorating angle. I thought I would be done with my bookshelves when I had everything organized by subject and alphabet, but now I see I will have to buy even more phrasebooks. It’s amazing how people who would probably never think of buying an alcoholic a drink will casually suggest more books to other LH readers, and even click on Hat’s Amazon links.
Nijma, you don’t have to get more phrasebooks. You can always hide the one you have if it does not go with your decor. For instance, it would not take much space in your kitchen cupboard, mixed in with cookbooks of various sizes.
Those who would buy drinks for an alcoholic are probably alcoholics themselves. Most of us here are addicted to books, especially language books. Books are not bad for your health, both mental and physical. If you choose right, they are “the gift that keeps on giving”.
It’s not exactly a twelve-step program for bookaholics in here, is it. I can see I’m in the right place. I just hope my guests won’t wonder if the books are going to fall through onto the floor below. But cookbooks in the kitchen? Hmmm….
I nearly ordered the Tschenkéli dictionary, but I stopped myself.
Someone please tell me I made the right decision. :'(
I don’t believe Mac fonts (specifically Lucida Grand) support Georgian. This page lists input methods and fonts that you can download. I haven’t tried them.
The Quivira seems to work for me.
For various other fonts that are not included in the standard Mac fonts, Alan Woods’ Unicode Resources are useful.
Not particularly relevant, but I thought I’d share with everything the shock of recognition I just had as I realized that the layout of the Georgian keyboard is based on the Russian keyboard layout.
This is going to save me a lot of grief…
‘Everything’? … -> ‘everyone’
For those resisting the 12-step program and in need of the equivalent of a half-pint bottle of something with a screw-off cap, I should follow-up on my post of several days back with the tidbit that the Strand is selling that Hani-English / English-Hani dictionary for only $4.95 – 97% off original. (The copy I got was marked $14.95, but they seem to have taken off an additional $10 at the register.) Maybe you still think you can afford not to buy it, but as the saying goes, Alwuv qiq siq niq taq maq diq. It perhaps can’t fully compete with the sophisticated bouquet of Georgian being swirled around in the snifter, but no prior knowledge of German is required.
The Strand? As in, England?
Is it written in Chinese characters?
The Strand.
which at my advanced age isn’t going to happen
Since I’m feeling nitpicky over decimal places of kilogram conversions today, I have to say something about Hat’s old codger meme. For a long time I thought he really was old, until a few days ago when he mentioned which years he was an undergraduate student and realized that if I had been able to continue my studies on schedule, they would have overlapped my own undergraduate years. Ask AJP if that is old.
Also a belated thanks for the upthread discussion of the Rasulids; I have put the Bosworth on my wish list. I had assumed some sort of claim about descent from the Prophet, many leaders have claimed such ancestry, from Jordan’s Hashemite kings to Saddam Hussein, but it never occurred to me a rasool might be an emissary from a human source as well.
The horror of bookaholism is that it is an addiction that doesn’t stop: if your financial situation/significant other prevents you from buying more books, well, the addiction isn’t over: you simply read/re-read what you already have, without spending a penny (and organize your bookshelves better in the process)…and as you do it you want *more*…come on, for the price of just a few cups of coffee a mere phrasebook can’t do any harm…and for the price of *just* a dozen phrasebooks a grammar will do…I mean, think of all the coffee you won’t be drinking in the future, buying these books makes you healthy…and it doesn’t cost anything to enter a bookstore, just browsing, just looking, just ssmelling that dusty old paper, and if a title should call out to you, well, that really can’t be helped…
Changing topics: I found Thomas Widmann’s comment on Georgians being more interested in Basque than vice-versa very interesting: are there any Basque reference/teaching grammars in Russian? If not, that might explain why so few Georgians have learned it: if fluent French or Spanish is required, that does make for a reduced pool of Basque learners in Georgia…
I met (online) one of the few Georgian translators (through a mailing list for Russian translators) once, and he told me something similar — the Georgians are very interested in any possible Basque connection.
you are really good at this topic. I think you are great.
Aww, iPhone mac is my secret admirer! Too bad I hate spam, but it’s nice to be loved sometimes. 😉