Economic Success Drives Language Extinction.

Tatsuya Amano, Brody Sandel, Heidi Eager, et al. have a paper in Proc. R. Soc. B 22 of obvious LH interest; the abstract begins:

Many of the world’s languages face serious risk of extinction. Efforts to prevent this cultural loss are severely constrained by a poor understanding of the geographical patterns and drivers of extinction risk. We quantify the global distribution of language extinction risk—represented by small range and speaker population sizes and rapid declines in the number of speakers—and identify the underlying environmental and socioeconomic drivers.

The whole paper is online (way to go, Royal Society!); there’s also a summary at ScienceDaily that boils it down conveniently: “Thriving economies are the biggest factor in the disappearance of minority languages and conservation should focus on the most developed countries where languages are vanishing the fastest.” Thanks for the links, Paul and Martin!

Comments

  1. David Eddyshaw says

    It’s not obvious that higher GDP is a *cause* of language loss, rather than simply being highly correlated with it.

    The results seem to be highly skewed by cases like North America and Australia where, apart from outright deliberate destruction of the speakers themselves and their cultures there have also been, until only a generation or so ago, sustained deliberate efforts to destroy their languages via the educational system. This really only correlates with GDP by historical accident, or at least very indirectly.

    It does, I suppose reflect the ideology of the modern nation state, based on a fictitious concept of “nationality” intimately linked with language, with which it is declared to coincide; the same as has led in France (for example) to many decades of deliberate official suppression of Breton, Occitan, Basque …

    In purely economic terms, this ideology has paid off (let’s set aside the millions dead at its altar in European wars.) So that, at least, doubtless goes with higher GDP.

  2. My impression is that the decline of Mongolian in Inner Mongolia has accelerated in the last 20 years or so as a result of economic development. The cities are overwhelmingly Chinese, and if young Mongols want to do well for themselves in a non-herding lifestyle they have to learn Chinese. Their parents are willing to facilitate their switch over to Chinese in order to get on in the world. Previously to that Mongolian was holding up reasonably well. Of course, attempts by the Chinese government to phase out non-Chinese-language education have also played a part.

  3. Just like Bathrobe did not lost his English despite learning Mongolian and Chinese, Inner Mongolians learning Chinese will not automatically forget their native language.

    However, if a particular Inner Mongolian town remains overwhelmingly Chinese for a generation or two or three, in that time Mongolian minority living in that town will tend to lose its language, that’s true.

  4. “Inner Mongolians learning Chinese will not automatically forget their native language”

    Theoretically, maybe. But the language tends to become ‘impoverished’ as Chinese takes over more and more areas of life (decline in vocabulary and expressive ability). The consolidation of Mongolian schools in Inner Mongolia also leads to more and more people being unable to send their children to Mongolian schools because of distances involved. The fact is that Mongolian is on the defensive in Inner Mongolia. More and more people are losing their ‘native language’ and being brought up in Chinese.

  5. There are also no jobs for graduates of Mongolian schools and people majoring in Mongolian. The incentive to maintain Mongolian alongside Chinese is steadily declining. The result is language loss.

  6. You forget about differences in birth rates between Mongolian and Chinese population of Inner Mongolia. Ethnic Mongolians number about 21 percent of population of Inner Mongolia, but account for something like 40-45 percent of all births.

    This is for entire province and as you are aware, Inner Mongolia can be very diverse in Han/Mongolian ratio.

    So, in a town like Huhhot where 10% of population are ethnic Mongols, about 20-25% of all schoolchildren would be Mongols, But in, say, Tongliao city where Mongols were about 45% in 2000 and now probably number over 50%, the overwhelming majority of schoolchildren would be ethnic Mongols.

    In a class like the latter where native Chinese speaking children are a minority, it would be very difficult to lose one’s Mongolian, even if all classes are taught in Chinese.

  7. “In a class like the latter where native Chinese speaking children are a minority, it would be very difficult to lose one’s Mongolian, even if all classes are taught in Chinese.”

    Unless, of course, Chinese became the prestige language and people felt stigmatized by Mongolian, etc,

  8. Just like Bathrobe did not lost his English despite learning Mongolian and Chinese, Inner Mongolians learning Chinese will not automatically forget their native language.

    No, Bathrobe did not lose his English, because (I presume) he acquired those languages deliberately as an adult, they were not imposed on him by a government, school or other circumstances, and his parents and other adults did not feel compelled not to speak English to him as a small child.

    In such situations as Mongolian/Chineese in Inner Mongolia, it is not so much that individual people “forget their language” but that bilingual parents (often those schooled in the dominant language) choose to raise their children in the that language so that they will do better at school and later. If the parents still converse in their language with older people such as their own parents, and children hearing such conversations acquire a passive knowledge, parents might not realize that their children do not in fact speak the language and will not be able to function in it as adults. Restriction of vocabulary, etc does not seem to me such a huge problem compared to the lack of basic facility due to not speaking the language TO children.

    I have experienced this situation in my own family, since my maternal grandparents were Occitan speakers who had learned French in school, moved to Paris for work, and did not speak Occitan to their children (they used it as a secret language until the children told them they understood everything). On vacation in their native village I tried to learn it from some of their relatives, but that was disapproved: ‘People will think you are a peasant’. When I later found myself living and working in a First Nations community where many people my age (30’s at the time) were bilingual, often spoke their language to the remaining elders but hardly ever to their children, and only occasionally between themselves, I was puzzled at first by their attitudes and behaviour in relation to their language until I realized the similarity of their situation with my grandparents’ and my own.

  9. David Eddyshaw says

    What about Amazonia? OK, Brazil as a whole is not a byword for desperate poverty, but the groups speaking the indigenous languages don’t participate in the wealth. The important factor must surely be the powerlessness and poverty of the language community rather than the country as a whole. Or PNG, where children are more and more growing up speaking Tok Pisin instead of their many traditional languages? Or the Solomon Islands?

    The idea that efforts to preserve endangered languages should be concentrated on richer countries seems to result from underestimating the extent of the great economic inequalities characteristic of poor countries, and from inadequate information about the real linguistic complexity to be found in areas where linguists don’t go so much.

    There seems to be a certain self-serving quality to the conclusion too, though undoubtedly it has been arrived at in good faith. But the logic of their argument would also seem to lead to the conclusion that efforts to support minority languages in rich countries are likely to be unavailing anyway; more would probably be achieved by supporting languages not already mortally wounded by the effects of economic progress.

  10. Following up on Marie-Lucie’s comment, I can easily see a causal connection between economic growth and language loss, because economic growth happens through massive rural-urban migration. I apologize for not digging up a citation, but I’m pretty sure this isn’t an idea of my own.

    In a pre-industrial society, people are overwhelmingly peasants with little geographic mobility. Thus, there are no significant downsides if they speak languages that vary markedly across space. Industrialization moves people from distant places into large cities, where they have a pressing need to understand each other in a common language.

    I suppose this intra-national migration doesn’t differ all that much from the international variety. The first-generation immigrants speak mostly the language of their place of origin and probably acquire an imperfect working knowledge of dominant language of their destination. Each successive generation has practical and social incentives to learn ever less of the ancestral language.

    Needless to say, a national authority actively promoting the dominant language and suppressing the others is going to make the process faster and more complete. But I doubt it’s a necessary condition. Also, while I do believe that nationalist ideology plays a big role in such language policies, I also believe there are efficiency gains to be reaped by having everyone speak the same language.

    In fact, I wonder if the ideological rationale of creating a common national identity is entirely distinct from the military efficiency rationale of being able to recruit armies by nation-wide conscription and have them function in a single army-wide language.

  11. I wonder if the ideological rationale of creating a common national identity is entirely distinct from the military efficiency rationale of being able to recruit armies by nation-wide conscription…
    It was also linked to the rise of democracy, but I heard this theory only in the usual handwaving manner of political philosophers.

  12. The common language of the Ottoman military was (doubtless somewhat broken) Turkish, but Turkish did not become the common language of the empire. Even in the Roman Empire, Greek and the Semitic languages very much held their own against Latin.

  13. JC: “the common language of the empire” (or not)

    I wonder if this has to do with literacy, along with religion. Greek and Semitic languages, some of which had a long written tradition, persisted in the Eastern Roman Empire, but Celtic and most other languages, which did not, largely disappeared from the Western Empire, since most of the inhabitants (including Germanic conquerors) switched to Latin over a period of a few centuries: administration and religion, not to mention higher learning, went hand in hand with the Latin language both spoken and written. In the Ottoman empire, the Turkish language could not match the religious and literary prestige of Arabic among the populations outside of Turkey, even those which did not actually speak Arabic. Instead, Turkey eventually adopted the Arabic alphabet already used in Persian, the prestigious language which was that of the Ottoman government for a time.

  14. The common language of the Ottoman military was (doubtless somewhat broken) Turkish

    I suspect it wasn’t, at least until very late in the Empire. At the higher levels, sure, but units were raised locally, and fought as local units. So a cohort of Syrian peasants would all come from the same village, or neighboring villages and would communicate with each other in Syrian Arabic, and their leaders – local nobility – would communicate with their own troops in Arabic (or Albanian, or Greek or whatever) and maybe Turkish with the other nobility. I don’t think the Ottoman military was so sophisticated that, say, Bosnian recruits would be plugged en masse into a Thracian unit to replace losses, resulting in communication issues. The Habsburgs were similar – for most of the history of the Empire there was usually no pressing need for ordinary infantry soldiers or NCOs to have to communicate with soldiers from other units, as long as the officers could communicate. Units were raised locally, and losses generally replaced with people from the same region, or not replaced. In fact, I suspect that was seen as a feature, not a bug, because it made it easier when necessary to pacify rebellious regions with troops that spoke a different language. In WWI that became a problem because the casualties were amazingly high during the beginning of the war, in particular among the junior officer class, and those were the people who generally knew both the language of command (German or Hungarian) and the unit’s local language. The result was chaos as officers or soldiers were plugged into units where they couldn’t speak to anyone.

  15. On vacation in their native village I tried to learn it from some of their relatives, but that was disapproved: ‘People will think you are a peasant’.

    There is something profound in this observation that rural -< urban migration may have a disproportionate effect on language survival, while entrenched minority languages in the cities may fare somewhat better over time.

    One anecdotal case of a young friend from rural Eastern Latvia, who grew up speaking Latgalian as a child, learned Latvian in school, then moved to Riga for economic and education opportunities and picked excellent Russian there. Thus in the trio of nation’s major languages – official Latvian, rural and neglected Latgalian, and systematically eradicated urban Russian – the language of the countryside is endangered, while the language of the cities finds new speakers.

  16. J. W. Brewer says

    The good news for those interested in preserving linguistic diversity is that due to trial and error over the course of the last century we know a lot more than we used to about which sorts of economic policies are sufficiently disastrous if adopted that they can safely be relied upon to keep poor nations poor. The even more ambitious policies are those that affirmatively increase linguistic diversity, for example those that were so successful at decreasing prosperity, education, and long-distance trade and communication among the western bits of the former Empire as to transform regional dialect variation in Vulgar Latin into a dazzling array of mutually incomprehensible Romance languages.

  17. J. W. Brewer says

    To Dmitry’s rural/urban distinction, one counterexample that comes to mind is that spoken Dutch survived with some degree of health much longer in the rural parts of the former New Netherlands (into the 19th century, although Dutch had ceased to be the rulers’ language in 1664) than it did in New York City – I think the ability of minority languages to survive in an urban environment where people are going to be interacting on a daily basis with speakers of the majority language as compared to in a rural village where the “minority language” is the dominant local tongue but is low-prestige and not spoken in the far-off capital just involve different factors. In the urban environment, one key factor is going to be how permeable versus impermeable the social barriers between the speakers of the different languages are, as seen in e.g. exogamy taboos. In 18th century New York, marriage between the children of Dutch- and English-speaking families was common, so the minority language got assimilated into eventual oblivion — in 18th century Wilno/Vilna, marriage between the children of Yiddish- and Polish-speaking families was not so common, so both languages survived (even if most of the Yiddish speakers had pretty good L2 fluency in Polish).

  18. we know a lot more than we used to about which sorts of economic policies are sufficiently disastrous if adopted

    So we do. But we also have international institutions to see to it that those policies are adopted and adhered to right up to the point of absolute ruin and beyond. (Sorry, I’m feeling peevish today.)

  19. JWB, true, after thinking some more about my strange anecdote, I suspect that Riga may be a very special case in urban multilingualism. There may be strong specific causes of improved survival of Russian language there. This a city which reportedly lost 1/3rd of its population in a generation, frequently to migration to the more prosperous European countries. But Russian speakers may be less affected by the emigration prospects because many of them aren’t Latvian citizens? In a sense, economic opportunity may be causing disproportionately stronger exodus of the national language speakers, a situation probably not unparalleled in the world but still very unusual.

  20. I see a potential moral dilemma. This might be somewhat like the issue of building dams in undeveloped countries in areas with all sorts of endangered species. We, in developed countries, decry the loss of habitat (and rightfully so). However, the local people stand to benefit greatly from the development.

    So it might be with some endangered languages. The speakers can potentially improve their lot in life by switching to a dominate language. However, humankind loses the linguistic diversity.

    I recognize that it isn’t a simple black and white, preserve or starve issue. But, the economic allure, as discussed, can contribute to language demise.

  21. J. W. Brewer says

    I think viewing language “extinction” as akin to biological extinction is a category error (and/or confuses a metaphor with reality), at least when the language goes away because the biological descendants of its former speakers choose to speak some other language(s) as opposed to its former speakers having left no living descendants. And even in biological terms it’s not clear to me that there’s some objectively optimal number of total species or that more is always better – is it really “worse” for there to be a single given species that is sufficiently flexible to thrive over a fairly wide geographical range than for there to instead be 10 or 20 or 100 separate-but-related species each with a teensy range and each at greater risk of extinction if something goes awry with the local habitat?

  22. However, the local people stand to benefit greatly from the development.

    Temporarily, just as you temporarily benefit from dealing crack.

    the biological descendants of its former speakers choose to speak some other language(s)

    You seem to be assuming that this is a free choice that they just happen to make, divorced from all other circumstances. Sort of like choosing to deal crack.

  23. J. W. Brewer says

    Why would I believe that? All human beings make choices among what they perceive as their available alternatives, which are often unsatisfactory and always constrained by circumstances, with those circumstances themselves subject to unpredictable change. (E.g. the geopolitical situation of Latvian and the factors that make particular individuals more or less likely to speak it have changed rather dramatically since my childhood, and the linguistic consequences of the way in which the removal of the Soviet yoke has been followed by freer opportunities for emigration to the EU might not have been foreseen by anyone back then – although fwiw however many locally-born folks have left, one of my US-born-and-raised high school classmates now lives in RIga, even though we did not have a “Most Likely to End Up Moving to a Formerly Captive Nation” category in the yearbook.) The endangered-species parallel implies that we should take speakers of endangered languages into captivity and put them behind glass in a zoo or museum, to insulate them from the risks of contact with the wider world and to prevent them from breeding with unapproved partners. And/or at a minimum turn their present geographical range into a “national park” which still has rules and regulations to prevent unauthorized interaction between the indigenes and outsiders, which raises all sorts of moral quandaries. Part of the whole issue is that speakers of all human languages are in fact members of the *same* species and fully interfertile, which is why the strength or weakness of taboos against exogamy (and/or barriers to geographical migration or social mobility, both of which tend to reduce the isolation and thus “purity” of any language community) is such an important factor.

    At least in NYC, as far as I can tell (from inter alia some modest experience representing pro bono clients who had been involved in that line of work) you generally can’t deal crack successfully without a certain amount of fluency in English or at a bare minimum maybe Spanish. There are other sorts of money-making opportunities available to recently-arrived L1 speakers of less common languages like Albanian or Mixtec or whatnot who haven’t yet mastered minimal competency in one of the more locally common languages.

  24. “The endangered-species parallel implies that we should take speakers of endangered languages into captivity and put them behind glass in a zoo or museum, to insulate them from the risks of contact with the wider world and to prevent them from breeding with unapproved partners.”

    No one would suggest such a thing and it certainly was not implied, although it seems that you did infer it.

    The comparison was the benefit to one group might involve a loss to a larger group. The economic status of small-language speakers can be improved, at least in some cases, by adopting the dominate language thereby benefiting these people while the original language becomes threatened with extinction with a loss to humankind. Nothing more was intended or implied.

  25. Why would I believe that?

    Because in these discussions you keep saying “It’s their choice” as if that were the end of the matter — they’ve chosen to learn a majority language and we should support them in their choice and be happy instead of wringing our hands about some supposed loss. But the loss is real, and a lot of people would prefer to be able to pass their language on to their kids rather than biting the bullet and bringing them up as speakers of a language that is not theirs so they will have a better chance in life. Some of us think that’s an unfortunate situation, much as it is unfortunate that people find themselves in situations where selling crack (in inner-city America) or growing opium poppies (in Afghanistan) is the best available option for getting out of poverty or even just staying fed. It’s one thing to say “Yeah, it sucks, but that’s life in the twenty-first century” and another to say “Don’t worry, be happy!”

    And yeah, nobody’s talking about putting them behind glass, that’s a straw man.

  26. J. W. Brewer says

    Also worth considering that language change is rapid and that the size of the underlying human population has exploded. Back when some unknown number of people in some unknown location spoke something kinda sorta resembling reconstructed Proto-Indo-European (let’s say 5,000 B.C. just to have a round number), the whole human population of the earth was a tiny fraction of what it is now (wikipedia’s best estimates run from 5 million to 20 million – we could call it 7 million for a handy ratio to today’s 7 billion). How much linguistic diversity did that size of global population manifest? Was there really a time (without going all the way back to the African savannah) when there were only a dozen or two human languages out there – PIE, Proto-Sino-Tibetan, Proto-Semitic, Proto-Austronesian, and the other products of modern reconstructed scholarship? Probably not; when you think about how widely dispersed the small population was geographically, it seems equally plausible that there were lots of languages then extant that have left no marks whatsoever in the historical record, even if we assume everything currently called an “isolate” is the sole survivor of a once-larger family. The massive spread of Malayo-Polynesian did not *necessarily* displace/destroy previously-extant languages, but it seems implausible to assume the same for e.g. the massive (and quite recent) spread of the Bantu languages, and we can infer that languages moved around quite a bit in North America in the centuries and millenia before European contact and we should likewise assume that an unknown number of unattested languages became extinct as a side-effect of that process. Heck, we can’t even figure out all that much about now-extinct non-IE languages (e.g. Etruscan) spoken in Europe in historical times. There was never a golden age of simultaneous linguistic diversity and linguistic stability. Life is change / How it differs from the rocks, as the hippies put it back when hat was a younger fellow. Or, perhaps more pointedly, “Soon you’ll attain the stability you strive for / in the only way that it’s granted / in a place among the fossils of our time.” It’s not a question of language extinction being a good thing or a bad thing — it’s an inevitable thing. Language change is part of nature, and nature is red in tooth and claw, as someone (probably not a hippie) put it.

  27. Actually, growing opium poppies in Afghanistan for sale as medical morphine for the Third World would be a huge win; the shortage of opiates outside developed countries is a cruel scandal.

  28. Let’s use pre-contact North America as an example of relatively undeveloped and sparsely populated region just starting transition to agriculture.

    There are about 300 indigenous languages spoken or formerly spoken north of Mexico. Pre-contact population of North America north of Mexico is disputed, but it is unlikely to have exceeded 10 million.

    So, let’s take 10 million and divide by 300. We get 3000 as average number of speakers per language.

    Apply this to 7 million which you suppose were living in the world back in 5000 BC and we get about 200 languages.

    Since Ethnologue lists more than 6000 languages today, we have to reach a surprising conclusion that numbers of languages spoken have increased thirtyfold since then.

  29. Sorry for wrong calculation. Dividing 10 million by 300, of course, gives 30 000, not 3000. It doesn’t change the subsequent results, though.

  30. Another example is aboriginal Australia. It was very undeveloped region which probably corresponds to the world in 10 000 BC rather than 5000 BC.

    Pre-contact population of Australia is estimated variously between 300 thousand and one million and it spoke some 250-300 languages (divided into over 600 dialects).

    It gives us average number of speakers per language at between 1000-4000.

    Application of this number to 7 million in 5000 BC would give 1300-7000 languages compared to 6000 languages now.

    Again, we are forced to conclude that the numbers of world’s languages have increased.

  31. Numbers of languages

    There are a number of problems in estimating both the size of human populations in a given period and the number of languages spoken by these populations. For the linguistic aspect, one problem is that of defining ‘language’ versus ‘dialect’, since the line, or rather the area, between total intelligibility and total unintelligibility is a rather fuzzy one. Ethnologue errs on the side of multiplying languages. I looked at its section on French a few years ago and I think it listed at least 7 different French languages! all of which were mutually intelligible and differing only in minor points. Native speakers are often unreliable sources about intelligibility, which is not an absolute criterion but often reflective of the prestige or lack thereof of given varieties: speakers often overestimate the resemblance of their own speech variety to a prestigious one, but claim not to understand a more closely related variety spoken by a socially disfavoured population.

    About North America, it is likely that the estimate of 300 original languages is an overestimate. Pre-contact cultures showed a lot of diversity, some of them heavily dependent on true agriculture, some building impressive cities and monuments, etc. Only a few were nomadic hunter-gatherers (a lifestyle which has its own sophistication).

    Even if some estimates for numbers of current or attested individual languages can be considered reasonably accurate, these numbers say nothing about the number of language families, or higher-level groups of languages. In North America (North of Mexico) the ‘mainstream’, conservative estimate counts approximately 60 different families (about the same as in mid 19th C when the first comprehensive survey was undertaken). In fact many of these families appear to be related to each other, so that their actual number is most likely much smaller. (The recent proposal by Greenberg combining most of the families of the entire Americas into a single “Amerind” group of related languages leaves A LOT to be desired and has been rightly rejected by almost all Amerincanist linguists). The number of distinctive families of course has a bearing on the number of distinct populations which may have spoken each of the language ancestors.

  32. Oops: Amerincanist

    My fingers crossed Americanist and Amerindianist. The latter specializes in Amerindian languages, those of people formerly called “American Indians”. The resemblance with Greenberg’a “Amerind” is unfortunate.

  33. It would be very interesting to discuss whether the numbers of language families have decreased since 5000 BC.

    Let’s apply Australian example here. Some 200-300 languages there are divided into 12 language families plus 5 isolates. This is for population of half a million.

    So the world in 5000 BC with population of 7 million would have had about 200-300 language families and isolates.

    This doesn’t differ much from current situation ( Wiki list of language families lists about 200 language families and isolates)

  34. Ethnologue errs on the side of multiplying languages. I looked at its section on French a few years ago and I think it listed at least 7 different French languages! all of which were mutually intelligible and differing only in minor points.

    The only other langue d’oïl currently listed by Ethnologue as spoken in France is Picard/Chtimi, which is recognized as one of the regional languages of Belgium. Outside French territory, there are Cajun French and Walloon. Somewhat more remote linguistically is Arpitan (or Franco-Provençal), still marginally spoken in both France and Switzerland.

    Beyond these. the native languages listed by Ethnologue as spoken in France are only distantly related or unrelated to French: Alsatian, Basque, Breton, Catalan, Corsican, Greek (on Corsica), Standard Italian, Ligurian (on Corsica and also in Monaco and east to the Italian border), Luxembourgeois, Occitan, Portuguese, Spanish, Vlaams, and five Romany languages. In addition, there are of course many languages spoken by immigrants, of which various Arabic varieties have the most speakers, about 1.2 million in total.

  35. A simple rule of thumb I’ve seen upheld by both anthropologists and linguists is that the number of languages spoken on earth has been relatively constant independently of the number of humans. The explanation is the economic system. Even a primitive agricultural economy can feed up to 40-50 times as many people on a given area of land compared to a hunter-gatherer one.

    This means, that as hunter-gatherers we lived in small, isolated groups, where dialects/languages necessarily must have driven apart through the generations. The areas in today’s world where at least remnants of pre-agricultural economies exist (New Guinea, Brazil, Australia, parts of Africa, circumpolar Eurasia and America) seem to confirm such a, from the economy derived, diversity.

    Today five million people speak just 10 or so languages. Ten thousand years ago the same amount of people may have spoken some 6000 ones just as we do today. That those languages eventually may have evolved from just one or a few is another story.

  36. Oops! I counted like a hunter gatherer myself : one, two three, many.. Today five billion… and ten thousand years ago five million.

  37. David Marjanović says

    Ethnologue errs on the side of multiplying languages.

    It probably does more often than it errs in the other direction – but from all I’ve heard it’s very heterogeneous, not using any consistent criterion to decide what counts as a language.

  38. To make educated judgements about 6000 languages, Ethnologue probably needs 6000 experts who know these languages.

    I somehow doubt they have that many experts

  39. Here are the official criteria used by ISO 639-3 (which is the international standard based on Ethnologue and vice versa):

    Two related varieties are normally considered varieties of the same language if speakers of each variety have inherent understanding of the other variety at a functional level (that is, can understand based on knowledge of their own variety without needing to learn the other variety).

    Where spoken intelligibility between varieties is marginal, the existence of a common literature or of a common ethnolinguistic identity with a central variety that both understand can be a strong indicator that they should nevertheless be considered varieties of the same language.

    Where there is enough intelligibility between varieties to enable communication, the existence of well-established distinct ethnolinguistic identities can be a strong indicator that they should nevertheless be considered to be different languages.

    In addition, it’s noted that only autonomous sign languages are listed, not signed transcriptions of spoken languages. There is also a concept of macrolanguages, defined as ““multiple, closely related individual languages that are deemed in some usage contexts to be a single language”. This covers things like Arabic (30 languages, but excluding Maltese) and Chinese (13 languages, but excluding Dungan).

  40. I looked att Ethnologue under ‘Sweden’ and found it quite accurate. So is Saami not one but five languages since they are not mutually (fully) understandable. It illustrates that the Saami until recent times lived in small isolated groups as reindeer herders or hunter gatherers. One could argue that Meänkieli is a Finnish dialect and Romani Tavringer a Swedish sociolect. But as everybody admits it’s impossible to define an exact distinction between dialects and languages. You could as well discuss whether dawn and dusk are parts of the day, the night or neither.

    In this sense Ethnologue’s estimations of the number of spoken languages are far better than nothing.

  41. If we were to apply intelligibility criteria honestly, how many different English languages could be found in Britain alone?

  42. JC: number of French languages: I consulted Ethnologue quite some time ago, so its entries have probably been upgraded since. I was looking at French, not the minority languages of France.

  43. Checked Ethnologue for Mongolian. It lists Mongolian Halh, Mongolian Peripheral (Inner Mongolian), Oirat and Buryat.

    While I agree that this is a correct description of four principal literary norms of Mongolian, from purely linguistic point of view, the dialects of Mongolian falling under Mongolian Peripheral differ from each other as much as Oirat and Buryat.

    And the literary standard Mongolian is essentially the same both in Mongolia and Inner Mongolia (differences are on the scale of British English and American English).

    And Bathrobe no doubt would be astonished to learn that according to Ethnologue language use of Mongolian Peripheral (Inner Mongolian) is described as

    Vigorous. Chinese living in the area can also speak it. All domains. All ages. Positive attitudes.

  44. J. W. Brewer says

    The Mongolian/ethnologue/lumper/splitter point raise another facet of this. I think most of us might think it was in some sense a loss (maybe historically inevitable, maybe offset by other gains, but nonetheless in some sense a loss) if the ethnic-Mongol population of Inner Mongolia ceased over time to use “their” traditional language and their kids and grandkids all became monolingual Mandarin (or bilingual Mandarin-English!) speakers. (Maybe less of a loss if an increased rate of marriage between ethnic-Mongol and Han means it becomes increasingly difficult to tell who is who, or at least that would be my idiosyncratic take.) But assume that happens but Mongolian remains a flourishing/dominant language right across the border in formerly-outer Mongolia proper. Is that “language extinction”? Does the answer depend on whether we agree with ethnologue that what is spoken on the PRC side of the border is a different language as opposed to simply a regional variant?

    Similarly, it’s sorta cool that there’s still a Welsh-speaking enclave in Patagonia, but if the Welsh-ancestry population there were ultimately assimilated and became monolingual Spanish-speakers, Welsh would still exist in the world (unless ethnologue already thinks “Patagonian Welsh” has diverged enough to be a different thing?). French’s status as a major world language was not affected by the loss of French in the one time Francophone enclaves in colonial New York (here in the close-in NYC suburbs, I presently live less than two miles from a church that reportedly still kept its parish register in French circa 1740 when one of my ancestresses was baptized there), and probably would not be affected all that much if the grandkids of the current Francophone population of Canada all ended up monolingual Anglophones (or maybe bilingual in Mandarin, for all I know). But the social/historical dynamics that affect whether or not Welsh-in-Patagonia (or French-in-New-Rochelle) survives in that particular community are not all that different from those that affect whether a language uniquely spoken in a similarly-sized enclave survives. Of course, for more marginal languages with historically wide geographical range, you may never know which is the enclave that might be the last one. So e.g. if you are interested in the survival of Yiddish in the world you ought to root for the success of the remaining Yiddish-speaking enclaves in New York (assuming you are willing to tolerate the highly illiberal means that appear to be necessary to that success), because it might be imprudent to assume that any other geographically-separated group of Yiddish speakers will necessarily keep the language going if they don’t, whereas it seems unlikely at present (though not impossible) that the Patagonia enclave will be where Welsh makes its last stand.

  45. Patagonian and rectangular (to coin a term) Welsh are still mutually intelligible. Systematic differences include deaspirating p, t, c and pronouncing mh, nh, ngh, rh, h as m, n, ng, r, and zero. The main problem is the Red Queen’s: “Speak in English when you can’t think of the Welsh for a thing”, which doesn’t work when your interlocutor’s fallback language is Spanish. The maintenance of Welsh in Patagonia depends a good bit on tutors and television programs from Wales, though.

  46. J. W. Brewer says

    http://www.nytimes.com/2014/08/29/science/study-offers-clues-to-arctic-mystery-paleo-eskimos-abrupt-extinction.html was a story last month reviewing a new study claiming that the old population of much of the Canadian Arctic until approx 7 centuries ago (not that far before 1492, and they *might* have been among the mysterious Skraelings the Norse encountered earlier on), the so-called “Dorset People,” disappeared without survivors, i.e. the genetic evidence is that they did not mix reproductively with the new arrivals to become part of the ancestry of the more recent Inuit-speaking inhabitants of that area, but were totally replaced in their former territory by the Inuit speakers, by inference after themselves having all died off by famine, epidemic, massacre or some combination of the above (probably too little GDP per capita than too much). I assume we don’t have a frickin’ clue what language they spoke. There is no particular reason to believe it was anything closely related to Eskimo-Aleut, whose own distant connections with other things that were floating around Siberia 3 or 4 millenia ago are themselves speculative/disputed. We just don’t know. And this was the dominant human language over a decent chunk of the earth’s surface not all that long ago in a big-picture sense, contemporaneously with when Middle English was the dominant language in much but not all of Great Britain. But the NY Times story doesn’t even mention language, and the loss w/o documentation of what might have been a totally cool and awesome language with who knows what fascinating phonological or morphosyntactic features is, I think, fairly far down the list of things to be regretted or thought tragic about the disappearance without any genetic descendents of an entire people. For me, it brings home how “language extinction” is really not at all the same thing as literal physical/biological extinction, although as in this instance it might be a secondary consequence of it.

  47. But the NY Times story doesn’t even mention language

    Are you proposing NY Times reporters as proper judges of what is and isn’t worth attention?

    and the loss w/o documentation of what might have been a totally cool and awesome language with who knows what fascinating phonological or morphosyntactic features is, I think, fairly far down the list of things to be regretted or thought tragic about the disappearance without any genetic descendents of an entire people.

    Just out of curiosity, what other things would be on that list, and in what order?

    For me, it brings home how “language extinction” is really not at all the same thing as literal physical/biological extinction, although as in this instance it might be a secondary consequence of it.

    Did anyone say or imply it was? Do you really think anyone thinks it is? Once again, you seem to be dragging out straw men to slap around. Something doesn’t have to be The Worst Thing Ever to be worth regretting, but you seem to feel that if you’ve demonstrated something isn’t as bad as the Holocaust, we might as well forget about it and move on to something more important.

  48. I read the article in question.

    The disappearance of the Dorset people reminded me of the diseappearance of the Scandinavian Greenlanders (as described in Jared Diamond’s Collapse), mostly through famine brought about directly or indirectly by climate change for the worst. JD makes the point that the Greenlanders coexisted with Eskimo people who were used to living with ice conditions and to fish and hunt under those conditions, but people practicing a European type of agriculture and unable or unwilling to adopt the Eskimo lifestyle were doomed once the climate no longer allowed them to continue the way of life they were used to.

  49. J. W. Brewer says

    Hmm. Does bringing up the Holocaust (for the first time in the thread) while simultaneously saying the thing you’re concerned about is not quite as bad as the Holocaust sufficiently implicate Godwin’s Law that I get to declare victory?

    The Amano et al. paper linked in the original post consistently treats language extinction as more or less the same sort of phenomenon as species extinction, without spending a lot of time saying “of course, we know that this is a very loose and poetic metaphor that is not to be taken overseriously.” Sentences like “This contrast might show that linguistic conservation has been less successful and/or has attracted less attention even in economically developed temperate regions, compared to biodiversity conservation” suggest to the contrary that they’re taking the metaphor quite seriously and running with it and/or running it into the ground. But if hat or anyone else wants to distance themselves from that style of talking about things, I’m all for it.

  50. David Marjanović says

    Eskimo-Aleut, whose own distant connections with other things that were floating around Siberia 3 or 4 millenia ago are themselves speculative/disputed

    I find this pdf intriguing, though.

    sufficiently implicate Godwin’s Law that I get to declare victory?

    Godwins law only states that, as a discussion goes on, the probability that someone will mention Nazis approaches 1. It does not say anything about Final Victory.

  51. I agree that language and biological extinctions are two separate types of phenomena and that drawing exact parallels between the two situations is often a mistake, since the circumstances that lead to preservation or extinction, and the countermeasures that might be taken, are so different. When a species disappears, the animals or plants are no longer around, able to be seen, heard, smelled, hunted, eaten, etc by observers (apart from no longer living their own lives in which they were seeing, smelling, eating, etc on their own). Preserving a species largely means preserving the natural habitat to which it is adapted, in terms of both quantity and quality: “if you provide it, they will come back”, if there are enough of “them” around. When a language disappears, it means that nobody speaks that language any more, so the language is no longer heard or said, although many of its features may be preserved in writing. But the people who might have acquired and spoken that language under different, sometimes completely changed circumstances are still around, though (usually) speaking a different language. They have not disappeared, and unless the last speaker has also been the last of his people (as in the case of “Ishi” in California), they have not stopped communicating through speech, let alone stopped living. Preserving a language on the verge of extinction does not mean preserving its actual speakers, who will die eventually, it means making sure that the children in the relevant community are acquiring the language in normal settings, meaning living from infancy with people who speak it as their normal means of communication. But trying to recreate a family or community speaking language X when most of its members except the oldest one(s) have limited if any facility in it is even more difficult than preserving an animal habitat since not only the people’s ability but their attitudes to this language and to the one they normally use are not natural results of their own makeup (as with animals’ relationship to their environment) but depend largely on a variety of social circumstances, including their background and upbringing. Some communities are making heroic efforts to revive a language now known only through written documentation, with the whole community attending courses, camps and other social gatherings conducted in the language, but communities where there are only a few speakers left are not necessarily doing better at preservation, because some old people are proud of their command of a once-despised language and (perhaps from a barely conscious desire for revenge) unwilling to help younger people who are struggling to acquire a language with very different features from the one they are used to. Nor is it obvious for the speaker of any language to “teach” it to potential learners: immigrants, like children, learn from being around speakers much more than if one of those speakers undertook to teach them. (Linguists are eager learners, and figure out how to get taught).

    People who for some reason have not spoken their language to their children when they were small are overwhelmingly bilingual, so not speaking that language has not prevented them from speaking to those children. In many cases the adults seem to assume that the children will start “speaking their own language” in good time. The loss of language starts being obvious to a community when a couple of generations have not acquired the language in question, not because a form of mute communication was used instead (because of deaf parents, for instance) but as a result of being spoken to in a different language. Old people may deplore the fact that their children and grandchildren “don’t want to speak the language” (that in many cases they are incapable of speaking), but they don’t necessarily realize the language is on its way out until there are only a handful of speakers left, or they are indeed the ONLY speaker left and have no one left to speak it with.

  52. I think it would help if everyone made distinction between languages spoken by tiny groups and languages spoken by millions of people.

    Both categories can be threatened, both can go extinct, but in very different ways.

    Mongolian in Inner Mongolia and native languages/dialects in southern France fall in the latter category. These languages are highly developed and associated with a lot of high culture unlike dying tribal languages spoken by tiny groups of people.

    I also would add a third category of language extinction where a national government is actively trying to eradicate a language or dialect spoken by absolute majority of people in that region (or even in the country as a whole).

    Chinese government appears to be successful in eradicating Chinese dialects/languages different from Mandarin, even though they are still spoken by hundreds of millions of people.

    And Indonesian government is determined to destroy Javanese (spoken by majority of Indonesia’s population) along with other Indonesian languages (in favor of Bahasa Indonesia which is spoken as native language a small minority)

  53. David: thanks for the link to the pdf on Siberia and Eskimo-Aleut. I have looked at it briefly and frankly I am not impressed. A few brief comments:

    – The author relies on a 100-word Swadesh list (of “basic” words likely to exist in every language). This means relying almost exclusively on lexical items, which are only part of what is needed for historical study (morphology must also be considered in so far as the data allow). In addition, the author allows a wide semantic range for the meanings of those items, so that forms allegedly related might not be so when further study narrows down the relevant meanings (something always a problem with superficial language studies).

    – He then uses the percentage of common items between languages to determine a) the degree of linguistic relationship and b) the precise date of the divergence between two languages. These were Swadesh goals in compiling his lists, but attempts to use this method (“glottochronology”) has shown it to be highly inaccurate when applied to language families with a known history (eg Germanic, Romance, etc), especially as a dating method. As for degree of relationship, having had a look at lists for Koryak, Chukchi and Alyutor (Itelmen) myself, I found Itelmen very, very different from the other two.

    – The alleged phonological correspondences are all over the place, with very few examples of regularity. The reconstructions are mosly the author’s own, and even assuming that the lexical comparisons are reasonably accurate (which is not a given), I would want to consult other opinions (perhaps in the references given) in order to evaluate correspondences and reconstructions, some of which require what seem to be rather convoluted explanations on the part of the author.

    In other words, something to keep on file, but not to rely on for sweeping generalizations.

  54. SFR: … native languages/dialects in southern France fall in the latter category. These languages are highly developed and associated with a lot of high culture

    The Occitan varieties in Southern France (especially those in the Southwest) were associated with high culture in the Middle Ages, but from the 16th century approximately their use declined among the upper class and they were soon fragmented into many different local dialects used mosly in rural districts (as in my grandparents’ case) although some local poets continue to use them (but not in the sophisticated manner of the medieval poets). Nowadays there are very few, mostly very old native speakers although many younger people are trying to relearn them.

  55. Re: Mudrak’s article on Proto-Altaic/Proto-Eskimo comparison.

    He is trying to prove that they are related at the level slightly below Nostratic/Eurasiatic.

    Which is, of course, very problematic since separation goes back to about 10-15 thousand years ago. He might as well try proving Amerind.

  56. Re:Alyutor (Itelmen)

    Alyutor and Itelmen are different languages. Alyutor is a Chukotkan language closely related to Koryak and Chukchi, while Itelmen is an Kamchatkan language, related more distantly to Chukotkan languages.

  57. SFR: thanks for the correction about Alyutor and Itelmen. I was relying on my memory and had the impression that the two words had been used for the same or closely related languages.

  58. To make educated judgements about 6000 languages, Ethnologue probably needs 6000 experts who know these languages.

    Not really. Most languages are not “isolates”, they have “relatives” in the same family, and many linguists specialize not just in a single language but in a family, such as Germanic, Slavic, Romance, Semitic, Malayo-Polynesian, etc. When dealing with classification and language history it is much better to consider a family than a single language, since some features of each of the component languages are strange from the point of view of that particular language but easily explained by comparison with the other family members.

  59. Ethnologue claims that it relies on “mutual intelligibility” criteria for language classification. This requires either native speakers or experts specialising in that dialect/language to make an educated judgement on whether these dialects/languages are mutually intelligible or not.

    To use the above mentioned example of Inner Mongolian, its numerous dialects can range from virtually indistinguishable from Halh Mongolian to essentially unintelligible.

    Ethnologue’s choice to divide Mongolian in two languages is not based on mutual intelligibility criteria, but on extra-lingustic considerations

  60. @David Marjanović: (This is off topic, but I wanted to mention it, since Godwin’s Law came up.) When Godwin originally stated his law in 1990, it merely stated that eventually, Hitler or Nazis would come up in a discussion. However, within a few years (no later than 1994, I’m sure), it had been extended to say that once Nazis were mentioned, the discussion was supposed to be over, and whoever had first mentioned them could be taken to have lost the debate. In this 1995 email from Godwin, he implicitly accepts that this is part of the law by quoting Quirk’s Exception, that you cannot force a discussion to end by bringing up the Nazis just for that purpose.

  61. Ethnologue’s choice to divide Mongolian in two languages is not based on mutual intelligibility criteria, but on extra-lingustic considerations

    I think that is the main reason for it to divide French into several languages. This division implies unintelligibility, which is grossly exaggerated.

  62. Godwin’s Law extension: in a long enough internet discussion the Nazis or Hitler will be brought up with probability approaching 1, after which discussion will be steered toward the meaning of Godwin’s Law.

  63. Bah, this is Languagehat, where etymological debates trump semantic ones evvery day of the week! It seems the only thing Godwin invented about Godwin’s law is its name. The phrase argumentum ad Hitlerum ‘Hitler was a vegetarian, so vegetarianism is bad’ goes back to Leo Strauss (1951), and there is someone (I can’t track him down right now) who was able to show a Usenet posting predating Godwin and with the same purport.

  64. Say, what’s the etymology of the name Hitler, anyway?

  65. The closest I can find with the help of google is Hungarian hit and hittel, both giving ‘faith’ in English, ‘Glauben’ in German and ‘tro’ in Swedish. The ‘-er’ part is of course the same in German as in English, i.e. a male performer of something (baker, painter, worker). So ‘believer’ might be a suggestion. That he didn’t appreciate Christianity (or didn’t look very ‘aryan’) is another story.

  66. David Marjanović says

    and b) the precise date of the divergence between two languages.

    Yeah, I’m not intrigued by that part. Grottoclonology, as it’s been called, is a somewhat embarrassing attempt to invent molecular dating while only having 1950s computing power at one’s disposal…

    He is trying to prove that they are related at the level slightly below Nostratic/Eurasiatic.

    He’s arguing that they’re very, very close relatives, with Altaic having about one innovation (a /q/-/g/ merger) compared to its common ancestor with Eskimo-Aleut.

    …without any Aleut data, because Proto-Aleut and Proto-Eskimo-Aleut still haven’t been reconstructed. There’s obvious room for improvement all around!

    In this 1995 email from Godwin

    …Oh.

    Say, what’s the etymology of the name Hitler, anyway?

    The variant Hüttler appears to exist; Hütte = hut. “Can’t afford a proper house” or something like that.

    There’s no evidence for ancestors anywhere near Hungary.

  67. Both of Hitler’s putative grandfathers were named Hiedler, but what a Hiedel might be, I don’t know.

  68. Could it be from heide – heather?

  69. @John Cowan: Wikitionary gives the seemingly absurd: “a surname applied to those who resided near a Hiedl (‘subterranean river’)” in its etymology of “Hitler.”

  70. Uncyclopedia s.v. Godwin’s Law.

  71. David Marjanović says

    Both of Hitler’s putative grandfathers were named Hiedler, but what a Hiedel might be, I don’t know.

    Neither do I. But what the Standard German form of that would be depends on 1) whether ie is meant to represent the dialectal diphthong [ɪɐ̯], which usually corresponds to Standard ie [iː] (with exceptions both ways), or some kind of monophthong (more or less [i]), and 2) how far gone the /d/-/t/ distinction was in whichever dialect this name was first written down in this form: geographically, that was almost certainly a Central Bavarian one, and in those (like mine) /t/ generally changes into /d/ at the ends of words and between vowels – even though it remains voiceless, making it very similar to the unaspirated [t] and creating more possibilities for confusion and hypercorrection. In short, it’s all a mess. 🙂 – There’s no phonemic vowel length in Central Bavarian; like in Russian, stressed vowels are generally longer than unstressed ones, and that’s pretty much it.

    Could it be from heide – heather?

    Wrong vowel.

    Hiedl (‘subterranean river’)

    I don’t know any such word. Doesn’t mean it doesn’t exist; on the other hand, subterranean rivers (that anyone knows about, so not moving groundwater) only occur in the Calcareous Alps, which would narrow down the geographic origin of the name quite a lot. – The -l has to be a diminutive suffix.

  72. David Marjanović says

    I just love the jab at the Kansas State Board of Education. 🙂

Trackbacks

  1. […] Hat summarizes a paper suggesting that language death and economic success are […]

Speak Your Mind

*