The Prehistory of Prehistory.

A 2006 paper (pdf) by Peter Rowley-Conwy, “The Concept of Prehistory and the Invention of the Terms ‘Prehistoric’ and ‘Prehistorian’: The Scandinavian Origin, 1833–1850” (European Journal of Archaeology 9:103–130), not only antedates by twenty years the OED’s first citation for the English word (1871 E. B. Tylor Primitive Culture II. 401 “The history and pre-history of man take their proper places in the general scheme of knowledge,” in an entry updated in March 2007), it provides a fascinating look at how the term and the concept developed. Here’s the abstract:

It is usually assumed by historians of archaeology that the ‘concept of prehistory’ and the terms ‘prehistoric’ and ‘prehistorian’ first appeared in Britain and/or France in the mid nineteenth century. This contribution demonstrates that the Scandinavian equivalent terms forhistorisk and förhistorisk were in use substantially earlier, appearing in print first in 1834. Initial usage by Molbech differed slightly from that of the present day, but within three years the modern usage had been developed. The concept of prehistory was first developed at the same time by C.J. Thomsen, though he did not use the word. It was used more frequently in the nationalism debates of the 1840s, particularly by J.J.A. Worsaae. One of the other protagonists, the Norwegian Peter Andreas Munch, was probably responsible for introducing the concept to Daniel Wilson in 1849, and suggesting that an English equivalent to forhistorisk was required.

And here’s the beginning of the introduction:

Modern archaeologists clearly grasp the concepts of ‘history’ and ‘prehistory’. We have no difficulty envisaging a historical period extending a certain distance back, and before this a much longer prehistoric period extending into the deep past. For the historical period there is documentary evidence; but we accept that there was a prehistoric period, studied by prehistorians, for which there is (by definition) no documentation – material evidence is the only means by which we can examine it. But until the 19th century there was no concept of prehistory. The origin of the concept is one of the key developments in our understanding of the human past, and has seen considerable discussion in the recent English and French literature. This discussion has explored the 19th century archaeologies of the two countries, examining both the concept of prehistory, and the terminology used to discuss it.

Terminology is the more clear-cut. The first use of the word ‘prehistoric’ in English was not by an Englishman at all, but by the Scot Daniel Wilson, in his Archaeology and Prehistoric Annals of Scotland (Wilson 1851). It is generally thought likely that he invented the term himself (Daniel 1964, Chippindale 1988, Kehoe 1991, 1998, Trigger 1999, Kelley 2003). Clermont and Smith (1990:98-99) however point out that the French archaeologist Gustav d’Eichthal however used préhistorique as early as 1845, but add that it remains unclear whether Wilson ever came across this and (consciously or unconsciously) adopted it. Préhistorien first appeared in French in 1872, almost 20 years before ‘prehistorian’ first occurs in English (Clermont and Smith 1990:97).

The concept of prehistory is less clear-cut. In English, Daniel Wilson was probably the first to demonstrate a clear grasp of it; certainly no English or Irish archaeologist did so before him (Rowley-Conwy in press). In France, Paul Tournal used anté-historique as early as 1833 (Coye 1993, Stoczkowski 1993), but it is open to question whether he understood this in the way ‘prehistoric’ is now used. Tournal excavated caves and recovered human artefacts and bones of extinct mammals in the same layers, believing them to be contemporary. […]

The purpose of this contribution is to argue that the Anglo-French focus of the recent discussion has been misplaced. The Danish word for ‘prehistoric’ was first published in 1834, over a decade before even d’Eichthal’s monograph. This has passed almost entirely unnoticed in the recent Anglo-French literature […]

That gives you the gist; if you’re as interested in this stuff as I am, you’ll want to visit the link for more. (Christian Molbech, by the way, is perhaps best known today for having been nasty to Hans Christian Andersen.)

Comments

  1. This looks great – thanks, Hat. I’ve downloaded it for later. Earlier this year I read a book about the idea of prehistory (whose author may be the Daniel 1964 in your post) and quoted a nice passage from it that you might enjoy.

  2. Thanks for this – very cool. The late Bruce Trigger (of Trigger 1999, cited above) was my PhD supervisor, so this discussion brought back some great memories of my time at McGill anthropology in the late 90s. I’m going to share your post with my current undergraduates, who are working on word history projects like this right now.

  3. How widely accepted is the concept of prehistory? Is there any kind of deconstructionist debate questioning its validity?

    It seems to me that China’s national ideology rejects the notion of prehistory because it isn’t conceived as cutting off at the start of writing; it seems to extend back further. Apart from the Chinese, is there anyone who seriously suggests that history should include pre-written ‘history’?

  4. “How widely accepted is the concept of prehistory? Is there any kind of deconstructionist debate questioning its validity?”

    It has been my understanding, which could well be wrong, that the term simply is used to designate a period before which there is any written record. If that is correct, then the time period would differ from place to place.

    It is also my understanding, that historical archaeology correlates and compares artifacts with the documentary record. As I recall, for example, that artifacts found in some places challenge and dispute the history written record.

  5. the time period would differ from place to place

    Indeed. As Wikipedia notes, Egyptian prehistory ends in the -32C, whereas New Guinea prehistory extends to the 20C, a difference of five millennia. Throughout this period, we also have the problem that early written records may be less history than myth (when people write about themselves, like the Romans) or libel (when people write about their illiterate neighbors, like the Romans writing about the Celts), and therefore may need correcting against other sources of information.

  6. “we also have the problem that early written records may be less history than myth (when people write about themselves, like the Romans) or libel (when people write about their illiterate neighbors, like the Romans writing about the Celts), and therefore may need correcting against other sources of information.”

    I think that is one of the values of historical archaeology.

  7. John Emerson says:

    A Dane, C.J. Thomsen, also originated the Stone Age /Bronze Age / Iron Age succession. To say nothing of their bog people (P.V. Glob, The Bog People. Glob was also a sityationist and a supporter of the kandinavisk Institut for Sammenlignende Vandalism.

  8. I think “prehistory” is considered an outdated 19th century term in Russian.

    Like “natural history”…

  9. marie-lucie says:

    It is understandable that people whose recent “history” is recorded in others’ written documents would not be happy with the word “prehistory” which suggests that either nothing of interest ever happened to them or that their traditions are not considered valid at all. When in elementary school I learned about les hommes préhistoriques, who lived in caves, hunted mammoths, etc. Calling the recent ancestore (a few generations ago) of some of the world’s peoples “prehistoric” might be perceived as lumping them together with stereotypical European “cavemen” (eg according to the still popular image of Neandertals).

    In linguistics, “historical linguistics” started from the study of old written documents, from which it was possible to extrapolate and reconstruct (as much as possible of) the unwritten common ancestor of some language families (PIE and a few others), so the term is currently applied to the historical study of languages both with and without such documents.

  10. — “historical linguistics” started from the study of old written documents

    Not quite. Vedic Sanskrit was preserved orally for one thousand years before it was written down.

  11. It used to be amibiguous, but is now less so, whether “history” applies to oral history or not.

    I once heard Malcolm Ross, a great Austronesian historical linguist, speaking of how he’d gotten his start in the field. He was working in Papua New Guinea as a school teacher and principal, and was stricken by how the history curriculum discussed Australian and British history, but nothing local. When he asked his students (or maybe it was another teacher?) why that is, he was told, “because we have no history!” That gave him the impetus to go to graduate school and study historical linguistics, aiming at the prehistory of New Guinea.

  12. There are two West Papuan languages which were spoken by peoples who had states, sultans, literacy (in Arabic) and written history dating back to 13th century.

    These Papuan states controlled world access to spices and played a key role in early modern global trade and world history.

    Columbus accidentally discovered America trying to reach these spice islands.

  13. I assume you’re referring to the Moluccas (Maluku), which are technically not ‘Papuan’, although their languages definitely belong to the Papuan languages.

  14. Wikipedia says “Papuan is an umbrella term for the various indigenous peoples of New Guinea and neighbouring islands, speakers of the Papuan languages.”

    Sultanates Ternate and Tidore were certainly Papuan by this definition.

    These states also occupied parts of island of New Guinea (though they never reached eastern part New Guinea which now forms independent nation of Papua New Guinea, but surely PNG doesn’t have monopoly on term Papuan)

  15. marie-lucie says:

    – “historical linguistics” started from the study of old written documents
    – SFR: Not quite. Vedic Sanskrit was preserved orally for one thousand years before it was written down.

    I think this is a misunderstanding. Oral preservation of older texts through memorizing them is not “historical linguistics” any more than memorizing a poem or play comes under the definition of “linguistics”. Indian scholars writing down and studying Vedic Sanskrit and writing grammars of it (the most famous being Panini’s) were not doing “historical linguistics” in the modern Western sense. Indeed Panini’s work has been described as a masterpiece of “descriptive linguistics”.

    Historical linguistics did start from the comparative study of of written documents of various ages in order to establish rules of both correspondences (as in Latin – Greek – Sanskrit) and evolution (as from Latin to Spanish, Italian, French etc, Sanskrit from the related Indian languages, etc). Where old documents were not available from languages considered to be related (as with German, Dutch, English etc), linguists worked on using such comparisons to reconstruct the common language which the modern ones must have derived from. This method actually started with the reconstruction of the (at first hypothetical) Proto-Indo-European language which involved considering and comparing all the component language families). To my knowledge nobody has suggested that this was the type of work that Panini or anyone else at the time was engaged in.

    For languages without a long written tradition, such as in Africa, the Americas and some other places, it was once thought that the same method could not be applied in the absence of documents predating European arrival, but in America this idea was put to rest by Bloomfield’s work on Proto-Algonquian reconstruction.

  16. I think William Jones in the 18th century is considered the beginning of historical linguistics.

  17. –I think William Jones in the 18th century is considered the beginning of historical linguistics.

    And he started with Sanskrit, the language which goes back all way to 2nd millenium BC (and lacked writing half of its age)

  18. marie-lucie says:

    SFR: Sir Wililam Jones … the father of historical linguistics … started with Sanskrit …

    Sir William Jones did not set out to start historical linguistics! or even to study the history of Sanskrit. Like many educated Englishmen, he had been taught Latin and Greek in his youth, and when found himself a judge in India, he started to learn Sanskrit to help him professionally, since he realized he needed to understand the local traditions of jurisprudence and literary allusions). He was astonished by how much Sanskrit resembled Latin and Greek, especially in its noun and verb morphology (declensions, conjugations), and deduced that the three languages must have derived from a common ancestor, “which perhaps no longer exists” (in contrast to a prevailing opinion that “the earliest language” was still to be found somewhere on earth). Whether Sanskrit was older (or at least attested from a much earlier period) than Latin and Greek was not relevant to whether the three languages were genetically related or not.

    Jones’ lectures and writings about his linguistic work spread his ideas among European scholars, with great success. The methods of systematic comparison and proto-language reconstruction were developed over the next century or so, using not just these three languages but most of the ones now considered Indo-European, and were also shown to be applicable to other language families. These methods, not the age of the oldest documents (or of orally preserved traditional texts), are the heart of historical linguistics.

  19. I think William Jones in the 18th century is considered the beginning of historical linguistics.

    This is an outdated view. Wikipedia has a good summary:

    Early Indo-European studies

    In a publication of 1647, Marcus Zuerius van Boxhorn proposed the existence of a primitive common language he called “Scythian”. He included in its descendants Dutch, Greek, Latin, Persian and German, and in a posthumous publication of 1654 added Slavic, Celtic and Baltic. The 1647 publication discusses, as a first, the methodological issues in assigning languages to genetic groups. For example, he observed that loanwords should be eliminated in comparative studies, and also correctly put great emphasis on common morphological systems and irregularity as indicators of relationship. A few years earlier, the Silesian physician Johann Elichmann (1601/1602–1639) already used the expression ex eadem origine (from a common source) in a 1640 study published relating European languages to Indo-Iranian.

    The idea that the first language was Hebrew continued to be advanced for some time: Pierre Besnier (1648–1705) in 1674 published a book which was translated into English the following year: A philosophical essay for the reunion of the languages, or, the art of knowing all by the mastery of one.

    Leibniz in 1710 proposed the concept of the so-called Japhetic language group, consisting of languages now known as Indo-European, which he contrasted with the so-called Aramaic languages (now generally known as Semitic).

    The concept of actually reconstructing an Indo-European proto-language was suggested by William Wotton in 1713, while showing, among others, that Icelandic (“Teutonic”), the Romance languages and Greek were related.

    In 1741 Godofredus Henselius (1687–1767) published a language map of the world in his Synopsis Universae Philologiae. He believed that all languages were derived from Hebrew.

    Mikhail Lomonosov compared numbers and other linguistic features in different languages of the world including Slavic, Baltic (“Kurlandic”), Iranian (“Medic”), Finnish, Chinese, “Hottentot” and others. He emphatically expressed the antiquity of the linguistic stages accessible to comparative method in the drafts for his Russian Grammar published in 1755:

    Imagine the depth of time when these languages separated! … Polish and Russian separated so long ago! Now think how long ago [this happened to] Kurlandic! Think when [this happened to] Latin, Greek, German, and Russian! Oh, great antiquity!

    Gaston-Laurent Coeurdoux (1691–1779) sent in 1767 a Memoire to the Academy of Sciences (France) in which he demonstrated the similarity between the Sanskrit, the Latin, the Greek and even the German and Russian languages.

    Despite the above, the discovery of the genetic relationship of the whole family of Indo-European languages is often attributed to Sir William Jones […]

  20. “Papuan” is a wastebasket category for any language from the right part of the world that is neither Austronesian nor Australian. It isn’t even paraphyletic, really; there is simply no evidence at all for the relatedness of the Papuan languages, which divide into about 60 first-order families and lots of isolates.

  21. Like “Siberian,” then.

  22. Only more so, because there are fairly few languages in Siberia and there are a lot in New Guinea and its offshore islands. The area has two to three times the language diversity of Nigeria, but with about 6% of the population.

  23. Like “Paleo-Siberian”, which bundles the Siberian isolates, and is not used much anymore.

    Some linguists working in that corner of the Pacific use “non-Austronesian languages” to avoid the implication of relatedness in ‘Papuan’.

  24. New Guinea prehistory extends to the 20C… Throughout this period, we also have the problem that early written records may be less history than myth (when people write about themselves, like the Romans) or libel (when people write about their illiterate neighbors, like the Romans writing about the Celts), and therefore may need correcting against other sources of information.

    In Russian, prehistory is defined with an important additional nuance. Yes, it’s largely “before written sources” but also it’s about the times when the ethnonyms, toponyms, and names of cultural and political entities are merely conditional / invented postfactum, and circumspect. So we may be discussing e.h. Andronovo culture or the Denisovans, but we make no pretense of knowing their actual names and boundaries. OTOH in the presence of oral tradition, or untrustworthy early written records, we usually make claims of knowing the non-conditional names of peoples (distorted and non-native though they may be).

  25. This is pretty funny:

    The first use of the word ‘prehistoric’ in English was not by an Englishman at all, but by the Scot Daniel Wilson

    Whoda thunk? A word in English, coined by some guy who wasn’t even from England!

  26. marie-lucie says:

    LH, thank you for the WIkipedia citation (it never occurred to me to look there). Mallory’s In search of the Indo-Europeans cites two or three precursors but not this whole list, and nobody whose scope was as wide.

    It looks like the work done on the “continent” was not widely known. Jones’ work may have confirmed this work rather than pioneered it. In any case his brief statement, made for a general audience rather than a small circle of scholars, attracted much attention, at a time when the British conquest of India made the study of Sanskrit available to many more people.

    The insights of those early historical linguists about the proper methods of comparison (emphasis on morphology, irregularities, etc rather than just the lexicon) makes me aware once more of the sorry state of affairs in some areas of American Indian linguistics! (“A language family is either obvious, … or it is forever unknowable”, or words to that effect, as I keep quoting).

  27. Marie-Lucie–

    Sadder still is the fact that in some parts of the Americas American Indian linguistics actually had a head start: apparently the first claim of the existence of the Mayan language family was made in the sixteenth century by a Spanish missionary who wrote that all of the Mayan languages had once been a single language, adding that this was not so amazing, as the same was true of Spanish, which was simply Latin that the Spaniards had corrupted in their own way, just as the French and the Italians had each corrupted Latin in their own way.

    So in effect this missionary was saying that the Mayan languages bore the same relationship to one another that the Romance languages do. And we now know that this missionary was right. Yet despite this early discovery Mayan historical and comparative linguistics cannot be said to be more advanced than the historical and comparative linguistic work on scores of other language families…

    More and more it seems to me that the history of linguistics demonstrates one thing with crystal clarity: any claim that a science always progresses, that the current generation of researchers always builds upon and improves the work of the previous generation of researchers, is utter nonsense.

  28. David Marjanović says:

    Also, “the closer you get to humans, the worse the science gets” (biologist proverb that holds within biology and more generally).

  29. Counterexample: string theory.

    I’ve just fixed a bug in the script that generates the recently commented-on posts, so that it doesn’t truncate a title that has internal markup, like the Prehistory in the title of this post. Hurrah for Perl regular expressions!

  30. marie-lucie says:

    Etienne: Mayan languages

    If these languages are as close to each other as the Romance languages, it means that they form a single, “first order” family. In Eurasia such families were identified (even by other names) quite early, as some of the references above show. In North America alone (North of Mexico), approximately 60 such families were identified in a large-scale survey in the mid-19C, and the current classification is still basically the same. Comparative-historical work within those families (often in the absence of older documents) has not been too bad, but most linguists, with very limited training in historical methods, have been reluctant to look beyond the obvious groups and to look for larger groupings similar in scope to Indo-European. The six “phyla” or “superstocks” proposed by Sapir have mostly been dismantled (except Eskimo-Aleut and Na-Dene) but my feeling is that Sapir’s smaller number of less closely related anguage groups is probably more realistic (although a few isolates may remain unaffiliated).

    any claim that a science always progresses, that the current generation of researchers always builds upon and improves the work of the previous generation of researchers, is utter nonsense.

    Certainly there have been a number of examples in the 20th century, in varied disciplines! Part of it is the modern cult of originality: new proposals are often accepted simply because they are new, and the “old”, tried and true methods end up no longer being taught or even understood. I think this is less likely to occur in the hard sciences (barring totalitarian government intervention), where new “discoveries” spark replication attempts which confirm or destroy the new insights or theories: when some physicists announced they had achieved “cold fusion”, a number of labs immediately tried to reproduce the results and failed. The necessity of very specialized, often extremely expensive equipment for experiments also prevents fanciful explorations.

    In linguistics of course we have seen Chomkyan syntactic theory devour the teaching of more solid basics such as morphology. In comparative linguistics, there are so many languages, and the scholarship on them is so uneven in both quantity and quality, that it is difficult for any one person, let alone the general public, to evaluate proposals outside of their own corner of the field (and even inside it): witness the popularity of Greenberg’s classifications (Amerind, Eurasiatic, etc) in France, and the current trend for biologists to take over what had been the province of historical linguists, using a not even elementary understanding of historical methods. In my opinion, the problem is not that the traditional comparative method is outdated, or incapable of further development: it is that it has not often been used to its fullest extent.

  31. One thing which always bothered me.

    Romance languages all descend from Latin which was spoken two thousand years ago, so they are a first order family.

    But Latin itself was one of several Italic languages which are as close to Latin as Romance languages to each other.

    So Italic languages should be a second order family?

    And as I understand, Italic and Celtic families are quite close and related too, so we could talk about Italo-Celtic third order family.

    So it stands to logic that Indo-European is a fourth order family!

  32. marie-lucie says:

    SFR, you get the idea! I use “first order”, etc for convenience so as to distinguish families of closely related languages (like the Romance ones) from superfamilies such as Indo-European, which are groups of families, but “first”or other orders are not official terms used by all historical linguists. Beyond groups such as IE, others have been proposed, such as Nostratic and (even higher) Eurasiatic, but those have not been generally accepted. Of course, the more “orders” one recognizes, the more difficult it would be to set the criteria between the orders in one superfamily and another, especially since the attestation or alleged reconstruction of proto-languages (and often proto-proto-X or even proto-proto-proto-.. -X_ depending on the possible depth of reconstruction.

  33. marie-lucie says:

    (Sorry, I did not quite finish my last sentence)

    I meant that the approximate dates of the proto-language of one family (let’s say Proto-West-Germanic) might be quite different from the dates of the proto-language of another family even in the same “order”.

  34. “the closer you get to humans, the worse the science gets” (biologist proverb that holds within biology and more generally).

    Especially in sociology, although most people might consider this paradoxical or wrong-headed. Luhmann, the nicest guy you could ever hope to encounter (as one sees in recorded interviews with and lectures by him), still had a special brand of razor-sharp humor. An importunate interviewer once asked him if there was any subject that left him completely cold. He replied: ‘Ich lehne alle Einladungen ab, die mich veranlassen wollen, über den Menschen zu sprechen. Menschenbilder, so was Grausliches. Also der Mensch interessiert mich nicht, wenn ich das so hart sagen darf.’ [I always refuse invitations to speak on human nature. Discussions about the essence of man give me goosebumps. Mankind is not my thing, if I may put it thus harshly.]

  35. –I use “first order”, etc for convenience so as to distinguish families of closely related languages (like the Romance ones) from superfamilies such as Indo-European, which are groups of families, but “first”or other orders are not official terms used by all historical linguists.

    And this causes all the confusion, because no one is quite sure what kind of time depth is involved when we are discussing, say, Mayan languages.

    And closer to us, I don’t know how many linguists which don’t specialize in Altaic languages actually realize that, say, Turkic languages (without Chuvash) are a group of very closely related languages (partially mutually intelligible) and equivalent to Slavic (closer to each other than Romance languages).

    And that Mongolian (with possible exception of Daur) languages are actually either a single language or a dialect chain (essentially mutually intelligible and equivalent to East Slavic or Scandinavian languages without Icelandic-Faroese).

  36. So Italic languages should be a second order family?

    Romance is really a zeroth-order family, because we have actual records not indeed of the common ancestor but at any rate of a very close relative of it. The same is true of Hellenic, so much so that it is usually just called “Greek”, as if it were a single language, though by world standards Cappadocian, Pontic, Tsakonian, and Standard Greek are four separate languages. Italic then is a first-order family, because it is plain that Latin/Faliscan and the Sabellian (Osco-Umbrian) languages are as closely related as, say, the Slavic languages, though we have no recorded common ancestor for either group.

  37. —Romance is really a zeroth-order family, because we have actual records not indeed of the common ancestor but at any rate of a very close relative of it.

    Using similar logic, we could declare Indo-Aryan languages a zeroth-order family, since Vedic Sanskrit comes very close to being its common ancestor.

  38. David Marjanović says:

    Counterexample: string theory.

    Not sure if that counts; it may or may not be bad – the question is whether it’s science. I suppose it has the potential to become testable once there’s a collider that’s big enough.

  39. marie-lucie says:

    zeroth-order families

    There is a misunderstanding somewhere. By first, second, etc order families I mean a group of closely related known languages with a common ancestor, whether this ancestor is known or has been reconstructed. So Romance is a first-order family because the languages, although not mutually intelligible (otherwise they would be considered dialects of a single language), are close enough that their relationship is quite obvious even to untrained persons. Whether its ancestor is known (even if that ancestor was a Latin sociolect or a standard variety – the two were mutually intelligible) or reconstructed (like Proto-Germanic or Proto-Slavic) is irrelevant to the obvious relationship between the current languages descended from that ancestor. Now it is true that Latin had “sisters” within the Italic group, and from that point of view you could call Italic a first-order family. This is why the terms I have been using are unofficial shortcuts, not proposals for accurate classification.

  40. It is always difficult to say whether it’s a single language or several closely related languages.

    Chinese is considered a single language, but it is actually several mutually unintelligible languages (on the same order as Romance languages)

    But Mongolian is considered a family, but actually it is a single language (or was until reforms of 20th century)

  41. Chinese is considered a single language

    Not by linguists (i.e., the people who you would expect to know about these things).

  42. With Mongolian, some qualifications should be made. Besides Daur which is very likely a descendant of sister language of Common Mongolian, several rather peculiar Mongolic languages exist in Gansu province of China.

    These languages are without doubt Mongolian and of quite recent origin, but they are unintelligible to Mongolian speakers.

    The reason is very simple. The Gansu-Qinghai languages are spoken by strange Sunni Muslim population of sedentary peasants, very much unlike Mongols (racially, they resemble a mix of Han Chinese with Persians or Turks). These people are not Mongolians and likely never were. The circumstances under which these peoples switched to Mongolian are not very hard to guess.

    In a big Caribbean island, a population of apparent African origin currently speaks a Romance language resembling French. Quite similarly, ancestors of speakers of Mongour, Dongxiang, Kangjia, Bonan, etc were brought to Gansu and Qinghai in 13th century as slaves and ended up speaking pidgin Mongolian as a common language (but some related groups speak varieties of Uyghur Turkic or Mandarin Chinese).

  43. —Not by linguists

    Not by Western linguists.

    Just ask any Chinese linguist what he thinks on this subject….

  44. Nationalism is a sad business. Linguistics has similarly been distorted in the Balkans.

  45. Sadly, Scandinavian linguists are also victims of nationalism and maintain against overwhelming evidence to the contrary that Danish, Norwegian and Swedish are three different languages….

  46. The Scandinavian languages can be considered an edge case; the languages of China are not. There is no way an intellectually honest linguist could say that Mandarin, Shanghainese/Wu, and Cantonese/Yue are all the same language.

  47. Imagine for a moment that French, Spanish, Portuguese and Italian never developed separate literary languages and continued to use Latin for this purpose.

    Then we would have several dozen dialects divided into several large and mutually unintelligible dialect groups with the single literary language – Latin.

    That’s the situation with Chinese and Arabic now (and with some East and South Slavic languages before 17th century when they had a single literary language – Church Slavonic).

    These are edge cases too – plenty of honest linguists could argue that we are witnessing a break of common language into several different languages, but the process is not over yet due to existence of common literary language.

  48. I’m sorry, but that’s just nonsense. The fact that people who speak different languages are forced through circumstances of history and politics to use a common literary language has no bearing on the status of the languages they speak. I’ve heard the arguments before, believe me; you can put my refusal to be convinced down to pigheadedness if it makes you feel better, but I’m not going to be convinced, any more than you could convince me by brilliant rhetoric that the sun actually goes around the earth.

  49. For a long time the minor languages of the Caucasus were unwritten, and people had to use Arabic or Russian (depending on period and region) to express themselves in writing; does that mean they all spoke the same language?

  50. Here is an interesting situation.

    Imagine an Algerian speaking with Yemeni in Modern Standard Arabic.

    Or native of Shansi conversing with Fujian resident in Mandarin.

    Western linguists would classify these conversations as speakers of two different languages speaking to each other in a third language.

    But the speakers themselves would think that they are simply speaking in literary register of their common language.

    Which view is right?

    I don’t know. It depends on willingness to tolerate rather wide range of what constitutes a single language.

  51. J. W. Brewer says:

    Except the Sinitic languages/dialects/whatever lost their culturally-unifying single literary language (obviously historically known only to a minority of the population) about a century ago. It’s as if Napoleon suppressed Latin throughout Europe and made Italian and Spanish speakers write in French instead.

  52. Western scientists (incidentally, I don’t know why you keep using the qualifier “Western” — I’m pretty sure Yuen Ren Chao, say, didn’t buy into the “one language” myth) say the earth goes around the sun, but the natives with their native wisdom say the reverse. Which view is right? Who can say? The truth probably lies somewhere in between.

  53. —For a long time the minor languages of the Caucasus were unwritten, and people had to use Arabic or Russian (depending on period and region) to express themselves in writing; does that mean they all spoke the same language?

    Of course not.

    However, in a very similar case, several Turkic peoples of Central Asia used a common Turkic literary language based on Volga Tatar before 1920s.

    They understood it fine, it was intelligible to them (except for some Arabic and Persian learned vocabulary).

    But in the end Bolsheviks opted to create separate literary languages for Kazakh, Uzbek, Kirghiz and Turkmen (and Tatar and Bashkir as well).

    Again, another edge case. It could have easily remained a single language with Soviet educational policy, erasng “dialectal” differences over time.

  54. marie-lucie says:

    SFR: Your Romance example corresponds to the situation in the former Roman possessions in Western Europe a thousand years ago: there were no standard languages, the de facto official, supranational language was Latin, a language no one spoke natively any more. Even if there were already a few writings in the local languages, they were considered entertainment not serious literature: anything serious was written in Latin. Yet the languages, or dialect groups if you like, were already quite differentiated from Latin and from each other, clustering respectively within France (French and Occitan), Italy and the Iberian peninsula. The people were quite conscious of not speaking Latin, to which they were exposed every time they went to church, but which they could not understand. Learning Latin meant spending years in school, something few people did. Latin was a foreign language, even in Italy.

  55. And of course, Mongolian. Until 1920s, Buryats in Russia used Classical Mongolian as a literary language,

    Then in Russia, a choice was made to develop Buryat into a separate literary language, while Buryat in Mongolia and China continued to use Mongolian literary language.

    Buryat and Mongolian are essentially mutually intelligible despite grammar and pronunciation differences.

    Mongolian linguists had a very difficult problem with this. They had to explain why Buryat in Russia is a separate language, but Buryat in Mongolia is just a Mongolian dialect.

    For political reasons, they could not argue that Soviet decision to make Buryat a separate language was wrong (as linguists in Inner Mongolia did).

  56. —marie-lucie

    If modern educational system and television existed back then, everyone would have learned Latin in no time!

  57. To clarify my point.

    I am arguing that language/dialect issue is relative and depends on what the people themselves think. Of course, there are some linguistic limits to how far it can be stretched – you can’t make Kurdish a dialect of Turkish, for example.

    But closely related languages with a single literary standard can be claimed with some justification to be a single language.

    And vice versa, mutually intelligible languages with different literary standards can be regarded separate languages.

    I am perfectly fine with both positions.

  58. Yuen Ren Chao, say, didn’t buy into the “one language” myth

    He certainly didn’t. But he did devise General Chinese as a common written form for the Sinitic languages to replace Classical Chinese. It comes in two flavors, one of which is a diaphonemic romanized orthography making all the distinctions of Middle Chinese (much like the Regularized Inglish I have occasionally discussed hereabouts, only more drastic), except for those which have been merged away essentially everywhere (like vain-vein in English). For example, his own name 趙元任 is written Dhyao Qiuan-Remm, and pronounced Zhào Yuánrèn in Mandarin, Jiuh Yùhn-Yahm in (Yale) Cantonese, Tiō Gôan-Jīm in Taiwanese, and even Jō Gan Nin in Japanese and Triệu Nguyên Nhẩm in Vietnamese. Note that this is a tonal spelling, like his better-known Gwoyeu Romatzyh (which is for Mandarin only), and like GR it is rather more complicated than absolutely necessary (in my opinion) in the name of brevity.

    The other flavor is a convention for using exactly 2082 traditional hanzi, one for each possible syllable of the romanization. 80% of these are already in use with unique meanings (90-95% in running text), so someone who knows Traditional Chinese only has to learn which specific hanzi to use for the 20% of cases where more than one hanzi is traditionally used, but for someone who does not, it reduces the burden of both learning and use substantially. It essentially extends the phonetic-loan principle already in use for some hanzi to the whole of the system: as YRC said, if an ancient writes a hanzi with the same sound as the correct hanzi, it is a phonetic loan; if a modern child does it, it is a mistake!

  59. @SFReader: Great! Now, do speakers of different Sinitic languages learn Modern Standard Mandarin in no time?

  60. SFReader, do you have a sense as to how different Moghol (of Afghanistan) is from the rest of Mongolian?

  61. A very nice discussion of the ambiguities in term Papuan (even as a non-genetic term) is in this article by Mark Donohue.

  62. “Imagine an Algerian speaking with Yemeni in Modern Standard Arabic.”

    Diglossia.

  63. These Chinese, Arabic, and Balkan cases are why Peter Constable introduced into the ISO standards the concept of macrolanguages, which are groups of languages that are treated as a single language for some but not all purposes. Thus Arabic is a macrolanguage including Standard Arabic and 29 colloquials, and Chinese is a macrolanguage including Standard Chinese and 13 other Sinitic languages, but not Dungan, even though Dungan is on pure descriptive grounds actually inside the Mandarin dialect continuum. There are more than 60 macrolanguages currently recognized by ISO worldwide, of which the largest is Zapotec (58 languages). Serbo-Croat is considered a macrolanguage (but this is a place where ISO is arguably wrong even on its own terms), as are Norwegian, Quechua, Guarani, Buryat, and many other cases we have discussed.

  64. I can buy the macrolanguage idea, but I doubt those in the one-language camp would.

  65. Now, do speakers of different Sinitic languages learn Modern Standard Mandarin in no time?

    Nobody knows, because heard through a Chinese filter, that question amounts to “Do speakers of Chinese learn Chinese in no time?” There is nowhere you can go in the Sinosphere today where you don’t hear Mandarin from age five onwards. In older times, it is reported that you could learn a different topolect in a matter of months rather than years, but that would probably only amount to passive and partial understanding. Also, since a topolect boundary may be a language or a dialect boundary, there would naturally be easier and harder cases.

  66. David Marjanović says:

    Mongolian linguists had a very difficult problem with this. They had to explain why Buryat in Russia is a separate language, but Buryat in Mongolia is just a Mongolian dialect.

    Oh, that should be easy: if it’s spoken on the Dutch side of the border, it’s a Dutch dialect, and if it’s spoken on the German side of the border, it’s a German dialect! ^_^

    (A whole series of isoglosses runs at 90° to the border.)

    Note that this is a tonal spelling, like his better-known Gwoyeu Romatzyh (which is for Mandarin only), and like GR it is rather more complicated than absolutely necessary (in my opinion) in the name of brevity.

    The idea was to leave the most common tone unmarked – and that’s a different one for each syllable. It greatly reduces the amount of required letters per text, but puts a somewhat English-like burden on the learner.

  67. From what I remember reading about Moghol, it had mostly Dari vocabulary and was spoken by older generation only as a secret language. I think it is already extinct now.

    Unlike the Gansu languages, it was spoken by real Mongols, at least. But heavy contact with Dari made it very distinct from Mongolian.

  68. By the way, it just occured to me that when circa 100BC-100AD, various Italic and Celtic populations in Italy, Gaul and Iberian peninsula were switching to Latin, they could have easily regarded it just as another (more developed) variety of their own language.

    Proto-Italo-Celtic was spoken sometime in third or even second millenium BC, so in 100 BC, common ancestor of Gaulish and Latin was spoken less than 2000 years ago, similar to distance between current Romance languages and Latin.

  69. —if it’s spoken on the Dutch side of the border, it’s a Dutch dialect, and if it’s spoken on the German side of the border, it’s a German dialect! ^_^

    Yes, similar approach was tried and it was suggested with some linguistic justification that southernmost Buryat dialects (Tsongol and Sartuul) are actually dialects of Khalkha Mongolian.

    But unfortunately Tsongol dialect is spoken on both sides of the border, so the situation again became unbearable – how can one dialect be a dialect of both languages at the same time?

  70. Re: ISO macrolanguages.

    Idea is good, but unfortunately this list suffers from same lack of proportion.

    To say that Buryat is a macrolanguage and Chinese is a macrolanguage, seems rather strange.

    If we used objective mutual intelligibility criteria, Buryat macrolanguage would be equivalent to a group of Mandarin dialects and certainly not to the Chinese as a whole!

  71. About Mongolian:

    Inner Mongolia and Xinjiang have a lot of different Mongolian dialects. When people from the west and the east are thrown into contact, they initially have a lot of difficulty understanding each other. With time this improves greatly.

    Mongolians from Mongolia find even ‘standard’ Inner Mongolian speech, which is close to the standard in Mongolia apart from pronunciation differences, hard to understand. One person told me that Inner Mongolian in movies (historical movies) was only 50% intelligible, although I suspect this was a bit of an exaggeration. Mongolian from the east is almost completely unintelligible to Mongolians. (I heard one person from the east complain that Mongolians did not understand her at all when she said /xədi: ʤɔ:s/ ‘how much (money)?’ on a trip to Mongolia. The problem in this case is (1) Mongolians do not normally use хэдий зоос /xedi: ʣɔ:s/ to say ‘How much?’ and (2) the differences in pronunciation — /ə/ for /e/, /ʤ/ for /ʣ/ — would have made it more difficult to figure out the words she was using.)

    With familiarity many of these difficulties are lessened. Mongolians have little opportunity (or inclination) to listen to Inner Mongolian dialects, with the result that they have difficulty understanding them when they do encounter them.

    Buryat has a number of linguistic features which are different from mainstream Mongolian, including syntax (notably, if I remember rightly, postposed pronouns), morphology (final ‘n’ is retained, unlike Mongolian where it only appears in certain inflected forms and is now prone to be inserted unetymologically, a little like English intrusive ‘r’), pronunciation, and vocabulary. Taken together, this means that Buryat as written in Cyrillic looks rather different from Mongolian written in Cyrillic. It forms a typical case where language standardisation based on political reasons has indeed resulted in two ‘different’ languages.

  72. Classical Mongolian was as different from Khalkha Mongolian as Buryat (including grammar).

    So I suppose before 1940, in order to become literate, Mongolians in Mongolia were forced to learn what amounted to foreign language 😉

  73. Re: inability of Mongolians to understand different dialects

    I think this is relevant to Khalkha Mongolians from central Mongolia who are not very exposed to different varieties of Mongolian.

    Khalkha Mongolians from the western provinces understand various Oirat Mongolian dialects just fine.

    * and some even understand Kazakh

    Similarily, Khalkha Mongolians in northern provinces have no problem with understanding or even speaking Buryat. (they sing songs in Buryat all the time)

  74. Well, as a person who is trying to master the Classical Mongolian script (but not the Classical language), I would say that there is indeed an element of mastering a foreign language!

  75. I pity Inner Mongolian schoolchildren who have to learn in school both Classical Mongolian and Mandarin Chinese!

  76. Well, different, not really foreign in the sense of associated with another polity.

  77. I pity Inner Mongolian schoolchildren who have to learn in school both Classical Mongolian and Mandarin Chinese!

    Both tend to suffer. I believe that there is a continuing decline in Mongolian standards in Inner Mongolia (loss of language richness). At the same time, people who attend Mongolian schools do not always have a fully adequate level of Chinese for finding jobs in China. More and more Inner Mongolians now go to Chinese schools (partly a result of government policy, partly a result of parental choice) and are losing their language.

    Most of the students at the school I attend in Beijing are Mongols who can speak the language to some extent but can’t write it (because they attended Chinese school), or Mongols who can speak no Mongolian and want to reclaim some of their heritage.

  78. Stefan Holm says:

    A dilemma in maintaining languages are the children supposed to maintain them. The aunt of a friend of mine moved to USA already in the sixties. There she tried to speak Swedish with her children, e.g. calling them in for dinner through the window. -Stop using that monkey language, they replied (possibly mocked by their playmates).

    The same complaint I have heard from the seventies until today among lots of immigrant co-workers in Sweden. Despite deliberate attempts to pass their native tongues on to their sons and daughters the latter often oppose to it even at home. -You know Swedish, then use it! they say.

    So in spite of their parents and of a generous policy by the authorities in supporting education in immigrant languages, those it concerns, the subjects who are thought to benefit from it, seems to be the least interested.

    The one and only exception are the Finns, who came here mainly during the fifties and sixties. They often instead stubbornly refused to learn Swedish. In fact our closest neighbours are the single immigrant group understanding and speaking absolutely least of Swedish. But that has political reasons, Swedes historically being the upper class in Finland and the immigrants being not. (It’s a growing problem today, when this generation is growing old and in need of institutional care-taking – and can’t communicate).

    Maybe a ‘critical mass’ is needed to preserve a tongue? I don’t know if any studies have been made about Spanish in southern US or Arabic in Marseille and Paris, whether they are upheld mainly through continued immigration or also passed on to the younger generation. In any case the situation in inner Mongolia ought to basically be same.

  79. David Marjanović says:

    But unfortunately Tsongol dialect is spoken on both sides of the border, so the situation again became unbearable – how can one dialect be a dialect of both languages at the same time?

    Hasn’t stopped anyone in the Netherlands and Germany; but perhaps the isoglosses are more bundled in the Mongolian case, so it’s easier to identify distinct dialects?

  80. A dilemma in maintaining languages are the children supposed to maintain them. The aunt of a friend of mine moved to USA already in the sixties. There she tried to speak Swedish with her children, e.g. calling them in for dinner through the window. -Stop using that monkey language, they replied (possibly mocked by their playmates).

    The same complaint I have heard from the seventies until today among lots of immigrant co-workers in Sweden. Despite deliberate attempts to pass their native tongues on to their sons and daughters the latter often oppose to it even at home. -You know Swedish, then use it! they say.

    My mother was Norwegian-American and grew up in a heavily Norsk town (Roland, Iowa) where pretty much the entire adult generation spoke Norwegian and wanted the kids to learn it. My mother’s parents offered her and her brothers a penny for every Norwegian word they learned (and this was back in the 1920s and ’30s, when a penny was actual money!), and yet they refused; they only wanted to speak English.

    Yesterday I had coffee with a guy who was originally from India but had married a Mexican woman (whom he met in Australia when they were fellow grad students in biology — what a world!) and is now living in Mexico. They’re trying to bring their daughter up bilingual, but when he talks to her in English, she says (in Spanish) “Dad, we’re in Mexico, speak Spanish!” I fear there’s not much hope for her learning Kannada (his native language).

  81. And yet some people do succeed. We had Germans across the street where the kids spoke German with their parents, no problems that I could see.

  82. Stefan Holm says:

    I was only trying to illustrate that All animals are equal but some animals are more equal than others. The social context makes all the difference. There was a time, when in Swedish schoolyards Saami children were even physically punished for using their native tounge.

    Today this is a collectively felt shame in my country. It certainly is connected to the starting point of this thread, prehistory as a word with Scandinavian origin. The Saga and/or propaganda litterature of the Icelanders, Saxo Grammaticus, Tacitus, Jordanes, Adam of Bremen, the poet behind Beowulf and many others served as a basis for a false, ridicolous but in the end dangerous mythology about the proud and ancient history of the ‘Germanics’.

    It may have started in Scandinavia but our kinsmen weren’t late to follow. Queen Victoria herself claimed to be a direct descendent of viking king Harald Bluetooth. In Germany was raised monuments like http://en.wikipedia.org/wiki/Hermannsdenkmal in praise of Arminius, the defeater of Roman legions. Even when it comes to eugenics, ‘scull measuring’ or in plain English, racism, we the Scandinavians are far from innocent. Finally these myths formed a basis for the ideology of the Dritte Reich.

    Today we are, hopefully, trying to make up for all this. Germany and Sweden allow for more international refugees than all other EU countries (well, an ulterior motive to maintain a surplus number in the labour force can’t be excluded). We try in every way to promote the languages we once oppressed. What happens? The former victims are not interested.

    Vanity of vanities, saith the Preacher, vanity of vanities; all is vanity. What profit hath a man of all his labour which he taketh under the sun?

  83. Today we are, hopefully, trying to make up for all this.

    Friends recently visited Greenland, that tiny speck of Denmark off Baffin Island. They came back with stories of how the locals (i.e., mostly Inuit) are looked down on as wretched primitives by the sophisticates of Copenhagen.

  84. marie-lucie says:

    City sophisticates everywhere have a tendency to regard everyone outside the big city as wretched primitives!

  85. parents offered her and her brothers a penny for every Norwegian word they learned (and this was back in the 1920s and ’30s, when a penny was actual money!), and yet they refused; they only wanted to speak English.

    wretched primitives

    of course this is how languages end up being lost, through stigma morphing into self-hatred. But being ashamed of one’s heritage has a whole slew of additional negative consequences, apart from loosing ancestral languages. That’s why, when we teach Russian to its “legacy speaker” children, it’s almost as if we care less about acquiring proper age-appropriate vocabulary and grammar than about learning that the culture of the old country is cool and fun and great – about learning to appreciate one’s roots. True, the parents usually care the most about the reading / writing levels of their children, and that’s what they expect, but we preach the holistic approach to the union of language skill, love, and pride, and that’s what keeps the kids’ engaged. Paying a dime a learned word would seem a whole lot more straightforward approach, but it doesn’t quite work.

  86. marie-lucie says:

    a dime for each word learned

    It sounds like the parents had not been speaking to the children in Norwegian as a matter of course but tried to teach the language to them as they got a little older, one word at a time. Of course that method does not work.

    Most people don’t realize that children don’t start speaking out of the blue at a certain age but start using words they recognize and understand from hearing them from parents and other caregivers in situations which are important to the child.

  87. I don’t think of non-New Yorkers as wretched primitives: after all, das Ausland is where all the (im)migrants come from, and where would we be without them?

  88. Have just noticed the confusion over “Papua” in this thread three days ago. It doesn’t help that the term is a political minefield, and is vague/contested as an ethnic and linguistic classification, but:
    – nobody from Maluku (Moluccas) would self-identify as “Papuan”, now or at any time in the past (any more than, say, someone from England or Sweden would self-identify as German because they speak a Germanic language).
    – the sultanates of Ternate and Tidore never occupied any part of New Guinea, although in the 16th century the heads of several communities in (for want of a better word) “Papuan” islands off the western New Guinea coast accepted the nominal overlordship of those sultanates. A third sultanate, Bacan, claimed overlordship of more islands than Ternate and Tidore, but not of anywhere on the New Guinea mainland AFAIK. One of Tidore’s sultans, possibly trying to impress a Portuguese interlocutor, claimed dominion over all New Guinea, but this was fantasy. The Java-based Majapahit kingdom had also claimed New Guinea as part of its territory in the 14th century, to similar effect.

    On prehistory: Baliem in Indonesian Papua is one of the seven (I think) locations in the world where plants and animals were domesticated at some point before 5000 BCE, leading to the development of intensive agriculture. This evolved in Baliem largely without external influences until the mid-20th century: surely that makes Baliem nirvana for some aspects of the study of prehistory.

  89. Raising American kids in Austria, I can see that there will be no issues with any of them, even the youngest who moved when he was three, maintaining their English. It has clearly become a prestige language, and even other Austrian children will often try to speak English to them. As will grown-ups at times, which I find not particularly helpful for the kids’ German. A lot of children of immigrants from non-English speaking countries – Italy, ex-Yugoslav lands, Russia – also prefer speaking English with my kids to German if they have the chance, at least in the Gymnasien my kids attend,

  90. marie-lucie says:

    JC: I don’t think of non-New Yorkers as wretched primitives:

    Of course not, but I am sure you are not the average New Yorker. I was thinking “hillbillies” stereotypes, rather than immigrants.

  91. Oh, that should be easy: if it’s spoken on the Dutch side of the border, it’s a Dutch dialect, and if it’s spoken on the German side of the border, it’s a German dialect!

    I’ve heard the same can be said of Paraguayan : Brazilian :: Guarani : Tupi.

    By the way, it just occured to me that when circa 100BC-100AD, various Italic and Celtic populations in Italy, Gaul and Iberian peninsula were switching to Latin, they could have easily regarded it just as another (more developed) variety of their own language.

    The Romans, for their part, were more impressed by resemblances between Celtic and Greek. Caesar sometimes renders Celtic words with Celtic endings when they coincide with Greek, e.g. druides with short e.

  92. By the way, it just occured to me that when circa 100BC-100AD, various Italic and Celtic populations in Italy, Gaul and Iberian peninsula were switching to Latin, they could have easily regarded it just as another (more developed) variety of their own language.

    I’d never thought of that, and it’s an intriguing idea.

    Caesar sometimes renders Celtic words with Celtic endings when they coincide with Greek, e.g. druides with short e.

    How do you know whether the e is long or short?

  93. marie-lucie says:

    Is Italo-Celtic a generally accepted group nowadays?

    When I was teaching the history of French I bought a book on Gaulish. It seemed to me that the Gauls would not have found Latin (at least the basics) terribly difficult to learn.

    The Romans, for their part, were more impressed by resemblances between Celtic and Greek.

    Which Romans are we talking about? Is Caesar considered an authority on language? Untrained persons often form opinions on the basis of a few details that struck them, so for instance if Gaulish had the suffix -os where Latin had -us but Greek had -os, presto, Gaulish is very similar to Greek!

  94. Is Italo-Celtic a generally accepted group nowadays?

    No. As far as I know, the only non-controversial subgroups of IE are IE-without-Anatolian, IE-without-Anatolian-or-Tocharian, and Indo-Iranian. In my extremely non-professional opinion, Italo-Celtic, Balto-Slavic, and Helleno-Armenian are also correct, but I certainly can’t say they are well-established.

  95. Ferdinand Lot was first to propose this idea

    “Ces notions permettent, si insuffisantes soient-elles, de comprendre que les Gaulois n’aient pas éprouvé de graves difficultés à apprendre le latin. Est-ce à dire qu’ils n’ont eu à opérer que ces substitutions partielles qui permettent aux paysans français usant encore du patois local d’acquérir sans grande peine le français littéraire et officiel ? Non!
    Il y a eu changement de langue plutôt qu’adaptation. Mais ce changement n’a pas fait plus de difficultés au Gaulois qu’à l’Ombrien, au Samnite, au Campanien, c’est-à-dire aux Italiques, dont le parler avait fini par diverger considérablement du parler latin. En tout cas, un Gaulois a eu infiniment moins d’obstacles à vaincre pour apprendre le latin que le Toscan, tout voisin de Rome, mais dont le parler, l’étrusque, différait des langues indo-européennes.”

  96. How do you know whether the e is long or short?

    I thought of that when I was typing it. I know I’ve seen it rendered that way.

  97. Which Romans are we talking about?

    Exactly the kind of untrained persons you invoke, and for that sort of reason. IIRC the idea comes down to Geoffrey of Monmouth, who invokes it anent the Trojan origin of the Britons.

  98. Stefan Holm says:

    if it’s spoken on the Dutch side of the border, it’s a Dutch dialect, and if it’s spoken on the German side of the border, it’s a German dialect!

    I suppose it is an internationally known story although I’ve only read it in a Swedish version. But anyway: Start a journey from the southernmost part of Italy, go all the way north to the French border, continue along the Mediterranean coast to Spain, travel south passing Catalonia, follow the coast line south and then west, turn northwards through Portugal and stick to the Atlantic coast passing France again and end up in Belgium!

    Along this long journey you won’t find one single village, in which the language is not perfectly understood by the people in the previous or next village. Still – a Wallonian will understand next to nothing communicating with a Calabrian.

    There simply is no scientific definition of what is a dialect vs a language – and can never be. In the Swedish town Strömstad close to the Norwegian border, where hordes of Norwegians come every day to spend their oil money on cheap liqour and food in the Banana Kingdom of Sweden the inhabitants understand Norwegian better than they understand, say, Scanian or Gutnic (which they of course also understand but with more attention required).

  99. Along this long journey you won’t find one single village, in which the language is not perfectly understood by the people in the previous or next village. Still – a Wallonian will understand next to nothing communicating with a Calabrian.

    It’s a frequent paradox in geographical divides, where two sharply distinct localities can sometimes be connected by a line along which the changes are merely gradual, and one might never encounter a sharp transition if traveling from point A to point B along such a path.

    For example, two neighboring phytocenoses might have a sharp dividing line between them, yet there may be a line along which you don’t experience any abrupt transitions between two types of plant communities. They are internally variable and mosaic in structure (tessera / parcellae of the small-scale variation abound), and if you pick a certain path, the changes are never more than small-scale and gradual.

    It doesn’t mean that the concept of phytocenosis is all fiction – any more so than a concepts of liquid vs. gas phases are fictions (even though one can convert liquid into gas without any abrupt phase transitions, by carefully “choosing a path” through higher pressures and temperatures around the critical point). It’s just, there may be situations where the two states continuously morph into one another, but typically there is an abrupt transition, and typically there are great many important differences between the stuff on both sides of this divide.

  100. Day and night are tolerably distinct even if no one can say exactly where the boundary is (which is a different point from the existence of boundaryless paths: here all the paths are boundaryless).

  101. David Marjanović says:

    Is Italo-Celtic a generally accepted group nowadays?

    There are big names like Don Ringe behind it, and I’m not aware of any argument against it. Little research has been done on it, though; the list of proposed common innovations is rather short so far.

    The only people who don’t accept Balto-Slavic nowadays, however, are traumatized Lithuanian nationalists who apparently remember some weird Soviet propagandical misuse of the results of linguistics and want to put as much Deep Time between themselves and the Russian language. Proto-Balto-Slavic is very well understood; for comparison, I don’t know of any attempt to reconstruct Proto-Italo-Celtic, and it is much less clear whether West Baltic and East Baltic share a common ancestor that they don’t share with Slavic.

  102. Stefan Holm: actually, for Romance the story is partly wrong. Even a couple of centuries ago, before the spread of the standard languages, most of the border between Spain and Portugal was also a sharp language border: to the East of Portugal/Portuguese you would have found Spanish, with no intermediate varieties. This is due to the fact that over Central and Southern Spain and Portugal Spanish and Portuguese are languages transplanted from the North: as a prestige variety each spread South, and thus both spread side by side. Originally, in their original homelands in the Northern part of the Peninsula there existed (and still exists, in some rural areas) a dialect continuum.

    Outside of the mountainous far North the same is true of Catalan: to its West and South you found Spanish, and there do not exist any intermediate varieties whose genetic classification might be problematic. The cause is the same: over much of its territory Catalan is a transplanted language which, like Spanish further West, eliminated the pre-Catalan Romance varieties and Arabic originally spoken. As a result we’ve very sharp language borders separating Spanish from Portuguese and Spanish from Catalan over most of the Iberian peninsula, and these sharp borders have been there for centuries.

    There are a few other parts of Romance-speaking Europe where similar such sharp divides exist because of various instances of pre-modern linguistic expansion (Gallurese and Sassarese versus Logudorese in Northern Sardinia, Poitou-Saintongeais versus Gascon in South-Western France…)

    Closer to your own neck of the woods, there is a sharp linguistic border between North Germanic and West Germanic in Schleswig-Holstein: Southern Jutish is North Germanic, Northern Low Saxon is West Germanic, and no intermediate variety exists. A Slavic scholar of my acquaintance has suggested that the area was Slavicized in the Early Middle Ages, and that this Slavic language was gradually lost to Germanic, with North Germanic spreading from North to South and West Germanic spreading from South to North…and thus, both branches of Germanic “met” on formerly Slavic-speaking territory, each variety of Germanic by then being sharply distinct from the other.

  103. Proto-Balto-Slavic is very well understood

    Thanks, that looks like a fun paper, and props to Carrasquer Vidal for illustrating it with Repin’s Reply of the Zaporozhian Cossacks!

  104. most of the border between Spain and Portugal was also a sharp language border

    Well, it was and is sharp enough, but the coincidence with the political boundary is I suppose more recent, a product of mass education in the standard languages. Historically, the various isoglosses dividing Scots from local English (notably the /ʊ/ > /ʌ/ change, which appears in southern English and Scots but not in the intervening varieties) ran north of the political border at its eastern end, but nowadays they hew far more closely to the border.

  105. David Marjanović says:

    A Slavic scholar of my acquaintance has suggested that the area was Slavicized in the Early Middle Ages, and that this Slavic language was gradually lost to Germanic, with North Germanic spreading from North to South and West Germanic spreading from South to North…and thus, both branches of Germanic “met” on formerly Slavic-speaking territory, each variety of Germanic by then being sharply distinct from the other.

    Huh. I thought the “textbook explanation” was that the Angles and especially the Jutes might have spoken something intermediary, but they left, and the vacuum was gradually filled from both sides. But in any case it is apparently uncontroversial that the Slavic expansion occasionally reached the North Sea…

    and props to Carrasquer Vidal for illustrating it with Repin’s Reply of the Zaporozhian Cossacks!

    You will truly appreciate that if you read the Wikipedia version first!

  106. David Marjanović says:

    Forgot to mention that Frisian (with all its internal diversity) has a sharp border to Low German and Danish; probably to Dutch as well, but I don’t really know.

  107. John Cowan: no, the Spanish-Portuguese political border was a linguistic border (except to the North) long before mass literacy and/or schooling had any tangible impact: most work on Iberian dialectology used illiterate rural informants (of which there were plenty on both sides of the border), and what was spoken in Portugal was indubitably Portuguese, what was spoken in Spain was just as indubitably Spanish (leaving aside a few enclaves). The same is true of the Northern Sardinian/ South-Western French cases I gave above.

  108. John Cowan: I’m curious how the distribution of the foot-strut split – present in southern England and Scotland, but absent in northern England – came to be. It’s not surprising to see an area of innovation with conservatism on both sides (e.g. akanye in most of Russia, with okanye in both northern Russia and Ukraine), but the situation with foot-strut is the reverse.

  109. >Etienne
    Do you know our « Fala »? We understand better Galician than Portuguese.
    http://en.wikipedia.org/wiki/Fala_language

  110. A Slavic scholar of my acquaintance has suggested that the area was Slavicized in the Early Middle Ages, and that this Slavic language was gradually lost to Germanic, with North Germanic spreading from North to South and West Germanic spreading from South to North…and thus, both branches of Germanic “met” on formerly Slavic-speaking territory, each variety of Germanic by then being sharply distinct from the other.

    Related to this, I learned recently that the German Ostsiedlung of the Middle Ages proceeded in general directly west to east, so that, e.g. the Plattdeutsch of Hamburg and that of East Prussia were easily mutually comprehensible, and as you went from north to south you encountered the same rough dialect borders and comprehensibility issues as you would have closer to France. (Am on a filtered web connection so I can’t easily Google up the place I read this, but e.g. this paragraph in Wikipedia supports it.)

  111. Stefan Holm says:

    Controversial Finnish phonetician Kalevi Wiik some ten years ago launched the idea that large parts of Europe once was inhabited by proto-Uralic or proto-Fenno-Ugric speaking hunters and gatherers. When the Indoeuropeans arrived with their superior economic system (agriculture) their dialect was gradually adapted by the FU speakers. This was, says Wiik, a process taking place in stages during a very long period. In that process the FU:s couldn’t help applying their own phonology on pre-Gmc and thus give rise to the distinctive features of Germanic.

    He mentions twenty such features, among them expiratory stress on first syllable (typical for FU languages) and the umlaut (reflecting the ‘vowel harmony’ in all FU languages). The stages in his time scale are:

    1. The LBK culture in the Oder-Vistula area in 5500-4200 BC.
    2. The Ertebølle-Ellerbeck culture in the northern Germany-
    Denmark-Skåne area in 4200-3500 BC.
    3. The TRB culture (the Funnel-Beaker culture) in the inland area of southern Sweden in 3500-2800 BC.
    4. The Battle-Axe culture in central Scandinavia and the coastal regions of inland Sweden in 2800-2300 BC.
    5. During the Copper-Stone Age in central Scandinavia in 2300-1800BC.
    6. During the Bronze Age (the Scandinavian Bronze Culture) to the north of the last mentioned area in 1800-800 BC.

    His idea is that the further north the IE speakers got the stronger the FU phonetic element grew. That’s why the umlaut gets more significant in the north (e.g. a- and-u-umlaut). Being an elite language it’s not strange that Gmc borrowed very few words from FU but that the phonological impact was heavy.

    Combined with a possible Balto-Slavic intrusion in northern Poland/Germany at some stage (and separating North from West Gmc) this at least doesn’t sound all lunatic to me. But everyone shall know that Wiik is highly controversial – and it hasn’t helped him, that he occasionally has given support in public to political ideas in the direction of a ‘Greater Finland’.

    His twenty points by the way are:

    1. Initial stress or the fixing of stress on the first syllable of the word (Akzentverschiebung/Akzentwandel).
    2. Change in the quality of stress from a relatively tonal type to a more expiratory type (more dynamic) and from a
    relatively even to a more centralized type.
    3. Foot isochrony.
    4. Grimm’s Law (die erste/germanische Lautverschiebung).
    4a. Aspiration of plosives.
    4b. Tone of plosives.
    5. Verner’s Law (das Vernersche Gesetz).
    6. Merger of the palatal and velar places of articulation in
    plosives.
    7. Palatalization of consonants.
    7a. i-umlaut.
    7b. Diphthong ei shortened to í.
    7c. Vowel change e > i.
    7d. Diphthongization of some monophthongs.
    7e. a- and u-umlaut.
    8. Dissolution of the syllabic resonants into an u followed by the resonant in question.
    9. Merging of ∂ and a into an a.
    10. Merging of o and a into an a.
    11. Merging of ó and á into an ó.
    12. Apocope of a and e.

  112. Lazar: It’s complicated.

    The first thing to note is that there is no FOOT-STRUT split in Scots (including Scottish English); the exception that preserved /ʊ/ after labial consonants, as in put, did not operate there, which is why the word putt, borrowed from a specific meaning of the Scots pronunciation of put, has /ʌ/. So the change is not really the same in Scotland as in Southern England (which for this purpose includes Wales and all overseas areas except parts of Northern Ireland).

    The second thing to note is the spelling of foot. As it indicates, all the FOOT words not resulting from the the /ʊ/ > /ʌ/ sound change had long o in Middle English, and would have been expected to have /u/ today, except that some of them were shortened (and laxed) to /ʊ/ on a word-by-word basisafter original /ʊ/ > /ʌ/. Indeed, there are only six uncontested minimal pairs in Southern accents for the /u/ : /ʊ/ contrast, all of which can be accounted for by other changes: fool/full, pool/pull, who’d/hood, suit/soot, kook/cook, Jewry/jury. The historic pronunciation of woof ‘crosswise threads’ has /u/ and contrasts with woof ‘dog noise’, but it is not uncommon to use /wʊf/ for both of the last (as I do). The loss of yod after coronals also gives nuke/nook in most North American accents; the dictionaries say that Luke/look is not a minimal pair even in RP, but I don’t know why the former isn’t /ljuk/. A few accents merge only before /l/, thus eliminating the first two pairs.

    Now this irregular shortening, crucially, did not happen in any non-Southern accent, resulting in what I have called the “wedding-cake” structure of accents in Britain: come down (with ME short and long u respectively) is /kʌm dun/ in Scotland, /kum dun/ in the North of England (with merger), /kʊm dun/ in the Midlands, and /kʌm daʊn/ in the South (because the Great Vowel Shift did not change /uː/ except in the South). So in Scots there is not only no foot-strut split, there is no foot-strut contrast either, because foot has /u/. Indeed, Scots proper has no [ʊ] at all.

  113. Stefan Holm says:

    One should add that there are phenomena which Wiik does not mention at all. One is the very common occurrence of initial consonant clusters all over Germania (impossible in Fenno-Ugric). Screw, stream, spring. How did they survive (or even got extended, *sraum- → *straum-)? Another is the time scale. The Gallehus horns (200 AD?) has still un-umlauted ‘hlewagastiR’ (‘holtijaR is a more problematic issue).

  114. marie-lucie says:

    JC: the dictionaries say that Luke/look is not a minimal pair

    What??? I have never heard them pronounced the same.

  115. David Marjanović says:

    as you went from north to south you encountered the same rough dialect borders and comprehensibility issues as you would have closer to France

    Pretty much.

    The Wikipedia article mentions “Lichtervelde in Flanders” out of the blue. And indeed, there’s a Lichterfelde right here in Berlin, a few bus stations away. 🙂

    He mentions twenty such features, among them expiratory stress on first syllable (typical for FU languages) and the umlaut (reflecting the ‘vowel harmony’ in all FU languages).

    I wouldn’t speak of Wiik in the present tense; he’s very much history of science.

    Yes, the stress on the first syllable is shared with Uralic, but also with Italic and apparently Celtic.

    As processes, umlaut and vowel harmony are sort of opposites – umlaut consists of analogical changes to the first syllable based on the second one, vowel harmony goes in the other direction. Vowel harmony can probably evolve from umlaut if the triggers of umlaut aren’t lost, but it’s hard to imagine the reverse. – And note that the only umlaut process that may have operated before the basal split of Proto-Germanic is the change from e…i to i…i; all others are restricted to Northwest Germanic and completely lacking in Gothic.

    Being an elite language it’s not strange that Gmc borrowed very few words from FU but that the phonological impact was heavy.

    I’d agree on the words; but I’d also expect that phonology would not be borrowed in such a situation – syntax would.

    3. Foot isochrony.

    Is that what’s called “a stress-timed language”? Because the Germanic languages I’ve heard enough of are stress-timed, but Finnish and Hungarian are as syllable-timed as Spanish or Italian.

    6. Merger of the palatal and velar places of articulation in plosives.

    The merger of palatalized and plain velars hardly needs any explanation. The plain ones were rather rare, so the distinction didn’t carry a high functional load.

    7. Palatalization of consonants.

    …Of vowels, right?

    7b. Diphthong ei shortened to í.

    [ei̯] > [eː] > [iː] is too common to need an explanation. And what’s Uralic about this change anyway?

    7d. Diphthongization of some monophthongs.

    Again, what’s Uralic about that?

    8. Dissolution of the syllabic resonants into an u followed by the resonant in question.

    Syllabic resonants are rare enough on a global scale that this doesn’t need an explanation.

    9. Merging of ∂ and a into an a.

    “ə”? What “ə” – syllabic laryngeals?

    10. Merging of o and a into an a.
    11. Merging of ó and á into an ó.
    12. Apocope of a and e.

    All of these made Pre-Germanic less like Uralic than it was before.

    What??? I have never heard them pronounced the same.

    Me neither, for what that’s worth. But Luke seems to be a very rare name in England, so maybe I’ve only heard Americans say it?

    (‘holtijaR is a more problematic issue)

    For i-umlaut, yes. It does, however, show Northwest Germanic a-umlaut (u…a > o…a), before which short /o/ didn’t even exist.

  116. Sorry, I meant to say that they are a minimal pair even in BrE according to the dictionaries. I would expect this to be true only in North America, where /ju/ > /u/ after coronals. But apparently /lju/ > /lu/ elsewhere by a separate change, although /dju tju/ have either remained the same or become /dʒu tʃu/, and /nju/ has remained unchanged.

  117. Eli Nelson says:

    The change /lj/ > /l/ (and its counterpart /rj/ > /r/, which is shared among all dialects of modern English as far as I know) is pretty old, actually. Fowler discusses it in his Dictionary of Modern English Usage in 1926; he also mentions that the traditional /sj/ in words such as “Susan” was going out of fashion in his time in favor of a simple /s/.

  118. marie-lucie says:

    Luke/look

    Even if you say lu not lju, the u is not the same, any more than in the pair boot/foot (is there a dialect where these two words rhyme?).

  119. the u is not the same

    Precisely my point: Luke/look IS a minimal pair for /u/ ~ /ʊ/. I’m surprised that this is true everywhere, that’s all. And yes, in Scots and Scottish English, boot and foot do rhyme (though their realization is fronted from cardinal [u]): there is no /ʊ/ or even [ʊ] at all.

  120. There are some speakers in the north of England (mostly in Lancashire, I think) who pronounce -ook words with /uː/, making look homophonous with Luke.

    On the yod-dropping thing: I have an English-German dictionary from the 60s that gives pretty conservative RP pronunciations, and as best I can tell it attests /lj/ only in lure, allure and lurid (something to do with /r/?), with the other lu- words like lukewarm, luminous and lucrative all using /luː/. Maybe /lj/ died out in these words at the same time as the /rj/ in rude.

  121. David Marjanović says:

    There are people who pronounce evolution with [lj].

  122. Surely /rju/ > /ru/ is older; for one thing, it is universal across all accents. Indeed, anglophones have a lot of trouble saying words like “Ryukyu” without an epenthetic vowel for that reason.

  123. I don’t think I’d distinguish ‘Jewry’ and ‘jury’ (although I think ‘Jewry’ is a reading pronunciation for me — it’s not something that I ever discussed with anyone when I was growing up).

    I think that ‘hoof’ has both pronunciations. While they are not in distinctive contrast semantically, I think they are perceived as separate pronunciations.

  124. There are a number of words that can have either /u/ or /ʊ/: hoof, roof, soot. This is partly dialectal, partly idiolectal. I was mocked as a child for saying /ruf/, but when I told my wife about this, she said “/rʊf/ is what a dog says!”

    The point about Jewry is that it doesn’t have coda /r/ (or a centering diphthong in non-rhotic varieties), and so belongs to the GOOSE lexical set. “God Rest You Merry, Gentlemen” we sing “In Beth-le-hem in Jew-ry”, where “Jew” carries two notes but is not a diphthong.

    Per contra, jury has ambisyllabic /r/ and belongs to the CURE lexical set. This is a relatively sparse set, and its words are undergoing lexically specific shifts, especially in non-rhotic varieties: poor and friends often winds up in the FORCE lexical set (homonymous with pore), whereas other words become disyllabic. So it is certainly not the best example of /u/ ~ /ʊ/, and I should have left it out.

    Wikipedia implies that /lju/ > /lu/ is about the same age as /sju/ > /su/, /zju/ > /zu/, and /θju/ > /θu/, mid-20C or so, and so younger than the general American /ju/ > /u/ after all coronals. But /Clju/ > /Clu/ (where C = consonant) is much older: no accent still uses /blju/ for blue.

  125. Stephen Bruce says:

    And in Norfolk, apparently, the yod has been dropped from all ”u” vowels: “bootiful” for “beautiful,” etc.

    In Russian, the name Bruce is transliterated as Брюс instead of Брус; this could be merely based on the spelling, but it might also reflect a time when it was pronounced as /brjus/.

  126. Forgot to mention that Frisian (with all its internal diversity) has a sharp border to Low German and Danish
    With Northern Frisian that’s due to the fact that it’s a transplant as well, the Northern Frisians settled the area only in the Middle Ages. East Frisia was isolated by bogs, creating a sharp linguistic border (now East Frisian is only spoken in the small Saterland area outside of Ostfriesland proper, while in Ostfriesland it was replaced by Low German, which, like Low German everywhere in Northern, is currently being replaced by Standard German).

  127. David Marjanović says:

    the small Saterland area

    which is itself an island in a bog.

  128. Stefan Holm says:

    Low German everywhere in Northern, is currently being replaced by Standard German

    So what we learned in school as an example will soon not be understood in Berlin: eene jut jebratne Jans ist eene jute Jabe Jottes?

  129. Richard Burton had a nice Welsh contrast between /ɪu/ in blew and /uː/ in blue. Compare:

    5:20 The fair breeze bl/ɪu/, the white foam fl/ɪu/

    and

    6:34 Burnt green, and bl/uː/ and white.

    The Rime of the Ancient Mariner

  130. Well found!

  131. So what we learned in school as an example will soon not be understood in Berlin: eene jut jebratne Jans ist eene jute Jabe Jottes?

    Berlinerisch doesn’t really count as Low German, it’s more of a colloquial Standard German with some Low German Features (e.g. /j/ for /g/, /e:/ for old /ai/, /o:/ for old /au/, merger of dative and accusative in pronouns – some of these features are also shared by other Eastern German regional variants). It’s close enough to Standard German to be understood anyway. When I last lived in Berlin, in the late 90s, it seemed to be still in use among “native” Berliners. OTOH, while many of my age-peers spoke Platt in Ostfriesland when I grew up (I didn’t, as my family had moved there and didn’t speak Platt), I don’t hear young people speaking it much nowadays. Most you’ll hear is some Kind of “Missingsch”, i.e. a mixture between Platt and colloquial Standard German.

  132. And in Norfolk, apparently, the yod has been dropped from all ”u” vowels: “bootiful” for “beautiful,” etc.

    The yod has dropped, true, but that doesn’t actually mean beautiful is pronounced with the local vowel of boot; it just sounds that way to outsiders. Unfortunately, Peter Trudgill’s papers that give a proper phonetic explanation seem to be offline now.

  133. “There are some speakers in the north of England (mostly in Lancashire, I think) who pronounce -ook words with /uː/, making look homophonous with Luke.”

    That would be particularly true in Liverpool (or “Liverpewul”)

  134. The yod has dropped, true, but that doesn’t actually mean beautiful is pronounced with the local vowel of boot; it just sounds that way to outsiders.

    It seems to me that in the “broad” Norwich accent beauty and boot do have the same vowel — a strongly fronted high rounded /ʉː/ (and moot = mute). It’s boat that is likely to confuse outsiders. The reflex of ME ǭ is a fully back, slightly diphthongal /uː/ in most words of the GOAT set (including goat itself), but with unpredictable lexical exceptions like home, road, post, in which the vowel has been shortened to /ʊ/ (reminiscent of “New England short o“). Most words spelt with ow, ou, however, contain a reflex of an old vowel+glide sequence, which is a diphtong in Norfolk. The starting point is a rather open central vowel close to “mainstream” /ʌ/.

    In a nutshell: knows /nʌʊz/ ≠ nose /nuːz/, and boat /buːt/ ≠ boot, beaut(y) /bʉːt/.

    There are further complications, too. In some varieties of Norwich English (and elsewhere in East Anglia, especially in Suffolk) boat joins the HOME set with /ʊ/, and many speakers have a fronter variety of /uː/, leading to a near-merger with /ʉː/, though apparently a small contrast is normally maintained.

    If Peter Trudgill visits this blog, he would of course be the most competent person to comment. He is a native speaker of Norwich English, and proud of it.

  135. Okay, I found Trudgill’s paper that refers to bootiful, but I can’t get access to the critical part. It’s “Dedialectalisation and Norfolk dialect orthography” in I. Tavitsainen, G. Melchers & P. Pahta (eds.) Writing in Nonstandard English. Amsterdam: Benjamins, 1999, p. 323- 330. I can see it from a mangled Internet Archive copy of a probably-illicit copy of a possibly earlier version, however, and the key point Trudgill makes is that oo in Norfolk dialect writing represents the non-dialect phoneme /u/, whereas ew, eu, ue etc. represent /ʉ/. Because there are no yods, this distinction has been repurposed, and so the appropriate way to spell beautiful is bewtiful, not bootiful.

  136. David Marjanović says:

    So what we learned in school as an example will soon not be understood in Berlin: eene jut jebratne Jans ist eene jute Jabe Jottes?

    Understood, yes – but the people I’ve heard don’t speak like that anymore themselves. What I’m observing is that the phonological features of the dialects become lexicalized, remaining only in the most common morphemes. The prefix ge- is regularly [jə], the neuter adjective ending -es is often [ɛt], there are people who pronounce gut with [j] at least sometimes, and stressed ein- gets [eː], but you’re not likely to hear the complete sentence as you wrote it. (As a mere article, eine is just [nə] or thereabouts.)

    This has unexpected consequences. The usual dialect maps say Berlin is north of the line where final /k/ became /xː/, so you’d expect ich and auch to show up as [ɪk] and [oːk]. [ɪk] is indeed quite common; but auch is consistently [oːχ].

    The t in ist, BTW, was already gone in Old Saxon.

    Richard Burton had a nice Welsh contrast between /ɪu/ in blew and /uː/ in blue. Compare:

    Perhaps interestingly, this [ɪ̯uː] is quite different from the [ɪʊ̯] some Americans (like Obama) have, which sounds a lot like the Mandarin [ɪʊ̯] in the syllables liu and niu.

Trackbacks

  1. […] Hat considers the origins of the term […]

  2. […] How prehistory  – the idea and the word – developed. […]

Speak Your Mind

*