Elfdalian Runes.

We discussed Elfdalian back in 2016; now Dmitry Pruss sends me Miranda Bryant’s Guardian report on it:

It is a distinct language that has survived against the odds for centuries in a tiny pocket of central Sweden, where just 2,500 people speak it today. And yet, despite bearing little resemblance to Swedish, Elfdalian is considered to be only a dialect of the country’s dominant language. Now researchers say they have uncovered groundbreaking information about the roots of Elfdalian that they hope could bolster its standing and help it acquire official recognition as a minority language.

Elfdalian is traditionally spoken in a small part of the region of Dalarna, known as Älvdalen in Swedish and Övdaln in Elfdalian. But using linguistic and archeological data, including runes, Elfdalian experts have tracked the language back to the last phase of ancient Nordic – spoken across Scandinavia between the sixth and eighth centuries. They believe it was imported to hunter-gatherers in the Swedish region of Dalarna from farmers based in the region of Uppland, which became an international base for trade, who started adopting the language. At the time, the hunter-gatherers of Dalarna spoke a language referred to by linguists as “paleo north Scandinavian”.

Yair Sapir, the co-author of a new book on Elfdalian grammar, the first to be published in English, said: “There is research that compares the distance between Elfdalian vocabulary and it shows the distance is as large [between Swedish and Elfdalian] as between Swedish and Icelandic. So there is higher mutual intelligibility between speakers of Swedish, Norwegian and Danish than between Swedish and Elfdalian.”

Until around 1400, as a trade and transit area, the region was influenced linguistically and culturally from Norway and other parts of Sweden. But when the Kalmar Union was established and trade patterns dramatically changed, innovations in the language suddenly stopped. It was not until about 1900, with the arrival of schools, industrialisation and urbanisation, bringing with it a strong Swedish influence, that the language started to change again. This, in effect, said Sapir, made it “a medieval language that survived up to modern times”.

Before then there were multiple highly specific dialects that varied between villages and sometimes even within villages. “People did not move so much, there was not so much mobility and the units were quite self-reliant. They didn’t need to have so much contact with the outside world.” While runes had became obsolete in most of Sweden as early as the 14th century, there is evidence of runes being used in Älvdalen as late as 1909, making it the last place in the world where they were used. […]

“The linguistic landscape has also changed in the last 20 years or so, you see many more signs in Elfdalian in Älvdalen. You can also see that the feelings of shame have been replaced with feelings of pride.”

But as the influence of Swedish on the language grows even stronger, weakening the structure of the language and replacing Elfdalian words, greater protection is needed. “Sometimes it’s difficult to know if a word is Swedish or Elfdalian because they are related to each other.” Bringing back some of the linguistic features of the pre-1900 version, known as Late Classical Elfdalian, is helping native speakers to reclaim the language and allow new speakers in, argue Sapir and his co-author Olof Lundgren in their book A Grammar of Elfdalian. But it would benefit even more from official recognition as a language, they write.

Thanks, Dmitry!

Comments

  1. The book, which came out a month ago, is here, and it’s Open Access, Free-As-In-Free-Beer.

  2. Excellent, thanks for finding that!

  3. Lars Mathiesen (he/him/his) says

    Didn’t I report on the Swedish officialdom’s stance here within the last year? Too late at night to find it again, but their stance is that anything North Germanic spoken in Sweden is ipso facto to be regarded as a dialect of Swedish (unless it’s another national language, or some such). Hey, if they can do it in Italy…

    I don’t think any politicians at the national level have cared so far but the ECHR might, especially given that new book. Gutnish is also said to have been a branch separate from the East Norse that became Danish and Swedish, but there may be too much mixing with Standard Swedish now to make a case. And Scanian is not a variety of Danish any more.

  4. Didn’t I report on the Swedish officialdom’s stance here within the last year? Too late at night to find it again, but their stance is that anything North Germanic spoken in Sweden is ipso facto to be regarded as a dialect of Swedish (unless it’s another national language, or some such).

    That does ring a bell.

  5. David Eddyshaw says

    Isn’t the eighth century rather late for hunter-gatherers in those parts?

  6. Trond Engen says

    David E.: Isn’t the eighth century rather late for hunter-gatherers in those parts?

    Maybe they are better understood as specialized hunters and trappers since they produced iron, kept some animals, and were integrated into the wider economy. Which seems to be part of the argument.

    I’ll read the book. Until then I’ll note that “paleo North Scandinavian” is an interesting construction. I wonder if there’s evidence of something that wasn’t Sami. Sami is thought to have spread from Southern Finland a few centuries earlier, so maybe they think Proto-Norse became the lingua franca in the southern parts of Scandinavia, Sami in the north.

  7. Trond Engen says

    It will also be interesting to see if there’s firm evidence on continuous runic literacy in Övdalen. However sympathetic I am to the idea, 1909 is so late that every living Swede would have had history classes at school and introductions to the runic alphabet had been published regularly in popular magazines for decades.

  8. Among other gems that UCLPress has published as Open Access is The First Hebrew Shakespeare Translations; the Preface to Ithiel the Cushite of Venice is perhaps relevant to the other thread.

  9. David Eddyshaw says

    paleo North Scandinavian

    Perhaps they mean this:

    https://en.wikipedia.org/wiki/Paleo-Laplandic_languages

    Who doesn’t love a hypothetical substrate?

  10. “Who Doesn’t Love a Hypothetical Substrate?” sounds like a Gilbert and Sullivan tune (cut from Princess Ida, but originally to have been performed by the men’s chorus).

  11. Who doesn’t love a hypothetical substrate?
    Who doesn’t love a basis they’ve never seen?
    If you can bend an iron-clad circular tub straight,
    Then you can create a language that’s never been!

  12. FWIW a short reddit thread is skeptical of the distinctiveness of Elfdalian. Not sure how many of them have actually read the grammar.

  13. David Eddyshaw says

    Seems I guessed right, though the supposed palaeo-whatever substrate is actually only invoked indirectly:

    (Page 11, fn 23):

     A Palaeo North Scandinavian substrate is not discernable in Elfdalian on the surface, although this topic has not been studied systematically. A Palaeo North Scandinavian substrate is, however, discernable in Proto-South Sámi. More than a thousand words in Sámi which are not shared with other Uralic languages and are believed to originate from Palaeo North Scandinavian (Aikio 2012; Piha 2018). A Sámi substrate or influence on Elfdalian, namely in the domains of morphology and syntax, is possible in the following domains: (a) clause-initial negation adverbial; (b) diphthongisation of Old Norse ó and ei into uo and ie, respectively; (c) pro drop in some verb declensions; (d) the highly productive compounding of adjectives with their head nouns (‘incorporated adjectives’); (e) a suffix in nouns for marking non-delimited use; (f) a rich system of adverbials and verbs conveying the speaker’s attitude; (g) the scarce use of the infinitive particle; (h) the scarce use of the subordinating conjunction at (a word borrowed into Elfdalian from Swedish), as well as (i) a separate verb for ‘to speak Elfdalian’. Likewise, expressions for cardinal points in Elfdalian are not absolute but stand in relation to the terrain, similar to the Sámi system

    [Hatters will immediately recognise that the incorporation of adjectives with their head nouns is simply an inheritance from Scandi-Congo; the construction remains regular in Oti-Volta.]

    One thing that I hadn’t realised is that the language described here is partly an artficial de-Swedified revival. The actual current spoken language is apparently much closer to Swedish. (Perhaps this explains some of the disagreements about the degree of distinctiveness of the language.)

  14. @Hat: Your first two lines have “Maria” (“How do you solve…”) from The Sound of Music going through my head.

  15. I smell a Broadway hit!

  16. Who is not enamoured of a substrate hypothetical,
    That answers each and every dialect anomaly
    With the ghosts of past inflections of a language quite theoretical
    Constructed from vague fragments out of quite a different family?

  17. Brilliant!

  18. Bengt Pamp’s book Svenska Dialekter (Natur och Kultur, 1978), which I bought on 20th October 1978 and have just managed to find, introduces the chapter on Dalarna with quite a long piece in Elfdalian, with a parallel translation into Swedish. He mentions that the area contains some of Sweden’s most unusual dialects, and that ‘in hardly any other district of Sweden has dialect fragmentation gone so far; dialects differ not just from parish to parish but from village to village, and in some cases one can even find different dialects within the confines of a single village’.

  19. David Eddyshaw says

    Brilliant indeed. I am honoured to have been the occasion of it.

    It occurs to me that two rather different questions are in danger of getting confused with one another in this business of whether Elfdalian is a distinct language from Swedish.

    One is whether the current language is really not mutually comprehensible with (other?) Swedish dialects. This is the usual can of worms, further complicated by the fact that the language described in this book is artificial, and has been deliberately “purified” of some “Swedish” elements. It also sounds rather as if the proponents of separatism are greatly underestimating the dialect variation within the boundaries of what is usually and uncontroversially regarded as Swedish.

    The other is whether Elfdalian is, or is not, part of the same primary branch of proto-Norse as Swedish historically. This is a quite different question, particularly in the context of Scandinavia. Gordon’s grammar (which is all I know about this, and is presumably very outmoded here) describes a primary division between Norwegian-Faroese-Icelandic and Danish-Swedish; this is of course quite different from how things now look after a millennium of mutual influences (or not) among the various languages.

    If Elfdalian really does show significant substrate effects from Sámi (which needs a lot better evidence than a simple list of similarities), this would actually undermine the argument that it was a “different language” from Swedish in this second sense, by creating differences which had nothing to do with historical branching.

  20. Jen in Edinburgh says

    As an outsider in the sense of Knowing Nothing About This, it seems to me that claims are regularly made that two or more distant things are actually related, and that these are generally nonsense.

    Are claims that two apparently close things are actually distant more likely to be true? Claims that two close things are actually close wouldn’t get in the papers, after all – but maybe it is like cryptic species and there’s work still to be done.

  21. David Eddyshaw says

    As an outsider in the sense of Knowing Nothing About This

    Me too.
    Basically, I’m just waiting for Trond to finish the book …

    Are claims that two apparently close things are actually distant more likely to be true?

    In Oti-Volta too (not that it is very relevant, though at least I actually do know something about that) a fair number of distant things thought to be different actually turn out to be close, but not so much the opposite. Though I think I have managed to show that Waama actually isn’t closely related to its Eastern Oti-Volta neighbours, as traditionally supposed, so it can happen. And there are a couple of cases of close things thought to be different turning out to be close after all …

    I suppose the moral is that commonly accepted opinions actually are pretty often quite right after all, on account of having become commonly accepted for plausible reasons in the first place.

  22. David Marjanović says

    The other is whether Elfdalian is, or is not, part of the same primary branch of proto-Norse as Swedish historically.

    There used to be a presentation here that answered a lot of questions. Alas, it’s *poof* gone. Why do people do that…

    Anyway, the answer seems to be no: the division into West Norse and East Norse remains accepted, but Elfdalian is outside that… but the question remains where the other Dalecarlian dialects go; apparently they remain woefully underresearched.

    Gutnish, BTW, may well have started out as not even Northwest Germanic, lacking the *u…a > *o shift entirely. It’s quite stunning.

  23. Trond Engen says

    I’m finally reading it.

    On a South Sami substrate (pp. 5-6) – very similar to David’s quote above:

    The lexicon, grammar and phonology of Elfdalian appears to be predominantly North Germanic. A non-Germanic language substrate has not been attested, and interference from non-Germanic sources appears to be negligible: some syntactic interference from the neighbouring Sámi languages might have taken place. Such interference may be explained by the long history of neighbourhood of the two speech communities, as well as the fact that some nomadic Sámis in Central and Northern Sweden have settled down permanently, thus becoming farmers and assimilating into local Swedish communities.9

    Footnote 9:

    Sámi influence might have affected the following features in Elfdalian: clause-initial negation adverbial, diphthongisation into uo and ie, pro drop in some verb conjugations, the highly productive compounding of adjectives with their head nouns (‘incorporated adjectives’), a suffix in nouns for marking non-delimited use, a rich system of adverbials and verbs conveying the speaker’s attitude, the scarce use of the infinitive particle, the scarce use of the subordinating conjunction at (a word borrowed into Elfdalian from Swedish), as well as a separate verb for ‘to speak Elfdalian’. Likewise, expressions for compass points in Elfdalian are not absolute, but stand in relation to the direction of the Sturövę (East Dal River, Swe. Österdalälven), in resemblance to the Sámi system south of the Northern Sámi area (see also Söder and Parkvall 2010). For more on compass points in Elfdalian, see Chapter 8. For more on possible Sámi influence on North Germanic and Swedish varieties, see also Rießler (2002).

    If these are South Sami substrate features, some of them have certainly set their mark much wider than in Övdalska,

    clause-initial negation adverbial

    A clause-initial negation adverb is used for emphasis, but in Ôvdalska it might rather be an archaism. The common Scand. negation adverbials ikke and inte have different origins, and their convergence as variant forms in the same syntactic position could be recent.

    the highly productive compounding of adjectives with their head nouns (‘incorporated adjectives’)

    This a general feature of Northern Scandinavian.

    expressions for compass points in Elfdalian are not absolute, but stand in relation to the direction of the Sturövę (East Dal River, Swe. Österdalälven), in resemblance to the Sámi system south of the Northern Sámi area (see also Söder and Parkvall 2010).

    I haven’t looked up the references, but no doubt north is upriver. This is shared with traditional Eastern Norwegian. It would be more of a surprise if it weren’t the same on the Swedish side.

    More later.

  24. Trond Engen says

    I like this (p. 16):

    Attempts to classify Dalecarlian linguistically face several challenges. The first is the difficulty of navigating between a phylogenetic model based on (implicitly standard) North Germanic languages on the one hand and North Germanic dialects on the other. These two dimensions are rather different, as standard languages, for example in the case of Scandinavia, often coincide with political or administrative borders, whereas dialect groups do not. For this reason, even the term ‘Dalecarlian’ may be misleading in this sense. Secondly, the origin of Dalecarlian was implicitly assumed from a phylogenetic analysis. For instance, Dalecarlian could be defined as a sub-branch of the Sveamål dialect group, which would imply that Dalecarlian branched off from the Sveamål group. However, due to the lack of written records from the early history of Dalecarlian and its neighbouring dialects, diachrony and synchrony can easily be mixed up with each other and challenge this type of assumption. The third challenge lies in the importance of written and/or standard norms, which do not always reflect the spoken language of the area where texts were produced. Fourthly, there is a lack of written records from Dalarna prior to the fourteenth century (when they are still scarce). Written records from Dalecarlian that are sufficiently long for substantial linguistic analysis are virtually non-existent prior to the seventeenth century. Fifthly, conclusions about Proto-Norse based on material from one area were sometimes generalised to the whole Proto-Norse-speaking area. Finally, classifying dialect groups located on a dialect continuum is a rather challenging task, and such a classification requires explicit methodology. As Gooskens (2020, 766) points out, the motivation for Bandle’s ([1973] 2011) selection of isoglosses for the classification of the North Germanic dialects is rather subjective and remains unclear.

    According to this hypothesis, Classical Dalecarlian constitutes a separate sub-branch of the Nordic languages within the North Scandinavian dialect continuum, taking into account the influence from both the west and north-west and from the east and south-east and the idiosyncratic features of Dalecarlian – strong conservatism on the one hand and strong innovation on the other. More systematic linguistic research is deemed necessary to establish a thesis on the position of Dalecarlian within the North Germanic language family.

    Due to influences from both west and east, an attempt to force Dalecarlian into the west/east dichotomy of the North Germanic languages appears demanding. A classification of the Nordic languages and dialects into insular and mainland Nordic and of the mainland Nordic languages into North (Sweden, Norway and Swedish Finland) on the one hand and South (Denmark) on the other, as proposed by Torp (2004), appears more appropriate for the classification of Dalecarlian.

    It should follow that there might not be that many unique features to be found, even if the language as a whole stands out. Footnote 34 (pp. 18-19):

    Some archaic features preserved in UpSD from Ancient Norse, differently from all or almost all North Germanic varieties, include the following: […] (h) in terms of syntax, Garbacz and Johannessen (2015, 17) mention several features that Elfdalian shares with Insular Nordic languages which appear to be archaic, such as embedded V-to-I movement (as in eð ir biln so an will int åvå ‘it is the car that he doesn’t want to have’) as well as present participle construction expressing modality (as in wattneð ir it drikkend ‘the water may not or should not be drunk’). Besides Icelandic, this construction (elsewhere called ‘gerundive’) with a similar use is also known from Sanskrit, Old Persian, Classical Greek and Latin;

    I can’t document this without more research than I’m able to put into it now, but the former construction wouldn’t sound off in some Swedish-adjacent Norwegian dialects.

    The latter construction is pretty common in Norwegian, also including the standard language, and with or without a final -s in the dialects. Vanne ække drekkans where I live. Dette er ikke noe blivende sted. “This is not a place to stay (or a place where we should stay)” is bog standard bordering on literary, maybe rather hær erekke blivans in dialect.

    (Not a major point. The book goes on to list shared features in footnotes 35-39, showing clearly that Elfdalian was not linguistically isolated.)

    Going on to cultural features, we get this:

    For instance, this border constitutes the southern limit of the transhumance system, that is, the seasonal movement of cattle (Swe. fäbodväsendet), which was practised until the middle of the twentieth century in Northern Dalarna (Dahl 2015, 12). This husbandry could be a relic from the hunter-gatherer culture.

    It most certainly is not. Seasonal transhumance is advanced animal husbandry. Also, it’s all over Norway. And the summer pastures were where you would meet people from different parishes or counties, renewing and rearranging social bonds.

    Geographical isolation was no more significant than in other inland regions in Scandinavia. We’re just a short trip across Lake Siljan from the proto-industrial powerhouse of Bergslagen, after all. What I think happened is that the inhabitants of Övdaln, because of their specialized economy, were more prone to travel than most. Trading iron and furs in all directions in the winter, they early became bilingual in Standard Swedish (and maybe Eastern Norwegian), leaving the dialect to develop along its own separate path from a very early stage – and eventually fragmentizing into hamlet or household varieties whose speakers were barely exposed to eachother.

    Evidence from the book:

    P. 25:

    As mentioned in Section 1.1, the split of UpSD into various dialects is reportedly the most radical in the whole North Germanic language area. Mutual intelligibility between inhabitants of the three parishes of Upper Siljan was not always easy and was at times virtually impossible when Elfdalian was the target language, as described by Björklund (1956). Among the vastly varied parish vernaculars spoken in Upper Dalarna, the dialect of Övdaln has an absolutely special position. Even to people in the closest neighbouring parishes, Elfdalian may at times appear both archaic and difficult to understand.

    P. 26:

    Travelling and trade were usually limited to the beginning and the end of the winter, when it was possible to slide horse carriages on the soft snow. Labour mobility to other parts of Sweden had been going on for some centuries. However, this mobility usually took the shape of groups of people working in mansions, which probably did not affect Elfdalian considerably (see also Heed 1978).

    .
    Pp. 26+28:

    As Swedish was generally conceived as the standard language that was fit for writing, most runic inscriptions from Övdaln are in Swedish, sometimes with a local Elfdalian influence. However, inscriptions, especially early ones, are in Elfdalian.

    The remarkably archaic dialects were fossilized under a stratum of Standard Swedish. Whenever Dalacarlians went out of the valley, or to church, or, eventually, to the next hamlet over, they normalized, while Elfdalian was the language of the home, of farmwork and hunting, and of storytelling.

  25. David Eddyshaw says

    Very interesting, Trond. Thanks!

  26. David Marjanović says

    Makes sense.

    Standard German can imitate the gerundive thing when it is attributive: das nicht zu trinkende Wasser “the water that is impossible to drink”. But to make this predicative, you have to replace the present participle with the infinitive: das Wasser ist nicht zu trinken – this absolutely does not work with trinkend.

    The infinitive construction can be nominalized: das Wasser ist nicht zum Trinken “the water isn’t for drinking”.

    I don’t know how that varies in dialects beyond shouting distance of mine. Mine 1) lacks present participles altogether* and 2) avoids zu + inf., so the nominalized version is used to express all the above.

    * …after they almost became an inchoative aspect, it seems. Even that is now lexicalized to between 2 and 5 verbs.

Speak Your Mind

*