English or Engelsk?

« previous post | next post »

A recent article in Science Daily has the headline `Linguist makes sensational claim: English is a Scandinavian language'. The claim in question is Jan Terje Faarlund's conclusion that `English is in reality a Scandinavian language' — that `Old English quite simply died out while Scandinavian survived, albeit strongly influenced of course by Old English.' The core of Faarlund's argument is that, in addition to many words that originally belonged to Norwegian and/or Danish, English has syntactic structures that are Scandinavian rather than West Germanic in origin. Specifically, Faarlund argues that `wherever English differs syntactically from the other Western Germanic languages — German, Dutch, Frisian — it has the same structure as the Scandinavian languages.' Faarlund then gives a few examples of syntactic parallelism between English and Scandinavian [that is, the Germanic languages of Scandinavia] and concludes that `the only reasonable explanation' for this parallelism `is that English is in fact a Scandinavian language, and a continuation of the Norwegian-Danish language which was used in England during the Middle Ages.'

Extraordinary claims require extraordinary evidence, as the saying goes. The evidence cited in the article is nowhere near extraordinary. Assuming that he is quoted accurately, there are some serious problems with Faarlund's claims.

First, some general comments about language contact. Faarlund says that `[e]ven though a massive number of new words are on their way into a language, it nevertheless retains its own grammar. This is almost a universal law….It is highly irregular to borrow the syntax and structure from one language and use it in another language.' He's mistaken in his belief that languages in contact can be counted on to retain their own grammar: there are hundreds of convincing examples of structural diffusion — including phonological, morphological, syntactic, semantic, and discourse features — in contact situations all over the world.

Here are a couple of examples. Probably the most famous case of all is Kupwar, a village in India in the border area between Indic languages in the north and Dravidian languages in the south. Morphosyntactic diffusion has been multidirectional in Kupwar, but the most extensive changes have affected the Kupwar variety of the Indic language Urdu, which has borrowed from the Dravidian language Kannada and from Marathi, the other Indic language spoken in the village. The changes include adoption of an inclusive/exclusive `we' distinction, subject-verb agreement rules in four different constructions, word order features, and about a dozen other features (details can be found in the 1971 Gumperz & Wilson article). Another striking case was reported by Andrei Malchukov in 2002: the Tungusic language Evenki has borrowed a volitional mood suffix and an entire set of personal endings from the Turkic language Yakut. It's worth noting that word order is the most frequently borrowed type of syntactic feature — a relevant point because two of Faarlund's examples of Scandinavian structure in English are word order features.

In short, there's a whole world of language contact phenomena out there. People borrow structures as well as words from their neighbors' languages, and people who give up their native language and shift to someone else's language make "learners' errors" that introduce foreign structural features into the target language.

Next, consider Faarlund's belief that the only reasonable explanation for parallelism between syntactic structures of Scandinavian languages and English is that English is a Scandinavian language: what about the possibility that English and Scandinavian languages independently developed some similar syntactic features? Parallel but independent innovations in closely-related languages are well known and reliably attested. The process is known as drift, and (apologies for oversimplifying slightly here) it results from structural imbalances that make particular bits of grammar hard to learn. English and the Germanic Scandinavian languages are all changed later forms of Proto-Germanic, and a thousand years ago (the relevant period when we're looking at intensive Norse-English contact) they were very closely related.

One of Faarlund's examples, for instance, is the split infinitive, which occurs in English and Scandinavian but not in West Germanic languages such as German, Dutch, and Frisian. But both English and Scandinavian have innovated an infinitive construction introduced by a particle (e.g. the to of the English infinitive phrase to go and the att of the Swedish infinitive phrase att regna `to rain'), while West Germanic languages other than English have no such thing — they have only an infinitive suffix on the verb, as in German Ich will gehen `I want to go'. There is no possible way to insert an adverb or any other word between a stem and a suffix; only in languages with at least a two-word infinitive phrase does the possibility of a split infinitive even arise, so it is trivially true that German, Dutch, and Frisian lack that possibility. It may well be that English and Scandinavian independently developed infinitive phrases introduced by a particle and then independently innovated split infinitives. (I don't know enough about the histories of these languages to know when the phrasal infinitives arose, so I don't know whether the innovation in English coincides with the period and location of intensive Norse-English contact.) Note too that the constructions are not completely parallel: Scandinavian languages retain an infinitive suffix, while English has lost its suffix entirely; and in Swedish, at least, the distribution of att is not uniform across either verbs or kinds of infinitive constructions.

Now, about old English-Norse contacts. In the following comments I'm relying primarily on the English case study that Terry Kaufman wrote for our co-authored 1988 book Language Contact, Creolization, and Genetic Linguistics — specifically his section `Norse influence on English' (pp. 275-306).

As Faarlund says, Viking chiefs and their descendants ruled the Danelaw for half a century (longer, in one area). The number of Norse settlers is estimated to have been in the hundreds (not thousands), so that they would have been far outnumbered by English speakers. How did the Norse and English speakers communicate (when they weren't killing each other)? The two languages wouldn't have been mutually intelligible at that period, but they would have been similar enough to make becoming bilingual easy. Or maybe they each spoke their own language and developed merely passive understanding of the other language. It is highly probable, at least, that the Norse had the prestige in the Danelaw — they were the rulers, after all. That's the easiest way to account for the hundreds of Norse loanwords in English.

After the period of Norse rule, when the former Danelaw was once again under English control, the available evidence indicates that Norse ceased to be spoken after just a few generations, about sixty years. One major piece of evidence for this conclusion is that all the objects with Norse inscriptions found in England were inscribed within sixty years after the end of Norse rule.

Faarlund comments on the striking differences between Old English and Middle English. One reason for these differences is that the vast majority of Old English materials were written in the West Saxon dialect, while most written Middle English was based (as is modern Standard English) on the dialect of London, a Midland dialect. These two dialects were quite divergent already in Old English times. The West Saxon kingdom was destroyed by the Normans in 1066, and with it went West Saxon as a written form; when writing in English picked up again, its center was London. It's quite true, though, that the dialect of London — in spite of being in an area that was not Norsified — later acquired a number of Norse features from people who immigrated from the former Danelaw.

Kaufman's survey of Norse influence on the English of the Danelaw focuses exclusively on phonological and morphological features. He counts 57 structural traits of Norse origin in Norsified English dialects, out of a total of at least 260 grammatical traits. That is, no more than 20% of the total set of comparable structural features of the most Norsified English dialects came from Norse. Even if Faarlund's percentage for the syntax turns out to be higher, his syntactic Norse features are unlikely to raise the overall percentage of Norse-origin structures to an unusually high level, compared to other instances of structural diffusion in intense contact situations. Moreover, 38 of the 57 Norse structural traits in those English dialects are (according to Kaufman) `mere phonological variants of what English had in the first place' — which makes them look like fashionable "accent" shifts rather than wholesale borrowings. The same is true of many of the Norse loanwords in English, among them sister, skirt, die, give, and guest.

Kaufman's conclusion: `The Norse influence on English was pervasive, in the sense that its results are found in all parts of the language; but it was not deep, except in the lexicon.'

Ironically, Faarlund's own scenario is a counterexample to his claim that grammar doesn't get borrowed: if, as he claims, English is a Scandinavian language, how did it get its numerous English structures? They would have to be borrowed from English, right? He emphasizes the borrowing into English of Norse grammatical morphemes; but the majority of English grammatical morphemes are native English forms, not from Norse. When he focuses on Scandinavian parallels to modern English syntactic structures, he says that these parallels exist `wherever English differs syntactically' from the other West Germanic languages; but this implies that there are syntactic structures in which English matches the rest of West Germanic — so again, if English is a Scandinavian language, it must have borrowed those English syntactic features. And English basic vocabulary items, in spite of all those Norse loanwords (and in spite of some French loanwords in the basic vocabulary) are mostly West Germanic in origin.

So it's English, not Engelsk. English, a West Germanic language whose closest relative is Frisian, with a substantial (but not an extreme) component of Norse-origin features in the lexicon and in all areas of the grammar. An interesting contact situation, with interesting results, but not out of the ordinary.


  1. RP said,

    December 4, 2012 @ 12:08 pm

    I am not clear whether it is true that German has no infinitive particle ("zu") (there are certainly sources that say it has – e.g. http://clasfaculty.ucdenver.edu/tphillips/grammar/syntax_sentence_field.pdf ) – although I guess the salient point is that it used in fewer constructions than the English and Swedish equivalents.

  2. Nik Berry said,

    December 4, 2012 @ 12:22 pm

    I seem to recall that English borrowed the use of 'do' to form questions from a Brythonic source.

  3. Neil said,

    December 4, 2012 @ 12:29 pm

    Irrespectively of whether certain contact phenomena are theoretically common or not, isn't the larger issue that the transition from Old English to Middle/Modern English is in any case well enough documented for there not to have been any real doubt in the first place?

  4. Brett said,

    December 4, 2012 @ 12:30 pm

    Wow. Of course German has a split infinitive! "To go" in German is "zu gehen." ("Zu" is also cognate to English "to" in other senses as well.) It is true that German has more modal verbs that don't use the two-word infinitive than does English, and there are other contexts in which "zu" can be dropped that do not have analogues in English, but basing an argument on the supposed lack of a split infinitive in German is strikingly absurd.

    This is the kind of mistake that might be made by somebody who only ever took one year of high school German and never saw more complicated verb constructions than, "Ich mag spielen." It is usual to refer to the verb form without the "zu" as the "infinitive" in discussions of German grammar, but that is an inconsistency of terminology, not a substantive difference in grammatical structure.

  5. D-AW said,

    December 4, 2012 @ 12:32 pm

    The Norwegians have been doing some related research recently which may shed some light:

  6. bulbul said,

    December 4, 2012 @ 12:36 pm

    the point is that German has an actual way of morphologically marking the infinitive (when you see an infinitive, you know it's an infinitive) and "zu" is therefore optional which is not the case with "to" or "att" or "å".

    My first reaction to this piece of news was "But haven't they readThomason and Kaufman?". I'm glad you decided to weigh in.

  7. Eirik Hektoen said,

    December 4, 2012 @ 12:42 pm

    I also have always been under the impression that German zu is an infinitive-marking particle just like English to, Norwegian/Danish å and Swedish att (this is based on School German lessons, not linguistics studies).

    In any case, the example used for showing that German does not have an infinitive particle, Ich will gehen 'I want to go', doesn't work, as the Scandinavian languages do not include the particle after modal auxiliary verbs either: No. jeg vil gå, etc. Come to think of it, nor does English, e.g. in I will go!

  8. bulbul said,

    December 4, 2012 @ 12:50 pm

    German has a suffix to turn a verb into an infinitive, English (et al.) doesn't. "Zu" often shows up with the infinitive, but it doesn't make an infinitive.

    "I will go", but "I need to go", "I want to go" etc. Additionally, there is "wants, needs" but not (at least in these structures) "*wills" – "He will go", not "He wills go". You are making an excellent point, but one that has to do with the nature of "will" in English (one could argue that it's a particle rather than a modal verb).

  9. Brett said,

    December 4, 2012 @ 12:53 pm

    @bulbul: While German has more way of distinguishing verb forms directly than English, it is certainly not possible to identify an infinitive as such without the accompanying "zu." Fuly half of the finite verb forms are generally identical to the infinitive.

    In the terminology of English, we refer to the "infinitive" as including the "to"; in the terminology of German, "infinitive" typically refers to the verb without the "zu." People are being confused by this difference in terminology, when there is no fundamental difference in the way the verb is structured (except that there are more situations in German than in English where the "zu" either can or must be dropped).

  10. Nelson said,

    December 4, 2012 @ 12:59 pm

    Old English had two forms of the infinitive, a shorter form made with the suffix -an (e.g. lufian 'to love'), and a so-called inflected infinitive with a longer suffix and the particle (e.g. tó lufienne 'to love'). There were some distributional differences, roughly similar to those in Modern German. Obviously the infinitive with won out, but there's no need to invoke Norse: the loss of inflectional endings would be enough reason to generalize the particle as a marker.

    I'm really curious to see if Faarlund can demonstrate some plausible contact effects on Norse syntax on Anglian English (preferably with the West Germanic points of comparison being Old Frisian, Old Saxon, and Old High German, not Modern German and Dutch). But his claims about Old English dying out sound more like something to grab headlines than a serious scholarly claim.

  11. Jonathon said,

    December 4, 2012 @ 1:09 pm

    @Neil: Exactly. It's true that there's something of a break between Old English and Middle English, but it's still not hard to see the transition from late Old English to early Middle English. And a quick glance at core vocabulary like personal pronouns and the be verb makes it obvious that the Middle English forms must have come from Old English, not Old Norse.

  12. Howard Oakley said,

    December 4, 2012 @ 1:10 pm

    I think the basic claim is more true of Norn – it was derived from Norse, and did die out, unlike Old English.


  13. bulbul said,

    December 4, 2012 @ 1:15 pm

    Fuly half of the finite verb forms are generally identical to the infinitive.
    Now if I recall correctly…
    Ich arbeite du arbeitest er arbeitet wir arbeiten ihr arbeitet sie arbeiten
    That's two out of six = 30% and that's just the present tense indicative, i.e. not counting imperative or subjunctive. So more like less than a third.

    in the terminology of German, "infinitive" typically refers to the verb without the "zu."
    My copy of "Schülerduden Grammatik" (2006 edition, p. 41) defines the infinitive as follows:

    "Der Infinitiv besteht bei allen Verben aus dem Stamm und
    der Endung -en oder -n"
    "For all verbs, the infitinive consists of the stem and the suffix -en or -h"

    It goes on to say (p. 56):

    "Der Infinitiv hat in manchen Konstruktionen die Partikel zu bei sich…"
    "In many structures, the infinitive is accompanied by the particle zu…"

    So to repeat, in German, there is no infinitive without "-(e)n", but there are infinitives with "zu", whereas in English, an infinitive is not marked morphologically and always shows up preceded by "to". A fundamental difference, wouldn't you say?
    Of course, that's assuming the validity of theuse of a Latin-based term like "infinitive" for a language like English.

  14. marie-lucie said,

    December 4, 2012 @ 1:15 pm

    Syntactic borrowing from English has been occurring at a fast pace in the French spoken in France, especially in written texts. I think that is because of hurried, unidiomatic translations from the English-language press appearing in the French press, and from there leaking into original French-language texts. It is often possible to guess the age of a writer, especially in the press: older writers use a more traditional French syntax, while younger ones are more influenced by English syntax. Many people, French and foreigners alike, mock the Académie for trying to ban English words, but syntactic borrowing is more insidious than lexical borrowing: an English word is usually quite easy to notice, but complex English sentence structure used with French words is more difficult to spot by readers who do not know English very well, and who simply think that "French has evolved" since they were young. It has indeed, but the unconscious contribution of English has played a large role in the modern evolution.

    The situation in Canada is quite different: younger people, better educated in French (and in English) than most of their elders, try to avoid anglicisms as much as possible.

  15. merijn said,

    December 4, 2012 @ 1:33 pm

    There is a paper by Joseph Emonds, who is probably the same as the Joseph Emmonds named in the article, here: http://conference.uaa.utb.cz/TheoriesAndPractice2010
    I think the discussion is not what the status of German zu/English to/Dutch te is, but if you can put something in between zu/to/te and the infinitive. It is possible in English (to boldly go) but as a native speaker I am quite certain that it isn't possible in Dutch, and probably not possible in German.

  16. Daniel von Brighoff said,

    December 4, 2012 @ 1:33 pm

    I seem to have understood the bit about "split infinitives" differently than most of the posters here. German does make common use of infinitive clauses introduced with the preposition zu. What it doesn't do, however, is allow anything to intervene between this preposition and the infinitive. Thus, the German equivalent to the listed examples:

    I promise to never do it again.
    Eg lovar å ikkje gjera det igjen.

    is this:

    Ich verspreche es nie wieder zu tun.

    As far as I know (native speakers please correct me if I have this wrong) the word order *Ich verspreche es zu nie wieder tun is never acceptable in any form of the language, no matter how colloquial.

  17. Jonathon said,

    December 4, 2012 @ 1:36 pm

    @bulbul: "In English, an infinitive is not marked morphologically and always shows up preceded by 'to'."

    The English infinitive isn't marked morphologically, but it isn't always preceded by 'to', as in I will go or I made him go.

  18. bulbul said,

    December 4, 2012 @ 1:52 pm


    true, my bad. But then again, that's the question – is it an infinitive?
    Huddleston and Pullum, for example, argue that "There is no form in the English verb paradigm called 'the infinitive'" and instead they speak of "non-finite clauses" and use the terms "plain form" and "to/for infinitival". In this context, they refer to "to" as a "special marker".
    And the larger point still stands – the German infinitive is an actual morphological form and no adverbs or what have you can come between the stem and the suffix.

  19. LDavidH said,

    December 4, 2012 @ 2:03 pm

    As a Swede, I quite liked the idea of Engelsk! :-) but I find it unlikely. I prefer to think of English as a category of its own… but obviously closer to Frisian – Dutch – Afrikaans than any Scandinavian language.

    Having said that, I find that I understand the King James Version of the Bible more easily than my English wife, because its syntax is more similar to Swedish ("say not" etc instead of "do not say" etc).

    And no, I didn't mean the KJV is easier to understand than my wife!

  20. bulbul said,

    December 4, 2012 @ 2:17 pm

    the issue is with the terminological confusion: to some people, "to" is an integral part of the infinitive, obviously not morphologically, but still in some way. Of course it makes sense to speak of infinitival clauses like you or H&P do, but then we have to decide whether English has an infinitive at all. German, btw, indubitably does.

    As for your question, one could check one of the German corpora like DWDS using a query like "zu $p=ADV $p=VVINF" (so zu + adverb + infinitive). Turns out that this way, you get a lot of hits for structures where "zu" modifies the adverb (meaning "too, excessively"), so you'd have to wade through them manually.

  21. Jonathon said,

    December 4, 2012 @ 2:22 pm

    @bulbul: I haven't read up on CGEL's argument, but it seems to me that it's mostly terminological. What I was taught to call a bare infinitive they call a plain form. But yes, you're right that the German infinitive (as a single inflected word) is not splittable.

    But Daniel von Brighoff raises a good point: German does have zu + infinitive combinations, and they're still not splittable. If we're comparing English infinitives with Scandinavian and German ones, I think this is the better argument. It still doesn't mean that English is really a Scandinavian language, of course.

  22. Brett said,

    December 4, 2012 @ 2:35 pm

    @bulbul: To get the half figure for the finite verb, I was counting seven declarative forms (ich, du, es, wir, ihr, sie, und Sie), of which three are identical to the infinitive. (Grammatical sources differ as to whether sie and Sie should really count as separate, but I'm used to counting them separately.) However, there are also three imperative forms, which are not generally the same as the declaratives (du, wir, und Sie); of these, the last two are identical to the infinitive (for normal verbs). So that gives five out of ten (or four out of nine) that are the same as the infinitive.

    Off the top of my head, I can think of one more non-finite form that is always the same as the infinitive (even for sein)–what I was taught to call the "gerund" form (although from reading Language Log, I now suspect that's probably not a good term for it).

  23. bulbul said,

    December 4, 2012 @ 2:45 pm

    so that gives us:
    – 6 forms in present indicative (there is no reason to count Sie as a separate one) – 2 identical to the infinitive.
    – 3 forms in the imperative (that includes Sie), of which one is identical to the infinitive, the rest are not (bleib – bleibt – blieben Sie, arbeite – arbeitet – arbeiten Sie).
    Three out of nine, 30%.

    the "gerund" form
    The grammar I cited calls them "nominalisierte oder substantivierte

  24. Andrew (not the same one) said,

    December 4, 2012 @ 2:49 pm

    Strunk (writing in 1918) says that 'bid' 'takes the infinitive without to'. (This, strikingly, is the only use of 'infinitive' in his – the original – Elements of Style.)

    I wonder if the use of 'infinitive' for the form with 'to' has become more widespread because of the frequent use of the term 'split infinitive'.

  25. gnaddrig said,

    December 4, 2012 @ 3:35 pm

    @ RP, Brett, bulbul: The "zu" does exist in German, but it is certainly not part of the infinitive. In German, the infinitive is identified by the suffix -en.

    "Zu" is used in a number of constructions together with the infinitive. In these cases it is not optional, it can't be dropped, and it doesn't mark the infinitive.

    When "zu" is used with the infinitive, one can't put anything between the zu and the infinitive, and this insertion is what "splits" the infinitive in English. German does definitely not have anything like the English split infinitive.

    A few examples for "zu" with infinitive:

    I don't have to goIch brauche nicht zu gehen, but Ich muss nicht gehen. In both cases, gehen is the infinitive, but brauchen + infinitive always requires "zu" while müssen + infinitive doesn't.
    He promised to helpEr hat versprochen, zu helfen
    He forgot to eatEr hat vergessen, zu essen

    There is sort of an inverted split infinitive in German, though. In certain cases, the "zu" is inserted between the prefix and the verb. Take an example with to look sth. up (in the dictionary)etwas (im Wörterbuch) nachschlagen:
    I promised to look it upIch habe versprochen, es nachzuschlagen

    And sometimes verbs are separated from their prefixes and stuff is put in between, but then a finite verb form is used, not the infinitive:
    I will look it up in the dictionaryIch schlage es im Wörterbuch nach. But both cases have nothing to do with the English split infinitive.

  26. RP said,

    December 4, 2012 @ 3:52 pm

    I personally do not see why the "to" is widely considered part of the infinitive in English. I think it is an unfortunate idea And obviously a lot of English infinitives don't have the "to". I think that Brett was right to say that counting "to" as an infinitival particle but not "zu" is just a terminological difference in convention.

    But the point that is relevant to the OP is that German does have a parallel to the English to+infinitive construction, and that German doesn't have split infinitives – which adds up to a point in Faarlund's favour (even though I suspect Thomason is right to reject his overall claim). The point that the German infinitive is morphologically marked ("-en") – while the English one is plain – is interesting but is not really relevant. The Swedish infinitive is morphologically marked ("-a"), and Faarlund's point was that in terms of split infinitives, English and Swedish have them in common, while German doesn't have them (but not because it doesn't have an equivalent to "to" – because it does; its "zu" is not used in as wide a range of constructions as "to" or "att", but again, this fact is in Faarlund's favour). I think I remember reading that splitting the infinitive is sometimes obligatory in Swedish, but this is a side-issue, albeit an interesting one.

  27. Lane said,

    December 4, 2012 @ 4:17 pm

    Daniel, you said what I would: as a 2L speaker of German, I'd say you can't put anything between particle zu and the verb. So if we agree with Duden that "zu gehen" is an infinitival construction, we can't split it. Oder?

  28. Oliver said,

    December 4, 2012 @ 4:19 pm

    But the infinitive in German can take more particles than "zu" and it can take an article. In these cases the infinitive must be split.

    Um kleine Kinder zu erschrecken, kaufe ich Feuerwerk.

  29. Lane said,

    December 4, 2012 @ 4:41 pm

    Hm. Good point, Oliver, but I don't think of the um as part of the infinitive, but part of a larger construction meaning "In order to… '

    Nice example, btw!

  30. Peter said,

    December 4, 2012 @ 5:30 pm


    There is no possible way to insert an adverb or any other word between a stem and a suffix; only in languages with at least a two-word infinitive phrase does the possibility of a split infinitive even arise

    In the period after the appearance of the “to”/“att” marker for infinitives but before the appearance of splitting, what makes us analyse the “to”/“att” as a separate word rather than simply as a prefix?

  31. Ran Ari-Gur said,

    December 4, 2012 @ 6:07 pm

    @Peter: Do you think there was such a period? I would assume that "to"-infinitives started out as two clearly-separate words (the preposition "to" plus the bare infinitive functioning as a nominal), and the greater tendency to combine them (e.g., the abandonment of all other prepositions besides "to", and the use of the "to"-infinitive itself as a nominal) coming later. No? So "to"-infinitives would have been "splittable" even before they existed, really.

  32. David Eddyshaw said,

    December 4, 2012 @ 6:11 pm

    @Nik Berry:

    "I seem to recall that English borrowed the use of 'do' to form questions from a Brythonic source."

    I think you're remembering John McWhorter's book on the history of English. He's very expert in his field, but his ideas on Brythonic influence on English seem very farfetched to me (who am not expert at all.)

    Welsh (for example) actually *doesn't* form questions with its "do" verb. Northern Welsh has a periphrastic *past* tense (only) with the "do" verb, which doesn't really match English at all, isn't found in South Welsh, and doesn't seem a very old feature. Welsh doesn't make a present tense with "do" + infinitive. Modern Welsh (but not the older language) extensively uses "be" + preposition + infinitive as a present, but even that isn't quite like English as it corresponds not only to the English "continuous present" but also to the simple present. So

    Wy'n canu = "I am singing" AND "I sing."
    (The old present tense "canaf" has metamorphosed into a future form.)

    As I understand it (not very far), JMcW's idea is that periphrastic constructions with "do" are very uncommon so it's hardly conceivable that Welsh and English both have them by accident. But the forms don't really match up well, and the timescales in Brythonic and English don't seem to work unless there were substantial bilingual communities speaking Brythonic and Middle English. Not only does this seem unsupported by much historical evidence, but it's hard to square with the fact that there are remarkably few Brythonic loanwords in English. Contrary to what I've seen written sometimes, Welsh is *not* like English syntactically; the whole group of Insular Celtic languages are pretty weird for Indoeuropean, with VSO word order, conjugated prepositions …

    I don't think periphrasis with "do" is all that rare cross-linguistically come to that, though there are plenty of LL readers who know much more about all this than me (John McWhorter for one …)

  33. J.W. Brewer said,

    December 4, 2012 @ 6:58 pm

    LDavidH: I suspect that the various now-archaic syntactic features of the King James Version that might seem congenial to a modern Swedish speaker would be at least as congenial to a modern German or Dutch speaker, so that doesn't get us very far on the question at hand, other than reminding us that English-specific innovation in more recent centuries is no less significant than more ancient divergences between West and North Germanic.

  34. Jim said,

    December 4, 2012 @ 7:03 pm

    The use of "do" to form negatives and questions may or may not reflect any influence or connection with Welsh, but it sure does parallel the situation in Irish. In modern Irish there is the tá/fuil – bhí/raibh binary, and this is relictual frro the situation in Old Irish where all verbs had this kind of split, and where the second form was used for negatives and interrogatives.

    English was a minority language in Ireland for centuries and there is quite a lot of lexical borrowing, words that no English speaker would ever guess were not native since they have bled back into Britain and burrowed so deep into the language. Bilingualism was such a threat to the persistence of English that the English community felt a need to pass a law to forbid English from learning Irish, the linguistic equivalent of a kosher law. It is not a stretch to think that some syntactic features were borrowed too.

    Something else. Faarlund may be restricting himself to too narrow a period of influence. The conventional wisdom is that English come to Britain in the form of Anglo-Saxon after Roman power collapsed. That is based on a reading of tacitus where he always refers ot Celts, all over the island. But in fact that tells us only about what the tribal elites that the Romans were interested in spoke. it tells us exaclty nothing abuot what the peasantry spoke. It doesn't even tell us if they spoke the same language.

    The whole southwestern coast of the island was held by "Belgae". What did they speak? We do know the Romans were not very clear themselves on whther the continental Belgae were Celts of Germans. What prevents them from being Germanic and speaking some Germanic language? It would absolutely not have to be ancestral to what the Germanic people in that region speak now. It would not have to have much to do with Dutch at all. There was a lot of disruption in those centuries.

    As for Britain, the contacts with Scandinavia were close and ancient. they go back to the bronze age. You can track a change in styles as it spreads from Denmark on finally to Ireland. There is no reason there could not have been Norse settlement in Britain all along.

    The Saxones were no more numerous in relation to the native population than the Norse were in the Danelaw. The gentic evidence shows only small islands of Frisian and coastal Saxon DNA. And by the way, there is nothing trivial about any feature of Frisian when you are discussing AS. The Saxons were no more a demic tsunami that swamped all and establsihed English than the Norse were. If they could establish, arguendo that the ancestor of Englsih was not already in place, then why would it be so impossible for Norse to have a similar effect?

  35. Ian Loveless said,

    December 5, 2012 @ 12:30 am

    @Jonathan: Looking at personal pronouns, Norse-derived "they, them, and their" supplanted the Old English versions, which is described as unusually rare and an example of the "profound impact" of Old Norse on English, according to the etymology of these words described in the American Heritage Dictionary.

  36. Lars Arvidsson said,

    December 5, 2012 @ 2:09 am

    What about the Jutes, one of the tribes that made up the English? They are supposed to have come from Jutland in present-day Denmark. What was their language like? We really have very few sources, so many scenarios are possible.
    Calling English a Scandinavian language is surely going too far, but obviously English is very much affected by Old Norse. For me as a Swede, that's interesting enough.

  37. Interested said,

    December 5, 2012 @ 5:40 am

    Getting away from the details of infinitives …. a lot of the vocabulary evidence they proffer is very bad indeed.

    For instance, on their list of supposedly-unborrowable everyday words are several where *both* cognate forms from Old English and Old Norse occur in Middle (and in some cases Modern) English. For instance, "raise" (OE > rear), "though" (thaw), "egg" (ey), "skirt" (shirt) – in all of these both the OE and ON forms were current well after 1100. E.g. it wasn't till Early Modern English that "ey" gave way to "egg"; "rear" and "shirt" are of course still around.

    So clearly these words don't count as evidence that Middle English as of circa 1100 is a descendant of ON, which is what they are arguing for – the presence of "egg" is evidence of ON descent to exactly the same extent that that the presence of "ey" is evidence of OE descent, that is, the evidence as observed is equally supportive of either case and thus of neither.

    Other points of their evidence contradict themselves. For instance, they (incorrectly) make a big deal of the fact that bits of the core grammar don't get borrowed. But earlier on the article says:

    >For example, all the lexical words in this sentence are Scandinavian: He took the knife and cut the steak. Only he, the and and come from Old English.<

    … which, if they were taking the "can't-borrow-core-grammar" argument seriously, would necessarily be a major stoke AGAINST their thesis since "he/the/and" are elements of the core grammar!

    Maybe – maybe – these inadequacies are the result of the fact that this theory is being put forward in the in-house magazine of the University of Oslo's press office ( http://www.apollon.uio.no/english/articles/2012/4-english-scandinavian.html ). Maybe there is a peer-reviewed article in the works which will present a non-dumbed-down version of the argument that will not suffer from these flaws.


  38. Nelson said,

    December 5, 2012 @ 6:00 am

    @Ian Loveless, those plural pronouns do indeed come from Norse, and are exactly what the AHD says. They are not really evidence for Middle English somehow being a descendent of a variety of Old Norse. Against those examples, I, he, it, you, me, we, and our are all unambiguously from Old English (a few, like thou or mine, could in theory be from either language). Faarlund's view also requires Anglo-Norse to have borrowed rather a lot of other morphology from English: a-stem plurals in -s, the verbal endings -eth and -est, the adverbial suffix -ly, the abstract suffix -ness, etc.

    It's kind of strange to see dream on Faarlund's list. The semantics agree with Norse (Old English dréam meant 'joy'), but the phonology is very distinctly English (the early Norse form of the word was *draumʀ. Even if we assume Norse influence (rather than an unattested Anglian *dréam 'dream'), this is hardly strong evidence of Norse vocabulary in English – it's just evidence of a close contact situation.

  39. Marion Crane said,

    December 5, 2012 @ 6:03 am

    J.W. Brewer: Yes, as a native Dutch speaker I have to say that archaic English syntactic strunctures do feel more like Dutch than English at times, but I agree that that doesn't really help the original argument. Likewise, one of my college professors advised us to read Chaucer as though it were Old Dutch rather than a form of English (and out loud, too, because the similarities with Dutch become even more apparent then), because English changed in so many ways from Middle English onwards, and to his credit that did work quite well. Easier than trying to translate it to Modern English and then to Dutch, at any rate.

    I would be curious to see if this works for Scandinavian speakers as well. But still, all that tells us, I feel, is that all our languages were once closely related and then evolved in different directions, some more so than others. Not like this is news, so I agree with whoever it was who already said that it sounds more like someone trying to get a good headline than actual useful or decent research.

    For the record, Dutch does have infinitives with a suffix, and does allow 'te' in combination with an infinitive in certain situations, but no split infinitives.

  40. Oliver said,

    December 5, 2012 @ 6:10 am

    The verb is also split from the article in German:

    Das kleine Kinder mit Feuerwerk Erschrecken wird mir zu laut.

    And I guess the "zu" also arose as a part of an "in order to" construction. This also explains why "zu" is not used with true auxillary verbs and dropped in the colloquial language when verbs, like "brauchen" develop into auxillary verbs.

  41. RP said,

    December 5, 2012 @ 6:31 am

    In Herbert, "The Germanic Languages", we read that:
    "All of the Germanic languages distinguish between 'bare' infinitives and prepositional infinitives – that is, infinitives preceded by prepositions. It is not clear whether this opposition can be reconstructed for the parent language, since the different dialects use different prepositions in the latter, and differ from each other as well with respect to whether the infinitive inflects for case after the preposition… The earlier West Germanic infinitives inflect for dative case after the preposition."

    This practice of inflecting for dative, he says, has since disappeared – but is still present "vestigially" in Frisian.

    The split infinitive, he says, is present in English and Swedish, but not in Danish.

  42. RP said,

    December 5, 2012 @ 6:32 am

    Sorry, I meant to add also that English and German have in common in the West Germanic practice of using "to" or a cognate, whereas Scandinavian languages use a cognate of "at". This is a point against Faarlund.

  43. Alon Lischinsky said,

    December 5, 2012 @ 7:06 am

    @Lars: the Jutes were an Ingvaeonic people, and their language therefore technically West Germanic.

    There is no reason to suppose it was especially influenced by North Germanic languages prior to settlement in Britain; the Kentish dialect of Old English is supposed to be a reflex of Jutish influence, but the differences with Anglian and West Saxon are minor (such as voiced initial /v/ for Anglian and Saxon /f/, as in vader, or the preservation of initial /ɬ/ long after it had merged into /l/ in the remaining dialects).

  44. dainichi said,

    December 5, 2012 @ 7:07 am

    I'm not enough of a linguist to judge whether Faarlund's claim is valid, but as a native Danish speaker, I've always found English grammar to be at least as similar to Danish grammar as to German grammar, with the quoted examples being prime examples.

    By the way, the danish infinitive particle is "at", not "å". For what it's worth "at" is also the conjunction used to nominalize subclauses (like English "that"). The English "to" and German "zu" both seem to share the infinitive particle and preposition function (on top of the fact that they're probably cognates), so this is an obvious example of English and German being more similar.

    Finally, is it possible that English is a Scandinavian/West-Germanic creole?

  45. Minne said,

    December 5, 2012 @ 8:07 am

    For me as a Frisian speaker, the outcome of this study seems odd.

    We have some many frequently used words in Frisian which are more or less the same in English (we call it Ingelsk). You don't find a comparable translation in the Scandinavian languages, Dutch or German.

    The syntactic structures may have more in common with the Scandinavian language, but don't underestimate the influence from Frisian on the English language.

  46. Charles Hall said,

    December 5, 2012 @ 8:39 am

    It was interesting to see the origin of the periphrastic "do" being discussed again. In my now ancient dissertation from 1983, I discuss the evidence that suggests that the periphrastic "do" first arose in the Cornish speaking areas of the UK and was used in Cornish but not in Welsh.

    Hall, Charles Edward, Jr. (1983). Periphrastic do: History and Hypothesis. Univ. Florida, PhD dissertation/UMI.

    haven't thought about that in a couple of decades.. Thanks!

  47. Robert van der Hall said,

    December 5, 2012 @ 9:05 am

    I am a historian not a linguist, but I miss in this most interessting discussion the influence of the Normans (Normandy) who spoke a Norwegian/Danish language strongly influenced by the French language or even a typical French greatly influenced by their own Norwegian/Danish.

  48. Oliver said,

    December 5, 2012 @ 9:07 am

    Some dialects of German also use a construction with to do, especially for expression of subjunctives.

  49. Jason said,

    December 5, 2012 @ 10:03 am

    I don't think we have good vocabulary for this sort of influence.

    I'm learning Tok Pisin at the moment, which is an English lexified melanesian/austronesian creole. One thing that's happening to this language at the moment is increasing decreolisation as its speakers learn English, introducing new grammatical forms previously alien to it. For example, a new instrumental sense of the preposition "wantaim" under the influence of its English equivalent, "with", and a locative use of the previously exclusively temporal adverb "behain" under the influence of its English cognate, "behind". "Classical" or "Biblical" Tok Pisin from the 1960s never uses these words in this way.

    What's notable is that you can actually watch, almost in real time, English grammar "leak" into Tok Pisin. A lot of it occurs from code-switching by bilingual speakers, who will drop in an English word or whole constituent into a stream of Tok Pisin. Thus "bikos" has become interchangeable with the equivalent conjunction "bilong wanem", through code switching. Moreover, just as a previous poster has mentioned for French, many "translationisms" from clumsily translated English text have produced new models of syntax for speakers of Tok Pisin to internalize.

    Watching this gives me a much better understanding of how French grammar must have leaked into English after the Norman conquest.

    So when Faarlund says that `[e]ven though a massive number of new words are on their way into a language, it nevertheless retains its own grammar. This is almost a universal law….It is highly irregular to borrow the syntax and structure from one language and use it in another language", he is, unfortunately, flat wrong. He should study a creole undergoing decreolization. He'll observe grammar leaking into it like a sieve.

  50. Lane said,

    December 5, 2012 @ 10:21 am

    RP, Herbert is wrong about Danish. Google "jeg lover at ikke" ("I promise to not…") and you'll get hundreds of hits from native speakers. You can definitely split an infinitive in Danish.

    What confuses me is that we're only seeing examples from modern Norwegian (Nynorsk, at that), and not from Old Norse. I'm sure this is to make the research readable to Norwegians, but until we learn more about when "att/at/å" arose in Scandinavian and "to" in English, the case is at the least seriously incomplete.

  51. Gav said,

    December 5, 2012 @ 10:35 am

    @David Eddyshaw the do + verb construction is quite common in modern Wenglish, as for example "you do sing betterer and betterer every time we do hear you". As you say the "do" isn't a direct translation from the Welsh, but it may be standing in for the particle "yn", or something like that.

  52. RP said,

    December 5, 2012 @ 11:03 am

    Thanks for that. (Also, apologies to Wayne Harbert as I misspelt his name.)

  53. Mark F. said,

    December 5, 2012 @ 11:05 am

    Is it reasonable to take away from this that anything can be borrowed? And, if that is true, is it reasonable to say that, at least in principle, there can sometimes be no right answer for what family a language belongs to? Suppose the Norman conquest had never happened but there had been more waves of Scandinavian language contact so that, eventually, English really was more like a Scandinavian language than to any (other) West Germanic language. This is kind of a Ship-of-Theseus situation, except that I'm imagining that ship being made, board by board, into a different ship. In a case like that, I can imagine contexts where either categorization would make sense. Is this utterly fanciful? My take is that linguists would go with the historically motivated "once a WG language, always a WG language" classification, but it seems like an interesting question.

  54. J.W. Brewer said,

    December 5, 2012 @ 11:11 am

    The key problem is perhaps that the metaphorical family trees traditionally used by historical linguists differ from actual human family trees in that they assume parthenogenesis at every generation. The ultimate "mother" PIE bore fatherless "daughters" (ignoring possible intervening stages) like Proto-Germanic and her sisters, and Proto-Germanic in turn had fatherless daughters like Proto-West-Germanic and early Old Norse. There is no way in this sort of tree to cleanly categorize languages (pidgins/creoles being an extreme example) where there has been substantial (and not limited to loanwords) input/cross-contamination from sources other than whatever is conventionally considered the "mother" language, thus leading to wacky proposals like this one arguing that we have misidentified the true mother of Middle English, because maternity has to be an all-or-nothing choice between Old Norse or Old English. Putting the pre-1066 ON input to the side, let's stipulate that in actual fact the quite substantial French input into Middle English did not reach the point where it would be more useful/meaningful to describe the result as a "Romance" language rather than "Germanic" language, but I don't know that we have a coherent way of even talking about these sorts of phenomena such that we could agree how much more French-driven change would have been necessary before that tipping point would be reached.

  55. John F said,

    December 5, 2012 @ 11:28 am

    apologies in advance to Matt http://languagelog.ldc.upenn.edu/nll/?p=4341#comment-295318

    Is this the part where we complain about the word "fernsehen"?

  56. RP said,

    December 5, 2012 @ 2:14 pm

    @J. W. Brewer,
    Lexicostatistics could be a relevant tool. If you used a Swadesh list then you'd surely find that the relationship between English and any Germanic language was much closer than that between English and Romance language. According to Wikipedia, such comparisons can in fact be used to construct language family trees. Whether this gives the same results as the traditional classifications, it doesn't say.

    In terms of grammar, there are a number of shared characteristics of Germanic languages, such as the fact that they have periphrastic futures instead of morphological future tenses, or the fact that verbs are either strong or weak, with the former inflected for the past tense via internal vowel changes and the latter by adding dental suffixes. If we had only just discovered English, we would be struck by the fact that it too has these characteristics. So, it seems likely that both grammatically and lexically, English would have to have been a good deal more heavily influenced by French in order for it to be reclassified as Romance. But that said, I think I'd agree with your observations about the simplism of the family tree approach.

  57. Norse Nonsense | The Tulip Institute said,

    December 5, 2012 @ 2:35 pm

    […] Mind you, Faarlund's claim is not being reported in some reputable academic journal. It's essentially a press release from the University of Oslo. It's also completely bogus. […]

  58. Dan M. said,

    December 5, 2012 @ 4:16 pm

    Note that the analogy between language families and parthenogenesis suggests that the appropriate biological model is something in which children come from a single direct parent, but also can exchange genes with moderately distant cousins. Bacteria spring to mind, and the inter-cousin exchanges are called "lateral gene transfer". Unfortunately, I don't know what solution biologists have formed for describing species and phyla in bacteria.

  59. John W. Kennedy said,

    December 5, 2012 @ 5:52 pm

    I wouldn’t even exalt myself to the rank of “amateur linguist”—“less ignorant than the average layman” is all I dare aspire to—but I have had a few days to think about this new hypothesis, and I have to say that wherever I have seen it discussed, there has been much too much MnE and MnHG dragged in where LOE and EME would be more to the point. What on earth has the interrogatory "do" to do with this, for example?

    That said, I find it hard not to believe that, if there were any real meat to this, people like, say, J. R. R. Tolkien would have noticed it yonks ago.

  60. David Eddyshaw said,

    December 5, 2012 @ 7:21 pm

    @Charles Hall:

    Your dissertation sounds interesting. Don't suppose it's available online at all? Presumably not if it dates from primeval 1983. Did Cornish use "do" as an auxiliary like English? How would you know if the influence was actually rather from English to Cornish, especially given the fact that the language was evidently seriously threatened by English much earlier than Welsh?

    Anyhow, it's good to be reminded that Welsh is not all there is to Brythonic, even now, and all the more so in previous centuries.

    Still doesn't seem plausible to me that my other Celtic cousins would have constituted a large enough mass of Brythonic/English bilinguals to produce a change like this among the majority of poor monoglot Saxons, especially as it's a fairly late change in the history of English AFAIK, from well after those guys had pushed us out of our historic cities of Llundain and Rhydychen and Efrog … perhaps they just wanted to sound cool by imitating the indigenes,

  61. marie-lucie said,

    December 5, 2012 @ 8:32 pm

    RP: If you used a Swadesh list then you'd surely find that the relationship between English and any Germanic language was much closer than that between English and Romance language. According to Wikipedia, such comparisons can in fact be used to construct language family trees.

    Swadesh lists are very useful for languages which are demonstrably closely related, which usually means having not only much vocabularly in common but also similar morphology. A Swadesh list for English, German and French may show that English shares more verbs with German than French, but it will not show that English is much more similar to German than to French in the patterning of verb forms, and also in the various forms taken by adjectives. Morphology was important to the early historical linguists, a fact which is not always recognized by modern ones who rely to a much larger extent on lexical (= vocabulary) comparisons which tend to use only the bare or basic forms of words (eg good : gut, not also better : besser, best : best).

  62. Sally Thomason said,

    December 5, 2012 @ 8:35 pm

    @Mark F.: I do think (and Kaufman & I argued in 1988) that anything can be borrowed. For instance, there's at least one example in the literature — unfortunately without a lot of linguistic detail, as far as I know — of a language that has borrowed its entire grammar, bit by bit, from another language: Laha (spoken in Indonesia), borrowing from another Austronesian language, Ambonese Malay. Laha has kept (most of?) its own lexicon, though.

    @J.W. Brewer: It *has* in fact been suggested that English is — well, not a Romance language, but a French-based creole. Not by me; I think that case is even weaker than Faarlund's for English as Scandinavian, because in spite of the fact that English has thousands of loanwords from French, many many more than its stock of Scandinavian loanwords, the structural influence of (Norman) French on English, while not trivial, was not extensive, much less deep. In contrast, the Norse influence on English structure was substantial, even though it left the bulk of English structure in full West Germanic mode.

    Apologies to all, by the way, for my stupid mistake about German and other West Germanic infinitives. I should know better; in fact, I do know better, but for some reason my brain apparently switched off after right after it processed the German modal + infinitive construction. The basic point I was making with that example stands, however: if split infinitives, where the "to" word is separated by another word from the infinitive verb form, are innovative in either or both of English and Scandinavian, the changes could easily be due to independent parallel innovation. It's not true, as Faarlund claims with reference to all his syntactic examples, that "these kinds of structures are very unlikely to change within a language".

  63. marie-lucie said,

    December 5, 2012 @ 8:51 pm

    RvdH: the influence of the Normans (Normandy) who spoke a Norwegian/Danish language strongly influenced by the French language or even a typical French greatly influenced by their own Norwegian/Danish.

    The Norman conquest of England by the descendants of Vikings who settled in Normandy took place just a few generations after that settlement. Since the vast majority of the settlers were younger men, they married local, French-speaking women. As often happens in such circumstances, the children must have been more or less bilingual, the grandchildren perhaps passively so, and later generations were monolingual in the local French, For a while, upper-class Normans sent their young sons to a special school in Bayeux, the site of a bishopric [where the famous "tapestry" is still preserved), to learn the language of their ancestors, because most of those boys' fathers and grandfathers no longer spoke it and could not pass it on. Local French dialects adopted a number of Norse words, mostly dealing with the vocabulary of the sea, some of which later found their way into Standard French when it became fashionable to go sea-bathing on Norman beaches (which are the closest to Paris). I am not aware of other recognized Norse contributions to local dialectal peculiarities.

  64. Marja Erwin said,

    December 5, 2012 @ 10:46 pm

    Isn't it common Germanic to use the plain infinitive, or du+infinitive, to express purpose? Wright notes it in Gutrazda, which is about as far from English as any of the Germanic languages.

    Is it possible that the English use of to+unmarked for the infinitive is derived from the common Germanic du+infinitive, although it has expanded from marking purpose to marking most other infinitive roles?

  65. Anders Stallemo said,

    December 6, 2012 @ 12:28 am

    @Sally Thomason: It’s a bit unfortunate that the debate has centred around only one syntactic construction. I doubt that Faarlund & Emonds would have made such a claim on a single piece of evidence. However if there are several cases where the syntax of Germanic Scandinavian languages and English concur, and at the same time they behave differently than West Germanic languages, wouldn't such a body constitute an argument against independent parallel innovation in Scandinavian languages and English? Allow me to update merijn’s link as the URL provided above leads nowhere: http://conference.uaa.utb.cz/TheoriesAndPractice2010.pdf#page=13. On pages 22-25 there are 7 morpho-syntactic phenomena discussed by Emonds. I can’t vouch for all of these, and I'm not really trying to advocate Faarlund's view here, but it’s important to notice that the discussion should be something more than if German can split infinitives or not.

    Also notice that the assertion in the paper is “Middle English (and therefore also Modern English) originates as an amalgam of West Germanic Old English and North Germanic Old Norse”. I’m guessing the whole ‘English is Scandinavian’ claim (albeit logically correct if we buy the fusion theory) is only fielded due to the home crowd advantage.

    Finally I’d like to point out that Faarlund has been dealt no favours by the translator here, he comes across as pigheaded and ignorant in the English version. But the original, found here: http://www.apollon.uio.no/artikler/2012/4-engelsk-er-skandinavisk.html, shows that his reluctance to syntactic borrowings and changes are not as deeply rooted as one might think. The sentence that goes: “It is highly irregular to borrow the syntax and structure from one language and use it in another language” could better be rendered “Syntax and structures are not as easily borrowed [as words and morphemes] between languages” (in Norwegian: “Syntaks og struktur låner ein ikkje så lett frå eitt språk til eit anna”).

    And a correct translation of what you cite “these kinds of structures are very unlikely to change within a language” would be “it’s highly unlikely that all these structures can [simultaneously] be in the process of change in a language” (Norwegian reads: “det er svært lite sannsynleg at alt dette kan vera i endring i eit språk”).

  66. John Walden said,

    December 6, 2012 @ 5:10 am

    Is it possible that a trading creole existed on Continental Europe? The melting pot didn't have to be what is present-day England.

    The speakers of the cognate languages had plenty of contact with each other. That they got around a lot is evident, and rather the point.

  67. Daniel von Brighoff said,

    December 6, 2012 @ 12:29 pm

    The problem with theories of "creolisation" is that Middle English retains too much irregular morphology to fit any model of creole genesis developed from languages which can be shown with certainty to have undergone this process. How many verbs with suppletive forms are there in basilectal Tok Pisin? Which adjectives in Haitian Creole have irregular comparative forms?

    I have to agree with Jason that we don't have a good vocabulary for contact phenomenon which don't fit into traditional models of linguistic development. As a result, people seem to apply the term "creolisation" rather indiscriminately to a whole range of types of influence which fall well short of it.

  68. marie-lucie said,

    December 6, 2012 @ 8:21 pm

    Attested creoles derive from two (perhaps more) languages which are vastly different from each other in almost every respect. This was certainly not the case with Old Norse and Old or Middle English.

  69. Thomas said,

    December 7, 2012 @ 2:23 am

    Sorry for my terrible research skills, but where does one find information about the number of Norse people in England?

    Sally Thomason's post says

    The number of Norse settlers is estimated to have been in the hundreds (not thousands), so that they would have been far outnumbered by English speakers.

    This surprised me. I'm ashamed to say I've read about English history before, without it ever occurring to me to find out how many Norse settlers there were, or how many Normans later, or Saxon settlers earlier, etc.

    So where does one find such information?

  70. This Week’s Language Blog Roundup: Dictionary scandal, names, 30 Rock cocktails | Wordnik said,

    December 7, 2012 @ 10:03 am

    […] ­vibrancy” in the film, Lincoln. Sally Thomason at Language Log took a look at the claim that English is a Scandinavian language. Johnson explored internet language, dictionaries and finding the right format, and Christmas […]

  71. Gabe Ormsby said,

    December 7, 2012 @ 5:04 pm

    Sort of a tangent, but I see this at the tail end of the article:

    "English, a West Germanic language whose closest relative is Frisian…"

    Every time I see this I think, "What about Scots?"-Is it the consensus, or at least contested, that Frisian is closer to English that Scots is? Or is this a case of the mental boilerplate not having caught up with the elevation of Scots to the level of a fully qualified Anglic language on par with English?

  72. marie-lucie said,

    December 7, 2012 @ 5:07 pm

    The number of Norse settlers is estimated to have been in the hundreds (not thousands), so that they would have been far outnumbered by English speakers.

    It seems incredible that such a small number of people would have had such an influence on the language of the country they conquered, unless the native population in the region had been very small and sparsely distributed to begin with. It is possible that the original invaders/settlers were joined in short order by many of their fellow Danes, including both men and women, so that the language continued to be spoken for quite a while within families before the newcomers merged with the native population. Since the languages were very close in structure and vocabulary, there must have been a fair amount of near-bilingualism, with passive understanding on both sides leading to a fair amount of borrowing. Compare this with the situation in Normandy, where the original settlers cannot have been very numerous, but the Norse language did not survive for more than a couple of generations before giving way to the linguistically unrelated French spoken by the majority of the population. (Of course, Norse and French were not totally unrelated as both were Indo-European, but they were not close enough for easy communication).

  73. Phil Kosta said,

    December 8, 2012 @ 10:02 am

    A number of interested readers have already reacted on the use of the split infinitive in German. A similar situation exists in the Dutch language where the English "to", the German "zu" and the Dutch "te" are used in conjection with certain verbs.
    A few examples: 'I wish to stay' translates into 'Ik wens te blijven'. And also 'I will stay' or in Dutch 'Ik wil blijven' (in both cases without to/te). A parent may say to a child: 'You have to behave yourself' which becomes 'Je hebt je te gedragen'. Another example: 'One ought to honour their parents' translates into 'Men behoort zijn ouders te eren'. Many other examples can be given. The bottom line is that English is a Anlo-Saxon-Frision language with a modest Norse influence and ~ a 30% inheritance of words derived from Latin and French.

  74. Sally Thomason said,

    December 8, 2012 @ 10:48 am

    @Thomas and Marie-Lucie: Estimated numbers of Norse settlers in England are discussed in Terry Kaufman's case study in our 1988 book Language Contact, Creolization, and Genetic Linguistics (see especially p. 360, footnote 13). One of Terry's main sources was P.H. Sawyer's book The Age of the Vikings, 2nd edition (London: Edward Arnold, 1971).

    @Gabe Ormsby: I doubt if anyone thinks that Frisian is closer to English than Scots, a.k.a. Scots English, is. The question would be whether Scots should be considered a separate language from English; these decisions, in the case of very closely related language varieties, are never straightforward linguistically, because the only linguistic criterion we have, mutual intelligibility, is often not usable: attitudes can and do have a *major* impact on people's judgment of mutual intelligibility. Political considerations, which are (almost?) always in play when such questions arise, muddy the waters and make the decision ultimately non-linguistic. For people who believe that Scots is a separate language, I assume (without any expert knowledge on this question!) that your rubric "Anglic" would include both Scots and English, and that the closest relative of Anglic would then be Frisian.

  75. marie-lucie said,

    December 9, 2012 @ 4:48 pm

    Thanks for the references, Sally.

  76. Wickie und der starke William – Sprachlog said,

    December 10, 2012 @ 3:14 am

    […] amerikanische Linguistin und Expertin für Sprachkontakt, Sally Thomason, stellte letzte Woche im Language Log unter English or Engelsk? dementsprechend treffend fest: „Außergewöhnliche Behauptungen […]

  77. The day I took my mask off my face was missing « my nerves are bad to-night said,

    December 10, 2012 @ 5:16 am

    […] Extraordinary claims require extraordinary evidence, as the saying goes.  read more […]

  78. Link love: language (49) « Sentence first said,

    December 10, 2012 @ 9:46 am

    […] English is not a Scandinavian language. […]

  79. Daniel von Brighoff said,

    December 11, 2012 @ 4:06 pm

    Attested creoles derive from two (perhaps more) languages which are vastly different from each other in almost every respect.

    Generally more. Where only two languages are involved, you're more likely to get either stable bilingualism or language shift. (Only in some highly unusual circumstances do you get a truly mixed language like Medny Aleut or Michif.)

  80. Me Towk Gud engish said,

    December 14, 2012 @ 2:01 pm

    This is just ridiculous.

    No one doubts that Danish had a strong influence on the development of Middle English from Old English. But the fact remains that there are plenty of words and syntactic structures from Old English that remain in Middle English – Old English did not die off, it evolved.

    And Norman French had a strong influence after Danish. This means that Modern English is not a direct descendant of the Danish-influenced Middle English.

    Furthermore, there was an intentional expansion of English in the modern period for several purposes. Artistic innovators like Shakespeare created many new words. Scientists and others imported Latin and ancient Greek terms to use as a technical vocabulary.

    So modern day English consists of Anglo-Saxon-Jute dialects + Danish + French + Poetical Licence + Latin and Greek.

    Which is what linguists believed already. I have never seen any linguist claim that Modern English is a direct descendant of Old English. Our scholar is attacking a straw man that he created.

    If you want to have an example of language replacement, you can also look to England. Old English almost entirely replaced the Romano-Celtic dialects that were present in England before the Anglo-Saxon-Jute invasions. But even there, much survived, pushed to Wales, Cornwall, Scotland.

  81. Indian History Carnival–61: Linguistics, Sernigi, Babur, Ramanujam, Thirumalapuram | varnam said,

    January 15, 2013 @ 3:58 am

    […] close contact with each other borrow just words or do they borrow grammar as well.? Sally Thomason mentions an example from India Probably the most famous case of all is Kupwar, a village in India in the border area between Indic […]

  82. David Marjanović said,

    January 27, 2013 @ 2:32 pm

    Wow. Of course German has a split infinitive! "To go" in German is "zu gehen." ("Zu" is also cognate to English "to" in other senses as well.)

    To boldly go where no [hu]man has [been] before.
    Kühn dorthin( )zu( )gehen, wo noch nie zuvor ein Mensch gewesen ist.

    "Der Infinitiv hat in manchen Konstruktionen die Partikel zu bei sich…"
    "In many structures, the infinitive is accompanied by the particle zu…"

    It may be cognate to many, but manch- means "some".

    I suspect that the various now-archaic syntactic features of the King James Version that might seem congenial to a modern Swedish speaker would be at least as congenial to a modern German or Dutch speaker


    Bilingualism was such a threat to the persistence of English that the English community felt a need to pass a law to forbid English from learning Irish, the linguistic equivalent of a kosher law.

    The best part, from today's point of view is how that law, from Dublin, is worded: ITEM ORDINE EST que chescun engleys use la langue engleyse… not a word in English in it! :-)

    Even if we assume Norse influence (rather than an unattested Anglian *dréam 'dream')

    Perhaps that's exactly what we should assume, given German Traum "dream".

    Unfortunately, I don't know what solution biologists have formed for describing species and phyla in bacteria.

    It's very simple: they haven't found a solution. People who work on bacteria often use a couple of very simple criteria that aren't very closely connected to phylogeny and… well… one of them is "70 % genome identity or more = same species", which would put all primates into the same species – mouse lemurs and men.

  83. On Scandinavian Roots, the Goodreads Choice Awards, the Publishing Person of the Year, the Business Side of News and Social Media Wars | Project Chiron (Beta) said,

    April 3, 2013 @ 8:46 am

    […] Sally Thomason on Language Log refuted Faarlund’s theories and countered his claims that “languages in contact can be counted […]

  84. Pia said,

    May 18, 2013 @ 9:48 am

    I have the feeling that if Chomsky would have made such a weird claim on some other kind of topic, several journals would have already published the work.

    I think that Faarlund's claim is unreasonable starting from the first step – not to consider the rest.

    Structural diffusion, as Sally says, can happen on several levels. Saying that just one of them has the privilege sounds a little bit like blasphemy to me.

    I am not qualified enough to comment on the relationship between Norse, Anglic, or other dead languages with English, but what I keep thinking is 'no evidence, no party'.

  85. Leonard said,

    October 12, 2013 @ 1:18 pm

    @Sally Thomason

    I speak English, Swedish and German.

    So is English closer to Swedish or German? No contest, much closer to Swedish/Norwegian/ Danish.

    This is what you have to refute: http://conference.uaa.utb.cz/TheoriesAndPractice2010.pdf#page=13

    A scientific article by Joseph Emonds. Not a simple newspaper story.

    Nota bene: Following general practice throughout linguistics, a language is classified by its syntactic decent.
    Emonds gives us 7 scandinavian. Can you give us one Oldenglish/West-Germanic?

  86. K International | Is English a Scandinavian Language? said,

    October 15, 2013 @ 7:54 am

    […] to Sally Thomason at Language Log, the answer is "no." She points out there have been a number of documented cases of one […]

  87. Norse Nonsense | The Tulip Institute said,

    December 25, 2013 @ 9:33 am

    […] Mind you, Faarlund's claim is not being reported in some reputable academic journal. It's essentially a press release from the University of Oslo. It's also completely bogus. […]

RSS feed for comments on this post