Updated June 27, 2016. This post will be regularly updated for some time. Warning! This essay is very long; combining all 3 parts together, it runs to 255 pages.
This post has been broken into three parts: Part 1: Low German, Part 2: Middle German and Part 3: High German.
This classification divides German from 20 languages to 137 languages based on the criteria of intelligibility. >90% intelligibility = dialect, <90% intelligibility = language.
The German languages, and German is not a single language, but, like Chinese and Italian, a family of languages, have been in need of a good reclassification for some time now. Ethnologue has done an excellent job, dividing German into 20 separate languages.
However, Ethnologue’s treatment does not go nearly far enough, as they themselves admit in the entry for Standard German: “Our present treatment in this edition is incomplete.” The entry for Low German itself formerly stated that LG is made up of 20-30 separate mutually unintelligible dialects, although this has been revised to “differing intelligibility, depending on distance” in the latest edition.
Hence, this treatment will attempt to expand the 20 German languages listed at present into a higher number. Many of the language divisions noted below are arbitrary, and admittedly based on more intuition than hard evidence. In many cases, intelligibility testing could clear up a lot of confusion. This treatment, like my prior treatment of Chinese, is best seen as a series of often very tentative hypotheses rather than a set of conclusions.
The classification scheme (for instance, the decision to include Low Saxon as a part of Macro-German rather than a part of Macro-Dutch) is fairly arbitrary and is not the purpose of this treatment, which deals mainly with intelligibility. This treatment makes no statements about classification, generally just following Wikipedia and Ethnologue. There are others doing major work on classification, and I will leave that up to them.
So far, this classification expands German from 20 separate languages to 142 separate languages. It is incomplete, and it is also a pilot study intended to spur further research, analysis and especially evidence-based criticism.
Criticism is welcome, as long as it is rational and evidence-based. Keep in mind that we have valid intelligibility data for a lot of these languages, so wild claims of widespread intelligibility are likely to be ignored. Further splitting is certainly warranted, and lumping may be too. Both will require evidence before proceeding.
Method: Literature and reports were examined to determine the intelligibility of the various dialects of German. Native speakers of various lects were also interviewed, and the results of scientific intelligibility testing were examined. There was an appeal to authority – if states or the ISO recognized that a lect was a language, this determination was simply accepted.
>90% intelligibility was considered to be a dialect of German. <90% intelligibility was considered to be a separate language from Standard German.
The emphasis was on intelligibility rather than structural factors. Certain sociolinguistic factors also went into the calculation, but their use was minimized. Overtly political argumentation was ignored.
This piece may be seen as a companion piece to my other similar pieces. A reclassification of Chinese expands Chinese from 14 languages into 343 languages. A reclassification of Catalan reanalyzed it from 1 language to 2 languages. A reclassification of Occitan changes it from 6 languages to 12 languages. A reclassification of Dutch changed it from 15 to 30 languages.
As far as my qualifications for writing this, I have a Masters Degree in Linguistics, and I have been employed as a linguist for an American Indian tribe where I created an alphabet, ran the language program, worked on a dictionary and phrasebook and did fieldwork with native speakers.
German, like Chinese, is a pluricentric language, with a standard version and many typically mutually unintelligible major dialects surrounding it. Interdialectal comprehension is achieved via the use of Standard German.
Hence, the intelligibility estimates by the ignorant are going to be biased. What these people mean when they say that everyone in Germany can understand each other is that they can when they speak Standard German to each other. However, there are still a few older folks in Germany who cannot speak Standard German and can only speak their regional form of German.
There are 27 main German dialect families, and all are considered to be separate languages.
The German dialects exist as a dialect chain where dialects are normally intelligible to the dialect regions next door, but not to those more distant. At the same time, it is frequently stated that the major German dialects are not mutually intelligible. This makes delineating languages from dialects quite difficult and is why intelligibility testing is needed.
Most German “dialects” have low intelligibility (below 90%) to speakers of Standard German, because they are quite divergent and hence hard for a Standard German speaker to understand. There is a strong suggestion that all of the strong forms of the regional lects are not intelligible to a speaker of Standard German.
Germany is awash in dialects. There are over 4000 (!) different dialect groups within Low German alone, and there are 150 dialects in Ripuarian Franconian that were different enough to have dictionaries written for them.
In addition to not being intelligible with Standard German, the major German dialects are in general not mutually intelligible with each other either. Inside of that, there is the even more alarming suggestion that many of the major dialects are so diverse that they are not even completely intelligible among themselves.
A graph of the major German languages is here, and an even better one is here.
Separate languages or suspected separate languages are bolded below. Dialects or extinct languages are generally italicized. Macrolanguages that do not deserve separate language designation are generally printed in normal typeface. All languages and dialects are spoken in Germany unless otherwise noted.
Languages or dialects marked by an asterisk were definitely full languages 50 years ago, but whether they still are today is less certain. Some may still be languages, others may have dwindled to dialects and others may have disappeared. 50 years ago, those languages were still alive and well and probably even being taught to children. Most or all still have speakers, though the youngest speakers may be over 50 in some cases. These starred lects are very tentative additions to this classification.
This research takes a lot of time, and I do not get paid anything for it. If you think this website is valuable to you, please consider a a contribution to support more of this valuable research.
A Reworking of German Language Classification
Updated June 27, 2016. This post will be regularly updated for some time. Warning! This essay is very long; combining all 3 parts together, it runs to 255 pages.
40 thoughts on “A Reworking of German Language Classification”
Doesn’t the German dialect chain extent into the Netherlands and include Dutch (i.e. the many Dutch dialects, many of which are also mutually unintelligible if I’m not mistaken)? If so, shouldn’t that be included in your analysis?
To be honest, though, I think you’re way too into this whole thing about separating “dialects” from “languages” – there’s obviously a lot of gray area, and any strict division is going to be mostly arbitrary, especially when you have phenomena like dialect chains complicating matters.
This is one of the lies of the sociolinguists. If you look in this piece around the languages called Limburgs and Gronings you can start to understand what these sophists are talking about. There are a variety of separate languages in this region that are frankly intermediate between Dutch and German. It is these languages that these liars are referring to when they speak of this wonderful dialect chain between German and Dutch.
The truth is that there *is* a dialect chain starting in Belgium and going to Austria, but all that means is that one village can pretty much understand the one next door! So, if you follow this chain, yeah, they can all understand each other. So what! You need to understand the concept of the dialect chain.
I honestly doubt if any competent Germanic linguist would seriously disagree with this latest assessment. For the most part, the people who freak out about are almost all non-linguists, many of whom are really upset that folks are splitting off “dialects” into “languages.”
For instance, my Chinese reclassification, which expands from 14 languages to 343, is being supported by some of the top Sinologists in the world. They don’t exactly agree with my conclusions, but they don’t disagree with them either.
There is a massive gulf here between us linguists and everyone else.
Keep in mind that the modern field of linguistics is splitting dialects to languages at 90%. Below 90% and it’s supposed to go to new language.
That all said, it’s a fact that I am a maniacal splitter. But I’m with Ethnologue on the wild splitter team, and they are the guys handing out ISO codes these days.
You have some interesting data. I am acquainted with the Ethnologue and understand the basic premises of their classification by personal interaction with editors. I have also lived in Germany myself (and speak German fluently) and have a good comprehension of the concept of intelligibility. I believe this concept needs to be defined more clearly for any such criterion to be generally accepted. However, in my recent combing of German linguistic literature, I have confirmed that German linguists do NOT generally accept the upgrade of ‘Dialekt’ or ‘Mundart’ to languages. As it was told me, they often challenge the Ethnologue’s present classification into languages (not necessarily denying the varieties – Varienten).
Yes but they are just wrong. A lot of those “dialects” have only 40% intelligibility with Standard German. They must be separate languages. One usually splits language from dialect at 90% intelligibility.
It is already very well defined. Intelligibility is what Ethnologue uses to split languages from dialects. You can do intelligibility testing or you can just ask native speakers what they can understand and what they can’t and to what extent they understand neighboring lects. What I found is that they will define “our language” (their dialect) within a circumscribed region, and list all of the cities where “our language” is spoken. Once you start getting outside of that though, they say, “That is not our language anymore.” You ask if they can understand them and they say, “No, not completely.” You ask them how much and they tell you something like 60%.
This stuff is not that controversial. You just need cooperative native speakers.
Dutch and german is as different as is danish to german. You can understand less then one percent of the words. A dane, a dutch and a german can NOT understand each other.
I speak a little german, give me six months in Germany and I would speak fluent german. The north germans speak slow and very clear, the south german and Shweiz german speaks down in the troath, just like the cook in Sesam Street does:
They speak that way but its no problem for a north german to understand what they are saying.
When I speak to foreigners I always listen for the dialect, not what country they are from, but what part of the country they are from.
I’m not sure I get your point – if everyone can understand people from the next village from Belgium to Austria, then where exactly do you draw the line between “languages”? Seems like whether you want to call it all one language or 1000 different language is just a matter of opinion, regardless of what your standard is for mutual intelligibility (because no matter how you slice them, you’re inevitably separating people with near 100% mutual intelligibility at the geographic borders of the “languages”).
Minor correction: you wrote Wikipedia instead of Ethnologue and that confused me for a minute.
Where to draw the line in dialect chains is a very interesting question with no easy answers. However, people like Ethnologue do it all the time. If you go to the Ethnologue page about Mexico, look at how many languages they are splitting off. Every little Indian village speaks its own language. Are they dialect chains running through Indian Mexico? Who knows? Maybe.
There is said to be another gigantic dialect chain running from Portugal all the way to Sicily or so and another one running through the Slavic lands.
To answer your question briefly, take Slovak and Czech. Two languages, however, the west dialects of Slovak can understand the eastern dialects of Czech. Towards the other ends though, they can’t understand each other. If there are some dialects of Lect A that cannot understand Lect B, it’s reasonable to assume two separate languages. Language A and Language B. If all dialects of Lect A and Lect B are intelligible, then you cannot split into separate languages and there is Language C with dialects A, B and subdialects.
It’s well-known and often stated (just look around the Net) that the German dialects are not intelligible with Standard German, nor in general are they with each other. Now they say that, but what exactly do they mean by “German dialects?” Do they mean the main ones? That’s what I’m assuming.
You can see above that I split Badisch and Swabian. That was a tough call, because people were saying that they are intelligible. But it’s more that the West Swabian dialects can talk to the East Badisch dialects. Once you get to the West Badisch and East Swabian dialects, they can’t talk anymore. So then you can split, even though many Swabians and Badisch speakers can talk.
Luxembourgian can talk to Moselle Franconian, but only the far eastern dialects of Luxembourgian can talk to the far western dialects of Moselle Franconian. So you need to split.
If you can see above, I did not split Thuringian off. That is because I currently lack evidence that Thuringian is unintelligible with Standard German. Thuringian has many dialects within it.
I also did not split Central and Northern Bavarian, since it seems that all dialects can talk to each other. As long as all dialects can talk to each other, there’s no split. Once you start running into unintelligible dialects, you need to split somewhere.
A very large number of those 6,800 languages in Ethnologue are dialect chains.
You are right though. Splitting dialect chains is hard to do! But it can be done, and linguists have been doing it for a long time now.
There are other ways beyond intelligibility to split languages nowadays. I wonder why those Dutch Low Saxon lects are split into 8 or so languages. But the Dutch government has decided that they are regional languages, so we accept that. My splitting is solely on intelligibility.
That’s actually a conservative assessment. Swiss German itself may be 20-50 separate languages, and I think I only split 4.
This is a pilot study. It’s intended to spur research, debate and criticism. Do you know what a pilot study is?
You can say it’s a matter of opinion and arbitrary all you want, but we linguists don’t really argue about this stuff too much.
Nobody argues for one language from Belgium to Austria.
Can all dialects of Westphalian talk to all dialects of Eastphalian? Yes? One language (make up a name). No? Split – two languages, Westphalian and Eastphalian.
Mr. Lindsay, I can understand that using a 90% mutual intelligibility criterion, many erstwhile languages would have to be split into several languages. However, in relation to the comment of “Reader” that the divisions are arbitrary, it seems to me that there might be some circular reasoning involved, at least in the treatment of particular cases.
For example, in your answer to Reader, you said: “You can see above that I split Badisch and Swabian. That was a tough call, because people were saying that they are intelligible. But it’s more that the West Swabian dialects can talk to the East Badisch dialects. Once you get to the West Badisch and East Swabian dialects, they can’t talk anymore. So then you can split, even though many Swabians and Badisch speakers can talk.” Furthermore, “Luxembourgian can talk to Moselle Franconian, but only the far eastern dialects of Luxembourgian can talk to the far western dialects of Moselle Franconian. So you need to split.”
If I understand your analysis correctly, there were already two groups of dialects to begin with, Badisch and Swabian. Since not all speakers of Badisch and not all speakers of Swabian understand one another, the two are considered separate. My question is: What was the basis for grouping the dialects into Badisch and Swabian in the first place? By asking “what basis” I do not question the criterion for splitting—since we know that the criterion is mutual intelligibility—but the criterion for lumping. In other words, using the mutual intelligibility criterion, how were the groups “Badisch” and “Swabian” arrived at? If the goal is to revamp the classification while adhering strictly to the “90% mutual intelligibility” criterion, why should we accept the “predetermined” division of the dialects in this area into “Badisch” and “Swabian”? Assuming that there are no predetermined groups, and that intelligibility data are available for the entire stretch of dialects in Baden-Württemberg and west Bavaria, is it not possible that another linguist may discover that “West Swabian dialects can talk to the East Badisch dialects” and conclude that they belong to one language (and maybe come up with a new name). Is it not possible that, studying the data further, the linguist would discover that to the west of his new language, dialects in west Baden are intelligible with dialects in east Baden (near the western border of his new language), and that to the east, dialects in east Swabia are intelligible with dialects in west Swabia (near the eastern border of the new language). Would he not consider lumping all these dialects into one language…until he finds out that dialects in west Baden and east Swabia are unintelligible with each other? It seems that there are only two logical choices for the linguist: lump all the dialects into a single language, or separate them into several languages, NOT JUST “Badisch” and “Swabian”. I think this is the point of Reader’s comment: that division of a group of dialects into a particular number of languages is arbitrary, particularly when dialect chains are involved.
“We linguists”… that’s pretty funny. So, if “you linguists” don’t dispute this stuff much, then why isn’t your proposed classification already the accepted standard? Why hasn’t the issue already been definitively resolved, and instead requires your expertise? In any case, I know for a fact that linguists dispute this kind of thing, as do biologists struggling with the analogous problem of species classification. Any classification of almost anything is generally pretty arbitrary – nature generally doesn’t break things down into neat little categories for us.
Anyway, split as much as you like. To me, mapping out the linguistic variation is interesting… coming up with arbitrary divisions between languages and dialects and splitting hairs over the precise numbers of each is not.
Here’s a potentially more rigorous and interesting question: what is the minimum number of dialects/languages/whatever, within the German dialect chain, that one would need to speak in order to communicate with everyone in the dialect chain (at some very high level of mutual intelligibility, corresponding to a very natural level of conversation)? Same for Chinese.
So, if “you linguists” don’t dispute this stuff much, then why isn’t your proposed classification already the accepted standard? Why hasn’t the issue already been definitively resolved, and instead requires your expertise?
Well, there already are a lot of people saying something similar to this, but it’s just that usually people don’t sit down and start trying to split like this.
Most linguists are professors at universities and shy away from big splitting projects. It’s seen as kind of super-bold, sleazy, tacky and controversial.
Anyway, it’s already an accepted maxim that the principal German dialects are not intelligible with each other. Now all that needs to be done is to start drawing lines.
Nobody really wants to do this kind of work. It’s such a critical lightning rod in the field that most linguists don’t want to mess with it, and it’s not real controversial in the field anyway. Most linguists are off arguing about this really theoretical stuff all the time, and they don’t bother too much with dialectology. Ethnologue just splits or lumps wherever they do, and most people say, “Ok.”.
If Ethnologue is really messed up, everyone pretty much agrees, and they change it. There’s really way more languages out there than Ethnologue acknowledges, but you need to petition them via a fancy ISO-3 form for each one, and Ethnologue lately are real dicks about making new languages. I submitted some forms a few years ago and they shot down each one, so I’m never going to do it again.
Usually it’s some guy who has done really extensive work in the field who submits change forms to Ethnologue and then they often get accepted. I think Hmong got split from 2 into 40 languages by some guy a while back and there is a similar petition to split Zuang or Li from like 2 into like 40 separate languages. Those are all spoken in China. But those guys have been working on those languages for years and they have an interest in that petition.
Otherwise, in some cases, speakers are trying to get an ISO code for their language, but usually speakers don’t care. Wikipedia is now requiring ISO codes for all Wikipedias in regional languages, so speakers are now submitting ISO apps, but they are often amateurish and they don’t look very good. SIL requires all this documentation and sourcing and academic refs and it’s all a great big pain in the ass.
Why hasn’t someone else already done it? Isn’t that the sort of question one can ask about all sorts of projects? In many cases, someone just has not. That is why new works, new patents, etc. occur all the time.
Species classification is not that controversial nowadays either. Science is science, new papers get published and it’s pretty clear that Owl A is really Owl Subspecies A of Owl Species B, or Owl Subspecies A of Owl B is actually Owl Species B and Owl Species B. There’s tons of controversy about speciation in the press, mostly for political reasons, but I read the biological journals a lot, and in the field, there’s not a lot of fighting going on. Like with languages, it’s mostly being done in distance now. Genetic distance usually.
Splitting dialects into languages reduces lots of non-linguists to sputtering fury (check around the Net for the wild fights), but in the field, most linguists just shrug their shoulders and most differences are just settled amicably and with considerable consensus.
Reader, the answer to your question is really the number of languages in that chain. If you would need to speak 40 dialects to talk to everyone in the chain, it really does look like there are a *maximum* of 40 languages in the chain.
And I guess that’s the dirty secret to how we split up all these chains.
I am not really surprised at your carving up of the German language area. But what is lacking is a definition of what counts as a German dialect. . This may differ from case to case. The Low Saxon dialects of the Netherlands may have been included because they are an extension of Low Saxon Niederdeutsch, even though they are under the roof of Schrift-Dutch rather than Schrift-Deutsch. That Limburgish has been included may be due to a couple of High German Low German isoglosses running through that area. But if even South Guelderish and Veluws are included (in so far as I know Veluws is Franconian but for the eastern strip along the river IJssel) I don’t see any reason why Brabantian-East Flemish, West Flemish, Zealandic and Hollandic should be excluded. Furthermore, the exclusion of West and North Frisian also needs some justification. Referring to Anglo-Frisian isoglosses will not be enough because unlike English these languages share a lot of syntax with the rest of Continental West Germanic, since they are SOV cum Verb Second. And if we try to set the two Frisians apart by referring to the idiosyncrasies of the verbal cluster (no Ersatz-Infinitiv, strict head-final order but for the Frisian “kortstaarten” litt. ‘short tails’) then — maybe — Gronings should be taken out of your list of German dialects/languages because it shares a lot of verbal cluster syntax with West Frisian.
And zum Schluss — I am not joking — what about Yiddish?
Hans den Besten
Yeah, what about Yiddish as one of the colonial varieties of German? What about Colonial German, East of the Elbe, anyway? Zaynen mir da?
Hello Hans. I answered you in a new post.
Probably in the pingback above.
I rather would see the ‘east friesian’ as a dutch dialect!
And Gronings and Drenths are both more dutch than German.
Of course, at the border they are mixed with the ‘border’ german.
In your view, you could call dutch a german dialect.
It’s all quite interesting, but what do you mean with Holland? The state is called the Netherlands. You should know that. Holland is a region in the Netherlands, on line with Limburg, Friesland and all the other provinces and regions you mention.
I fixed that, Wim.
Here in the US, we don’t know that it’s called the Netherlands officially and we often refer to it as Holland. I had no idea that Holland was merely a region in the nation.
Edward, the intention is here is a discussion of intelligibility, not classification. Here I treat Macro-Dutch as Low Franconian only. Everything else I am tossing to Macro-German, except I have no idea what to do with East Frisian.
It is a fact. But I admit, many citizens in the Netherlands call it ‘Holland’ themselves as soon as they are abroad. But that’s mainly because the word is shorter and easier, I would guess. Nevertheless irritating. By the way, once I heard mr. Burns in ‘the Simpsons’ speaking about ‘queen Beatrix of the Netherlands…..’ Well informed man!
Edward Rousou prefers calling East Frisian a Dutch dialect and also c0nsiders Gronings and Drents nearer Dutch than German. But that is a repetition of the exchange of views Robert and I had before. Of course East Frisian and all the other Low Saxon dialects/languages sound more Dutch-like than German-like. But don’t single out East Frisian. Either the whole set of Low Saxon dialects is under the umbrella of Macro-German or it is not.The sole thing that is special about East Frisian is that it is Low Saxon on Frisian substrate just like Gronings.
PS Maybe Edward is thinking of the time that Schrift-Dutch was used as the written medium in East Frisia. But if that is a criterion we also have to set Bentheim apart and parts of the Lower Rhine area in Germany and maybe also the Lingen area. The mind boggles at these perspectives.. I’d rather have them all together — either inside or outside Macro-German.
Interesting essay Robert, but now that everybody seems to be splitting hair here, why do you repeatedly refer to Oldenburg as Oldenberg, wheras you spell other city names correctly? Why do you spell Kurpfälzisch with an Umlaut and Pfälzisch without an Umlaut? Why does Vorarlbergerisch (second choice to Vorarlbergisch) link to the Wikipedia entry for High Allemannic, rather than to the Wikipedia entry for Vorarlbergerisch? Within the context of your essay I would expect a reference to the Benrather Linie?
Well just wondering when reading the first reader post and the first answer: isn’t this 90% at which you split anything but arbitrary?
(I was searching for the difference between “dialect” and “language”, i.e. a definition of a “language” as opposed a definition of a “dialect” when finding your blog)
Well, you have to split somewhere, and Ethnologue seems to be splitting at 90% these days, and they are the guys giving out ISO codes. ISO codes determine that something is a language and not a dialect.
Further, at <90% intelligibility, one starts to run into serious problems in conveying complex and complicated ideas, so there is some basis for it.
The whole what’s a language and what’s a dialect thing is never going to end really.
It is worse : I have a mother from Hamburg and a father from Antwerp. My grandfather in Hamburg had a grandmother from Brasschaat (near Antwerp). My mother’s family came originally from Alsace – My father’s family from West Flanders (Blankenberge) all these different sounds and dialects you can find in my language..
I ‘m now 64 and learned in my lifetime also to speak (besides Dutch and German and the different dialects..) French, English and Esperanto…
My point is : dialects and language varieties do not merely exist in a given region, village ktp.. but can be also different in cities and in different families & individuals…
Language/dialect is not a static phenomenon, since people move around on this planet, the languages they carry move around accordingly.
I am a student at GIAL and am involved in researching the justification for the present classification, and considering whether it should remain as it is. I have not looked yet in depth at your posts, so, I apologize if I am asking you to re-post what you’ve already made clear. But my greatest questions are these: what sources have you used for your thesis and proposition of “splitting up” German into 140 languages? Have you personally done fieldwork in Germany? What is your Level of knowledge of the German language(s)? You may reply to my e-mail instead of the post, if you wish.
I searched online to determine the extent to which various lects were intelligible with other lects, and to the degree I could, I also spoke with native speakers of different regions and I also spoke with some scholars working on these lects. This is just a pilot project intended to stimulate further research into this area. The German people who live in these areas know very well what is a separate language and what is not. They even know the boundaries of these languages, and they know who can understand who and even to what extent. They often gave me intelligibility figures like 60%, 70%, 80% and 90%. People have an intuition about these things a lot better than we give them credit for.
Some of my determinations are in error, but I really just wanted to stimulate discussion among scholars about this whole area.
There should be links in the various listings where I describe the basis for splitting off a language. The links will say that it is not intelligible with neighboring related lects, etc.
Undoubtedly the number will have to be reduced from 140. However, Ethnologue’s 20 different languages is completely wrong, completely wrong. There are many more than that.
What is GIAL?
I do not speak German and I have done no fieldwork in Germany.
I showed this article to a professor of German in the Netherlands when I only had about 80-90 languages, and he said, “I hate to say it, but I think you are right!”