On Shibbolethisation

In Scandinavia, we are used to being able to almost understand each other´s languages. This situation may be unique in Europe today, but probably wasn't earlier. One could compare it to the situation in a language with several dialects. Here's an example:

The english printer Caxton (late 15th century) tells the following story in a preface to one of his books to illustrate how much the English language had changed in his time: Two friends of his wanted to go to Flanders for business purposes. Because of unfavorable wind they had to wait. Since they were low on provisions they walked up to a nearby farm where they asked the farmer woman for some "eggs". Whereupon she became very angry and said that she was not French. So the two gentlemen asked her for "eieren", and got their eggs. So, Caxton continues, what is one to say today, "eggs" or "eieren"?

A Dane trying to communicate with a Swede today (yesterday, rather, the present-day generation have not been watching Swedish TV as pre-cable-TV generations did) will be aware of some general rules of the "we say A - they say B" type

Perhaps you could compare it to Crotian "ije" vs. Serbian "e", but I don't know much about the sociolinguistics there.

The extended version of that expression is "we say A - they say B - you say B, and therefore we'll kill you". I know of three examples.

• The famous one is of course the one from the Old Testament: The israelites do battle with some other tribe and prevail. After the battle straying soldiers are stopped and asked to pronounce the word "shibboleth". Those that say "sibboleth" are killed.

• Most people tend not to believe it about the peace-mongering Scandinavians but we have had some nasty wars. After a battle in one of them, the survivors were told to pronounce a sentence, I forgot which, but the word "hvit"(white) occurred in it. At the time Danes had stopped pronouncing the "h", but the Swedes hadn't yet. I forgot who won the battle, so I don't know who was killed.

• When I lived in Holland a Dutchman told me that in the last war, if the Dutch captured someone they would ask him to pronounce the name of the town of Scheveningen (Du. sxe:-, Ge. she:-). Wrong pronounciation and he was shot.

Obviously some pretty strong feelings are involved here. We also all know the following exchange of letters to the editor:

Non-linguist: Many people people today say B, and they should be saying A. This is horrible and would some linguist expert please set them right.

Linguist: Bla-bla and in the 15th century and on the other hand...

Non-linguist: You traitor! Impostor! You job is to help us! etc.

So how come, if these distinctions are so important in everyday life, that awareness of how "the significant other tribe" speaks is never taken into consideration in linguistic accounts of how some particular language developed?

Below are the stages I think language communities go through when breaking up (for the sake of argument, let us assume the community breaks into two):

1. Peace and quiet. Diglossia. An elite knows the rules of the language (since it retains the knowledge of that people by memorizing it, they have to be particular about grammar and inflection). Over time, because of phonetic developments, inflectional paradigms become more and more opaque to the non-elite speakers of the language. Since they don't know the intricate rules of derivations, the inflectional paradigms look suppletive to them, as if built on several opaquely related stems, and they make frequent, but unsystematic paradigm-levelling mistakes.

2. Crisis. The elite is killed or loses authority.

3. Reorganization. Groups of the tribe generalize the opaque paradigms on the basis of different stems in the perceived suppletive inflectional paradigm. A shibboleth, "we say A - they say B" based on the relation between the surviving stems has now arisen in the two languages.

4. War. Purges. Each of the new tribes undertake a purge in their language based on the "we say A - they say B" relation. Thus, a phonetic sound change (as linguists will later see it) spreads from inflectional paradigms to the rest of the language. But some words escape detection and survive in their old form. (To surviving members of the elite, this whole process appears as a terrible pidginisation and - later - creolization).

5. Much later. Linguists ascribe that particular difference between the two languages as caused by autonomous, independent change in one (or both) of the languages.

Here is an example. The Old Icelandic verb "bjóða" (to bid) of the IE eu/ou/u ablaut type.

By stages:

1. Old Icelandic inf. bjóða, (South Eastern Norse *bjúða?), pres. ind. 3.sg. byðr

2. Tension arises in the Kalmar union. The Swedes do not want to be part of the hopeless Danish venture of pushing the Germans away from the southern shores of the Baltic.

3. War. Reorganization. In eu/ou/u ablauting verbs the Danes generalize the pres. ind.: at byde, byder; the Swedes the infinitive.: att bjuda, bjuder.

4. Purge. The ju/y shibboleth spreads outside the paradigms: Da. dyr, Sw. djur "animal".

5. Linguists claim a rule ju->y in Danish.

Another example: linguists claim there was a rule in North Germanic #wo -> #o, #wu -> #u (loss of w before o and u). Example Old Icelandic:

The interesting thing is that the we/wa/u/o alternation can be explained in two ways, both as a result of the above rule and as the regular result of the ablaut series wer-/wor-/ur-. So perhaps the shibboleth mechanism is a better way of understanding the loss of w in North Germanic than a phonetic rule? On the other hand, the spreading, as regards nouns (e.g. Da/Sw ord, Eng. word, Ge. wort), could be understood also as paradigm levelling (the North Germanic form from gen. urdó-), although of course the noun paradigms have been levelled in the oldest documented Germanic languages (but nothing prevents us from claiming that Proto-Germanic had retained a we/wa/u ablaut in noun inflections).

This is all Scandinavian stuff and probably of minor interest to outsiders. But suppose that ablaut series of early IE -ke-/-ko-/-k-, -ge-/-go-/-g- turned (via ke -> kye -> che -> tse -> se, ge -> gye -> je -> dze -> ze) to se-/ko-/k- and ze-/go-/g-. Suppose that a shibbolethisation took place: two groups each levelled the paradigm, one based on -k- and one based on -s-, and then the -k-/-s- opposition became a shibboleth and spread within each group outside the ablaut series. Now we have an explanation for the kentum and satem groups. On the other hand we now have to assume some early political conflict between the kentum and satem groups which makes little sense since the kentum languages are geographically peripheral to the satem languages.

A very important shibboleth is the guttural sound. This is of course not a linguistic term, but it is very important to non-linguists. Example, quoted by memory from a TWA 1948 guide to Europe: "The Danish language, being a Teutonic tongue, is a very guttural language". I was of course saddened to hear my beloved mother tongue described thusly, and wondered to myself which of the phonemes of Danish were guttural. The uvular r that we have borrowed from German and French? But then the French language would be guttural, and I felt this was not what the authors intended. The glottal stop? But then Cockney would be a guttural dialect and Hawai'ian a guttural language. This could not be right either. The German "ch" then? But we haven't had that sound for about a thousand years. Finally I realized that the statement was not meant to analytical at all, it had nothing to do with observable reality, it was a synthetic statement; Danish, being Teutonic, was necessarily guttural, since Teutonic languages are guttural. Teutonic, BTW, is a non-linguist term which means non-English Germanic.

And then I wondered: Didn't the Dutch generalize the "voiced ch" pronounciation of "g" and the English abandon that same pronounciation of the letter group "gh" at a time where the two nations were at war? Didn't Chomsky with relish proclaim that although the English speakers seemingly did not pronounce "gh" they really meant to deep in their deep phonetic structure? And didn't it take two world wars to make the Germans accept the laryngeal theory? (You may object here that considerations of early linguistic origins did not enter into the picture in the world wars. Some will disagree.)

There is a related phenomenon in Danish that results in some curious retrograde development. A development k -> ky -> ch is considered normal. In Danish the opposite development has taken place. In Norwegian and Swedish there has been a development k-, g- -> kj(ky)-, gj(gy)- -> tj(ch)-, dj(j)-, sk- -> skj(sky)- -> (sj)sh- before front vowels. A similar development took place in Danish, as reflected in 19th century spelling (Kjøbenhavn, today København, Copenhagen) and preserved in the dialects of Bornholm and North Jutland (djæss "geese"). In the beginning of the 19th century there was a sizeable German-speaking minority (one third) in Copenhagen, but towards the middle of the century relations grew increasingly hostile between German- and Danish-speakers, ending in the wars in 1848 and 1864 over Schleswig after which time all things German were extremely unpopular in Denmark.

Palatalization wreaks havoc with ablauting inflections e.g. "to cut" in Swedish:

to remedy which some paradigms are levelled e.g. "to do"

It seems that the sh- pronounciations by the Danes was identified as German (as in Holland), and consequently the sk- and hence k- and g- pronounciation was felt to be Danish. In spite of intense Scandinavism, the complicated Scandinavan inflection paradigm could not be restored, k-, g- and sk- prevailed everywhere and this phenomenon is today an important shibboleth (in the sense of tentative phonemic translation rule for a Dane speaking to a Swede or Norwegian and vice versa; there is plenty to differentiate the languages these days, and we don't kill each other any more). The effect of this was an "unnatural" development in Standard Danish before front vowels kj -> k, gj -> g. This was good for some paradigms:

but not so good for the levelled ones:

A famous example, known by any teacher of Danish, of a word that got caught in this upheaval is the word sky "aspic". The word was originally borrowed from French jus "juice" and pronounced shü (zh -> sh, there are no voiced sibilants in Danish). But the word was mis-identified in the linguistic war as being related to Danish sky "cloud" and therefore had to undergo the change sh- -> sk-.

The point of this last example is that speakers in the heat of a linguistic war sometimes mis-identify the words to reform and therefore generalize the wrong way. Could this be a way to explain away those forms in IE that have necessitated the k2, g2 etc series in IE?

Some more examples:

	sg.	pl.
Old Norse	nom. hest	nom. hestaz, acc. hesta
Danish	hest	heste
Swedish	häst	hästar

Danish has generalized the acc.pl. Swedish has generalized the nom.pl.

Parallel example from romance:

	sg.	pl.
Latin	nom. populus	nom. populi, acc. populos
Spanish	pueblo	pueblos
Italian	popolo	popoli

Spanish has generalized the acc.pl. Italian has generalized the nom.pl. But did -s/-i become a shibboleth?

I have seen various phonetic explanations for the Italian development -s -> -i. None of them has convinced me. Is this a case of: "If they say -s, we say -i"?

Pasting over shibboleths

This is the Danish (long) vowel system:

This is the Swedish (long) vowel system:

Actually there is a difference between the two vowel systems. They are not really as presented here, except for the written language (with the substitutions æ/ä, ø/ö) . The left hand side of the Danish vowel system has been twisted clockwise: a->ä, (and I won't do an approximation of rest of the changes on the left side, since i remains unchanged, and therefore that side is squished together); the right hand side of the Swedish system is twisted counter-clockwise (i becomes (at least in Stockholm - and this is not textbook stuff, but my own opinion) something close to a Russian "y") and u->y, o->u, å->o, a->å (approximately). This makes every wovel of the system a shibboleth between Danish and Swedish.

In 1658, at the peace of Roskilde the eastern Danish province Scania (Skåne) (with Halland and Blekinge) was ceded to Sweden. Since then, the Danish dialect of Scania has been replaced by Swedish. Also, for a long time, there was a guerilla war going on.

About 150 years ago, the railway from Stockholm arrived in Scania. Until then, Scania had been isolated from the original Sweden by the barren province of Småland (remember The Immigrants?). Then the northern Scanian dialect which came into direct contact with "real" Swedish.

The northern Scanian dialect now developed diphthongs and triphthongs (y(->?yi)->øi, å-(->?åo)->äåo, please note: approximately). In my opinion, the purpose of these diphthongs and triphthongs are to paste over the shibboleths. Don't forget how dangerous a shibboleth can be: By just pronouncing a word you are coming down on one side or another. Here's an example (using Danish vowels in the notation):

• If you say: [skån:e], you declare yourself to be Danish (or "a true" Scanian)
• If you say: [sko:ne], you declare yourself to be a ("true") Swede
• But if you say [skäåone], you are avoiding the issue. But you are also creating a sneaking suspicion regarding your loyalty.

Loyalty and Trade Languages

A long time ago somewhere I found a typology of societies that I liked. I forgot where. These thoughts of mine are actually some ten years old, but I never bothered to write them down then. This I why there are no references here. Supposedly, this typology is famous. Anyway, here goes: a typology based on wealth.

• Just about anybody has just about nothing. Not very interesting.
• A few people have a lot. They exchange stuff with their equally wealthy neighbours. They are surrounded by a large retinue of enforcers , concubines, court poets and family. The relations between people are regulated by gifts. There is a strict code of honor, loyalty and obligation.
• Everyone has a lot. Everything can be bought and sold. No honor. No obligation, except by contract.

Yes, I know, types II and III look a lot like Marx' feudal and capitalist stages (let's disregard type I; it's boring, nothing happens). Maybe they are the same. The point is: can we use them to characterize languages?

The difference between these societies is based on degrees of wealth.

What is the basis of wealth? In my opinion, it is the quality of the access roads (I hesitate to use the word infrastructure, since that is a rather misleading denotation for a sea lane or river). In the classic situation, that is before the invention of the train and the truck, the mode of transportation was by water and the access roads were rivers and (later) sea lanes.

There is always hostility between type II and type III societies. They don't understand each other. A type II society sees a type III as lax and semi-criminal, without sense of honor and duty. A type III society sees a type II society as despotic, aggressive and oppressive. Conflict seems inevitable.

Occasionally, a river is opened for transportation, a mountain pass is opened or a whole new mode of transportation is invented. Since transportation is the basis of wealth, this may cause a society, or parts of it, to change from type II to type III. This is usually an ugly sight, given the inherent hatred between the two types.

French and Dutch: II and III?

In the middle ages, the Flemish markets thrived. There were also very unpleasant conflicts between the French-speaking nobility and the Dutch-speaking merchants and artisans. You might also include the English coming to market as, at least, tentatively Dutch-speaking.

There is a famous (Dutch) study (the name of which I also forgot) on the development of Dutch -ui- (so written). Supposed the development was u: ->ü -> üi (and further, which need not interest us here). The study found (except for the fact, known to all Dutch, that there are relic areas with u: and ü:) that some areas had huis, but mu:s. This was ascribed to a sociological effect: "mouse" belongs to a local, private sphere, and "house" to the public sphere, and is therefore changed to fit the prevailing taste coming from the prosperous Western Netherlands.

Maybe there is another way to interpret the supposed sequence u: -> ü: -> üi. We know that ü: for u: pronounciation (actually part of whole left hand side, clockwise shift, as in Swedish) originated in Normandy, whence it spread to French in general. The original Dutch pronounciation, as in Plattdeutsch, Dutch being originally Lower Franconian, of the vowel, is u:. We now have a perfect shibboleth: French inluenced ü: vs. "good old" Low German u:. The conflicts, as usual ended with a muddled compromise, the result of which is Belgium today. Therefore: perhaps the diphthong üi is a "pasting over" of the u:/ü: shibboleth?

Do traders paste over?

It would seem that the willingness to paste over shibboleths is a prerequisite for trade and hence for traders. So the question is: are there "pasting-over" languages in Europe? Let's look at vowel systems. We want a vowel system with a lot of glides, but not just that, we also want "hypocrisy" in the language: the speakers of the language should be claiming to be pronouncing pure "straight" vowels when they are actually pronouncing glides.

I knw of three such European languages: English, Dutch and Plattdeutsch. Strangely enough, the communities which speak them are all trade-oriented.

Reaction: straightening glides?

Plural -s - an inflectional shibboleth?

English has generalized plural -s everywhere.

Dutch has -s in bisyllabic words ending in -el, -en, -em, and in words of French origin. But in the glory days of the Netherlands plural -s spred to monosyllabic words: "arms" vs. modern "armen". But Dutch poets tried to keep even the old -n termination of dative plural.

Plattdeutsch which since the decline of the Hanse has sunk to the level of substandard variety speech in Northern Germany has plural -s approximately where Dutch has it.

German has -s in a few words of Plattdeutsch, and in words of French (and English) origin.

French has (used to have) -s everywhere in the plural (after Old French having a more complicated two-case system).

Obviously trade languages prefer the -s plural, since they are "creolized". When you speak a foreign language, you want one with one regular plural ending.

Before the St. Gotthard pass opened, trade passed thru the Northern Fench towns. This must be what "pidginized" French, so that it has only one plural.

About music

You can sing a tune in a major key. It means you are glad.

What's it all about?

This has become rather long. The questions I wanted to ask are these:

• Is shibbolethisation a valid way to describe processes of linguistic change?
• Should students of historical development of a language take into account developments in that language's favorite foe (assuming they spoke a understandably similar language)?