Étienne de La Vaissière’s Acta Orientalia article “A Military Origin for New Persian?” (open access) attracted my attention because of my long-standing interest in Persian and its history. The abstract:
The question of the transition from Middle Persian to New Persian has been hotly debated. This article attempts to answer two questions: who spoke New Persian before it was put into writing in the middle of the 9th c.? This social group is identified with the soldiers of the armies of Abū Muslim, i.e. peasants from Marw and their descendants. They came during one century to the forefront of Abbasid political and administrative life and imposed their specific dialect as a political language, in the shadow of Arabic. The second question is: what could have been the origins of the spoken language in the Marw oasis of the first half of the 8th c.? The article tries to demonstrate, on a much more tentative basis, that the demographic history of an oasis twice manned by soldiers from the South, first Middle Persian-speaking ones and then Arabic ones, both groups added to the local, Parthian-speaking population, is well reflected in the unique combination of Middle Persian, Arabic and Parthian characteristic of Early New Persian. Early New Persian is the language of 8th c. Marw, or more generally Outer Khurāsān. This Marw hypothesis, based on the presence of Parthian vocabulary, is however very cautious, as nothing is known of the grammar of spoken late Middle Persian and many of the linguistic differences between Middle and New Persian might have evolved separately in different historical processes.
Ignorant as I am, I find the idea plausible, and I like his modesty:
For the time being, the argument must rely primarily on historical analysis. The grammar and phonology of spoken Middle Persian are not known, as its written forms are largely archaizing—this represents a major limitation. With regard to vocabulary, I had hoped that a sociolinguistic analysis of Early New Persian, particularly of its distinctive Parthian layer, might yield results. Unfortunately, very little research has been conducted in this area.
The final paragraph:
In the absence of more data, especially from earlier texts, many complex scenarios remain possible for the linguistic situation in early 8th-century Marw. More comprehensive studies of Early New Persian vocabulary—or new textual discoveries—are needed to confirm or refute this plausible hypothesis.
But the details he provides are intriguing and help fill out my picture of the situation in that time of rapid change.
Modern Persian is clearly not a creole (unless you’re John McWhorter), but I was reminded of a real Army Creole (as opposed to the one in The Right Stuff*):
https://en.wikipedia.org/wiki/Nubi_language
* https://en.wikipedia.org/wiki/Army_creole
DE, “military koine” in Arabistics.
[Ferguson’s?] hypothesis that vernacular Arabic has developed from military koine.
____
“Although I am not a linguist…” (from the article)
Unexpected:)
Also on the LLog, but without any comments so far.
“military koine” in Arabistics
Not KiNubi, though, which is very definitely Not Arabic, just as Nigerian Pidgin is Not English (or even more so.)
I was under the impression when I lived in Nigeria that Shuwa Arabic was a creole, but it very much isn’t: it’s a pukka Arabic dialect (much the same as Chadian Arabic), though with a great deal of influence from sundry West African languages.
Actual creolists (or some of them) can get a bit shirty about things like this though. And I do agree with John McWhorter that languages which have spread rapidly and widely, presumably at the expense of “substrate” languages, can do some creole-y things.
In Oti-Volta, I’m sure that there is a strong inverse correlation between morphological complexity and number of speakers, though a lot of that is driven by the relatively simple morphology of Western Oti-Volta, far and away the biggest single branch in terms both of numbers of speakers and of geographical spread. Not sure that the numbers would be convincing without WOV.
Yes, not KiNubi wherever the Ki comes from (I think not from Nubi).
Rather a term similar in form to yours (and designating something analogous to Persian… or partly analogous)
wherever the Ki comes from
Swahili.
The Nubi were favoured by Amin, and have consequently had a hard time since his fall; that’s why a lot of them live in Kenya now.
Modern Persian is not hugely different from Pahlavi in either morphology or syntax, which is why I say it’s definitely no creole. (Obviously it’s borrowed a great amount of Arabic vocabulary.) It is a strikingly simple language morphologically, but the simplification largely antedates Islam.
That is not what the paper is on about, though: I was just contrasting it with an actual army creole, because why not?
@LH, thank you! I was dissatisfied by accounts that I found (they sounded as an attempt to dress up very vague understading in precise words) and made a mental note that the question – I mean, the one the article is trying to answer – is both important and must be partly answerable. And I wanted to read more about it. I’m surely going to read the article and some of its references.
(and Marw is in my list of things “big in the first millenium”. Like Syriac)
David Eddyshaw: languages which have spread widely CAN indeed do creole-y things, but tellingly, most of them do not. In your own neck of the woods, Bantu is morphologically more elaborate than Oti-Volta, despite its having spread over a vast territory once dominated (if the surviving pre-Bantu languages are a faithful sample of the typical pre-Bantu language of the Southern half of Africa, that is) by languages which were in key respects more isolating.
And the most striking thing about the world’s best-known geographically extensive languages families is how the correlation between creole-like linguistic evolution and distance from the URHEIMAT seems non-existent in most.
Greenlandic, Navajo, Moroccan Arabic, Sorbian, Micmac, Nahuatl and Hungarian are, each of them, the (or one of the) geographically most eccentric (=geographically furthest removed from the location of the proto-language) member of its family. In no instance can its evolution be characterized, globally, as more creole-like than that of geographically more central members of the language family each belongs to. Or, tellingly, less creole-like either. Which makes me skeptical of the claim that, if we observe creole-like language change in the case of a language expanding at the expense of others, this must mean that creolization (i.e. pidginization followed by nativization) must be used as an explanatory tool.
Considering the lack of any correlation in most language families between creole-like diachronic change and history of expansion (=distance from the original location of the proto-language), the fact that such a correlation can be observed in SOME language families might be a coincidence, pure and simple. I wonder whether such might be the case for Persian within Iranian.
Bantu is morphologically more elaborate than Oti-Volta
I’d dispute that, actually. I think this (partly) reflects the Bantuist tail-wagging-the-dog misunderstanding of the position of Bantu within Volta-Congo.
The Bantu verb has agglutinated a lot of preverbal clitics as affixes, but in proto-Bantu (and indeed in many modern Bantu languages) the resulting system remains notably transparent. And the noun-class system of proto-Oti-Volta was at least as complex as that of proto-Bantu; so too, with the other major locus of morphological complexity in Volta-Congo, verb-stem derivation.
Even in the modern languages, Ditammari (for example) has eighteen of what are called “genders” by the Bantuists, and Nawdm verb derivation makes yer typical Bantu language look positively isolating.
Phonologically, proto-Bantu had already radically simplified the inherited system, and many (most, probably) modern Bantu languages are very much simpler phonologically than e.g. Kusaal.
There is also the point that complexity doesn’t merely reside in the number of morphologically marked distinctions. but also in things like the degree of fusion and suppletion (and also, of course. in syntax.) “Creolisation” might in principle result in extensive regularisation rather than loss of distinctions. (Indeed, that very thing seems to have happened in Lingala verb derivation.)
Although I don’t agree with your specific example, I do largely agree with your overall point, though.
Even the striking simplicity of e.g. Western Oti-Volta verb flexion compared with pretty much every other branch of Oti-Volta looks like a pretty natural internal development, with many parallels cross-linguistically (essentially, making an imperfective finite form out of a periphrasis with an originally nominal deverbal form, as in, um, English. And Welsh. And Hausa.)
I think the trouble is that assertions about “semicreolisation” tend to end up circular, a vice very evident in some of McWhorter’s pronouncements: we know what creoles look like (according to him) and anything that looks like that must be the result of “creolisation” regardless of whether there is any actual historical evidence or any actual data regarding supposed “substrates.”
Incidentally, I look forward to Xerîb’s view on all this.
Ironically, one comes across the assertion from the Larry Hyman school, who claim that the agglutinative verb structure of proto-Bantu should be projected back to “proto-Niger-Congo”, that linguists working on West African languages have frequently misanalysed preverbal affixes as proclitics …
Puts me in mind of that horror-show post that DM linked to the other day, about all phonology being syntax. One of the links led to a paper which apparently claimed that the polysynthetic structure of Tlingit verbs is all syntax too …
A positive development. Maybe even Chomskyites will one day come to see that it’s Constructions All the Way Down.
Maybe even Chomskyites will one day come to see that it’s Constructions All the Way Down.
Is Construction Grammar the Unitarian Church of syntax?
Construction Grammar is the Panpsychism of syntax: meaning inheres in all levels of syntax, down to the most elementary …
David Eddyshaw: Incidentally, the case for the Persian creole-like features (=loss of grammatical gender, of any agreement between noun and adjective, of any case-marking on the noun) only being coincidentally similar to the changes separating creoles from their source language is made more likely to my mind by the fact that Persian is the LEAST creole-like Iranian language in one key respect: In verb morphology Persian has six suffixes indicating the three persons and two numbers unambiguously. This is unusual among languages of the Iranian family spoken today, where your typical language has perhaps three or four person- and number-marking verb endings (having a single ending for all plural persons seems to be especially common).
(I learned the above fact at a conference on Iranian linguistics, so it will take a while -i.e. until the proceedings are published- before I can refer you (or anybody who is interested) to a written source for the above point).
Now, inasmuch as creoles differ from their source languages in losing all morphology, including person-marking morphology, it may fairly be said that in the case of this feature Persian is the least creole-like of Iranian languages.
Y: Chomsky’s linguistics as a Church…you are not the first to observe the parallel:
https://specgram.com/LP/40.metalleus.fragment.html
Etienne, I think your thought experiment is a bit of a strawman. Why would there be a correlation between distance of migration and creole-like features? Far-spread languages can move into little-populated places (in fact that is arguably easier and therefore likelier), where they won’t be subject to much language contact.
Y: Yes indeed, far-spread languages *can* move into little-populated places. But if there was any truth to the claim that creolization and untutored L2 acquisition are fundamentally the same process, we would expect languages which have gone through the most instances of language shift (which, in pre-modern times, did not involve any significant learned influence), i.e. the ones which spread furthest from the location of the proto-language of whatever family they belong to, to be the most creole-like.
This is practically never the case. You will notice, incidentally, that several of the instances I quoted upthread involved the spread of language family members from low population density to high population density locations (Nahuatl, Hungarian, Navajo). I am unaware of anything in their diachrony indicative of more creole-like language change than what could be found in any randomly-selected Uto-Aztecan, Uralic or Athabaskan language, respectively.
Mooré, by far the most widely spoken Oti-Volta language, and one which has almost certainly attained its present geographical spread only over the past four centuries or so, is not by any stretch the most innovative Western Oti-Volta language. Traditional history suggests that this was not at all a case of migration into a thinly-populated zone, but a takeover by a relatively small group of invaders/migrants.
The title of “most innovative” WOV language surely goes to Dagaare, geographically separated from the rest of WOV by a migration to the west which traditional accounts suggest only happened at the end of the eighteenth century. (I don’t know what these accounts say about previous inhabitants.)
Dagaare has shed most of its verb-derivational morphology and undergone quite a bit of reshuffling of noun flexion, but all of that is fairly obviously the consequence of some quite far-reaching phonological changes. (Phonologically, it’s the French of Western Oti-Volta.) On balance, all this has produced a morphological system which is more complicated than before, because of a significant increase in synchronically-opaque irregularities.
Didn’t Navajo spread after everyone except the Hopi had starved? (The other two examples definitely hold, though.)
I don’t understand why Sorbian.
The distance to Morocco is greater than to Egypt. But I don’t think that means “greater degree of dilution of Arabs”.
Egypt speaks Arabic, and many in Maghreb… I say “Maghreb” because I don’t know who of Berbers when came to where in Morocco from where, many in Maghreb speak Berber and many (Tunisia) shifted when the French came.
But what does that tell? Say, someone is given two Arab countries. One is the size of Egypt (and she doesn’t know its structure, anything but the name and size). One is same size but named differently “Morocco and a piece of Algeria” (and she does not know how it is structured). And she’s told that the first speaks Arabic and the other Arabic and Berber and that it is 21st century (15th Hijri century). Can she say anything about these two Arabics? I don’t know what can be said:)
But Egypt was not Arabophone when the Muslim army invaded Morocco.
For an attempt to calculate when Muslims reached majority in Roman Africa refer to Conversion to Islam by Richard W. Bulliet (google it). Of course wildly imprecise (it is based on known names). The 9th century. I don’t know if anyone tried same for langauge shift.
@David Marjanović: If you are right about the American Southwest being depopulated when the linguistic ancestors of the Navajo arrived, then either take any other Southwestern Athabaskan language or any one of the Athabaskan languages of the Pacific coast instead.
Drasvi: Sorbian is (along with Slovenian) the Westernmost Slavic language. Neither expanded over uninhabited territory, and neither looks at all creole-like in its diachrony compared to other Slavic languages (with their preservation of the Proto-Slavic dual they are in one key respect the LEAST creole-like Slavic languages).
And yes, Arabic reached Morocco before Egypt shifted to Arabic, but the further away from the Urheimat you are the more L2 speakers are potentially liable to be part of the founding population of the group of what a few generations later will be L1 speakers of the expanding language…
Well, following the Migration Period, the population in the previously Germanic-speaking places does seem to have been very thin – barely dense enough to transmit a few place names (Silesia, Vienna – and the latter is not Germanic).
David Marjanović: Also Dresden. EDIT: oh, you meant the opposite. EDIT2: as in Dresden is Slavic.
Can’t say I like this account.
1. People speak with confidence about what languages East Germany spoke in, say, the 4th century. That’s unscientific.
2. “Germans went to the west and also to the east and also to the south and Slavs came to the empty land”
Is that normal? If not, I would love to see a detailed model.
(I know about the idiotic argument between Slavs (Poles in particular) and Germans regarding “who came first” to Poland. Of course among Slavic-speaking lands in 600s Slavs came “to” some and “from” others and of course I don’t need Germany to be “from” rather than “to”. But 1. annoys me and 2 perplexes me)
drasvi: I meant that the name “Dresden” is etymologically slavic? nothing more, nothing less? If that’s what you were refering to? I’m not sure. I’m kind of confused.
If you are right about the American Southwest being depopulated when the linguistic ancestors of the Navajo arrived…
Parts of the Southwest, but I don’t think the part I live in (north-central New Mexico) was depopulated.
the further away from the Urheimat you are the more L2 speakers are potentially liable to be part of the founding population of the group of what a few generations later will be L1 speakers of the expanding language…
So this theory that you’re providing evidence against is: In previously populated places where migrants settled and people ended up speaking the migrants’ language, there should have been lots of descendants of people who learned the migrants’ language without being taught it in school and thus spoke a creolish form, and that should have influenced the language as a whole to make it more creolish?
That last step seems very strange.
The situation is very similar to the situation in my town, which has a majority with Hispanic ethnicity but English L1. Anglos started settling after 1846 but started seriously suppressing Spanish only after WW II. The difference I see is that of course English was taught in schools, which was part of the suppression.
Toponyms may not tell the whole story. “London”, for example, has a good Celtic name, and Celts were certainly there before the English, but the city seems to have been pretty much completely abandoned long before the English arrived.
Keen as I am on Welsh irredentism, I think the case for reclaiming Llundain is a bit on the weak side. Perhaps if we’d looked after it a bit better in the first place …
DE’s comment led me on a brief wiki-excursion that ended in my unsettling realization that the “Tower of London” is only about 3.5 stories, and about as wide as it is tall. WTH? I’m going to start calling our house the Tower of Ryan.
Also, it’s on “Tower Hill”, a name of which I as an Illinoisan can fully approve. It reminds me of Grand Ridge, IL.
Can’t say I like this account.
1. People speak with confidence about what languages East Germany spoke in, say, the 4th century. That’s unscientific.
2. “Germans went to the west and also to the east and also to the south and Slavs came to the empty land”
It may be that you don’t like it, but it’s the account that still seems to be generally accepted by historians. AFAIK, it’s based on the reports of Germanic tribes living in the area in contemporary Greek and Roman sources, on toponyms (i.e., Silesia being named after the Silingi, a sub-tribe of the Vandals), and on the fact that Slavs and Slavic names suddenly start to show up all over Southern and Eastern Europe from ca. the 6th century, while there is almost no linguuistic evidence for them before that. Other interpretations are possible, but while I remember seeing serious discussions about the location and extent of the Slavic homeland, I haven’t seen any serious challenge to the narrative that the territory where Slavic languages are spoken greatly expanded as the last act of the Great Migrations.
I once read an archaeological study about the settlement patterns in the Angeln peninsula in late antiquity and the early Middle Ages. Apparently after the 5th century, the peninsula seems to have been largely uninhabited; there was no continuity between the settlements of the late antique period and the medieval settlements after the Jutes moved in (except for a few names that seem to date to the first centuries CE). There apparently was hardly anyone living there for the first few centuries of the Middle Ages. Although the study concentrated on Angeln, it was mentioned that there were similar patterns in the area north of the Elbe, except that the Slavs seem to have moved in there already in the 6th century.
Anyone interested in the topic should check out The Early Slavs: Culture and Society in Early Medieval Eastern Europe, by P.M. Barford.
There also seems to be a strange hiatus in the Runic record further North in Denmark, Proto-something up to AD 400, mostly in Jutland IIRC, and then all over the country from 700 or so in the younger futhark. I’ve seen speculations that the Danish tribes went to Svealand to be nobility there, but got thrown out or became so numerous they had to colonize their old neighborhoods/petty kingdoms. (Presumably there were still people tilling the land in the intervening centuries, but they couldn’t be bothered with runes. Too poor or busy or something). Supposed to explain dǫnsk tunga and danmark since it really was the southern border of the Danish realm for a time.
Specifically, *siling- borrowed completely regularly as *sьlęd͡z- as mentioned, and on the southern side the Longobardi settling in eastern Austria and Hungary, then packing up and moving through Slovenia into northern Italy (according not only to somewhat later historians but also archeology), where they pumped the place full of (apparently North) Germanic names and other words.
As Bede claimed in the 8th century.
Also Proto-Saami *tańćə “Norse” – turn *dani- into *danjV- somehow, pass it on to that mysterious Germanic branch with fortition of *w and *j that left a bunch of loans (with p and t) in Finnic, and get that into Saami.
But I thought “Denmark” was from Charlemagne’s Danish march?
One thing to keep in mind is how lightly populated Europe was back then. Josiah Russell’s estimates (described in what source as “quite speculative, but not insane”) have 3.5 million human beings spread through all of what we would now call Germany/Scandinavia as of A.D. 500, then going down to 2 million by 650. That’s consistent with large chunks of territory that could easily have supported human life in terms of their agricultural potential etc. simply being vacant.
Regardless of total population, we know from archeological evidence that a lot of land that had been under cultivation in Roman times had fallen fallow by 650. The Younger Fill is a layer of topsoil that washed away and accumulated around river mouths during the Migration Period.
Charlemagne’s Danish march. That’s one story. Whoever the speculator I referred to was, they didn’t buy it. And anyway, if there were Danes in Denmark it wouldn’t be the march innit, that would be like Holstein maybe. But metonymy.
(ulr’s peninsula is just Angeln, this speculation was about all of Jutland).
@Etienne,
Sorbian, Slovenian: I mean, they are not really far from where Slavs came from…. depending on where they came from, of course. And one can’t be sure the languages they were in contact with (Gothic?) were different from langauges of East Germany.
Morocco: I think your model is that of a snowball, slowly rolling. And I think they threw this snowball (of Arabic speakers, some L2): an army, conquering lands thousands kilometers away. In the Maghreb (Andalusia is Maghreb too) with converted Berberophone armies.
Also the Hilalan invasions of the 11th century: Egypt sent HIjazi Bedouin tribes to subjugate the Maghreb. Somehow, Europeans know about the first Muslim conquest of North Africa but not about this second conquest (is it “Arab internal affairs”? Conquests of Mongols and Timur are known in Europe, even in Muslim lands…). This was a large snowball and I don’t even know which conquest was worse for (Christian, Jewish or Muslim) locals.
Thus Arabic dialects of Maghreb are either “Hilalan” or not, based on certain (phonological) variables. The division basic for the dialectology of Maghreb (but it happens that values of these variables differ for men and women in some cities). Despite this, of course there have been convergence.
Can it be so that the number of L2 speakers greatly exceeded the number of L1 speakers in the Andalusian snowball of Arabic users – and that it didn’t exceed the number of L1 speakers in the Egyptian snowball – and that it was so because of the greater distance to Morocco and Andalusia?
Yes, I can imagine that. But I don’t want to make predctions without data, because of all other variables. The Nile valley is enormous. I think Egypt is some 90 million people in the 21st century (by memory, and by memory some 10 in Tunisia, 5 in Libya, 23 in Syria). I don’t know what it was, but I think a very large crowd of Greek and Coptic speakers…
@Hans, my 1 and 2 are not about Slavic.
Slavic can have to do with them, if
– East Germany was mixed or even a system of people speaking several langauges
– Slavic one of them
1. and what Romans tell about known languages of the empire? I think they’re a very poor guide to those.
And “east of Elbe” is not even their empire.
They list names of dozens peoples. Those are usually referred to as “Germanic tribes”. Which I think is usually understood as “Germanic speakers” (with “Germanic” traditions) even though often nothing is known but the name.
Some of these names can be read in Germanic (langobardi). Some also names of certain peoples who invaded the empire in a different century and are known to speak Germanic.
But how anyone can be confident about others and also about languages of East Germany?
2. empty Germany sounds strange. I don’t mean it can’t be so.
But IF it is so, I need a detailed explanation.
@drasvi: I can only tell you what I have read in general accounts of history; this is not a period about which I have read more than some general (and mostly popular science) accounts, so I can’t even point you to scientific articles by historians or archeologists. Others have pointed out some archeological evidence for depopulation during the period of the great migration; maybe you can start with the sources they quote and take it from there to confirm or overcome your doubts.
Look, for several centuries there one of the most important things going on in Europe was the so-called Völkerwanderungen, in which largeish groups of Germanic origin relocated themselves via migration-or-invasion to other parts of Europe (and sometimes North Africa) that had not hitherto had any Germanic-speaking residents. So that by definition means there would be fewer of those Volk remaining in the locations that had been Wandered away from. If the source territory was already overcrowded, you might not immediately notice, but there’s no particular reason to think it was. And a particular “lifestyle” (in terms of mode of agriculture, what is necessary to support cultural functions people value etc.) tends to require a minimum population density just as much as it can’t exceed a maximum population density without a crisis. So if you have let’s say 30-40% fewer people over territory that was at that minimum-or-perhaps-optimal density w/o being overcrowded it won’t be that it becomes uniformly less densely populated, but that the portions of the territory that are actively occupied will shrink and the rest will be vacant/fallow from a human-use perspective until things change. Which could ultimately include new arrivals, such as Slavic-speakers, taking advantage of farmable land with no one farming it.
It is not exactly doubts. I can’t make sense of it…
P.S. Tacitus speaks of nomadism among Germanic peoples…
Drasvi, I can offer a parallel. During the colonial period in North America, the area that is huge parts of present day Ohio, Indiana and Illinois were depopulated by a dynamic encapsulated as the Beaver Wars. The Iroquois, with superior access to European weapons, implements and supplies, laid waste to the area and made it an occasional resource extraction and hunting zone for themselves. This fed a cycle of increased access to beaver, allowing the purchase of more ammunition and metal implements. They still primarily lived in upstate New York, where they could maintain their connections to Dutch and English traders. They overran or incorporated most speakers of Iroquoian languages. Other native groups clustered around French trading routes north of the Great Lakes and along the Mississippi and its portages to the Lakes.
One could imagine something similar in northern Europe in the Migration Period, where Germans and Slavs clustered south and west on the edges of the empire where they had access to things the empire could sell them, and the northern Germanic peoples remained in Scandinavia in part for the furs and amber they could trade down eastern river corridors. Occasional violent forays into the German interior may have been enough to discourage scattered, weak settlement.
I read de La Vaissière’s article.
Or “La Vaissière’s article”?
I expected a flood of facts and references. To my surprise (pleasure as I was reading it which gave way to dissatisfaction when I read it) I know a large part of both. E.g. I even cite in this thread a [different] book by Bulliet.
I wonder if that means I (not a specialist!) know a large part of what is known. (Or does it reflect his choice of facts and literature? That’s better…)
P.S. I didn’t know Zhang Zhan 2023. ‘Two Judaeo-Persian Letters from Eighth-Century Khotan.’
It speaks about Bactrian “taught” in Marw.
The reference… does speak about someone’s attempt to learn it.
Which often means someone’s attempt to teach it too.
No evidence of such that I’m aware of. The distances, even from the big rivers, are also considerable. But there were a few plague outbreaks, IIRC.
That would have made it easier to understand. Usually models have someone (Huns) or something (desertification) push people from where they are.
Islam did not “push” Arabs from Arabia.. but Arabs are in Arabia. Or if you find there Indians instead, that’s because Arabs hire them.
It’s a large land with numerous “tribes” in it, one would think. Where did each of them go? I think even for a partial understanding the answer is necessary.
Normally I imagine “partial understanding” like this: you tell me “they all joined the Frankish invasion in France. Such and such texts lists them, such and such facts confirm the account”. I understand when, how and why, but I have questions: “why all? Why not a half of the tribes? Why not some people from some tribes?” etc.
But I don’t have even this:/
Well, the Hunnish invasion sounds somewhat like the Iroquois invasion, a group devastating those who oppose while also incorporating those willing. And bringing many along, outside of the area of their previous settlement. At that point, is the previous economy still viable?
Once the land is relatively depopulated, periodic violent forays might encompass hunting parties, groups crossing to other areas and recons in force that make it clear that smallholders are in danger — incursions too small to leave much archaeological record.
Yes, but the Huns were *poof* gone some 200 years before the first Slavs showed up. At that time the Avars were playing that role, but they don’t seem to have gone north of the Pannonian Basin even.
David M.: Also Proto-Saami *tańćə “Norse” – turn *dani- into *danjV- somehow, pass it on to that mysterious Germanic branch with fortition of *w and *j that left a bunch of loans (with p and t) in Finnic, and get that into Saami.
That reminds me that Ante Aikio just uploaded a conference presentation on Sami ć < Gmc. *j to Academia. He pinpoints the location of the Germanic dialect with fortition of *j to roughly current Troms in Northern Norway based on two toponyms, one of them the island of Senja. I’ll have to reread it to see if it says anything of timing, but if I were to guess, from a settlement of Northernmost Norway that was aborted or samified after the crisis of the mid-6th C.
I don’t think he goes into that other fortition, and a (literally) marginal dialect in that part of Norway certainly couldn’t have left loanwords in Finnic. But I do think we should compare and contrast the different patterns of fortition to those attested in Gothic and Old Norse and see if some insights emerge.
But I thought “Denmark” was from Charlemagne’s Danish march?
I like that hypothesis so much that I may have given it more publicity in these quarters than it objectively deserves.
This word makes me think about Franks and their empire rather mechanically.
But I don’t have even this
Welcome to much of ancient history outside the areas that left a lot of written records. You get some mentions by those writing cultures, some archaeological and perhaps linguistic evidence, and you try to derive a picture from that. Not very satisfying, but that’s all we have.
Luobbal Sámmol Sámmol Ánte (Ante Aikio): Two strata of Proto-Norse loanwords in Saami: Evidence for an extinct Norse variety (conference presentation uploaded to academia.edu):
We may speculate whether the lack of Sievers’s law was a dialectal feature or because of the age of the loans.
@Hans, yes.
Ooh, that part is new! Neat.
The Fennicists have known since 1986 that Sievers’s law had ceased to operate by Proto-Northwest-Germanic times. It’s routinely missing in loans into Finnic, and also in High German words like Weizen, heizen, Flötzer… I posted the Google Books link once here, in a thread so long Google refuses to search it apparently, found it once or twice again and never since. Hm. Anyway, I’m pretty sure it’s by Jorma Koivulehto. (And it’s in German.) Google Books itself finds this, which isn’t what I mean but is by Koivulehto and says the same thing in a less pointed, more long-winded way.
Hans: It may be that you don’t like it, but it’s the account that still seems to be generally accepted by historians. AFAIK, it’s based on the reports of Germanic tribes living in the area in contemporary Greek and Roman sources, on toponyms (i.e., Silesia being named after the Silingi, a sub-tribe of the Vandals), and on the fact that Slavs and Slavic names suddenly start to show up all over Southern and Eastern Europe from ca. the 6th century, while there is almost no linguuistic evidence for them before that. Other interpretations are possible, but while I remember seeing serious discussions about the location and extent of the Slavic homeland, I haven’t seen any serious challenge to the narrative that the territory where Slavic languages are spoken greatly expanded as the last act of the Great Migrations.
I have a long backlog of interesting stuff from the last few months. One of them is these two papers from last month (by courtesy of Dmitry’s Facebook news service):
Gretzinger et al: Ancient DNA connects large-scale migration with the spread of Slavs, Nature (2025):
Schultz et al: Ancient genomes provide evidence of demographic shift to Slavic-associated groups in Moravia, Genome Biol 26 (2025):
David M.: The Fennicists have known since 1986 that Sievers’s law had ceased to operate by Proto-Northwest-Germanic times.
His point (as I read it) is not that Sievers’s law had been abolished, but that these loans show no sign of it having operated on the source. There’s another set of loans (the other stratum, presumably) that do show the effect of Sievers’s law. Aikio:
Nobody is denying that Sievers’s law, and even its converse, happened at some point in common Germanic history. High German, too, has plenty of words that have undergone it (as shown by lack of West Germanic consonant stretching). Maybe it’s my innovation to put the few but disparate sources together and conclude that the law had stopped operating by NWGmc times because some of the loans in question, in both Finnic and Saami, show *ā from PGmc *ē (though I see one *ā is from Latin; blessed indeed are the cheesemakers), and I haven’t seen counterexamples.
Consider another episode of North American depopulation more recent than the Beaver Wars. Many parts of the American West (past the Mississippi but still east of the Divide) have experienced dramatic population decreases over the last century-plus, with 2020 populations being 60%, 70%, or even in a few instances 80% below a recorded local historical peak falling some time in the 1890-1930 range. https://en.wikipedia.org/wiki/Depopulation_of_the_Great_Plains#Counties_with_large_population_losses And there are many many places not on that list because the population decreases were more modest but still dramatic in absolute terms (e.g. the county in Iowa where my paternal grandfather was born, down 47% from 1900 to 2020).
Was this a catastrophe? Well, not from an outside perspective, because it was a predictable side effect of dramatic increases in the productivity of agricultural labor – it now requires dramatically fewer human beings to get optimal production of food from X thousand acres planted in wheat or Y thousand acres of range where cattle are being raised than it did back then. And a well-functioning market economy means that absent new non-agricultural sources of employment in those areas the now surplus-to-requirements workers will predictably migrate elsewhere where new non-agricultural employment is available.
But that economic efficiency has some more negative side effects for those still in the region growing the crops and raising the livestock, as decreased population density means that individual communities can no longer sustain their own schools, churches, shops, medical practices etc., and now the remaining inhabitants need to travel 25 or 30 or 40 miles for what they used to travel 5 or 10 miles for. I suspect that without good written records archeologists of the future might find the traces of abandoned schools / churches etc. easier to identify than the increased economic productivity that had led to them.
The same thing has been happening in Russia for decades (Reddit, Moscow Times, Wikipedia).
I said I didn’t know Zhang Zhan 2023. ‘Two Judaeo-Persian Letters from Eighth-Century Khotan.’
It seems i didn’t know Zhang Zhan 2016: https://languagehat.com/a-judaeo-persian-letter-from-dandan-uiliq/ Zhang Zhan 2023 is a commented edition of two letters from one place, one of which is Zhang Zhan 2016.
(I like the third paragraph, about a stranger who knocked on Duan Qing’s door in the morning)