Having finished The Reconstruction of Nations
This difference was clearly reflected in political nomenclature. A careful observer of early-sixteenth-century Muscovy, Baron Sigismund Herberstein, noted that the Nogays had no tsar (i.e., a khan), but only a princely chief (i.e., a beg). The beg (referred to in Russian as a grand prince, bol’shoi kniaz’) was the ruler of the Nogays. The next in line of succession was the nureddin (a personal name of Edige’s eldest son which evolved into a title), an heir apparent and the second highest title, followed by the keikuvat (a title derived from the name of Edige’s younger son) and the toibuga. […]
The candidates for the four princely titles had to be confirmed in the Nogay Grand Council, known as the körünüsh (korniush in Russian transliteration), which consisted of the members of the ruling house (mirzas), tribal aristocracy (karachis), distinguished warriors (bahadurs), the beg’s retinue (imeldeshes), and Muslim clergy (mullahs). The beg had his own administration (a treasurer, a secretary, scribes, tax collectors) and a council comprising the best and most trusted people. Yet his authority as projected through this rudimentary official apparatus was greatly circumscribed by the powerful and independent mirzas and karachis.
I might note that as a result of the long and at times heated discussion in this thread, I was forced to acknowledge that even I would have a hard time calling keikuvat and toibuga English words, even though Khodarkovsky drops the itals after first mention and talks about “the keikuvat” just as though he were talking about “the vice president.” I still think such use is a good rough-and-ready criterion, but each case has to be examined on its own; if a lot of people started writing about nureddins and keikuvats in English, these sentences would be early attestations for OED citations, but by themselves they do not create new English words.
Similar situation with Oirats. Their rulers were not descendants on Genghis Khan in direct male line (but they sure had great khan’s blood, since their earliest known ruler, Oirat chief Khutuga-beghi* was married to Genghis Khan’s daughter Chichigen)
For this reason attempts by Oirats to gain power or claim title of khan (surely this is an English word now?) were fiercely resisted by eastern Mongols who regarded their rulers unsuitable to rule – only Altan Urug – Golden kin of Genghis Khan were worthy enough to be khan.
Later, in 17th century, according to a new doctrine developed in Mongol tradition, Dalai Lama was thought to have power to grant this title to devout Buddhist rulers. Thus, Galdan-Boshigt-Khan of Dzunghar Khanate and Ayuka-Khan of Volga Kalmyk Khanate became khans.
* Note that beghi is a same Turkic title as “the beg” you mention.
the soviets were careful to not use the term colonialization, wasn’t it always osvoenie sibiri, ili tam otkrytie, if to think about though sprashivaetsya, kto kogo otkryl-to, well, to bylo in medieval the middle ages, and this is the modern age, so now they are newly capitalistic and the interpretation becomes imperialism and colonialization which doesn’t sound like a very novel idea, breakthrough something
i liked more their eurasia and mirnoe sosushestvovanie cohabitation, brotherhood etc. explanations, soviet propaganda of course, but wasn’t it nice or at least original
The term colonization was definitely used in Russian sources. In tsarist times, it was thought to be a good word and Russian colonization was something to be proud of – a part of advancement of European civilization and culture into savage lands.
In early Soviet period, the word gained a negative connotation of subjugation of freedom-loving peoples and was now used to condemn Tsarist colonialism.
The change came in 1950-60s, I think. Condemnation of Tsarist colonialism became risky, because it could put in danger legitimacy of Soviet borders, particularly in the Far East where Maoist Chinese leadership claimed that pretty much all of Siberia and Mongolia too were stolen by Tsarist Russian colonizers from China.
So Soviets came up with explanation that there was no colonialism and no conquest involved in the process of acquisition of Siberia by Russian states. It was all just peaceful exploration and osvoenie (assimilation, mastery, acculturation).
“So Soviets came up with explanation that there was no colonialism and no conquest involved in the process of acquisition of Siberia by Russian states. It was all just peaceful exploration and osvoenie (assimilation, mastery, acculturation).”
whatever was the reasons, couldn’t it have been for once something progressive and good for everyone concerned, in history, even if just in the explanations, it just has to revert to something ugly and “truthful”, i think there were many advances in many areas of people’s social life during the socialist period, it’s just such a pity that everything what was good, together with of course a lot of things which were not, has to be lost too soon
Does the book describe the social structure of the Bashkir society too? Another one without a top-line Genghizide claim, and with a wonderful variety of social structure terms, mescheryaki / mishari, teptyari, bobyli of all languages and faiths.
Read – the word “discovery” is rarely used for Russian exploration of Siberia, and then only for its most remote corners. The usual word is покорение, subjugation or conquest. Certainly that’s the word used for the proper Siberian Khanate.
It’s a relatively widespread point of view in Russia that Russia itself morphed from the colonial dependence from the Steppe, from its own subjugation. But of course the preferred term for the Steppe Rule is Иго, the Yoke.
SF – of course the word “colonization” existed in imperial Russian, but it had a very specific narrow meaning: “colonies” were settlements of the foreigners, typically of German farmers, most numerous in lower Volga and in Ukraine, and far post-dating the subjugation of the nomads.
As The Reconstruction of Nations remarks, no state can afford to admit that its former borders were illegal, even if (as Poland did) it has relinquished them forever. That would undermine the legitimacy of all actions taken in that time, including private affairs such as contracts and the title to lands.
I was very impressed, by the way, by Poland’s early-1990’s western strategy (unilaterally abandon all claims west of the Oder-Neisse line even though Germany had not yet reciprocated) and even more by its eastern strategy, about which I knew nothing before reading Snyder. Since Poland was a modern independent ethnic state, it would treat Lithuania, Ukraine, and Belarus as equally modern independent ethnic states, even before they actually were — not only abandoning territorial claims, but the much older and deeper and more threatening claim that there was no high culture of any sort between Poland and Russia, and that all the “borderlands” were naturally either Polish or Russian.
From the quoted passage: “an hear apparent and the second highest title” ?
On the other hand, “heir-say” would be a pleasing variant of “heir apparent”.
—It’s a relatively widespread point of view in Russia that Russia itself morphed from the colonial dependence from the Steppe, from its own subjugation
I suppose Poland also could count as a former colonial master of a large portion of historical Russian lands.
*And exactly 400 years ago, Moscow itself was under Polish yoke…
L. N. Gumilev’s views are the most likable for the indigenous people of the steppe and siberia, his interpretation was that it was not that much of igo, but something like voluntarily taking sides within the orda and rus, and later peaceful exploration of the great land mass with not that much antagonism but cooperation of the native populations
and if it could be interpreted that way, this colonization and pokorenie stuff sounds just like pleasing the imperialist ambitions of the new russians, which could be perhaps avoided just simply because nobody denies greatness and attractiveness of the russian culture which could be even greater asserting its more tolerant progressive open and equal among equals declaring qualities, without raising at someone else’s expense, not reinventing the colonialist ideology and start sounding as if like copying the west with all those subjugation claims, one could expect in some time they’ll start including some more racist terminology in there perhaps too, doesn’t sound as if like improvement in any sense, political scientific or diplomatic
and if it was really in there, the colonialist ideology of the past, the most number of new settlers were sent out pretty late only during the stolypin reforms, no? maybe that was really like the official tzarist colonization period only
I was very impressed, by the way, by Poland’s early-1990’s western strategy (unilaterally abandon all claims west of the Oder-Neisse line even though Germany had not yet reciprocated) and even more by its eastern strategy, about which I knew nothing before reading Snyder.
Same here! Really mind-blowing (as we used to say).
an hear apparent
Thanks, fixed.
historical Russian lands
Only historically Russian if you are committed to identifying Kievan Rus’ with Russia, a shaky proposition at best.
Read,
“L. N. Gumilev’s views are the most likable for the indigenous people of the steppe and siberia, his interpretation was that it was not that much of igo, but something like voluntarily taking sides within the orda and rus, …”
This makes sense, even if it flatters the indigenous people, because it takes a lot more troops and others to subdue and adminster conquered teritory than it does to just conquer it, and I doubt any of the hordes had that kind of manpower. They had to get a least some degree of willing submission. And it may not have been too hard if they offered better government than the indigeonous rulers in Russia, just as the Ilkhans were valued as rulers in Iran.
“this colonization and pokorenie stuff sounds just like pleasing the imperialist ambitions of the new russians,…”
That makes sense. It would be analogous to the British fascination and self-identification with Rome.
Incidentally, the Nogai Keikuvat had some important namesakes: Kaykaus I and Kaykaus II of the Seljuq Sultanate. The name of the Turkic Gagauz people, who live north-west of the Black Sea and are Christians, is said to be derived from the name of the second Kaykaus.
distinguished warriors (bahadurs)
a lovely definition for what must be one of the sources for Russian bogatyr (богатырь).
Well, actually, it looks like Mongolian баатар ‘hero’ (as in Ulaanbaatar). The old spelling in Uighur script was ‘bagatur’.
must be a cognate, Russian etymological dictionaries mark it as of Turkic and Mongolian origin.
must be a cognate
Bahadur and baatar (formerly bagatur may be cognate, though not Russian bogatyr (богатырь) which is a borrowing from Turkic or Mongolian (which are considered to be related). Saying “a cognate” is like saying “a cousin”. You have to make it clear who is who’s cousin, what word in what language is cognate with what word in which other language. On the other hand, bahadur and bagatur may be the same word in different dialects, and it may have started in one dialect and been borrowed into another dialect, rather than going back to the proto-language. (I can’t tell, because I am not familiar with these languages, but these are likely alternatives).
Incidentally, the Nogai Keikuvat had some important namesakes: Kaykaus I and Kaykaus II of the Seljuq Sultanate. The name of the Turkic Gagauz people, who live north-west of the Black Sea and are Christians, is said to be derived from the name of the second Kaykaus.
Excellent, I love that kind of connection!
Perdue’s China Marches West covers the closing of the steppe from the Chinese side. I haven’t finished it but it’s good, with a few odd errors however.
In Anglo India (per Hobason Jobson) a cognate of bagatur, presumably via the Moghuls, means “bully” or “thug”.
JE: a cognate of bagatur…
Is the Anglo-Indian word a cognate of “bagatur”, borrowed from a language related to that of the Moghuls, or an adaptation of “bagatur”, itself a word borrowed through the Moghuls?
Obviously a borrowing.
i am willing to ‘cede’ the word ‘bahatur’ to turkic language, cz it can’t be separated into smaller units to get to the root word, baatar is baatar, means hero, so must be it is a borrowing from turkic or could be even sanskrit
while nuudelchin for example is derived from nuu(to go, migrate) – nuudel (moving, migration)- nuudelchin (nomad, so if one can follow the clear meaning of the root word maybe that word is native to the language it seems to me, no?
if in hindi? the word was used before the moghul conquests it could be a borrowing the other way around too, if later, than the newly acquired negative meaning is understandable too
Read, that etymon could well come out of a Turkic language that borrowed it at some point from some Indo-Iranian language up there on the steppe, or even a lot earlier, proto-Turkic from Indo-Iranian itself, and then it spread wherever else. There is a Sanskrit word ‘bhag-, bhagava’ that means lord or something along those lines, and it is the same etymon as ‘bog’ in Slavic that means God, so it’s been around a while and any Iranian language in Central Asia was very likely to have it too. If Uralic languages were borrowing from Indo-Iranian, there is no reason Turkic languages cannot have done the same. And vice versa.
The origin of Russ. богатырь is never seriously hypothesized to be “бог” god (of course there exists a folk etymology to dispel, linking богатырь with богатый “rich” which is ultimately cognate with “бог”). But folk etymologies aside, nobody doubts that Russian “bogatyr'” (as well as Hungarian bátor) is a Turkic / Mongolian borrowing. But could it, ultimately, be a re-borrowing as Jim suggests?
The name of the Turkic Gagauz people, who live north-west of the Black Sea and are Christians, is said to be derived from the name of the second Kaykaus
In the North Caucasus region, the other set of trans-regional ethnonyms which always fascinated me are the tribal names of Ossetians. One of the two tribal groupings of Ossetians is Iron ( ~ Iran), and within the Iron, one finds Kurttati ( ~ Kurds + Tats?)
from bhag – bhagava to bahatur and bogatyr’ sounds as if like, plausible, yes, though russian bog – bogatyi bogatyr’ also seems like probable too
we have pretty many sanskrit borrowings mostly as if like some, more like abstract concepts like nouns, sansar is cosmos for example, when our own original word is ogtorgui- with a root word ogt – none or absolute, ogt oor – other, ogtorgui – cosmos
Something tells me that baatar goes way back earlier than sansar.
why it should be so, i think any language vocabulary has the most generic words like water-us, mountains- uul, ogtorgui is also the same category as nar-sun, sar-moon, so it could be the borrowing from sanskrit of sansar was not that recent but pretty ancient since the concepts and connections existed long before, at least not when our only cosmonaut have chosen it as his surname
sansar means “world” in hindi they say, while samsara is universe, or the cycle of existence and non-existence, so maybe you meant, with adoption of buddhism it came into our language, maybe
our word for world is ertonts, have no idea how world becomes cosmos though when borrowing a word, with their different combinations can get all the meanings though sansar cosmos, ertonts world, ogtorgui ertonts – universe, sansar ogtorgui – outer space
Read: The English word mountain is in fact borrowed. Of the 100 very basic words in the English version of the Swadesh list, the following are borrowed from other languages: person, grease, path, mountain, round.
JC, five of those words are from French, but path? The word seems to occur in all West Germanic languages. Perhaps Proto-West-Germanic borrowed it from somewhere else, but that would have been before its differentiation into English, Dutch, etc.
OED sez:
path and and russian put’ then are of the same origins must be
it’s no wonder maybe for english to inherit such basic words from PIE, since the language is itself continuation of it within the language family with borrowings from other IE languages, ancient and recent
if our language is an isolated language family itself, its basic words are perhaps those, native to it with the borrowings along the way too from turkic, sanskrit, chinese etc used as synonyms of those generic basic words must be
Thanks for the OED quotation, LH. Somewhere I had read that Latin pontus (hence French pont) ‘bridge’ might have referred originally to raised wooden paths built across swamps, such as are still built in parks over swampy or fragile ground. I don’t know much about the history of Celtic, but Welsh pant ‘valley’ might well fit in here, since apparent opposites imply each other: the existence of a structure that bridges a space implies a space that needs to be bridged. You never see a picture of a bridge without also seeing the water beneath (or traces of it in a dried-up bed).
I did not know about the unshifted p- class in West Germanic. The “pre-Germanic IE substratum” would most likely be Celtic.
Does anyone know about the etymology of Russian put’ that read mentions?
About a possible Iranian borrowing into Germanic, weren’t the Alans (some of whom ended up in Celtic lands) Indo-iranian speakers? and Germanic languages were once spoken in some currently Slavic-speaking lands (Gothic being the attested language but probably not the only form of East Germanic that ever existed).
The fun etymologies are “sack” and “tank”.
Yes, read is correct: put’ is from Old Slavic pǫt’ (with nasal o from *an), which is cognate with Sanskrit panthās (and the Iranian words cited by the OED), Latin pons ‘bridge,’ Greek patos ‘path’ (with a from zero-grade n), etc.
LH, you are right, Latin for ‘bridge’ is pons/pontis, and pontus is ‘sea’, from Greek pontos. Would that word refer only to ‘bounded sea’, as opposed to ‘open sea, ocean’?
Back to path, could the “unshifted p’s” be due to borrowing after the operation of the first sound-shift, after all the original p’s had changed to f’s?
JE: The fun etymologies are “sack” and “tank”.
We dealt extensively with “sack” here not too long ago. That was great fun!
Here’s the complex etymology of tank in full, from the OED2. Different senses may have different original etymologies, but they have clearly influenced each other, too, so I have interleaved the definitions of the older senses.
Etymonline basically tells the same story, except that it attributes the Romance forms to OSp OPt estancar, Fr estanchier ‘hold back a current of water’ (> E stanch) < VL *stanticare < *stagnicare and back to stagnum.
So the basic disagreement is this: Did the Neo-Indic languages inherit their tank-like words from Sanskrit and pass them to the Romance languages, where they mixed with native words of the same form and similar sense, and thence to English? Or are the Indian words themselves borrowed from Romance forms?
m-l: Yes, it’s the combination of /p-/ (which is very unlikely to be native) and /-θ/ (which certainly could be) that marks path as a borrowing at the Proto-Germanic level that came in after the First Consonant Shift. Poking about for languages which have shifted PIE *pent- (regularly > Eng find) to something like /paθ-/, we hit on the Iranian languages. Bingo! A similar borrowing, now lost, is Gothic páida, OE pád ‘coat’ (distinct from OE pæð > path), attested in Beowulf in the compound here-pád ‘soldier’s coat, coat of mail’.
There are at least two separate pieces of evidence that Germanic languages were once spoken far east of where they are found today. One comes from the Old Norse saga King Heidrek the Wise, which preserves very old Germanic names in its account of a 4th-century battle between Goths and Huns, specifically Harvath-fells (ON Harvaða fjöllum), which when the First Consonant Shift is run in reverse, becomes the peaks of the Carpathians (both peaks and Carpathians being borrowed). For the other, see the comments at the end of this LH post.
John, I reread the LH post in question, your summary of some phylogeneticists’ conclusions, and also Etienne’s comment doubting some of those conclusions.
Do you mean the classification of Germanic branches together with Albanian? that’s one of the things that Etienne finds very doubtful, given the short time period since written evidence of Albanian, along with its numerous borrowings from neighbouring languages. Not knowing any more, I can’t support one conclusion rather than another, but I will keep an open mind.
I do like the Carpathian peaks though. I would not have recognized them offhand under their Germanic disguise. Does the word Carpath(ian) have an etymology?
Greek Καρπάτης is the immediate origin (note tau, not theta; the th was falsely added by German), with a possible connection to Albanian karpe ‘rock’.
As for the Albanian position in the IE tree, Ringe et al. agree that the data are no good; the oldest Albanian available is only 15th-century, when most of the PIE morphology has already disappeared, and the language is almost all loanwords anyway, to the point where it was originally thought to be an aberrant branch of Indo-Iranian. Their important result about Germanic is that it is the sister group to what they call the Satem Core, the immediate parent of Balto-Slavic and Indo-Iranian, despite its heavy modification by western IE languages.
Something tells me that baatar goes way back earlier than sansar Why should it be so?
Well, as a feeling:
1. My strong suspicion is that baatar went back to the time of Genghis Khan, so it was a very old word.
2. Mongolians were not Buddhist at the time of Genghis Khan, so Sanskrit loanwords would likely have come later.
I realise that this is simplistic — apparently Mongolians’ first linguistic contact with Buddhism was via Turkic, since Buddhist scriptures were translated into Mongolian via Turkish in the late 12th/early 13th centuries. But the idea that baatar ‘hero, warrior’ preceded sansar ‘universe, cosmos’ seems somehow intuitively more plausible.
Marie-Lucie: I found the entire exercise doubtful, not just their classification of Germanic. Simply put, there were too many inaccuracies in the data. But even with impeccable data I would regard this work as being of very limited use (I can elaborate, if anyone is interested).
John Cowan: I have never heard of anyone calling Albanian a form of Indo-Iranian. Armenian was considered an Iranian language once, on the basis of its many Iranian loanwords: could this be the language you had in mind?
Yes, Etienne, please elaborate. I am sure I won’t be the only one interested.
About the name of Albania: is there any relationship with the Scottish Alba and the English Albion? (“la perfide Albion”, as it was known in France for centuries).
People have postulated a dubious IE root alb- ‘white’ and an even more dubious pre-IE root *alb- ‘mountain’ and used them to explain Alps, Alba ‘Scotland’, Albania, Albion. Of course, high mountains often have white tops. Nobody really knows.
What I never understood was why Albion was called “perfidious”, i.e. treacherous. France and England had been straight-up enemies since the Hundred Years’ War, so where does the treachery come in?
What I never understood was why Albion was called “perfidious”, i.e. treacherous.
Apparently, the idea was popularised by Republican France in its wars with Britain. It was an attempt to associate the British Empire with Carthage, which the Romans frequently accused of being untrustworthy (“Punic faith”). Of course, the implication is also that France is the Roman Empire and thus will triumph over Carthage-Britain. Carthage and Britain were both regarded as commercial empires (a “nation of shopkeepers”).
(I got all this from Richard Miles’s Carthage Must Be Destroyed).
Wikipedia is on the case:
Thank God for the Entente Cordiale.
For read:
Amritas has a discussion of the history and derivation of 筆, bitig, бичиг etc. here.
” I thought it might be from Middle Chinese”
i don’t know who he is and he might think what he might think
as for me I’ve accepted SFR’s explanation that it’s from Jujan or Kidan or some other mongolic language meaning scribe iirc as pretty convincing
as for biir – brush – yes that might be a borrowing from Chinese and let’s not start that discussion again
“why Albion was called “perfidious”, i.e. treacherous.”
i recall it’s called tumannyi Albion in Russian poetry mostly, so could it be foggy could mean something similar to illusory therefore “treacherous”
^the
The endless wars between England and France were an aftermath of the conquest of England by the Vikings established on French soil in “Normandy”. After William (a bastard son of the duke of Normandy but a great-nephew of the King of England, himself a Viking descendant) asserted his hereditary claim by force at the battle of Hastings, intermarriage and competing rules of royal inheritance meant that the kings of England had close links with the kings of France, and sometimes possibly legal claims to the throne, or at least to a large part of the country. The 1904 “Entente Cordiale” (lit. ‘heartfelt understanding or agreement’) was intended to put an end to centuries of distrust and warfare. So the role of France’s ennemi héréditaire stopped being England’s and became Germany’s (Bismarck’s trick of baiting Napoleon III into declaring war in 1870 over a trivial matter prepared the way).
Miles quotes Thomas Jefferson, writing about Britain in 1809-10: “Her good faith! The faith of a nation of merchants! The Punica fides of modern Carthage.” Miles notes: “A nation of shopkeepers could not be trusted to keep its word.”
a nation of merchants/shopkeepers
But that nation is where the Founding Fathers or their forebears came from. They and their descendants obviously could not fully shake off their heritage: who said the business of America is business?
But Jefferson was not of that party; he loved farmers and hated merchants (except, of course, the ones who sold him his Bordeaux).
Calvin Coolidge, but that’s not quite what he said, and the difference is significant. His actual words in a speech to the American Society of Newspaper Editors were “After all, the chief business of the American people is business”, and he was stating it as a fact, not as a quasi-moral claim implying that American had no other proper business. He was, in fact, defending the position that making money (by creating useful stuff, not by finance capitalism) is a means to various good ends rather than either good or bad in itself, and therefore as long as newspapers defended the public interest, it did not matter that they themselves were big businesses. I suspect he would change his emphasis today — he was, after all, a Vermont Republican of his time.
1-John Cowan: Albanian “karpe” may or may not have anything to do with the etymology of “Carpathian”.
The reason for this is that similar words are found in a number of Romance languages spoken in Mountainous areas (Modern Romansch has “crap” (rock), for instance). This word is non-Latin in origin, but its distribution suggests that it entered spoken Latin from some unknown language early and thence became part of spoken Latin throughout the Empire. Meaning that the word may not be native to Albanian, and may instead have entered the language from Latin (a similar situation is found with Basque: as a great linguist, Koldo Mitxelena (AKA Luis Michelena), once pointed out, many pre-roman elements in Basque are not ancient in that language and must have been transmitted through Latin).
2-Read: “Amritas” is the blog of a historical linguist whose specialty is historical linguistics involving East Asian languages. I have read his blog and corresponded with him a few times, long ago, and I assure you that he is a competent historical linguist. If you wish to learn something about historical linguistics and the linguistic history of East Asia I recommend his blog unreservedly.
3-Marie-Lucie: The core problem is that no distinction is made between “states” that can be expected to continue in the history of a language and those that cannot. Thus, I had originally criticized them for coding both Old Irish and Welsh as having an -i genitive (in o-stem nouns and adjectives) when in fact there is no clear evidence that Welsh ever had an i-genitive.
Now, they have coded Latin accurately, as having once had an -osyo and an -i genitive. But if we compare Old Irish and Middle Welsh to CONTEMPORARY “Latin”, i.e. Early Medieval Romance languages, then we would have to code them as lacking either ending. In fact, their morphological coding would yield clearly different results for Latin and Romance (even Early Medieval Romance!), as the latter lacks any of the case-endings they encode, any of the superlative-marking endings, any inherited mediopassive-marking morphology.
Similar such patterns of loss have affected a number of other Indo-European languages over time, and thus, if instead of taking the earliest data from each branch, you compare CONTEMPORARY data between different branches, results vary to an alarming degree. Enough so to cast doubt on their results. Combine that with the actual mistakes, and frankly, I find this article fatally flawed.
But not worthless. It would be interesting to use this method using states which, upon being reached, are basically immutable. Take the merger (word-initially) of aspirate and non-aspirate voiced stops. The beauty of this is that those branches of Indo-European which merged them (Celtic, Albanian, Baltic, Slavic, Iranian) and those which kept them separate (Germanic, Latin, Indo-Aryan) have the same state, NO MATTER WHAT TIME PERIOD THE DATA COME FROM.
Thus, Proto-Germanic and Latin would be coded as having preserved the distinction. So would Modern English (TREE, THREE, DOOR, whose etyma in Indo-European had initial *D, *T and *DH, respectively) or Modern French (DON, TROIS, FAIRE, whose etyma in Indo-European had the same three initial consonants). Conversely, Slavic would be coded as having merged aspirate and voiced stops: this is as true of Proto-Slavic as it is of Modern Russian (DAT’, TY and DYM whose etyma originally had the same three initial consonants).
A classification of Indo-European languages following the above lines would neutralize the difficulty we have in dealing with data from different time-periods. Who knows, I may try doing so myself someday.
Etienne:
I was wrong to say that Albanian (rather than Armenian) had been thought to be Indo-Iranian because of its Iranian loanwords, but the rest of my references to Albanian are correct: there are no texts available before the 15th century, when the morphology has already been lost. Barring the discovery of some much older Albanian texts, we will never have an accurate idea of where it stands in the IE tree, though its status as an IE language is secure. The same can be said mutatis mutandis of the extinct Palaeo-Balkan languages.
Ringe et al. looked at more than 400 characters to start with, and rejected most of them as parsimony-uninformative, often because they might possibly be not shared characters but primitive characters lost in all other branches. In the end, only a few characters were genuinely diagnostic. Stripping them down further as you propose will probably produce a null result altogether.
But granted that you are right about Old Welsh, I do not understand your reasoning about contemporary vs. oldest-available data: can you elucidate? Thanks.
Etienne: their morphological coding would yield clearly different results for Latin and Romance (even Early Medieval Romance!), as the latter lacks any of the case-endings they encode, any of the superlative-marking endings, any inherited mediopassive-marking morphology.
Different branches of the same family usually change in different ways, so that what is obvious in some members (or even just one) may have been weakened or lost in others. In terms of morphology, even in the worst cases there can still be some traces of earlier structures in now irregular, even apparently aberrant forms.
For instance, English has lost most of the older case-endings, but the m in him and them is the single remaining link to the dative in its less eroded relatives (German) as well as the accusative in more distantly related ones (eg Latin). Similarly, the comparative ending (in -er) is well-preserved (although unproductive) in English or German, but in Romance they are still recognizable in a few special forms (eg Fr meilleur, Sp mejor, It migliore, etc ‘better/best (adjective)’). Their own origin in -s- forms is itself attested by other apparent irregularities (eg Latin melius ‘better/best (adverb)’). Shared retentions like these link the more innovative with the more conservative descendants of a single language, and suggest features of the common ancestor even if that language is unattested by any actual record (like PIE).
Marie-Lucie: yes, traces CAN remain. In the case of the genitive in -i and the medio-passive in -r there isn’t a trace to be found in any Romance language, modern or ancient. Please note, however, in this connection, that finding such residual features in an extinct language means having a corpus that is large and very diverse, stylistically: for a majority of Ancient Indo-European languages this is not the case.
Also, even when they are found, residual traces of a given feature are liable to be interpreted in different ways, making the resulting trees less reliable. By the way: English and Germanic -M is indeed a dative marker, but it is not a cognate of the accusative -M found in Latin.
John Cowan: my point is that the dates of first attestation for various Indo-European languages are basically due to various historical accidents. And looking at some of the features used by the authors, it is clear that their trees would be substantially different if data from some of the branches were drawn solely from later stages (i.e. from Romance languages instead of Latin, from Prakrits instead of Sanskrit). Since the goal of this article is to create a tree of Indo-European which presumably reflects historical reality, the fact that different trees would result from using data from later time periods (for those branches with a sufficiently long history) is not reassuring. It suggests that conversely, if we had data on earlier stages of all branches of Indo-European, especially those attested late (Baltic, Albanian…) the tree in question would also be different. Hence my suggestion that features which are known to be diachronically stable be used for classificatory purposes.
There’s yet another factor which they do not take into account, by the way: that is that different words/morphemes did not stand an equal chance of surviving in different languages. Take, again, the Old Irish genitive in -i. Technically speaking, the /i/ morpheme did not survive into Old Irish: its effects are visible in the palatalization of final consonants and the fronting/raising of vowels. This is a consequence of the large-scale apocoope of final syllables which turned Primitive Irish into Old Irish, from 500-900 AD, more or less.
But let’s use our imaginations here. Let us imagine that Primitive Irish originally had a final /i/ AND a final /od/ (drawn from the ablative: as I wrote earlier some believe this is the etymology of the Celtiberian genitive in /o/)) as genitive singular markers of o-stem nouns and adjectives. Now, in the wake of large-scale apocope the former would expand at the expense of the latter, since post-apocope the genitive singular in (original) /od/ would be indistinguishable from the nominative singular. Thus, to code Old Irish as having an /i/ genitive instead of a genitive drawn from the ablative (/od/) is true, but misleading. The latter simply could not, in the wake of the sound changes which turned Primitive into Old Irish, serve as a genitive singular marker.
For those branches of Indo-European which underwent major sound changes (Old Irish, Welsh, Slavic, Armenian especially come to mind) I thus suspect that, for short lexemes and grammatical morphemes especially, there were far fewer possibilities to choose from.
Marie-Lucie, Etienne: One of my favorite bits of Indo-European conservatism in English is the fact that older and more conservative Indo-European languages always have syncretism in the nominative and accusative endings of neuter nouns, adjectives, and pronouns. Though noun gender and adjective agreement have utterly vanished from Modern English, we still see that it, unlike all the other personal pronouns, takes the same form in subject and object positions.
Etienne, you write:
[M]y point is that the dates of first attestation for various Indo-European languages are basically due to various historical accidents. And looking at some of the features used by the authors, it is clear that their trees would be substantially different if data from some of the branches were drawn solely from later stages (i.e. from Romance languages instead of Latin, from Prakrits instead of Sanskrit).
Indeed. In point of fact, Ringe et al. experimented with the same methods using data from the modern West Germanic languages only. No stable tree appeared, consistent with the vast amounts of homoplasy (borrowing and convergence) known to exist around the North Sea and the Western Baltic.
Since the goal of this article is to create a tree of Indo-European which presumably reflects historical reality, the fact that different trees would result from using data from later time periods (for those branches with a sufficiently long history) is not reassuring.
I would count that fact as an argument for rather than against the method. We are looking for the faint traces of very ancient events, 4000-6000 years old. (If they weren’t faint, the classical comparative method unaided would have turned them up long ago.) The more modern homoplasy we can eliminate a priori by using sources as old as we can get, the more likely we are to see the signal beneath the noise. In particular, we know that the last two millennia have seen large Sprachbund effects not only in the Balkans but in most of Europe as a whole, and independently in India. Weeding this out can only help clarify the relationships produced by common descent.
It suggests that conversely, if we had data on earlier stages of all branches of Indo-European, especially those attested late (Baltic, Albanian…) the tree in question would also be different.
Doubtless it would be: the method is statistical, and present anomalies might vanish in the presence of wider and older data. Still, by the same token we would expect the basic tendencies to strengthen as more noise disappears.
I’m not clear on whether you are attacking the underlying data only, or whether you are also attacking the method.
John Cowan: I am attacking the data and the method. As I wrote, there are some genuine errors in the data. But even with impeccable data I don’t think it would matter, because it is increasingly clear to me that this entire article *cannot* deliver what it promises.
Basically, along with their other errors (see my comments above), the authors also commited a methodological blunder.
I can summarize this blunder simply: THE CLOSER YOU GET TO THE DATE OF THE BREAK-UP OF PROTO-INDO-EUROPEAN, THE LESS SENSE IT MAKES TO TREAT DESCENT AND CONTACT AS SEPARATE PHENOMENA.
Let me illustrate. English has a large number of Latin/Romance loans: these are easily detected, because at the time when contact took place English/Germanic on the one hand, Latin/Romance on the other, were clearly separate entities. Hence these and other borrowed elements are not relevant to a genetic classification of English.
Now, let’s look at different dialects of English. There’s a set of sound changes which separates North American from British varieties of English. Thus, in principle, whether a given variety of English is North American or British (assuming for the sake of simplicity that we are operating with this two-way classification) should be as clear-cut as determining whether a given language is Romance or Germanic. Right?
Wrong. The problem is that dialect contact/mixture can yield new varieties which are unclassifiable genetically, only typologically. Is a given Mid-Atlantic accent a British-influenced North American accent or an American-influenced British accent? The answer is yes. And no.
You see, when language contact between two clearly distinct languages takes takes place it is easy to determine what kind of a mixture it is. That is to say, English or Luxemburgish are both heavily Romance-influenced Germanic languages, and Romantsch or Dolomitic Ladin heavily Germanic-influenced Romance languages. Thus, while Romance and Germanic have influenced one another greatly, there can be no doubt whether a given language is Germanic or Romance, genetically.
This matches what research on language contact has shown: bilingual speakers are quite conscious that they are using one language or the other, and “language mixture” in fact involves a MATRIX LANGUAGE, as Carol Myers-Scotton has called it. A “matrix language” is the language which is the base to which elements from another language are added. However Romance-influenced it may be, English is a Germanic language which has absorbed a huge number of Romance loans. The Germanic-Romance contact which has yielded Modern English always involved (some earlier varieties of)English as the matrix language. Conversely, Romansch, however German-influenced it is, is the product of Romance-Germanic contact with (some earlier varieties of) Romansch as the matrix language.
But this breaks down if you are not dealing with two clearly separate systems. Such is the case of two varieties of English in contact, yielding a third. Here we cannot distinguish a donor versus a recipient language, we cannot speak of a matrix language receiving elements from another. This is because these two dialects are variations upon a single system, not two clearly separate subsystems. To ask whether a fluent speaker of a Mid-Atlantic variety of a English speaks British-influence North American English or North American-influenced British English would be like asking whether Zebras have black or white stripes. And indeed, a native speaker of a British variety and a native speaker of a North American variety, each seeking to make their speech more “unmarked”, might well each end up speaking with accents so similar as to be indistinguishable from one another. In other words, dialect mixture can yield similar results, *even if the dialectal starting point is not the same*. This is very different from what happens with separate systems: English and Romansch remain quite visibly Germanic and Romance, respectively, despite the heavy contact history of each language.
Crucially, Indo-European must have originally been a dialect continuum rather than a collection of clear-cut dialects. What this means is that ANY common innovation which goes back early enough *cannot* be used as an argument for genetic subgrouping, as in principle any such common feature could also have spread as a contact feature.
The above point, by the way, was made by Antoine Meillet (I think in LES LANGUES DANS L’EUROPE NOUVELLE), to whom I would refer any interested reader, not least because he made the point far more clearly and succinctly than I did (or could).
Marie-Lucie: While the choice between synthetic and analytic comparatives in English is hard to nail down (various factors contribute to the choice in varying amounts), I would not call synthetic comparison unproductive. When fun became a full adjective, roughly around 1970, people started to say (though not yet to write) the novel forms funner and funnest.
Etienne:
I grant that your argument is correct in principle. But it’s a long way from Sanskrit back to the breakup of IE into first-level language families. However, if you look at this later paper from the same group, you’ll see that it postulates three hitherto unknown non-genetic contact events in order to create a perfect phylogenetic tree: one between Proto-Germanic and Proto-Italic (6 characters), one between Proto-Germanic and Proto-Baltic (9 characters), neither of which is surprising, and one between Proto-Italic and Proto-Greco-Armenian (2 characters), which is more surprising but can be accounted for if Proto-Italic developed outside Italy. This tree with its sideways links appears on p. 403 as Fig. 12.
In addition, I think your English example is incorrect as to facts. I do not think there exist any Mid-Atlantic English dialect communities — that is, ones that cannot be clearly assigned to North America or the British isles — even though there are individuals with unclassifiable dialect or accent mixtures, and there are certainly isoglosses that cross the Atlantic. My own accent retains the English feature of not having the LOT=CLOTH merger, and there are parts of England (to say nothing of Scotland and Ireland) that retain the now mostly North American feature of rhoticity. But nobody hearing me would suppose that I speak with any sort of British accent, or that a Bristol accent is somehow North American.
(For the record, evidence from embryology, teratology, and comparative anatomy clearly establishes that zebras are black with white stripes, and this despite their white underbellies, a shared primitive or convergent character common to many animals.)
vanya: A lot of Italians I’ve known seem very resistant to the idea that classical Latin was actually pronounced differently than Tuscan.
I did not mean things like the pronunciation of ci as “chi” rather than “ki”, but the placement of stress and the alternations in vowel length. Both these things are totally absent from the pronunciation of most French people (listen to ROSA ROSAM, etc), but should come quite easily to Italian speakers.
I remember an old Latin teacher telling us this anecdote of his youth: the food in the school dining room (where the boarders ate) was pretty bad, and one day when the main dish consisted of a rather watery lentil soup, one student called the supervisor, showed him his plate, and said, quoting Virgil: appaRENT raRI nanTES in GURgiTE vasTO (‘rare swimmers appear in the vast abyss’). I think the proper stress pattern should be apPArent RAri NANtes in GURgite VASto, uttered not in even-length syllables as in French but with extra length and stress where needed, as in Italian.
JC: I would not call synthetic comparison unproductive:
OK, not 100 percent, but generally so. In this respect it is similar to the formation of noun plurals and verb past tenses through internal change: these processes affect very few new forms, and when they do, it is usually in a jocular context (eg meese as the plural of moose).