Home > DNA, Indo Europeans > Slavs Along the Frankish Border

Slavs Along the Frankish Border

Bronze Age skeletons in South Siberia have been sampled for Y-DNA. Christine Keyser did a great job in mapping recent close matches that are stored in the YHRD database. Most were concentrated in Central Europe.

The remains are attributed to the Andronovo Culture (2300–1000 BC), that was centred in Siberia and Kazakhstan. The culture is assumed to be ancestral to Scythians and closely related to the European Kurgan complex and was also proposed as Ugric. Andronovo is considered a prehistoric culture of mixed composition, although virtually all samples were of haplogroup R1a.

Christine Keyser et al., 2009. This graph intends to map the few close matches that were found in the YHRD database with the paleogenetic samples of Bronze Age skeletons in South Siberia, attributed to the Andronovo Culture.

There might be some bias because of the limited number of STR that was used for the correlation, though the matches show a clear division running north to south through Central Europe: the virtual absence of corresponding haplotypes west of this division is as notable and intriguing as the steep decline to the east.

If these concentrations are ancient, then why do they correspond so neatly to the eastern borders of the Frankish empire of Charlemagne?

Boundaries of the Frankish Empire.

Do we see here the remainder of Frankish genocide, or the successful Frankish defense against a powerful, early medieval Slavic advance?

Note that Austria (the Frankish Eastmark) was colonized and (re-?)germanized by the Franks at the cost of Slavic Carinthia. Much of Eastern Germany was overrun by West Slavic tribes. During the Frankish expansion to the north and north-east the Abodrites, speakers of the polabian language, became allies of Charlemagne in his fight against the Saxons and were rewarded with properties in historically Germanic territories.

The western dots in C.Keysers map followed the Frankish borders of the east in a way that suggests a medieval Slavic advance and reinforcements along a fortified zone. However, being concentrated as a fringe that contrasts sharply with both the Frankish territories and the eastern Slavic backlands, the map doesn’t reveal an obvious source in the neighbourhood for this specific set of haplotypes. Any Slavic connotation, however likely, does not correlate to expectations about a single geographic origin of Slavic people.

Then where the haplotype came from inmediately before the event of spreading along the western borders of the Slavic world? The Andronovo culture was already long ago and a related origin has to cope with a gap of thousands of miles running through territories without archeological traces of an ancient proto-Slavic presence. This issue touches the core of the Slavic origin problem. Where and how did Slavs originate?

Possible Slavic homelands.

The main contenders for the Slavic origin are the Zarubintsy culture (“Zarubinec”), the Chernoles culture and the Przeworsk culture. However, the Zarubintsy option, centered by the Pripet river, knew some influential opponents that consequently identified this culture as (East?) Baltic. The C.Keyser graph gives nothing but some superficial support to the Przeworsk culture and none to the Chernoles culture nor the Zarubintsy culture as the origin of the Slavic ethnicity. Earlier Y-DNA studies based on haplogroup frequencies and FST values asserted a genetic origin of Slavs in Ukraine (e.g. Rebala et a., 2006), thus assuming Slavic replacements that never occurred, being too inconsistent with genetic distances and RST values to make this a valid approach and in contradiction with the low R1a1a diversity measured by Underhill et al. (2009). The latter measured a R1a-M458 gradient west to east instead. Though an important new marker that may be congruent with earlier Balto-Slavic expansions to the east, the Slavic match doesn’t extend to the Balkans and thus still can’t account for all important Slavic migrations.

To discover the nature of the genetic patterns described above and a possible association to Slavic expansion, we are forced into an indepth assessment of the Slavic identity.

Departing from the assumption of a genetic relation between Slavic and Baltic languages, the change to Slavic appears to be drastic. The exact location of this change remains inconclusive, though must have been somewhere in the Praque-Korchak-Penkovka triangle. A foreign superstratum could have triggered this change, using a Baltic dialect as lingua franca. This may explain the apparent lack of overall consistency of the genetic signature of Slavic territories compared to any of the proposed geographic homelands. Archeological traces of steppe nomads in Przeworsk territory are less telling than historical references of Alans in the Great Migration period. In the Zarubintsy and Chernoles cultures inmediately to the east this archeological link is easier to discern.

If the assumption of a foreign superstratum holds we should expect a genetic input caused by elite dominance and also a lexical input in some degree that is shared by Slavic languages while being foreign to Baltic languages.

Having this as a working hypothesis I will attempt to identify a possible ethnical candidate that may approximate the unknown superstratum and match the preliminary results of this investigation with a tentative genetic affiliation of this hypothetical ethnicity as a lead to further investigation.

List of numbers in Slavic languages.

One glaring lexical difference between Baltic and Slavic languages can be found in their word for number one. The Baltic versions neatly derive from Proto Indo European *oino in vienas (Lituanian, viens (Latvian) etc., while in Old Church Slavonic “jedinu” the PIE root was only preserved in suffix -inu (van Wijk, 1912, eg. cited by Jadranka Gvozdanović, 1991). The prefix “jed” is exclusively Slavic.

List of numbers in Baltic languages.

Let us assume this Slavic prefix of number one is indeed a lead to the identity of a non-Baltic superstratum that accelerated the development of Slavonic languages from a Baltic substratum language. Where did it come from?
Comparing the languages in the wider region, the best phonetic equivalent I could find that fits the orgin of the prefix in question, is in the Hungarian language. Number one corresponds here to “egy“, pronounced “edj“, whose Uralic origin becomes clear in comparing equivalents in other Uralic languages.

Number one in various Uralic languages.

Interestingly, Indo-Iranian versions of number one appear to come closer to Finnish. This feature won’t unambiguously exclude a direct genetic relationship between the Slavic language group and the Indo Iranian language group, though the genetic transition between both forms seems to be much better documented in Uralic languages. I consider parallel lexical evolution of a single number between different groups of languages virtually inconceivable, and suffice in proposing a separate event that accounts for Uralic contacts much further back in time, i.e. between the forebears of Finnish and Indo-Iranian speakers. However, this possible event is irrelevant for the purpose of this article.

List of numbers in Indo Iranian languages.

This preliminar investigation above may indicate an Ugric superstratum in the formative stage of Slavic, c.q. an elite that spoke a language that approximated Hungarian. How does this fit within the historical context? The accepted version of Hungarian history is their arrival from the steppes in the Hungarian Conquest (896 AD), being triggered by Turkish pressures of Pechenegs and Danube Bulgars. Before they migrated wholesale to Hungary as a confederation of Ugric and Turkic “Magyar” tribes (On-Ogur, the Ten Arrows), they are reported to have roamed for a while in their mythical homeland Levedia, usually located somewhere in the neighbourhood of the Dnjeper and Don, 7th century AD, where they must have endured a profound Turkic influence. There is no doubt about it that before this episode they had their dwellings further east, in the steppes south of the Ural mountains. Archeological continuity of this area with the Bronze Age Andronovo culture has been suggested many times in literature. The Magyar migration is thus part of a much wider movement of people from that direction, involving Scythians, Sarmatians, Alans and Huns, that already supplied the classical world a complicated succession of invading tribes to worry and write about. More arrivals from the east were still to come.
The Slavs are generally assumed to have started their expansion in the wake of the Migration Period, in the footsteps of the Avars (568-800 AD) when the Huns had already disbanded. The formative stage of Slavic in time and space within the wider Prague-Korchak-Penkovka triangle thus appears to be congruent to the previous Alan expansion on the eve of the Great Migration. Polish Sarmatism may be a reminiscent memory of these events, even though this tradition conveniently forgot about the incorporation of the classical Sarmatians into the Alan might. If the Ugric hypothesis holds, at least one of these nations from the steppes must have been Ugric already before the Magyars appeared on the scene.

According to Nekhleh, Ringe and Warnow (2005), “Baltic is most unlikely to have begun diverging from Slavic by 1000 B.C.E., because Proto-Slavic seems still to have been more or less uniform in the 8th c. C.E., and Proto-Baltic and Proto-Slavic are so similar that they had probably been diverging for less than two millennia.” Their study did not take into consideration the possibility of an accelerated development due to a foreign substratum, though even like this the development of Slavic from Slavo-Baltic thus fits into the timeframe of historic Scytho-Sarmatian expansions that are recorded since Herodotus (5th century BC).

Building further on the assumption of an originally Ugric superstratum of the Slavic nation that amassed along the Frankish border, the purported central asiatic origin of this elite may now be verified with the genetic results of the Andronovo samples.

Christine Keyser et al. This table gives an overview of the matches that were found in the YHRD database with the paleogenetic samples of Bronze Age skeletons in South Siberia, attributed to the Andronovo Culture.

What if the Andronovo culture, or at least an important part of it, was indeed Ugric, then wouldn’t the arrival of a superstratum from this particular direction match corresponding haplotypes in Europe? Hence I calculated the percentages of the matching haplotypes per country in YHRD.

Country # Samples (YHRD) # Matches % all-over R1a1 freq. % of R1a1
Ireland 155 0 0% N.A. 0%
Austria 715 1 0.14 % 14% 1%
Norway 1766 3 0.17% 18% 0%
Poland 4581 27 0.59% 56% 1.1%
Czechia 2331 13 0.56% 33.7% 1.7%
Slovakia 856 3 0.35% 27% 1.3%
East Germany + Hamburg 900 15 1.7% 15% 11%
Slovenia 180 7 3.9% 38% 10.2%
Ukraine 586 6 1.0% 42% 2.4%
Hungary 116 4 3.4% 20.4% 17%

The Hungarian percentage would drop to 6.5% if the 184 “Romani” (=Roma gypsies) samples are included. Although the Romani samples don’t reflect population proportions because of a sampling bias, the overall matches still result in a significant “Ugrian” (i.e. Hungarian) component, and thus wouldn’t compromise an important Ugrian indication of the Andronovo culture.


  • Christine Keyser et al. – Ancient DNA provides new insights into the history of south Siberian Kurgan people, Human Genetics, Saturday, May 16, 2009, doi: 10.1007/s00439-009-0683-0, 2009, link
  • List of numbers in various languages, Wikipedia, link
  • YHRD.org, link
  • István Fodor – The Main Issues of Finno-Ugrian Archeology, in Ancient Cultures of the Uralian People by Peter Hajdú, 1976, ISBN 963 13 3019 2
  • Pál Lipták – Anthropology of the Finno-Ugrian Peoples, in Ancient Cultures of the Uralian People by Peter Hajdú, 1976, ISBN 963 13 3019 2
  • Luay Nakhleh, Don Ringe & Tandy Warnow – Perfect Phylogenetic Networks: A New Methodology for Reconstructing the Evolutionary History of Natural Languages, 2005, link
  • Underhill et al.- Separating the Post-Glacial Coancestry of European and Asian Y Chromosomes Within haplogroup R1a, 2009, link
  • Jadranka Gvozdanović – Indo-European numerals, 1991, link
Categories: DNA, Indo Europeans
  1. Usman
    January 27, 2010 at 20:59

    I like your blog

  2. dpwes
    April 9, 2010 at 15:01

    Modern Hungarians are North-Central Europeans in terms of Y-DNA and genome-wide SNPs.

    Also, Andronovo tribes were closely related to Timber-Grave tribes, who were largely of Corded Ware (Fatyanovo) and Yamnaya origin.

    So the Andronovo R1a1 and Slavic R1a1 come from North-Central and Central-East Europe, not from any Ugric migration.

    • April 9, 2010 at 17:10

      This article investigates the facts that link paleogenetic Andronovo samples and the people just east of the historic Frankish border. The ultimate origin of Andronovo is likely to be more into the direction of the Ural. Any likeliness of modern Hungarians with neighboring people is not an issue in this article.

  3. dpwes
    April 27, 2010 at 03:25

    One of your arguments in this article is that the Andronovo R1a1 is Ugric because Hungarians carry a lot of similar haplotypes. But the haplotypes that the Hungarians carry belong to Central European R1a1 (Germanic and West Slavic). Try and prove otherwise.

    BTW, as per above, the Andronovo tribes derive from several migration waves from Central Europe. Just because they formed near the Urals doesn’t make them genetically Uralic.

    • April 27, 2010 at 06:14

      This is only about the subset of Central European R1a1 haplotypes that is similar to paleogenetic Andronovo samples. The peak, indeed, is in Central Europe: East Germany (11% of all R1a1), Slovenia (10.2%) and Hungary (17%). The rest of R1a1 in the neighbourhood does not show the same affinity and, like you say, could be local Central European.

  4. dpwes
    May 15, 2010 at 00:26

    There’s no point counting these haplotypes as a percentage of R1a1. What you have to do is count them as a percentage of all the haplotypes in the database for each population.

    From the above study…

    “As shown in Table 5, and particularly in Fig. 3, the current distribution of the ancient mtDNA haplotypes can be broadly divided into three different geographic poles. The first is represented roughly by eastern and northern Europe, the second by the Volga–Ural region and the third by southern Siberia. It is interesting to note that the distribution of the paternal and maternal lineages is close. Indeed, except for the Volga–Ural region, both maps overlap.”

    • May 15, 2010 at 08:20

      That is what I did. The YHRD database only stores haplotypes, so you have to calculate the matches as a fraction of the current R1a frequency of the population.

      We already know there is a big geographic hole in the Y-DNA matches between South Siberia and the Frankish border. This hole hardly exists for mtDNA. The most recent movements involve Y-DNA, in this case the most shallow marker that shows the best contrasts. We don’t need to take this case even further back than Andonovo. If we do, we can’t ignore the Andronovo mtDNA matches with the Volga-Ural region either.

      This article takes the peculiar distribution pattern along the North-South fringes of the Frankish Empire as an indication of a distinct ethnic element, a military “Sarmatian” elite during the medieval Slav-related expansions, whose ultimate origins may be sought in the Andronovo horizon. The tentative affiliation of this ethnic element with Hungarians does not confirm the matches are native to Central Europe.

  5. tibor
    June 20, 2010 at 21:05

    You did not calculate with Czech in your table. It does not match with your theory? I think, finno-ugrians are Y-DNA N, not R1a. Haplogroup N has a wide geographic distribution throughout northern Eurasia, from Norway to China. Its highest frequency occurs among the Finnic and Baltic peoples of northern and eastern Europe, the Ob-Ugric (linguistically the nearest to Hungarian language by scholars) and Northern Samoyedic peoples of western Siberia (wikipedia). They are original finno-ugrian speakers, or not? Hungarian scholars fight about the N haplogrouph very strongly. I saw, in hungarian head it was only matter of time to skip from N to R1a, percentage of it is much better. A new theory is coming, super. I have read, Germans are of R1a origin too, because excavations with R1a haplotypes were done on the German territory.

  6. June 20, 2010 at 23:03

    I inserted the Czech rates as well, assumed the YHRD database still reflects the data of Christine Keyser et al.
    The high concentration of matches in the Czech Republic drawn on Keyser’s map are hampered by a high sampling bias, to the effect that the actual percentage is not as high as other Slavic strongholds on the Frankish borders. Like the situation in Austria, this might be caused by an early breach at the Slavic defense in Moravia and an extensive Frankish involvement.
    Still, also here the number of matches just east of the Frankish demarcation line is significant.

    BTW. These paleogenetic “German” Corded Ware R1a samples are of a different type and nothing to do with this article.

  7. comod
    September 16, 2010 at 11:17

    Map of Prague-Korchak-Penkovka is incorrect, mistaken for the Oder to the Elbe. Here is more correct -> http://www.ctdlib.org/images/hd3img/slavicultures.png, just for variety mistaken Kolochin Korchak culture.
    Here the same Prague-Korchak -> http://img121.imageshack.us/img121/6686/83702753.jpg

    • September 16, 2010 at 22:18

      Does this sound familiar?

      “The boundaries of these cultures were transformed considerably in the late sixth century and into the seventh century. Although the areas occupied by the Kolochin and Penkovka cultures remained the same, the Prague culture spread widely to the west: it encompassed the basin of the Middle Danube and the upper and middle Elbe. At the same time a new phenomenon arose in the basin of the Oder and on the southern coast of the Baltic Sea: the Sukow culture, most likely the younger stage of the Prague culture.”

      This illustrates that the maps are more or less correct just before the Franks and Slavs met. At some locations inbetween the Slavs were definitely first, but all of this was the result of movements at the frontiers.

  8. Zima
    September 20, 2010 at 00:07

    Rokus, please check „A Vedic Grammar for Students” Arthur Anthony Macdonell 2004 page 136. Root ‘yuj’-“to join” (in Vedic notation ‘j’ is pronounced as ‘dz’) is cognate to Slavic root ‘jed’-“to join” (in Slavic notation ‘j’ is pronounced as ‘y’ so it is the same root). Conjugation of that root in Vedic is close to Slavic with identical umlaut and one metathesis ‘n’ ‘d’: Slavic ‘jednotem, jednocisz, jednoti’–>Vedic ‘yunajmi, yunakszi, yunakti’ (it should be ‘yujnatemi’ in PIE and Proto-Slavic). Slavic conjugation of verbs is more archaic, conservative and closest to PIE, Vedic has evolved and changed but it is still close, not as Baltic conjugation, which is completely different and does not resemble PIE.
    Slavic often uses diminutive of ‘jeden’ which is ‘jedynkam’, ‘jedynka’ and from this comes Vedic ‘ekam’, ‘ekah’, ‘eka’.

    Polish language doesn’t have any not-IE substratum. Baltic has a lot of not-IE substratum, it has plenty of Ugro-Finnic features in grammar and lexicon, for example four cases in Lithuanian: inessivus, adessivus, illativus and allativus are not IE but Ugro-Finnic. There are many Ugro-Finnic and Paleoeuropeand words in Baltic lexicon. Which is quite understandable as half of Baltic population is hg. N1c, an Ugro-Finnic marker absent in Poland.

    Anthropologists have shown by cranium analysis that the population of Przeworsk culture was identical to the population of Wielbark culture and both those populations were identical to Early Middle Ages Slavic population of Poland. This very well agrees with recent Underhill 2009 results which show genetic continuity of Polish population since at least the Bronze Age. There was no population replacement in Poland. It is also confirmed by linguists: Polish is the most archaic language among Slavic languages and does not have non Slavic substratum and also Hydronyms in Poland are Slavic. You know very well that presence of Germanic tribes in Poland like Goths is a fantasy of German and Scandinavian nationalists and presence of Balts is a fantasy of Baltic nationalists like Gimbutas. There are no genetic traces of them. So you hypothesis seems be wrong and based on mostly outdated information.

  9. September 20, 2010 at 11:53

    I appreciate your efforts, but first of all Arthur Anthony Macdonell 2004 page 136 doesn’t say any such thing and does not link ” yuj” to IE “one” forms, not even to a possible Slavonic *ed (“only”?). He just gives a list of conjugations. Besides, to derive Sanskrit “ekam” from a regular suffix is highly controversal.

    The most common view can be read at Jadranka Gvozdanović – Indo-European numerals page 726 etc. (1991):
    “In Slavonic, the simple form *ein- survives only with shifted meaning, as “other” […] Elswhere, a prefix ed-, jed-, od- is added […] The prefix CSI. *ed- is not fully clear in origin”

    Racial continuity does not contradict linguistic reforms, that for sure don’t imply the linguistic/population replacement scenarios that you have in mind.

    Ultimately, the Finnish influence on Baltic you refer to only concerns the north, that didn’t become slavonized. This influence is anachronistic to the Baltic substratum referred to in this article.

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: