Corded Ware substrate hypothesis

From Old European


The common traits found between Indo-Iranian and Balto-Slavic – not related to each other – and between Balto-Slavic and Germanic – not more related between them than to Italo-Celtic –, have puzzled Indo-Europeanists for more than a century.

While common substrate vocabulary and isoglosses have been proposed, there is no coherent picture to date of their actual relationship.

In this paper we connect recent genetic investigation with the potential substrate language common to the three branches, represented by Corded Ware culture groups of the North Caspian region, central Europe, and Scandinavia. Furthermore, we argue that populations of the Corded Ware culture may have spoken Uralic proto-languages.

Different communities with a common origin

Both European cultures (mainly of R1b-L51 subclades) and eastern Pontic-Caspian steppe cultures (mainly of R1b-Z2103 subclades) underwent an evolution represented by their absorption into populations of Corded Ware lineages (represented mainly by R1a-Z645 subclades). This evolution happened roughly at the same time, so it could be argued that these northern-eastern European peoples of the Corded Ware culture happened to speak similar dialects that could have influenced their adoption of Indo-European languages.

In Europe, North-West Indo-European communities speaking Pre-Germanic merged with peoples from the Battle Axe culture during the Dagger Period of the late Nordic Neolithic (ca. 2400-1700), while Pre-Balto-Slavic probably merged with Corded Ware cultures in the Únětice or Mierzanowice/Nitra cultural regions (ca. 2300-1600). In the steppe, Graeco-Aryan dialects spoken in the eastern Yamna and Poltavka cultures were replaced by peoples of Abashevo origin forming the Potapovka and Sintashta cultures (ca. 2100-1800 BC). Because both dialects, a Northern and a Southern IE one, already developed quite differently, evolved in a similar manner, their changes may be explained by a common Corded Ware substrate language.

The nature of this proposed substrate language may thus be a priori non-Indo-European, Pre-Indo-European, or Indo-European.  

Uralic as the language of the Corded Ware culture

It has been classically proposed that a Mesolithic language of eastern Europe is to be identified with a Uralic community, and a date ca. 4000 BC has been proposed for the common reconstructible Proto-Uralic language[Parpola 2012][Kortlandt 2002]. Furthermore, Finno-Ugric has been shown to have developed in close contact with Proto-Indo-Iranian[Kallio 2002].

A common Indo-Uralic[Kortlandt 2002][Kloekhorst 2008] community is probably to be traced back to the formation of early Sredni Stog and early Khvalynsk cultures at the end of the 6tʰ millennium, and their development as Uralic and Indo-European respectively is traced to their independent evolution during the Eneolithic in the Pontic-Caspian steppe, west and east of the Don River, respectively[Quiles 2017].

Recent genetic investigation has shown that the expansion of the third Corded Ware horizon was closely related to the cultures of the north-west Pontic steppe, heirs of the early Sredni Stog culture. This is therefore to be related to the expansion of the main Proto-Uralic dialects.  

Indo-Iranian and Balto-Slavic

It has been argued that similarities found in Indo-Iranian and Balto-Slavic languages – like the peculiar phonetic ruki development, a similar satem trend in both groups[Meier-Brügger 2003] – suggest a sort of west-east continuum between both languages, with certain features running through them[Mallory and Adams 2007].

From a linguistic point of view, the characteristic palatalization of the consonant system in Proto-Uralic - including palatalised *ć, *ś (and postalveolar *č, *š) alongside plain velar *k and dental *s –, is compatible with the similarly transposed velar and sibilant system adopted for Late Indo-European dialects by Balto-Slavic and Indo-Iranian speakers, thus explaining the strongest phonetic connection between these dialectally diverse Indo-European languages. Differences in the Baltic and Slavic satemization processes also point to an early split of the North-West Indo-European dialect ancestral to both, before or during its assimilation by different Uralic-speaking communities of late Corded Ware cultures.

The potential satem influence argued to be behind certain phonetic developments of Anatolian (especially Luwian) and certain Paleo-Balkan languages can also be posited to be the result of adoption of these traits during the crossing of territories of the Sredni Stog / Corded Ware horizon, during the migration of Indo-Hittite and Late Indo-European speakers respectively, although they most likely represent independent satemisation processes (see above).

This model supports thus the reconstruction of two series of velars: the traditional reconstruction of dorsovelars and labiovelars[Lehmann 1952], which is usually ignored in common textbooks in favour of the older reconstruction of a third series of palatovelars[Bomhard 2015]; but also Martinet’s glottalic consonants[Gamkrelidze and Ivanov 1995].

The developments of Proto-Finno-Ugric *ō, *ē - → Proto-Ugric *a, *ä – merging with original *a, *ä –[Häkkinen 2009] could be related to the phonetic changes found between Late Indo-European and Proto-Indo-Iranian, i.e. *e/o → *a, *ē/ō→*ā. That would suggest that the easternmost part of the domain, including probably the Abashevo-Balanovo cultures, spoke Proto-Ugric or a related Finno-Ugric language, at roughly the same time as the assimilation of the (Pre-Indo-Iranian-speaking) Poltavka population happened within the Sintashta and Potapovka cultures, ca. 2100 BC.

Balto-Slavic and Indo-Iranian share a special position among Indo-European languages regarding their rather conservative nominal case system. It has been argued that languages with more second language speakers lose nominal cases[Bentz et al. 2015]. It has also been shown that forces driving grammatical change are different (stronger) than those driving lexical change[Greenhill et al. 2017]. These natural human trends would explain the higher simplification of the declension system in Late Indo-European dialects of west and south-east Europe, compared with the conservation of the original system by speakers of Uralic dialects, known for their large set of grammatical cases. At the same time, the greater stability of lexicon would support the close relationship of European languages of the North-West Indo-European group.

On the other hand, this could also give support to the theory that Late Proto-Indo-European had in fact a simpler nominal system, derived from a still simpler one of Middle Proto-Indo-European[Adrados, Bernabé, and Mendoza 2016]. In this case, Indo-Iranian and Balto-Slavic morphological differences would be later innovations. However, that would need an explanation as to how Uralic speakers adopting Late Proto-Indo-European added complexity to the language, instead of simplifying it.

Germanic and Balto-Slavic

A western Corded Ware substratum could also be argued to be the origin of certain common isoglosses found between Germanic and Balto-Slavic.

According to Kortlandt[2016], the similarities between both dialects must be due to a common Indo-European substrate, since there is no reason to assume early contacts between Germanic and Balto-Slavic. In terms of the “Temematic hypothesis”, which favours a satem or Indo-Slavonic group, Germanic and Temematic would share common western Corded Ware isoglosses, and only later would Proto-Balto-Slavic – already separated from Proto-Indo-Iranian – absorb Temematic as a substratum language.

On the other hand, the expansion of East Bell Beaker peoples must have happened in different waves: one to the east, through Moravian and Bohemian groups into Polish lands and north European lowlands; and one to the west, including Middle Elbe/Saale and Dutch groups, which later migrated into southern Scandinavia, maybe absorbing some common linguistic traits in their north-eastern migration through the European lowlands.

Especially important is the peculiar dative plural in *-m- shared by Germanic and Balto-Slavic, which can’t be explained by late influences. Because of that, Kortlandt[2016] has argued that LIE dative plural *-mus must have been replaced by the ablative ending *-bhos in Italo-Celtic and Indo-Iranian (where *-bhi̯os may reflect the attachment of *-os to the instrumental forms in *-bhi-). Nevertheless, on one hand there is a general consensus that the original form behind Sla. *-mъ and O.Lith. -mus (maybe influenced by Old Prussian) must have come from a dative-ablative plural *-mos (cf. PGmc *-maz), and not from *-mus as suggested by Georgiev[1966] and Kortlandt[Halla-aho 2006]. Similarly, the common instrumental in *-mi- behind Germanic and Balto-Slavic forms contrasts with the rest of the Late Indo-European domain, which shows *-bhi-.

The Uralic declension system of genitive in *-n, locative in *-n- (with three-way systems in later periods), as well as the lative in -ŋ, may have influenced the change of dative-ablative and instrumental forms in *-bʰ- → *-m-. Judging from samples of potential Indo-Uralic cognates, the correspondence between Proto-Indo-European and Uralic forms has been tentatively reconstructed by Kroonen[2015] as follows: PIE *d – PU *n; PIE *bʰ – PU *ŋi; PIE gʰ – PU *ŋ; PIE *gʰʷ – PU *uŋ. Interesting in this respect may also be the Livonian dative in -n, only partially stemming from the Uralic genitive in *-n, and which has strong links to the Latvian dative in *-m-[Seržant 2015].

It could then be hypothesised that North-West Indo-European had the old dative-ablative and instrumental forms in *-bʰ- during the initial migration of East Bell Beaker groups into Corded Ware territories of the northern lowlands. There, the declension system would have undergone a slight phonetic change (adapted to the somehow similar Uralic case system), e.g. ins. sg. *-bʰi → *-ŋi, ins. pl. *-bʰis → *-ŋis, and (maybe by assimilation with the other two forms) dat.-abl. pl. *-bʰos →*-ŋos. Such a change would obviously need an additional, ad hoc explanation for the change *-ŋ- to the reconstructed common *-m-. An explanation may be found in the lack of the phoneme /ŋ/ in the definitive phonetic system adopted, thus compelling for the eventual adoption by the next generations of speakers of a different phoneme, /m/, already present in the declension system (in the accusative singular ending), and – in contrast with /n/ – without the possibility of confounding these forms with the accusative plural in *-ns. This (now fully Indo-European) substrate language of north-central European Bell Beakers would have later influenced western groups during their migration into Scandinavia through the northern lowlands, and they would have remained as a part of the eastern Bell Beaker groups that later formed the Únětice and the Iwno-Mierzanowice cultures.

Witness to this intermediate substrate may also be the common forms of Indo-European origin found in Germanic and Baltic, and to some extent in Slavic, limited to social phenomena and especially to technical terms for wooden tools and utensils[Kortlandt 2016].

The same substrate could be argued to be behind certain traits common to Germanic, Balto-Slavic, Uralic, and other Eurasian languages[Klesment et al. 2003] – although many are constrained to Balto-Slavic and Uralic, which probably developed late in neighbouring territories.

The Germanic passive ending in *-i, in contrast with the original PIE ending in *-r-, may also be related to a common loss of the middle-passive endings in both Germanic and Balto-Slavic (or in the substrate language). It would have then been remade later with the common primary ending *-i, during the development of a Germanic community in Scandinavia after the Dagger Period, and only traces of the ending *-r with an impersonal value are left in Germanic.

Supporting the presence of an intermediate Indo-European substrate before the formation of a Pre-Germanic community would be the lack of a strong phonetic influence from Uralic, as found in Indo-Iranian and Balto-Slavic. Its development in Scandinavia during the unification represented by the Dagger Period must have been influenced by different regional cultures.  

Common traits and other substrate hypotheses

Common traits between Germanic, Balto-Slavic, and Indo-Iranian, not coincident with other Indo-European branches, are quite difficult to find. After all, any Uralic traits common to the three branches may be related to an original Indo-Uralic community.

One such example is found in the reconstructed IE *manu-, ‘man’, cf. O. Ind. mánu-, Avestan Manuš-čiɵra-, Gmc. *mann-, Sla. *mǭ̀žь (< **mon-gio-?, with suffix similar to Lith. žmo-g-ùs, ‘man’). It has its parallel in PU *mańć-, ‘man’, cf. Hu. magyar, Finn. mies, Khanty mańt́, Mansi mäńćī. Its connection with an Indo-Uralic word may be made through a potential PToch *mänśu-, ‘prince’, as reconstructed by Adams, although this etymology (from Toch. A mäśkit, B mäñcuṣke) is dubious at best, and such a frozen use could have been influenced by the Indo-Iranian expansion in the region. The natural use of the word ‘man’ to describe an ethnic group would place it in a good position to survive in superstrate languages replacing the Uralic languages of cultures remaining in close contact with Uralic-speaking peoples. The lack of such an essential word – also strongly connected to basic mythological cosmology – in the other attested Indo-European dialects is difficult to justify.

Many of the Indo-Iranian substrate words and word forms described by Lubotsky[2001], most of them probably of non-Indo-European origin, may have been in fact of Uralic origin: “as is well known, Uralic has heavily borrowed from Indo-Iranian, but I agree with those scholars who believe that many of the apparent early borrowings rather reflect an etymological relationship between Uralic and Indo-European”.

Kroonen’s agricultural substrate hypothesis relates the substrate vocabulary and noun inflection traits to a Middle Eastern language, potentially related to Proto-Semitic[Kroonen 2012], which he has only recently related to the adoption of the language of the Funnelbeaker culture in Scandinavia[Iversen and Kroonen 2017], presupposing that Corded Ware peoples spoke Indo-European dialects. However, the same substrate could be argued to have influenced the third Corded Ware horizon, from the interaction of Balkan and steppe communities in the north-west Pontic steppe, since it is known that there is a strong genetic and cultural (and thus probably linguistic) connection of Balkan Chalcolithic cultures to Neolithic Anatolian farmers.

This is compatible with the idea that no words for domesticated animals can be reconstructed for Proto-Uralic, safe for dog[Pereltsvaig and Lewis 2015]. Hence peoples from the western steppe (mainly late Sredni Stog and Kvitjana cultures) might have borrowed them during the formation of the Third Corded Ware Horizon (through the influence of Trypillian, GAC, and Baden cultures), and they would have expanded with initial migration to the north-west.

That substrate, common to the western Uralic dialects spoken by Corded Ware groups across northern Europe, would have then been assimilated to different degrees by both Pre-Germanic and Balto-Slavic communities absorbing Corded Ware groups – and even Pre-Greek communities because of contacts in the Balkans –, as the examples in Kroonen[2012] show.


  • [Adrados, Bernabé, and Mendoza 2016] ^ Adrados, F.R., A. Bernabé, and J. Mendoza. 2016. Manual of Indo-European Linguistics II: Nominal and Verbal Morphology. Edited by P. Swiggers. 3 vols. Vol. 2, Orbis Supplementa. Leuven: Peeters.
  • [Bentz et al. 2015] ^ Bentz, Christian, Annemarie Verkerk, Douwe Kiela, Felix Hill, and Paula Buttery. 2015. Adaptive Communication: Languages with More Non-Native Speakers Tend to Have Fewer Word Forms. PLOS ONE 10 (6):e0128254.
  • [Bomhard 2015] ^ Bomhard, Alan R. 2015. A Comprehensive Introduction to Nostratic Comparative Linguistics. With special reference to Indo-European. Second revised, corrected and expanded edition (as of May 2017) ed. 4 vols. Vol. 1. Charleston, SC.
  • [Gamkrelidze and Ivanov 1995] ^ Gamkrelidze, T. V., and V. V. Ivanov. 1995. Indo-European and the Indo-Europeans: A Reconstruction and Historical Analysis of a Proto-Language and Proto-Culture. Part I: The Structure of Proto-Indo-European. Part II: Semantic Dictionary of Proto-Indo-European Language. Vol. 80. Walter de Gruyter, 1995. Edited by W. Winter. Vol. 80. Berlin / New York: Mouton de Gruyter.
  • [1966] ^ Georgiev, Vladimir Ivanov. 1966. Introduzione alla storia delle lingue indeuropee. Vol. 9: Ed. dell'Ateneo.
  • [Greenhill et al. 2017] ^ Greenhill, Simon J., Chieh-Hsi Wu, Xia Hua, Michael Dunn, Stephen C. Levinson, and Russell D. Gray. 2017. Evolutionary dynamics of language systems. Proceedings of the National Academy of Sciences 114 (42):E8822-E8829.
  • [Häkkinen 2009] ^ Häkkinen, Jaakko. 2009. Kantauralin ajoitus ja paikannus: perustelut puntarissa. SUSA/JSFOu 92:9-56.
  • [Halla-aho 2006] ^ Halla-aho, Jussi. 2006. Problems of Proto-Slavic Historical Nominal Morphology On the Basis of Old Church Slavic, Faculty of Arts, University of Helsinki, Helsinki.
  • [Iversen and Kroonen 2017] ^ Iversen, Rune, and Guus Kroonen. 2017. Talking Neolithic: Linguistic and Archaeological Perspectives on How Indo-European Was Implemented in Southern Scandinavia. American Journal of Archaeology 121 (4):511-525.
  • [Kallio 2002] ^ Kallio, Petri. 2002. Prehistoric Contacts between Indo-European and Uralic. In Proceedings of the Thirteenth Annual UCLA Indo-European Conference, edited by K. Jones-Bley, M. E. Huld, A. D. Volpe and M. R. Dexter. Washington, DC: Institute for the Study of Man.
  • [Klesment et al. 2003] ^ Klesment, Piret , Ago Künnap, Sven-Erik Soosaar, and Rein Taagepera. 2003. Common Phonetic and Grammatical Features of the Uralic Languages and Other Languages in Northern Eurasia. JIES 31 (3 & 4):1-27.
  • [Kloekhorst 2008] ^ Kloekhorst, Alwin. 2008. Some Indo-Uralic Aspects of Hittite. JIES 36 (1 & 2).
  • [Kortlandt 2002] ^ 1 2 Kortlandt, Frederik. 2002. The Indo-Uralic verb. In Finno-Ugrians and Indo-Europeans: Linguistic and literary contacts. Maastricht: Shaker.
  • [2012] ^
  • [Kroonen 2012] ^ Kroonen, Guus. 2012. Non-Indo-European root nouns in Germanic: Evidence in support of the Agricultural Substrate Hypothesis. In A Linguistic Map of Prehistoric Northern Europe, edited by R. Grünthal and P. Kallio. Helsinki: Suomalais-Ugrilaisen Seura.
  • [2015] ^ Kroonen, Guus. 2015. Indo-Uralic lookalike sets, an etymological quick scan. In The Precursors of Proto-Indo-European: The Indo-Hittite and Indo-Uralic Hypotheses. Leiden University, 9-11 July 2015.
  • [Lehmann 1952] ^ Lehmann, W. P. 1952. Proto-Indo-European Phonology. Austin, TX: University of Texas Press.
  • [2001] ^ Lubotsky, Alexander. 2001. Early Contacts between Uralic and Indo-European: Linguistic and Archaeological Considerations. Paper read at Papers presented at an international symposium held at the Tvärminne Research Station of the University of Helsinki 8-10 January 1999. (Mémoires de la Société Finno-ougrienne 242.), at Helsinki.
  • [Mallory and Adams 2007] ^ Mallory, J., and D.Q. Adams. 2007. Reconstructing the Proto-Indo-Europeans. In The Oxford Introduction to Proto-Indo-European and the Proto-Indo-European World. Oxford: Oxford University Press.
  • [Meier-Brügger 2003] ^ Meier-Brügger, Michael. 2003. Indo-European Linguistics. Berlin, New York: Walter de Gruyter.
  • [Parpola 2012] ^ Parpola, Asko. 2012. The problem of Samoyed origins in the light of archaeology: On the formation and dispersal of East Uralic (Proto-Ugro-Samoyed). Mémoires de la Société Finno-Ougrienne (264).
  • [Pereltsvaig and Lewis 2015] ^ Pereltsvaig, Asya, and Martin W. Lewis. 2015. The Indo-European Controversy. Facts and Fallacies in Historical Linguistics. Cambridge: Cambridge University Press.
  • [Seržant 2015] ^ Seržant, A. 2015. Dative experiencer constructions as a Circum-Baltic isogloss. In Contemporary Approaches to Baltic Linguistics, edited by P. Arkadiev, A. Holvoet and B. Wiemer. Berlin/Boston: Walter de Gruyter.