Indo-Uralic is a hypothetical language family consisting of Indo-European and Uralic, which was already proposed in the 19th century and has been the subject of discussion ever since. Most linguists are not convinced of the relationship, but few would opine that Indo-European and Uralic could not be related.
History of the hypothesis
The tremendous success of comparative works on the Indo-European languages in the 19th century naturally led to considerations whether further, more distant language relationships could be established by the same methods. One of the connections considered was one between Indo-European and the neighbouring Uralic family.
A brief history of early Indo-Uralic studies can be found in Holger Pedersen’s Linguistic Science in the Nineteenth Century (1931:336-338). As early as 1869, the Danish linguist Vilhelm Thomsen proposed a genealogical relationship between Indo-European and Uralic, but this received little enthusiasm, and Thomsen did not pursue the idea further. Nikolai Anderson published a work on the matter in 1879. The phonetician Henry Sweet argued for kinship between Indo-European and Finno-Ugric in his semi-popular book The History of Language in 1900. Further studies followed, among others by Björn Collinder and recently by Frederik Kortlandt and Alwin Kloekhorst. Yet, despite all evidence adduced to the hypothesis, most historical linguists working in this field still consider the matter undecided.
The Indo-European and Uralic language families are neighbours in Europe and probably always have been. The homeland of Proto-Indo-European is now believed to have been located north of the Black Sea by most relevant scholars, that of Proto-Uralic between the middle Volga and the Ural mountain range by many specialists. The Dutch linguist Frederik Kortlandt has proposed that the Proto-Indo-Uralic homeland may have been located north of the Caspian Sea.
Indo-Uralic and larger proposed families
If Indo-European and Uralic are related to each other, this bears the possibility that further languages are related to the two. It cannot even be ruled out that the latest common ancestor of Indo-European and Uralic is also an ancestor of further languages. On the Indo-European side, there are few good candidates for such a relationship, the best perhaps being Etruscan, though the resemblances between Indo-European and Etruscan are not better than those between Indo-European and Uralic. On the Uralic side, Yukaghir, Chukotko-Kamchatkan and Eskimo-Aleut have been proposed as relatives of Uralic (see Michael Fortescue, Language Relations across Bering Strait, who proposes a Uralo-Siberian family consisting of Uralic and these three groups). All of these are spoken in areas far to the east, and most of the morphological resemblances between them and Uralic are also seemingly shared by Indo-European, and Kortlandt has proposed that Indo-Uralic has to take the place of Uralic in Fortescue's Uralo-Siberian family (which then should receive a different name - perhaps Euro-Siberian?).
The most commonly advanced arguments for a relationship between Indo-European and Uralic are based on similarities in morphology that are difficult to dismiss as either chance resemblance or the result of borrowing. Most prominent among these are the pronomninal roots *m- (1st person), *t- (2nd person), *kʷ-/k- (interrogative) (see also Mitian for a larger proposed unit based on these). The verbal personal endings are also similar between both families, but as they are obviously connected with the pronouns in both families, these do not actually constitute a further data point.
Both language families further share an accusative ending *-m, and the Uralic ablative *-ta has been compared to the IE ablative *-d. The Uralic genitive *-n may be related to the IE denominal adjective suffix *-no-. Also, a relationship between Uralic dual *-k and plural *-t, and IE dual *-eh₁ and plural *-es has been proposed.
There is a number of lexical resemblances between Indo-European and Uralic (see list below). However, lexical resemblances between geographically adjacent languages may be due to borrowing, and it remains to be shown that this is not (always) the case here. Many resemblances can indeed be shown to be loanwords by the presence of post-PIE developments being visible in the forms seen in Uralic languages. For example, the numeral *śëta/*śata '100' of Finno-Ugric distribution shows a typically 'satem' development and thus was probably borrowed from (Pre-)Proto-Indo-Iranian into the Finno-Ugric languages.
Some of the lexical resemblances, however, form parts of the basic vocabulary which is widely considered resistant to borrowing. These show no known post-PIE innovations in the Uralic forms of the words, and a descent from a common ancestor cannot be ruled out.
Some of the most promising items may be the following:
- 'I' (1PSg): PIE *h₁me- — PU *minä ~ mun (alternates in PIE with *(h₁)eǵ-)
- 'thou' (2PSg): PIE *ti- — PU *tinä ~ tun (the PIE form attested only in Anatolian, levelled in favor of the alternant *tu- elsewhere)
- 'who': PIE *kʷi-s — PU *ku-
- 'name': PIE *h₁nōmn̥ — PU *nimi
- 'water': PIE *wodr̥ — PU *weti (contrasts in PU with a secondary root *śäčä)
- 'to drink': PIE *(h₁)ēgʷʰ- — PU *jëxi-
- 'to fear': PIE *pelh₁- — PU *peli-
- 'to go': PIE kʷelH-e/o- — PU *kulki- (secondary PU root: more widespread is *meni-)
- 'to hear': PIE *ḱlew- — PU *kuwli- (or *kowli-?)
- 'to plait': PIE *pn̥H-e/o- — PU *puna-
- 'to wash': PIE *mezg- — PU *mośki-
- PIE *dʰeh₁- 'to put' — PU *teki- 'to do'
- PIE *deh₃- 'to bring' — PU *toxi- 'to give' (contrasts with a secondary PU root *amta-; cf. also the next)
- PIE *h₂mey- 'to exhange' — PU *mexi- 'to sell'
- PIE *wedʰ- 'to lead' — PU *wetä- 'to pull'
- PIE *weǵʰ- 'to carry' — PU *wixi- 'to take (from)', *wexi- 'to take (to)'
Kortlandt (2001) adds 'to drive': PIE *h₂aǵ- — PU *aja-, but here the correspondence *ǵ ~ *j suggests loaning from Indo-Iranian to Uralic.
The five (pro)nominal matches and the verbs 'to hear', 'to drink' are present on the 100-item Swadesh list. Kassian, Zhivlov & Starostin (2014) argue that these seven matches should be considered sufficient evidence to conclude a relationship. The longer 207-item list also includes 'to fear', 'to give', 'to wash'.
Koivulehto (2002) analyzes both nominal matches here and the verbs 'to fear', 'to go', 'to plait', 'to wash', 'to exchange/sell', 'to lead/pull' (as well as some non-core vocabulary items not mentioned here: 'to bore', 'to must', 'pole', 'price') as a layer of loanwords from PIE to PU. This has been criticized as semantically implausible by Helimski (2002).
Some possible cognates
(from the Wikipedia article)
|first person singular||*-m||*-m|
|first person plural||*-me||*-me|
|second person singular||*-s (active), *-tHa (perfect)||*-t|
|second person plural||*-te||*-te|
|nominative/accusative plural|| *-es (nominative plural)
*-n̥s (accusative plural) < *-m̥ (acc.sg.) + *-(e)s (pl.)
|oblique plural||*-i (pronominal plural, as in *we-i- 'we', *to-i- 'those')||*-i|
|'and' (postposed conjunction)||*-kʷe 1||*-ka ~ *-kä 2|
|negative particle 'not'||*ne 3||*ne 4|
|'I, me'|| *me 'me' (accusative) 5
*mene 'my' (genitive) 6
|*mun, *mina 'I' 7|
|'you' (singular)|| *tu (nominative) 8
*twe (accusative) 9
*tewe 'your' (genitive) 10
|*tun, *tina 11|
|demonstrative pronoun||*so 'this, he/she' (animate nominative singular) 12||*sä 'he/she, it' 13|
|demonstrative pronoun||*to- 'this, that' 14||*tä 'this', *to 'that' 15|
|'who?' (interrogative pronoun)|| *kʷi- ~ *kʷe- ~ *kʷo- 'who?, what?' 16
*kʷi/e/o- + -ne 'who?, what?' 17
| *ki ~ *ke ~ *ku ~ *ko 'who?, what?' 18 |
*ken 'who?' 19
|'to give'||*deH₃- 20||*toHi- 21|
| 'to moisten',
| *wed- 'to wet', 22
*woder- 'water' 23
|*weti 'water' 24|
|'name'||*nomen- 'name' 25||*nimi 'name' 26|
|'fish'||*kʷalo- 'large fish' 27||*kala 'fish' 28|
|'sister-in-law'||*galou- 'husband's sister' 29||*kälɜ 'sister-in-law' 30|
|'much'||*pḷlu- 'much' 31||*paljɜ 'thick, much' 32|
Notes to table
1 Latin -que, Greek te, Sanskrit -ca, etc.
2 Finnish -kä in ei ... eikä 'neither ... nor', Saami -ge, Mordvin (Moksha) -ka, Votyak -ke, Komi / Zyrian -kȯ, etc.
3 Latin ne-, Greek ne-, Sanskrit ná, Old High German and Old English ne ~ ni, etc.
4 Hungarian në, Cheremis / Mari nõ-, ni-, Votyak / Udmurt ni-, etc.
5 Greek me (enclitic).
6 Old Persian mana, Old Church Slavic mene, Welsh men, etc.
7 Finnish minä, Estonian mina, Nenets /mønʲə/. Uralic reconstruction *mun.
8 Latin tū, Greek sú (Attic), tu (Dorian), Lithuanian tù, Old English þu > archaic English thou, etc.
9 Greek sé, Sanskrit tvā (enclitic), Avestan θwā (enclitic), Old Church Slavic tebe, etc.
10 Sanskrit táva, Avestan tava, Proto-Celtic *towe (< PIE *tewe, with complex developments in the individual languages, Lewis and Pedersen 1989:193-217).
11 Finnish sinä (< *tinä), Saami ton, tú-, Mordvin ton, Votyak ton, Zyrian te, accusative tenõ, Hungarian të 'you' (singular), ti 'you' (plural), etc. Samoyed: Tavgi tannaŋ, Yeniseian Samoyed tod'i, Selkup tan, tat, Kamassian tan.
12 Gothic sa, Sanskrit sá, etc.
13 Finnish hän (< *sä-n), Saami son, Udmurt so. Samoyed: Nganasan syty.
14 Greek tó, Sanskrit tá-, Old Church Slavic to, etc.
15 Finnish tämä 'this' and tuo 'that (one)', Cheremis ti 'this', Mordvin te 'this', etc.; Udmurt tu 'that', Mordvin to 'that', etc. Cf. Hungarian tétova 'hesitant' (i.e. reluctant to choose between this and that).
16 *kʷi-: Hittite kuis (animate nominative singular), kuit (inanimate nominative-accusative singular), Latin quis, quid, Greek tís, tí, etc.
*kʷe-: Greek téo (Homeric), Avestan čahmāi (dative singular; ča < PIE *kʷe), etc.
*kʷo-: Latin quod, Old Latin quoius > Latin cuius (genitive singular), Old English hwæt > English what, etc.
17 E.g. Latin quidne.
18 Saami gi ~ gä 'who?, which?, what sort of?' and gutti 'who?', Mordvin ki 'who?', Cheremis and Mari ke, kö, kü 'who?', Hungarian ki 'who?', Finnish kuka 'who?', Komi / Zyrian kod 'which?', Ostyak koji 'who?', kŏti 'what?', etc.
19 Finnish ken ~ kene 'who?', Votyak kin 'who?', Udmurt kin 'who?', Komi / Zyrian kin 'who?'. Samoyed: Yurak Samoyed kin 'who?', Southern Nenets kin 'who?'.
20 Hittite tā-, Latin dō, Greek dídōmi, Sanskrit dā-, etc.
21 Finnish tuo 'bring', Estonian too- 'bring', Saami duokə- 'sell', Mordvin tuje- 'bring'. Samoyed: Tundra Yurak taš 'give, bring', Enets ta- 'bring', Tavgi tətud'a 'give, bring', etc.
22 Sanskrit ud-.
23 Hittite wātar (instrumental wēdanda), Umbrian utur (ablative une < *udne), Greek húdōr (genitive húdatos < *hudn̥tos), Sanskrit ud-án- (oblique cases only, nominative-accusative defective), Old Church Slavic voda, Gothic watō (n-stem, dative plural watnam), Old Norse vatn, Old English wæter > English water, etc.
This word belongs to the r / n stems, a small group of neuter nouns, from an archaic stratum of Indo-European, that alternate -er (or -or) in the nominative and accusative with -en in the other cases. Some languages have leveled the paradigm to one or the other, e.g. English to the r, Old Norse to the n form.
24 Finnish vesi / vete-, Estonian vesi, Mordvin wət, Udmurt vu, Komi / Zyrian va, Vogul wit, Hungarian víz. Samoyed: Forest Yurak wit, Selkup üt, Kamassian bü, etc.
25 Latin nōmen, Greek ónoma, Sanskrit nā́man-, Old English nama > English name, etc.
Indo-Europeanists are divided on whether to reconstruct this word as *nom(e)n- or as *H₁nom(e)n-, with a preceding "laryngeal". See Delamarre 2003:50 for a summary of views, with references. The o timbre of the root is assured by, among others, Greek ónoma and Latin nōmen (with secondary vowel lengthening). As roots with inherent o are uncommon in Indo-European, most roots having e as their vowel, the underlying root is probably *nem-. The -(e)n is an affixal particle. Whether the e placed in parentheses is inherently part of the word is disputed but probable.
26 Finnish nimi, Saami nama ~ namma, Mordvin lem, Cheremis lüm, Votyak and Zyrian ńim, Vogul näm, Ostyak nem, Hungarian név. Among the Samoyed languages: Yurak nim, Tavgi ńim, Yenisei Samoyed ńii’, Selkup nim, nem. Compare, in Yukaghir, Kolyma niu and Chuvan nyva.
27 Latin squalus (with s-mobile) 'large sea fish', Old Prussian kalis 'sheatfish', Old English hwæl 'whale' > English whale, etc.
28 Finnish kala, Estonian kala, Saami kuollē, Mordvin kal, Cheremis kol, Ostyak kul, Hungarian hal; Enets kare, Koibal kola, etc.
29 Latin glōs (genitive glōris), Greek gálōs, Old Church Slavic zŭlŭva, all meaning 'husband's sister'.
30 Finnish käly 'sister-in-law', Estonian kälī 'husband's brother, wife of husband's brother', Saami kāloji 'sister-in-law', Mordvin kel 'sister-in-law', etc.
31 Greek polú-, Sanskrit purú-, Avestan pouru-, Gothic filu, Old High German filu > German viel, all meaning 'much'.
The ḷ in Indo-European *pḷlu- represents a vocalic l, a sound found in English in for instance little, where it corresponds to the -le, and metal, where it corresponds to the -al. An earlier form of the Indo-European word was probably *pelu-.
32 Finnish paljon 'much', Cheremis pülä 'rather a lot', Vogul pāľ 'thick', Yurak palɁ 'thick'. Cp. Tundra Yukaghir pojuoŋ 'many'.
An asterisk (*) indicates reconstructed forms.
A tilde (~) means 'alternating with'.
There are three main objections against the Indo-Uralic hypothesis, one based on typology, and one based on the nature of the observed sound correspondences. The third objection is not against the relationship as such, but questions the validity of the Indo-Uralic node.
It is often objected against the Indo-Uralic hypothesis that the two families are typologically dissimilar. The reconstructed phonologies of Proto-Indo-European and Proto-Uralic are very different. The Indo-European languages are mostly fusional, while the Uralic languages have a more strongly agglutinating character (though not as extremely as, for instance, the Turkic languages). However, as much as typological similarity provides no evidence in favour of a relationship, typological dissimilarity provides no evidence against it, because languages can change their typology considerably within a few thousand years.
The sound correspondences found in lexical resemblances between the two families are not uniform, which points at different loanword layers. The largest number of lexical resemblances show the sound correspondences tabulated below (after Kallio 2001:223).
These correspondences look exactly like the sound substitutions one would expect from a layer of PIE loanwords in Proto-Uralic. When the PIE phoneme also exists in PU, it is preserved unchanged; otherwise, a nearest Uralic equivalent occurs, often with some vacillation between different possibilities. This is not what one would expect from true cognates inherited from a common ancestors spoken several thousand years earlier. Consider especially how the Late PIE vowels, with ablaut grades and even laryngeal colouring, are reflected faithfully in Uralic. Also, the reflexes of the PIE palatovelar may point at an early stage of satemization (the two most closely neighbouring IE branches to Proto-Uralic, Balto-Slavic and Indo-Iranian, are both satem branches).
This of course does not rule out the possibility of the existence of actual cognate pairs with different, less obvious sound correspondences (as to be expected from two languages that diverged several thousand years ago). A possible parallel is Armenian, which was initially classified as an Iranian language due to many lexical resemblances which later turned out to be loanwords; however, the language is nevertheless related as a separate branch of Indo-European, as can be shown by a different (smaller) set of lexical and morphological cognates with a different set of sound correspondences.
Validity of the Indo-Uralic node
It may be that Indo-European and Uralic are related to each other but do not form a valid node, i. e. there are further languages which are more closely related to one of them. There are not many candidates on the Indo-European side; Tyrsenian, Kartvelian and Afrasian have been proposed, but few scholars take any of these proposals seriously. More seriously considered are such proposals on the Uralic side. While not many scholars support the once very popular Ural-Altaic hypothesis any more, relationships to Yukaghir, Chukotko-Kamchatkan and Eskimo-Aleut are seriously discussed (see Uralo-Siberian).
Also, some of the morphological resemblances between Indo-European and Uralic are shared by further languages. Thus, it may be that Indo-European and Uralic are members of a larger family such as Nostratic, Eurasiatic or Mitian, but not the closest relatives of each other. On the other hand, such wider relationship hypotheses do not rule out the possibility of Indo-European and Uralic forming a valid node within such a unit. The impression of many languages being more closely related to Uralic than to Indo-European may simply result from Proto-Uralic being more conservative than Proto-Indo-European, making relationships to the former more easily discernible than to the latter.
- A roadmap to Indo-Uralic research, by Juho Pystynen
- Anderson, Nikolai. 1879. Studien zur Vergleichung der ugrofinnischen und indogermanischen Sprachen ('Studies on the Comparison of the Ugro-Finnic and Indo-Germanic Languages'). Dorpat: Heinrich Laakmann.
- Carpelan, Christian and Asko Parpola. 2001. "Proto-Indo-European, Proto-Uralic and Proto-Aryan." In The Earliest Contacts between Uralic and Indo-European: Linguistic and Archeological Considerations, edited by C. Carpelan, A. Parpola, and P. Koskikallio. Mémoires de la Société Finno-Ougrienne 242. Helsinki. ISBN 952-5150-59-3
- Collinder, Björn. 1934. Indo-uralisches Sprachgut ('The Indo-Uralic Linguistic Heritage'). Uppsala.
- Collinder, Björn. 1954. "Zur indo-uralischen Frage" ('On the Indo-Uralic question'), Språkvetenskapliga Sällskapets i Uppsala Förhandlingar Jan. 1952 – Dec. 1954, 79–91.
- Collinder, Björn. 1960. Comparative Grammar of the Uralic Languages. Stockholm: Almqvist & Viksell.
- Collinder, Björn. 1965. "Is the Uralic family isolated?" in An Introduction to the Uralic Languages, pages 30–34. Berkeley and Los Angeles: University of California Press.
- Čop, Bojan. 1970–1989. Indouralica.
- I.1974. Slovenska Akademija Znanosti in Umetnosti 30.1.
- II. 1972. Ural-Altaische Jährbucher 44:162–178.
- III. (Not published.)
- IV. 1973. Linguistica 13:116–190.
- V. 1978. Collectanea Indoeuropaea 1:145–196. Ljubljana.
- VI. (Not published.)
- VII. 1970. Zeitschrift für vergleichende Sprachforschung (KZ) 84:151–174.
- VIII. 1974. Acta linguistica Academiae scientarum hungaricae 24:87–116.
- IX. 1989. Linguistica 29:13–56.
- X. (Not published.)
- XI.( Not published.)
- XII. 1987. Linguistica 27:135–161.
- XIII. (Not published.)
- XIV. 1970. Orbis 19.2:282–323.
- XV. 1974. Zeitschrift für vergleichende Sprachforschung (KZ) 88:41–58.
- XVI. 1973. Orbis 22:5–42.
- XVII. (Not published.)
- XVIII. (Not published.)
- Joki, Aulis J. 1973. Uralier und Indogermanen. Mémoires de la Société Finno-Ougrienne 151.
- Reviews: Kiparsky, Valentin. 1975. Finnisch-Ugrische Forschungen 41.
- Kallio, Petri. 2001. Phonetic Uralisms in Indo-European? In The Earliest Contacts between Uralic and Indo-European: Linguistic and Archeological Considerations, edited by C. Carpelan, A. Parpola, and P. Koskikallio. Mémoires de la Société Finno-Ougrienne 242. Helsinki. ISBN 952-5150-59-3.
- Koivulehto, Jorma. 1999. Verba mutuata: quae vestigia antiquissimi cum Germanis aliisque Indo-Europaeis contactus in linguis Fennicis reliquerint. Mémoires de la Société Finno-Ougrienne 237.
- Lewis, Henry and Holger Pedersen. 1989. A Concise Comparative Celtic Grammar. Göttingen: Vandenhoeck and Ruprecht.
- Paasonen, Heikki. 1907. "Zur frage von der urverwandtschaft der finnisch-ugrischen und indoeuropäischen sprachen." Finnisch-Ugrische Forschungen 7.
- Pedersen, Holger. 1933. "Zur Frage nach der Urverwandschaft des Indoeuropäischen mit dem Ugrofinnischen" ('On the question concerning the original relationship of Indo-European with Ugrofinnic'). Mémoires de la Société Finno-Ougrienne 67:308–325.
- Rédei, Károly (editor). 1986a. Uralisches etymologisches Wörterbuch, 3 volumes, translated from Hungarian by Mária Káldor. Wiesbaden: Harrassowitz.
- Rédei, Károly. 1986b. "Zu den indogermanisch-uralischen Sprachkontakten." Sitzungberichte der Österreichischen Akademie der Wissenschaften, philosophisch-historische Klasse 468.