A genetic relationship between Indo-European and Uralic was first proposed by the Danish linguist Vilhelm Thomsen in 1869 but was received with little enthusiasm (Pedersen 1931:336). Since then, the predominant opinion in the linguistic community has remained that the evidence for such a relationship is insufficient. However, a minority of eminent linguists has always taken the contrary view (e.g. Henry Sweet, Holger Pedersen, Björn Collinder, and Jochem Schindler), making it hard to dismiss the relationship out of hand.
There are two distinct questions here:
(1) Are Indo-European and Uralic genetically related?
(2) If so, do Indo-European and Uralic constitute a valid genetic node? The Eurasiatic and Nostratic hypotheses both consider Indo-European and Uralic (or Uralic-Yukaghir) to be genetically related. However, the Indo-Uralic hypothesis in the strict sense is something different from this: it holds that Indo-European and Uralic have an especially close genetic relationship; it does not necessarily include assertions that Indo-European and Uralic are related to any other language families.
At the same time, most of the prominent supporters of a relationship between Indo-European and Uralic have also supported their relationship to additional language families, leading some to regard Indo-Uralic as a subset of the larger Nostratic hypothesis.
The following article focuses on question (1), genetic relationship, and only treats incidentally of question (2), possible relation to other language families.
The Dutch linguist Frederik Kortlandt supports a model of Indo-Uralic in which the original Indo-Uralic speakers lived north of the Caspian Sea, and the Proto-Indo-European speakers began as a group that branched off westward from there to come into geographic proximity with the Northwest Caucasian languages, absorbing a Northwest Caucasian lexical blending before moving farther westward to a region north of the Black Sea where their language settled into canonical Proto-Indo-European. Allan Bomhard suggests a similar schema in Indo-European and the Nostratic Hypothesis (1996). Alternatively, the common protolanguage may have been located north of the Black Sea, with Proto-Uralic moving northwards with the climatic improvement of post-glacial times.
The most common arguments in favour of a relationship between Indo-European and Uralic are based on seemingly common elements of morphology, such as the pronominal roots (*m- for first person; *t- for second person; *i- for third person), case markings (accusative *-m; ablative/partitive *-ta), interrogative/relative pronouns (*kw- 'who?, which?'; *y- 'who, which' to signal relative clauses) and a common SOV word order. Other, less obvious correspondences are suggested, such as the Indo-European plural marker *-es (or *-s in the accusative plural *-m̥-s) and its Uralic counterpart *-t. This same word-final assibilation of *-t to *-s may also be present in Indo-European second-person singular *-s in comparison with Uralic second-person singular *-t. Compare, within Indo-European itself, *-s second-person singular injunctive, *-si second-person singular present indicative, *-tHa second-person singular perfect, *-te second-person plural present indicative, *tu 'you' (singular) nominative, *tei 'to you' (singular) enclitic pronoun. These forms suggest that the underlying second-person marker in Indo-European may be *t and that the *u found in forms such as *tu was originally an affixal particle. Indeed, the Finnish verb conjugation system does appear suspiciously Indo-European-like, and similarities to the verb conjugation systems of Latin, Russian, the Baltic languages, and so forth have been noted for years.
A second type of evidence advanced in favor of an Indo-Uralic family is lexical. Numerous words in Indo-European and Uralic resemble each other. The problem is to weed out words due to borrowing. Uralic languages have been in contact with a succession of Indo-European languages for millennia. As a result, many words have been borrowed between them, most often from Indo-European languages into Uralic ones.
An example of a Uralic word that cannot be original is Finno-Ugric *śata 'hundred'. The Proto-Indo-European form of this word was *km̥tóm (compare Latin centum), which became *ćatám in Indo-Iranian (compare Sanskrit śatám, Avestan satəm). This is evidence that the word was borrowed into Finno-Ugric from Indo-Iranian or Indo-Aryan. This borrowing may have occurred in the region north of the Pontic-Caspian steppes around 2100-1800 BC, the approximate floruit of Indo-Iranian (Anthony 2007:371-411). It provides linguistic evidence for the geographical location of these languages around that time, agreeing with archeological evidence that Indo-European speakers were present in the Pontic-Caspian steppes by around 4500 BCE (the Kurgan hypothesis) and that Uralic speakers may have been established in the Pit-Comb Ware culture to their north in the fifth millennium BCE as well (Carpelan - Parpola 2001:79).
Another ancient borrowing is Finno-Mordvinic *porćas ‘piglet’. This word corresponds closely in form to the Proto-Indo-European word reconstructed as *porḱos, attested by such forms as Latin porcus 'hog', Anglo-Saxon fearh (> English farrow 'young pig'), Lithuanian par̃šas ’piglet, castrated boar’, Kurdish purs 'pig', and Saka pāsa (< *pārsa) 'pig'. In the Indo-European word, *-os (> Finno-Mordvinic *-as) is a masculine nominative singular ending, but it is quite meaningless in Uralic languages. This shows that the whole word was borrowed as a unit and is not part of the original Uralic vocabulary.
A later borrowing, from well after the Proto-Indo-European period, is Finnish kuningas 'king'. This word corresponds almost exactly in form to the reconstructed Proto-Germanic word *kuningaz (> Old English kyning > English king). Again, we have fascinating confirmation for a reconstructed proto-form. Here, both the stem suffix -ing and the masculine nominative singular ending *-az (< PIE *-os) are meaningless in Finnish, making it clear that the whole word was borrowed as a unit, like *porćas.
Such words as those for 'hundred', 'king', and 'pig' have something in common: they represent "cultural vocabulary" as opposed to "basic vocabulary". They are likely to have been acquired along with a more complex number system, the domestic pig, and the institution of kingship from the more advanced Indo-Europeans to the south. Similarly, the Indo-Europeans themselves had acquired such words and cultural items from peoples to their south or west, including possibly their words for 'ox', *gwou- (compare English cow) and 'grain', *bhar- (compare English barley). In contrast, basic vocabulary – words such as 'me', 'hand', 'water', and 'be' – is much less readily borrowed between languages. If Indo-European and Uralic are genetically related, they should show agreements in basic vocabulary, with more agreements if they are closely related, fewer if they are less closely related.
Advocates of a genetic relation between Indo-European and Uralic maintain that the borrowings can be filtered out by application of phonological and morphological analysis and that a core of vocabulary common to Indo-European and Uralic remains. As examples they advance such comparisons as Proto-Uralic *weti- (or *wete-) : Proto-Indo-European *wot’er- (or *wodr̥), oblique stem *wet’en-, both meaning 'water', and Proto-Uralic *nimi- (or *nime-) : Proto-Indo-European *nomen- (or *H₁nōmn̥), both meaning 'name'. In contrast to *śata and *kuningas, the phonology of these words shows no sound changes from Indo-European daughter languages such as Indo-Iranian. In contrast to kuningas and *porćas, they show no morphological affixes from Indo-European that are absent in Uralic. According to advocates of the Indo-Uralic hypothesis, the resulting core of common vocabulary can only be explained by the hypothesis of common origin.
It has been countered that nothing prevents this common vocabulary from having been borrowed from Proto-Indo-European into Proto-Uralic.
For the old loans, as well as uncontroversial ones from Proto-Baltic and Proto-Germanic, it is more the rule than the exception that only the stem is borrowed, without any case-endings. Proto-Uralic *nimi- has been explained according to sound laws governing substitutions in borrowings (Koivulehto 1999), on the assumption that the original was a zero-grade oblique stem PIE *(H)nmen- as attested in later Balto-Slavic *inmen- and Proto-Celtic *anmen-. Proto-Uralic *weti- could be a loan from the PIE oblique e-grade form for 'water'. Proto-Uralic *toHį- 'give' and PFU *wetä- 'lead' also make perfect phonologic sense as borrowings.
It is also objected that some or all of the common vocabulary items claimed are false cognates—words whose resemblance is merely coincidental, like English bad and Persian bad.
The items concerned represent basic vocabulary – unlikely to have been borrowed – or items appropriate to a Mesolithic level of culture and therefore plausible as shared terms.
With regard to the postulated equivalence of Uralic -i and Indo-European -en, we need a little more explanation on how "sound laws", which are regular by definition, can be equivalent to "substitutions in borrowings", which are by definition analogical and therefore not regular, phonologically speaking. Koivulehto’s position may well be possible; the issue is whether it is the most compelling explanation of the data.
The points raised concerning the words for 'name’, 'water', and 'give' require a glance at the possible relations of Indo-European and Uralic with other language families, in particular the languages hypothetically grouped as Uralo-Siberian by Fortescue, Eurasiatic by Greenberg, and Nostratic by Holger Pedersen and various successors of his. While it is perfectly true that the Uralic words for these things could be derived from the Indo-European ones (or vice versa), the Uralic words have apparent equivalents among other languages variously identified as "Uralo-Siberian" or "Eurasiatic". For example, according to Fortescue (1998), Proto-Finno-Ugric *toɣe- 'bring, take, give' is cognate with Proto-Chukotko-Kamchatkan *teɣiŋrə- 'pull out' and Proto-Eskimo *teɣu- 'take'. He reconstructs these forms to a Proto-Uralo-Siberian *toɣə- 'take'.
If the Uralic word is borrowed from Indo-European, why is it found in nearly identical form right across Siberia? Possible cognates are also found for the words for 'name' in Chukchi nənnə 'name' and Old Japanese na 'name' and for 'water' in Evenki udun 'rain', Even udən 'rain', and Ainu owata 'water' (Greenberg 2002). Thus, alongside the hypothesis of borrowing from Indo-European, another possibility is that Indo-European and Uralic themselves belong to a larger grouping.
Finally, the claim that all such forms are "false cognates" is not widely accepted. The disagreements between e.g. Koivulehto and Kortlandt do not turn on whether the forms under discussion are true cognates, which is generally accepted, but on whether they result from borrowing or genetic inheritance. This is thus the key point at issue.
|'I, me'|| *me 'me' (accusative),|
*mene 'my' (genitive)
|*mVnV 'I' 1|
|'you' (singular)|| *tu (nominative),|
*tewe 'your' (genitive)
|demonstrative pronoun||*so 'this, he/she' (animate nominative)||*ša (third person singular)|
(animate interrogative pronoun)
| *kʷi- 'who?, what?' |
*kʷo- 'who?, what?'
| *ken 'who?'|
| 'who, which'|
|*yo-||*-ja (agent noun)|
|nominative/accusative plural|| *-es (nominative plural),|
*-m̥-s (accusative plural)
|oblique plural|| *-i (pronominal plural)|
(as in *we-i- 'we', *to-i- 'those')
|first person singular||*-m (first person singular active)||*-m|
|second person singular||*-s (second person singular active)||*-t|
|stative|| *-s- (aorist),|
*-es- (stative substantive),
*-t (stative substantive)
|negative particle||*ne 2||*ne 3|
|'to give'||*deH₃- 4||*toHi-|
| 'to moisten', |
| *wed- 'to wet',|
*wódr̥ 'water' 5
| 'to assign', |
| *nem- 'to assign, to allot',|
*H₁nōmn̥ 'name' 6
2 Latin ne-, Greek ne-, Sanskrit ná, Old High German and Old English ne ~ ni, etc.
3 Hungarian në, Cheremis / Mari nõ-, ni-, Votyak / Udmurt ni-, etc.
4 Latin dō, Greek dídōmi, Sanskrit dā-, etc.
5 Hittite wātar, instrumental wēdanda; English water.
6 Latin nōmen, Greek ónoma, Sanskrit nāman-, Old English nama > English name, etc.
An asterisk (*) indicates reconstructed forms.