The Uralic tree

Proto-Uralic was spoken by people of the northern forest. Its reconstructed vocabulary has words for fish, nets, bows and arrows, skis and reindeer, and none for crops, ploughs or domestic cattle. Péter Hajdú and others pointed to inherited names of conifers, such as the Siberian fir and the Siberian stone pine, whose modern ranges lie mostly east of the Volga, as a clue to where the speakers lived.

For much of the 20th century the favoured homeland lay west of the Urals, around the Volga and Kama rivers, and the family was often dated to 4000 BCE or earlier. Two developments have moved the consensus later and further east. First, the oldest Indo-Iranian loanwords, shared by branches as far apart as Finnic and Ugric, suggest that the proto-language was still a single language in the late third or early second millennium BCE, when Indo-Iranian itself took shape. Petri Kallio, Jaakko Häkkinen and Ante Aikio have argued for a date around 2500–2000 BCE. Second, a group including Juha Janhunen and Riho Grünthal (2022) placed the homeland in southern Siberia and linked the spread westward to the Seima-Turbino phenomenon, a network of bronze-working and trade that crossed the forest zone around 2000 BCE.

Ancient DNA fits a movement from the east. A distinctive Siberian ancestry component, most concentrated today in the Nganasan, appears in northern Europe by about 1500 BCE, in remains from the Kola Peninsula studied by Thiseas Lamnidis and colleagues (2018). Genes and languages need not travel together, though. Hungarians today carry little of this ancestry, and many Finns and Saami descend largely from earlier Europeans who shifted language. Whether the homeland lay east or west of the mountains is still argued, but few specialists now accept a date much before 3000 BCE.

A reader meeting Finnish, Estonian and Hungarian side by side notices first that words grow at the end. Uralic languages are agglutinative: grammatical information is added as a chain of suffixes to a stem that changes little. Hungarian ház 'house' gives házam 'my house', házamban 'in my house' and házaimban 'in my houses'. Finnish talo gives talossa 'in the house' and talossani 'in my house'. Most of the languages have many cases, usually including a set of local cases for 'in', 'into', 'out of', 'on', 'onto' and 'off': Finnish has about fifteen, Hungarian around eighteen by most counts.

Several features recur across the family. There is no grammatical gender, and a single pronoun covers 'he' and 'she' (Finnish hän, Hungarian ő). Stress falls on the first syllable in most of the languages. Many have vowel harmony, in which suffix vowels match the stem: Finnish talossa 'in the house' but kylässä 'in the village', Hungarian házban but kertben 'in the garden'. After a numeral the noun often stays singular, as in Hungarian két ház 'two houses'. Finnic, Saami, Mari and Permic express negation with an auxiliary verb that takes person endings, while the main verb stays uninflected: Finnish en tiedä 'I don't know', et tiedä 'you don't know'.

The original word order was probably subject–object–verb, with postpositions rather than prepositions, and this order survives in Samoyedic and Ob-Ugric. The western languages, after centuries of contact with Germanic, Baltic and Slavic, have moved towards subject–verb–object. The three national languages, and Saami, are written in the Latin alphabet; the languages of Russia use Cyrillic, with extra letters for sounds Russian lacks.

That the Uralic languages are related is not in doubt. How the family divided is. The classic tree, set out by Otto Donner in 1879 and refined over the following century, splits off Samoyedic first and then breaks the rest, Finno-Ugric, into ever smaller pairs: Ugric against Finno-Permic, Permic against Finno-Volgaic, and so on. Tapani Salminen (2002) argued that most of these intermediate nodes rest on too few shared innovations, and that the family is better drawn as a 'rake' of about nine branches from one root. Others have proposed different groupings: Jaakko Häkkinen, for example, has suggested an 'East Uralic' group joining Ugric and Samoyedic. Much of the difficulty is that neighbouring branches kept borrowing from each other after they separated, so shared features may reflect contact rather than common descent.

Dating is linked to the same problem. The later chronology now favoured by many Finnish scholars leaves little time for a long Finno-Ugric period, while the traditional chronology needs one. The discovery of old loanword layers keeps adjusting the picture, and so does the question of what unknown languages Saami and Finnic replaced as they spread west.

Beyond the family, no deeper relationship has been demonstrated. The 19th-century 'Ural-Altaic' grouping with Turkic and Mongolic, based on typological similarity, has been abandoned. A link with Yukaghir in north-eastern Siberia has had serious supporters, among them Irina Nikolaeva, but many specialists interpret the shared words as ancient loans. Proposals for an Indo-Uralic family, or for membership in a larger Nostratic or Eurasiatic grouping, rest on pronouns such as *me 'I' and *te 'you' and a handful of other forms. Most Uralicists regard them as unproven.