1.3. Early Uralic

1.3.1. Early Uralic evolution

Common traits of Uralic languages, which can be traced back to the parent language[ii], include the following (Janhunen 1982; Comrie 1988; Sammallahti 1988; Raun 1988):

·       General SOV order. Noun phrase basic order is attribute (adjctive, genitive, numeral) before the head noun. Postpositions instead of prepositions.

o   Subjects of finite clauses in the nominative, of nonfinite verb forms in the genitive (or appear as possessive suffixes, for pronouns).

o   Noun phrase structure: Attribute precedes the head noun, with no agreement between attributive adjective and head noun.

o   Direct object in the accusative.

o   Possession may be expressed by two bare nouns standing adjacent to one another (attributive use of a noun), but a specialised structure with possessor in the genitive + head noun unmarked is also common to this early stage.

o   No distinction between nominal and verbal stems.

·       Number: Singular, plural, and probably dual.

o   Dual suffix *-ka- + -n ~ *-kä- + n/ń⁠.

o   Plural marker originally probably *-t, ancient ones including *-t, *-i, and *-k.

·       Three grammatical cases (nominative, accusative, genitive) and three local cases (locative, allative, ablative):

o   Subjects of finite clauses in the nominative, of nonfinite verb forms in the genitive (or appear as possessive suffixes, for pronouns).

o   Direct object in the accusative (if no distinct form, the nominative form is used). General accusative ending sg. *-m.

o   Oblique plural case suffix *-i, possibly from a possessive adjective.

o   Subordinative suffix (eventually functioning as genitive / prenominaliser with nouns, or adverb-formant with verbs) in *-n.

o   Ablative or separative as *-tV (*-tə, *-tu), for example found with the verb ‘to fear’ in all Uralic languages, and in the adverb ‘from under’ *al-ta.

o   Locative in *-na/-nä, with the original local meaning appearing in adverbs and postpositions.

o   Lative suffix *-k(V) ‘moving toward, moving along; becoming somebody or something’.

o   Dative-lative suffix *-n or -ń⁠.

·       Possession:

o   Possessive suffix 1p *-mV/me, 2p. *-tV/te, 3p. *-sV/se. In plural a *-t or *-k is added, in dual cases an *-n.

o   Oblique cases *-n before a possessive suffix in singular or plural, probably from the genitive ending.

·       Diminutive *-mpV.

·       Ancient ordinal *-ntV.

·       Pronouns *mi/me-nä ‘I’, *me ‘we’, *ti/te-nä ‘thou’, *ti/te ‘you’, *ći/ će‘this’, *e ‘this’, * ‘this’, *to ‘that’, *u/o ‘that’, *ke/ki ‘who’, *ku/ko ‘who’, *-me ‘what’. Functions of third person personal pronouns are covered by demonstrative pronouns.

·       Originally, neither imperative nor indicative were marked.

·       Verb agreement in person and number with the subject, but less usual in the third person.

·       Original verb form probably a kind of aorist, neutral as regards time. Present marker *-k would develop initially as an emphatic.

·       One tense distinction: past and non-past tense (subsuming present future).

·       Verbal personal suffixes 1p. *-m, 2p. *-t (or *-n?), 3p. *-s, plus number suffixes.

·       Deverbal suffixes in nouns: *-kV, *-V (to designate the actor), *-mV (different meanings), *-nV (infinitives and participles), *-tV/ttV, and *-pV (predominantly participial), which can be traced back to an Indo-Uralic participial suffix, cf. Pre-PIE *-bo (Hyllested 2009).

·       Deverbal suffixes in verbs: *-l- (frequentative or momentary), *-tV (frequentative and causative), *-ttV (momentary, causative), *-ktV (causative), *-ntV (frequentative or causative), and reflexive *--.

·       Denominal verbs: *-j-, *-lV-, *-mV-, *-nV-, *-tV-, and reflexive *--.

·       The verb ‘to have’ was expressed with the owner in the locative (possibly also genitive), and the thing owned (grammatical subject) in the nominative, with the verb ‘to be’ acting as the predicate.

·       Phonetically, Uralic vowels were divided in two exclusive harmonic categories. Front and back vowels could not occur together in a (non-compound) word (Sammallahti 1988; Janhunen 1982):





stressed positions


















unstressed positions





















o   *a is typologically more likely than the traditionally reconstructed *a, hence the more modern notation will be used here.

o   The traditionally reconstructed unrounded *ɨ (also *ï) will also be replaced by the proposed mid vowel *ë /ɤ/, so the traditional reconstruction of the Uralisches etymologisches Wörterbuch (UEW) by Rédei (1988) can be used for consistency purposes. The true original value of certain vowels is disputed (Abondolo 1998), and different probably between EPU and LPU.

·       About consonants, the following is the common repertoire:






















































o   The phonetic nature of *ɣ is unclear. It could have been two different sounds, a laryngeal (*h or *ʔ) and a velar (*g or *ǥ); or even a mere syllable boundary between two successive heterosyllabic vowels.

o   Phonemes *d and *ð were probably spirants, and *ć was retroflex (cacuminal).

o   Consonants could be combined to form geminates, at least *pp, obstruent + obstruent, sonorant + obstruent, and sonorant + sonorant.


1.3.2. Early Uralic–Indo-Anatolian contacts

If one assumes no genetic relationship between Proto-Indo-Anatolian and Early Proto-Uralic, then one should interpret the above Indo-Uralic roots and words as wanderwords, or loanwords from one language into the other, at a very early stage of both, probably during the Neolithic and Early Eneolithic in the Pontic-Caspian steppes. In this case, it is very difficult to say in each case which one was the donor language, if only because of the scarce material available on Uralic comparative grammar.

If one assumes a genetic relationship between PIA and PU, and thus an Indo-Uralic trunk, it would be very difficult to differentiate an early loanword from a common root, since the phonological rules involved in borrowing would have been quite similar to those described here for derivation.

Suggested loans include:

·       PIA *ɣwneɣwmn- ‘name’ PA *ʔlóʔmn, cf. Hitt lā́mn (Kloekhorst 2008), cannot be the origin of PU *nime ~ PYuk *nime, unless a late, non-laryngeal PIE stage is proposed for its adoption. Most likely, Cavoto’s interpretation of an original PIA root *(H)nem- +*-men accounts for PIU *ɣnem yielding PU *nem- due to the phonotactically illicit sonorant + obstruent + sonorant in PU (Hyllested 2009). It is, therefore, more likely an ancient, shared IU noun.

·       Similarly, PIU *medu   PU *mete ‘honey’ ~ PIA *medu (cf. hitt. mitgaimi- ‘sweetened (bread)’) ‘mead, sweet; honey’ seems also an ancient root, and it could have originally been a borrowing from Semitic into Indo-Uralic, cf. Semitic *mVt -, ‘sweet’, NE Caucasian miʒʒV ‘sweet’ (Bjørn 2017). Since beekeeping may have spread (replacing the previous honey gathering techniques) with the domestication of Apis mellifera ca. 9000 BC in the Near East (Bloch et al. 2010), the arrival of a foreign word could have reached Indo-Uralic with Neolithisation via the North Pontic to the west, or via the Caucasus to the south.

·       PIU *ede ‘water’   PU *ete ~ PIA *ed-, is in e-grade not only in Hitt. obl. eten-, but also in other Late PIE languages (Kloekhorst 2008), and could thus suggest a very old loanword, but it is unlikely that this kind of word would be easily borrowed (Kortlandt 2010), and the vocalic alternation suggests a more complex nom. *uód-r, obl. *ud-én- paradigm (Kloekhorst 2019 fthc.).


1.3.3. Schleicher’s fable in Early Proto-Uralic

For a dynamic Proto-Uralic phonological evolution, an intermediate stage has been selected before the full development of de-voicing of voiced stops and affricates in PU (Hyllested 2009), the palatalisation before front vowels, and the harmonic categories.



geu̯e – luɣɨt

geu̯e e aptɨ ne

luɣɨi̯ gokɨ;

kje gauga u̯eɣem u̯eɣetä,

kje enä kanem,

kje u̯irkäm suɣɨm kanɨta.

keu̯e luɣɨi̯ u̯akɨ:

kjedä-mi tjumtɨ

ikräm luɣɨi̯ kajatam äntetä.”

luɣɨt u̯akɨt: “keulɨ, geu̯e!

kjedä-me tjumtɨ u̯äntetä:

ikrä, u̯äke, geu̯ei̯ aptɨn

su eci esem teke,

geu̯ei̯ aptɨ ne .”

kje keulɨtɨ geu̯e kanjgak bugä.


Early Proto-Uralic

keu̯i – luɣët

keu̯i e aptën e-u̯olë

luɣëi̯ kokë;

će kauka u̯iɣim u̯iɣitä,

će enä kanëm,

će u̯irkäm suɣëm kanëta.

keu̯i luɣëi̯ u̯akë:

ćiðä-mä tumtë

irkäm luɣëi̯ kai̯atam äntitä.”

luɣët u̯akët: “keulë, keu̯i!

ćiðä-mät tumtë u̯äntitä:

irkä, u̯äki, keu̯ii̯ aptën

su eči esim teki,

keu̯ii̯ aptën e-u̯olë.”

će keulëtë keu̯i kaŋkak pukä.

The following changes are tentatively proposed, where an Early Proto-Uralic community is probably closer chronologically to CIE than to PIA (i.e. after the separation of Anatolian):

·       Word-initial PIU *ɣ → PPU **g → EPU *k.

·       Word-initial PIU *χ → PPU **k→ EPU *k.

·       PIU *ɠ → PPU **g→ EPU *k (except for *χaɠa- →**kaja-).

·       PIU *d → PPU **d→ EPU *ð(j).

·       PIU *k(e/i/a) → PPU **kj→ EPU *ć.

·       PIU *g(e/i/a) → PPU **gj→ EPU *ś.

·       PIU *b → PPU **b → EPU *p.

·       In vocalic evolution, it is posited a system where harmonic categories are already being developed: PIU *e → EPU **e → LPU *i and PIU *i → EPU **ɨ → LPU *ë.