Consonants

The following consonant sounds are generally posited for the early stages of Uralic: *p, *t, *č (pronounced as the ch in chip), *k, *s, *š (pronounced as the sh in ship), *ð (pronounced as the th in then), *l, *r, *m, *n, *ŋ (pronounced ng as in sing), *j (pronounced as the y in yet), *v, and the palatalized alveolar sounds *t′, *ś, *ð′, *l′, *ń, plus a few others less well established. Modern Finnish has a much smaller inventory of consonants, having lost the palatalized alveolar sounds and *č, *š, *ð, and *ŋ. Hungarian, on the other hand, has a larger number of consonants by virtue of a newly introduced distinction between sounds made with and without vibration of the vocal cords (voicing), such as voiceless p, t, s as opposed to voiced b, d, z; e.g., dél ‘noon’: tél ‘winter.’ Other Uralic languages, such as Komi, have also acquired a voicing contrast (e.g., doj ‘pain’: toj ‘louse’), but the geographic distribution of those languages in which the voicing contrast plays an active role leaves little doubt that it originated under the influence of Indo-European and Turkic languages.

Vowels

Essentially nothing is known of the Proto-Uralic vowels, and there is little agreement about the nature of the Proto-Finno-Ugric vowel system. It is clear, however, that, in contrast to a relatively limited number of consonants, Finno-Ugric must have had a fairly large number of vowels (nine to 11 are usually posited). One hypothesis is that the original vowel system was essentially like that of Finnish, which has eight vowel sounds: i, ü, u, e, ö, o, ä, a (ü—spelled y in the standard orthography—and ö are front rounded vowels, as in German; ä is a low front vowel, as a in cat). Hungarian has a similar system, although not all dialects have a separate ä sound, which is not distinguished from e in the orthography. A second approach posits a Proto-Uralic vowel structure closely resembling that of Khanty, with seven full vowels and three reduced vowels.

The early Finno-Ugric system of vowels most likely possessed quantitative vowel contrasts (long versus short, or full versus reduced). Such contrasts are present in Baltic-Finnic, Sami, and Ugric and within Samoyedic—e.g., Finnish tulen ‘of fire’ and tuulen ‘of wind,’ tuleen ‘into fire,’ and tuuleen ‘into wind’; Hungarian szel ‘slice’ and szél ‘wind,’ szelet ‘wind’ (accusative case), and szelét ‘its wind’ (accusative). The possibility of influence by neighbouring languages cannot be ruled out in the case of vowel length, because western Finno-Ugric languages have been in close contact with Slavic and Germanic languages with similar vowel contrasts, and the eastern languages form an areal group among themselves. The remaining languages lack vowel quantity and are in intimate contact with Russian, which has lost the original contrastive vowel quantity of Indo-European. The Izhma dialect of Komi, adjacent to Nenets, has superficial contrasts such as pi ‘son’ versus pī ‘cloud,’ but this vowel length is the result of a change of an l at the end of the syllable to a vowel.

Stress

In numerous Uralic languages—including Finnish, Estonian, Hungarian, and Komi—stress is automatically on the first syllable of the word; it is likely that Proto-Uralic also had word-initial stress. Closely related to this initial stress is the apparent severe limitation on early Finno-Ugric noninitial vowels; the full range of contrasts was permitted only in the first syllable. In certain languages, such as Eastern Mari and the Yazva Komi dialect, stress is not bound to a given syllable, and determining the place of stress requires information concerning vowel quality as well—e.g., Yazva śibdinə ‘to bind,’ líććina ‘to descend,’ l′iśn̥na ‘wood’ (the i’s, which receive stress, were long at an earlier period; ś, ć, l′ are palatalized consonants). Stress at the end of a word is also found—e.g., in Eastern Mari and Udmurt. Nganasan has a mora-counting stress, falling on the third unit of vowel length from the end of the word (where short vowels count as one unit, long vowels as two).

Vowel harmony

Vowel harmony is among the more familiar traits of the modern Uralic languages. Although most Uralic scholars trace this feature back to Proto-Uralic, there is good reason to question this view. Vowel harmony is said to exist when certain vowels cannot occur with other specific vowels within some wider domain, generally within a word. For example, of the eight vowels of Finnish, within a simple word, any member of the set ü, ö, ä prohibits the use of any member of the set u, o, a, but i and e may occur with either set. That is, within a word, vowels that are either rounded (such as ü, ö, u, o) or low (such as ä, a) must agree with each other in frontness or backness. (The distinction is marked phonetically by putting two dots over the front vowels.) The unrounded front vowels, i and e, may occur with any of the other vowels. Thus, from talo ‘house’ one may form talossa ‘in (the) house,’ but for kynä ‘pen’ the comparable form is kynässä ‘in (the) pen’; similarly, talossansa ‘in his house’ contrasts with kynässänsä ‘in his pen’ and talossansako ‘in his house?’ with kynässänsäkö ‘in his pen?’, whereas taloni ‘my house’ and kynäni ‘my pen’ have the same ending because i can occur with either of the two sets of vowel classes. Hungarian has essentially the same system, differing only in certain minor details (short e is the front vowel counterpart of a)—e.g., asztal ‘table,’ asztalok ‘tables,’ asztalokban ‘in the tables,’ but föld ‘land,’ földök ‘lands,’ földökben ‘in the lands.’ Similar though less general front-back vowel-harmony systems are found in given dialects of Mordvin, Mari, Mansi, Khanty, and Kamas.

Frequently confused with the true harmony situations above are partial and total assimilations of vowels in adjacent syllables. These assimilations illustrate a universal tendency of vowel interaction and are of relatively recent origin; they are best held apart from the question of vowel harmony. Examples of vowel assimilations abound. In Finnish an unstressed e in the illative case (“place into”) is totally assimilated to a preceding vowel, even across an intervening h: talo + hen becomes taloon ‘into the house,’ talo + i + hen yields taloihin ‘into the houses,’ työ + hen becomes työhön ‘into the work.’ The Hungarian allative case (“place to or toward which”) shows an assimilation of the phonetic feature of lip rounding with front vowels in addition to the standard vowel harmony; thus, ház-hoz ‘to the house,’ kéz-hez ‘to the hand,’ betű-höz ‘to the letter.’ Apart from such nonharmony alternations, no support for rounding harmony is found in Uralic.

Considered from an areal viewpoint, two aspects of Uralic vowel harmony must be considered. First, those languages that show productive or active vowel harmony, with the exception of Baltic-Finnic, have had recent Turkic neighbours whose languages exhibited vowel harmony. For languages such as Mansi and Khanty, dialects with vowel harmony are located close to Tatar groups. Second, the original homeland of Uralic lies in the centre of an enormous hypothetical areal grouping, labeled by the Russian-American linguist Roman Jakobson as the “Eurasian language union.” The languages of this “union” are said to be characterized by two features: (1) the absence of a tonal accent (changes in pitch that change meaning, as is found in Chinese, Swedish, or Serbian) and (2) the contrast of plain and palatalized consonants (as in Russian). The distinction between palatalized and nonpalatalized consonants has the same acoustic basis as the contrast of front and back vowels (i.e., palatalized consonants and front vowels share a heightened tonal quality). Indeed, in Erzya Mordvin, vowel harmony and palatalization appear to be conditioned by essentially the same rules. Instead of seeking a genetic explanation of vowel harmony in Uralic, a somewhat more recent areal origin—in part under Turkic influence—must be considered. Of significance is the further consideration that, among the northwestern languages, far from Turkic influence, it is precisely Sami and the Baltic-Finnic Estonian and Livonian that do not have vowel harmony and that have developed special syllable-accent systems (thus, they lack both traits of the Eurasian union).

Consonant gradation

The alternation of consonants known as consonant gradation (or lenition) is sometimes thought to be of Uralic origin. In Baltic-Finnic, excluding Veps and Livonian, earlier intervocalic single stops were typically replaced by voiced and fricative consonantal variants, and geminate (double) stops were shortened to single stops just in case the preceding vowel was stressed and the following vowel was in a closed syllable; that is, *p alternated with *v and *b; *t with *ð and *d; *k with *ɤ and *g; *pp with *p; and so on. Finnish thus shows pairs such as mato ‘worm’ and madon ‘of the worm,’ matto ‘rug’ and maton ‘of the rug,’ poika ‘boy’ and pojan ‘of the boy,’ lintu ‘bird’ and linnun ‘of the bird,’ selkä ‘back’ and selän ‘of the back.’ Estonian shows the same type of alternation, with considerable difference in detail—e.g., sada ‘hundred’ and saja ‘of a hundred,’ madu ‘snake’ and mao ‘of the snake,’ lind ‘bird’ and linnu ‘of the bird,’ and selg ‘back’ and selja ‘of the back.’ Most of the Sami languages exhibit similar alternations, but the process applies to all consonants and, moreover, works in reverse: single consonants are doubled in open syllables—e.g., čuotte ‘hundred’ and čuoðe ‘of a hundred,’ borra ‘eats’ and borâm ‘I eat.’ The change of t to ð, however, is not a part of Sami gradation but rather a general process that voices and weakens all single stops between voiced sounds (in this case, vowels).

Despite their essential differences, the Baltic-Finnic and Sami gradations appear to be areally related. The Baltic-Finnic type, which represents a more plausible phonetic change, indicates that early Sami may have acquired its gradation under Baltic-Finnic influence. The evidence within Baltic-Finnic points to a relatively late, post-Proto-Baltic-Finnic origin. The existence of analogous consonant weakening in various Samoyedic languages (Nganasan, Selkup) is the result of independent innovation.

Syllable-accent structures

Closely related to the gradation phenomena is the development of syllable-accent structures in Estonian, Livonian, and Sami. Estonian is known for its unique quantity alternations of three contrastive vowel and consonant lengths—thus, vara ‘early’ versus vaara ‘of the hillock’ (aa = long ā) versus vaara ‘hillock (partitive)’ (here aa = extra-long â); lina ‘linen’ versus linna ‘of the city’ (nn is pronounced as two short n’s) versus linna ‘into the city’ (here nn is pronounced as long n̄ plus short n; the contrast with the previous nn is not shown in the standard orthography). The extra-quantity contrast is in fact found with all stressed syllable types containing at least one vowel or consonant following its first vowel; thus, taevas ‘sky’ (with short e) versus taevas ‘in the sky’ (with long ē); osta ‘buy!’ (with short s) versus osta ‘to buy’ (with long s̄), whereas a two-syllable form such as osa ‘part’ (o/sa) with only a single vowel in the first syllable is incapable of such a quantity contrast. A multitude of analyses of Estonian quantity have been proposed, although not all have recognized the phenomenon as a function of whole syllables bound to stress—in other words, that it is an accent phenomenon. One orthographic dictionary (by E. Muuk), for example, utilizes this principle, placing a grave accent mark before syllables with extra quantity. Otherwise, Estonian orthography marks the three degrees of duration only for stops: b, d, g indicate single short (voiceless lenis) stops (tuba ‘room’); p, t, k are plain geminates, or double consonants (tupe ‘of the sheath’); and pp, tt, kk mark extra-long geminates (tuppa ‘into the room,’ tuppe ‘into the sheath’). Because the extra quantity is in part tied to an original open next syllable, it frequently operates together with gradation alternations—e.g., linnu ‘of the bird’ versus lindu ‘bird (partitive),’ with extra quantity.

The syllable quantity accent in Sami superficially resembles that in Estonian and, like the former, occurs only under stress and is in part conditioned by the openness of the next syllable. In North Sami (Utsjoki), alternations in paradigms involve three grades of quantity shaping: mânâm ‘I go’ (â is a Sami letter for a somewhat rounded a) versus mânna ‘he goes’ versus mân′ne ‘goer’; dieðam ‘I know’ versus dietta ‘he knows’ versus diet′te ‘knower’; juol′ge ‘leg’ versus juolge ‘of the leg.’ This series of contrasts shows a three-stage decrease in initial-vowel duration and a three-stage increase in the duration of the first consonant after the first vowel or vowels. The other northern and eastern Sami languages display similar alternations, but there is considerable diversity in the phonetic details.

Word order

The grammatical structures of the various Uralic languages, despite numerous superficial differences, generally indicate a basic Early Uralic sentence structure of (subject) + (object) + main verb + (auxiliary verb)—the parenthesized elements are optional, and the last element is the finite (inflected) verb, which is suffixed to agree with the subject in person and number. This pattern has been best preserved in the more eastern languages, especially Samoyed, Yukaghir, and Ob-Ugric—e.g., Nenets t́iky pevśumd’o-m saravna t′eńe-vaʔ ‘we well remember that evening’ (literally, ‘that evening-[accusative] well remember-we’); Mari joltaš-em-blak lum tol-mə-m buč-aš tüŋal-ət ‘my friends begin to wait for the coming of snow’ (literally, ‘friend-my-[plural] snow coming-[accusative] wait-to begin-they’); Yukaghir met Tolstoj-wiejuol-knigleŋ juonumeŋ ‘I see a book written by Tolstoy’ (literally, ‘I Tolstoy-written-book see-[present auxiliary]’). This order is common but optional in the languages of central Russia. Sami, Baltic-Finnic, and Hungarian now show the typical European subject–verb–object order: e.g., Finnish isä osti talo-n ‘father bought a house(-genitive),’ Hungarian János keres egy ház-at ‘John seeks a house(-accusative).’ Although the latter languages have relatively “free” word order, the object precedes the verb only for special emphasis—e.g., Hungarian János egy házat keres ‘John is looking for a house (and not something else),’ Estonian ma ta-lle leiba ei anna ‘I won’t give him any bread’ (literally, ‘I him-to bread not give’). Estonian sentence structure somewhat resembles that of German, with its tendency to place the finite verb in second position while the rest of the verb complex remains at the end of the sentence—e.g., mehe-d ol-i-d ammu koju jõud-nud ‘the men had got home long ago’ (literally, ‘man-[plural] be-[past]-they long-ago home arrive-[past participle]’).

The verb “be”

In place of a verb “have,” the Uralic languages use the verb “be,” expressing the agent in an adverbial (locative or dative) case—e.g., Finnish isä-llä on talo ‘father has a house’ (literally, ‘father-at is house’), Hungarian János-nak van egy ház-a ‘John has a house’ (literally, ‘John-to is one house-his’). In Proto-Uralic the copula verb “be” was lacking in simple predicate adjective or noun sentences, although the predicate was probably marked to agree with the subject. The following Hungarian sentences reflect this situation: a ház fehér ‘the house [is] white,’ a ház-ak fehér-ek ‘the houses [are] white.’ In Nenets and Mordvin such nonverbal predicates, even nouns, are conjugated for subject agreement and tense in the manner of intransitive verbs—e.g., Nenets mań xańenadmʔ ‘I am a hunter,’ pydaŕỉ xańenadiʔ ‘you two are hunters,’ mań xańenadamź ‘I was a hunter,’ pydaraʔ xańenadać ‘you (plural) were hunters.’ Otherwise, a wide range of grammatical usage is found. In Baltic-Finnic and Sami the use of a copula verb is obligatory, in Permic it is optional, and in Hungarian the copula is absent only in the third person (“he, she”) in a nonpast tense.

Negative sentences and questions

Negative sentences in Early Uralic were indicated by means of a marker known as an auxiliary of negation, which preceded the main verb and was marked with suffixes that agreed with the subject and perhaps tense. This is best reflected in the Finnic, Samoyedic, and Yukaghir languages—e.g., Finnish mene-n ‘I go,’ e-n mene ‘I don’t go,’ mene-t ‘you go,’ e-t mene ‘you don’t go’; Yukaghir met elūjeŋ ‘I didn’t go’ (with negative prefix el- [äl- in Finnish]; compare met merūjeŋ ‘I went’). Ugric employs undeclined negative particles (e.g., Hungarian nem), and in Estonian only negative imperative forms are still conjugated, although colloquial Estonian has initiated a tense distinction—e.g., ma/sa ei tule ‘I/you don’t come’ and ma/sa e-s tule ‘I/you didn’t come.’

In Proto-Uralic, questions were formed with interrogative pronouns, beginning with *k- and *m-, illustrated by Finnish kuka ‘who,’ mikä ‘what’ and Hungarian ki ‘who,’ mi ‘what.’ Yes–no questions were formed by attaching an interrogative particle to the verb, as in Finnish mene-n-kö ‘am I going?’ and e-n-kö minä mene ‘am I not going?’ (in Finnish the verb also shifts to initial position). The use of intonation (changes in pitch) in interrogative sentences is currently widespread. In Hungarian it is the only way to form direct yes–no questions, although in indirect questions a particle -e is used—e.g., a házak fehérek? (with sharply rising intonation of the next to the last syllable, dropping again on the final syllable) ‘are the houses white?,’ nem tudom, fehérek-e a házak ‘I don’t know if the houses are white.’

Conjunction

Conjunction, the connecting of clauses, phrases, or words, was formerly without the aid of specialized conjunctions. In the modern languages the conjunctions are largely borrowings from Germanic (Finnish ja ‘and’) and Russian (Mari da ‘and; in order to,’ a ‘but,’ ńi…ńĭ ‘neither…nor,’ jesle ‘if’). Both coordination and subordination in sentences are marked by a wide range of constructions, especially by means of infinitive verbs, participles, and gerunds—e.g., Mari keče peš purgəžan poranan ulmaš ‘the weather was very stormy and snowy’ (literally, ‘weather very stormy snowy was’), ača-ž aba-št ‘their father and mother’ (literally, ‘father-his mother-their’), nuno batə-ž-ẖḥn ‘he and his wife’ (literally, ‘they wife-his with’); Finnish kirja-n lue-ttu-a-ni ‘when I had read the book…’ (literally, ‘book-[genitive] read-[past passive participle-partitive case]-my’), luke-akse-ni kirja-n ‘in order for me to read the book’ (literally, ‘read-to-[translative case]-my book-[genitive]’).

The case system

Verb inflections

The Proto-Uralic verb was inflected for tense-aspect (*-pa indicated “nonpast,” *-ka indicated “perfect nonpast; imperative,” *-ja indicated “past”) and mood (*-ne indicated “conditional-potential”). The use of auxiliary verbs to indicate tenses was unknown, although Sami, Baltic-Finnic, and Hungarian now have essentially a Germanic-type tense system, with perfect formations based on the “be” verb; e.g., Finnish mene-n ‘I go,’ ole-n men-nyt ‘I have gone’ (‘be-I go-[past participle]’), men-i-n ‘I went,’ ol-i-n men-nyt ‘I had gone,’ men-isi-n ‘I would go,’ ol-isi-n men-nyt ‘I would have gone.’ Under Germanic and Slavic influence both Estonian and Hungarian have developed separable verbal prefixes with adverbial and aspectual meanings; e.g., Estonian ära söö- ‘eat (perfective)’ and ta sõ-i kala ära ‘he ate the fish’ versus ta sõ-i kala ‘he was eating fish,’ ta hakkas kala ära söö-ma ‘he began to eat (up) the fish’; Hungarian meg-tanul ‘learn (perfective)’ and János megtanul-t magyar-ul ‘John learned Hungarian’ versus János tanult magyarul ‘John was learning Hungarian,’ János tanult meg angolul ‘John learned English,’ János nemetül tanult meg ‘John learned German’ (with special emphasis as indicated).

Proto-Uralic did not have specialized voice markers, such as the Indo-European passive; rather, the function of voice was interwoven with topicalization (a way of indicating the main subject of a sentence), emphasis, and definiteness of the subject and object as well as with verbal aspect. An indefinite subject of an intransitive verb or an indefinite object were marked with the ablative case (*-ta), but a definite object took the accusative marker (*-m) and other subject situations were unmarked (nominative). This system is best preserved in Finnish: vesi (nominative) juoksee ‘the water is running’ versus vettä juoksee ‘there is water running,’ juon vede-n ‘I will drink the water’ (-n is from older *-m) versus juon vettä ‘I drink water.’ (Note that aspect as well as tense is affected by these case distinctions.)

The widespread use of separate subjective and objective conjugations among the Uralic languages (as in Mordvin, Ugric, and Samoyedic) are the result of an original system for singling out the subject or object for emphasis (focus), and not simply a device for object–verb agreement (similar to subject agreement). For example, Nenets tymʔ xada-v ‘I killed a deer (focus on the agent)’ versus tymʔ xada-dmʔ ‘I killed a deer (focus on the object),’ in which -v signifies ‘I…it’ (the objective conjugation) and -dmʔ signifies ‘I’ (the subjective conjugation). Note also the objective forms xada-n ‘I killed [them],’ xada-r ‘you (singular) killed [it],’ xada-d ‘you (singular) killed [them],’ and so on for nine possible subjects (three persons times singular, dual, plural) times two object numbers (singular and nonsingular [not actually distinguished with third-person subjects]); and the subjective forms xad-n ‘you (singular) killed’ and so on, for nine subject agreements. Yukaghir similarly employs distinct conjugations to reflect sentence focus; e.g., met ai ‘I shot (focus on subject),’ met meraiŋ ‘I shot (focus on verb),’ met ileleŋ aimeŋ ‘I shot the deer (focus on object).’ Hungarian opposes definite and indefinite conjugations: two different sets of personal endings are used—one with transitive verbs with definite objects and the other elsewhere—e.g., olvas-om/od a level-et ‘I/you read the letter’ versus olvas-ok/ol egy level-et ‘I/you read a letter.’ Along with its subjective and objective conjugations, Khanty has added a so-called passive conjugation (compare kitta-j-m ‘I am being sent,’ -j- = “passive”) as an extension of the earlier focus-topicalization system. Mari and Komi have two past tense formations with related function. Again, the westernmost languages have passive constructions similar to those in both Slavic and Germanic.

Verbal derivation was richly developed already in Proto-Uralic with a wide variety of verbal nouns, infinitives, and participles. Each of the three tense-aspect markers was apparently used as a participial formative (compare Finnish lähde from *läkte-k ‘source,’ lähtijä ‘one who leaves,’ lähte-vä from *-pa ‘leaving’).

Several of the modern Uralic languages make extensive use of their native derivational processes to eliminate foreign loanwords; e.g., for ‘telephone’ Finnish has puhelin, which is derived from puhel- ‘talk,’ just as soitin ‘musical instrument’ comes from soitta- ‘to play.’ The Uralic finite verb originally may have been based on participial constructions parallel to the noun-plus-predicate-adjective sentences (like Hungarian a ház fehér ‘the house [is] white’). Thus, one may reconstruct sentences like *ema tumte-pa ‘mother [is] knowing,’ *ema tumte-pa-ta ‘mothers [are] knowing’ (with subject number expressed only in the predicate [agreement]) to explain the close similarity of participial and finite verb constructions such as Estonian tundev ema ‘knowing mother,’ tundvad emad ‘knowing mothers,’ ema tunneb ‘mother knows,’ emad tunnevad ‘mothers know.’