||This article includes a list of references, but its sources remain unclear because it has insufficient inline citations. (June 2013)|
Proto-Finnic or Proto-Baltic-Finnic is the common ancestor of the Finnic languages, which include the national languages Finnish and Estonian. Proto-Finnic is not attested in any texts, but has been reconstructed by linguists. Proto-Finnic is itself descended ultimately from Proto-Uralic.
- 1 Background
- 2 Phonology
- 3 Grammar
- 4 Later developments
- 5 Notes
- 6 References
Three stages of Proto-Finnic are distinguished in literature.
- Early Proto-Finnic, the last common ancestor of the Finnic languages and its closest external relatives — usually understood to be the Samic languages, though also the Mordvinic languages may derive from this stage (see Finno-Samic languages). This reconstruction state appears to be almost identical to Proto-Uralic.
- Middle Proto-Finnic, an earlier stage in the development on Finnic, used in Kallio (2007) for the point at which the language had developed its most characteristic differences from Proto-Uralic (mainly: the loss of several consonant phonemes from the segment inventory, including all palatalized consonants).
- Late Proto-Finnic, the last common ancestor of Finnish and Estonian. Most other Finnic languages derive from this stage as well. It however appears that the South Estonian language and the Livonian language had already diverged earlier.
Changes up to Late Proto-Finnic
|This section is incomplete. (April 2014)|
- *ë > *a. This change is shared by several other western Uralic languages, including the Samic and Mordvinic languages.
- Word-initial deaffrication:
- *č > *š.
- *ć > *ś.
- *ŋ and *x are lost as phonemes. Between vowels they are usually lost entirely, triggering lengthening of a preceding vowel.
- In certain cases, this may have proceeded thru vocalization to *w. Compare e.g. PU *mexi- > PF *möö- "to sell"; PU *sewi- > PF *söö- "to eat".
- Before dental/alveolar consonants, both consonants usually vocalize to *w.
- The cluster *ŋk remains, but *ŋ in this case is now simply an allophone of *n rather than an independent phoneme.
- *ć > *c.
- *ś > *s.
- *δ́ > *δ.
- *ń > *n.
- *ĺ > *l.
- Lengthening of open vowels:
- *a > *oo (including former *ë) and *ä > *ee, when the vowels appear
- in an open syllable, and
- followed by a non-semivowel sonorant consonant (*m, *n, *l, *r, *δ), and
- followed by an original non-open vowel *i (also denoted *ə, *e).
- E.g: PU *ńäli- > PF *ńeele- "to swallow"; PU *ńëli > PF *nooli "arrow"
- *a > *oo (including former *ë) and *ä > *ee, when the vowels appear
- *δ > *t.
- In non-initial syllables, vowels followed by *j are modified in various ways. In particular, low vowels are raised.
- *äj > *ej > *ij > *i
- *aj > *oj when the preceding syllable contains a non-rounded vowel.
- *aj > *ej > *ij > *i elsewhere.
- Word-final *-e becomes *-i.
- *ti is assibilated to *ci. The change was blocked if another coronal obstruent preceded, i.e. *tti, *cti, *sti, *šti (thus Finnish kaksi "two" ~ kahden < Pre-Proto-Finnic *kakti, but lehti "leaf" ~ lehden < *lešti).
- Apocope of final *-i when at least two syllables preceded. This occurred after assibilation, which created alternations between final *-c and medial *-t- in some nouns (e.g. Finnish nouns in -us, genitive -uden, essive -utena).
- Syncope of medial *-e- between certain consonants and *t. This almost universally happened after a coronal consonant (Finnish vesi "water", partitive vettä < Pre-Proto-Finnic *vetetä). The sequence *-met- was also often syncopated (Finnish lumi "snow", partitive lunta < Pre-Proto-Finnic *lumeta) and occasionally even *-ket- (Finnish nähdä "to see" < Post-Proto-Finnic *nähtäk < Proto-Finnic *näktäk < Pre-Proto-Finnic *näketäk). It is possible that in at least some cases that show apparent syncope in the later Finnic languages, the -e- was never there to begin with, and was epenthetically added to most other forms instead. Finnish ohut and kevät, for example, seem to have been genuine consonant stems, as any stem-final *-e would have become *-i and would have then triggered assibilation (like in the -us nouns noted above), giving instead *ohus and *keväs in Finnish.
- Application of radical gradation in closed syllables, causing voicing of short obstruents and shortening of geminate stops. This occurred after apocope, or was still productive at the time, as the newly consonant-final syllables resulting from apocope triggered gradation as well.
- *š > *h
- The clusters *tš and *kš lose their first component to also become simple *h.
- Loss of glides before vowels:
- *ji > *i. This is seen in Finnish veli "brother", genitive veljen.
- *je > *e word-initially.
- *vu > *u.
- *vü > *ü.
- *vo > *o. This change must date until after Proto-Finnic broke up, as Estonian and Võro võtma "to take" preserved the consonant until after the dialectal unrounding of *o to *õ (which prevented the change from affecting it). Compare Finnish ottaa, Veps otta, where it did apply as there was no unrounding in those dialects.
The sounds of Proto-Finnic can be reconstructed through the comparative method.
Reconstructed Proto-Finnic is traditionally transcribed using the Uralic Phonetic Alphabet. The following UPA and related conventions are adopted in this article for transcribing Proto-Finnic forms:
- Front vowels are denoted with a diaeresis, following Estonian and (partly) Finnish orthography: ä ö ü.
- The affricate /t͡s/ is written as c.
- The sound /x/ is written as h.
- Long consonants and vowels are written doubled: aa ee ii pp tt kk cc etc.
- Half-long consonants are written with a following apostrophe: p' t' k' c'.
- The labial semivowel /ʋ ~ w/ is written as v.
- Diphthongs are written with two vowel letters when a consonant follows: au ai (not av aj).
The Proto-Finnic consonant inventory had relatively few phonemic fricatives, much like that of the modern Finnic languages. Voicing was not phonemically contrastive, but the language did possess voiced allophones of certain voiceless consonants.
The table below lists the consonantal phonemes of Late Proto-Finnic. Phones written in parentheses represent allophones and are not independent phonemes. When a consonant is notated in this article with a symbol distinct from the corresponding IPA symbol, the former is given first, followed by the latter.
|Voiced||(β [β])||(δ [ð])||(γ [ɣ])|
- *h had evolved fairly late from the Middle Proto-Finnic postalveolar sibilant *š ([ʃ]). It may have been realised as [h] before another consonant.
- *v was perhaps realised as labiodental [ʋ] when a vowel followed, rather than a true bilabial
- [ŋ] was an allophone of *n before *k. The original Proto-Uralic phoneme *ŋ had been lost and changed into other sounds, except in this position.
- [b β], [d ð] and [ɡ ɣ] were allophones of *p, *t and *k respectively, and developed as a result of consonant gradation.
- Final *-k was probably unreleased [k̚].
Proto-Finnic possessed two phonemic levels of consonant duration, short and long (geminate). The contrast itself had been inherited from Proto-Uralic, but was considerably expanded: all consonants except *r, *h, *j and *w could be short or long. The three plosives also possessed a half-long duration ([pˑ], [tˑ] and [kˑ]), but these were in complementary (allophonic) distribution with fully long consonants, and therefore were not phonemic. They appeared in predictable positions as a result of consonant gradation, like the voiced fricatives.
|Strong grade||Weak grade|
|cc [tt͡s]||c' [t͡sˑ]|
Voiced plosives occurred after nasals (mb nd ŋg), voiced fricatives in all other weak grade environments.
Gradation occurred in two different environments, and can therefore be split into two types:
- Radical gradation affected consonants that appeared at the beginning of a closed syllable (a syllable that ended in a consonant). It affected consonants preceded by a vowel or sonorant, but not those preceded by another obstruent.
- Suffixal gradation affected consonants that appeared at the beginning of a non-initial odd-numbered syllable. It only affected consonants preceded by a vowel and did not affect the geminates[clarification needed].
It is unclear whether consonant gradation was a Finnic innovation, or a retention of an old Uralic feature that was lost in most other Uralic branches. It is likely that it was inherited from an earlier stage that was also the ancestor of the Sami languages, which have gradation that is very similar to that found in the Finnic languages. However, it was still productive after certain sound changes specific to Finnic, such as the apocope of final *-i, so it was probably present as a phonetic "post-processing" rule (Surface filter) over a long period of time. It is no longer fully productive in any Finnic language, but most languages still retain large amounts of words preserving the earlier alternations.
The Proto-Finnic vowel inventory was very similar to that of modern Finnish, although the distribution of the sounds was different. The following table lists the monophthong vowels reconstructable for Proto-Finnic.
|o, oo(, ë)
/o/, /oː/(, [ɤ])
All vowels could occur both short and long. In Proto-Uralic, rounded vowels (*u, *ü, *o) could not occur in non-initial syllables, but because of sound changes, they did appear in Proto-Finnic.
The short unrounded mid back vowel *ë was not an independent vowel, but appeared as the counterpart of the front vowel *e in the system of harmony. It merged with *e in most Finnic languages, but not in South Estonian or Votic. See below under vowel harmony for more details.
The status of short *ö is unclear. It was not present in ancestral Proto-Uralic, and many instances of ö found in modern Finnic languages have only developed after Proto-Finnic, due to various sound changes. For example, Finnish has öy from *eü: löytä- 'to find' < Proto-Finnic *leütä-, while Estonian has unrounded the diphthong instead, giving leida-. Short ö was also generally added to the system for reasons of symmetry, to complete the system of vowel harmony (see below). This happened in Finnish näkö 'sight' < Proto-Finnic *näko, but not in Votic näko.
The existence of long öö is clear, as this sound had regularly evolved from other combinations of sounds, in words of Uralic origin (e.g. *söö- 'to eat' ← Proto-Uralic *sewi-).
Proto-Finnic also possessed diphthongs, which were formed by combinations of a short vowel with the vowels /i/, /y/ and /u/, or equivalently with the semivowels /j/ and /w/. Long vowels did not form diphthongs; if they were followed by a (semi)vowel, they were shortened.
|Front + *i||Front + *ü||Front + *u||Back + *i||Back + *u|
|Mid to close||*ei(,*öi)
|Open to close||*äi
For *öi, see the note on *ö above.
Proto-Finnic possessed a system of vowel harmony very similar to the system found in modern Finnish. Vowels in non-initial syllables had either a front or a back vowel, depending on the quality of the vowel of the first syllable. If the first syllable a contained front vowel, non-initial syllables would contain such vowels as well, while back vowels in the first syllable would be matched with back vowels in the other syllables. Thus, all inflectional and derivational suffixes came in two forms, a front-harmonic and a back-harmonic variety.
In non-initial syllables, the vowels e and i were originally a single reduced schwa-like vowel in Proto-Uralic, but had become differentiated in height over time. i arose word-finally, while e appeared medially. These vowels were front vowels at the time, and had back-vowel counterparts ë and ï. In Proto-Finnic, ï had merged into i, so that i was now neutral to vowel harmony and could occur in both front-vowel and back-vowel words, even if it was phonetically a front vowel. The vowels e and ë appeared to have remained distinct in Proto-Finnic, and remained so in South Estonian (as e and õ) and Votic. In the other Finnic languages, they merged as e.
Stress was not phonemic. Words were stressed in a trochaic pattern, with primary stress on the first syllable of a word, and secondary stress on every following odd-numbered syllable.
Root words were normally bisyllabic, and generally followed the structure CVCV, CVCCV CVVCV. A word could begin and end with at most one consonant; consonant clusters were permitted only word-internally. Any consonant (as far as allophones allowed) could begin a word, but only the alveolar consonants (*n, *t, *r, *s and perhaps *c) and the velars *k and *h could appear word-finally. Final *-k and *-h were often lost in the later Finnic languages, but occasionally left traces of their former presence.
Word-internal consonant clusters were limited to two elements originally. When impermissible clusters appeared (that is, any word-initial or word-final cluster, or a word-internal cluster of more than two consonants), this was generally solved by deleting one or more elements in the cluster, usually the first. This led to alternations that are still seen in, for example, Finnish:
- laps-i ("child", nominative) + -ta > las-ta (partitive), with medial simplification -pst- > -st-
- stem tuhant- ("thousand"):
- > tuhat (nominative), with final simplification -nt > -t
- > tuhatta (partitive), with medial simplification -ntt- > -tt-
- stem kolmant- ("third")
- > kolmas (nominative), with final simplification (but note assibilation t > s, so this reflects earlier *-nci before apocope)
- > kolmatta (partitive), with medial simplification -ntt- > -tt- (as assibilation shows that this had a stem-final vowel originally, this reflects earlier *-ntet- before syncope)
Note in the examples of tuhatta and kolmatta that Proto-Finnic did not initially tolerate clusters of a sonorant plus a geminate consonant. These have sometimes become permissible in the later Finnic languages, and they are particularly abundant in modern Finnish.
All inflectional and derivational endings containing a or u also had front-vowel variants with ä and ü, which matched the vowels in the word stem following the rules of vowel harmony. o did not follow this rule, as noted above.
Endings which closed the final syllable of a word triggered radical gradation on that syllable. An ending could also open a previously closed syllable, which would undo the gradation. Suffixal gradation affected the endings themselves. For example, partitive -ta would appear as -da when added to a two-syllable word ending in a vowel (e.g. *kala, *kalada "fish"), but as -ta after a third syllable or a consonant (*veci, *vettä "water").
Nouns and adjectives
Proto-Finnic nouns declined in at least 13 cases. Adjectives did not originally decline, but adjective-noun agreement was innovated in Proto-Finnic, probably by influence of the nearby Indo-European languages. The plural of the nominative and accusative was marked with the ending -t, while the plural of the other cases used -i-. The genitive and accusative singular were originally distinct (genitive -n, accusative -m), but had fallen together when final -m became -n through regular sound change. Some pronouns had a different accusative ending, which distinguished them.
|Nominative||∅||*-t||Subject, object of imperative|
|Accusative||*-n (also -t)||*-t||Complete (telic) object|
|Genitive||*-n||*-ten (-den, -ðen)
|Partitive||*-tA (-ðA)||*-itA (-iða)||Partial object, indefinite amount|
|Elative||*-stA||*-istA||Motion out of|
|Illative||*-sen (-hen)||*-ihin||Motion into|
|Allative||*-len / *-lek||*-ilen / *-ilek||Motion onto/towards|
|Essive||*-nA||*-inA||Being, acting as|
|Translative||*-ksi||*-iksi||Becoming, turning into|
|Comitative||*-nek||*-inek||With, in company of|
|Instructive||*-n||*-in||With, by means of|
The genitive plural was formed in two different ways:
- The "western" type was formed by adding the singular ending to the nominative plural, with an additional vowel: *-t-en. This then became *-den after a vowel through consonant gradation.
- The "eastern" type was formed by using the general plural stem in -i- as the base: *-i-en. This became *-jen after a vowel.
Both types are still found in Finnish, although one or the other typically dominates. In the western type, the regular loss of -d- after an unstressed (even-numbered) syllable has created forms such as -ain (< *-a-den), which are now archaic, however.
Proto-Finnic also had special possessive suffixes for nouns, which acted partly as genitives. The following are reconstructable:
|First person||-ni||-mme / -mma|
|Second person||-ci||-nne / -nna|
- Indicative mood - suffix: present none (but -k- in the passive), past -i-
- Imperative mood - suffix: -k-
- Conditional mood - suffix: -ksi- or -isi-
- Potential mood - suffix: -ne-
The indicative mood distinguished between present (which also functioned as future) and past tense, while the other moods had no tense distinctions. New "perfect" and "pluperfect" tenses had also been formed, probably by influence of the Indo-European languages. These were created using a form of the copula *oldak "to be" and a participle.
There were six forms for each mood, for three persons and two numbers. In addition, there were two more forms. One was a form that is often called "passive" or "fourth person", and indicated an unspecified person. The second was the "connegative" form, which was used together with the negative verb to form negated sentences.
|First person||*-n||*-mmek / -mmAk|
|Second person||*-t||*-ttek / *-ttAk|
|Third person||*-pi (-βi), -||*-βAt, -|
|Passive||*-ttA- + (tense/mood suffix) + *-sen (-hen)|
The variation between forms with *-e- and forms with *-A- in the 1st and 2nd person plural reflects a former distinction between the dual and the plural (respectively), although this has not been attested from any Finnic variety.
The third person forms only had an ending in the present indicative. In all other tenses and moods, there was no ending and the singular and plural were identical.
The 3rd person singular was entirely unmarked in South Estonian: the Late Proto-Finnic ending had evolved from the participle *-pA during the Middle Proto-Finnic stage, and this innovation had not reached the already separated South Estonian.
The imperative had its own set of endings:
|First person||-kadamme/a / -kademme/a|
|Second person||-k||-kada / -kade|
|Third person||-kahen, -kohen||-kahen, -kohen|
|Passive||-ttakahen / -ttakohen|
|Connegative||-kak / -kok|
In addition, there were also several non-finite forms.
|Infinitive I||*-tAk (-dAk, -ðAk) : *-tA- (-dA-, -ðA-)|
|Infinitive II||*-te- (-de-, -ðe-)|
|Gerund ("Infinitive III")||*-mA|
|Action noun ("Infinitive IV")||*-minen : *-mise-|
|Present active participle||*-pA (-βA)|
|Present passive participle||*-ttApA (-ttAbA)|
|Past active participle||*-nUt|
|Past passive participle||*-ttU|
The following is an overview of the more important changes that happened after the Proto-Finnic period.
Changes to consonants
- Obstruent cluster assimilation. This change happened very late in the Proto-Finnic period, but as South Estonian developed somewhat differently, it shows that dialectal diversification was beginning to occur around this time.
- In South Estonian:
- *kt, *pt > *tt (*kakteksa "two" > katõssa). Compare *lehti "leaf" > leht' where *ht originates from *št; these two clusters remain distinguished only in South Estonian, and merge everywhere else due to the change below.
- *kc, *pc > *cc (*lapci "child" > lats')
- *ks, *ps > *ss ("two" above, and also *maksa "liver" > mass)
- In other Finnic dialects, the effect is much more limited:
- *kt, *pt > *ht after a stressed syllable.
- *kt, *pt > *tt elsewhere.
- Note that in all Finnic dialects, *pt and *kt have the same reflex. It is therefore impossible to distinguish them in reconstruction, unless there is additional internal evidence (in the form of grammatical alternations) or external evidence (from non-Finnic languages).
- In South Estonian:
- *c > *s generally: Proto-Finnic *ükci "one", *veci "water", *lapci "child", *cika "pig" > Finnish yksi, vesi, lapsi, sika.
- This often makes it impossible to distinguish the two sounds using Finnic evidence alone, if internal reconstruction is not viable (e.g. via t ~ s alternations from assibilation, or South Estonian ts versus other Finnic ps/ks).
- South Estonian preserves an afficate in some cases, such as Võro tsiga "pig" versus Estonian siga, Finnish sika < Proto-Finnic *cika.
- The geminate affricate *cc generally remains. Its gemination was no longer contrastive due to the previous change, so it often lost its status as a phoneme and "degraded" into the cluster *ts.
- In early Finnish it is fronted to interdental *θθ, which later changes into a variety of other dialect-specific sounds such as tt or ht. The sound is written as 〈ts〉 in the earliest records, and in Standard Finnish this eventually causes the pronunciation to shift back to ts.
- Loss of final consonants
- Final *-k was generally lost. It is preserved in some dialects:
- In Eastern Votic as -g
- In Võro as -q (a glottal stop /ʔ/)
- In Finnish it became a sandhi effect, assimilating to the initial consonant of the following word and lengthening it. This effect does not occur in all dialects and is not represented orthographically, but is often noted with a superscript "ˣ" in reference works.
- Final *-h is widely lost as well. It is preserved:
- In Karelian and Veps as -h.
- In Southern Estonian as either -h or as a glottal stop -q.
- In Finnish with sandhi effects like those of final *-k. In Western dialects there was also metathesis, which preserved the original *h along with sandhi lengthening, e.g. Proto-Finnic *mureh "sorrow" > Western Finnish murheˣ (Karelian mureh, Võro murõh/murõq) and Proto-Finnic *veneh "boat" > Western Finnish venheˣ (Karelian/Veps veneh, Võro vineh/vineq).
- Standard Finnish inconsistently adopts some words in their Western Finnish shape (e.g. murhe; perhe "family", valheellinen "untrue"), some in their Eastern Finnish shape (e.g. vene; vale "lie").
- Final *-n is lost in most of the South Finnic area (as well as widely in modern-day colloquial Finnish). In Votic this triggers compensatory lengthening of the preceding vowel.
- The 1st person verbal ending resists the change, and generally remains as -n.
- Final *-k was generally lost. It is preserved in some dialects:
- The reintroduction of palatalized consonants in most varieties other than Western Finnish.
- The most widespread source of palatalization are a lost word-final or word-medial *-i (a form of cheshirization), and consonant clusters with *j as a 2nd member.
- In several varieties, a diphthong ending in *-i and the long vowel *ii also cause palatalization of a following consonant.
- In Votic, *k and *γ are palatalized to tš and j adjacent to all front vowels.
- Postalveolar sibilants š [ʃ] and possibly ž [ʒ] are also reintroduced to most varieties.
- In Estonian, Votic and Finnish, š occurs almost solely in loanwords, most commonly of Russian or German origin.
- In Livonian, palatalized *ś and *ź generally shift to postalveolar š and ž.
- In Veps, *s affected by progressive palatalization (and possible intervocalic voicing) yields š or ž, while *s affected by regressive palatalization yields ś or ź. E.g. Proto-Finnic *viici "five" > *viisi > Veps viž; but Proto-Finnic *kuuci "six" > *kuusi > Veps kuź.
- In Northern Karelian, a general shift *s > š occurs, except blocked in progressive palatalization contexts.
- Loss or modification of most of the voiced fricatives that resulted from consonant gradation.
- In the marginal languages Livonian and Veps, all three spirants *β *δ *γ are reflected as plain voiced stops b d g.
- *β merges with *v in all languages other than Livonian and Veps.
- *mb and *nd assimilate to mm and nn almost as widely, though they are also retained in Olonets Karelian and Livvi.
- *ŋg is assimilated to the corresponding geminate ng /ŋː/ in several of the Finnish dialects. However, given the lack of a pre-existing *ŋ, the cluster widely "un-gradates" back to *ŋk.
- *δ is widely assimilated to a preceding liquid consonant (*l or *r), creating the characteristic gradation patterns lt : ll and rt : rr.
- In other positions, *δ *γ are lost early on in the other languages of the Eastern Finnic group (Eastern Finnish, Karelian, Ludic and Ingrian) as well as in Estonian. After a stressed syllable, they remain until relatively late in Western Finnish, being lost or assimilated to other sounds only during the 2nd half of the 2nd millennium.
- Standard Finnish adopted the spelling 〈d〉 early on, which eventually led to a standard pronunciation /d/. Individual words may follow particular dialects instead, e.g. zero in *nauδetta > navetta "cowshed", /l/ in *taδikkoi > talikko "manure fork".
- *γ is fortified to g in Votic, when not palatalized (see above). *δ is lost, much as in Estonian.
- Loss of spirants often created new long vowels and diphthongs, particularly in non-initial syllables (e.g. Finnish auttaa "to help" < Proto-Finnic *aβuttaδak; cf. the unmodified strong grade in apu "help, assistance").
- Lenition of the voiceless (strong grade) obstruent consonants *p, *t, *k, *s between voiced sounds.
- These become voiced in Veps and Livonian (the stops merging with their weak grade counterparts), as well as in Southern Karelian.
- Estonian b, d, g are lax voiceless consonants [b̥], [d̥], [g̊].
- Intervocalic obstruents remain unlenited in Finnish, Votic, and Northern Karelian.
- Vocalizations of coda consonants
- *n before *s (including that from former *c) is lost in the Southern Finnic group, with compensatory lengthening of the perceding vowel. E.g. *kanci "lid" > Estonian kaas, Finnish kansi; *pensas "bush" > Es. põõsas, Fi. pensas.
- In Western Finnish, stop consonants before a sonorant are vocalized to u. E.g. *kapris "goat", *atra "plough", *kakra "oats" > Finnish kauris, aura, kaura; Estonian kaber, ader, kaer; Karelian kapris, atra, kakra.
- Coda *l is vocalized in Veps at a late date, creating u-final diphthongs in the northern and central dialects, long vowels in the southern.
Changes to vowels
- In the southern Finnic languages, a new back unrounded mid vowel develops from *e in words with back vowel harmony (e.g. Estonian võlg "debt" versus Finnish velka).
- In Estonian and Votic, instances of õ also develop from original *o under certain conditions (e.g. Estonian õlg "hay" versus Finnish olki).
- Vowel length is lost before h early on in the southern Finnic group and in Veps.
- Later on, all long vowels are shortened in Veps.
- Diphthongization of long vowels
- The long mid vowels *oo, *öö and *ee become opening diphthongs in Finnish (uo, yö, ie), Karelian, and several marginal dialects of Northern Estonian.
- Diphthongization also occurs in Livonian and Southern Estonian, but only under certain conditions. The mid back unrounded long vowel õõ of these languages is not affected.
- In Livonian, the short vowels *o and *e may also diphtongize, leading to a contrast of short uo ie vs. long ūo īe.
- In Eastern Finnish and Karelian, the low vowels *aa and *ää also diphthongize (Karelian oa, eä; Savonian ua, iä).
- Assimilations of diphthongs
- The diphthong *eü is fully labialized to öü in Northern Finnic and South Estonian. North Estonian instead unrounds all diphthongs ending in -ü to -i (*eü > ei, *äü > äi).
- In northern dialects of Veps, this is followed by the creation of new long close vowels by the raising of several diphthongs: *ei and *öi > ii; *iu, *iü, *eu and *öü > üü ~ üu; *au and *ou > uu.
- In Savonian Finnish, the 2nd element of all labial diphthongs is lowered: *äü > äö ~ ää, *au > ao ~ aa, etc.
- In Livonian, *au is labialized to ou, and *äi is palatalized to ei. Following this, the mid diphthongs are smoothed to long vowels under certain conditions: ou > oo; õi, õu > õõ; ei > ee.
- Loss of short final vowels after long syllables in Veps, both Northern and Southern Estonian, and Livonian (e.g. Estonian kaks "two", viis "five" versus Finnish kaksi, viisi < Proto-Finnic *kakci, *viici).
- Loss of vowel harmony in Estonian, Livonian and partly Veps, but not South Estonian or Votic (e.g. *külä "village" > Estonian küla versus Finnish kylä, Votic tšülä, Võro külä). Languages that retain vowel harmony often extend it to *o as well.
- Kallio 2007, p. 231.
- Posti 1953, pp. 10–21.
- Kallio 2007, pp. 231–233.
- Itkonen 1949.
- Janhunen 2007, pp. 221–222.
- Kallio 2007, p. 233.
- Posti 1953, pp. 26–29.
- Aikio 2012.
- Posti 1953, pp. 83–85.
- Laakso 2001.
- Lehtinen 2007, p. 137.
- Lehtinen 2007, pp. 138-139.
- Petri Kallio, The non-initial-syllable vowel reductions from Proto-Uralic to Proto-Finnic, 2012
- Lehtinen 2007, pp. 124-125.
- Lehtinen 2007, pp. 125-126.
- Lehtinen 2007, pp. 134-135.
- Viitso 1998, p. 108.
- Aikio, Ante (2012), "On Finnic long vowels, Samoyed vowel sequences, and Proto-Uralic *x", Suomalais-Ugrilaisen Seuran toimituksia 264, ISSN 0355-0230.
- Itkonen, Erkki (1949), "Beiträge zur Geschichte der einsilbigen Wortstämme in Finnischen", Finnisch-Ugrische Forschungen 30.
- Janhunen, Juha (2007), "The primary laryngeal in Uralic and beyond" (pdf), Suomalais-Ugrilaisen Seuran toimituksia 253, ISSN 0355-0230, retrieved 2010-05-05.
- Kallio, Petri (2007), "Kantasuomen konsonanttihistoriaa" (pdf), Mémoires de la Société Finno-Ougrienne 253.
- Laakso, Johanna (2001), "The Finnic languages", The Circum-Baltic languages volume 1: Past and Present, John Benjamins.
- Lehtinen, Tapani (2007), Kielen vuosituhannet, Tietolipas 215, Suomalaisen Kirjallisuuden Seura, ISBN 978-951-746-896-1.
- Posti, Lauri (1953), "From Pre-Finnic to Late Proto-Finnic", Finnische-Ugrische Forschungen 31.
- Viitso, Tiit-Rein (1998), "Fennic", in Abondolo, Daniel, The Uralic Languages.