{1. Ostiak.
A. Ugrian. {2. Vogul.
{3. Magyar.
{i. Permian { Syrjenians, Permians, Votiaks.
{
{ {_a._ Volga group {Cheremissian, Mordvinian.
B. Finnish. { {
{ii. Volga { {1. Lappish.
{ Baltic {_b._ West Finnish {2. Esthonian, Livish,
{ { { Votish, Vepsish.
{ {3. Finnish.
All these languages have a certain common vocabulary, and a common
grammatical substratum, though many of them possess constructions unknown
to the others. The pronouns, numerals, and in a less obvious degree the
pronominal affixes of nouns and verbs are also identical. The Ugrian
languages however, seem to have parted company with the rest before a
system of declension had been fixed. Their nominal suffixes seem to be
mostly later formations, though we find =t=, =tl=, or =k= for the plural,
and traces of =l= as a local element. Corresponding phenomena appear in
the conjugation of verbs, as noticed above.
The remaining languages—or Finnish group—have not developed any striking
differences from the Ugric division, but they show greater resemblance
to one another in details. They all have local cases characterised
by the letter =s= (unknown in the Ugric group), others characterised
by =l=, an abessive ending in =ta= or =tak=, and negative adjectives
characterised by the syllable =tem=, or =tom=. They mostly agree in
having a peculiar form for the negative conjugation. The present of
the positive conjugation has =p= (or =v=) regularly or sporadically in
certain persons, and the remaining verbal forms, though far from agreeing
absolutely, show a sufficient resemblance to warrant us in regarding them
as the results of a common development.
From the accounts given of Syrjenian and Votiak it would seem that they
were the first to cease to participate in this common development. They
appear to be characterised by few striking peculiarities, but to show
a less degree of conformity to a common standard than the remaining
languages. The phonetic system of Syrjenian seems to be much the same as
that of Cheremissian or Mordvinian, except that it has a great fondness
for the sound of =i̱=. The pronouns of both numbers, and the pronominal
affixes of the singular, show much more resemblance to the Finnish
than to the Ugrian, but the pronominal affixes of the plural (=ni̱m=,
=ni̱d=, =ni̱s=) are curious, and obviously represent the singular
affixes in combination with a syllable =ni̱=, which may perhaps be akin
to the Finnish plural demonstrative =ne=. In the verb, the 1st person
singular has no personal termination. In the plural we find =m=, =ni̱d=,
=ni̱s=, as in nouns. Generally Syrjenian seems to be more thoroughly
agglutinative, as opposed to inflected, than the Finnish language. In
this it may be compared with Cheremissian, where the plural is formed
by the syllable =vlja= (or =vi̱lä=) added between the stem and case
termination, just like =jas= in Syrjenian. The Permian languages have
some close analogies in detail with the Ugrian group. Thus twenty is
=ki̱s= in Syrjenian, =kōs= or =χūs= in Ostiak, =husz= in Magyar; the
reflexive pronoun is =as= or =ats= in Syrjenian and Votiak, =at= in
Ostiak.
Of the remaining languages Mordvinian is in many ways the most
remarkable. Though generally admitted to be nearly akin to Cheremissian,
it has many constructions peculiar to itself. Thus it has a fully
developed object conjugation and two forms of declension, the definite
and indefinite. It has a great fondness for the letter =f=. On the
other hand, such phenomena as the suffix =n-za= for the 3rd person, the
word =kemen=, _ten_, the regular formation of the imperfect with =i=,
infinitives in =ma=, and participles in =f= (Finn. =va=) show a near
approach to Finnish. Cheremissian has also developed some new singular
negative forms, by which the verbal root is negatived (in the preterite)
by a suffix =te + l=, after which the personal suffixes are added.
The connection of the West Finnish languages is much more striking
than that of those which have already been discussed. The resemblance
of Lappish to Finnish and Esthonian is a little disguised owing to the
elaborate phonetic system of this language, which has an extraordinary
fondness for diphthongs, and also a very extensive provision of
consonants, including some (e.g. =γ=, =θ=, =δ=) which do not occur in any
of the cognate languages. The nominal declension is very similar to that
of Finnish. Compare the singular forms =tšalme=, =tšalmen=, =tšalmesn=,
=tšalmest=, =tšalmetaka= with =silmä=, =silmänä=, =silmässä=, =silmästä=,
=silmättä=. In the singular the genitive and partitive have lost their
termination (cf. the Esthonian forms, nom. =silm=, gen. and part.
=silma=). The analogy in the plural is still closer; the nominative ends
in =h= or =k=, but the other cases are characterised by the insertion of
=i=, =tšalmeh= or =tšalmek=, =tšalmiti=, =tšalmi=, =tšalmin=, =tšalmisne=,
=tšalmist=, =tšalmitaγa= corresponding to =silmät=, =silmiä=, =sílmien=,
=silminä=, =silmissä=, =silmistä=, =silmittä=. There are also forms in
=l=, =le=, =lt=, or =ld=. The partitive case appears to be peculiar to
the West Finnish languages. The pronouns of Lappish point the other way,
and are all but identical with those of Mordvinian, =mon=, =ton= or
=don=, =son=, =mi=, =ti= or =di=, =si=; Mordvinian =mon=, =son=, =ton=,
=min=, =sin=, =tin=. The singular affixes are =m=, =d=, =s= (Mordv. =n=,
=t=, =nza=), but those of the plural take the characteristic =k= (=mek=,
=dek=, =sek=). There are also dual forms. The verb is peculiar in using
different affixes for the present and preterite. Otherwise it is much the
same as the Finnish forms. We have =b= or =p= as a sign of the present,
=je= or =i= for the preterite, =ket=, =kus=, etc. in the imperative;
comparatives formed with =fč-=, =kč-=, =č-= (=ksi=), or =le=. The
negative conjugation also shows close analogies. Lappish has thus a great
resemblance to Finnish, but is much nearer than the remaining western
languages to Cheremissian and Mordvinian, thus connecting these latter
with Finnish and the Esthonian languages.
There is no need to insist on the close connection of Livish, Votish,
Vepsish, and Esthonian with Finnish. The grammatical structure and
vocabulary of these languages is so alike that the fact is obvious not
only to a philologist, but to the most casual learner. Esthonian is now
far the most important of these languages, and has a certain amount of
literary culture. It has two chief dialects, that of Revel and that of
Dorpat[3].
In a few cases Esthonian shows older forms than Finnish, but on the
whole it is less primitive. Besides the Finnish consonants it possesses
=b=, =g=, =z=, and =ṅ=. The accent is on the first syllable, and has led
to weakening or loss of final syllables. The vowel harmony is known only
in the Dorpat dialect, and there not perfectly. Consonants are weakened
in much the same way as in Finnish, =pp=, =tt=, =kk= are reduced to
single consonants; =p=, =t=, =k=, =s= become =b=, =d=, =g=, =z=; =b=
becomes =w= or disappears, and =d=, =g=, =z= disappear. These changes,
however, do not in the present state of the language take place only when
syllables are closed, although the original form generally ended with a
consonant. Thus =tīb=, =urk=, =rind=, =toit=, =rid= form the genitives
=tīwa=, =urga=, =rinna=, =toidu=, =riu=; and the verbs =pühkima=,
=uskuma=, =prūkima=, =hoidma= form =pühin=, =uzun=, =prūgin=, =hoian=.
The noun has only eleven cases, the six local (exterior and interior)
the genitive, partitive, abessive, and translative, all almost identical
in form with Finnish, except that the genitive has lost the =n= of the
termination, and the other cases the final vowel. The plural takes =d= in
the nominative and =i= in the other cases. The pronouns resemble Finnish,
but the 3rd person singular and plural is the demonstrative =tema= or
=nema= in Esthonian, Livonian, and Votish, but Vepsish has =hän=, =hö=.
Esthonian has almost lost the pronominal affixes, which are used only
in adverbial forms, and replaced in most cases by the genitive of the
personal pronoun.
The affirmative verb closely resembles Finnish, except that the
concessive formed with =ne= has, except in the Dorpat dialect, been
almost entirely lost. The 3rd person singular present ends regularly in
=b=, and the 3rd plural in =wad=. The preterite is sometimes formed with
simple =i=, but generally with =si=, and the conditional with =ksi=. The
imperative 2nd singular has no termination, but as in Finnish the root
is weakened. The other persons are characterised by =gu= or =ge=. The
passive is formed by affixing =ta + k + se= in the present, and =ta + i=
(=ti=) in the preterite. The negative verb for the imperative is =ärä=,
=ärgu=, =ärge=, but in the other forms Esthonian does not affix personal
endings to the negative, but uses =ei= with all persons. Similarly
Livish has =äb= or =ab= for all persons except the second, where =ad=
is sometimes used. Votish and Vepsish follow Finnish. It will thus be
seen that Esthonian, closely allied as it is to Finnish, has lost many
peculiarities which it once no doubt possessed, but occasionally (e.g.
the passive present in =takse= and the conditional in =kse=) preserves
forms which in Finnish have been weakened or disguised.
It is easier to discuss the relations of the Finno-Ugric languages
to one another than to decide what are their affinities with other
groups. They are generally considered to be connected with the Samoyede,
Turkish, Mandchu, and Mongolian languages, that is to say, the ancestor
of each of these groups was related to the ancestor of the Finno-Ugrian
languages. When, however, identity of vocabulary cannot be proved, it
is dangerous to make comparisons on the ground of general grammatical
resemblances, because the grammar of agglutinative languages offers few
striking peculiarities, and represents a stage of development through
which may other languages, certainly the Aryan, have passed. The only
general description which can be given of the Finno-Ugric group is that
they are languages without gender, whose grammatical structure consists
entirely in appending suffixes. To these characteristics is generally
added another, the vowel harmony, but this exists very partially in the
Finnish group. Now without denying the possible relationship of Turkish,
Mongol, and Mandchu to Finnish, it must be admitted that they have only a
very general resemblance, and very many and precise differences. Mandchu
and Mongol, with their uninflected verbs, would have been put into quite
another class were they not undoubtedly akin to languages with a more
developed system. And why should Japanese be excluded? It presents no
phenomena incongruous with the grammar of the languages above cited, and
the want of vowel harmony cannot be alleged as a difficulty.
If, however, we turn to the Samoyede languages the case is very
different. They are usually mentioned as if they stood no nearer to the
Finno-Ugric group than Turkish or Mandchu, and yet the resemblances in
detail are numerous and striking.
They possess the vowel harmony, and apparently a law for the weakening
of consonants analogous to that of Finnish and Esthonian, e.g. =kinta=,
=kindan=; =mat=, =maden=; =sok=, =sogon=. The noun is strikingly like
Finnish. We find a genitive in =n=, an accusative in =m= or =p= (cf.
Lappish), a dative in =ni= or =n=, a locative in =nan= (cf. Finnish
=na=), and an instrumental in =se=. There are three numbers, the dual is
characterised by =g= or =hā= (cf. gen. in Ostiak, =ag= in Vogul), and the
plural is formed with =la= or =t= (=d=). The pronouns show a good deal of
variety, and it is evident that some forms (e.g. =pudar=, _thou_; =puda=,
_he_) are not real pronouns but substantives used as pronouns. Still we
find a general consensus for =man= as the 1st person singular, and some
form of the same word for the plural; =tan= for the second singular,
and =te=, =ten=, or =ši= for the plural (cf. Lappish, Cheremissian,
Syrjenian, and Mordvinian). The 3rd person exists in several forms,
=tep=, =sete=, =di=, etc., which may perhaps be compared with the Ostiak
=ten=, Cheremissian =tidä=, and Syrjenian =si̱a=. The pronominal affixes
are also extraordinarily numerous and varied, but they seem to include
=m= for the 1st singular, =t= or =d= for the 2nd, =jea= (cf. Magyar =ja=)
for the 3rd; =mu= or =met= for the 1st plural, and =ta= or =tet= for the
2nd. The interrogative pronouns are =kutö=, =hübea=, =hoke=, etc., _who_,
and =ma=, =mi=, _what_. The verb and noun are not sharply distinguished.
The verb takes two sets of pronominal suffixes, the predicative, with
which an intransitive verb is always conjugated, and the possessive,
which are used with both transitive and intransitive verbs. The
preterite is formed by adding =s=, which occurs as a preterite suffix in
Cheremissian, Mordvinian, Vogul, and Ostiak. The conjunctive is formed
with =nji=, =ne=, =na= (cf. the Finnish =ne=). The imperative is formed
with the syllable =kar=, =gar=, =har= (cf. Finnish =ka=). There is also
an optative with the suffix =rava=, so perhaps =kar= is =ka-ra=.
Besides this similarity of grammatical forms there is a large common
vocabulary. The following examples are taken at haphazard out of
Castren’s lexicon:—_River_, =jaha=, =joha= (Finn. =joki=); _hill_,
=pirda= (Finn. =vuori=); _dark_, =paebi=, =paevuda= (=pimeä=); _to blow_,
=pu’u= (=puhua=); _half_, =pealea=, =fealla= (=puoli=); _good_, =sava=,
=sova= (=hyvä=); _fire_, =tu= (=tuli=); _fish_, =kole=, =hale= (=kala=);
_tree_, =po=, =pe=, =pea= (=puu=); _bear_, =korg=, =kuerg= (=karhu=);
_earth_, =mon= (=maa=); _narrow_, _small_, =tîjea= (=tyhjä=, _empty_);
_to place_, =puenan= (=panna=); _live_, =jileadm= (=eleä=); _come_,
=tû’am=, =töak= (=tulla=)[4].
On the other hand, the Samoyede languages differ in many ways from
the Ugro-Altaic group. The numbers are entirely different. The Ostiak
Samoyede =ōker=, _one_, shows a faint resemblance to =yhte=, and
=sidea=, or =sede=, _two_, has been compared with =kahte=, but such
analogies are doubtful. Only the number _seven_, =sin=, =sjelde=,
=sjaibua= shows a resemblance with the Finno-Ugric forms. But it is
clear that the Samoyede numbers represent very primitive attempts at
numeration (e.g. Jurak. =hâsava-ju=, _Samoyede big number_, for _nine_,
=lutsa-ju=, _Russian big number_ for _ten_), and that in many cases
Turkish numbers have been borrowed (cf. =tjet=, =tet=, _four_, with Yakut
=tüört= and =kamass=, =khera= 40, =iliχ= 50, =althon= 60, with Turkish
=kirk=, =elli=, =altmish=). Samoyede is also more like the Turkish than
the Finno-Ugric language in its power of adding predicative and temporal
suffixes to nouns (which implies a want of distinction between the verb
and noun). Thus =lūtsa= means _a Russian_; =lūtsam=, _I am a Russian_;
=lūtsamsʻ=, _I was a Russian_.
The above sketch of Samoyede has no pretence to be exhaustive, and
may be charged with inaccuracy, inasmuch as words and forms are cited
indifferently from all the dialects. For a proper investigation of the
question it is no doubt necessary to thoroughly study the relations of
the Samoyede languages to one another, to establish the original forms,
and in particular to determine the influence of foreign languages,
whether Turkish or Finno-Ugric, on the Samoyede vocabulary. But unless
there is something strangely misleading in the superficial character of
these dialects, it appears to me that they undoubtedly stand far nearer
to Finno-Ugric than do Turkish or Mongolian, and should indeed be classed
as outlying members of the Finno-Ugric group. The want of similarity in
the numbers is certainly very strange, but the other resemblances in
vocabulary seem to me to be conclusive, unless all the words in question
are borrowed.