Ago Künnap                
University of Tartu  
              

Possible Language Shifts in the Uralic Language Group


I am convinced that Uralic languages do not descend from one, more or less unitary proto-language, spoken about 8,000–4,000 years ago in the Uralic proto-home in West Siberia, South Urals or in the Volga area from where the speakers of that proto-language began to spread out, primarily westward, just as the traditional Uralicists usually suppose. The proto-home of the Uralic languages and the speakers of these languages could conventionally be regarded as locating in the area of the Ukraine, to the north of the Black Sea where during the Last Glacial Maximum (23,000–19,500 years ago) the eastward refuge of the European human population – the Ukrainian refuge – could have taken place. The westward refuge could have located in Iberia (Franco-Cantabria) and the middle, a relatively small one – in the Balkans.
In the cold and dry climate of the Last Glacial Maximum both humans and animals found shelter in those refuges as the latter offered tolerable climatic conditions, also for the vegetation. As suggested by Kalevi Wiik, concerning the Ukrainian refuge it is possible to conventionally observe the Uralic proto-language or rather a Finno-Ugric-type intermediary lingua franca that made the languages spoken in the refuge similar. Similarly, in the Iberian refuge a Basque type lingua franca and in the Balkan refuge a possibly Indo-European type lingua franca were operating much the same way. (See Wiik 2002.) After the Last Glacial Maximum the flora, fauna and human populations moved in the wake of the retreating icecap to ice-free areas whose humid climate fostered the plant growth, in particular. From these refuges human populations spread out in the fan-shaped form to North Europe, over the whole ice-free area (see Fig. 1).
>From the Ukrainian refuge the people spread over the area that eventually extended from the so-called North Sea Land on the site of the present-day British Isles as far as the Urals. It should be regarded as probable that the Uralic lingua franca was functioning in the area also before the Last Glacial Maximum, consequently before the people gathered in the refuges. It continued functioning in the Ukrainian refuge and, in addition to this, also in the ice-free area, creating a foundation by making languages into the present-day Uralic languages. As a result of people’s spreading out from the larger refuges the whole northward Europe was shared by the Basque and Uralic type of languages (see Fig. 2).
Genetically, the human population of Europe is regarded as Europoid and has mainly remained Europoid to date. It is also held valid for the whole European Finno-Ugric-speaking population. The Mongoloid population, arriving at the area on the Volga and its neighbourhood within the last 1,600 years from the direction of Asia, has turned part of speakers living there somewhat Mongoloid. At the same time we can observe that the speakers of Samoyed languages are strongly Mongoloid. Richard Villems’ data confirm that: in the maternal lineages (mtDNA) Nenetses inherit ca 53 %, Nganasans 88 % and Selkups 36 % from Asia (Villems 2002). Thereby, e.g., Selkups’ European-related maternal lineages have few varieties, indicating that those reached the Selkups relatively recently, giving no branches as yet. Villems also indicates that when observing the “classical set” of the Mongoloid maternal lineages, as, for instance, in the Chinese, Mongols or Kirghizes, then it is formed from the groups A, B, F and M of maternal lineages. At the same time, from the four groups practically only M has spread among Lapps, Finns, Karelians and Nenetses. He supposes that the Uralic-speaking peoples could have had contacts with the maternal lineages of the Paleosiberian settlers among whom very possibly predominated single variants of Asian M-lineages. (Villems 2002.)
Therefore it is fully conceivable that the ancestors of Samoyeds came from Asia, having originally been Mongoloid. Besides, the Mongoloid M-lineages in Nenetses (having, to a lesser degree, spread from them to the Lapps, and from them to Finns and Karelians) show that Nenetses and, as expected, other Samoyeds descend from Paleosiberians. The Samoyeds replaced their earlier, a Paleosiberian type of language (languages) by a language (languages) of a Finno-Ugric type. The language shift is testified by the Paleosiberian substratum in Samoyed languages. The substratum is so strong that, based on respective linguistic material, János Pusztay has persuasively indicated a particularly close typological connection with Paleosiberian languages (see Pusztay 1995: 11–37, 123–125; 1998). The linguists, who as expected deny the suggested language shift, opine that the postulated human population of the Uralic proto-home in West Siberia, South Urals or the Volgaic area need not have been genetically unitary. Their opinion is conceivable, principally, but the research results of the population genetics and archaeology provide not the least basis for the postulation of such an eastward proto-home and the subsequent migration, primarily westward. Considering the Ukrainian refuge as the starting point, the conception of the original Europoid and Mongoloid Uralic-speaking population fails to prove anything.
It has been known for some time already that Finnic-Lapp and Samoyed languages share a number of exceptional common features, lacking in other Finno-Ugric languages. Their evidence has been explained by a well-preserved heritage of the Uralic proto-language in the western and eastern peripheries of the Uralic language group. This explanation has no considerable foundation. First, such a theory of proto-language and periphery is no more regarded as universal in linguistics. Second, there is no reason to regard Finnic-Lapp and Samoyed languages as belonging to the western and eastern peripheries, respectively, because the linguistic area of the most widely used Samoyed language – Nenets – extends as far as the White Sea in the west. The place names, mainly, prove that formerly the area of use of the Finnic-Lapp type of languages extended to the north-eastern corner of Europe (see, e.g., Saarikivi 2000), probably as far as the Northern Urals (see Matveev 1976: 120–121; 2001; see also Fig. 3). I have noted before that there are many more exceptional Finnic-Lapp-Samoyed common features than believed heretofore and, based on that, I have supposed that in the course of their language shift, the ancestors of Samoyeds took over namely the Finnic-Lapp type of language form (see, e.g., Künnap 1996; 2000: 48–52). 
There is a special reason to observe Lapps (Saamis). In accordance with the estimation by the population geneticists’ research group headed by Villems, Lapps descend from common maternal lineages of Europe, although 5 % of their genetic heritage is related to Mongoloid (Villems 2002). I regard the 5 % share as the impact of their Samoyed Nenets neighbours because other sources of influence are practically impossible to detect. (Similarly, Villems supposes the arrival of the Mongoloid maternal lineages in the westernmost Finno-Ugrians proceeding namely along the shores of the Arctic Ocean.) Namely, in the Late Glacier the Lapps’ ancestors moved along the Norwegian western coast that was ice-free thanks to the Gulf Stream, up to the north and reached the northernmost point of Scandinavia 11,000 years ago at the latest. In Scandinavia the Lapps remained isolated from the rest of the European population behind the ice field for thousands of years, both genetically and linguistically. (See Fig. 4.)

Good examples of the consequences of the Lapps’ genetic isolation is the haplogroup V of the maternal lineages (mtDNA), originating from the Iberian refuge (see Torroni etc. 1998: 1146–1149) as well as its high concentration among the present-day Lapp-speaking population (see Fig. 5).
Accordingly, it may be concluded that whichever the language – a Uralic or non-Uralic –, brought behind the ice field by the Lapps’ ancestors, it certainly preserved well under the conditions of isolation.
Observing found exceptional Finnic-Samoyed linguistic-structural common features, lacking in Mari and Permic (see Fig. 6), it is noteworthy that, besides Mordvin languages, a great number of them are also lacking in Lapp and Ob-Ugric languages. Ignoring the possibility that the Samoyeds’ ancestors did learn a Finnic type of language (languages), two questions still remain: why do Lapp languages differ so considerably from Finnic languages and, on the other hand, Ob-Ugric languages from Samoyed languages? After all, Lapp languages are spoken in the immediate Finnic neighbourhood just as Ob-Ugric languages are spoken in the immediate Samoyed neighbourhood. In addition to this, it is customarily supposed that Finnic and Lapp languages descend from a common proto-language. The differences under observation indicate rather the variety of the Finnic and Lapp original language forms and only a partial assimilation.  It may have been caused by the fact that the language form carried by the Lapps’ ancestors behind the ice field primarily differed either more or less from the Finnic language form. After the icecap had conclusively thawed, both language forms came into contact and, based on all the factors mentioned above, it is reasonable to suppose that a language shift actually took place – the transition of the Lapps to the Finnic language form with a strong substratum of their own earlier language form, preserved in their new language. This supposition is backed up by the whole picture showing what happens if we observe the attempts to take Finnic and Lapp languages back to a common proto-language.
The “fans” of the population from the Iberian and the Ukrainian refuges, in the Lapps’ area of departure on the North Sea Land (see Fig. 2 and especially Fig. 4) towards Scandinavia, partly overlap each other (see Fig. 1). Therefore Wiik supposes some substratum of the Basque type in the then Lapp language admitting, though, that he cannot concretely show it. Such substratum is still worth looking for in the present-day Lapp languages by way of comparison with a single preserved Basque type language – namely Modern Basque. Michel Morvan has shown (Morvan 199?: 36) that already about the middle of the 19th c. Louis-Lucien Bonaparte indicated the possibility of connection  of the Basque and Lapp (and Hungarian) plural marker -k (Bonaparte 1862). The possibility of connection of the plural marker -k in Basque and Lapp substantives is noted also by Morvan himself, cf. e.g. Basque guk Euskaldunok ‘nous les Basques’ (Morvan 199?: 192–193). However, from the aspect of the Uralic language group I regard the plural marker of the Lapp substantives rather as a detached phenomenon of foreign origin – why not the Basque substratum – than some common-Uralic suffix. Besides Lapp languages the plural marker -k of substantives occurs only in Hungarian but even this is exceptional. Angela Marcantonio notes about the Hungarian plural marker -k (Marcantonio 2002: 234–235), “Unlike most U[ralic] languages, Hungarian has a different Plural ending, used both for nouns … and for verbs: the ending -k. A Plural -k is found also in Lapp, although this is generally considered as deriving from *-t … […] The origin of *-k is disputed. Some researchers believe that it derives from a derivational suffix *-kkV … This explanation looks a bit far-fetched. […] Aalto … considers the possibility of connecting -k with the Samoyed  co-affixal  element *-k(¸)- …, as well with the Tungus, Turkic and Mongolian collective ending -g.  […] A Plural -k exists also in Dravidian.”
Another example of the Basque-type substratum in Lapp languages may be the evidence of the North-Lapp word-initial semi-voiced or voiceless medial plosive stops b-, d-, g-. Namely, Morvan links the Basque word er(h)i  ‘doigt’ and the variant b¿lge of the Lapp word ‘thumb’, considering the word form *b˚er(e)xî as a proto-Basque reconstruction (Morvan 199?: 242–243). Morvan’s concrete word etymology is erroneous already from the viewpoint of incompatible consonants r – l (cf. also Rédei 1986–1991: 383), but what is interesting here is attracting attention to the Lapp word-initial consonant b- against the Basque language-historical background.
A concise but exhaustive overview of the research into the history and structure of Lapp languages (the author regards them as dialects of the Lapp language) can be found in Mikko Korhonen’s book “Johdatus lapin kielen historiaan” (An introduction into the history of the Lapp language) (Korhonen 1981). Explanations about the origin of Lapp languages can roughly be divided into three groups: 1) a hypothesis of language shift, 2) a hypothesis of loan-contact and 3) a hypothesis of the Finnic-Lapp common proto-language. The first hypothesis was fathered by K. B. Wiklund by whom originally the Lapps spoke an unknown language. Konrad Nielsen thought that the Lapps were of the Samoyed descent. This hypothesis was supported also by E. N. Setälä and Y. H. Toivonen. The latter proceeded from the words that occur only in Lapp, Ob-Ugric and Samoyed languages. The loan-contact hypothesis by no means excludes the hypothesis of language shift. Korhonen himself was in favour of the hypothesis of a common proto-language with Finnic languages but he also considered a possibility that there could be substratum of an unknown language in Lapp languages. (See ibid. 23–26.) Consequently, the concept about the Lapps’ language shift and the evidence of non-Uralic substratum in their languages is not at all new.
Leaving aside the vocabulary with its relatively changeable nature, it should be mentioned that concerning the structure of Lapp languages their phonetics differs from that of Finnic languages in the way that instead of the common proto-phonetic divergence, i.e. branching off, one feels compelled rather to think about the difficulties the Lapps faced when trying to learn Finnic phonetics: in Lapp languages it is possible to trace a sufficient number of strangely modified Finnic phonetic elements. I would like to present only a few of them. Among them there are, e.g., North-Lapp word-initial semi-voiced or voiceless medial plosive stops, cf. e.g. North-Lapp b¿sse ~ Finnish pesä ‘nest’, dollâ ~ tuli ‘fire’, giettâ ~ käsi ‘hand’. Explaining them by the influence of Norwegian and Swedish (see ibid. 127) is not convincing because no other Uralic language was influenced to that extent by foreign languages with word-initial voiced plosive stops. It seems rather a relict of the original Lapp pronunciation that has disappeared towards the south thanks to closer contacts with Finnic languages. Further, the North-Lapp word-initial è instead of Finnic s is unknown in the other Uralic languages, cf. e.g. èâlb´me ~ silmä ‘eye’. More strictly, Finnic languages do not share the gradation as, for instance, nom.sg. suovvâ : gen./acc.sg. suovâ ~ savu : savun (‘smoke’), although here it is possible to suppose a recent generalisation from a type, common to both languages (see ibid. 142–149), e.g. nom.sg. lam´pa : nom.pl. lampak ~ lamppu : lamput (‘lamp’). Tiit-Rein Viitso has drawn my attention to the fact that Finnish dialects, too, have the type nom.sg. koirra : gen.sg. koiran (‘dog’).
The plural marker -k of the Lapp substantives as an equivalent to the Finnic -t is inexplicable, cf. e.g. guolek ~ kalat ‘fishes’ (see ibid. 208–209). The supposition of the change -t > -k in this case is a phonetic-historical nonsense. The 2Psg suffix -k of the Lapp conjugation belongs here, too, thus the derivation of -k from the common with Finnic form -t (see ibid. 271) is absurd. No faith can be inspired by the vast, often multi-staged system of the derivation of the Lapp phonetic peculiarities from the Finnic-Lapp proto-language as if this were the right way to explain them. It is evident morphologically that in the regular case paradigm of substantives of the Lapp absolute declension there is no equivalent to the Finnic partitive suffix -tA (see ibid. 214–216): it may be a simplification of the declension system. There are numerous similar simplifications in the Lapp word-changing system compared to Finnic languages as shown by the following eloquent example. Namely, the regular case paradigm of the Lapp substantives does not discriminate interior and exterior local cases (ibid. 216–224), besides – depending on the language – the cases meaning ‘where?’ and ‘where from?’ (ibid. 221–224), e.g. riddost ‘in the coast, at the coast, from the coast’. (In some Finnic languages the inessive and elative are also indistinguishable but in those the phenomenon occurs in the plural, in Lapp languages in the singular; Tiit-Rein Viitso considers the inessive-elative indiscrimination as a possible Finnic-Lapp common areal feature – see Viitso 1996: 118–119. Samoyed languages do not discriminate interior and exterior local cases either and this fact may indicate the Paleosiberian substratum in Samoyed languages.)
I would separately touch upon the fact that up to now I have found 19 Finnic-Samoyed clearly observable exceptional language-structural common features, 7 of them are lacking in Lapp languages (see Fig. 6). The reduction of consonants to a laryngeal stop (common feature 3) is not generally typical of Finnic languages, occurring only in Finnish (e.g. tulet_tänne’ ‘come here!’ < tule’ tänne’ < *tulek tännek) and in Estonian South-Estonian dialect (e.g. tulõ? ‘come!’ < *tulek). Consequently it may be a relatively recent special development in Finnic languages.  Lapp languages have no double marker *i + *t (4) of nouns in the plural, co-affixes of local cases (in Finnic there are *-s-,  *-l- and in Samoyed languages *-k-) (6), the use of the object of 2P imperative in the nominative (10), 3P suffixes of the reflexive conjugation, sg. *-sen, pl.   *-set (14), the use of the suffix *m of the infinitive (supine) together with a case suffix (15) and the double suffix *p + *j of the participle (16). All these 7 are lacking also in Mordvin and Ob-Ugric languages as most of the 19 markers in general. The latter fact may show that these are so very recent Finnic features that only Samoyeds could borrow them. 
I would add a few more comments to the above mentioned 19 exceptional common features, bearing in mind the Lapp evidence, in the first place. The aim of the table of the common features mentioned (see Fig. 6) is to bring the common features forth clearly. In a number of the common features suggested the Uralicists have some information that makes them doubt as far as my viewpoints are concerned. In the case of the reduction of consonants to laryngeal stops (common feature 3) one should recall Tibor Mikola’s paper (1996) in which the author summarises his research results on the reduction of consonants in Uralic languages. Mikola shows that the consonants in Lapp, Mordvin and Khanty languages may have reduced to voiceless consonants or to an h-type sound, e.g. Lapp Karesuondo Vx3Pl ĸulléH (‘to hear’). (Mikola has connected this reduction or neutralisation in Lapp with the supposition about the origin of the Lapp plural marker -k from the form -t.) I limit myself only to the evidence of the transition to an actual laryngeal stop (’) in Uralic languages, although, the consideration of Mikola’s data would not substantially change the picture because, in addition to Finnic and Samoyed languages, there are again only Lapp, Mordvin and Ob-Ugric languages to be taken into account.
Speaking about the co-affixes of the endings of local cases (6), one should not forget Károly Rédei’s paper (1996), first of all, in which he concisely gives an excellent overview touching upon the evidence of the former in Uralic languages, presenting the language material that earlier had received no due attention. Besides Finnic languages he points out the co-affix s also in Lapp, Mordvin and Mari languages (ibid. 258–259). However, I would draw attention to the fact that in the latter there is not always any evidence particularly in ‘where?’ cases about the co-affixation of the s, e.g. in the Erza inessive pak-úa (‘field’) the reconstruction as *-s-na is only a supposition of little credibility. Roughly the same is valid also in case of the co-affix l, e.g. the Mari ablative wüt-leC (‘water’) < (!?) *-l-tä. For the clarity’s sake I should have worded the common feature (6) as follows: the  c o n s i s t e n t  use of the coaffixes of local cases. As far as the use of the *n-genitive as a case of the direct object is concerned (9), it is a common feature that needs a more profound explanation (for this explanation see Künnap 2002).  At any rate, in Finnic there is no evidence about the accusative ending *-m, although there is a definite evidence about the use of the *n-genitive as a case for the object. Similarly, in North-Samoyed languages there is some evidence about the use of the *n-genitive as a case of the object, e.g. Nganasan ñömu-n-tu tEpta ñelhoδaamu ‘he let his rabbit free, too’ (Skazki 1980: 49, 51). The use of the marker *m of the infinitive (supine) with case endings (15) may bring about contradictions as to its evidence in Uralic languages. It is possible to make an admission here but again only in case of Lapp and Mordvin languages, e.g. Lapp mon leggjim èalle-mi-n ‘I was writing’ (see Korhonen 1981: 293–294), Mordvin kando-m-s ‘to carry’ (illative?). In any case, it does not change the principal division of its evidence among Uralic languages.
Proceeding to the problems of Samoyed languages, I also pass by the vocabulary with its relatively changeable nature (however, I have touched upon the vocabulary earlier, see e.g. Künnap 2001: 179–181). I would point at Sven-Erik Soosaar’s recent comment on Samoyed languages, “If we assume, that pre-P[roto-]S[amoyed] was actually a lingua franca in the form of a pidgin spoken by various pre-historic tribes who later abandoned their original language in favour of PS in process of creolization, we should find a number of strange words (not F[inno-]U[gric]) and phonological and morphological simplification as substrate features. Indeed, PS is phonetically only a little simpler than PU, it has somewhat less consonant clusters than PFU has (PS *puEj ~ PFU *polwe ‘knee’).” (Soosaar 2000: 163). Leaving aside Soosaar’s specifications about proto-languages, it needs to be said that his remark is very precise.
Likewise, Steven Berbeco (2002) presents a few reconstructed equivalents of Proto-Finno-Ugric and Proto-Samoyed word roots which, in my opinion, reveal a certain phonetic simplification of the Samoyed inventory:
     PFU *uke((t)E) ‘path’ ~ PS *uEt ‘path, track’
     PFU *koke  ‘see’ ~ PS *ko- ‘see, find’
     PFU *mçke-  ‘sell’ ~ PS *mi- id.
     PFU *ole- ‘be’ ~ PS *åE id.
     PFU *kçle ‘tongue’ ~ PS *keEj id.
There are numerous phenomena in Samoyed phonetics that can be explained by difficulties one has to face when learning the Finnic-Lapp type of language form. One of the most conspicuous among them is s > t (typical of Ob-Ugric languages as well), cf. e.g. Finnish pesä ~ Enets pide, Selkup pitta, Kamass phidä ‘nest’. In case of Finnic-Samoyed linguistic-structural common features, the equivalents from Lapp, Mordvin and/or Ob-Ugric languages can be observed. Consequently, they are lacking in Mari and Permic languages as well as in Hungarian (see Fig. 6). (The item 20 in the table is based on Lehtisalo 1936 and needs more checking.)

Figure 6. The evidence of exceptional language-structural common features in Finnic, Lapp, Mordvin, Ob-Ugric and Samoyed languages (lacking in Mari and Permic languages)



Finnic

Lapp

Mordvin

Ob-Ugric

Samoyed

1.

Persistence of word-final vowels                               

+

+

-

-

+

2.

Gradation of consonants                                            

+

+

-

-

+

3.

Reductionof consonants into laryngeal stop 

+

-

-

-

+

4.

Plural marker *i                                                          

+

+

+

+

+

5.

Double marker  *i + *t of the plural of nouns            

+

-

-

-

+

6.

Co-affixes of the endings of local cases                      

+

-

-

-

+

7.

Consistent sequence of Cx + Px                                  

+

+

+

-

+

8.

Main word of postpositions consistently in the *n-genitive                                                                   

+

+

-

-

+

9.

Use of the *n-genitive as an object case                     

+

+

+

-

+

10.

Object of 2P imperative in the nominative                 

+

-

-

-

+

11.

Object of 3P imperative in the genitive/accusative    

+

+

-

-

+

12.

Imperative marker *k                                                  

+

+

+

-

+

13.

Double marker  *k + *m of 3P imperative                  

+

+

-

-

+

14.

3P ending sing. *-sen, plur. *-set in the reflexive conjugation                                                                 

+

-

-

-

+

15.

Marker of the infinitive (supine) *m + case ending     

+

-

-

+

+

16.

Double marker *p + *j of the participle                      

+

-

-

-

+

17.

Derivation suffix *k of pronouns                                 

+

+

+

-

+

18.

Translative marker *-kse                                              

+

+

+

-

+

19.

Coaffix *n of possessive suffixes                                

+

+

-

-

+

20.

(?) Common derivational suffix *ß, derivational suffix *n of denominal verbs, derivational suffix *t of deverbal substantives and reflexive translatives, derivational suffix *tt of deverbal reflexives.                                         

±

±

-

-

+


>From the aspect of the problems under consideration it is again very interesting what Soosaar writes, “Of course, one would still have to find out the regularities pertaining to the development of pidgins and creoles in order to prove the hypothesis that Proto-Samoyed or PU was a creole language. In the following I will compare some features, known to be widely shared by creole languages ... with the data of Samoyed languages. Most creole languages share the following grammatical features: a) lack of forms of the word ‘to be’ of the use of the so-called ‘zero-copula’. This is also the case at least in North Samoyed, where e.g. in the Nenets predicative flexion of nouns the 3Psg is the same as nom. sg form: xasawa ‘man’ or ‘is man’. […] c) preverbal negation, which is also present in Samoyed (as well as in FU). d) a simple form of the verb refers to whatever time in focus, which is not the case in present Samoyed languages but could be in PS. e) each creole tends to have three verbal markers: 1) anterior tense marker, 2) irrealis mood marker, 3) non-punctual aspect marker. For PS two tense markers are reconstructed: an aorist *-ng(a)- and preteritum (anterior) *-sa- … and 3 moods: indicative with zero suffix,  imperative-adhortative  with  several  suffixes  (-k, -j, -mt) and conjunctive with *-ni. If variants of imperative are left aside, there are not too many verbal markers for a creole language.” (Soosaar 2000: 162–163).
It should also be mentioned that in accordance with Eugen Helimski the metrics of North-Samoyed Nenets and Nganasan shaman songs is similar to the metrics of the Finnic Kalevala-song – which is octametre (see Helimski 1998: 44–45; Helimskij 2000).
Based on everything said above I think there is sufficient linguistic evidence to make it possible to assume that Samoyeds learnt their present-day language from the speakers of the Finnic-Lapp languages or a language of a similar type. The time and place of the Samoyed language shift remain, however, open-ended. But it may have happened thousands of years ago and somewhere in the northeastern part of Europe or its neighbourhood where Nenetses are still residing. If the Lapps’ ancestors replaced their earlier language by a Finnic type of a language, then, as shown above, it could probably have taken place earlier than the Samoyeds’ language shift.
One may wonder why there are only a couple of suffixal case forms in regular case paradigms of some Ob-Ugric dialects, while in other dialects as well as Hungarian the paradigm is built up mainly as a result of grammaticalisation of postpositions, i.e., with recent suffixal case forms, developed from postpositional constructions. The declension paradigm with suffixal case forms in Ugric languages seems to have been reduced at some stage and replaced by postpositional constructions. It reminds of the consequence of pidginisation. For comparison it can be emphasised that under the influence of languages in contact the word-changing system has been simplified, for instance, in English, Swedish and partly in French. Wiik has supposed that Ob-Ugric languages emerged so that Samoyeds learned the Finno-Ugric language form (Wiik 2002). The supposition is in keeping with the fact that the speakers of Ob-Ugric languages are genetically strongly Mongoloid – 37 % of Mansi and 31 % of Khanty maternal lineages (mtDNA) come from Asia (Villems 2002) –, although not to the extent as Samoyeds. Thus Ob-Ugric languages could contain some kind of substrate and the pidginisation of the Finno-Ugric language form. Likewise, the pidginisation includes also Hungarian whose speakers are not the least Mongoloid. We can only suppose that the Hungarians’ ancestors, in turn, learned the Ob-Ugric language form that was constituted in the way described above.
The migration wave of the Turkic-speaking Mongoloid population from Asia to Europe carried with it the Ob-Ugric-speaking ancestors of Hungarians. Somewhere in the region of Magna Hungaria or after leaving it for the western areas, the language of Europoid Hungarians influenced that of the ancestors of Europoid Mordvins since the Mordvins were supposedly the southernmost people among the Volgaic and Permic peoples who spoke Finno-Ugric (see Fig. 7).
This is how the Mordvin language-structural common features could be explained as regards Ugric and Samoyed languages. Pusztay supposes even that perhaps Mordvins moved westward from the east together with Hungarians (Pusztay 2000). I consider the “influence of by-passing Hungarians” on Mordvins more probable because, apart from Hungarians, there are no distinct traces available to the effect that Mordvins moved together with Turkic-speaking nomads.
It should also be noticed what Angela Marcantonio is writing about the Uralic cases (2002: 212): “The Case endings (and/or related postpositions) formed through Grammaticalisation are present in almost all the U[ralic] languages. […] However, they are particularly well attested, transparent in structure and clearly observable in Vogul …, some Samoyed languages, Old Hungarian, and Ostyak. The chronology of the formation of postpositions and related Case endings in the U languages is quite ‘late’ (to use the Korhonen’s definition … In other words, these Case endings were not formed sometime in the unrecorded past.”
Actually, the whole picture is more complicated because, first of all, based on the place names, the region of Mansis’ settlement has been supposed to occupy a rather vast area to the west of the Urals (almost as far as Archangelsk and Kazan). In more recent times the extension of the Mansis’ settlement area seems to be confirmed also by historical records. Z. P. Sokolova’s map (see Figure 8) presents a compendious picture of the area. Can we similarly suppose the Mansis’ itinerary from the east to the west (along with Hungarians?) at the very beginning? In any case, this option cannot be excluded.
It should be regarded even as more probable that the Hungarians’ ancestors learned the language of their northward neighbours Mansis in Magna Hungaria. Both Mansis and Hungarians had settled in Europe at that time and were Europoids. Only after a later movement of Mansis away to Asia (to West Siberia) could partly turn them Mongoloid there. Bearing in mind the Mansis’ ancestors onetime residence in the west of the Urals, they could with a greater probability than the Hungarians’ ancestors bring about peculiarities in Mordvin languages. Perhaps even so that the Mordvins’ ancestors were first the speakers of Mansi who later transferred to the Finnic type of language form, preserving the Mansi morphosyntactic substratum in Mordvin languages. Eventually, there is a possibility that Hungarians were first the southward Mansi-speakers who lived in Magna Hungaria. It should also be kept in mind that all the modern Ugric languages testify to the pidginisation process, first of all in the form of the reduction of the word-changing system of substantives. If this was caused by  Samoyeds, one can suppose that the pidginised Ugric language form spread from West Siberia to the western side of the Urals as well. Proceeding from the supposed course of affairs, Ob-Ugric languages were finally formed only in West Siberia in the way suggested by Wiik: the Samoyeds’ ancestors (or part of them) who had so far spoken a Paleosiberian language form, learned the Finno-Ugric language form, spoken then by the ancestors of Mansis and Khanties. Through contacts in West Siberia besides the language shift a partial Mongoloidisation of Ob-Ugrians as well as a partial Europoidisation of Samoyeds took place, continuing later under the influence of Finnic- and Lapp-speakers of Northwestern Europe.
There is another mystery concerning the Ob-Ugrians. Namely, Villems draws attention to the fact that the Khanties and Mansis as well as Yenisey-speaking Kets with a considerable frequency display such maternal lineages that have not yet been found in other Siberian peoples or in Central Asia and Europe. The nearest peoples who have such maternal lineages reside at a distance of 3,000 km in Turkey, Iran, Jordan and Lebanon. Thereby it is undoubtedly possible to trace the migration of concrete women although their direction of movement is not yet known. (Villems 2002.) Considering the Hungarians’ migration one cannot ignore the possibility of the movement of Ob-Ugrians together with them. The Ob-Ugrians, however, reached more southward areas in comparison with the migration route of Hungarians. (Another problem is connected with the Kets, presenting various explanations as their langauge and genes. At least the supposition about the southward origin of all Yenisey languages should not be overlooked.)
I will try to range the order of the supposed language shifts in the Uralic language group. 1) First of all Lapps took over the Finnic language form; 2) Mordvins took over the Mansi language or primarily spoke Mansi; 3) (part of)  Samoyeds took over the Ugric language form (Wiik); 4) the Mansi-speaking Mordvins and (Ugric-speaking?) Samoyeds took over the Finnic-Lapp language form. Lapps may have preserved some substratum from some other (Basque type?) language, Ugrians – some substratum from a Paleosiberian type (from Samoyeds?), Samoyeds – some substratum from a Paleosiberian (partly also Ugric?) type, Mordvins – some substratum from an Ugric (Mansi?) type of language. I would underline that such a picture can be sketched thanks to the following more or less certain data about the populations’ residence as well as movement: 1) the Lapps’ ancestors’ itinerary from the onetime North Sea Land behind the ice field as far as the northernmost point of Fenno-Scandia and in the wake of the thawing icecap to the southern areas of Fenno-Scandia, 2) the itinerary of the ancestors of the Finnic populations northward after the thawing ice until meeting the Lapps, 3) the residence of the ancestors of Finnic and Lapp populations also in Northeastern Europe, 4) the residence of the Ugrians’ (first of all Mansis’ and Hungarians’) ancestors in an expansive area to the west of the Urals, 5) the residence of the Mordvins’ ancestors in the neighbourhood of the Ugrians, 6) the residence of the Ob-Ugrians’ ancestors in the neighbourhood of the Samoyeds’ ancestors in West Siberia, 7) the itinerary of the Samoyeds’ ancestors to the neighbourhood of the Finnic and Lapp populations in the region of the North Urals or further westward, to Northeastern Europe. Here I would rather keep back from temporal determinations as there are only a few fixed points for them, such as the time of the thawing of the icecap and the time of the departure of (part of) Hungarians from Magna Hungaria.
Against the background of my suppositions it is interesting to observe János Pusztay’s most recent viewpoint about this treatment, worded by him in his private letter as follows (translated from Estonian): “Barely based on a linguistic analysis – leaving out anthropological, genetic and archaeological considerations – I came to the following conclusions:
1) Mordvins cannot be original Finno-Ugrians; in their languages we find many such features that do not occur in the neighbouring Finno-Ugric and Turkic languages, however, they do occur in (Ob-)Ugric and Samoyed as well as in Paleosiberian languages (the truth is that in Mordvin languages there are numerous Finnic elements but those result from more recent intensive contacts);
2) Samoyeds are original Paleosiberians, their language structure displays various features which occur only in Paleosiberian languages (and partly also in (Ob-)Ugric and Mordvin languages); I believe that Samoyeds came into intensive contacts with genuine Finno-Ugrians (Permic, Mari, Finnic peoples) as a result of external pressure (e.g., the immigration of the present-day Yeniseyans about 8,000–10,000 years ago);
3) Ugrians are an transitional group between genuine Finno-Ugrians and Paleosiberians (e.g., Samoyeds) but they are closer related to Samoyeds than Finno-Ugrians; their vocabulary is closer to Finno-Ugrians than that of the Samoyeds’ but from the grammatical point of view they, too, are Siberians; […]
4) the language structure of Ugrians, Mordvians and Samoyeds resembles to that of Paleosiberians; on the other hand, Altaic languages are similar rather to other Finno-Ugric languages.
In accordance with my suppositions that 10,000–20,000 years ago the western area of the North-Eurasian belt was inhabited by genuine Finno-Ugrians (the predecessors of modern Finnic, Mari, Permic peoples), the eastern area of the mountain range of the Urals was inhabited by various Paleosiberians (the predecessors of modern Ugrians, Mordvins, Samoyeds belonged to their northernmost group); about 8,000–10,000 years ago the predecessors of the present-day Yeniseyans, a linguistically completely different group of people, invaded the basin of the Yenisey River and separated the western group of Paleosiberians (i.e. the present-day Mordvins, Ugrians and Samoyeds) from other Paleosiberians. Thus the western group sought for closer contacts with genuine Finno-Ugrians. When living among Finno-Ugrians it was probable that Maris and Permians could have some contacts as with the Samoyed-Mordvin-Ugric group as well as with the present-day Finnic people.”
My suppositions differ from those presented by Pusztay to a certain extent because I try to take into account the state of art of the population-genetic research results.

References


Berbeco, S. 2002, Comments on the Shape of Proto-Uralic Forms. – P. Klesment (ed.), Fenno-Ugristica 24. Keelekontaktidest keelevahetuseni. From Language Contacts to Language Shifts, Tartu, pp. 64–76.
Bonaparte, L.-L. 1862, Langues basques et langues finnoises, Londres.
Helimski, E. 1998, Samojedit ja ðamanismi. Viisi luentoa samojedeista, ðamanismista ja uralilaisesta kulttuurista (Samoyeds and shamanism. Five lectures on Samoyeds, shamanism and Uralic culture), Tampere.
Helimskij, E. A. 2000, Komparativistika, uralistika. Lekcii i stat’i (Comparativistics, Uralistics. Lectures and articles), Moskva.
Janhunen, J. 1989, Samojedin predestinatiivisen deklinaation alkuperästä (On the origin of the Samoyed predestinative declension). – Journal de la Société Finno-Ougrienne 82, pp. 298–301.
Korhonen, M. 1981, Johdatus lapin kielen historiaan (A introduction into the history of the Lapp language) (= Suomalaisen Kirjallisuuden Seuran toimituksia 370), Helsinki.
Künnap, A. 1996, On Some Similarities Between Finnic-Lapp and Samoyed Languages. – A. Künnap (ed.), Fenno-Ugristica 20, Tartu, pp. 65–72.
Künnap, A. 2000, Contact-induced Perspectives in Uralic Linguistics (= LINCOM Studies in Asian Linguistics 39) [München-Newcastle].
Künnap, A. 2001, Keneltä samojedit ovat oppineet kielensä? (From whom did Samoyeds learn their language?). – R. Kasik (ed.), Keele kannul. Pühendusteos Mati Erelti 60. sünnipäevaks 12. märtsil 2001 (= Publications of the Department of Estonian of the University of Tartu 17), Tartu, pp. 177–185.
Künnap, A. 2002, Uralic Accusatival and Genitival Direct Object. (Forthcoming.)
Larsson, L.-G. (ed.) 1996, Lapponica und Uralica. 100 Jahre finnisch-ugrischer Unterricht an der Universität Uppsala. Vorträge am Jubileumssymposium 20.–23. April 1994 (= Studia Uralica Upsaliensia 26), Uppsala.
Lehtisalo, T. 1936, Über die primären ururalische ableitungssuffixe (= Mémoires de la Société Finno-Ougrienne 72), Helsinki..
Marcantonio, A. 2002, The Uralic Language Family. Facts, Myths and Statistics. (= Publications of the Philological Society 35), Oxford UK & Boston USA.
Matveev, A. K. 1976, K probleme domansijskogo substrata v gidronimii (About the problem of the pre-Mansi substratum in hydronymy). – Proishoþdenie aborigenov Sibiri i ih jazykov, Tomsk, pp. 118–121.
Matveev, A. K. 2001, Substratnaja toponimija Severa Rossii (Substratal toponymy of Northern Russia) I, Ekaterinburg.
Mikola, T. 1996, Zur Geschichte der auslautenden Tenues im Lappischen. – Larsson 1996, pp. 99–102.
Morvan, M. 199?, Les origines linguistiques du basque, Presses Universitaires de Bordeaux.
Pusztay, J. 1995, Diskussionsbeiträge zur Grundsprachenforschung. (Beispiel: das Protouralische) (= Veröffentlichungen der Societas Uralo-Altaica 43), Wiesbaden.
Pusztay, J. 1998, Zur Entstehung des sprachlichen Bildens Nord-Eurasiens. – K. Julku, K. Wiik (eds.), The Roots of Peoples and Languages of Northern Eurasia I. Turku 30.5.–1.6.1997, Turku, pp. 161–167.
Pusztay, J. 2000, A “megzagkított egyensúly” elmélete és az uráli alapnyelv kialalakulása (The theory of “punctuated equilibrium” and the origin of Proto-Uralic). – Folkloristika 2000-ben. Folklór – irodalom – szemiotika. Tanulmányok Voigt Vilmos 60. születésnapjára II, Budapest, pp. 550–556.
Rédei, K. 1986-1991, Uralisches Etymologisches Wörterbuch, Budapest.
Rédei, K. 1996, Zu der Geschichte des PU-PFU Kasussystems. Die Rolle der Koaffixe in der Herausbildung der Deklination. – Larsson 1996, pp. 257–271.
Saarikivi, J. 2000, Kontaktilähtöinen kielenmuutos, substraatti ja substraattinimistö (Contact-induced language change, substratum and substratal place names). – Virittäjä, pp. 393–415.
Skazki narodov Sibirskogo Severa (North-Siberian peoples’ fairy tales) 1980, Tomsk.
Soosaar, S.-E. 2000, The Relation of Samoyed Languages to the Neighbouring Languages. – Études Finno-Ougriennes 31, pp. 157–166.
Torroni, A., H.-J. Bandelt, L. D’Urbano, P. Lahermo, P. Moral, D. Sellitto, C. Rengo, P. Forster, M.-L. Savontaus, B. Bonné-Tamir, R. Scozzari 1988, mtDNA Analysis Reveals a Major Late Paleolithic Expansion from Southwestern to Northeastern Europe. – American Journal of Human Genetics 62, pp. 1137–1152.
Viitso, T.-R. 1996, Ostseefinnisch und Lappisch. – Larsson 1996, pp. 113–121.
Villems, R. 2002, Uurali keelkonna rahvaste geenitiigist (On the genetic pool of the peoples of the Uralic language family). – Õpetatud Eesti Seltsi Aastaraamat, Tartu. (Forthcoming.)
Wiik, K. 2002, Eurooppalaisten juuret (The roots of Europeans), Jyväskylä. (Forthcoming.)