CC BY-NC-ND
62
Kirill Reshetnikov
Institute of Linguistics of the Russian Academy of Sciences (Moscow)
Mikhail Zhivlov
Russian State University for the Humanities (Moscow)
Studies in Uralic vocalism II:
Reflexes of Proto-Uralic in Samoyed, Mansi and Permic1
The paper presents a set of rules that explain the complementary distribution of the reflexes of alleged Proto-Uralic vowels *i and *a in Proto-Samoyed, Proto-Mansi and Proto-Permic. Our aim is to show that the reconstruction of one vowel *a instead of these two phonemes is sufficient for the explanation of all the reflexes in the languages under consideration.
Keywords: Uralic languages, Finno-Ugric languages, Samoyed languages, Mansi language, Komi language, Udmurt language, comparative phonology, linguistic reconstruction, Proto-Uralic vocalism.
Most present day works dealing with historical phonology and etymology of Uralic languages accept that in the first syllable of some Proto-Uralic lexical stems there must have been a high unrounded back vowel. This phoneme, usually marked as *i and sometimes as *i, is absent in most early variants of the Proto-Uralic reconstruction as well as in the reconstruction system tentatively used in UEW, but is present in works by Juha Janhunen, Pekka Sammallahti and their followers [Sammallahti 1979; Janhunen 1981; Sammallahti 1988; Aikio 2002; Aikio 2006]. A version of Finno-Ugric historical phonology giving the same vowel the status of a proto-language phoneme was suggested already by Wolfgang Steinitz [Stei-nitz 1944].
The reason why the vowel system of Proto-Uralic was enlarged by *i is that Balto-Finnic *a of the first syllable together with its counterparts in Mordvinic, Saami and Mari has a variety of correspondences in other Uralic languages, in particular in Samoyed.
As demonstrated in [Janhunen 1981], Proto-Samoyed counterparts of the West Uralic *a include, besides *a, also *i and *e. Considering this fact, Janhunen suggests that Samoyed *a, corresponding to Finno-Permic *a, goes back to Proto-Uralic *a and Samoyed *i (as well as *e), as a counterpart of Finno-Permic *a, reflects Proto-Uralic *i. Thus the original system distinguishing *a and *i would have been preserved in Samoyed, while in Finno-Permic the two phonemes in question would have merged in *a (although some traces of Proto-Uralic *i would have remained in Ugric) [Janhunen 1981: 227-229, 233-234, 236-237].
In works by Pekka Sammallahti, particularly in his fundamental article, dedicated to the historical phonology of Uralic languages, evidence for Proto-Uralic *i (*i in Sammallahti's notation) is found, besides Samoyed and Ugric, also in Permic. According to Sammallahti (as can be inferred from [Sammallahti 1988], although the phonetic rules assumed by the author are not given an explicit wording), the development of Proto-Finno-Ugric *a, *i, *oo (< PU *ax) and *ii (< PU *ix) in Ugric and Permic can be formulated as follows.
1 This study was carried out with the support of Russian Foundation for the Humanities, project № 09-04-95202a/F.
Journal of Language Relationship • Вопросы языкового родства • 5 (2011) • Pp. 96-109 • © Reshetnikov K., Zhivlov M., 2011
In Hungarian, PFU *i and *ii in PFU ¿-stems yield i or i, whereas PFU *a and *oo in this type of stems yield a or a. In PFU a-stems both *i and *a yield a or a (after a labial PFU *a yields o or a in both i- and a-stems) [Sammallahti 1988: 500-501, 513-515].
In Ob-Ugric languages PFU *i and *ii are always reflected as Proto-Mansi *ii, Proto-Khanty *aa. PFU *a in *a-stems yields Proto-Mansi *uu, Proto-Khanty *aa, while PFU *a and *oo in ¿-stems yield Proto-Mansi *aa, Proto-Khanty *oo (after a labial PFU *a gives Proto-Mansi *uu or *u, Proto-Khanty *uu in both i- and a-stems) [Sammallahti 1988: 500-507].
In Proto-Permic PFU *i yields *u (PP *u in our reconstruction2, see [Zhivlov 2010]), and PFU *a yields *u (PP *o in our reconstruction) irrespective of the second-syllable vowels. PFU *ii and *oo merge and yield PP *u (*5 in our reconstruction) or *i (*e in our reconstruction), but PP *wu (*wo in our reconstruction) in the initial position [Sammallahti 1988: 522-531].
Thus, the distinction of PFU *i and *a is best of all preserved in Proto-Mansi and Proto-Permic. In the present article we will try to show that precisely in these languages, as well as in Proto-Samoyed, the reflexes of alleged *i and *a are in complementary distribution.
Before we proceed to the examination of the relevant evidence, a few preliminary remarks are in order.
Following Meri Lehtinen's ideas [Lehtinen 1967], we do not reconstruct Proto-Uralic long vowels (or combinations of vowel + *x that have replaced long vowels in Janhunen's and Sammallahti's reconstruction). Proto-Balto-Finnic *o is, in our view, the result of regular lengthening of *a in Balto-Finnic e-stems before intervocalic *r, *l, *m, *n and *6. The only clear exception to this rule is PU *pani- 'to put', reflected as *pane- in Balto-Finnic. This verb, however, has irregular vowel reflexes in Permic and Ob-Ugric as well and so can hardly be used as evidence for the opposition 'shortness vs. length'. As for the alleged correspondence between Balto-Finnic long vowels and Proto-Samoyed vowel sequences with *d [Janhunen 1981: 240], it does occur in several reliable etymologies3: PBF *keli ~ PS *kddj 'tongue', PBF *peli ~ PS *ped(j) or *pid(j) 'edge, side', PBF *kusi ~ PS *kadt 'spruce' and PBF *kole- ~ PS *kad- 'to die'. Only the first two of these examples can be regarded as certain, since the reflexes of the word for 'spruce' in daughter languages, especially Ob-Ugric (Proto-Mansi *kayt3, Proto-Khanty *kal), point rather to PU *kowsi (cf. PU *taw5i > Proto-Mansi *tayl, Proto-Khanty *tal), while in the verb 'to die' PS *d may in fact reflect PU *l [Janhunen 2007: 219]. It is not excluded that one of the sources of Samoyed postvocalic *d is a process similar to, but independent from, Balto-Finnic lengthening.
We concur with Janhunen and Sammallahti in reconstructing PU high vowel in non-initial syllables instead of the traditional *e. The fact that in Saami and Mordvinic *e of the first syllable merges with *i if the following syllable contains the vowel in question, whereas in Mordvinic *o of the first syllable merges with *u in the same position, tells us that Balto-Finnic *e in non-initial syllables must be the result of secondary development.
In a number of words, where Janhunen and Sammallahti reconstruct PU / PFU *i or *a, evidence of West Uralic languages (Balto-Finnic, Saami and Mordvinic) points rather to *u or *o. The following etymologies can be mentioned (reconstructions are given according to [Sammallahti 1988]): PU *jiqsi 'bow' (cf. Finnish jousi), PU *kixi 'moon' (cf. Finnish kuu, Erzya and Moksha kov), PU *kaxli- 'to die' (cf. Erzya kulo-, Moksha kuld-), PU *kaxsi 'spruce' (cf. Erzya and Moksha kuz), PU *lixi 'bone' (cf. Finnish luu, Erzya and Moksha lovaza), PU *naxli- 'lick' (cf. North Saami njoallut, Erzia and Moksha nola-), PFU *a)]i 'mouth' (cf. Erzya or/kst, Moksha
2 In the present article we will use the version of Proto-Permic reconstruction suggested in [Zhivlov 2010].
3 We omit cases where internal Samoyed evidence does not allow an unequivocal reconstruction of sequences with *3.
ovdst), PFU *niri 'marsh' (cf. Finnish noro). Such words are not examined in the remainder of the present article.
Janhunen and Sammallahti, as well as many other scholars, systematically distinguish intermediate levels of reconstruction between Proto-Uralic and lower-level protolanguages, namely Proto-Finno-Ugric and Proto-Finno-Permic. However, from the viewpoint of phonological reconstruction it is actually quite difficult to point out significant differences between Proto-Uralic, Proto-Finno-Ugric and Proto-Finno-Permic. Concerning consonantism, no differences come into question at all, and at least some of the vocalic distinctions, postulated in [Sammallahti 1988], are, in fact, mere fiction. In this paper we do not intend to join the discussion on internal Uralic taxonomy and, therefore, do not make any arguments for or against any particular variant of subgrouping for Uralic languages. In dealing with reconstructed protoforms, we will tentatively proceed from only one high level of reconstruction, identifying it as Proto-Uralic. Nevertheless, below we do note protoforms as PU, PFU or PFP. These notes do not actually refer to different levels of reconstruction. Identifying a word in this way, we only indicate how widely it is distributed in Uralic languages. On the other hand, Proto-Samoyed, Proto-Mansi, Proto-Permic, Proto-Komi, and Proto-Udmurt, as well as other low-level protolanguages, are regarded by us as absolutely real idioms with their own phonological and morphological features.
Below we demonstrate a set of rules systematizing the reflexation of Proto-Uralic *a (corresponding to both *a and *i in Janhunen's and Sammallahti's reconstruction) in Samoyed, Mansi and Permic.
Samoyed
In Proto-Samoyed4, reflexation of PU *a depends on stem type, i. e. on second syllable vowels as well as consonant surrounding. It is important to note that, while rules based on consonant surrounding sometimes have to do with the etymological nature of Proto-Samoyed consonants, i. e. with their Proto-Uralic sources, rules that concern stem type do not refer to Proto-Uralic stem-final vowels, but only to Proto-Samoyed ones.
First, we will consider a somewhat rare PS reflex of PU *a. In some specific positions, PU *a yields PS *a (*a in Janhunen's system). This happens if *a is followed by a cluster beginning with a PU palatal or palatalized consonant or is surrounded by two stops or affricates.
1) PU *a > PS *a before clusters beginning with a PU palatal or palatalized consonant:
PU *aska-5 'to step', *asMi 'step' > PS *asdl- (only PSlk *asal-) [UEW: 19 (*aske(~h), *acke(~l3)); Sammallahti 1988: 542 (*aski/ali); Aikio 2002: 40-41 (PU *askili-); SlkWb: 38]; PU *kajwa- 'to dig, to throw', *kajwa 'spade' > PS *kajwa 'spade' [UEW: 170-171 (*kojwa-); Sammallahti 1988: 552 (PFP *ka/ojwa-); SW: 63 (PS *kajwa); Aikio 2002: 41-42 (PU *kajwa-)];
PU *wajji 'breath' > PS *wajj- [UEW: 552-553 (*wajje); Sammallahti 1988: 541 (PU *wajji-); SW: 173 (PS *wajj-)].
4 For Proto-Samoyed we use Juha Janhunen's reconstruction [SW] with modifications by E. Helimski [Helim-skij 1978; Helimski 2005].
5 Reflexes of this root in Mansi and Permic indicate that it was an a-stem. The vowel e of the second syllable in the Finnish derived noun askel 'step' is apparently part of the suffix. Whether the noun in question should be reconstructed as *askili on the PU level is not clear.
Apparently, this group is to be enlarged by a non-trivial case with a unique PU cluster, the first element of which seems to have been a palatalized *n:
PU *anti 'horn' > PS *amtd [UEW: 12-13 (*arjt3 / *oqt3); SW: 20 (PS *amtS)\.
Here we reconstruct the medial nasal as *n in the light of two circumstances. First, its reflexes are generally different from those of the nasal also followed by *t in PU *ort3 'sting, lance', where we definitely deal with a velar: *ort3 > Proto-Mansi *awta, Proto-Khanty *art3-, PS *artd, but *anti > Proto-Mansi *fnt3, Proto-Khanty *ardt and PS *amtd. Second and most important, the Mansi reflex is simply *n. Note that we have no other examples of PU words with a cluster like that in the word 'horn'.
2) PU *a > PS *a between two stops or affricates:
PU *caca- 'to be born, to grow' > PS *cac3 (only PSlk *caci) [UEW: 52 (*cac3- ~ *canc3-); Sam-
mallahti 1988: 552 (PFP *cëcV-), 554 (PFP *sëcV-); SlkWb: 154-155]; PU *caca- (~ *c-) 'to move' > PS *cac3- (only PSlk *cacd-) [UEW: 53 (*canc3- ~ *cac3-); SlkWb: 155];
PU *pata- 'to pot', *pata 'pot'6 > PS *patd- [UEW: 358 (*pata); Sammallahti 1988: 548 (PFU
*pâtâ); SW: 118 (pS *pat- ~ *pat3-)]; PU *taka 'behind' > PS *takd [UEW: 506-507 (*taka); SW: 154 (PS *tak(-) ~ *takd-)}.
The remaining cases of Samoyed reflexation of PU *a can be divided into two groups. In the first group the Samoyed reflexes depend on the quality of the syllable onset, in the second — on the vowels of the second syllable.
Let us look at the first group. Reflexation of PU *a depends on the syllable onset in the following cases: a) in word-initial position PU *a yields *i or *e7, b) in the position after *k-, PU *a yields PS *â, c) in the position after PS dentals (except *t- < PU *s-) PU *a also yields PS *â. Let us list the respective cases.
3) PU *a > PS *i/*e in word-initial position:
PU *akta- 'to hang' > PS *itâ- (tr.), *ito- (intr.) [UEW: 5-6 (*akta-); Sammallahti 1988: 536 (PU
*iktâ-); SW: 25-26 (PS *zt3-)]; PU *ala 'under' > PS *ila [UEW: 6 (*ala); Sammallahti 1988: 536 (PU *ilâ); SW: 24 (PS *№)]; PU *ana, also in compound *ana-appi8 'mother-in-law' > PS *inapd [UEW: 9-10 (*an3(-pp3));
Sammallahti 1988: 536 (PU *inâ-); SW: 24-25 (PS *inSpS)]; PU *ani 'tame' > PS *ind [UEW: 340 (*on3); Sammallahti 1988: 536 (PU *ini); SW: 25 (PS *lnd)\;
6 Comparison of PFU *pata with PS *pat3- is mentioned in [Napol'skih 1997: 124] and [Aikio 2002: 50].
7 As noted by Janhunen, the general rule for the distribution of *i and *e in Samoyed words going back to Proto-Uralic is that *i appears in an open syllable and *e in a closed one [Janhunen 1981: 233-234, 236]. Indeed, in all of the words containing *e, it occurs in a closed syllable (*epta 'hair', *pen- 'put', *je(a)m 'bird-cherry', *ne(a)j 'arrow', *ner 'cartilage', *jeps3 'cradle', *cen 'sinew'), and in the words where *a > *i, the first syllable is open. The only exception is *limpa 'marsh'.
8 Janhunen reconstructs this word as PU *ina + *ippi > *inajppi, suggesting that *aj yielded o in the Finnish cognate anoppi [Janhunen 1982: 227-228]. This reconstruction cannot explain why we see o and not oi in Finnish. We suggest another line of development: *ana + *appi > *anaappi > *anappi > *anoppi (see above on the lengthening of *a to *o in Balto-Finnic) > anoppi (*o was shortened because of the general ban on long vowels in non-initial syllables in Proto-Balto-Finnic).
PU *apti 'hair' > PS *eptd [UEW: 14-15 (*apte); Sammallahti 1988: 536 (PU *ipti); SW: 21 (PS *eptd)].
4) PU *a > PS *a in position after *k-:
PU *ka8a- 'to leave' > PS *kaja- (tr.), *kajo- (intr.) [UEW: 115-116 (*ka8a-); Sammallahti 1988:
537-538 (PU *kad'a-); SW: 58 (PS *kaja- (tr.) ~ *kajd- (intr.))]; PU *kaja 'sun' > PS *kaja [Sammallahti 1988: 538 (PU *kaja); SW: 58 (PS *k&jS)\, PU *kala 'fish' > PS *kala [UEW: 119 (*kala); Sammallahti 1988: 538 (PU *kala) ; SW: 59 (PS *kala)];
PU *kani- 'to go', *kanta- 'to carry' > PS *kan- [UEW: 124 (*kanta); Sammallahti 1988: (PU
*kani-); SW: 59 (PS *kan-)]; PU *kari 'bark' > PS *kar ~ *kar9 [UEW: 184-185 (*kore / *kore); SW: 64-65 (PS *kar)]; PU *kasa- 'to present' > PS *kaso [UEW: 111 (*kac3); Sammallahti 1988: 538 (PU *kasi-); SW: 61 (PS *kSsaj)\.
5) PU *a > PS *a in position after PS dentals (except *t- < PU *s-):
PU *nataw 'sibling / (younger) relative of spouse' > PS *nato [UEW: 299-300 (*nat3); Sammallahti 1988: 539 (PU *natiw); SW: 98-99 (PS *nat1S-)]; PU *saba- 'to rain' > PS *sara- 'to rain', *saro 'rain' [Sammallahti 1988: 540 (PU *sada-); SW:
135-136 (PS *sara- ~ *sard-)]; PU *sala- 'to flash, lighten' > PS *sald- [UEW: 459 (*sala); Sammallahti 1988: 549 (PFU *sila-);
SW: 135 (PS *sald-); Aikio 2002: 27-29 (PU *sala-)]; PU *sajka- 'to sting, stick' > PS *sajk3- 'id.', *sajkd 'penis' [Aikio 2006: 24 (PU *sajka-)]. PU *talja 'skin' > PS *tadjdw [UEW: 508-509 (*talja); SW: 150 (PS *tiaSj1S)]; PU *tanti- 'ground; to tread' > PS *tantd- [SW: 151 (PS *t1ant1d-); Redei 1998]; PU *tasa 'even, exact' > PS *tasd (only Nenets tas) [UEW: 513 (*tasa)].
The fact that PS *t- < PU *s- does not behave like the other Proto-Samoyed dentals11 can be explained if we suppose that at some stage, intermediate between Proto-Uralic and Proto-Samoyed, this phoneme had another place of articulation, e. g. interdental or lateral. Development s > t is found in several branches of Uralic (Samoyed, Mansi and Southern Khanty), but the only branch where we can observe intermediate stages of this development is Khanty. PU *s yields Proto-Khanty voiceless lateral *n that later develops into t in Southern Khanty, but is preserved as a lateral in other dialects. The situation that we assume for Pre-Proto-Samoyed is analogous to that of Surgut dialect of Khanty, where PU *t, *s, and *n are reflected as dentals t, s, and n, but PU *s yields lateral fricative a.
In all the remaining cases, the quality of Samoyed reflexes of PU *a depends on the vocal-ism of the second syllable. The rules can be formulated as follows: a) PU *a yields *i or *e12 in stems whose second syllable contains *d or *e and in stems ending in a consonant, b) PU *a yields PS *a in stems containing PS *a, *a or *o in the second syllable. Below we list the respective cases.
9 The comparison with Samoyed was suggested by E. Helimski in [Helimski s.a.], see also [Aikio 2002: 50].
10 PS *3 in the first syllable of this word may reflect PU *l in the position between PS *a and a consonant (this rule was suggested by J. Janhunen [Janhunen 2007: 219], although he did not mention PS *taaj3).
11 Note that PS *c is a retroflex affricate, not a dental one.
12 See above on the distribution between *i and *e.
6) PU *a > PS *i/*e in stems whose second syllable contains PS *d or *e and in stems ending in a consonant:
PU *8ami 'bird-cherry' > PS *je(d)m [UEW: 65-66 (*8eme / *8ome); Sammallahti 1988: 536537 (PU *d'ixmi)];
PU *lampi 'marsh' > PS *limpd13 (underived stem is preserved in Tundra & Forest Enets lubo 'clay', Nganasan lur/hd 'marshy place', see [Helimski s.a.]) [UEW: 235 (*lampe); SW: 83-84 (PS *limpâ- 'sumpfig sein' ~ *limpd- '(in den Sumpf) sinken')]; PU *lapsi 'cradle' > PS *jepsd ~ *lepsd [UEW: 260 (*lapc3 / *laps3); SW: 41 (PS *jepsd ~ *teSpsS)]; PU *maksa 'liver' > PS *mitd [UEW: 264 (*maksa); Sammallahti 1988: 538 (PU *miksâ); SW: 93-94 (PS *mitS)];
PU *nali 'arrow' > PS *ne(d)j [UEW: 317 (*nele / *nole); Sammallahti 1988: 539 (PU *nixli);
SW: 108-109 (PS *ne1j)]; PU *nari14 'cartilage' > PS *ner [UEW: 317 (Чвткз / *n0rk3); Sammallahti 1988: 546 (PFU
*nirki) ; SW: 108 (PS *ner)]; PU *pani- 'to put' > PS *pen- [UEW: 353-354 (*pane-); Sammallahti 1988: 539 (PU *pini-); SW: 118 (PS *pen-)];
PU *saksa 'cedar' > PS *titer [UEW: 445-446 (*soks3 / *saks3 / *seks3); Sammallahti 1988: 540
(PU *siksi) ; SW: 160 (PS *titSjSr)]; PU *sani 'vein, sinew' > PS *cen15 [UEW: 441 (*sene / *sone); Sammallahti 1988: 548 (PFU *siini); SW: 32-33 (PS *cen)].
7) PU *a > PS *â in stems containing PS *â, *ä or *o in the second syllable:
PU *ca/a- 'to strike, beat, rub' > PS *câ/â- (tr.), *câ/o- (intr.) [UEW: 53-54 (*ca/3-); SW: 151
(PS *t1â/â- (tr.) ~ ЧЧ/д- (intr.)); Aikio 2002: 11-12 (PU *ca/a-)]; PU *lapta 'flat, thin' > PS *jâptâ ~ *japtâ [UEW: 238 (*lapta); SW: 38 (PS *jâptâ ~ *jâptâ)]; PU *pala- 'to bite', *pala 'piece' > PS *pâlâ- [UEW: 350 (*pala); Sammallahti 1988: 540 (PU
*pâlâ-); SW: 116 (PS *pâlâ-)]; PU *sala- 'to steal' > PS *tâlâ- [UEW: 430-431 (*sala); Sammallahti 1988: 540 (PU *sâlâ-); SW: 150-151 (PS *tâlâ-)];
PU *sarka 'branch, fork' > PS *târkâ [Sammallahti 1988: 540 (PU *sârkâ); SW: 152 (PS *t1ârkâ
~ *t1älk3-)];
PU *wala 'word' > PS *wâlâ [UEW: 812 (*wala); Aikio 2006: 26-27 (PU *wala)];
PU *wanca 'root' > PS *wânco [UEW: 548-549 (*wac3 ~ *wanc3); Sammallahti 1988: 541 (PU
*wâncâ) ; SW: 171 (PS *wânc3(-))]; PU *wara 'hill, ridge' > PS *wârâ [Sammallahti 1988: 551 (PFU *wârâ); Aikio 2006: 27-28 (PU *wara)].
It seems possible to add to this group also the following case:
PU *naca 'moss; a kind of grass' > PS *nâcâ [UEW: 311 (*nac3); SW: 105 (PS *nâc ~ *nâcd ~ *nâcâ)].
13 PU *l was apparently regularly preserved in front of *i and *e in Samoyed.
14 Despite UEW, this word should be reconstructed without *-k-, as indicated by Moksha nar 'cartilage' (PU *-rk- would have been preserved in Moksha as -rg-). Curiously enough, the Moksha word is absent in all etymological sources.
15 The consonantal correspondence here is admittedly irregular. Nevertheless, the comparison with Samoyed can be accepted much in the same way as, e.g., comparison of PIE *krd- 'heart' with Old Indic hrd- 'id.' (with equally irregular consonant correspondences) is accepted in modern Indo-European linguistics.
The Nenets and Enets forms (Tundra Nenets nada 'reindeer moss', Tundra & Forest Enets naba 'id.') clearly point to PS *naca or *nata, and the Selkup form (PSlk *nucd 'grass' [SlkWb: 237]) may have also had an original a-stem. The Nganasan form (nota 'grass') reflects *naca, but it may still be a result of a secondary (derivational?) process. On the other hand, we may be dealing here with reflexes of two different PS words: *naca (or *nata) 'reindeer moss' and *nacd 'grass'. In this case, only the former word (if it, indeed, contained PS *c rather than *t) can be included in the Uralic comparison.
Besides this, there are two cases for which we cannot determine the quality of the Proto-Samoyed vowel of the second syllable.
PU *wanca- 'to walk slowly' > PS *wancs- [UEW: 557 (*wanca-); Sammallahti 1988: 551 (PFU *wanca-); Aikio 2002: 36-38 (PU *wanCa-)];
PU *wasa- 'left' > PS *wat3- [UEW: 559 (*wasa); Sammallahti 1988: 541 (PU *wasa-); SW: 172-173 (PS *wat3-)].
In addition to the groups considered above, let us mention a separate case where the re-flexation of first-syllable *a may be due to a specific word structure:
PU *jayari 'cross-grained wood' > PS *jar [UEW: 90 (*jar3); SW: 38 (PS *jar)].
Here the long aa in the Finnish reflex jaarun 'cross-grained wood' points to a deleted consonant that may be tentatively identified as *y (cf. PFU *maya 'earth, land' > Finnish maa).
It seems possible to conclude that all the material which has, up till now, served as Sa-moyed evidence for Proto-Uralic *i, i. e. the cases where West Uralic *a has non-low Samoyed counterparts such as *i and *e, can be interpreted in terms of secondary Samoyed processes caused by specific phonological positions. Some of the postulated rules may seem unusual from the typological point of view. Nevertheless, we think that such rules can be accepted as valid if they are confirmed by a sufficient number of examples and there are no counterexamples to speak of.
As noted above, Mansi and Permic have also been regarded as distinguishing between Proto-Uralic *a (which stands for Proto-Uralic *a in the modern variant of phonological reconstruction) and *i [Sammallahti 1988]. Sammallahti assumes that in Proto-Mansi, *a yields *uu (= *u in the notation that we use in the present article) in *a-stems and *aa (= *a) in i-stems (after a labial in both types of stems *a yields *uu (= *u) or *u), while *i always yields *ii (= *i). In Proto-Permic, on the other hand, *a yields *u (*o in our reconstruction), while *i yields *u (*u in our reconstruction) irrespective of the type of stem (i. e. second-syllable vowel). If this really were so, we could observe the following regular types of correspondences between Mansi and Permic:
1) PMs *u ~ PP *o (PFU *a in a-stems);
2) PMs *a ~ PP *o (PFU *a in i-stems);
3) PMs *u or *u ~ PP *o (PFU *a after a labial);
4) PMs *i ~ PP *u (PFU *i).
In actual fact, we also find Mansi *u (and *u) corresponding to Permic *u and Mansi *i corresponding to Permic *o. Cf. the evidence: PMs *mkc3m ~ PP *ndkcim 'gills'; PMs *slt3 ~ PP *so 'hundred'; PMs *kulap ~ PP *kul- 'net'; PMs *kun- ~ PP *kund- 'to dig'; PMs *nuns- ~ PP *nuz- 'to stretch'; PMs *puns- ~ PP *puj- 'to open'; PMs *uns- ~ PP *vuj- 'to walk slowly'.
Meanwhile, it turns out that at least in those cases where West Uralic languages show evidence for first syllable *a, the appearance of *u or *i in Mansi, as well as that of *o or *u in
Permic, is determined by historical phonological contexts. In all respective words, the Proto-Uralic source of the first syllable vowels would be *a. Like in Samoyed, the actual vowel changes depend both on consonant surrounding and on second syllable vowels, which are mostly stem-final. But, unlike in Samoyed, here it is the Proto-Uralic stem types that are relevant, which makes the picture somewhat different.
Mansi
1) PU *a > PMs *i in i-stems:
PU *anti 'horn' > PMs *mt3 [UEW: 12-13 (*ajt3 / *ojt3); Honti 1982: 128];
PU *apti 'hair' > PMs *it3 [UEW: 14-15 (*apte); Sammallahti 1988: 536 (PU *ipti, PFU *ipti); Honti 1982: 128];
PFU *caci 'wild duck' > PMs *sis [UEW: 58 (*cec3 ~ *cenc3); Honti 1982: 132];
PU *8ami 'bird-cherry' > PMs *llm3 'bird-cherry' [UEW: 65-66 (*8eme / *8ome); Sammallahti 1988: 536-537 (PU *d'ixmi, PFU *d'iimi); Honti 1982: 140];
PFU *kanti 'tree stump' > PMs *kmta [UEW: 123 (*kanta); Sammallahti 1988: 543 (PFU *kinta); Honti 1982: 153];
PFU *nakcimi 'gills' > PMs *mkc3m [UEW: 311-312 (*najkce); Sammallahti 1988: 546 (PFU *nikcimi); Honti 1982: 169];
PU *nali 'arrow' > PMs *mh [UEW: 317 (*nele / *nole); Sammallahti 1988: 539 (PU *nixli, PFU *niili); Honti 1982: 171];
PU *nari 'cartilage' > PMs *mrdy [UEW: 317 (*nerk3 / *nork3); Sammallahti 1988: 546 (PFU *nirki); Honti 1982: 172];
PFU *pasi 'mitten' > PMs *pisya [UEW: 376 (*pes3); Sammallahti 1988: 547 (PFU *piisi); Honti 1982: 179];
PU *sani 'vein, sinew' > PMs *tln3 [UEW: 441 (*sene / *sone); Sammallahti 1988: 548 (PFU *siini); Honti 1982: 137];
PFU *sami 'fish scales' > PMs *slm3 [UEW: 476 (*seme / *some); Sammallahti 1988: 549 (PFU *siimi); Honti 1982: 184].
In the word *anti 'horn', we reconstruct an i-stem (rather than an unknown stem-final vowel, despite UEW), because the development PU *a > PMs *i, if not after *s- or before a velar, happens only in i-stems (whereas in a-stems PU *a > PMs *u, see below). The same can be said of PFU *kanti 'tree stump'. We suggest that the latter word should be compared only to Finnish kanto 'tree stump', and not to Finnish kanta 'foundation' (reflexes of this etymon in all other language groups refer only to a kind of stump, fallen tree or post, not to 'foundation').
In PU a-stems the reflexation of *a depends on surrounding consonants: a) an unmarked reflex is PMs *u, but we also have b) PMs *u after PU *w-, c) PMs before a velar and d) PMs *i after PU *s-.
2) PU *a > PMs *u in a-stems:
PFU *aja- 'to drive' > PMs *ujt- [UEW: 4-5 (*aja-); Sammallahti 1988: 542 (PFU *aja-)];
PFU *ama- 'to scoop' > PMs *um- [UEW: 7-8 (*ama-); Honti 1982: 126];
PU *aska- 'to step', *ask3li 'step' > PMs *us-, *usdl [UEW: 19 (*aske(~l3), *acke(~l3)); Sammallahti 1988: 542 (PFU *aski/ali)];
PU *caca- (~ *c-) 'to move' > PMs *sus- [UEW: 53 (*canc3- ~ *cac3-); Honti 1982: 181];
PU *ka8a- 'to leave' > PMs *kM- [UEW: 115-116 (*kada-); Sammallahti 1988: 537-538 (PU, PFU *kad'a-); Honti 1982: 146];
PU *kala 'fish' > PMs *kul [UEW: 119 (*kala); Sammallahti 1988: 538 (PU, PFU *kala); Honti 1982: 148];
PFU *kala- 'net' > PMs *kulap [UEW: 120 (*kah); Sammallahti 1988: 545 (PFU *kalaw); Honti 1982: 149-150];
PFU *kala- 'to stay overnight' > PMs *kul- [UEW: 120-121 (*kal3-); Sammallahti 1988: 545 (PFU *kala-); Honti 1982: 148];
PFU *kana- 'to dig' > PMs *kun- [UEW: 125 (*kan3-); Sammallahti 1988: 545 (PFU *kana-); Honti 1982: 152];
PU *kanta- 'to carry' > PMs *künt- [UEW: 124 (*kanta); Sammallahti 1988: 538 (PU *kani-, PFU *kanta-); Honti 1982: 153];
PFU *ñanca- 'to stretch' > PMs *nuns- [UEW: 323 (*nonc3 / *ñoc3); Sammallahti 1988: 546 (PFU *ñanca-); Honti 1982: 172];
PU *pala 'piece' > PMs *pul [UEW: 350 (*pala); Sammallahti 1988: 540 (PU, PFU *pala-); Honti 1982: 176];
PFU *panca- 'to open' > PMs *püns- [UEW: 352 (*pance- / *pace-); Sammallahti 1988: 548 (PFU *panca); Honti 1982: 177];
PU *pata 'pot' > PMs *püt [UEW: 358 (*pata); Sammallahti 1988: 548 (PFU *pata); Honti 1982: 180];
PU *sala- 'to steal' > PMs *tuldm- [UEW: 430-431 (*sala); Sammallahti 1988: 540 (PU, PFU *sala-); Honti 1982: 137];
PFU *sar¡ca- 'to stand' > PMs *tüñc- [UEW: 431-432 (*sar¡ca-); Sammallahti 1988: 549 (PFU *sansa-); Honti 1982: 138].
3) PU *a > PMs *u in a-stems after *w-16:
PFU *waja- 'to sink' > PMs *uj- [UEW: 551 (*waja-); Sammallahti 1988: 551 (PFU *waja-); Honti 1982: 192];
PU *wanca- 'to walk slowly' > PMs *uns- [UEW: 557 (*wanca-); Sammallahti 1988: 551 (PFU *wanca-); Honti 1982: 194];
PU *wara 'hill, ridge' > PMs *ur [Sammallahti 1988: 551 (PFU *wara); Honti 1982: 196].
4) PU *a > PMs *l in a-stems before a velar17:
PFU *a¡a- 'to open' > PMs *¡roko- [UEW: 11 (*a¡a-); Sammallahti 1988: 542 (PFU *ira-); Honti 1982: 127-128];
PU *ca¡a- 'to strike, beat, rub' > PMs *s¡roko- [UEW: 53-54 (*ca¡3-); Honti 1982: 133];
PFU *maya 'earth, land' > PMs *mi (*mfy-) [UEW: 263-264 (*maye); Sammallahti 1988: 546 (PFU *mixi); Honti 1982: 202];
PU *maksa 'liver' > PMs *mljdt (< *mlydt) [UEW: 264 (*maksa); Sammallahti 1988: 538 (PU, PFU *miksa); Honti 1982: 163];
PFU *palya 'village' > PMs *piwdl (< *piyodl) [UEW: 351 (*paly3); Sammallahti 1988: 548 (PFU *palwa); Honti 1982: 175];
16 But not before a velar, see below on PMs *wfydl- < PFU *walka-.
17 The following velar becomes labialized if the word does not begin with a labial. The only exception is the word for 'village'.
PFU *pa¡ka 'mushroom' > PMs *pí¡k3 [UEW: 355-356 (*pa¡ka); Sammallahti 1988: 547 (PFU *pir¡ka); Honti 1982: 178];
PU *saksa 'cedar' > PMs *tít (< *tiydt or *tlyodt) [UEW: 445-446 (*soks3 / *saks3 / *seks3); Sammallahti 1988: 540 (PU *siksi, PFU *siksi); Honti 1982: 136];
PFU *walka- 'to descend' > PMs *wiyal- [UEW: 554 (*walka-); Sammallahti 1988: 551 (PFU *wilka-); Honti 1982: 193].
5) PU *a > PMs *í in a-stems after *s-18:
PU *sala- 'to flash, lighten' > PMs *sll- [UEW: 459 (*sala); Sammallahti 1988: 549 (PFU *sila-); Honti 1982: 184];
PFU *salka 'pole' > PMs *siyla [UEW: 460-461 (*salk3); Sammallahti 1988: 549 (PFU *silka); Honti 1982: 183];
PFU *sata 'hundred' > PMs *sit3 [UEW: 467 (*sata); Sammallahti 1988: 549 (PFU *sata); Honti 1982: 186].
6) PU word-initial *a > PMs *a in stems with more than two syllables:
PU *ana-appi 'mother-in-law' > PMs *amp [UEW: 9-10 (*an3(-pp3)); Sammallahti 1988: 536 (PU, PFU *ina); Honti 1982: 127];
PFU *asira 'lord' > PMs *atdr [UEW: 18 (*assr3); Sammallahti 1988: 542 (PFU *asira)].
Cases of irregular reflexation of PU *a:
PU *ala 'under' > PMs *jal- [UEW: 6 (*ala); Sammallahti 1988: 536 (PU, PFU *ila); Honti 1982: 199];
PFU *ancara 'tusk, fang' > PMs *ancdr [UEW: 340 (*onca-r3); Sammallahti 1988: 541 (PFU *¿nsara)];
PU *pani- 'to put' > PMs *pina- [UEW: 353-354 (*pane-); Sammallahti 1988: 539 (PU *pini-, PFU *pini-); Honti 1982: 177].
Permic
1) PU *a > PP *o in i-stems:
PFU *caci 'wild duck' > PP *cjz > PK *coz ~ -o-, PUd *coz [UEW: 58 (*cec3 ~ cenc3); KESK: 310];
PU *8ami 'bird-cherry' > PP *lom > PK *lom (*lomj-), PUd *lom [UEW: 65-66 (*8eme / *8ome); Sammallahti 1988: 536-537 (PU *d'ixmi, PFU *d'iimi, PFP *d'oomi); KESK: 164-165];
PU *nali 'arrow' > PP *nol > PK *nol (*nolj-), PUd *nol [UEW: 317 (*nele / *nole); Sammallahti 1988: 539 (PU *nixli, PFU *niili, PFP *nooli); KESK: 197];
PFU *pasi 'mitten' > PP *pjz > PK *ke-pus19, PUd *poz [UEW: 376 (*pes3); Sammallahti 1988: 547 (PFU *piisi, PFP *poosi); kesk: 121];
PU *sani 'vein, sinew' > PP *son > PK *son ~ -o-, PUd *son [UEW: 441 (*sene / *sone); Sammallahti 1988: 548 (PFU *siini, PFP *sooni); KESK: 263];
PFU *sami 'fish scales' > PP *som > PK *som, PUd *som [UEW: 476 (*seme / *some); Sammallahti 1988: 549 (PFU *siimi, PFP *soomi); KESK: 270].
18 This rule was suggested by Ante Aikio [Aikio 2002: 28-29].
19 Komi has *u instead of expected *o, because the latter vowel was not allowed in non-first syllables.
There is also one PP stem with *i in the second syllable, where PU *a > PP *o (not *j):
PFU *asira 'lord' > PP *ozir > PK *ozùr, PUd *uzir [UEW: 18 (*as8T3); Sammallahti 1988: 542 (PFU *âsirâ, PFP *asira); KESK: 203].
Note that here, unlike cases in which PFU *a > PP *o, the vowel change happens in wordinitial position and, besides, the word has a trisyllabic structure where *i of the second syllable is not a stem-final vowel.
2) PU *a > PP *u in a-stems before non-palatalized consonants:
PU *ala 'under' > PP *ul > PK *ul, PUd *ul [UEW: 6 (*ala); Sammallahti 1988: 536 (PU, PFU *ilâ, PFP *ëla); KESK: 295];
PFU *amta- 'to give' > PP *ud- > PK *ud-, PUd *ud- [UEW: 8 (*amta-); Sammallahti 1988: 541 (PFU *imtâ-, PFP *ëmta-); KESK: 295-296];
PU *caca- 'to be born, to grow' > PP *cuz- > PK *cuz-, PUd *cizi [UEW: 52 (*cac3- ~ *canc3-); Sammallahti 1988: 552 (PFP *cëcV-), 554 (PFP *sëcV-); KESK: 312];
PFU *kala- 'net' > PP *kul- > PK *kulom [UEW: 120 (*kal3); Sammallahti 1988: 545 (PFU *kâlâw, PFP *kalama); KESK: 145];
PFU *kana- 'to dig' > PP *kun- > PK *kund- [UEW: 125 (*kan3-); Sammallahti 1988: 545 (PFU *kânâ-, PFP *kana-); KESK: 146];
PFP *karwa 'bitter, sharp' > PP *kur- > PK *kurùt, PUd *kurit [UEW: 128-129 (*karwa); KESK: 147];
PFP *lamta20 'low, lowland' > PP *lud > PK *lud, PUd *lud [UEW: 235-236 (*lamte); Sammallahti 1988: 552 (PFP *lënti); KESK: 163];
PFU *maya 'earth, land' > PP *mu > PK *mu, PUd *mu [UEW: 263-264 (*maye); Sammallahti 1988: 546 (PFU *mixi, PFP *mëxi); KESK: 177];
PU *maksa 'liver' > PP *mus > PK *mus (*musk-), PUd *mus [UEW: 264 (*maksa); Sammallahti 1988: 538 (PU, PFU *miksâ, PFP *mëksa); KESK: 179];
PFP *marta 'sterile' > PP *muri > PK *mur, PUd *muri ~ -ù- [UEW: 699 (*marta); KESK: 178];
PFU *nanca- 'to stretch' > PP *nuz- > PK *nuz-, PUd *nuz- [UEW: 323 (*nonc3 / *noc3); Sammallahti 1988: 546 (PFU *nâncâ-, PFP *nanca-); KESK: 200];
PFU *panca- 'to open' > PP *puj- > PK *puj-, PUd *puz- [UEW: 352 (*pance- / *pace-); Sammallahti 1988: 548 (PFU *pâncâ, PFP *panca-); KESK: 231];
PFP *para 'good' > PP *bur > PK *bur, PUd *bur [UEW: 724 (*para); Sammallahti 1988: 553 (PFP *përa); KESK: 42];
PU *saksa 'cedar' > PP *sus > PK *sus, PUd *susi ~ -ù- [UEW: 445-446 (*soks3 / *saks3 / *seks3); Sammallahti 1988: 540 (PU *siksi, PFU *siksi, PFP *sëksi); KESK: 267];
PFU *salka- 'to stand' > PP *sul- > PK *sul-, PUd *sil- [UEW: 431 (*salk3-); Sammallahti 1988: 553 (PFP *sëlkV-); KESK: 265];
PFU *salka 'pole' > PP *sul 'sledge runner'2i > PK *sul, PUd *sul [UEW: 460-461 (*salfo); Sammallahti 1988: 549 (PFU *silkâ); KESK 273];
PFP *tanta- 'to hinder, to be obstinate' > PP *dud- > PK *dud- [KESK: 97];
20 We reconstruct *a in the second syllable of this word because of Saami data (North Saami luovdit 'to lie down flat'). The e-stem in Finnish lansi 'low, lowland' can be explained as resulting from a well-known but poorly understood process of secondary transfer of some PU a-stems into the class of Balto-Finnic e-stems, cf. cases like PFP *komta 'lid' > North Saami goavdi, Finnish kansi, PFU *sorwa 'horn' > North Saami coarvi, Finnish sarvi etc.
21 We suggest this etymology instead of the traditional comparison of PFU *salka with PK *jal 'lath'.
PFP *tarna 'grass' > PP *turin > PK *turùn, PUd *turin [UEW: 792 (*tarna); Sammallahti 1988: 554 (PFP *tërna); KESK: 287];
PU *wanca 'root' > PP *vuji > PK *vuz (*vuzj-), PUd *vi3i [UEW: 548-549 (*wac3 ~ *wanc3); Sammallahti 1988: 541 (PU, PFU *wâncâ, PFP *vanca); KESK: 69-70];
PU *wanca- 'to walk slowly' > PP *vuj- > PK *vuj-, PUd *vÎ3- [UEW: 557 (*wanca-); Sammallahti 1988: 551 (PFU *wâncâ-, PFP *vanca-); KESK: 69];
PFP *wajka 'handle (of vessels)' > PP *vugi > PK *vug (*vugj-), PUd *vugi ~ -ù- [UEW: 814 (*wajka); Sammallahti 1988: 554 (PFP *vëjka); KESK: 69];
PFU *warka- 'to sew' > PP *vur- > PK *vur-, PUd *vur- [UEW: 584-585 (*work3-); Sammallahti 1988: 551 (PFU *wirkâ-, PFP *vërka-); KESK: 70].
3) PU *a > PP *wô-, -ô- in a-stems before palatalized consonants and *-rj-:
PFU *aja- 'to drive' > PP *woj---o- > PK *woj---o-, PUd *ùj- [UEW: 4-5 (*aja-); Sammallahti 1988: 542 (PFU *âjâ-, PFP *aja-); KESK: 61];
PFU *ancara 'tusk, fang' > PP *wôj3r ~ -o- > PK *wojir ~ -o-, PUd *wajer [UEW: 340 (*onca-r3); Sammallahti 1988: 541 (PFU *insârâ, PFP *ënsara); KESK: 60];
PU *aska- 'to step', *ask3li 'step' > PP *wock3l > PK *woskol, PUd *ùckil [UEW: 19 (*aske(~h), *acke(~l3)); Sammallahti 1988: 542 (PFU *âski/âli, PFP *askili); KESK: 64];
PU *kaSa- 'to leave' > PP *kol- > PK *koï-, PUd *Ш- [UEW: 115-116 (*ka8a-); Sammallahti 1988: 537-538 (PU, PFU *kâd'â-, PFP *kad'a-); KESK: 131];
PU *kajwa- 'to dig, to throw' > PP *koj- > PK *koj-, PUd *kùj- [UEW: 170-171 (*kojwa-); Sammallahti 1988: 552 (PFP *ka/ojwa-); KESK: 128];
PU *kasa- 'to present' > PP *kôj- > PK *kojin, PUd *kùjim [UEW: 111 (*kac3); Sammallahti 1988: 538 (PU *kâsi-, PFU *kâsi, PFP *kasi-); KESK: 127-128];
PFP *sarja 'lath, pole' > PP *sôri > PK *sor (*sorj-), PUd *sùri [UEW: 770-771 (*sarja); Sammallahti 1988: 553 (PFP *sarja); KESK: 253].
Despite Sammallahti and [Janhunen 1981: 231], the final vowel in PFU/PU *kaca- 'to present' is to be reconstructed as *-a rather than *-i (or, according to our reconstruction system, *-i). We can make this conclusion on the basis of Permic *kôj-, because *a > PP *ô in a-stems and not in i-stems (in an i-stem, *a of the first syllable would have changed into *э — see above the group 1).
4) PU *a > PP *ô before clusters beginning with *-k-:
PU *akta- 'to hang' > PP *ôkt- > PK *okt- [UEW: 5-6 (*akta-); Sammallahti 1988: 536 (PU, PFU *iktâ-, PFP *ëkta-); KESK: 204];
PFU *nakcimi 'gills' > PP *nôkcim ~ -o- > PK *nokcim ~ -o- [UEW: 311-312 (*najkce); Sammallahti 1988: 546 (PFU *nikcimi, PFP *nëkcimi); KESK: 189].
5) PU *a > PP *wô-, -ô- before a deleted dental:
PFU *aöi 'year' > PP *wô > PK *wo, PUd *wa [UEW: 335-336 (*oôe / *ööe); Sammallahti 1988: 552 (PFP *ooti); KESK: 59];
PFU *sata 'hundred' > PP *sô > PK *so, PUd *su [UEW: 467 (*sata); Sammallahti 1988: 549 (PFU *sâtâ, PFP *sata); KESK: 252].
It is interesting that before a deleted dental or a cluster beginning with *-k-, *ô appears even in i-stems, where PU *a is otherwise regularly reflected as *э (cf. above the group 1)): PFU *aôi 'year' > PP *wô, *nakcimi 'gills' > PP *nôkcim.
Cases of irregular reflexation of PU *a22:
PFU *kala- 'to stay overnight' > PP *kol- > PUd *kol- [UEW: 120-121 (*kah-); Sammallahti 1988: 545 (PFU *kala-, PFP *kala-); KESK: 61];
PU *pani- 'to put' > PP *pan- > PK *pon---o-, PUd *pon- [UEW: 353-354 (*pane-); Sammallahti 1988: 539 (PU *pini-, PFU *pini-, PFP *peni-); KESK: 228];
PFP *saca- 'to reach' > PP *suj- > PK *suj-, PUd *suj- [UEW: 748 (*sac3-); Sammallahti 1988: 553 (PFP *seci-); KESK: 265];
PFU *sarjca- 'to stand' > PP *sij- > PK *sij- [UEW: 431-432 (*sarjca-); Sammallahti 1988: 549 (PFU *sansa-); KESK: 254];
PFP *sarna 'incantation; speech' > PP *sorn- > PK *sorni [UEW: 463-464 (*sarna); Sammallahti 1988: 553 (PFP *sarna); KESK: 253-254];
PFP *tarka 'tassel, fringe' > PP *tug > PK *tug (*tugj-), PUd *tug [UEW: 791 (*tarka); KESK: 285];
PFU *waja- 'to sink' > PP *vej- > PK *voj-, PUd *vij- [UEW: 551 (*waja-); Sammallahti 1988: 551 (PFU *waja-, PFP *vaja-); KESK: 66].
Investigation of the relevant material brings us to the conclusion that the vocalic distribution, which Janhunen and Sammallahti interpreted as reflexes of PU *i and *a, is, in fact, completely dependent on phonological environment of the respective vowels in Samoyed, Mansi and Permic. The vowels in question thus seem to reflect only one Proto-Uralic phoneme, which, in the light of phonetic and typological considerations, should be reasonably reconstructed as *a. Thus, at least these particular three branches offer no evidence for the reconstruction of a non-low unrounded back vowel in Proto-Uralic.
Abbreviations
PBF — Proto-Balto-Finnic; PFP — Proto-Finno-Permic; PFU — Proto-Finno-Ugric; PIE — Proto-Indo-European; PK — Proto-Komi; PMs — Proto-Mansi; PP — Proto-Permic; PS — Proto-Samoyed; PSlk — Proto-Selkup; PU — Proto-Uralic; PUd — Proto-Udmurt
Literature
Aikio, A.: New and Old Samoyed Etymologies // Finnisch-Ugrische Forschungen 57 (2002), 9-57.
Aikio, A.: New and Old Samoyed Etymologies (Part 2) // Finnisch-Ugrische Forschungen 59 (2006), 9-34.
Helimskij, E. A.: Rekonstrukcija prasevernosamodijskih (PSS) labializovannyh glasnyh nepervyh slogov // Konfer-
encija «Problemy rekonstrukcii». 23-25 oktjabrja 1978 g. (Tezisy dokladov). Moskva, 1978. Pp. 123-126. Helimski, E.: The 13th Proto-Samoyedic vowel // Wagner-Nagy, B. (ed.): Mikola-konferencia 2004. Szeged, 2005. Pp. 27-39.
Helimski, E.: Marginalia ad UEW. Notizen zu Karoly Rédeis Uralisches Etymologisches Wörterbuch. S.a. [http://helimski.
com/Marginalia_ad_UEW/Marginalia ad UEW.doc] Accessed at 30.10.2010. Honti, L.: Geschichte des obugrischen Vokalismus der ersten Silbe. Budapest, 1982.
Janhunen, J.: Uralilaisen kantakielen sanastosta // Journal de la Société Finno-ougrienne 77 (1981), 219-274. Janhunen, J.: The primary laryngeal in Uralic and beyond // Ylikoski, J. & Aikio, A. (eds.): Samit, sanit, satnehamit. Riepmocala Pekka Sammallahtii miessemanu 21. beaivve 2007. (MSFOu 253). Helsinki, 2007. Pp. 203-227.
22 One more apparent exception from the rules postulated above, the word for 'salt' (Finnish suola, Erzia and Moksha sal, PP *sol > PK *sol, PUd *silal), was, in our view, independently borrowed from Indo-European languages into Balto-Finnic, Mordvinic, and Permic.
KESK — Lytkin, V. I. & Guljaev, E. I.: Kratkij etimologiceskij slovar' komi jazyka. Moskva, 1970.
Lehtinen, M.: On the Origin of the Balto-Finnic Long Vowels // Ural-Altaische Jahrbücher 39 (1967), 147-152.
Napol'skih, V. V.: Vvedenije v istoriceskuju uralistiku. Izhevsk, 1997.
Redei, K.: Suomen tanner // Grünthal, R. & Laakso, J. (eds.): Oekeeta asijoo. Commentationes Fenno-Ugricae in honorem Seppo Suhonen sexagenarii 16.V.1998. (MSFOu 228). Helsinki, 1998. Pp. 384-385.
Sammallahti, P.: Über die Laut- und Morphemstruktur der uralischen Grundsprache // Finnisch-Ugrische Forschungen 43 (1979), 22-66.
Sammallahti, P.: Historical phonology of the Uralic languages // Sinor, D. (ed.): The Uralic Languages: Description, History and Foreign Influences. Leiden e. a., 1988. Pp. 478-554.
SlkWb — Alatalo, J. (comp. & ed.): Sölkupisches Wörterbuch aus Aufzeichnungen von Kai Donner, U. T. Sirelius und Jarmo Alatalo. (LSFU XXX). Helsinki, 2004.
Steinitz, W.: Geschichte des finnisch-ugrischen Vokalismus. Stockholm, 1944.
SW — Janhunen, J.: Samojedischer Wortschatz: Gemeinsamojedischen Etymologien. (Castrenianumin toimitteita 17). Helsinki, 1977.
UEW — Redei, K.: Uralisches Etymologisches Wörterbuch. Bd. I-III. Budapest, 1986-1991.
Zhivlov, M.: Studies in Uralic vocalism I: A more economical solution for the reconstruction of the Proto-Permic vowel system // Journal of Language Relationship 4 (2010), 167-176.
В статье предлагаются правила дополнительного распределения между рефлексами гипотетических прауральских гласных *i и *ä в прасамодийском, прамансийском и прапермском языках. Цель авторов — показать, что реконструкции одной гласной *a вместо *i и *ä достаточно, чтобы объяснить все рефлексы в рассматриваемых языках.
Ключевые слова: уральские языки, финно-угорские языки, самодийские языки, мансийский язык, коми язык, удмуртский язык, историческая фонетика, лингвистическая реконструкция, прауральский вокализм.
