STUDIES IN URALIC ETYMOLOGY I: SAAMI ETYMOLOGIES

Abstract

This paper is the first part in a series of studies that present additions to the corpus of etymological comparisons between the Uralic languages, drawing data from all the major branches of the language family. It includes both previously unnoticed cognates that can be added to already established Uralic cognate sets, as well as a few completely new reconstructions of Uralic word roots. In this first part new Uralic etymologies for the following Saami words are discussed: North Saami báhtarit 'flee, escape' (< PU *pätäri-), cáhkit 'put (into), stick into' (< PU *cäkä-), coska 'block of wood' (< PU *cucki), faððut 'beat, lash, slap' (< PU *widi-), guodja 'flower stalk or seed shell of a sedge' (< PU *kaji), guolmmas 'soft, white inner bark of conifers' (< PU *kolmis), vuoððu 'bottom, foundation' (< PU *adÍi-), and South Saami muhtsies 'slovenly, untidy, messy' (< PU *muca) and vïekedh 'grab, take hold of' (< PU *wexi-), and Ter Saami caoÎÎped 'make corner joints (in building logs)' (< PU *c[asymptotically =]ippa-).

Headnote

Abstract. This paper is the first part in a series of studies that present additions to the corpus of etymological comparisons between the Uralic languages, drawing data from all the major branches of the language family. It includes both previously unnoticed cognates that can be added to already established Uralic cognate sets, as well as a few completely new reconstructions of Uralic word roots. In this first part new Uralic etymologies for the following Saami words are discussed: North Saami báhtarit 'flee, escape' (< PU *pätäri-), cáhkit 'put (into), stick into' (< PU *cäkä-), coska 'block of wood' (< PU *cucki), faððut 'beat, lash, slap' (< PU *widi-), guodja 'flower stalk or seed shell of a sedge' (< PU *kaji), guolmmas 'soft, white inner bark of conifers' (< PU *kolmis), vuoððu 'bottom, foundation' (< PU *adÍi-), and South Saami muhtsies 'slovenly, untidy, messy' (< PU *muca) and vïekedh 'grab, take hold of' (< PU *wexi-), and Ter Saami caoÎÎped 'make corner joints (in building logs)' (< PU *c[asymptotically =]ippa-).

Keywords: Uralic languages, Saami languages, etymology, historical phonology.

Headnote

Abbreviations

Fi - Finnish; Hung - Hungarian; Kam - Kamas; KhE - East Khanty; KhN - North Khanty; KhS - South Khanty; MariE - East Mari; MariNw - Northwest Mari; MariV/U - Volga/Upçsa Mari; MariW - West Mari; MdE - Erzya Mordvin; MdM - Moksha Mordvin; MsE - East Mansi; MsN - North Mansi; MsS - South Mansi; MsW - West Mansi; NenT - Tundra Nenets; Ngan - Nganasan; PKh - Proto-Khanty; PMari - Proto-Mari; PMd - Proto-Mordvin; PMs - Proto- Mansi; Pre-PFi - Pre-Proto-Finnic; PSaa - Proto-Saami; PSam - Proto-Samoyed; PSlk - Proto-Selkup; PU - Proto-Uralic; SaaI - Inari Saami; SaaK - Kildin Saami; SaaL - Lule Saami; SaaN - North Saami; SaaP - Pite Saami; SaaS - South Saami; SaaSk - Skolt Saami; SaaT - Ter Saami; SlkK - Ket Selkup; SlkTa - Taz Selkup; SlkTy - Tym Selkup; Udm - Udmurt.

(ProQuest: ... denotes non-US-ASCII text omitted.)

Introduction

During the 1980s there have been significant advances in Uralic phonological reconstruction, especially due to the studies conducted by Janhunen (1981) and Sammallahti (1988). The prerequisites for these developments were provided by E. Itkonen's studies on vowel correspondences between Finno-Permic languages (esp. 1946; 1954), as well as successive studies on phonological reconstruction in the central and eastern lower-level branches: Permic (Lytkin 1964), Samoyed (Janhunen 1977) and Ob-Ugric (Honti 1982). As a result of this work, the highly complex vowel correspondences between branches of Uralic had become much clearer by the end of the 1980s. On the other hand, this progress required a critical reexamination of the etymological stock, and numerous Uralic lexical comparisons that had been considered reliable in earlier etymological references were rejected by Janhunen (1981) and Sammallahti (1988).

The framework of Uralic sound laws established during this phase of research has well endured critical examination. While there is ongoing debate concerning many details of phonological reconstruction (e.g., Häkkinen 2007; Reshetnikov, Zhivlov 2011; Aikio 2012), the differences of opinion do not encompass basic sound correspondences and the major outlines of Proto-Uralic phonological structure. Further confirmation for Janhunen and Sammallahti's results has been reached through the discovery of new etymological comparisons that abide by the proposed sound laws, especially in the case of Samoyed (e.g., Aikio 2002; 2006), but also in more western branches (e.g., Saarikivi 2007).

The present paper is the first part in a series of studies which put forward new etymological explanations for various Uralic words. In addition to the self-evident purpose of etymologizing individual word-roots in the Uralic languages, there is a more general aim: it will be shown that the adherence to a strict framework of regular sound correspondences has a significant predictive value in the etymological study of all Uralic languages, and it is possible to establish previously undiscovered Uralic etymologies for word roots from every branch of Uralic. In order to demonstrate this, new etymologies will be presented for many words from lower-level branches of Uralic, i.e. Saami, Finnic, Mordvin, Mari, Permic, Khanty, Mansi, and Hungarian; a number of new Samoyed etymologies have been presented by Aikio (2002; 2006). In this first part of the series of papers, etymologies for Saami words are discussed.

The sound correspondences on which the etymologies are based are, for the most part, identical to those assumed by Janhunen (1981) and Sammallahti (1988). Some modifications to their framework of vowel correspondences, in particular with regard to the Finno-Permic or Finno-Ugric long vowels supposedly reflecting a Proto-Uralic sequence *-Vx-, have recently been suggested by Aikio (2012). In the interpretation of sound correspondences there will be some minor differences to Sammallahti 1988; these will be explained in the connection of the individual etymologies. In the case of already established or proposed Uralic etymologies, the "Uralisches etymologisches Wörterbuch" (UEW) is referred to, but the phonological reconstructions do not necessarily accord with those presented in this dictionary. The sources of lexical data are not separately referred to, but the dictionaries used are listed among the references. A list of the abbreviations for the names of languages is provided at the end of the paper.

1. SaaN báhtarit 'flee, escape' ~ MsN pat?r- 'disappear'

< PU *pätäri- < Aryan *patar-

SaaN báhtarit has cognates in all Saami languages except for Ter Saami: cf. e.g. SaaS baataridh ~ baateridh, SaaL báhtarit, SaaI patârið, SaaK pαdd.reδ. In Skolt Saami the verb has an irregular vowel (puåt'tred < *poat≠er≠e-), but this is evidently secondary, as all the other forms point to PSaa *a; an irregular change *a > *oa in Skolt was probably caused by the labializing influence of *p-. The meaning is the same in all languages: 'flee, escape, run away'. In North Saami also a noun *báhtu occurs in adverbs that are fossilized local case forms: LOC báðus 'in flight', ILL báhtui 'escaping, taking flight'. These are best explained as retrograde formations of the verb báhtarit, as no other Saami language shows forms based on a shorter root *patV-.

No etymology has been proposed for the Saami verb. However, there is a strikingly similar verb in Mansi: MsN pat?r-, MsE pöät?r-, MsW p t?r- 'disappear', MsS p?tur- 'hide' (< PMs *p?t?r-). While the meanings are not exactly the same as in Saami, the link between 'escaping', 'hiding' and 'disappearing' is self-evident. The consonantism of the Saami and Mansi words is identical, and the first-syllable vowel correspondence PSaa *a ~ PMs *? is the regular outcome of the Uralic vowel combination *ä-ä (Sammallahti 1988 : 504). Aminor phonological problem, however, is caused by the second syllable vowel in Saami: most of the forms presuppose PSaa *pat≠er≠e-, and the PSaa vowel combination *a-≠e is unetymological - i.e., it generally does not appear in word-roots of Uralic origin. In this case, however, the unstressed vowel *≠e is in all likelihood the result of secondary development: in South Saami also the form baataridh is found, and this presupposes a PSaa reconstruction *patar≠e-, which is exactly the regular cognate of PMs *p?t?r-. Hence, the Uralic proto-form can be reconstructed as *pätäri-.

The trisyllabic form of the verb *pätäri- is anomalous in regard to Uralic root structure. One would expect a trisyllabic verb stem to be a derivative of some disyllabic root, but no such derivational suffix as *-ri- is known. In Saami the verb is even synchronically unusual: báhtarit is one of only three North Saami verbs with a stem-final -r-, the others being guohkkárit 'get nicks (of a knife blade)' and nimmorit 'grumble, whine about something'. These two others look like recent formations, as their only cognates are found in Inari Saami. There is a straightforward explanation for the non-canonical shape of the verb báhtarit: it appears to be borrowed from Aryan *patar-, cf. Sanskrit patará?, patáru? 'flying, shooting through the air', Avestan patar?-ta- 'flying' (related to Sanskrit pátram 'wing, feather' ~ English feather, etc.; Mayrhofer 1986-2001 : 71).

2. SaaN cáhkit 'put (into), stick into' ~ MariE cije- 'dress, put on (clothes)'

< PU *cäkä- 'put, stick in'

Cognates of SaaN cáhkit are attested in all Saami languages (Lehtiranta 1989 : 73). The basic meaning of the Saami verb is 'put into, stick into', but also the sense 'dress, put on (clothes, a piece of clothing)' is widely attested. The verb would regularly reflect PU *cäkä-, and hence MariE cije-, MariW cie- 'put on (clothes)' (< PMari *cie-) can be analyzed as its cognate. The vowel shiftPU *ä > PMari *i is regular in Mari *CV(V)-type stems: cf. PU *jäni 'ice' > PMari *i > MariE ij, W i (UEW 93), PU *täji 'louse' > PMari *ti > MariE tij, MariW ti (UEW 515), PU *wäki 'strength' > PMari *wi > MariE wij, MariW wi (UEW 563). Also the loss of intervocalic *k is regular in Mari, cf. PU *ikä 'year, age' > PMari *i > MariE ij, MariW i 'year' (UEW 98), PU *luka 'ten' > PMari *lu > MariE, MariW lu (UEW 253). Note that MariE -jin cije- is a secondary hiatus-filling glide, not a direct reflex of PU *k.

Semantically the equation of SaaN cáhkit and Mari cije- is flawless, as the meaning 'dress, put on (clothes)' is attested in both branches. The original meaning of the verb is probably 'put into, stick into', from which the meaning 'dress, put on (clothes)' secondarily developed.

3. SaaT caoÎ Îped 'make corner joints (in building logs)' ~ MdE capo-, Komi cup-, Udm cup[asymptotically =]i- 'make a cut or notch (in wood), make corner joints (in building logs)'

< PU *c[asymptotically =]ippa- 'make corner joints'

SaaT caoÎ Îpe- would go back to a PSaa form *cappe-. Despite the lack of cognates in other Saami languages, the verb can be equated with the following already established Uralic cognate set: MdE èapo-, MdM sap.- 'make a cut or notch (in wood), make corner joints (in building logs)', MdE èapo, MdM sap ~ sapa 'cut, notch (in wood)', Komi èup-, Udm èup.i- 'make a cut or notch (in wood), make corner joints (in building logs)' (UEW 618- 619). The proto-form of these words can be reconstructed as PU *è.ippa-. The vowel *-.i- (as opposed to PU *-a-) is suggested by the Komi cognate, as the regular reflex of PU *-a- is Komi -o- instead of -u- (Sammallahti 1988 : 530-533). According to Reshetnikov's and Zhivlov's recent theory (2011) the vowel reflexes subsumed by Janhunen (1981) and Sammallahti (1988) under PU *a and *.i can be explained as developments of a single phoneme (PU *a) under different phonological environments; for Komi they assume a delopment PU *a > Komi o before palatalized consonants and clusters of the type *-Cj-, whereas in other contexts the development would have been PU *a > Komi u. If this is correct, the original form could be reconstructed as *èappa-, but the question has no bearing on the validity of the comparison.

Previously the Mordvin and Permic verbs mentioned above have been compared to SaaN cáhpat 'chop', but the equation has been considered uncertain (UEW 618). This suggestion must be rejected, as SaaN cáhpat (< PSaa *cap.e-) shows the unetymological vowel combination *a-.e which does not normally occur in roots of Uralic origin, and moreover its single stop *-p- does not match the geminate *-pp- presupposed by the Mordvin and Permic forms. SaaT caoÎ2 Îped, in contrast, shows a phonologically straightforward match, and the equation is also semantically more exact. It is true, the fully regular reflex of PU *è.ippa- (or *èappa-) would be PSaa *cuoppeinstead of *cappe-, but there are a few cases where the regular development Pre-PSaa *a > PSaa *uo did not take place, but the reflex of Pre-PSaa *a (< PU *a, *.i) is PSaa *a instead; all of these cases involve Uralic *a-stems (Korhonen 1981 : 90-91; Sammallahti 1998 : 183). Previously known examples include, e.g., PU *wanèa- 'move cautiously' > PSaa *vance- > SaaN vázzit 'walk' (UEW 557; Aikio 2002 : 36-38), PU *kaèka- 'bite' > PSaa *kacke- > SaaN gáskit (UEW 641), and PU *wara 'hill, mountain' > PSaa *vare > SaaN várri (Aikio 2006 : 27). Also SaaT caoÎ2 Îped may be counted in this group of exceptions. As the equation is otherwise both semantically and phonologically precise, it would be difficult to classify it as a chance correspondence. Some of the proposed exceptions to the development *a > *uo result from borrowing from Finnic (e.g. SaaN báhti 'caldron' < PSaa *pate < Fi pata 'caldron' < PU *pata 'pot'; UEW 358), but this cannot be the case with SaaT caoÎ2 Îped due to its initial affricate c-. Moreover, no reflex of PU *è.ippa- is attested in Finnic.

4. SaaN èoska 'block of wood' ~ MdE èoèko 'timber, log' ~ ? SlkTa tot.i 'pole'

< PU *èuèki 'log'

SaaN èoska has cognates in almost all Saami languages, e.g. SaaL tjasska 'piece of firewood', SaaI èuskâ 'log', SaaSk èockk 'block of wood' (Lehtiranta 1989 : 25). The Proto-Saami form of the words was *èock.e, which would regularly reflect a PU form *oeuèki or *æ uèki. This closely resembles MdE èoèko, MdM soèka 'timber, log', which would presuppose a Uralic form *cucki. The correspondence between the Saami and Mordvin initial affricates is not regular, but the Saami representation can be quite naturally explained as due to a dissimilation of two identical affricates: PU *cucki >> Pre-PSaa *cucki. This change even has an exact parallel, PU *cecä 'uncle' >> *cecä > PSaa *ceace > SaaN ceahci 'paternal uncle (younger than father)' (Janhunen 1981 : 225). The same dissimilation is attested in the Finnic cognate: cf. Fi setä 'paternal uncle' < Pre-PFi *cecä. Notably, UEW 34-35 reconstructs the PU form as *cecä, and assumes that the forms with two identical affricates arose through assimilation; this does not seem plausible, however, as in that case an identical assimilation process would have to be postulated for Mordvin, Mari, Udmurt, Mansi and Samoyed.

Previously SaaL såsså 'beam (for hanging nets to dry on)' and PSlk *coc? (SlkTa tot[asymptotically =]i, SlkTy co?, SlkK co??) 'pole' have been presented as possible cognates of MdE cocko, MdM socka (UEW 61). SaaL såsså would presuppose a PSaa form *sos≠e, which renders the comparison entirely impossible: PSaa *s developed from PU *s and *s, and it does not reflect an earlier affricate. The vowel of the Selkup word is also slightly irregular: PSlk *coc? would presuppose a PSam form *coc?, but PSam *-o- is not a regular reflex of PU *-u-; one would expect a PSam form *cuc?. However, the correspondence is otherwise quite precise; the development PU *ck > PSam *c is regular, cf. PU *pucki 'tube; Angelica plant' > PSam *puc? (UEW 397- 398; Janhunen 1981 : 236). Hence, it seems plausible that also PSlk *coc? belongs in this cognate set despite its deviant vowel.

5. SaaN faððut 'beat, lash, slap' ~ Komi vi(j)-, Udm vij[asymptotically =]i-, KhE wel-, MsE äl-, Hung öl 'kill'

< PU *widi- 'beat; kill'

The Saami verb *v≠edo- is attested in North Saami and in the Eastern Saami languages; in more western languages no cognates are found. The f- in SaaN faððut is a secondary development motivated by foreign influence; the phoneme /f/ became established in North Saami through Scandinavian borrowings, and has since spread to many native words as well through an irregular change *v- > f- (Korhonen 1981 : 132). The original v- is preserved in Eastern Saami, e.g. SaaSk vââððad 'beat, lash, slap'. The meaning of the verbs is the same in all Saami languages: 'beat (with a supple object), lash, whip; slap (with the hand)'.

The Saami word can be analyzed as a reflex of PU *widi- 'kill'. The previously known members of this cognate set are Komi vi(j)-, Udm vij[asymptotically =]i-, KhE wel-, w l-, KhS, KhN wet- (< PKh *wäl-),1 MsN al-, MsE, MsW, MsS äl- (< PMs *äl-), and Hung öl, all with the meaning 'kill' (UEW 566-567). Sammallahti (1988 : 551) has reconstructed the PU form as *wilä- and proposed another Saami cognate for these words, SaaN fallehit 'attack'. This comparison needs to be abandoned for phonological reasons, however, as the Permic forms show that the verb must have originally had the consonant *-d-; intervocalic *-l- is preserved as such in Permic, whereas *-d- > Ø is a regular development. Moreover, this etymology is not even semanti- cally satisfying, as fallehit 'attack' is a causative derivative of fallet 'jump up, leap up', which in turn is a derivative of falli 'fast (runner, skier, etc.)'. Sammallahti (1998 : 124) has later abandoned the comparison.

Some of the verbs meaning 'kill' also display other meanings which come very close to the Saami word family. In Komi dialects the verb vi(j)- also has the meanings 'hit' (Sysola, Luza, Letka), 'thrash, give a thrashing' (Letka, Içzma). As for Udm vi[asymptotically =]i-, the meaning 'whip' has been attested in the Jelabuga dialect. In Mansi the verb also has the meanings 'hunt, fish, catch (game, fish); hit, wound, torment'. Hence, the comparison of PSaa *v≠edo- to the words discussed above is semantically quite precise. It can be added that an etymological connection between verbs meaning 'kill' on the one hand and 'hit, strike, beat' on the other is very common. Parallels could be cited almost endlessly, beginning with English slay ~ German schlagen, Swedish slå 'hit, beat'. Other examples discovered through a quick search are Fi tappa- 'kill; (dial.) thresh' ~ MdE, MdM tapa- 'beat, crush, trample' (UEW 509); MdE kera- 'hit, chop off, fell down; (dial.) kill' (< PU *kirä- 'hit'; UEW 666); MdE cavo-, MdM sav?- 'strike, hit, beat; kill' (< PU *cana-; UEW 53-54; Aikio 2002 : 11-12); Tok Pisin kil-im 'hit, beat' (< English kill; -im is a marker of transitive verbs); Hawaiian pepehi 'beat; kill' (Pukui, Elbert 1986 s.v.). Numerous further examples from Oceanic languages are cited by Ross, Pawley and Osmond (1998 : 267-273).

The equation is also phonologically straightforward. The predictable reflex of PU *widi- is PSaa *v≠ed≠e-, and the second-syllable labial vowel in the form *v≠edo- is a derivational suffix that forms frequentative verbs. UEW 566 reconstructs the proto-form as *wedV- instead of *widi-, but the reason for postulating *-e- remains unclear, as none of the reflexes actually point to this vowel. The regular reflexes of PU *e in Permic and Ugric are entirely different from those attested in this cognate set; the vowels in this cognate set clearly point to an original *i instead (Sammallahti 1988 : 504, 530- 533).

6. SaaN guodja 'flower stalk or seed shell of a sedge' ~ Komi, Udm k[asymptotically =]i 'awn' ~ MsE koj 'hair (on the head), tuftof hair, mane', Hung haj 'hair' ~ SlkTa qu 'slender object'

< PU *kaji 'grass, stalk?'2

SaaN guodja has the following cognates in other Saami languages: SaaS goeje 'sedge with a seed shell; leaf-bud', SaaP guojjâ 'cottongrass (Eriophorum)', SaaL guodja ~ guovva 'sedge with a seed shell', SaaI kuojâ 'sedge' (< PSaa *kuoj≠e).

SaaS goeje (along with the verb goejedh 'strip tree bark') has been equated with Fi (dial.) koja 'tree bark', Komi k≠eja 'bracket fungus', KhE koj 'skin on the face or on the forehead (of an animal; used for making shoe soles)', and Hungarian héj 'skin, shell, rind, crust'; the proto-form of the words has been reconstructed as *koja (UEW 166). This etymology must, however, be rejected due to the irregular sound correspondences: the vowels of PSaa *kuoj≠e could not regularly reflect the reconstructed proto-form *koja, and the Permic and Ugric forms cannot be regularly derived from such a form either (Sammallahti 1988 : 504, 530-531). Moreover, the semantic correspondences are not satisfying either, especially regarding the Saami noun for which the primary meaning 'sedge' or 'seed shell of a sedge' must be assumed.

Instead, the Saami words can be further equated with Komi and Udm k[asymptotically =]i 'awn', and a proto-form *kaji or *koji can be reconstructed for the words. The phonological development is the same as in PU *ka/oji 'dawn' > Komi k[asymptotically =]ia ~ k[asymptotically =]iva and PU *koji-ra 'male' > Komi k[asymptotically =]ir 'male dog' (Aikio 2002 : 235- 236; UEW 168-169). Formerly the Permic word meaning 'awn' has been compared to Fi käpy 'cone' (SSA s.v.), but this equation is not phonologically satisfying: Finnish ä does not regularly correspond to Komi and Udmurt [asymptotically =]i in this kind of environment (Sammallahti 1988 : 530-533; Aikio 2012 : 240).

The word seems to have reflexes in the Ugric languages as well; one can derive from this form Hung haj 'hair (on the head)' and MsE koj 'hair (on the head), tuftof hair, mane' (cf. UEW 854). The Mansi cognate demonstrates that the Uralic form was *kaji, as MsE o (< PMs *a) is the regular reflex of PU *a in an *i-stem (Sammallahti 1988 : 504). From the sequence *-oji- a different vowel developed in Mansi: cf. MsE kuj- 'male' < PU *koji (UEW 166-167) and MsE suj 'sound, noise' < PU *soji (UEW 482-483).

Even though the difference between meanings such as 'sedge', 'awn' and 'hair' may seem substantial, this kind of semantic development has many parallels. A quite exact parallel is provided by Irish folt 'hair', which corresponds to Old Prussian wolti 'ear (of corn)', Lithuanian váltis 'ear (of oats)', Russian ?o?o?? 'fibre; blade of grass; ear (of corn)' (< PIE *walti-). Also Russian ?o?oc and Avestan var?sa- 'hair' are derived from the same Indo-European root, and their exact cognate is Sanskrit valsa- 'branch, bough' (Vasmer 1953-1958 s.v. ?o?oc, ?o?o??); Fi verso 'shoot, sprout' and KhE wär?s 'horse-hair', KhN war?s 'thin rod, shoot of a leaf-bearing tree' have been borrowed from this Aryan word (Parpola 1999 : 201; Joki 1973 : 336). Another good semantic parallel is provided by the creole language Tok Pisin, where the word gras 'hair; grass' (< English grass) is used in such expressions as gras nogut 'weed' (nogut 'bad'), gras bilong het 'hair on the head' (het 'head'), gras bilong fes 'beard' (fes 'face'), gras bilong pisin 'feathers' (pisin 'bird') (Romaine 2003 : 600). Yet another example is Icelandic hár 'hair', which also has the secondary meaning 'blade of grass', and a closely resembling semantic development is also encountered in Hawaiian oho 'hair (on the head); leaves (of plants); fronds (of ferns)' (Pukui, Elbert 1986 s.v.).

Moreover, the etymology also has a Samoyed cognate which bridges the semantic gap between Saami-Permic and Ugric: PSlk *qu > SlkTa qu 'stalk, stem, slender object'. The word is used in reference to both 'hair' and 'grass' as well as other kinds of slender objects: nut[asymptotically =]it qu 'a blade of grass' (nut[asymptotically =]i 'grass'), opt[asymptotically =]it qu '(a single) hair on the head' (opt[asymptotically =]i 'hair on the head'), polÍ qu 'beam, timber' (po 'tree, wood'), tulÍpot qu 'conifer needle' (tulÍpo 'conifer branch'). The Selkup word apparently reflects PSam *kå?; the phonological development of the word in Samoyed is discussed in Aikio 2012 : 245.

7. SaaN guolmmas 'soft, white inner bark of conifers' ~ MariE kum?z 'birch bark'

< PU *kolmis 'tree bark'

Cognates of SaaN guolmmas are found in all Saami languages except for Ter Saami. The meaning is nearly everywhere the same - the word refers to the softand white inner bark of conifers from which a kind of flour substitute used to be formerly made. Only SaaS goelmese shows a divergent meaning 'peel (e.g., of potatoes)'. No etymology has been proposed for the Saami word. However, there is a noun in Mari that can be analyzed as cognate: MariE kum?z, MariW k ??m z 'birch bark' (< PMari *kum?z). Semantically the match is straightforward. The development *lm > *m is regular in Mari at least after back vowels, cf. PU *ko/ulmi 'three' > PMari *kum > MariE kum, MariW k ??m (UEW 174). After front vowels the cluster *lm was preserved, though, as in PU *külmä > PMari *kilm? > MariE, MariW k?lm? 'frozen' (UEW 313-314), PU *nälmä > PMari *jilm? > MariE, MariW j?lm? 'tongue' (UEW 663).

The vowel correspondence between Saami and Mari requires closer scrutiny. The Mari word shows the correspondence East u ~ Volga/Upçsa u ~ Northwest o ~ West : MariE kum?z, MariV/U kumuz, MariNw komoz, MariW k ??m z 'birch bark'. This correspondence reflects PMari *u, which is normally a reflex of PU *u. PU *o, in turn, is normally reflected as PMari *u (> u in all dialects) or *o (> o in all dialects). Compare the following examples:

PU *u:

* PU *kunsi- 'urinate' > PMari *kuza- > MariE kuza-, MariV/U kuza-, MariNw koza-, MariW k ??za- (UEW 210)

* PU *lupsa 'dew' > PMari *lups > MariE lups, MariV/U lups, MariNw lops, MariW l ??ps (UEW 261)

* PU *muna 'egg' > PMari *mun? > MariE muno, MariV/U mun?, MariNw mono, MariW m ??n (UEW 285)

* PU *purki 'blizzard' > PMari *purg?ste- > MariE pur??ste-, MariNw por?oste-, MariW p ??r? ste- 'cover with snow; whirl' (UEW 406)

* PU *sula- 'melt' > PMari *sule- > MariE sule-, MariV/U sule-, MariNw sole-, MariW s ?? le- (UEW 450)

* PU *tuli 'fire' > PMari *tul > MariE tul, MariV/U tul, MariNw tol, MariW t ?? l (UEW 535) PU *o:

* PU *kota 'tent, house' > PMari *kud? > MariE kuδo, MariW kuδ ?? 'Mari summer house' (UEW 190)

* PU *ora 'squirrel' > PMari *ur > MariE, MariW ur (UEW 343)

* PU *oksi- 'vomit' > PMari *uks?nca- > MariE uks?nca-, MariW uks nca- (UEW 716)

* PU *soksi 'worm' > PMari *suks > MariE, MariW suks (UEW 764)

* PU *sorwa 'horn' > PMari *sur > MariE, MariW sur (UEW 486)

* PU *sodÍa- 'war' > PMari *sudala- > MariE, MariW suδala- 'curse' (UEW 777)

However, there seems to be a conditioning factor which accounts for PMari *u as a reflex of PU *o. There are three examples of the developluobbal ment PU *o > PMari *u in contexts where the labial stop *p followed the vowel:

* PU *kopa 'skin' > PMari *kuw? > MariE kuwo, MariV/U kuwo (UEW 180)

* PU *koppala 'wood grouse hen' > PMari *kuw?lc? > MariE kuw?lc?, MariV/U kuwulzo (UEW 181)

* PU *koppi- 'get mouldy' > PMari *kup-a 3SG > MariE kupa, MariV/U kupa, MariW k ??pa (UEW 680)

The case of PMari *kum?z is similar, as there is a labial nasal *m following the vowel. Hence, we may assume that the development PU *o > PMari *u took place before *p and *m; there do not seem to be any counterexamples to this sound law.

8. SaaS muhtsies 'slovenly, untidy, messy' ~ Komi m[asymptotically =]iz 'guilt, sin, crime', Udm m[asymptotically =]iz 'an evil spirit', KhE m?c 'sickness; fault', MsE m?s 'hole; injury' < PU *muca 'fault, defect, sickness'

SaaS muhtsies has no cognates in other Saami languages, but it would regularly reflect a PSaa form *moces. The final *-s can be interpreted as an adjective suffix, even though the root *moce- is otherwise unattested. This suffix is very common in Saami adjectives, and it has been secondarily attached to many adjectives of Uralic origin: cf. SaaN njuoskkas 'wet' < PU *nacki (> MdE nacko, MariW nack 'damp, wet', KhN nasa 'raw'; UEW 311), SaaN oðas 'new' < PU *wudÍi (> Fi uusi, MdM od, Komi, Udm v[asymptotically =]ilÍ, Hung új 'new'; UEW 587), SaaN vuonjas 'tame (of birds)' < PU *[asymptotically =]ini (> KhE ån[asymptotically =]i 'tame (of birds)', NenT nij? 'tame'; UEW 340).

The PSaa root *moce- presupposes a Uralic proto-form *muca-. Such a word-stem has, in fact, been reconstructed on the basis of other branches. This form is reflected in Komi m[asymptotically =]iz 'guilt, sin, crime', Udm m[asymptotically =]iz 'an evil spirit that brings sickness; sickness (caused by an evil spirit)', KhE m?c, KhN mus, mos 'sickness; fault (in an object)' (< PKh *m ??c),3 MsE m?s, MsN mos 'hole; physical injury', MsW mos?n 'sick' (< PMs *mas) (UEW 283). The Saami word can be semantically quite naturally included in this set: the previously attested cognates show meanings related to some kind of weakness, defect, fault or sickness, and also the meaning 'slovenly, untidy, messy' can be counted in the same semantic field.

MariE muzo, MariW m ??z 'sickness; an evil spirit that causes sickness' (< PMari *muz?) has been equated with the Permic and Ob-Ugric words cited above (UEW 283). This equation is phonologically irregular, however, as PU *c has been regularly preserved as an affricate in Mari. Instead, the Mari word must have been borrowed from Permic, where an assibilation of the affricate (*c > *? > *z) had occurred. Also Hung -máz in the compound hagy-máz 'typhoid, severe fever (with hallucinations)' has been mentioned as an uncertain reflex of PU *muca (UEW 283). This comparison is, however, also irregular because the development *c > Hung z would contradict known sound laws. The origin of Hung -máz remains unclear.

9. SaaS vïekedh 'grab, take hold of' ~ KhE we(j)-, MsE wi(?)- : wäj-, Hung vësz 'take'

< PU *wexi- 'take, grab'

SaaS vïekedh has no cognates in other Saami languages, but it would regularly reflect a PSaa form *viek≠e-, which would further presuppose PU *wäki-, *wäxi- or *wexi-. The last of these forms gives reason to compare the word with a semantically very close Ugric etymological set: KhE, KhS we(j)-, KhN wu- 'take (= nehmen); buy' (< PKh *wi(j)-), MsN wi(?)- : woj-, MsE wi(?)- : wäj-, MsW wi- : wäj-, MsS wü- (< PMs *wi(?)- : *wäj-) 'take', Hung vësz (vëv-) 'take (= nehmen); buy'. The proto-form of the Ugric verbs can be reconstructed precisely as *wexi-; the Ob-Ugric vocalism is the same as in KhE, KhS me(j)-, KhN ma-, mij- (< PKh *mi(j)-), MsN mi(?)- : maj-, MsE mi(?)- : mäj-, MsW mi- : mäj-, MsS mä(j)- 'give' (< PMs *mi(?)- : *mäj-) < PU *mexi- (> SaaSk miõkkâd, Fi myy-, MdM mij?- 'sell', Ngan miési 'give') (UEW 275). Hence, one can reconstruct a PU verb *wexi- 'take, grab', rhyming with *mexi- 'give, sell'.

Previously the Ugric verbs have been included in a different but both semantically and phonologically very close Uralic cognate set. They have been considered the reflexes of PU *wixi- 'take (= wegbringen, führen)' and thus cognate with SaaSk viikkâd, Fi vie-, MdE vije-, MdM vij?- 'take (= wegbringen, führen)', Hung visz (viv-) 'take (= wegbringen, führen), carry, transport' (UEW 573). However, there seem to have been two phonologically and semantically close but distinct Uralic verbs: PU *wexi- 'take (= nehmen)' and *wixi- 'take (= wegbringen, führen)'. This is most clear in the case of Hungarian, where both verbs have been retained: Hung vësz 'take (= nehmen), buy' (< PU *wexi-) vs. visz 'take (= wegbringen, führen), carry, transport' (< PU *wixi-). Contrary to what is maintained by etymological dictionaries (e.g., UEW 573; BenkéÓo 1992-1997 s.v. visz; SSA s.v. viedä), there seems to be no obvious reason to assume that these two distinct Hungarian verbs reflect the same Uralic proto-form. In order to advance such a hypothesis, it should at least be explained how the assumed split into two entirely distinct verbs has come about in Hungarian. UEW merely obscurely states that Hung vësz is a variant of visz, and BenkéÓo 1992-1997 speaks of lexical split ('Wortspaltung'), but neither statement actually provides any explanation as to how the difference between two such semantically very basic verbs would have arisen.

It is easier to assume that Hungarian has preserved an original Proto- Uralic lexical distinction in the verbs visz and vësz. Further evidence for this interpretation is found in Saami. Even though both verbs are not attested in the same modern Saami language, there is a clear phonological distinction between these two Uralic items: SaaS vïekedh 'grab' (< PSaa *viek≠e- < PU *wexi-) vs. SaaSk viikkâd 'take (= wegbringen, führen)' (< PSaa *vik≠e- < PU *wixi-). Moreover, in the old South Saami literary language there still was a verb wiike-, wijke- 'take (= wegbringen, führen)'; this is a regular cognate of SaaSk viikkâd (< *vik≠e-), and would correspond to *vïjkedh in modern South Saami.

Due to their phonological and semantic closeness it is not surprising that the PU verbs *wexi- 'nehmen' and *wixi- 'wegbringen, führen' have become intermixed in etymological references. However, there is a consisluobbal tent semantic difference between the reflexes of *wexi- 'nehmen' and *wixi- 'wegbringen, führen' in nearly all attested cognates. The only exception is Mansi, where the same verb has both meanings. However, this has a straightforward phonological explanation. PMs *wi(?)- : *wäj- belongs to a small group of verbs which show morphophonological alteration between a high and a low vowel. The verbs showing the alteration *i : *ä regularly developed from Uralic verb stems of the shape *Cexi-: cf. PMs *mi(?)- : *mäj- 'give' < PU *mexi- (UEW 275), PMs *li(?)- : *läj- 'throw, shoot' < PU *lexi- (UEW 247). However, the reflex of Uralic verb stems of the shape *Cixi appears to be an unaltering PMs *Cäj-stem: cf. MsN kaj-, MsE, MsW k?j-, MsS kij- 'court (of birds); shamanize' < PMs *käj- < PU *kixi- 'be in heat' (UEW 143). Hence, the expected developments would have been PU *wexi- > PMs *wi(?)- : *wäj- 'nehmen' and PU *wixi- > PMs *wäj- 'wegbringen, führen'. As regular development would have caused the paradigms of the two verbs to partially merge in Proto-Mansi, it is only natural that they analogically fused into a single verb which exhibits both meanings.

It can be added that PU *wexi- and *wixi- may have distinct reflexes in Samoyed as well. The latter of these developed into PSam *ü- (< Pre-Proto- Samoyed *wü- < *wi-), which is reflected in SlkTa ?- 'pull, drag' (Helimski 1999). However, there is also a PSam verb *i- 'take' (> SlkTa i- 'take (= nehmen)', Kam i- 'take (= nehmen), take away; get, receive'). This could perhaps be the reflex of PU *wexi- 'nehmen', but this is uncertain because the loss of *w- without a labialization of the following vowel remains problematic; cf. also PU *weti 'water' > PSam *wet > SlkTa üt, Kam b"u.

10. SaaN vuoððu 'bottom; layer (under something), foundation; sole (of a shoe)' ~ ? Fi vuode 'bed', Komi volÍ 'reindeer hide', Udm walÍ[asymptotically =]i- 'spread; make the bed', MsS alÍat, Hung ágy 'bed'

< PU *adi- 'bed; spread'

SaaN vuoððu has cognates in all Saami languages (Lehtiranta 1989 : 150), all of which show the same basic meaning. The PSaa form can be reconstructed as *vuodo. This can be further analyzed as a reflex of the previously reconstructed PU stem *adÍi- 'bed; spread',4 which is reflected in Komi volÍ 'reindeer hide (for sleeping on it)', Udm walÍ[asymptotically =]i- 'spread out; make (the bed)', MsS alÍat, Hung ágy 'bed' (UEW 4). Also Fi vuode 'bed' may be of the same origin, unless it is instead a derivative of vuota 'hide', which is a Baltic loan (cf. Lithuanian óda 'skin, hide'). SSA (s.v. vuode) supports the latter alternative, but it seems difficult to exclude either of these etymologies, as both are semantically and phonologically quite impeccable.

Regardless of whether Fi vuode belongs in this Uralic cognate set, the Saami word *vuodo can be added to it as a new member. The comparison is phonologically quite unproblematic. PSaa *vuodo can be interpreted as a nominal derivative of a lost verb stem *vuod≠e-, which would be the regular reflex of PU *adÍi-. Semantically the Komi derivative volÍ≠es 'shoe sole' comes especially close, as the same meaning is widely attested in Saami. Parallels for the assumed development 'bed' > 'bottom, foundation' are found in Germanic: cf. English layer ~ German lager 'camp; storeroom; bed', Gothic ligrs 'bed' (Kluge s.v. lager), and English bed, which also has the meanings 'bottom (of a body of water)', 'foundation, base', 'layer, stratum'.

Sidebar

...

Footnote

1 The Proto-Khanty vocalism applied in this paper follows Zhivlov's reconstruction (Wivlov 2006), which is slightly modified from that of Helimski (2001); there are major differences to the Proto-Khanty vowel system proposed by Honti (1982).

2 This etymology was briefly suggested in Aikio 2012 : 245, but without detailed semantic and phonological argumentation.

3 The PKh vowel * is not a regular reflex of PU *u (Sammallahti 1988 : 504), but the expected PKh vowel *a (= *o in Sammallahti's reconstruction) occurs in V Vj moè..l.-, Trj moè..l.- 'be sick; cry for someone (e.g. a child for mother)' (< PKh *maè..l.-).

4 Sammallahti (1988 : 542) reconstructs a Finno-Ugric long vowel for this word (*odÍi), but as argued by Aikio (2012), Finnic long vowels can be analyzed as a result of secondary development, and there are not sufficient grounds for reconstructing long vowels into proto-languages older than Pre-Proto-Finnic.

References

S O U R C E S O F L E X I C A L D A T A

A l a t a l o, J. 2004, Sölkupisches Wörterbuch aus Aufzeichnungen von Kai Donner, U. T. Sirelius und Jarmo Alatalo, Helsinki (LSFU 30).

B e r g s l a n d, K., M a t t s s o n M a g g a, L. 1993, Åarjelsaemien-daaroen baakoegæ rja - Sydsamisk-norsk ordbok, [Lakselv].

F o k ó s - F u c h s, D. R. 1959, Syrjänisches Wörterbuch, Budapest.

G r u n d s t r ö m, H. 1946-1954, Lulelapsk ordbok, Uppsala (Skrifter utgivna genom dialekt- och folkminnesarkivet i Uppsala. Ser. C:1).

H. Paasonens Mordwinisches Wörterbuch. Zusammengestellt von Kaino Heikkilä. Bearbeitet und herausgegeben von Martti Kahla, Helsinki 1990-1999 (LSFU 23. Kotimaisten kielten tutkimuskeskuksen julkaisuja 59).

H e l i m s k i, E. 1997, Die matorische Sprache. Wörterverzeichnis, Grundzüge der Grammatik, Sprachgeschichte. Unter mitarbeit von Beáta Nagy, Szeged (Studia Uralo-Altaica 41).

H e r r a l a, E., F e o k t i s t o v, A. 1998, Moksalais-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 58).

I t k o n e n, E. 1986-1991, Inarilappisches Wörterbuch, Helsinki (LSFU 20).

I t k o n e n, T. I. 1958, Koltan- ja kuolanlapin sanakirja, Helsinki (LSFU 15).

Kai Donners KamassischesWörterbuch nebst Sprachproben und Hauptzügen der Grammatik. Bearbetet und herausgegeben von A. J. Joki, Helsinki 1944 (LSFU 8).

L e h t i s a l o, T. 1956, Juraksamojedisches Wörterbuch, Helsinki (LSFU 13).

M a k s i m o v, S., D a n i l o v, V., S a a r i n e n, S. 2008, Udmurttilais-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 79).

M i k o l a, T. 1985, Morphologisches Wörterbuch des Enzischen, Szeged (Studia Uralo-Altaica 36).

M o i s i o, A., S a a r i n e n, S. 2008, Tscheremissisches Wörterbuch, Helsinki (LSFU 32).

N i e m i, J., M o s i n, M. 1995, Ersäläis-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 48).

O r s z á g h, L., M a g a y, T. 2006, Angol-magyar nagyszótár, Budapest.

S a m m a l l a h t i, P., M o r o t t a j a, M. 1993, Säämi-suomâ sänikirje - Inarinsaamelais- suomalainen sanakirja, Ohcejohka.

S a m m a l l a h t i, P. 1989, Sámi-suoma sátnegirji - Saamelais-suomalainen sanakirja, Ohcejohka.

S t e i n i t z, W. 1966-1993, Dialektologisches und etymologisches Wörterbuch der ostjakischen Sprache 1-15, Berlin (= DEWOS).

Suomen sanojen alkuperä. Etymologinen sanakirja 1-3, Helsinki 1992-2000 (SKST 556; Kotimaisten kielten tutkimuskeskuksen julkaisuja 62) (= SSA).

U o t i l a, T. E. 1942, Syrjänischer Wortschatz nebst Hauptzügen der Formenlehre. Aufgezeichnet von Yrjö Wichmann, Helsinki (LSFU 7).

Wogulisches Wörterbuch. Gesammelt von Bernát Munkácsi, geordnet, bearbeitet und herausgegeben von Béla Kálmán, Budapest 1986.

Wotjakischer Wortschatz. Aufgezeichnet von Yrjö Wichmann. Bearbeitet von T. E. Uotila & Mikko Korhonen. Herausgegeben von Mikko Korhonen, Helsinki 1987 (LSFU 21).

... B. B. 2005, Ce??κ??cκo-p?ccκ?? ??a?eκ???? c?o?ap?, ToMcκ.

...

...

...

H e l i m s k i j E. 2007 [unpublished manuscript], ...

R E F E R E N C E S

A i k i o, A. 2002, New and Old Samoyed Etymologies. - FUF 57, 9-57.

_____ 2006, New and Old Samoyed Etymologies II. - FUF 59, 5-34.

_____ 2012, On Finnic Long Vowels, Samoyed Vowel Sequences, and Proto-Uralic *x. - Per Urales ad Orientem. Iter polyphonicum multilingue. Festskrifttillägnad Juha Janhunen på hans sextioårsdag den 12 februari 2012, Helsinki (MSFOu 264), 227-250.

B e n k éÓo, L. 1992-1997, Etymologisches Wörterbuch des Ungarischen, Budapest.

H ä k k i n e n, J. 2007 [unpublished], Kantauralin murteutuminen vokaalivastaavuuksien valossa. (Graduate thesis, University of Helsinki. https://www. doria.fi/handle/10024/7044)

H e l i m s k i, E. 1999 [unpublished], Basic Vocabulary in PU and PFU - Remarks on Etymology and Reconstruction [A handout delivered at the Budapest Uralic Workshop 2: Etymology, September 1999, organized by the Nyelvtudományi Intézet, Budapest].

_____ 2001, Ablaut als Umlaut im Ostjakischen: Prinzipien und Grundzüge der lautgeschichtlichen Betrachtung. - Fremd und eigen. Untersuchungen zu Grammatik und Wortschatz des Uralischen und Indogermanischen in memoriam Hartmut Katz, Wien, 55-76.

H o n t i, L. 1982, Geschichte des obugrischen Vokalismus der ersten Silbe, Budapest.

I t k o n e n, E. 1946, Zur Frage nach der Entwicklung des Vokalismus der ersten Silbe in den finnisch-ugrischen Sprachen, insbesondere im Mordwinischen. - FUF XXIX, 222-337.

_____ 1954, Zur Geschichte des Vokalismus der ersten Silbe im Tscheremissischen und in den permischen Sprachen. - FUF XXXI, 149-345.

J a n h u n e n, J. 1977, Samojedischer Wortschatz. Gemeinsamojedische Etymologien, Helsinki (Castrenianumin toimitteita 17).

_____ 1981, Uralilaisen kantakielen sanastosta. - JSFOu 77, 219-274.

J o k i, A. J. 1973, Uralier und Indogermanen, Helsinki (MSFOu 151).

K l u g e, F. 1995, Etymologisches Wörterbuch der deutschen Sprache. Bearbeitet von Elmar Seebold. 23., erweiterte Auflage, Berlin-New York.

K o r h o n e n, M. 1981, Johdatus lapin kielen historiaan, Helsinki (SKST 370).

L e h t i r a n t a, J. 1989, Yhteissaamelainen sanasto, Helsinki (MSFOu 200).

M a y r h o f e r, M. 1986-2001, Etymologisches Wörterbuch des Altindoarischen, Heidelberg.

P a r p o l a, A. 1999, Varhaisten indoeurooppalaiskontaktien ajoitus ja paikannus kielellisen ja arkeologisen aineiston perusteella. - Pohjan poluilla: suomalaisten juuret nykytutkimuksen mukaan, Helsinki (Bidrag till kännedom av Finlands natur och folk 153), 180-206.

P u k u i, M. K., E l b e r t, S. H. 1986, Hawaiian Dictionary. Revised and Enlarged Edition, Honolulu.

R e s h e t n i k o v, K., Z h i v l o v, M. 2011, Studies in Uralic Vocalism II. Reflexes of Proto-Uralic *a in Samoyed, Mansi and Permic. - Journal of Language Relationship / Bo?poc? ???κo?o?o po?c??a 5, 96-109.

R o m a i n e, S. 2003, Germanic Creoles. - The Germanic Languages, London- New York, 566-603.

R o s s, M. D., P a w l e y, A., O s m o n d, M. 1998, The Lexicon of Proto- Oceanic. Volume 1. Material Culture, Canberra.

S a a r i k i v i, J. 2007, Uusia vanhoja sanoja. - Sámit, sánit, sátnehámit. Riepmoçcála Pekka Sammallahtii miessemánu 21. beaivve 2007, Helsinki (MSFOu 253), 325-347.

S a m m a l l a h t i, P. 1988, Historical Phonology of the Uralic Languages with Special Reference to Samoyed, Ugric and Permic. - The Uralic languages. Description, History and Foreign Influences, Leiden-New York-København- Köln, 478-554.

_____ 1998, The Saami Languages. An Introduction, Kárásjohka.

V a s m e r, M. 1953-1958, Russisches etymologisches Wörterbuch, Heidelberg.

... M. 2006 [unpublished],

...