Printer Friendly

Studies in Uralic etymology II: Finnic etymologies.

1. Fi aita'fence' ~ KhE at 'fence' < PU *ajta'fence'

Fi aita 'fence' has cognates in all Finnic languages except for Livonian, and its PFi form can be reconstructed as *aita. The word has no known etymology beyond Finnic (SSA s.v. aita). It has remained overlooked, however, that there is a strikingly similar noun in Khanty with the same meaning: KhE at', at', KhS of, KhN js'fence, enclosure'< PKh *ac.

The vowel correspondence PFi *a ~ PKh *a is quite regular, so to prove the comparison we need to establish that PKh *c can reflect an earlier cluster *-jt-. First, one can note that there seem to be no Khanty word roots with a cluster *-jt-, nor with the sequence *-jet- which would be the expected result if the cluster had been broken by the addition of an epenthetic schwa. This shows that if the cluster *-jt- occurred in Pre-PKh, it must have developed into something else in Khanty. That the reflex of Pre-PKh *-jt- is PKh *c is demonstrated by the following two Ob-Ugric comparisons that seem to have remained unnoticed by previous research:

* MsN xujt-, MsE kujt- 'agitate, lure, tempt'(< PMs *kujt-) ~ KhE kut-, KhS xut'-, KhN xus- 'lure, tempt'(< PKh *kuc-).--The Khanty verb has been considered cognate with SaaN gohccut'order, call' and Fi kutsua'call, invite', but apparently erroneously (UEW 192; SSA s.v. kutsua). The vowel correspondence between the Saami-Finnic verb and PKh *kuc- is irregular, and the previously overlooked Mansi cognate shows that medial consonant (*c < *jt) does not match either.

* MsS wajten 'beautiful'(< PMs *wajten) ~ KhE wit', KhS wet', KhN wes 'beauty'(< PKh *wec), KhE wit'en, KhS wet'an, KhN wesan 'beautiful' (< PKh *wecan)--Note also MsN wes 'beauty', wesen 'beautiful', which are Khanty loanwords.

Thus, PKh *ac 'fence'can be considered a regular cognate of PFi *aita 'fence', and the PU noun *ajta 'fence' can be reconstructed. No cognates seem to be found in other branches, but the fully regular sound correspondence and the identical meanings of the compared forms leave little room for doubt regarding the validity of the etymology.

2. Fi ammottaa 'gape open'~ MariW omesta- 'yawn'~ PSam *ammV- 'yawn' < PU *ammV- / *anmV- 'yawn'[left arrow] *ani 'mouth, opening'

Fi ammottaa gape open has cognates in southern Finnic: Est ammutama 'gape open; open one's mouth wide'and Liv amtlo'yawn'. The PFi shape of the verb can be reconstructed as *ammo-tta-; the Livonian cognate reflects a frequentative derivative *ammo-tt-ele-. In Finnish there is also a noun *ammo in expressions such as olla ammollaan'be wide open, be agape'; this may be a back formation based on the verb. SSA (s.v. ammottaa) equates the Finnic words with SaaSk ammsed ~ aammsed and MariW omesta- 'yawn', and characterizes the words as "descriptive-colored" (Finn. "deskr.-savyinen"). It is also mentioned that there are similar words in Samoyed languages, but they are not considered cognate with the Finnic word.

The Skolt Saami word must, however, be a loan from Finnic due to its vowel a'- (< PSaa *a-); in an inherited cognate one would expect a development PU *a- > PSaa *vuo-. The sense 'yawn'in Saami seems to show that this meaning was once more widespread in Finnic, even though it is now only found in Livonian. The equation of Finnic *ammo-tta- and MariW omesta- is, however, phonologically and semantically flawless and one can thus reconstruct a common root *ammV- for the Finnic and Mari verbs; the derivational suffixes in the items differ. Also some of the Samoyed items mentioned by SSA can be quite naturally included in this etymology: Slk *ammu- and Kam amoi- 'yawn' can be straightforwardly equated with *ammV-reconstructed on the basis of Finnic and Mari. The Selkup and Kamas words have a further cognate in northern Samoyed: Ngan yamialasa 'yawn'.

Thus, one could reconstruct a PU verb *ammV- 'yawn'. Ultimately, however, this verb very probably is a derivative of PU *ani 'opening, mouth' (> Komi vom, Udm im, KhE on, PSam *an) (UEW 11-12). According to Alatalo (2004 : 46), Slk *ammu- is a derivative of Slk *an 'mouth', and hence *ammV- must reflect an earlier form *an-mV-. Such an analysis is supported by another Samoyed cognate set, NenT nankem-, SlkK angu- 'yawn'(< PSam *ankV-), which can be explained as a parallel derivative of the same noun. But importantly, the Finnic and Mari cognates of PSam *ammV- show that this derivative was already formed in Proto-Uralic. Also the simplification of the consonant cluster (*-nm- > *-mm-) may have occured at this stage already, but it may have taken place independently in the daughter branches as well.

3. Fi kaiho 'longing, yearning '~ MdE kaz 'bad, poor; accident, misfortune ' ~ Komi kiz 'stillborn child', Udm kiz 'sickness; a spirit that causes sickness '~ MsN xojt- 'be sick '~ PSam *kajte 'sickness'

< PU *kajsV 'sickness'

In standard Finnish kaiho is mainly known in the meaning longing, yearning, wistfulness, but in dialects the word also has meanings such as sorrow, need, lack, worry, fear, temper, anger, and envy. The word has the following cognates in other Finnic languages: Kar kaiho grief, sorrow; poor thing', Est kahju, Liv ka 'i 'damage, injury, loss, harm '(< PFi *kaiho). No further Uralic cognates have been suggested.

It has been proposed that Fi kaiho is of the same origin as Fi kaihi cataract (SSA s.v. kaihi, kaiho), but the semantic difference makes this etymology difficult to substantiate. Despite the skeptical view of SSA and LAGLOS (Kylstra, Hahmo, Hofstra, Nikkila, 1991-2012, s.v. kaihi), Fi kaihi 'cataract' seems best explained as a loan from Proto-Germanic *xaixaz (> Gothic haihs 'one-eyed '; cognate with Latin caecus 'blind '). This etymology presupposes that there was a period of borrowing when PFi *k- was substituted for Germanic initial while at the same time PFi *-h- was substituted for Germanic medial *-X-. This would be possible if there was some notable allophonic difference between the pronunciation of Germanic inital and medial *X at this point (e.g., *[x] initially vs. *[h] medially). In any case, the assumed substitutions have a plausible parallel in Fi kahle chain (cf. Old High German hahal 'kettle hook '< *xaxlaz < Proto-Germanic *xjanxlaz). Fi haahla 'chain for hanging a caldron over the fire is a later loan from the same word.

Fi kaiho can, however, be equated with MdE kaz, kazo bad, poor; accident, misfortune, MdM kaz bad; weak, thin, wretched (person or animal) . The semantic connection to the Finnic word family is obvious, and the equation is straightforward phonologically as well. PFi *kaiho presupposes a Pre-PFi form *kajsV--(the final -o may be a suffix), and also the Mordvin word can be regularly derived from the same form. The voicing of the sibilant (*s > z) is regular in a voiced context, and the loss of preconsonantal *j appears to be regular as well, at least before coronal consonants. There are not many examples of the latter sound change, as clusters of the type *-jC- seem to have been quite rare in Uralic. However, Aryan loan-words support the assumption of this sound law: cf. MdE vano-, MdM vane- 'look, watch '< Pre-PMd *vajna- < Aryan *vaina- 'see, look, watch MdE sed', M sad' 'bridge '< Pre-PMd *sajti < Aryan *seitu- 'bridge '(Koivulehto 1999 : 230-231). The same development is also attested in some derivatives based on Uralic stems: MdE vade-, MdM vad'e- grease, smear < Pre-PMd *vaj-ta- (cf. MdE oj, MdM vaj 'fat, butter, oil '< *vaj < PU *waji; UEW 578), MdE pid'e-, MdM pid'e- 'cook (tr.), bake '< Pre-PMd *pej-ta-(cf. MdE pije-, MdM pije- 'cook (intr.)'< PU *peji-; UEW 368). Hence, MdE kaz reflects PU *kajsV quite regularly.

Further cognates can be found in Mansi and Samoyed. As both branches show a change *s > *t, the Finnic-Mordvin root can be quite naturally compared to MsN xojt-, MsE kojt-, MsW kwojt- 'be sick '(< PMs *kajt-) and NenT xaedo 'epidemic', EnF kad'e(d ') 'ache', Ngan koca 'sickness', koce- 'be sick', SlkTa quti 'sickness', quti- 'be sick '(< PSam *kajta-; Janhunen 1977: 58-59). PSam *a is a regular reflex of PU *a. According to the Ob-Ugric vowel developments postulated by Sammallahti (1988 : 504) one would, however, expect PMs *u, or perhaps PMs *a if the word was originally an *i-stem (*kajsi). However, the development PU *a > PMs *a is also attested in MsN oj- 'run away '(< PMs *dj-), MsN ojt-, MsE ajt-, MsW ojt-, MsS ajt- 'unbind, unharness; shoot (with an arrow) '(< PMs *aj-t-), which derive from PU *aja- 'drive; escape '(UEW 4-5). Hence, the development PU *a > PMs *a is probably regular before *j). (1)

It appears that also the following Permic words can be derived from PU *kajsV: Komi (obsolete) kiz 'stillborn child', Udm kiz 'sickness; an evil spirit that causes sickness. The consonant correspondences are quite regular: in Permic sibilants became regularly voiced in voiced contexts, and preconsonantal *j was lost as in PU *ajma 'needle '> Komi jem (UEW 22) and PU *kojra 'male '> Komi kir (UEW 168). As regards vocalism, however, the normal reflexes of PU *a are Komi o, u and Udm u (Sammallahti 1988 : 530, 533; Reshetnikov, Zhivlov 2011 : 105-107). But there probably was a conditioned development *-aj- > Komi and Udm i; a parallel is provided by PU *kaji > Komi, Udm ki 'awn '(Luobbal Sammol Sammol Ante (Aikio) 2012 : 245; 2013 : 166-167). Previously Komi and Udm kiz have been compared to Fi kitu- be in severe pain, linger in pain, languish and KhE, KhS keca, KhN kasi 'sickness '(< PKh *kici), but the vowel correspondences are not regular. Moreover, the Khanty word is quite evidently a loan from Permic *kiz. The vowel substitution PPerm *i (Komi i) > PKh *i is attested in several dozen Permic borrowings (Toivonen 1956 : 138; PKh *i = *a in Toivonen s notation). Consider the following examples:

* KhE, KhS ker 'mortar '< PKh *kir < PPerm *gir (> Komi, Udm gir mortar)

* KhE, KhS mert, KhN mart 'with difficulty '< PKh *mirt < PPerm *mird- (> Komi mirden 'by force', Udm mirden 'hardly; with difficulty')

* KhE, KhS mecak, KhN masek 'fist '< PKh *micak < PPerm *mizik (> Komi, Udm mizik fist)

* KhE, KhS pem, KhN pam 'heat '< PKh *pim < PPerm *pim (> Komi pim hot, burning)

* KhE pera-, KhS per-, KhN par- 'pass (of time); end '< PKh *pira- < PPerm *bir- (> Komi bir-, Udm biri- 'run out; end')

* KhE, KhS ser, KhN sar 'spade '< PKh *sir < PPerm *zir (> Komi zir spade)

The semantic relationship between damage, harm, and misfortune found in Finnic and Mordvin and sickness in Udmurt, Mansi and Samoyed is rather obvious, and English ill serves as a parallel; in addition to meaning 'sick', it may mean e.g. 'very bad '(ill weather), 'unlucky '(ill fate), 'harmful '(ill effect), 'immoral '(ill deed), and 'malevolent '(ill intent). It can be assumed that in Finnic and Mordvin a semantic abstraction from sickness to misfortune, accident, damage took place. Hence, there is no semantic obstacle to reconstructing a PU root *kajsV- 'sickness; be sick'.

Furthermore, Finnic shows a possible trace of the meaning sickness . The root *kajsV- may have yet another uncertain reflex in Finnic. In addition to PFi *kaiho, also an adjective *haikeda with very similar semantics is attested in the Finnic languages: cf. Fi haikea sad, wistful, melancholy, Est haige 'sick'. It appears possible that *haikeda developed through an irregular metathesis from an earlier form *kaiheda; the ending *-eda is an adjective suffix. On the other hand, there are also the similar nouns haika faint smell; rumor and haiku smoke, reek, coal gas; haze, mist; scent, smell', which are cognate with SaaL suojgge 'draft '(< Pre-PSaa *sajka). Hence, it might be that there originally were two distinct words, PFi *haika 'smoke, smell, haze '(< *sajka) and PFi *kaiho 'sickness; sadness; damage, harm '(< *kajsV), but these became folk-etymologically mixed through a metathesis *kaiheda >> *haikeda in Proto-Finnic. This etymology of Fi haikea is, of course, speculative, but this has no bearing on the etymology of Fi kaiho and its cognates in Mordvin, Permic, Mansi and Samoyed.

4. Fi katkera 'bitter', katku 'burnt smell', katketa 'break in two' < PU *kacka- 'bite'

In UEW (113) Fi katkera bitter is considered cognate with SaaN guohca (gen guohccaga) 'rotten, spoiled ' (< PSaa *kuoccek), MariE koco, MariW kace 'bitter ' (< PMari *kace) and Udm kes 'bitter, tasteless, unsalted', possibly also KhE kicim, KhS xecem, KhN jisam 'mold '(< PKh *kicim), MsS kasak, MsW kassa, MsE kesay, MsN xassi 'mold '(< PMs *kisyi); the Uralic protoform is reconstructed as *kacke-. This etymology is unacceptable for phonological reasons, however: none of the proposed cognates of the Finnic word suggest a cluster *-ck-. In Saami the word has a geminate affricate *-cc- (< PU *-cc-), and the single affricates in Mari and Ob-Ugric can be explained as regular results of geminate shortening. Sammallahti (1988 : 552) equates the Saami, Mari and Udmurt forms and reconstructs the proto-form *kac(c)V-. There is no reason to leave the Ob-Ugric words for mold out of this cognate set, however, as they can be regularly derived from the form *kiccV-. The regular reflexes of PU *i are PMs *i and PKh *a. The high unrounded vowel *i in the Khanty cognate is the high ablaut grade of an original *a (Helimski 2001; [TEXT NOT REPRODUCIBLE IN ASCII] 2006 : 42); the ablaut was triggered by the vowel in the suffix *-im. However, Udm kes shows an obscure sibilant as the assumed reflex of an earlier affricate, and also the vowel -e- is irregular; for this reason it is best excluded from the cognate set.

Fi katkera cannot reflect PU *kiccV- due to its consonant cluster -tk-, and it must thus have a different etymology. Curiously, even though it has been hypothetisized that katkera goes back to a proto-form *kacke-, it has apparently never been connected with the reconstructed verb *kacka- (> SaaN gaskit 'bite', MariE kocka-'eat', Komi gecki- 'ruminate, chew the cud '; UEW 641). The derivation of an adjective meaning 'bitter ' from a verb meaning bite is semantically straightforward, and an exact parallel is provided by English bitter (< Proto-Germanic *bitra-), an obscured derivative of bite (< Proto-Germanic *bita-). The only problem in this etymology is the second-syllable vowel -e- instead of expected -a- (*katkara). This slight irregularity need not invalidate an otherwise straightforward etymology, however. The ending -ra can be analyzed as an adjective suffix; even though Finnish adjectives ending in -rA are rare, there are a couple of adjectives derived from Uralic roots with this suffix: Fi kumara 'stooping, bent '< PU *kuma' fall over; knock over '(UEW 201-202), Fi avara 'spacious, wide, open '(< PU *ana- 'open, take off '; UEW 11). A noteworthy case is Fi (dial.) vihera ~ Veps, Vote viher 'green '(< *vihera), derived from PU *wisa- 'green '(UEW 823-824), which serves as a parallel for the irregular change *a > *e in the second syllable of an adjective derived with *-ra. The standard Finnish adjective for 'green', vihrea, is a later irregular development of *vihera; note also dialectal viheria green .

In addition to Fi katkera bitter there is a similar noun katku burnt smell, fumes, stink. This noun, in turn, has been considered cognate with MdE kacamo smoke and Komi kocis burnt smell, and the proto-form is reconstructed as *kackV- in UEW (641). This etymology is not satisfactory because the Mordvin and Komi forms do not support an original cluster *-ck-. Instead, it is tempting to analyze the Mordvin form as a derivative of the PU root *kiccV- 'rotten, moldy ' discussed above. While it might seem far-fetched to include a noun meaning smoke in this cognate set, there is also a parallel verbal derivative with a meaning closer to the Uralic word family: MdE kacado-, MdM kacade- smoke (intr.); smell (intr.)'. A parallel for the connection of words for bad smell and smoke can be found in Germanic: cf. English reek (earlier 'smoke, vapor, mist ') ~ German Rauch, Swedish rok 'smoke'. The concept that connects the Mordvin words to PU *kiccV- 'rotten; moldy' is some kind of unpleasant or bad smell; SaaN guohcistit smell rotten or bad and SaaL guohtsa foul smell come semantically particularly close. Furthermore, there is a previously unnoticed Samoyed cognate that exhibits the meaning of bad smell : SlkTa qeti bad smell, stink', SlkK qeci- 'smell, stink'. These words reflect PSam *kece-, which suits phonologically exactly as the reflex of PU *kiccV-.

As regards Komi kocis, both SSA and UEW present it as an uncertain cognate. As a reason for this uncertainity it is mentioned that the word had Proto-Komi *o in the first syllable, which is the regular reflex of PU *e(-a) (Sammallahti 1988 : 530): cf., e.g., PU *pesa 'nest '> PKomi *poz(j-) (UEW 375), PU *ela- 'live '> PKomi *ol- (UEW 73); PU *ceca 'uncle' > PKomi *coz (UEW 34-35). UEW states the irregularity could be explained by the influence of the following affricate *c, but this is an ad hoc explanation, as there appear to be no parallels for the assumed sound change. Due to the irregular vowel it remains unclear whether Komi kocis has any relation to PU *kiccV- 'rotten, moldy'.

Let us return to Fi katku burnt smell, fumes, stink'. This word can have nothing to do with the PU *kiccV- 'rotten, moldy', as it shows an unmatching consonant cluster -tk-. Fi katku also has a south Finnic cognate which demonstrates that the word originally had an affricate: Est katk, Voro katsk 'plague, epidemic' < *kacku. (2) Semantically it is not at all difficult to connect a noun meaning burnt smell and fumes to a verb meaning bite. One can note that both the Saami and Mari cognates are contextually used in reference to the unpleasant effects of smoke: cf. SaaN suovva gaska calm-miid 'the smoke irritates ("bites") the eyes', MariE sincam siks kockes 'the smoke irritates ("eats") my eyes'. Hence, also Fi katku 'burnt smell, fumes, stink can be included in the same word family.

Also the Finnish verbs katketa (katkea-) break in two (intr.) and katkaista (katkaise-) break in two (tr.) need to be considered in this connection. These also reflect a PFi root *kacke-; the affricate is preserved in Voro katski 'broken '(~ Fi katki 'broken in two '). UEW (641) derives these words from yet another reconstruct *kacka-, this time supposedly reflected also in MariE kuske[??]a-, MariW keske[??]a- 'tear off, tear in two' and Udm kwackatear (intr.)'. At least the Mari verb can have nothing to do with the others, as Mari -sk- does not reflect earlier *-ck-, and also the vowel (MariE u, MariW e < PMari *u) is irregular. Instead, the Mari verb must be a reflex of PU *kisko- 'tear '(> SaaN gaikut, Fi kiskoa 'tear', Komi kos-, Udm kesirip, KhE kos- tear down, break up ; cf. UEW 162; SSA s.v. kiskoa). The development PU *i > PMari *U occurred in disharmonic stems, cf. PU *wisa-ra 'green' > PMari *uzar > MariE uzar, MariW ezar ~ zar (this is cognate with the Finnish formation vihera green discussed above).

Whether Udm kwacka- can be analyzed as a cognate of Fi katketa is not clear. The vowel correspondence is unusual, as in other cases Udm kwa- corresponds to Fi ko-: cf. Udm kwas 'shallow '~ Fi koski 'rapids '(SSA s.v. koski; cf. UEW 674), Udm kwala 'summer cabin '~ Fi kota '(Saami) tent' (UEW 190), Udm kwaldi- 'split' ~ Fi kolo 'hole, hollow' (UEW 174). Hence, Udm kwacka- would be a regular reflex of the form *kockV-, not *kacka-. Regardless of the origin of Udm kwacka-, however, it can hardly be denied that also Fi katketa and katkaista must be related to PU *kacka- 'bite'. Here we seem to have yet a different semantic development in Finnic from bite loose to break in two .

Finally, it can be added that the PU verb *kacka- 'bite' also has previously unnoticed Ob-Ugric reflexes: KhE kic-, KhS xec-, xes-, KhN xis- hurt, ache; sting (of a nettle) 'and KhE kic, KhS xes 'nettle '(< PKh *kic-) and MsS kos-, MsW kus-, MsN xus- 'sting (of a nettle) '(< PMs *kus-). The semantics of these verbs comes very close to the secondary meanings of the Saami reflexes of PU *kacka-: SaaN gaskit, I kaskid 'bite; sting, smart, burn'. It is also noteworthy that SaaN gaskalas nettle is a semantic development of the adjective gaskalas apt to bite, a derivative of the verb gaskit. As regards the historical phonology of the Ob-Ugric forms, the vowel correspondence is regular: the regular reflexes of PU *a(- a) are PKh *a and PMs *u, and the attested PKh *i is the high ablaut grade of *a (Helimski 2001; [TEXT NOT REPRODUCIBLE IN ASCII] 2006 : 42), implying that there has been a lost suffix in the stem that triggered the ablaut. Also the loss of PU *k in the cluster *ck may be regular, but no other examples of the development of this cluster Ob-Ugric seem to be known that would allow this to be verified.

To sum up, it is not justified to reconstruct the four near-homonymous roots *kacka- 'bite', *kacka- 'tear in two', *kacke- 'bitter', and *kackV- 'smoke, burnt smell '(cf. UEW 113, 641--642). Instead, the following two cognate sets can be postulated:

PU *kacka- 'bite ': SaaN gaskit 'bite; sting, smart, burn', Fi katkera 'bitter', katku 'burnt smell, fumes, katketa break in two, MariE kocka- eat', Komi gecki- ruminate, KhE kic- hurt, ache; sting (of a nettle), MsN xus- sting (of a nettle)

PU *kiccV- 'smelly, rotten, moldy ': SaaN guohca 'rotten', MdE kacamo smoke, kacado- smoke; smell, MariE koco bitter, KhE kicim, MsN xassi mold, SlkK qeci- smell, stink

5. Fi korpi 'dense forest, wildwood '~ MdE kuro 'bush, shrub '~ MsW korp 'forest, woods'

< PU *korpi 'woods'

Fi korpi has cognates in most Finnic languages, e.g. Veps korb, Est korb 'woodland, wildwood '(< PFi *korpi), but no generally accepted cognates outside Finnic. SSA (s.v. korpi) maintains that the cognates proposed from more distantly related languages are quite uncertain. UEW (217) notes that Finnish korpi has been compared to MdE kuro, MdM kur(a) bush, shrub; bushes', KhS, KhN xar 'forest (esp. as a hunting ground)', MsW korp 'forest' and NenF kure 'dense forest on the shore of a river', but rejects the inclusion of the Finnish word in this etymological set. Also the equation of the Mordvin, Khanty, Mansi and Nenets words is considered uncertain by UEW.

The equation between the Finnic and Mansi words was first presented, with hesitation, by Liimola (1956 : 243-244). This comparison can be rehabilitated; the sound correspondence between the items is fully regular. On the basis of MsN xcrp, MsW xcrp, MsE korep, MsS kcrp, kcrep a Proto-Mansi form *karp can be reconstructed. As regards vowels, Proto-Uralic *o has developed into PMs *a in stems of the type *(C)oCCi- (Sammallahti 1988 : 504), so the match between PMs *karp and PFi *korpi is fully regular. The consonants in the two forms are identical, and the correspondences *k ~ *k, *r ~ *r and *p ~ *p are of course regular. Even so, UEW states the comparison between Finnic and Mansi is rejectable due to the Finnic consonant cluster *rp. This statement is not logical, as the Mansi word has the same cluster. UEW maintains that Mansi -p is a derivational suffix, but this suggestion is not substantiated in any way. The meanings of the words are nearly identical: the Finnic words generally mean dense forest, wildwood, and in Mansi meanings such as birch forest and fir forest have been attested; the MsS form korp, korap means 'grove'.

On the basis of Finnic and Mansi a Proto-Uralic noun *korpi' forest' can be reconstructed. Also MdE kuro, MdM kur(a) bush, shrub; bushes (< PMd *kure), which was already mentioned above, can be considered a reflex of this Uralic word. Semantically the comparison is quite natural; cf., e.g., MariE codera 'forest, woods' ~ MariW ca[??]ra 'twigs, withered branches'. The vowel correspondence requires more detailed argumentation, however, because according to Itkonen (1946 : 205) the regular reflex of PU *o(-i) is PMd *o, not *u. Itkonen gives eight examples of the correspondence PFi *o(--i/e) ~ PMd *o, on which he bases his assumption of regular development:

MdM Jov (name of the river Moksha) (< PMd *jov) ~ Fi joki 'river'

MdE, MdM jonks (< PMd *jonks) ~ Fi jousi 'bow'

MdE kolmo, MdM kolma (< PMd *kolme) ~ Fi kolme 'three'

MdE loksej, MdM loksti (< PMd *loksten) ~ Fi joutsen 'swan'

MdE nolgo, MdM nolga (< PMd *nolge) ~ Fi nolki 'snot'

MdE ovto, MdM ofta (< PMd *ovte) ~ Voro ots : GEN ote 'bear'

MdE olgo (< PMd *olge) ~ Fi olki 'straw'

MdE pongo, MdM pov (< PMd *pone) ~ Fi povi 'bosom'

Not all of these examples are convincing. As regards Fi joki river and kolme 'three', their Saami cognates point to an original *u instead (SaaN johka 'river' < *juki, golbma 'three' < *kulmi). As the cognates of these words in more eastern Uralic languages also show irregular vocalism (cf. UEW 99, 174; Sammallahti 1988 : 537, 543), the reconstructions *joki 'river' and *kolmi 'three' cannot be substantiated; it is possible that the Mordvin forms derive from *juki and *kulmi instead and thus reflect the regular change PU *u > PMd *o. The words for 'swan' are phonologically even more obscure; Fi *j- ~ PMd *l- is a completely irregular correspondence, and also the other suggested cognates show strange deviations (e.g., SaaN njukca swan < PSaa *nukce has an initial nasal and *u < Pre-PSaa so this etymology cannot serve as an example of regular vowel development. Two of the proposed examples, bear and straw, are only attested in Finnic and Mordvin, and they may have been borrowed between Pre-Proto-Finnic and Pre-Proto-Mordvin.

The only widespread Uralic words not following the sound law *o(- i) > PMd *u are PMd *jonks 'bow '(< PU *jonsi), PMd *pone 'bosom '(< PU *poni), and PMd *nolge 'snot '(< PU *nolki). In contrast, there are at least twelve cases showing the vowel development *o(-i) > PMd *u:

PU *oksinta- 'vomit '> PMd *uksende- > MdE uksno-, MdM uksende (UEW 716)

PU *omti 'cavity, hollow '> PMd *unde > MdE undo, MdM unda (UEW 338)

PU *colmi 'knot '> PMd *sulma > MdE sulmo, MdM sulma (UEW 38)

PU *korpi- 'blaze '> PMd *kurva- > MdE kurva- (UEW 186)

PU *moski- 'wash '> PMd *muska- > MdE muske-, MdM muske- (UEW 289)

PU *soksi 'worm, maggot '> PMd *suks > MdE, MdM suks (UEW 764)

PU *sormi 'finger '> PMd *sur > MdE, MdM sur (UEW 765)

PU *soski- 'chew '> PMd *suska- > MdE susko-, MdM suske- 'bite' (UEW 448-449)

PU *solki 'clasp' > PMd *sulgame > MdE sulgamo, MdM sulgam (UEW 774-775)

PU *sodi- 'leak, flow '> PMd *cude- > MdE cud'e-, MdM sude- (UEW 786)

PU *totki 'tench '> PMd *tutka > MdM tutka (UEW 532)

PU *woli- 'be '> PMd *ul'e- > MdE ul'e-, MdM ul'e- (UEW 580-581)

Thus, the regular development is clearly PU *o(-i) > PMd *u, contra Itkonen (1946). The lack of the change *o > *u in the words *jonks 'bow' (< PU *jonsi) and *pone 'bosom '(< PU *poni) seems to be conditioned by the following velar nasal. There is also a third etymology that serves as an example of this rule: MdE onkstt (pl), MdM ovs ~ ovks 'bit (in bridle) '(< PMd *ons) ~ SaaN vuonas 'muzzle of a dog' (< PSaa * vuones), the protoform of which can be reconstructed as PU *onis. UEW (11) and Sammallahti (1988 : 542) further equate these words with PU *ani 'opening, mouth', but this is not supported by the vocalism of the Mordvin form: the development PU *a > PMd *o would be irregular. This leaves only PMd *nolge snot as an unexplained exception; the word is affective, which might account for its irregularity.

We can now return to the case of MdE kuro, MdM kur(a) bush, shrub; bushes'. The regular vowel correspondence established above supports deriving this word from PU *korpi 'woods'. As for the consonant correspondence, PU *-rp- > Md -r- is apparently not fully regular, as the reflex of this cluster is Md -rv- in at least two cases: cf. PU *korpi- 'scorch '> MdE kurva- (UEW 186), PU *turpa 'lip '> MdE turva, MdM terva (UEW 801). However, the complete loss of *p in the cluster *-rp- is attested in one already well-established etymology: PU *orpas 'orphan '> MdE uros, MdM ures (UEW 343). Thus, there is no obstacle to analyzing MdE kuro, MdM kur(a) as a reflex of PU *korpi. However, the remaining proposed cognates (KhE -kari 'place '(in compounds), KhS jar, KhN xar, xari 'forest (esp. as a hunting ground); clearing '(< PKh *kiri) and NenF kure) show no regular correspondence to PU *korpi; hence, they must be excluded from this etymology.

6. Fi ohut 'thin '~ KhE woyel', MsN wowta 'thin' < PU *woksi- 'thin'

The standard Finnish word for thin is ohut. In dialects also parallel forms with different derivational suffixes are attested, such as ohea (< *oheda), ohukainen and its syncopated variant ohkainen (< *ohukkainen). All three forms have cognates in other Finnic languages as well: cf. Votic ehud (< *ohut) and Est ohe (< *oheda), ohukene (< *ohukkainen) 'thin'. A Proto-Finnic root *ohe- can be reconstructed on the basis of these forms; a similar pattern of parallel derivatives is attested in some other Finnic adjectives as well, e.g. Fi kevyt (< *kebut) ~ kepea (< *kepeda) ~ koykainen (< *kebukkainen) 'light'. The root *ohe- has had no further etymology so far (SSA s.v. ohut). However, it is of Uralic origin: it has fully regular cognates in the Ob-Ugric languages, which have so far remained unnoticed due to the rather non-transparent sound correspondences involved.

Honti (1982 : 193) equates the following Ob-Ugric adjectives: KhE (V Vj) woyel', (Trj) woy ev, KhS woxet, KhN (Ni) uxet, (Kaz) oxel, (O) oxel 'thin (of flat things) '(< PKh *wayal), MsW wayta, MsN wowta 'thin' (< PMs *wayta). These words show no superficial resemblance to PFi *ohe-, but the sound correspondence is actually fully regular. The initial *w- presupposed by Ob-Ugric is lost in Finnic before the vowel *o, as in PU *woli- 'be '> Fi ole- (UEW 580), PU *wolka 'shoulder '> Fi olka (UEW 581) and PU *wosta- buy > Fi osta- (UEW 585). As for the medial consonants, the correspondence PKh *-yel- ~ PMs *-yt- implies Ob-Ugric *-yil-, which can reflect four different PU clusters: *sk, *ks, *sk and *ks. The sibilants *s and *s have merged in Ugric languages, and a regular metathesis also took place in clusters of the type *sk and *sk in Ob-Ugric. The last one of the four alternatives, PU *ks, accounts for the correspondence between Finnic and Ob-Ugric, as the regular reflex of this cluster is *h in Finnic: cf. PU *meksi 'bee' > Fi mehi-lainen (UEW 271), PU *maksa- 'rotten (wood) '> Fi mahi (UEW 698). Hence, a Proto-Uralic root *woksi- 'thin' can be reconstructed on the basis of Finnic and Ob-Ugric.

As regards vowels, PFi *ohe- presupposes a Uralic vowel combination *o - i. PU *o in stems of the type *(C)oCCi- is usually reflected as PKh *a and PMs *a. Examples include:

PU *omti ' cavity '   > PKh *ant, PMs *antar (UEW 338)
PU *ponci ' tail '    > PKh *pac, PMs *pans (UEW 353)
PU *sorsi ' span '    > PKh *saras, PMs *taras (UEW 448)
PU *wonki ' den '     > PKh *wank, PMs *warka (UEW 583)
PU *korpi ' woods '   > PMs *karp (see etymology 5)

The reflexes of *woksi- 'thin', however, display the correspondence PKh *a ~ PMs *a. The difference in Mansi vowel length results from a conditioning factor: if the following consonant was PMs *y, a shortening *a > *a has taken place. Compare the following two parallels:

PU *jonsi ' bow '     > PKh *jayel, PMs *jayt (UEW 101-102)
PU *soski- ' chew '   > PKh *Layel-, PMs *tayt-(UEW 448-449)

The otherwise very common vowel correspondence PKh *a ~ PMs *a apparently never occurs in words where the following consonant is PMs *y; no such case can be found in the list of Ob-Ugric cognate sets presented by Honti (1982 : 123-198). This offers further support for the hypothesis that a secondary vowel shortening *a > *a took place before PMs *y. In addition to the well-established Mansi reflexes of PU *jorsi 'bow' and *soski 'chew', the word *woksi 'thin '> PMs *wayta provides a third example of an Uralic item that follows this sound law.

7. Fi puhjeta (puhkea-) 'burst; open (of flowers), come out (of leaves)' and putkahtaa 'emerge, come up, appear (suddenly), pop up' < PU *pucki 'tube; stalk (?)'

The Fi verb puhjeta (puhkea-) has cognates in all Finnic languages, while the similar word putkahtaa is only attested in Finnish and Karelian (SSA s.v. puhjeta, putkahtaa). SSA hesitatingly equates these words with MdE pockod'e- 'burst', MdM pockod'e- 'burst; open (of sprouts)', Komi and Udm bicki- sting, pierce. In addition, the dictionary describes the verbs puhjeta and putkahtaa as descriptive. This characterization appears misleading, as it remains unclear what sound symbolic conventions might be manifested in these words.

In any case, the equation of the Finnic, Mordvin and Permic words is quite convincing, and the underlying Uralic root can be reconstructed as *pucki-. The dual representation of the consonant cluster (-hk- ~ -tk-) in Finnic appears to result from an irregular split of the preconsonantal affricate *c in Pre-Finnic. The cluster -tk- is the regular reflex of PU *-ck-; cf., e.g., Fi kotka'eagle '(< PU *kocka; UEW 668) and Fi notko 'depression '(< *nocko; UEW 714). The form puhkea- can be explained as the outcome of an irregular assibilation in Pre-PFi (*pucki-ta >> *puski-ta > PFi *puhke-da-). Exactly the same phonological split has occurred in another Finnic word-family as well: Fi potka 'shank' (< *pocka) and Fi pohje (gen pohkeen), dial. pohkea 'calf (of the leg) '(< PFi *pohkeda < Pre-PFi *poskita << *pockita); the word is cognate with SaaN boaski 'ankle (of an animal)', SaaS baetskie 'heel '(< PSaa *poacke) (SSA s.v. potka).

The initial b- in Komi and Udm bicki- is not regular, but there are several other, well-known examples of sporadic voicing of initial stops in Permic: e.g., Komi bez, Udm biz' tail' (< PU *ponci 'tail '; UEW 353), Komi dor, Udm dur 'edge (< PU *tera'edge, blade '; cf. UEW 522, 795), (3) Komi giz, Udm gizi 'nail'(< PU *kunci 'nail '; UEW 157). In the verb bicki- also the affective semantics might have played a role in the voicing.

It is notable that the reconstructed root *pucki- 'burst' is homonymous with another well-established PU root *pucki 'tube; stalk; Angelica (a plant with a hollow stalk)'. This word has widely attested reflexes, e.g., SaaN boska garden Angelica (Angelica Archangelica), Fi putki tube; Angelica, MdE pocko tube; hollow stalk; Angelica, MariE puc stalk; tube, NenT puda 'small metal tube', Slk *puca 'tube; soft inner part of plants; inside; middle', Kam put 'marrow; inside, inner part; groove'. As the reconstructed stems are homonymous, it is in order to examine whether they could also be etymologically identical.

It is noteworthy that the Finnish verb puhjeta also has meanings related to the growth of plants: cf. lehdet puhkeavat the leaves come out, puhjeta kukkaan blossom, open its flowers--literally the leaves burst, to burst into flowers. A similar usage is attested in MdM pocked'e- open (of sprouts)'. This already brings the comparison closer to the word *pucki 'tube; stalk; Angelica'. One can thus reconstruct a derived verb *pucki-ta-open (of sprouts, flowers), blossom, which already at an early stage developed the secondary meaning burst ; as parallels, compare Lithuanian sprogti 'burst; blossom' and Hungarian feslik 'rip (intr.), get torn (e.g., of clothes); blossom. The meaning burst, then, gave rise to its transitive equivalent pierce in Permic.

Notably, also the nominal reflexes of PU *pucki 'tube; stalk' have developed a wide range of derived and more abstract meanings. In the Samoyed reflexes one encounters, in addition to tube and the like, also more abstract meanings connected to inside, inner part of something. This is evident in the Selkup reflexes, for instance; Bykonya 's Selkup dialect dictionary gives the forms ObSh, ObCh, Ty puz, ObS, Vas puze, Tur, Elputi, Ket puca ([TEXT NOT REPRODUCIBLE IN ASCII] 2005 : 197), with the diverse meanings 'inside, interior; soul; stomach; navel; seed; core; pipe (instrument); barrel (of a gun)'. Also the meaning of a type of plant with a hollow stalk is found in SlkEl put 'Angelica (Russ. [TEXT NOT REPRODUCIBLE IN ASCII])' 2005 : 198) and SlkTa puti 'Siberian hogweed (Heracleum sibiricum)' ([TEXT NOT REPRODUCIBLE IN ASCII] 2007). (4)

Semantic abstraction similar to Selkup is also found in the other Samoyed cognates: Kamas put 'groove; inside; marrow', bun put 'riverbed' (bun 'water, river' GEN), kuzan put 'human innards' (kuzan 'person' GEN), Ngan hudee 'core; middle of a river '(Helimski 1997 : 249). According to Lehtisalo (1956 : 369), the NenT cognate pude means 'small metal tube (used as an ornament in a girl s cap), but a more abstract meaning of core is found in expressions such as pam pude 'core of a tree '(pa 'tree, wood '), nuwem pude 'the Milky Way '(num 'sky, heaven '), and jam pude 'the core of the earth '(ja'earth '). Tereshchenko ([TEXT NOT REPRODUCIBLE IN ASCII.] 1968 s.v. nyd) gives only an abstract meaning that which keeps something stable or balanced (e.g. spine, core of a tree). The derivative pudo means spinal cord, as does its Forest Enets cognate puduj. A further degree of abstraction is found Ngan hute[??]e 'body, figure' and EnF pudodo 'body; the person himself '; from the meaning 'body' the word was finally grammaticalized as a reflexive pronoun in Mator (Helimski 1997 : 249); apparently, the Tundra Nenets pronoun root pid- (< *pite-) is of the same origin, even though it shows an unexpected illabial vowel. Thus, the long path of semantic divergence has created a semantically absurd cognate relationship between the westernmost and easternmost Uralic languages: South Saami batske flower stalk of Angelica turns out to be etymologically identical with the Mator reflexive pronoun hudu!

The semantic abstraction (hollow) stalk, stem (of a plant) > inside, core has apparently begun already in Proto-Uralic, as the meaning inside is also attested in Komi pic (pick-) and Udm puc (puck-) ~ pus (pusk-), which can be included in this cognate set; local case forms of these Permic nouns also function as adverbs and postpositions in the senses in, into . In Udmurt also the more concrete meanings insides, intestines, womb and stomach are attested. The meaning stomach is found in Selkup, too, and also the Kamas expression kuzan put human innards is notable in this connection. The Permic forms had been included in the cognate set by Sammallahti (1979 : 35), but for some reason they are no longer cited in his later paper (Sammallahti 1988 : 539). UEW (397-398) makes a distinction between two both phonologically and semantically distinct etymological sets: on the one hand, the Permic words meaning inside and the Samoyed items with similar meanings are derived from a PU word *puckV inside, and on the other, the words with concrete meanings such as tube, 'stalk' and 'Angelica' are derived from another reconstruct *pucke. The latter reconstruction is, however, clearly incorrect because none of the cognates show any evidence for a palatalized affricate *c; the cognate set quite unambiguously points to the PU cluster *-ck-.

As the two correspondence sets yield an identical Proto-Uralic reconstruction, and their assumed reflexes in Samoyed languages also show meanings that bridge the semantic gap between the two groups, there are no grounds for treating them as etymologically separate sets. The remaining question is how exactly the semantic heterogenity of the attested cognates is to be accounted for, or more precisely, how the duality between the meanings hollow stalk; Angelica, plant with a hollow stalk and inside, interior, core originally developed. As the semantic bifurcation seems to date to a very early phase--likely to Proto-Uralic already--the details of the process can hardly be reliably reconstructed any longer. One possibility, however, is offered by the meaning soft inner part of plants attested in Selkup dialects. One could surmise a development plant stalk > the soft inner part of a plant stalk and further > some kind of soft inner part in general (e.g. marrow, intestines) > inside, core. On the other hand, one could also think of a development hollow stalk; tube > marrow bone, and a further metonymic shift yielding the sense of marrow.

Regardless of the how exactly each semantic shift in this word family has taken place, all the attested meanings of the cognates can be rather straightforwardly derived from a relatively narrow axis of basic meanings, as shown in Figure 1.

8. Fi sato 'harvest, crops', sataa 'yield harvest' < PU *caca- 'grow'

According to SSA (s.v. sataa, sato), Fi sato harvest, crops is etymologically the same word as Finnic *sato 'falling; rain '(> Fi comp. sulka-sato 'molting' (sulka 'feather '), Vot sato 'heavy rain', Est sadu, Liv sa 'd 'rain '), and a derivative of PFi *sata- 'rain; fall '(> Fi sataa 'rain', Est sadama 'rain, fall', etc.). In Finnish dialects there is also a verb sataa 'yield harvest (subj.: field, corn)'. The verb *sata- 'rain' derives from PU *sada- and is cognate with Proto-Samoyed *sara- 'rain '(Janhunen 1981 : 221). From a semantic perspective it is far from obvious, however, that Fi sato harvest and sata- yield harvest' are reflexes of PFi *sato 'falling; rain' and *sata- 'fall; rain', and not merely coincidentally homonymous with them.

Setala (1902 : 222) was probably the first to connect sato 'harvest' with sataa rain, and he suggested that the Finnic word is cognate with SaaN cuohcit get stuck (of fish in a net); have an effect on, MdE caco-, MariE soca- 'be born; grow, yield harvest', Komi cuz- 'be born; grow', Udm cizi-vizi 'relative' and KhN sose 'local, native'. Toivonen (1928 : 87-88), however, treated the two Finnic words as etymologically distinct: he considered Fi sataa 'rain' the cognate of SaaS tjuetsedh 'snow' and SaaI cuoccad molt, whereas sataa yield harvest and sato harvest were in his view cognate with the Mordvin, Mari, Permic and Khanty words mentioned above, as well as MsW sosiy, MsE sase, MsN sossa 'homestead, homeland'. Ravila (1938 : 19), in turn, considered Fi sataa yield harvest and sato harvest etymologically identical with Fi sataa rain, and maintained that they are in no connection with the Mordvin and Mari verbs meaning be born; grow'. UEW (52) reconstructs the Proto-Uralic form *cacV- ~ *cancV, and states that Fi sato harvest cannot be included in this cognate set due to its meaning and its initial s-. In addition to the words mentioned above, the dictionary also presents Samoyed cognates: NenT tence 'kind, sort; tribe', EnF tiz, Ngan tanse 'tribe, clan', and Slk *caci (SlkTa tati, SlkTy caz, SlkK cazi) 'family, tribe'. Sammallahti (1988 : 552) includes only the Mordvin, Mari and Permic words in the cognate set.

As the semantic connection of Fi sataa yield harvest and sato harvest to PFi *sata- 'rain; fall' is not obvious, the alternative connection of these words to PU *cacV- needs to be reconsidered. Semantically the comparison is flawless, as MdM sace- means not only be born but also grow and yield harvest; the derived noun sacama means both birth and harvest . Also MariE soca-, MariW saca- mean both be born and grow (of plants), yield harvest. Hence, the reference to semantic problems made in UEW (52) remains incomprehensible. The only problem in the etymology is phonological: Fi s- is not a normal reflex of PU *c-.

However, the development *c- > Fi s- is accounted for by the fact that the verb originally contained two identical affricates. A dissimilation of the affricates first took place (*caca- > *caca-), after which there was a regular change *c > Fi s. This dissimilation has one well-established parallel, Fi seta 'paternal uncle '< *ceca < PU *ceca (Janhunen 1981 : 225; Sammallahti 1988 : 536; cf. UEW 34). In this case there is also a Saami cognate exhibiting the same dissimilation: SaaN ceahci paternal uncle (younger than father) < *ceca; note the regular changes PU *c > PSaa *c and PU *c > PSaa *c. That the word for uncle originally contained two identical affricates is evident from cognates in other branches: e.g., MdE cice brother-in-law (one's sister s husband), MariW caca maternal uncle, Udm cuz-murt uncle, MsE sas 'uncle '(Ms s < PU *c), SlkK cicce 'uncle'. In Komi, however, a dissimilation identical to Finnic and Saami took place: Komi coz maternal uncle. This is a regular development found in several other words as well, including cuz' grow; be born'. Other examples include Komi cuzji- ~ Udm cizi- 'kick', Komi cuz ~ Udm cuzjem 'malt', and Komi cez ~ Udm cez 'wild duck'.

Moreover, in Saami yet a third example of the same dissimilation has been discovered: SaaN coska 'block of wood '< *cucki < *cucki, cognate with MdE cocko 'timber, log '(Luobbal Sammol Sammol Ante (Aikio) 2013 : 164-165). The dissimilation *c - c > *c - c can be considered a regular sound change in Finnic and Saami, which removes any problem in connecting Fi sataa 'yield harvest 'and sato 'harvest 'to PU *caca- 'grow'. From a semantic perspective it is obviously much more plausible to analyze these words as cognate with MdM sace-, MariW saca- be born; grow, yield harvest and MdM sacema birth; harvest instead of connecting them with a verb meaning rain and 'fall'.

The suggested Ob-Ugric and Samoyed cognates, which are not cited by Sammallahti (1988 : 552), are also worth closer examination. The inclusion of NenT tence 'kind, sort; tribe', EnF tiz and Ngan tansa 'tribe, clan' (UEW 52) in the cognate set is obviously incorrect, as they presuppose a PSam form *tense or *cense; the cluster *-ns- does not match PU *-c-, but would instead presuppose PU *-ns- (cf. PU *kunsi- 'urine '> PSam *kunsa; Janhunen 1981 : 236). However, Slk *caci would continue a PSam form *cacV-, which matches PU *caca- as far as consonantism is concerned. The same is true of the KhS cace, KhN sosa, sosi, sasi 'local, native '(< PKh *cicay) and MsW sosiy, MsE sase, MsN sossa 'homestead, homeland '(< PMs *sjsya); in Mansi there was a regular sound change *c > *s. The vowel correpondences require closer scrutiny, however.

Sammallahti (1988 : 552) reconstructs the verb as Finno-Permic *cecV-, which corresponds to PU *cicV- in the present notation. The reason for not reconstructing the vowel *i is the Komi cognate cuz-. According to Sammallahti' s theory of Permic historical vocalism PU *a developed to PPerm *u (> Komi o, Udm u), whereas PU *i is reflected as PPerm *u (> Komi and Udm u). In his framework the reconstruction *cicV- is incompatible with the Khanty, Mansi and Selkup forms, as PU *i should yield PMs *i, PKh *a and PSam *i or *e (Sammallahti 1988 : 484, 504). However, recently Reshetnikov and Zhivlov (2011) have suggested that *a and *i in Janhunen 's (1981) and Sammallahti s (1988) system of PU vocalism can actually be identified as a single phoneme (*a), and its varying reflexes in Permic, Mansi and Samoyed can be explained by conditioned sound changes. As for Permic, they assume that there was vowel a shift PU *a(- a) > PPerm *u before palatalized consonants and the cluster *-rj-, whereas a shift *a(-a) > PPerm *u occurred before other types of consonants. Reshetnikov's and Zhivlov's hypotheses regarding Mansi and Samoyed historical vocalism cannot be assessed here, but their sound laws appear to account for the Permic data, and thus Permic does not seem to offer evidence for reconstructing the opposition *a Hence, the verb root can be reconstructed as PU *caca- instead of *cica-, which allows also Slk *caci 'family 'to be included in the cognate set. Semantically the Selkup form comes close to the Udm cognate cizi-vizi relative .

The vocalism in the Ob-Ugric forms (PKh *cicey, PMs *sjsya) is less clear. As PKh *i is the high ablaut grade of original *a (Helimski 2001; [TEXT NOT REPRODUCIBLE IN ASCII] 2006 : 42), the underlying Ob-Ugric vowel correspondence is PKh *a ~ PMs *c. According to Sammallahti (1988 : 504) the Uralic source of this correspondence is roots of the shape *(C)uCa- and *(C)iCa-, whereas the reflexes of PU *(C)aCa-roots display the vowel correspondence PKh *a ~ PMs *u. Hence, these forms are irregular, and they must be excluded from the etymology because there is another Ob-Ugric word family which shows both phonologically and semantically a better match with the reconstructed Proto-Uralic verb *caca-: MsN sus--suns- 'grow, increase (intr.); have cubs (of a bear) '(< PMs *su(n)s-), MsN susm-, MsW susm- 'grow, increase (intr.) '(< PMs *susm-), MsN sust-, MsW sust-, MsS sost- 'grow, increase (tr.) '(< PMs *sust-), MsN susam, MsE susam 'one-year old bear cub' (< PMs *susam), KhE cicim, MsN sisam 'bear cub in its first year '(< PKh *clclm). The PKh vowel *i is the high ablaut grade of *a, so the underlying vowel correpsondence is PKh *a ~ PMs *u, exactly as expected in the reflex of PU *caca-. The only irregularity is the unexpected nasal in MsN sus--suns-, but this is in all likelihood a secondary development: the verb has analogically acquired the morphophonological alteration -s- : -ns-, which developed in roots with an original cluster *-nc- (Honti 1999 : 49-51). The secondary origin of n is also evident from the lack of the nasal in all the derived forms as well as in the Khanty cognate.


EnF--Forest Enets; Fi--Finnish; Hung--Hungarian; Kam--Kamas; KhE--East Khanty; KhN--North Khanty; KhS--South Khanty; MariE--East Mari; MariW--West Mari; Mat--Mator; Md--Mordvin; MdE--Erzya Mordvin; MdM--Moksha Mordvin; Ms--Mansi; MsE--East Mansi; MsN--North Mansi; MsS--South Mansi; MsW--West Mansi; NenF--Forest Nenets; NenT--Tundra Nenets; Ngan--Nganasan; PFi--Proto-Finnic; PKh--Proto-Khanty; PMari--Proto-Mari; PMd--Proto-Mordvin; PMs--Proto-Mansi; PPerm--Proto-Permic; Pre-PFi--Pre-Proto-Finnic; Pre-PMd--Pre-Proto-Mordvin; Pre-PSaa--Pre-Proto-Saami; PSaa--Proto-Saami; PSam--Proto-Samoyed; PSlk--Proto-Selkup; PU--Proto-Uralic; Saal--Inari Saami; SaaL--Lule Saami; SaaN--North Saami; SaaS--South Saami; SaaSk--Skolt Saami; SaaT--Ter Saami; Slk--(Proto-)Selkup; SlkK--Ket Selkup; SlkTa--Taz Selkup; SlkTy--Tym Selkup; Udm--Udmurt. 10.3176/lu.2014.1.01


Alatalo, J. 2004, Solkupisches Worterbuch aus Aufzeichnungen von Kai Donner, U. T. Sirelius und Jarmo Alatalo, Helsinki (LSFU XXX).

Benko, L. 1992-1997, Etymologisches Worterbuch des Ungarischen, Budapest.

Bergsland, K., Mattsson Magga, L. 1993, Aarjelsaemien-daaroen baakoegaerja--Sydsamisk-norsk ordbok, [Lakselv].

Fokos-Fuchs, D. R. 1959, Syrjanisches Worterbuch, Budapest.

Grundstrom, H. 1946-1954, Lulelapsk ordbok, Uppsala (Skrifter utgivna genom dialekt- och folkminnesarkivet i Uppsala. Ser. C:1).

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

Helimski, E. 1997, Die matorische Sprache. Worterverzeichnis, Grundzuge der Grammatik, Sprachgeschichte. Unter Mitarbeit von Beata Nagy, Szeged (Studia Uralo-Altaica 41).

Herrala, E., Feoktistov, A. 1998, Moksalais-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 58).

Itkonen, E. 1986-1991, Inarilappisches Worterbuch, Helsinki (LSFU XX).

Itkonen, T. I. 1958, Koltan- ja kuolanlapin sanakirja, Helsinki (LSFU XV).

Kai Donners Kamassisches Worterbuch nebst Sprachproben und Hauptzugen der Grammatik. Bearbetet und herausgegeben von A. J. Joki, Helsinki 1944 (LSFU VIII).

Lehtisalo, T. 1956, Juraksamojedisches Worterbuch, Helsinki (LSFU XIII).

Maksimov, S., Danilov, V., Saarinen, S. 2008, Udmurttilais-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 79).

Mikola, T. 1985, Morphologisches Worterbuch des Enzischen, Szeged (Studia Uralo-Altaica 36).

Moisio, A., Saarinen, S. 2008, Tscheremissisches Worterbuch, Helsinki (LSFU 32).

Niemi, J., Mosin, M. 1995, Ersalais-suomalainen sanakirja, Turku (Turun yliopiston suomalaisen ja yleisen kielitieteen laitoksen julkaisuja 48).

Orszagh, L., Magay, T. 2006, Angol-magyar nagyszotar, Budapest.

Sammallahti, P., Morottaja, M. 1993, Saami-suoma sanikirje--Inarinsaamelais-suomalainen sanakirja, Ohcejohka.

Sammallahti, P. 1989, Sami-suoma satnegirji--Saamelais-suomalainen sanakirja, Ohcejohka.

Steinitz, W. 1966-1993, Dialektologisches und etymologisches Worterbuch der ostjakischen Sprache 1-15, Berlin (= DEWOS).

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

Uotila, T. E. 1942, Syrjanischer Wortschatz nebst Hauptzugen der Formenlehre. Aufgezeichnet von Yrjo Wichmann, Helsinki (LSFU VII).

Wogulisches Worterbuch. Gesammelt von Bernat Munkacsi, geordnet, bearbeitet und herausgegeben von Bela Kalman, Budapest 1986.

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







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

Honti, L. 1982, Geschichte des obugrischen Vokalismus der ersten Silbe, Budapest.

--1999, Az obi-ugor konszonantizmus tortenete, Szeged (Studia Uralo-Altaica Supplementum 9).

Itkonen, E. 1946, Zur Frage nach der Entwicklung des Vokalismus der ersten Silbe in den finnisch-ugrischen Sprachen, insbesondere im Mordwinischen.--FUF XXIX, 222-337.

Janhunen, J. 1977, Samojedischer Wortschatz. Gemeinsamojedische Etymologien, Helsinki (Castrenianumin toimitteita 17).

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

Koivulehto, J. 1999, Varhaiset indoeurooppalaiskontaktit: aika ja paikka lainasanojen valossa.--Pohjan poluilla. Suomalaisten juuret nykytutkimuksen mukaan, Helsinki (Bidrag till kannedom av Finlands natur och folk 153), 207-236.

Kylstra, A. D., Hahmo, S.-L., Hofstra, T., Nikkila, O. 1991-2012, Lexikon der alteren germanischen Lehnworter in den ostseefinnischen Sprachen, Amsterdam--Atlanta (= LAGLOS).

Liimola, M. 1956, Etymologische bemerkungen.--FUF XXXII, 226-264.

Luobbal Sammol Sammol Ante (Aikio, A.) 2012, On Finnic Long Vowels, Samoyed Vowel Sequences, and Proto-Uralic *x.--Per Urales ad Orientem. Iter polyphonicum multilingue. Festskrift tillagnad Juha Janhunen pa hans sextioarsdag den 12 februari 2012, Helsinki (MSFOu 264), 227-250.

--2013, Studies in Uralic Etymology I: Saami Etymologies.--LU XLIX, 161-174.

Ravila, P. 1938, Uber die entstehung des tscheremissischen konjugationssystems.--FUF XXV, 1-25.

Reshetnikov, K., Zhivlov, M. 2011, Studies in Uralic Vocalism II. Reflexes of Proto-Uralic *a in Samoyed, Mansi and Permic.--Journal of Language Relationship /[TEXT NOT REPRODUCIBLE IN ASCII] 5, 96-109.

Sammallahti, P. 1979. Uber die Laut- und Morphemstruktur der uralischen Grundsprache.--FUF XLIII, 22-66.

--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--Kobenhavn--Koln, 478-554.

Setala, E. N 1902. Zur finnisch-ugrischen lautlehre.--FUF II, 219-280.

Toivonen, Y. H. 1928, Zur geschichte der finnisch-ugrischen inlautenden affrikaten.--FUF XIX, 1-270.

--1956. Uber die syrjanischen Lehnworter im Ostjakischen.--FUF XXXII, 1-169.




Ante Aikio

University of Oulu


(1) A different vowel development occurred in MsN, MsS uj- 'sink '(< PMs *uj- < PU *wajV- 'sink '; UEW 551), but this was probably caused by the initial *w-. Sammallahti (1988 : 500) maintains that in Proto-Ugric *a was labialized to *o after an initial labial consonant (*p- or *w-); hence the development was probably *wajV- > *woji>- PMs *uj-. The development PU *-oji- > PMs *-uj- is regular: cf. PU *koji 'male '> PMs *kuj > MsE, MsW kuj-, MsN xuj- 'male '(UEW 166-167), PU *koji 'dawn '> PMs *kuj > MsS koj, MsN juj 'dawn '(UEW 167), PU *soji 'sound '> PMs *suj > MsW suj, MsE soj, MsN suj 'voice '(UEW 482-483).

(2) The semantic development 'burnt smell, fumes, stink '>'plague, epidemic' can be understood in context of the so-called miasmatic theory, i.e. the belief that diseases and epidemics are caused by bad air emanating from rotting organic matter. This belief was prevalent in Europe, India and China since ancient times, and was only gradually displaced in the 19th century by the germ theory of disease. As a parallel to the semantic development of Finnic *katku one can mention malaria from medieval Italian mala aria 'bad air'. It is also noteworthy that in some dialects Est katk means rotten spot in a swamp, deep puddle of mud. It was commonly held that disease-causing bad air emanated from swamps, and also malaria has been formerly called marsh fever due to its association with swamps and marshlands.

(3) UEW (522, 795) distinguishes the cognate sets for 'blade' and 'edge ': SaaN dearri, Fi tera, MariE tur, MariW ter 'blade', Udm tir 'ax', Hung tor 'dagger' (< PU *tera) and MariE tur, MariW tar, Komi dor, Udm dur'edge; shore '(< PU *terV). This analysis is clearly erroneous; in Mari the words are homonymous, and since there are parallels for the semantic relationship'edge' ~ 'blade '(such as English edge < Germanic *agjo- '(edge of a) blade '), it is reasonable to postulate only one underlying etymon, PU *tera'edge/blade'. The distinction between Udm tir 'axe' and dur edge; shore does not support the reconstruction of two originally distinct roots, as Udm u is a regular reflex of PU *e but Udm i is not. Therefore, Udm tir 'axe' most probably has another etymology, and is not related to Fi tera blade and its cognates.

(4) It is not altogether clear whether the gloss of some variants in Bykonja 's dictionary ([TEXT NOT REPRODUCIBLE IN ASCII] 2005) should be read as 'Angelica' instead of 'pipe', as Russian [TEXT NOT REPRODUCIBLE IN ASCII] means both. SlkEl put is unambiguously glossed as '[TEXT NOT REPRODUCIBLE IN ASCII]', however.
COPYRIGHT 2014 Estonian Academy Publishers
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2014 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Aikio, Ante
Publication:Linguistica Uralica
Article Type:Report
Geographic Code:4EXHU
Date:Mar 1, 2014
Previous Article:Valdek Pall in memoriam 30. Juni 1927-17. April 2013.
Next Article:On the possible Baltic origin of Est varn, Fin vaarna/Uber eine mogliche baltische herkunft von est. varn, FI. vaarna.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |