Are there Proto-Slavic loanwords in Saami?

The impetus for this etymological case study was provided by the paper Wie alt sind die Kontakte zwischen Finnisch-Ugrisch und Balto-Slavisch? by Jorma Koivulehto (2006). One of the ideas argued for in that paper is that the Saami languages possess a couple of Proto-Slavic loanwords. The presence of Russian loanwords in Eastern Saami has been known for long, of course, but possible older loans from Proto-Slavic into Saami have been little discussed by research. Only a very small number of etymologies suggest such prehistoric language contact, so the topic is quite marginal even in the field of Saami historical linguistics. On the other hand, the scrutiny of this small and very specific etymological problem does raise some methodological questions of more general interest as well.

Koivulehto (2006) presents a Proto-Slavic loan etymology for two Saami words: *multtē ‘soap’ and *kuompe̮r ‘mushroom’. The two etymologies have interesting implications, as the alleged borrowing correlates phonologically with the Proto-Saami level of reconstruction and thus suggests direct prehistoric contact between Proto-Saami and some stage of Proto-Slavic. The Slavic etymology of *multtē ‘soap’ had already been proposed over half a century earlier by Y. H. Toivonen (1949: 346–347), but after that the proposal appears not to have been commented upon prior to Koivulehto’s contribution. The handbook of Saami historical linguistics by Korhonen (1981: 52–53) does not mention the possibility of early contacts with Slavic; the only Slavic loanwords discussed there are Russian ones, which at least for the most part have been adopted after the founding of the Pechenga monastery in the 16^th century.

In addition to the two etymologies mentioned, Koivulehto also suggests an early Proto-Slavic or Proto-Balto-Slavic origin for Saami *muottē- ‘snow’ (verb), *ćuorpme̮s ‘hail’ and *ve̮ljē ‘abundance’. These etymologies presuppose a much earlier date of borrowing because the words have undergone the Pre-Proto-Saami vowel changes *a > *ō > *uo and *i > *e̮; the last word also has a cognate in Finnic (cf. Fin vilja ‘grain, cereal; abundance’). As these words seem to belong to a different and older lexical stratum, I will omit them from consideration here. If the etymologies are correct, the sound correspondences imply that they are roughly equal in age with the earliest Germanic and Baltic loans in Saami.

The concept of “chance correspondence”

As is well-known, there are many chance similarities between words in different languages, and it is usually not very difficult to distinguish them from genuine etymological correspondences – at least when they occur in languages whose historical development is well understood. However, there is also another related phenomenon that has not been clearly distinguished from the concept of “chance similarity” in the theory of historical linguistics. I will refer to this phenomenon as “chance correspondence”. Chance similarities between words are based on an impression of likeness of form and meaning, and they are therefore always subjective in nature; it is usually easy to show the coincidental nature of such similarities by applying proper methods of historical linguistics and etymological research. Chance correspondences, on the other hand, differ from the former in that they involve word-forms that show a formally regular or predictable correspondence despite of not being etymologically related at all.

To illustrate chance correspondences, let us look at the following word pairs that Campbell and Mixco (2007: 29) cite among their examples of chance similarity between languages:

• English much, Spanish mucho ‘much’
• Hungarian fiú, Romanian fiu ‘boy; son’

These examples do not show mere vague similarity of phonological form and meaning, but a systematic match between two nearly identical forms. The compared words have the same meaning, and the very small phonological differences between them could be easily explained on the basis of the differences between the phonological systems of the two languages:

• (Old) Spanish mucho /muʧo/ ~ Middle English muche /muʧe/ > modern English much /mʌʧ/ (dialectally /mʊʧ/). The minor difference of second-syllable vowels could be naturally attributed to the very limited vowel inventory of Middle English unstressed syllables: unstressed /e/ was probably realized as [ǝ], and it represented the Middle English outcome of all Old English unstressed vowels.
• Hungarian fiú /fi(j)uː/ ~ Romanian fiu /fiu/. Here, too, the small difference of form could be attributed to general differences of phonological structure: unlike Romanian fiu, Hungarian fiú is bisyllabic, but Hungarian lacks diphthongs altogether.

The key issue here is this: if we were to ignore everything else besides the formal criteria of phonological and semantic correspondence, then we could convert these word comparisons into formally “flawless” loan etymologies by claiming that Middle English muche was borrowed from Old Spanish mucho, and Romanian fiu from Hungarian fiú (or vice versa).

These “etymologies” are obviously wrong, but that is not because of any flaw in phonological or semantic correspondences. Instead, we know they are wrong because we know where the words actually come from. Spanish mucho reflects Latin multus ‘much’, whereas English much reflects Old English myċel ‘big; much’; Hungarian fiú is formed from a stem fi- which reflects PU *poj-ka ‘son, offspring’, whereas Romanian fiu reflects Latin fīlius ‘son’.

It is not an altogether uncommon occurrence to come across this kind of formally good-looking quasi-etymology for some word. Actually, my impression is that chance correspondences are more common than most professional historical linguists tend to assume, and that they do pose some real challenges for etymological research. In the case of the two examples above it is of course very easy to distinguish between the real etymology and the chance correspondence, but this is not always so. Every once in a while a word turns out to have two quite well-formed alternative etymologies (and sometimes even more than two), and it is difficult to decide which of them is the correct one.

Consider, for example, the case of Fin tuhto ‘thwart (rower’s seat in a boat)’, for which two etymologies have been proposed:

• Borrowing from PGerm *þuftōn- (> ONo þopta, OEngl þoft, þofte, OHGerm dofta ‘thwart’).
• Inheritance from PU *tukta (> MariW tǝ̑ktǝ̑ ‘boat rib’, Komi ti̮k ‘crossbar; spoke of a wheel’, KhVVj tŏɣǝt ‘crossbar of a boat’, MsLK tɔxt ‘thwart’, Hung tat ‘stern (in ship)’, NenT tǝdeʔ̰, SlkTa tati̮ ‘crossbar in a boat or dugout’).

Both of these etymologies are phonologically and semantically unproblematic and straightforward, and in the absence of one there would be no reason at all to doubt the other. Nevertheless, one of the two etymologies must be wrong, but it is quite hard to decide which one. The Finnish word does show a somewhat limited dialect distribution and it has no cognates in other Finnic languages, which is more typical of Germanic loans than of inherited words, but this is merely suggestive. At face value the second-syllable vowel of tuhto appears to show a better match with Germanic than with Uralic, but -o could be explained as a derivational suffix, and in Finnish dialects we also find the variant tuhta- in the compound noun tuhta-lauta ‘thwart’ (lauta ‘board’).

The existence of chance correspondences has an important implication for the evaluation of etymologies. We cannot blindly trust that an individual etymological comparison is correct even if it looks convincing by phonological, morphological and semantic criteria. The word might later turn out to have an alternative etymology, or it might be of a different origin even if we could not discover the correct etymology at all. Even convincingly argued etymologies involve some uncertainty; the degree of this uncertainty might be quite small, but that does not make it negligible.

Luckily, such slight uncertainties do not usually have broader implications for our conclusions and theories. Whatever the true etymology of Fin tuhto ‘thwart’ is, it will not alter our knowledge that in the Finnish lexicon both Germanic loanwords and inherited Uralic words number in the hundreds.

The situation is quite different, however, when the existence of a lexical stratum is inferred from a very small number of etymologies. The alleged Proto-Slavic loanwords *multtē ‘soap’ and *kuompe̮r ‘mushroom’ in Proto-Saami are a case in point. As the two proto-languages do not date very far back in time, the reconstructed lexicons that can be compared against each other amount to over 2000 words. There is a quite real possibility that a comparison of two such reconstructed lexicons will produce two formally good-looking etymological matches by sheer chance; for this to happen, the probability of an accidental match only needs to be no greater than 1 : 1000. Therefore, we need to carefully assess how solidly argued each of the two etymologies actually is, and to examine whether alternative etymologies for the words could be found.

Saami *multtē ‘soap’

This word is attested in North Saami and in Eastern Saami languages: SaaN *multi (Leem 1768 and Friis 1887: ‹multte›), SaaI multte, SaaSk muʹltt, SaaK muʹlht, saaT mïʹlhte ~ muʹlhte ‘(a kind of) soap’. The common proto-form of the words can be reconstructed as *multtē. The unrounded vowel ï = [i̮] in the Ter Saami variant mïʹlhte is irregular, but it could have recently developed by influence of the similar vowel in Russian мыло ‘soap’; also a phonologically regular variant mu´lhte has been attested by Genetz (1891). In modern varieties of North, Inari and Skolt Saami the word seems to be obsolete, and according to dictionaries the word has been used of various kinds of old-fashioned and usually home-made soaps, made with ingredients such as lye, reindeer fat, etc.

The phonological form immediately reveals that *multtē cannot be an inherited word in Proto-Saami: neither the vowel combination *u–ē nor the three-consonant cluster *ltt has any regular Proto-Uralic or Pre-Proto-Saami source. Of course, also the meaning ‘soap’ suggests that we are not dealing with an ancient inherited word. Thus, to the trained eye of a Uralic historical linguist *multtē stands out as a particularly obvious candidate for a loanword.

The source proposed by Toivonen and further argued by Koivulehto is PSlav *mỳdlo ‘soap’, or more precisely its immediate predecessor *mūdlo prior to the regular change of *ū into an unrounded vowel (Proto-Slavic *y = UPA [i̮] / IPA [ɨ]). The etymology is semantically and phonologically quite straightforward, and the only detail requiring an explanation is the metathetic substitution of Saami *ltt for the Slavic cluster *dl. This has a straightforward structural motive: Saami phonotaxis did not allow clusters of the alveolar stop *t followed by a sonorant. Although we do not have a precise parallel involving the cluster *ltt, there are several well-established examples of an analogous substitution of Saami *rtt for the Norse clusters *dr and *þr (Koivulehto 1988):

• SaaN fiertu ~ viertu ‘fine weather’ < PSaa *vierttō < PNo *wedra- (> ONo veðr ‘weather’)
• SaaL liertte ~ riertte ‘leather’ < PSaa *lierttē < PNo *leþra- (> ONo leðr ‘leather’)
• SaaN gurti ‘meat on the neck’ < PSaa *kurttē < PNo *kuþran- (> Icel koðri ‘scrotum’)
• SaaN vierca ‘ram’ < PSaa *viercce̮ < *viertte̮s < PNo *weþru-z (> ONo veðr ‘ram’)

Thus, the Slavic loan etymology of Saami *multtē is phonologically and semantically quite unproblematic, as long as it is assumed that borrowing took place prior to the unrounding of *ū in Proto-Slavic. Also the cluster *ltt suggests a relatively early date of borrowing, because the cluster *dl was retained only in West Slavic (cf. Polish mydło, Czech mýdlo ‘soap’, etc.), and in other branches of Slavic the stop was lost (cf. OCSlav мꙑло, Slovene mílo, Ukrainian мило, Russian мыло, etc.). At face value this etymology looks quite compelling, and would thus seem to provide good evidence of prehistoric language contact between Saami and an archaic form of Proto-Slavic. Furthermore, we also know that the same Slavic word has been independently borrowed into Finnic: cf. Fin (dialectal) muula, Veps mugl, Võro mugõl ‘(a kind of) lye’ < PFi *mukla < PSlav *mỳdlo. The Finnic word, too, preserves the stop in the cluster *dl; as the cluster *tl did not occur in Finnic, the existing cluster *kl was substituted for it. This pattern, too, is confirmed by Germanic loanwords: cf. Finnish neula ~ Veps negl ~ Võro nõgõl ‘needle’ (< PFi *ne̮kla < PGerm *nēþlō-) and Finnish seula ~ Veps segl ~ Võro sõgõl ‘sieve’ (< PFi *se̮kla < PGerm *sēþla-). Because of this different nativization strategy the Saami word *multtē could not have been mediated by Finnic: the PFi cluster *kl could not have yielded *ltt in Saami.

Although the Slavic loan etymology of Saami *multtē looks quite flawless, the word also has an alternative etymology which has remained overlooked by previous research. The word can be straightforwardly compared to PNo *smulta-, reflected in the following forms:

• ONo / Icel smolt ‘grease floating on top of hot water’
• Far smoltur ‘liquid fat (fat from web-footed birds when boiled)’
• ODan smolt ‘melted fat’
• Sw (dialectal) smult ‘melted or purified lard or goose fat (used in cooking or on bread)’
• Nw smult ‘lard (used for cooking and soap production)’

In phonological terms the Norse loan etymology is completely straightforward. The vowel correspondence is paralleled, e.g., by SaaN gurti ‘meat on the neck’, which was mentioned above. Other borrowings showing of the same vowel correspondence include: SaaN rudni ‘hole in the ice’ < PSaa *runnē < PNo *brunna- (> ONo brunnr ‘spring, well’); SaaSk uʹrmm ‘botfly larva’ < PSaa *urmē < PNo *(w)urma- (> ONo ormr ‘worm, snake’); SaaN durdi ‘filth, dirt’ < PSaa *turtē < PNo *turdã- (> ONo torð- in torðýfill ~ tordýfill ‘dungbeetle’; cf. OEngl tord ‘dung’). The substitution of Saami *ltt for Norse *lt is paralleled by the well-known borrowing SaaN sálti ‘salt’ < PSaa *sālttē < PNo *saltã- (> ONo salt). Also a new etymology displaying the same substitution can be presented: SaaN boltut ‘rummage’ < PSaa *poltt-ō- < PNo *bultia- (> Icel bylta ‘throw to the ground, overturn, overthrow’) or PNo *bultō- (> Far bólta ‘turn over, tumble, upset, overturn, roll (down)’). No etymology has been previously proposed for this Saami verb.

A notable phonological feature is the simplification of the Norse word-initial consonant cluster *sm- into *m- in Saami. In most Proto-Norse borrowings foreign clusters of the type *sC- have been retained in Western Saami languages, and for the most part also in Inari and Skolt Saami (Aikio 2012: 78). The only other widespread loanword featuring the cluster *sm- in Norse retains the cluster everywhere except for Kola Saami: cf. SaaS smaave, SaaN smávis (attr. smávva), SaaSk smaavâs (attr. smaavv), SaaK mååvv ‘(very) small, tiny’ (< PSaa *(s)māve̮s ~ *(s)māvēs : *(s)māve̮). This borrowing must be younger than Proto-Norse, however: the intervocalic consonant *-v- in Saami reveals that it was not borrowed from PNo *smāxa-z, but from a form postdating the loss of *x / *h (cf. ONo smá-r ‘small, little’); the syllable *-ve̮- was added because Saami morpheme structure does not allow monosyllabic content word-stems. The word *multtē ‘soap’ looks like an older borrowing which dates back to a period when foreign word-initial clusters of the shape *sC- had not yet become established in Saami. At least two other such loans have been discovered by previous research:

• SaaS -gaejmie in tjeada-gaejmie ‘shimmer, dawn’ (tjeada ‘twilight’), SaaN (dialectal) gáibmu ‘dawn, dusk’ (< PSaa *kājmē ~ *kājmō) < PNo *skaima- > dialectal Swedish skäim ‘dawning, dim twilight’. — The Saami noun was further borrowed into Far-Northern dialects of Finnish as kaimo ‘dawn, first light’.
• SaaN gáiru ‘great black-backed gull’ (< PSaa *kājrō) < PNo *skairȭ- > ONo skári ‘young gull’

What is more, at least seven new etymologies can be presented which display the same kind of simplification of a word-initial consonant cluster:

• SaaS baakoe, SaaL báhko ‘word’ (< PSaa *pākō) < PNo *spaxō- > ONo spá ‘prophecy’. — The assumed semantic development in Saami is admittedly not a common one, but on the other hand, the Saami noun has no other plausible etymology although the vowel combination *ā–ō suggests that it is a loanword. According to SSA it could be cognate with (Northeastern) Finnic *pakise̮- ‘speak, talk, chat’, or alternatively a borrowing from Finnic. Cognation is out of the question, however, because the vowel correpsondence is irregular. Neither is borrowing from Finnic plausible, because there is no noun in Finnic that would suit as the source of PSaa *pākō. In Saami there is also a verb that comes somewhat closer to the Finnic forms (SaaU bååhkadit ‘speak (about)’, SaaN báhkkodit ‘say out loud, express’ < PSaa *pākuje̮nte̮-), but this verb is clearly a derivative of the noun *pākō and as such it cannot be direct borrowing from Finnic. SSA also mentions another Kola Saami verb in this connection: SaaK paaʹgge, SaaT paaʹgged ‘quarrel, argue’ (< PSaa *pākē-). This, however, seems to be an etymologically unrelated borrowing from PNo *bāga- (> Far bága ‘harm, injure’). The verb is scarcely attested in Nordic languages, but it must be old because it is cognate with OHGerm bāgan ‘quarrel, argue; engage in a lawsuit’ (which shows the same meaning as the Kola Saami verb). As a side note, SaaN biehkut ‘complain, grumble, whine (about something)’ (< PSaa *piekō-) seems to be an older borrowing from the same Germanic verb, adopted from an archaic form *bēga- prior to the vowel change *ē > *ā in North and West Germanic.
• SaaU bädtjet, SaaL bádtjit ‘incite’ (< PSaa *pāńć-e̮je̮-) < PNo *spanja- > ONo spenja ‘attract, allure’. — The etymology entails the substitution of PSaa *-Cć- for PNo *-Cj-. Although the phonological and phonetic motivation for this substitution pattern remains unexplained, many examples of it are well-established: cf., e.g., SaaL ávttja ‘bird-cherry’ (< *āvće̮ < PNo *hagja- > ONo heggr ‘bird-cherry’), SaaL ávttjit ‘urge, incite’ (< *āvć-e̮je̮- < PNo *agja- > ONo eggja ‘incite; sharpen’), SaaL guolltje ‘cold north wind’ (< *kuolćē < PNo *kōljōn- > Icel kæla ‘light, cold wind’), SaaL ruvttja-suodna ‘back sinew’ (< *ruvće̮- < PNo *hrugja- > ONo hryggr ‘backbone, spine; ridge’), SaaL sjlávttjá ‘horsefly’ (< *(s)lāvćā < PNo *klagjã- > ONo kleggi ‘horsefly’), SaaL sjkálltjo ‘seashell’ (< *skālćō < PNo *skaljō- > ONo skel ‘shell, seashell’), SaaL skávttjá ‘beard’ (< *skāvćā < PNo *skagja- > ONo skegg ‘beard’), SaaL sjtádtjo ‘casting ladle; frying pan’ (< *stāńćō < PNo *stainjō-; not attested in Norse, cf. OHGerm steina ‘stone or earthenware pot’).
• SaaS baenie, SaaN bátni, SaaSk pääʹnn, SaaT paaʹnne ‘tooth’ (< PSaa *pānē) < PNo *spānu- > ONo spánn, spónn ‘chip, shaving; spoon’. — Published references present no etymology for the Saami word, but it is obviously an innovation because it has replaced the reflex of PU *piŋi ‘tooth’, which is retained by almost every other branch of the family. The Norse loan etymology was suggested to me by Jorma Koivulehto in the course of personal communication in 2003, but to my knowledge he never publicly presented the etymology. At first sight the comparison looks semantically far-fetched, but in Saami also meanings much closer to the Germanic forms are attested: SaaN bátni refers not only to ‘teeth’, but also to tooth- or rod-like parts that occur in rows in various kinds of objects: the teeth of a saw, a rake, or a comb; the rungs of a ladder; the steps of a staircase; the spokes of a wheel; the horizontal bars of a gate; the vertical strips of a weaver’s reed. Notably, the predecessor of the Norse word, PGerm *spēnu-, was borrowed into Finnic: cf. Fin piena ‘wooden slat, crosspiece’. On the other hand, a later borrowing of the same Norse word into Saami is SaaS spaanese, SaaN spánas ‘wood shaving’ (< PSaa *spāne̮s). The semantic correspondence displayed by the earlier loan *pānē ‘tooth’ is similar to that in Ancient Greek γόμφος ‘peg, bolt, nail’ ~ OEngl camb ‘comb’ ~ Sanskrit jámbha-, Old Church Slavonic zǫbъ, Latvian zuobs, Albanian dhëmb ‘tooth’ (< PIE *ǵómbʰo-) (note also Fin hammas ‘tooth’ < Pre-PFi *šampas < Proto-Baltic *žámbas, a borrowing which replaced the inherited Uralic word pii in the primary anatomical sense).
• SaaU gáddiet ‘suspect; accuse’, SaaN gáddit ‘think (mistakenly), believe (falsely); mistake (something for something else)’ (< PSaa *kāntē-), SaaL gáddalit ‘suspect, mistrust’, SaaSk kaddled ‘slander, abuse verbally; quarrel’ (< *kānt-e̮le̮-) < PNo *skandia- (not attested in Norse; cf. OEngl scendan ‘put to shame, abuse, insult, harm’, MDu schenden ‘stain, dishonor; ruin (someone’s standing or happiness)’, Germ schänden ‘desecrate, dishonor; violate’). — Among the Saami forms the semantically most archaic one is SaaSk kaddled. T. I. Itkonen (1958: 77) suggested that this Skolt Saami verb could have been borrowed from Finnish kannella (: kantele-) ‘tell on, go tell about (someone’s illicit behaviour to a superior); file a complaint’. This appears unlikely, however, because SaaSk kaddled is formally identical to SaaL gáddalit ‘suspect, mistrust’ and therefore it can be straightforwardly analyzed as a derivative within Saami. Although Fin kannella is also attested in the meaning ‘quarrel; scold’, this meaning only occurs locally in Northern Savo dialects and it is not attested in Finnish dialects that have been in direct contact with Skolt Saami. The Norse loan etymology for the Saami verb presupposes the original meaning ‘slander, abuse verbally’; this can easily develop into ‘accuse’ (which is attested in Ume Saami and also in the 1780 “Swedish Saami” dictionary by Lindahl & Öhrling), and that further to ‘suspect (someone of something)’ (which is attested in most of the Saami cognates). The loan original itself was lost in Norse, and reflexes of PGerm *skandja- are only attested in West Germanic; this is not a problem, however, as also many other such Proto-Norse loans are known (Aikio 2020).
• SaaN givdnjut, SaaSk ǩeunnjad, SaaK kïvvnjâ ‘shimmer, flicker, appear by glimpses, appear briefly and repeatedly’ (< PSaa *kivńō-) < PNo *skiuma- (< *skeuma-) > Nw (dialectal) skjoma ‘flicker, shine with flickering light’ (cf. ONo skjómi ‘flickering light’). — The cluster *vm does not occur in Saami, which explains the substitution of Saami *ń for Norse *m in this context. The same pattern is attested in at least two other loanwords as well, although in these cases the eastern Saami languages partially show *ŋ: SaaN rávdnji, SaaSk räuʹnnj ~ räuʹŋŋ, SaaK raʹvvnj ‘current, stream’ (< *rāvńē ~ *rāvŋē) < PNo *strauma- > ONo straumr ‘stream’; SaaN sávdnji, SaaSk säuʹnnj ~ säuʹŋŋ, SaaK saʹvvŋ ‘seam’ (< *sāvńē ~ *sāvŋē) < PNo *sauma- > ONo saumr ‘seam’.
• SaaS laekedh ‘hit, strike, beat; strike dead, beat to death’ (< PSaa *lākē-) < PNo *slaxa- > ONo slá ‘hit, strike, beat’. — The verb occurs in South Saami only, but the sound substitutions *sl- > *l- and *-x- > *-k- reveal that the borrowing must be quite old. One parallel for the latter substitution is PSaa *pākō ‘word’ (< PNo *spaxō-) which was discussed above; another is SaaL láhko, SaaN láhku ‘highland plain, gently sloping valley in highlands’ < PSaa *lākō (< PNo *flaxō- > ONo flǫ́ ~ flá ‘rock ledge; gently sloping valley in highlands’, Nw flå ‘plateau; valley in highlands’) (Aikio 2020: 21). Note, by the way, that there is another isolated verb somewhat similar to SaaS laekedh in the opposite geographic end of the Saami language area: SaaT lïïʹgged ‘chop (wood)’ (< PSaa *luokē-). This seems to be an even earlier borrowing from Germanic that has undergone the regular change of Pre-PSaa *a to PSaa *uo: it was probably borrowed from PGerm *slaxa- into Pre-PSaa as *laka-.
• SaaS laehpedh ‘leave (transitive)’, SaaN láhppit ‘lose; shed (antlers, hair)’, SaaSk läʹpped, SaaT laaʹhhped ‘lose’ (< PSaa *lāppē-) < PNo *slãppia- (< *slampia-) > ONo sleppa (past.3sg sleppti) ‘make slip, let slip’, Icel sleppa ‘release, let go’. — This weak verb is a causative of the strong verb sleppa (past.3sg slapp) ‘slip, slide’ (< PNo *slẽppa- < *slempa-). A parallel for the consonant correspondence can be found in another new etymology: SaaN ráhpis, SaaSk rääʹppes, SaaK raaʹbbes ‘rough and rocky (of terrain)’ (< PSaa *rāppēs) < PNo *krãppa-z (< *krampa-z) > ONo krappr ‘narrow’, Icel krappur ‘narrow, scarce, difficult, dangerous’, Far krappur ‘acute-angled, extremely bent, bowed, curved’. This word has also been reconstructed as PNo *krappa-z without a nasal, but Kroonen (2013: 301) reconstructs PGerm *krampa-z and considers the Nordic words cognate with OHGerm krampf ‘bent, curved, crooked’.

In light of the etymologies discussed above, there is no doubt that the Saami languages possess a stratum of early Proto-Norse loanwords in which the sibilant was dropped in word-initial clusters of the type *sC-. The word *multtē ‘soap’ could thus have been borrowed from PNo *smulta- and belong in this stratum of loanwords.

Now that two possible sources of borrowing have been identified for Saami *multtē ‘soap’, we have to evaluate which of them is likely to have been the actual source of the word. However, the basic criteria of phonological and semantic correspondence do not help in settling the issue, because in this regard both etymologies are completely straightforward and unproblematic. The phonological shape of the Saami word is perfectly well explained by Proto-Norse *smulta- and early Proto-Slavic *mūdlo- alike. As regards semantics, the Saami word shows a more precise match with the Slavic one, but there is no real semantic problem in the Norse etymology either; the assumption of an unremarkable semantic shift like “key ingredient of soap” > ‘soap’ could not, by itself, provide a serious argument against the etymology.

There is another key criterion, however, which is independent of the features of the individual etymologies themselves. On the one hand, there are literally hundreds of ancient Norse loanwords in Saami, including dozens upon dozens of Proto-Norse ones; on the other hand, there are next to no plausible candidates for Proto-Slavic loanwords. In addition to *multtē ‘soap’, the only other promising example of such a borrowing is the word *kuompe̮r ‘mushroom’, and that etymology is not without problems either, as will become clear in the discussion below. Therefore, the very existence of Proto-Slavic loans in Proto-Saami is doubtful, and this serves as a very strong argument for the Norse loan etymology of *multtē. Although the alternative Slavic etymology cannot be totally disproved, it looks far less probable than the Norse one, and therefore the word *multtē does not provide any real evidence of contacts between Proto-Saami and Proto-Slavic.

Saami *kuompe̮r ‘mushroom’

All Saami languages share a common word for ‘mushroom’: SaaS goebpere, SaaU guabbar, SaaP SaaL SaaN guoppar, SaaI kuobâr, SaaSk kuõbbâr, SaaK kuumbâr, SaaT kïïmbâr (< PSaa *kuompe̮r). Two etymologies have been proposed for the word. On the one hand, it has been considered cognate with Komi gob and Udm gubi̮ ‘mushroom’; on the other, it has been regarded as a borrowing from PSlav *gǫba ‘mushroom, fungus’ (> OCSlav gǫba ‘sponge’, Serbo-Croatian guba, Bulgarian гъ́ба, Czech houba ‘mushroom’, Russian губа ‘lip; (dialectal) bracket fungus’, etc.), or from its Balto-Slavic predecessor.

The comparison to the Permic forms has been supported by Sammallahti. In an earlier paper (1988: 552) he reconstructed a common proto-form *ko/ampV, but he had marked the etymology with a question mark and also mentioned the possibility that the Permic forms have been borrowed from Chuvash; later (1998: 121), however, he included SaaN guoppar in a list of words inherited from “Proto-Finno-Permic”, expressing no uncertainty.

The suggested etymological connection of the Saami and Permic forms appears untenable because the vowel correspondence is quite irregular. According to Zhivlov’s (2023: 135–138) model of Permic historical phonology, the vowel correspondence between Komi gob : gobj-, (Upper Sysola dialect) go̭b and Udm gubi̮ points to PPerm *göbi̮. The vowel *ö has two regular sources: 1) PU *e before second-syllable *-i or *-äj; 2) PU *a before palatal or alveolo-palatal consonants. In the case of the word for ‘mushroom’ only the first alternative would be possible. Therefore, had the word been inherited, it would regularly go back to PU *kempi or *kempäj; cf., e.g., PU *keri ‘crust, bark’ > PPerm *kör(j-) > Komi kor : korj-, (Upper Sysola) ko̭r, Udm kur; PU *penä-j ‘dog’ > PPerm *pöni̮ > Komi pon : ponj-, (Upper Sysola) po̭n, Udm puni̮. This excludes cognation with PSaa *kuompe̮r, as PSaa *uo cannot reflect PU *e. Yet another problem is that the stem-final *-r in Saami has no correspondent in Permic; I will return to this issue further below.

In addition to these problems, the Permic word for ‘mushroom’ has also alternative comparanda: the resemblance to the aforementioned PSlav *gǫba ‘mushroom, fungus’ is especially striking, and also Tatar gömbä and Chuvash kămpa ‘mushroom’ come formally close. The mutual etymological relations of these words are unclear, but at any rate the Permic word cannot be expained as a Chuvash loan as was passingly suggested by Sammallahti (1988: 552). The Chuvash word itself has been regarded a borrowing from Tatar gömbä (< *gümbä), but the origin of the latter is not quite clear. As regards Permic *göbi̮, the assumption of borrowing from a Turkic source is problematic also because the change of nasal+stop clusters into voiced stops (*NT > *ND > *D) had in all likelihood taken place in Permic already before the earliest Permic-Turkic contacts (see the discussion by Metsäranta 2020: 201–202). The possibility of borrowing from Slavic seems a more attractive hypothesis, but a proper evaluation of the etymology would require the features and chronology of the earliest East Slavic loanwords in Permic to be worked out first.

As regards Saami *kuompe̮r, its alleged cognation with Permic *göbi̮ must be false, so there is no competing hypothesis for the Slavic loan etymology of the former. Nevertheless, the word does not offer unambigous evidence of contact between Proto-Slavic and Proto-Saami even if the loan etymology is correct, because Koivulehto presents two chronologically different interpretations of the etymology. According to the first one, PSaa *kuompe̮r was borrowed from PSlav *gǭba (> *gǫba); in this version the nativization of the nasal vowel and the following stop would be analogous to that in Fin kuontalo ‘roll of wool or flax fiber (for spinning yarn from)’ < PSlav *kǭdělь > *kǫdělь (> OCSlav kǫdělь, Rus кудель, Czech koudel, Polish kądziel ‘sliver, tow’). According to the second interpretation the loan would be considerably older: PSaa *kuompe̮r could go back to Pre-PSaa *kampir, which was borrowed from Pre-Proto-Slavic (or Proto-Balto-Slavic) *gambā.

The latter alternative, in fact, appears a priori more likely than the first one. It would make the loan roughly equal in age to the earliest layers of Proto-Baltic and Proto-Germanic loans in Saami, which in turn would imply that the borrowing had taken place in a quite different geographic setting. It is well-established that Pre-Proto-Saami was originally spoken at a more southerly latitude, somewhere in Southern Finland and Karelia (see the discussion by Aikio 2012 and references therein), and it would not be difficult to assume that also a stray Pre-Proto-Slavic borrowing had entered the language at that time. What is more, in this scenario it would not even be necessary to assume that the borrowing was adopted from Slavic in the first-place: in terms of phonological reconstruction Proto-Balto-Slavic is nearly synonymous with Proto-Baltic, and one could alternatively assume that the word was borrowed from Baltic *gambā-, i.e., from a cognate form of the Slavic word which just happened to be later lost in the Baltic branch. The same expanation could apply also to a couple of other alleged Pre-Proto-Slavic or Proto-Balto-Slavic loanwords, at least to the aforementioned SaaN vallji ‘abundance’ ~ Fin vilja ‘grain; abundance’ (< *wilja < Proto-Balto-Slavic or Proto-Baltic *wīl(i)ja-; cf. Koivulehto 2006: 187–188). In fact, the existence of some such Baltic borrowings is entirely predictable: it would be quite contrary to expectations if the source form of every single prehistoric Baltic borrowing had been preserved in the Baltic branch itself. In a similar way, the much more numerous Proto-Norse borrowings in Saami also include ones whose sources have not been preserved in the attested Nordic languages; 18 such cases are discussed in a recent paper of mine (2020).

However, one problem still remains that provides an argument against both the Saami-Permic comparison and the Slavic (or Baltic) loan etymology: the assumption that PSaa *kuompe̮r contains a suffix *-(e̮)r. Koivulehto (2006: 184–185) comments the issue as follows: “From a morphological point of view it should be noted that the Saami word must be a suffixed stem in *-er(e) [= *-ir according to the present reconstruction]. Suffixations are not uncommon in loanwords.” (Quote translated from German.) This is a purely ad hoc hypothesis, however: there is no derivational suffix *-(e̮)r in Saami, so nothing in PSaa *kuompe̮r itself suggests that the word could be a derivative, let alone that it “must” be one. The assumed suffixation is circularly based on the loan etymology itself, which can only explain the part *kuomp(e̮)- but not the stem-final consonant *-r.

To substantiate this counterargument, let us look at the origins of Lule, North and Skolt Saami noun-stems ending in *-r. Such nouns can be etymologically grouped in four broad categories. Many of the words are borrowings from Nordic languages; most are recent loanwords, but there are also some earlier borrowings from Old Norse and even from Proto-Norse:

• SaaN áittar (gen./acc. áitara) ‘caretaker, owner’ < PSaa *ājtte̮r < PNo *aixter- (> OSw -attari in iorþ-attari ‘landowner’)
• SaaL áldar (acc. álldarav) ‘age’ < Nw alder
• SaaN bolsttar (gen./acc. bolstara) ‘matress; pillow’ < Nw / Sw bolster
• SaaN dimbbar (gen./acc. dimbara) ‘timber’ < ONo timbr
• SaaN eappir (gen./acc. eabbára) ‘wooden pail, bucket’ (< *eampēr) < OSw æmbar
• SaaN fáttar (gen./acc. fáddara) ‘godparent’ < Nw / Sw fadder
• SaaN fuođar (gen./acc. fuođđara) ‘fodder’ < ONo fóðr
• SaaN gufihtar (gen./acc. gufihttara) ‘gnome (a kind of anthropomorphic underearth being in Saami mythology)’ < ONo *góð-vættr (góð ‘good’ + vættr ‘a supernatural being’) > Nw godvette ‘a kind of benevolent gnome-like being’.
• SaaN keallir ~ geallir (gen./acc. keallára ~ geallára) ‘cellar’ < Nw kjeller / Sw källare
• SaaN meašttir (gen./acc. meaštára) ‘master, expert, champion’ < Nw mester / Sw mästare
• SaaN minsttar (gen./acc. minstara) ‘pattern, model’ < Nw mønster / Sw mönster
• SaaN sohkar (gen./acc. sohkkara) ‘sugar’ < Nw sukker / Sw socker
• SaaN šláttar (gen./acc. šláddara) ‘gossip’ < Nw / Sw sladder
• SaaN uŋggar (gen./acc. uŋgara) ‘craving (for a particular food, etc.)’ < Nw / Sw hunger
• SaaN viesttar (gen./acc. viestara) ‘west wind; west’ < ONo vestr

There are also some borrowings from Finnic, although they are much fewer in number:

• SaaN gágir (gen./acc. gáhkira ~ gáhkára) ‘lump of reindeer feces’ < PSaa *kākēr < PFi *kakara (> Fin kakara ‘turd, lump of animal feces; brat’). — As a side note, the Finnic word has an obvious but previously unnoticed cognate in Mordvin: MdE kavoŕks, MdM kavǝŕks, kavǝrks ‘lump, clod (of earth, etc.)’ (< PMd *kavǝŕ-ks < Pre-PMd *kakarV-).
• SaaN ságir (gen./acc. sáhkára) ‘tang of a scythe blade’ < Fin sakara ‘jag, spike, tang’
• SaaN máttar (gen./acc. máddara) ‘ancestor’, SaaL máttar ‘ancestor; base, lower and wider part of something’ < PSaa *mānte̮r < PFi *mande̮r (> Fin manner ‘mainland, continent’). — PFi *mande̮r, in turn, is cognate with Komi (obsolete) mude̮r ‘floor, bottom of a house’, Udm mudor ‘icon; altar or sacred shelf in a prayer hut; deity or sacred center of a tribal territory’. One can reconstruct the common proto-form *me̮ntVr, although the second-syllable vowels in Finnic and Permic do not quite seem to match. In this case there is actually a reason to view the part *-(V)r as a derivational suffix: there are also related forms pointing to a simplex stem *mantV-, e.g. Fin (dialectal) mantu ‘land, area; farm’ (< Pre-PFi ?*mantV-w), which has also been borrowed into Saami (cf. SaaS maadtoe ‘birthplace; ancestry; kindred’, SaaL máddo ‘ancestry; kindred’, SaaN máddu ‘oldest known ancestor; a mythological ancestral form of an animal species’ < *māntō). In addition, there are two related Saami nouns which look like borrowings from (Proto-)Finnic forms that were not preserved in Finnic itself: SaaN máddi ‘south’, máttá-s ‘southwards’, SaaT maanda-s ‘landwards, towards the mainland’ (< *māntē, from unattested PFi *manta?) and SaaS maadtege ‘foot of a tree or a mountain; older generation’, SaaN mátta ‘foot of a tree’ (< *mānte̮k, from unattested PFi *mand-e̮k?). Moreover, one can add that the stem *me̮ntV has a previously overlooked regular reflex in MariE möδǝ-wuj, müδǝ-wuj ‘hummock, tussock’ (< PMari *müdǝ-wuj); the head of the compound is wuj ‘head’ (< PU *ojwa), so the word can be traced back to a metaphoric expression “head of land”, or the like. In light of this data it seems possible that Fin manner is a denominal derivative with a suffix *-(V)r. The analysis remains uncertain, however, as one can also reconstruct a related verb stem *me̮nta- on the basis of MariE MariW müδem ‘cover; bury, cover with earth’ and Udm mudi̮- ‘shovel earth around the foundation of the house (for insulation)’.

Seven words appear to be inherited items with cognates in other Uralic languages:

• SaaSk čuõmâr (gen./acc. čuõmmâr) ‘grain, crumb’ < PSaa *ćuome̮r < post-PU *ćomir (> Fin somero, somer ‘coarse gravel’)
• SaaL dabár (acc. dahparav) ‘prattle, nonsense’ < PSaa *te̮pe̮r < post-PU *tüpir (~ *tüpirä > Fin typerä ‘stupid, foolish’). — The etymological connection between the Saami and Finnish words does not appear to have been previously noticed. In addition, one could tentatively suggest a further connection to Proto-Khanty *tepǟr ‘dust, waste, garbage’ (> V Vj tewǝr, Sur tȧ̆pǝr, Irt tĕpǝr, tȧ̆pǝr, Ni Kaz tăpǝr, O tȧ̆pȧr). In this case the PU form would have been *tipVr(V), and the change *i > *ü in Finnic would be paralleled by Fin tyven ‘calm, windless weather’ < Pre-PFi *tüwin < PU *tiwin (> Kh V Vj teɣǝn, Sur tȧ̆ɣʷǝn ‘calm, windless’).
• SaaN duottar (gen./acc. duoddara) ‘tundra’ < PSaa *tuonte̮r < (post-)PU *tanti/ar (> Fin tanner ‘hard trampled ground; yard, field, open space’, Veps tandar ‘hard trampled ground’). — Kaheinen (2022) has recently suggested that the word has a further cognate in Samoyed: Ngan ćintǝ ‘crest (of hill or moutain), ridge’ and ćintǝrǝǝ ‘mountain’ (? < PSam *ti̮ntǝ̑, *ti̮ntǝ̑rǝ̑jǝ̑). If this is correct, then Saami-Finnic *tanti/ar could indeed originally be interpreted as a derivative of a shorter PU noun stem *te̮nti. However, there is no evidence for such a segmentation outside Nganasan, and also another quite attractive etymology for Saami-Finnic *te̮ntir has been proposed by Rédei (1998). In his view, it was derived from a PU verb stem *tanta- ‘tread, trample’ which has a reflex in Samoyed: NenT tanǝ-, EnF tadu- ‘tread on, step on’ (< PSam *tåntǝ̑-), NenT tanǝʔ-, EnF taduʔ- ‘trample’, Kam tōnuʔ- ‘walk, tread’ (< PSam *tånt-ut-). The meaning shows a quite precise match with a derived verb in Finnic: Fin tannertaa, Veps tandarta ‘trample’ (< PFi *tande̮/arta-). Moreover, this etymology can be corroborated with further cognates in Permic and Mordvin: Komi dud- ‘be obstinate, balk, move backwards (e.g., of horses)’, MdE tandadoms ‘get frightened’, MdM tandadǝms ‘get frightened; buck, bolt (of horses)’. According to Zhivlov (2014: 143) the Komi and Mordvin verbs can reflect a proto-form *tanta-; thus, the phonological match with PU *tanta- ‘tread, trample’ is precise, and even the semantic correspondence is fairly straightforward.
• SaaL guoŋar (acc. guogŋarav) ‘boat rib’ < PSaa *kuoŋe̮r < PU *ke̮ŋir (> Fin kaari ‘curve; boat rib’, KhNi xuŋxarǝ ‘palm of the hand’, MsK kē̮ŋǝr ‘hollow of the knee’)
• SaaSk kõõddâr (gen./acc. kõddâr) ‘hock’ < PSaa *ke̮nte̮r < post-PU *kintir (> Fin kinner ‘hock’)
• SaaSk kõõnjâr (gen./acc. kõnnjâr) ‘elbow’ < PSaa *ke̮ńe̮r < PU *küńir ~ *küńär(ä) (> Fin kyynärä, kyynär-, MdE keńeŕ, MariE kǝ̑ńer, Udm gi̮r-). — Although this is undoubtedly a Proto-Uralic word, the etymology involves some morphological and phonological unclarities. First, only the Skolt Saami form is regularly comparable to the cognates cited above, whereas all other languages show a quite aberrant form: cf. SaaS gernjere, SaaL garŋŋel, SaaN gardnjil, gargŋil, SaaK kâʹrrŋel ‘elbow’ (< PSaa *ke̮rń/ŋēl/rē, as if from pre-PSaa *kürń/ŋäl/rä). Second, there are apparently related words in the Ugric languages which lack the consonant -r- and instead show other obscure the stem-final elements: cf. KhSur kö̆nʾŋi ~ kö̆nŋi, MsK kʷänɣǝľ, Hung könyök ‘elbow’. These words share a common Ugric stem *küṇV- < *künV-; note that the change *kVn- > *kVṇ- is regular in Ugric (Zhivlov 2016). This stem is no doubt of the same origin as the part *küńV- in the word for ‘elbow’ in the more western Uralic branches, despite the irregular correspondence *n ~ *ń. However, since the shared part *künV- / *küńV- is not attested as an independent word-stem anywhere in Uralic, it is not clear whether the part *-r(ä) can really be analyzed as a derivational suffix.
• SaaN muogir (gen./acc. muohkára ~ muohkira) ‘blackfly’ < PSaa *muokēr < post-PU *makar ? ~ *mäkärä (> Fin mäkärä ‘blackfly’). The irregular vowel correspondence makes it is uncertain whether the Saami and Finnic words are direct cognates; they might also be parallel borrowings from some unknown source, for example.

Quite a few Saami noun stems in -r lack an etymology; many of them probably belong to the substrate lexicon Saami has acquired from unknown “Palaeo-Laplandic” languages (see Aikio 2012: 80–88):

• SaaSk aautâr (gen./acc. ahttâr) ‘storm’ < PSaa *ākte̮r / *āvtte̮r. — To venture a speculation, one could think of borrowing from PNo *austrã ‘east’ (> ONo austr) in the meaning ‘east wind’; cf. the aforementioned SaaN viesttar ‘west wind; west’ (< ONo vestr < PNo *westrã). There is no evidence of the proposed semantic shift, however, so this is a mere conjecture.
• SaaN čagar (gen./acc. čahkara) ‘cartilage, gristle; penis (of an animal)’ < PSaa *će̮ke̮r
• SaaN čiegar (gen./acc. čiehkara) ‘winter pasture (where reindeer have dug up lichen under the snow) < PSaa *ćieke̮r
• SaaN čuokkar (gen./acc. čuoggara) ‘lump’ < PSaa *ćuoŋke̮r
• SaaN dieskkar (gen./acc. dieskara), SaaL diesŋar (acc. diessŋarav) ‘fur lining on the inside of mittens’ < PSaa *tiesŋe̮r
• SaaN duogur (gen./acc. duhkora) ‘children’s game’ < PSaa *tuokōr
• SaaN feaskkir (gen./acc. feaskára) ‘porch, entry (of a house)’ < PSaa *feaskēr
• SaaN giegir (gen./acc. giehkira ~ giehkára) ‘windpipe’ < PSaa *kiekēr
• SaaN miegar (gen./acc. miehkara) ‘sleeping tent, shelter from mosquitoes (under which one sleeps outside)’ < PSaa *mieke̮r
• SaaN muttar (gen./acc. muddara) ‘sod, peat (used in the construction of sod huts and houses)’ < PSaa *munte̮r
• SaaN nagir (gen./acc. nahkára) ‘sleep, sleepiness’ < PSaa *ne̮kēr
• SaaL sájger (acc. sájggárav), SaaS saajkere ‘sharp stick (made of wood or antler)’ < PSaa *sājkēr ~ *sājke̮r. — A similar Finnish dialect word saikkara ~ saikara ‘(a kind of) pole or rod; dry branch’ occurs in Ostrobothnian and Far Northern dialects; due to its northern distribution it looks like a borrowing from Saami, although the possibility of borrowing in the opposite direction cannot be completely ruled out. In the Northern Ostrobothnian subdialect the word also has a variant form saikka; this looks like a contamination of saikkara and the etymologically unrelated saitta ‘(a kind of) pole or rod’. The latter is related to SaaN sáiti ‘spear’ (< PSaa *sājttē) via borrowing in one direction or the other.
• SaaN skážir (gen./acc. skáhčira) ‘blade of grass’ < PSaa *skāćēr
• SaaN spáiddar (gen./acc. spáidara), SaaS svaajhtere ‘torch’ < PSaa *svājte̮r ~ *svājtte̮r
• SaaN suonjar (gen./acc. suotnjara) ‘ray, beam (of light)’ < PSaa *suońe̮r. — The word may be somehow etymologically connected to SaaSk šuõnjsted ‘loom, shimmer (of a distant object)’, SaaK šuunjse ‘shine between clouds (of the sun)’ (< PSaa *śuońe̮stē-); the word-initial *ś- could be have developed via assimilation to the following alveolo-palatal nasal *ń. Nevertheless, it would be circular to analyze SaaN suonjar as an example of a derivational suffix -r, as the morphological relationship between this noun and the verb *śuońe̮stē- is not regular and the ultimate origin of both words remains unknown. For all we know, they could also be parallel borrowings from related forms in an unknown source language, or simply coincidentally similar words.
• SaaN šimir (gen./acc. šipmára), SaaI šomer, SaaSk šâʹmmer ‘back of a knife or ax blade’ < PSaa *śimēr ~ *śe̮mēr ~ *śomēr. — The relationship to the similar Fin hamara ‘back of a knife or ax blade’ (< Pre-PFi *šamara) is unclear. The sound correspondence does not support either cognation or direct borrowing between Finnic and Saami, so some kind of indirect etymological connection seems more likely. Moreover, the irregular vowel variation within Saami seems to suggest post-Proto-Saami origin.
• SaaL viettar (acc. vieddarav) ‘high or steep sandy bank’ < PSaa *viente̮r

To come to the point, the lexical material analyzed above offers very little evidence for the existence of a derivational suffix *-r in Saami, or even in Proto-Uralic. Only in the case of SaaN duottar ‘tundra’ and máttar ‘ancestor’ there are any reasons to assume that the consonant -r at the end of the noun stem could be a derivational suffix. At least it must originate in a separate morpheme of some type, but the details remain unclear, as the word-formation has taken place at a very early date and the morphological makeup of the words has become obscured since. Moreover, the two words are more likely to be deverbal than denominal formations, and the hypothesied suffix *-(V)r remains unidentified in any case. What is more, the word máttar was not even directly inherited in Saami, but instead borrowed from Finnic *mand(-)e̮r (< PU *me̮nt(-)Vr).

The scarcity and ambiguity of this evidence implies that denominal nouns with a suffix *-(V)r are extremely rare at best, and more probably do not exist at all. Even if they do, the pattern of word-formation is obviously archaic: the possible examples predate the existence of Saami as a separate branch, and within the Saami branch there is no evidence at all suggesting the reconstruction of such a derivational pattern. In any case it would not have been productive in Proto-Saami anymore, and therefore it could not have been applied to a Slavic loanword thought to have been adopted at that stage.

For the sake of completeness, there is also a slightly different morphological explanation that ought to be discussed: one could think of modifying Koivulehto’s explanation by reconsructing PSaa *kuompe̮rē instead of *kuompe̮r, and assuming that the word then became analogically restructured as a consonant-stem in *-r in Saami. In this way one could hypothesize that the word originally contained a derivational suffix *-rē (< PU *-rA). In fact, it is not always easy to tell the difference between Saami noun stems in *-rē and *-r, and in individual languages one does find some examples of the latter having originated by analogy from the former. One such word, in fact, has been suggested by Nikkilä (1993: 96) to be a Germanic loanword with a suffix *-rē added in Saami:

• SaaU guöhpiere, SaaP guohper, SaaL guober (acc. guohperav), SaaN guobir (gen./acc. guohpira ~ guohpára) , SaaI kyeppir, SaaSk kueʹpper ‘hoof’ < PSaa *kuopērē (~ *kuopēr) < Pre-PSaa *kapa(-)ra ? < Pre-PGerm *kāpa- (> PGerm *xōfa- > OEngl hōf, OHGerm huof, ONo hófr ‘hoof’). — Note that SaaS guehpere ‘hoof’ looks like a borrowing from Ume Saami due to its unexpected -hp-; the expected inherited reflex would be *guepere.

This word does not offer an exact parallel for the word *kuompe̮r ‘mushroom’, however. It is quite obvious that the word for ‘hoof’ must be reconstruced as PSaa *kuopērē: only SaaL guober and SaaN guobir are gradating r-stems, and they must have secondarily developed via analogy, as non-gradating cognates occur in the Saami languages both to the southwest and to the east of Lule and North Saami. In the case of *kuompe̮r, however, there is no evidence of an earlier stem in *-rē; the word behaves everywhere as a gradating r-stem.

Besides this, the loan etymology suggested for *kuopērē is also problematic in itself. One certainly can question whether there even was a productive derivational suffix *-rA in Pre-Proto-Saami at the time when this Germanic word is assumed to have been borrowed; and even if there was, it is not clear at all what kind of semantic function it served and what types of bases it attached to. In fact, the entire material suggesting the reconstruction of the derivational suffix *-rA in Proto-Uralic is scarce and opaque. The only thing that looks clear is that such a suffix did once exist, as it is attested in two derivatives reconstructible into Proto-Uralic; one of them is denominal, the other deverbal:

• Fin koira ‘dog’, koira-s ‘male’, Võro koir ‘male dog’, KomiJ kȯr ‘male dog’, KhVVj kar, MsK xār, kē̮r ‘male; reindeer bull’, Hung here ‘testicle’, NenT xora, SlkTa qori̮ ‘male; reindeer bull’ < PU *koj-ra ← *koji ‘man, male’ (> SaaSk kuõjj ‘(young) husband’, KhVVj ku ‘man, husband; male animal’, MsK xŏj, kŏj ‘male animal’).
• Fin kumara ‘stooped, bent over’, MariNw kŏmŏr ‘brushwood, windfall, driftwood’, Komi ki̮mi̮r ‘slouching’, Hung homor-ú ‘concave’ < PU *kuma-ra ← *kuma- ‘fall or bend over’ (> MdE koma-, Komi ki̮m-, MsK xåm-, kåm-, NenF kămă-, EnT koo-, Ngan kǝmǝ-). — The appurtenance of MariNw kŏmŏr has not been previously noticed; as regards semantics, a parallel is provided by another noun derived from the Samoyed reflex of this verb: NenF kăm°xa, EnT kooxa, Ngan kǝmǝgu ‘fallen tree’, Mat kamaga ‘block of wood, driftwood’.

In addition, there are some examples of derivatives in *-rA in Finnic:

• Fin avara ‘wide and open, spacious’ ← PU *aŋa- ‘open / take off’ (> KhIrt eŋx-, MsN āŋkʷ- ‘take off’); cf. also Fin avata : avaa- ‘open’ (verb) < PFi *avaida-, a parallel derivative of the same base.
• Fin kamara ‘hard crust (esp. crust of the earth); pork rind’ ← PU *kama (> MariE kom ‘crust (of bread), peel (of fruit)’, NenT śaw°, SlkTa qāmi̮ ‘fish scale’).
• Fin katkera ‘bitter’ ← PU *kačka- ‘bite (?)’ (> SaaN gáskit ‘bite’, MariE kočka- ‘eat’, KhVVj ki̮č- ‘hurt, ache; sting (of nettles)’, MsN xūs- ‘sting (of nettles)’) (Aikio 2014: 5–8).
• Fin pisara ‘drop’ ← (post-)PU *pi(ń)ća- (> MdE piźe- ‘rain (verb)’).
• Fin tappura ‘rough hemp or cotton fibers, waste fibers’ ← PU *tappa (> MdE tapo ‘shaggy, tangled linden bast or tow’, MariW top ‘ball, bundle, sheaf (e.g., of wool or bast)’). Note also the derived verb MdE taparda- ‘wrap, swaddle, wind’ ~ Komi tupir̮t- ‘wind, reel, coil’ (< *tapparta-), and Komi tupi̮ľ ‘ball (of yarn), coil; lump’ (with an opaque stem-final element –ľ). This is a new etymology; the background of the second-syllable u in Finnic is obscure, but otherwise the comparison is quite straightforward.
• Fin (obsolete) viherä ‘green’ ← PU *wiša ‘green / yellow’ (> MdE ožo ‘yellow’, Komi vež ‘green, yellow’, Udm vož ‘green’); cf. also Fin vihanta ‘lush, green (of vegetation)’, a parallel derivative of the same base.

In Saami, however, I have found only a single example of this kind of derivative:

• SaaSk tooppâr, SaaT toʹbbear ‘weather with sticky snow’ < PSaa *tope̮rē ~ *topērē ← *tope̮- (> SaaL dåhpåt, SaaN dohpat, SaaI tuuppâđ ‘stick on (skis, etc.; of sticky snow)’). There is also a derived verb with -r-: SaaSk toppred, SaaK tobbre (< PSaa *tope̮rV-), SaaI toperuđ ‘stick on (skis, etc.; of sticky snow)’ (< *topēr-uove̮-).

As far as I am aware, this example is completely unique in Saami: there appear to be no other noun stems in *-rē (or verb stems in *-rV-) where that stem-final element could be segmented as a suffix. Furthermore, the derivative has an archaic look to it, as it appears to have developed an irregularly syncopated variant already in Proto-Saami, which then served as the base for further derivatives. SaaT tobbrnânndâd ‘stick on (of sticky snow)’ (< PSaa *topre̮-n-e̮nte̮-) is a frequentative verb formed from an otherwise unattested PSaa *topre̮-ne̮- (an intrasitive transformative verb formed from a syncopated stem *topre̮-). This verb comes phonologically and semantically extremely close to another verb stem which has lacked an etymology: SaaS dabraanidh, SaaU SaaL dabrránit, SaaN darvánit, SaaI tarvaniđ ‘stick on, get stuck’ < PSaa *te̮prā-ne̮-. The irregular vowel change *o > *e̮ has occasionally taken place adjacent to labial consonants in Saami: cf. such cases as SaaN bohčit ~ bahčit ‘squeeze, wring; milk’ (< PSaa *poćē- ~ *pe̮ćē- < PU *puća-), botnit ~ batnit ‘twine, spin’ (< PSaa *ponē- ~ *pe̮nē < PU *puna-), monni ~ manni ‘egg’ (< PSaa *monē ~ *me̮nē < PU *muna), SaaK kâʹppse, SaaT koʹppsed ‘go out (of fire)’ (< PSaa *kopsē- ~ *ke̮psē- < PU *kupsa-), SaaN bordit ~ bardit ‘stack up; load (a boat)’ (< PSaa *portē- ~ *pe̮rtē- < PNo *burdja- > ONo inn-byrða, Icel byrða ‘pull onboard’, Far byrða ‘burden’; a new etymology for the Saami verb).

Therefore, it is reasonable to assume that PSaa *tope̮-rē is an archaic formation inherited from an earlier language stage when the suffix *-rA was still productive. On the basis of Saami we can thus tentatively reconstruct post-PU *tupi- and *tupi-ra, and indeed, for the derived form a plausible cognate is found in PMari *tŭwǝr-. This stem underlies the derivatives MariE tuwǝ̑rɣem, MariW tǝ̑wǝ̑rɣem ‘curdle (of milk); clot, coagulate (of blood)’ (< PMari *tŭwǝr-g-e-) and MariE tuwǝ̑rtem-, MariW tǝ̑wǝ̑rtem ‘make (milk) curdle, make curds’ (< PMari *tŭwǝr-t-e-). Although the meanings in Saami and Mari are different, the connection is transparent.

The conclusion regarding PSaa *kuompe̮r ‘mushroom’ is that it is very unlikely to have been derived with a suffix *-r or *-rē from a stem borrowed from PSlav *gǫba ‘mushroom, fungus’. The question, then, is whether we can accept a loan etymology which entails the assumption that an obscure suffix-like element without any recognizable semantic function was added to the word-stem. In fact, such an assumption does not inevitably invalidate an etymology; in inherited Uralic vocabulary one finds a few examples of the phenomenon that are well-established and difficult to dismiss. In Saami, I managed to find the the following nine examples:

• PSaa *ćeapēttē ~ *ćeapōttē ‘neck’ (> SaaU tjiäbuote, SaaL tjiebet, SaaN čeabet, SaaSk čeäppat, SaaT čeabbad) ← *ćeapē ‘neck’ (> SaaS tjiepie, SaaL tjiehpie). — No noun suffix *-ttē is known. The stem reflects PU *ćepä ‘neck’ (> Veps seba, MdE śive ‘collar’, MariE šüj, MsN sip ‘neck’).
• PSaa *earttiŋkV ‘ribs (meat cut)’ (> SaaS eerhtege, SaaL hiertig, SaaN erttet, SaaI eertig) ← *earttē ‘side (body part or meat cut)’ (> SaaI ertti, SaaSk jeäʹrtt, SaaT jieʹrhte). — No suffix *-(i)ŋkV is known. The stem reflects Pre-PSaa *erttä, cf. MdE iŕďes ‘rib’, MariE erδe ‘thigh’, Udm urd ‘side; rib’ (< *ertä). The correspondence between *-rtt- in Saami and *-rt- in the other branches has not been explained. The word is a well-known borrowing from Pre-Proto-Indo-Iranian *(H)érdʰo- (> Proto-Indo-Iranian *(H)árdʰa- > Sanskrit árdha- ‘side, part, half, place’, Avestan arǝδa- ‘side, half’).
• PSaa *jievje̮mē ~ *jievjōmē ‘lichen on trees’ (> SaaN jievjun, SaaI jievjâm). — The part *jievj(e̮)- regularly reflects PU *jäwji (> KhVVj jej ~ jĕj, NenT juj°, Ngan ďiǝ ‘lichen on trees’). However, no denominal noun suffix *-(ō)mē is known (although *-mē < PU *-mA is a fully productive deverbal noun suffix that forms action nouns). Note also that in southwestern Saami the word shows an irregular first-syllable vowel: SaaS joevjeme ~ jovjeme, SaaU jåvjjamah (pl.).
• PSaa *kāme̮k ‘shoe’ (> SaaS gaamege, SaaN gáma, SaaI kaamuv, SaaT kaamâg). — The word must have been derived with a suffix *-e̮k from the otherwise unattested stem *kāmē, which is the regular reflex of PU *kämä ‘(a kind of) shoe’ (> MdE keme, MariE kem ‘boot’, Komi ke̮m ‘bast shoe’). However, PSaa *-e̮k is not a known denominal noun suffix (although it is a highly productive deverbal noun suffix). Note that the former existence of the stem *kāmē is also implied by the parallel derivative *kām-e̮s ‘leg skin’ (> SaaS gaamese, SaaN gámas, SaaI kaamâs, SaaT kaams). The suffix *-e̮s forms denominal nouns that denote materials used in the manufacture of the referent of the base noun. The underlying meaning of *kām-e̮s ‘leg skin’ is thus “material for shoes”; traditional Saami fur boots are sewn from reindeer leg skins. As a side note, this morphological analysis contradicts Koivulehto’s (2007: 584–587) proposal that the Uralic word was borrowed from PGerm *xammō- ‘shank’: the original meaning of the former was obviously ‘(a kind of) shoe’ and not ‘leg skin’, and therefore the semantic development ‘shank’ > ‘leg skin’ > ‘shoe’ that was suggested by Koivulehto cannot be assumed. The etymology also involves other difficulties: the Uralic front vowel *ä would be an unexpected substitute for PGerm *a, and moreover, it is a priori unlikely that a word having regular cognates in as far east as Permic could have been borrowed from a language representing the Proto-Germanic level of reconstruction. This combination of semantic, phonological and chronological problems implies that the etymology must be wrong.
• PSaa *koackēmē ‘eagle’ (> SaaL goasskem, SaaN goaskin, SaaSk kuäʹcǩǩem, SaaT kïõʹcckem). — The part *koackē- is the regular reflex of PU *kočka ‘eagle’ (> Fin kotka, Komi kuč). Regarding the suffix *-mē see *jievje̮mē ~ *jievjōmē ‘lichen on trees’ above.
• PSaa *luompe̮l ‘small lake (through which a river runs)’ (> SaaS loebpele, SaaN luoppal, SaaSk luubbâl, SaaK luumbâl). — No suffix *-l is known, but the part *luompe̮- regularly reflects PU *le̮mpi (> Fin lampi ‘pond, small lake’, Ngan ľüŋhǝ, SlkK li̮mbi̮ ‘boggy place, quagmire’) (Aikio 2014: 86).
• PSaa *oalōl ‘lower jaw’ (> SaaL oalol, SaaN oalul ‘lower jaw’, SaaI uálul-tähti ‘jawbone’). — No suffix *-(ō)l is known (cf. the case of *luompe̮l above), but the part *oal(ō)- goes regularly back to (post-)PU *ola- and matches MdE ulo ‘chin’. Probable further cognates include MsK ūlǝś ‘chin, lower jaw’ (< PMs *ūlǝć, with an obscure stem-final element *-ć) and Hung áll ‘chin’ (with -ll < *-lCV, thus also originally containing an obscure stem-final element). Note that SaaSk vuål-täʹhtt, SaaK vual-taaʹhht ‘cheekbone’ probably do not preserve the shorter stem without *-(ō)l: the compound is otherwise identical to the aforementioned SaaI uálul-tähti, so its modifier has probably undergone haplology (PSaa *oalōl-tāktē > *voalal-tākte > *voal-tākte).
• PSaa *peane̮k ‘dog’ (> SaaN beana, SaaI peenuv, SaaT pienâg). — The word must have been derived with a suffix *-e̮k from the otherwise unattested stem *peanē, which is the regular reflex of PU *penä ‘dog’ (> MdE pińe). Another derivative (with a known suffix) is PU *penä-j ‘dog’ (> Fin peni, MariE pij, Komi pon : ponj-, Udm puni̮). Regarding the suffix *-e̮k see *kāme̮k ‘shoe’ above. Note, moreover, that in southwestern Saami the word appears in an irregular form: SaaS bïenje, SaaU biäŋŋa ~ biägŋa ~ biädnja (< *pieńe̮ ~ ?*pieŋe̮). Although this form seemingly lacks the suffix *-e̮k, it cannot be a direct reflex of the simplex stem *penä either because of its irregular vowels and the irregular place of articulation of the nasal; the predicted regular reflexes of PU *penä would be SaaS *bienie and SaaU *biennie ~ *biednie. The development of this form remains unexplained.
• PSaa *pe̮ŋkōj ‘hazel grouse’ (> SaaL bakkoj, SaaN bakku, SaaSk pââgg). — The Saami word regularly reflects Pre-PSaa *püŋkäw, so it looks like an archaic consonant-stem derivative of PU *püŋV (? ~ *pi/eŋV) (> Fin pyy, MdE povo, KhVVj pĕŋk ‘hazel grouse’, Hung fogoly ‘partridge’; the last with obscure stem-final element -(o)ly). The denominal noun suffix *-kA can be reconstruced into Proto-Uralic, but its semantic function is obscure (Aikio 2022: 19). There are also some possible traces of an opaque denominal noun suffix *-w, for example SaaL guottoj ‘fallen tree’, SaaN guottu ‘tree stump’, Fin kanto ‘tree stump’, MsN xānta ‘horizontal floor beam of a storehouse’ < PU *ke̮ntaw ← *ke̮nta (> Fin kanta ‘base; heel’; note also MdE kando ‘fallen tree’ and KhVVj kant ‘pillar of a storehouse’, which could reflect either the simplex or the derived stem). Regarding *pe̮ŋkōj ‘hazel grouse’, however, it remains totally unclear why two opaque derivational suffixes would have been added to the base.

In the cases discussed above an inherited noun stem has been augmented by adding an element that looks like an opaque derivational suffix, but the process is not accompanied by any semantic change: the meaning of the derived form in Saami matches that reconstructed for the Proto-Uralic simplex stem. One could therefore ask whether the same kind of process could also have affected some loanwords. The answer is, of course, that it might; but it is quite another issue whether such cases could be reliably identified, and whether invoking such an obscure process could be methodologically justified in loanword research.

The key issue here is probability. According to my calculations there are over 300 inherited noun stems in Saami with cognates in other branches of Uralic, so the nine examples discussed above amount to less than 3% of all inherited nouns. If we had a large stratum of Slavic loans in Saami that contained, say, 100 borrowed nouns, it would then make perfect sense to expect a couple of them to contain an obscure stem-final element. But the situation is completely different when we are dealing with merely two nouns alleged to be stray borrowings from Proto-Slavic. In such a situation we also need to establish the existence of the alleged loanword stratum itself, and for this purpose unambiguous and impeccable etymologies are needed as evidence. If we permit ad-hoc postulation of unknown “suffixes” to account for non-matching stem-final material, the possibility of finding chance correspondences increases, which in turn weakens the hypothesis of the very existence of a Proto-Slavic loanword stratum in Proto-Saami. In this regard the Slavic loan etymology of PSaa *kuompe̮r ‘mushroom’ differs crucially from the Uralic etymology of, e.g., PSaa *peane̮k ‘dog’: we already know that Saami has inherited several hundred words form Proto-Uralic, and moreover we can reconstruct the noun *penä(-j) ‘dog’ on the basis of Finnic, Mordvin, Mari and Permic forms, completely regardless of the origin of PSaa *peane̮k ‘dog’. Therefore, there is a quite different set of facts that leads us to conclude that PSaa *peane̮k consists of an inherited stem *pean(ē)- and an opaque suffix *-e̮k.

To conclude this analysis of PSaa *kuompe̮r ‘mushroom’, the etymology deriving it from Proto-Slavic *gǫba turns out to be quite weak because it fails to explain the stem-final consonant *r. This does not completely disprove the hypothesis, as there are some well-established parallels for the addition of an obscure stem-final element on Saami nouns, but the rarity of such processes is a significant weakness in the etymology. Moreover, even if the hypothesis of a link between the Saami and Slavic words were correct after all, there would still not be compelling reasons to interpret the word as a borrowing from Proto-Slavic: the word could also have been adopted from Proto-Baltic or Proto-Balto-Slavic into Pre-Proto-Saami. Therefore, the etymology proposed for PSaa *kuompe̮r does not offer support for direct contacts between Proto-Slavic and Proto-Saami.

Conclusion

In this study the two Proto-Slavic etymologies that have been proposed for Saami words were critically reviewed. The word *multtē ‘soap’ turned out to have an alternative and far more plausible Proto-Norse loan etymology, and the Slavic loan etymology of *kuompe̮r ‘mushroom’ turned out to be weak because it does account for the stem-final consonant *r. Therefore, it is very proable that the striking matches of these words with PSlav *mỳdlo ‘soap’ and *gǫba ‘mushroom, fungus’ are examples of the phenomenon of ‘chance correspondence’: even though the phonological and semantic match beween these forms can be described as ‘regular’ by criteria of historical linguistics, it nevertheless in all likelihood stems from pure coincidence.

There is an important methodological lesson in this. Although historical linguists are well aware of how common chance similarities are across languages, the possibility of finding seemingly regular chance correspondences is largely not taken into account. The latter are, of course, much rarer than the former, but apparently not as rare as is generally thought. Every once in a while a professional historical linguist will encounter an etymology which, at face value, looks phonologically and semantically impeccable, although other criteria may strongly suggest or even prove that it is wrong. This means that such etymologies occasionally also occur when there are no criteria to prove them wrong. Because of this, no far-reaching conclusions should be drawn or broad generalizations made on the basis of just a couple of etymologies, no matter how convincing they seem.