2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Discussion Delete this Discussion 129 Participant Discussion ended on April 23, 2021 +125 Kirill Reshetnikov 3 hrs ago In Yenisseian, there is also a word-initial correspondence “Kott h- ~ Ket, Yugh, Arin Ø” (in Pumpokol, no respective data is attested), e.g. Ket Λľa ~ Yugh Λrej ~ Kott hili ‘outwards’ or Yugh εхtaŋ / aхtaŋ ~ Arin ittä ~ Kott hītēg ‘belt’. There are not so many examples, but this is obviously still regular and it is just Proto-Yen. *hthat is reconstructed word-initially; S.A. Starostin’s Proto-Yen reconstuctions for the above etymons are *hər1and *hΛqtΛ respectively (see Праенисейская реконструкция и внешние связи енисейских языков, 175; the work can be easily found, although there seem to be some technical problems with a direct link). So the Yenisseian word for ‘island’ is tentatively considered to belong just to this group: *h- is reconstructed because of the supposed relation between Ket ēje1, ēj1 ‘island’, ei-tu ‘Flussbusen’ and Kott hau-tu ‘Flussbusen’. As for the monosyllabic shape of the Proto-Yen. reconstruction, it is indeed not quite correct and probably derives from a hyphenated notation like *h[e]j- meaning that the given part of the form is followed by some further elements in the daughter languages (cf. Starostin’s Сравнительный словарь енисейских языков, 230, 288 in Ketskij sbornik published in 1995; this is just CCE 230!). Of course, all this can be contested, but my point is that all the “obscurities” with this reconstructed form do have their explanation. And now something on the main subject: why still “Proto-Samoyed *kat ‘to bind, sew’” and “*kät in SW 68”, while it is *ket1 in SW and therefore should be just *kät in accordance with the updated reconstruction system? As it was already mentioned here in the discussion, this word has some FU parallels. I actually have some more remarks of the same type, but I unfortunately had no time to share them. Briefly, I think that at least some of the Sam words in question should be additionally checked for Uralic cognates. Like Peter S Piispanen 7 days ago The etymology of Proto-Samoyed *woej 'island' (SW 177) is not at all clear to me. Attestation includes Nganasan uai 'island'; Enets n'ue, nuij 'id.'; Tundra Nenets ŋo 'id.'; Forest Nenets ŋoo 'id.'; Selkup ko~ku 'id.'. Now, is this reconstruction correct? The Enets and Nenets n'- and ŋ- looks epenthetic, and only Nganasan suggests something like *w- to me. An original PU *w- is usually found as a retained w- in Nenets, b- in Enets and Nganasan, and as q- or k- in Selkup AFAICT, and this is not what we see here for 'island'. There is actually partly similar Proto-Yenisseian *hej 'island', attested as Ket ēje 'island'; Yug ēj 'id.'; Kottish https://www.academia.edu/s/010b2ad552#comment_809468 1/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu hau-tu 'Flussbusen' (CCE 230). Clearly, an assimilation has produced the long vowels in Ket and Yug. Could the Yenisseian and Samoyedic words be connected through borrowing? Like Eugen Hill 6 days ago 'There is actually partly similar Proto-Yenisseian *hej 'island', attested as Ket ēje 'island'; Yug ēj 'id.' ... Clearly, an assimilation has produced the long vowels in Ket and Yug.' What is your source for Ket and Yugh? In many sources the macron doesn't mark length but only the first (high-even) intonation. Ket ēje must be either North or Central Ket because of the preserved -e which would be regularly dropped in the 'standard' South Ket dialect as it is in Yugh. I wonder how the ProtoYenisseain word can be monosyllabic but still produce ēje in dialects of Ket. What do you mean by ‘assimilation’? Why *h- in the reconstruction? Because of Kott? Doesn’t Kott h regularly correspond Ket q, Yugh χ? Like 1 Juho Pystynen 6 days ago *wo- > *o- (> *ŋo-) is apparently regular across northern Samoyedic, attested also in *wota 'berry' > TN ŋōďā, FN ŋōćā, TE ore, FE ode, Ng ŋuta (~ Selkup ⁽*⁾kotə; ? Kamassian mōdo with irregular m-). Nganasan ŋuai (Castrén's ‹˜uai›; not **uai) is simply from *oaj < *oəj. Unlike 2 Ante Aikio 6 days ago At least in Kosterkina, Momde & Zhadanova's dictionary ("Slovar' nganasansko-russkij i russkongansanskij", 2001) the form of the word is <ӈүай> = /ŋüaj/, which is problematic in regard to the proposed PSam reconstruction *woǝ̑j. While *w- > *Ø- (> *ŋ-) seems to be regular begfore *o and *u, the vowels are harder to reconstruct. Nganasan /ü/ suggests PSam *u, moreover, and /a/ as the second component of the Nganasan vowel sequence does not support the idea that we are dealing with a PSam vowel sequence with *ǝ̑ as the second member. Because Nganasan second-syllable /a/ goes back to PSam *a (in Helimski's revised reconstruction), so at least on the face of it, this would seem to suggest something like PSam *(w)uaj. I'm not enttirely sure, however, whether the assumed Selkup cognate and the other Samoyed forms are compatible with this reconstruction. Like 1 Juho Pystynen 6 days ago We find *o >> ü also in *koəj > küə 'birch' (C ‹kụa›), *poəj > hüə 'year' (C ‹fụa›), presumably by the effect of the stem-final *j, given the contrast with *mo(ə) > muə-ďə 'branch' (C ‹môja›). At least the first of these clearly has original *o, coming from PU *kojwV > Fi. koivu etc. It's very possible of course that maybe much or all of the PS vowel sequences need general https://www.academia.edu/s/010b2ad552#comment_809468 2/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu reorganization, e.g. Wagner-Nagy in FUM 26/27 already suggests *kuå, *poə̈, *moə̑ for the words above (though without any detailed argument). Like Peter S Piispanen 6 days ago @Eugen: the reconstruction is not mine, but the materials are taken from someplace referred to as CCE 230, whatever that may be in the starling-database. But these words are discussed in Werner's volumes of course, and where the Kottish form hau-tu 'Flussbusen' is described as cognate with Ket ei-tu 'Flussbusen' (I am not completely convinced by this). However, both of these words may be entirely unrelated to the Ket and Yug words at hand because I too agree that the regular reflex of Kott h should be Ket q, Yugh χ. Therefore, the proper Yenisseian reconstruction might be something very close to *ēje 'island'. With assimilation I simply mean *he- > ee- if the macron signifies a long vowel sound. You are also correct in that the macron might mean first high-even intonation, but I cannot verify this as I don't have access to Werner at the moment. In any case, the Ket and Yug words for 'island' appears to be quite close to the Samoyedic words, but the latter words may not all be traced back regularly to one root as others have suggested above. Like B. Blasebalg 6 days ago Peter, as far as the Yenisseian words are concerned, I wonder how you would vindicate a change in meaning from "island" to "Flussbusen". Germanic *ahwo:, *agwijo: and their large set of cognates only seeminly sets up a comparable case: There we have a meaning "water body" for the source word and "land near the water" for the derivation. This logic does not work when you start with the meaning "island" and try to derive a certain part of a water body. Is "-tu" known to be a Ket/Kott derivational component, do the words look like composites, or are even the segmentations 'ei-tu/hau-tu' open to doubt? Unlike 2 Peter S Piispanen 5 days ago @Blasebalg: I agree, that exactly is the problem I had - I am not convinced that 'island' is to be necessarily connected with 'Flußbusen', which I take to mean some type of 'riverbed' or 'river bay' (perhaps throught the idea of a 'peninsula'?). Werner (2 262) suggests that the second word -tu means 'back, place between shoulders' (although this seems to go back to *-tuGv), but I am not certain that this either is the correct interpretation of the compound meaning 'Flußbusen' although it could be the case. In any event, I think it is best to treat Ket ēje 'island' and Yug ēj 'id.' as entities independent from the Kott and Ket compound forms for the sake of comparison with the Samoyedic forms (which in themselves may not all originate in one common etymon). Like https://www.academia.edu/s/010b2ad552#comment_809468 3/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu 1 Eugen Hill 5 days ago I checked the entry in StarLing database. The Ket and Yugh word is indeed spoken with high-even intonation (marked with 1 following Werner’s system). Ket ēj1 is explicitly characterised as ‘South’. The macron means length. Nevertheless, the picture is quite streightforward. In Ket all vowels spoken with high-even tone are always half-long. In Yugh only those which are in words having lost a second syllable, all other vowels bearing high-even tone are short. Here are the same data in the new transcription (i.e. following Vajda and Georg): North or Central Ket ēˑje, South Ket and Yugh ēˑj. I still don’t know what CCE 230 means and why the Proto-Yenisseian reconstruction beginns with *h and lacks *-e. If I remember it right, Vajda once proposed that such vowels as the second -e in Central and North Ket ēˑje are recent (he calls them ‘excrescent’). This is certainly wrong for North Yenisseian (i.e. Ket and Yugh) because: (a) in South Ket and Yugh all short vowels are dropped at word-ends, including for instance the verbal root -a ‘to eat’, (b) in South Ket and Yugh word-final k, q and p are phonteically realised like their intervocallic counterparts if the relevant word in C and N Ket ends in a vowel. It follows that such words as ‘island’ must have been disyllabic in Proto-North-Yenisseian (i.e. the ancestor of Ket and Yugh). Unlike 4 Peter S Piispanen 5 days ago Excellent analysis, Eugen - thanks! I take it then that the Proto-North-Yenisseian form should have been something akin to *ēˑje 'island'. The only question then that remains is, can this root be connected to certain groups of Samoyedic words meaning 'island'? Like B. Blasebalg 3 days ago Peter: OK, "back of an island" as a motive for naming a river bay ('Flussbusen') is mildly plausible. Like Eugen Hill 3 days ago @ Peter ‘The only question then that remains is, can this root be connected to certain groups of Samoyedic words meaning 'island'?’ I don’t know. It depends on two factors: (a) the PSam and/or Proto-North-Sam and/or Proto-South-Sam reconstruction (i.e. PSam *uaj? ProtoNorth-S ?, Proto-Nenets-Enets ?, Proto-Selkup ?), https://www.academia.edu/s/010b2ad552#comment_809468 4/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu (b) whether it can be shown that one of these potential reconstructions is compatible with Proto-NorthYen *ēje (i.e. Proto-E form had *ö and it can be shown that in other Proto-North-Yen lexemes supposedly borrowed from Proto-E this *ö is also substituted by *e etc.). I.e. the matter should be investigated systematically. The problem is, it is essential to take into account both the Samoyed and the Yenisseian side. But who is competent in both? I happen to know some Yenisseian but my knowledge of Samoyed is by far not sufficient. Like 2 Peter S Piispanen 3 days ago @Eugen: yes, in this case, as we have seen by the posts by Aikio and Pystynen, the reconstructed ProtoSamoyed form is not secure in itself: it may be *wuaj 'island', but the Selkup form may not quite fit there, so this may be a Proto-North-Samoyed word (with regular loss of *w- in Northern Samoyedic). I am myself starting to get away from the whole North/South dichotomy in favor of the groups Nganasan, Mator, Nenets-Enets and Kamassian-Selkup, but this has no real bearing on the hypothesis at hand. On the Yenisseian side, we appear to have Proto-North-Yenisseian *ēje 'island' based on your argumentation. So, the forms are quite close, but not a perfect match. Factors speaking in favor of a borrowing: a. the Samoyedic vowel cluster could reflect the Yenisseian tone, b. both words have the *-j-, c. the meanings are identical, d. the word appear to be geographically limited, however it is so in both language groups! d. there does appear to exist some phonological peculiarities on the Samoyedic side, which is a diagnostic of it being borrowed. Factors speaking against a borrowing: e. the Yenisseian word is disyllabic, while the Samoyedic word is monosyllabic: this is not a huge hurdle however because Samoyedic has more than a little tendency to reduce the number of root syllables. f. we do not really have secure chronological details when said borrowing could have taken place, and we therefore cannot say which sound laws should be applied and which not (for example already Proto-Samoyedic had undergone a lot of changes since Proto-Uralic). As far as I can tell, there have been no major papers dealing specifically with prospective lexical borrowings between Samoyedic and Yenisseian. If we could find additional attestation of this word in some Samoyedic or Yenisseian languages it could perhaps clarify matters. Like Ante Aikio 3 days ago I took a closer look at Nganasan words/forms with the vowel sequence /üa/ in the dictionary by Kosterkina & al. There are very few, and they did not make the issue any clearer to me at least. For /hüǝ/ 'year', two alternative PL.GEN forms are given: /hüǝʔ/ ~ /hüaʔ/. If we assume that the former is analogical and the latter a more archaic form, then perhaps this could suggest that /a/ somehow https://www.academia.edu/s/010b2ad552#comment_809468 5/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu developed under the influence of following /j/ (because the pl.GEN suffix goes back to *-jʔ)? However, no other noun of the type /CVǝ/ has a PL.GEN variant of type /CVaʔ/ in the dictionary. Then there was also one derived verb with variation, /hüǝlǝ-/ ~ /hüalǝ-/ 'blow' derived from PSam *puǝ̑-; the background of /a/ here is not at all clear to me. In the parallel derivative /hüar-/ 'blow', the vowel /a/ is perhaps a part of the suffix (? PSam *pu-e̮r-, with regular *e̮ > Nganasan /a/?). Then there's Nganasan /ďüa-/ 'interfere, distract, prevent', and Castrén also has <jụai> = */ďüaj/ 'fence', which correspond to Tundra Nenets /jū-/ 'dam a river, block a river with a fish weir' and /jū/ 'dam, fish weir'. But here the vowel correspondence is different from that in Nganasan /ŋüaj/ ~ Tundra Nenets / ŋo/ 'island'. There's also the bird name /ďüari̮ ǝ/ 'гусь-пискулька' (~ Forest Enets /ďora/ 'гагара-крохаль'), but this seems to have no Nenets cognate. Moreover, in the last two words /ü/ could in priciple also secondarily go back to earlier */u/ < PSam *o, because in Nganasan also an underlying /u/ of any origin changed to /ü/ after word-initial /ď/ (< *j). Unlike 1 Ante Aikio 3 days ago @Peter Note that the (North) Samoyed word clearly is not monosyllabic, but disyllabic instead; Nganasan /VV/sequences are really sequences of two vowels that belong to separate syllables (they are not diphthongs or long vowels), and the vowel sequences reconstructed for Proto-Samoyed must be interpreted in the same way. And in the case of this 'island' word, we also have an Enets cognate that apparently preserves the bisyllabic structure: modern Forest Enets has /nuj/, but Castrén documented the word as "ńue". Note that PSam *a regularly developed into Enets e, so this could support the reconstruction of a sequence *-ua-. However, the initial nasal in Enets is strange. Like 1 Peter S Piispanen 3 days ago @Ante: thank you for this and the above comments on the reconstruction! I was kind of hoping that the Proto-Samoyed form could also have been disyllabic to place it more in line with the Yenisseian form. I am thinking that, in Castrén, the Enets word given with ń- actually ortographically reflects an ŋ- instead, that is the epenthetic consonant. Clearly, there are some oddities going on with the vowels in these words, but you may be onto something with the assumed final *-j, which certainly could influence a vowel change in the preceeding position. With the disyllabilicity of the Samoyed root for 'island', do you suggest that there used to exist a now eroded sound between these two vowels? Clearly, in the word for 'to blow' it did (which I take is the usually assumed *x). Seemingly, the best reconstruction we can make is Proto-North-Samoyedic *wuaj https://www.academia.edu/s/010b2ad552#comment_809468 6/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu 'island', no? But alternatives do exist. That is, even though this does not capture all the attested forms; the Selkup form, for example, is, I am guessing, non-related to the other ones. Like Ante Aikio 3 days ago As far as I can see, Proto-Samoyed vowel sequences have indeed always developed through loss of an intervocalic consonant (as I've argued in my 2012 paper "On Finnic long vowels, Samoyed vowel sequences, and Proto-Uralic *x"). The details are of course complicated, but it appears that the lost consonant may have been *w, *l, or *j. Words with PU *x, however, appear not to have vowel sequences in Samoyed, but they are reflected as monosyllabic vocalic stems (e.g., PU *mexi- 'sell' > PSam *mi-, PU *ńoxi- 'chase' > PSam *ńo-). I would reconstruct PU *puwa- 'blow', as this also accounts for MdE /puvams/ 'blow'. However, these processes are only known to have produced Proto-Samoyed vowel sequence of the shape *-Vǝ̑-. We can, perhaps, also reconstruct sequences of the type *-Va- (as in PSam ?*(w)uaj 'island'), but I have no idea what their earlier development and ultimate source was. Note, however, that Nganasan also has numerous secondary vowel sequences that have developed through regular loss of intervocalic *j, as in e.g. Ngan /kou/ 'sun' < PSam *kåjå, Ngan /muaŋ/ 'torment, trouble' < *måjan (cf. NenT /majaʔ̰/). In some types of stems also alternation between sequences /Vj/ and /VV/ developed: e.g. Ngan /ŋoj/ 'foot' : PL.NOM /ŋuǝʔ/ 'feet' (< PSam *åj : *åj-ǝ̑t). Unlike 2 Peter S Piispanen 2 days ago Ah, interesting! So, attempting to synthesize something from the suggestions thus far, we could have: *wujǝ̑j or *wulǝ̑j or *wuwǝ̑j > *wuǝ̑j > *wuaj 'island'. Now, a possibility that comes to mind is that this word for 'island' is an oblique Pre-Proto-Samoyedic compound. Something like: *vij- 'to stream' and *jåǝ̑ 'land, place' - thus *vij-jåǝ̑- - with a literal interpretation as 'river land', which is exactly what an 'island' is. This phonologically complex mix of semivowels and vowels could perhaps have turned into *wujǝ̑-j. This root *vij- is from KESK 59 (which I take it describes a Proto-Samoyed root), and appears to find a correspondence in Yukaghir *wojo- 'to stream; stream'. PS *jåǝ̑ 'land, place' also has a derivative *jåǝ̑rå 'sandbank' and in Kamassian the meaning is 'small stones at the bank of a sea or river', which is getting semantically closer to 'peninsula' or even 'small island'. This is reminiscent of the Yenisseian idea of 'river bay = back of an island', so I'm saying that there appears to be some predecent for the possibility that the word that has become PS *wuaj 'island' started out as a compound of traceable elements. A vague possibility with unsolved problems, but it may be worth mentioning at least. Like 1 https://www.academia.edu/s/010b2ad552#comment_809468 7/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter Kitson 2 days ago Peter: “exactly what an ‘island’ is”. Only some islands. In English the default referent for general words for “island” was historically land surrounded by standing water (e.g. in the Fens) or by the sea; river islands were a subset felt by at least some speakers to need specification as “eyot” or “ait” (both variants of a derivative of the most widespread word for island in Old English). Are you assuming that the semantic salience of the types to your Proto-Samoyeds was the other way round, and/or that seas, or at any rate oceanic islands, were not part of their environment? Like Juho Pystynen 2 days ago Proto-Permic surely, Samoyedic is not even reconstructed with *v-. Enets ń- is unlikely to be a graphical error for ŋ-, when Enets does not regularly have word-initial ŋ- in native vocabulary. Most cases seem to be loanwords from Tundra Nenets. In some palatalizing environments there is still a prothetic ń- though (presumably thru earlier *ŋ́, as also in Nenets). This could be the case here too, as something like pre-Enets *üe > *ŋüe > *ńüe > ńue. Needing to route an *ü into Enets would compound further the reconstruction problems however. Unlike 3 Peter S Piispanen 2 days ago @Kitson: interesting distinction between islands in rivers and in the seas - I was not aware of that. As for the Samoyeds, well, I would harbor an educated guess, actually a fact, that the speakers of ProtoSamoyed had rivers (PU *joke 'river') and lakes in their native habitat. One general word for 'sea' (PS *jäm 'sea'), however, appears to be a Turkic borrowing; perhaps this means that the Turkic population meeting (and probably admixing) with them were seafarers (lake boat people?) or originally from a larger sea area probably close by). However, this word means 'river' or 'great river' in some Samoyedic languages. A parallel is found with PU *toxi 'sea' > PS to 'sea', so Uralic people and Samoyeds also knew what a sea were, as dinstinct from a lake or a pond. Selkup has a word, kêĺ~kueĺ 'Flußbusen, Seebusen' which finds cognate forms in Ugric and Permic languages, so that too is of Uralic origin with the meaning 'bay'. So, the Proto-Samoyeds had rivers, lakes, and bays, and boats, and fishing nets (PU *kalV 'net') and other fishing gear (PU *kulta- 'to fish with fishing gear; to scoop in Samoyedic) and islands in rivers, and deep places in water (PU *jurma 'deep (of water)', etc. but perhaps at least originally no larger sea bodies closeby? So, perhaps their islands were located in rivers and lakes, but originally not in the seas? In any event, I would be interested in getting further feedback on the reconstructive issues and the very hypothetical compound idea for 'island'. Like Ante Aikio 21 hrs ago https://www.academia.edu/s/010b2ad552#comment_809468 8/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu I am not aware of any reflex of PU *toxi that would have the meaning 'sea' - what's the motive for this semantic reconstruction? AFAIK, all the reflexes mean 'lake', in some cases specifically 'small lake, pond'. Proto-Permic *vi(j)- has no cognates outside Permic, so it is not at all probable that it would be of ProtoUralic origin. Even if it was, there is a huge number of Uralic proto-forms that it could reflect; at least intervocalic *p, *t, *k, *x, *w, *δ, and *j are possible here. Considering, furthermore, that the front vowel -i- does not match Samoyed -u-, and that Samoyed *jåǝ̑ does not have any Uralic etymology either, I can't see any possibility of analyzing the Samoyed noun for island as a compound word with these elements. Like 1 Peter S Piispanen 11 hrs ago The semantic reconstruction of the UEW (giving this as *towV 'sea', but Sammallahti as *toxi 'see') is based on Mansi tō 'sea'; Hungarian tó 'sea'; Nenets tō 'sea', and all the other Samoyedic forms, so I'm guessing there is predecence for regarding the root meaning as 'sea' already in PU. But correctly you suggest that 'pond' and 'small lake' are more common meanings in the languages bearing cognates of this root. Frankly, doesn't your suggestion that this Proto-Permic root could hardly be of Proto-Uralic origins go against your own methodology in the Studies in Uralic Etymology paper series? Therein you methodically reconstruct reasonable roots in numerous different Uralic language groups, and always present the final result as a Proto-Uralic root. Those words aren't attested in all branches either as far as I can tell. Did I miss something there? True, PP *vi(j)- is most likely a later innovation - I agree with this - but history tells us that numerous PP roots were later discovered to actually be of Proto-Uralic origins (as Samoyed or Ugric cognates were found). In other words, we just don't know until more studies have been carried out. You are absolutely right about the huge phonological hurdles associated with the PP root! Without other external comparanda we are quite in the dark here. Thank you for this analysis of my most tentative compound hypothesis. Like Ante Aikio 11 hrs ago @Peter: I don't know any source giving the meaning as 'sea'; this is a misunderstanding, and the German word "See" in the glosses of these words means 'lake' (= Russian 'озеро'), not 'sea'. Unlike 2 Mikhail Zhivlov 10 hrs ago @Peter: you cite Sammallahti's reconstruction as *toxi 'see' [you mean 'sea'], but Sammallahti actually has *toxɨ 'lake'. Like https://www.academia.edu/s/010b2ad552#comment_809468 9/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Tom Martin 10 hrs ago Yes, my Hungarian-English dictionary translates tó as 'lake', not 'sea'. The Hungarian word for 'sea' is tenger, which is borrowed from a Turkic language. German See is 'lake', when it is a masculine noun, and 'sea', when it is a feminine noun. Unlike 1 Mikhail Zhivlov 10 hrs ago @Peter: Also, Proto-Samoyed *jam 'sea' is not a Turkic loanword. I do not know what Turkic form you have in mind. Anikin & Helimski compare it to Tungusic *lāmu 'sea; lake Baikal'. Unlike 1 Peter S Piispanen 10 hrs ago Ah, yes, it's a Tungusic loanword, not a Turkic one, and a possible one only at that - misremembered there. Well, regarding the 'sea' translation - it is given as such for this word here: https://starling.rinet.ru/cgi-bin/response.cgi? single=1&basename=%2fdata%2fnostr%2fnostret&text_number=++83&root=config It appears to be give 'sea' as a mistranslation from German 'See', which then is incorrect given that the actual attestation is 'lake' and 'pond' as you all mention above (plus German See is lake, as a sea or ocean is probably closer to Meer instead). Could this mean that there was no Proto-Uralic word for 'sea', only 'lake' and 'pond' and 'river'? If so, and that's a big if, that would give us some additional information about the location of the Urheimat, namely that there were no seas close-by. Like Tom Martin 9 hrs ago German has two words for 'sea', namely See, when See is a feminine noun, and Meer. When See is masculine, it means 'lake' Like Peter S Piispanen 9 hrs ago Ah, yes (I don't remember this fact from school...), and because the source I gave doesn't give the root gloss with grammatical gender the mistranslation is quite understandable. Peter signing over and out... see you in another Session! Like Ante Aikio 3 hrs ago https://www.academia.edu/s/010b2ad552#comment_809468 10/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu @Peter Indeed, it seems that no PU words with meanings connected to sea can be reconstructed, so it's not that big of an "if" in my opinion. Moreover, it is well-known that Finnic and Saami words with meanings connected to sea are largely borrowings from Baltic and Germanic. Thanks for hosting this interesting session! Unlike 3 Arnaud Fournet 20 days ago yes, but what about loss of stem-final -s in Samoyedic ? Like 1 Annotation: Page 4 Peter S Piispanen 20 days ago Uralic or Samoyedic roots generally do not end in *-s, and so it is automatically eradicated in most cases of borrowing. You will not find an *-s ending root in the entirety of the SW. An rare example of a ProtoUralic root that actually had a final *-s is *kolmis 'tree bark' (Aikio 2013:168), but this is not found in Samoyedic, which prefers simplified, shortened roots. Another example is PU *tejnis 'pregnant (of animals)', but this appears to have been borrowed as such from Indo-European (Aikio 2014:90), and is not found in Samoyedic either. It is possible that both of these roots were actually borrowed into ProtoUralic. Here the above words are nouns, and I have not seen any verb ending in *-s, so some form of prosodic restriction may have been in action immediately eradicating the final *-s- of the Turkic root with this borrowing into Samoyedic. There are also certain suffixes in the various language branches that do end in -s, but as they were productive in older times they are found added in a number of places rather systemtically only as far as I can tell. Examples: Finnish -os~-ös, a resultative marker or -kas~-käs forming adjectives from nouns. Like 1 Juho Pystynen 19 days ago Loss of absolute final *s might be possible (not that there's any evidence for final *ś in Proto-Uralic either!), but general loss of merely stem-final *s from a verb seems less motivated. Maybe this could have been lost first in the preterite stem: *CVs-sA- > *CV-sA-? Unlike 1 Robert Lindsay 18 days ago https://www.academia.edu/s/010b2ad552#comment_809468 11/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu I really dislike these Proto-Uralic and Proto-IE borrowing scenarios. I've read that there are layers and layers of these borrowings! Right, except they were never next to each other in space-time once again. PIE is in Anatolia from 9,000 YBP until 5,500 YBP when it's between the Caspian and Black Seas. Both are very far from the PU homeland in Western Siberia. Granted, Proto-Indo-Iranian is around Samara 4,500 YBP and is wellpositioned to borrow with PU in the PU homeland as they are time-space adjacent. I would reject all PIE-PU "loans" as irrational because they were never close enough to trade words. I hate to bring up Kortlandt's Indo-Uralic, but that's exactly where all of these "loans" may end up deriving from. Like 1 Juho Pystynen 18 days ago These days we tend to call words like *tejnəš "Proto-Uralic" as long as they can be reconstructed to a common proto-form, though it is clear that this for example is a loanword that we do not need to assume to have ever existed in branches where it's actually found, viz. Finnic and Mari. (A few decades ago this still would have been called "Proto-Finno-Volgaic" instead.) It's also definitely not from PIE, probably instead from early Balto-Slavic or Indo-Iranian. Like 2 Arnaud Fournet 18 days ago Well, *tejnəš does look like an IEan word, from *dheH1- "to breast-feed". Like Eugen Hill 17 days ago PFi *tīneh reconstructed by Aikio and its Mari counterpart indeed look somewhat close to what is reflected as Lith dienì in Baltic and Skt dhenú- in IIr. However, PFi *tīneh can’t be borrowed from ProtoBaltic or Proto-Balto-Slavonic predecessor of Lith dienì, because this predecessor must have been (roughly) nom.sg. *deinī or *dainī (end-stressed, with a surprising circumflex intonation of the first syllable). I’m not sure if IIr is of much help here because Skt dhenú-, Avestan daēnu- presuppose ProtoIIr nom.sg. *dhainúš whose *ai and *u don’t match PFi *ī and *e. Or am I mistaken here? Like Mikhail Zhivlov 17 days ago Aikio reconstructs "Proto-Uralic" [in fact, post-PU] *tejniš, where at least second syllable *i is a known substitution of Indo-Iranian *u, cf. post-PU *meti 'honey' from *medhu-. Such a substitution is expected, since there was no *u in non-first syllables in Proto-Uralic. Finnic second syllable *u goes back to *iw. Like 2 https://www.academia.edu/s/010b2ad552#comment_809468 12/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Jaakko Häkkinen 17 days ago Robert Lindsey, it is true that Proto-Uralic and Proto-Indo-European are no more contemporaneous. Still, the labels of loanword layers - even the "wrong" labels - cannot make the loanwords false. We just need a more accurate label, like Archaic Indo-European. It is widely seen that in the northwest, the IE language remained archaic up to 2nd millennium BC. So, this AIE was contemporaneous with ProtoUralic and still close to PIE reconstruction stage. On your views about the IE datings: here you seem to have some very own hypothesis, diverging from the best-argued view. Linguistically it is difficult to defend PIE/PIH in Anatolia at 7000 BC. Unlike 2 Robert Lindsay 17 days ago Isn't an Anatolian homeland for PIE at 8-9,000 YBP the latest from glottochronology and particularly Renfrew's view? I thought that was their dating. That glottochronological dating and location was the latest study done on the efforts. Let me look up the paper. Also I think Nostraticists think that all these PIE-PU loans are genetic. Like Jaakko Häkkinen 16 days ago Robert Lindsay, there are tons of counter-arguments against both Renfrew's view (based on the erroneous method of seeing the linguistic continuity from the archaeological continuity) and glottochronology. http://www.elisanet.fi/alkupera/Problems_of_phylogenetics.pdf The problem considering the shared PIE-PU words inherited from the alleged common protolanguage is, that the languages including their phonological systems would have developed into very different directions, yet their inherited w... Read More Like 1 Arnaud Fournet 16 days ago I must say that I do not share Robert's tenderness for glottochronology, which I believe is thoroughly unreliable, and also to a large extent is a head-reducing and brain-shrinking idiotic method. That being said, your claim that "there are tons of counter-arguments against Renfrew's view" is entirely false. The Anatolian origin of PIE is quite certainly correct. Like 1 Robert Lindsay 16 days ago https://www.academia.edu/s/010b2ad552#comment_809468 13/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Arnaud, if we can't use glottochronology to date protolanguages, what can we use? Lexicostatistics? Jakko doesn't like computational phylogenetics either. Great. So how on Earth can we date any protolanguages at all then if all of our tools are faulty? Like Arnaud Fournet 16 days ago I'm afraid we (I mean as linguists and paleolexilogists) have no reliable way of providing absolute datings. About everything we can say depends on external information, most of which is archeological. Let's take an example: the word "apple", which is more or less diversely reconstructed as pseudo-PIE *ab(H)l-. This word has a variant in Latin and Greek as *ma:l-. If we combine both words, we may hypothesize that a preform of the type *(a)-m(a:)l- (either *aml- or *ma:l-) intruded in a number of Indo-European languages. I assume that *a-ml- regularly was rearranged as *abl-. The morphology of *(a)-m(a:)l- is un-Indo-European. Now, when did that intrusion happen? Absolutely no idea on linguistic grounds. Now, wikipedia informs that wild apples are found in Kazakhstan and the fruit wandered west and reached Italy about 6000 years ago. Next, Turkic for "apple" is *alma, which is a near hit for *(a)-m(a:)l-. Interestingly, dialectal Chuvash has omla with the phonemes in the order of pseudo-PIE *(a)-m(a:)l-. So it appears that pseudo-PIE *(a)-m(a:)l- is probably related to Turkic *alma (< more archaic *amla, miraculously preserved in dialectal Chuvash omla). Nothing in this can be dated on linguistic grounds. The same is true with arrows and bows, the appearance of which linguistics is clueless to date. Unlike 3 Arnaud Fournet 16 days ago Quote form wikipedia: "At the Sammardenchia-Cueis site near Udine in Northeastern Italy, seeds from some form of apples have been found in material carbon dated to around 4000 BCE." => incidentally, this proves that PIE no longer existed at that dating, otherwise Latin and Greek should not have a variant word. Like Rasmus G Bjørn 16 days ago Renfrew himself has conceded the veracity of Gimbutas' Steppe hypothesis, primarily due to genetic evidence. Only Anatolian remains difficult, as you say for want of genetic markers, and I'm happy to entertain any possibility. But Afanasievo is an exact Yamnaya match, so we're pretty far in terms of aligning the aDNA and linguistic evidence on the Steppes. I have recently submitted a manuscript to HS, basically suggesting that the "Semitoid" traits in PIE are due to the Balkan Neolithic language (expanding https://www.academia.edu/s/010b2ad552#comment_809468 14/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu on the numeral spread hypothesis, https://brill.com/view/book/edcoll/9789004416192/BP000003.xml?rskey=lm7OxZ&result=1). In my humble opinion it answers more questions than it raises, but there are certainly still rough edges. The *only* thing I disagree with Renfrew's Anatolian hypothesis on is the language that spread. Let me also add that Russell Gray and other Bayesiophiles wish to see a Transcaucasian homeland, but cannot seem to pass peer-review. They further posit Indic and Iranian as the earliest and separate offshoots from south of the Caspian, based on the idea that they "only" share some 57 % of (central?) vocabulary. Suffice it to say that I am less than convinced, but there's a hypothesis for anything if you're interested. Apples, I'm very interested, but I hardly think that apples (what kind?) at 4000 BC in Northern Italy has much bearing on the Steppe Hypothesis. Lastly, and most to the point of the current debate, is the question of Uralic and Samoyedic homelands. I agree with many statements made, and would like to direct attention to Peyrot's and Abel Warries' recent demonstration of significant Samoyedic substrate in Proto-Tocharian. Squared with the Indo-Iranian substrate in Fenno-Ugric (however you slice it), I think the original bifurcation of Uralic has to be found east. I have also abandoned the archaeologically motivated "neighboring homelands" of PU and PIE on the Volga, especially due to the lack of culturally relevant borrowings for the envisaged time frame. Rather, I stand by my analysis in (https://brill.com/view/book/edcoll/9789004409354/BP000003.xml?rskey=vo8yrr&result=2) of old affinity, but how old is obviously debatable. I can agree that some forms are suspiciously similar, but I try to show that it cannot deter us from making an honest comparison. At any rate, proof of relationship exactly lies in words that appear to be different on the surface. Unlike 4 Arnaud Fournet 16 days ago @ Rasmus, 1. Genomic evidence disproves the Pontic-Caspian archeofable. No steppic input in Anatolian speakers. 2. What is HS ? "hors service" = out of order ? 3. Peer-review is not at all a guarantee of quality... and this is an understatement... Like Peter Kitson 15 days ago Rasmus, “Renfrew himself has conceded the veracity of Gimbutas’ Steppe hypothesis.” Really? Where do you claim that such concession was made? I have read several (not all) of Renfrew’s later papers, and the closest he gets in them is positing a sort of prehistoric Balkan Sprachbund, in which however the north European farmers (not the steppe people nor the Balkan townsfolk) still had the major part. Something similar is postulated from the steppe side by Asko Parpola, but as far as my reading goes he seems to have no followers even among his fellow-proponents of the steppe hypothesis. Can the Finns present tell me if he has garnered any in Finland? Like 1 https://www.academia.edu/s/010b2ad552#comment_809468 15/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter Kitson 12 days ago Specifically, Renfrew’s piece in Celtic from the West 2, which must be one of his last (2013), still stresses the deficiencies of the ‘Kurgan’ hypothesis and reaffirms the basic soundness of the Anatolian one. Like 1 Arnaud Fournet 11 days ago @ Peter Thank you for this. I also heard that Renfrew had allegedly given up his Anatolian Neolithic scenario. I'm glad to see that this is not true. Anyway, the Pontico-Caspian archeofable crashes against a number of basic facts: 1. archeologically, it does not work and cannot account for all the western half of IEan languages, Alexander Häusler: Nomaden, Indogermanen, Invasion. Zur Entstehung eines Mythos tears apart the myth of steppic influences on the western part of Europe, 2. genetically, there is no trace of steppic genome in Anatolian speakers, 3. what is the substrate of Luvian, if that language came from somewhere else? I've never read about a substrate in Luvian. 4. the arguments proferred by Anthony-Ringe are either false, or lies or just plainly idiotic. Like Rasmus G Bjørn 11 days ago @Peter Kitson, thank you for you question. While I may have used a bit of hyperbole, he does call Marija Gimbutas "triumphant" in this recent video: https://www.youtube.com/watch?v=pmv3J55bdZc , where you conveniently can move the timer to 10:00. I don't have time to watch it all now to see his exact position as of three years ago, but it is no understatement that this speech is an admirable testimony of academic discourse. I would also like to stress the silliness of ad-hominem or conspiratorial attacks on any homeland hypothesis. I don't think anyone goes to the table with malignant intent, and I do, indeed, in my own work, draw on the strengths of the explanatory frameworks of proponents of other homelands, especially Gamkrelidze & Ivanov and Renfrew. I have also looked over the evidence for a purported connection between Indo-European and Hurro-Urartian as presented by Fournet & Bomhard, and tried to include it in my attempts at a survey of potential immediate extra-familiar relations of PIE. Unlike 2 Rasmus G Bjørn 11 days ago @Peter Piispanen, sorry for the tangent. I am a big fan of your work and hope to be able to contribute more directly in the future :) Like Peter S Piispanen 11 days ago https://www.academia.edu/s/010b2ad552#comment_809468 16/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu @Rasmus: not a problem - these tangential discussions are interesting, and thanks for the vote of confidence! :-) Indeed, I wish that there were a forum here at Academia where various discussions could continue. Meanwhile, I am considering setting up a blog page somewhere (like where Juho Pystynen, for example, does) to discuss some ideas with interested parties. First and foremost, however, I suppose I should work primarily on attempting to publish 10-15 draft papers or so for the sake of the scientific community. Like 2 Peter Kitson 11 days ago Arnaud, To your points I would add 5. The minimum claim to be made from onomastics is that alteuropäisch and Rig-Vedic river-names are exponents of the same underlying place-naming system. That fits seamlessly with the Anatolian/north European hypothesis but cannot readily be squared with any of the usual versions of the steppe hypothesis. Parpola’s scenario logically allows the language (or language group) of any of north Europe, the steppe, or the Balkans to be the matrix language of IndoEuropean, with the other two as sub-, ad-, and/or superstrates. 6. The distinguishing aspects of that system would arise easily from normal linguistic causes at an identifiable stage in the generally accepted prehistory of Indo-European, and not as far as I know in that of any other language group. 7. The dialect geography of the existing Indo-European branches (including Indo-Iranian, but perhaps excluding Anatolian) took shape in Europe. This is more economically explained on the north European hypothesis than on the other two. Kortlandt’s version in ‘The expansion of the IEur languages’ (2018) is triply impossible. Unparsimonious consequences of the steppe model include reification of ‘Temematic’ (Kortlandt 2010.79 after Holzer), whose combination of features would occur as part of an ordinary dialect continuum in Europe. Aspects of Germanic which the Leydeners either invent substrates for or don’t notice also arise naturally in the model of an already existing dialect geography disrupted by effects of steppe incursions. 8. My preferred model has (up to a point, and mutatis mutandis) steppe people as an analogue to Vikings, Corded Ware as an analogue to Normandy, and Beaker Folk as an analogue to Normans. To turn devil’s advocate for a minute, a weakness of your (2) is that very little relevant archaeogenetic work has yet been done in Anatolia. In ten years’ time the picture may well be a lot more complicated. (Think how much more complicated the picture in Europe has become in the last ten years—not to mention how fatally for theories like Vennemann’s.) But I agree that Damgaard et al. provide what on the face of it is strong evidence for either Renfrew’s Indo-European or ‘Indo-Hittite’, and against the steppe hypothesis, and it’s a disgrace that they presented it on the explicit assumption that the steppe hypothesis is true. (Rather like Friedrich’s trees, which though presented on the same explicit assumption have, taken as a set, a north European centre of gravity, and don’t at all fit an area inhabited by Indo-Iranians millennia more recently than other Indo-Europeans.) As for (3), not only has Woudhuizen argued for an alteuropäisch-type substrate in Luwian, on the basis of about a dozen names all of which are individually debatable but which may collectively for all I know add up to something, Villar et al. (2011) starting from different premisses assert that there is such a substrate in all the IEur languages (and some other languages) of Europe and west Asia. https://www.academia.edu/s/010b2ad552#comment_809468 17/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Re (1), I’ve read some of Häusler’s things (he sent me some himself), but hadn’t come across the one you mention. Thank you for the reference. Like Peter Kitson 11 days ago Rasmus Bjørn, thank you for the reference, which I shall follow up if I can find it, and let you know what I think. Do you really perceive no malignant intent in Kortlandt’s (2018 as above) equating disagreement with him about how linguistic evidence should be interpreted as “denial of the linguistic evidence”? Or in the repetitiveness of Bichler’s attacks on Udolph for consistency in a methodology different from his own? I think it naïve not to recognize that there is a good deal of it in the Urheimat debate historically, and that there is a significant element of groupthink in current orthodoxies. How else do you explain Damgaard et al. (2018) not declaring openly that what they have come up with is a piece of archaeological evidence, arguably the first significant one for a generation, favouring Renfrew’s hypothesis and/or ‘Indo-Hittite’ and strongly disfavouring any existing version of the steppe hypothesis? Again, a variant of Parpola’s could be constructed to fit it (but one that K. Kristiansen, whose archaeological views Kortlandt prefers to Gray’s linguistic ones, would reprobate). Peter, this forum works because the people on it by and large are able and willing to recognize the limitations of their own expertise and the extent of other people’s. An open-access blog would attract a lot of cranks and participants both ignorant and unwilling to learn how to learn. So do some sessions even on Academia, e.g. currently on Peter Revesz’s (rather shallow) computeristic approach to the Indus Valley script. Like 1 Arnaud Fournet 11 days ago The old draft Fournet-Bomhard on Hurrian is now quite obsolete. People interested in the issue of the links between Hurrian and PIE should read: 1. https://www.thebookedition.com/fr/pie-and-hurrian-ii-ten-years-later-p-375829.html The book more or less reconstructs the ancestor of PIE and Hurrian, which I call Proto-Anatolic and explains how PIE came to become PIE. 2. https://www.thebookedition.com/fr/pie-roots-attested-in-hurrian-p-377092.html Lists Hurrian words with clear comparanda in IEan l... Read More Like 2 Arnaud Fournet 11 days ago I will add that Pre-Indo-European Europe from the Atlantic to the Caucasus seems to have spoken Basquo-Caucasic languages, so PIE cannot originate in the middle of an area where a continuum of related Basquo-Caucasic languages is spoken. I've begun to build a word-list that involves Basque, Caucasic and substratic words in IEan languages. My next billet d'humeur will be on the words shared by Albanian and Romanian, and quite obviously, https://www.academia.edu/s/010b2ad552#comment_809468 18/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu some of the words have comparanda either in Basque or Caucasic. I hope I can live long enough to write the Pokorny of Basquo-Caucasic. Like Peter Kitson 11 days ago Bichler] Bichlmeier — sorry. Like Peter S Piispanen 11 days ago @Peter: yes, I was afraid of such risks on public linguistic forums. Perhaps I should stick to Academia Sessions: there is no lack of materials to discuss - my own main problem is to actually complete a project and publish the materials reasonably quickly before being swayed away by new ideas altogether. Many of these sessions have easily become my most rewarding linguistics discussion ever (even beating courses, conferences, etc.) in terms of historical detail, ethnolinguistic situation, available documentation, richness of suggestions, pointing out of previously done research and useful references, crossdisciplinary information, function as a first peer review step, etc. and not to mention the useful contacts and connections that are naturally established here as well. The limitations you mention are actually richness of focus, dedication to specific fields, and, in this format, a willingness to discuss, share, improve and argue around new findings and suggestions. So, yes, I guess we can (all?) agree that these sessions can be extremely fruitful endeavors! Like 4 Peter Kitson 11 days ago Peter, Yes. Like Peter Kitson 11 days ago Arnaud, I do not believe in your Basque-Caucasian Europe because what the geneticists seem to have shown (on considerably more evidence, I seem to remember, than the ones in Anatolia) is that the ancestral speakers of Basque were not autochthonous since the repopulation of Europe after the Ice Age but were (like ancestors of Indo-Europeans according to some of us) a population intrusive with the spread of agriculture, namely users of Impressed Ware alias Cardial pottery (and that the immediately pre-Indo-European inhabitants of Britain were an offshoot of that population, not anything that might be Semitic-speaking as Vennemann would like). I can’t lay my hands on the reference at the moment, but if you want me to I can probably manage it in a day or two. The general probability in my opinion is that the language of mesolithic European hunter-gatherers was not closely related to anything spoken nowadays, and arguments from suppositions about it are pretty well guaranteed to be unsound. Like 1 https://www.academia.edu/s/010b2ad552#comment_809468 19/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Arnaud Fournet 11 days ago @ Peter Kilson => never mind, so we disagree in the worst possible manner !! Like Rasmus G Bjørn 9 days ago I'll be happy to revisit the hypothesis when time permits. Kortlandt is certainly an eccentric in his own right, but we simply must refrain from seeing conspiracies. I can be perfectly open about my own biases, coming from the Copenhagen school and basically assuming the Steppe hypothesis from received wisdom, but our work as researchers must strive to test these assumptions; a major problem with the Steppe homeland, vis-a-vis the Caucasus/Anatolia, was the Semitoid traits, and in particular the numeral seven. Having now spent some time looking into, first, the numeral spread hypothesis, and, more recently, Anthony's flimsy suggestion of an Afro-Asiatic language in the ancient Balkans, the viability of the Steppe hypothesis (at least for post-Anatolian, but probably the lot), actually provides answers to questions not even addressed in the original question, e.g. the presence of both "Berber" and "Egyptian" substrate words in the European IE languages. I HOPE to attract a good discussion, and hopefully also dissenting opinions to move the field forward, but there are multiple different academic outlets available where proper, and sometimes scathing, peer-review can be met; fortunately the roster is large enough to go elsewhere if the ideas are robust enough. Renfrew does concede the massive migration from the Steppe, which does include Afanasievo, so most likely the Tocharians, so what we have left to discuss in terms of homeland is Anatolian. Arnaud has pointed to some linguistic indications (lack of Luwian substrate, I believe), and others may point to the elite status of Hittite, as other languages appear to have been spoken as the more common vernaculars of Eastern Anatolia. So there is room for discussion Damgaard et al. 2018 doesn't prove anything linguistically. It adds colour to Eurasian pre-history against which we may try our varying hypotheses. The absence of Yamnaya DNA in Hittite graves is a problem for the original Steppe origins, but not more so than the situation of present day Hungary where elite dominance induced a language shift without change in local DNA. Their dates for South Asia are also perfectly in line with the Steppe origin of the Indo-Iranian languages -- in my humble opinion at least. But let's defer these points to another venue, hopefully, in due time, at a conference :) Like Arnaud Fournet 9 days ago The situation is more complex than what you seem to describe. After a first wave of Neolithic Anatolian intrusion, a second phase occurred when pre-existing Mesolithic Hunter-Gatherers converted to Neolithic, so that Neolithic is no longer recognizable genetically. Like Arnaud Fournet 8 days ago https://www.academia.edu/s/010b2ad552#comment_809468 20/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu @ Rasmus, I'm curious to see what these supposedly Berber and Egyptian words in PIE look like. Word "seven" belongs to what I would call "miroir aux alouettes" (skylark-catching mirror trap). I don't believe Post-Anatolian *sept-m has anything to do with Semitic *sabˁ- which would be PIE *sebhH3-. My opinion about *sept- is that it's a borrowing from a kind of para-Kartvelian where *kw > p. Kartvelian can be reconstructed as *skwid "seven", as you can see: *skwid is *spid if you accept *kw &gt; p, as in the pair *akwa / *apa. So we don't need Semitic, because neighboring Kartvelian will do the job better than Semitic, which is an impossible "solution", both in terms of chronology and geography. Like Ante Aikio 8 days ago @Peter: Just as a side note regarding the word *kolmis 'tree bark' which you mentioned way above in this thread: I now notice that this reconstruction should actually be corrected to *kolmiš instead. The reason for this is the Malmyzh dialect form /kumuž/ in Mari; this dialect has preserved the opposition between PU *s on the one hand and PU *š/*ś on the other; the form *kolmis would be expected to yield Malmyzh Mari */kumuz/ instead of the actually attested /kumuž/. In the rest of Mari varieties, and in Saami, the reflexes of PU *s and *š have merged and the two phonemes cannot be distinguished. Unlike 3 Peter S Piispanen 11 hrs ago @Ante: this is an interesting root, *kolmiš 'tree bark'. The Malmyzh Mari diagnostic is fascinating, and I think that is correct too. There is nothing like it in any other Siberian language AFAICT, but it doesn't really look or feel Uralic to me. What do you think is the etymological origin of this root? Are there any other words having this distinction in Mari dialects, and which are etymologically from Proto-Uralic? Could it be that this is phonotactically, prosodically and morphologically Proto-Uralic, and that it actually describes another hitherto unknown and valid way of reconstructing Proto-Uralic roots (i.e. ending in *-š). I can't think of any other root ending in *-š, do they exist? Regardless, the evidence, as you have presented, does point to this reconstruction as being correct IMO. Like Ante Aikio 11 hrs ago *-iš is a common noun suffix in Finnic and Saami, and probably has a cognate in Mari as well, so *kolmiš could have originated as a derivative of a stem *kolmV-. Unlike 2 Peter S Piispanen 10 hrs ago https://www.academia.edu/s/010b2ad552#comment_809468 21/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Excellent point - I believe this is entirely correct: a derivative it is! Like Tom Martin 10 hrs ago I have seen now your updated version of your paper. I have the following comments: You have Proto-Turkic 'winter' ending in *l, this cannot be correct, see for example Uzbek qish, etc. So the final consonant in ProtoTurkic was either *š or perhaps some kind of *l but not normal *l. Proto-Turkic and Common Turkic did not have initial *g, so in one of your words for 'navel, the initial *g is not correct, it must have been *k. And the same is true of the Common Turkic word for 'come'. Likewise Common Turkic did not have initial *d, so the initial *d in your compound word for 'navel', the *d in 'four', the *d in "branch, willow' is not correct, it is initial *t. as it is still preserved in some Turkic languages. Initial *t changed in some Turkic languages to d, but not always,there are exceptions, like for example tan 'dawn', tepe 'hill' in Turkish, where Azerbaijani and Turkmen have d, though in many proto-Turkic words Turkish changed initial *t to d, like in'four', 'branch'. The Karagas language is not a Samoyedic language, it is a Turkic language. So the Karagas word for 'raven' is probably not borrowed. Like Peter S Piispanen 1 day ago Good people, Thank you for an enormously productive and fun draft paper session! The entire discussion will be summarized in a pdf-file for those interested, and for posteriority, and I will employ the materials to improve on the argumentation and details of the updated paper before eventual submission. Other draft paper sessions are in the works of course as meager time permits, either dealing with the East Siberian Languages, Finno-Permic, Elamite or the South American indigenous languages ... all good things to those who wait, I hope! Best regards, / Peter :-) Like Peter Kitson 22 hrs ago And congratulations to you as ringmaster. Like Peter S Piispanen 12 hrs ago Thank you, my namesake, thank you. I am honored to have gathered such a great group of experts and other interested parties, and also humbled by the attention that my work has received. So, what else https://www.academia.edu/s/010b2ad552#comment_809468 22/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu remains to say except: see you in another Session! :-) Like B. Blasebalg 24 hrs ago How do you motivate the semantic change from 'kettle' to 'spoon'? Like 2 Annotation: Page 5 Peter S Piispanen 13 hrs ago I do not motivate it - it is not my suggestion, but my comment therein should have read that the comparison is phonologically flawless, although not always semantically. As Aikio suggested, one of the forms cannot be an early borrowing into Ugric because of the phonology in Hungarian (and the word is missing in Khanty and Mansi), and in this case I am also a bit perplexed - I haven't seen this suggested semantic change before elsewhere. Sure, both are containers of liquid, one larger and one smaller, and the phonology is a perfect math, but suggestion too is not as secure or promising as the other suggestions (three better ones are still remaining). Like Alvah M Hicks 2 days ago 04-21-2021 From my Papers at Academia.edu Amerindian mtDNAs and Admixture in Siberian Populations: Examining Alternatives to Traditional Models of Ancient Human Migrations Abstract This is a "positional paper" according an Amerindian contribution to Athapaskan, Eskimo/Aleut and, as well, Siberians Population formation since these groups share "distinct genetic affinities with Native Americans (Torroni et al. 1993b, pg. 591)." Presumably, as Emoke Szathmary identified in the Plenary Session for Chacmool 1998, Boas believed the removal of glacial barriers precipitated human contact between the Longitudinal Hemispheres that was, until-this-time, geographically encumbered. Boas further believed that Amerindians migrated into Siberia with widespread Holocene acculturation (also see Ackerman 1982; Dumond 1983; and Heizer 1943; and others), since earlier American Indian Tribal Populations were present before deglaciation. A second migration out of the Americas provides a backdrop for the later formation of Sea Mammal Hunting Cultures, an idea Franz Boas identified as "Eskimo wedge theory" (Boas 1905 and 1910; also see Steven Ousley and others in, Human Biology, June 1995). Does the implied presence in Northeast Asia of founding or nodal mitochondrial DNA (mtDNA) haplotypes for each Amerindian haplogroup (Torroni et al. 1993a) preclude https://www.academia.edu/s/010b2ad552#comment_809468 23/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu evidence of Holocene admixture between Northeast Asians and Amerindians? “Reverse migration” provides an alternative explanation to mtDNA analyses, challenging the idea that Amerindian mtDNA Lineages (AAM1, CAM43, and DAM88) found in Siberians are ancestrally linked to the initial colonizers of the Americas. Rather, mtDNA analysis could be seen to support the Boas data in that the formation of contemporary Circumarctic Populations in Siberia may have been influenced by post-glacial Amerindian movements into Beringia, Siberia, and Northeast Asia. Archaeologically based chronologies may imply that “back migration”, evidenced by the presence of common Amerindian mtDNAs in Siberians, links Circum-arctic Populations to the Americas. Like Peter S Piispanen 11 days ago Because we here have a rare collection of linguists and other experts, allow me to briefly mention another little tangential discovery I made recently. I am at the moment attempting to summarize what appears to be Koreanic borrowings into Manchu (there are some ten entries therein). The main problem for me is reading Korean, but it will work out in the end I'm sure. In that line of research, something interesting popped up, namely the possibility of para-Mongolic borrowings in Manchu as per the following: Another Para-Mongolic borrowing in Manchu? Given that Manchu clearly had borrowed lexicon from neighboring languages, we might want to discuss one new suggestion possibly related to Para-Mongolic. Indeed, Vovin (2007b) suggest that Manchu gida 'lance, spear' was borrowed from a Para-Mongolic source derived from Pre-Proto-Mongolic *ghida, which was later found as Proto-Mongolic *ǯida. Here, I note, we might want to add Jurchen gida ‘spear’ (Kane, D. 1989:250), suggesting that this is an old borrowing. A phonological parallel was presented with Pre-Proto-Mongolic *ghiamcin 'official at a post station', later found as Proto-Mongolic *ǯamucin 'post office people', but borrowed from a Para-Mongolic source as Manchu giyamun 'relay post station'. Indeed, these suggestions sound reasonable to me to explain the unexpected, word-initial Manchu g- of these borrowings. We thus appear to have: Pre-Proto-Mongolic *ghida > (Para-Mongolic > Manchu gida +) Proto-Mongolic *ǯida and Pre-ProtoMongolic *ghiamcin > (Para-Mongolic > Manchu giyamun +) Proto-Mongolic *ǯamucin. Now, in the same vein, I am considering if Manchu gabtan 'archery', and numerous other derivatives, could be borrowed from a Para-Mongolic source derived from a tentative Pre-Proto-Mongolic *ghabtan 'arrow', which would later be borrowed from Proto-Mongolic as Tungusic Ulcha ǯabdụ(n) 'arrow' and Nanai ǯabdu 'arrow'. The Proto-Mongolic form, however, is *ǯebe 'end of an arrow', but I am unconvinced that this is a fully correct reconstruction (see below; *ǯebten is possible). I am assuming Pre-Proto-Mongolic *gh- > Proto-Mongolic *ǯas per the above. Thus, I am further assuming the change of Pre-Proto-Mongolic *gha- (borrowed into Manchu as ga-) > Proto-Mongolic *ǯa-, similar to what happens with *ghi > *ǯi-, etc in Vovin’s suggestion. Semantically, 'arrow' is in the same category as 'lance, spear', i.e. weaponry. Unlike the Manchu form (derived from the Para-Mongolic lineage having *gha-), the Nanai and Ulcha forms look like they were instead borrowed from *ǯa- from the main Proto-Mongolic lineage. Thus, Pre-Proto-Mongolic *ghabtan > (ParaMongolic > Manchu gabtan +) Proto-Mongolic *ǯebten > Ulcha ǯabdụ(n) & Nanai ǯabdu. It is curious that the https://www.academia.edu/s/010b2ad552#comment_809468 24/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu first syllable vowel is -a- in all of the attested forms, but *-e- in the Proto-Mongolic reconstruction: perhaps it should really be *ǯabten instead. Para-Mongolic is an interesting entity. I have recently also found what appears to be direct Mongolic borrowings in Yukaghir - forms that are wholly missing in the usual transmitting languages (Ewen, Ewenki and Yakut). How did they end up in Yukaghir? Related to this question is: could they be from a Para-Mongolic source instead of a main-sequenced Proto-Mongolic source? Any comments or suggestions regarding these matters? Like 1 Geoffrey Caveney 10 days ago Whatever the result of these investigations, I insist that you please do incorporate the phrase "Pre-ProtoMongolic post office people" into your paper, or if possible into the title. It will make for an interesting alternative to the usual semantic fare of proto-form glosses! Like 1 Peter S Piispanen 3 days ago I have now been reading up on the subject properly with works, for example, by Ki-Moon Lee, A Comparative Study Manchu and Korean, Ki-Moon Lee & S. Robert Ramsey, A History of the Korean Language (2011), and will move on to G.J. Ramstedt, A Korean Grammar (1997) and others, and G.J. Ramstedt, Studies in Korean Etymology (1949) if I can find it; already, Poppe's review of it (1950) is illuminating. Frankly, I must say that I am astounded - I never realized that there were so many clear correspondences to be found between Korean and Tungusic, and then, in particular, between Korean and Manchu! I do now much better understand the whole view about the Macro-Altaic languages. Perhaps it should come as no great surprise, still, because they are neighbors after all. In any case, there could really be something here - these details cannot just be dismissed out of hand, but at least for now I will base my hypotheses on the Manchu words that are identical or near-identical in both phonology and semantics with Korean words to be borrowings. Clearly, one can discern regular phonological traits between these two, and these are absolutely not accidental look-alikes. In the above works a few hundred correspondences are listed, with some sixty or so very secure. Among those, there is a subgroup of correspondences between Korean and Manchu specifically, and these I intend to build upon further with new findings, that is loanword suggestions apparently not suggested before. Ki-Moon had already noted seven of my noted correspondences. I am sure a few more are given in the EDAL as cognates. Some new suggestions should remain, and then additionally two tentative ones of expected regular forms albeit with a phonological peculiarity. So, I suppose those notes should be finalized into something readable and discussable, perhaps as an Academia draft paper. Also, in parallel, there are the materials about b. Mongolic borrowings in Yukaghir https://www.academia.edu/s/010b2ad552#comment_809468 25/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu (apparently hitherto entirely unknown materials), as well as c. additional Indo-Iranian borrowings in Finno-Permic (our Finnish colleagues have made some heavy strides in that department in recent years!), and d. Elamite borrowings in Indo-Iranian (which oddly enough, according to Lubotsky and others, doesn't appear to be a as of yet fully researched field). Like Arnaud Fournet 3 days ago You know, there are plenty of things that remain to be done, so I'm not surprised by your to-do list. Unlike 1 ian stiver 6 days ago Hi Peter, Is it possible to summarize all the discussion and comments in one single pdf/doc document ? All the stuffs out here are really Interesting... Thanks Like Annotation: Page 1 Peter S Piispanen 5 days ago Hi Ian! Absolutely - I intend to finish this Session by making a pdf-document of the entire discussion, which is then uploaded onto my page under Sessions for your perusal and download :-) Like 1 ian stiver 5 days ago Thank you ! Like Csaba Barnabas Horvath 7 days ago I am not a linguist, but in my research on identifying prehistoric migration patterns based on archeogenetics, ( https://www.academia.edu/44981646/_How_Eurasia_Was_Born_HOW_EURASIA_WAS_BORN_A_Provi sional_Atlas_of_prehistoric_Eurasia_based_on_genetic_data_supporting_the_farming_language_dispersal _model_CSABA_BARNAB%C3%81S_HORV%C3%81TH )I got to the conclusion that bbefore the spread of Turkic languages, samoyeds populated even the steppes of the western half of Mongolia as steppe horse nomads. In this case I would suggest that it would be justified to reconsider the possibility whether if some of the common words in Turkic and Samoyedic are in fact Samoyedic words borrowed by Turkic, and not the other way around. Unlike 1 https://www.academia.edu/s/010b2ad552#comment_809468 26/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Kristóf István Szegedi 6 days ago With all respects, some of your conclusions are published in potential predatory journals, like "European Scientific Journal" or "Asia Pacific Journal of Advanced Business and Social Studies" (see: https://predatoryjournals.com/journals/) or political science journals, like International Relations Quarterly. So, their peer reveiw - in archaeology and prehistorical sciences - is at least debetable. As a reminder concerning predatory journals: "Ezeknek a lapoknak tudományos értékük nincs, szakmai bírálatot nem végeznek" see: https://lib.semmelweis.hu/oa_parazita_lapok. By the way: your results are based on archaeological and archaeogenetic data and you cannot ask linguists to reconsider the ways of loanword borrowings, because linguistic concepts can be proved by linguistic data, not by archaeological or archaeogenetic. What you say is a complete missunderstanding of methodology. Like 1 Marcel Erdal 7 days ago To Tom Martin: There are NO Chuvash words which find correspondences in Mongolic but not in "Common Turkic"; ALL Chuvash words which have correspondences in Mongolic are also found in other Turkic languages. We just know that they were borrowed from Early Chuvash-Bulgar and not from some other Turkic language because of their specific phonological features. Like 2 Tom Martin 6 days ago Well, Peter Piispanen above wrote two days ago that there are Chuvash words which find correspondences in Mongolic but not in 'Common Turkic'. So that is what I responded above. Like Arnaud Fournet 9 days ago This is a bit tangential to the session, but how many cognates do Chuvash and Mainstream Turkic share? Like Peter S Piispanen 9 days ago Good question, and not tangential at all IMO! If my numbers are correct, the EDAL holds a total of some 2001 Common Turkic roots, and 925 of those have Chuvash cognates. So, that's a considerable number actually, and more Chuvash cognates could quite easily be added here as well. Chuvash lexicon is truly extensively well documented, and knowing the phonological rules from Protohttps://www.academia.edu/s/010b2ad552#comment_809468 27/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Turkic it is possible to find the regular or semi-regular Chuvash cognates. It is my understanding that such cognates can sometimes be quite semantically different from the meanings found in Common Turkic. Then there are Chuvash words also, of course, which have been borrowed from other Turkic languages, and then the group of Chuvash words which find correspondences in Mongolic for whatever reason, but not in Common Turkic. I am certain our local experts could complete this picture to greater satisfaction, and perhaps also agree that not everything is as clear as one would like them to be. For example, I understand that Chuvash vocalism is a field of notoriously difficulty in comparative Turkology. Like 2 Arnaud Fournet 8 days ago That's indeed quite enormous. I did not expect the figure to be so high. What I had in mind is this: - there are about 450-500 roots shared by Hurro-Urartian and PIE, and personally I would date the split at about 10,000-8,000 BCE a bit before Neolithic. Besides, Hurrian is an imperfectly attested language, so the figure is underestimated by our limited knowledge of Hurro-Urartian. - maybe it would be possible to find about 100 roots shared by Basque and PIE, and I think the split belongs to the Paleolithic, maybe something like 30,000-40,000 years ago. - the figure between Chuvash and Common Turkic suggests that the split is not very ancient, a few thousand years, not more. Turkic therefore appears to be quite young. What I find really troublesome is the very low figure between Samoyedic and Finno-Ugric at about 250. If I remember well, FU has about 500-600 good words, which makes it about the same age or slightly younger than PIE. My conclusion would be that PU at 3000-4000 BCE is completely impossible. The split between Samoyedic and Finno-Ugric must be much much older than that. Obviously, PU must be significantly older than PIE. Unlike 1 Juho Pystynen 8 days ago ~900 cognates sounds like about what I would expect, when Proto-Turkic is close to the same age as Proto-Samoyedic where we can currently also reconstruct about this same number. Unlike 2 Juho Pystynen 8 days ago Compared to IE on the other hand, I'd expect the numbers of cognate roots to be naturally lower across families like Uralic and Turkic, where a lot more of the lexicon is made of transparent derivatives. For one this means that if we counted all applicable derivatives, we could reconstruct easily twice as much lexicon for every protolanguage. For two, this often leads to compounding lexical replacements, across the entire "derivational paradigm". https://www.academia.edu/s/010b2ad552#comment_809468 28/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu E.g. *jëxə- 'to drink' is replaced in Mordvinic by *śimə- (via semantic shift; cognate to 'to sip' in Finnic), and this has triggered *jëxə-ma 'drink' (which would be in principle reconstructible from Finnic *jooma ~ Komi juem) also being replaced by the equivalent derivative *śimə-ma. Likewise *jëxə-ja 'drinker' (> ? Finnic *jooja, Sami *jukëjē) has been be replaced by *śiməj; *jëxə-kta- 'to give to drink' (> ? Finnic *jootta-, Mansi *äjt-, Hung. itat) by *śimə-ftə-; etc. So loss of one root has led to the loss of what could be maybe half a dozen lexemes altogether in a less transparently suffixing language. (This effect can be seen even within subgroups, e.g. between Estonian and Finnish or Finnish and Karelian.) Unlike 2 Peter B Golden 8 days ago On Chuvash, see Klara Agyagási. "Chuvash Historical Phonetics" (Wiesbaden: Harrassowitz, 2019). For a thorough discussion of the Old Chuvash (Oghuric/West Old Turkic) preserved as loanwords in Hungarian, see András Róna-Tas and Árpád Berta, "West Old Turkic. Turkic Loanwords in Hungarian" (Wiesbaden: Harrassowitz, 2011) 2 vols. Unlike 2 Alexander Savelyev 8 days ago It is, of course, only the roots with known Altaic parallels that are included in the Altaic dictionary. The actual number of the Chuvash-Common Turkic cognates is substantially higher. The current rough estimate is ca. 1300. Unlike 4 Ante Aikio 8 days ago @Arnaud: The dating of PU and the question of the Samoyed vs. Finno-Ugric split are two quite distinct issues. I suppose every informed specialist agrees that PU has existed, and thus it must have existed at some particular date. But it is far from clear that "Finno-Ugric" (as a distinct intermediste protolanguage and a real sub-branch of Uralic) has ever existed. To securely establish this, one would have to demonstrate innovations (e.g., sound changes) that are shared by all Finno-Ugric languages, but not by Samoyed. But there do not seem to be any such reliably attested innovations, and because of this there is no broad agreement on the internal taxonomy of Uralic at the moment. Among currently active scholars, Juha Janhunen may be the only one who has continued defending the idea of a primary split between Samoyed and Finno-Ugric by presenting actual arguments (rather than by only subscribing to a tradition). Like 1 Tom Martin 8 days ago So Pekka Sammallahti is wrong then, when he proposed five innovations for Proto-Finno-Ugric? Like *o > *u in open syllables before a second syllable *i, PU *noxi > PFU *nuxi 'to pursue', or *VV > *V in a closed syllable, like in the word for 'feather'? https://www.academia.edu/s/010b2ad552#comment_809468 29/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Tom Martin 8 days ago Interesting that there is a group of Chuvash words which find correspondences in Mongolic, but not in Common Turkic. So perhaps the so-called West Turkic branch of Turkic originated to the north of the socalled East Turkic branch. So that could also explain why the borrowings into Mongolic and Tungusic from Turkic share the r and l with the so-called West Turkic, corresponding to *z and *š respectively in the so-called East Turkic branch. So perhaps the consonants of the so-called East Turkic are original, from Proto-Turkic. While the so-called West Turkic experienced mergers of the consonants here, like *z merging with original *r, and *š merging with original *l of Proto-Turkic. Like Ante Aikio 7 days ago @Tom: Regarding the development of the assumed long vowels (*VV), see my paper "On Finnic long vowels, Samoyed vowel sequences and Proto-Uralic *x" (link below). It appears that no long vowels can actually be reconstructed for "Finno-Ugric" (or to any proto-language stage prior to Proto-Finnic) at all. https://www.academia.edu/1959258/On_Finnic_long_vowels_Samoyed_vowel_sequences_and_Proto _Uralic_x The proposed change *-oCi-> *-uCi- in Finno-Ugric is a more complicated issue that has not been fully resolved, but it is not clear that we are actually dealing with a change in "Finno-Ugric" at all. It could instead be a matter of a change *u > *o in Samoyed, although the conditions of such a change are not transparent. In any case, the sound law *-oCi-> *-uCi- proposed by Janhunen (1981) and supported by Sammallahti (1988) does not work because there are counterexamples. In some cases "Finno-Ugric" fails to display the expected change (at least in FU *koki- 'check, go and see' ~ Samoyed *ko- 'see, find', FU *ćoji- 'sound' ~ Samoyed *so- 'be heard'). On the other hand, in some cases we find this correspondence, even though there is a consonant cluster following the vowel: FU *čučki 'block of wood' ~ Samoyed *čočǝ̑. Samoyed also has a couple of stems which show irregular variation between *u and *o, which also suggests it may be a question of a conditioned change *u > *o in Samoyed: at least *tuj- ~ *toj- 'come', *tuj ~ *toj 'fire', *num ~ *nom 'heaven'. Like 3 Arnaud Fournet 7 days ago @ Aikio It's because your Pokornyan system for PU is false. And your claim that PU did not have long vowels is also fundamentally false and flawed. The word "fire" is obviously cognate to PIE *dhuH- "smoke", while "to see" is obviously cognate to PIE * (s)keu-, you cannot handle these sound correspondences because your system with no laryngeal and fancy symbols like *o, *ü, *ë, *ä is garbage. https://www.academia.edu/s/010b2ad552#comment_809468 30/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu PU should be handled with three basic vowels *a *i *u, four laryngeals and diphthongs. And all the fancy symbols must be canned. But, you're just as stubborn as Kümmel who liked your comment and refuses to accept *H4. Coming to Kümmel, I wanted to say that even Basque proves that PIE had two a-coloring vowels: Basque hartz = PIE *H2rt-k^o- "bear" Basque an(d)re "woman" = PIE *H4ner "man oh yeaaah, I know, Basque is isolated, blablabla... By the way, Dear Prof. Kümmel, your list of Iranian words for "tooth" which look like *haK- does not mention that Basque hagin means "(molar) tooth". As you can see, Basque supports *H2 in that word. So I repeat my question: Re *H4, wann können Sie das Licht sehen? Like Ante Aikio 7 days ago @Arnaud: Whatever. I suppose the one thing we agree upon is that I "cannot handle" the correspondences I mentioned, meaning that I cannot explain them - and as far as I am aware, neither can anyone else at the moment. Unfortunately, I cannot find anything that would help in explaining them in word salad where the only discernible meaning seems to be that everyone not agreeing with some unexplained, way-out Uralic-Indo-European-Basque comparative scheme is a "stubborn" person advocating "garbage" based on "fundamentally false and flawed" ideas. But, of course, based on my previous experience I should have guessed that any comment from me to anything you write is likely to trigger such a reaction; and in the absence of any actual arguments or other input for productive discussion, I'll be happy to leave it at that. Like 1 Arnaud Fournet 7 days ago @ Ante Aikio [Quote is empty blather] Let's put it short, simple and blunt: Your Pokornyan way of reconstructing PU is fundamentally flawed, false, inadequate and doomed to fail, => your castle of cards will never work, nor reach the functional Swiss-Clock level of Laryngealic PIE, Your fancy symbols are a kind of esoteric abstruse garbage, whose main result is that serious sincere people give up trying to reconstruct a clean acceptable PU, [needless to say that I refuse to read and try to understand what your esoteric abstruse apparatus of fancy symbols might stand for => you're basically a crook...] What you call "conditioning factors" is in fact symptoms that your esoteric abstruse pseudo-system of fancy graphemes is a fundamentally flawed heap of garbage, There are no "conditioning factors", the issue is just that your esoteric abstruse pseudo-system of fancy graphemes does not work in the first place, and will never work, no matter what, I'm not the one who is chickening out of discussion, As a temporary conclusion, I would say that you're the problem, not me. https://www.academia.edu/s/010b2ad552#comment_809468 31/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Peter S Piispanen 7 days ago Ok, Arnaud, simmer down. Save that discussion and tone for elsewhere. This is not kindergarden where sand is thrown in the face of those not in agreement or thinking differently. This is a scholarly forum and discussion where argumentation follows known, if not always fully accepted by everyone, facts, and we seek further knowledge by our combined ideas, information, interests and hypotheses. Your suggestions are usually helpful - that is acknowledged, but take this as an only warning: keep it down, and stick to the subject matter at hand please. Like Tom Martin 6 days ago Thank you, Ante Aikio. That article of yours explains well the origin of Finnic *ee and *oo. Like Marcel Erdal 7 days ago Hurro-Urartean was mentioned before. How about looking a bit at grammar, and not just at lexemes, which can be loans (especially since the Hurrians lived in South-East Anatolia and the Urarteans a bit further in the east)? Hurrian is a purely suffixing language for its rich derivation and inflection. Like Turkic, it has possessive suffixes, and a suffix for verbal negation (which do not exist in Mongolic, e.g.). How about connecting it with Turkic? Numerous cognates (not necessarily to my taste) have been produced. Like Marcel Erdal 8 days ago There are not 2001 but tens of thousands of Common Turkic simple lexical stems ("roots" is the wrong term; Semitic and Proto-Indo-European have roots; Altaic languages don't) and the great majority have ChuvashBulgar cognates. Chuvash-Bulgar is just one of the branches of Turkic and does not differ more from the other branches of Turkic than the differences between the other branches. It has its peculiar features (like rhotacism and others) but features of other branches are just as peculiar. It also has many Finno-Ugric loans. ChuvashBulgar caught the attention of comparatists because all early Turkic loans in Mongolic are from this branch the ancestors of the speakers of Chuvash appear to have been neighbors of the speakers of Proto-Mongolic. Chuvash is an aberrant Turkic language, but not more than Khalaj, e.g.. Like 1 Ante Aikio 8 days ago https://www.academia.edu/s/010b2ad552#comment_809468 32/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu The usage of the term 'root' in linguistics is not really limited to the special type of lexical content morphemes that occur in e.g. Semitic languages. Often 'root', 'word-root' or 'root word' are used as terms referring to any monomorphemic (morphologically simple) lexical content morpheme. The term 'stem', in turn, has a quite different meaning: stems can be polymorphemic, and they may contain any number of morphemes. It is true that this usage of the term 'root' has also been criticized, but nevertheless it is relatively common. Like 1 Marcel Erdal 7 days ago Semitic and Proto-Indo-European roots consist of consonants and the vowels are introduced for the grammar. In that situation, consonants are sufficient for determining etymological connections. In the Altaic languages, vowels are an integral part of lexemes (though not of pronouns) and if they don't fit, an etymological proposal is wrong. The term 'root' has been misused in this way. Stems can of course be polymorphemic but if you say 'simple stem' they are not. Like 1 Arnaud Fournet 7 days ago Well, Turkic *bir "one" and *barmak "thumb" are in my opinion etymologically related, so no matter what a "root" or stem actually is in the different lexicographic traditions, Turkic does have roots like *b_r and possibly used to have active vocalic apophony. Like Marcel Erdal 7 days ago In all the Turkic languages where it is (and was) attested (including Chuvash, where the form is a bit different) barmak does not signify 'thumb' but 'finger'; it denotes ANY finger. So there is no semantic connection. Even if there were, you would need a denominal suffix "-mak", which exists nowhere in these languages. So barmak is an unanalysable stem. Like 1 Arnaud Fournet 10 days ago Re https://www.youtube.com/watch?v=pmv3J55bdZc and Renfrew's appraisal of Gimbutas, The word "triumphant" is indeed pronounced twice at 11.41, I suppose he might want to be very polite and also bait the audience, but Renfrew emphasizes that he has his own preferred solution at 10.36. So nothing really new under the sun. Like Peter Kitson 9 days ago https://www.academia.edu/s/010b2ad552#comment_809468 33/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Rasmus Bjørn: your picture looks as if you may be too young to remember, but Renfrew spoke from the perspective of a ‘New’ alias ‘Processual’ archaeologist who with his fellow-processualists denied for decades that the waves of migration M.G. postulated from the steppe to Europe were likely ever to have happened. When he says that work on ancient DNA “has completely rejuvenated Marija Gimbutas’s ‘Kurgan’ hypothesis because it has supported strongly some—not all, but some—of the elements which she emphasized”, and that “She ends up as the triumphant precursor of much current work”, the fact of that migration is what is mainly meant. A couple of minutes before the end, “Now I still feel myself that the Hittites have an important rôle in all of this, and it’s the absence of good ancient DNA material from Anatolia that is obscuring the problem, and I think that the Anatolian part of the picture is at the moment devalued by the lack of that material.” By implication, the linguistic elements are ones DNA work has not supported strongly. One of the reasons why it has not is that Haak et al. did not take into account the practicalities of intermingling of patchily distributed populations preferring different kinds of real estate in landscapes with enormously lower population densities than now, so they misinterpreted the historical meaning of their own figures. If in bones of an early stage of the relevant culture the proportion of DNA from the new population is 70% and at a later stage 50%, that does not mean that there has been a historically unexplained “resurgence” of the old population. It means that at the early stage only people with 70% ancestry in the new population had adopted enough of its material culture to be recognizable archaeologically as belonging to it, whereas by the later stage its fashions had spread to the extent that people with only 50% of such ancestry were; and the lower later figure is the maximum positable for the new element in the original mixture. (If the new élite had special privileges in access to women, as seems to have been customary in later steppe cultures of more than one linguistic group, the original proportion might have been very much smaller.) When I taxed Dr. Haak with this at a conference at Aberystwyth in 2014 he admitted it. He stated that what he thought he had done was not support the steppe hypothesis but “level the playing-field” between the Anatolian and steppe hypotheses, a position to which he adhered scrupulously in more formal language in subsequent written disputation with the late L.S. Klejn and others including his fellow-geneticist Allentoft. He added that the published version of the paper of which he is named as lead author does not fully reflect his own views (it was written by his archaeologist colleague Professor Anthony). Like Juho Pystynen 10 days ago I noticed that Róna-Tas e.g. in his 1988 handbook article on Turkic–Uralic contacts considers this borrowed specifically from Samoyedic (as "Ancient Chuvash" *puyu > *huyu > *uyu). Sounds like an awfully late date for *p > *h though. Unlike 1 Annotation: Page 7 Peter S Piispanen 10 days ago Interesting! So the idea about the Chuvash form being a borrowing has already been suggested. And I agree, the full change from Samoyedic to the modern Chuvash form seems too late. https://www.academia.edu/s/010b2ad552#comment_809468 34/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Could the Chuvash form be another one of those Mongolic borrowings / Mongolic correspondences because the Mongolic words were already *h-initial to start with (although likely *p- in Pre-ProtoMongolic)? The problem then, however, is that all modern Mongolic languages are very eroded forms from that Proto-Mongolic root and these eroded forms appear to focus on the original final syllable, which does not exist at all in any form in Chuvash. In Mongolic we have: Written Mongolian: uɣula (L 866: ujil 'excrescence on a tree'; ?L 14: aɣli id.); Khalkha ūl 'tinder'; Buryat ūla 'пробка'; Dagur xuāĺ 'tinder'; Shary-Yoghur χū 'id.'; Monguor fula 'amadou'. So, I don't know what exactly has been going on with this word that it ended up this way in Chuvash. Hmmm... Like Peter S Piispanen 16 days ago Another suggestion: Proto-Turkic *tut- 'to grasp' (VEWT 502, EDT 451, Егоров 268-269, Федотов 2, 268-269, Stachowski 233) > Proto-Samoyed *tǝtǝ- 'haften = to stick' (SW 149). To my understanding the PS *-ǝ- should go back to *u, just as is found in the PT form. Semantically, to grasp and to stick are fully comparable. The PT root is extremely well-attested - and has Altaic correspondences while the PS root is attested in Nganasan, Tundra Nenets, Forest Nenets, and Selkup, justifying the PS root reconstruction. Any comments or thoughts? Like Peter S Piispanen 12 days ago I take it the lack of comments on this suggestions may mean that the comparison is accepted. The PS meaning is perhaps better translated into English as 'to cling' or 'to adhere' The Turkic meaning of 'to grasp' aka 'to grab' or 'to clutch' is thus well-suited semantically to match the Samoyed meaning. The comparison is also a seemingly perfect match phonologically, and so this suggestion should probably be added to those in this draft paper. One or two current suggestions will be demoted to near-hit status only, and thus removed from the main part of the paper. Like 1 Ante Aikio 11 days ago In my view, this etymology does involve major problems. The first is semantics. English 'to stick' is an ambiguous glossing because English shows widespread homonymy of transitive and intransitive verbs. The key issue here is that Nenets /tǝdǝ-/ and Nganasan /tǝtǝ-/ are intransitive verbs, so they only mean https://www.academia.edu/s/010b2ad552#comment_809468 35/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu 'to stick' in the sense 'to get stuck on something, become or remain attached to something' - and this is of course rather far from 'to grasp', as there is not only a difference of meaning but also a difference in argument structure and in the semantic role of the subject argument: the subject of a verb meaning 'to be or get stuck on something' is the object that is stuck, not a human or other conscious agent 'grasping' or otherwise manipulating something. Note, moreover, that the verb actually occurs in Northern Samoyed only and the Tym Selkup cognate cited by Janhunen (Samojedischer Wortschatz p. 149) is wrong. The Tym Selkup verb (1SG /tǝd-ap/) actually does have a meaning closely corresponding to the Turkic form you mentioned: 'greifen; anfangen' = 'grab; begin'. However, it cannot be a cognate of the Northern Samoyed verbs because Tym Selkup /ǝ/ does not reflect Proto-Samoyed *ǝ̑. According to Alatalo (Sölkupisches Wörterbuch 1074) the Proto-Selkup form of the verb was *ti̮ t(ǝ)-, which is incompatible with Northern Samoyed *tǝ̑tǝ̑-. As Alatalo notes, the Selkup verb instead reflects Proto-Samoyed *ti̮ t-; the verb itself was apparently only preserved in Selkup, but the deverbal instrument noun *ti̮ tsan 'pliers' also has a cognate in Nenets (see Samojedischer Wortschatz p. 160). As we are only left with the Nenets and Nganasan forms to compare, we face an additional phonological problem. In the absence of a Selkup cognate we cannot distinguish between Proto-Samoyed *t and *č, so we have several possible Proto-Samoyed reconstructions. Moreover, Proto-Samoyed *t has multiple regular Pre-Proto-Samoyed sources: word initially it can reflect PU *t and *s, and between two vowels even clusters like *sk, *ks, *tk, *kt. So we have extremely many potentially valid source forms for the Northern Samoyed verbs: they could regularly reflect at least PU *susV-, *sutV-, *suskV-, *suksV-, *sustV-, *sutkV-, *suktV-, *sučV-, *sučkV-, as well as all any form otherwise identical but with initial *tand *č- (a total of at least 27 forms, and there are possibly even others). For this reason, it is particularly easy to find coincidentally matching comparanda for any Northern Samoyed stem of the shape *tVtV-. One could note that this illustrates the problems of working primarily on the basis of mere etymological references (like Janhunen's Samojedischer Wortschatz, for example). It is relatively easy to come up with superficially plausible-looking comparisons by matching reconstructed forms in two proto-languages, accompanied by one- or two-word glosses of the supposedly reconstructed original meaning of the word given by etymological references. However, when one carefully goes through the actual lexical data, one often finds many small details that yield a more elaborate and precise understanding of the reconstructed words in question. And it is not rare, in my experience, that these details end up making the initially promising etymological hypothesis look much less plausible. Unlike 6 Daniel Nikolić 16 days ago Would this be a borrowing from some Turkic language that underwent l' > š? If so, we have to explain only -kaif the consonant was at that time something not existing in Proto-Samoyed and substituted with -rLike Annotation: Page 3 https://www.academia.edu/s/010b2ad552#comment_809468 36/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter S Piispanen 16 days ago I see your point, but it is starting to look like - based on other input in this Session - as if the Samoyedic root may still be of Uralic etymology, and in that case no loanword hypothesis would be required at all. Like Tom Martin 13 days ago The idea that Proto-Turkic had *l' which then changed in most Turkic languages to š, is an idea that is maintained by supporters of the Altaic theory. Others, who believe that the Turkic family has not been proven related to any other family, say that here Proto-Turkic had š, which then changed in one branch of Turkic to l, causing a merger of š and l to l, and then Mongolic and Tungusic languages borrowed some words from that branch of Turkic, with l. There is no evidence of any language having l' in such words. Like Arnaud Fournet 13 days ago *l' is a cover symbol, and possibly more technical symbols like ɬ or ɮ would be preferable. Unlike 1 Christopher Culver 13 days ago The idea that Proto-Turkic had ɬ is now maintained by more than just Altaic supporters. See Antonov & Jacques’ “Turkic kümüš ‘silver’ and the lambdaism vs sigmatism debate" for evidence of a Proto-Turkic lateral that does not require accepting claims of Altaic genetic relatedness. Like 3 Peter S Piispanen 13 days ago Ah, the lambdaism vs. sigmatism debate is an interesting question indeed! You are correct, Christopher, in that there is no Altaic genetic relatedness requisite for accepting a possible Proto-Turkic lateral *ɬ. The same is true of a possible Proto-Turkic *ŕ, The EDAL reconstructs these where the Chuvash form displays the retained -l- and -r- for such roots, whereas all other living Turkic languages (i.e. Common Turkic instead display -z. Let's take but two examples: Proto-Turkic *bāl-dɨŕ 'a man's wife's younger sister' > Middle Turkic baldɨz 'id.', Yakut balɨs 'id', but Chuvash poldъr 'id.'. (perhaps comparable to Proto-Mongolic *balčir 'very young' and Proto-Tungusic *baldi- 'to bear, to be born'; not the best example, there are far better). Indeed, we do have many words where Chuvash displays these, and perhaps the most interesting external comparison is then to the available Mongolic and Tungusic forms. As most of us know, these very often display the exact same consonantism here as Chuvash does. Without having any preconceptions, this would suggest in any basic linguistic study, that West Turkic regularly displays https://www.academia.edu/s/010b2ad552#comment_809468 37/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu certain phonological traits extensively shared with both Tungusic and Mongolic, whereas East Turkic does not. Why would this be so? Considering the very extensive lexical sharing on every possibly conceivable level it COULD suggest that these three language groups are genetically related (i.e. the Altaic language theory). It would also seem, based in external comparanda, as if Proto-Turkic had lexicon which was phonologically closer to the forms attested in the later Mongolic and Tungusic languages: it is completely logical to assume that West Turkic retains the traits of Proto-Turkic the closest regarding the whole *ŕ and *ɬ. phonemes, while these very regularly changed in all East Turkic languages because they are still retained in Tungusic and Mongolic. The opposite assumption, that East Turkic retains the original forms, which for some unforseen reason changed dramatically in West Turkic to exactly fit the Tungusic and Mongolic consonantism does not really make any sense at all IMO. From this point of view, I will claim that *ŕ and *ɬ are probably the original Proto-Turkic phonemes, still present in closely matching forms in West Turkic (i.e. Lir-Turkic), Mongolic and Tungusic, but changed fairly radically in all East Turkic (i.e. Shaz-Turkic) languages (to z and š respectively). Like 1 Juho Pystynen 13 days ago Note also *r and *j being the established reflexes in borrowings into Samoyedic: *jür '100' ← *jǖr₂, *jür 'fat' ← *ür₂, *kïj 'winter' ← *qïl₂, *pajmå 'boot' ← bal₂maq. These could simply continue *r and *l in the loan-giving Turkic variety, or also something intermediate like *ð and *ʎ. Potentially a pre-Samoyedic *ð (< PU *d, conditionally *t) could have been substituted even for a Turkic *z, but nothing similar is feasible for *j ~ *š. Unlike 2 Peter S Piispanen 13 days ago Yes, exactly, Juho! Which indeed strongly suggests that the Turkic borrowings made into Samoyed was either from Proto-Turkic or from a Oghur/West Turkic variety! This is actually something that needs to be discussed within this very topic, and I will add notes to that effect to the updated paper as well. Thanks for this important suggestion! For the words you mention, we then - based on the existence of these Samoyedic words- have solid evidence for the correctness of the reconstructed Proto-Turkic *jǖŕ 'hundred' (> Chuvash śǝʷr 'id.'), Proto-Turkic *ǖŕ 'fat' (> Chuvash jor-var 'скоромная пища'), Proto-Turkic *Kɨĺ 'winter' (> Chuvash xǝl 'id.') and Proto-Turkic *bAĺ-mak 'boot, shoe' (is there a known Chuvash cognate?). Like 1 Tom Martin 12 days ago Well, the argument from geography is not that decisive. Suppose West Turkic originated to the north of East Turkic? That could explain why the borrowings into Samoyedic, Mongolic, and Tungusic have the https://www.academia.edu/s/010b2ad552#comment_809468 38/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu West Turkic consonants r and l, with of course Samoyedic changing the l to j. So I think the situation of the consonants in Proto-Turkic is not resolved. Chuvash is of course one of the most northern Turkic languages. Sakha (Yakut) of course reaches further north, but that is due to a later expansion of Sakha to the north. Like Ante Aikio 16 days ago I don't think a strong case can be made for the claim that the cluster *-lp- was not allowed in Proto-Uralic. The lack of examples may simply be an accidental gap in the reconstruction, because otherwise clusters of the type "liquid + stop" seem to have been well permitted: we know examples of *rp, *rt, *rk, *lt, *lk, so the absence of *lp is an unexpected gap. In general, it would be unrealistic to expect every single cluster that actually occurred in the proto-language to be preserved in the etymological material, so mere absence of evidence cannot be interpreted as evidence of absence. Like 2 Annotation: Page 7 Peter S Piispanen 16 days ago Ante, NOBODY has claimed that *-lp- wa not allowed in Proto-Uralic; on the contrary, it most probably was even though it may not have been reconstructed anywhere. Just as you present there are several clusters which probably did exist, but which we haven't seen preserved in the etymological materials. The same is true of many clusters in Yukaghir, Tungusic, etc. However, the voiced variant, *-lb-, cannot have been allowed in Proto-Uralic. That is the only point that was given in this entry. Like Ante Aikio 15 days ago Well, I don't think that was quite clearly put in the manuscript, and I also do not quite see the point here: of course, there was no *-lb- in Proto-Uralic because there was no *b. But it still does not really explain why a foreign cluster *-lb- would turn up as *-p- in Uralic/Samoyed. Like 1 Peter S Piispanen 15 days ago Ok, I understand and will make it clearer in the text. Yes, the *-lb- > PS *-p- does require some similar or parallel examples or argumentation to be fully believable - I will work on that as well. I had something specific in mind, which is not here, and will look it up again, thanks. Like Ante Aikio 15 days ago https://www.academia.edu/s/010b2ad552#comment_809468 39/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu It is of course difficult to know for sure what the reflex of *-lp- in Samoyed would be, as we do not have any clear examples of this cluster. However, if it existed in PU (which appears probable), my first hypothesis would be that it developed into Samoyed *-jp-, because PU *l generally developed into Samoyed *j in the syllable coda, and the development *lm > *jm is in any case attested (cf. PU *ćilmä 'eye' > Sam. *sǝjmä). Unlike 2 Peter S Piispanen 15 days ago This is a reasonable working hypothesis indeed. So, Pre-Samoyedic *telbe would possibly become *täjpə, although this is still some ways off PS *təpə. Earlier *-äj- would become forms like -e- and -i- (as long as the long voweled forms) in various Samoyedic languages, but this is not the case here. How would we account for vocalism of the last step, *täjpə > *təpə ? Like Ante Aikio 15 days ago The question is whether it even can be accounted for in the first place; I don't see how that could be done. PSam first-syllable *ǝ̑ goes back to PU *u, so in regard to known sound laws, PSam *tǝ̑pǝ̑ / *čǝ̑pǝ̑ would imply an earlier form *tup(p)V / *sup(p)V / *čup(p)V (note that we cannot even distinguish between PSam *t and *č here, as we have no data from Selkup where the opposition of the two phonemes was retained). The correct default hypothesis is that the comparison is wrong and there is no etymological connection between the two forms; I'll be happy with that as long as there is no plausible account of how *telbe could yield PSam *tǝ̑pǝ̑. Like 1 Peter S Piispanen 15 days ago A good analysis and explanation, Ante. Your zero hypothesis could be correct at that, as the comparison could present two accidentally similar forms only. It is true that the phonology does not appear to be regular regarding the vocalism, and possible not either regarding the consonant cluster. Given the limited number of phonemes in the human languages, and the shortness of roots, there are bound to be a number of coincidental non-related similarities to be found across languages, and the details do suggest that this is one of those cases. Therefore I might have to relocate this suggestion to one of those near-hits mentioned. Like Ante Aikio 14 days ago It is indeed very easy to find such accidental similarities. For example, I could immediately suggest another formally even better-looking "etymology": this Proto-Tungusic *telbe 'dirt, dirty' shows quite a neat correspondence with Proto-Saami *tuolve̮ 'dirt, dirty' (< Pre-Proto-Saami *talwi / *tolwi), which so far lacks an etymology. But of course, also this similarity is purely coincidental and there cannot be any true etymological connection between the two words. https://www.academia.edu/s/010b2ad552#comment_809468 40/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like 2 Peter S Piispanen 14 days ago I suppose the same could be said about Proto-Tungusic *tikte 'louse' and North Saami tikke 'louse'. At any rate, I am sure everybody agrees on that there needs to be more evidence, reason and attestation to make a valid loanword suggestion based on scant evidence found in two distantly located, non-related languages. Like Onno Hovers 14 days ago "Proto-Saami *tuolve̮ 'dirt, dirty' (< Pre-Proto-Saami *talwi / *tolwi), which so far lacks an etymology" Perhaps IE *dʰelbʰ 'to dig, to hollow out' > West-Germanic *delban 'to dig' > English to delve, Dutch delven; Lithuanian delbti 'to drop', Slavic dьlbiti 'to hollow out, to chisel',..)? The semantic development would be 'dig' > 'that what is digged out' (cf. Dutch delfstof 'mineral') > 'earth, dirt'. Like Ante Aikio 12 days ago @Onno: English "dirt" is of course polysemous, but the Saami word does not mean 'dirt' in the sense of 'soil of the earth' - it only means 'dirt which stains or sticks on objects', so there is no semantic connection at all to digging. What is more, the vowels do not match either, and there are no plausible examples of substituting Saami *v for Germanic *b in old loanwords; the regular substitute is *p. Like Juho Pystynen 12 days ago Speaking of *tōlve̮, a much better candidate for a Siberian cognate would seem to be Samoyedic *t¹ålwə 'darkness'. The correspondence seems rather regular, though the retention of *lw is unusual (this is indeed the only known case of this cluster Proto-Samoyedic). Like Ante Aikio 12 days ago Semantically the comparison to PSam *tålwǝ̑ would be quite a stretch. Moreover, there actually are two examples of *lw > *jw in Samoyed: PU *tolwa 'wedge' > PSam *tajwå (cf. Mordvin *tulǝ, Permic *tul(j-)) and *talwa- 'lead, take, transport' > PSam *tåjwå-. Besides, the reconstruction of the cluster *lw for the word *tålwǝ̑ does not seem quite clear to me. The only form suggesting this is Tundra Nenets /talw°/ 'dark part of the night'. but on the other hand, this is contradicted by Forest Nenets /tanʟ°/, which points to *lt or *lč instead. Kamas /tōlu/ is not helpful in solving the problem, and there seem to be no cognates in other Samoyed languages. https://www.academia.edu/s/010b2ad552#comment_809468 41/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like 1 Marcel Erdal 13 days ago I am convinced that Proto-Turkic had *l' / ɬ / ɮ and do not support the Altaic hypothesis; there is NO causal relationship between the two matters. Like 5 Arnaud Fournet 13 days ago yes, of course these are two different issues. Like Kamil Kartal 15 days ago Hello friends, greetings from Antalya, Türkiye. I could have a quick look at the article and found it very interesting, Congratulations for that. I would like to attract your attentions to an important point about Turkish language, i.e. her most fundamental difference than any other language living or dead. As one of the borrowed words (perhaps the second one) indicates the onomatopoeic feature, it should be well noted that at least half of the Turkish vocabulary (some 100K words) are originally onomatopoeic - that is, imitation of sounds in nature. So when you look at the proto Samoyedic or Proto (any -nd-all) languages comparison between Turkish and tracing back to the roots, please keep in mind that nly in Anatolian Turkish there are more than 15K words listed by Hamza Zülfikar (https://emagaza-tdk.ayk.gov.tr/detay/628/turkcede-sesyansimali-kelimeler-inceleme-sozluk-2018) and there is known to be more than 10 times greater vocab than this in all Turkic languages together - not less than 150K words ALL and PURE imitations. This is the minimum amount we can confidently assume to be existing in Turkish. And it simply means a lot of gravity for the language emergence from the nostratic perspective. Letting aside other heavenly theories, mankind invented language himself by carefully listening to the mother nature and her established + complicated communication language. This was then a need as important as water and food for survival. IF there was no language, there would be no human existence, no linguistic studies, no churches and no gods! We are all and deeply indebted to the first language. Tihs is my first point about the article. Secondly, all languages have proto versions and there are great gaps between now and then of English, or Arabic, or Hindu, or Chinese. But this does not apply to Turkish for Turkish has NOT changed from its very early so called-proto version to today. The words are still the same in Turkish of today and Turkish of, let's say, 2000 years ago; or even much older times. Words, pronunciations, meanings, uses, are still the same. Not only the words but in general the structure of the language has not changed at all. So there is not a PROTO Turkish language of ancient times establishing the roots of today's Turkish language but which is much different in form compared to today. There is a Turkish language of all times spoken by hundreds of millions of peoples in a variety of large geography with various accents and dialects - yet they are all one and the same Turkish language of all times. So let me repeat that there is no PROTO TURKIC or PROTO TURKISH language. Thanks for your kind attention. Regards, https://www.academia.edu/s/010b2ad552#comment_809468 42/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Orhan Erdoğan 14 days ago I totally agree with you. Like 1 Juho Pystynen 19 days ago Even earlier apparently: probably already pre-Smy *pukəja or *puxəja (or *-la, cf. Mongolic?) with regular contraction to bisyllabic *pujå. Unlike 1 Annotation: Page 7 Peter S Piispanen 19 days ago Yes, I was wondering about that, but forgot to mention the possibilities! The forms you suggest do make sense according to our understanding on Samoyed historical phonology. In particular we could have: (Pre-)Proto-Tungusic *pug(i)ju > Pre-PS *pukəja > Pre-PS *puxəja > PS *pujå. These changes are certainly possible and the final contraction would even be very expected and regular. I am not certain why the vowels would change in this way but perhaps there are prosodic reasons for that. As you mention, Pre-Proto-Mongolic *pughula is another possible source for Pre-Proto-Samoyedic like you mention. Like 1 Robert Lindsay 18 days ago Problem: Proto-Samoyedic goes back to the Sayans 5,300 YBP. There is no pre-proto-Mongolic at that time. There is Turkic-Mongolic in Outer Mongolia after 5,200 YBP after Proto-Tungusic takes off to go to the mouth of the Amur. Pre-Proto-Mongolic has to date from 3,400 YBP after Proto-Turkic takes off to go to Western Siberia and Northern Kazakhstan where their homeland lies. This problem runs though a lot of these suggestions. These groups were never close enough to each other in space-time (go get relativistic here) to trade words unless they had jets back then, and I don't think they did. Unlike 1 Peter S Piispanen 18 days ago Robert, you are spot on regarding the chronologic problems - these appear at times to present insurmountable obstacles to various loanword suggestions and etymologies. What it means, however, is that our view of the historical situation is far from fully understood - in fact, certain facts are probably completely off. I see what you write here, and won't protest against these ideas - and as you know we are very much on the same page regarding much about the historical situation, various proto-languages, the impossibilities and possibilites of borrowings, etc., such as the impossibility of four different chronological waves of migrations from thousands upon thousands of miles away of different groups of https://www.academia.edu/s/010b2ad552#comment_809468 43/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Uralic speakers to lend their vocabulary to the utmost northeastern Siberian corner of the Yukaghir lands; no amount of historical twisting to make the impossible possible, no assumption of non-existent trade routes, no jet planes can explain such facts where the similarities simply represent an ancient genetic affiliation and nothing else. Chronological problems are a major part of the problems with the comparisons of the Altaic languages they are extremely far separated in time, and quite far separated in space as well. This is a topic that could fill volumes, and actually already have, and I realize with your input here that I must mention and discuss the chronological problems with these suggestions here as well. Clearly there is much we don't know or understand about the historical situation and it has to be acknowledged. Like 1 Juho Pystynen 18 days ago Here I would agree with Robert's gist: if this were indeed to date already to pre-Samoyedic, I'm not very convinced that the direction is Tungusic → Samoyedic (and of course definitely not Mongolic → Samoyedic!); maybe it's rather e.g. para-Samoyedic → Tungusic, or a loan from a common Central Siberian substrate in both. I could even offer a somewhat speculative native etymology: there is a West Uralic *pukta- 'to wake (tr.)', in principle analyzable as a causative *puk-ta- from an otherwise unattested root **pukə- 'to wake (intr.)'. Pre-Samoyedic *pukə-ja could then be seen as a parallel actor noun 'waker' > 'kindler'. Unlike 1 Peter S Piispanen 18 days ago Hmm.... at least phonologically and morphologically Juho's suggestion works out, but it seems a bit shakier on semantic grounds. I assume that 'tinder' would be what awakens the fire - I suppose it is a possibility (also considering that fire actually 'eats' whatever it burns in some Uralic languages), but then again, do we know of any parallels to this view in another language? In any event, it is worth mentioning it in the paper itself under this entry. Also, the idea of para-Samoyedic →Tungusic or even para-Samoyedic → Pre-Proto-Tungusic is certainly possible. However, as noted in this draft the Tungusic root is phonologically more complex with a cluster whereas the Samoyedic is not; still, of course, the para-Samoyedic root could have retained a more complex root, which has been lost in Proto-Samoyedic but which is still visible in Proto-Tungusic. Difficult questions! One way or another these chronological problems must be solved - it seems as if the more one looks at such matters the further back in time one must place different proto-languages. This may not be a healthy trend. It has led to the fairly disturbing notion of Pre-Pre-Proto-languages that I have read at places or other odd nomenclature to explain the chronology of older languages where certain sound changes appear to predate other known sound changes. I used Early, Middle and Late to describe certain sound changes in Yukaghir in my sibilant and semivowel papers, and these I took from Häkkinen's somewhat earlier paper (despite me not at all agreeing with his borrowing thesis). https://www.academia.edu/s/010b2ad552#comment_809468 44/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Of course, languages do not appear out of a vacuum - Mongolic, Tungusic and Turkic did exist in some form already 5-6 thousands of years ago, but what then should we call these predecessors? It depends who you ask, I guess. Like 1 Juho Pystynen 17 days ago Yes, at minimum I would have to look for parallels for this semantic development to 'tinder' before proposing this fully seriously. It would be also good to know if Tg. *pug(i)ju or Mg. *huɣula are amenable to any morphological analysis. Like Onno Hovers 17 days ago There is even a well-known case of what looks like a Tungusic borrowing into Proto-Uralic proper: Tungusic siŋgere > PU *šiŋiri 'mouse'. Of course this could also be explained as 'Central Siberian substrate'. When it comes to the dates and homelands of proto-languages there is a lot of uncertainty. In general I am not of the school that 'wants it big and early'. Proto-Uralic is probably from around ~2500-2000 BCE. Clearly there were extensive contacts between east and west around that time (Seima-Turbino). Like Mikhail Zhivlov 17 days ago @Onno: Apparently, Uralic-Indo-Iranian contacts started around 2000 BCE, if we associate early IndoIranians with Sintashta culture; and by the time of these contacts Uralic was already not a uniform protolanguage, as is evident from uneven distribution of IIr loans across Uralic branches and various irregular correspondences between Uralic branches in these loanwords. So 2000 BCE as a date for the breakup of Proto-Uralic seems too late for me. As for the word for 'mouse', why not the other way round - from some unattested Uralic language into Tungusic? Like Onno Hovers 17 days ago "As for the word for 'mouse', why not the other way round - from some unattested Uralic language into Tungusic?" 1) Aikio identified PU *š as being a sort of loan-word phoneme in "The Finnic ‘secondary e-stems’ and Proto-Uralic vocalism" pages 44-46. 2) A 3 syllable structure is unusual for a basic Uralic word, and PU *-ri is not a widely attested nominal suffix as far as I know. 3) There would be an unexplained loan substitution PU *ŋ > Tungusic *ŋg. (Although I myself think that https://www.academia.edu/s/010b2ad552#comment_809468 45/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu most of what is reconstructed as PU *ŋ was actually PU *ŋg (> Ugric *ŋk; rest *ŋ). But this is based on 'external reconstruction' and not relevant to this discussion). Unlike 1 Ante Aikio 15 days ago @Onno: there are actually quite a few trisyllabic Uralic noun roots that cannot be explained as derivatives of any bisyllabic root: in addition to *šiŋiri 'mouse', one could mention e.g. *jikini 'gums', *ćijili 'hedgehog', *jäsini 'joint', *ńi̮ kćimi 'gill', *epiki 'owl', *wVdimi 'marrow', *ki̮ ŋiri 'curve, bend'. Although these were infrequent compared to bisyllabic vocalic roots, there's plenty of good examples nevertheless. It is merely some kind of axiom of earlier research that Uralic roots were always bisyllabic. Unlike 5 Onno Hovers 15 days ago I have looked into it in more detail. The problem for me is that there is no good up-to-date list of derivational suffixes. Older works on Uralic morphology do list an -r suffix but are unclear about its function. Hakulinen writes: "The original connotation appears to have been diminuitive". So regardless of whether any third syllable has to be viewed as originally being some suffix or not, on closer inspection my previous point (2) is not valid anyway. Unlike 2 Arnaud Fournet 14 days ago Among these three-syllable words, some can be derived from *CvC- Nostratic roots. The word *ńi̮ kćimi 'gill' is conspicuous for having four consonants, which suggests that the word is either suffixed twice or that it's a compound, the latter being probable. Like Juho Pystynen 19 days ago UEW's compares this with *śiŋe ' (> Khanty 'bend in wood', Finnic & Permic 'ceiling beam'), maybe worth noting as a better phonetic match though clearly worse semantically. Janhunen's *siŋ probably needs some sort of amendment anyway, besides /a/ in Nganasan also long /ī/ in Nenets and mid /e/ in Enets are hard to explain. Unlike 1 Annotation: Page 3 Peter S Piispanen 19 days ago https://www.academia.edu/s/010b2ad552#comment_809468 46/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Yes, indeed, a new reconstruction is required! But I am not certain if it is possible if we assume that all of these Samoyedic words are related. It is possible, as mentioned, that we are talking about two different origins here. I have noticed that words meaning 'to bend', 'curve', 'round' etc. appear to very often have *-ŋ- as part of the root not only in Uralic, but also in Turkic, Mongolic, Tungusic, Yukaghir and other languages. Why is that? Is there some sort of sound symbolism I'm missing here? Like Juho Pystynen 18 days ago Good question… [ŋ] as onomatopoetic for the creak of wood bending or twisting? Like Peter S Piispanen 18 days ago Maybe so! The sound is present also in words meaning 'stomach', 'circle' and 'opening'. Perhaps there used to exist some sort of onomatopoetic short root (even a Wanderwort?) in Pre-PU and Pre-Altaic times, which became derivative to function to describe a lot of different round things, concepts or objects. Like Robert Lindsay 18 days ago Or maybe they're all related? ;) Uralo-Altaic theory goes back to 1820. Unlike 1 Juho Pystynen 18 days ago Words as otherwise different as PU *peŋərä 'wheel' and *joŋsə 'bow' don't seem to have a snowball's chance in hell to be related unless you suggest something like starting to break them down to onephoneme "roots"… Like 1 Peter S Piispanen 18 days ago Of course nobody suggests that ALL words being even remotely related to anything round must be connected to one another. No, merely that there are SOME words with the engma with meanings related to roundness in several different, nonrelated languages. Sometimes these words are similar to each other. It is odd, but we must probably assume coincidence for such matters. Like Ante Aikio 17 days ago https://www.academia.edu/s/010b2ad552#comment_809468 47/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu The Nganasan form in actually <сыӈ> = /si̮ ŋ/, which would imply Proto-Samoyed *seŋ in Helimski's revised reconstruction of Proto-Samoyed vocalism. However, it does not match the long vowel in Nenets. Unlike 3 Peter S Piispanen 16 days ago Yes, so there really appears to be two different groups of words in Samoyedic with this meaning; it seems quite likely that they have different origins, and that two roots should be constructed to capture them all. Like Ante Aikio 15 days ago I don't think we are dealing with words of different origin: although the vowel correspondence is irregular, the forms are otherwise very close, and the meanings belong to a very specific semantic field: Tundra Nenets śīʔ : śīN- 'сторона чума против входа (<считается священной>)' Nganasan si̮ ŋ 'чистая часть чума напротив входа' Kamas si̮ ŋ 'Zeltwand der Tür gegenüber' I think it is more likely that we are dealing with some kind of loanword from an unknown source; the irregular correspondences could have resulted from the same word being adopted independently into different Samoyed languages. Like 1 Peter S Piispanen 15 days ago Yes, that is the point exactly: probably two independent loans - no monogenetic origin, so to say, on the Samoyedic side, but probably ultimately from the same source, albeit independently. Like Geoffrey Caveney 14 days ago Since you are discussing the possibility of [*-ŋ-] in roots with meanings related to creaking / bending / twisting, I may also mention here Proto-Yupik-Sirenik *caŋuʀ- 'be twisted or bent down' (CED 75) and Proto-Eskimo *ciŋquʀ- 'crack or crackle' (CED 89), among others. Unlike 1 Jonathan N Adsit 15 days ago https://www.academia.edu/s/010b2ad552#comment_809468 48/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Hello, thanks for inviting me first of all. I greatly enjoyed reading this piece, and I was unaware of some of these borrowings you suggest. Absolutely fascinating. That said, the only thing I would mention is that which Ante already did, which you've already discussed with him. I think that could be made a littler clearer what the intentions were with that. Like Peter S Piispanen 15 days ago I am glad you found the piece to be an interesting and fascinating read! And yes, agreed, more than a few matters need to be clarified (and even exemplified in a few places) to improve the manuscript to publishable standards. Like Peter S Piispanen 16 days ago A third consideration: Proto-Turkic *dar- 'to go apart, scatter, spread; to branch, be forked; branch; claw; finger' (VEWT 463, EDT 529, ЭСТЯ 3, 150-151, Дыбо 312, Лексика 256, Федотов 2, 251, Stachowski 218) > Proto-Samoyed *tar(ǝ)- 'to divide' (*tär- in SW 154) ? A problem with this suggestion is the supposed existence of PU *šurV 'to cut, to divide', which has all of the Samoyedic representatives listed in it. Any comments or thoughts? Like Alexander Savelyev 16 days ago One odd thing about your way to cite Altaic data (here and -- what is more crucial -- recurrently in the paper) is that you cite the reconstructed forms and meanings as given in the Altaic dictionary while avoiding to mention the source. The works that you refer to instead are mentioned in the Altaic dictionary to cover the research history. Most of those works do not contain the phonetic and semantic reconstructions that you rely on. Unlike 2 Peter S Piispanen 15 days ago You are fully correct: the references in combination give the necessary data to reconstruct the protoform, but the proto-form itself is not to be found in all of those references, only in a few of them, and then in different forms at times. Just like has been suggested - and agreed with - I will separate the references and give the variations of proposed proto-forms in each entry so that it is clear what is presented and where. I probably should also refer to the EDAL which is the source that usually has https://www.academia.edu/s/010b2ad552#comment_809468 49/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu combined the data into (their own) proto-forms as well (sometimes, seemingly, to better fit the protoforms presented for other Altaic languages actually)! Like Juho Pystynen 15 days ago Cf. also PIE √der- 'to tear (intr.)' which seems a bit closer semantically. Unlike 2 Peter S Piispanen 15 days ago Semantically 'to go apart' and 'to be forked' surely is close to 'to divide'. Semantics, however, is all too often a subjective matter, and what is obvious or clear to one, is not clear or connected at all to another, so I concede that 'to tear (INTR)' could be close to 'to divide' as well. If correct, this PIE root must have been borrowed from some daughter language into PS due to chronological issues. However, in my view the Turkic root is closer both phonologically and semantically, and almost contemporary, with the PS form than any PIE root could be. In any event, the PIE root is another worthy comparison of interest and will be added to the entry as well. Thanks! Like Ákos Bertalan Apatóczky 16 days ago According to MNyTESz its an uralic heritage in FU languages (attested forms Old Hungarian ket- ’to bind’ (modern: köt- ’to bind, to knit’), Man’shi köt-, Udmurt kėtki-, Finnish kytkeä- etc.) they conclude the proto FU form might have been *kitke- or *kütke. MNyTESz pp. 625-626. Unlike 2 Annotation: Page 2 Peter S Piispanen 16 days ago I never considered Finnish kytkeä 'to connect' to be a possible cognate of the Samoyedic rooot, but this is an interesting comparison at the very least worthy of mention. Thanks, Ákos! Like Ante Aikio 15 days ago I don't see how Uralic *kütki- (> Finnish kytkeä, etc.) could be in any way connected with this Samoyed verb: the vowel correspondence is entirely irregular. Besides, also the Mansi verb mentioned in this connection cannot be related to the Finnish/Udmurt forms because of the completely irregular sound correspondence (it does not show the expected reflex of the cluster *-tk-, and the vowel is irregular, too), although it is erroneously cited as cognate by all etymological dictionaries. Like 1 https://www.academia.edu/s/010b2ad552#comment_809468 50/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter S Piispanen 15 days ago Well, I actually agree fully with this analysis. I don't believe they are from a common etymon either, but still the similarity is peculiar and noteworthy, and so I will at least address these matters in a footnote. Like Ákos Bertalan Apatóczky 15 days ago @Ante Aikio The MNyTESz doesn't say anything about the Samoyedic connection, what it says is that on the FU side these are cognates. If FU experts say they're not, I will accept it, I'm far from this field. Like Arnaud Fournet 16 days ago @ Häkkinen http://www.elisanet.fi/alkupera/Problems_of_phylogenetics.pdf => As an aside, it must be emphasized that there is no indication of any contact between Semitic and PIE, and especially with the Non-Anatolian branch. Now I completely disagree with this part: "The widely known support for the Copper Age steppe homeland comes from linguistic paleontology. Words connected to wheeled vehicle (wheel, convey, axle, thill), secondary products (milk, butter, wool), animal traction (yoke, harness, harness pole), and metals (copper, gold, silver) have been reconstructed in Proto-Indo-European, [most of this is fairy tale ! It must also be emphasized that the Caucasic lexical input is grossly ignored or overlooked.] and they give the upper limit (terminus post quem) for the dispersal of Proto-Indo-European: the language cannot have dispersed before the cultural development had reached the certain level seen in the meanings. (Mallory 1989; Mallory & Adams 2006.) => I've written a detailed survey, which exposes why the Pontico-Caspian theory and its low dating is an archeofable. https://www.academia.edu/42452209/Getting_rid_of_the_Pontic_Caspian_archaeofable Besides, it can be added that: 1. archeology does not support the Pontico-Caspian archeofable, 2. Anatolian genomic analyses do not support intrusion of Steppic features into Anatolian speakers. 3. Where is the substrate of Luwic, if Luwic came from somewhere else? Show words of substratic origin, please. Like Jaakko Häkkinen 15 days ago Arnaud Fournet, I'm aware of your views. Only relevant arguments count. But I will not terrorize this discussion with off-topic anymore. :) https://www.academia.edu/s/010b2ad552#comment_809468 51/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Arnaud Fournet 15 days ago I agree that only relevant arguments count. I will add that it matters as well that the relevant arguments should be true. Like Peter S Piispanen 16 days ago A second suggestion: Proto-Tungusic *tōli 'belt, strap for trousers' (TMS 2 232-233; Vasilevic, G.M. 1958:392) > Proto-Samoyed *tolǝ 'belt' (SW 165). The Tungusic root - attested in Ewenki, Manchu and Udighe - has a correspondence in Proto-Mongolic *telej 'belt for trousers'. The Samoyedic root is attested in Nganasan, Selkup and Kamassian justifying the PS root reconstructed. Any comments or thoughts? Like Mikhail Zhivlov 16 days ago This comparison was already suggested in Anikin & Helimski's book, pp. 91-92. Unlike 1 Peter S Piispanen 15 days ago Ok, good, so it is correct then, but already known. I have yet to dig up that volume again and didn't recall it. Thanks, Mikhail! Like Ante Aikio 17 days ago I think a rather strong case can be made for the Uralic (or, at least, Ugric) origin of Hungarian <oldal> 'side': it could be a reflex of Ugric *aŋti(-lV) 'rib / side', as I've argued in my paper "Notes on the development of some consonant clusters in Hungarian", 2018, p. 87. Like 2 Annotation: Page 5 Alexander Savelyev 15 days ago https://www.academia.edu/s/010b2ad552#comment_809468 52/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu In addition, the idea of Tungusic-Hungarian contact cannot be taken for granted. The Tungusic etymologies for Hungarian words proposed by Futaky and Helimski were considered controversial or directly refuted, e.g., by Kara, Róna-Tas, and recently by de la Fuente. Like 1 Kamil Kartal 16 days ago Hello, I received this invitation but couldn't find suitable time to read the article and your comments yet. If that's possible, please don't close the discussion a few days more, I will join with my comments. Best... Like Peter S Piispanen 16 days ago Hello Kamil, no problem, the discussion is still active for 16 more days. You are also welcome to contact me privately by email regarding this matter if you please! Like 1 Juho Pystynen 19 days ago By the way, it seems like one or both of Khanty sŏŋ ~ Hungarian zug 'nook, corner' might be also connected with this Turkic etymon. Unlike 1 Annotation: Page 3 Peter S Piispanen 19 days ago Why, yes, it does! Excellent catch. Perhaps we are dealing with an independent Turkic borrowing into Ugric here? Like Juho Pystynen 18 days ago Independent of Samoyedic certainly, I wonder even if independent in both Khanty and Hungarian, since this is one of the words where they fail to agree on "common Ugric" *ŋ > *ŋk (> Hu. g). Like Peter S Piispanen 18 days ago So perhaps this was the spreading of some new type of basic architectural knowledge then? Surely the Proto-Uralians could already build cabins because they had words for smoke hole, to fit logs together, door, etc., but you are right in that this appears to be independently borrowing into Khanty and Hungarian, and possibly also into two groups of early Samoyedic populations. https://www.academia.edu/s/010b2ad552#comment_809468 53/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like 1 Robert Lindsay 18 days ago Except again Proto-Turkic and Proto-Samoyedic are not parallel in space-time. Proto-Samoyedic is in the Sayans at 5,300 YBP while Proto-Turkic does not even exist until 3,400 YBP on the Central Kazakhstan border with Russia. Granted there was a trade route from Proto-Samoyedic homeland back and forth to the Proto-Uralic region, trading mostly in precious gems like rubies being mined along the route in the Altai. But when did Proto-Samoyedic break up. It shows up at 5,300 YBP, but when does it split up? Also Pre-Proto-Turkic splits with PT-M Mongolic at 3,500-4,000 YBP, and they migrate from the Khitans to around Nur-Sultan and Omsk 3,400 YBP. This journey, which may have taken some time, would take them right across the PS homeland in the Sayans, that is, if PS still exists at that time and has not yet split up. The migrants could have traded words along the way. If these two cannot be shown to be parallel, all of the Turkic-Samoyedic borrowings suggested in this paper are simply false. If you want to trade words between Proto-Samoyedic and Proto-Turkic, you must do so between PS and Proto-Turko-Mongolic in the Khitans at 5,200 YBP. This calls into question most of the "borrowings" in this paper and suggests that they theory may need a redo where you look for PS borrowings into PT-M instead and you need to find the word in both T and M. That is unless you can make the trade route scenario in paragraph 1 make sense... Like Juho Pystynen 18 days ago This alleged date of 5300 BP for Proto-Samoyedic simply seems to be wrong by a millennium or two. Like 2 Mikhail Zhivlov 17 days ago @Robert: I really envy your having access to time machine. Like 1 Jaakko Häkkinen 17 days ago Robert Lindsay, again your datings seem very peculiar. What is the method behind them? If it is the computational phylogenetics, then I must warn you, that the lexical level is prone to some distorting effects: one cannot reliably date protolanguages merely on the basis of lexical retentions. https://www.academia.edu/s/010b2ad552#comment_809468 54/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu And once again you should see the reality behind the labels: if the labels are too nasty for your eyes, you could just replace them with better labels. Samoyedic-Turkic loanword connection cannot be disproved on the basis of the labels used. You should try to disprove the cognates themselves. Like Robert Lindsay 17 days ago Then the date of 4,200 BP for Proto-Samoyedic is correct. That date sticks in my mind too. The early date for PT, see my paper, "Mutual Intelligibility among the Turkic Languages." See my paper for an explication of that. And I put the homeland at Omsk and Nur-Sultan on the Siberia-Kazakhstan border. The dates of 5200 BP for Proto-Tungusic homeland at the mouth of the Amur and the same date for Proto-Turkic-Mongolic in the Khitans are from Robeets. Of course this is all within an Altaic theory perspective. I have no idea where anti-Altaicists put these homelands or when they date them. I'm not sure what methods Robeets used to date her homelands, but if Altaic exists, 7,000-8,000 YBP seems about right, that's for sure - that is if you look at all the languages separately and then theorize that they were all together once, that is how far you have to push back the proto-language. If you are not going to use phylogenetics (or glottochronology I guess?) what method are you going to use for dating proto-languages? You are saying that we don't need to put PT and PS adjacent in space-time to even suggest these borrowings? I would think that would be mandatory. How can you even posit these borrowings if you cannot put these proto-languages adjacent in space-time? Like Robert Lindsay 17 days ago Mikhail, proto-languages exist and can be dated. Their homelands exist and can be placed. Acting as if such things do not exist and then postulating borrowings ad infinitum across space-time strikes me as reckless. You all going to put some limits on how far apart these languages and Uhrmeits need to be to posit borrowing? Otherwise you can go pretty crazy with borrowing scenarios very far removed in spacetime. You going to put any limits on that at all? Like Mikhail Zhivlov 17 days ago @Robert: Yes, proto-languages and their homelands exist and can in principle be located in space and time. But any such work can be based only on what can be reconstructed for these protolanguages using the comparative method. And our knowledge of different protolanguages is very uneven. Now, you pretend to know the ultimate truth in every case and you seem to, e.g., exclude Samoyed-Yukaghir contacts because you "know" where the speakers of these languages lived. But what are the facts this knowledge is based on? Actually, it's the other way round: we know that these contacts took place https://www.academia.edu/s/010b2ad552#comment_809468 55/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu because otherwise we can't explain why certain Yukaghir words are specifically similar to Samoyed words, i.e. they show the results of known sound changes from Proto-Uralic to Proto-Samoyed. Such words cannot be inherited from a Uralo-Yukaghir protolanguage irrespective of whether such a language existed or not. Therefore, these words are loans, and we must postulate that speakers of Yukaghir and Samoyed were in contact. Like 6 Arnaud Fournet 16 days ago Besides, proto-languages are not always, and often are not, linguistic objets where everything has the same dating. Some features are older than others. And the bigger the family, the bigger the issue. So, in all cases, claiming to date proto-languages is highly dangerous. Like Robert Lindsay 16 days ago Where are there no existing Samoyedic languages adjacent to Yukaghir except Nganasan in the far west? Did they all pack up and leave? Why are all of the "borrowings" in core vocabulary with nothing whatsoever in cultural borrowings? There are no shared cultural terms between Samoyedic and Yukaghir, zero. And there should be lots. Instead lots of shared core vocabulary. That goes against common sense. And why couldn't Samoyedic-Yukaghir have been part of some genetic unity based on a mutual Uralic split? As in Eastern Uralo-Yukaghir or Far Eastern Uralo-Yukaghir? The Samoyedic languages have been traced back through time to their homeland in the Sayans 4,200 YBP. At no time has anyone ever postulated them as far to the east as Yukaghir. We have traced and reconstructed the splits of Southwest Samoyedic, South Samoyedic, and North Samoyedic. Seems we can trace this language family back through time pretty well. When can we ever show it on the Tamyr Peninsula? We not only know the homelands of old language families and the positions of their current speakers, but in many cases we can trace their movements through time, such as Proto-Indo-Hittite speakers moving through the Caucasus and picking up North Caucasian substratum along the way over to the area north of Sebastopol in the Kurgan Culture to core PIE. We can trace the movements of Proto-Indo-Iranian speakers in rather precisely time as they move south via archeology. So this idea that we only know where languages exist now and maybe 1,000-2000 years back and then where the homeland was (some people on this thread seem to be disputing that we can infer Uhrmeits at all), but we have no idea where they were in between seems dubious and rife for theoretical exploitation. https://www.academia.edu/s/010b2ad552#comment_809468 56/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu It seems we can trace the movements of proto-languages and their descendant splits through time with accuracy. A lot of times we can spot that a proto-language traveled through some territory by the loans that they picked up on the way. I'd think we should exercise extreme caution in positing borrowings going back thousands and thousands of years all the way to proto-languages. Of course note recent borrowings or in the past 2,000 years. But we are talking about proto-languages whose reconstructions themselves are vague and fraught with error. And we have 6,000 year time blockades after which nothing cognate remains. Except somehow these miraculous borrowings violate this boundary, as they abound everywhere we look. Extrapolating back to see borrowings 4-8,000 years ago would seem to be something we should proceed towards with the utmost of caution. Like 1 Arnaud Fournet 16 days ago The Rule is that irregular patterns of sound correspondences indicate probable borrowings, while regular patterns combined with canonical morphology indicate probable inheritance. I do not believe in the 6 kyr limit BS. Unlike 2 Juho Pystynen 16 days ago It does seem to be a good possibility that some unattested Samoyedic groups did "pack up and leave" from eastern Siberia under the Yakut and Evenki expansions. Already the Seima-Turbino phenomenon that likely brough Samoyedic to central Siberia reaches further east still to Lake Baikal (though further archeological support would be good to have too). This may not even require permanent settlement: we know that before being pushed back the Russian expansion, the Evenki had trade routes stretching all the way to the (north)west of the Urals. Like Peter Kitson 16 days ago Robert Lindsay: by Uhrmeit do you mean Urheimat? Your paragraph beginning “We not only know” assumes as fact half a dozen propositions that are (to put it mildly) very debatable. Like 1 Arnaud Fournet 16 days ago https://www.academia.edu/s/010b2ad552#comment_809468 57/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu @ Häkkinen http://www.elisanet.fi/alkupera/Problems_of_phylogenetics.pdf => (this is a bit off topic, but I suppose it will be accepted.) I roughly agree with this in your paper: "Based on all the relevant arguments, Proto-Uralic is located in the taiga zone in the Volga-Ural region, from where its expansion began only ca. 2000 BC (Kallio 2006; Häkkinen 2009)," [this might be a bit late though] "but Pre-Proto-Uralic seems to have been spoken in Southern Siberia, north from the Sayan Mountains, where it shared typological developments with the proto-languages of the Altaic type (Janhunen 2001; 2007)" [yes, that sounds extremely plausible and reasonable Note that it tends to disprove a close relationship between PIE and PU (smiley)] "and donated loanwords to Pre-Proto-Yukaghir (Häkkinen 2012b)." The Aryan developments must have taken place in the vicinity of Proto-Uralic, that is in the North Caspian Steppes, as extensively argued by Carpelan & Parpola (2001)". [I suppose "Aryan" means Indo-Iranian]. Like 1 Alexander Savelyev 19 days ago Why "surely borrowed"? Like 1 Annotation: Page 7 Peter S Piispanen 19 days ago I am basing the assumption that the Chuvash word for 'tinder' is a borrowing because of the extremely limited attestation: AFAICT there are NO Turkic cognates to be found for this word at all, and thus I assume that it was borrowed from Mongolic. Like Alexander Savelyev 18 days ago Because of the binary structure of the family, the lack of cognates in Common Turkic does not prove anything. One single event, i.e. the loss of a root from Proto-Common Turkic, would lead to an isolated status of the Chuvash lexeme in Turkic. There are of course numerous Chuvash roots with parallels in Mongolic or Tungusic, but not in Common Turkic. Like 1 Peter S Piispanen 18 days ago Yes, this is of course a possibility, but given that there are very similar words in Mongolic, and many other Mongolic borrowings in Chuvash are already known, it seems to be the most plausible explanation that this Chuvash word too is a Mongolic borrowing. https://www.academia.edu/s/010b2ad552#comment_809468 58/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu As you state, there are many Chuvash roots with correspondences in Tungusic and/or Mongolic, and these are extremely interesting. Not all can be borrowings, and some are probably, as you suggest, ProtoTurkic remnants now entirely lost in East Turkic. In any event, this field is really something that should be studied properly, summarized and evaluated phonologically. As I have suggested before, you, Alexander, should publish more about Chuvash, and its interconnectedness with East Turkic, Mari, Mongolic and Tungusic - every time we have had a Session, you have presented new findings. Like Robert Lindsay 18 days ago It's not really possible for there to be Chuvash borrowings from Mongolic or Tungusic anyway. Let's look at space-time. Proto-Bulgaric homeland is probably somewhere around Samara and the southern end of the Urals? Time depth? 2,500-3,000 YBP. Meanwhile, Proto-Tungusic is hanging out at the Amur mouth and Proto-Mongolic is in the Khitan Range. They're not near enough to each other for any of this borrowing to occur. This is the problem with a lot of mass borrowings of core vocabulary anti-Altaic argument. Aside from the fact that nothing like this has been documented anywhere, the proto-languages were typically not that close to each other in spacetime to be able to trade words. Ditto with Aiko's anti-Uralo-Yukaghir argument which has Proto-Samoyedic in the Sayans at 5,300 YBP trading words with Yukaghirs on the Tamyr Peninsula 2,000 miles away. Because the theory and the facts don't make sense, instead of going back and fixing the theory, Aiko says the facts must be wrong, that is, the facts are not the facts. So he ends up creating "alternate facts" to support his theory, which obviously must be 100% true. This bassakwards stuff is all through these anti-long range mass borrowing arguments. Granted the longrangers haven't proved their stuff but their theories are at least plausible, unlike anti-long-rangers who don't even have plausible theories. 1st and 2nd person pronouns got borrowed 450 times (!) in the Americas? Really? Keep in mind that this is the standard theory everyone believers. The "kook theory" is Amerind and is actually at least logical. Sorry for derail into long-range stuff. Unlike 1 Juho Pystynen 18 days ago Does anyone propose *Proto*-Bulgharic borrowings into Mongolic? The usual scenario assumed seems to be borrowings from some later and more eastern offshoot of Bulgharic. Like https://www.academia.edu/s/010b2ad552#comment_809468 59/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Robert Lindsay 17 days ago Who says there was an eastern offshoot of Bulgaric? is there any evidence for it? And how does a borrowing into an eastern offshoot of Bulgaric end up in Chuvash so far away? Unlike 1 Arnaud Fournet 16 days ago You know, things and words can travel very far away. Blench has shown that the New Guinean word for "banana" ultimately reached Medieval Latin, after thousands of km of journey thru southern Eurasia. Now, we can see that Germanic Sleipnir is a horse with eight legs, which is a feature of Ugric cosmic Elk, with six legs. Unlike 1 Juho Pystynen 19 days ago Even better actually: both Moksha ľäpä and Khanty lewət etc. < *lä̆pət point to *läppV- rather than *leppV-. Mari *lewə 'lukewarm' is probably rather from *lämpə 'warm' as argued in recent times by Metsäranta, and Finnic *leppedä primarily means 'kind, pleasant, appropriate' rather than physical 'soft' and might not be related either. Unlike 1 Annotation: Page 3 Peter S Piispanen 19 days ago Oh, so we could then actually have the regularly changed PU *läppV > PS *japV? Which then in turn could mean that these Samoyedic words are actually of Uralic etymology anyway and not at all borrowed from Turkic! Like Juho Pystynen 18 days ago Yes, plus also here too this should be PS *jäpɜrkɜ = pre-05 *je¹pɜrkɜ. On further thinking even *-r might reflect the PU adjective suffix *-ətA as continued in the Khanty word (though more regularly I'd expect *-rV). Unlike 1 Ante Aikio 16 days ago https://www.academia.edu/s/010b2ad552#comment_809468 60/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu I wonder if the ending *-rkV might be cognate with the North Samoyed suffix *-rǝkA, which forms denominal similative adjectives (*X-rǝkA = 'X-like, resembling X, similar to X'). Like 2 Ante Aikio 17 days ago I cannot understand why this should be considered a "near-hit". The suggestion that a prothetic *j- was irregularly added in Samoyed is an arbitrary ad hoc hypothesis. Since the vowel does not match either, the only "match" we are left is the medial consonant (a rhotic). Like 1 Annotation: Page 8 Mikhail Zhivlov 17 days ago Also, UEW compares PSmd *jürə- to Udmurt ji̮ romi̮ - ‘get lost’, Mansi *jɔrəɣl- ‘to forget’ and Khanty *jur‘to forget’. Semantically the comparison in UEW is quite straightforward. As for phonology, PSmd *ü is indeed irregular, but anyway this comparison is better than the comparison of the Samoyed word with Turkic. Like Peter S Piispanen 17 days ago Surely nobody has suggested that the prothetic *j- in Samoyed is irregular? Instead it is very regular given the right phonological conditions. The UEW comparison is interesting - and the point of the "near-hit" between the Samoyed and Turkic words is exactly that the comparison is NOT valid. Then, the UEW here provides good evidence for the Samoyedic root being Uralic after all. Like Ante Aikio 16 days ago I am not aware of any regular prothesis of *j- in Proto-Samoyed. In some Samoyed languages we find a protethic *j- that was added before word-initial *e- or *i-, e.g. *elä- 'live' > *jelä-, *iǝ 'belt' > *jiǝ. However, these developments clearly occurred after the breaking up of Proto-Samoyed, because only a part of the languages show a reflex of this prothetic *j-. Moreover, I am not aware of any examples of prothetic *j- before the vowel *ü. Like 1 Ante Aikio 17 days ago https://www.academia.edu/s/010b2ad552#comment_809468 61/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu The Selkup form in Alatalo's dictionary points to Proto-Selkup *ćapǝ/urka, and the first-syllable vowel *-aimplies Proto-Samoyed *-ǝ̑-, so this invalidates the reconstruction earlier proposed by Helimski. Since ProtoSamoyed first-syllable *ǝ̑ goes regularly back to Proto-Uralic *u, the comparison to *leppV- (or the like) cannot be correct. Like 2 Annotation: Page 3 Peter S Piispanen 17 days ago In other words: the Samoyedic form is too different to be related to the Turkic root at hand, the entry therefore will be deleted as the suggestion is no longer a valid hypothesis. Like Ante Aikio 17 days ago It is contradictory to assume that a Hungarian word with initial b- (beteg) could go back to the Ugric stage. Like 1 Annotation: Page 5 Peter S Piispanen 17 days ago In the case of beteg this is true. It should be a later borrowing, if it indeed is a borrowing, and we probably won't find it in Khanty and Mansi. Like Eugen Hill 19 days ago Why are some Russian titles transliterated and others not? Like Annotation: Page 11 Peter S Piispanen 19 days ago Merely because I did not yet take the time to do that - I wanted all the references in place, but I was fully aware that some of them needed more work (formatting, spelling, publisher info, etc.), as your corrections indeed do show. I will transliterate all Russian titles in the updated version for clarity and consistency. Like Robert Lindsay 18 days ago Actually that is fairly common in a lot of papers that quote Russian linguistic materials if I am not mistaken. A lot are left in Cyrillic and but others are transliterated. Unlike https://www.academia.edu/s/010b2ad552#comment_809468 62/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu 1 Geoffrey Caveney 19 days ago This is just a general comment: It seems to me that the field is in great need of an up-to-date and comprehensive etymological dictionary of the Samoyed languages. I suppose that Janhunen 1977 is the closest thing to such a work, but it is now almost a half century old. Many more recent works such as Aikio's seem to focus mainly on those Samoyed roots that can be shown to have Uralic etymologies. But I would not know where to look, for example, to find a comprehensive account of Samoyed roots that are *not* found elsewhere in Uralic. Unlike 3 Peter S Piispanen 19 days ago Absolutely so, Geoffrey. Janhunen's old stuff is the best thing we have thus far, but work in improvement has been done, additional roots presented in scattered publications, and, yes, Aikio's current UED project will hopefully upgrade materials with Uralic connections. But you are correct: where are those non-Uralic Samoyed roots? Again, only in scattered publications (and some in Pystynen's WIP database as well). Actually, there are quite a lot of good dictionaries on various Samoyedic languages, and it would certainly be possible to create new reconstructed roots, but do note that the SW is fairly extensive. Extremely detailed studies were done with Mator, for example, and those data are still referenced heavily in Samoyedology. Still, reconstructing new PS roots it is not as extreme or hopeless as it is in Yukaghir studies - where all available resources seem to have been exhausted already in Nikolaeva's Historical Dictionary of Yukaghir - I could recreate but only very few new Late Proto-Yukaghir roots in that paper I published in Ural-Altaic Studies. So, if time and the will is at hand, it would be possible to create Samoyedic-only roots by classic bibliographic studies. Some of such would then later likely also be found in a few other Uralic branches, and voilá, we would have new PU roots as well. Naturally, this would be a ripe field also for finding additional borrowings into Samoyedic, and actually, in the long run, tell us quite a lot about the ancient language contacts of central and east Siberia. Samoyedic phonology is not simple by any means, and yet it is systematic enough to allow for such an endeavour, which might also additionally show us how the various Samoyedic languages are interrelated in greater detail. So yes, this would be a fascinating and rewarding project all over! Like 2 Geoffrey Caveney 18 days ago Thank you for this very informative reply, Peter. I will note here in passing another curiosity: In a recent survey study exploring the possible location of the Proto-Uralic homeland, Johanna Nichols rather surprisingly to my mind raised as a serious possibility the idea that the Samoyed-only lexicon could have been the original Proto-Uralic lexicon, and *all* Uralic lexicon not found in Samoyed could represent a later layer of "Proto-Finno-Ugric" borrowing from outside this conception of "Proto-Uralic"!! Naturally I https://www.academia.edu/s/010b2ad552#comment_809468 63/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu reject this hypothesis, and I suspect almost all serious specialists in Uralic historical linguistics will reject it as well. Nichols' suggestion seems to be based, among other things, on the now outdated concept of a distinct "Proto-Finno-Ugric" stage, rather than the simple division of Uralic into 9 separate branches, one of which is Samoyed. I suspect this idea may have been an example of taking the argument that "a lot more lexicon is borrowed than historical linguistics has traditionally realized" rather too far. Like 3 Peter S Piispanen 18 days ago Fascinating, Geoffrey! I had not heard about this new idea by Nichols. I do not believe it either, but won't dismiss such ideas automatically out of hand either. And I certainly don't believe that all the major branches of the Uralic languages split off at the same time from one original source either: there exists some sort of language tree even for Uralic, which appears to be more complicated than expected due to extensive language contacts. Once upon a time I took the time to really read up on population genetic studies made on various Uralic populations (noteworthy at that time perhaps some 60 papers or so - and that is a field that gets updated every other year it seems now). I plotted out various haplogroups and convalescence times on a huge map, and could therewith trace what appears to be population migrations. Some facts if I remember them correctly without digging up all of those hundreds of pages of notes: There are some very clear-cut population genetic splits between certain Uralic populations. This suggests very strongly that some groups took on certain forms of Uralic languages, switching away their own original language, and then ran with their new Uralic form until they were quite unique forms in the end (i.e. today). The greatest population genetic split is found between what is today termed West Uralic and East Uralic (Samoyeds and Ugric) folks. For example, IIRC, Finns are an Uralic genotype mixed with Scandinavians, but the Saami are not mixed very extensively at all with Scandinavians. The Inari Saami, however, specifically, are genetically different from all other Saamic populations and they also do have certain peculiarities in their language, which then could be explained by the fact that they are an original aboriginal population of northern Finland that switched to a Saamic language from something else. There is another great population split isolating the Ugric folks from all the others, and the Permic folks are divided 50/50 by genes from the east and from the west. Hungarians share genes with the other populations of those great "mercenary" hoardes of ancient times: Persians and populations much further eastwards. The Saami share some genes with Volga Urals only some 2000 years old or less, and also with some old Scandinavian populations, just as if the Saami had arrived up in the north by groups both from Finland and from Sweden/Norway, intermixing with a fairly unique aboriginal group. Some Samoyedic groups share genetics with the Yukaghirs. Certain other traits are shared between some Altaic populations, running through some Uralic groups all the way to the Saami and Finns, but these links are very old. I have seen a few studies in the recent years discussing such matters, but nobody has made a comprehensive survey of all such population genetic studies yet. It is a complicated field, and requires certain unique skillsets for a linguist. I have my old notes on these connections handy somewhere, but they would need to be updated with data from the last couple of years. I traced the Yukaghir populations back some twenty thousand years or so, and various Uralic https://www.academia.edu/s/010b2ad552#comment_809468 64/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu populations back to the last Ice Age. The whole point of this rant is this: there are enormous findings to be made by the combined efforts of population genetic studies, archaeology and historical linguistics. Like 2 Juho Pystynen 18 days ago As he mentions e.g. in his 2015 SEC paper, Janhunen's original plan indeed was to follow up SW with a second volume with common North Samoyedic and South Samoyedic lexica. This would be still interesting today, all the more since now it seems that neither of these is much of clear primary subgroup. E.g. already Nenets–Nganasan cognates would be basically Proto-Samoyedic. Last I heard at the 7th Samoyedology Conference in Tartu, at least one North Samoyedic project is currently in the works by a new Russian team. Most languages definitely have room for basic research too, maybe Selkup most of all, which has quite a lot of diversity scattered across dozens of varieties and sources. Alatalo's Sölkupisches Wörterbuch is still a good start, aside from organizing Donner's materials it actually also mentions several evidently new comparisons esp. with Kamassian. Unlike 3 Peter Kitson 19 days ago There’s something wrong with “convalescence”. I think you must have meant “convergence”, but even that’s a bit odd, implying a perspective backward along the graph from the present. Would it not be better to write “divergence”, respecting the direction of events in time? Like Annotation: Page 9 Peter S Piispanen 19 days ago Actually it is 'convalescence' as it is used this way in population genetic studies. Odd it may seem, but that's the common useage. Indeed, it implies a perspective backwards to a time when two populations started sharing a common genetic trait (or, rather, haplotype). So if we know the convalescence time for a given haplogroup, which can be calculated by known rates of mutation, then we will find out a point in time after which two populations must have intermixed, and from that point on started sharing this haplogroup. This is potentially, if done right, a very powerful method for dating "at the latest" epochs when two populations met and intermixed genetics. It seems to rarely be used this way, but thinking about it - if two populations share a common, old gene in a large part of their respective populace, then by default they must once upon a time have orinated from a common ancestor or being transmitted from one group to the other in older times Like 1 Peter Kitson 18 days ago You and I have evidently read different population geneticists! https://www.academia.edu/s/010b2ad552#comment_809468 65/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like 1 Marcel Erdal 19 days ago Just one remark on Turkic soŋ and sɨŋarsuk: You write that "there may have been some early variation in the root vowel among the subsequent Turkic languages", but there is never variation in the vowels of Early Turkic nouns (there is variation only in the vowels of pronouns). Therefore soŋ and sɨŋarsuk (or indeed sɨŋar, the source of the latter) can not be connected. I find everything else in your paper to be fine from the Turcological point of view. Marcel Erdal Unlike 5 Peter S Piispanen 19 days ago Thanks, Marcel, that is a good an important point and I will correct this detail. It was from Janhunen's work, and there he handled exactly pronouns. Thanks for the vote of confidence! Like 1 Juho Pystynen 19 days ago You seem to have double-corrected this: it's *ket¹- in SW = *kät- in Helimski's system. *k > *s/š is also probably post-Proto-Samoyedic even in cases like this, given discrepansies like *ka- > Nenets & Enets *kä- > *sä-, vs. > Nganasan ka-, Selkup & Kamassian qā-. Unlike 2 Annotation: Page 2 Peter S Piispanen 19 days ago Hah, you are right - I should have remembered this from our correspondence! It is properly reconstructed according to our current state-of-the-art of Samoyedic phonology as *kät-, and that exactly was the possible obstacle to connecting it to the Turkic *kat-. So, the front- vs. back-vowel problem and discussion is still valid. That whole *k > *s/š mess appears far from clear to me in the literature. It is an unusual sound change in the first place, albeit possible, and we clearly do not know all the conditioning factors for it. If it is postProto-Samoyedic that would imply that different Samoyedic groups may have used different conditioning factors for this change. Still, because practically all did it we could invoke heavily linguistic contacts (less likely) or assuming a change at the Proto-Samoyed stage anyway. I would be very interested in your view on the chronology and possible conditioning factors for this change, Juho! I should now actually separate the languages into the groups: Enets-Nenets-Yurats, Mator, Kamassianhttps://www.academia.edu/s/010b2ad552#comment_809468 66/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Selkup and Nganasan, respectively, instead of the mere North Samoyed and South Samoyed, because this seems more accurate and correct for a number of reasons. Like Peter S Piispanen 19 days ago Also, wouldn't Janhunen's *ket¹- suggest that his reconstructed root is either *ket- och *kec-? Although I see little motivation or possibility for the latter! Like Juho Pystynen 19 days ago I don't know of any special studies in recent times on the palatalization of *k, but Janhunen's data in SW shows this to be quite regular but depending on the following vowel. By Helimski's vowel system we have: – Nganasan: *k > s before *i *e *ä – Selkup–Kamass: *k > š before *i *e *ü *ä – Nenets–Enets: *k > ⁽*⁾s before *i *e *ü *ä *a (or rather, after the chainshift *e *ä *a > *i *e *ä) (mostly thus also in overviews: Sammallahti 1988: 497–8 and in Mikola 2004) Plausibly the early common Samoyedic change was just *k > *ć, with *ć > ś > š ~ s as later independent decay, but I don't know of any evidence that would help in dating these intermediate steps. At minimum this *ć > š needs to be younger than the development of original Uralic *ć > *ś to Selkup & Kamass /s/, though. Donner's original 1920 article on this (in JdSFU 37) mentions also a few cases in apparently post-ProtoSamoyedic loanwords from Turkic/Mongolic (sorry, don't have them on hand), so reasons to think of this as not quite Proto-Samoyedic have been there all along. Unlike 3 Peter S Piispanen 19 days ago Thanks, Juho, for another - as usual - supreme analysis of a complex phonological process. I will now of course include these conditioning factors into the paper for the sake of the reader (and myself!) - thanks also, then, in particular for giving a reference for them. I agree that the first change might have been *k > *ć (which is, what, a spirantization in linguistic terms?), followed by what you call later independent decay (same thing, that is the decay, happened in Yukaghir). I fail, however, to see why such a change would occur and what could possibly motivate it. The whole branch of Samoyedic is indeed full of super-odd changes like this; due to substrate influences perhaps? Further, why would this change occur in parallel in post-Proto-Samoyed in different language groups with almost the same conditioning factors? Such a situation would usually indicate that the change should have been early enough to have spread out to all of them - but then you actually present data to the contrary as well. Very complex indeed! https://www.academia.edu/s/010b2ad552#comment_809468 67/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Mikhail Zhivlov 19 days ago The change "velar stop > affricate > sibilant fricative" before front vowels is one of the most typologically trivial and widespread changes in languages of the world. It certainly can happen without any substrate influence. Like 2 Juho Pystynen 19 days ago PKh *kăŋkar is indeed usually derived from PU *kaŋərə 'bow, 'bend' (> Fi. kaari etc.) Unlike 1 Annotation: Page 6 Peter S Piispanen 19 days ago Ah, yes, and I seem to recall having seen the change of *-ŋa- > Proto-Khanty *-ŋka- discussed in one of Aikio's papers as well. Ok, good, then the Ob-Ugric forms are not related to the Samoyedic forms, and the latter could then be Turkic borrowings as suggested. Like Juho Pystynen 19 days ago Another double-correction: *er- in SW, *är- by Helimski. Unlike 1 Annotation: Page 4 Peter S Piispanen 19 days ago Ah, yes, I see that you are right at that. I must have gone through the same data twice, resulting in something of an odd, erroneous vowel shift! So, there still remains a discrepancy in the vocalism of this suggestion. The forms will naturally be corrected. Like Peter S Piispanen 19 days ago should have read "one of the other two micro-Altaic", but even this is no longer applicable. I should rewrite this section because I do not think that the Proto-Tungusic root given in the TMS exists at all, and instead believe that the Manchu and Ulcha words are Koreanic borrowings - which will be explained in a footnote. Thus, the only suggestion here is a Turkic loanword etymology for the Samoyedic words. Like Annotation: Page 4 https://www.academia.edu/s/010b2ad552#comment_809468 68/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Eugen Hill 19 days ago непервых? Like Annotation: Page 10 Peter S Piispanen 19 days ago Ah, yes, as written it looks a bit ungrammatical to me - but it was actually непервый in my source too. I need to check it up in the original (and also hope for some input from a native Russian speaker here!). Like Juho Pystynen 19 days ago A few of these typos were I think due to me rather than Peter, thanks for pointing them out. Like Peter S Piispanen 19 days ago Perhaps, but in any case the responsibility of checking all of them falls upon me for using them in this work! And good to have you on-board, Juho! Like Eugen Hill 19 days ago <Čuvašsko-> Like Annotation: Page 11 Peter S Piispanen 19 days ago That would be the literal transcription, yes, and I probably should go with that (along the original Cyrillic given). Like Eugen Hill 19 days ago лабиализованных Like Annotation: Page 10 https://www.academia.edu/s/010b2ad552#comment_809468 69/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter S Piispanen 19 days ago Yes, only that makes sense - will correct. Like Eugen Hill 19 days ago Wortschatz? Unlike 1 Annotation: Page 1 Peter S Piispanen 19 days ago Yes, that should indeed be Wortschatz, thanks. I have apparently read too many Wörterbücher (=dictionaries) and wrote that by reflex, but this is instead Wortschatz (=vocabulary). Thanks for all of these reference corrections, Eugen, much helpful! Like Eugen Hill 19 days ago Not сравнительно-историческая? Like Annotation: Page 11 Peter S Piispanen 19 days ago Absolutely, that is the only thing making any sense! Like Eugen Hill 19 days ago Typo: probably <jazyk-osnova> Like Annotation: Page 11 Peter S Piispanen 19 days ago Yes, <jazyk-osnova> it is.. will correct this too. Like https://www.academia.edu/s/010b2ad552#comment_809468 70/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Eugen Hill 19 days ago Typo: <Etymologien> Like Annotation: Page 11 Peter S Piispanen 19 days ago Yes, all German nouns come with a capital initial letter... will correct. Like Eugen Hill 19 days ago Typo: <etymologischen> Like Annotation: Page 12 Peter S Piispanen 19 days ago Yes, wrong case used - will correct to <etymologischen> Like Eugen Hill 19 days ago It should be <Gosudarstvennoe izdatel’stvo inostrannyx i nacional’nyx slovarej> Like Annotation: Page 12 Peter S Piispanen 19 days ago Ah, yes, of course - typo - I probably should write it out in Cyrillic as well for maximum clarity. Thanks, Eugen. Like András Róna Tas 19 days ago Dear Peter, it is essential what you reconstruct for Proto Turkic. It is important among others for the phonology and the chronology. Please read what I wrote on pp. 1107-1114 in West Old Turkic (2011). Otherwise I found many of your etymologies as acceptable, to which you will find backgound material in the same work.. your https://www.academia.edu/s/010b2ad552#comment_809468 71/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu András Unlike 3 Peter S Piispanen 19 days ago Dear András, thanks you for this suggestion. Yes, I should pinpoint the given reconstructions clearer as they indeed do differ slightly in some of the sources, and that too should be mentioned. Naturally, I will re-read the part you refer too; such background work is vital and I am sure this too will come in handy here, thanks again. Indeed, I am hoping to improve some of these Proto-Turkic reconstructions if only a Chuvash cognate can be found (like was done in the first part of this paper series; I know that, for example, our colleague A. Savelyev has a superior knack for this), all too often lacking for many of them, meaning they may not be proper Proto-Turkic reconstructions at all, but merely, perhaps Common Turkic, or East Turkic only. Of course, knowledge of West Old Turkic is of utmost importance when discussing Turkic borrowings made into Samoyedic, Ugric, etc.! Like 2 Eugen Hill 19 days ago Probably <finnischugrisch-samojedischen> Like Annotation: Page 11 Arnaud Fournet 20 days ago I think Vogul-Mansi also has a word nyawarak with that meaning. Like Annotation: Page 3 Peter S Piispanen 20 days ago Hmm... in that case a reference would be great. If so, that surely would have an interesting string of consonants in parallel with the Samoyedic root. The closest I find are Khanty ńämǝk (V) 'soft', ńȧmǝk (DN) 'warm weather, thawing weather', ńamǝk (Kaz.) 'light; mild' (, borrowed as Selkup ńami̮ k 'weich')'; Mansi ńɔ̄̈mkǝm (KU KM) 'soft', ńāmėk (N) 'soft'. Like Arnaud Fournet 19 days ago I think you should be able to find the word in a lexicon of Vogul-Mansi. https://www.academia.edu/s/010b2ad552#comment_809468 72/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Like Mikhail Zhivlov 20 days ago What do you mean by Proto-Samoyed *t'? Unlike 1 Annotation: Page 7 Peter S Piispanen 19 days ago Good question - I could not include the right sign for this in google doc where I wrote this draft, but will correct it in MS Word when I get the opportunity to adjust this in the future. It reflects Janhunen's *t (plus 1 in superscript) - which, in practice means either *c- or *t-. I should make this much clearer of course... good catch. As it reads now it looks like as if it is a palatalized t, and this is not necessarily the case at all. Like Mikhail Zhivlov 20 days ago There is a whole monograph by Anikin and Helimski dedicated to Tungusic-Samoyed lexical contacts. Unlike 2 Annotation: Page 5 Peter S Piispanen 20 days ago Good to have you on-board, Mikhail! Ah, yes, and that's a good one too - of course I need to include it among the essential reading materials... thanks! Like Arnaud Fournet 20 days ago PU *telpe is allowed, or is it not ? Like Annotation: Page 7 Peter S Piispanen 20 days ago Yes, PU *-lp- was allowed but *-lb- was not and that is what we have here. Also, both *-lp- and *-lbwould be expectedly simplified into Samoyedic *-p- so the suggestion should work out well both phonologically and semantically. Like https://www.academia.edu/s/010b2ad552#comment_809468 73/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Arnaud Fournet 20 days ago What about a connection with PU *puwi "tree" ? Like Annotation: Page 6 Peter S Piispanen 20 days ago PU *puwi 'tree' already has known cognates throughout Samoyedic. Therein, the *-w- is expectedly eradicated resulting in words monosyllables, on occasion with the vowel long in other other languages the vowel a short one. The Proto-Samoyed root was no doubt monosyllabic. While it may be tempting to compare these, I believe my suggestion hits closer to spot so to say. Like Arnaud Fournet 20 days ago Semantics is extremely lax ! Unlike 1 Annotation: Page 4 Peter S Piispanen 20 days ago Extremely lax may be an overstatement: surely there is a connection between 'to scrape' > 'to pick up', and 'to scoop', and also from the ultimately Koreanic etymology I suggested from an original 'to wipe clean' > 'to scrape'. Like Arnaud Fournet 20 days ago That kind of words looks Nostratic, but I've not checked in Bomhard. Like Annotation: Page 2 Peter S Piispanen 20 days ago Good Arnaud, thanks for these quick notes! Well, it looks like as if the root was originally created onomatopoetically, but was then suffixed in different ways in different language groups. The EDAL appears to think that this is a secure Altaic root, but I do not intend to go into the whole Altaic debate in this paper. IIRC, this root has even been borrowed elsewhere, such as into Yukaghir. Indeed, the whole situation might be more complex than one initially believes - the word is found in identical form for example in English, so it is difficult to know what are locally created and what was borrowed. Like https://www.academia.edu/s/010b2ad552#comment_809468 74/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu Peter S Piispanen 20 days ago Typo: should of course read Manchuria. Like Annotation: Page 6 Peter S Piispanen 20 days ago This is the etymological suggestion for the Manchu/Ulcha words, actually not included at all in this draft paper: Proto-Korean *sìs-, *s ś ‘to wash; to cleanse, to wipe off’ (Nam Kwang U. 1960:322, 328) > Manchu siša- > Ulcha sịsa-. This Proto-Korean root was based on the attestation as Modern Korean: s:it- [s:is-] [ 씻어] ‘to wash to cleanse, to wipe off’ (Martin, S.E., Yang Ha Lee & SungUn Chang 1967:1065; Middle Korean sìs- ‘to wash’, s ́s- ‘to cleanse, to wipe off’, but the significance of this in comparison to the Turkic root for ‘to scoop’ mentioned in this paper, if any at all, is not clear. We may attempt to instead connect the Tungusic words to this Korean root. Could, therefore, the Manchu word at hand, siša- ‘рыться (в земле); точить = to dig (in the ground); to whet’, and, therefrom by extension (i.e. subsequent borrowing), the Ulcha form, sịsa- ‘скоблить, скрести = to scrape, to scratch’, be ultimately of Korean origin and therewith be given loanword etymologies as such? Both the vocalism and consonantism as well as semantics do suggest that this is a historically sound case. Note: This suggestion was removed from the draft at an early stage (but might return in the form of a footnote) to make it into a paper of its own on "Koreanic loanword etymologies for Manchu", which includes no less than ten suggestions, and which might be presented as a draft paper of its own in case of any interest! Like 1 Annotation: Page 4 Peter S Piispanen 20 days ago Dear colleagues and other interested parties, Please find for your perusal my recent draft paper on additional Turkic (and Tungusic) lexical borrowings into Samoyedic! This second part, and likely last part, is the continuation of part one, which was printed in 2018, and I hope that it offers new acceptable loanword etymologies for the discussed PS roots. The formatting is a bit off in places because I had to write this directly into a google doc on my Chromebook as I currently, and sadly, do not own a serviceable computer with MS Word. But research must go on! Your comments, insights and suggestions are most welcome - indeed, I expect at least a few of the suggestions to be shot down successfully. So, without further ado, welcome to my new Draft Paper Session and happy reading! Like 2 https://www.academia.edu/s/010b2ad552#comment_809468 75/76 2021-04-23 Discussion: Turkic lexical borrowings in Samoyed, pt.2 - Academia.edu https://www.academia.edu/s/010b2ad552#comment_809468 76/76