Of precious swords and Old Sinitic reconstructions, part 3
« previous post | next post »
Previous posts in the series:
"Of precious swords and Old Sinitic reconstructions " (3/8/16)
"Of precious swords and Old Sinitic reconstructions, part 2 " (3/12/16)
The following post is not about a sword or other type of weapon per se, but in terms of its ancient Eurasian outlook, it arguably belongs in the series:
"Of felt hats, feathers, macaroni, and weasels " (3/13/16)
Today's offering was suggested by this note from Ron Kim, commenting on the first post in the series:
This would hardly be a unique case of a so-called Wanderwort for a word denoting useful weaponry: one thinks at once of the 'ax'-word, which in various forms, mainly the metathesized pVrVt and tVpVr, has spread over much of Eurasia. The p-initial variant is attested in Greek pelekús, Old Indo-Aryan paraśú-, Old Persian paraθu-, and Ossetic færæt, Toch. B peret, A porat (probably a very early borrowing from Iranian; see my first ever published paper in Tocharian and Indo-European Studies [TIES] 8 [1999]), and outside IE in Akkadian parašu- (I think); the t-initial variant underlies Russian topor and many other forms in Finnic languages, if I recall correctly.
As for this word for 'sword', am I not right that the loanwords in Balto-Slavic and Ostseefinnisch point to an early borrowing from Germanic, which suggests the possibility that a certain kind of technologically innovative sword may be involved? Early (East) Germanic borrowings into Balto-Slavic were already investigated in Neogrammarian times, and I think of them whenever I use words like 'bread' or 'kettle'.
Asko Parpola remarks:
Topór is the Russian word for ax. It has been borrowed into Finnish in the form tappara. In Russian the accentless first o is pronounced as a.
In Chinese, we have at least three words for ax: fǔ 斧, qī 戚, and yuè 鉞. Even before considering the latter two and reflecting upon what kind of ax they might be, I immediately thought that fǔ 斧 is very likely to be the same Eurasian word as that discussed by Ron Kim above.
Axel Schuessler's OS reconstruction is *paʔ and that of Baxter and Sagart is *p(r)aʔ. In his Minimal Old Chinese and Later Han Chinese: A Companion to Grammata Serica Recensa, p. 60, Schuessler gives the Tibeto-Burman root *r-pa 'axe'.
More than three decades ago, before I knew that so many languages share pVrVt / tVpVr words for "ax", I was already on the trail of a common word for this weapon / tool because I had noticed the prevalence of what archeologists call the "socketed celt" from one side of Eurasia to the other during Bronze Age times. I don't know for certain what the ancient Chinese term for "socketed celt" was, but I suspect that it might have been fǔ 斧 (*paʔ / *p(r)aʔ). Likewise, the words for "ax" in all of the other languages mentioned above may have actually been referring to the socketed celt.
One thing we can be sure of is that the Bronze Age weapon / tool in question could not have been called a "celt" in any language of the time, since that would be both anachronistic and a lapsus calami. See the explanation here.
Whether the pVrVt / tVpVr words and fǔ 斧 (*paʔ / *p(r)aʔ) are actually referring to the co-called socketed celt or to some other kind of "ax", it seems to me that there is a very good likelihood that they are related to each other.
More to come.
[Thanks to Asko Parpola, Axel Schuessler, Rod Campbell, and Chris Button]
Christopher Henrich said,
March 16, 2016 @ 4:15 pm
When (according to the best modern chronology) were the Neogrammarian times?
ayac said,
March 17, 2016 @ 1:58 am
I had never heard of this Eurasian pVrVt / tVpVr before. But I think it's worth mentioning that words for 'axe' in southern Uto-Aztecan languages are strangely similar:
Northern Tepehuan tupúúra(i)
Tarahumara ripurá
Mayo-Yaqui tepua(m)
Cora tepuaíj
Huichol tepḯa 'iron; metal'
Nahuatl tepo(ztli) 'axe; metal; metal object'.
The correspondences are not entirely regular — there's no reason for -r- in some languages to correspond to zero in others — so it's probably not an inherited word. (Also note that the Nahuatl word has been reborrowed into some of these languages, so e.g. in Mayo-Yaqui tepuam 'axe' exists alongside tepohti 'branding iron' from Nahuatl.)
languagehat said,
March 17, 2016 @ 9:57 am
Vasmer mentions an alleged Old English tæpperæx 'hatchet', sourcing it to Holthausen, Aengl. Wb. 343, but I can find no other reference to it (and don't have access to Holthausen).
Victor Mair said,
March 18, 2016 @ 6:13 am
From Alain Matthey:
Found in Falileyev 2000 : Etymological Dictionary of Old Welsh
pelechi (probably [peleki], AM) pl. gl. Clauae, J. 94: ‘clubs’, ‘cudgels’, ‘staffs’. Comment : this an -ach derivative of *pal < IE *kuel; cf Middle Welsh pal ‘spade’, ‘shovel’. My emphasis: Id. mod. Breton: pal (yr bal) ‘a shovel.’
January First-of-May said,
March 18, 2016 @ 7:46 am
The Papuan language Bongu borrowed the Russian word, as taporr.
(It also borrowed several other words from Russian, through a Russian expedition that stayed there in the late 19th century.)
Victor Mair said,
March 18, 2016 @ 9:54 am
From Peter B. Golden:
In re: Russ. “topor” you might want to add Iranian (Pers.) täbär “ax.” Topor is well-represented in Slavic. Whether it is an Iranian loanword in Proto-Slavic, native to Slavic or simply a coincidence is unclear. In addition to Vasmer (Russ. ed. IV:79-80), see also P. Ia. Chërnykh, Istoriko-etimologicheskii solver’ sovremennnogo russkogo iazyka (8th ed.) MoscowL Media, 2007), II:250-251.
Victor Mair said,
March 18, 2016 @ 9:55 am
From Alexander Vovin:
Given Ir. -b- : Slavic -p- as well as the fact that in Ir. is word attested well beyond Persian, it looks like an Iranian loan in Slavic to me, but a very old one given the quite IE correspondence of Ir. a to Slavic o.
Francesco Brighenti said,
March 18, 2016 @ 7:20 pm
R. Kim’s reference to “the ‘ax’-word, which in various forms, mainly the metathesized pVrVt and tVpVr, has spread over much of Eurasia,” is misleading. This statement harks back, I believe, to V. Abaev, who explained Persian tapar ‘axe’ as a result of metathesis in Saka *parat(a)- ‘axe’ (Ossetic færæt). However, H.W. Bailey has convincingly shown that such a metathetic relationship is unnecessary:
“The Iranian base tap- ‘to strike’ is in Zoroastrian Pahlavi tapah ‘injury’, tapahenitan ‘to destroy’ […], New Persian tabah, and in Iranian tapar ‘axe’, Zoroastrian Pahlavi tapar, Armenenian loanword tapar, Middle Persian (Turfan) tbr, tapar, New Persian tabar, [+ Balochi tapar, Waxi təpār,] Čeremis [= Mari, an Uralic language] loan-word taβar, Slavic: Russian topor ‘axe’. Old Slavonic has tep- ‘to strike’. This tabar is formed by suffix -ar-a. A confirmation of this connexion can be seen in the Tabarsarani [a.k.a Tabas(s)aran, a Northeast Caucasian language of the Lezgic branch] tabuš, tawuš […] ‘hammer’ […], which reflects the [Iranian] suffix -uša used of tool and agent. […] The Caucasian tabuš confirms the derivation of tapar from tap- ‘to strike’ and excludes a metathetic connexion with *paraθu-, in Ossetic færæt ‘axe’, Khotanese Saka paḍe ‘axes’ from *parat’u-.” (H.W. Bailey, “Iranica in Caucasian”, in G. Cardona & N.H. Zide (eds), Festschrift for Henry Hoenigswald, Tubingen, Gunter Narr, 1987, p. 35.)
Old Persian *tapara ‘axe’ is now thought to be attested in Achaemenid Elamite cuneiform sources as “Da-ba-ra” (personal name) – see Jan Tavernier here: http://tinyurl.com/z9b8up8 . Thus, this word seems to be old in Iranian. As regards its verbal base tap-, Bailey holds the view that the latter corresponds to Iranian tap- ‘to press, oppress’ (according to him, a reflex of Pokorny’s PIE root *tap- ‘to press down, press together’) as in Zoroastrian Pahlavi tapāh, New Persian tabāh ‘to harm’, Manichean Sogdian ṯp- ‘to annoy’. This tap- would be – just by chance? – homonymous with yet another Iranian verbal base tap- ‘to be hot’ tapas- ‘heat, glow; mortification, asceticism, penance’ “seems to have two origins,” thus, paralleling its Iranian counterpart (see H.W. Bailey, Dictionary of Khotan Saka, Cambridge, Cambridge University Press, 1979, p. 135 s.v. ttaus-). Finally, to close the circle, Bailey includes in the alleged PIE root *tap- (unconvincing for many IEists) both “his” Iranian base tap- (to which he now assigns the meaning ‘to strike’) with its supposed Persian derivative tapar ‘axe’ and Old Slavonic te(p)ti ‘to beat’ with its supposed – and rather doubtful – Russian derivative topor ‘axe’ (see H.W. Bailey, “Hyaona”, in H. Pilch & J. Thurow (eds), Indo-Celtica. Gedächtnisschrift für Alf Sommerfelt, Munich, Hueber, 1972, p. 24 n. 17).
However, Rick Derksen’s Etymological Dictionary of the Slavic Inherited Lexicon reconstructs Proto-Balto-Slavic *tep- > Proto-Slavonic *teti (tep-Ø+ti) ‘to beat’ > Old Slavonic teti, 1sg. tepõ. This etymon is reconstructed by Derksen to Balto-Slavic only; it is, thus, not warranted it has a PIE pedigree, pace Bailey.
Some Slavicists, assuming an etymological connection of Russian topor ‘axe’ and its Slavic cognates with Old Slavonic te(p)ti ‘to beat’, even hold Persian tapar ‘axe’ to be a prehistoric Slavic loanword in Iranian. This is a complete reversal of the direction of the borrowing – from Iranian to Russian thanks to the migrations of the Scytho-Sarmatians – opined by the majority of philologists-cum-linguists.
It is furthermore germane to say here that Bailey’s hypothesis about two tap- roots existing in Indo-Iranian has been contested as unnecessary by some leading specialists in the field. For instance, S.W. Jamison writes:
“As is well known, the Vedic root tap- shows two distinct types of meaning: physical ‘heat, burn’, mental ‘vex, harm’. Though Bailey (1956) favors separating these two into originally distinct Indo-Iranian roots, analogous semantic developments are widespread (witness in English the metaphorical meanings of both ‘heat’ and ‘burn’), and there seems no reason why both meanings cannot have developed from a single original root.” [+ Fn. 84 at p. 148: “There is no way of knowing whether the metaphorical meaning had already developed in Indo-Iranian times or whether the Indic and Iranian developments are independent (see the later Iranian forms cited by Bailey 1956: 99). It does seem likely that the metaphorical application of a root of this meaning is always a possibility, and that no long period of development need be posited.”] (S.W. Jamison, Function and Form in the -áya-Formations of the Rig Veda and Atharva Veda (Göttingen: Vandenhoeck & Ruprecht, 1983), pp. 147-148)
Likewise, in the opinion of M. Mayrhofer (EWAia I p. 624 s.v. TAP) the recognition of an Indo-Iranian root *tap-, etymologically distinct from *tap- ‘heat, be/become hot, burn’ and meaning ‘press, oppress, vex’, is unnecessary despite Bailey’s attempts.
Summing up, there exists, in my opinion, an Old Persian word *tapara ‘axe’ that can be derived from tap- ‘to strike’, an Iranian verbal base representing a specialized meaning of the PIE-inherited Indo-Iranian root *tap- ‘to be hot, burn etc.’ The “oddly metathesized resemblance” (R. Kim in TIES 8 (1999), p. 127 n. 40) of Persian *tapara to such reconstructed Saka forms as *parat(a), *par(a)t(a) ‘axe’ (which were perhaps ultimately borrowed from Akkadian pilaqqu ‘axe’ via Proto-Indo-Iranian *parac’u- vel sim.), is immaterial. The Iranian word for ‘axe’, tapar, migrated to South Russia following the movements of the Scytho-Sarmatian tribes and was borrowed into Russian as topor; from Russian it passed into other Slavic languages and even into Old English and Old Icelandic (through the Vikings/Varyags – but this is another story: see A. Liberman at http://tinyurl.com/hma6hqb ).
Francesco Brighenti said,
March 18, 2016 @ 7:50 pm
I also have a comment re: Victor’s suggestion that Chinese fǔ 斧 (*paʔ / *p(r)aʔ) ‘axe’ could be related to a “ghost” metathetic form pVrVt / tVpVr ‘axe’ – which, as I think I have shown in my previous post, is, to put it simply, not real. Any attempt at tracing an etymological relation of Chinese fǔ 斧 to other Eurasian language families must keep in view that this Chinese term has a plausible Tibeto-Burman etymon, *r-pwa ‘axe’, earlier reconstructed as *r-wa – see the STEDT Database at http://stedt.berkeley.edu/~stedt-cgi/rootcanal.pl/etymon/2772 . The reflexes of this noun root are even found in archaic Tibeto-Burman languages spoken in northeastern India. Consequently, any direct borrowing or any transmission (as a Wanderwort) of this word from other Eurasian language families must be placed at the Proto-Tibeto-Burman stage (still in the Mesolithic!) in a homeland that recent research (Blench, van Driem, etc.) situates near the eastern Himalayas. This would automatically exclude the Iranian words I have discussed in my previous post.
Francesco Brighenti said,
March 18, 2016 @ 8:15 pm
For unknown reason, a sentence included in my earlier long post has been deleted in the process. Let me therefore repost the concerned paragraph (with apologies):
——–
Old Persian *tapara ‘axe’ is now thought to be attested in Achaemenid Elamite cuneiform sources as “Da-ba-ra” (personal name) – see Jan Tavernier here: http://tinyurl.com/z9b8up8 . Thus, this word seems to be old in Iranian. As regards its verbal base tap-, Bailey holds the view that the latter corresponds to Iranian tap- ‘to press, oppress’ (according to him, a reflex of Pokorny’s PIE root *tap- ‘to press down, press together’) as in Zoroastrian Pahlavi tapāh, New Persian tabāh ‘to harm’, Manichean Sogdian ṯp- ‘to annoy’. This tap- would be – just by chance? – homonymous with yet another Iranian verbal base tap- ‘to be hot’ tapas- ‘heat, glow; mortification, asceticism, penance’ “seems to have two origins,” thus, paralleling its Iranian counterpart (see H.W. Bailey, Dictionary of Khotan Saka, Cambridge, Cambridge University Press, 1979, p. 135 s.v. ttaus-). Finally, to close the circle, Bailey includes in the alleged PIE root *tap- (unconvincing for many IEists) both “his” Iranian base tap- (to which he now assigns the meaning ‘to strike’) with its supposed Persian derivative tapar ‘axe’ and Old Slavonic te(p)ti ‘to beat’ with its supposed – and rather doubtful – Russian derivative topor ‘axe’ (see H.W. Bailey, “Hyaona”, in H. Pilch & J. Thurow (eds), Indo-Celtica. Gedächtnisschrift für Alf Sommerfelt, Munich, Hueber, 1972, p. 24 n. 17).
Francesco Brighenti said,
March 18, 2016 @ 8:18 pm
No way to repost it correctly! Apologies, again…
David Marjanović said,
March 19, 2016 @ 6:48 pm
Why "very old"? The Slavic *a > o shift only happened in historical times, long after the geographic expansion and the arrival of the Avars; it may have been the last part of the Great Slavic Vowel Shift. Someone's "name" is recorded in an early Old High German source as Tagazino – transparently *togo synъ "that one's son" (the son of the man mentioned immediately before), with already shifted *y ( < *ū) but still unshifted *o. It's too late at night for me to look up sources now, but AFAIK this has been mainstream for a while.
(In OHG writing, z was used both for [ts] and for laminal [s], as opposed to apical/retracted [s] as found in northern Spanish today, which was written s. The /s/ of all Slavic languages is laminal and must have been so for a long time.)
David Marjanović said,
March 19, 2016 @ 6:49 pm
Early ninth century, IIRC. I'll try to look it up tomorrow.
Victor Mair said,
March 24, 2016 @ 2:36 pm
From Askold Ivantchik:
If I remember well, topor is not just Russian, but common Slavic *toporъ (well represented also in other Slavic languages), which is borrowed from Iranian *tapara-. The same word is borrowed from Iranian also in Armenian and in some Finno-Ugric languages of Volga (Mari), but not in Finnish: the Finnic word is borrowed from the East-Slavic (Old Russian).
Victor Mair said,
March 25, 2016 @ 7:34 am
See also the lengthy comment by András Róna-Tas here, under "Of precious swords and Old Sinitic reconstructions, part 4" (3/24/16).
Victor Mair said,
March 25, 2016 @ 9:20 am
From Marcel Erdal:
‘Axe’ is balto / baltu / balta in all Turkic languages, showing the medial /l/ found in Greek and Celtic but the third-syllable /t/ found in Iranian.
Victor Mair said,
March 25, 2016 @ 1:33 pm
From Ron Kim, commenting on his paper in Tocharian and Indo-European Studies [TIES] 8 [1999]) in the o.p.:
=====
My point there (which I'm not sure any specialists have ever explicitly responded to) is that the vowel correspondence in the first syllable of the Tocharian forms (B e ~ A o) suggests that the Iranian word was borrowed as Proto-Tocharian *e, but after palatalization had become fully phonemic, hence as PT *peret@. (Compare Russian бизнесмен, Интернет with palatalized n', m', t' vs. Polish biznes, Internet with unpalatalized consonants; to simplify slightly, Russian consonants are automatically palatalized before /e/, whereas Polish has a full phonemic contrast.)
….
Abaev's Ossetic etymological dictionary s.v. færæt gives some forms from Finno-Ugric languages: Komi purt, Udmurt purt 'knife', Hungarian bart 'sword', Turkish balta, Chuvash porda 'ax', plus the Tungusic Evenki has parta, porta, purta 'knife'.
….
Yhe Proto-Slavic forms *katilu (> Old Church Slavonic kotilŭ, Russ. kotël, Pol. kocioł) and *xlēbu (> OCS xlěbŭ, Russ. xleb, Pol. chleb, Serbo-Croatian hl(j)eb [though the Croats usually use kruh], Bulg. xljab) are clearly borrowed from Germanic, where the protoforms are *katilaz and *hlaifaz. Given the chronology and geography, East Germanic speakers are the obvious source, and indeed Gothic has katils and hlaifs. There are several other good examples found in all Slavic languages which are for that reason regarded as early borrowings into Proto-Slavic, e.g. *lēku 'medicine' (> OCS lěkŭ, Russ., Pol. lek, Serbo-Croatian l(ij)ek), while others are later and/or confined to only part of the Slavic world. A book about the topic just appeared, by Saskia Pronk-Tiethoff, but I haven't seen it yet.
=====
languagehat said,
March 25, 2016 @ 2:24 pm
Russian is perfectly capable of borrowing words without palatalizing consonants before /e/; this is one of the main reasons you need a good dictionary of pronunciation to render words properly, since the distinction is invisible is spelling (I use the Словарь ударений для работников радио и телевидения). As it happens, бизнесмен is /biznesm'en/ with unpalatalized n, and all words beginning интер- are /inter/, with no palatalization: интервал, интервенция, интервью, интернат, интернациональный, etc. I assume this holds for Интернет as well, though my dictionary is too old to include it; Russian-speakers will correct me if that word is an anomaly, but the general point about borrowing without palatalizing stands. (I well remember my surprise when I learned the /m/ in Гомер 'Homer' was not palatalized; that's what opened my eyes to this problem.)
NV said,
March 29, 2016 @ 8:11 am
@Victor Mair, languagehat
As a native speaker, I've never heard anyone pronounce бизнесмен and Интернет with palatalized n. I imagine only an elderly person who doesn't know these words would read them this way (and would probably be promptly corrected).
languagehat said,
March 29, 2016 @ 8:41 am
I'm glad to have my point confirmed by a native speaker.