"Horse" and "language" in Korean

« previous post | next post »

A Korean student was just in my office and saw this book on my table:  mal-ui segyesa 말의 세계사.

She said, "Oh, a world history of words!"

But I knew that couldn't be right because the book is a world history of horses.  It's actually a Korean translation of this book by Pita Kelekna:

The Horse in Human History (Cambridge:  Cambridge University Press, 2009)

So what happened?  Did the student make a mistake?

A colleague who is a Korean specialist remarks:

세계사 is indeed "world history."  And 의 is the genitive marker.

말 is both "horse(s)" and "language, speech, word(s)",  i.e., a simple homonym.  She didn't do anything wrong, although the English rendering should have been "world history of speech" or some such.  As always, context counts and she probably figured anything in your office would be language related.

From another colleague who is a specialist in Korean language teaching:

I would also think the book is about the history of language rather than a world history of horses. I never heard of a horse history book.

I asked Bob Ramsey:

말 means both "horse" and "word", right?

He replied:

Sure does. But 'word, speech, language' has a long vowel, while the vowel of 'horse' is short. (And the long vowel, btw, was a rising tone in Middle Korean.)

I continued:

But these things are not reflected in the orthography, are they?

Bob replied:

No, they're not. They're written exactly the same way in today's orthography. Thus your student's reaction.

Juha Janhunen observed:

Yes, indeed, I never thought of it, but in modern Korean mal 'horse' and mal 'language' (as in Ilbon-mal 'Japanese language', Goryeo-mal 'Koryeo language' = the language of the Central Asian Koreans) sound the same. To my knowledge mal 'horse' goes back to măl = măr < *mor(V), while mal (mar) 'language, word' derives from mal < *mar(V). The vowels a (< *a) and ă (< *o) were written with different letters in the original version of Hangeul (1446), but in most Korean dialects today they have phonetically merged, which is why the letter for ă (a dot) is no longer used in the standard orthography, so the two words have become both homonyms and homographs.

Since we have often touched upon Eurasian words for "horse" and will be coming back to them in short order, it's worth mentioning one of the nine different etymologies for "mal 말" listed in Wiktionary that means "horse":

The ma sound denoting "horse" is common to a number of languages of central Asia, where horses were first domesticated, suggesting a possible cognate root. Compare Manchu ᠮᠣᡵᡳᠨ (morin, "horse"), Mongolian морь (morʹ, "horse"), Mandarin (, "horse"), Japanese (uma, "horse") and Proto-Indo-European *márkos ("horse") and descendants such as Irish marc ("horse", archaic) or English mare ("female horse"). More at *márkos.

This experience with a word in the title of a Korean book on my office table raises many interesting and important questions concerning orthography, phonological evolution, homonymy and homophony, and so forth, but above all context.

 

Selected readings

"An early fourth century AD historical puzzle involving a Caucasian people in North China" (1/25/19)

"Of horse riding and Old Sinitic reconstructions" (4/21/19)

"Of reindeer and Old Sinitic reconstructions" (12/23/18)

 

[Thanks to Bill Hannas, Haewon Cho, and Yishu Ma]



48 Comments

  1. Phillip Minden said,

    October 30, 2019 @ 7:38 am

    Ko'rean student or Korean 'student? (Unless a Korean Korean student, of course.)

  2. Victor Mair said,

    October 30, 2019 @ 7:47 am

    She's a native speaker of Korean.

  3. Chau said,

    October 30, 2019 @ 9:40 am

    In Old Norse marr is 'horse, steed', and mál 'speech' with á being a long 'a'. Thanks to Bob Ramsey's comment, "But 'word, speech, language' has a long vowel, while the vowel of 'horse' is short," it makes perfect sense to me now.

  4. Victor Mair said,

    October 30, 2019 @ 11:07 am

    From Brian Spooner:

    =====

    In my early days wandering about the countryside in Iran I was surprised to hear people using the word "mal," which anyone who knows Persian will tell you means "property," for any animal they or others were riding.

    =====

    This reminds me (VHM) of the "cattle-capital-property" nexus in English:

    =====

    cattle (n.)

    mid-13c., "property" of any kind, including money, land, income; from Anglo-French catel "property" (Old North French catel, Old French chatel), from Medieval Latin capitale "property, stock," noun use of neuter of Latin adjective capitalis "principal, chief," literally "of the head," from caput (genitive capitis) "head" (from PIE root *kaput- "head"). Compare sense development of fee, pecuniary.

    in later Middle English especially "movable property, livestock" (early 14c.), including horses, sheep, asses, etc.; it began to be limited to "cows and bulls" from late 16c.

    https://www.etymonline.com/search?q=cattle

    =====

  5. Victor Mair said,

    October 30, 2019 @ 11:10 am

    From Robert Drews:

    The borrowing of an IE "mare" word into east Asia seems possible to me, but I would also think that east Asians in a northern latitude should have had a word for horse already in the Paleolithic period, long before any domestication.

  6. Peter B. Golden said,

    October 30, 2019 @ 1:28 pm

    Russ. мерин (мерин) "gelding" Cf. also the expression врёт как сивый мерин (lit."he lies like a gray gelding") "he is an out and out liar." The origins of this expression are much debated.

    мерин is considered a loanword from Mongol (morin)

  7. J.W. Brewer said,

    October 30, 2019 @ 1:41 pm

    In the King James Version, the very last verse of the Book of Jonah has the Lord saying to Jonah: "And should not I spare Nineveh, that great city, wherein are more than sixscore thousand persons that cannot discern between their right hand and their left hand; and also much cattle?" I've always liked that line, which somehow seems solemn and goofy at the same time. But the point here is that some more recent translations substitute "many animals" for "much cattle," suggesting that the KJV may have retained that broader-scope meaning of "cattle" not yet limited to "cows and bulls."

  8. Victor Mair said,

    October 30, 2019 @ 2:29 pm

    The following account is not directly related to the main issues being discussed in this post, but it does have to do with horses, so I can't resist mentioning it here.

    A student has been wrestling with a somewhat thorny situation involving the bureaucracies of two universities that had been dragging on for months. I was not even aware there was a problem until three days ago. When I found out, I felt very bad about it and immediately got involved in trying to straighten the matter out.

    I was just now informed that the difficulty is being ironed out. When I told the student, she said, "hái děi yào nín chūmǎ 还得要您出马" ("after all, it was necessary for you to take the field").

    The interesting phrase she used is "chūmǎ 出马", lit., "go out [on one's] horse". This is an old expression from the literature of military engagements which means "go into action; take the field; to set out (on a campaign)", but in recent times is used figuratively in the political arena:​ "to stand for election; to throw one's cap in the ring". When she said that about me, it made me chuckle, since I'm actually afraid of horses.

  9. David Morris said,

    October 30, 2019 @ 2:53 pm

    Attempting to learn Korean as a second language, I have encountered mal=word/language many times, but never mal=horse. For some reason it never crops up in textbooks focusing on daily life in modern Korea.

    Was there a photo or other graphic on the cover of the book?

  10. Victor Mair said,

    October 30, 2019 @ 3:47 pm

    Since the book was buried under a pile of other books, all she could see was the spine, hence no picture of a horse, though there was one on the front cover.

  11. AmyW said,

    October 30, 2019 @ 4:10 pm

    J.W. Brewer,

    Regarding KJV use of the word "cattle," Genesis 30 (esp. verse 32) is a good example where "cattle" appears to refer to livestock in general, apparently sheep in particular.
    "I will pass through all thy flock to day, removing from thence all the speckled and spotted cattle, and all the brown cattle among the sheep, and the spotted and speckled among the goats: and of such shall be my hire."

  12. AntC said,

    October 30, 2019 @ 4:33 pm

    This reminds me (VHM) of the "cattle-capital-property" nexus in English:

    Aha! That sense persists in the legal term '(goods and) chattels', which is still used (in BrE at least) in house sales and tenancies to mean the semi-fixed furnishings, as opposed to the bricks-and-mortar. (Buyers/renters will often bargain with the sellers/landlord to for example avoid outlay on drapes and carpets.)

  13. Chris Button said,

    October 30, 2019 @ 4:59 pm

    This reminds me (VHM) of the "cattle-capital-property" nexus in English

    Jim Matisoff wrote a paper about that from a Sino-Tibetan perspective: "Universal Semantics and Allofamic Identification–Two Sino-Tibetan Case-studies–'straight/flat/full' and property/livestock/talent' " (1988)

    The ma sound denoting "horse" is common to a number of languages of central Asia, where horses were first domesticated, suggesting a possible cognate root. Compare Manchu ᠮᠣᡵᡳᠨ (morin, "horse"), Mongolian морь (morʹ, "horse"), Mandarin 馬 (mǎ, "horse"), Japanese 馬 (uma, "horse") and Proto-Indo-European *márkos ("horse") and descendants such as Irish marc ("horse", archaic) or English mare ("female horse"). More at *márkos.

    Probably worth mentioning here again the velar component /ɣ/ in OC 馬 *mráɣʔ that unfortunately goes missing in most reconstructions today (it's a velar nasal in some of its Tibeto-Burman cognates).

    @ Professor Mair

    I'm still looking forward to that post on the external connections of 熊 "bear"…

  14. David Marjanović said,

    October 30, 2019 @ 6:31 pm

    All I can add is the usual warning about "Proto-Indo-European *márkos": it's found exclusively in Celtic and Germanic, and may well be a loan from one to the other or into both from yet another source (provided all that happened before Grimm's law, probably); the *a is highly suspicious in a PIE reconstruction (a problem that disappears if the Celtic version is a loan from Germanic); and it does happen that words with the same meaning are completely identical, let alone merely similar, between different languages for no reason other than random chance.

  15. Victor Mair said,

    October 30, 2019 @ 7:24 pm

    This is not just an isolated word in one language here and in another language there, but a whole cluster of words with similar sounds and the same meaning in numerous languages that were in contiguous contact.

  16. Victor Mair said,

    October 30, 2019 @ 7:29 pm

    @Chris Button

    Thanks for the reminder.

    The preliminary research I did for 熊 "bear" is buried thousands of e-mails ago. If you can recall the month and perhaps even the approximate day when I mentioned 熊 "bear", I might be able to dig the notes up within a couple of weeks. Otherwise, I probably won't be able to do it till Yuletide.

  17. Chris Button said,

    October 30, 2019 @ 9:46 pm

    This is not just an isolated word in one language here and in another language there, but a whole cluster of words with similar sounds and the same meaning in numerous languages that were in contiguous contact.

    Indeed. I think Pulleyblank put it quite nicely 20+ years ago: "…Celtic and Germanic are not usually thought to be especially close to one another and it could that they have preserved an old word lost elsewhere. In Old English eoh, derived from *eḱu̯o-, means specifically 'war horse.' One might suppose that *marko- was an older generic term that was replaced by *eḱu̯o- at first as a term for 'war horse,' later extended to the generic meaning in most dialects. The fact that Tocharian as we know it only has words derived from *eḱu̯o- is a problem but less so when we recall that the attested forms in Tocharian are at least 2000 years later than the presumed time of borrowing into Proto-Chinese."

    Incidentally, Mallory and Adams suggest "wild horse" versus "domesticated horse" as a possibility.

    If you can recall the month and perhaps even the approximate day when I mentioned 熊 "bear"

    I believe it was Sunday September 15th.

    Like 犬 "dog", I find 熊 "bear" to have a somewhat unstable reconstruction in Old Chinese which suggests an external connection. In the meantime, the issue was touched on here:

    https://languagelog.ldc.upenn.edu/nll/?p=8050

  18. Philip Taylor said,

    October 31, 2019 @ 5:41 am

    "I believe it was Sunday September 15th" — Also June 28th 2017 ('In Chinese Rilakkuma is called Sōngchí xióng 鬆弛熊 (lit., "relaxed / loose / flaccid / flabby bear")').

  19. Victor Mair said,

    October 31, 2019 @ 5:44 am

    From Joshua Katz:

    I'm about to board a flight to London and so am not in the best position to be authoritative! However, if the word is found outside Germanic and Celtic, I don't know it — unless it is somehow connected as a loan or Wanderwort with such words in the Far East as Chinese ma3, which I can't evaluate but I'm sure you have an opinion on (and I expect I should know it but don't remember, sorry…). I'm not bothered about *a in the way the Leideners are, but I don't suppose a very early borrowing from the one to the other (cf. iron in Germanic, from Celtic) is out of the question.

  20. Victor Mair said,

    October 31, 2019 @ 5:47 am

    From Brian Joseph:

    There are some IE-ists, especially of the "Leiden School" (headed by Beekes and Kortlandt, though both are retired now), who deny that PIE had an *a vowel. And regardless of what one thinks about the reconstruction, it is certainly the case that correspondences involving an *a that is **not** connected to a laryngeal consonant (the second laryngeal, the *H2 that appears in Hittite as ḫ initially and medially) are rare and generally (Meillet noted this, and maybe others before him) are in the vicinity (before or after) *s, or a guttural, and maybe *r (I can't remember offhand). For instance, the root *Hyaǵ- 'worship', which gives Sanskrit yaj- 'worship' and Greek hagios 'holy' (the initial *H- is to explain the Greek #h- — PIE #y- by itself gives a Greek < ζ > (zeta)) would be one such case of *a. The Leiden folks treat all of those as involving laryngeals, so their reconstruction would be something like *HyeH2ǵ- and then to "explain" the fact that there are **no** forms with a long vowel for this root (one would expect *eH2 to give *ā), they resort to a "rule" (Beekes refers to it as "Lubotsky's Rule" or "L's law", in the old-style nomenclature for these sound changes) that somehow eliminates the laryngeal when it is not convenient. As you can see from my presentation, I am not convinced. But admittedly the distributional facts surrounding *a that Meillet commented on are a bit disturbing no matter how you slice it. Notice that a root shaped *HyH2eǵ- would be problematic because the *y should vocalize between consonants and there is no evidence that I am aware of for a syllabification *iaj- in Sanskrit for this root and the Greek outcome would be odd (#ια- is an acceptable onset for roots – there are lots of verbs that start with that sequence).

    I hope this helps.

    What does Don Ringe think about this?

  21. Victor Mair said,

    October 31, 2019 @ 5:50 am

    From Michael Weiss:

    Yes, AFAIK, it's only Celtic and Germanic and indeed *a is rare in PIE words, but not impossible. It's also true that since *o > *a in PGmc. you could start with *morkos > PGMc. *marhaz and loan it into Celtic where it could be naturalized as *markos. But loans between Celtic and Germanic mainly go the other way.

  22. Victor Mair said,

    October 31, 2019 @ 5:53 am

    From Hannes Fellner:

    The Indo-European word has a conspicuous geographic distribution: it is exclusively attested in the two westernmost branches of Indo-European, i.e. Proto-Germanic *marha- (e.g. OE mēarh, OHG marh, ON marr) and Proto-Celtic *marko- (e.g. Gaulish markos, OIr. marc, OBret. marh).

    Furthermore, the phonological shape of *marko- raises doubts as to whether the word ever existed in non-western branches of IE. First and more importantly, a as a root vowel is rare in Proto-Indo-European. Second and less importantly, if *marko- is further analyzed as *mark-o-, the coda -rk is rare in IE roots (the Lexikon der indogermanischen Verben only has three such roots). However, if *marko- is analyzed as *mar-ko-, e.g. with the adjectival suffix *-ko- that designates origin (cf. Ved. síndhu-ka- 'from Sindh', Gk. Libu-kós 'Libyan', Latinized Gaulish Are-mori-cī 'those by the sea, Aremoricans', Goth. staina-hs 'stony'), the word would not exhibit a rare root coda, but then the derivational semantics are unclear. Under either analysis, of course, the root remains unidentified.

    It may be significant that the verb from which English ride is derived, sometimes reconstructed for PIE as *reidh-, is only securely attested in Germanic (OE rīdan etc.) and Celtic (Old Irish -réid). If both *marko- and *reidh- are Celtic-Germanic isoglosses, it is worth considering the possibility that the two words were introduced into Celtic and Germanic together withhorseback riding. That it is highly likely that *marko- originally referred to the riding horse is evidenced by the Gaulish denominal verb marcosior 'I should like to ride (sexually)' and Galatian trimarkisia 'set of three horsemen'.

    Another potential Celtic-Germanic isogloss originally likely from semantic sphere of horseback riding is the word for 'breeches' Proto-Celtic *brōk- (in Gaulish brāca whence Lat. brāca) and PGmc. *brōk- (OE brōc, OHG bruoh, ON brōk).

  23. Victor Mair said,

    October 31, 2019 @ 5:57 am

    From Juha Janhunen:

    I haven't checked it in etymological handbooks, but, indeed, Russian mérin could be a rather recent loanword from Western Mongol (Oirat) mörin, in which the original velar *o is umlauted to front [ö], replaced in Russian by /e/, but the *i of the second syllable is still preserved as a segment (as you know, in modern Oirat the form is /mörn/). Somebody should check the age of merin in Russian – when is it first recorded?

  24. Victor Mair said,

    October 31, 2019 @ 5:58 am

    From Melanie Malzahn:

    Much ink has been spent on that problem, but no solution possible. Since there is also the PIE word for horse, it is likely that *markos is indeed a loan into both or one of them from a pre-IE European language.

  25. Victor Mair said,

    October 31, 2019 @ 6:01 am

    From Stefan Georg:

    Acc. To Sreznevskij (Mat. Drevner.) in 1571: Правая грамота о возвратѣ во владѣніе Николаевском Карельскому монастырю деревни и полутора въ Картиной курҍ 1571 г. To be found in (so he says): Акты юридическіе или собранное форма старинного дѣлопроизводства. СПг. 1838.

    [VHM: I have an image of Sreznevskij's entry from Stefan if anyone wants to see it.]

  26. Victor Mair said,

    October 31, 2019 @ 6:03 am

    From Stefan Georg:

    Sreznevskij is, surprisingly, not precise about the 1st attestation, since the said collection of documents (1838) has the word in some act of 1547. Vasmer is more cautious (and obviously rightly so) and says „since about 1500").

  27. Victor Mair said,

    October 31, 2019 @ 6:04 am

    From Stefan Georg:

    Though I will stop this unexpectedly fascinating research here (for the moment), I might add that I found, in some other collection of old court proceedings, occurrences of merin in a piece dated „около 1500". I do not know, at the moment, where these proceedings happened, but it seems to be Moscow-ish; the 1547 document mentions the word several times, and it seems to have been comprehensible to everybody involved.

  28. Chris Button said,

    October 31, 2019 @ 7:32 am

    I might add that several people have quite reasonably suggested that e/o in PIE goes back to ə/a, as we see in Old Chinese and Proto-Tibeto-Burman–the Burmese inscriptions in particular being very instructive for the latter, and for which the Northwest Caucasian languages provide a modern example in terms of phonology (but not in surface phonetics of course, in spite of common misconceptions)–which raises further problems for any notion of a separate original PIE "a" as something that wasn't conditioned.

  29. Victor Mair said,

    October 31, 2019 @ 10:48 am

    From Tsu-Lin Mei:

    I have a lot to say about "horse" in Sino-Tibetan and 'horse' in Altaic, to which Korean belongs. I have nothing to say about "language" in Korean.

    Sino-Tibetan. Gong Hwang-cherng 2002; 8,200, 221 gives OC *mrag 馬, WT rmang < *mrang, WB mrang. Yakhontov 1960 "Consonant combinations in Archaic Chinese" showed that Second Division words have medial –r- in Old Chinese. In Yakhontov 1960 he reconstructed 馬*mla and compared it with WB mrang. WT rmang "horse" is an Old Tibetan word for 'horse' (the current word is rta) discovered by Coblin in Dunhuang manuscripts; see Coblin 1974 "An early Tibetan word for 'horse'", JAOS 94:1:124-125. Beyer, The Classical Tibetan Language, p.8 footnote has a long discussion of rmang in Central Asian manuscripts and it relation to WT rta "horse". The –g in OC *mrag is reconstructed for all 魚部陰聲words, and this *-g is confirmed by the homorganic nasal –ng in WT and WB. Sino-tibetan *mrag/*mrang is wanderwort since horse is not native to East Asia and it must have been borrowed from IE at the Sino-Tibetan stage. Sino-Tibetan is monosyllabic . Altaic , which includes Mongolian, Manchu, Korean, and Japanese, is poly-syllabic. So when ST *mrag/*mrang was borrowed by Altaic , it acquired an –o- between m- and –r-, e.g. "horse" Manchu morin, Mongolian mor, and Korean mal < *mor. That is to say, the final –r in Altaic is the medial –r- in Sino-Tibetan. I gave a course "Chinese historical phonology" at Harvard in 1968 & 1969. At the time I knew very little about Chinese historical phonology, but I had just learned from Jerry Norman Yakhontov 1960 and the Manchu word for horse "morin". I mentioned all this to my class. There was a Korean visiting scholar in my class and he reported my theory about East Asian "horse" in writing. Many years later my colleague John Whitman at Cornell told me about my course at Harvard. So your Korean student should really talk to John Whitman, who is an expert of Korean-Japanese & Altaic linguistics.

  30. Victor Mair said,

    October 31, 2019 @ 11:23 am

    From Don Ringe:

    I agree with you, Brian. I'm not sure that *a is really as rare as some claim, and there are examples that not only can't be explained with laryngeals, like yours, but don't fall within Meillet's distribution *and* ablaut to boot. The two obvious ablauting ones are *wā́stu ~ *wástu- 'settlement' (Skt. vā́stu 'settlement' and Toch. A waṣt, B ost 'house' with *ā, Greek ἄστυ, Myc. wa-tu 'citadel' with *a) and *nā́s- ~ *nás- ~ *n̥s- 'nose' (the zero grade in Dutch neus and possibly in OE nosu).

    As for *márkos, it clearly isn't PIE just because of its distribution, and I think a loan from Celtic into Germanic is more likely.

  31. BZ said,

    October 31, 2019 @ 12:29 pm

    The original Hebrew word translated as "cattle" in Jonah is "בְהֵמָ֖ה" which can refer to any domesticated animal (not just cows or sheep), or less frequently, any animal in general. It's in the singular here, but is used as a mass noun.

  32. Victor Mair said,

    October 31, 2019 @ 2:51 pm

    From Jay Jasonoff:

    It's a huge exaggeration to say "the *a is highly suspicious in a PIE reconstruction"; this is in fact the weakest possible indication. The real thing, in my view, is that a word denoting a breed or variety of horse is just not likely to go back to IE times – and is even less likely to if it only occurs in Celtic and Germanic. A Celtic to Germanic loan is probably likelier than Germanic to Celtic; the Germans were more often on the receiving end back then.

  33. Nelson said,

    October 31, 2019 @ 3:39 pm

    Adam Hyllestad discusses Celtic-Germanic words in a 2010 paper:

    'The Precursors of Celtic and Germanic', in Proceedings of the 21st Annual UCLA Indo-European Conference, ed. Jamison, Melchert, & Vine, pp. 107-128.
    https://www.academia.edu/377059/The_Precursors_of_Celtic_and_Germanic

    He argues that there is a set of ten equestrian-related innovations (shared words, shared semantic innovations, or shared morphological peculiarities) found in these two branches but not elsewhere in Indo-European. The relevant forms are on pages 119-120 of the article (category VII). He attributes these to early contacts, of course, not to a Celto-Germanic linguistic branch (for which there could be no evidence).

  34. Bathrobe said,

    October 31, 2019 @ 7:06 pm

    Interesting that mal is used for property and livestock in Persian. Is there any connection with Mongolian mal, which means 'livestock'?

  35. Chris Button said,

    October 31, 2019 @ 8:36 pm

    Gong Hwang-cherng 2002; 8,200, 221 gives OC *mrag 馬, WT rmang < *mrang, WB mrang

    A voiced stop *-g in 馬 has now been convincingly discredited, but the evidence lying behind it has not simply disappeared, in spite of some modern reconstructions trying to ignore it. Pulleyblank's *-ɣ (or *-ɰ) corresponds to Li Fang-Kuei's *-g and hence Gong's *-g. Li notably commented that it may well not have been a stop per se. In his OC Loanwords in Tai (1945) article he assumes a shift of -əg > -əɣ > -əɯ. Pulleyblank does not assume the earlier stopped phase and treats Li's vocalic /ɯ/ as the velar glide /ɰ/ instead (or lightly fricated /ɣ/).

    So we have 馬 *mráɣʔ which compares well the Celtic and Germanic forms in PIE.

    As for Old Burmese mrɐŋ² "horse", Shorto and Luce associate it with Mon-Khmer (Luce suggests MK as the source, although notably Hla Pe does not include it in his list of such words; Shorto does not speculate). Comparisons like Kuki-Chin *rɐŋ² are noted by Benedict as possible evidence that the *m- may not be original. On this basis, Pulleyblank suggests that OB mrɐŋ² and OC 馬 *mráɣʔ may actually be unrelated in spite of their superficial similarity.

    The problem as I see it is less about whether or not any of these forms are actually related, but rather about some of the "modern" (and largely erroneous in my opinion) reconstructions of OC that produce forms that no longer bear any resemblance to their putative cognates in PIE or elsewhere. The assumption seems to be that we should just ignore the possibility now since these new OC reconstructions must be correct…

    So…

    Yes, 巫 *màɣ had a velar coda just like 馬 *mráɣʔ. No, 天 *tʰə́ɲ did not have a liquid onset, and neither 犬 *kʰʷə́ɲʔ nor 雁 *ŋráns had liquid codas. And no 車 *kɬàɣ did not have a "t- prefix. I could go on…and on… and on…

  36. Paul said,

    November 1, 2019 @ 9:52 am

    Curious story, I didn't know this fact. Korean and Japanese for me are of the most difficult languages to study. Because my parents have property in Phuket Thailand https://tranio.com/thailand/phuket/ I can speak some Thai language as I visit them quiet often but I don't understand much of Korean.

  37. David Morris said,

    November 1, 2019 @ 2:55 pm

    I wonder what she would have thought 배의 세계사 was about.

  38. Victor Mair said,

    November 1, 2019 @ 6:11 pm

    Today when I went to the dry cleaner to pick up my pants, I told the Korean lady working there about this post, and she immediately said what Bob Ramsey said above in the o.p.: "'word, speech, language' has a long vowel, while the vowel of 'horse' is short."

  39. Chris Button said,

    November 2, 2019 @ 8:57 am

    As for Old Burmese mrɐŋ² "horse", Shorto and Luce associate it with Mon-Khmer (Luce suggests MK as the source, although notably Hla Pe does not include it in his list of such words; Shorto does not speculate). Comparisons like Kuki-Chin *rɐŋ² are noted by Benedict as possible evidence that the *m- may not be original. On this basis, Pulleyblank suggests that OB mrɐŋ² and OC 馬 *mráɣʔ may actually be unrelated in spite of their superficial similarity.

    I wonder of the relationship if any in Burmese between မြင်း *mrɐŋ² "horse" and မြည်း *mrɐɲ² "donkey, ass". The former would go back to earlier *mraŋʔ and the latter to *mrjəŋʔ if it could indeed be taken back that far. The -rj- combination in *mrjəŋʔ suggests that the m- results from prefixation or compounding.

  40. Chris Button said,

    November 2, 2019 @ 9:00 am

    Hla Pe tentatively notes an association of မြည်း *mrɐɲ² "donkey, ass" with Mon, but he is uncertain of the direction if it is even valid at all.

  41. Chris Button said,

    November 2, 2019 @ 9:02 am

    Actually မြည်း *mrɐɲ² could go back to *mrjəŋʔ or *mrjənʔ . A coronal -n coda instead of a velar -ŋ distances it somewhat more from မြင်း *mraŋʔ

  42. R. Fenwick said,

    November 6, 2019 @ 1:52 am

    @Victor Mair:

    This reminds me (VHM) of the "cattle-capital-property" nexus in English:

    As the entry you cite in your post already notes, the nexus is an old one indeed in Indo-European, visible also in Latin pecūnia "money" (from pecū "cattle"), but with a further cognate pair in Avestan that may suggest the view of livestock and wealth is original to Proto-Indo-European: Avestan fšūmant- "householder" and pasu "livestock" have formally diverged phonologically, but both come from PIE (*péḱu "livestock" yielding, with PIE-style ablaut, *pḱumento– *"livestock owner" → "householder, landholder"). Within the Indo-Iranian context I also can't help but think of the Sanskrit example used in the film Arrivalgáviṣṭi- "desire (especially for battle); (by extension) battle"—and wondering if its etymology might also have some relevance here.

    FWIW, in Ubykh the same relationship pertains: landʷá "livestock" also means "property, wealth, riches" (note also the compound landʷatʷəʥá "gift, present"; literally, "given livestock"). The etymological origin of the term isn't clear, but it seems to be a deverbal noun of some sort standing in opposition to ʃʷandʷá "wild animals, game", which is itself a straightforward patientive deverbal noun from a now-extinct verb *ʃʷa "to hunt" (which doesn't survive alone in Ubykh, but is visible also in ʃʷạ́kʲ'a "hunter" and a derivational suffix –ạʃʷa "hunting for [an animal]"—e.g. ɬ-ạʃʷá "hunt for wild goat", l[a]-ạʃʷá "hunt for hare", χʷˁ[a]-ạʃʷá "hunt for boar"—as well as good cognates in Abkhaz á-ʃʷaraχ "wild animals, game" and a-ʃʷaráʦa-ra "to hunt"). Indeed, there's a lovely pseudo-reduplicative dvandva compound attested in Ubykh landʷaʃʷandʷa "the animal kingdom", literally "wild and domestic animals".

  43. Philip Anderson said,

    November 6, 2019 @ 8:05 am

    The same nexus was found in Old English 'feoh', a cognate. Modern English 'fee' comes from another Germanic cognate.
    https://en.m.wikipedia.org/wiki/Fehu

  44. Victor Mair said,

    November 7, 2019 @ 12:40 am

    From Jay Jasanoff:

    It's a huge exaggeration to say "the *a is highly suspicious in a PIE reconstruction"; this is in fact the weakest possible indication. The real thing, in my view, is that a word denoting a breed or variety of horse is just not likely to go back to IE times – and is even less likely to if it only occurs in Celtic and Germanic. A Celtic to Germanic loan is probably likelier than Germanic to Celtic; the Germans were more often on the receiving end back then.

  45. David Marjanović said,

    November 10, 2019 @ 6:32 am

    Thank you, Prof. Mair, for soliciting all these comments from IEists!

    It just occurred to me that, from a purely phonological perspective, the Germanic word could have been – like so many others – borrowed from Celtic after Germanic shifted its consonants. In that case, as far as I understand, we could be looking at a PIE *mr̩kós… which could have the *-kó- suffix mentioned above, but it wouldn't make semantic sense, because the root *mer- meant "die". OK, in Hittite it meant "disappear", which should logically be the older meaning; but *mr̩kós being "the one associated with disappearing" because it runs so fast strikes me as a stretch.

    If the whole thing is a loan, I can see two options. Either it's from a local non-IE language that has disappeared; this hypothesis is hardly testable. Or it's from the east, a Wanderwort related to the East Asian words – but that would have to have arrived through Iranian, one should think, and Iranian traces of this word have not been reported so far.

    ================

    For the record, I agree (rather unsurprisingly) with Prof. Ringe that there are a few PIE roots for which *a remains the most parsimonious reconstruction; Lubotsky's law works in the sense of not contradicting the data, but the fact that it seems to have no applications outside these few roots makes it less parsimonious. Still, *a must have been rare enough that any reconstructed root with *a should be suspected of being either wrong or a loan.

    (The Balto-Slavic "acute" tone of the "nose" root does show that there was a laryngeal somewhere in it. But it doesn't tell us where; it could be *h₁nas- or something.)

    ================

    I think Pulleyblank put it quite nicely 20+ years ago: "…Celtic and Germanic are not usually thought to be especially close to one another

    Not that it matters in this case, but they are usually thought to be pretty closely related, even though they're clearly not each other's closest known relatives. There is some evidence for a "West IE" group of Italo-Celtic, Germanic and Albanian, such as Dybo's Law (shortening of long vowels in syllables before the stressed one; not sure if it's been tested on Albanian data, which has to be difficult) or the unusual change of *-tst- to *-ts- (later on to *-ss- in all four branches separately; still ts in some early Celtic inscriptions).

    Continuing the same quote:

    and it could that they have preserved an old word lost elsewhere. In Old English eoh, derived from *eḱu̯o-, means specifically 'war horse.' One might suppose that *marko- was an older generic term that was replaced by *eḱu̯o- at first as a term for 'war horse,' later extended to the generic meaning in most dialects. The fact that Tocharian as we know it only has words derived from *eḱu̯o- is a problem but less so when we recall that the attested forms in Tocharian are at least 2000 years later than the presumed time of borrowing into Proto-Chinese."

    Sure, there's nothing impossible about this. But it's improbable. This holds even more strongly given the fact that *eḱu- (without the derivative suffix -o-!) is attested in several Anatolian languages, while *mar(k)- is not.

    Incidentally, Mallory and Adams suggest "wild horse" versus "domesticated horse" as a possibility.

    Again, that's possible, but is there any further evidence for this?

  46. pamela said,

    November 10, 2019 @ 10:50 am

    I am fascinated by this discussion (came here from marko). I have nothing to add in the linguistic line apart from historical context. I have always assumed that the pockets of Eurasia who had early horses, say before 1000 BCE, in some cases stenonids or surviving stenonid populations, had their own words. the spread of -mar words for horse (which I assume are loanwords in almost all cases) seems connected to the spread of horse types, probably those connected to chariot warfare or some forms of transportation. The questions for myself are: We know from genetic studies and a little bit of historical commentary that changes in horse type were associated with the introduction of stallions to local mare populations. So shouldn't -mar be associated in some way with stallions or imported horses (and what happened with OE?); instead, it seems associated with the horse type produced. In that case, shouldn't -mar words spread in the general pattern that the horses themselves did? This partly fits, since there seems a general sort of western-central-Eurasia radiant point, and places like Egypt that had been important horses centers but generally on the receiving end of horse use and technology don't seem to have spread their own horse-related vocabularies very far. but for some reason we tend to see a lot of -mar impetus coming from west to east, while we know that an early and influential genetic pool for the modern horse was actually in eastern Eurasia. so perhaps the -mar series of continental scope was inspired less by the type of horse and more by the style of use.

  47. R. Fenwick said,

    November 11, 2019 @ 2:21 am

    Following from David Marjanović, a further potential issue with the supposition of *eḱu̯o- as "war horse" may be raised by Udi eˁkʷ ~ eˁk (pl. eˁk-ur) "horse". John Colarusso has suggested the Udi term is a loan from Indo-European, with the Udi pharyngealisation preserving a relic of the laryngeal in PIE *h₁éḱu- and suggesting that this has been the basic term for a very long time indeed. Of course, it's possible the Indo-Europeans were trading specifically in war horses and that the Udi term is therefore evidence for that, but that seems to be an even longer bow to draw. (It'd be nice to see what the cognate is in Old Udi, but I don't believe the surviving texts from the Mount Sinai palimpsests include the word for "horse".)

  48. Megan H. said,

    November 11, 2019 @ 10:36 am

    David Morris, I would imagine that contextually, one would assume 배의 세계사 meant the world history of boats rather than of pears, stomachs, or trophies. :)

RSS feed for comments on this post