188.8.131.52:36: AN ESEN-TIAL SEMANTIC SPECTRUM
Here is my attempt to diagram the semantic overlap between the words in the last two entries and the Japanese root yasu as an unrelated external 'control' for comparison:
|Khitan esen||Mongolian esen
(Lessing 1960: 333)
(Clauson 1972: 248)
|Middle Persian āsān
(MacKenzie 1986: 12)
|Arabic ḥasan||Japanese yasu|
|安い・廉い yasu-i 'cheap'|
|easy||易い yasu-i 'easy'|
|at rest||休む yasu-m-u 'to rest'|
|calm, quiet||peaceful||安らか yasu-raka 'peaceful'|
|sound, safe||安らか yasu-raka 'sound, safe' (obsolete)|
|healthy, good health||in good health||(康 yasu 'health')|
I have tried to arrange the various meanings on a color-coded rainbow spectrum. Of course, there is no reason for semantics to be easily described in terms of a single dimension (or even two!), and other arrangements are possible.
'Good' is so broad that I assigned no color to it. Unlike John Tang, I don't think ḥasan 'good' has anything to do with the s-n words to the east.
I wonder if Japanese yasu- 'cheap' originated as 'easy to obtain'. The semantic range of yasu- and its derivatives in modern times is between 'cheap' and 'peaceful', but in earlier Japanese bodies could be yasu-raka 'safe and sound' (arguably even 'healthy'), and even in modern Japanese the name element yasu can be written with the Chinese character 康 'health' (but not 壽/寿 'longevity'!). The entire range (apart from 'good') could be defined in terms of the absence of ...
cost, which is a kind of ...
difficulty, which is a common characteristic of ...
work, which may lead to (is there a better transition?) ...
conflict, which can involve ...
danger, which may result in ...
injury, which is like ...
disease, which also harms the body, and may lead to ...
death, which is the ultimate harm.
The Persian and Turkic meanings do not intersect, though that by itself is not sufficient to argue that āsān and esen are unrelated, since 'peaceful' could shift to 'safe'. However, the semantic difference coupled with the vocalic mismatch make me skeptical about esen as a borrowing of āsān.
Mongolian shares 'health' in common with Turkic but not 'calm, quiet'. What is the earliest attested meaning? If it is 'healthy', then the Turkic word was borrowed in a narrow sense and the range of the Mongolian word later expanded into the 'green zone'.
I assume 'longevity' is a Khitan-internal extension of 'health' which might have been the original meaning of esen in the common ancestor of Khitan and Mongolic that borrowed the word from Turkic. Or was the direction of borrowing in the other direction? A third possibility is that both borrowed the word from Xiongnu, but of course that hypothesis cannot be tested because the Xiongnu word for 'health' is unknown (along with 99.9% of the Xiongnu lexicon). Is there anything like esen 'health' in the Yeniseian languages which may be related to Xiongnu?
184.108.40.206:55: PEACEFUL BUT NOT ESEN-TIALOn Sunday I asked,
Does [Middle Persian] āsān have an Indo-European or at least Iranian etymology?
On Monday I found the entry for that word in Nyberg's A Manual of Pahlavi (1964-74 II: 30) which derived it from āsāy- 'to rest' and led me to Bailey (1930: 16) who derived it from an Iranian root *sam, equivalent to *samH in Cheung's (2007: 330) modern Proto-Iranian reconstruction.
The Sanskrit cognate of *samH is śami. (I like Cheung's use of superscript i to indicate the Sanskrit i that is a reflex of *H.)
Proto-Iranian *samH and Sanskrit śami in turn come from Proto-Indo-European *kemʕ. The only English word with that root in Watkins' (2011: 41) American Heritage Dictionary of Indo-European Roots is nosocomial. I never even saw the word until yesterday.
There is one problem with that internal etymology, though. I don't understand how Bailey got from *sam to -sāy-. (Ā- is an Indo-Iranian prefix.)
Cheung (2007: 328) avoided that problem by deriving Manichaean Middle Persian sy-* (cognate to Pahlavi [Cheung's "Zoroastrian Middle Persian"] -sāy-) and Pahlavi āsān (a present participle) from Proto-Iranian *saiH 'to lie down, go to sleep'. The Sanskrit cognate of that root is śayi, and the Proto-Indo-European source of that root is *keiʔ with the following descendants in English (Watkins 2011: 39):
- Native: hind 'British farm assistant', hide 'unit of land area' (both as unknown to me as nosocomial!)
From Irish: ceilidh
From Latin (directly or otherwise): city, civic, civil; incunabulum (yet another new word for me - I'm referring to the last one, of course!)
From Greek: cemetery
From Sanskrit: Shiva
Given that internal etymology of āsān, there is no need to derive it from Arabic ḥasan.
I am still not entirely convinced that Khitan/Mongolic/Turkic esen is from Middle Persian becaue of the vowels. Are there any other cases of Turkic e corresponding to Middle Persian ā? Clauson (1972: 248) wrote that "The spelling asan, which is common in Uyğ. is prob. an aberration." What if asan is an Uyghur borrowing from Middle Persian while esen is an unrelated native soundalike in Turkic?
*The Manichaean script is an abjad, so I guess sy- was pronounced [saːj], but I don't really know.
220.127.116.11:55: AN ESEN-TIAL ETYMOLOGY
On Friday I found John Tang's paper "On the Terms Concerning Longevity in Khitan and Jurchen Languages". He proposed that the Khitan word
<es.en> (in the large and small scripts) 'longevity'
a probable etymological explanation: [Mongolian] esen 'healthy, good health; calm, quiet' [see Lessing 1960: 333] < Uig. äsän 'in good health, sound, safe' [see Clauson 1972: 248] < Pahl. 's'n [âsân] 'at rest, easy, peaceful' [see MacKenzie 1986: 12]. It is related to the personal name Ḥasan, which is common in Arabic and Persian languages (Rybatzki 2006: 176-177). This connection suggests that Khitan, the so-called “Para-Mongolic” (Janhunen 2003: 391-402), had been partly influenced by Near Eastern culture, as well as by the significant Uighur compacts. (p. 483)
I am not sure all of that holds up.
I do think the Khitan/Mongolic and Turkic terms are connected. Perhaps esen is an early loan from Turkic into the common ancestor of Khitan and Mongolic.
But is esen native to Turkic? Clauson (1972: 248) wrote that Turkic esen was "[n]ot to be confused with Pe[rsian] āsān 'easy' ": i.e., the descendant of Pahlavi/Middle Persian āsān. I suppose he did not regard the Turkic word as a borrowing from Persian. I am hesitant to do so, as I don't understand why Persian ā would have been borrowed as e instead of a in Turkic.
Then again, Clauson also wrote "but see Doerfer II 478": i.e., Türkische und mongolische Elemente im Neupersischen. According to Doerfer, the Middle Persian word was not only borrowed into Turkic but also borrowed from Turkic back into New Persian as اسن esän (not isan in his Persian romanization)?
In any case, I doubt Arabic ḥasan has anything to do with the above words for four reasons (revised 7.21.1:03):
1. Ḥasan means 'good', not 'healthy' or 'safe'. A semantic shift of 'good' to 'healthy' is not impossible (e.g., English feel good), but the semantic fit is nonetheless loose.
2. MacKenzie (1986: 12) did not list Middle Persian āsān as a borrowing from Arabic. (Does āsān have an Indo-European or at least Iranian etymology?)
3. I would expect Arabic ḥ- to correspond to Middle Persian h- rather than zero. (Cf. how Arabic ḥ- is pronounced [h] in New Persian.)
4. Esen is widespread within Turkic, so it might be reconstructible at the Proto-Turkic level. And esen might also be reconstructible for the common ancestor of Khitan and Mongolic (unless it was borrowed from Turkic by Khitan and Mongolic after their breakup). If Persian āsān were a borrowing from Arabic, it would have to date from the Islamic conquest in the seventh century AD or later: i.e., long after the breakup of Proto-Turkic and the split of the common ancestor of Khitan and Mongolic. (7.21.7:24: Is āsān attested in Persian before the seventh century?)
I conclude that ḥasan and āsān are unrelated lookalikes.