The Tocharian A word for "rug" and Old Sinitic reconstructions

« previous post | next post »

There's a Chinese character 罽 (Mandarin jì, Old Sinitic *kràts), which means "rug, carpet; woolen textile; fish net").  On the basis of its sound, meaning, place, and date of occurrence, it would seem to be related to Toch. A kratsu "rug".

This raises two questions:

1. Does this Tocharian word have cognates in other IE languages?

2. Who borrowed it from whom?   Sinitic from Tocharian or Tocharian from Sinitic?

I knew this character long ago as part of an old (medieval), disyllabic Chinese designation for Kashmir, viz., Jìbīn 罽賓.  See Paul Kroll, A Student's Dictionary of Classical and Medieval Chinese, p. 189b and Wiktionary.    But I think this is a transcriptional usage of the first character and semantically has nothing to do with the older (at least 100 AD) word meaning "rug", etc. that I consider to be linked somehow to the Tocharian A word.  Kroll says that the first syllable by itself means "haircloth; coarse felt".

In 2015, Hannes Fellner gave a talk at the School of Oriental and African Studies (London) that touched on this subject:

"Hannes A. Fellner (Vienna): Linguistic Contact between Indo-European and Old Chinese" (29:03)

The talk focused on Tocharian, and mentioned the question of the relationship between Old Sinitic *kràts ("rug, carpet; woolen textile; fish net") and Toch. A kratsu ("rug") (see at 19:32-20:48).  In the context of all the other evident Tocharian loanwords in Sinitic that Fellner assembles, it would appear that Toch. A kratsu ("rug") may indeed be the source of Old Sinitic *kràts ("rug, carpet; woolen textile; fish net").

Doug Adams offers the following pertinent observations:

TchA kratsu has an exact equivalent in TchB kretswe, both reflecting something like PTch *krætsäwæ or the like.  The <æ> is sort of a graphic compromise between B <e> and A <a> (I usually use *<e> however), but it has some actual phonetic support in that thus is the vowel which is used in borrowings from early Iranian into Tocharian for Proto-Iranian short /a/.  This vowel has a history in Iranian of being quite front, or æ-like, even unto Modern Persian.  Certainly the PTch form was trisyllabic, as /ts/ only develops before some sort of front vowel.

I don't have any particular insight into the meaning of the TchA word, though the most reasonable assumption is that it means much the same as the B word—all translators have agreed on that.  In B (and A) it is glossed as (Germ.) 'Lappen' or (Eng.) 'rag/piece of cloth,' and not 'rug.'  The one good example in B is: ṣamānentse yśel[mi pä]lskone tsaṅkaṃ kwipe-ike keuwco kaltärr-ne [sic] tu kretswesa yaṣtär '[if] sexual desires should arise to a monk and his "shame-place" stand tall and he should stimulate it with a rag' (334b2/3E/C).  This is a text which describes various (prohibited) forms of masturbation.  One can easily imagine masturbating with a rag or piece of cloth, but surely not a rug.

If the Chinese word is related, the semantic change must be from 'piece of cloth' to 'rug' rather than vice versa.  A Navajo rug is after all a piece of cloth woven from stout yarn used as a coverlet or rug.  The history of rug itself, I believe, shows a similar semantic development.

Wiktionary:

Origin uncertain; probably of North Germanic origin, compare dialectal Norwegian rugga ("coarse coverlet"), Swedish rugg ("rough entangled hair"), from Old Norse rǫgg ("shagginess; tuft"), from Proto-Germanic *rawwō ("long wool"), related to English rag and rough.

Etymonline:

1550s, "coarse fabric," of Scandinavian origin; compare Norwegian dialectal rugga "coarse coverlet," from Old Norse rogg "shaggy tuft," from Proto-Germanic *rawwa-, perhaps related to rag (n.) and rough (adj.). Sense evolved to "coverlet, wrap" (1590s), then "mat for the floor" (1808). Meaning "toupee" is theater slang from 1940. Cut a rug "dance" is slang first attested 1942. To sweep (something) under the rug in the figurative sense is from 1954. Figurative expression pull the rug out from under (someone) "suddenly deprive of important support" is from 1936, American English. Earlier in same sense was cut the grass under (one's) feet (1580s).

The oldest surviving Chinese dictionary, Ěryǎ 爾雅 (Approaching Elegance) (circa 3rd c. BC) has the character *kràts 罽  with the apparent meaning ("[textile made of] coarse wool"), which would put it around the same time as Tocharian A kratsu and with roughly the same meaning.  The semantic drift to "rug" in both languages would have occurred in parallel fashion.

I'm writing this at 35,000 feet as I fly from Seattle to Dallas, so I do not have access to my library.  I'm hoping that earthbound colleagues will provide more precise data that will enable us to determine more definitively when the borrowing took place and in which direction.  All things considered (including importation of rug-weaving technology) at this point and from this vantage, my inclination is to opt for a borrowing from Tocharian into Sinitic during pre-Han (3rd c. BC and earlier) times, the same period as the famous borrowing of the Tocharian word for "honey" into Old Sinitic (see this post on "mead", etc.).

Readings

[Thanks to Chris Button]



10 Comments

  1. Chris Button said,

    January 3, 2020 @ 6:34 pm

    All things considered (including importation of rug-weaving technology) at this point and from this vantage, my inclination is to opt for a borrowing from Tocharian into Sinitic during pre-Han (3rd c. BC and earlier) times,

    It's worth adding in this regard that the Shuowen glosses the variant form as 西胡毳布也 "Western foreigner/barbarian hairy/furry cloth".

  2. Chris Button said,

    January 3, 2020 @ 6:36 pm

    The variant form should have appeared as 糸 on the left of 罽

  3. Chris Button said,

    January 3, 2020 @ 10:11 pm

    Regarding the interesting graphic form, I wonder if the base under ⽹ at the top is a corruption of 毯 as the variant form listed here would suggest: https://zh.wiktionary.org/wiki/%F0%A6%8C%93 .

    I'd sort of assumed 毯 was a native word on the basis of possible word-family associations, but Zhang He suggests otherwise here: http://sino-platonic.org/complete/spp257_carpet_terminology.pdf . The conflation of 氊 with 毯 on p.11 there unfortunately ignores the earlier *-m coda in 毯 (at least a reconstructed Old Chinese lateral onset *ɬ- based on the phonetic 炎 *làm/*lám, could plausibly have been tʰ- by the time of its earliest attestation), so the comparison of a Persian root tan- with 氊 by Laufer (whose typo 檀 is duly noted) probably can't be equally applied to 毯.

  4. Chris Button said,

    January 3, 2020 @ 10:14 pm

    Regarding the interesting graphic form, I wonder if the base under ⽹ at the top is a corruption of 毯 as the variant form listed here would suggest: https://zh.wiktionary.org/wiki/%F0%A6%8C%93 .

    I'd sort of assumed 毯 was a native word on the basis of possible word-family associations, but Zhang He suggests otherwise here: http://sino-platonic.org/complete/spp257_carpet_terminology.pdf . The conflation of 氊 with 毯 on p.11 there unfortunately ignores the earlier *-m coda in 毯 (at least a reconstructed Old Chinese lateral onset *ɬ- based on the phonetic 炎 *làm/*lám, could plausibly have been tʰ- by the time of its earliest attestation), so the comparison of a Persian root tan- with 氊 by Laufer (whose typo 檀 is duly noted) probably can't be equally applied to 毯.

  5. Chris Button said,

    January 4, 2020 @ 6:55 am

    Continuing a little more on the graphic form of 罽…

    The Shuowen associates 㓹 with 銳 *lʷàts "sharp". Although phonologically unrelated, 剡 *ɬàmʔ/*làmʔ means "sharp" and has the same phonetic 炎 as in 毯. In 剡, the phonetic 炎 is really being used for 覃 *lə́m, which is phonetic in 簟 lʲámʔ "(bamboo) mat", albeit itself really being used for 㐁 which was originally a depiction of a mat…

  6. Victor Mair said,

    January 4, 2020 @ 9:02 am

    From Håkan Wahlquist:

    I may mention that there is s Swedish verb related to rug; "rugga". A word for how birds shed their old feathers for new ones, but also for how the surface of a woven glove can be treated to make its surface less smooth. The same, in the olden days, to treat a pair of skis underneath to make them better suited for climbing uphill (not slipping back).

    http://www.saob.se/artikel/?unik=R_2647-0312.kUc5

  7. Chris Button said,

    January 4, 2020 @ 10:47 am

    In 剡, the phonetic 炎 is really being used for 覃 *lə́m, which is phonetic in 簟 lʲámʔ "(bamboo) mat", albeit itself really being used for 㐁 which was originally a depiction of a mat…

    Still going a little off topic into other rug-type items, but one of the concerns I have with associating 毯 *ɬámʔ "blanket" directly with 簟 lʲámʔ "(bamboo) mat" is the palatal -j- in the latter, which suggests external (dialect?) influence or chance convergence. The word "blanket" in English is associated with "blaze" (via "blanc" as white cloth) which does fit 毯 nicely within the 炎 word family (if it pertained to white cloths there too?). Meanwhile 簟 fits better with words like 恬 "tranquil" and 甜 "sweet", both *lʲám, via the concept of softness (cf. douce, dulcet/dolce and French doucette "fine cloth")

  8. Ivan Zakharyaschev said,

    January 4, 2020 @ 10:18 pm

    As for a possible IE etymology, the following proposal is mentioned in the dictionary by Douglas Adams (found online at https://www.win.tue.nl/~aeb/natlang/ie/tochB.html ):

    kretswe* (n.) '± rag'
    [-, -, kretswe//] ∎TchA kratsu and B kretswe reflect PTch *kretswe but extra-Tocharian cognates are uncertain. VW:233 suggests a connection with PIE *ker-t- 'cut'; H:180-181 a PIE preform *krodhiwo- and compares Sanskrit kṛdhú- 'stumped, shortened' (comparative kradhīyas-), where *(s)kredh- is presumably an élargissement of *(s)ker- 'cut.'

    The suggested preforms make me wonder a bit, although one should note that I'm not professional in this field. Let me tell what I find strange with PIE preform *krodhiwo- from the passage cited by me or PTch *krætsäwæ cited by you with the note "Certainly the PTch form was trisyllabic, as /ts/ only develops before some sort of front vowel". Elsewhere https://en.wikipedia.org/wiki/Tocharian_languages#Palatalization , one can read that ts develops only before y, but not the vowels:

    "the result of palatalization of t and dh before y is different from palatalization before e, ē and i, while other consonants do not show such a dual outcome. (A similar situation occurred in the history of Proto-Greek and Proto-Romance.)"

    An example of the different development before i (rather than y) can be found in the dictionary by Douglas Adams. Consider an example (that I have just found by chance, without a very thorough search):

    kerccī (n.[pl. tantum masc.]) 'palace'
    [//kerccī, -, kerccīyeṃ] ∎Etymology uncertain. It is possible (with VW:215) that we have a putative PIE *ghortiyo- related to Greek khórtos (m.) 'enclosed place, feeding place,' Latin hortus (m.) 'garden,' Old Irish gort 'standing crop,' Latin cohors 'enclosure, yard,' possibly English garden (cf. P:442-443; MA:199). However, there seems to be no reason on phonological grounds to reject a putative PIE *ghordhiyo- (as suggested by both Meillet, in Hoernle [1916:379] and Lidén [1916:21-2]) and thus a relationship with OCS gradъ (m.) 'city,' Sanskrit gṛhá- (m.) 'house, habitation, home,' Gothic gards (m.) 'house,' Old Norse garδr (m.) 'fence, hedge, court,' Old English geard (m.) 'enclosure, yard,' Lithuanian gar̃das (m.) 'fold, pen,' Phrygian -gordum 'city,' and Górdion 'Gordium' (P:444; MA:199). Except that Tocharian kerccī is plural it would match Phrygian Górdium exactly. In any case PIE *ghort- and *ghordh- are likely to be phonologically conditioned variants of what was originally a single paradigm with a nominative singular *ghórts (with automatic devoicing) and a non-nominative stem *ghordh-. In a variation of this proposal, Isebaert (apud Thomas, 1985b:150) suggests that we have here a borrowing from a Middle Iranian *gardiya-).

    So, here we see a development into c rather than ts before i.

  9. Ivan Zakharyaschev said,

    January 4, 2020 @ 10:42 pm

    I have another hypothetical suggestion of a possible IE etymology of TchA kratsu and B kretswe
    after having looked around and having read the proposal mentioned in
    the dictionary by Douglas Adams (found online at
    https://www.win.tue.nl/~aeb/natlang/ie/tochB.html ):

    kretswe* (n.) '± rag'
    [-, -, kretswe//] ∎TchA kratsu and B kretswe reflect PTch *kretswe but extra-Tocharian cognates are uncertain. VW:233 suggests a connection with PIE *ker-t- 'cut'; H:180-181 a PIE preform *krodhiwo- and compares Sanskrit kṛdhú- 'stumped, shortened' (comparative kradhīyas-), where *(s)kredh- is presumably an élargissement of *(s)ker- 'cut.'

    We might consider the root *ger "to weave" found in the StarLing
    database
    http://starling.rinet.ru/cgi-bin/response.cgi?single=1&basename=%2fdata%2fie%2fpiet&text_number=++1681&root=config
    :

    Proto-IE: *ger-

    Meaning: to weave

    Old Indian: guṇá- m. `a single thread or strand of a cord or twine, string, rope';
    jaṭā f. `hair twisted together; fibrous root';
    jāla- n. `net, snare, cob-web, any woven texture';
    gala- `reed, rope of reed (L.)';
    gárta- m. `high seat, throne'

    Baltic: *geran̂k=, *garan̂k= (2); *gerg=, *gerk=

    Germanic: *kar-m-a- m.

    Russ. meaning: плести

    References: WP I 593 f

    Technically, deriving from *ger "to weave" mustn't be worse than deriving from
    *(s)ker- "cut". (My wonder about the conditions for palatalization in
    TchA kratsu and B kretswe was expressed in the previous comment by
    me.)

    But semantically "to weave" might be nicer (you also mention weaving
    as the thing that could be borrowed also culturally; so the term might
    have designated specifically weaving). Also, one of the
    derivations supposed by the authors of StarLing is quite close in
    meaning:
    Old Indian jāla- n. `net, snare, cob-web, any woven texture'.

    Their supposed way of deriving jāla- is not quite clear to me, of
    course, and hence doubtful.

    And it can't be identical to the Tocharian form anyway, particularly
    due to the Old Indian palatalization g > j before a proto-IE front
    vowel, which doesn't match the Tocharian form, which lacks the
    palatalization of the first consonant.

    (As for the possible way of deriving jāla-, I've just written a bit more
    details at https://linguistics.stackexchange.com/a/34774/443 ; I've
    found that question during the googling inspired by your blog posting.

    Other IE forms that are assumed to be related in the StarLing database
    and have somewhat close meaning are also mentioned in my post there.
    This could lead someone to some further thoughts.)

  10. Lugubert said,

    January 5, 2020 @ 6:09 pm

    Wiktionary's Swedish rugg ("rough entangled hair") was quoted above. Given that gloss, I'd first think of 'ragg', especially on goats ('getragg'). The pronunciation is close to English rug.

RSS feed for comments on this post