Sino-Semitica, part 2: of massage and Old Sinitic reconstructions

« previous post | next post »

As part of our research on the dictionary of Middle Vernacular Sinitic (MVS) that Zhu Qingzhi and I have been working on for more than two decades, I was tickled by this quaint poem (below on the second page) by the medieval Buddhist poet, Wáng Fànzhì 王梵志 (Brahmacārin ब्रह्मचारिन् Wang; fl. first half of 7th c.).

I have been an avid fan of Wáng Fànzhì's unique poetry for nearly half a century.  Quaint, indeed, and also quirky.  Wang Fanzhi is self-demeaning in a funny, adorable way.  The poem I'm about to introduce you to is a good example of his trademark self-abnegation.

What attracted me particularly to this poem for the purposes of our research on MVS is the first word in line 2, chǎngtóu 長頭 ("for a long time"), which does not exist with this meaning in Literary Sinitic (LS) / Classical Chinese (CC).  Finding chǎngtóu 長頭 ("for a long time") in Wang Fanzhi's poem was already enough of a treat, but when I got to the last word of the couplet, I was even more delighted.  As you will momentarily see, what Wang says about his wife's tummy is funny by itself, but the word he uses to describe what the wife does to her tummy made me even more excited.

But let's read the poem first, then I'll talk about the word in question, namely, méisuō 沒娑 ("massage").

Jiāzhōng jiànjiàn pín
The Family Gradually Grows Poor

jiāzhōng jiànjiàn pín; liáng yóu yōnglǎn fù
家中漸漸貧; 良由慵懶婦。
The family gradually grows poor; indeed it is because of my lazy wife.

chǎngtóu ài chuáng zuò; bǎo chī méisuō dù
長頭愛牀坐; 飽吃沒娑肚。
She loves to sit in bed for a long time; eats and eats until she is so full that she massages her belly.

yǐnjiǔ wǔ fū dí; bù jiě fèng shān kù
飲酒五夫敵; 不解縫衫褲。
As for drinking she can compete with five men; as for sewing she knows nothing of upper or lower garments.

shì dāng hǎo yīshang; dé biàn zǒu chūqù
事當好衣裳; 得便走出去。
Whenever she is dressed in good clothes; she would seek a chance to go out for a walk.

bùyào nán wèi bàn; xīnlǐ héng pānmù
不要男為伴; 心裏恒攀慕。
She does not want a man to be her companion; (yet) inside her heart she constantly curries their favor.

dōng jia néng niè shé; xī jiā hào hé dòu
東家能涅舌; 西家好合鬬。
She is good at starting rumors at the neighbors on the east; she likes to fight in groups at the neighbors on the west.

liǎng jiā jì bù hé; jiǎo yǎn xiāng qū dù
兩家既不合; 角眼相蛆妒。
After which the two neighbors can no longer get along; they gaze at each other in anger and mutual jealousy.

bié mì hǎo shí duì; chèn què mò jiāo zhù
別覓好時對; 趂卻莫交住。
I will look for a good counterpart for myself elsewhere; chase her away and never interact with her anymore.

Translation by Diana Shuheng Zhang

All right, what's so special about méisuō 沒娑 ("massage")?  It's a plain, quotidian word, n'est-ce pas?  Well, no, it is not just a simple, everyday word.  What triggered my interest about méisuō 沒娑 is that it's a disyllabic word with multiple orthographic forms:  沒娑, 沒挲, 摩挲, 摩娑, 摩莎.  All five forms — and there are probably others — represent one word.  These come across in Modern Standard Mandarin (MSM) with various pronunciations:  méisuō, m éisā, mósā / māsā / māsa / mósuō, mósuō, mósuō / móshā (I can't guarantee that these pronunciations are infallible, because Chinese will pronounce them slightly differently by region and individual).

The third form is already in the historical work known as the Hòu Hànshū 後漢書 (Book of the Later Han), written in the 5th c. AD, but covering the years 6-189 CE.

Readers of Language Log will know that, when I see a long list of variant orthographies for a disyllabic lexeme, I suspect that the word in question may be a borrowing from a foreign language or arose within Sinitic from a non-mainstream source or sources.  In the long list of posts on Old Sinitic reconstructions (see "Selected readings"), the most persuasive cases — including especially where we are able to adduce ancillary supporting evidence from archeology, art, genetics, history, and other fields — are of this type.

For finer grain phonological detail, I asked Diana to comment on the relationships among 沒娑, 沒挲, 摩挲, 摩娑, and 摩莎.  She began by stating that they all represent one word, but not the same one as mōsuǒ 摸索 ("grope, fumble, feel, feel about, feel around"), which bears a superficial resemblance in sound and meaning.

1. 摩 "ma" has an open ending. The reason why it was written as 沒 in 沒娑 may be because of place assimilation. When one makes a slight stop in the middle of "ma-sa" it's easy to pronounce it as "mat-sa" (think of Japanese). However, 摸 "mak" has a -k ending and does not fit into this at all.

2. 沙 "sra" has an open-syllable ending; same for all those with 沙 as phonophore. But 索 "sat" is a ru-tone word with the -t ending, in the word family with 冊, something that "binds".

3. 沒挲, 摩挲, 摩娑, 摩莎 all mean "touching, caressing, massaging, stroking" while 摸索 means "investigating, probing, getting to know (by means of touching)". Therefore I don't think that they are related either in sound or in meaning.

Enough for the ancient sounds of 沒娑, 沒挲, 摩挲, 摩娑, and 摩莎 for now.  I invite Language Log readers who are interested in Old Sinitic reconstructions to comment further.

Casting about for parallels to this old Sinitic lexeme meaning "massage", it didn't take long for me to light upon "massage", which is the very word we use to translate méisuō 沒娑, etc.  When I looked up the etymology of "massage", I was amazed to learn that "massage" derives from an Arabic word.  Even my Arabic colleagues were unaware of the Semitic etymology of the English word.

French, from masser, to massage, from Arabic masaḥa, to stroke, anoint; see mšḥ in Semitic roots or massa, to touch; see mšš in Semitic roots. (American Heritage Dictionary of the English Language, 5th ed.)


To anoint.

1. Messiah, from Aramaic məšiḥā, anointed (one) (from məšaḥ, to anoint) or Hebrew māšîaḥ, anointed (one) (from māšaḥ, to anoint).
2. massage, masseur, from Arabic masaḥa, to anoint, stroke, rub (or massa, to touch; see mšš).


To feel, grope for, touch.

massage, masseur, from Arabic massa, to touch (or masaḥa, to anoint, stroke, rub; see mšḥ).

The m-s root is also in l-m-s in Arabic, which likewise means to touch.

Food for thought.

Selected readings

[Thanks to Joe Lowry]


  1. Victor Mair said,

    February 16, 2020 @ 9:10 am

    The usual, proper word for "massage" in MSM (Modern Standard Mandarin) is ànmó 按摩. I recall that, when I was living in Taiwan from 1970 to 1972, they had another — transcriptional — word for "massage", mǎshājī 马杀鸡 (lit., "the horse kills the chicken"), with salacious undertones. Such things also go on in disreputable barber shops.

    The same usages pertain to the mainland.

  2. y said,

    February 16, 2020 @ 12:15 pm

    Anybody knows the etymological origin of "massage" in Thai and Malay? I checked, and the words seem very different.

  3. Chris Button said,

    February 16, 2020 @ 12:22 pm

    Regarding the word originally represented by 摩 *mál, Pulleyblank (1975) compares the variant 磨 *mál with Proto-Indo-European *melh₂- "grind" as evidence for a common ancestor of Old Chinese and Proto-Indo-European, since it is unlikely to be a loanword. Not buying his hypothesis for a common ancestor (far more evidence would be needed), I personally would tend to regard it as a coincidental similarity and nothing more.

    I wonder of the use of 摩 for 沒 in the above discussion is semantically influenced?

  4. Jinrui Zhang said,

    February 16, 2020 @ 1:25 pm

    This is such an interesting blog! I enjoyed reading this very much.

    There is one little question here:

    '索 "sat" is a ru-tone word with the -t ending, in the word family with 冊, something that "binds".'

    From what we have in the Korean and Cantonese readings of these characters, I believe that 索 and 冊 are actually -k ending words. In Cantonese we have sok and tsaak, and in Korean, we have 색 saek and 책 chaek.

  5. Victor Mair said,

    February 16, 2020 @ 4:11 pm

    From Diana Shuheng Zhang:

    -t and -k endings in Early Han Chinese have many merging instances (not applicable to all -t/-k ending words!). One could look at Yang Xiong 揚雄's Fu rhapsodies for instances; Luo Changpei and Zhou Zumo have a book that lists the rhyming words in the Han Fu. So if the merger occurred that early it would not surprise me about the appearance of only -k.

  6. cameron said,

    February 16, 2020 @ 4:58 pm

    Persian has "massage" as ماساژ. That's an example of a Persian word of Arabic origin that was not borrowed into Persian from Arabic. It's borrowed from the French. You can tell at a glance that it's not from Arabic, because it's spelled with the letter ژ.

  7. Jinrui Zhang said,

    February 16, 2020 @ 6:31 pm

    Thank you very much for the reply! But I am still a little bit confused:

    Since an early instance of this word is in 後漢書, what we have in 漢賦 is probably not that relevant here. In the 切韻 system, t is t and k is k. The Japanese 音読み readings of these end in ku. The Hokkien readings end in k. Here what we have in the Chinese language of 後漢書 and onward is most likely a k coda for 索 and 冊, or else the principles of comparative reconstruction is being compromised.

  8. Chris Button said,

    February 16, 2020 @ 9:21 pm

    I believe that 索 and 冊 are actually -k ending words.

    Yes, they are.

    索…in the word family with 冊, something that "binds".

    hmm…I'm not sure if I buy any semantic relationship between these two. 冊 seems to be operating in a different semantic field. More to the point, the vocalism is off so it wouldn't be a straightforward relationship in any case.

  9. John Walden said,

    February 17, 2020 @ 3:28 am

    I was reminded of 'masa', the Spanish for 'dough' and then 'mass' in its non-religious sense:

    mass: late Middle English: from Old French masse, from Latin massa, from Greek maza ‘barley cake’; perhaps related to massein ‘knead’.

  10. Chau said,

    February 17, 2020 @ 5:29 pm

    In echoing what John Walden has said, I am also inclined to think that mósuō 沒挲 (摩挲, etc.) is related to Gk μάσσειν ‘to knead’ with an a>o vowel sound change. It is understood to mean ‘to knead dough.’ In mósuō 沒挲, the aching body becomes the dough. A related Latin word is massa ‘a lump of dough’ as in the Vulgate, Galatians 5:9, “totam massam corrumpit” ‘spoils the whole lump’ / [New International Version] ‘leavens the whole dough’.

RSS feed for comments on this post