Written Mongolian čamča ‘shirt’ etimology

Per Urales ad Orientem. Iter polyphonicum multilingue.

Festskrift tillägnad Juha Janhunen på hans sextioårsdag den 12 februari 2012.

Suomalais-Ugrilaisen Seuran Toimituksia = Memoires de la Societe Finno-Ougrienne 264.

Marek Stachowski

Uniwersytet Jagielloński

Written Mongolian čamča ‘shirt’ and

its etymological counterparts in Europe

Seventy five years ago G. J. Ramstedt wrote on the origin of Kalmyk tsamtsɒ

‘shirt’ as follows: “viell. *samča zu ma. samsu ‘hanfgewebe’, kor. sam ‘hanf’”

(KW 421b).1 Approximately a decade later, in 1947, Leonardo Olschi suggested

that the Written Mongolian word čamča ‘shirt’ possibly reflected a Greek etymon

and was connected with French chemise and Italian camicia id. (cited after

de Rachewiltz 2004: 309). Two years later, G. J. Ramstedt repeated his opinion

that this word is a reflex of a proto-form *samča, being a derivative of a *sam in

the proto-language, attested only in Korean. Thus, *sam (> Kor. sam ‘hemp’) >

*samča > Mo. čamča ‘shirt’ (~ Ma. samsu ‘thin hempcloth’) > Ma. Nan. čamči

‘shirt’ (SKE 222).

Neither Ramstedt nor Olschi explained the sound changes and morphological

problems involved. In this situation, readers had at their disposal no precise

arguments for or against Ramstedt and Olschi. It could thus be expected that

some of them preferred the European etymology and others the Altaic. Indeed,

this was the case.

Pavel Poucha (1956: 47sq.) devoted a discussion of some length to this Mongolic

word. However, some of his formulations are not really clear and unequivocal

or easily acceptable:

(1) “[…] nach Ramstedt vielleicht aus sam-ča zu manǰ. samsu ‘Hanfgewebe’,

kor. sam ‘Hanf’, dann könnte man das mong. čamča aus dem

Chinesischen herleiten, wo ṣan < ṣām ‘Hemd’ vorkommt.” (Poucha

l.c.).

There can be no question that Ramstedt’s task when publishing

his SKE was to show the Altaic genetic unity, rather than Chinese

loanwords in Korean and Mongolic.

(2) “Obwohl Ramstedts Etymologie verführerisch aussieht, so ist doch

zu erwägen, daß dieser gemeinmongolische2 Ausdruck dem europäischen

‘Hemd’ sehr nahe kommt: neugriech. (ὑπο) κάμισον aus

1. Ramstedt used the sign () to indicate that the preceding part of a word is a nominal stem. Thus, his

*samča is what we would today write *sam+ča, in contrast to the deverbal derivative *sam-ča. However, in

the fragment cited after Poucha below, the word is traditionally transcribed *sam-ča.

2. Poucha (l. c.) adduces the following Mongolic forms: Dörbet-bejse šams(e), “udschumtsin”

(= Üǯümüčin) čamč(i), “dschastu” (= Ǯasagtu) šamži, Ordos čamča ‘shirt’ and Dagur čhančhi ‘overall, coat’.

The modern Khalkha form camc, Buryat samsa and Kalmuk camcɒ ‘shirt’ (mentioned as camca in Poucha

l.c.) are to be added here.

Helsinki 2012. 445–451.

446 Marek Stachowski

dem Romanischen, spanisch camis(ol)a, franz. chemise aus vulgärlat.

camĭsia […] ‘Männerhemd’ < Gall. < Germ. (urgerm. *kamitja > ahd.

hemidi > nhd. Hemd) […]. Und so wäre es nichts Außerordentliches,

wenn man annehmen wollte, was ich anzunehmen geneigt bin, nämlich,

daß das mongolische Wort čamča letzten Endes mit deutschem

‘Hemd’ verwandt ist […]” (Poucha l.c.).

Poucha, like Olschi, does not explain the sound differences

observed here, nor does he settle the order of transitional languages.

One cannot even determine what specific European word is to be

regarded as the etymon of the Mongolic word.

In 1969, Martti Räsänen (VEWT 98a) continues, as was only to be expected, the

“Altaic tradition” in that he repeats Ramstedt’s suggestions, albeit in a somewhat

more cautious way and without reconstructed forms: “mo. […] čamča ‘Hemd’

(> ma. čamči ‘Weiberhemd’) ~ ma. samsu ‘Hanfgewebe’ ~ kor. […] sam ‘hemp’”

(VEWT 98a). The real relationship between these forms remains unclear.

In the early 21st century Igor de Rachewiltz (2004: 309) derived the Mongolic

word ‒ with a question mark ‒ from Chinese shan-tzu ‘woman’s dress,

shirt’, and this explanation was accepted by V. Rybatzki (2006: 307b). Thus, a

third etymological suggestion emerged.

An element common to all these explanations is the fact that Turkic data

are always reported to be loans from Mongolic. This does not of course mean

that words like Oyr., Leb. čamča ‘Hemd’, Tel. čamča ‘Rock’, Brb. camca ‘Kleid’,

Eastern Tkc. čamča ‘hemdartiges Kleidungsstück’, Saryg-Uyg. čamǯa ‘кафтан,

верхняя одежда’, Čag. čumča ‘Hemd’ (VEWT 98a); Sal. čamǯa ‘surtout simple

des femmes; chemise; pèlerine’ (Drimba 1976: 418) have not been borrowed

from Mongolic. They certainly have. Nevertheless, the opposite borrowing direction

seems possible as well, if one is ready to consider one further source of

the Mongolic word, namely the word čamašyr ‘underwear’,3 present in numerous

Turkic languages, i.e. Turkic > Mongolic > Turkic.

The Uygur language with its y > i palatalization4 and loss of word-final

r (čamašyr > ºčamaši)5, as well as vowel raising (ºčamaši > ºčamiši)6 seems to

3. For the semantics cf. Fr. chemise and its English reflex chemise.

4. In Turkic words, the letter ‹y› stands for the velar counterpart of i (i.e. = Tksh. ı, Russ. ы), often rendered

also by ‹ï› in other Turkological works. In non-Turkic examples, ‹y› = i̯.

5. A degree symbol (º) is used to signal that a form is a modern and perfectly possible, although unrecorded

variant, rather than a protolinguistic reconstruction (see Anikin 1997).

6. Both phenomena can easily be observed in the Uygur name of the desert in the Xinjiang Uygur

Autonomous Region in China: (Täklimakan < and ~) Taklimakan < *Taklamakan < *Taklarmakan, lit. ‘place

of arches’ (< taklar ‘arches’ [pl. < tak ‘arch archit.’] + makan ‘place’) because winds occasionally blow the

sands away and expose some remains of old buildings, namely arches jutting out of walls (Jarring 1997: 447).

The modern pronunciation Täkli… (instead of Takli…) results from a secondary vocalic harmonization. It

cannot possibly be interpreted as the result of the so-called “Uygur umlauting” because this process does

not work “before an i that is the result of raising” (Hahn 1991: 51). Even if one assumes that i in Takli… was,

at some stage, no longer perceived as secondary (easily imaginable in an old compound), was no longer etymologically

transparent and used only as a geographical name, the umlauting would yield an e ‒ i sequence

(as in *baš ‘head’ + -im ‘my’ > Uyg. bešim ‘my head’), rather than ä ‒ i (Hahn l.c.).

Written Mongolian čamča ‘shirt’ and 447

its etymological counterparts in Europe

best fit into the imaginable borrowing channel and the train of sound changes,

because the subsequent syncope of a narrow vowel in the second syllable of

a three-syllable word, usually called Mittelsilbenschwund in Turkic linguistics

(ºčamiši > ºčamši) is quite a regular tendency on the brink of being a rule. The

frequent alternation -a ~ -y (> -i) makes the occurrence of ºčamši and ºčamša side

by side quite possible. The only problem is that neither ºčamši nor ºčamiši could

actually be found in Uygur. On the other hand, Eastern Turkic (“turc oriental”)

forms like ǯumǯār (Zenker 375a: ǵ umǵ ar راجموج ‘chemise, vêtement / Hemd,

Kleid’) and ǯumǯāh (Zenker 365b: ǵ umǵ ah هاجمج ‘chemise / Hemd’) seem

to support our conjecture about the existence of a former Persian-Turkic form

ºčamašy(r) ~ ºčamša or the like.

Provided that we accept this etymological possibility for the time being,

we may go a step further back because the Turkic word čamašyr ‘underwear’

is a loan from Persian. Interestingly enough, most sources adduce only Pers.

ǯāmašuy ‘laundress’ as the source of the Turkic word (the non-trivial semantic

change has probably never been explained or at least discussed7). A. Tietze

(2002: 471) even resorts to an inorganic (i.e. non-etymological) r that as hypercorrection

occurs after a vowel and refers to the Turkish word pair alengilli ~

alengirli (argot) ‘distinguished, noble’. Some mistakes have to be corrected here:

(3) In his Turkish formulation (“inorganik bir /r/” = ‘an inorganic /r/’),

Tietze, for unknown reasons, uses a phonological notation /r/, which

certainly is incorrect in this context. As a matter of fact, an etymologist

does not care whether an inserted unetymological consonant is a

phoneme or an allophone in the given linguistic system.

(4) In case of alengilli ~ alengirli one should invoke a dissimilation (ll >

rl) or assimilation (rl > ll), i.e. focus on the consonant cluster, rather

than the postvocalic position of the r. Besides, there is no “inorganic

r” in alengirli, even if this variant really goes back to alengilli.

(5) Since the etymology of alengilli ~ alengirli remains unknown, one

cannot decide whether -r- actually is secondary here (cf. Tietze 2002:

149ab).8

Additionally, the assumption of a hypercorrect insertion of r is, in point of fact,

totally unnecessary. The Persian word consists of ǯāma ‘clothes, dress; clothing,

apparel’ and šuy, the present tense stem of šustan (~ šostan) ‘to wash’. But this

stem has, in reality, three forms: šu, šuy and šur (PRS 2: 101), so that one can

expect ǯāmašuy to have two other variants as well. Indeed, Pers. ǯāmašur ‘laun-

7. However, this is not the only case of such a change. An interesting parallel is Turkish kaşar ‘(wheel

of) fat sheep milk cheese’ < Romanian cășár ‘a shepherd who produces cheese’ (< caș ‘fresh, i.e. unsalted

sheep milk cheese; single wheel of cheese’). I would like to sincerely thank Corinna Leschber (Berlin) for

her help with the Romanian word material.

8. The explanation of /r/ (written again as a phoneme) given by Tietze sub alengilli requires further critical

commentary (for one, Turkish çamaşır is the only example where r is not followed by a consonant). This

would, however, lead us too far astray.

448 Marek Stachowski

dress’ is attested too (PRS 1: 425; this form is also given in PLOT). The only

form I could not find is the variant ºǯāmašu but this seems to be nothing more

than a phonetic variant of ǯāmašuy.

It is of no great importance whether we take Pers. ǯāmašur, ǯāmašuy or

ºǯāmašu as our starting point. The Uygur reflex would have probably always

been ºčamašu, and this would in its turn change, according to rules of the Turkic

vowel harmony, into ºčamašy. The further phonetic evolution of the word is suggested

above.

In view of these data, Germ. Hemd ‘shirt’ cannot be considered a European

cognate of Mo. čamča id. This does not, however, mean that no correspondence

of čamča is known in Europe. Another Persian derivative of ǯāma ‘clothing’

is ǯāmadān ‘1. wardrobe; 2. portmanteau, suitcase’. This word was borrowed

into Russian as чемодан ‘portmanteau, suitcase’, probably via some Turkic

language(s). Thus, čem- in the Russian word чемодан is the European etymological

counterpart of čam- in the Written Mongolian word čamča.

On the other hand, Pers. ǯāmadān was also borrowed into Manchu as

čamda ‘portmanteau, suitcase’ (Anikin 1997, 2000 s.v. чемодан), so that this

language has reflexes of two derivatives of Pers. ǯāma ‘clothing’:

Russ. чемодан = Ma. čamda ‘suitcase’ < Pers. ǯāmadān id. < ǯāma

‘dress’ > ǯāmašu(y) ~ ǯāmašur ‘laundress’ > various Turkic languages

čamašyr ‘underwear’ ~ Uyg. ºčamašu ~ (*čamašy >) ºčamaši ~

ºčamaša > ºčamišu ~ ºčamiši ~ ºčamiša > ºčamšu (> Ma. samsu ‘hemp

fabric’) ~ ºčamši (> Üǯümüčin čamč(i), Ǯasagtu šamži, etc.) ~ ºčamša

(> Written Mo. čamča ‘shirt’ > Khamnigan Evenki camca id. [Janhunen

1991: 104]).

Yet another trace of Pers. ǯāma is hidden in English pyjamas ~ pajamas,9 a word

borrowed ‒ via Urdu ‒ from Pers. pāǯāma (PRS 258a) ~ pāyǯāma (PRS 271a) ~

payǯāma (PRS 321a) < Pers. pā(y) ~ pay ‘foot; leg’ + ǯāma ‘clothing’.10 Since this

English word was afterwards borrowed into numerous languages, an etymological

counterpart of Mongolic čamča can easily be found virtually all over Europe.

An additional problem to be solved in the future is whether the Siberian

Turkic forms like Oyr. čamča ‘shirt’ should be better derived directly from Uyg.

ºčamša or via Mongolic čamča. At least the č ‒ č sequence in Oyrot etc. seems

to point towards Mongolic mediation.

Yet another problem is whether both the phonetic form and the meaning of

Ma. samsu ‘hemp fabric’ actually were influenced by Kor. sam ‘hemp’ ‒ a question

that I do not feel competent to answer.

9. I would like to warmly thank Andrzej Pisowicz (Kraków) for directing my attention to this reflex of

the Persian word.

10. The fact that Pers. pāy+ǯāma originally was a piece of clothing that covered legs, i.e. a sort of trousers

is also reflected in the structure of Engl. pyjama+s like trouser+s, drawer+s, breech+es.

Written Mongolian čamča ‘shirt’ and 449

its etymological counterparts in Europe

In any case, if these words are cognates in an Altaistic spirit, the wordinitial

s- is original, and the č- of all other variants must be explained. If they are

not, the origin of the Manchu s- is to be explained because Manchu does tolerate

a word-initial č-, and this appears to be a case more complicated than the former

one.

Furthermore, there exists a homonym čamča (~ čömče) ‘spoon, ladle’ in

Turkic, attested also in Persian (čumča id.). The etymology and the conduits of

transmission (TMEN 3: 95 Nr. 1121; Doerfer 1968–69 Nr. 68: Ar. čímča) are not

ultimately settled, and two aspects are possibly of special importance to čamča

‘shirt’. One is the phonetic shape of the word: Can our understanding of one

čamča word be effectively used in order to explain the origins and evolution of

the other čamča word? The other aspect is of contactological nature: Is it possible

that these words affected each other, e.g. in phonetic terms?

All in all, I do not actually think that the Persian-Turkic word čamašy(r) is

the only source of the Mongolic word čamča. Rather, čamašy(r) was one of the

forms involved, and the whole word family of Siberian čamca ~ samsa ~ čamži

and so on, is arguably to be viewed as the result of blending of different words11

whose more detailed analysis requires further research (although it is not certain

that this etymological knot can ever be ultimately untied).12

Even if Mongolic čamča and its counterparts in other languages are no Oriental

reflexes of German Hemd and French chemise, they build a set of phonetic

and semantic variants that certainly merit our attention.

11. This concerns both the genesis of the Mongolic word and the precise establishing of a source of its

reflexes in Turkic. It is thus easily understandable that V. Drimba (1976: 426) adduces Sal. čamǯa in the

context characterized in the following way: “Il existe un assez grand nombre de mots qui ne nous permettent

pas de préciser à quelle langue mongole ils ont été empruntés […]”.

12. A good example of such a special blending (fortunately, a solved one) is the semantic history of

Siberian words with the meaning ‘1. Russian; 2. monster’ or, sometimes, ‘1. monster; 2. Russian’ (Janhunen

1997).

450 Marek Stachowski

Abbreviations

Ar. = Arabic

Brb. = Baraba

Čag. = Chagatay

Engl. = English

Fr. = French

Germ. = German

Leb. = Lebed

Kor. = Korean

Ma. = Manchurian

Mo. = Mongolic

Nan. = Nanay

Oyr. = Oyrot

Pers. = Persian

Russ. = Russian

Sal. = Salar

Tel. = Teleut

Tkc. = Turkic

Tksh. = Turkish

Uyg. = Uygur

References

KW = Ramstedt, Gustaf John 1935: Kalmückisches Wörterbuch. Helsinki.

PLOT = Stachowski, Stanisław: Osmanlı Türkçesinde Yeni Farsça Alıntılar Sözlüğü

– Wörterbuch der neupersischen Lehnwörter im Osmanisch-Türkischen

(ed. Mehmet Ölmez), İstanbul 1998 [originally published as a series

of articles with the title “Studien über die neupersischen Lehnwörter im

Osmanisch-Türkischen”. – Folia Orientalia 14 [1972–73] – 20 [1979]).

PRS = Rubinčik, Jurij Aronovič et al. 1985: Persidsko-russkij slovaŕ. Vol. 1–2.

Moskva.

SKE = Ramstedt, Gustaf John 1949: Studies in Korean etymology. Helsinki.

TMEN = Doerfer, Gerhard 1967: Türkische und mongolische Lehnwörter im

Neupersischen, vol. 3. Wiesbaden.

VEWT = Räsänen, Martti 1969: Versuch eines etymologischen Wörterbuchs der

Türksprachen. Helsinki.

Zenker = Zenker, Julius Theodor 1866: Dictionnaire turc-arabe-persan / Türkisch-

arabisch-persisches Handwörterbuch, vol. 1–2. Leipzig.

Anikin, Aleksandr Evgen’evič 11997: Ètimologičeskij slovaŕ russkich dialektov

Sibiri. Zaimstvovanija iz ural‘skich, altajskich i paleoaziatskich jazykov.

Novosibirsk.

Anikin, Aleksandr Evgen’evič 22000: Ètimologičeskij slovaŕ russkich dialektov

Sibiri. Zaimstvovanija iz ural’skich, altajskich i paleoaziatskich jazykov.

Moskva & Novosibirsk.

Doerfer, Gerhard 1968–69: Die özbekischen Lehnwörter in der Sprache der

Araber von Buchara. ‒ Central Asiatic Journal 12: 296–308.

Drimba, Vladimir 1976: Remarques sur les mots d’emprunt mongols de la langue

salare. ‒ Revue Roumaine de Linguistique 21(3): 417–427.

Written Mongolian čamča ‘shirt’ and 451

its etymological counterparts in Europe

Hahn, Reinhard F. 1991: Spoken Uyghur. Seattle & London.

Janhunen, Juha 1991: Materials on Manchurian Khamnigan Evenki. Helsinki.

— 1997: The Russian monsters. On the etymology of an ethnonymic complex.

‒ Studia Etymologica Cracoviensia 2: 159–165.

Jarring, Gunnar 1997: Central-Asian Turkic place names. Lop Nor and Tarim

area. Stockholm.

Poucha, Pavel 1956: Die Geheime Geschichte der Mongolen als Geschichtsquelle

und Literaturdenkmal. Ein Beitrag zu ihrer Erklärung. Praha.

Rachewiltz, Igor de 2004: The Secret History of the Mongols. Leiden.

Rybatzki, Volker 2006: Die Personennamen und Titel der mittelmongolischen

Dokumente. Eine lexikalische Untersuchung. Helsinki.

Tietze, Andreas 2002: Tarihi ve etimolojik Türkiye Türkçesi lugatı, vol. 1. İstanbul

& Wien.