On Some Manchu Etymologies

The following HTML text was created from the original print version. Although care has been taken to transcribe the original correctly, errors may remain. Please refer to the PDF as the text of record for citation. If you encounter a mistranscription, please report it to the editor ([email protected].)

With the publication of the monumental Sravnitel'nyj slovar' tunguso-man'čžurskix jazykov by the late V. I. Cincius and her colleagues in 1975 (volume two in 1977), a new era in the comparative study of the Tungusic languages began. Not only can the etymological roots of many hundreds of forms be traced within the Tungusic family itself, the Altaic connections of more and more Tungusic forms are becoming clear. The present paper, which attempts to delve into the origins and connections of several Manchu words, is heavily indebted to the TMS (the editors' suggested abbreviation of the Sravnitel'nyj slovar') and to the other writings of the late Professor Cincius. Manchu forms and their meanings are based on Norman (1978).

1. jancuhūn 'sweet, pleasant, agreeable'; jancuhūkan 'rather sweet'; jancuhūnje 'sugar cane'; jancuhūri orho 'licorice'.

Jancuhūn is the primary form of this set of forms; jancuhūkan is a regular diminutive formation. The forms jancuhūnje and jancuhūri orho represent neologisms coined to translate Chinese words referring to things originally unknown to the Manchus. These formations are all too transparent to reveal anything further about the origin of the word jancuhūn 'sweet'.

The word itself clearly consists of a root jancu- plus a suffix -hūn. This suffix, which has the harmonic variants -hun and -hon, is perhaps the single most common adjectival suffix in Manchu. The following words are all formed by means of this suffix: banuhūn 'lazy' (cf. ban- 'to be lazy'), fulahūn 'pink, reddish' (cf. fulara- 'to turn red'), getuhun 'awake' (cf. gete- 'to awaken'), milahūn 'wide open' (cf. mila 'open, wide open'), sohon 'deep yellow' (cf. soro- 'to turn yellow').

Roots ending in voiced consonants often devoice when they occur before a suffix beginning with the consonant h as the following examples illustrate: waci-hiya- 'to complete' (cf. waji- 'to finish'), fonto-ho 'small hole' (cf. fondo-lo- 'to pierce, make a hole'), yafa-han 'on foot' (cf. yabu- 'to walk'), tofo-hon 'fifteen' (cf. Mongolian tabun 'five'), ufu-hi 'part, share' (cf. ubu 'portion, share'), afa-ha 'sheet (of paper)' (cf. abdaha 'leaf')[1],fecu-hun 'low, base' (cf. fejergi 'underneath').

Another fact that is important in the present case is that the second vowel in a Manchu trisyllabic word is inherently weak and is easily altered under the influence of the vowel of the following syllable beginning with a velar or uvular: gete- 'to wake up', getuhun 'awake (adj.)'; ete- 'to overcome', etuhun 'strong'; hala- 'to be hot', halukan 'warm'; hatan 'pungent, strong tasting', hatuhan 'salty'; seri 'sparse', seruken 'rather sparse'; tafa- 'to ascend', tafukū ≈ tafakū 'steps'; juše- 'to be sour', jušuhun 'sour (adj.)'. Note that in all the cases cited above the vowel preceding the suffix in the derived form is a u. In view of the two tendencies described above, it is clearly justified to think that the original root of the word jancuhūn was *janji- and not jancu-, since both the change of j > c and the change of i > u are well attested in many other forms in Manchu.

From the comparative work of Cincius (1949) and Benzing (1955) it is well known that a proto-Tungusic *l becomes n in Manchu before a dental or alveopalatal consonant Evk. baldī- 'to be born', Ma. banji- 'id.'; this fact can also be seen internally in Manchu itself: ulu 'empty, unfertilized (of an egg)', untuhun (< ul-tu-hun) 'empty'. Moreover, Manchu ji and ci in many cases regularly derive from earlier *di and *ti respectively (Cincius 1949: 184, 188; Benzing 1955: 979, 980). Hence, it is possible to see our hypothetical *janji coming from an earlier *jaldi. We are now in a position to see the origin of jancuhūn. The following two Evenki forms, attested in different dialects, are glossed as 'tasty, sweet': daldi (Sakhalin and Kachug dialects) and dalli (Sakhalin and Ayan dialects). The standard Even form is dalra with dialectal variants dalda and dalci with the same meaning; Negidal has a comparable form daligdi (TMS 1.13a). The Evenki and Even forms can be derived from a proto-Tungusic form *daldi. The evolution of the Manchu form can be plotted hypothetically as follows: *daldi> *dalji > *danji > *janji. The shift of initial *d > *j is an assimilatory change, apparently sporadic, seen also in the word jabjan 'python, large snake', cf. Evenki. jabdar, Even jabda 'snake'.[2]

Does our proto-Tungusic *daldi have any further Altaic connections? It is tempting to see a link between proto-Tungusic *daldi and Early Written Turkic (EWT hereafter) tatlïg, Turkish tatlï, Modern Uyghur tatliq 'sweet, tasty'. These forms derive from a root tat '(good) flavor' plus the common adjectival suffix -lïg (or its dialectal counterpart).[3] But what about tat itself? Might it not somehow be related to *daldi? The loss of the final vowel presents no problem since this is a regular feature of Turkic development (Poppe 1960: 117). Altaic *d ordinarily becomes Turkic y (Ramstedt 1957: 38, Poppe 1960: 22) but there are also a significant number of cases where Mongolian and Tungusic d corresponds to Turkic t: Mongolian dörben 'four', Evenki digin 'id.', EWT tört 'id.'; Evenki doŋoto- 'to freeze over", EWT toŋ- 'to freeze', Nanai dalan- 'to overflow', EWT taš- 'id.'; Evenki dulin 'middle', Mongolian düli 'half, middle, noon', EWT tüš 'noon'. This shows that Altaic *d may actually represent two different proto-correspondences, one yielding Turkic y, the other t. Since Turkic languages do not allow two consonants at the end of a root, the original *l would have regularly dropped out, giving the following hypothetical development: *daldï> *tald > *tād > *tāt. The long vowel in Common Turkic is indicated by Turkmen dāt (Sevortjan 1980: 162); is this long vowel not perhaps due to a compensatory lengthening which occurred with the loss of the *l? Compare the following forms: Evenki alga 'net', Turkmen (dialectal) āq 'id.', Mongolian aldar 'name (honorific), fame', Turkmen āt 'name'; Mongolian elde- 'to dress a hide, curry, soften, knead', Modern Uyghur ät- 'to make, do, knead (dough), EWT etik 'a soft shoe made of leather', Kazakh etik 'boot', Turkmen ētik 'id.' (Räsänen 1969: 52).[4]

To sum up, I propose that Manchu jancuhūn 'sweet' derives from proto-Tungusic *daldï meaning 'tasty, sweet' and this form is in turn comparable to Common Turkic *tāt.

2. farhūn 'dark, obscure, confused, dim, muddled'. farhūda- 'to act in a muddled way'.

The verb farhūda- is a derivative of farhūn formed with the common denominal verb suffix -da which generally has the sense of 'to act in such and such a way'; farhūn itself is composed of a root far- and the common suffix -hūn (described in the previous section). The root far- appears in a number of other words which together form an extensive word family in Manchu: 1)farfa- 'to be confused', probably best viewed as a semireduplicate 2)farila- 'become dark'; a derivative of fari (see below) and the common denominal verb suffix -la (≈ le ≈ lo) 3) geri fari 'indistinct, dim'; a binome formed from the roots seen in gere- 'to become bright, to dawn' and the far- seen in farhūn. The above three forms clearly reflect a meaning 'dim, dark, indistinct, muddled, confused'. It is likely that fara- 'to faint, lose consciousness' also belongs to this word family.

The Manchu root far- is comparable to a number of Tungusic forms: Evenki harū- 'to grow dim (of the eyes), feel dizzy, feel drunk', harūn- 'to grow dark, to become intoxicated'; Even hāru- 'to be dizzy, lose consciousness'; Nanai pargan 'muddled, stupid', pargāci- 'to act in a muddled or stupid manner'. Here once more we encounter the notions of 'dim, dark, confused, muddled, and faint'. There is an interesting semantic parallel with the Chinese word 昏 hūn which means 'dark, dim, confused, muddled'. On the basis of the forms cited above, we can reconstruct a proto-Tungusic *pār-.

I propose that the following Turkic forms are related: EWT az- 'to get lost, lose one's way, go astray', Turkish. az- 'to go astray', Turkmen āz- 'id.', Khalaj hāz- 'id.', Modern Uyghur az- 'id.'. From these forms one can infer a Common Turkic *hāz- which derives from an earlier proto-Altaic *pāŕ-.

Clearly the Turkic sense of 'go astray' comes from the notion of 'confused' which is well attested in the Tungusic forms. This connection is further strengthened by several other Turkic forms derived from this root: Tksh. azgïn 'astray', Uyg. azɣaq 'easily confused' and Uyg. azγin 'confused, perplexed, at a loss'.

Among the Tungusic languages Manchu seems to be alone in using the root *pāŕ- in the meaning 'dark'. The more usual Tungusic root is illustrated by the following forms: Evenki haktirā- 'dark', Even hātar- 'id.', Nanai pakcï 'id.'. These forms are from a root *pak- also seen in Nanai pakala 'dark'. Manchu has a perfect cognate in fahala 'dark purple, thick, viscous, opaque (of liquids, specifically wine or swill)'.

Finally, I would like to point out that an alloharmonic variant exists for Manchu fara- 'to lose consciousness', namely, fere- 'to become giddy, become dizzy'. Other cases of such alloharmonic pairs can be found in Manchu and other Altaic languages: Manchu: ulan 'ditch', ulen 'irrigation ditch'; fiyele- 'dry oneself by a fire', fiyala- 'dry by a fire'; makara- 'become decrepit', mekere- 'id.'; saci- 'chop off, seci- 'cut away' or 'chop off'; baša- 'chase away', bošo- 'id.'. Mongolian: dorgi-/dörgi- 'tremble, shake'; ɣatul-/getül- 'cross a river'; dabsi-/debsi- 'advance'; qabta/kebte 'lying flat'; ulbayi-/ülbeyi- 'become flabby'.

Now we are in a position to examine the Mongolian reflexes of the Tungusic and Turkic forms cited above. For the back vocalic variant *pāŕ we have Mo. argi- 'speak in a muddled way, talk nonsense (of old people)'. For the alloharmonic variant *pēŕ- we can cite Mo. ergigüü 'foolish, muddled, stupid'. Finally we should cite the following relevant form from the Secret history of the Mongols: herü baru 'darkness, twilight' (Haenisch 1962: 75).

3. gefehe 'butterfly'. No derivatives.

This word consists of a root gefe- plus the common noun suffix -he (≈ -ha ≈ -ho) which occurs in scores of nouns: fodoho 'willow', fulha 'poplar', saksaha 'magpie', niyehe 'duck', usiha 'star', etc. Manchu f in almost all cases derives from an earlier pTg. *p, so gefe- must go back to an earlier *gepe. The only Tungusic cognate recorded for this Manchu word is Udehe gepte 'a kind of butterfly' (TMS I.180b). I suspect that gepte is a contracted from *gepe-kte where -kte is a collective suffix referring to things occurring in large numbers (Benzing 1955: 72, 80).

The following Turkic forms can be compared to proto-Tungusic *gepe-: EWT kepäli 'moth' , Modern Uyghur kepinäk 'butterfly', Khirgiz köpölök 'id.', Kazakh köbelek 'id.', Salar kegelex 'id.', Turkish kepenek 'moth', Chuvash kewe 'id.'.[5]

4. fer seme 'fluttering, floating', fer far seme 'fluttering like a butterfly in flight', ferehe singgeri 'bat'.

The first and second forms are examples of a very large number of descriptive adjectives and adverbs in Manchu which are formed with a root plus a form of the verb se- 'to say (usually the converb seme or the participle sere). The second form, fer far seme, can be considered a semireduplicate, for which numerous similar examples can be cited. The term for 'bat', ferehe singgeri, contains the word singgeri 'mouse' (cf. German Fledermaus) plus ferehe which must be the specific form referring to 'bat'. This latter form consists of the same suffix -he discussed above and a root fere- which I believe is related to the notion of 'fluttering' found in the two forms fer seme and fer far seme.

Related to the Manchu root fer(e) are Evenki perē 'to flit (of birds)' and possibly Nanai pēr 'lightly (leaping)'. For proto-Tungusic we can reconstruct a form *pēr(e) 'fluttering, flitting about' (TMS II.48a, Onenko 1980: 347).

Related to this proto-Tungusic root are a number of Monglian forms: erbelje- 'to flit, flutter', erbegen 'flitting back and forth, floating', erbegci 'small butterfly', erbekei 'butterfly'.

That these forms derive from forms with an earlier initial *p can be seen from the Middle Mongolian form for 'butterfly' preserved in the dictionary of Makaddimat Al-Adab, herbegei (Poppe 1938: 184).

The root of the Mongolic forms goes back to a proto-Altaic root *pēr 'flit, flutter'. That the following syllables of the cited forms are all separate morphemes can be seen from the following comparisons:

With this information we are in a position to analyze the Mongolian word for 'butterfly', erbegekei. The common noun suffix -kei (≈ -qai) generally has a diminutive meaning (Ramstedt 1952: 207). In Ramstedt (1912: 68 ff.) a suffix -ba (≈ -ma) is identified; although it is of uncertain meaning, Ramstedt notes that it generally has an intransitive or middle function. The element -ge(n) (≈ -ga(n)) is in origin a deverbal noun or adjective suffix (Poppe 1927: 97, 98).

References

• Alderson, A. D. and Fahir Iz. 1959. The concise Oxford Turkish dictionary. Oxford: Clarendon Press.
• Benzing, Johannes. 1955. Die tungusischen Sprachen: Versuch einer vergleichenden Grammatik. Akademie der Wissenschaften und der Literatur, Abhandlungen der Geistes- und Sozialwissenschaftlichen Klasse, Jahrgang 1955, nr. II. Wiesbaden: Franz Steiner Verlag.
• Cincius, V. I. 1949. Sravnitel'naja fonetika tunguso-man'čžurskix jazykov. Leningrad: Nauka.
• Cincius, V. I. et al. 1975-77. Sravnitel'nyj slovar' tunguso-man'čžurskix jazykov. 2 vols. Leningrad: Nauka.
• Haenisch, Erich. 1962. Manghol un niuca tobca'an. Teil II, Wörterbuch. Wiesbaden: Franz Steiner Verlag.
• Hu Zhenhui. 1986. Kē'ěrkèzǐyǔ jiǎnzhì. Beijing: Minzu Chubanshe.
• Jīn Bǐngzhé et al. 1979. Hǎ-Hàn cidiǎn. Urumqi: Xinjiang Renmin Chubanshe.
• Kane, Daniel. 1989. The Sino-Jurchen vocabulary of the bureau of interpreters. Uralic and Altaic series, vol. 153. Bloomington: Research Institute for Inner Asian Studies.
• Lessing, Ferdinand D., et al. 1960. Monglian-English dictionary. Berkeley: University of California Press.
• Lín Liányún. 1985. Sālāyǔ jiǎnzhì. Beijing: Minzu Chubanshe.
• Nadeljaev, V. M., et al. 1969. Drevnetjurkskij slovar'. Leningrad: Nauka.
• Norman, Jerry. 1972. The evolution of Proto-Tungusic *t to Manchu s. Central Asiatic Journal 21: 229-233.
• Norman, Jerry. 1978. A concise Manchu-English lexicon. Seattle: University of Washington Press.
• Onenko, S. N. 1980. Nanajsko-Russkij slovar'. Moscow: Izdatel'stvo "Russkij Jazyk".
• Partridge, Eric. 1958. Origins, a short etymological dictionary of modern English. New York: Macmillan Publishing Company.
• Poppe, N. 1927. Die Nominalstammbildungssuffixe im Mongolischen. Keleti Szemle 20: 89-126.
• Poppe, Nicholas. 1938. Mongol'skij slovar' Mukaddimat Al-Adab. čast' I-II. Trudy Instituta Vostokovedenija. Moscow: Izdatel'stvo Akademii Nauk SSSR.
• Poppe, Nikolaus. 1960. Vergleichende Grammatik der altaischen Sprachen, Teil I, Vergleichende Lautlehre. Porta Linguarum Orientalium, neue Serie IV. Wiesbaden: Harrassowitz.
• Ramstedt, G. J. 1912. Zur Verbalstammbildungslehre der Mongolischen-Türkischen sprachen. Journal de la Societe Finno-ougrienne 28: 1-86.
• Ramstedt, G. J. 1952. Einführung in die altaischen Sprachwissenschaft. Formenlehre. Helsinki: Suomalais-Ugrilainen Seura.
• Ramstadt, G. J. 1957. Einführung in die altaischen Sprachwissenschaft. Lautlehre. Helsinki: Suomalais-Ugrilainen Seura.
• Räsänen, Martti. 1969. Versuch eines etymologischen Wörterbuchs der Turksprachen. Helsinki: Suomalais-Ugrilainen Seura.
• Sevortjan, E. V. 1980. Ètimolgičeskij slovar' tjurkskix jazykov: obščetjurkskie i mežtjurskie osnovy na bukvy "B" i "D". Moscow: Nauka.
• Skvorcov, M. I. 1985. Cuvaško-Russkij slovar'. Moscow: Russkij Jazyk.
• Xīnjiāng Dàxué. 1982. Wéi-Hàn cídiǎn. Urumqi: Xinjiang Renmin Chubanshe.

1. The fact that b alternates with f in these forms suggests that intervocalic b had a fricative pronunciationin Manchu as it does in Sibe: Manchu yabu-, Sibe yav-.

2. These cases of assimilation at a distance brought about by the presence of alveopalatals in the same stem may be related to the change of proto-Tungusic *t to Manchu s in stems containing a j in the second syllable; see Norman (1972). In the Sino-Jurchen vocabulary of the bureau of interpreters studied by Daniel Kane (1989), the word for 'sweet' in Jurchen is given as *dancu.

3. In various Turkic languages tat- is also a verbal root meaning 'to taste": EWT tat-, Chuvash tutan- (Skvorcov 1985: 499).

4. It would appear that EWT etük and related froms in other Turkic languages are deverbal nouns from et-;this shows that et- at an earlier time must have had a meaning similar to Mongolian elde-. The now widely attested meaning of 'to make, to do' in all likelihood derives from this earlier and more concrete sense. It is interesting to note that English make and German machen are thought to be related to Greek magis 'dough' and massein (for an earlier *magsein) 'to knead (dough)' (Partridge 1958: 373).

5. Sources for the Turkic forms are as follows: Early Written Turkish — Nadeljaev et al. (1969), Modern Uyghur — Xīnjiāng Dàxué (1982), Khirghiz – Hú (1968), Kazakh — Jīn (1979), Salar — Lin (1985), Turkish — Alderson and Iz (1959), Chuvash — Skvorcov (1985).