*m-ts(y)(u/i)p

Many forms descending from this etymon have voiced initials, a fact which we might capture by setting up a separate allofam *m-dz(y)up ⪤ *m-dz(y)ip. Some of these cases can be ascribed to the influence of the nasal prefix. However, since there is so much variation in voicing even within a single subgroup, we have chosen to group all the reflexes under the prototype with voiceless root initial.

This widespread etymon is set up as *dzoːp in STC #69, on the basis of forms from four Barish and Kuki-Chin languages: (Barish) Dimasa dźop and (Kuki-Chin) Lushai (Mizo) foːp, Thado tsop, and Siyin tuop ‘suck; kiss’. (Cf. also Kom Rem məǰop, Tiddim toːp, and Lai Chin doop.) However, the rhymes *-op and *-oːp are extremely rare in TB, with virtually no other examples (see HPTB, pp. 381-2), and it seems preferable to consider these forms with -o(ː)- vocalism to reflect a localized secondary development. (On the other hand, there is a parallel form in NW rGyalrong kantsʰrop, alongside Ma’erkang rGyalrong kɑ mə scçup.) The vast majority of the reflexes point to a high proto-vowel, either *-u- or *-i-. The alternation between these two vowels is especially frequent in the environment of labial consonants, and indeed this is one of the best attested patterns of variation in TB, as well as in Chinese. (See VSTB pp. 41-2, HPTB pp. 493-505. See also #3450 PTB *m-pup ⪤ *pip KISS / SUCK / SUCKLE and #1613 PTB *(t)s(i/u)p NEST / WOMB / SCROTUM, below.) Several languages show internal variation between these vowels (e.g. Dimasa dźop ‘suck’ ⪤ sep ‘milk’; Tiddim teːp¹ ‘suck’ ⪤ toːp¹ ‘kiss’).

The nasal prefix *m- has been added to the reconstruction since it occurs in many languages of the NE Indian Areal Group (e.g. Sema mtsü, Ao Mongsen mechep, Kom Rem məǰop, Moyon njúp, etc.), as well as in Qiangic (e.g. Ergong ndʑip⁵³, Namuyi ntshu⁵³), and WT ḥjib(s)-pa, where a-chung “ḥ” is interpreted as a nasal prefix (see HPTB pp. 115-6).

Reflexes of the allofam with *-u- vocalism are presented in #5541 PTB *m-ts(y)(u/i)p SUCK / KISS / BREAST / MILK, and those from the *-i- variant in #5541 PTB *m-ts(y)(u/i)p SUCK / KISS / BREAST / MILK.

I also recognize an opened-syllable allofam *dz(y)əw MILK/BREAST, treated separately below #5539 PTB *dzy(ə/o)w SUCK / KISS / BREAST / MILK.

A few languages show secondary variants with final stops other than -p. Jingpho has a doublet, one with the expected -p (tšùp ‘suck, kiss, absorb’), and one that reflects final *-k (tšúʔ ‘breast’; -ʔ is the regular Jg. reflex of *-k). Similar forms appear in several other languages: Sherpa (Helambu) chuk pekin ‘kiss’, Tamang (Sahu) cyok-pa ‘kiss’, Lashi tʃu̱:k⁵⁵ ‘nurse; suckle’, Bisu tʃhɔ̀k ‘suck’. All these languages seem to have doublets with both -p and -k. WB has an aberrant form with a final dental (cut ‘suck’), paralleled by Dulong tsut⁵⁵ ‘suckle’ and Tamang (Sahu) nyeh cʰuT Tai'◦ti-pa.¹

For other etyma with similar variation in final position, cf. #3450 PTB *m-pup ⪤ *pip KISS / SUCK / SUCKLE and #1613 PTB *(t)s(i/u)p NEST / WOMB / SCROTUM.

The semantic connection between SUCK and KISS is paralleled by the vulgar English expression suck face for ‘kiss’.

There is a possible Chinese comparison, 噆 (Mand. cǎn, zǎn); this character has two OC readings, *ts'əm ‘have in the mouth’ and *tsəp ‘bite; sting and suck (as a mosquito)’ [GSR #660f, #660o]; see also Coblin 1986:144. TB also shows some variation between final stop and nasal in this etymon: cf. Sunwar cim-cā ‘milk a cow’ ⪤ yup- ‘suck’; Tsangla (Motuo) jum ‘suck (milk)’ ⪤ tɕup⁵⁵ a⁵⁵ ‘kiss’. The final nasal in the Milang form ɟim-ma seems clearly to be due to assimilation of the initial to the suffix. See the discussion by ZJH, below.

An undoubtedly related root (not in STC or HPTB) is #86 PTB *(t)sy(i/u)p ⪤ *(t)swap SQUEEZE / WRING / CRUMPLE, the semantic connection being the squeezing action required to express milk from a breast or udder. See DL:557-8.

噆 OC *tsʼəm/tsəp, GSR #660f, 660o ‘have in mouth; bite; sting’; Li 1971: (*tshəm/p); Baxter 1992: (*tshɨm/p); B & S 2011: *tsˁəp {tsˁ[ə]p}; Mand. cǎn, zǎn.

Neither Li nor Baxter (1992) reconstructs this word. As Baxter notes (1992:555), this phonetic series presents unusual difficulties for OC reconstruction. Baxter reconstructs some words with *o and others with *i/ɨ.

It is not clear why Karlgren lists the character twice in GSR #660. At GSR #660f he reconstructs *tsʼəm for the meaning ‘have in mouth’ and *tsəp for the meaning ‘bite’, while at GSR #660o he reconstructs *tsəp meaning ‘sting and suck’. Karlgren’s usual practice in GSR is to group multiple readings of a single character together, listing that character only once.

On the apparent mismatch between the voiceless PTB initial and the voiced OC initial, see the discussion under #1654 PTB *pu EGG.

Vowel correspondences between OC and PTB before labial codas are difficult to pin down. In Li’s system, OC *-ə- regularly corresponds to TB *-u- and/or *-i-. Examples include 三 OC *səm ‘three’, PTB *gsum, Mand. sān; 入 OC *njəp ‘enter/sink’, PTB *nup ⪤ *nip, Mand. rù; and 寢 OC *tshjəm ‘sleep’, WT gzim, Mand. qǐn (Coblin 1986:134). Using Baxter’s system, however, where six vowels (as opposed to just *a, *ə in Li’s system) occur before bilabial consonants, correspondences are less regular. This is in part because in many cases there is ambiguity in the reconstruction of vowels before bilabial codas in his system. At the present state of our knowledge, the proposed cognate set is viable.
[ZJH]