More Chinese milk, lick, tongue

And I ask permission for a short(?) quote:

from Sagart
ANCESTRY OF CHINESE
SOME REMARKS ON THE ANCESTRY OF CHINESE

(.. stands for a Chinese sign)
"
1.1.3. Correspondences of the borrowed layer
Some phonological criteria may already be used with confidence to
distinguish Chinese loanwords in TB. There are three main
correspondences for OC *l- among TB words reconstructed by Benedict
in the STC:
1. OC (h)l- = TB l-
.. hlot 'peel off, take off, take away' =
PTB (Benedict) *g-lwat 'free, release'
.. hljur? or t-ljur? 'river, water' =
PTB (Benedict) *lwi(y) 'flow; stream', *twiy (< t-lwiy) 'flow;
water'
.. Lje? 'to lick' = PTB (Benedict) *(m-)lyak, (s-)lyak 'tongue'
.. ljap 'leaf = PTB (Benedict) *lap 'leaf'
.. ljan, 'hawk' = PTB (Benedict) *lan, 'falcon, vulture, eagle,
kite, hawk'
2. OC l- = TB d-~t-
.. luk/duok (pronunciation see infra) 'poison' =
PTB (Benedict) *duk~tuk (or *dok~tok, cf. STC note 232) 'poison'
.. lïj?/dieiB 'younger sibling' =
PTB (Benedict) *toy~doy 'younger sibling'
.. lon,-s/dun,C 'cave' =
PTB (Benedict) *dwa:n, 'cave'11
.. lïm/dâ,mA 'deep, deep place, pond' =
PTB *tum~dum 'pond'12
3. OC lj- = TB y-
.. ljïn/j&n,A 'fly (n.)' =
PTB *yan, 'fly' (n.).
Loaned to Vietnamese as l&n,2 'large green fly'.
.. s-ljor/zwjeA 'follow' = Kuki-Naga *ywi (Benedict);
PTB *s-yuy (Matisoff) 'follow'. Note Newari li (< lwi) 'follow'13,
representing an earlier loan or a true cognate.
.. ljak-s/jaC 'night' = PTB (Benedict) *ya 'night'
.. ljak/jak 'armpit' =
PTB (Benedict) *yak 'armpit'. Compare .. k-lak/ kâk 'armpit', with
k- prefix.
.. lju?/juB 'wine' =
PTB (Benedict) *yu 'wine'. Loaned to Buyi as lau wine'.

The TB reconstructions under correspondences 2 and 3 match the MC
pronunciations better than they do OC. OC *l- regularly changed to d-
, and OC *lj- to j- (.. ..) in Han times. No such change occurred in
TB. Doublets with l- initial at times occur in TB languages (Newari
li 'follow'), representing an earlier layer (whether genetic or not
is still uncertain); or corresponding l-initialled forms occur in
other languages (Vietnamese, Buyi, etc.), usually to be regarded as
early Chinese loanwords. This is a clear indication that the
comparisons under correspondences 2 and 3 belong to a layer of
relatively recent Chinese loanwords.

The fact can be confirmed semantically: three items under
correspondence No 2: 'poison', 'younger sibling' and 'pond', have
semantics which are secondary in Chinese.

Poison
PTB *duk~tuk (but *dok~tok in STC, note 232) - .. MC duok. This word
is not attested epigraphically but occurs 11 times in received texts
of Western Zhou and early Chun-Qiu period. Although standard
translations (Karlgren, Legge, Couvreur, etc.) usually render the
word as 'poison', none of these early occurrences refers to poison
as a material substance causing harm or death: instead, most refer
to anger, harm, suffering or worry. In the meaning 'suffering',
there is an interesting collocation with words meaning 'bitter' (..
or ..) in several of the earliest occurrences of .. (Shi Jing: Daya,
ode 257, 11; Xiaoya, Ode 207,1; Shu Jing, .. .. 12,3). Furthermore,
in one instance (Yi Jing, hexagram #21, ca. 900 B.C.E.), .. refers
to a material substance one risks to find in dried meat. That
substance appears to be unpleasant (bitter ?), but harmless, and a
translation as 'poison' is particularly awkward:

'Eating dried meat, find [duok] Small inconvenience. No
catastrophy.'

Both the the Yi Jing line and the collocation with 'bitter',
common in early Zhou occurrences of the word, but absent later, find
an explanation if we assume that the original meaning of [duok]
was 'gall'. Finding bits of dried meat made bitter by gall may have
been be a common experience in early China, especially if the meat
was that of small animals (such as birds, poultry, etc.), in which
the gall bladder is small and special care is required to avoid
breaking it. Through a very common semantic shift 'gall'
gave 'anger', 'harm', 'suffering', 'worry', etc. More recent
meanings, attested in the Zhan Guo period, revolve around the notion
of 'care': 'to care for, govern' (Yi Jing, hexagram 7); 'nourish'
(Lao Zi); 'give medical care' (Zhuang Zi); 'medicine, drug, cure'
(Zhou Li, in the binom .. ..). These late meanings (which have been
an endless source of perplexity to Han scholars) are derivable
from 'worry', as NHG Sorge.14
The meaning 'poison' (as a material substance causing harm or
death...) represents a distinct line of development, being directly
from 'gall', most likely through 'animal venom', a shift having
parallels in Greek, Germanic and Latin15. [duok] in the
meaning 'poison' has its first two occurrences in the Zuo Zhuan: it
is noteworthy that in one of these (Zuo Zhuan, 22nd year of Duke
Xi.3) [duok] means specifically 'animal venom':
.. .. .. [duok] 'bees and scorpions have venom'
The graph, incorrectly analyzed as a xing-sheng in Shuo Wen16,
consists of .. 'give birth' above .. 'mother', as can be seen from
the earliest exemplars in the paleographical record17. This
combination fits neither 'gall' nor 'poison', and a Jia Jie (loan
character) origin must be envisioned. A character of very similar
composition is .. [1021a] ljuk/jwok 'to produce': it consists of
a .. mother on the left, with a head-down child on the right: 'an
illustration of child-birth' (Karlgren; see also Boltz 1994). This
character, which writes the same word as .. [1020a] id. 'to rear,
breed; to nourish' and .. [1024b] id. 'to nourish; young (n.)',
involves a word-stem OC *luk, with infixed -j-. Probably cognate are
the uninfixed .. [1024i] luk/duok 'abortion' and .. [10241]
id. 'calf'. I would propose that the character .. originally wrote
an uninfixed member of that word family, and was applied through Jia-
jie to the homophonous word for 'gall' (which must have been OC *luk
rather than *duk), later 'poison' etc.
Support for OC *l- in [duok] comes from the dialects, modern and
ancient. From *luk/duok 'poison' was obtained by tone derivation in
the OC period a verb *luk-s/dâuC 'to poison', the latter form being
reflected only in modern Min and south Hunan dialects. The shift of
OC *l- to MC d- occurred in the standard Han language (Yakhontov
1976) and most modern reflexes are based on a prototype with d-
initial (so the list in Norman 1988: 213). However the *l- initial
was preserved dialectally in post-Han times in Hunan around
Changsha, and in the northeast, two peripheral areas settled by
Chinese speakers before the shift of *l- to d- took place18. For the
northeast, cf. Fangyan
3: .. .. .. .. .. .. .. , .. .. .. .. .. .. .. .. and Shuo
Wen: .. .. .. .. .. .. The Guang Yun reads .. as MC lâu (Qu tone)
with reference to the Shuo Wen passage. The same form .. lau21 (Yang
Qu) is the colloquial word for 'to poison' in modern Changsha and
other Xiang dialects. From interior south Chinese dialects the word
was loaned to Miao-Yao (Chuanqiandian Miao has Iao8 'to poison
fish') and to some southern TB languages as
well (cf. the set assembled by Shafer 1966-74: 195)
Irregularity in the sound correspondences of the word 'poison' in TB
languages is another clear indication of borrowing: note in
particular the pronunciation tuk in Jingpo, with final -k instead of
expected -? (Jingpo reflects PTB -k as -?). Workers in ST have been
all too willing to overlook such irregularities when 'excellent
roots' are at stake.


14In my previous work (including the first version of this paper), I
had assumed, wrongly it now appears, that the meaning 'poison' had
been derived from 'cure, medicine', as Fr. poison, NG pharmaki,
etc., the meaning 'cure' itself being from the notion of 'care for',
which has an occurrence in Yi Jing (hexagram 7). However, the
passage in question comes from the Tuan Zhuan .. .., a late addition
to the Yi Jing text. The meaning 'care for' thus cannot be shown to
be ancient. Moreover the meaning 'cure' is slightly later than the
meaning 'poison', and under my previous explanation the collocation
with 'bitter' in the early examples is unexplained.
"


Semantically, the best fit (for *luk/duok) would be "bodily fluid",
specifically one of the four cardinal bodily fluids (blood, phlegm,
black bile, yellow bile). Then in 'eating dried meat, find [duok]
Small inconvenience. No catastrophy.' it means "fluid, non-dried
matter" (as that theory predicts there would be). I think Sagart's
previous assumption, in his note 14, is the right one. This would
mean that milk was used at first for medicinal purposes, possibly to
redress a 'lack of phlegm'?


Note the fact that Old Chinese *l- can appear in loans in other
languages as *l-, *d- and *j-. Cf. Latin lingua, Old Latin dingua,
Russian jazykU, an old problem that can only solved ad hoc within IE.


Torsten