While looking up the reconstruction of Proto-Mongolic 'six' for my last post in Starostin's Altaic database, I found that Starostin reconstructed Proto-Altaic *ɲu- from

Proto-Mongolic *jirgu-ɣa- (without the *l- which might be implied by Khitan 'six' resembling Chinese 來 *lai)

Proto-Tungusic *ɲu-ŋu-

Proto-Japonic *mu-

I am not convinced.

I know of no Tungusic-internal evidence for dividing 'six' into a root-suffix sequence. What is the function of the suffix *-ŋu-? It's not in other numerals: e.g., PT *tunga 'five' and *nadan 'seven'. (Could Jurchen and Manchu ninju 'sixty' be a contracton of an earlier *NVNV-ju with a disyllabic root?)

However, I suppose PM *-ɣa- is an allomorph of the numeral suffix *-ɣ/gVn found in other Classical Mongolian numerals:

nigen 'one'

doluɣan 'seven'

But how does one get from monosyllabic PA *ɲu- to disyllabic PM *jirgu-?

One could also object to the PM affricate *j- (not IPA [j]) coming from PA *ɲ-, but there is an areal precedent for such a change: Mandarin r- [ʐ] is from Middle Chinese *ɲ-. Sino-Japanese j- is partly from borrowings of Middle Chinese syllables with *ɲ- via *ɲj-.

Nonetheless, PA *-u- > PM *-irgu- is impossible without positing another suffix *-rgu which still leaves PM *i unaccounted for. (Starostin reconstructed PA *ɲu- > PM *ju- in this dubious etymology with loose semantics ['plant glue' > 'broth'!?], so I would expect PM 'six' to also have *ju-.)

Moreover, is there any attested case of *ɲ- > *m-? It's not entirely implausible for a palatal consonant to become labial before a labial vowel, but I've never seen this change elsewhere.

I could try to salvage this etymology by reconstructing PA *mju- with a cluster *mj- that simplified in three different ways (> PM *j-, PT *ɲ-, PJ *m-). PM *-i- and the -i- in Jurchen *niŋgu and Manchu niŋgun 'six' could be a trace of the original *-j-. However, this still fails to account for the -rgu- in PM *jirgu-.

I prefer to regard the Mongolic, Tungusic, and Japonic words for 'six' as unrelated.

There are even some Chinese words for 'six' that superficially look like Starostin's PA *ɲu-:

Mandarin dialects

Hankou nou, Chengdu nu ~ niəu

Xiang dialects

Changsha nou, Shuangfeng nəu

But these forms are actually all descendants of Middle Chinese *luk with *l- and *-k. Similarity in one time period does not guarantee similarity in an earlier time period.

Many Austronesian words for 'six' contain syllables like no. One could try to cut up, say, Maori ono into a 'prefix' o- plus a root no to derive the latter from PA *ɲu-, but arbitrary segmentation and loose phonetic similarity of single syllables are excessively powerful tools for historical linguists.

I require all segmentation to be based on language-internal variation, not a desire to link a word in one language with a shorter word in another language.

Loose and even perfect phonetic similarity can be coincidental. COMING UP WITH A BETTER SMALL PROPOSAL FOR KHITAN *L

I think Andrew West has done just that. He suggested to me that the Khitan small script character

for *l could be from Chinese 來 'come' which would have been *lai in the tenth century. That is a much simpler and therefore more likely explanation than my convoluted proposal in which was derived from the shape of the the large script character

for 'six' and its sound value was taken from the initial *l- of Chinese 六 'six'. (The Khitan word for 'six' is unknown.)

Andrew also suggested that

for 'six' could be from Chinese 來 'come'. Why would 'six' be written like 'come'? Could the Khitan word for 'six' have had initial *l- (like both Chinese 來 'come' and 六 'six'?). Could it even have sounded like Chinese *lai 'come'?

I doubt the Khitan word for 'six' was borrowed from Chinese since I'd expect other numerals to be borrowed as well. The two Khitan numerals known to me, *tau 'five' and *jau 'hundred', are clearly not from Chinese and are cognate to Proto-Mongolic *tabun 'five' and *jaɣun 'hundred'. A Khitan *l- word for 'six' would not be cognate to Proto-Mongolic *jirguɣan 'six'.

10.29.0:05: The Khitan large script character

for 'six' should not be confused with the large script character 来 whose meaning is unknown. 来 looks exactly like Chinese 来, a variant of 來 'come'. You can see Khitan 来 in the first line of the Duoluoliben langjun memorial. THE SIXTH SOURCE?: A SMALL PROPOSAL

While looking at the Duoluoliben langjun memorial at Andrew West's blog, it occurred to me that the Khitan small script character

for *l could be derived from the Khitan large script character

for 'six' by removing its horizontal stroke and bottom hook.

in turn may be derived from Chinese 六 'six' (pronounced with initial *l- in the 10th century as well as in modern Mandarin). The Khitan word for 'six' is unknown. Kane 2009 does not reconstruct it. But the creator of the small script would have known that 六 and both meant 'six' and could have chosen to represent the initial consonant of the former with a simplification of the latter.

'Six' has another large script character which also appears in the Duoluoliben langjun memorial. This second 'six' bears much less resemblance to 六 and I wonder if it was derived from rather than directly from 六.

1.28.1:28: The Khitan small script character for 'six' is which must be derived from 六. has a dotted masculine form implying that 'six' may have been an adjective. See this table for Khitan masculine color adjectives in the small script. KHITAN KHOLORS (PART 6)

All previous examples of Khitan small script might lead one to conclude that Khitan small characters sound nothing like Chinese characters. So one might be surprised by Kane's (2009: 176) transcription of


as yô.u. The first character resembles Chinese 由, pronounced you in modern Mandarin. (The tenth century pronunciation of 由 was similar.) I wonder if is a placeholder name like yỏ for

an exact lookalike of Chinese 由. yỏ has the superscript question mark that Kane uses for placeholders.

There are no other yV syllable graphs in Kane's (2009: 301-305) index of Khitan small characters, though there is a y and several characters for iV: iu, ii, ie, etc. Should iV be reinterpreted as yV? (iu could also be [iw] as well as [yu].)

The second character of 'black'


resembles Chinese 及 (pronounced something like *giʔ in the tenth century) but sounded nothing like it.

'Black' has no known gender differentiation. Is this an artifact of a small corpus? Would a masculine form have a different second character?

The Khitan large script graph for 'black'

resembles Chinese 而 'and' and 南 'south' (neither of which sound like yô.u) more than 黑 'black' at first glance, though I wonder if the Khitan graph is based on the 土 + 灬 shape at the bottom of 黑.

This table (expanded from Kane 2009: 176) sums up what is known about the large and small script forms of Khitan colors:

Large script: Does it resemble a Chinese character? Small script feminine / gender unmarked Small script masculine Gender differentiation? Transcription
green Resembles 火 'fire' (not green!) over 木 'tree' (with green leaves) Yes s.iau.qú (f.), s.iau.qu (m.)
red < Chinese 赤 'red'? (?) None known to me; might have same f. suffix as 'green' l.iau.qu (f.?); is m. l.iau.qú?
gold < Chinese 黄 'yellow'? Yes (but phonetic difference, if any, between f. and m. forms unknown) unknown
white No similar Chinese character for 'white'; doubt it's made up of the bottom plus top of the graph 金 for its associated element 'metal'; resemblance to 夹 'be on both sides' not meaningful
black < bottom of Chinese 黑 'black'?


(Originally written 1.24.21:11; not posted until now.)

To celebrate the launch of W. Robinson Mason's hadanus.com site, I've written Hadan, the name of one of his main characters, in Khitan small script using 'false friend' symbols that look like Chinese:

xa-da-án (in Kane 2009's transcription)

I wanted to include án since it looks like 'gold' from my last post.

looks like Chinese 不 'not' but is a phonetic symbol for xa (pronounced like a cross between ha and kha; Khitan had no ha)

looks like Chinese 久 'long time' but is a phonetic symbol for da.

不久 is 'not long' or 'soon' in Chinese but is just the two syllables xa-da in Khitan.

Hadanus, the name of Hadan's planet could be


with us added. Unlike the other symbols, it doesn't resemble any Chinese character.

I don't know if a Khitan scribe would choose these four symbols. Although xa has only one symbol , there are three different da-symbols:

The first and second are obviously variations of each other and the third may also be related to the first two.

I could have written the -n of Hadan as

n resembling the Chinese character 公 'duke' (also a Khitan large script character)

but I chose to write it with án since some Khitan closed syllables are written with vowel-consonant syllable symbols. Perhaps da-án would be pronounced with a long vowel as [daan]. The phonetic difference, if any, between án and ~ an is unknown. Kane's accents indicate the potential presence of unknown phonetic distinctions.

No Khitan small script character for nu is identified in Kane 2009. If such a character existed, -nus might be written as

nu + us

rather than as


While some closed syllables are written with what may be redundant vowels: e.g., Chinese 蘭 *lan 'orchid' was borrowed into Khitan as

l-a-an (for *[lan] or *[laan]?)

others were written with no apparent vowels at all (cf. Arabic and Hebrew): e.g.,

s-l-b (for *sVlVb?)

Was this option chosen when there were no preexisting syllabic symbols?

Hadanus could be written in this vowelless style as


There are no true vowelless symbols for x and d since x (the resemblance to the letter x is coincidental) can also be xe and d can also be de (but which de were written with d(e) rather than de?). KHITAN KHOLORS (PART 5)

Today I realized that looking at the Khitan scripts is like looking at Cyrillic (from the point of view of someone accustomed to Chinese characters and the Roman alphabet). Some characters have familiar shapes and sounds and even meanings, others are false friends, and still others are totally alien.

The Khitan small script character for 'yellow, gold' looks like Chinese 山 'mountain' but has an entirely different meaning. Its reading is unknown. Could the Khitan word for 'yellow' or 'gold' have sounded like Chinese 山 'mountain' (which would have been something like shan in the 10th century) rather than Proto-Mongolian *sira 'yellow'?

Like 'white' in part 4, has a masculine form with a dot: . Dots do not mark gender in Chinese writing.

The genders of 'green' and presumably 'red' were indicated by different symbols for the final syllable:

'green' (masc.) <> 'green' (fem.)

'red' (masc.?) <> ? 'red' (fem.; hypothetical)

It's not clear how masculine and feminine forms of adjectives differed.

I wonder if

'white' (fem.) <> 'white' (masc.)

'gold' (fem.) <> 'gold' (masc.)

had no gender distinction in speech, or if they were monosyllables with different vowels depending on gender.

could also be a noun referring to the 金 Jin 'gold' dynasty of the Jurchen.

The Khitan large script equivalent of was which might be a simplification of Chinese 黄 'yellow'.

Tangut fonts by Mojikyo.org
Tangut radical font by Andrew West
All other content copyright © 2002-2010 Amritavision