What looks 'one' + 'water' to Chinese eyes equals 'white' in the Khitan small script:

一 + 水 =

Adding a dot makes it masculine:

These small script graphs look nothing like the large script graph

I presume that the large script graph did double duty for both the feminine and masculine forms of 'white'.

None of the above Khitan graphs look like Chinese 白 'white', though a wild imagination might see an extremely vague resemblance to 金 'metal' (the element that 'white' corresponds to), 庚, or 辛 (the two Heavenly Stems corresponding to and ).

According to Kane (2009: 36), readings based on Classical Mongolian čaɣan 'white' have been proposed for . I presume that would also share this reading. It is impossible to determine the reading of this word without any other Khitan spellings or foreign transcriptions. I suspect that its reading was monosyllabic since the vast majority of known small script graph readings are monosyllabic or subsyllabic. Given that the disyllabic Classical Mongolian word qaɣan 'emperor' has a disyllabic Khitan equivalent


I would expect the Khitan cognate of čaɣan to be


I suspect that the Khitan word for 'white' was a monosyllable unrelated to čaɣan. KHITAN KHOLORS (PART 3)

Why did Khitan *siauqu 'green' have two spellings with different final elements?


This variation between and is also found in the spellings for masqu 'first':


Kane (2009: 143) observed that the first spelling precedes masculine nouns whereas the second spelling precedes feminine nouns. This implies that and represent adjective endings for different genders (plus a stem-final *-q if that's not part of the ending). Kane uses an acute accent to distinguish these elements in transcription:

qu vs.

The phonetic difference, if any, between the two is unclear. (Kane [1989: 19] reconstructed the latter as *e!) I am inclined to reconstruct the masculine ending with [ʊ] since 'first' has a third (erroneous?) spelling with qo:

[qʊ] is closer to [qo] than [qu].

Although I've only seen one spelling of *liauqu 'red' so far

I wonder if another spelling with was used before grammatically feminine nouns. 'Tiger' was preceded by

*siauqu 'green'

with the feminine ending in the date equivalent to Chinese 甲寅 on the Da Jin huangdi dutong jinglüe langjun xingji inscription, implying that 'tiger' was inherently feminine (or that 'tigress' was used in Khitan dates).

I was surprised to learn that Khitan may have had grammatical gender since such a feature is otherwise unknown in East Asia. Wu Yingzhe (2005) proposed that the dot in elements like indicates grammatical gender.

Next: One + water = white? KHITAN KHOLORS (PART 2)

Here's the solution to yesterday's question:

*liauqu 'red' and *siauqu 'green' have one or two elements in common:

'red' vs. and 'green'

They are phonetically identical except for the first consonant. They are always written with different top left elements. Therefore

= *l ... (at least; at this point *li, *lia, *liau, etc. cannot be rejected)

= *s ... (at least; at this point *si, *sia, *siau, etc. cannot be rejected)

If represented *li, *lia, *liau, etc.,

*lan 'orchid' (< Chinese 蘭)

should be read as *lian or *liaun. But it isn't. If we correlate the shared first consonants of *liauqu and *lan with the shared first element of their spellings, we can conclude that

= *l (rather than *li, *lia, *liau, etc.)

and guess that

= *s (rather than *si, *sia, *siau, etc.)

The final elements of *siauqu 'green'


also appear in the spellings for *masqu 'first':


Since *siauqu and *masqu share the syllable *qu in common, we can conclude that

and = *qu

So far we have identified the first and third elements of *liauqu

as *l and *qu. By process of elimination, the second element at the top right must be

= *iau

Thus *liauqu can be broken down into

= *l + *iau + *qu

Similarly, the spellings for *siauqu can be broken down into

= *s + *iau + *qu

= *s + *iau + *qu

Next: Why are there two spellings for *siauqu and *masqu? KHITAN KHOLORS (PART 1)

I thank Andrew West for giving me a tantalizing glimpse of Daniel Kane's 2009 book The Kitan Language and Script (only $190 with free shipping!). The lowest price I've seen so far is $165.52 for 304 pages (54 cents a page!). Google Books has a summary:

The Qidans [= Kitan / Khitan] established the Liao dynasty in northern China, which lasted for over two centuries (916-1125). In this survey the reader will find what is currently known about the Qidan language and scripts. The language was very likely distantly related to Mongolian, with two quite different scripts in use [the 'large script' and the 'small script']. A few generations after their state was defeated, almost all trace of the Qidan spoken and written languages disappeared, except a few words in Chinese texts. Over the past few decades, however, inscriptions from the tombs of the Liao emperors and the Qidan aristocracy have been at least partially deciphered, resulting in a significant increase of our knowledge of the Qidan lexicon, morphology and syntax.

Unfortunately, no online preview is available.

To see what a text in Khitan large script looks like, check out the Khitan memorial text on Andrew's blog (page 1 / page 2).

Since (1) I've been talking about Indo-European words for 'fire' in my last two posts and (2) what I've seen of Kane's book has set my brain on fire, I thought it'd be appropriate to look at the Khitan word for 'red which also represented the element 'fire'..

*liauqu 'red' was written as

in the large script. This character may be a distortion and simplification of Chinese 赤 'red'. (But not all large script characters are distortions of Chinese characters. Some are identical and some are totally different. The reason for this diversity is unknown.)

'Red' was written as a combination of three elements

in the small script. Can you figure out what the three components stand for?

= ?

= ?

= ?


1. Kane reconstructed 'blue'/'green' (also the element 'wood') as *siauqu which has two spellings:


The latter spelling is in the Da Jin huangdi dutong jinglüe langjun xingji inscription.

2. Chinese 蘭 'orchid' was borrowed into Khitan as *lan:

3. The word for 'first' has three different spellings:

and and

Kane reconstructed the first as *masqo and the others as *masqu. (I am deleting his diacritics for reasons I'll explain later.)

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