What does it mean?

September 07, 2006

Understanding Tangut

I've been following Amritas's Tangut analyses for quite some time now. From Wikipedia:

Tangut (also Xixia) is an ancient northeastern Tibeto-Burman language spoken in the Tangut Empire. By some linguists it is classified as belonging to the Qiangic languages. It is only distantly related to Tibetan and Burmese, and possibly also to Chinese.

Among the Qiangic languages one also notably finds Qiang and Rgyalrong.

This is the ancient official language of the Tangut empire (known in Tibetan as Mi-nyag and in Chinese as Xixia 西夏) which obtained its independence from the Chinese Song dynasty at the beginning of the 11th century, and was annihilated by Činggis Qaɣan (commonly known as Gengis Khan) in 1227.

The Tangut script, which Sofronov (1968) considered with reason to be one of the most complex in the history of humanity, was created by a decree of the emperor Li Yuanhao (李元昊) in 1038. The invention of the script was bestowed on Yeli Renrong (野利仁榮), a scholar close to the imperial family. After the destruction of the empire, the writing did not completely disappear, and it was used at least until the end of the 15th century.

The weird thing about Tangut is that we have a tremendous amount of knowledge about the language: 10,000 volumes of literature, most of which are translations of works we know from other languages, plus a native tradition of linguistic/grammatical analysis! But we still can't figure out how the Tangut script works! According to the Wikipedia link:

The script is presumed to have been designed by "The Teacher, Iri" under the supervision of the Emperor of the Tangut state, Li Yuanhao. It consisted of approximately 6,600 logographic characters built from radicals, in much the same way as they are in the Chinese script.

If the script were designed, you wouldn't expect it to consist of 6,600 random symbols. Take a look at Amritas's latest Tangut post. As you can see, the characters look much less like pictures than Chinese characters. But can you imagine memorizing 6,600 random characters like those chicken-scratches? Moreover, if you look at the characters long enough, vague patterns begin to emerge. Too vague to be definitive but too suggestive to be random. Amritas's favorite hypothesis, and the one that I'm convinced is correct, is that there are really two Tangut languages, which he calls Tangut A and Tangut B. Tangut A is the one for which we have phonetic knowledge. But Tangut B is the one which the characters represent phonetically. This isn't as unlikely as it sounds - Japanese has that sort of relationship with Chinese. What makes it less likely is that Chinese is the pricipal language of East Asia, while Tangut B is even more obscure than Tangut A.

However, there are several appealing things about this theory. For one thing, it would explain how Tangut writing could persist for hundreds of years after the Tangut empire was destroyed - in such a circumstance it must be relatively easy to learn.  But most of all, it would exactly explain the vague patterns we do, in fact, see - which I have tabulated below. Of course, the big problem with this theory is that there's no independent evidence of Tangut B. Oh well.

Phenomenon for similar characters Tangut A feature Tangut B feature
Similar sounds in Tangut A Words inherited from Tangut B The words sound similar in Tangut B "by chance" - i.e. the similar sounds don't reflect similar meanings
Words borrowed from Tangut B
Similar meaning in Tangut A Words not borrowed or inherited from Tangut B Similar sounds in Tangut B reflect similar meanings
Words evolved to the point where their phonetic relationship is unclear
Similar sound and meaning in Tangut A Words inherited from Tangut B
Words borrowed from Tangut B
Neither sound nor meaning similar in Tangut A Words not borrowed or inherited from Tangut B The words sound similar in Tangut B "by chance" - i.e. the similar sounds don't reflect similar meanings
Words evolved to the point where their phonetic relationship is unclear
Posted by David Boxenhorn at 10:48 PM  Permalink | Comments (0)
Trackback URL: http://blog2.mu.nu/cgi/trackback.cgi/185227

September 05, 2006

Alumni, of course!

For some time I have been wondering about how to organize an organization so that it doesn't become a self-perpetuating priesthood. How do you build in outside control that won't be captured by the bureaucrats? How do you make sure that the organization stays on course, doing what it was set up to do, rather than serving its own bureaucracy? At least for universities, I think I've found the answer: alumni. Of course! (Via Instapundit.)

Posted by David Boxenhorn at 07:46 AM  Permalink | Comments (0)
Trackback URL: http://blog2.mu.nu/cgi/trackback.cgi/184831