Michael Everson everson at
Tue Sep 9 09:32:10 CEST 2008

On 9 Sep 2008, at 03:37, John Cowan wrote:

> Doug Ewell scripsit:
>> The question of whether to assign "zh" or "zh-Latn" as the Prefix for
>> 'pinyin' has NOTHING to do with generic variants.  Where did that  
>> come
>> from?  Nobody has seriously proposed making 'pinyin' a generic  
>> variant
>> like 'fonipa'.  At most they have suggested that a small set of
>> languages be added as multiple Prefix fields.
> On the contrary, that is exactly what Michael has proposed: making the
> subtag 'pinyin' mean 'any romanization called pin1yin1 in Chinese,  
> except
> for wei1tuo3ma3 pin1yin1, otherwise known as Wade-Giles romanization.
> I think this proposal is totally wrong-headed, to be sure.

I don't. I don't see an argument for setting the prefix, as we did not  
do so for similar Latin orthographic entities. I don't see Pinyin as  
significantly different from IPA or UPA. All three of them use Latin  
letters for orthographic purposes based on a set of agreed  
conventions. In the case of Pinyin, the value of <q> and <x> for  
instance are iconic. IPA may not be a perfect analogy here, but UPA  
certainly is. I grant that more of you know something about IPA than  
about UPA, but even so. There is a lot of variation in IPA usage.

In the first place, what does IPA mean? IPA 1949? IPA 1999? IPA broad  
or narrow, phonetic or phonemic? All of those choices offer the same  
kind of variation that we find between Mandarin Pinyin and Tibetan  
Pinyin and Tongyong Pinyin (Tongyong Pinyin and Hanyu Pinyin are 80%  
the same). In Phonemic transcription, IPA conventions are looser than  
narrow phonetic transcription: In Cornish for instance [eː] and [ɛ]  
are /eː/ and /e/.

If you have a transducer  to convert from zh-cmn-pinyin to zh-cmn- 
wadegile and then to zh-TW-pinyin the transducer will be using three  
different dictionaries in any case. The tags are mnemonic, not  

22:10:32 John Cowan: I still think that 'pinyin' should refer solely  
to Hanyu Pinyin, which is what 95% of all references to 'Pinyin' in  
English refer to, and that other pin1yin1s should get their own subtags.
22:11:58 Michael Everson:

There are 28 Sino-Tibetan nationalities in China, and at least 25  
others of different language groups. A fair percentage of these have  
romanizations (whether transliterations or transcriptions) based on  
Pinyin conventions. This *is* a generic subtag like "fonupa" and  
"fonipa" are (even though the scope of "fonipa" is wider than that of  
either "pinyin" or "fonupa"). Restricting the subtag "pinyin" to  
Mandarin Chinese would be like restricting the subtag "fonupa" to  
Sami. I do not agree that Tibetan Pinyin should have a subtag of its  
own "bopinyin"; it shold be "bo-pinyin" alongside "zh-pinyin", with or  
without "Latn" as required. Tongyong pinyin text too should be  
specified by an appropriate concatenation of subtags, not by something  
like "typinyin".

Michael Everson *

More information about the Ietf-languages mailing list