Link Roundup – 6 Dec 2010

Since ’tis nearly the season: Chinese lyrics to Santa Claus is Coming to Town with fully tone-marked Pinyin. has scanned a section of a Pinyin book with a period (1963) blend of somewhat-simplified Hanzi orthography.

John Wells points to vowel inventories on the World Atlas of Language Structures (WALS) online. If you don’t think this sounds cool, check out the map here and mouse over some of the enormous differences you can see in vowel counts of languages across the region of Sinoglot interest.

edit: Don’t miss this awesome Rubick’s cube with Chinese movable type. I want one so bad. -kp

4 responses to “Link Roundup – 6 Dec 2010”

  1. Louis Platz says:

    This comment speaks to the vowel inventory link.

    Is it better to construct a broad vowel inventory for standard Chinese (SC) or a narrow inventory? That is, should we write 寫 as [ɕjɛ] or [[ɕje]?
    Some background information:

    (1) broad vowel inventory for SC
    i y – – u
    ɛ – – – ɔ
    e – ɘ ɤ o
    a – – – ɑ

    (2) narrow vowel inventory for SC
    i y – u
    – – ə –
    a – – –

    It’s not hard to derive (1) from (2).

    /ə/ becomes (>) [e] / before a high front vowel. 给 [kei] 鬼 [kwei]
    /ə/ > [ɛ] / after a high front vowel. 谢 [ɕjɛ] 雪 [ɕɥɛ]
    /ə/ > [o] / before a high back vowel 楼 [lou]留[liou]
    /ə/ > [ɔ] / after the back glide 博[pwɔ]
    /ɘ/ > [ɤ] / in syllable final position 乐 [lɤ]
    /ɘ/ > [ɘ] / before a nasal coda 更 [kɘŋ] 跟 [kɘn]
    (There is a difference between [ɘ] in [ɘŋ] and [ɘ] in [ɘn] due to assimilation towards the nasal codas.)

    /a/ > ɑ / before the high back vowel and the velar nasal [ŋ]. 光 [kwɑŋ] 老 [lɑo] 教 [tɕjɑu]
    /a/ > a / everywhere else 班 [pan]] 八 [pa] 管 [kwan]挂 [kwa]
    (I have a hard time differentiating [a] and [ɑ].)

    (In “The Sounds of Chinese” (2007), Yen-Hwei Lin says that /a/ > [ɛ] between a high front glide and the dental nasal (选 [ɕɥɛn]) . But, as we can find minimal pairs distinguishing [a] and [ɛ] (假 [tɕʰja] and 谢 [ɕjɛ]), I don’t see how one can assign [ɛ] as an allophone of the phoneme /a/. The professor teaching the SC phonology class I’m in ignores this observation.)

    I suppose the best method is to make your assumptions clear and then be consistent in your IPA spellings.

  2. Louis Platz says:

    The IPA spelling of the main vowel in 寫 could also be written as [ɕjə]. [ɕjə] underlies two phonological processes. The first is the fronting of [ə] when it is next to a high front vocoid — that is /ə/ > [e] when next [-i], [j-] or [ɥ-]. The second is the upward assimilation of [e] to [ɛ] when it follows the high glides [j-] or [ɥ-]. It seems to me that the separation of these two processes is too artificial, as they occur simultaneously. The writing [ɕje],then, is to avoid recognizing the tense-lax distinction in SC mid vowels. Doing so isn’t unreasonable, since the this distinction is phonetic — the lax variant [ɛ] does not contrast with the tense variant [e]. To summarize:

    [ɕjə]: underlying form.
    [ɕje]: a surface representation (SR) that shows fronting, but does not include the tense-lax distinction.
    [ɕjɛ]: a SR that shows fronting and the tense-lax distinction.

    Which one is best for general purposes? My feeling is [ɕjɛ], because it best represents what people (/I) say.

  3. Kellen says:

    I suppose the best method is to make your assumptions clear and then be consistent in your IPA spellings.

    Couldn’t agree with this more. Much discussion was had on the topic over at Annals of Wu a while back. It’s especially difficult and thus especially important when it comes to non-standardized languages like Wu.

  4. Louis Platz says:

    I came across a relevant paragraph in:

    Martin, Samuel E. 1957. “Problems of hierarchy and indeterminacy in Mandarin phonology”. Bulletin of the Institute of History and Philology 29.1: 209-229.

    I’ll reproduce it here.

    The analyst at work often faces the dilemma “Should I conceal some of the data to reveal pattern, or should I conceal pattern in order to include all of the data. It seems to be axiomatic that the more data acquired (and the more closely it is examined), the more ragged becomes the pattern. The inescapable conclusion: language patterns are more complex than linguists’ pictures of them suggest. Perhaps an analogy with cartography is valid: no map includes all the facts about a given piece of terrain, but only those facts thought relevant for the purpose of the map.

Leave a Reply