Nuosu language

Dialects

Lama (2012)

Lama (2012) gives the following classification for Nuosu dialects.

• Nuosu
  • Qumusu (Tianba)
  • Nuosu proper
    • Nuosu
      • Muhisu
      • Nuosu ()
        • Yinuo
        • Shengzha
    • Niesu ()
      • Suondi
      • Adu

The Qumusu (曲木苏, Tianba 田坝) dialect is the most divergent one. The other dialects group as Niesu (聂苏, Suondi and Adu) and as Nuosu proper (Muhisu 米西苏, Yinuo 义诺, and Shengzha 圣乍). Niesu has both lost voiceless nasals and developed diphthongs.

Adu (阿都话), characterized by its labial–velar consonants, is spoken in the Butuo and Ningnan counties of Liangshan Yi Autonomous Prefecture, Sichuan province, and also in parts of Puge, Zhaojue, Dechang, and Jinyang counties.

Nyisu or Yellow Yi (黄彝) of Fumin County, Yunnan may either be a Soundi Yi (Nuosu) dialect or Nisu dialect.

Zhu and Zhang (2005) reports that the Shuitian people (水田人) reside mostly in the lowlands of the Anning River drainage basin, in Xichang, Xide, and Mianning counties of Liangshan Prefecture in Sichuan. They are called Muhisu (mu33 hi44 su33) by the neighboring Yi highland people. Shuitian is spoken in the following locations. Shuitian belongs to the Shengzha dialect (圣乍次土语) of Northern Yi.

• Mianning County: Jionglong 迥龙, Lugu 泸沽, Hebian 河边; Manshuiwan 漫水湾
• Xichang: Lizhou 礼州, Yuehua 月华
• Xide County: Mianshan 冕山镇 (including Shitoushan Village 石头山村), Lake 拉克

Bradley (1997)

According to Bradley (1997), there are three main dialects of Nosu, of which the Southeastern one (Sondi) is most divergent.

• Northern
  • Tianba 田坝 a.k.a. Northwestern
  • Yinuo 义诺 a.k.a. Northeastern
• Central (Shengzha 圣乍)
• Southeastern (Sondi)
  • Sondi
  • Adur

Chen (2010)

Chen (2010) lists the following dialects of Nosu. Also listed are the counties where each respective dialect is spoken.

• Nosu 诺苏方言
  • Senza, Shèngzhà 圣乍次方言
    • Senza, Shèngzhà 圣乍 (): 1,200,000 speakers primarily in Xide, Yuexi, Ganluo, Jinyang, Puge, Leibo, Xichang, Dechang, Mianning, Yanyuan, Yanbian, Muli, Shimian, Jiulong, and Luding; also in Huaping, Yongsheng, Ninglang, Lijiang, Jianchuan, Yongshan, and Qiaojia
    • Yino, Yìnuò 义诺 (): 600,000 speakers primarily in Meigu, Mabian, Leibo, and Ebian, Ganluo; also in Yuexi, Zhaojue, and Jinyang
    • Lidim, Tiánbà 田坝 (): 100,000 speakers primarily in Ganluo, Yuexi, and Ebian; also in Hanyuan
  • Sodi, Suǒdì 所地次方言 (): 600,000 speakers primarily in Tuoxian, Huili, Huidong, Ningnan, Miyi, Dechang, and Puge

Phonology

Consonants

Gerner (2013) and Edmondson, Esling & Ziwo (2017) use the Sinological symbol to transcribe the alveolo-palatal nasal.

Eatough (1997)'s chart and transcriptions slightly differ from the later sources:

• The retroflex fricatives and affricates are notated as plain postalveolar (i.e. with instead of )
• The alveolo-palatal series is notated as palatalized (post)alveolar (i.e. with instead of )
• The alveolo-palatal (palatalized) nasal includes a voiceless pairing in concordance with the alveolar nasals (i.e. with in addition to )

Eatough (1997) and Gerner (2013) transcribe the voiceless lateral as an approximant , while Edmondson, Esling & Ziwo (2017) transcribe it as a fricative . See § Assimilation for descriptions of syllabic realizations of the laterals.

Gerner (2013) segments the bilabial trill as an allophone of the following set of consonants before the vowel phonemes , while Eatough (1997) and Edmondson, Esling & Ziwo (2017) segment it as a syllabic realization of the same vowels (see § Fricativized vowels):

• as an allophone of
• as an allophone of
• and as allophones of
• as an allophone of

Vowels

Eatough (1997)

Gerner (2013)

Edmondson, Esling & Ziwo (2017)

Phonation

Nuosu has five pairs of phonemic vowels, contrasting in a feature Eatough (1997) calls loose throat vs. tight throat, while Edmondson, Esling & Ziwo (2017) call it lax vs. tense; Gerner (2013) simply describes the contrast as with and without laryngealization. Underlining ⟨⟩ is used as an ad hoc transcription for tight/tense phonation, used both phonemically and phonetically by Eatough (1997), and phonemically by Edmondson, Esling & Ziwo (2017). Gerner (2013) treats the contrast both phonetically as and with the transcription of creaky voice ⟨⟩, while Edmondson, Esling & Ziwo (2017) treat it phonetically as laryngeal register and with the transcription of retracted tongue root ⟨⟩. According to Eatough (1997), vowel quality (height) is a byproduct of loose vs. tight throat, working as a reinforcement; the phonation type is the primary phonemic distinction, rather than the quality. Edmondson, Esling & Ziwo (2017)'s analysis shows this reinforcement likewise extends to the fricativized vowels (syllabic fricatives). According to Gerner (2013) and Walters (2022), a syllable-final ⟨-r⟩ is added in the Nuosu Pinyin orthography to indicate vowels with tense/tight/creaky phonation.

Quality

Eatough (1997) notes as sounding slightly closer (higher) than their cardinal values. Edmondson, Esling & Ziwo (2017)'s analysis found this to be true for , while they found to be much closer (higher) and analyze them instead as more similar to , and found to be slightly more open (lower) than its cardinal value, the opposite of Eatough (1997)'s conjecture.

According to Eatough (1997), the tight throat vowel occurs as a phonetic realization of the phonemically loose throat (Edmondson, Esling & Ziwo (2017)'s ) in the high tone, due to raising of the larynx to produce high pitch; it only occurs in this tone. The loose throat never occurs in this tone, and as such there is no three way contrast known to exist in any tone between .

According to Edmondson, Esling & Ziwo (2017), vowels are often nasalized after nasal consonants. The tense (tight) vowels often have a schwa offglide .

Fricativized vowels

The fricativized vowels (also called fricative vowels or syllabic fricatives) in Nuosu are the close vowel phonemes:

• unrounded or
• rounded or

The initial fricative elements of the fricativized vowels (which Edmondson, Esling & Ziwo (2017) state have also been notated as ) may be voiceless or voiced bilabial trills after bilabial and alveolar plosives; both the lax and tense variants may be either type of voicing, depending on the context. Eatough (1997) suggests that the realization is more lenis after bilabial plosives, and more fortis after alveolar plosives. The vowel elements may be a range of , being fronted especially after more forward consonants. The fricative (or trill) elements may occasionally be reduced to offglides , or entirely neutralized. Eatough (1997) transcribes this process instead as bilabially trilled onglides while the fricativized vowel retains its form .

The fricativized vowels (which Eatough (1997) notates as ) are described by Edmondson, Esling & Ziwo (2017) as 'voiced alveolar fricative syllabic continuants', which are apical and approximated (more open), and with the tongue position between and ; they are more accurately transcribed as , or with the Sinological symbol . They may also be fronted to or become fully rhotic after alveolo-palatal consonants. The retroflex pair are described as retroflex equivalents of this articulation; they are more accurately transcribed as , or with the Sinological symbol , and are shown to be allophones of after retroflex consonants. In addition to the assimilated forms described below, Eatough (1997) suggests complementary distribution of (or per the later authors' transcriptions ) and equivalent for the tight (tense) form, based on place of articulation of consonant onsets, which is roughly corroborated by Edmondson, Esling & Ziwo (2017) in further detail.

Assimilation

The fricativized vowels show systematic assimilation to preceding lateral continuants and bilabial nasals, resulting in the formation of syllabic lateral and nasal consonants.

According to Edmondson, Esling & Ziwo (2017), the fricativized vowels may assimilate with the laterals to form syllabic lateral–median approximants (roughly for and for ), which they transcribe as to indicate quality between and ; the voicing type is determined by the voicing of the laterals, not by vowel tenseness (tightness), and the combined forms in the case of the voiceless laterals indicates lengthened (geminated) with changing voicing. The initial voiceless onsets may be reduced to , resulting in the forms . Eatough (1997) shows almost the same process, but treats them as simple laterals . Gerner (2013) provides a similar but narrower analysis, stating only unrounded central assimilates with to form syllabic . Both Eatough (1997) and Edmondson, Esling & Ziwo (2017) show both lax and tense (loose and tight) forms of the syllabic laterals, while Gerner (2013) does not.

In a similar case as the laterals, all authors show some degree of fricativized vowel assimilation with the bilabial nasals . Gerner (2013) treats this identically to the laterals; only unrounded central assimilates to form syllabic . Eatough (1997) also treats this closely to the laterals, where are labialized, and the voiceless variants are lengthened with changing voicing. However, rather than forming simple nasals, Eatough (1997) reports them as forming co-articulated nasals and laterals ; this results in the forms . Edmondson, Esling & Ziwo (2017) do not include (extra-)labialized forms, but do note the same quality of changing voicing, and still show as assimilating under the process just as do; this results in the forms . Just as with the laterals, both Eatough (1997) and Edmondson, Esling & Ziwo (2017) show both lax and tense (loose and tight) forms of the syllabic nasals, while Gerner (2013) does not.

Nuosu syllable structure is (C)V.

Tones

According to Gerner (2013) and Walters (2022), the Nuosu Pinyin orthography indicates tones with the following letters at the end of syllables:

• high / – written ⟨-t⟩
• high-mid / or mid falling / – written ⟨-x⟩ (written with the diacritic ⟨◌̑⟩ over symbols in the syllabary)
• mid / – unmarked
• low falling / – written ⟨-p⟩

The high-mid tone is only marginally contrastive. Its two main sources are from tone sandhi rules, as the outcome of a mid tone before another mid tone, and the outcome of a low-falling tone after a mid tone. However, these changes do not occur in all compounds where they might: for instance ꊈ wo "bear" + ꃀ mop "mother" regularly forms ꊈꂾ wo mox "female bear", but ꃤ vi "jackal" + ꃀ mop "mother" forms ꃤꃀ vi mop "female jackal" without sandhi. The syntax creates other contrasts: tone sandhi applies across the boundary between object and verb, so is present in SOV clauses like ꃅꏸꇐꄜꎷ mu jy lu ti shex "Mujy looks for Luti", but is absent in OSV clauses like ꃅꏸꇐꄜꎹ mu jy lu ti shep "Luti looks for Mujy". A few words, like ꑞ xix "what?", have underlying high-mid tone.

Writing system

Classic Yi is a syllabic logographic system of 8,000–10,000 glyphs. Although similar to Chinese characters in function, the glyphs are independent in form, with little to suggest a direct relation.

In 1958 the Chinese government had introduced a Roman-based alphabet based on the romanized script of Gladstone Porteous of Sayingpan. This was later replaced by the Modern Yi script.

The Modern Yi script (ꆈꌠꁱꂷ nuosu bburma 'Nosu script') is a standardized syllabary derived from the classic script in 1974. It was made the official script of the Yi languages in 1980. There are 756 basic glyphs based on the Liangshan dialect, plus 63 for syllables only found in Chinese borrowings. The government requires the use of the script for signs in some designated public places.

Vocabulary and grammar

Nuosu is an analytic language. The basic word order is Subject–object–verb. Vocabularies of Nuosu can be divided into content words and function words. Among content words, nouns in Nuosu do not perform inflections for grammatical gender, number, and cases, classifiers are required when the noun is being counted; verbs do not perform conjugations for its persons and tenses; adjectives are usually placed after the word being fixed with a structural particle and do not perform inflections for comparison. Function words, especially grammatical particles, have a significant role in terms of sentence constructions in Nuosu. Nuosu does not have article words, but conjunctions and postposition words are used.

Numbers

Classifiers are required when numbers are used for fixing nouns.