Tsou language
Encyclopedia
Tsou is a divergent Austronesian language
spoken by the Tsou people
of Taiwan
.
branch of Austronesian. However, several recent classifications, such as Chang (2006) and Ross (2009) dispute the Tsouic branch, with Tsou more divergent than the other two languages, Kanakanabu and Saaroa.
The Tsou language is spoken in the following villages (Li 1979, Zeitoun 2005). All of the villages are located in Alishan Township 阿里山鄉, Chiayi County
嘉義縣 except for Mamahavana 久美 (Jiumei), which is located in Hsinyi/Xinyi Township 信義鄉, Nantou County
南投縣. Both the native Tsou names and Chinese names are given.
Tapangu (Tapaŋʉ)
Tfuea (Tfuya)
Duhtu (Luhtu)
Iimucu - extinct
rather than long vowels or diphthongs. Vowels, especially back vowels, are centralized
when flanked by voiceless alveolar consonant
s (/t, ts, s/). This may involve a central offglide, so that /o/ is pronounced as a diphthong [öə̯] or [ɵə̯] in this environment.
The approximants
/w/ and /j/ may surface as non-syllabic mid vowel
s [e̯] and [o̯], even (for /j/) in initial position (/jo~joskɨ/ [e̯oˈe̯oskɨ] "fishes"; /w/ does not occur in initial position), explaining the spelling Tfuea (/tfuja/) for the name of the dialect. However, stress assignment ([ˈtfue̯a]) and restrictions on consonant clusters (see stress and phonotactics below) demonstrate that they behave as consonants.
The plosives are not aspirated. Phonetically aspirated stops are actually sequences of stop plus /h/, as can be seen by the fact that they cannot cluster with a third consonant (see phonotactics below), and by morphological alternations such as /phini/ ~ /mhini/ "to trade".
According to spectrum analysis
, /h/ appears to be a glottal fricative in most environments, but approaches a velar [x] next to the central vowel /ɨ/, as in /tsaphɨ/ 'palm, sole'. However, the fact that the sequences /hʔ/ and /ʔh/ occur, when no other homorganic sequence is allowed, suggests that /h/ and /ʔ/ may not both be glottal. (Additional evidence that /h/ might best be analyzed as velar is the fact that */kh/ is not found, and that /hk/ is only found medially, in the single known word /kuhku/ "fox".)
The voiceless sibilants, /ts/ and /s/, are palatalized to [tʃ] and [ʃ] before the front vowels /i/ and /e/. However, the voiced sibilant /z/ is not affected by this environment.
The implosives /ɓ/ and /ɗ/ are uncommon. Both may be glottalized ([ʔɓ], [ʔɗ] or maybe [ʔb], [ʔd]) in intervocalic position. In addition, alveolar /ɗ/ has some unusual allophony: About a third of speakers pronounce it with a lateral release
, or before /a/ as a lateral approximant [l], as in /ɗauja/ [lauja] "maple". Indeed, Tsuchida (1976) transcribed it as a preglottalized lateral, [ˀl].
if a moraic analysis is adopted. That is, a final heavy syllable (double vowel) receives stress ([eˈmoo] "house"); otherwise, stress falls on the penultimate syllable ([oˈkosi] "his child"). Additional stress falls in a trochaic pattern: Every other light syllable (single vowel) also receives stress. Unstressed vowels are deleted, except at word boundaries (initial or final vowel) and unless doing so would create a forbidden consonant cluster (see below).
For example, the verb //seʔe-nətəh-a// "to cut with a bolo" takes stress on the syllables //tə// and //ʔe//, and is realized as [sʔenˈtəha]. However, this does not explain all consonant clusters, many of which are lexically determined.
CV
V
. Tsou is unusual in the number of consonant cluster
s that it allows. Homorganic clusters are not allowed, unless one is a nasal consonant, and a maximum of two consonants may occur together, but otherwise about half of possible sequences are known to occur. For example, all non-homorganic sequences starting with /t/ and /ts/ are found. Missing clusters may not be allowed, or may simply be accidental gaps due to limited knowledge of the lexicon.
In clusters of oral stops, both have a release burst. This is true even between vowels, an environment where the first stop is unreleased
in most languages, supporting an analysis of these clusters as part of the syllable onset, with no syllable coda
s occurring in the language.
Stops, oral or nasal, may or may not have a release burst before a nasal stop, depending on the speaker. The initial clusters /hp, ht, hʔ/ are unusual cross-linguistically. The spectrum shows that the tongue moves towards an alveolar articulation during the /h/ of /ht/, demonstrating that it is not articulated as a velar. The initial clusters /pʔ/ and /tʔ/ are sometimes realized as two released stops, but sometimes with a single release, resembling ejective consonant
s in other languages. (/kʔ/ is again notably missing, except intervocalically, despite the fact that [kʼ] is the most common ejective cross-linguistically.)
, Tsou displays a predicate-initial syntax.
Tsou has three main types of questions (Zeitoun 2005:282).
Tsou has the following types of clauses:
Important function words are:
Case markers are as follows, with nominative forms placed before slashes and oblique forms placed after them (Zeitoun 2005:274). The nominative form is given when there are no slashes.
Tsou is unique for not having any preposition-like elements, instead using nouns or verbs to express these notions.
Tsou verbs can be divided into five major classes (I, II, III-1, III-2, IV, V-1, V-2) based on morphological alternations (Zeitoun 2005:285). Tsou verbs do not have as many morphological distinctions as other Formosan languages do, since the Tsou language makes more extensive use of auxiliary verbs. For instance, there are no temporal/aspectual distinctions, separate markings for imperatives, and stative/dynamic distinctions. Nevertheless, Tsou still preserves the causative poa- (allomorphs: p-, pa-).
Tsou auxiliary verbs can carry temporal/aspectual and modal information as well as voice. They are marked for the following voices:
These auxiliary verbs can be divided into three classes:
Tsou has the following aspectual suffixes:
Tens are derived with the circumfix
(confix) m- -hʉ. There is also a u/ʉ vowel harmony phenomenon.
Formosan languages
The Formosan languages are the languages of the indigenous peoples of Taiwan. Taiwanese aborigines currently comprise about 2% of the island's population. However, far fewer can still speak their ancestral language, after centuries of language shift...
spoken by the Tsou people
Tsou people
The Tsou are an indigenous people of central southern Republic of China . They are spread across three administrative entities of the Republic of China — Nantou County, Chiayi County and Kaohsiung City....
of Taiwan
Taiwan
Taiwan , also known, especially in the past, as Formosa , is the largest island of the same-named island group of East Asia in the western Pacific Ocean and located off the southeastern coast of mainland China. The island forms over 99% of the current territory of the Republic of China following...
.
Classification
Tsou has traditionally been considered part of a TsouicTsouic languages
The Tsouic languages are three Formosan languages, Tsou proper and the Southern languages Kanakanabu and Saaroa. The Southern Tsouic languages of Kanakanabu and Saaroa have the smallest phonemic inventories out of all the Formosan languages, with each language having only 13 consonants and 4 vowels...
branch of Austronesian. However, several recent classifications, such as Chang (2006) and Ross (2009) dispute the Tsouic branch, with Tsou more divergent than the other two languages, Kanakanabu and Saaroa.
Dialects
The dialectal variation of Tsou is not great. There are four recorded dialects, Tapangu, Tfuea, Duhtu, and Iimcu, of which Tapangu and Tfuea are still spoken. Iimcu is not well described. The grammar of the other three dialects is nearly identical, and phonological variation is marginal: In certain environments, Tapangu /i/ corresponds to Tfuea and Duhtu /z/ or /iz/, and Duhtu had /r/ for Tfuea and Tapangu /j/. (Actually, older speakers were recorded to vary between [r] and [j], but at that point the dialect was moribund.)The Tsou language is spoken in the following villages (Li 1979, Zeitoun 2005). All of the villages are located in Alishan Township 阿里山鄉, Chiayi County
Chiayi County
Chiayi County is a county in southwestern Taiwan surrounding but not including Chiayi City. Its historical name in Taiwanese Hokkien derives from Tsirosen in the Formosan languages...
嘉義縣 except for Mamahavana 久美 (Jiumei), which is located in Hsinyi/Xinyi Township 信義鄉, Nantou County
Nantou County
Nantou County is the second largest county of Taiwan. It is also the only landlocked county in Taiwan. Its name derives from the Hoanya Taiwanese aboriginal word Ramtau. Nantou County is officially administered as a county of Taiwan....
南投縣. Both the native Tsou names and Chinese names are given.
Tapangu (Tapaŋʉ)
- Tapangu 達邦 (Dabang)
- Nia'ucna/Nibiei 里佳 (Lijia)
- Saviki 山美 (Shanmei)
- Sinvi 新美 (Xinmei)
Tfuea (Tfuya)
- Cayamavana 茶山 (Chashan)
- Dadauya 樂野 (Leye)
- Ranguu/Punguu/Dadangia 來吉 (Laiji)
Duhtu (Luhtu)
- Mamahavana 久美 (Jiumei)
Iimucu - extinct
Vowels
Tsou has six vowels, /i ɨ u e o ɑ/. Vowel sequences occur, including sequences of like vowels (/ii/ etc.), but these are separate morasMora (linguistics)
Mora is a unit in phonology that determines syllable weight, which in some languages determines stress or timing. As with many technical linguistic terms, the definition of a mora varies. Perhaps the most succinct working definition was provided by the American linguist James D...
rather than long vowels or diphthongs. Vowels, especially back vowels, are centralized
Centralization
Centralisation, or centralization , is the process by which the activities of an organisation, particularly those regarding planning and decision-making, become concentrated within a particular location and/or group....
when flanked by voiceless alveolar consonant
Alveolar consonant
Alveolar consonants are articulated with the tongue against or close to the superior alveolar ridge, which is called that because it contains the alveoli of the superior teeth...
s (/t, ts, s/). This may involve a central offglide, so that /o/ is pronounced as a diphthong [öə̯] or [ɵə̯] in this environment.
Consonants
| labial | alveolar | velar | glottal | |
|---|---|---|---|---|
| nasal | m | n | ŋ | |
| plosive | p | t | k | ʔ |
| implosive | ɓ | ɗ~ˀl | ||
| affricate | ts~tʃ | |||
| fricative | f v | s~ʃ z~ʒ | h | |
| approximant | w | ɹ~j | ||
The approximants
Approximant consonant
Approximants are speech sounds that involve the articulators approaching each other but not narrowly enough or with enough articulatory precision to create turbulent airflow. Therefore, approximants fall between fricatives, which do produce a turbulent airstream, and vowels, which produce no...
/w/ and /j/ may surface as non-syllabic mid vowel
Mid vowel
A mid vowel is a vowel sound used in some spoken languages. The defining characteristic of a mid vowel is that the tongue is positioned mid-way between an open vowel and a close vowel...
s [e̯] and [o̯], even (for /j/) in initial position (/jo~joskɨ/ [e̯oˈe̯oskɨ] "fishes"; /w/ does not occur in initial position), explaining the spelling Tfuea (/tfuja/) for the name of the dialect. However, stress assignment ([ˈtfue̯a]) and restrictions on consonant clusters (see stress and phonotactics below) demonstrate that they behave as consonants.
The plosives are not aspirated. Phonetically aspirated stops are actually sequences of stop plus /h/, as can be seen by the fact that they cannot cluster with a third consonant (see phonotactics below), and by morphological alternations such as /phini/ ~ /mhini/ "to trade".
According to spectrum analysis
Spectrogram
A spectrogram is a time-varying spectral representation that shows how the spectral density of a signal varies with time. Also known as spectral waterfalls, sonograms, voiceprints, or voicegrams, spectrograms are used to identify phonetic sounds, to analyse the cries of animals; they were also...
, /h/ appears to be a glottal fricative in most environments, but approaches a velar [x] next to the central vowel /ɨ/, as in /tsaphɨ/ 'palm, sole'. However, the fact that the sequences /hʔ/ and /ʔh/ occur, when no other homorganic sequence is allowed, suggests that /h/ and /ʔ/ may not both be glottal. (Additional evidence that /h/ might best be analyzed as velar is the fact that */kh/ is not found, and that /hk/ is only found medially, in the single known word /kuhku/ "fox".)
The voiceless sibilants, /ts/ and /s/, are palatalized to [tʃ] and [ʃ] before the front vowels /i/ and /e/. However, the voiced sibilant /z/ is not affected by this environment.
The implosives /ɓ/ and /ɗ/ are uncommon. Both may be glottalized ([ʔɓ], [ʔɗ] or maybe [ʔb], [ʔd]) in intervocalic position. In addition, alveolar /ɗ/ has some unusual allophony: About a third of speakers pronounce it with a lateral release
Lateral release (phonetics)
In phonetics, a lateral release is the release of a plosive consonant into a lateral consonant. Such sounds are transcribed in the IPA with a superscript ⟨l⟩, for example as . In English words such as middle in which, historically, the tongue made separate contacts with the alveolar ridge for the ...
, or before /a/ as a lateral approximant [l], as in /ɗauja/ [lauja] "maple". Indeed, Tsuchida (1976) transcribed it as a preglottalized lateral, [ˀl].
Stress
With a few exceptions, stress is not only predictable, but shifts when suffixes are added to a word. It falls on the penultimate vowel, or on the penultimate moraMora (linguistics)
Mora is a unit in phonology that determines syllable weight, which in some languages determines stress or timing. As with many technical linguistic terms, the definition of a mora varies. Perhaps the most succinct working definition was provided by the American linguist James D...
if a moraic analysis is adopted. That is, a final heavy syllable (double vowel) receives stress ([eˈmoo] "house"); otherwise, stress falls on the penultimate syllable ([oˈkosi] "his child"). Additional stress falls in a trochaic pattern: Every other light syllable (single vowel) also receives stress. Unstressed vowels are deleted, except at word boundaries (initial or final vowel) and unless doing so would create a forbidden consonant cluster (see below).
For example, the verb //seʔe-nətəh-a// "to cut with a bolo" takes stress on the syllables //tə// and //ʔe//, and is realized as [sʔenˈtəha]. However, this does not explain all consonant clusters, many of which are lexically determined.
Phonotactics
The most complex syllable in Tsou is CConsonant
In articulatory phonetics, a consonant is a speech sound that is articulated with complete or partial closure of the vocal tract. Examples are , pronounced with the lips; , pronounced with the front of the tongue; , pronounced with the back of the tongue; , pronounced in the throat; and ,...
CV
Vowel
In phonetics, a vowel is a sound in spoken language, such as English ah! or oh! , pronounced with an open vocal tract so that there is no build-up of air pressure at any point above the glottis. This contrasts with consonants, such as English sh! , where there is a constriction or closure at some...
V
V
V is the twenty-second letter in the basic modern Latin alphabet.-Letter:The letter V comes from the Semitic letter Waw, as do the modern letters F, U, W, and Y. See F for details....
. Tsou is unusual in the number of consonant cluster
Consonant cluster
In linguistics, a consonant cluster is a group of consonants which have no intervening vowel. In English, for example, the groups and are consonant clusters in the word splits....
s that it allows. Homorganic clusters are not allowed, unless one is a nasal consonant, and a maximum of two consonants may occur together, but otherwise about half of possible sequences are known to occur. For example, all non-homorganic sequences starting with /t/ and /ts/ are found. Missing clusters may not be allowed, or may simply be accidental gaps due to limited knowledge of the lexicon.
| Initial or medial | Medial only |
|---|---|
| /pt, pts, ps, pn, pk, pŋ, pʔ, ph/ | /pz/ |
| /ft, fts, fk, fŋ, fʔ/ | /fn/ |
| /vts, vh/ | /vn, vʔ/ |
| /ɓn/ | /ɓk/ |
| /mp, mf, mts, ms, mz, mn, mʔ, mh/ | /mɓ, mt/ |
| /tp, tf, tv, tm, tn, tk, tŋ, tʔ, th/ | /tɓ/ |
| /tsp, tsf, tsv, tsm, tsn, tsk, tsŋ, tsʔ, tsh/ | /tsɓ/ |
| /sp, sv, sɓ, sm, sn, sk, sŋ, sʔ/ | — |
| — | /zʔ/ |
| /nm, nt, ns/ | /np, nv, nts, nz, nk, nʔ, nh/ |
| /ks, kn/ | /kts, kʔ/ |
| /ŋv, ŋh/ | /ŋm, ŋt, ŋts, ŋs, ŋz, ŋk/ |
| /ʔp, ʔv, ʔm, ʔt, ʔts, ʔs/ | /ʔf, ʔɗ, ʔn, ʔk, ʔh/ |
| /hp, hv, hm, ht, hts, hn, hŋ/ | /hs, hz, hk, hŋ/ |
In clusters of oral stops, both have a release burst. This is true even between vowels, an environment where the first stop is unreleased
Unreleased stop
An unreleased stop or unreleased plosive is a plosive consonant without an audible release burst. That is, the oral tract is blocked to pronounce the consonant, and there is no audible indication of when that occlusion ends...
in most languages, supporting an analysis of these clusters as part of the syllable onset, with no syllable coda
Syllable coda
In phonology, a syllable coda comprises the consonant sounds of a syllable that follow the nucleus, which is usually a vowel. The combination of a nucleus and a coda is called a rime. Some syllables consist only of a nucleus with no coda...
s occurring in the language.
Stops, oral or nasal, may or may not have a release burst before a nasal stop, depending on the speaker. The initial clusters /hp, ht, hʔ/ are unusual cross-linguistically. The spectrum shows that the tongue moves towards an alveolar articulation during the /h/ of /ht/, demonstrating that it is not articulated as a velar. The initial clusters /pʔ/ and /tʔ/ are sometimes realized as two released stops, but sometimes with a single release, resembling ejective consonant
Ejective consonant
In phonetics, ejective consonants are voiceless consonants that are pronounced with simultaneous closure of the glottis. In the phonology of a particular language, ejectives may contrast with aspirated or tenuis consonants...
s in other languages. (/kʔ/ is again notably missing, except intervocalically, despite the fact that [kʼ] is the most common ejective cross-linguistically.)
Syntax
Like most other Austronesian languagesAustronesian languages
The Austronesian languages are a language family widely dispersed throughout the islands of Southeast Asia and the Pacific, with a few members spoken on continental Asia that are spoken by about 386 million people. It is on par with Indo-European, Niger-Congo, Afroasiatic and Uralic as one of the...
, Tsou displays a predicate-initial syntax.
Tsou has three main types of questions (Zeitoun 2005:282).
- Yes-no questions
- Alternative questions
- Wh-questions (information questions)
Tsou has the following types of clauses:
- Verbal
- Declarative
- Imperative
- (Verbal) interrogative
- Equational
- Existential (no auxiliary verbs are allowed)
Important function words are:
- zou - "to be"
- a - "it is in case that"
- o'ta - (it is) not (in case that)"
- pan - "there is" / existential
- uk'a - negative existential (usually followed by ci)
- o'a - negation of a fact or event
- ci - relativizer
- o - prohibition (AV constructions)
- av'a - prohibition (UV constructions)
Case markers are as follows, with nominative forms placed before slashes and oblique forms placed after them (Zeitoun 2005:274). The nominative form is given when there are no slashes.
- e - visible and near speaker
- si / ta - visible and near hearer
- ta - visible but away from speaker
- o / to - invisible and far away, or newly introduced to discourse
- na / no ~ ne - non-identifiable and non-referential (often when scanning a class of elements)
Word classes
Tsou nouns are distinguished from verbs by the presence of case markers and suffixed genitive pronouns, both of which cannot be applied to verbs (Zeitoun 2005:264). Verbs, on the other hand, have elaborate voice marking. Adjectives and certain adverbs actually function as verbs, since they also undergo voice inflection and are placed at the same positions within clauses as verbs (i.e., predicate-initial).Tsou is unique for not having any preposition-like elements, instead using nouns or verbs to express these notions.
Verbs
Main verbs can take on three types of voices (Zeitoun 2005:284).- Patient voice: -a
- Locative voice: -i
- Instrumental/benefactive voice: -(n)eni
Tsou verbs can be divided into five major classes (I, II, III-1, III-2, IV, V-1, V-2) based on morphological alternations (Zeitoun 2005:285). Tsou verbs do not have as many morphological distinctions as other Formosan languages do, since the Tsou language makes more extensive use of auxiliary verbs. For instance, there are no temporal/aspectual distinctions, separate markings for imperatives, and stative/dynamic distinctions. Nevertheless, Tsou still preserves the causative poa- (allomorphs: p-, pa-).
Tsou auxiliary verbs can carry temporal/aspectual and modal information as well as voice. They are marked for the following voices:
- Actor voice (AV)
- Undergoer voice (UV)
These auxiliary verbs can be divided into three classes:
- AV constructions - mio, mo, mi-, moso, mo(h)-
- UV constructions - i-, o(h)-
- AV/UV constructions - te, ta, tena, nte, ntoso, nto(h)-, la
Tsou has the following aspectual suffixes:
- -cu/-c'u - already
- -n'a - still, just, about to
- -la - once
Pronouns
The personal pronouns below are from the Tfuya dialect of Tsou, and are sourced from Zeitoun (2005:265). Note that third-person pronouns are distinguished between those that are visible (abbreviated vis. below) or non-visible.| Type of Pronoun |
Free (neutral) |
Bound (nominative) |
Bound (genitive) |
|---|---|---|---|
| 1s. | a'o | -'o/-'u | -'o/-'u |
| 2s. | suu | -su/-ko | -su/-ko |
| 3s. (vis.) | taini | -ta | -taini |
| 3s. (not vis.) | ic'o | - | -si |
| 1p. (incl.) | a'ati | -to | -to |
| 1p. (excl.) | a'ami | -mza | -mza |
| 2p. | muu | -mu | -mu |
| 3p. (vis.) | hin'i | -hin'i | -hin'i |
| 3p. (not vis.) | hee | - | -he |
Numerals
Tfuya Tsou numerals are (Zeitoun 2005:265):- coni ; 10. m-as-kʉ
- yuso ; 20. m-pus-ku
- tuyu ; 30. m-tuyu-hu
- sʉptʉ ; 40. m-sʉptʉ-hʉ
- eimo ; 50. m-eimo-hʉ
- nomʉ ; 60. m-onmʉ-hʉ
- pitu ; 70. m-pʉtvʉ-hʉ
- voyu ; 80. m-voyvʉ-hʉ
- sio ; 90. m-sio-hʉ
Tens are derived with the circumfix
Circumfix
A circumfix is an affix, a morpheme that is placed around another morpheme. Circumfixes contrast with prefixes, attached to the beginnings of words; suffixes, that are attached at the end; and infixes, inserted in the middle. See also epenthesis...
(confix) m- -hʉ. There is also a u/ʉ vowel harmony phenomenon.
Further reading
- Tsuchida, S. (1976). Reconstruction of Proto-Tsouic phonology. Tokyo: Institute for the Study of Languages and Cultures of Asia and Africa, Tokyo Gaikokugo Daigaku.