Tsou at Luhtu [1]

- Fredric F. Weingartner - 

	In order to read the entire contents of this essay, you must first download and install SILDoulosIPA.
	How to Install
	Download

Introduction

For a synchronic study of an unwritten minority language which is a relic of a larger speech community that has ceased to exist, the linguist will collect data and process them applying established methods of linguistic research.  When, however, the speech community from which the isolated group is an offshoot is known, a comparative study can be made.  Such a study will be easier, if both the speech of the parent community and its offshoot have been studied and described by linguists, who have discovered significant differences between them.  New data will then have to be consulted from both speech communities for comparison with the earlier descriptions.  If a new description of the parent language but not of the daughter language exists, the investigator will only have to provide recent data of the daughter language.  The comparison can then be expected to yield information on the amount and the direction of linguistic change within a well-known time span.  The situation just described was found for Tsou at Luhtu [2]. 

 1. Sources

A short description of Tsou as parent and Luhtu as daughter language was published by Tung Tung-ho in 1963.  He describes the Luhtu people as “ the survivors of a once flourishing branch of the Tsou tribe” which was in the late fifties already “a very small minority in the village where they are now living”.   

In 1979 Li R.K. published his study on 'Variations in the Tsou Dialects' wherein he considers Duhtu – as he calls it [3] – as the older variety of the three northern Tsou dialects he describes.  When the author recorded Luhtu data in 1998 that minority had shrunk to a handful of individuals, most of them advanced in age.  They still recognize the names of Tung’s informants and Mamahavana, the original name of the village, when Tsou was still dominant in the region.

The same year the Crane Company published a description of the Tsou language for speakers of Tsou, mainly at Tfuye [4], hence called The Primer.

The division according to age groups at Luhtu is similar to that of other indigenous tribes in Taiwan [5].  People under 20 speak only Taiwan Mandarin, those above 70 know Tsou and Japanese.  Natives between 20 and 70 besides Mandarin know the two aboriginal languages, Tsou and Bunun, the latter being the dominant language.  All of them can be called multilinguals even though the fluency with which they speak these languages varies greatly.   

Fortunately Tung provides a great wealth of background information on Luhtu, in 29 texts of varying length.  Twenty-eight of them begin with the words <moso> and/or <noanao> [6] 'past'.  They tell myths handed down through the ages, similar to those related by other Formosan tribes [7]. 

The last text begins with <maitan/e> 'today' or 'nowadays'.  There the informant Tasura Luheatsana, then 50 years old, in speaking with his investigator, compares the old days with the present time.  He mentions among other things that he has learned the Chinese language.  The text may have been recorded around the year 1960, that is fifteen years after the Japanese occupation of Taiwan ended. 

2. Informants 

Only elderly people were recommended, as good speakers of Luhtu.  Table one gives their data of our informants. 

Table 1 : Informants 

Li Ren-kuei's informant Avai Tenayana, 吉太平 [8], was a 74 years old man.   

3. Methods of collecting data  

In a situation where the investigator does not speak the language, from which she or he is attempting to collect data, and the informants only speak their own language, a primitive method of pointing to things may bring some results.  There is, however, the danger that asking for the word glass may yield the answer for window. 

We adopted the method of the Primer which shows pictures to the informant [9].  This method also has its limitations, which became evident when we tried to elicit the words for single digit numbers.  The Primer arranges flowers on a string [10], but does not ask for the number one, which most likely would have yielded the answer 'flower'.  Nor does the picture show two flowers in order to elicit the number two.  For although there is a free form<euso> 'two' it takes many different forms, depending on what is counted : 'Two trees' is <mehaeso> [11] and 'two days' is <mipsohi> [12].  Once the picture was explained by a native, the counting of free forms went without a hitch and even included the number 'ten'.  See table 5, page 26.   

The Primer does not contain abstract items such as 'peace' or 'health', and the recording became difficult when grammatical terms such as 'singular' 'third person', tenses and moods were asked. 

For time adverbs the Primer provides an ingenious method showing an arrow [13] moving up and down from a position meaning 'today' to 'tomorrow' or 'yesterday' and so on.  <Hucma> 'short time span' can than mean 'tomorrow' or 'yesterday', depending on which particle precedes <ho> or <ne>.  

We supplemented the above with further questions and focused our study on the 144 items listed in Table 2, here called "Basic Vocabulary".   

Table 2 : Basic Vocabulary  

 4. Processing data

a. Consonants :

Uncommon consonant symbols in the Primer. 

[/] Glottal Stop

[j]  Palatal Approximant

[ts] Alveolar Affricate

[º]      Bilabial Implosive

[ë] Alveolar Implosive

[R]  Uvular flap

1. Glottal stops are formed by aligning the vocal cords. They are sounds, not diacritics.  

2. In approximants the tongue approaches its target but withdraws before it has reached it.  In the case of [j] the target is the front palate.  [j] is only found before vowels, the Primer  writes  <y>. 

3. Affricates are stops followed by a fricative.  They count as one segment written in the Primer as <c>. 

4. Tung Tung-ho calls his voiced 'stops' [14] preglottalized, describing a process in which "the vocal cords make a closure of the glottis before vibrating" [15].  Von Essen calls that 'hard insertion' [16].  Scholars in Taiwan describe it as 'first close then open' 緊閉後開 [17], but others direct the intention to what makes that articulation special, using "inhaled" 吸氣 [18] and 吸入 [19] indicating that it is the ingressive air stream and not pulmonic air, what is crucial for the articulation of these voiced sounds. 

5. Using [ l ] as a symbol for the implosive [ë] seems rather inappropriate, if the symbol denotes a lateral articulated with pulmonic air. 

6. A uvular vibrant 搖音 is a single flap of the loosely hanging tiny continuation of the back palate, caused by the air streaming into the oral cavity.  In fact it is an approximant.  The phonetic symbol is an inverted small capital [Ò] [20].  Such vibrants are considered specific for Tsou at Luhtu.  They were recorded by two of our informants.  Wcd in item zou : [Ò] for <z> in [Òuo] and Kam in item naveu : [Ò] for [e] in navÒo [21]. 

7. Consonant clusters in Luhtu consist of only two sounds.  Except for the glottal stop they may belong to different morphs.  /c/ [ts] is not considered a cluster and <ng> is written <N>. 

b. Vowels :

The Primer lists six vowels.  Besides the traditional sounds a, e, i, o, and u there is / ¬ /, described as a high back vowel pronounced with lips spread.  Our informants pronounce the [¬] almost like [u], in agreement with the findings of Tung Tung-ho[22].  All sequences of vowels up to five in a row are found, but their number of instances decreases as the number of vowels increases. 

c. Speech Units :

Speech units are meaningful elements.  They may consist of one or more syllables which make up words.  Units which consist of one or more words, which then are connected in a special manner, are called compounds.  Each unit has only one primary stress. 

 [1] Syllable

Tsou syllables consist of several elements.  They begin with one or two consonants or with a single vowel.  They have only one vowel.  The syllable nucleus is one of the six vowels of the Tsou language.  Tsou syllables never end in consonant clusters and seldom in consonants other than a glottal stop, which is always a syllable boundary. 

 [2] Stress

In describing Tsou it does not seem necessary to distinguish various degrees of stress (primary, secondary, tertiary) as against un-stressed.  All isolated forms are stressed including monosyllables. 

The location of stress is defined in terms of syllables counted from the end of a speech unit.  Polysyllabic units of speech have only one stress which in Tsou is normally placed on the second -to- last (penultimate) syllable.  The peak of a stressed syllable is called its nucleus.  It is a vowel, which is pronounced louder than the other elements of the syllable. 

There follow some samples of stress location in Tsou units.  In contrast to the Primer, grave accents marks are placed before the first sound of syllables.  Table 3 contains many non-basic items which were collected at random.   

Table 3 : Samples of Stress Location  

 d. Answers

The answers of informants showed either full agreement with the Primer or not.  The negative responses were in either zero or partial agreement with the Primer as in Table 4. 

 Table 4 : Answers Given by Informants  

Detailed answers are given in Table 5.  It has six columns.  Column One gives the item in the sequence of the Primer.  Column Two contains the transcription of the Primer.  Column Three names the informants.  Column Four reports these answers.  The final column explains elements in these answers and tries to give reasons for them based on general linguistic knowledge. 

 The following symbols and spellings are used :

 [  ]  phonetic  /  / phonemic  <  > orthographic

< y > before vowels is < j >, after vowels it is < i >. 

< u > before vowels is < w >, after vowels it is < u >. 

 Table 5 : Different Renderings