The language treated in this volume is spoken by a group of Formosan aborigines of the Mt. Ali area who call themselves /cou/. In Japanese Kana /cou/ is generally transcribed as ???, and in Chinese writing it is in most cases rendered as 鄒 according to the pronunciation of the character in standard Mandarin, /tsou/, but rarely as 曹 according to the South?Min pronunciation of the character, /tso/, which is obviously less preferable. There is another transcription found in Chinese literature as 朱歐 (/tsu ou/), apparently a second-hand derivation from Japanese. In romanization “Tsou” seems to be the commonly accepted form.

The term Tsou is further used in a broader sense among some anthropologists. It designates also the people who call themselves /la?alua/ and those who call themselves /kanakanavu/, each living in a different area to the south of Mt. Ali. According to those anthropologists, the Tsou people of Mt. Ali are the “Northern Tsou” while /la?alua/ and /kanakanavu/ are the “Sothern Tsou". We are not even in a position to comment on such a classification merely from a linguistic point of view, since the languages of the other two groups of people are still very inadequately known to us. For our purpose we should like to make only one point, that the term Tsou refers to neither /la?alua/ nor /kanakanavu/ in our present study.

The main peak of Mt. Ali is about two thousand meters above sea level. The Tsou people now scatter to the west of it, in villages or settlements around one thousand meters in altitude. Distances among them vary greatly. While it takes only one how or less to walk from some settlements to the next one, it takes one whole day from some others to their neighbor. In terms of the present-day administrative districts, all Tsou settlements are in Wu-feng Hsiang, Chia-yi Hsien, except one belonging to Hsin-yi Hsiang, Nan-t'ou Hsien.

At present the Tsou people of Wu-feng Hsiang, Chia-yi Hsien form the main body of the tribe. There are about two thousand or more of them. Regardless of the Japanese administrators and school masters of the former days and their Chinese counterparts at present, no other people seem to have lived together with the Tsou in this area. Though life has considerably changed, they still retain their old tradition to a certain extent. Only very few intellectual old men and the young generation can speak Japanese or Chinese as their second language.

The Tsou people of Hsin-yi Hsiang, Nan-t'ou Hsien, have a population of only a few dozens, being the survivors of a once flourishing branch of the Tsou tribe. They are a very small minority even in that village where they are now living. The people dominant in that village and that whole area are the Bunun, the most populous aborigines in middle and southern Taiwan. It seems to us that the Tsou people there have ceased to keep any of their old traditions. Furthermore, owing to intermarriages between the two peoples and the dominance of the Bunun, there are now but very few Tsou who do not know the Bunun language and it is not infrequent that a Tsou man has a wife and children who do not speak his ancestor language at all. Another feature of the Tsou people of that place is that the comparatively few purely Tsou speaking families generally have some connections with their tribesmen of Wu-feng Hsiang, especially with those of Tfuea which is said to be the oldest settlement of the Tsou tribe. They visit Tfuea or have visitors from there once in a while. They are perhaps the conservatives, some with housewives recently from Tfuea.

According to the tradition handed down from generation to generatfon, the Tsou tribe formerly had four major branches, namely: Tapangu (/tapa?u/), Tfuea (/tfuea/), Luhtu (/luhtu/) and Iimutsu (/iimucu/) each with its own dialect. No Iimutsu descendants are still existing. The Luhtu people have but one small portion of them left, known among the Tsou people specifically as Mamahavana (/mamahavana/) which is also the native name of the village in Hsin-yi Hsiang, Nan t‘ou Hsien, where they live at present. People of the Tapangu and Tfuea branches are those now scattered in Wu-feng Hsiang, Chia-yi Hsien. The Tapangu settlements are found to the left of a deep valley and the Tfuea ones to its right. With a suspension bridge recently built, the main settlement of Tapangu and the main settlement of Tfuea have come quite near to each other.

To our knowledge the three dialects now existing differ from one another but very slightly. In general, there is hardly any difficulty in their communicating with one another. Some natives say that the differences of the three dialects are largely “accentual.” To use our terminology they are of some minor phonological and lexical features only, and there is scarcely any grammatical discrepancy being observed. In so far as our material is concerned, the Luhtu dialect does not show any trace of the influence exercised by the Bunun language surrounding it. That is perhaps because of the conservatism of our informants. On the other hand, there are apparently a lot of Tfuea elements in their speech.

In all three dialects, Japanese or Chinese loan-words are frequently used for things or concepts of modern civilization.

Our investigation of the Tsou language started in the summer of 1957 when a group of two dozen students of anthropology led by Professor E. W. Tu and the present author were practising field methods in the Mt. Ali area. Though not very much usable linguistic material was obtained on that occasion, the present author was convinced that there should be a more thorough and systematic study of the language besides the works previously done by Professors A. Ogawa and K. Asai in The Myths and Traditions of the Formosan Native Tribes (Texts and Notes) (Taihoku University, 1935, pp. 669-692) and Professors H. L. Wei and H. L. Lin in The Tsou People (Taipei, 1952).

In the next year a three-month investigation was carried out, jointly sponsored by Academia Sinica and Taiwan University. Working together with the present author were three young men: Mr. S. H. Wang, Mr. T. K. Kuan and Mr. T. F. Cheng. All Tsou settlements except two newly established ones were visited. In each place at least two informants, mostly old folk over fifty, were interviewed. When some of them could speak Japanese or South-Min Chinese freely, we communicated directly with them. In case someone spoke but his native language, young people knowing standard Chinese were asked to be interpreters. A list of the informants and interpreters will be found in the in the introductory remarks of Part II: Texts and Translation (pp.235-237). Most of our records obtained from them are in notes. There are also some tape recordings.

In the summer of 1959, after we had completed a preliminary study of the material, our informant/interpreter Mr. An was kind enough to come to us, helping in the final check of the phonemic transcription of the texts, in the transcribing of the tape recordings, and in testing the validity of our interpretations of the structure of the language.

The procedures in our study of the language may not differ very much from those of other linguists. We nevertheless would like to make the following points which, we believe, might lead to a clearer understanding of the results of our study.

1) No pre-arranged “word list” of any sort is used.-We do, however, start our investigation by asking questions as: “What is...called in your language?” (pointing to something), “How do you say...in your language?” (referring to some Chinese or Japanese expression). It should be noticed that in noting down what we hear from the informant we take them as some actually occurring utterances of the language; but not as “words” or “phrases” in abstract which hardly mean anything when we have not yet any idea about the structure of the language. To borrow Leonard Bloomfield’s term, they are “completive answers.” In our case, specifically, they occur after the questions of the investigator put in a different language. Thus we make a rule to note down their intonations in addition to the “vowels” and “consonants” we have discerned in them. Yet the purpose of having a certain amount of them in the first stage of our investigation is primarily for the convenience in our recognizing distinctive and non?distinctive sound features of the language. They are usually short utterances, not too complicated for us to operate with.

2) We do not base ourselves upon any pre-arranged questionnaire aiming at the discovery of the grammatical features of the language -We try to transcribe longer speeches step by step soon after we are somewhat familiar with the sound features of the language. In fact we already learned a little about the grammatical features and found some of the linguistic elements primarily of grammatical significance by comparing the short utterances. In the present stage those features and elements naturally come to us in abundance. Effort is made only to let the informant speak freely on any topic. The more material we get, the more such features and elements we can extract from them. Sometimes we try to induce the informant to express himself in some particular way in order to get information about some point we have in mind. That is done in a roundabout way. We never ask him to make a literary translation of a sentence constructed in our language for some purpose.

3) We have had very little success in making use of the tape recordings of longer speeches.-The battery-operated recorder we had with us in the field did not work well. Satisfactory transcriptions of longer speeches became unattainable. As a remedy, two informants were afterwards asked to go to a city to make more recordings. It turned out, however, once coming to the modern world and apparently being excited, they preferred to talk about nothing but the new things. That made most of their speech full of Chinese of Japanese expressions which .are quite undesirable. In consequence, the only thing for us to do was to base our study chiefly on our notes taken in dictation in the field. With regard to the notes we must point out that there were unnatural stops inevitably made in the original speech. That was for us to catch up. Although we can check and recheck the material with the informant and succeed in healing up the injuries fairly well, what we finally have is no longer the original. Instead, it is the edited notes. They are in general inadequate for our study of certain aspects of the language, especially its intonation.

4) In many cases, it is not only very difficult for us to learn the meaning of a speech form, but it is no easy job for the informant to explain it either. Generally the informant can help us learn the meaning of an utterance only by giving a very rough translation in Chinese or Japanese. We try to understand what it exactly means by comparing similar instances. As for those elements that seldom occur alone, we try to understand their significance from the utterances including them. A translation of an informant, howsoever it is done, is used only as a guide.

5) The analysis of the language is carried out according to purely descriptive principles. That means: (a) we try to know the system of the language only as it is revealed in what we have in our records: (b) all interpretations are based on formal features﹔(c) no grammar of any other language, has been used as reference; and (d) every single term for any element of this language is defined in terms of its position in the framework of the language. To emphasize the first point, we make it a rule to cite examples only from the published material in Part II and Part III of this volume which in fact includes approximately all material we have. (See the introductory remarks of Part II.) When the example is a “phrase” or a “sentence”, reference in the texts is given. We are compelled to break our rule only on such occasions when the examples we want do not occur in the published texts, but are originally elicited from our informants during our discussing the language with them. They are marked with asterisks. As the reader will see, we have used but very few of them.

6) Attached to Part II is “An Index to the Contents of the texts”, and to Part III “A Translational Index to the Glossary”. For the uses of them, see the introductory remarks of the two parts respectively.

Of the four young persons participating in this study, Mr. Wang Mr. Kuan and Mr. Cheng have helped in field recording, in editing and translating the texts, and in preparing the glossary. Miss Yan has helped in the preparation of the indices and in proof readings.