J. Collins A note on cultural vocabulary in the Moluccan Islands

In: Bijdragen tot de Taal-, Land- en Volkenkunde 139 (1983), no: 2/3, Leiden, 357-362

This PDF-file was downloaded from http://www.kitlv-journals.nl

Downloaded from Brill.com09/29/2021 06:16:58PM via free access KORTE MEDEDELINGEN

JAMES T. COLLINS, UNIVERSITI KEBANGSAAN MALAYSIA l A NOTE ON CULTURAL VOCABULARY IN THE MOLUCCAN ISLANDS In recent years M. A. Chlenov has played a unique and important role in Austronesian linguistics. In his writings (1973,1976) he has made available some of the results and data of his extensive field work in East . His stimulating article (1980) represents still an- other contribution reflecting a familiarity with a number of diverse languages as wel1 as a deep humanistic concern with extracting relevant conclusions from such a diversity. It is certainly not the intention of this brief note to discourage these efforts or to disparage such a wide knowledge. On the contrary, Dr. Chlenov's work remains of permanent importance in Moluccan linguistic studies. Nonetheless, it is appro- priate to comment on the procedure and some of the examples in his most recent article. Consideration of this procedure may benefit our attempts at reconstructing proto-languages and discovering the outlines of ancient cultures. There is an implicit methodology employed by Chlenov in sepa- rating Austronesian from non Austronesian vocabulary. Confronted by a large amount of data, apparently his first step was to group seemingly related words together. There followed a search for possible cognates in other languages and in reconstructed proto-languages. Thus, al1 the entries which are provided with a Proto-Austronesian "pedigree" are considered inherited reflexes when there is no evident borrowing. Other words are either reflexes of a more recent Austronesian proto- language (Proto-Oceanic, etc.) or "pre-Austronesian" - again when they are not apparent loanwords. Based on this categorization, cal- culations comparing the proportion of Austronesian and non- Austronesian vocabulary items are possible. Chlenov's lexicostatistical orientation is obvious in this approach, if we assume that lexicostatistics involves "procedures of applying statistics to the genetic comparison of the lexicon . . ." (Dyen 1975: 137). Preliminary, prima facie allocations of a large number of data precede a comparison with other languages, in this case including the proto-languages. Cognates are sorted from non-cognates and cal- culations begin. Certainly such a procedure is of value when ap- proaching a great deal of still disorganized information. Working hypotheses can be arrived at so that the process of theorizing and testing can continue. It is important, however, to apply such a poten- tially powerful procedure with great rigor. Each step must be carefully controlled. Setting aside the problems which are inherent in initia1 allocations and groupings of data, such as the possibility of intimate borrowing, we are confronted with the serious problem of searching out Proto-

Downloaded from Brill.com09/29/2021 06:16:58PM via free access 358 Korte Mededelingen

Austronesian reflexes. Apparently Chlenov (p. 437, Note 2) chiefly relied on the citations in Wurm and Wilson (1975). Several years have passed since the publication of that book and these years have seen a considerable increase in our knowledge about Protc-Austronesian and its descendant languages, particularly those in the Pacific. But even if our knowledge of Austronesian vocabulary were limited to what is contained in the pages of Wurm and Wilson, we note some cognates overlooked by Chlenov. The relationships of these cognates have been obscured by the occurrence of well-known linguistic changes. First, we note an apparent case of semantic shift, that is, narrowing a genera1 concept to a more specific one. Under "SagoJJ (p. 430) we read that "the eastern dialects of Ambon have the root *pa(l)u which in other languages means 'something made of sago"'. More precisely, in othe~.languages this word is quite specific: "roasted sago bread" l. (See Wallace for a description of the process and the product.) Note the following entries: Village Sago tree Sago bread Hila (Ambon) lapia pa?a puti Hitu (Ambon) lapia paputi Tulehu (Ambon) lapia paputin Iha-Kulur (Seram) lepia pa% Kulur (Saparua) lipia pa3ulo Qta (Seram) ripia pa?uro Hunitetu (Seram) liki P""? Kasie (Seram) likia PaPa u Eti (Seram) lapia pa?u Wanasa (Seram) ipia paua Masiwang (Seram) bai bau Buru bia baku In Eli, a surviving descendant of the now spoken on Kei Besar, we find baun "mold for baking sago bread". Throughout the Centra1 Moluccas, from Buru to the mouth of the Masiwang river in East Seram, we note reflexes of a word which can be reconstructed as **mbaku "roasted sago bread". This reconstruction should be compared to Dempwolff's 1929 entry for Proto-Austronesian *baku "geronnen" which became PPN *faku "to roast" and *paku "crust" (sec also Wurm and Wilson 1975:173). The addition of a stative prefix *ma- to *faku would yield *ma+faku "that which is roasted". By deletion of the first vowel and subsequent sound changes **mbaku appears as a clear reflex of *ma+ (b,p)aku in an older proto-language. By narrowing the meaning of "something baked" to apply only to that chief baked good of the Moluccas and its basic food stuff, **mbaku came to mean "roasted sago bread". In some languages this term was also applied to the tree which supplied the meal for this bread. In this connection we should probably consider Chlenov's citation of Makian Dalam bakoé. My fieldnotes indicate that baku means "roasted sago bread" as wel1 as "sago treeJJ in Makian Dalam (and Makian Luar). Sohyo in nearby Taliabo also displays baku "roasted sago' bread".

Downloaded from Brill.com09/29/2021 06:16:58PM via free access Korte' Mededelingen 359

A second well-known sound change apparently is involved in reflexes of *niuR "coconut". Chlenov asserts tha *mu is "obviously non-Austronesian". He mentions that reflexes of this root (*mu) are found in Batuasa, Liambata and Kilmuri in Seram and in the . Unfortunately he does not cite the words in these languages. We are forced to refer to the entries in dialects of each of these languages. Werinama, differing only slightly from Batuasa, has nua. Adabai, closely related to Liambata, also displays nua. Geser, mutually intelligible with Kilmuri, has niu. Saparua, as it is spoken in Latu (Seram), displays nuwollo. In none of these dialects do we observe a reflex of *mu. Rather we note clear reflexes of PAN *niuR. It is not inconceivable that in other dialects, including those mentioned by Chlenov, nu- or niu became mu by assimilation of the initia1 nasal to the following vowel; that is, the alveolar nasal, n, (or perhaps at that stage the palata1 nasal, ñ (<%i)) became a labial nasal when followed by a "labial", that is, rounded, vowel, u. It is difficult to say with certainty what occurred, because we do not know what the actual entries are. Nonetheless, unusual cases of assimilation in this lexical item are not unheard of. In Kailolo (Haruku), near Saparua, we find nimel. In this case a rounded semi- vowel became a labial nasal following a nasal (**niwel > nimel). Again, without the data no conclusion about the languages said to reflect 'mu can be reached. However, it seems rash to say that forms like mu are "obviously non-Austronesian". Assimilation is a common enough phenomenon. Assimilation is probably involved in a third example as well. Chlenov lists among "unique roots" for fowl (p. 432) "Manusela mailahu". Data from Manusela are not available, but a comparison with three mutually intelligible dialects is revealing. In al1 three dialects manua(m) means "bird". Under the entry for fowl we find Wanasa (near Teluti Bay, south coast) manlohua Sadar (adjacent to Kobi, north coast) manlohua Huaulu (central mountain range) malohuam In Sadar and, presumably, the other dialects of Centra1 Seram, lohu means "space below the house". Thus, we have an analytic con- struction: "bird-below-the-house" = "fowl". Undoubtedly mailahu (as Chlenov records it 2, is a local variant of this compound word. Perhaps the assimilation of n to the following l has resulted in the cluster y1 (manlahu > mallahu > maylahu). At any rate, in view of the entries from three other dialects of the Same language we can only assume that mailahu is a reflex of *manuk and not a "unique" root. Applying statistics to lexical comparison requires that the starting point in such comparisons be firm. It should be clear that the iden- tification of some words in a given corpus as Austronesian reflexes does not mean that al1 the other words are non-Austronesian (or "pre- Austronesian"). Linguistic reconstruction is too complex to admit such a decisive dichotomy. Commonly recognized linguistic phenomena such as semantic shift, consonantal assimilation and analytic constructions

Downloaded from Brill.com09/29/2021 06:16:58PM via free access 360 Korte Mededelingen

with subsequent phonetic synthesis have obscured some Austronesian cognates. Comparative linguistics has proceeded beyond the stage of simply shuffling certain straightforward lexica1 items. The pitfalls inherent in incomplete analysis of the existing data are apparent in the conclusion to Chlenov's article. Reiterating his con- tention (p. 433) that one of the words for clove "goes back to the PA *bulav-an, 'gold"', Chlenov maintains that the terms for clove and nutmeg "are borrowed from the consumers of the product and coincide with the words for gold or wealth". It would be an interesting semantic shift if it were true that an agricultural product came to be called by the term "gold" after that product had become the chief cash crop. However, the data do not support such a contention. Chlenov says that the words for "clove" are derived from the pre- nasalized form of 'bulav-an, that is "*mbulav", whereas the words for "gold" derive from the form "*bulavJ'. That such derivational doublets could occur in the Centra1 Maluku languages is not unusual. Nonetheless, Chlenov cites no examples. A perusal of some of the available data suggests that from a strictly phonetic perspective Chlenov's theory is questionable. Village 'rclove" 'rgold" Laha (Ambon) pulawag bas) Kaitetu (Ambon) pulawan halawan Asilulu (Ambon) pukalawan halawan Luhu (Seram) pukalawane halawane Kailolo (Haruku) po?olawan halawane Buru buglawan flawan In addition to the different initia1 consonants, accounted for by Chlenov by referring to prenasalization, there are other discrepancies. First, the vowels are different. In al1 the languages cited here the vowel of the first syllable in "gold" is a. This contrasts with the first vowel of "clove", which is u. If these words are derived from the Same proto-word, why are the vowels different? Second, in al1 but a small, closely related group of East Ambon languages (including Kaitetu and Laha) the reflexes suggest a jour-syllable word for clove, perhaps **pukalawan. Consequently a side-by-side comparison suggests ihat "gold" and "clove" have nothing in common except a superficial similarity in the last two syllables. An identification of **pukalawan with * * balawan (<*bulav-an) seems unlikely. There are, then, three phonetic inconsistencies apparent in the com- parison of **pukalawan and the Proto-Austronesian word for 'rgold". There is, in fact, a more acceptable source for **pukalawan. Burkill (1966:977) notes that throughout the archipelago "from India east- wards" clove was known by forms of the word lavanga. The ultimate origin of this word is disputed. Burkill assures us it is Sanskritic but Gonda (1973:24 and elsewhere) suggests that it may have been bor- rowed from an int0 both Sanskrit and Tamil. In any case it is likely that **pukalawan "clove" should be assc- ciated with lawayl not *bulav-an. What then of the first two syllables,

Downloaded from Brill.com09/29/2021 06:16:58PM via free access Korte Mededelingen

puka-? Tentatively this is considered a loanword from Malay buga "bloom, flower". We know that in Centra1 Maluku languages some loanwords display the change b > p. For example in Asilulu (Ambon) we observe that Malay buka "open, release" became puka "the first day of the tenth Islamic month" (via buka puasa "break the fast") and Makassar bela "friend" (from Sanskrit?) became pela "allied village". Similarly in some loan words yl became k. In Asilulu and numerous other languages (Stresemann 1927:74) we find Malay siyla "lion" (from Sanskrit) became sika "cat", although this etymology may be doubtful (J. Anceaux, p.c.). The phonetic and semantic evidence suggests that the Malay com- pound noun buyla lawayl, still in use in peninsular Malaysia, is the source for **pukalawan "clove". So it remains true, as Chlenov ob- serves, that the "designations are taken from foreign languages", but the loanword source has nothing to do with "gold" and consumers. Again the critica1 importance of comparing a word as carefully as possible with entries in other closely related languages and dialects is apparent. Weak etymologization yields a weak abasis for comparative computations. Still Chlenov's attempt is an importan't one. It is hoped that in future studies he might include more cultural vocabulary such as the words for tubers, millet, Inocarpus sp. and other traditional food sources as wel1 as the words for thread, tin, weapons and tools. Tsuchida's study (1977) on botanica1 vocabulary in Formosan lan- guages provides a good model for such wider studies. Similarly Blust's (1976) careful analysis of data from numerous to reconstruct some of the cultural vocabulary of Proto-Austronesian suggests a methodology which can be used for the reconstruction of lower level proto-languages, As long as the linguistic underpinnings are secure, reliable information about the ancient societies of the Moluccas wil1 be forthcoming.

NOTES 1 In fact, in most of the languages mentioned by Chlenov for which I too have collected data the entries which he cites mean "roasted sago bread" not "sago". Village Sago tree Sago bread Nila pria 13ieta Damar Ipi prisa Geser (Kilmuri) kel suat Kei er mana babau Sanana sáa bak? Yamdena orum bui ke:de Babar (Wetang) rotna rotna sker Babar (Yatoki) ro?na ro?na Serua skera skera It should be noted that on Serua, where one term is used for both meanings, there is only one sago tree!'This was recently introduced by villagers who had been to Seram. It is interesting too that in Kei maga appears in a compound word with a reflex of **mbaku. If- maga is related .to Proto-Polynesian *maqaga "mouthful" or Proto-Austronesian *ma-kaen 'feat", we can see that

Downloaded from Brill.com09/29/2021 06:16:58PM via free access Korte Mededelingen

here too we are dealing with an analytica1 construction: maga babau = "roasted food". In any case, it is not surprising that a specific culinary prepa- ration should have numerous local names. 2 But in Chlenov and Sirk (1973:77) it is not clear that either of the Chlenovs personally transcribed the 500-word vocabulary of Manusela. Was this "trans- lated from Malay by a Manusela native"? If so, there may be some problems with interpreting the orthography of the "translation". The occurrence of lahu instead of lohu is odd. Note that Tauern (1931:109) lists "lahua Asche:' but "lohua Platz unter dem Hause" (p.111). Chlenov and Sirk considered Tauem's Wahinama material the Same as Manusela. Similarly Boot's Nisawele dialect is considered "almost identical to Manusela"; in that material (Boot 1893:1186) we find "rnänoewif mäleehoewa" for "kip". It is surprising then that Chlenov did not recognize mailahu as a compound word.

BIBLIOGRAPHY Blust, R. k 1976 'Austronesian culture history: some linguistic inferences and their re- lations to the archaeological record', World Archaeology 8.1:19-43 (reprinted with minor additions in NUSA, part III:25-37, 1977). Boot, J. 1593 'Korte schets der Noord-Kust van Ceram', Tijdschr. v. h. Ned. Aard. Genootschap 10(2):1183-90. Burkill, I. H. 1966 A dictionary of the economic products of the Malay peninsula, Kuala Lumpur: Ministry of Agriculture and Cooperatives. Chlenov, M. A. 1969 Ocherki po etnicheskoi istorii narodov tsentral'nykh molukk (Indoneziia), Moscow: Dissertation. 1976 Naselenie Molukkskikh Ostrovov, Moscow: Nauka. 1980 'Cultural vocabulary as an indicator of inter-ethnic relations: Eastern Indonesian evidence', BK1 136:426-39. Chlenov, M. A. and U. Sirk 1973 'Merger of labial phonemes in Ambonese languages', Acta et commen- rationes Universitaris Tartuensis, Oriental Studies 111. Collins, J. T. n.d. Unpublished fieldnotes of languages in the Moluccas, Ms.

Dempwolff, O. F 1929 'Das austronesische Sprachgut in den polynesischen Sprachen', Feest- bundel van het Bataviaasch Genootschap van Kunsten en Wetenschap- pen. Dyen, I. 1975 Linguistic subgrouping and lexicostatistics, The Hague: Mouton. Gonda, J. 1973 Sanskrit in Indonesia, New Delhi: International Academy of Indian culture. Stresemann, E. 1927 Die Lauterscheinungen in den Ambonischen Sprachen, 10 Beiheft der Zeitschrift fur Eingeborenen-Sprachen, Berlin. Tauem, O. D. 1931 'Beitrag zur Kenntnis der Sprachen und Dialecten von Seran', Part 4, Anthropos 26:109-39.

Downloaded from Brill.com09/29/2021 06:16:58PM via free access Korte Mededelingen 363

Tsuchida, S. 1977 'Some plant names in ', Cornputational analyses of Asian and African languages 7:79-119. Wallace, A. R. 1869 The Malay archipelago, London: Macmillan. Wurm, S. A., and B. Wilson 1975 English finderlist of reconstructions in Austronesian languages (Post- Brandsieiter), Pacific Linguistics C/33.

LOKESH CHANDRA REMARKS ON KUR JARAKARNA The edition and translation of the Kuñjarakarna Dharmakathana by A. Teeuw and S. O. Robson constitutes a new landmark in the study of Indonesian Buddhism. The interpretation of some of its passages becomes clearer if we bear in mind that the centra1 deity of the Buddhist system put forward in this work is Vairocana. Vairocana belongs to vajrayäna or Tantric Buddhism. There are two iconographic forms of Vairocana and these correspond to two different divisions of tantras. The first is the Vairocana of the caryä-tantra division, with his hands in the dhyäna mudrä. The second Vairocana pertains to the yoga-tantra division, and his hands exhibit the bodhyagri mudrä. He is the Lord of the universe Vajradhätu and his mandala is called the Vajradhätu mandala. From the Sutasoma and Sang Hyang Kama- häyänikan we know that the hands of Vairocana in Indonesia are in the bodhyagri mudrä. His consort (prajñä) is Vajradhätviivari in the Sang Hyang Kamahäyänikan. Al1 of them point to the prevalence of the Vajradhätu Vairocana of the yoga-tantras in Indonesia. Teeuw/Robson (1981:23) say: "There is, however, one further figure, a rather mysterious one, who turns up in this Canto: in 2.1 even a whole stanza is devoted to someone who is twice called bharäli (la and 2a) and who obviously in terms of protocol has precedence over al1 the other gods . . . The word bharäli has also been found in inscriptions on statues of gods (Brandes 190485, 94). Brandes has given a lengthy survey of the Buddhist pantheon, especially the iaktis. He is rightly of the opinion that these variants (bharäla and bharäli) of bhaflra and bhafäri cannot pbe autochthonous in Java." The editors rightly maintain that bharäli is bhafäri. They concur with Brandes in maintaining that the variant bharäli cannot be autochthonous in Java. Bharäli is a vajrayäna term which is of Indic origin. The form bharäli is actually found in Nepalese sketch-books of vajrayäna iconography which are being edited by the writer of this note. The Kuñjaraka~amust be compared with the totality of the Buddhist tradition for precise comprehension. It may be worth while to illustrate this by means of two stanzas. In stanza 1.7 it is evident that

Downloaded from Brill.com09/29/2021 06:16:58PM via free access