Published online by Cambridge University Press: 22 April 2026
Trukese is a Malayo-Polynesian language spoken by about 10,000 inhabitants of the islands of the Truk lagoon and by a few thousands more in the Hall Islands to the north and in the Mortlock Islands to the south. This paper is the first attempt (as far as I know) to apply the comparative method to Trukese. A large number of reconstructions based on the comparison of some eleven different Malayo-Polynesian languages are available. This paper attempts to deal with the results obtained by applying these reconstructions to Trukese. This procedure should give us some insight into the history of the Trukese vowels.
1 During the summer and fall of 1947 I spent four months at Truk as a participant in the Coordinated Investigation of Micronesian Anthropology (now called the Scientific Investigation of Micronesia) organized by the Pacific Science Board of the National Research Council with funds supplied by the Office of Naval Research. I spent nearly all of my time on the island of Romonum, where a western Lagoon dialect is spoken. I use the word Trukese (Tr.) to designate words in the Romonum dialect, unless some other qualification is added.
2 Names of languages are abbreviated as follows: Bs., Bisayan; Fi., Fiji; Fu., Futuna; Hv., Hova (Malagasy dialect); Jv., Javanese; Ml., Malay; NgD. Ngaju-Dayak; PMP, Proto-Malayo-Polynesian; pre-Tr., pre-Trukese; Sa., Sa?a; TBt., Toba-Batak; Tg., Tagalog; To., Tonga; Tr., Trukese.
3 Collected and systematically treated by Otto Dempwolff in the three parts of his Vergleichende Lautlehre des austronesischen Wortschatzes, in Zeitschrift für Eingeborenensprachen (Berlin): 1. Induktiver Aufbau einer indonesischen Ursprache, Beiheft 15 (1934); 2. Deduktive Anwendung des Urindonesischen auf austronesische Einzelsprachen, Beiheft 17 (1937); 3. Austronesisches Wörterverzeichnis, Beiheft 19 (1938).
Although most of the reconstructions I cite are Dempwolff’s, I find it convenient to use a set of symbols partly different from his. In the following list my symbol precedes the colon, Dempwolff’s follows it: c : k’, D : d, e : a, intervocalic h : ’, ? : intervocalic and final
. I write initial ? if a Tonga word can be cited with this phoneme; otherwise I write nothing for Dempwolff’s initial as well as his final symbol’. I use initial h as he does. The following symbols are the same in his transcription and in mine: a b d g i k l m n ŋ p t u.
4 Only in the reconstruction *Dewha ‘two’; see my article The Malayo-Polynesian word for ‘two’, Lg. 23.50-5 (1947).
5 I reconstruct *-ey to explain final correspondences such as those of Tg. paláy, Ml. padi as against the reconstruction of *-ay to explain the correspondences of Tg. á:nay> Ml. anay-anay ‘termite’, To. ane ‘moth’. Dempwolff reconstructs *-ay in both cases.
6 Tagalog has the suffixes -an, -in; Bisayan has the respectively corresponding suffixes -an, -on; Javanese has -an, -ĕn. From a comparison of these suffixes we can reconstruct Proto-Indonesian -an, -en, with the probability that the same suffixes existed in Proto- Malayo-Polynesian.
7 The PMP consonants are listed in fn. 3.
8 Although Trukese has no double consonants in absolute initial position, I have adopted the practice of writing with initial double consonant those words which appear with a double consonant after a word ending in a vowel; e.g. ca ‘blood’ represents the actual pronunciation when isolated of what I write as cca; the latter form appears after another word ending in a vowel.
9 Published by the U. S. Naval Government, 1947. I write ö for his ȯ, y for his ü, and g for his ng, among other differences. Elbert does not usually indicate double (or long) vowels. Wherever it is useful, I cite the form he gives in parentheses. When his form is the only available source, I reinterpret it into the shape I believe it would have in the Trukese that I recorded and place the reinterpretation after an asterisk.
10 An explanation of the appearance of a final ? in this and many other Tagalog words related to words in other languages ending in a vowel (particularly Malay and Javanese) will be presented in an article dealing with PMP laryngeals.
11 Cf. Dempwolff 2.135 ff. and 2.179 ff. for similar assumptions to explain Melanesian and Polynesian phenomena.
12 Like Dempwolff I employ brackets to enclose two or more reconstructed phonemes which are not distinguished by the reflexes of available cognates.
13 See fn. 10.
14 For -ey, see fn. 5.
15 See Dempwolff 3, sub kati R.
16 The reconstruction beyi reconciles Dempwolff’s doublet *bayi and *binay, if Jv. bayi ‘suckling’ is excluded from the cognates because of its different meaning. This reconstruction cannot however explain the element of To. fa-fine ‘woman’ as a reduplicating first syllable unless either an otherwise unexemplified development of *e to To. a or an exceptional type of reduplication is posited.
17 See fn. 10.
18 For this reconstruction see the reference in fn. 4.
19 The double vowel in the base is taken to be analogical in origin. We should have expected a single vowel.
20 See §46.
21 Dempwolff made this reconstruction ambiguous with respect to the presence of PMP s or c. The t of Tr. jëty-n excludes the possibility of *-c; see §23.
22 For an indication that a PMP prefix ma- can be posited, see my article The Tagalog reflexes of Malayo-Polynesian D, Lg. 23.233 (§3.10) and fn. 32.
23 The final q of the Malay word (which elsewhere usually reflects *k) has not yet been explained.
24 Elbert’s ü in his ȯsün is for my y; see fn. 9. I take it that he either misheard u as y, or that y occurs in the dialect which he mainly recorded (eastern Lagoon Trukese) as a stem-vowel after back-rounded vowels. This type of paradigm does not occur on Romonum.
25 See §45 and fn. 16.