The creation of a dictionary often strikes people as an extraordinary undertaking, although it is more of a commonplace at The University Museum than elsewhere because of the Sumerian and Aramaic dictionaries now being compiled there.
Putting together the Trukese-English dictionary called for a somewhat different approach, for the initial recording of the language had not been made by the people who spoke it. Trukese, which belongs to the Austronesian language family (see box), is spoken on the islands of Truk, an atoll in Micronesia in the western Pacific (Fig. 1).
Compilation of the many words, their meanings, and other necessary information was a complicated process greatly aided by computerization. Having all the data computerized also means that further work on the dictionary can continue easily and indefinitely. The present product—the dictionary itself—is a two-volume work. The first volume, published in 1980, contains 55 pages of introductory material describing alphabetic representation of sounds and the grammar required to use the dictionary, followed by 399 pages of Trukese words with English glosses and related information. All told, there are about 12,000 entries.
The second volume is now in press and will appear early in 1990. The main section is a 450-page listing of all the English glosses in the first volume, with the Trukese entries under which they appear there. In addition, there is a concordance of Trukese “roots,” the basic units with meaning—words or affixes—from which the Trukese words in the first volume are constructed. After each root a simple gloss is given and then a listing (without glosses) of all the entries in the first volume in which the root appears. Many roots, listed separately, represent borrowings from English, Japanese, and other languages. Much of the volume, especially the concordance of roots, is of interest mainly to specialists, but the process of evolution and preparation is of wider concern.
Genesis
The dictionary project had a practical basis. My first visit to Truk took place in 1947, when I was part of a group conducting ethnographic research. The team included a linguist, Professor Isidore Dyen of Yale University, who was preparing a grammar of Trukese. At the time, no one in Truk spoke English well enough to interpret for us, and none of us knew Japanese, which was a second language for many people there. Therefore, we had to learn to speak Trukese in order to do our work. A dictionary prepared the previous year by Professor Samuel Elbert of the University of Hawaii, then an officer in the U.S. Navvy, was an immense aid, but it had been put together hastily and did not meet the standards of phonological accuracy required by the comparative method on which historical linguistics depends.
After leaving Truk, I became interested in examining the relationship of Micronesian languages (Fig. 2) to others in the vast Austronesian language family. At that time the languages of Truk and the Gilbert Islands (now Kiribati) were the only Micronesian languages on which there was much information. It therefore seemed reasonable to start with them and to upgrade Elbert’s dictionary for comparative scholarly purposes. This I undertook as a sideline when I returned to Truk for a year of ethnographic research in 1964-65 (Fig. 3).
Gathering the Data
My first step was to go over all the entries in Elbert’s dictionary with a mature native speaker, Eiwe Rewi (Fig. 4), with whom I had worked in 1947. After his death halfway through the year, I continued with the late Boutau K. Efot (Fig. 5), another native speaker from the same community. It took us about four hours a day, five days a week, over nine months to work our way through it. We also went through all the technical terms in a monograph describing traditional Trukese arts and crafts.
An ethnographer is exposed to specialized and topical vocabularies that deal with different aspects of a people’s culture. Collecting stories and texts is an invaluable way to obtain lexical and grammatical information, but their subject matter is usually limited. They are not likely to be a good source of information on cooking techniques, for example (Fig. 6), or canoe building (Fig. 7). In addition, stories contain casual references to many things familiar to the audience, which are thus not made explicit. For a dictionary, oblique statements need to be reviewed systematically in order to determine the various things that they mean to the listeners.
In working with Elbert’s dictionary, a similar effort had to be made in exploring synonyms, antonyms, and other related terms. Drawing on both sources—oral tradition and Elbert’s dictionary happened was that as I discovered the need to set up new entries for such roots as I went along, I had no easy way to go back and find words already listed in which these roots appeared. Retrieving roots and adding other necessary items were becoming increasingly more difficult as the data accumulated. I had thus worked my way through several letters of the alphabet, learning as I went about what I should be trying to do, when I heard that the University of Hawaii had obtained a contract from the U.. Trust Territory of the Pacific Islands to prepare a series of dictionaries and grammars of the languages spoken in Micronesia, including the language of Truk. I offered to join the Hawaii project, and was accepted. Thus began my Trukese-English volume, and much of my year as a visiting professor there in 1982-83 was devoted to doing the same for the supplementary volume.
Alphabetic Representation
Toward the end of the last century, missionaries had developed an alphabet for Trukese and thee dialects of neighboring atolls that had become well established in schools (Fig. 8) and in ordinary usage before the middle of this century. It was, however, inadequate: it used five vowel symbols to represent nine different vowels. it represented semivowels as if they were vowels, it failed to distinguish adequately between two different sets of labial consonants, and it took no account of the crucial distinction between long (double) and short vowels and between consonants that were held for a syllabic beat (now written as double consonants) and those that were not. Thus what the dictionary shows as mwaan (`man, male’), system of information marking that served us well. With the help of a grant from the National Science Foundation, I was able to employ graduate student assistants to code the material in accordance with the system thus created. I reviewed everything they did for accuracy and for further editing. The edited information was then typed onto scrolls, later cut up into convenient working lengths. The scroll paper made an automatic carbon for my own file, and the original was mailed to Hawaii, where a student entered the material into the mainframe computer. (Personal computers were not available yet.) Each summer in Hawaii, I proofread computer printouts of what had been entered. (I must have proofread the whole dictionary five or six times before we were done.) We were also able to use the computer effectively to highlight typographical errors in format. Nevertheless. some errors appear in the publication, but few serious ones.
Errors, but it is an important advance over what was available for both practical and scholarly purposes before. We hope that soon it will be superseded by an expanded and upgraded edition; but that will probably have to be someone else’s task. Whoever undertakes it will find the present dictionary easy to build on, thanks to its availability on diskette for use on a personal computer.