ADJUSTMENTS TO THE METHODOLOGY FOR

242

CHAPTER 8 ADJUSTMENTS TO THE METHODOLOGY FOR

MORONENE In this chapter I will discuss how the methodology presented in Chapter 2 needs to be adapted to cope with the specifics of Moronene syntax as presented in Chapter 7. I will also describe the texts which make up the Moronene corpus. Defining Moronene Clause Types As far as possible, the same features which I used to distinguish Hebrew clause types will be used for Moronene clause types. Although some adjustments need to be made because of differences between Hebrew and Moronene clause syntax, in a number of ways the syntax of the two languages are similar. In Moronene, just as in Hebrew, clauses may or may not have explicit nominal subjects and objects. There is also considerable variation in constituent order: attested orders in Moronene include SV, VS, VO, OV, SVO, VSO, and VOS. Just as in Hebrew, the object may occur as a suffix on the verb. In Moronene a suffixed object and nominal object often both occur in the same clause referring to the same entity, whereas this is rare in Hebrew. In Hebrew, clause types are distinguished according to the verb conjugation used, as well as according to whether the verb is morphologically passive. Moronene has undergoer voice verb forms, marked by a ni - prefix. For the purposes of this research, these are grouped together with resultative verb forms, marked with a te - prefix. Like the forms with a ni - prefix, the resultative forms have the undergoer as the subject, and the 243 actor is usually implicit. Hence they are syntactically passive. Ideally, the topicality patterns of ni - forms and te - forms would be investigated separately, but in the present corpus, the sample sizes would be too small. Therefore they are grouped together under the label undergoer voice verb UVV. Unlike Hebrew, which expresses tense and aspect distinctions through the choice of verb conjugation, in Moronene, such distinctions are primarily expressed through optional auxiliaries. However, as described in Chapter 7, there is an important distinction in Moronene transitive verbs between forms with the action focus prefix moN - or poN -, and forms without the prefix. This distinction will be reflected in the variables determining the clause types for analysis. A significant difference between Moronene and Hebrew syntax is how the subject is marked by affixes or clitics. Whereas in Hebrew, the form of subject affixes is completely determined by the choice of verb conjugation, in Moronene there are three options for the grammatical case of the subject affix or clitic, which is either affixed to the verb or cliticized to a conjunction or auxiliary. These are either nominative NOM, which can index the subject in intransitive clauses or the actor in transitive clauses, absolutive ABS which can index the subject in intransitive clauses or the object in transitive clauses, or possessive POS which can index the actor in transitive clauses with undergoer voice. The variables which determine the clause types in Moronene can be understood as the answers to the following eight questions:  Is there an explicit nominal subject or object? S, O  For undergoer voice clauses, is there an explicit nominal actor or undergoer? A, U  What affix or clitic set is used to mark the subjectactor? NOM, ABS, POS 244  Does the verb take an action focus prefix? monV 1  Does the verb take a ni - prefix or te - prefix undergoer voice? UVV  Does the verb have an object suffix? Opro  Is there an auxiliary carrying a suffix? Aux 2  Is there a preposition governing an oblique undergoer or actor? prep Examples of how these variables are used to classify particular clauses were given in Chapter 7. Other Adjustments to Methodology The criteria for selection of Moronene clauses to be analyzed for the most part parallel the criteria used for Hebrew. I restrict the analysis to semantically transitive clauses. As in Hebrew, I exclude relative clauses and clauses which take a clause as object. But unlike Hebrew, I do not exclude participial clauses. According to Mead, the nonfinite form of the prefix, moN-, is a participle 1998:173. Whereas in Hebrew one factor leading to the exclusion of participial clauses from the analysis was their relative infrequency, in Moronene moN-V clauses are quite frequent. In Moronene, only narrative is analyzed. As in Hebrew, I exclude embedded quotations. One result of this is that there are no clauses with first or second person subjects or objects included in the analysis. Although it would be attractive to include an analysis of Moronene directive discourse to parallel the Hebrew investigation, there are three reasons I chose not to do so. First, the practical need to limit the time spent on research and analysis, and to limit the length of the dissertation. Second, the relative paucity of directive texts in my 1 I ignore the distinction between the finite form of the transitive prefix poN- and the nonfinite form moN-. 2 Only auxiliaries with an absolutive suffix are included in the clause type abbreviation. Others are ignored, so as not to unnecessarily increase the number of clause types. 245 collection of Moronene texts. In order to have an adequate sample size, it would have been necessary to record and translate many additional texts. Third, in Moronene, most of the major verb forms occur frequently in narrative, making a parallel investigation of directive discourse less urgent. This is unlike Hebrew, in which a parallel investigation of directive discourse is necessary in order to understand the functioning of important verb forms like weqatal and imperative, which are less frequent in narrative. With regard to defining clause units for counting purposes, I count serial verbs as separate clauses. Parallel to Hebrew, I include relative clauses in the same clause unit as the main clause they modify. Similarly I do not count certain preposed temporal clauses as separate clause units. Whereas in Hebrew the distinguishing characteristic of such temporal clauses is the verb  “and it was” or  “and it will be,” in Moronene the distinguishing characteristic is that no participants are mentioned. On the other hand, temporal clauses which mention participants are counted as separate clause units. Because the study of Moronene is limited in scope, I restrict the statistical analysis of topicality patterns to clause types with nine or more occurrences, 3 similar to what was done for Hebrew in Chapter 3. I do not define any amalgamated clause types based on particular syntactic features. Hence the type of analysis found in Chapters 4 to 6 is not done for Moronene. Texts I have chosen several texts, or large text extracts, and have investigated most of the clauses within the narrative portions of these texts. I have analyzed two narrative poetry texts, and four narrative prose texts. 3 Because of the relatively small size of the Moronene corpus compared to Hebrew, and the smaller numbers of each clause type, I have reduced the minimum sample size from ten to nine. 246 The first narrative poetry text is Kada , an epic poetry text recounting conflict between Moronene and Tolaki kings in bygone days. “Kada” is actually the name of the genre, rather than a specific text. The text is composed in traditional Moronene meter, which means that each line has seven syllables. There is a great deal of poetic synonymous parallelism, so that adjacent lines often express the same meaning using synonymous vocabulary. The text I am analyzing was recited by Nandi in the village of Rau-Rau, Rarowatu subdistrict, in the 1980s. It was recorded, transcribed and translated into Indonesian by Sahido Tambera. This work was part of a project sponsored by Koninklijk Instituut voor Taal-Lande en Volkenkude KITLV. The text is found in the KITLV library in Leiden, designated by the code KITLV Or 563. I am grateful to KITLV for giving me permission to use this text. I am also grateful to Sahido Tambera and Nandi‟s son Etu for helping me correct the transcription and translation. This text is a very long text; it was recorded over five successive nights, and the typed transcription and translation is over seven hundred pages long. Therefore I have chosen an extract of the text to use for analysis, consisting of 986 clauses. Within this extract, some passages particularly fighting scenes, repeat earlier passages almost word for word. I have eliminated these from the analysis. The remaining extract consists of 852 clauses. The second narrative poetry text is Kunini , a Christmas carol composed in the traditional heptasyllabic meter with much poetic parallelism. It was composed by Bawea Ferdinand Powatu, the first Moronene pastor, in the 1930s. It consists of 148 heptasyllabic bicolons. It recounts most of the Christmas story, including the annunciation, the journey to Bethlehem, and the visit of the shepherds and the Magi. It takes about twenty minutes to sing. The text was transcribed and translated into Indonesian by Ndasi. 247 The first narrative prose text is entitled Petampu’uno Ica Diu The Beginning of the Dugong. It is a traditional folk tale, 478 clauses long, about a woman who changed into a dugong. Tinano Maegani Miano Cantete Maegani‟s Mother; the Person in the Latrine, 943 clauses long, is about the rescue of a hostage imprisoned in a latrine. Wuu Sio Ropa Nine Fathom Hair, 285 clauses long, is about an orphan boy who married a princess with very long hair. These three stories were recited by Wede in the village of Taubonto, Rarowatu subdistrict. Colisi Pinkie is 229 clauses long, and is about a boy the size of a pinkie who was rescued after being swallowed by a horse and a tiger. It was recited by Ndasi in the village of Taubonto. I divide the corpus into a primary corpus and a secondary corpus. The primary corpus consists of the two narrative poetry texts, Petampu’uno Ica Diu, Colisi , and the first 512 clauses of Tinano Maegani Miano Cantete . The secondary corpus consists of the last 431 clauses of Tinano Maegani Miano Cantete as well as Wuu Sio Ropa . The reason for the division into two corpa is as follows. In order to be included in the analysis, it is necessary for a clause type to occur at least nine times in the corpus. When all of the 596 semantically transitive clauses in the primary corpus were analyzed, five common clause types occurred more than forty times, whereas all the others had sample sizes of seventeen or fewer. There were a number of clause types which had almost nine occurrences. Therefore I searched the secondary corpus for only the less frequent clause types, to increase the sample sizes and to find the additional examples needed to reach the minimum number of nine. The other more common clause types were skipped in the secondary corpus. I added an additional forty-four semantically transitive clauses from the secondary corpus, bringing the total to 640. Within these clauses, there are fifteen clause types with a sample size of nine or over. 248

CHAPTER 9 CLUSTERING OF MORONENE CLAUSE TYPES