Discourse chart Dooley and Levinsohn sequential default chart

57 literacy program. All three texts are used with the permission of ICC. Electronic copies of the texts were obtained in Legacy format from the ICC READ project office in Saen Monourom, Mondulkiri province. The resulting Legacy formatted files were converted to IPA Unicode format using TECkit v2.1 14 with mapping files developed by our colleague, Dr. Diethelm Kanjahn. Each of these texts was then imported into the computer program, The Field Linguist’s Toolbox v1.5.1 15 , and interlinearized. After adding a free translation for each clause, the texts were charted in Excel following Longacre and Levinsohn’s 1978 approach to charting discourse texts.

4.3 Charting

4.3.1 Discourse chart

The Excel charts enabled the text to be analyzed clause by clause, with the linear order of each constituent maintained. The texts were charted to identify and distinguish initial and final sentence elements. The chart is divided into four main sections: introducers and left-dislocated elements, preposed elements, independent clause, and right-dislocated elements Dooley Levinsohn 2001:68. The organization of the chart allows preposed elements and subordinate clauses to be easily distinguished from the main clause. The chart also requires that decisions be made regarding which syntactic elements are zero, a common occurrence in Bunong. Many times a decision had to be made whether or not a verb had a zero argument. 14 The software is available at http:www.sil.orgcomputingcatalogindex.asp. 15 The software is available at http:www.sil.orgcomputingcatalogindex.asp. 58

4.3.2 Dooley and Levinsohn sequential default chart

After charting each of the texts, four additional columns were added to the right for completing the Dooley and Levinsohn tracking method 2001:127. These additional columns have the following headings: 1 Subject environment and referring expression, 2 Coding, 3 Non-subject environment and referring expression, and 4 Coding. In the Coding column, after the environment for each subject and non-subject was identified, the coding for each referring expression was marked as either default D, under-coded U, or over-coded O. Marked word order was also considered in making this determination. For example, a motion presentational construction, where the subject follows a verb of motion, is a marked construction. While it is not an example of over- coding per se, I placed it in the over-coded category to signal that the referring expression occurs in a marked, topic-promoting construction cf. §2.2.4. Although Dooley and Levinsohn recommend indexing each participant with a number, I used full noun phrases to label the participants to keep better track of them. Bunong relies heavily on zero anaphora for both subjects and non-subjects that have already been activated, so it was less confusing to state the participant explicitly. Often a zero would refer to one or more participants, depending on the context. The environment of each reference was then identified according to the following descriptions provided by Dooley and Levinsohn 2001:130-31. 59 S1 same subject as preceding independent clause S2 subject is addressee of a direct or indirect speech reported in the preceding clause 16 S3 subject is a non-subject in the preceding independent clause S4 other changes of subject than those covered by S2 and S3 Figure 6: Sequential default subject environments N1 referent is in the same non-subject role as preceding independent clause N2 referent is addressee of reported direct or indirect speech and subject in the preceding independent clause N3 referent is involved in preceding clause in a different role other than N1 and N2 N4 non-subject referent does not occur in preceding clause Figure 7: Sequential default non-subject environments The first mention of a participant was given the label Intro; thereafter it was labeled according to one of the environments described above. Several times the data did not strictly match the subject and non-subject environments described in Figures 6 and 7. I extended the S3 environment to include a plural subject that encompasses both the subject and the object participants of the previous clause. Another instance where the four different environments for the subject did not fit the data well was when a subject changed from a plural to a singular subject. In this situation, two or more participants are the subject in the previous clause and then in the next clause, just one of the participants is referred to. When the subject changes from a plural subject to a singular subject, I assigned an S4 label because the subject was not the exact same as the subject in the previous clause, neither was it an addressee or the non-subject in the previous clause. 16 For both the S2 and N2 categories, “direct or indirect speech” was added to Dooley and Levinsohn’s 2001:130 original definitions of these categories. 60 I only kept track of subjects and non-subjects that occurred in independent clauses. After charting both of the third person narratives, very few references to participants occur outside of an independent clause. Only participants were tracked and not props, since props did not play a significant role in any of the texts. In the Fish story, while the fish and the ‘fish shirt’ could be considered props, they were tracked with the other references to Bachelor Smelt-fish since the ‘fish shirt’ is really a disguise for him. The material within quoted speech was not labeled, since quoted speech presents a different set of variables. However, the quote formula preceding the quoted material, when one existed, was included in the analysis. Sometimes the quote formula consisted only of a verb. In this situation, the speaker and the addressee were both given a zero reference. If no speaker, speech verb, or addressee was given, then this was not included in the counts.

4.3.3 VIP coding chart