| This study analyzes variation in lexical repetition in placement-type essays written by 134 nonnative writers and 57 native writers of English. Essays were written on either a descriptive or a persuasive topic and rated holistically for proficiency. The nonnative writers were enrolling in an intensive English program and represented four linguistic backgrounds: Arabic, Japanese, Korean, and Spanish. The native writers were high school students from two accelerated and one unaccelerated class. Four measures of lexical repetition were employed: (1) the overall frequency of paired relations; (2) the degree to which T-units are connected by clusters of repetition; (3) the proportion of repetition which is paraphrase; and (4) the ratio of unique semantic referents to repetition tokens. A multiple regression analysis was employed to determine the degree to which variation in topic, background (native language, academic class, or native/nonnative), and writing proficiency predicts changes in the measures.; Topic, background, and proficiency predict significant variation in the repetition measures. The persuasive texts exhibit greater repetition of propositions, whereas the descriptive texts tend to repeat only the object being described. The background and proficiency variables, on the other hand, relate more to the way a writer develops supporting information within a text than to the communicative goal. Among the nonnative speakers, East Asian writers use more repetition than the Arab and Spanish writers. The latter groups frequently use techniques such as embedding narratives or grammatical parallelism to provide supporting information. The East Asian writers, however, seem to use less-linear modes of development, and thus their essays require greater repetition to appear cohesive. With regard to proficiency, less proficient native and nonnative writers employ more repetition. This may result because less proficient writers lack the linguistic abilities and/or rhetorical strategies for developing supporting information. Instead they tend to repeat the central message(s) of the text. The only variable which was not a significant predictor for any of the measures was the native/nonnative distinction.; In addition to the greater understanding of variables affecting second language writing, this study also explores how repetition creates cohesion within a text and the relation between repetition and text effectiveness. |
