An analysis of argument structure in expert and student persuasive writing /

• Main Author: Crammond, Joanna G.
• Other Authors: Bracewell, R. J. (advisor)
• Format: Others
• Language: en
• Published: McGill University 1997
• Subjects: Reasoning.; Exposition (Rhetoric)
• Online Access: http://digitool.Library.McGill.CA:80/R/?func=dbin-jump-full&object_id=37709

This study investigated differences among student writers at three grade levels (i.e., 6, 8, and 10), and between expert writers and students, in terms of (a) the extent to which argument structures were used in their persuasive texts, (b) the complexity of these argument structures (as measured by depth and elaboration), and (c) the use of general semantic structures and conjunctive ties to represent argument substructures. In addition, the study determined the predictive relationship between the holistic scores assigned to student texts and argument structure measures. To identify and analyze argument structure a model was developed that could account for the variability in structure observed across a range of persuasive writing situations. The model was a modified version of Toulmin's (1958) schematic, and its characteristics were defined using categories derived from a theory of semantic representation in discourse. === Results of the structural analyses indicated that (a) argument was the predominant organizational structure for expert and student writers, (b) over 80% of students produced elaborated arguments involving some form of opposition, (c) experts produced more arguments and more complex arguments than students, and (d) expert texts contained relatively higher frequencies for warrants, countered rebuttals, and modals, and student use of these argument substructures increased with grade level. The general semantic and linguistic analyses revealed the following patterns particular to experts: (a) the use of identification types of claims, (b) an increased use of modals and decreased use of opinions as marks of argumentation, and (c) an infrequent use of causal conjunctions to mark data structures. Results of a forward stepwise regression analysis revealed that argument structure complexity accounted for 40% of the variance associated with quality ratings assigned to students' texts. Two other variables were significant predictors: number of supporting structures and number of opposing structures. === The results were interpreted from a rhetorical perspective: the developmental and expertise-related patterns of performance associated with the use of particular argument substructures, and the representation of these substructures were seen as reflecting an awareness of and ability to manipulate one's audience---skills that are necessary to achieve the goals of persuasive discourse.