Re‑examining Contrastive Rhetoric: A Quantitative Study of Arab ESL Graduate Students' Persuasive Discourse

This study quantitatively re-examined the contrastive rhetoric hypothesis by investigating the persuasive writing of advanced Arab English as second language (ESL) and US native English-speaking graduate students. The research aimed to determine if rhetorical performance could predict language/cultural background and identify specific rhetorical dimensions that might challenge Arab ESL writers due to first language (L1) transfer. The study utilized a quasi-comparative design with participants matched for educational level and disciplinary expertise. All writers responded to the same persuasive essay prompt under controlled conditions. Quantitative analyses included Multiple Discriminant Analysis (MDA) and Analysis of Variance (ANOVA) on holistic and four analytic measures of rhetorical performance (namely, argument superstructure, Toulmin’s informal reasoning, persuasive appeals, persuasive adaptiveness). Contrary to the contrastive rhetoric hypothesis, the MDA failed to produce a model capable of significantly classifying writing samples by language/cultural background. ANOVA revealed no statistically significant differences in rhetorical performance between the Arab and US groups on any measure. Observed performance issues were common to both groups, and within-group variation was larger than between-group variation. These findings challenge Kaplan’s (1966) assertion that rhetoric and logic are predetermined by native language and culture, suggesting that student writing problems are not necessarily use to negative transfer from L1, but are possibly shaped by a complex interplay of prior educational experiences, cultural influences, institutional contexts, and personal adaptation strategies. The study underscores the importance of methodological rigor in cross-cultural writing research and suggests pedagogical approaches might benefit from addressing the universal complexities of persuasive writing rather than focusing solely on presumed L1 interference.