A corpus-based and systemic functional analysis of syntactic complexity and nominal modification in academic writing
Writing quality has been considered a significant indicator of success in education (McNamara, Crossley, & McCarthy, 2010). One way to understand the predictors of proficient writing is to document the linguistic characteristics of quality writing. Syntactic complexity, defined as the sophistication of writing production through the range or variety of linguistic resources (Ortega, 2003, 2015), has been widely used to determine the complexity/maturity of writing in both first and second language writing research. Despite references to the variety, diversity, or richness of linguistic resources for syntactic complexity (e.g., Bulté & Housen, 2012; Norris & Ortega, 2009), a great majority of the past research on syntactic complexity has focused on the length of the production, viewing clausal subordination as a key indicator of syntactic growth (e.g., Hunt, 1970; Loban, 1976). Although syntactic complexity is a multi-dimensional construct, most measures have examined it at the sentence or T-unit level, ignoring the complexification at the phrase level (Biber, Gray, & Poonpon, 2011; Bulté & Housen, 2014; De Clercq & Housen, 2017). Yet the nominal style of academic writing characterized mostly at the sub-clausal level through embedded phrases has largely been overlooked in most of the operationalizations of syntactic complexity (De Clercq & Housen, 2017), and the grammatical features of phrasal complexity have been relatively ignored in educational practice. Thus, considering the construct multidimensionality of syntactic complexity (Norris & Ortega, 2009), and the critical role of phrasal complexity features (e.g., noun phrases) in the construction of academic writing such as science discourse, this dissertation adopted a complementary approach, incorporating both clausal and phrasal complexity measures to analyze written academic language across three disciplines.
Drawing on the premises of corpus linguistics methodology and systemic functional theory, this dissertation investigated syntactic complexity measures in science research writing contexts across three disciplines (agronomy, applied linguistics, and industrial and manufacturing systems engineering) and between two registers (journal research articles and master’s theses). A specialized corpus with nearly 1.9 million words was compiled. After the corpus preparation stage, the whole corpus was tagged, and normed rates of occurences for each linguistic feature in each text were obtained. A 2x3 factorial design was modeled, and multivariate analysis of variance (MANOVA) was performed to measure to what extent six syntactic complexity indices were distributed across the disciplines and between the registers. Moreover, a two-way between-subjects analysis of variance (ANOVA) was conducted to measure to what extent discrete linguistic features differed across the disciplines and between the registers. Finally, a more qualitative functional approach was employed, drawing on the premises of SFL-based discourse analysis and corpus-based phrasal complexity research to investigate the ways nominal modifiers are used in academic writing and the roles they take in the science discourse of these disciplines.
The overall findings of the study appeared to validate the arguments that science research writing is highly characterized by a dense nominal style. The results indicated that while discipline and register had a statistically significant main effect on syntactic complexity measures, the interaction between the two was minimal. For example, clause length was found to be significantly longer in agronomy than in the other disciplines. The results indicated that clausal subordination was mostly represented in master’s theses, while phrasal elaboration occurred more frequently in published journal articles. The findings also showed that discipline rather than register exerted a main effect on the occurrences of linguistic features in the corpus. The results showed that phrasal complexity features such as nouns and nouns as nominal premodifiers appeared more frequently in agronomy while clausal complexity features such as finite complement clauses occurred more frequently in applied linguistics. The analysis of individual linguistic features revealed that some features such as adjectives as prenominal modifiers and nominalizations were influenced by the combined interaction effect of discipline and register. The functional analysis of phrasal complexity features suggested that nominal modifiers allowed for the expression of reality in science by packaging informational content in phrases rather than clauses. The analysis revealed that nouns and NP modifiers were heavily used in the disciplines in highly technical and abstract terms, increasing the informational density of the texts.
In sum, the findings of this study contribute to the syntactic complexity research through its complementary approach to constuct definition and operationalization. While the findings present an overarching investigation of global syntactic complexity measures, the study addresses syntactic complexity via specific grammatical features, explicating the linguistic bases of phrasal and clausal complexity. A more functional analysis of nominal modification in academic writing complements the quantitative results of syntactic complexity measures. The findings in this study as well as the research methodology have the potential to inform future research on syntactic complexity in written academic discourse.