Task differences in writing versus speaking: Apples and oranges?

Early research on the effects of task complexity focused on oral language production such as Skehan and Foster’s (1999) study of narrative retellings. Tavakoli and Foster’s (2008) similar study drew on Levelt’s model of oral language production to explain differences in language output between two narrative task types. This line of research was extended to writing by Kuiken and Vedder (2008) who found accuracy but not syntactic complexity was related to task complexity. Whether or not written language is affected in the same way by task types as oral language is open to debate, but the fact that writing has different processing constraints from oral language is an important distinction and one noted by Jackson and Suethanapornkul (2013) in their meta-analysis of task-complexity. More recently, Johnson (2017) conducted a meta-analysis of task complexity in L2 writing and discussed a range of task features and outcome measures that have been investigated. We seek to extend Johnson’s work by conducting a meta-analysis of genre differences, as opposed to task differences, in both L1 and L2 writing. We detail the studied features of tasks and examine how they compare to differences in genre. For example, a task feature such as plus or minus reasoning demands can be compared to differences between argumentative and narrative tasks. A few studies of task differences in writing (Yoon & Polio, 2017 and Frear & Bitchener, 2015) have accounted for linguistic differences by considering the communicative requirements of the tasks, following Biber and Conrad (2009). The goal of this meta-analysis on written genres is to try to account for differences in the language produced in written tasks and determine if those differences align with task complexity features, as they might in speaking, or differences in the communicative functions of the task.