In the field of Task-Based Language Teaching, studies on the impact of task complexity on L2 learners’ linguistic performance are considerable. These studies have examined the impact of task complexity variables along different dimensions on different types of writing tasks, lending vigorous support to the theories of task complexity. Regretfully, up to now, far too little attention has been paid to the reading-writing continuation task—a type of writing task in which a reading comprehension section is embedded. In addition, recent years have seen a growing number of studies on the reading-writing continuation task. A great deal of research has explored from different aspects the factors that affect the strength of alignment effect. However, to date, few investigations have taken cognitive complexity of tasks into consideration.
To fill these gaps, this study endeavors to, based on Robinson's Cognition Hypothesis and the Triadic Componential Framework, investigate the effects of task complexity on Chinese EFL learners’ linguistic performance and alignment in the reading-writing continuation task. Specifically, the following questions are addressed: (1) What are the effects of task complexity on the complexity, accuracy, and fluency of Chinese EFL learners’ linguistic performance in the reading-writing continuation task? (2) What are the effects of task complexity on alignment in the reading-writing continuation task? (3) Are there any correlations between alignment and linguistic performance in terms of complexity, accuracy and fluency in the reading-writing continuation tasks of different cognitive complexity?
In order to answer the above questions, the study adopted a single-factor between-subject design. Two versions of reading-writing continuation tasks with different complexity were proffered. The complexity of the task was manipulated by the presence or absence of content support (+/-content support), which was regarded as a resource-dispersing dimension variable based on Robinson’s Triadic Componential Framework. In the simple task, participants were provided with 20 pictures which contained the subsequent plots of the story, while in the complex task, not any content was provided. A total of 64 intermediate level English sophomores from a key university in Hubei participated in this experiment, and 54 pieces of valid writings and questionnaires were collected. Through software analysis and manual annotation, the data were analyzed in terms of linguistic performance and alignment.
The findings were summarized as follows. (1) When no pictures with story content were provided, the syntactic complexity of the learner’s written output was significantly improved, specifically in the number of verb phrases per T-unit, while no significant effects were found with regard to the lexical complexity, accuracy and fluency; (2) Without pictures, learners tended to align more with the preceding text at lexical, phrasal and clausal levels, whereas with the presence of pictures, learners’ written output was highly aligned with the pictures; (3) Significant positive correlations were found between alignment and linguistic performance in terms of all aspects in the complex task group, and in the syntactic complexity and lexical complexity on the whole, though no significant correlations were found in the simple task group. 
These results indicated that the reading-writing continuation tasks without content support led to better linguistic performance and stronger alignment effect, while tasks that provide content support could elicit learners to produce linguistic expressions in line with the given content. Although the findings did not support Robinson's Cognition Hypothesis, the present study advanced research in both fields of task complexity and the continuation task. And it could provide valuable implications for teachers in terms of the design and implementation of the reading-writing continuation tasks in language teaching.
 

task complexity, the reading-writing continuation task, Cognition Hypothesis, linguistic performance, alignment