1.1 Definition and design

The reading-writing continuation task, namely, combining reading and writing, is a kind of writing task that closely follows a reading comprehension.As Wang (2014)put it,in this task,learners are provided with a text of an incomplete story with its ending deleted and they are then required to write a coherent and logical continuation to the story after reading it (Wang & Wang, 2014:5).

Actually, the reading-writing continuation task has long existed in language learning and teaching as a pedagogic task. It is unknown who invented this type of task. Since the proposal of the Length Approach(Wang et al,2000), this task has often been used to encourage students to write longer so as to promote English learning. Whereas,it was not until Wang (2009) proposed the principle of "learn-together-use-together"(LTUT)that the theoretical and pedagogical significance of the task was gradually recognized. And more research on the reading-writing continuation task sprang up ever since.

There has now been a large quantity of research towards the reading-writing continuation task by researchers inspired by Wang (2009).The reading-writing continuation task is found to have a great potential to promote L2 learning. Wang(2012)pointed out that the reading-writing continuation task serves to enhance EFL teaching and learning efficiency due to a combination of language learning and its use. This claim has been corroborated by Wang (2015b) in a qualitative study that probed into learners'thinking process through the think-aloud method. Began with the reading-writing continuation task, this line of research has expanded its research focus to other kind of continuation tasks. And these studies gave birth to the Extensional Hypothesis (Wang 2016), which proposes that learners can learn a language effectively in a route of comprehension, extension and production. Wang(2016)holds that a tight integration of comprehension and production is an effective way of L2 learning. Based on this hypothesis, different continuation tasks can be designed through the combination of different modalities of input (listening or reading) and output (speaking, writing or translating). 

In spite of the variety of continuation task types, the reading-writing continuation task,which is root of the Extension Hypothesis,remains the most studied by researchers and most widely used by instructors. 

1.2 Alignment in the reading-writing continuation task
 

Compared to other kinds of writing tasks, the reading-writing continuation task is found to have a greater potential to promote L2 learning. In Wang'(2012) view,to complete the text in the most logical and coherent way,learners have to dynamically interact with the given text, thus they will align their output with the given input both in linguistic features and in content. The phenomenon that L2 writers tend to use the same or similar language forms that appear in the reading text is referred to as alignment, which was proved to exist in the continuation task (Wang & Wang, 2014).

As a matter of fact, alignment is not a novel term and it has appeared very early, though not in SLA and L2 writing. Initially, the concept of alignment appeared in Pickering & Garrod's (2004)Interactive Alignment Model (IAM), a model that explains the mechanism of successful interpersonal interaction. According to this model, there tends to be a coordinated behavior between interlocutors in a dialogue. Underlying this behavior is the alignment of mental representations in the situational and linguistic level. Pickering & Garrod's (2004) hold that alignment exists in the interpersonal dialogue and is the key to successful interaction.

In contrast to the representative-oriented IAM, Atkinson et al. (2007) adopt a socio-cognitive approach to alignment. According to Atkinson et al. (2007), alignment not only occurs at the cognitive level but also at the social level in that there exists an interaction with the ever-changing environment. They extend the concept of alignment to SLA, holding that alignment is the key to L2 acquisition.

In light of the above theories about alignment, Wang & Wang (2014)speculate that alignment exists in the reading comprehension sincereading comprehension entails interaction between the reader and the text. Their research confirmed the existence of alignment in the reading-writing continuation task.They found that L2 learners tended to imitate the language forms in the provided text(Wang& Wang,2014).Consequently, Wang (2015) claimed that it was the alignment in the continuation task that promoted L2 learning,and the stronger the alignment effect, the more effective the L2 learning.

Inspired by the finding of the importance of alignment in the reading-writing continuation task, a series of studies that explore the factors affecting the alignment effect have been conducted. 

1.3 Empirical studies on factors influencing alignment in the reading-writing continuation task

The great potential of the reading-writing continuation task in promoting L2 learning was gradually recognized by researchers.In order to magnify the facilitating role of this task in L2 learning, a bunch of studies have made attempts to investigate factors that may affect the alignment effect from various aspects, including interestingness of the reading text (Xue,2013), presence of the reading text during the writing(Xiao,2013)task instruction(Yuan,2013), peer interaction (Pang, 2014; Xu,2016), linguistic complexity of the reading text (Peng,2015),textual complexity of the reading text (Xin & Li,2020), provision of a modal text (Chen, 2017), etc.

As one of the first attempts to investigate the alignment effect, Xiao(2013)probed into the effects of the presence of the reading text during the writing process. Her findings revealed that the alignment effect was stronger when learners continued a story with the reading text available during the writing process than when they were not accessible to the reading text.In addition to the stronger alignment, greater fluency and accuracy was also found in the linguistic output.

As Wang (2012) called for empirical evidence for his assumptions on the best implementation of the continuation task,studies that investigate the alignment effect from the perspective of the input material and writing instructions have sprung up in response.((Xiao,2013; Yuan,2013;Peng,2015). Xue(2013), Peng (2015)and Xin (2020)looked into different inner features of reading materials, an important element in the task. Xue (2013) investigated the effects ofinterestingness of the reading text on the alignment effect. She found that when performing the more interesting task, the learners aligned more linguistic structures with the preceding text and wrote longer continuations with fewer errors. A positive correlation was found between the alignment and interestingness of the story.

Peng(2015)looked into the linguistic complexity of the reading text. Two versions of reading materials with different linguistic complexity were designed,one of which tailored to the learners'comprehensive ability(the complex one)and the other their productive ability(the simple one).Significant difference was only found in the partial phrasal alignment. Specifically,those who read the simplified reading material showed more frequent partial alignment at the phrasal level. Notwithstanding the weak effect on the linguistic alignment, the simplified input text was found to have considerable positive impacts on the accuracy and fluency of learners'linguistic performance. A little different from Peng (2015), Xin &Li (2020)examined the textual complexity of the reading text, a more abstract characteristic which was controlled by syntactic complexity and text length. The same participants were asked to continue a short complex text (higher textual complexity)and a long simple text (lower textual complexity). It was found that the former ones produced shorter texts with higher syntactic complexity while the latter ones produced longer texts with lower syntactic complexity. These results indicated that alignment could exist in the textual level specifically, the syntactic complexity and text length. No significant differences were found in the accuracy of their language.

Yuan (2013)examined whether the writing instructions affect the alignment effect. In her study,one group was instructed to continue the story, while the other group was instructedexplicitly to imitate the language forms in the text and underlie them.Results showed that those given explicit instruction aligned more with the preceding text and made significantly fewer grammatical errors than those given implicit instruction.This study implied that the alignment effect could be affected by external enforcement. A fact the above studies unveiled is that through optimizing the input text and task instructions, the learner-text interaction could be intensified and thus lead to stronger

alignment effect. It lent support to Wang & Wang's (2014)contention that more intensive interaction brings about stronger alignment.