2.2 Constructivist discourse synthesis

Writing from sources or discourse synthesis is perceived as a“hybrid”reading（comprehension）-to-writing（production）process in which readers/writers are requested to select，organize and connect content from source texts when they construct their own products（Spivey 1990）.As Flower（1989）pointed out，writing from sources is an

act of literacy in line with recent educational emphasis on the development of academic discourse which places special value on integrating own ideas and knowledge into the written conversation with one’s sources（...）Such integration[is expected]as a move toward critical literacy and toward realizing writing’s epistemic potential to transform knowledge rather than to report knowledge（p.26）.

In this sense，therefore，reading-to-write tasks involves both reading and writing skills.Grabe（2001）held that in performing a reading-to-write task，reader/writers are preliminarily requested to use the same skills needed for reading comprehension（e.g.，accurate word recognition，fluency in processing meaning chunks，ability to build meanings from texts，ability to make necessary inferences to prior knowledge stored in long-term memory，and competence to monitor），and minimally expected to follow the same writing processes in traditional writing-only tasks（e.g.，planning，activating background knowledge，making rhetorical decisions，revising，and monitoring）.Spivey（1990）claimed that while completing reading-to-write tasks，reader/writers’competence of reading and writing influence each other.In view of the complexity of the reading-to-write performing processes，it would be inadequate to portray the intentional act of reading to write as a linear，two separate procedures wherein a person reads a source text simply for comprehension before beginning the process of writing.This was also echoed by Grabe（2001）who reiterated that having independent reading and writing abilities is not adequate to successfully perform a reading-to-write task.While these basic abilities are minimally necessary in integrating information from a source to construct a text，reading-to-write competence is apparently not their simple sum.Reading-to-write，thus，should be explicated as both integration and interaction of the two literacy skills in which reader/writers modify the processes and strategies according to their goals and abilities，and contextual factors，and the requirements of reading-to-write task itself.What is needed now，therefore，is an integrated framework that elaborates to what extent reading and writing interact with one another in reading-to-write performing process.

A constructivist approach theorized the literate acts of reading to write as constructive meaning-making processes，also called discourse synthesis.Constructivism explicates both composing and comprehending as meaning constructing processes（Kucer 1985；Nelson&Calfee 1998）.According to Spivey（1990），discourse synthesis is

complex processes involved in readers’and writers’construction of textual meaning：how people construct meaning from texts through reading and from text through writing.Building meaning through reading entails organizing，selecting，and connecting.Readers use previously acquired knowledge to operate on textual cues，organizing mental representations that include materials they select from the text and connect with materials they generate.This constructivist characterization of the reading process extends also to literate acts in which people are writers as well as readers，those acts in which they compose texts by drawing from textual sources（p.256）.

In other words，meaning is constructed from source texts while reading and for produced texts while composing，in which such three key operations as organizing，selecting，and connecting are involved（Spivey 1990，1997）.When composing from sources，reader/writers organize textual meanings，select relevant textual content for meaning representation，and connect ideas based on inferences from sources as well as generated from previously acquired knowledge.To put it simply，when reading to write，a reader/writer conducts the constructive operations of organizing，selecting，and connecting in an attempt to construct their own meaning.Taking the constructivist approach，previous researchers primarily analyzed the practical uses of organizing，selecting，and connecting by concentrating on the textual characteristics of the reading-to-write products（Spivey 1984，1991；Spivey&King 1989）.The constructivist approach proved plausible in an effort to explain the interaction of the two literacy skills when constructing meaning in reading-to-write tasks.

2.2.1 Organizing

Organizing is one of the most important elements in any kind of writing acts.Writers need to acquire skills about how to effectively organize the ideas generated from their long-term memory in writing-only tasks.But what happens to textual organizing when readers are also writers？In exploring their major concerns in writing，Spivey（1990）conducted a comparative study by means of a questionnaire survey，finding that most of reader/writers were engaged in organizing when synthesizing texts from sources.They reported that they expected to accomplish such goals as making explicit their ideas in a logical order，giving a meaningful sequence，and classifying relevant information effectively.This study of composing from other sources initially provided some insights into what reader/writers usually do when they take the role of writer while reading.

When organizing information from source texts for writing，reader/writers recognize or establish a framework that allows them to identify，recombine，and generate associations between different content units within a single text or across multiple texts by using their prior knowledge concerning discourse patterns and textual implications（Spivey 1990；1997）.This organizing framework，usually formed in reading，also applies to writing，especially in summaries；nevertheless，it is frequently adjusted in response to the writer’s goals and relevant information from source texts.Establishing a new framework through organizing information in the source text requires reader/writers to transform content units into topical chunks according to the logical relations between the source text and their own rhetorical plans.

2.2.2 Selecting

According to Spivey（1990；1997），selecting turned out the second important process in constructivist view of reading-to-write.Understanding a text involves an unconscious ranking of elements in the sequence of importance.In fact，text content can be easily divided into small units of meaning which can be arranged into hierarchical orders.Reader/writers select information when reading to write on the basis of their criteria of textual relevance，which allows them to attribute varying degrees of attention while approaching source texts.Hidi and Anderson（1986）also emphasized that the selecting process entails conscious judgments of the relative importance of information in meaning construction according to individual’s sensitivity to importance.This awareness of textual importance is associated with general reading ability and maturity（McGee 1982；Taylor 1981）.More proficient and mature readers make better choices of what they need based on the hierarchical nature of texts.Furthermore，in reading-to-write tasks，the extent to which readers focus their attention on either main ideas or details in the source text is dependent on what their writing purposes are，usually made explicit by task specifications，and on what specific information they need in it contributes the most to such purposes（Leki 1993）.

2.2.3 Connecting

Apart from organizing and selecting，a third operation in Spivey’s constructivist view is connecting，which involves reader/writers’integrating their prior knowledge with the information explicitly cued by the text.That is，they transform source texts by connecting content generated on the basis of their own stored knowledge with that selected from source texts.This integration allows them to make inferences that are not explicit in the text，and fill in gaps in comprehension by applying prior knowledge activated from long-term memory（Kintsch 1998）.Furthermore，we can clearly see evidence of the interplay between the two sources of content，source texts and the individual’s stored knowledge.

The connecting process can be correlated with Bereiter and Scardamalia’s（1987b）conceptualization of knowledge-telling as well as knowledge-transforming in writing processes.Writers tend to make use of the knowledge-telling strategy to elicit information from the source texts without depending on connections to previous knowledge when they have ample relevant content to meet the requirement of the task（Durst 1987）.Despite the possibility that knowledge-telling may also entails the ways to respond to a text，writers involved in the knowledge-telling mode are not engaging in a real reorganization of the source content and the knowledge domain in their writing.Instead，writers employing a knowledge-transformation strategy are believed to seek to solve problems while manipulating and transforming the information from both source texts and prior knowledge in innovative ways.It is through connecting information within source texts as well as with their stored knowledge that they can achieve this goal.

Content generation by means of background knowledge varies across task demands.Tasks requiring a condensed generalization of the source text（e.g.，summary writing）do not promote extensive integration of content through connecting the available information in source texts with previous knowledge（Ascención 2004）.Reading-to-write tasks，especially in academic writing context，however，call for more extensive use of prior knowledge to connect，organize，and transform the information elicited from source texts.Furthermore，content generation employing previous knowledge is also dependent upon such factors as the quality and quantity of the information elicited from the source text，familiarity with the target topic，and the discourse community whom the reader/writer expects to read the produced text.

Spivey’s findings indicated that writers were involved in a meaning-building process whereby they performed textual transformations by conducting operations of organizing，selecting，and connecting.Her constructivist approach provides insightful endeavor to elaborate on the relationship between the two modalities，reading and writing，in integrating the discourse synthesis composing process.However，this view fails to explicate the underlying metacognitive mechanisms which regulate the three operations conceptualized in the constructivist framework.There is no doubt that the utilization of the three operations（organizing，selecting，and connecting）requires the reader/writer to activate the metacognitive mechanism for the purpose of helping him or her become consciously aware of task goals，resource evaluation，and orchestration of the use of strategies they could use.Constructivist view，in other words，failed to provide a complete picture of what reading-to-write ability entails in that it didn’t include metacognitive components in this framework.Furthermore，by reviewing Spivey’s theory that conceptualizes discourse synthesis as a“hybrid”reading-writing task，which requests the reader/writer to select，organize and connect content from sources when they compose their own texts，Segev-Miller’s（2004）longitudinal classroom writing intervention research showed that writing from sources is very challenging for even most native speaking college students without effectively deploying a repository of strategies.Contrastive writing studies also revealed that the academic need to synthesize discourse from documented sources is often demanding to many ESL writers（Connor&Kramer 1995；Prior 1995；Casanave 1995）.In this sense，we essentially need a more comprehensive understanding of EFL readers/writers’strategy use when they engage in writing from source texts.