Font Size: a A A

Text Representation And Text Comprehension In EFL Context

Posted on:2008-02-16Degree:DoctorType:Dissertation
Country:ChinaCandidate:H W ZhanFull Text:PDF
GTID:1115360242976091Subject:Foreign Linguistics and Applied Linguistics
Abstract/Summary:PDF Full Text Request
It is of critical importance to grasp the main idea in reading. As a matter of fact, in EFL context, readers usually fail to understand the overall meaning of a text even though they have no difficulty in comprehending the meaning of each individual sentence of the text. They cannot see the wood from the trees. This study attempts to examine the relation between knowledge of text structure and the automatic processes of global comprehension in EFL context.There are two levels of comprehension involved in reading: local comprehension and global comprehension. The theories underlying global comprehension are Structure Building Framework (Gernsbacher, 1990), Macrostructure (van Dijk, 1980), and Construction-Integration Model (Kintsch, 1998). According to the Structure Building Framework, comprehension is the process of building mental structure, in which three cognitive processes are involved: laying foundations for mental structures, mapping coherent information onto developing structures, and shifting to initiate new structures. In the description of Macrostructure, global comprehension is the process of forming macrostructure by means of macro-rules. Macro-rules are semantic mapping rules: they relate proposition sequences with proposition sequences at a higher level, and thus derive the global meaning of an episode or a whole discourse from the local meanings of the discourse. In the construction-integration model, establishing an initial"textbase"gives rise to a semantically ordered hierarchical structure of information. This hierarchically structured textbase, along with the varied link strengths of its propositions, can explain global comprehension. The effect of top-level idea can be explained in terms of superordinate and subordinate effects in the hierarchy, which suggest that the ideas situated higher in the hierarchy are understood and recalled better than those that are lower. The effect can also be explained in terms of the degree and depth of cognitive processing of higher-level ideas. The higher-level ideas are more available for processing, due to their overarching connection with various lower-level ideas. Therefore, they are more active during encoding, and remain in working memory more often than other ideas.In the case of reading in a foreign language, research has indicated that, when nonnative readers are reading expository texts, readers'knowledge of a text's rhetorical structure can enhance comprehension and memory (Carrell, 1984, 1992; Lee & Riley, 1990; Vongpumivitch, 2004). Researchers have also found that there is interaction between L2 proficiency and text-structure variables. Horiba (1990) observed that high- and low-proficiency L2 learners of Japanese engaged in different levels of text processing. High-proficiency learners gave greater attention to global discourse features, whereas low-proficiency readers were more preoccupied with the extraction of local-text information. Salager-Meyer (1994) determined that text-structure variables diversely affect comprehension among ESL learners with varying L2 proficiency when reading domain-specific, content demanding texts. These findings suggest that L2 linguistic proficiency is related to structural sensitivity, thus accounting for the diverse impacts of text-structure variables on comprehension among high- and low-proficiency L2 learners.Although these studies are informative, there are very limited data that inform us about what happens in the processes of global comprehension. The relation between knowledge text structure and the automatic processes of global comprehension has not been investigated in a single study. Therefore, a research that directly examines this relation is needed.According to the model of text processing of Kintsch and van Dijk (1978) and van Dijk and Kintsch (1983), the formation of a macrostructure is an integral part of normal text comprehension. It occurs not only in response to special task demands, but is an automatic component of the process of comprehension. This generalization may not be true of Chinese students reading in English. There are many factors related to global comprehension: Inferencing, structure-building skills, working memory, knowledge of text structure. The following questions are addressed in this study:1. What are the characteristics of the processes of Chinese EFL readers'global comprehension?2. When the main idea is explicit, can L2 readers formulate the macrostructure automatically?3. What is the degree of automaticity in forming the macrostructures, when the main idea is implicit?4. In which ways readers'text-structure knowledge and L2 proficiency affect or facilitate their global comprehension?In order to explore these ideas, we designed an empirical experiment. 52 subjects took part in the experiment. Three tests were administered to measure subjects'ability of global comprehension, knowledge of text structure, and L2 proficiency. L2 proficiency is operationalized as the score on the tests of vocabulary and grammar. In the knowledge of text structure test, the subjects were asked to read a passage and given three tasks for measuring knowledge of text structure. The subjects read the passage at their own pace. The three tasks, SAQ--Graphic Organizer--Incomplete Outline, were given to the subjects one after another. Readers'ability of global comprehension was measured by a computer-based test. The subjects sat in front of the PCs, listening to the experimenter's oral instruction and demonstration. Subjects were asked to read a text and do the word recognition task. Four texts were used in the experiment. Each text contained a summarizing sentence either at the beginning or at the end of the paragraph. When the texts were used in the implicit-macrostatement condition, the summarizing sentence was deleted from the text. For the word recognition test of each text, five pairs of test words were used. The first member of a word pair was always the priming word and the second member of the pair was the target (the primed word). The five pairs of test words correspond to five types of priming condition: Macro, Micro-diff, Micro-same, Related, and Unrelated. The texts were presented on a CRT screen under the control of a computer software. Subjects read the text at their own pace. Recognition responses were made by pressing"Y"or"N"on the keyboard, with instructions to respond as quickly and accurately as possible. The time limit is 3000 milliseconds. Subjects'responses (including reaction times and accuracy) to the primed words were recorded as the data of the experiment. By comparing the difference between the reaction time for macro-words and control words, the priming effect was computed.Some of the detailed findings of this study are unique and deserve highlighting. In the GLM Repeated Measures analysis, the effect of text has been found. That is to say, subjects'performances differ from text to text. This result reveals the great difference between the reading passages and/or the test words. A further analysis (GLM Univariate) was carried out when the effect of text was controlled. In general, the effect of priming type is statistically significant. The significance values for"primtype"in both conditions are less than 0.05 (Sig.explicit=.032; Sig.implicit=.004). The effect of priming type proves the validity of the test method on the one hand; it also shows readers'ability of global comprehension on the other hand. The regression analysis tested the relationship between readers'global comprehension, KnTS (knowledge of text structure), and L2 proficiency. The result of the regression analysis shows that, the regression model is valid (F = 36.455, Sig. = .000) and that the L2 proficiency makes greater contribution (β= -.696, Sig. = .000) to the variation of global comprehension, in spite of the fact that the contribution of KnTS is not significant (β= -.269, Sig. = .090).This study is of great theoretical significance and practical significance as well. As for theoretical contribution, besides the theory of formal schemata, other frameworks are incorporated in the study to explicate the relation between knowledge of text structure and global comprehension. We have found that the Structure-building framework, notion of macrostructure, and construction-integration model offer adequate explanation for the comprehension of main ideas. When it comes to the research methods, this study examines the processes of global comprehension by means of on-line measuring, supplementing the deficits of product-oriented researches in the past. This study offers some implications for EFL reading instruction. Since knowledge of text structure contributes to the efficiency of global comprehension, we suggest two ways to improve readers'knowledge of text structure: classroom instruction, i.e., knowledge of text structure are presented to students explicitly; by increasing exposure to print, readers generalize the regularity of text genres and acquire knowledge of text structure implicitly.This study also shed lights on the design of future researches in this field. In the computer-based priming experiment, more texts for reading can be used, so as to counterbalance the effects of other variables. The experiment can be replicated on a larger sample of subject. Because the data collected are subjects'correct responses, more subjects are needed to obtain enough data for analysis. In the future research, some topics need further investigation: subjects of various grades should be investigated, so as to examine the individual differences of readers'knowledge of text structure and global comprehension; L2 proficiency and transfer of L1 text-structure knowledge; the process of inferencing, background knowledge, and global comprehension. In order to give an adequate account of L2 reading, more advanced experimental methods are needed, for instance, simulation model. By simulation model, the processes of text comprehension are demonstrated to verify a given model of text processing. For the statistical method, we suggest Structural equation model (SEM), which can be used as a more powerful alternative to multiple regression, path analysis, factor analysis, time series analysis, and analysis of covariance.
Keywords/Search Tags:EFL reading, Text representation, Global comprehension, Knowledge of text structure, L2 proficiency
PDF Full Text Request
Related items