- Email: [email protected]
Empirical examination of EFL readers' use of rhetorical information
English for Specific Purposes 21 (2002) 81±98
www.elsevier.com/locate/esp
Empirical examination of EFL readers' use of rhetorical information Ana Cristina Lahuerta MartõÂnez * University School of Technical Engineering, University of Oviedo, Manuel Llaneza 75, 33208 GijoÂn, Asturias, Spain
Abstract The present paper investigates the use of text structure as a tool to facilitate and improve EFL students' comprehension of a text written in a foreign language. It explains the results of an experimental study carried out to analyse the relationship between use of the rhetorical organization that a text employs, on the one hand, and the comprehension and the reproduction of information of the text on the other. We ®nd that it is only when reproduction and conscious recognition coincide in the reader, that the structure has a positive eect on reading comprehension and reproduction of the information presented in a text. When the reader does not recognize the organization of the text (even if he reproduces it), this structure does not aect the reader's performance. In this way making readers aware of the rhetorical organization becomes our ®rst criterion for an approach to structure as a teaching instrument. # 2001 The American University. Published by Elsevier Science Ltd. All rights reserved. Keywords: Rhetorical resources; Reading aid; Reading comprehension; Reproduction; Conscious recognition of structure
1. Introduction The present work investigates the use of text structure as a tool to facilitate and improve English as a foreign language (EFL) students' comprehension of a text written in English. We make a proposal for teaching reading comprehension in a non-native language based on the exploitation of text structure, making our readers aware of and capable of interpreting the rhetorical information a text presents.
* Tel. :+34-985-182277; fax: +34-985-182240. E-mail address: [email protected] 0889-4906/01/$20.00 # 2001 The American University. Published by Elsevier Science Ltd. All rights reserved. PII: S0889-4906(00)00029-6
82
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
As language teachers we often come across learners who do not understand either the individual sentences in a particular paragraph of a text, or the point or value which that paragraph conveys in the text as a whole. Among the possible reasons for this is that these learners presumably may not have hypothesized a clear rhetorical scheme for the text at hand, into which the group of sentences could ®t as a functional unit of discourse. Our exploitation of text structure as a tool to improve reading comprehension is part of an approach to the teaching of reading comprehension, the ultimate objective of which is to make readers capable of an intentional processing based on the adequate and eective use of two types of knowledge depending on the reading situation they face: the ®rst, knowledge of semantic relations, that is, how sentences may be joined by comparison, causality, etc., or how paragraphs may express two contrasting ideas, and the second, knowledge of the way these relations are manifested in the text, in other words the lexical, syntactic resources the foreign language employs to express them. This processing implies the procedural capacity to draw on relevant knowledge sources and select information from dierent levels (linguistic and nonlinguistic) in an adequate way. To learn a language means to learn words and sentences, and also to learn the procedures to retrieve and process them. That is, the process of learning needs both declarative and procedural knowledge. In this paper the focus will be on the kind of procedural knowledge that relates to comprehension: comprehension involves extracting the meaning in the light of all available linguistic cues in combination with the learner's general knowledge of the world. We distinguish between top-level and bottom-level comprehension to study which cues to meaning learners use. The top level is constituted by the knowledge system, that is, the schematic knowledge. Complementarily, the bottom level is constituted by the language system, i.e. the informant's knowledge of the L2 at the various linguistic levels (systemic knowledge). Our approach to the use of structure is integrated within a reading model which takes both levels into account. It is a discourse semantic approach to reading. The general structure of this model is particularly derived from Widdowson's (1983) theory of language use, in which for the ®rst time two traditions of research, schema theory (with authors like Rumelhart, 1980, and Carrell, 1988, 1991, 1992) and discourse analysis (e.g. Van Dijk and Kintsch (1983)), are put together in an appropriate framework. According to this model, readers make use of two types of knowledge in reading comprehension: systemic and schematic knowledge which correspond to the readers' linguistic and communicative competence respectively. And through procedures of interpretation, which are part of the readers' discourse competence, in top-down/bottom-up directions, they relate their schematic knowledge (content and formal) to the textual clues, actualizing in this way the text as discourse. The text is interpreted so that the linguistic elements have a communicative value in addition to a linguistic meaning. Consistent with this view of reading, we follow a type of analysis of discourse structure which looks into discourse structure from the point of view of the reader, trying to detect in the surface structure of discourse the elements which allow the
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
83
reader to interpret text. This view is represented by Winter (1977, 1982, 1992), Hoey (1979, 1983, 1991), Crombie (1985a,b) and Jordan (1984, 1992). According to this view of structure from the interpreter's point of view, discourse is framed at a lower level into binary discourse relations, that is relations such as cause-eect, conditionconsequence, etc., which are themselves part of a general level of organization (discourse macro-patterns). The rhetorical organization is manifested through resources in the language system. Linguistic resources have a schematic function because what they do is to prepare the reader for the recovery of discourse from the written text through interpretative procedures. In addition to schematic knowledge related to the local level of rhetorical organization of discourse structures, readers also possess formal schemata of a global nature. This relates to the global rhetorical organization of the text, i.e. how the dierent elements of the communicative dynamics of the text make it hang together as a whole. According to this view, the rhetorical information is interpreted from the actual interplay between local and global formal schematic knowledge and systemic knowledge. This interpretation assumes the selection and integration of information in an intentional way: the reader will focus on a global formal schematic level, on the local formal schematic level, or on the systemic level depending on the reading situation. The empirical studies presented in this paper are preliminary studies which aim to build up an approach to a teaching of reading comprehension based on the exploitation of the rhetorical information, understood as making readers aware of the rhetorical organization of texts, and making them capable of interpreting the rhetorical features. We analyse the eect of structure (the rhetorical organization of expository prose) on EFL reading performance through the study of the relation between use of structure, on the one hand, and comprehension and reproduction of information, on the other. This approach will give us information about the use our readers make of structure and the eect it has on their reading. Reading research has shown how formal schematic knowledge is activated in reading comprehension. More speci®cally, when the readers reproduce the rhetorical organization of the text they recall more information from the passage and understand the passage better. Carrell's (1984b) study is one of the few which centred speci®cally on the eect of text structure on ESL readers. She addresses the question of whether there are any dierences among non-native readers of English of varying native language and cultural backgrounds in their interaction with English expository texts of dierent rhetorical organization. She conducts an experiment to investigate the eects of these various dierent English rhetorical patterns on the reading recall of adult ESL students. The most interesting ®nding was that if ESL readers possess the appropriate formal schema against which to process the discourse type of the text, and, if they use that formal schema to organize their recall protocols, more information is retrieved. Block's (1986) study examines strategies used by second language readers and considers the use of text structure as a strategy. Block categorizes strategies into two levels: general comprehension and local linguistic strategies. General strategies include comprehension-gathering and comprehension-monitoring. Recognition of text structure is included among general strategies. Local strategies deal with attempts to understand speci®c linguistic units.
84
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
This study shows how many L2 readers possess strategic resources to control their reading. However, only some of them are able to use those resources as an aid; most apply them sporadically and unsystematically. In this experiment, the readers who used background knowledge of textual organization improved their reading comprehension and recall. Recent research on text has provided evidence of the relationship between coherence and the comprehensibility of a text. The concept of coherence has been used to describe the extent to which the sequencing of ideas in a text makes sense and the extent to which the language used to present those ideas makes the nature of the ideas and their relationship apparent. Studies that present readers with more or less coherent versions of a text have shown that the more coherent versions yield better comprehension (McKeown, Beck, Sinatra & Lexterman, 1992). Carrell (1985) reports a controlled training study designed to answer the question of whether ESL reading can be facilitated by teaching text structure explicitly. The training introduces the reader into the use of organization as a key for understanding. Students are given an explanation of description; causation; problem/ solution and comparison types of rhetorical organization and the signals that mark each type to show readers how to use the corresponding rhetorical organization to organize their writing. The results indicate that training in rhetorical organization of expository texts signi®cantly increased the amount of information that 25 intermediate-level ESL students could recall. The same conclusion was drawn by Chou Hare, Rabinowitz and Schieble (1989) and Berry, Scheer and Goldstein (1993) from studies which focused on teaching structure as an aid for reading performance (comprehension and recall of information). The previous studies show that the presence of recognized organizational patterns tends to increase comprehension and recall of information. The present study extends the previous research by examining whether similar results are found in another group of EFL students in the context of an ESP course. 2. Experimental study 2.1. Research questions For this experiment we asked the following questions: 1. Is there any relationship between presence of organization and absence of organization, and reading comprehension? 2. Is there any relation between the ability of EFL readers to use the rhetorical organization of a text and the understanding of the text? 3. Is there any relation between the ability of EFL readers to use the rhetorical organization of a text and the amount of ideas reproduced? 4. Are use and recognition of rhetorical structure dependent variables?
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
85
For this experiment, we compared the comprehension and reproduction of a text with no organization to the same text written with explicit signals of several semantic relations that give dierent organizations to the text (temporal (chronological sequence); causative (cause/eect); comparison), and problem-solution.1 We constructed a text with no organization because we wanted to compare the eect of the presence or absence of rhetorical organization on comprehension and reproduction of information. We hypothesized that ecient readers would use their knowledge about rhetorical organization to help their reading performance. 2.2. Subjects One way of eliminating extraneous variables in our experiment and of controlling its internal validity was to recruit a group of subjects which is as homogeneous as possible. In our research care has been taken to select a group of subjects who were as equivalent as possible in terms of age group and previous exposure to EFL. Sixty subjects participated in this study. The subjects were of a similar age, ranging from 20 to 25 years. They were Spanish students of English as a foreign language. At the time of the research they had been studying English for at least 4 years before entering university. They were all students of English (second year) at the Technical University School of Industrial Engineering in GijoÂn. The name of the course in which the study was conducted was English for Science and Technology. 2.3. Materials In order to control the structure and content of the information while investigating the eects of discourse type, ®ve versions of a single passage were written. The text was ``The Loss of Body Water'', taken from Carrell (1984a). We chose a text that was from a content point of view, relatively unfamiliar to the readers. In the literature we had found (e.g. Taylor & Beach, 1984; Birkmire, 1985; Roller, 1990), that the eect of text structure varies depending on the extent of the readers' knowledge of the topic of the text. The in¯uence of structure is exerted on moderately unfamiliar texts. Each version contained identical content information. Four of the ®ve texts had clear organizational patterns. The texts were written according to ®ve general semantic relations between propositions: chronological sequence that constituted a collection of descriptions; causation; comparison and problem/solution. One of them had no clear organizational pattern. For this version, we obscured the rhetorical structure of the text through the use of transformations and omitted all rhetorical signals.2 1 These correspond to the ®ve basic ways of organizing expository discourse for which authors like Meyer (1975, 1979) have gathered empirical evidence. 2 To make these transformations, we took into consideration Kintsch and Yarbrough's (1982); van Dijk and Kintsch's (1983) and Horowitz's (1982, 1987) opinion that reordering should be sensible, not random.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
86
The passages we used are in Appendix A1. The underlined words identify dierent ideas in the passages; the words in bold type explicitly signal the semantic relations that constitute the discourse types of the passage and the signals that mark them; the rest is identical information across the ®ve versions. 2.4. Procedure The experimental design involved the use of an experimental group and a control group. The 60 subjects who participated in the experiment were each randomly assigned to one of the ®ve text versions. In this way, each text was read by 12 subjects. Four 12-subject groups that received an organized version of the passage constituted the experimental group. Another 12-subject group constituted the control group. Those were the subjects who received the passage with no organization.3 The independent variables were the presence or absence of organization. The dependent variable was the subjects' scores in both a comprehension and reproduction of information test. The subjects were given the text; they then read the text at their own individual reading rates to ®nd out what it said about the topic. An immediate written recall was then requested. The subjects were asked to write down everything they could remember from the text, using their own words or words from the text. They were asked to try to write in complete sentences and not just to list isolated words or ideas. They were to try to show in their writing how the ideas from the text were related to each other. Finally, the subjects were asked whether they recognized the rhetorical organization of the passage they had been given. If they answered that they recognized it, they had to identify the type of organization. This test was administered as a normal activity in the students' academic life. The subjects were not made aware that they were taking part in an experiment. This was done to control the external validity of the experiment. We have no reason to suspect that the students' results in the test were a reaction to the experimental arrangement. The test was presented as an activity of the English Language for Science and Technology programme and the teacher of the group herself carried out the experiment. The identical information in all ®ve text versions was reduced to a total of 21 idea units (for the speci®cation of common ideas we followed Carrell (1984a)). The 21 common idea units were: 1. loss of body water is required by athletic coaches; 2. this is required of wrestlers; 3. this is required of boxers; 3 We investigated whether the use of an experimental group integrated by more subjects than the control group was statistically possible, and were told that such distribution was possible, and the results would be statistically valid.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
87
this is required of judo contestants; this is required of karate contestants; this is required of football team members; this is done so athletes will attain speci®ed body weights; these speci®ed body weights are lower than their usual weights; loss of body water causes damage to cardio-vascular functioning; this limits work capacity; a loss of 3% of body water; damages physical performance; a loss of 5% of body water; results in heat exhaustion; a loss of 7% of body water; causes hallucinations; a loss of 10% or more of body water; results in heat stroke; results in deep coma; results in convulsions; and if not treated, death will result.
Each recall protocol was scored for presence or absence of the 21 identical idea units. They were each classi®ed according to the semantic relation organizational type used by the subject to organize the protocol. We examined the overall rhetorical organization of the protocol and classi®ed it as either meeting the de®nition of one of the four original text types (collection of descriptions, causation, problem/ solution or comparison) or not. Appendix B contains examples of the recall protocols. To be classi®ed under the problem/solution type, the protocol had to state a problem and present a solution to the problem. The information in the problem/solution did not have to be the same as that in the original text. The requirement of a collection of descriptions type was a group of descriptions about a topic, where the descriptions were collectively organized. Protocols were classi®ed as causation if the structure consisted of an antecedent and consequent. In order to be classi®ed as a comparison type, the overall structure had to contrast opposing points of view, one of which may have been the subject's own point of view. We also scored the recall protocols to ®nd whether the students recognized the structural organization of the text. We used a YES/NO scale to indicate whether they identi®ed the organizational structure of the original text correctly, or not. In order to score the results of the comprehension test, we scored each recall protocol in terms of whether it showed that the reader had understood the passage or not, using the following scale to measure the overall comprehension of the passage: 0 points: The subject did not understand the passage. 1 point: The subject understood the passage in general. He/she, however, did not understand the whole passage. 2 points: Perfect comprehension of the passage.
88
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
We will explain the results obtained considering each research question separately. We explain the results obtained in each case, both in terms of the speci®c research question under consideration and in terms of our general experimental work. 2.5. First research question We wanted to see the eect of structure on EFL readers' comprehension, to check whether the readers who read a text with a certain organization comprehended it better. In order to approach our research question, we isolated the variables we wanted to work with: Text (A,B,C,D,E) and comprehension [the scale used was: 0 (good and complete comprehension); 1 (comprehension that is not complete); 2 (bad comprehension)], and analysed the relationship between these two variables. For the statistical analysis of that relation, we used the Chi-square signi®cance test. The level established for signi®cance was P<0.05.4 We calculated the experimental value of Chi-square of Pearson for the following comparisons: 1. Five texts together. 2. Every text with respect to Text E. 3. The four organized texts with each other. We ®rst considered all texts together, we then compared each text with Text E, since we hypothesized that, as this text had no organization or rhetorical signals, it would be more dicult to understand, as we have explained. We also compared the four organized texts with each other because we wanted to see whether there were any dierences between these texts in relation to comprehension. Previous studies (Meyer & Freedle, 1984; Carrell, 1984b) have shown this eect on recall. They have shown how some types of rhetorical organization are more facilitative of recall than others. The results obtained are shown in Table 1. We compared the result obtained in each case to that given by the tables for the distribution of Chi-square to see if it was signi®cant at the level of signi®cance accepted. When we consider the ®ve texts together we see that the result is signi®cant at a P<0.05 level (degree of freedom=8). When we consider each text with respect to text E the results are signi®cant at a P<0.05 level (degree of freedom=2), in every case. Finally, the comparison of the four organized texts with each other shows no signi®cant results at a P<0.05 level. We shall now explain all these statistical results. Both when we consider all texts together and when they are compared to the text with no organization, we observe a signi®cant relationship between text structure and comprehension. We see that an organized text has a positive eect on comprehension. Text E is signi®cantly less understood than the other texts. In fact, when we analysed in detail the protocols of the 12 students who read this text, we saw how they 4 This is the level accepted (a risk of 5% of error). However, the result may be signi®cant at a lower level of signi®cance. We will indicate those cases in which the result is signi®cant at a lower level of signi®cance in a scale from P<0.05 to P<0.005, and then, the result is stronger.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
89
Table 1 Textcomprehension (®ve texts together). Five texts together: chisquare=16.4* 1a textcomprehension (each text with respect to text E) Text A/text E: Chi-square=6.32* Text B/text E: Chi-square=6.86* Text C/text E: Chi-square=9.12* Text D/text E: Chi-square=8.16* 1b Textcomprehension (each text with respect to each other) Text A/text B: Chi-square=0.24 Text A/text C: Chi-square=1.68 Text A/text D: Chi-square=2.64 Text B/text C: Chi-square=1.68 Text B/text D: Chi-square=2.16 Text C/text D: Chi-square=0.48 *P<0.05.
were unable to understand what the text was about (with the exception of one subject). For them, the text was just a list of unrelated ideas. With respect to the four organized texts (A,B,C,D), unlike previous studies that measured recall, we did not ®nd signi®cant dierences in comprehension of the four texts, that is, there was not a speci®c organizational type that facilitated comprehension. We can say thus far that the organized text resulted in better comprehension. In conclusion, it seems that the rhetorical information helped the readers in the process of comprehension. We will return to this conclusion later in this discussion. 2.6. Second research question Our next research question was the following: is there any relationship between the use of the discourse organizational type of the original text and the comprehension of the text? We believed that such a relation did exist and those readers who made use of the rhetorical resources of the text would understand the text better. Before carrying out the statistical analysis, we ®rst analyzed the recall protocols that had the same structure as that of the original text, that is, the cases in which readers used the structure of the original text (we analyzed each subject's recall protocol using the YES/NO scale as explained previously in the procedure section). This is shown in Table 2. We noted that 28 out of 48 students used the structure of the original text. We again calculated the value of Chi-square for each text. We worked with the variables: use of original discourse type (measured in terms of YES/NO) and comprehension (measured by the scale: 0, 1 or 2). The results are shown in Table 3. The results show that when we consider all texts together, the relationship is signi®cant, at a P<0.005 level. When each text is considered separately, the relationship is
90
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Table 2 Number of recall protocols with same structure (n=12) Text A. Collection of descriptions B. Causation C. Problem/solution D. Comparison
8 6 7 7
Table 3 Use of original discourse organizational typecomprehension Text A. Chi-square=8.88** B. Chi-square=5.9* C. Chi-square=6.64* D. Chi-square=0.24 All texts together Chi-square=14.14*** *P<0.05; **P<0.025; ***P<0.005.
signi®cant at a P<0.05 level in the case of texts B and C and at a lower level (P<0.025) in the case of text A. This relationship is not signi®cant in the case of text D. Text D, a text organized on the basis of a comparative rhetorical organization, did not facilitate comprehension for readers. Text D shows no signi®cant relationship between use of organization and comprehension, unlike the other texts. We believe that this result is related to the readers' familiarity with the organizational type and their identi®cation of that type. The literature suggests that the comparison type of rhetorical organization appears to be less familiar to readers: Englert and Hiebert (1984) found that students performed less well in a task consisting of rating a target sentence as belonging with a certain structure when the structure type was comparison, providing some evidence that students may be less aware of this structure. Furthermore, Hiebert, Englert and Brennan (1983) and Horowitz (1987) found that students were less aware of the comparison structure than of the other structures. The relationship between organizational type and comprehension was especially signi®cant in text A. Text A is a text organized with a chronological sequence of semantic relations that makes up a collection of description-type rhetorical organization. The authors mentioned above also showed that this type of organization is one that readers are aware of. In the case of our own readers, this is a type of rhetorical organization that they encounter frequently in their reading. We observed a distinction in the eect of the type of organization a text uses that seems to result from the readers' recognition or identi®cation of the organizational type. In the next research question we will see whether the same result is obtained
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
91
with respect to the reproduction of information, and we will consider the results together. 2.7. Third research question Our third research question was the following: if readers used the discourse organization of the original text, would they reproduce more ideas? In order to ®nd out whether there was a relationship between the use of structure and the reproduction of ideas, we worked with these two variables (use of structure/ number of ideas reproduced) and calculated the value of Chi-square to see if the relationship was signi®cant. The results of the relationship between use of original discourse organizational type and reproduction of ideas are shown in Table 4. These results show that when we consider all texts together, the relationship is signi®cant (degree of freedom=2), at a low level (P<0.005), that is, the relationship is very strong. When each text is considered separately, the relationship is signi®cant in texts B and C at a P<0.05 level, at a lower level (P<0.025) in the case of text A. It is not signi®cant in text D. The relation was especially signi®cant in text A (constructed with a collection of description type of organization). The readers who used this type of organization reproduced more information from the text. The readers who used the comparison type of organization reproduced less information from the text. This is opposed to previous studies in L1 and L2 (e.g. Meyer & Freedle, 1984; Carrell, 1984b) that show that the description type of organization is less facilitative of encoding and reproduction of information. Meyer and Freedle and Carrell consider that this result is explained by the inherent characteristics of the loosely-structured collection of description type of organization, unlike the causation, problem/solution, and comparison structures which are more highly organized types of top-level structures, because of the relationships which hold between the top-level nodes. If readers possess the formal schemata against which to process the more highly structured types of organization, these facilitate encoding and retrieval. It must, however, be noted that in both these studies and in our own experiment use of structure is measured by examining the rhetorical organization utilised in the
Table 4 Use of original discourse organization typereproduction of ideas Text A. Chi-square=8.52** B. Chi-square=5.99* C. Chi-square=6.96* D. Chi-square=4.72 All texts together Chi-square=23.18*** *P<0.05; **P<0.025; ***P<0.005.
92
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
recall protocol. A reader is said to have recognized the top-level organization (i.e. possess the relevant formal schema) if the reader used the top-level organization in the recall protocol produced. Now a new factor appears: the organizational types that the readers are familiar with, identify and recognize seem to facilitate reading performance more. We have to take this variable into account and thus in order to solve these con¯icting results, we will dierentiate between use of structure as reproduction of the rhetorical organization (which may be unconscious), and use of structure as conscious recognition of the rhetorical organization. It may happen that it is not the characteristics of the organizational type but its recognition by readers that facilitates reading, that when readers do not consciously identify the rhetorical organization of the text (although they may reproduce it in their recall protocol), this structure is not an eective reading aid and does not have an eect on their reading performance. We will consider in the following research question the issue of whether those readers who reproduce the rhetorical organization of the text also consciously recognize it. 2.8. Fourth research question In order to investigate whether there was a relationship between use of structure and recognition of structure, we decided to measure recognition and use separately: recognition (possession of the relevant schema) is measured by having the students independently identify the rhetorical organization; use (activation of the relevant schema) is measured by examining the rhetorical organization utilised in the recall protocol. The results are shown in Table 5. The results show that when we consider all texts together, the relationship is signi®cant (degree of freeedom=1), at a low level (P< 0.005). When considering each text separately, the relationship is signi®cant in all texts at a P<0.05 level, with the exception of Text D where the relationship is not signi®cant. These variables are, thus, dependent. Those subjects who used the discourse type of the original text also reported that they recognized it. This result is important since it indicates that our subjects were aware of and recognized the dierent
Table 5 Use of structurerecognition of structure Text A. Chi-square=12* B. Chi-square=8.8* C. Chi-square=8.4* D. Chi-square=0.24 All texts together Chi-square=20.12** *P<0.05; **P<0.005.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
93
rhetorical organization patterns as they indicated in their answers. This relation was not signi®cant in Text D. We can say now that this type of rhetorical organization is not recognized by our readers. Consequently, it seems that, as we suggested in our previous research question, the eect of the use of structure has to do with the reader's recognition of the organizational type. The conscious identi®cation of the rhetorical organization of the text has a positive eect on the readers' reading performance. Thus, measuring use of structure by having the readers identify the rhetorical organization is more reliable than measuring it by examining the rhetorical organization utilized in the recall protocol, because this reproduction may be unconscious and the factor of the readers' conscious awareness of the organizational type cannot be taken into account. 3. Discussion We present the recognition of the rhetorical organization of texts and of the predictive character both of signals and formal knowledge, as our way of approaching and exploiting the rhetorical information that texts present in the English classroom. This brings us closer to a reading based on the activation and utilisation of cues from a greater number of levels, and increased interaction between levels. In order to develop our approach to a teaching of reading comprehension based on the exploitation of the rhetorical information (i.e. making readers aware of the rhetorical organization of texts, and make them capable of interpreting the rhetorical features), we carried out a preliminary empirical investigation. We started the empirical investigation by questioning the use our of own readers make of structure, and the eect such use has on their reading, and compared our results to those obtained in the literature. We have also clari®ed the notion of ``use'' of structure. The question of the unconscious and conscious utilization of the organization of a text has been considered and studied to establish the way we measure use and the way we present it in our instruction. The conclusions of our empirical study are as follows: The use of the rhetorical organization of the text has a positive eect on ESP readers' understanding and reproduction of the information of a text. We ®nd that several factors have to be taken into account when the use of structure is analyzed to see its eect on reading, as well as its possible use to facilitate reading. It is necessary to distinguish between reproduction of the rhetorical organization of the text and conscious recognition of that organization. As the reproduction may be unconscious, this measure does not dierentiate between a conscious and an unconscious utilization of structure. We distinguish between use of structure as reproduction and use of structure as conscious recognition. We ®nd that only when reproduction and conscious recognition coincide in the reader, does the use of structure have a positive eect on reading comprehension and reproduction of the information present in a text. When
94
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
the readers do not recognize the organization of the text (even if they reproduce it), this structure does not aect the readers performance. In this way making readers aware of the rhetorical organization becomes our ®rst criterion for an approach to structure as a teaching instrument in an ESP context. Thus, we make a proposal for teaching reading comprehension to ESP students based on the conscious use of structure understood as the interpretation of the rhetorical information and, of its predictive function. The activities will be expanded to introduce readers to the organizing principles of English, and to develop the procedures to process the rhetorical information through selecting and integrating information in an intentional way. The aim of the instruction is to develop in the readers the procedural capacity to make use of relevant knowledge sources, as well as awareness of the mental processes used to assimilate the information. The advantages we see in this approach are that it favours the interaction between reader and text, the readers' own knowledge is taken into account in the instruction, the development of the readers' awareness of their mental processes is also part of the instruction, and ®nally it promotes an independent and autonomous use of the resources at one's disposal, which is very important in the ESP context. Students on short ESP courses need to be able to continue their ESP learning without ESP teachers once they have moved on to their specialist studies. They will only be able to do this if they have been helped to do so while on the ESP course. Appendix A A1. Text AÐCollection of descriptions Several aspects of the loss of body water will be discussed. First, athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. Second, the loss of body water sustained by a 150-pound individual each day is three pints of water. Third, the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A2. Text BÐCausation It is true that athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
95
As a result, tragedies are unwittingly caused by the coaches who require this loss of body weight in these situations. These tragedies occur due to the fact that the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A3. Text CÐProblem/solution A serious problem is that athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. A solution to this problem is for school administrators to suspend coaches who require athletes to lose body water. This step must be taken due to the fact that the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A4. Text DÐComparison Athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. In contrast to the action taken by coaches, the American Medical Association strongly condemns the loss of body water for athletes. They comdemn loss of body water due to the fact that the the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A5. Text EÐText with no rhetorical organization and no rhetorical signals A loss of 3% of body water damages physical performance and a loss of 5% produces heat exhaustion. The loss of body water damages the cardio-vascular functioning, which limits work capacity. A loss of 7% of body water produces hallucinations. Losses of 10% or more of body water end in heat stroke, deep coma, and convulsions; without treatment, death will result.
96
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water and they will attain speci®ed body weights. Weights are considerably below the athletes' usual weights. Appendix B. Sample recall protocols B1. Collection of descriptions The loss of body water is discussed. First, athletic coach will ask boxers, judo, karate contestants and football team members to attain speci®ed body weights. Their weight is below usual athletic weight. Second, a loss of 150-pound each day is three pints. Third, loss of body water causes damage to cardio-vascular functioning. A loss of 3% of body water aects to work capacity, and a 5% produce exhaustion. Loss of 7% produce hallucinations. Losses of 10% or more, cause heart stroke, and if not revised can cause death. B2. Causation The athletic coaches have some restrictions in the weight of boxers, judo contestants, karate contestants or football players. So, some of them have to lose water of the body. The author explains the eect of the loss of water. If the loss is of 3%, this causes cardiovascular problems. Five percent causes heat exhaustion, 7% hallucinations, and if the loss reaches 10% or more, heat stroke, deep coma, or even death. B3. Problem/solution A serious problem in the athletic world is the loss of body water. Coaches require boxers, judo contestants, karate contestants, football team members etc. to lose body water. These athletes must have a speci®c weight. One solution is that they do not employ coaches who need that his athletes lose water. The loss of body water produces physical damage, heat exhaustion, hallucination, convulsions or death. B4. Comparison Athletic coaches frequently require boxers, judo contestants and football team members to lose body water so that they will attain speci®ed body weighs. In contrast to the idea taken from the coaches the Medical Association don't like this loss of body water because it damages the cardivascular function of the body. If you loss a 3% of body water you could be damaged physically. A loss of 5% results heat exhaustion. A loss of 7% produces hallucination, and a loss of 10% produces convulsions, so the work make by the coaches is dangereous for the body.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
97
B5. Text with no organization The loss of body water causes damage to cardiovascular functing. If you loss 3% of body water you note it. People who practice sport must control his weight. References Berry, C., Scheer, A., & Goldstein, C. (1993). Eects of text structure on the impact of heard news. Applied Linguistics Psychology, 7, 381±395. Birkmire, D. P. (1985). Text processing: the in¯uence of text structure, background knowledge, and purpose. Reading Research Quaterly, X(3), 314±326. Block, E. (1986). The comprehension strategies of second language readers. TESOL Quaterly, 20(3), 463± 494. Carrell, P. L. (1984a). Schema theory and ESL reading: classroom implications and applications. Modern Language Journal, 68(4), 332±343. Carrell, P. L. (1984b). Evidence of a formal schema in second language comprehension. Language Learning, 34(2), 87±112. Carrell, P. L. (1985). Facilitating ESL reading by teaching text structure. TESOL Quaterly, 19(4), 727± 752. Carrell, P. L. (1988). Some causes of text-boundedness and schema interference in ESL reading. In P. L. Carrell, J. Devine, D. Eskey, Interactive approaches to second language reading (pp. 101±114). Cambridge: Cambridge University Press. Carrell, P. L. (1991). Second language reading: reading ability or language pro®ciency?. Applied Linguistics, 12(2), 159±173. Carrell, P. L. (1992). Awarenesss of text structure: eects on recall. Language Learning, 42(1), 1±20. Chou Hare, V, Rabinowitz, M., & Schieble, K. (1989). Text eects on main idea comprehension. Reading Research Quaterly, XXIV(1), 72±88. Crombie, W. (1985a). Discourse and language learning. A relational approach to syllabus design. Oxford: Oxford University Press. Crombie, W. (1985b). Process and relation in discourse and language learning. Oxford: Oxford University Press. Englert, C. S., & Hiebert, E. H. (1984). Children's developing awareness of text structures in expository materials. Journal of Educational Psychology, 76, 65±74. Hiebert, E. H., Englert, C. S., & Brennan, S. (1983). Awareness of text structure in recognition and production of expository discourse. Journal of Reading Behaviour, 13, 359±365. Hoey, M. (1979). Signalling in discourse (discourse analysis monograph, no. 6) Birmingham: University of Birmingham. Hoey, M. (1983). On the surface of discourse. London: George Allen and Unwin. Hoey, M. (1991). Patterns of lexis in text. Oxford: Oxford University Press. Horowitz, R. (1982). Rhetorical structure in discourse processing. In R.SamuelsS. J. Horowitz, Comprehending oral and written language (pp. 115±130). London: Academic press. Horowitz, R. (1987). The eect of structure on ESL reading comprehension. Reading Research Quaterly, XIV, 1, 57±64. Jordan, M. P. (1984). Rhetoric of everyday English texts. London: George Allen and Unwin. Jordan, M. P. (1992). An integrated three-pronged analysis of a fund-raising letter. In W. C.ThompsonS. A. Mann, Discourse description. Diverse linguistic analyses of a fund-raising text (pp. 171±227). Amsterdam/Philadelphia: John Benjamins Company. Kintsch, W., & Yarbrough, J. C. (1982). Role of rhetorical structure in text comprehension. Journal of Educational Psychology, 74, 828±834. McKeown, M. G., Beck, I. I., Sinatra, G. M., & Lexterman, J. A. (1992). The contribution of prior knowledge and coherent text to comprehension. Reading Research Quaterly, 27(1), 79±99.
98
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Meyer, B. J. F. (1975). The organisation of prose and its eect on memory. Amsterdam: North Holland Publishing Co. Meyer, B. J. F. (1979). A selected review and discussion of basic research on prose comprehension. Technical Report. Department of Education Psychology, University of Arizona. Meyer, B. J. F., & Freedle, R. O. (1984). Eects of discourse type on recall. American Educational Journal, 21(1), 121±143. Roller, C. M. (1990). The interaction between knowledge structure variables in the processing of expository prose. Reading Research Quaterly, XXV, 2, 79±89. Rumelhart, D. E. (1980). Schemata: the building blocks of cognition. In R. J. Spiro, B. C. Bruce, W. F. Brewer, Theoretical issues in reading comprehension (pp. 33±58). New Jersey: Lawrence Erlbaum Associates Publishers. Taylor, B. M., & Beach, R. W. (1984). The eects of texts structure instruction on middle-grade students comprehension and production of expository text. Reading Research Quaterly, XIX, 2, 134±146. Van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension. New York: Academic Press. Widdowson, H. (1983). Learning purpose and language use. Oxford: Oxford University Press. Winter, E. O. (1977). A clause relational approach to English texts: a study of some predictive lexical items in written discourse. Instructional Science, 6(1), 1±92. Winter, E. O. (1982). Towards a contextual grammar of English. London: Allen and Unwin. Winter, E. O. (1992). The notion of unspeci®c versus speci®c as one way of analysing the information of a fund-raising letter. In W. C. Mann, S. A. Thompson, Discourse description. Diverse linguistic analyses of a fund-raising text (pp. 131±171). Amsterdam/Philadelphia: John Benjamins Company.
www.elsevier.com/locate/esp
Empirical examination of EFL readers' use of rhetorical information Ana Cristina Lahuerta MartõÂnez * University School of Technical Engineering, University of Oviedo, Manuel Llaneza 75, 33208 GijoÂn, Asturias, Spain
Abstract The present paper investigates the use of text structure as a tool to facilitate and improve EFL students' comprehension of a text written in a foreign language. It explains the results of an experimental study carried out to analyse the relationship between use of the rhetorical organization that a text employs, on the one hand, and the comprehension and the reproduction of information of the text on the other. We ®nd that it is only when reproduction and conscious recognition coincide in the reader, that the structure has a positive eect on reading comprehension and reproduction of the information presented in a text. When the reader does not recognize the organization of the text (even if he reproduces it), this structure does not aect the reader's performance. In this way making readers aware of the rhetorical organization becomes our ®rst criterion for an approach to structure as a teaching instrument. # 2001 The American University. Published by Elsevier Science Ltd. All rights reserved. Keywords: Rhetorical resources; Reading aid; Reading comprehension; Reproduction; Conscious recognition of structure
1. Introduction The present work investigates the use of text structure as a tool to facilitate and improve English as a foreign language (EFL) students' comprehension of a text written in English. We make a proposal for teaching reading comprehension in a non-native language based on the exploitation of text structure, making our readers aware of and capable of interpreting the rhetorical information a text presents.
* Tel. :+34-985-182277; fax: +34-985-182240. E-mail address: [email protected] 0889-4906/01/$20.00 # 2001 The American University. Published by Elsevier Science Ltd. All rights reserved. PII: S0889-4906(00)00029-6
82
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
As language teachers we often come across learners who do not understand either the individual sentences in a particular paragraph of a text, or the point or value which that paragraph conveys in the text as a whole. Among the possible reasons for this is that these learners presumably may not have hypothesized a clear rhetorical scheme for the text at hand, into which the group of sentences could ®t as a functional unit of discourse. Our exploitation of text structure as a tool to improve reading comprehension is part of an approach to the teaching of reading comprehension, the ultimate objective of which is to make readers capable of an intentional processing based on the adequate and eective use of two types of knowledge depending on the reading situation they face: the ®rst, knowledge of semantic relations, that is, how sentences may be joined by comparison, causality, etc., or how paragraphs may express two contrasting ideas, and the second, knowledge of the way these relations are manifested in the text, in other words the lexical, syntactic resources the foreign language employs to express them. This processing implies the procedural capacity to draw on relevant knowledge sources and select information from dierent levels (linguistic and nonlinguistic) in an adequate way. To learn a language means to learn words and sentences, and also to learn the procedures to retrieve and process them. That is, the process of learning needs both declarative and procedural knowledge. In this paper the focus will be on the kind of procedural knowledge that relates to comprehension: comprehension involves extracting the meaning in the light of all available linguistic cues in combination with the learner's general knowledge of the world. We distinguish between top-level and bottom-level comprehension to study which cues to meaning learners use. The top level is constituted by the knowledge system, that is, the schematic knowledge. Complementarily, the bottom level is constituted by the language system, i.e. the informant's knowledge of the L2 at the various linguistic levels (systemic knowledge). Our approach to the use of structure is integrated within a reading model which takes both levels into account. It is a discourse semantic approach to reading. The general structure of this model is particularly derived from Widdowson's (1983) theory of language use, in which for the ®rst time two traditions of research, schema theory (with authors like Rumelhart, 1980, and Carrell, 1988, 1991, 1992) and discourse analysis (e.g. Van Dijk and Kintsch (1983)), are put together in an appropriate framework. According to this model, readers make use of two types of knowledge in reading comprehension: systemic and schematic knowledge which correspond to the readers' linguistic and communicative competence respectively. And through procedures of interpretation, which are part of the readers' discourse competence, in top-down/bottom-up directions, they relate their schematic knowledge (content and formal) to the textual clues, actualizing in this way the text as discourse. The text is interpreted so that the linguistic elements have a communicative value in addition to a linguistic meaning. Consistent with this view of reading, we follow a type of analysis of discourse structure which looks into discourse structure from the point of view of the reader, trying to detect in the surface structure of discourse the elements which allow the
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
83
reader to interpret text. This view is represented by Winter (1977, 1982, 1992), Hoey (1979, 1983, 1991), Crombie (1985a,b) and Jordan (1984, 1992). According to this view of structure from the interpreter's point of view, discourse is framed at a lower level into binary discourse relations, that is relations such as cause-eect, conditionconsequence, etc., which are themselves part of a general level of organization (discourse macro-patterns). The rhetorical organization is manifested through resources in the language system. Linguistic resources have a schematic function because what they do is to prepare the reader for the recovery of discourse from the written text through interpretative procedures. In addition to schematic knowledge related to the local level of rhetorical organization of discourse structures, readers also possess formal schemata of a global nature. This relates to the global rhetorical organization of the text, i.e. how the dierent elements of the communicative dynamics of the text make it hang together as a whole. According to this view, the rhetorical information is interpreted from the actual interplay between local and global formal schematic knowledge and systemic knowledge. This interpretation assumes the selection and integration of information in an intentional way: the reader will focus on a global formal schematic level, on the local formal schematic level, or on the systemic level depending on the reading situation. The empirical studies presented in this paper are preliminary studies which aim to build up an approach to a teaching of reading comprehension based on the exploitation of the rhetorical information, understood as making readers aware of the rhetorical organization of texts, and making them capable of interpreting the rhetorical features. We analyse the eect of structure (the rhetorical organization of expository prose) on EFL reading performance through the study of the relation between use of structure, on the one hand, and comprehension and reproduction of information, on the other. This approach will give us information about the use our readers make of structure and the eect it has on their reading. Reading research has shown how formal schematic knowledge is activated in reading comprehension. More speci®cally, when the readers reproduce the rhetorical organization of the text they recall more information from the passage and understand the passage better. Carrell's (1984b) study is one of the few which centred speci®cally on the eect of text structure on ESL readers. She addresses the question of whether there are any dierences among non-native readers of English of varying native language and cultural backgrounds in their interaction with English expository texts of dierent rhetorical organization. She conducts an experiment to investigate the eects of these various dierent English rhetorical patterns on the reading recall of adult ESL students. The most interesting ®nding was that if ESL readers possess the appropriate formal schema against which to process the discourse type of the text, and, if they use that formal schema to organize their recall protocols, more information is retrieved. Block's (1986) study examines strategies used by second language readers and considers the use of text structure as a strategy. Block categorizes strategies into two levels: general comprehension and local linguistic strategies. General strategies include comprehension-gathering and comprehension-monitoring. Recognition of text structure is included among general strategies. Local strategies deal with attempts to understand speci®c linguistic units.
84
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
This study shows how many L2 readers possess strategic resources to control their reading. However, only some of them are able to use those resources as an aid; most apply them sporadically and unsystematically. In this experiment, the readers who used background knowledge of textual organization improved their reading comprehension and recall. Recent research on text has provided evidence of the relationship between coherence and the comprehensibility of a text. The concept of coherence has been used to describe the extent to which the sequencing of ideas in a text makes sense and the extent to which the language used to present those ideas makes the nature of the ideas and their relationship apparent. Studies that present readers with more or less coherent versions of a text have shown that the more coherent versions yield better comprehension (McKeown, Beck, Sinatra & Lexterman, 1992). Carrell (1985) reports a controlled training study designed to answer the question of whether ESL reading can be facilitated by teaching text structure explicitly. The training introduces the reader into the use of organization as a key for understanding. Students are given an explanation of description; causation; problem/ solution and comparison types of rhetorical organization and the signals that mark each type to show readers how to use the corresponding rhetorical organization to organize their writing. The results indicate that training in rhetorical organization of expository texts signi®cantly increased the amount of information that 25 intermediate-level ESL students could recall. The same conclusion was drawn by Chou Hare, Rabinowitz and Schieble (1989) and Berry, Scheer and Goldstein (1993) from studies which focused on teaching structure as an aid for reading performance (comprehension and recall of information). The previous studies show that the presence of recognized organizational patterns tends to increase comprehension and recall of information. The present study extends the previous research by examining whether similar results are found in another group of EFL students in the context of an ESP course. 2. Experimental study 2.1. Research questions For this experiment we asked the following questions: 1. Is there any relationship between presence of organization and absence of organization, and reading comprehension? 2. Is there any relation between the ability of EFL readers to use the rhetorical organization of a text and the understanding of the text? 3. Is there any relation between the ability of EFL readers to use the rhetorical organization of a text and the amount of ideas reproduced? 4. Are use and recognition of rhetorical structure dependent variables?
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
85
For this experiment, we compared the comprehension and reproduction of a text with no organization to the same text written with explicit signals of several semantic relations that give dierent organizations to the text (temporal (chronological sequence); causative (cause/eect); comparison), and problem-solution.1 We constructed a text with no organization because we wanted to compare the eect of the presence or absence of rhetorical organization on comprehension and reproduction of information. We hypothesized that ecient readers would use their knowledge about rhetorical organization to help their reading performance. 2.2. Subjects One way of eliminating extraneous variables in our experiment and of controlling its internal validity was to recruit a group of subjects which is as homogeneous as possible. In our research care has been taken to select a group of subjects who were as equivalent as possible in terms of age group and previous exposure to EFL. Sixty subjects participated in this study. The subjects were of a similar age, ranging from 20 to 25 years. They were Spanish students of English as a foreign language. At the time of the research they had been studying English for at least 4 years before entering university. They were all students of English (second year) at the Technical University School of Industrial Engineering in GijoÂn. The name of the course in which the study was conducted was English for Science and Technology. 2.3. Materials In order to control the structure and content of the information while investigating the eects of discourse type, ®ve versions of a single passage were written. The text was ``The Loss of Body Water'', taken from Carrell (1984a). We chose a text that was from a content point of view, relatively unfamiliar to the readers. In the literature we had found (e.g. Taylor & Beach, 1984; Birkmire, 1985; Roller, 1990), that the eect of text structure varies depending on the extent of the readers' knowledge of the topic of the text. The in¯uence of structure is exerted on moderately unfamiliar texts. Each version contained identical content information. Four of the ®ve texts had clear organizational patterns. The texts were written according to ®ve general semantic relations between propositions: chronological sequence that constituted a collection of descriptions; causation; comparison and problem/solution. One of them had no clear organizational pattern. For this version, we obscured the rhetorical structure of the text through the use of transformations and omitted all rhetorical signals.2 1 These correspond to the ®ve basic ways of organizing expository discourse for which authors like Meyer (1975, 1979) have gathered empirical evidence. 2 To make these transformations, we took into consideration Kintsch and Yarbrough's (1982); van Dijk and Kintsch's (1983) and Horowitz's (1982, 1987) opinion that reordering should be sensible, not random.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
86
The passages we used are in Appendix A1. The underlined words identify dierent ideas in the passages; the words in bold type explicitly signal the semantic relations that constitute the discourse types of the passage and the signals that mark them; the rest is identical information across the ®ve versions. 2.4. Procedure The experimental design involved the use of an experimental group and a control group. The 60 subjects who participated in the experiment were each randomly assigned to one of the ®ve text versions. In this way, each text was read by 12 subjects. Four 12-subject groups that received an organized version of the passage constituted the experimental group. Another 12-subject group constituted the control group. Those were the subjects who received the passage with no organization.3 The independent variables were the presence or absence of organization. The dependent variable was the subjects' scores in both a comprehension and reproduction of information test. The subjects were given the text; they then read the text at their own individual reading rates to ®nd out what it said about the topic. An immediate written recall was then requested. The subjects were asked to write down everything they could remember from the text, using their own words or words from the text. They were asked to try to write in complete sentences and not just to list isolated words or ideas. They were to try to show in their writing how the ideas from the text were related to each other. Finally, the subjects were asked whether they recognized the rhetorical organization of the passage they had been given. If they answered that they recognized it, they had to identify the type of organization. This test was administered as a normal activity in the students' academic life. The subjects were not made aware that they were taking part in an experiment. This was done to control the external validity of the experiment. We have no reason to suspect that the students' results in the test were a reaction to the experimental arrangement. The test was presented as an activity of the English Language for Science and Technology programme and the teacher of the group herself carried out the experiment. The identical information in all ®ve text versions was reduced to a total of 21 idea units (for the speci®cation of common ideas we followed Carrell (1984a)). The 21 common idea units were: 1. loss of body water is required by athletic coaches; 2. this is required of wrestlers; 3. this is required of boxers; 3 We investigated whether the use of an experimental group integrated by more subjects than the control group was statistically possible, and were told that such distribution was possible, and the results would be statistically valid.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
87
this is required of judo contestants; this is required of karate contestants; this is required of football team members; this is done so athletes will attain speci®ed body weights; these speci®ed body weights are lower than their usual weights; loss of body water causes damage to cardio-vascular functioning; this limits work capacity; a loss of 3% of body water; damages physical performance; a loss of 5% of body water; results in heat exhaustion; a loss of 7% of body water; causes hallucinations; a loss of 10% or more of body water; results in heat stroke; results in deep coma; results in convulsions; and if not treated, death will result.
Each recall protocol was scored for presence or absence of the 21 identical idea units. They were each classi®ed according to the semantic relation organizational type used by the subject to organize the protocol. We examined the overall rhetorical organization of the protocol and classi®ed it as either meeting the de®nition of one of the four original text types (collection of descriptions, causation, problem/ solution or comparison) or not. Appendix B contains examples of the recall protocols. To be classi®ed under the problem/solution type, the protocol had to state a problem and present a solution to the problem. The information in the problem/solution did not have to be the same as that in the original text. The requirement of a collection of descriptions type was a group of descriptions about a topic, where the descriptions were collectively organized. Protocols were classi®ed as causation if the structure consisted of an antecedent and consequent. In order to be classi®ed as a comparison type, the overall structure had to contrast opposing points of view, one of which may have been the subject's own point of view. We also scored the recall protocols to ®nd whether the students recognized the structural organization of the text. We used a YES/NO scale to indicate whether they identi®ed the organizational structure of the original text correctly, or not. In order to score the results of the comprehension test, we scored each recall protocol in terms of whether it showed that the reader had understood the passage or not, using the following scale to measure the overall comprehension of the passage: 0 points: The subject did not understand the passage. 1 point: The subject understood the passage in general. He/she, however, did not understand the whole passage. 2 points: Perfect comprehension of the passage.
88
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
We will explain the results obtained considering each research question separately. We explain the results obtained in each case, both in terms of the speci®c research question under consideration and in terms of our general experimental work. 2.5. First research question We wanted to see the eect of structure on EFL readers' comprehension, to check whether the readers who read a text with a certain organization comprehended it better. In order to approach our research question, we isolated the variables we wanted to work with: Text (A,B,C,D,E) and comprehension [the scale used was: 0 (good and complete comprehension); 1 (comprehension that is not complete); 2 (bad comprehension)], and analysed the relationship between these two variables. For the statistical analysis of that relation, we used the Chi-square signi®cance test. The level established for signi®cance was P<0.05.4 We calculated the experimental value of Chi-square of Pearson for the following comparisons: 1. Five texts together. 2. Every text with respect to Text E. 3. The four organized texts with each other. We ®rst considered all texts together, we then compared each text with Text E, since we hypothesized that, as this text had no organization or rhetorical signals, it would be more dicult to understand, as we have explained. We also compared the four organized texts with each other because we wanted to see whether there were any dierences between these texts in relation to comprehension. Previous studies (Meyer & Freedle, 1984; Carrell, 1984b) have shown this eect on recall. They have shown how some types of rhetorical organization are more facilitative of recall than others. The results obtained are shown in Table 1. We compared the result obtained in each case to that given by the tables for the distribution of Chi-square to see if it was signi®cant at the level of signi®cance accepted. When we consider the ®ve texts together we see that the result is signi®cant at a P<0.05 level (degree of freedom=8). When we consider each text with respect to text E the results are signi®cant at a P<0.05 level (degree of freedom=2), in every case. Finally, the comparison of the four organized texts with each other shows no signi®cant results at a P<0.05 level. We shall now explain all these statistical results. Both when we consider all texts together and when they are compared to the text with no organization, we observe a signi®cant relationship between text structure and comprehension. We see that an organized text has a positive eect on comprehension. Text E is signi®cantly less understood than the other texts. In fact, when we analysed in detail the protocols of the 12 students who read this text, we saw how they 4 This is the level accepted (a risk of 5% of error). However, the result may be signi®cant at a lower level of signi®cance. We will indicate those cases in which the result is signi®cant at a lower level of signi®cance in a scale from P<0.05 to P<0.005, and then, the result is stronger.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
89
Table 1 Textcomprehension (®ve texts together). Five texts together: chisquare=16.4* 1a textcomprehension (each text with respect to text E) Text A/text E: Chi-square=6.32* Text B/text E: Chi-square=6.86* Text C/text E: Chi-square=9.12* Text D/text E: Chi-square=8.16* 1b Textcomprehension (each text with respect to each other) Text A/text B: Chi-square=0.24 Text A/text C: Chi-square=1.68 Text A/text D: Chi-square=2.64 Text B/text C: Chi-square=1.68 Text B/text D: Chi-square=2.16 Text C/text D: Chi-square=0.48 *P<0.05.
were unable to understand what the text was about (with the exception of one subject). For them, the text was just a list of unrelated ideas. With respect to the four organized texts (A,B,C,D), unlike previous studies that measured recall, we did not ®nd signi®cant dierences in comprehension of the four texts, that is, there was not a speci®c organizational type that facilitated comprehension. We can say thus far that the organized text resulted in better comprehension. In conclusion, it seems that the rhetorical information helped the readers in the process of comprehension. We will return to this conclusion later in this discussion. 2.6. Second research question Our next research question was the following: is there any relationship between the use of the discourse organizational type of the original text and the comprehension of the text? We believed that such a relation did exist and those readers who made use of the rhetorical resources of the text would understand the text better. Before carrying out the statistical analysis, we ®rst analyzed the recall protocols that had the same structure as that of the original text, that is, the cases in which readers used the structure of the original text (we analyzed each subject's recall protocol using the YES/NO scale as explained previously in the procedure section). This is shown in Table 2. We noted that 28 out of 48 students used the structure of the original text. We again calculated the value of Chi-square for each text. We worked with the variables: use of original discourse type (measured in terms of YES/NO) and comprehension (measured by the scale: 0, 1 or 2). The results are shown in Table 3. The results show that when we consider all texts together, the relationship is signi®cant, at a P<0.005 level. When each text is considered separately, the relationship is
90
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Table 2 Number of recall protocols with same structure (n=12) Text A. Collection of descriptions B. Causation C. Problem/solution D. Comparison
8 6 7 7
Table 3 Use of original discourse organizational typecomprehension Text A. Chi-square=8.88** B. Chi-square=5.9* C. Chi-square=6.64* D. Chi-square=0.24 All texts together Chi-square=14.14*** *P<0.05; **P<0.025; ***P<0.005.
signi®cant at a P<0.05 level in the case of texts B and C and at a lower level (P<0.025) in the case of text A. This relationship is not signi®cant in the case of text D. Text D, a text organized on the basis of a comparative rhetorical organization, did not facilitate comprehension for readers. Text D shows no signi®cant relationship between use of organization and comprehension, unlike the other texts. We believe that this result is related to the readers' familiarity with the organizational type and their identi®cation of that type. The literature suggests that the comparison type of rhetorical organization appears to be less familiar to readers: Englert and Hiebert (1984) found that students performed less well in a task consisting of rating a target sentence as belonging with a certain structure when the structure type was comparison, providing some evidence that students may be less aware of this structure. Furthermore, Hiebert, Englert and Brennan (1983) and Horowitz (1987) found that students were less aware of the comparison structure than of the other structures. The relationship between organizational type and comprehension was especially signi®cant in text A. Text A is a text organized with a chronological sequence of semantic relations that makes up a collection of description-type rhetorical organization. The authors mentioned above also showed that this type of organization is one that readers are aware of. In the case of our own readers, this is a type of rhetorical organization that they encounter frequently in their reading. We observed a distinction in the eect of the type of organization a text uses that seems to result from the readers' recognition or identi®cation of the organizational type. In the next research question we will see whether the same result is obtained
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
91
with respect to the reproduction of information, and we will consider the results together. 2.7. Third research question Our third research question was the following: if readers used the discourse organization of the original text, would they reproduce more ideas? In order to ®nd out whether there was a relationship between the use of structure and the reproduction of ideas, we worked with these two variables (use of structure/ number of ideas reproduced) and calculated the value of Chi-square to see if the relationship was signi®cant. The results of the relationship between use of original discourse organizational type and reproduction of ideas are shown in Table 4. These results show that when we consider all texts together, the relationship is signi®cant (degree of freedom=2), at a low level (P<0.005), that is, the relationship is very strong. When each text is considered separately, the relationship is signi®cant in texts B and C at a P<0.05 level, at a lower level (P<0.025) in the case of text A. It is not signi®cant in text D. The relation was especially signi®cant in text A (constructed with a collection of description type of organization). The readers who used this type of organization reproduced more information from the text. The readers who used the comparison type of organization reproduced less information from the text. This is opposed to previous studies in L1 and L2 (e.g. Meyer & Freedle, 1984; Carrell, 1984b) that show that the description type of organization is less facilitative of encoding and reproduction of information. Meyer and Freedle and Carrell consider that this result is explained by the inherent characteristics of the loosely-structured collection of description type of organization, unlike the causation, problem/solution, and comparison structures which are more highly organized types of top-level structures, because of the relationships which hold between the top-level nodes. If readers possess the formal schemata against which to process the more highly structured types of organization, these facilitate encoding and retrieval. It must, however, be noted that in both these studies and in our own experiment use of structure is measured by examining the rhetorical organization utilised in the
Table 4 Use of original discourse organization typereproduction of ideas Text A. Chi-square=8.52** B. Chi-square=5.99* C. Chi-square=6.96* D. Chi-square=4.72 All texts together Chi-square=23.18*** *P<0.05; **P<0.025; ***P<0.005.
92
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
recall protocol. A reader is said to have recognized the top-level organization (i.e. possess the relevant formal schema) if the reader used the top-level organization in the recall protocol produced. Now a new factor appears: the organizational types that the readers are familiar with, identify and recognize seem to facilitate reading performance more. We have to take this variable into account and thus in order to solve these con¯icting results, we will dierentiate between use of structure as reproduction of the rhetorical organization (which may be unconscious), and use of structure as conscious recognition of the rhetorical organization. It may happen that it is not the characteristics of the organizational type but its recognition by readers that facilitates reading, that when readers do not consciously identify the rhetorical organization of the text (although they may reproduce it in their recall protocol), this structure is not an eective reading aid and does not have an eect on their reading performance. We will consider in the following research question the issue of whether those readers who reproduce the rhetorical organization of the text also consciously recognize it. 2.8. Fourth research question In order to investigate whether there was a relationship between use of structure and recognition of structure, we decided to measure recognition and use separately: recognition (possession of the relevant schema) is measured by having the students independently identify the rhetorical organization; use (activation of the relevant schema) is measured by examining the rhetorical organization utilised in the recall protocol. The results are shown in Table 5. The results show that when we consider all texts together, the relationship is signi®cant (degree of freeedom=1), at a low level (P< 0.005). When considering each text separately, the relationship is signi®cant in all texts at a P<0.05 level, with the exception of Text D where the relationship is not signi®cant. These variables are, thus, dependent. Those subjects who used the discourse type of the original text also reported that they recognized it. This result is important since it indicates that our subjects were aware of and recognized the dierent
Table 5 Use of structurerecognition of structure Text A. Chi-square=12* B. Chi-square=8.8* C. Chi-square=8.4* D. Chi-square=0.24 All texts together Chi-square=20.12** *P<0.05; **P<0.005.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
93
rhetorical organization patterns as they indicated in their answers. This relation was not signi®cant in Text D. We can say now that this type of rhetorical organization is not recognized by our readers. Consequently, it seems that, as we suggested in our previous research question, the eect of the use of structure has to do with the reader's recognition of the organizational type. The conscious identi®cation of the rhetorical organization of the text has a positive eect on the readers' reading performance. Thus, measuring use of structure by having the readers identify the rhetorical organization is more reliable than measuring it by examining the rhetorical organization utilized in the recall protocol, because this reproduction may be unconscious and the factor of the readers' conscious awareness of the organizational type cannot be taken into account. 3. Discussion We present the recognition of the rhetorical organization of texts and of the predictive character both of signals and formal knowledge, as our way of approaching and exploiting the rhetorical information that texts present in the English classroom. This brings us closer to a reading based on the activation and utilisation of cues from a greater number of levels, and increased interaction between levels. In order to develop our approach to a teaching of reading comprehension based on the exploitation of the rhetorical information (i.e. making readers aware of the rhetorical organization of texts, and make them capable of interpreting the rhetorical features), we carried out a preliminary empirical investigation. We started the empirical investigation by questioning the use our of own readers make of structure, and the eect such use has on their reading, and compared our results to those obtained in the literature. We have also clari®ed the notion of ``use'' of structure. The question of the unconscious and conscious utilization of the organization of a text has been considered and studied to establish the way we measure use and the way we present it in our instruction. The conclusions of our empirical study are as follows: The use of the rhetorical organization of the text has a positive eect on ESP readers' understanding and reproduction of the information of a text. We ®nd that several factors have to be taken into account when the use of structure is analyzed to see its eect on reading, as well as its possible use to facilitate reading. It is necessary to distinguish between reproduction of the rhetorical organization of the text and conscious recognition of that organization. As the reproduction may be unconscious, this measure does not dierentiate between a conscious and an unconscious utilization of structure. We distinguish between use of structure as reproduction and use of structure as conscious recognition. We ®nd that only when reproduction and conscious recognition coincide in the reader, does the use of structure have a positive eect on reading comprehension and reproduction of the information present in a text. When
94
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
the readers do not recognize the organization of the text (even if they reproduce it), this structure does not aect the readers performance. In this way making readers aware of the rhetorical organization becomes our ®rst criterion for an approach to structure as a teaching instrument in an ESP context. Thus, we make a proposal for teaching reading comprehension to ESP students based on the conscious use of structure understood as the interpretation of the rhetorical information and, of its predictive function. The activities will be expanded to introduce readers to the organizing principles of English, and to develop the procedures to process the rhetorical information through selecting and integrating information in an intentional way. The aim of the instruction is to develop in the readers the procedural capacity to make use of relevant knowledge sources, as well as awareness of the mental processes used to assimilate the information. The advantages we see in this approach are that it favours the interaction between reader and text, the readers' own knowledge is taken into account in the instruction, the development of the readers' awareness of their mental processes is also part of the instruction, and ®nally it promotes an independent and autonomous use of the resources at one's disposal, which is very important in the ESP context. Students on short ESP courses need to be able to continue their ESP learning without ESP teachers once they have moved on to their specialist studies. They will only be able to do this if they have been helped to do so while on the ESP course. Appendix A A1. Text AÐCollection of descriptions Several aspects of the loss of body water will be discussed. First, athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. Second, the loss of body water sustained by a 150-pound individual each day is three pints of water. Third, the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A2. Text BÐCausation It is true that athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
95
As a result, tragedies are unwittingly caused by the coaches who require this loss of body weight in these situations. These tragedies occur due to the fact that the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A3. Text CÐProblem/solution A serious problem is that athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. A solution to this problem is for school administrators to suspend coaches who require athletes to lose body water. This step must be taken due to the fact that the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A4. Text DÐComparison Athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water so that they will attain speci®ed body weights. These speci®ed weights are considerably below the athletes' usual weights. In contrast to the action taken by coaches, the American Medical Association strongly condemns the loss of body water for athletes. They comdemn loss of body water due to the fact that the the loss of body water causes damage to cardio-vascular functioning, which limits work capacity. More speci®cally, a loss of 3% of body water damages physical performance and a loss of 5% results in heat exhaustion. Moreover, a loss of 7% of body water causes hallucinations. Losses of 10% or more of body water result in heat stroke, deep coma, and convulsions; if not treated, death will result. A5. Text EÐText with no rhetorical organization and no rhetorical signals A loss of 3% of body water damages physical performance and a loss of 5% produces heat exhaustion. The loss of body water damages the cardio-vascular functioning, which limits work capacity. A loss of 7% of body water produces hallucinations. Losses of 10% or more of body water end in heat stroke, deep coma, and convulsions; without treatment, death will result.
96
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Athletic coaches frequently require wrestlers, boxers, judo contestants, karate contestants, and football team members to lose body water and they will attain speci®ed body weights. Weights are considerably below the athletes' usual weights. Appendix B. Sample recall protocols B1. Collection of descriptions The loss of body water is discussed. First, athletic coach will ask boxers, judo, karate contestants and football team members to attain speci®ed body weights. Their weight is below usual athletic weight. Second, a loss of 150-pound each day is three pints. Third, loss of body water causes damage to cardio-vascular functioning. A loss of 3% of body water aects to work capacity, and a 5% produce exhaustion. Loss of 7% produce hallucinations. Losses of 10% or more, cause heart stroke, and if not revised can cause death. B2. Causation The athletic coaches have some restrictions in the weight of boxers, judo contestants, karate contestants or football players. So, some of them have to lose water of the body. The author explains the eect of the loss of water. If the loss is of 3%, this causes cardiovascular problems. Five percent causes heat exhaustion, 7% hallucinations, and if the loss reaches 10% or more, heat stroke, deep coma, or even death. B3. Problem/solution A serious problem in the athletic world is the loss of body water. Coaches require boxers, judo contestants, karate contestants, football team members etc. to lose body water. These athletes must have a speci®c weight. One solution is that they do not employ coaches who need that his athletes lose water. The loss of body water produces physical damage, heat exhaustion, hallucination, convulsions or death. B4. Comparison Athletic coaches frequently require boxers, judo contestants and football team members to lose body water so that they will attain speci®ed body weighs. In contrast to the idea taken from the coaches the Medical Association don't like this loss of body water because it damages the cardivascular function of the body. If you loss a 3% of body water you could be damaged physically. A loss of 5% results heat exhaustion. A loss of 7% produces hallucination, and a loss of 10% produces convulsions, so the work make by the coaches is dangereous for the body.
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
97
B5. Text with no organization The loss of body water causes damage to cardiovascular functing. If you loss 3% of body water you note it. People who practice sport must control his weight. References Berry, C., Scheer, A., & Goldstein, C. (1993). Eects of text structure on the impact of heard news. Applied Linguistics Psychology, 7, 381±395. Birkmire, D. P. (1985). Text processing: the in¯uence of text structure, background knowledge, and purpose. Reading Research Quaterly, X(3), 314±326. Block, E. (1986). The comprehension strategies of second language readers. TESOL Quaterly, 20(3), 463± 494. Carrell, P. L. (1984a). Schema theory and ESL reading: classroom implications and applications. Modern Language Journal, 68(4), 332±343. Carrell, P. L. (1984b). Evidence of a formal schema in second language comprehension. Language Learning, 34(2), 87±112. Carrell, P. L. (1985). Facilitating ESL reading by teaching text structure. TESOL Quaterly, 19(4), 727± 752. Carrell, P. L. (1988). Some causes of text-boundedness and schema interference in ESL reading. In P. L. Carrell, J. Devine, D. Eskey, Interactive approaches to second language reading (pp. 101±114). Cambridge: Cambridge University Press. Carrell, P. L. (1991). Second language reading: reading ability or language pro®ciency?. Applied Linguistics, 12(2), 159±173. Carrell, P. L. (1992). Awarenesss of text structure: eects on recall. Language Learning, 42(1), 1±20. Chou Hare, V, Rabinowitz, M., & Schieble, K. (1989). Text eects on main idea comprehension. Reading Research Quaterly, XXIV(1), 72±88. Crombie, W. (1985a). Discourse and language learning. A relational approach to syllabus design. Oxford: Oxford University Press. Crombie, W. (1985b). Process and relation in discourse and language learning. Oxford: Oxford University Press. Englert, C. S., & Hiebert, E. H. (1984). Children's developing awareness of text structures in expository materials. Journal of Educational Psychology, 76, 65±74. Hiebert, E. H., Englert, C. S., & Brennan, S. (1983). Awareness of text structure in recognition and production of expository discourse. Journal of Reading Behaviour, 13, 359±365. Hoey, M. (1979). Signalling in discourse (discourse analysis monograph, no. 6) Birmingham: University of Birmingham. Hoey, M. (1983). On the surface of discourse. London: George Allen and Unwin. Hoey, M. (1991). Patterns of lexis in text. Oxford: Oxford University Press. Horowitz, R. (1982). Rhetorical structure in discourse processing. In R.SamuelsS. J. Horowitz, Comprehending oral and written language (pp. 115±130). London: Academic press. Horowitz, R. (1987). The eect of structure on ESL reading comprehension. Reading Research Quaterly, XIV, 1, 57±64. Jordan, M. P. (1984). Rhetoric of everyday English texts. London: George Allen and Unwin. Jordan, M. P. (1992). An integrated three-pronged analysis of a fund-raising letter. In W. C.ThompsonS. A. Mann, Discourse description. Diverse linguistic analyses of a fund-raising text (pp. 171±227). Amsterdam/Philadelphia: John Benjamins Company. Kintsch, W., & Yarbrough, J. C. (1982). Role of rhetorical structure in text comprehension. Journal of Educational Psychology, 74, 828±834. McKeown, M. G., Beck, I. I., Sinatra, G. M., & Lexterman, J. A. (1992). The contribution of prior knowledge and coherent text to comprehension. Reading Research Quaterly, 27(1), 79±99.
98
A.C.L. MartõÂnez / English for Speci®c Purposes 21 (2002) 81±98
Meyer, B. J. F. (1975). The organisation of prose and its eect on memory. Amsterdam: North Holland Publishing Co. Meyer, B. J. F. (1979). A selected review and discussion of basic research on prose comprehension. Technical Report. Department of Education Psychology, University of Arizona. Meyer, B. J. F., & Freedle, R. O. (1984). Eects of discourse type on recall. American Educational Journal, 21(1), 121±143. Roller, C. M. (1990). The interaction between knowledge structure variables in the processing of expository prose. Reading Research Quaterly, XXV, 2, 79±89. Rumelhart, D. E. (1980). Schemata: the building blocks of cognition. In R. J. Spiro, B. C. Bruce, W. F. Brewer, Theoretical issues in reading comprehension (pp. 33±58). New Jersey: Lawrence Erlbaum Associates Publishers. Taylor, B. M., & Beach, R. W. (1984). The eects of texts structure instruction on middle-grade students comprehension and production of expository text. Reading Research Quaterly, XIX, 2, 134±146. Van Dijk, T. A., & Kintsch, W. (1983). Strategies of discourse comprehension. New York: Academic Press. Widdowson, H. (1983). Learning purpose and language use. Oxford: Oxford University Press. Winter, E. O. (1977). A clause relational approach to English texts: a study of some predictive lexical items in written discourse. Instructional Science, 6(1), 1±92. Winter, E. O. (1982). Towards a contextual grammar of English. London: Allen and Unwin. Winter, E. O. (1992). The notion of unspeci®c versus speci®c as one way of analysing the information of a fund-raising letter. In W. C. Mann, S. A. Thompson, Discourse description. Diverse linguistic analyses of a fund-raising text (pp. 131±171). Amsterdam/Philadelphia: John Benjamins Company.