The Reading Matrix
Vol. 2, No. 1, April 2002

Developing Academic Reading at Tertiary Level: A Longitudinal Study Tracing Conceptual Change 1
Sima Sengupta

Abstract

This paper describes how classroom intervention, through developing process awareness of reading, helped change students' model of reading with a socially constructed, explicit one. It then shows the impact the changes in students' reading models had on their literacy experiences in other courses of their degree. The twenty-five participants were native speakers of Cantonese enrolled in an undergraduate degree in "Contemporary English Language" at a Hong Kong university. They attended tutorials every fortnight to discuss set readings of academic articles and book chapters for a course entitled "Language and Society". Data were gathered by audio recording classes, collating learning journal entries, and conducting in-depth interviews. The data were inductively analyzed to find emerging themes following a reiterative process of substantiating and elaborating the themes. Initially students considered word-for-word decoding the only way to read, and as they were shown how to read selectively and purposefully, students' implicit models of 'good' reading became increasingly explicit in which reading was seen as an active process entailing complex interactions between readers, writers and text. However, data also showed that such process models of reading needed contextual support if they were to have a role in students' literacy experiences outside the sheltered ESL classroom.

Introduction

Increasingly it is being argued that literacy is socially mediated and developed (See for example, Zamel and Spack, 1998; Hasan and Williams, 1996; Lankshear, and McLaren, 1993). Academic literacy evolves as the learners' understanding of the role literacy events (reading and writing) play within their academic contexts becomes clearer with constant interaction between reader and writer, an interaction mediated through the texts and discourses within a community (Luke and Freebody, 1997). The interactive mental processes student readers use to make meaning of a text are socially mediated by their experiences of appropriating the texts they read, discuss, and make their own as they write (Lotman, 1988; Wertsch, 1998). These social and mental processes could be brought home to students through making them question the models of reading they implicitly hold and helping them trace their own development as literate adults. But how does such critical awareness impact on their evolving academic literacy beyond the sheltered ESL environment?

In this paper I chart the literacy development of five groups of L2 learners in a Hong Kong tertiary institution. I will demonstrate how implicit models of L2 reading become explicit through classroom intervention and the ways in which students' awareness impacts (or fails to impact) on their literacy development. The focus of this paper is to report the findings from a study that attempted to see how the process approach in the ESL reading classroom change students' evolving models of reading within the classroom and whether this change goes beyond the ESL classroom. Such a focus will provide important information because studies in reading do not tell us the full impact of strategy-related instruction on the students' literacy development. Research into the nature of academic reading of content subjects (typically extended reading of long texts) for L2 tertiary students is not common in the ESL literature. This is surprising since a majority of L2 learners need to read a vast amount of text for their academic courses and may need help with the reading of long texts and the assimilation of information from a variety of sources.

In order to help our students, we need to know how the new strategies and pedagogies impinge on the ways in which students approach reading in other academic subjects and whether making students' implicit models of reading explicit is going to change students' reading experiences in the real-life academic context. As Leki (1995) points out, we need 'broader looks not only at their English classes but at their lives as they negotiate their way through higher education once they step outside the safe threshold of the ESL classroom' (Leki, 1995:236). Such a broader look would enable ESL teachers to examine the force of our individual theories in our classrooms.

The description in this paper constitutes a narrative of (students') experience. The importance of such reporting of ESL students' experiences has been characterised by Leki and Carson (1997) as an initial step in the direction of learning more about literacy development. This 'story' is re-told here in the belief that student experiences and student interpretations of these experiences form the core of the educational enterprise. But while as a narrative the story belongs to the learners, it is told by the researcher and thus it is inevitably an individual and in a sense 'inauthentic' representation. For this reason the genre of this research report should be seen as a combination of 'realist' and 'multivocal' texts (see Canagarajah, 1996). The study was conducted in an extended academic content reading classroom in which strategy-related pedagogy was combined with extensive content-based reading and prototypical tasks such as summarizing, presentation and discussion.

Background

Academic reading

Academic reading is complex, multi-level and different from other kinds of reading. Based on the course described below and the existing literature, I defined academic reading as purposeful and critical reading of a range of lengthy academic texts for completing the study of specific major subject areas. Academic reading is extended reading of a range of texts varying in length. Unlike the regular definition of extensive reading as close reading (see Leki, 2001:202), academic extended reading requires:

both extensive and intensive reading of texts that are discipline-specific,

careful synthesizing of material from a number of sources (Carrell and Carson, 1997) and

consciously finding authorial intentions and purposes (Huckin and Flower, 1990).

Spack (1993) points out that academic reading is often a process of actively engaging with what is read, gleaning information and then fitting this information by thinking things through and finally being able to interpret the content flexibly to suit the purposes of the academic writing assignment. However, in spite of the commonalties, exactly what counts as academic reading is subject to interpretation.

Reading research and pedagogy

Reading is a highly interactive phenomenon (Carrell et al, 1997). This interaction takes place through the activity of reading, which is itself a complex interplay between local level bottom-up strategies (identification of meaning from the level of word upwards) and increasingly more global levels of top-down, higher order mental processes and background knowledge (see Singhal, 2001 for a full discussion).

The interactions that develop literacy are often brought in through strategy-related instruction - essentially a process approach to reading which goes beyond a simple set of strategies. It includes knowledge about the processes and actions involved in reading as well as knowledge about how to monitor these processes. As Grabe (1991) puts it, metacognitive awareness in reading involves a number of abilities. One element in this process approach to reading is monitoring of cognition which entails an active, engaged and critical process of recognizing problems with information presented in texts and in this process understanding one's own comprehension problems. However, such an approach would need to situate the process within the social domain in order to be effective. Luke and Freebody's (1997) argue that "..It (reading) is about developing ways of seeing through texts, their descriptions of cultures and worlds, and how they are trying to position you to be part of these cultures and worlds" (1997: 219). Reading texts is therefore always necessarily a "complex conjoining of "word" and "world", "text" and "context". (Green, 1997: 231). Therefore, the process approach would need to teach reading strategies with reference to the overall social contexts of the texts that students read.

The process approach followed in the L2 reading classroom might make readers aware of the available reading strategies (Carrell, Pharis and Liberto, 1989). Indeed if suitably contextualised, L2 students could become aware of how these strategies could be used to read purposefully, actively and critically (Auerbach and Paxton, 1997; Kasper, 1994). Auerbach and Paxton (1997), for example, using narrative texts on literacy development, show that a process approach to reading leads to growing awareness, choice and control thus influencing the fundamental beliefs about reading. This paper starts with the notion of beliefs.

Internal models of reading

Beliefs about reading constitute each reader's 'epistemology of text' (Wineberg, 1991), which is invoked to mediate the relationship between the reader, writer and text (Wineberg, 1991 ; Schraw and Bruning, 1996). Different epistemologies, which imply different models of reading, may include a diverse set of beliefs and assumptions, from readers' assumptions about their roles as readers (Schraw and Bruning, 1996) to their self-perception of their reading ability (Shell et al. 1989). Different epistemologies can lead to different internal models of reading which predispose readers to read the same text differently.

Such models may well be shaped by the social and educational contexts in which readers function. Smagorinsky and O'Donnell-Allen (1998) found in their study of a literature group's reading of Hamlet that reading is a 'continually mediated process in which the context provides constraints that limit, channel and enable readers' ways of thinking about, talking about and representing the meaning they impute to the written sign' (Smagorinsky and O'Donnell-Allen, 1998:221).

These internal models are implicit for most readers. Schraw and Bruning (1996), using an 800-word narrative text and a questionnaire with university learners, found two types of implicit models in their study: the transmission/translation model, where the emphasis is on 'getting' the author's meaning, and the transactional model, which involves the reader in actively 'constructing' a text's meaning. Schraw and Bruning's (1996) definitions of implicit models arise from work done into reader response and are based on readers' beliefs about texts and their engagement with texts. They argue that the implicit models are influenced by three factors: experiential, sociocultural and formal instructional.

Extending internal models of reading through instruction

A reader's epistemology may be shaped and extended in the ESL reading classroom through metacognitive strategy training. Such awareness is an essential component of L2 reading pedagogy (Singhal, 2001), although reported research often only discusses how metacognitive awareness has helped both L1 and L2 readers' comprehension of short texts (Carrell, 1989; Nist and Mealey, 1991; Garner, 1987). For example, Carrell, Pharis and Liberto (1989) have shown the beneficial effect of the inclusion of comprehension-fostering explicit metacognitive training. However, some studies have looked at more extensive academic reading. Kasper (1994) reports the beneficial effect of a pedagogic approach that set out to pair academic ESL reading and content reading (i.e. an adjunct model of content-based instruction) in which one course complemented the other. Similarly Auerbach and Paxton (1997) demonstrate the value of strategy-related instruction where students become action researchers studying their own reading processes. These studies set the scene for taking reading research beyond the reading classroom and studying what happens in the arena of general academic reading at tertiary level.

The study

The course

The students in this study were enrolled in a Bachelor of Arts degree in "Contemporary English Language"; the degree course included a course on "Language and Society" (L&S) which was taught in one three-hour lecture a week, supported by tutorials. The lecturer-in-charge had set a textbook (Holmes, 1992), a reading list of 14 articles and book chapters (see Appendix 1), and an outline of the tasks for the tutorials. The tutor was expected to provide skills-based language support. According to the tutorials outline, the first two meetings were to focus on general reading of one article and one book chapter (see Appendix 1 for excerpt from guidelines), and in subsequent tutorial meetings three texts were assigned for each of three meetings. Of these three texts each student had to choose one for presentation and one for discussion. Students were expected to read the texts at home and give a short 10-minute presentation of each text, to be followed by a brief discussion of 5-10 minutes.

At the beginning of semester the author (as tutor) met with a few students informally and realized that they had not received any strategy-related instruction in secondary school or in their BA course, although it included a module named "Practical English Skills" (PES). I requested and received permission from the lecturer to introduce any aspect of reading as long as the basic outline was followed. The teaching plan started with developing purposive and selective reading strategies to help students recognize the salient features of a text, its authorial intentions, and to produce a gist of the text. This gist then lead to a reading plan, i.e. which parts of the text had to be read carefully, depending on the purpose for reading. In the tutorial the purposes included presenting a paper and discussing it, linking the ideas in different papers and questioning the main theme of the paper presented. Outside the tutorial, the purpose was to glean information for writing and discussion tasks assigned for the lectures. Tutorials 1 and 2 consisted solely of strategy-related tasks such as making a reading plan, constructing a gist while tutorials 3-5 consisted of presentations of assigned articles. In every session there were discussions of the articles read and the reading processes used to understand the writer's main arguments, e.g. strategies used, the reasons for any difficulties encountered, lessons learnt for writing assignment, and so on.

The sample

The sample consisted of twenty-five (23 females and 2 males) first year tertiary students completing a degree in "Contemporary English Language". All the students were native speakers of Cantonese, and all had received a grade C or D in the Use of English public examination in which a C is a credit and a D is considered an above average grade. Such grades were a prerequisite for entry into the BA degree. The tutorials were not remedial but compulsory for all year 1 students completing the degree. They were of intermediate proficiency and eager to talk and participate in the tutorial. The freedom of the university after the strict discipline of secondary school was still a novelty and they were ready to criticize and question. They had considerable demands on their time and the tutorial was not high on their list of priorities. No grades were given at the tutorial. They described their previous tutorial in the first semester as 'relaxing'. According to the students a typical tutorial consisting of reading out parts - discussing the difficult words and breaking up long sentences. Against this backdrop of their tutorial experience, some students were resistant to change initially (see Themes in the first lesson below) and some may not have taken the input seriously. I had told the students that I wished to systematically study the effect of an exciting new way of 'reading' and received lukewarm support. At the end, however, all wanted to provide data and a majority (22) said that they had realized that they were 'not bad' readers.

Research questions and methodology

The study, aiming to examine how students develop rhetorical consciousness through process-oriented reading instruction (Sengupta, 2000) started out with two research questions. Firstly, I attempted to find out how student readers revise their implicit models of reading when these models become explicit as a result of process-oriented reading instruction.

The data consisted of randomly audio-taped classroom lessons and field notes of all lessons, tutor's reflection evaluation of the class before and after listening to classroom tapes, and common points in student journal entries.

The data were transcribed and multiple columns with data from all sources created (see Green, Franquiz and Dixon, 1997) (see Appendix 2 for an example). Then recursive reading of the data helped me to identify the thematic clusters. Themes referred to the primary topic threads in an exchange and arose from the initial parameters that were set in studying models of reading. It was necessary to limit the concept of models of reading and based on prior research the following parameters were initially adopted to guide the themes:

• The focus of attention (word or meaning ), following Devine, 1988;
• Meaning, reader and writer arising from social constructionist tenets;
• Engagement with text, following Schraw and Bruning, 1996, and Wineberg, 1991.
  

However, in identifying the themes, these parameters had to be revised as a result of the inductive nature of the analyses. A research assistant also independently performed the thematic analysis (see Appendix 3). On the basis of the analyses, conclusions were drawn about changes in the students' models of academic reading.
A second question, arising from the research process itself, examined how students describe the impact of their changing models of reading on their academic life more generally, a context in which widespread reading is an unavoidable reality (Spack, 1997). I conducted in-depth interviews with nine of the twenty-five students almost six weeks after the last lessons in order to answer the question:

How do student readers who report substantial changes in their models of reading within a sheltered classroom see the impact of the change on their academic reading habit in general?

The 9 interviewees were chosen randomly from 22 volunteers. My aim was described to students at the time of setting up interviews as wanting to have an informal discussion of reading and university life. The key questions concerned the kinds of reading that they had done in Year 1 in all of their courses, and whether they had tried to use their ideas about reading as expressed in the tutorials. After describing their reading habits, all were shown the major themes that had emerged (see Appendix 4), and they explained their perceptions of the differences between reading in the tutorial and reading for other content courses. The audio tapes of the interviews were fully transcribed and themes detected on the basis of multiple reading by two independent readers. Common themes were added to the list of themes (arising from the transcribed lesson) wherever appropriate, but given the distinct focus of the interviews, a number of new themes emerged.

Interviews with the same nine students were conducted once again 18 months later on exactly similar lines as above. The aim was to see whether there were further changes in the reading models towards more active reading or whether the student readers had reverted back to their original models of reading.

Methodological theory and design

The study is situated within an interpretative framework in that it explores alternative ways of knowing (see Pierce, 1995; Davis, 1995). This interpretative research essentially depended on self-reported interview data to answer the second research question. We must be mindful that interview and classroom data are typically biased, as students are likely to say what they think is expected of them, especially to the teacher. This bias was minimized by the fact that interviews were conducted after the formal end of the semester and thus the students did not need to 'please' the researcher in any way as I was no longer teaching them

Findings: How implicit models became explicit

Themes found in the transcript of the first lesson

In order to understand the conceptual change, it would be interesting for readers to see what kinds of primary topic threads or themes were found in the data of the first lesson. Reading was seen as a linear additive process of reading the words, then the sentences, and so on from beginning to end by all participants. This narrow definition of reading has the following characteristics:
Linear process of reading: Everyone said that they started at the beginning and read page by page because that is how reading is supposed to be done.

Word-centred definition of meaning: Meaning was mentioned mainly at the word and sentence level although on questioning ideas were also mentioned. A typical exchange below shows this definition.

Reader's responsibility is to get the information and not engage with the text: The reader was seen as solely responsible for 'getting' the meaning by decoding words and sentences. The readers' job in academic reading was to get the information, and upon questioning, a purpose was articulated, e.g. remember it for tests, presentations and assignments as the following classroom exchange illustrates. The writer was not mentioned.

Reading process was perceived as entirely bottom-up: However, on probing, two top-down strategies were mentioned: guessing word meanings and skipping a paragraph with an obvious example. Yet when asked, 'Don't you want to know what the article is about?', the usual answer was that one needed to read the article carefully to know what it was about.

Confidence was seen as low for academic reading: Although most said they were quite good at comprehension exercises in PES, their perceived problems concerned technical vocabulary and the length and sentence structure of the reading text, but not a lack of knowing how to read purposively.

The themes from the first lesson, arising from the data of all groups (25 students) constructed after the first meeting, are presented below. The themes within the initial models arising from the first lesson are ranked, with common issues within each theme shown in brackets:

• 45% of themes concerned vocabulary, which was seen as the most important element in reading. (Unfamiliar and technical words were problematic so students needed to look up them up frequently in the dictionary.)
  
• 30% of themes related to the perception that sentences were another, almost equally central aspect of reading. (Complex content requires that all sentences are understood; sentences in academic texts were difficult to process as they contained too much embedding of new ideas.)
  
• 22% of themes were related to getting the 'correct' meaning in terms of memorizing, recalling and understanding their content and 20% were directly connected with either vocabulary or sentence structure.
  
• Only 3% concerned the important idea but vocabulary, sentences and 'accurate' meaning were always closely related with this.
  
  These themes were substantiated by the journals. In the journals most wrote about understanding important words though nobody mentioned how the importance of a word was determined. The themes indicated a tremendous respect for the written word - the texts. At the end of the lesson, for every group I asked a common question, "Do you sometimes read these articles and chapters - and think that this is not true - this is rubbish - I disagree?" This question met with incredulity, much amusement and the most common response was, "But it is in a book - it is written".
  
  The evolving model
  
  In the second lesson another major theme emerged, namely 'writer', which was added to the network of themes. The following extract shows this theme (bold), with '…' signifying another speaker taking over:
  
  

All subsequent meetings started with the presentation of texts and their discussion. However, often a presentation focused as much on the content of the text as on the writer's ability to write a reader-friendly text and the reader's process of getting to the summary content for presentation and discussion. The ensuing debates were often heated when two student readers may have arrived at different interpretation of a single text. By now the class were calling the kinds of selective purposive reading promoted in the class as 'lazy' reading as the idea of thinking through was considered "lazy" in comparison with laboured working out of subjects and embedding in each paragraph. The data shown in Table 1, illustrates the themes in bold, showing some of this change. Here we see students questioning Labov and looking at the rhetorical functions of textual information. Students were grappling with the central idea in a reading text in spite of the fact that they had not really understood all the arguments. The realization of the themes within the text is in bold and classroom processes are indicated in curly brackets.

In the above data we see the student readers beginning to challenge the writers, attempting to make meanings individually and even resorting to evaluations of writer's texts. We also see, within the second theme in Table 1 (entitled 'vocabulary/dictionary') that the apparent move away from vocabulary and other concerns may be only a way of meeting my (the teacher's) expectations.

The new themes resulting from these lessons were interesting in that the writers and their texts were taking centre stage - yet the themes that were central to the word-bound model had not been relinquished; instead, they had merely been relocated as both the journals and classroom discussions indicated. This is understandable since vocabulary and sentence structure are real concerns within the L2 reading context and just because students had started to think about the writers and interpret their texts in their individual ways, the perceptions regarding the importance of the initial themes (sentence, vocabulary) did not change.

The change was in the way students were interpreting the texts, with each student bringing individual elements to it. For example, one student wrote in her journal after the fourth lesson:

'What did I learn about reading this time? I am quite surprised. I understood all the words (even the word ethnography - which I looked up in a Chinese-English dictionary) and sentences but still I do not understand the article. I feel so frustrated.'

Vocabulary and sentences were often minimally mentioned in class at this stage but were apparent in the journals. Student models extended beyond the obvious themes - as one writer wrote in her journal: 'Reading lesson becomes a challenging time now.' So if we look at the themes in the 'explicit' process models arising from the final lesson, ranked, with the common issues within each theme shown in brackets, we see a considerable change:

• 78% of themes concerned the writer/meaning/reader combination in terms of writer's purpose or ability/inability to make the main points clear. Of these themes, 18% included an evaluation of the writer, e.g. agreeing with the writer's point of view. 45% of these themes explicitly related to the texts the writers wrote, which was seen as the most important element in reading, highlighting structural parts such as introduction and conclusion, or linguistic features such as signals (see Sengupta, 2000). 32% of themes mentioned reader purpose.
• 14% of themes mentioned reader's process and sometimes background knowledge.
• 8% of the themes still mentioned vocabulary and sentence-level complexities.

Of course, one may say that these were themes fed to the students in class - and that is essentially true. However, the above themes were part of natural classroom discussion and journal entries (for example, 8% on vocabulary and sentence structure were largely arising from the journals). At this stage my role was becoming more managerial in seeing that everyone was involved. Sometimes concepts or relevance were explained but the presentations and discussions were lead and sustained by the students themselves. In-depth interviews provided further evidence to support this new process model of reading in which the readers' focus seemed to have shifted to meaning and readers not only 'got' the meaning, as articulated at the beginning, but indeed actively made meaning by engaging in questioning the writers' text and ideas. This latter finding lends support to similar findings of other studies such as Auerbach and Paxton (1997), who report, 'Whereas before they saw their task as taking meaning from a text, at the end of the course they talked about how they brought their own thinking to a text' (Auerbach and Paxton, 1997:48). Similar to the experience of Auerbach and Paxton (1997), these lessons were steeped in texts, their writers and readers - with raising rhetorical awareness as the aim, naturally the discussions took on a definite flavour. It is clear that themes that surfaced in the last lesson were not remotely envisaged ten weeks earlier at the beginning of the tutorials. But what was novel was that these students, who had expressed such respect for the written text and such lack of confidence as readers, had started to challenge the texts read and critique the way information was presented and argued. They were daring to question Labov!

I started questioning whether as a result of these critical encounters with texts over a relatively short period of time, these students were able to extend their models this far or was this a clever ploy to give me what I wanted? As the teacher I needed to know whether there was a different model at work outside the confines of the ESL tutorial and my watchful eye.

Findings: The impact of the changing model beyond the ESL classroom

During the classroom discussions, I often asked about reading situations beyond the L&S classroom and the answers were not encouraging. Although three learners had mentioned in their journals the possibility of the reading lesson making other lessons easy, no student mentioned active construction of meaning or challenging the written word in the reading texts assigned in other courses.

In the interviews there were a number of combinations of multiple themes, including, for example, writer-reader interaction/use of vocabulary or sentence structure to create an effect. The data seemed to indicate some breadth in the description of reading in that even when talking about reading beyond the L&S classroom, the narrow description of word-for-word reading that typified the first lesson was not apparent. Although 7 of the 9 interviewees reported that they did try to get a feel of the texts they read by looking at the pages and /or subheadings or reading the beginning (introduction) and/ or end (conclusion) first, that is where the impact of the process model stopped operating and they resorted to passive word-for-word reading rather than active questioning and linking of information.

Risk taking

One central theme pertained to the 'risk' factor that students perceived in connection with active readiing. They believed that their teachers demanded word-for-word reading and 'correct' meaning, though not one of the nine who were interviewed could remember if the teachers had specifically asked for such reading. Two interviewees explained that from the way teachers reacted to 'wrong' answers and the way the grades suffered, it was clear that there was a 'correct' meaning in the teacher's mind. These in-depth interviews pointed to three major interrelated categories within the theme of 'risk-taking'. Firstly, students saw their reading as high-stake and thus lacked the confidence to evaluate and engage; secondly, they felt that their learning of reading had taught them to regurgitate rather than engage and question and finally, they believed that teachers expected a right interpretation. These categories further explained the learners' reluctance to be more active outside the classroom (see figure 1).

It appears that an active engagement with the writer of a given text was only a feature of the L&S classroom because this was expected - thus a separate model was at work there. However, within the larger educational context active engagement with the writer was seen as an unnecessary risk because the students' agreement and evaluation was not sought by their teachers. It is interesting that such a view was expressed since a random check of assignments set in all courses for year one showed that critical evaluation was usually demanded as students were often asked to critically comment on language data or review articles. However, applied linguists working in various disciplinary areas have started to question whether faculty expectations are made (and, in fact, can be made) completely clear to students (see, for example, Spinks, 1996 as cited in Candlin and Plum, 1999; Candlin and Plum, 1999). It is possible, even likely, that the meaning of 'analysis' or 'discussion' - words frequently featured in assignments - need to be elaborated and supported with examples in order to shape student perceptions of what is expected and valued by the discipline (see Atkinson, 1997).

These students believed that with their background knowledge it was unrealistic to expect that they will 'correctly' interpret a text and identify possible problems with it (see Carrell, 1988). One interviewee said: 'You know it was good the way we talked about these writers - but they are the - the - big people. If we think they are wrong then - what do you think? You think I am wrong - I have poor English and cannot understand.' Three interviewees went so far as to say that they were glad that active reading was not required because then they could not 'memorize' or 'quote', with one interviewee observing that, 'lazy reading was not lazy at all'. It is true that with evolving knowledge a critical stance may be easier to adopt. However, it might also be necessary for content teachers and ESL teachers alike to constantly draw their students' attention to the kinds of links that they can make across the texts they read and the ways in which each text builds on or brings into tension the existing knowledge they have. In the L&S tutorials this was made possible through constant critical comparison of texts, and that may have been the reason why purposive reading was reportedly done by the readers for this particular subject.

Changing models due to changes in expectations

Within the larger academic context, at year 1, these learners saw a clear role for gleaning every possible morsel of information from the texts set for reading, remember as much of this information as possible and then regurgitate it in their assignments and tests. Their models of reading were largely influenced by their lives as students, leading them to continue with their original models of reading (albeit in a slightly modified form by including writer, reader and text quite superfluously), with the models articulated in the L&S class being kept aside as a practical alternative. Similar to Spack's (1997) and Leki's (1995) participant/s, these learners' most pressing concern was managing competing demands within the constraints imposed by a second language as well as the countless other demands on their time. Risk taking was not seen as worthwhile beyond the comfort of the L&S classroom since it would, they felt, neither be appreciated nor expected.

In the subsequent interviews I discovered that the process model had indeed emerged over time by interviewing the same nine students in their Year 3, eighteen months after the study. All reported that a very similar model to the L&S model taught earlier had emerged - engaged purposive reading was reported with rare mention of vocabulary or sentence structure. The themes did not include vocabulary or sentence structure at all. Students were now talking about reading in a very different language. The reader must note that at this stage of their degree the students were completing a dissertation. Their purpose of finding literature supporting or relating to the dissertation was the major theme. There was description of how a general gist is made of all the texts that seem to be relevant and how only a handful are selected through critical evaluation of the writers and their texts.

Yet it was the context of Year 3 of their degree, in which they had to write a dissertation and read in order to get their ideas clear, which seemed to have extended the model taught previously in the Language and Society (L&S) module to their academic experience more generally. Six of the nine students interviewed seemed to think that it was possible to read in this way because of their past experience in L&S while the other three had to be asked if the work in L&S in their Year 1 had helped, the students having totally forgotten their work in the L&S module.

To summarize, by the end of year 1, the implicit models, that were predominantly word-focused as Devine (1988) would call them, had undergone changes only within the confines of the L& S classroom. Meaning, reader and writer were themes that were identified within the later classroom data and these themes continually mediated all discussions of the process. Yet as noted by Smagorinsky and O'Donnell-Allen (1998) the context, in this case the readers' perception of the teacher and course expectations, initially provided constraints that limited and somewhat disabled readers' ways of reading and thinking about reading. Yet at the end, it was their "reading" of the expectations and the nature of the assignments that extended the conceptual change. This very fact led me to question whether there are transmission or transactional models that are a fixed entity in student readers' epistemology, especially when these models become explicit or if the explicit model itself is continually shaped and reshaped by what we teachers ask them to do.

Conclusion

This experience of changing beliefs about reading, however, raises a set of new questions: Firstly, should I have taught literacy behaviours that could be sustained by the context and not imposed my theoretical beliefs on the first year students? The answer to this question depends on a related question: How do we frame our instruction in literacy development? Personally I felt that the study showed progress towards developing an informed model of reading. The reader may ask if I was trying to do a disservice to my students by getting them to think about meaning as a reader-based evolving construct? The answer to that question is 'no'. I was familiar with the curriculum as this was a Bachelor's degree course offered by my own Department. I knew that there was a strong emphasis on seeing discourse as socially constructed and I felt that I was preparing my students to take the critical stance that colleagues lamented was missing in our students' work. Unfortunately, my students did not see this emphasis at the initial stage of their study. In retrospect, I would agree that students' context of learning should have featured more prominently in that they should have been helped to see how tertiary educational demands go beyond their experience of reading in secondary schools. As teachers we may need to take students to new ways of knowing but unless the courses we offer fully support these ways, it may be a questionable venture.

Secondly, how far should we be slaves to our educational context? Schraw and Bruning's (1996) argument that the implicit model is influenced by three contextual factors, experiential, sociocultural and formal instructional seemed to hold even when the implicit became explicit. From experience these students, at their first year, felt that critical, purposive and engaged reading was an option that their formal-instructional context would not support. Freshly arriving from an examination-oriented secondary educational culture, they were not ready to take the risk of constructing their own knowledge, Schraw and Bruning's (1996) transmission model was by far the easiest model slightly modified with a dash of transactional element. Yet transactional elements took centre stage with the changing demands of the formal instructional setting. So should the students have benefited from a more contextually appropriate offering of these ideas? Or do these findings demand that we take more serious note of our context as suggested by Leki and Carson (1997) and deliberately manipulate the context in order to sustain in our students an 'epistemology of text' (Wineberg, 1991) that is involved and socially constructed?

As Hasan (1996: 410) points out, "If literacy is what education is about, and education is supposed to be truly egalitarian, and if the aim of education is to enable participation in the production of knowledge - and not just reproduction - then it follows that we would need to develop in all pupils the ability to reflect, to enquire, to analyse and to challenge" (my highlights). And if indeed education is about critical engagement, we see from this exploratory study that teaching critical engagement may result in apparently superficial conceptual change that can only be fully sustained by critically changing the context. Yet in the changes we see that the students can go beyond the imposed contextual constraints, if only they are helped with appropriate pedagogical support.

Acknowledgment: I am grateful to Dwight Atkinson for his insightful comments and useful suggestions for revising this paper. I wish to thank my colleagues, Liz Hamp-Lyons and Martin Warren, who have read and commented on earlier versions of this paper.

References

1 This paper reports research funded by a Departmental Research Grant from the Department of English, Hong Kong Polytechnic University.

i They had learnt the skills of skimming or scanning. They did not know when to skim or scan and how to use skimming, scanning, in conjunction with other strategies.

ii Sound-centred meanings were ignored as there was no reading aloud involved.