Second Language Literacy Development in the Context of Content-Based ESL
Thomas Destino (UC Riverside)

Background
Research has approached second language (L2) reading from a variety of angles. Following her extensive review of second language (L2) reading, Bernhardt (1991) reminds that research has focused on how readers use text-based data in order to comprehend, but has neglected the study of how second language readers "gather those data in order to construct their understandings" (p. 68; emphasis hers). Further, little is known about how reading "happens" in the context of classrooms. Research has neglected to study what reading lessons consist of and how they get "played out" in individual classrooms by groups of students. The dangers of such neglect may be obvious. This lack of observational research coupled with researchers insistence on inferring second language reading development instead of tracing it leaves little wonder as to why second language reading research remains in the shadow of L1 reading research and theory.

The present study responds by investigating, both qualitatively and quantitatively, the group context for literacy across four secondary-level ESL proficiency levels. This study used a sociocognitive heuristic for analyzing the way in which reading was played out in each proficiency level over the course of several months. The same heuristic was also used to analyze students' reading comprehension on written recall protocols of five authentic L2 texts (i.e., written for native English speakers, newspapers, magazines, textbooks, etc.). In this way, the observational data provides a context for the discussion of comprehension data and vice versa.

Theoretical Backdrop
On entering this study, it was assumed that both text-based and knowledge-based features must be considered simultaneously when studying L2 text-comprehension. Text-based features refer to words, grammar and syntax, and sound/symbol features. Knowledge-based features include readers' prior knowledge, metacognitive awareness, and perceptions of text. Hence, such an approach integrates the linguistic and structural elements of text with the pragmatic nature and intentionality of text. This heuristic reflects the sociocognitive stance toward L2 literacy development assumed in this study. We also assumed that reading performs a socializing function and simultaneously reflects group values and norms.

Setting and Participants
This study was conducted in an ethnolinguistically diverse high school in southern California. The vast majority of English language learners in the school were native speakers of Spanish, comprising approximately 40% of the entire student body. At the time of the study, the school's ESL teacher, Kathy, taught four different levels of ESL, each level was a separate class. All classes were included in the study resulting in a total of 75 native Spanish-speaking ESL students. Kathy, a US-born English/Spanish bilingual, has been teaching in California for many years, first as a social studies teacher before later switching to ESL. Kathy frequently uses academic content themes in her lessons as a source of motivation and for communicative opportunities.

Observational Data
Weekly videotapes were made of lessons across all four proficiency levels. One analysis investigated variations in the group context for literacy by proficiency level. Videotapes were analyzed around several themes; Kathy's teaching objectives, distribution of the modalities of reading, writing, speaking, and listening across ESL levels, features of teachers talk, and bilingual language use practices of students. A second analysis documented what reading "looked like" in each proficiency level. All tasks and assignments that incorporated reading across the four proficiency levels were categorized. Each reading task and assignment was then analyzed using the above-mentioned heuristic. The heuristic contains three text-based features; phonemic/graphemic, word recognition, syntactic structures and three knowledge-based features; prior knowledge, metacognitive awareness, and intratextual perception. This analysis addressed the process of how each task was carried out by locating each feature of the heuristic in instruction by level. In this way, it was possible to determine the extent to which the features of reading were differentiated by task and level.

Reading Comprehension Through Written Recall
One of Kathy's main interests in this study was to get a clearer picture of the variation in her students' reading comprehension. Therefore, late in the school year students were asked to read and recall in writing in the language of their choice five authentic texts. Recalls were scored according to the Johnson (1970) analytic system and scores were submitted to a repeated measures ANOVA. A qualitative analysis was also conducted on reading recalls using the six-feature heuristic. The focus of this study was the nature of readers' errors as they reconstructed text meaning.

Findings and Conclusions
Observational Data
Although Kathy integrated content-based themes into her instruction, the bulk of her lesson objectives were decidedly linguistic. Content-area objectives were broad and were mainly used for motivational value and communicative nature. Analysis of instruction in the various modalities across levels indicate that at the lower levels students were focused more on speaking and listening with less instructional time given to reading and writing. The emphasis somewhat shifted in the upper levels to more focus on writing and speaking, with reading receiving less attention than other modalities.

Teacher talk variables like words-per-minute (WPM), use of repetition, use of gestures, and pronunciation-emphasis reflect a clearly differentiated pattern of language use by level. Fewer WPM, more gestures, more repetition, and more instances of careful pronunciation characterized teacher talk in the lower levels. As students became more proficient, frequencies decreased on all variables except WPM, which rose steadily by level. When considering the language use practices of students across proficiency levels, substantial amounts of L1 were used in the three lowest levels on all aspects of classroom interaction. It was not until the second half of ESL III and into ESL IV that students began to use more English than Spanish.

The six features of the heuristic were intricately woven into all reading tasks and assignments. Most of the same tasks and assignments were used for all ESL levels. Tasks were differentiated by level based on content complexity and the manner in which the features of the heuristic were applied in instruction. Generally, instruction at the lower levels was more focused on text-based features like phonemic/graphemic characteristics, and word-level features. More instructional attention was explicitly paid at the higher levels to syntactic features and intratextual perception features. Interestingly, however, other features like prior knowledge, metacognitive awareness, as well as word recognition were included in lessons at all levels and in most tasks and assignments. As students became more proficient and the assignments became more academically complex Kathy continued to highlight word-level features of text. Her instructional focus was to help students access meaning in the text, regardless of text content. Very little time was devoted to instruction of text content per se.

Reading Recall Data
The ANOVA revealed statistically significant differences in scores by text type and level of instruction. These data illustrate a clear progression of reading development over the four years of ESL. The qualitative data confirm the quantitative findings indicating reading development as students were able to recall more and more accurately over time. These data provide support for Bernhardt's (1991) theory of L2 reading development, suggesting that as learners develop they make fewer text interpretation errors based on phonemic/graphemic and word recognition features while committing slightly fewer based on prior knowledge and intratextual misperception. In this study, students continued to make inappropriate meaning selections due to syntactic features throughout all levels. In short, although quantitative data indicate general reading development by proficiency level, qualitative analysis indicate that readers were plagued by misunderstandings due to word recognition, syntactic, and intratextual perception features through ESL IV. Finally, data supports Kathy's continued inclusion of both text-based and knowledge-based features of reading comprehension in her ESL instruction.

References
Bernhardt, E. B. (1991). Reading development in a second language: Theoretical, empirical, and classroom perspectives. Norwood, NJ: Ablex.

Johnson, R. E. (1970). Recall of prose as a function of the structural importance of the linguistic units. Journal of Verbal Learning and Verbal Behavior, 9, 12-20.