The Language Teacher
July 2002

Foundations for L2 Reading Instruction

William Grabe

Northern Arizona University

Introduction

Reading is one of the most important language skills in academic settings. It is also one of the most complex skills in which to develop strong second language (L2) fluency. Unlike speaking and writing, the reader is not able to control the message or the language used. It is also a skill that, like listening, must be carried out under real time pressure if it is done fluently. However, unlike listening, there are no opportunities to ask for clarification or additional information. Moreover, the range of vocabulary encountered in reading is much greater than is typically used in speaking and listening settings (Stanovich, 2000, p. 252-258). Given this starting point, it is fairly clear that the development of L2 reading abilities represents a serious challenge for both the learner and the teacher.

If teachers and curriculum developers are to help students make significant progress in reading instruction, they need to understand how reading works. Only in this way can we make informed decisions to guide effective reading instruction. One of the key starting points in this process involves understanding the fluent L1 reading process. While recognizing that there are major differences between L1 and L2 reading, a clear picture of fluent and effective reading must be a central consideration of all instruction, and there is far more information to draw on from L1 reading comprehension research.

At the same time that theory will inform us about the fluent reading process, and about distinct issues for the L2 reader, it does not tell us how to teach reading. At best, we can draw a set of implications for instruction. So, in addition to a strong knowledge base in reading, we need to know how to connect implications from theory to real instructional practices in a reading curriculum. Of course, spelling out all of these possibilities is beyond the scope of a single short article. The primary goal of this paper will be to address key research foundations and their implications for instruction and curriculum development, recognizing that the application of effective instructional practices within a coherent curriculum represents the other half of the picture.

The Nature of Reading and a Definition

Any description of reading abilities can begin with a simple purpose statement, such as the following: Reading is "the process of receiving and interpreting information encoded in language form via the medium of print" (Urquhart & Weir, 1998, p. 22). However, it should be evident that such a simple definition of reading will not take us very far. Complex skills and processes require more complex definitions, though such definitions must still be informative. Useful extended definitions of reading can be developed at two levels: 1) purposes for reading (why we read), and 2) components of reading ability (what skills are involved). A yet more complete picture is created by considering key processes involved in reading comprehension (how we read).

Purposes for reading

We read for a variety of purposes: Scanning, skimming, reading for general understanding, reading to learn, reading to integrate information, and reading to evaluate critically. There are several other types of reading purposes that could be considered: Reading as search process, expeditious reading, reading to write, reading while writing, and perhaps one or two other possibilities (see Alderson, 2000; Grabe, 2000; Urquhart & Weir, 1998). The key point is that there are multiple purposes for reading. As we read for different purposes, we often vary the ways that we use the cognitive processes and knowledge resources central to reading. At the same time, the actual processes and resources for reading themselves do not generally vary, just how they are used in combination. So we can still talk about reading as a single ability, while also recognizing levels of variability in response to differing purposes and tasks. To understand this consistency across purposes, a definition of reading must include a description of the component skills comprising reading abilities.

Components of reading ability

A definition of reading must recognize that a reader engages in processing at phonological, morphological, syntactic, semantic, and discourse levels, using the full range of linguistic knowledge bases. For example, fluent readers are rapid and efficient word recognizers: They do not guess upcoming words because their recognition skills are actually faster than those used to engage context information. To be fluent word recognizers, readers must also have a very large receptive vocabulary knowledge-base (Hulstijn, 2001). At another level, the reader engages continuously in goal setting, interpretive elaborating from knowledge resources, and goal monitoring. Fluent readers predict, on a general level, upcoming information and have strong expectations about the discourse organization of the text as they read. In addition, fluent readers make adjustments to enhance comprehension and carry out repairs to comprehension as needed. These, and other, components of reading are integrated as a set of activated processes and resources (in working memory) operating under intense processing-time constraints.

Fluent Reading Abilities

Fluent reading comprehension includes both lower-level and higher-level processing skills. This division is not meant to suggest that one set of skills is easier or harder than the other, only that the former set tends to be more automatized and the latter more accessible to conscious attention (Segalowitz, 2000). Lower-level skills include rapid and automatic word recognition, syntactic parsing, and semantic proposition formation (clause-level meaning units). Fluent readers must automatically recognize the vast majority of words they encounter in the text, at least 95% of the words in most cases (Hulstijn, 2001). Readers must also be able to draw key syntactic information from a text to establish accurate relations among the words and sentence parts; again, the initial parsing is usually done automatically and is not open to conscious reflection unless a problem arises with comprehension. Finally, readers must integrate lexical and syntactic information into clause-level meaning units (propositions), which can then be combined to generate textual meaning. These processes and information units are activated as part of working memory. (In fact, working memory is not some mental box that information moves to; rather, it is the sum of any given moment's pattern of activation across memory units in the brain. Information units and processes not sufficiently excited, electrically or chemically, are no longer "active" in working memory.)

Higher-level processing skills first involve the construction of a text-model of reading comprehension, representing a summary of the textual information that the reader believes is intended by the writer. As the reader progresses, the clause-level meaning units are integrated to form a general understanding of the text, with each new unit incorporated as it is created. When information is reinforced, it receives greater activation and is more central to the text model as a whole. Information that is not repeated or directly inferred loses activation and disappears through regular processes of pruning and restructuring. Inferencing is not used extensively for building the text model. The second major component of higher-level processing is the creation of the situation model of reading interpretation. A situation model expands upon the text model and incorporates the readers emotions, attitudes, background knowledge, motivations, and goals into a critical interpretation of the text--one that recognizes the author's views but also critically situates the text author from the reader's perspective (Grabe, 2000; Kintsch, 1998). Inferencing and reader knowledge play a strong role in building the situation model of text interpretation. Finally, higher-level processing requires some type of executive control processing, a monitoring of information activation, text construction, and reader goals, attitudes, and evaluations. (It should also be noted that higher-level processes and output are also networks within working memory.)

The simple sketch of reading comprehension provided above has many implications for reading development and reading instruction. Fuller details of the comprehension process and its implications are beyond the scope of this paper (cf. Alderson, 2000; Grabe, 2000; Grabe & Stoller, 2002; Stanovich, 2000; Thompson & Nicholson, 1999; for more detailed discussions). Nevertheless, several implications for instruction and curricular development will be outlined below. Before moving to instructional implications, there are two sets of issues that need to be addressed: 1) problematic issues in L2 reading research, and 2) differences between L1 and L2 reading.

Popular Notions in Reading Research that are Problematic

Four notions that are popular but problematic in reading theory deserve comment because they have been influential ideas. The Psycholinguistic Guessing Game view of reading is still popular, but it does not fit with the above description of reading comprehension for multiple reasons, and there are strong reasons for discarding the psycholinguistic guessing game. Evidence from eye movement research, context influences on readers, the time-course of word recognition, and longitudinal learning and training studies all argue that the view described in section III is a more appropriate synthesis of research findings (cf. Grabe, 2000; Grabe & Stoller, 2002). It is also clear that reading abilities are not all universally the same (Koda, 1996), as commonly claimed by the psycholinguistic guessing game. A second problematic notion, the assertion that reading develops naturally, much like speaking and listening, is often stated but seldom analyzed critically. There is obvious and overwhelming counter-evidence: One fifth of the world's population is illiterate, but nothing appears to stop this 1/5th of the world's population from speaking. Assuming that 1/5 of the world's population is not un-natural, then reading is not a naturally developing skill. Moreover, adult illiterates consistently have difficulties with basic skills required for reading (segmenting sounds to phonemes, recognizing words, making grammaticality judgments) despite having fluent speaking skills.

The role of context in reading is also a problematic issue, and one that needs to be understood better by teachers. A major distinction concerns claims about using context information to guess upcoming words (a misleading view) versus using context to build textual comprehension (a central notion) (Alderson, 2000; Grabe, 2000). At the level of word recognition, for example, poorer readers actually make greater use of context information than do more-skilled readers (Stanovich, 2000). A final problematic notion for reading comprehension is the role of authentic text resources in instruction. While many teaching experts state that only authentic texts should be used for reading instruction, there are many reasons to reconsider this advice. First, authenticity is not an easily definable concept, and what makes a text authentic is not usually spelled out in detail. Second, a classroom setting is an authentic setting in itself, and reading instruction must use those texts that more efficiently further the instructional goals of a curriculum. If pedagogically adapted texts work best for students, then they become, themselves, authentic in that context (see Widdowson, 2000). Third, students need to experience success while reading and engage with reading for extended periods of time. The criterion of authenticity becomes less important than the criteria of motivation and interest (Day & Bamford, 1998; Dornyei, 2001). Finally, a large amount of frustration-level reading, a common feature of authenticity in the reading classroom, can destroy motivation for reading, lead to negative self-esteem, and create poor environments for reading instruction.

Specific L2 Factors which May Influence the L2 Reader

A major issue for L2 reading research involves the different set of factors that influence L2 readers. L2 readers, first and foremost, do not have the same language resources as L1 readers at the outset of learning (see also Alderson, 2000; Grabe & Stoller, 2002, Urquhart & Weir, 1998). L2 readers have much lower levels of lexical, grammatical, and discourse knowledge at beginning stages of L2 reading than L1 readers do when they begin to read. In addition, L2 readers have much less overall exposure to L2 print (Day & Bamford, 1998); in contrast, L1 readers are consistently exposed to native language print from a very early age. L2 readers also vary considerably in their own L1 reading abilities, creating an added complexity.

Aside from linguistic differences, L2 readers often do not share all the social and cultural assumptions and knowledge bases that L1 readers use when reading in their own language. These contrasts include 1) differing socio-cultural backgrounds of L2 readers in comparison with the assumed audience of an L2 text, 2) differing kinds of text types and rhetorical patterns used in L2 tasks, and 3) differing assumptions about "how the world works" assumed by authors of L2 texts. Other cognitive factors can influence L2 readers in unique ways. Students often learn second languages for reasons that may be distinct from L1 literacy goals--to understand a new culture, to build more knowledge on an educational base that is already in place from L1 schooling, to go overseas, to have additional professional options, or to fulfill a seemingly irrelevant requirement. As a consequence, L2 students may have differing motivations for reading in the L2 than in the L1 (Dornyei, 2001). Moreover, L2 students work with cognitive and processing resources that involve two different languages, leading to various transfer phenomena. Working with two languages also implicates the use and control of two sets of word forms, text formats, and semantic concepts. Finally, working with two languages is likely to lead to greater metacognitive and metalinguistic awareness, particularly in contexts in which the L2 is learned well after L1 literacy skills have emerged.

These differences have at least four consequences. First, research in L2 reading cannot simply assume that results of research on L1 reading will apply in L2 contexts. Second, these differences suggest that L2 readers may employ cognitive resources in somewhat different ways from L1 readers, especially where there are clear differences between the L1 and the L2 (e.g., differing uses of phonological and morphological information from orthography while reading). Third, some of the observable differences between L1 and L2 reading may be due to proficiency limitations in the L2 (vocabulary, grammar, fluency, amount of exposure, etc.). Fourth, actual cognitive processes themselves may be somewhat different simply as a result of working with two languages (e.g., how words in the lexicon are stored and accessed; how transfer from the L1 impacts L2 reading).

Major Goals for Reading Instruction Supported by Recent Research: Implications and Applications

One of the major outcomes of changing views on reading research is the shifting array of instructional implications that arise for L2 reading instruction. While some of these implications have been well documented in earlier discussions of reading (e.g., focusing on general comprehension skills), the set of implications below reflect the more complex views of current reading research. Based on the research indicated in this article, there are at least 11 important implications for reading instruction and curricular development:

Unfortunately, proposing a set of implications for instruction does not ensure that actual instruction will lead to desired goals. So the task for teachers, curriculum planners, and materials developers is to move from implications to applications. Recognizing that teaching contexts vary by students, institutions, goals, proficiency levels, etc., it is the task of pedagogical researchers and teachers to determine how best to translate these eleven implications into effective classroom applications.

Conclusion

L2 reading instruction is a very complex undertaking, one that requires a considerable amount of experimentation and innovation. The research base discussed in this article is one informational source for curriculum planning (see also Kamil, et al., 2000; Pressley, 1998; Stanovich, 2000). A further source of information to complement the research base is teacher reflection and action research exploration in the classroom (Grabe & Stoller, 2002). A third base is institutional and student needs analyses. The latter two foundations for curriculum building require their own articles. For the present, I have outlined a way to understand reading and reading research that should offer useful options and alternatives for L2 reading instruction.

If there is one simple set of advice that can be drawn from this exploration of research and its implications, it is the following: Determine which aspects of reading instruction your students need the most help with, provide that help, provide students with many opportunities to read, and make sure that they read.

References

Alderson, C. (2000). Assessing reading. New York: Cambridge University Press.
Day, R. R., & Bamford, J. (1998). Extensive reading in the second language classroom. New York: Cambridge University Press.
Dornyei, Z. (2001). Motivation: Applied linguistics in action. New York: Longman.
Grabe, W. (2000). Reading research and its implications for reading assessment. In A. Kunnan (Ed.), Fairness and validation in language assessment (pp. 226-260). Cambridge: Cambridge University Press.
Grabe, W., & Stoller, F. (2002). Teaching and researching reading: Applied linguistics in action. New York: Longman.
Kintsch, W. (1998). Comprehension: A framework for cognition. New York: Cambridge University Press.
Hulstijn, J. (2001). Intentional and incidental second language vocabulary learning: A reappraisal of elaboration, rehearsal and automaticity. In P. Robinson (Ed.), Cognition and second language instruction (pp. 258-286). New York: Cambridge University Press.
Kamil, M., Mosenthal, P., Pearson, P. D., & Barr, R. (Eds.). (2000). Handbook of reading research: Volume III. Mahwah. NJ: L. Erlbaum.
Koda, K. (1996). L2 words recognition research: A critical review. Modern Language Journal, 80, 450-460.
Pressley, M. (1998). Reading instruction that works. New York: Guilford.
Segalowitz. N. (2000). Automaticity and attentional skill in fluent performance. In H. Riggenbach (Ed.), Perspectives on fluency (pp. 200-219). Ann Arbor, MI: University of Michigan Press.
Stanovich, K. (2000). Progress in understanding reading. New York: Guilford.
Thompson, G., & Nicholson, T. (Eds.). (1999). Learning to read. New York: Teacher's College.
Urquhart, S., & Weir, C. (1998). Reading in a second language: Process, product and practice. New York: Longman.
Widdowson,H. (2000). The mediating role of applied linguistics. Annual Review of Applied Linguistics, 20, 21-33.


All materials on this site are copyright © by JALT and their respective authors.
For more information on JALT, visit the JALT National Website