Educating Language-Minority Children (1998)

English-language learners in the United States are overrepresented among those performing poorly in school. An understanding of the cognitive challenges posed by learning to read and by acquiring new content knowledge, whether in a first or a second language, is a prerequisite to designing better instruction for these and indeed all children. Whereas the previous chapter focused primarily on acquisition of oral language skills, the focus in this chapter is on reading, writing, and subject matter knowledge. The emphasis is on research on the cognitive nature of the challenges inherent in learning to read or learning subjects such as math or history, and on the factors that facilitate success in learning. Most of this research has been conducted with monolingual native English-speaking children, but nonetheless is relevant to English-language learners. It should be noted that although this chapter includes some discussion of optimal instruction in the area of reading, most of the discussion regarding instruction is included in Chapter 7, on studies of school and classroom effectiveness.

There has been a vast amount of research related to literacy and literacy instruction. Here we can only provide examples of what has been

learned in the various domains of literacy development, focusing on concepts that are relatively well established for first-language reading and their potential relevance for understanding literacy development among English-language learners. We examine in turn prerequisites for the successful acquisition of reading skills, optimal early reading instruction, reading as a developmental process, and the comprehension of skilled readers.

It is clear that future successful readers typically arrive at school with a set of prior experiences and well-established skills conducive to literacy. The findings in this area are fairly consistent, though explanations of how those prerequisites function to foster literacy development are not. The key prerequisites include an understanding of the notion of literacy (both as a social process and as specific knowledge about letters, language, and symbolic systems that are prerequisites to full literacy); an abstract knowledge of the sound and structure of language (the ability to segment language into phonemic units, such as rhyming or focusing on similarities in sound rather than in meaning when grouping words); a certain level of vocabulary development; and conversational skills, such as the ability to adapt language to the needs of present or nonpresent listeners.

There is considerable controversy about the level of proficiency in a second language needed to support reading in that language. Wong Fillmore and Valadez (1986) argue that second-language reading for English-language learners should not be introduced until a fairly high level of second-language proficiency has been achieved. However, Anderson and Roit (1996), Gersten (1996), and others argue that instruction focused on second-language reading comprehension can be helpful to learners at all levels of second-language oral proficiency, even for those with learning disabilities (Klingner and Vaughn, 1996), and in fact that support of second-language reading comprehension can generate gains in second-language oral skills (see also Elley, 1981).

In general, positive correlations have been found between English second-language oral proficiency and English second-language reading ability, particularly at higher grade levels, but not equally across all first-language groups (Devine, 1987; see Fitzgerald, 1995, for a review). The mixed findings may well reflect differences in oral language proficiency measures used across the various studies and in conditions for literacy acquisition. For example, older, already literate second-language learners

acquiring English literacy through formal, foreign-language-type instruction may rely less on oral language as a route to English literacy than those acquiring their initial literacy skills in the second language.

Perhaps the most controversial area in reading research is the question of how best to teach initial reading—the whole-word method (Flesch, 1955), phonics/direct instruction methods, or whole-language methods (Chall, 1967, 1983; Adams, 1990).¹

While one can cite research findings in support of the value of certain of these practices over others, only recently has anyone officially sanctioned a mixed method of teaching reading—embedding direct instruction in component processes into meaningful, communicative, literate activities—that many experienced and successful teachers are in fact implementing in their classrooms (Adams and Bruck, 1995; Purcell-Gates, 1996).

With regard to reading instruction in a second language, there is remarkably little directly relevant research. Clearly one of the major intellectual stimuli to bilingual education programs has been the belief that initial reading instruction in a language not yet mastered orally to some reasonable level is too great a cognitive challenge for most learners. Studies of outcomes of bilingual programs, however, do not typically distinguish students who arrive at school already reading in their first language from those who learn to read only at school. The evidence that better academic outcomes characterize immigrant children who have had 2 to 3 years of initial schooling (and presumably literacy instruction) in their native countries (Collier and Thomas, 1989; Skutnabb-Kangas, 1979) is consistent with the claim that children should first learn to read in a language they already speak. However, it is clear that many children first learn to read in a second language without serious negative consequences. These include children who successfully go through early-immersion,

two-way, and English as a second language (ESL)-based programs in North America, as well as those in formerly colonial countries that have maintained the official language as the medium of instruction, immigrant children in Israel, children whose parents opt for elite international schools, and many others (see Christian, 1996; Feitelson, 1988).

What we know about early literacy acquisition suggests it is more likely than not to be successful under a wide variety of circumstances, but is nonetheless impacted by a long list of risk factors, including lack of explicit instruction in the local orthography, absence of the sort of background knowledge and skills acquired in highly literate environments, and unfamiliarity with the words one is trying to read. Exposure to any one of these and other risk factors may have no impact on literacy achievement, though the coincidence of several may lead to a greater likelihood of failure. The high literacy achievement of Spanish-speaking children in English-medium Success for All schools (Slavin and Yampolsky, 1992) that feature carefully designed direct literacy instruction suggests that even children from low-literacy homes can learn to read in a second language if the risk associated with poor instruction is eliminated.

There are rather different tasks and skills involved in reading at various points in the acquisition of skilled reading: learning about print versus nonprint, typically accomplished in the preschool years; learning to recognize and write letters; learning to decode words, which involves synthesizing phonological from graphemic sequences; reading relatively simple texts fluently; reading for comprehension texts that include new information and unknown lexical items; reading strategically, for specific information or purposes such as relaxation; and reading critically, to examine and compare the claims and arguments of different authors. The essential idea here is that the nature of reading skill needs to be defined somewhat differently at different points in its development, and thus that acquisition of prior skills does not always predict continued growth in reading ability; there are several points in development where novel skills need to be acquired.

The implications of this view for second-language learners are potentially enormous, as the task of learning to read in a second language is presumably quite different at different stages of first-language reading skill. Direct studies of the nature of what can be transferred from first- to second-language reading need to take into account not only the level of first-language reading, but also the level and content of the second-language material being read, as well as the nature of the orthographic,

linguistic, and rhetorical differences between the first and second languages.

Skilled readers are capable of reading with understanding in part because the component processes—letter recognition, word recognition, access to word meaning, syntactic parsing of the sentence—are fast and efficient (e.g., Adams, 1990). Those who have poor skills in word recognition can improve their comprehension by employing strategies such as reading the whole text for gist; self-monitoring for understanding; and using cues from titles, pictures, headings, and the like.

Explicit instruction in comprehension strategies such as prediction, summarization, and questioning—for example, the widely used "reciprocal teaching" (Palincsar and Brown, 1984) or Bereiter and Bird's (1985) think-aloud method—has been shown to be useful with poor first-language readers, and some evidence suggests it would also be useful with second-language readers who have comprehension difficulties (e.g., Barnett, 1989; Casanave, 1988; Cohen, 1990). Studies of the metacognitive strategies used by second-language readers of English reveal that such strategies are widely used (reviewed in Fitzgerald, 1995). The repertoire of those strategies includes some that may be specific to the second-language situation, such as using translation dictionaries or relying on information about cognates, but many are also typical of first-language readers as well, such as asking questions, predicting, and summarizing. However, some researchers have suggested that rather little attention is given to teaching or promoting comprehension strategies for English-language learners, even in the middle and later elementary grades when such instruction is important, because teachers tend to focus on word recognition and pronunciation (e.g., Gersten, 1996).

Skilled readers use syntactic information unconsciously to make the reading process more efficient, for example, by fixating on high-information items in the text (Rayner and Pollatsek, 1989). The fact that high-information items differ from language to language can lead to inefficient fixation patterns when reading in a second language (Bernhardt, 1987), thus perhaps disrupting the fluency that facilitates comprehension.

Skilled readers can tolerate a small proportion of unknown words in texts without disruption of comprehension and can even infer the meanings of those words from sufficiently rich contexts, but if the proportion of unknown words is too high, comprehension is disrupted. Word knowledge no doubt relates to reading comprehension both because encountering many unknown words slows processing and because lack of word

knowledge indicates absence of the relevant background knowledge that is crucial in reading texts of any complexity. Familiarity with content promotes reading comprehension when reading in either a second or a first language (Carrell, 1987; Johnson, 1981; see Fitzgerald, 1995, for a review), though knowledge of relevant background information may be less reliably indexed by second- than first-language vocabulary.

Comprehension is also supported by familiarity with larger structural patterns present in texts. Knowing that paragraphs have topic sentences on which other sentences are meant to elaborate, for example, aids the reader in integrating information across sentences. These macrostructures are culturally determined. For example, the writings of Michener, Allende, and Oe display wide variation in notions of plot and temporal sequence, of how much orientation is needed, and of how much interpretation should be supplied. In general, passages organized in a familiar structure are easier to comprehend and recall for second-language readers than those with a novel rhetorical structure (see Fitzgerald, 1995, for a review).

Considerable progress has been made over the last two decades in understanding the nature and processes of the learning and acquisition of knowledge of specific content information. This research has, for the most part, not concerned itself with issues of language per se, nor has it been incorporated into discussions about English-language learners.

Because the problem for the English-language learner has been considered as almost entirely language based, much of the research has focused on language acquisition issues. But the learning of school subject matter and work skills involves building intricate networks of concept relations, structuring and restructuring understandings, connecting them to other understandings, and practicing multiple skills in multiple environments. Therefore, more complex questions might fruitfully be asked about the nature of second-language students' learning, knowledge, and understanding of complex subject matter domains.

Discussion of complex questions of subject matter learning for English-language learners needs to be grounded in some assumptions about learning in general. The remainder of this section describes three assumptions drawn from cognitive analyses about school subject matter learning for primary-language content learning. These assumptions provide the context for much of the current research on school learning and apply to most students and most subject matter domains. First, we assume that different subjects have different core structures or epistemologies, thus making different demands on the learner. Second, we assume that there

are multiple forms or kinds of knowledge—for example, knowledge of facts and ideas, as well as knowledge of how to do something. Third, we assume that prior knowledge plays a significant role in learning, not only in terms of where to start, but also in terms of the actual meanings attached to new information. The discussion of these assumptions offers some examples and examines what a program of cognitive research that considered subject matter learning for English-language learners might look like.

Although learning, knowledge, and understanding differ across subject matter, these differences are embedded in larger general similarities. For example, understanding and learning about earth science or social studies requires the general ability to read, to construct meaning, and to understand and follow oral discussion in the language of instruction. It also requires general capabilities of inferencing, placing examples into overarching constructs, and building causal chains.²

Knowledge varies both across and within subject matter areas: it varies across because subjects have different arrangements of facts, concepts, notations, and patterns of reasoning; it varies within because some academic subjects have elaborate and importantly constraining notational systems (for example, algebraic and graphic systems).

The fundamental differences among subject areas necessitate highly differentiated systems of complex knowledge for both students and their teachers. While it is clear that at some level of abstraction, generalities across subject areas do exist, we believe these generalities are not sufficient to leapfrog the middle ground of differentiated knowledge. Further, a better understanding of this middle ground can enhance our understanding of the nature of both primary-language content learning and content learning in a second language.

In light of these differences among the various subject areas, certain disciplines may lend themselves more easily than others to the transfer of knowledge across languages, depending on the structure of knowledge within the domain, but the particular domains to which this would apply are not readily apparent. Because sophisticated knowledge in a given domain not only uses the terminology of that domain, but also builds upon

gradually developing concepts, learning such strands of content knowledge in one language and then shifting to another may be especially problematic.

Although there are substantial differences among subject matter areas, studies of English-language learners and their teachers seem to have ignored these distinctions, instead identifying a central problem facing these students as learning enough general language to enter English-only classrooms. We do not know what the advantages or complications are for English-language learners trying to learn the various disciplines themselves. However, we do suggest that it would be useful to learn how general language proficiencies interact with specific academic language proficiencies and with specific subject matter content.

Not only are there substantial differences among subject matter areas, but there are also different kinds of knowledge. One of the more common distinctions among types of knowledge is that between procedural knowledge, or knowledge of actions and skills, and declarative knowledge, or knowledge of concepts and principles (Chi and Ceci, 1987; Heibert, 1986; Lampert, 1986; Scribner, 1984). One task facing the student is to integrate these two types of knowledge.

Another distinction between types of knowledge is between knowledge of content and knowledge of that knowledge, referred to as metacognition. Brown (1980) discusses metacognition in terms of three features: knowing what you know and how well you know it, knowing what you need to know, and knowing the utility of active intervention. This self-awareness has been found to be a useful tool for learners across domains. Learners with such awareness are better able to organize the knowledge they have and identify the knowledge they need to acquire.

We do not have much information about the English-language learner with respect to multiple forms of knowledge. (See Chapter 7 for a review of studies that examine the effect of instruction in metacognitive skills on the subject matter learning of English-language learners.) However, issues of metacognition have been discussed for second-language learners in terms of the additive principle, which suggests these students have an advantage when learning new material.³ The argument that metacognitive

abilities facilitate learning by primary-language content learners lends support to the claim of the additive principle. Note, however, that in considering metacognition, the assumed advantage for second-language learners when learning new material has been focused strictly on linguistic awareness; the findings do not generalize to utility for learning particular subject matter knowledge.

The types and amount of knowledge available before encountering a new topic within a particular discipline affect how meaning is constructed. The knowledge structure can be thought of as nodes of information, such as concepts, that are linked to each other in particular ways depending on how and what information has been learned. Links between concepts can be acquired, reconstructed, or deconstructed, and particular learning outcomes are determined jointly by what was known before (the unique pattern of nodes and links) and the effects of instruction (additions to or rearrangements of that pattern).

The issue of prior knowledge can be considered one of depth, interconnectedness, and access. Depth of knowledge refers to the number of linked concepts a student has in a domain. In math, for example, students' depth of knowledge will influence their recognition of a problem, their sense of meaning associated with the problem, their ability to perform the appropriate mathematical operations, and their ability to recognize a reasonable answer. The extent to which concepts are interconnected reveals the coherence of a student's understanding of a particular domain. Finally, the existence of different kinds of knowledge poses a problem for both teaching and learning in that if the different types of knowledge are disconnected, they will be inert and unusable (Bereiter, 1984; Brown et al., 1983). A student may know what a long division problem is, but not know how to solve it. The development of deep, interconnected, generative knowledge instead of shallow, fragmented, inert knowledge needs to be a continuous process for both teachers and their students, with the interaction between the two forms of knowledge being explicitly taught.

Thus the depth, interconnectedness, and accessibility of prior knowledge all dramatically influence the processing of new information (Chi and Koeske, 1983; McKeown et al., 1992; Pearson et al., 1979). With respect to second-language learners, then, a number of questions arise. Under what conditions is content learning affected by the fact that a superordinate category and its instantiation (e.g., community and addition) are learned both tacitly and explicitly in one language, but are then to be used as a principle in a more complex instantiation in another language

(e.g., addition of algebraic polynomials)? How are "errors" that have a language base handled in a second language (e.g., in English, the confusion of "north" with "up" on a page versus in real space)? Naturally, the potential for interference in terms of access is also of concern—although this may be a vocabulary issue. A problem may arise if base examples are introduced at a young age in the child's first language (e.g., for social studies, notions of community, roles, freedom, and power) and are to be built upon in the second language at a later age (e.g., in learning about the French Revolution). Does this affect the second-language learner, and how?

At this point, we know next to nothing about how to answer these questions. We do not know, for example, whether (especially for the older new arrival) time should be taken to review existing knowledge that is available in the first language in a way that recontextualizes it in the second language, or whether the new knowledge (e.g., Algebra II) should simply be supported with back references to salient ideas known in the first language but now used in the second (e.g., Algebra I). The literature discussed here could be used to broaden the debate on content learning for English-language learners to address such issues.

Many of the findings regarding effective instruction and risk factors associated with reading for English-only students can be applied to English-language learners. An example is the benefit of a mixed method of teaching reading—embedding direct instruction of component processes into meaningful, communicative, literate activities. However, there are other important factors that must be taken into consideration in teaching English-language learners to read in English. For example, because there are rather different tasks and skills involved in reading at various points in the acquisition of skilled reading, the discussion of what can be transferred from first- to second-language reading needs to take into account not only the level of first-language reading, but also the level and content of the second-language material being read, as well as the nature of the orthographic, linguistic, and rhetorical differences between the first and second languages. Comprehension is supported by employing strategies such as reading the whole text for gist, self-monitoring for understanding, and using cues from titles, pictures, headings, and the like; by asking questions, predicting, and summarizing; by increasing word knowledge

using translation dictionaries or relying on information about cognates; and by gaining familiarity with larger structural patterns present in texts.

Three assumptions provide the context for much of the current research on school learning and apply to most students and most subject matter domains: (1) different subjects have different core structures or epistemologies, thus making different demands on the learner; (2) there are multiple forms or kinds of knowledge—for example, knowledge of facts and ideas, as well as knowledge of how to do something; and (3) prior knowledge plays a significant role in learning in terms not only of where to start, but also of the actual meanings attached to new information. Although most of this research has not been conducted with English-language learners, we can reasonably assume until proven otherwise that it applies to these students and points to the importance of understanding their acquisition of content knowledge.

Research is needed on language-literacy relationships. It is also needed on the nature of the relationship between first- and second-language literacy skill.

Research needs to investigate the optimal English literacy instruction for children of different ages, those with different native languages, those whose native language is not written, and those whose parents are not literate in English. Is there a single best way to teach all children to read, and if not, is there some way to identify child aptitudes so as to define optimal individualized instruction? What should be the role of writing in reading instruction, particularly for second-language learners? An important question to be addressed is whether literacy can be used as a route to language learning, and if so, under what circumstances and with what consequences.

There are four key research questions that address how those with limited English proficiency learn content. First, what are the effects of limited English proficiency on the acquisition of content knowledge at a fine-grained level? Specifically, what are the consequences of acquiring beginning-level content knowledge in one language and then switching languages for higher levels of the content domain? Second, what levels of English proficiency are prerequisite to the capacity to profit from content area instruction in English? Third, are there modifications to the language used by teachers that can make complex subject matters accessible even to second-language beginners? Fourth, how does the presence of a second language in the classroom affect the cognitive load for the content area teacher?