Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

The ability to read has always been of paramount interest to researchers and practicing teachers involved in L2/FL instruction. With the enhancements of electronic technologies and the growth of world globalization, the role of literacy (reading and writing) has even been strengthened, as it is a daily routine for many individuals to read different types of traditional and electronic texts for personal, educational and professional purposes. Understandably, whatever aspects of reading are considered, the main tendency is to seek some explanation for the complexity of the phenomenon of text comprehension. When defining reading, Grabe and Stoller (2002, p. 9) refer to a well-known simplistic view of reading as “the ability to draw meaning from the printed page and interpret this information appropriately”, only to show how incomplete it is. Conceptualizing reading, as they argue, has to address such issues as purposes for reading involving different combinations of skills and strategies, processes and knowledge bases in reading, the view of reading as a time-constrained cognitive process and its relationship with language proficiency. Bernhardt (2010, pp. 16–17) puts reading in a broader sociocultural perspective quoting the definition of literacy worked out by the Programme for International Student Assessment (PISA) by the Organisation for Economic Co-operation and Development (2006). Part of it states that: “[r]eading literacy is understanding, using and reflecting on written texts, in order to achieve one’s goals, to develop one’s knowledge and potential and to participate in society. This definition goes beyond the notion of reading literacy as decoding and literal comprehension: it implies that reading literacy involves understanding, using and reflecting on written information for a variety of purposes (…)”. Bernhardt draws attention to the significance of this definition in acknowledging the “intricacy of linguistic tools” used for comprehension and gaining information indispensable for one’s participation in modern social systems.

Likewise, the present paper adheres to the view that drawing solely on psycholinguistic or constructive-integrative models in explaining reading, or accounting for general comprehension skills in terms of language processing, prevents one from gaining a comprehensive view of the reading process characteristic of the L2 learner’s performance. Literacy has to be put in an adequate perspective in order to investigate it both as a linguistic activity and a social activity, and competently define how the two facets interact and constrain each other (cf. Wallace 1992; Colombi and Schleppegrell 2002; Grabe 2009; Bernhardt 2010). This is also the reason why defining the reading process in a generic way has its limitations. Reading comprehension comprises numerous activities that take place in diverse sociocultural contexts (including classroom settings) fulfilling learners’ purposes and motivations. They can entail a variety of reading materials, such as literary texts or multiple expository texts that contribute to the acquisition of the learner’s knowledge.

A significant development that has added new arguments against the universal approach to reading concerns the findings of research into different ways of discourse processing in selected disciplinary fields (e.g. history, science, mathematics). As a consequence, the concepts of disciplinary knowledge and disciplinary comprehension have received common acceptance, whereas the differentiation between content knowledge, domain knowledge and disciplinary knowledge has become more often embarked on (Shanahan 2009). The relationship between the three notions is not of a hierarchical type. While domain knowledge is defined as a field of study representative of some content knowledge, disciplinary knowledge is not just a part of content knowledge as it is constructed on the basis of some formal tradition, a selection of topics, and has some rhetorical and linguistic characteristics. This is the type of knowledge which is used in order to acquire some new content in the field (Shanahan 2009). Thus, considering L2/FL learners’ goals of knowledge acquisition, aiming at balancing content and language-focused dimensions in reading instruction, has to be seen as a most vital issue which awaits further theoretical and empirical examination.

It also needs to be remarked that in order to deal with L2 reading ability and comprehension processes properly, one has to consider the many issues in L1 and L2 reading theory, research and instruction that overlap. Despite differences between L1 and L2 contexts, numerous parallels between them can be found, in particular at higher levels of proficiency when analogous goals for L1 and L2 instruction are established. What is more, most learners taking up L2/FL courses have already developed their literacy skills in L1 to such an extent that they get transferred to L2 contexts activating cross-linguistic processing. In fact, explaining the complex L2 reading issues is not possible without referring to the advances reached by theory and research into native language reading development and use.

In the attempt at finding a dependable explanation for the importance of maintaining a steady focus both on language and content in L2/FL reading development and keeping the right balance between the two, the present paper discusses four major problems. First, the dominant text comprehension models are analyzed so as to pinpoint shifts in dealing with the most important aspects of reading comprehension of interest in the L2/FL context. Second, the discussion moves on to the interpretation of the relationship between text comprehension and L2 language acquisition processes. Third, some space is devoted to exploring schema theory, so influential for L2 reading instruction, with a view to presenting its criticism, as well as suggesting a better rendering of the problem of knowledge in the reading process. Finally, selected content-based reading frameworks are evaluated in light of their interpretation of language and content connections.

2 Conceptualizing Text Comprehension: Changing Models in Explaining L2 Reading

Despite the overwhelming amount of debate on the issues of reading comprehension processes both in L1 and L2, the key concepts providing an explanatory picture of the dominant views have not always been the same. Although text comprehension is of vital concern for L1/L2 educators, they have often been unable to deduce relevant information from the shifts in thinking brought about by new reading models. An interesting synthesis of the changing views on reading comes from Fox and Alexander (2009), who analyze three models of text comprehension: extraction-assembly, constructive-integrative, and transitional extensions models, by adopting four criteria: views of the text, typical texts utilized for instruction, and the reader’s activity and product. The common denominator for extraction-assembly models, popular in the 1960s and 1970s, is the assumption that texts provide static, simple and unambiguous messages coded in graphic symbols to develop reading skills. The information extracted and assembled by the reader is matched with the existing mental contents. In contrast, constructive-integrative models, dominant over the last 20 years, posit that although propositional network constitutes the textbase of a narrative or informational text, meaning is constructed on the basis on background knowledge, which enables the reader to arrive at a situation model of the text.

Interestingly, Fox and Alexander (2009, pp. 233–234) recognize the rise of a new trend in modeling reading, labeled transitional extensions, which sheds light on several significant aspects of text processing. First of all, text comprehension is treated as a connective activity, which is based not on one reader—one text processing, but on readers’ capabilities of navigating a hypertext, as well as looking for connections between multiple texts. Secondly, texts considered in this model are of different types, static or fluid, informational or argumentative, and, what is more, their writers are not invisible participants of the reading process but are evaluated for their credibility and accuracy. Thirdly, by allowing to create meaning across texts, the model highlights the importance of their content and knowledge developed on their basis. As a result, a profound role is played by readers’ interactive responsiveness to text comprehension, as well as by working out collaborative text interpretations. Although the name of this particular reading model is not commonly used by researchers, current publications on text comprehension discuss the problem of reading a vast variety of electronic texts and the need for developing the ability to navigate websites in search of the information required. Similarly, the problems of critical literacy, reading multiple texts (concerning one topic or domain-knowledge) and incorporating collaborative work on reading texts to replace silent reading have frequently been addressed (e.g. Klingner and Vaughn 2000).

However, the transitional extensions model representing a significant shift in modeling text comprehension has not been acknowledged in current L2 reading literature. After a long period of reliance on assembly-extraction models, constructive-integrative models seem to have become a major conceptualization of the reading process. Grabe (2009), for instance, assumes that the L2 reading process functions due to the interaction of components at two major levels: (1) lower-level processing, which covers word recognition, syntactic parsing and semantic-proposition encoding, and (2) higher-level comprehension processing, which helps the reader construct a coherent situation model of a text in memory. Efficient functioning of the reading process is also possible due to such mechanisms as strategies, goals, inferencing, background knowledge, and comprehension monitoring, all of which have attracted considerable attention of L2 reading specialists so far.

Many of the ideas associated with constructive-integrative reading models stem from Kintsch’s (1998, 2005) premise that in order to understand a sequence of clauses the reader applies his/her knowledge from long-term memory so as to integrate all the information in working memory. The comprehension of a text covers four levels: the surface code, the propositional textbase (i.e. meaning of the text expressed through logically organized propositions), the situation model (content of the text referring to the real world) and text genres, and is constantly modified in the process of reading (the connectionist view). All text perception and comprehension processes are based on a spreading activation network, and, in the case of a comprehension failure, the reader’s strategic behavior provides some repair for it (Kintsch 2005).

It is worth noting that the currently supported postmodern perspective in conceptualizing reading comprehension gives priority to the individual reader’s interpretation of the text, which means the acceptance of multiple text interpretations. This phenomenon is explained by the landscape model (Linderholm et al. 2004); as reading is based on multiple cognitive processing, particular elements in the text are activated in a fluctuating way (cohort activation), which is possible owing to the reader’s background knowledge and purpose. The standards of coherence readers impose on reading contributes to a high level of comprehension variation across readers and situations.

Even this very brief overview of basic models of text comprehension reveals that it is the concept of content that is central to explaining how the reading process functions. In terms of Kintsch’s (1998) model, one can talk about the content/meaning of textbase that leads to creating—with the help of background knowledge—a new quality content of a situation model of the text. Similarly, text content, as mentioned in the transitional extensions model, when utilized by the reader, contributes to the development of his/her topic-based knowledge. An important role is also assigned to reading purposes. This is not to say that the purposefulness of reading is a new concept—rather it has generated a revived interest among reading specialists and led to its fruitful reconsideration. Especially with reference to the academic level, the significance of reading to learn as well as to evaluate, critique and use the information gained while reading in the future has been underlined. With reading perceived as a tool for knowledge expansion, the concepts of information, content and knowledge, rather than globally described meaning or general comprehension, are associated with reading different genres of texts in socioculturally different contexts (Grabe 2009).

In developing her L2 reading model, Bernhardt (2005, 2010) recognizes the superiority of a compensatory model for defining the specificity of second language learning context. She refers to Stanovich’s (1980) compensatory model, which assumes that particular levels of processing in reading help the reader compensate for deficits occurring in other aspects of the reading process. Bernhardt proposes a model which comprises three sets of factors that affect L2 readers’ performance, and provides the percentages of variance they cause: (1) L1 literacy (alphabetics, vocabulary, text structure, beliefs about word and sentence configuration, etc.)—20 %, (2) L2 language knowledge (grammatical form, vocabulary knowledge, cognates, L1/L2 linguistic distance, etc.)—30 %, (3) unexplained variance (comprehension strategies, engagement, content and domain knowledge, interest, motivation, etc.)—50 %. Each set of variables can provide a fair amount of support for another if considerable flexibility is left to those processes. What is more, the role of factors involved in all the knowledge sources gets intensified with the development of proficiency. It is interesting to observe that the concept of content and domain knowledge has been inserted into the model as part of quite a large component whose variance is not specified, while at the same time the concept of background knowledge (reminiscent of schema theory, to be critically evaluated in Sect. 4) has been left out. The model surely points to a considerable significance of language-based aspects, some of which (cognates, L1-L2 distance) fall into the range of cross-linguistic properties, which is missing in general text comprehension models.

3 Towards Understanding the Relationship Between Reading Comprehension and Language Acquisition

In the last quarter of the twentieth century the area of ELT witnessed the emergence of two standpoints that not only found a steady place in heated debates in the mainstream methodology, but whose attempt at shedding new light on the role of information/content processing in language acquisition and use has not passed unnoticed. They were the Communicative Approach and Krashen’s Comprehensible Input Theory. The growth of the Communicative Approach was associated with the provision of a better theoretical background for a general description of language use as related to extralinguistic contexts and real-life situations. Due to the concept of speech events, language use started to be analyzed as socially and culturally determined activity. Dissatisfying, however, was the finding that meaningful communication does not guarantee that language input can be turned into intake and result in learning formal language features such as elements of morphology or syntax (Han and D’Angelo 2010). What is more, due to the main concern with oral interaction, literacy issues were not of particular interest to the advocates of communicative methodology. The problem of language comprehension as linked to text comprehension and L2 reading development was forcefully tackled by Krashen’s Comprehensible Input Theory, which could boast of considerable support from researchers and practitioners before its criticism was sounded. As remarked by Schleppegrell and Colombi (2002, p. 13), very critical of the impact of Krashen’s theory on language teaching, its “invisible pedagogy”—immersing learners in meaningful learning from input—suggested ignoring or even advised against explicit teaching of language form. Undoubtedly, the controversies about Krashen’s theory have given rise to a reconsideration of some basic issues in understanding L2 reading comprehension processes, and language acquisition.

Much scholarly discussion of the relationship between comprehension and language acquisition that revolved around Krashen’s Input Hypothesis, complemented with the Reading Hypothesis, concerned the direct claim that reading is a primary source of comprehensible input stimulating incidental acquisition of target language elements (Grabe and Stoller 1997). Despite its simplicity, attractiveness and the many arguments provided in support of it (cf. Mc Laughlin 1987), the hypothesis was found to be partially false due to the lack of congruity between comprehension and acquisition (Han and D’Angelo 2010). Some researchers (e.g. Sharwood Smith 1986) express the view that decoding text content does not have to result in intake, understood as code learning. As explained by Gass and Selinker (2008), even though some comprehension is necessary for language to be acquired at the syntactic level, comprehension is a very broad phenomenon involving top-down processing of the learner’s world knowledge, whereas contextual clues, which constitute the basis for inferential processes and attending to language form, require bottom-up processing. It is important to note that researchers examining the relationship between reading comprehension and language acquisition have tended to dissociate morphosyntactic and lexical levels of language processing.

Powerful argumentation in support of the distinct nature of comprehension and morphosyntax processing was provided by Lee (1998), who found no significant correlation between the two; it has been proved that L2 learners are able to understand texts with the help of extralinguistic information. Thus, he claims, language form can be acquired only when special attention is paid to it. One way of doing so is by means of enhancing language input with activities that help learners match meaning and form (cf. Izumi 2002; Yoshimura 2006). Similar conclusions were reached by Van Patten (2003, quoted by Young and Nokuma 2009) in his empirical studies, which revealed L2 learners’ preference for processing semantic information. His subjects tended to process content words in the input first, and, even in the case of processing morphological elements, they gave priority to the more meaningful ones. Van Patten recommends that special grammar instruction emphasizing form-meaning connection be offered to deal with language features that might cause misunderstanding in reading instruction.

A much more difficult problem concerns the relationship between text comprehension and incidental vocabulary acquisition, also described as vocabulary learning through reading. The Reading Hypothesis, an inherent component of Krashen’s Comprehensible Input Hypothesis, maintains that it is the comprehensibility of the text, that is recognizing the meaning of crucial words in utterances, that makes language acquisition effective (Krashen and Terrell 1983, p. 155). Since the 1980s incidental vocabulary acquisition while reading has become the key interest of a large body of research carried out by reading specialists both in L1 and L2. Yet, despite advances in the scholarship, the nature of the interdependence between reading comprehension and incidental vocabulary acquisition has not been adequately accounted for (cf. Hudson 2007; Pulido 2007; Hedgcook and Ferris 2009; Bernhardt 2010). An attempt to probe into the relationship between text comprehension results and vocabulary gains while reading in L2 was made by Chodkiewicz (2001). A large scale empirical study involving 222 advanced Polish learners of English, which showed an average 13.8 % incidental vocabulary gain as measured by a receptive test, did not reveal any significant systematic correlation between reading comprehension and vocabulary acquisition. Although different types of reading tasks helped the subjects get higher comprehension scores, no proportionate growth in incidental vocabulary learning was observed. On the contrary, the subjects reached higher vocabulary gains when working under the no intervention treatment (read only), not while performing tasks that required deeper purposeful processing of content and enhanced their comprehension. The findings of the study undoubtedly point to a very complex interrelationship between vocabulary learning and reading comprehension processes, and to the lack of a causal link between comprehension and vocabulary learning advocated by Krashen’s theory.

As already shown, vocabulary learning while normal reading is not an easy issue to examine. Empirical studies have proved that one of the reasons for difficulties in investigating incidental vocabulary acquisition is the inability to control for the sheer number of variables that influence it. Suffice it to mention the learner’s grade/age, level of reading proficiency, purpose for reading, the assessment methods used (few tests are sensitive to moderate gains in vocabulary learning), or text difficulty and frequency of exposure (Chodkiewicz 2000; Hedgcook and Ferris 2009). Attempts at the elucidation of the nature of incidental vocabulary learning as a language acquisition process, as well as the role of attention and awareness in it, have triggered controversy at the theoretical level. Despite this, some agreement on how L2 vocabulary acquisition functions during normal reading has been reached. The first step in learning an unfamiliar word while reading entails noticing it, then its meaning is to be inferred by using all the available contextual clues and attending to the connection between the word’s form and meaning. Finally, a particular item enters the learner’s lexicon (Pulido 2007).

In order to notice unfamiliar words and to process available contextual clues to infer words’ meanings while reading L2 learners have to deploy appropriate strategies. So far a great deal of effort has been devoted to classifying different types of contexts (e.g. general vs. global, external vs. internal) and even to offering detailed taxonomies of contextual clues such as those utilizing familiar expressions, synonyms, comparison or contrast, association, which can be pedagogically useful. However, a fundamental question relating to the quality of contextual clues and efficiency in using them has to be posed. The quality of contextual clues can differ from rich to poor, and, what is more, some clues can prove to be misdirective, while in some other cases contextual clues may be unavailable whatsoever (Chodkiewicz 2000). As regards the efficiency of using contextual information, Stanovich’s (1991) integrative-compensatory model of reading holds that good readers do not draw extensively on contextual information because they recognize words in a context-free manner. On the contrary, it is bad readers who tend to use contextual clues frequently to make up for their poor word recognition skills. The compensatory activation of contextual clues while reading automatically diminishes the quality of text comprehension and slows down the reading process.

On balance, as empirically documented, whereas no substantial gains in vocabulary learning from reading can be expected, reading passages are an important input source of vocabulary to be naturally acquired both in L1 and L2, with extensive reading being particularly recommended for L2 settings. Yet, it has to be noted that both the breadth and depth of vocabulary knowledge acquired in such a way is limited, and the effectiveness of the process depends on the many factors mentioned above, including the amount of reading done and the availability of lexical items that carry a significant meaning for the learner (Han and D’Angelo 2010). At the same time, numerous L2 reading studies imply that due to their limitations word meaning inferring strategies should be used with caution. Obvious to any practicing reading teacher is the advantage of employing a range of other well-known explicit strategies to help L2 learners in learning and retaining the target vocabulary, such as providing L1 equivalents, L2 definitions, or glosses (cf. Chodkiewicz 2000; Hedgcook and Ferris 2009).

Considering the characteristics of the meaning-based approach to reading, Han and D’Angelo (2010) recommend that Sharwood Smith’s (1986) dual relevance theory of language acquisition should constitute the basis for developing reading instruction. This means that reading for comprehension (semantic processing) and reading for syntactic processing should be kept apart. To adopt such an approach, teachers will first concentrate on interpreting the meaning of texts with the goal of developing reading efficiency, and then will continue to explore some grammatical features of the text, emphasizing meaning-form connections. At the early stages of reading instruction, one kind of processing can be focused upon at a time. Pedagogical procedures should draw on a combination of strategies helpful in making selected aspects of the input text more salient and ready to be noticed, such as enhancement strategies (e.g. capitalizing, underlying etc.), explicit grammar instruction, or so-called narrow reading, based on reading multiple texts as a way of increasing content-familiarity, recurrence of vocabulary/grammar structures and their saliency (Krashen 1981, quoted by Han and D’Angelo 2010).

4 Knowledge Bases for Reading: Making Up for the Weaknesses of Schema Theory

In the 1980s and 1990s some new insights into understanding how reading comprehension functions were provided by schema-theory due to the consideration of the reader’s general world knowledge. The prime goal of the theory was the explanation of the interrelationship between language, comprehension and the reader’s topic/world knowledge in text processing. In fact, it was the degree of topic familiarity that was assumed to determine the effectiveness of the reading process and contribute to the reader’s knowledge, represented by schemata or conceptual frameworks. The core premise was that new information is accommodated by the reader as a result of its integration with background knowledge—the schemata of world knowledge already accumulated in his/her mind. The reader’s knowledge of a given topic was found to be a better predictor of text comprehension than any other measure of reading achievement, and new language aspects appearing in the text were to be linked to the concepts of the reader’s background knowledge (Johnson and Pearson 1978).

Due to growing interest in schema theory both on the theoretical and practical grounds, the notion of background knowledge became a focal point. Bernhardt (1991), for instance, took a more elaborate view distinguishing between: (1) local knowledge—used by individuals for everyday purposes, (2) domain-specific knowledge—gained through education and by professionals who reach expertise, and (3) culture-specific knowledge. Moreover, a special kind of linkage was identified between the linguistic level and background knowledge, whose organization in long-term memory depended on the use of words and phrases corresponding to particular topic schemata. While the details of the morphosyntactic system were not taken into consideration, content vocabulary was found essential for understanding a text. Vocabulary work accompanying a reading task was to be topic-oriented and refer to the corresponding schemata (Nagy 1988). Some research studies were carried out to establish how the knowledge of content versus formal schemata affected the L2 reading comprehension process (Carrell 1987).

However, with the emergence of the construction-integrative models of reading, the role of background knowledge as explained by schema theory was cast into doubt. The main criticism concerned the fact that the theory regarded background knowledge as static, pre-existing extratextual knowledge stored in the reader’s long term memory, employed in a top-down way, whereas construction-integrative models treat it as belonging to text processing itself, and helpful in creating hierarchical patterns for organizing and upgrading the information gained while reading (Kintsch 2005). The information generated from multiple sources in the text (graphic, syntactic, semantic, rhetorical, pragmatic and thematic) is processed in the reader’s working memory and only then is it integrated into his/her knowledge. As demonstrated by the findings of empirical research, unfamiliar passages can be successfully read both by L1 and L2 readers despite the lack of background knowledge, provided they can establish meaning relations for a given text and create a coherent textbase for it. What causes the greatest reading problems for L2 readers is low efficiency in processing lower-level information due to their less fluent lexical and syntactic processing skills or insufficient sociocultural background, which makes it impossible for them to encode all the relevant text properties (Nassaji 2007).

It is important to note that the construction-integrative reading model has separated linguistic knowledge from background (domain-specific) knowledge to treat them as two components contributing differently to reading comprehension (Nassaji 2007): the former affecting the lower text-based level of lexical and syntactic processing, and the latter functioning at a higher level processing of semantic and conceptual information. Thus the acquisition of knowledge as a result of reading, is assumed to take place due to the integration of content from higher level processing of a particular text and the reader’s conceptual/prior knowledge.

5 In Search of Principles for Combining Language and Content in Content-Based Teaching

The exploration of the role of language and content in understanding L2 reading development would not be complete without a brief consideration of the ideas that have evolved within content-based teaching. Immersion courses introduced in the 1960s in Canada, the many alternative versions offered in the USA, most often labeled Content-Based (ESL) Instruction or Content-Based Language Learning/Teaching, and Content and Language Integrated Learning (CLIL), officially introduced in 1995 by the Council of Europe, have led to raising the question to what extent content should play a role in the process of language learning/teaching so as to make language instruction ‘contextualized’ and meaningful (e.g. Brinton et al. 1997; Larsen-Freeman and Freeman 2008; Wesche and Skehan 2002). As a consequence, establishing some common ground for the possibility of balancing the two approaches, that is focusing on language forms and structures and processing informational content of selected materials, still leaving some language aspects to incidental acquisition has started to be sought.

Principled integration of language and content has become the goal of numerous CBI and CLIL paradigms, in which non-language content (basically school or academic subjects) is to be integrated with language work focused on making up for L2 learners’ language deficiencies. Reading, representative of literacy skills, has a dominant role to play due to the natural dependence on a variety of written materials selected for content-based courses. Comprehending a text and interpreting its content is a basis for acquiring subject knowledge. It is to be looked upon as language input processed at grammatical, semantic, lexical and discourse levels, and a representative of a particular genre. This section sets out to analyze selected models developed with a view to finding a rationale for integrating language and content to achieve the targeted pedagogic purposes.

It is interesting to point out that one of the earliest ideas of integrating language and content was promoted in Poland by Marton (1978). Although motivated to a large extent by the practical need to make EFL teaching more attractive through textbook contents, the decision to introduce short texts based on knowledge related to school subjects was accounted for with psychological argumentation. Adding some new content suitable for learners’ life space was found important for their current needs and aspirations and conducive to their cognitive development. It was argued that EFL learners would appreciate dealing with aspects of disciplinary/general knowledge of biology, geography, physics, or film, music and travelling, etc., most of it being of concern to other school subjects. A series of EFL textbooks for a four-year long course for secondary school learners was supplied with short texts to be presented by teachers in the form of mini-lectures (Szkutnik and Marton 1977). The language of those texts was described as valuable for taking a deeper approach to language use through the categorization of concepts and relations between them (Szkutnik 1979).

An influential language-content integration framework designed by Mohan (1986, 2001) describes its theoretical underpinnings as based on Halliday’s (1985) functional theory of language, which defines the semantic potential of language and its social function. Language is not approached primarily as language acquisition but as a medium of learning content/subject matter and culture; that is why its use stretches beyond literacy areas and its basic function is to organize social practice. It is the act of learning that becomes a linguistic process. Defined at the level of discourse by such grammatical categories as classification, rules or evaluation, language can be rendered in a form of description, time, sequence and choice in practice (Mohan 2001). As noted by Colombi and Schleppegrell (2002), written texts, indispensable for the development of advanced literacy skills, characterized by grammar metaphor and a high lexical density, involve learners in an effective use of lexicogrammatical patterns through a range of different tasks that may also require a conscious study of language elements.

The potential of content learning is of central concern to the Cognitive Academic Language Learning Approach, intended for ESL learners, described as limited language proficiency learners, who need guidance in acquiring academic language competence. Working on selected content of school subjects, ESL learners develop their academic language skills with the help of metacognitive, cognitive and socio-affective strategies (O’Malley and Chamot 1990). Yet, their responses to content processing tasks are to be primarily evaluated in terms of content answers, although becoming familiarized with the properties of academic language, as different from the use of informal language in speech, is also a major objective to accomplish. Teacher assistance is indispensable also due to the fact that learners perform numerous context-reduced academic tasks whose difficulty is increased by the cognitive demands of learning new subject knowledge. The principles of CALLA-based courses have clearly been established in such a way that both content and language aspects can be efficiently benefited from at the course and classroom levels.

A turn towards a more precise rendering of the interdependence between the linguistic component and content teaching was made by Snow et al.(1989, p. 201) in their model for “a conceptual framework for the integration of language and content teaching in second and foreign language classrooms”. The researchers acted upon the premise that combining language and content can be beneficial as long as the linguistic aims are divided into obligatory and compatible ones. Language defined as obligatory for particular content comprises structural items (nouns, verbs, rhetorical devices), functional elements (narration, information, persuasion), as well as strategies to be employed in studying a particular text. On the other hand, compatible language does not have to refer to the aspects of content in question, although the context created by a particular content can be conducive to teaching selected language points. Despite it sounding innovative and appealing, Ellis (2002) finds such a solution difficult to implement, particularly to be followed throughout a multi-part language course. Instead he suggests that content and language components be separated from each other in such a way that the course starts with the communicative component emphasizing content and relying on incidental language acquisition, and only then is the direct study of the language code introduced.

Almost twenty years later, despite Ellis’s (2002) criticism of Snow et al. (1989), Gajo (2007) incorporates the concepts of obligatory and compatible language into his comprehensive CLIL model referring to L2, or even plurilinguistic contexts. The researcher takes an in-depth look at the integration of language and content as dependent on identifying different types of knowledge, which entails both the linguistic and the subject paradigms. By participating in the learning process—the discourse of pedagogic tasks, L2 learners get involved in complex relationships with language knowledge, which is remediated through discourse opacity (metalinguistic activities), and subject knowledge, which is mediated through discourse density (categorization process). Due to the relationship between linguistic and subject knowledge paradigms, the knowledge coming from the two sources can be negotiated by learning at the discourse level, through classroom interaction. Opaque in its nature, knowledge needs negotiation—some information can be potentially controversial, biased politically or culturally. Plurilinguistic competence of language learners is helpful in problematizing knowledge since the elements of linguistic knowledge and subject knowledge get interwoven, particularly at the level of lexis and discourse.

The development of knowledge can be described by means of three parallel categories in (1) language in language paradigm: content-obligatory language, content-compatible language and content-autonomous language, as well as in (2) language as embedded into the negotiated target content: content-embedded language (linguistic knowledge indispensable for communicating, establishing subject knowledge and subject paradigm), content-useful language (indispensable for performing a task or establishing or extending knowledge), and content-peripheral language (enhancing general links between language and subject knowledge) (Gajo 2007, p. 570). Gajo’s framework looks very elaborate, yet, it has to be treated as an attempt to implement some current ideas in explaining the structure and development of disciplinary knowledge. Of prime importance is also the emphasis on the need for a better conceptualization of the relation between content-language aspects and communicative competence—the key goal pursued in many L2 language classes.

Another recent model, called ‘4 cs’ (Coyle 2007), proposes that in view of language functioning as a communication tool, the learning process should be perceived as comprising content and cognition, as well as communication and culture. What is most crucial is that we move on from the use of language as a linguistic form to using language as representative of function and culture. The language-content relationship is conceptualized as consisting of three dimensions in language functioning: (1) language of learning—using a particular kind of language to aid content acquisition, (2) language for learning—language to be acquired by learners as well as a way of achieving the goal (e.g. awareness-raising, strategy training, language contextualization, checking understanding, etc.), (3) language through learning—activating language learning and thinking through different sociocultural contexts. Coyle’s model points to the multidimensional use of language, which is significant for educational settings where content acquisition is one of the main objectives. Her belief that knowledge is created through the activity of a community, gives support to the idea of communal constructivism.

In contrast, the Connections Model (Bigelow et al. 2006) goes in line with O’Malley and Chamot’s (1990) thinking already discussed. The model generally takes a similar path in the conceptualization of language-content relationship by introducing the same components as O’Malley and Chamot did, yet they are seen to function somewhat differently. The task for the teacher is to integrate contents (e.g. within a theme-based unit) with language functions depicted through reading and writing, as well as the organization of text structure with the help of strategies. The teacher’s decision as to the choice of language functions and text structures to work on can be based on learners’ problems, errors or their tendency to avoid some language aspects.

To sum up, some constructive suggestions for integrating language and content developed within content-based instruction, briefly dealt with in this section, have further shown that a rationale for a careful consideration of this issue is still much needed.

6 Conclusions

Despite limitations of space, the paper has made an attempt to demonstrate the importance of the conceptualization of language and content relationship in reading, and how important it is to assign a proper place to the two components in discussing L2 reading development issues. By language, with varying consistency though, researchers mean grammar, vocabulary, rhetorical structure and genre, as well as pragmatics. The term content is generally referred to text-based processing, but also—what is now more often the case—to knowledge-based reading processes and outcomes, which are closely associated with text meaning and comprehension. Explaining the interdependence between language and content in the process of text comprehension and their relationship to other variables involved in reading (e.g. individual learner differences interests, motivation, or text type) has proved to be equally problematic.

As indicated in this paper, presenting an overall picture of L2/FL reading development issues is a challenging task as different fields of scholarly discussion have to be examined in search of current information concerning text comprehension models, language acquisition and reading comprehension relationships, knowledge-based processing in reading, as well as conceptual frameworks for language-content integration within a range of CBI/CLIL approaches. The inadequacy of the interpretation of the role of background knowledge stemming from the schema theory, for example, could only be recognized, when a competing view of knowledge processing in reading was expounded by the constructive-integrative reading model. Following the growth of sociocultural theory, the consideration of the role of culture and knowledge in reading-based educational activities has helped to reshape the form and function focused view of reading.

As for a principled view of a dual focus on language and content in reading instruction, no consensus has been reached so far. On the one hand, research into reading and language acquisition points to the need of separation between morphosyntax and comprehension for teaching purposes; on the other hand, Coyle’s (2007) model, for example, seeks to establish a functional relationship between linguistic and content paradigms involved in classroom discourse, without the exclusion of some explicit consideration of language properties. As argued above, the dual focus on language and content will primarily call for some way of counterbalancing the use of the form-based and content-based facets in reading instruction so that the social side of learning is taken care of. On balance, to be truly updated on L2 reading development/instruction issues one has to be observant of the shifts in views on the reading process, constantly undergoing change and cross-fertilization of ideas coming from different information sources.