The role of grammar in the writing curriculum: A review of the literature

I Introduction

For most Anglophone countries, the history of grammar teaching over the past 50 years is one of contestation, debate and dissent: and 50 years on we are no closer to reaching a consensus about the role of grammar in the English/Language Arts curriculum. The debate has been described through the metaphor of battle and grammar wars (Kamler, 1995; Locke, 2005), frequently pitting educational professionals against politicians, but also pitting one professional against another. At the heart of the debate are differing perspectives on the value of grammar for the language learner and opposing views of what educational benefits learning grammar may or may not accrue. The debate has been well rehearsed elsewhere (Braddock et al., 1963; Cameron and Bourne, 1988; Christie, 2004; Hartwell, 1985; Hudson and Walmsley, 2005; Locke, 2009; Myhill and Jones, 2011; QCA, 1998), and only a brief overview will be offered here in order to frame the review that follows.

The Dartmouth Conference in the USA in 1966 signalled a turning point in educational thinking about grammar in the curriculum. Funded by the Carnegie Endowment, and organized by the Modern Language Association (MLA) and the National Conference of Teachers of English (NCTE), the Dartmouth Conference brought together teachers and educational researchers from the UK and the USA to discuss the grammar issue. The conference was prompted by growing dissatisfaction with classroom practice in grammar teaching, which was largely characterized by drills and exercises in labelling and identifying word classes and syntactical structures, and which to many education professionals had no educational relevance and no impact on language development. As Muller noted, the general consensus was that grammar teaching was ‘a waste of time’ (1967: 68). As a consequence of the Dartmouth Conference, many educational jurisdictions in the USA, the UK, Australia, New Zealand and Canada moved to exclude formal grammar teaching from the English curriculum, although, of course, there will always have been those teachers who continued to teach grammar, despite changed policy mandates.

The widespread abandoning of grammar teaching following the Dartmouth Conference activated two further strands in the grammar debate. First, whilst curriculum policy might have eschewed the teaching of grammar, an underlying sense of barely disguised outrage at its abandonment has surfaced repeatedly over the years, often from politicians or the general public. This strand of discourse tends to have been preoccupied with language standards and a pervading view that lack of attention to grammar was causing falling standards in language users (for example, Macdonald, 1995; Paterson, 2010; Truss, 2003; Uttley, 2008). It positions grammar as a form of ‘verbal hygiene’ (Cameron, 1995), which assumes a moral significance and acts as guardian to the moral standards of the nation (Cameron, 1994; Pullman, 2005). Second, the abandoning of grammar has fostered a strand of thinking about the role of knowledge about language in the curriculum. Here linguists have argued for the place of grammar knowledge, including its associated metalanguage, as a worthy and relevant body of knowledge in its own right (Hudson, 2004). At the same time, a broader view of knowledge about language places the emphasis on more inductive approaches, which investigate language in social contexts (for example, Barton, 1999; Carter, 1990; Denham and Lobeck, 2005; Keith, 1990; NATE, 1997).

This article is framed by this debate and by recent international developments that appear to be re-introducing grammar to the language curriculum. Through a systematic review of the literature, the article will provide a critical outline of research into the role of grammar in the English or Language Arts curriculum.

II Empirical studies on the teaching of grammar

Although the issue of the value of grammar teaching has often been framed in terms of the relevance of explicit and implicit grammatical knowledge, it is also true that one strong line of argument has been located in discussions of the impact of teaching grammar on students’ language use, particularly writing. In other words, does explicit knowledge of grammar support writing development and attainment in writing, or as Elbow argued, ‘nothing helps your writing so much as ignoring grammar’ (Elbow, 1981).

Over the past 50 years, there have been a number of research reviews or meta-analyses addressing the effect of grammar teaching on students’ learning (Andrews, 2005; Andrews et al., 2006; Braddock et al., 1963; Hillocks, 1984; Wyse, 2001), all of which have concluded that ‘the teaching of school grammar has little or no effect on students’ (Hillocks and Smith, 1991) or even that it has a ‘harmful effect on the improvement of writing’ (Braddock et al., 1963: 37). Indeed, Graham and Perin’s (2007) meta-analysis investigating effective strategies for teaching writing found a negative effect for ‘the explicit and systematic teaching of the parts of speech and structure of sentences’ and the authors argue that this finding challenges ‘some educators’ enthusiasm for traditional grammar instruction as a focus of writing instruction for adolescents’ (2007: 21). A review by the Evidence for Policy and Practice Information and Co-ordinating Centre (EPPI, 2004) concluded that it could find no evidence for any benefit of grammar teaching on writing quality, although the authors did also note that ‘the quality of seven out of ten of the primary studies included in the in-depth review on the teaching of syntax is a limitation on the review, as is the lack of recent research’ (Andrews et al., 2006: 51).

In the end, the EPPI review draws its conclusions based on the analysis of just three studies: Bateman and Zidonis (1966), Elley et al. (1975, 1979), and Fogel and Ehri (2000). It is worth looking at these three studies more closely. The Bateman and Zidonis study, in the USA, focused on sentence production in writing and investigated whether high school students could ‘learn to apply the transformational rules of a generative grammar in their writing?’ (1966: 3). The sample size was rather small: 50 students assigned to either the control or intervention, of whom only 41 were still in the sample at the end of the study. The intervention group were taught explicitly about syntactical structures for sentences, and written compositions from the two groups were analysed using an analytical tool that captured the generative transformational grammar rules of sentence construction and allowed them to compare development pre-treatment and post-treatment. An outcome of the analysis was a structural complexity score for each piece of writing (1966: 8). The statistical results were not conclusive, but the authors suggest that they did indicate a relationship between ‘a knowledge of generative grammar and an ability to produce well-formed sentences of greater structural complexity’ (1966: 39). In contrast to this small-scale study, the Elley et al. study is arguably the most robust research conducted in this field. They conducted a three-year longitudinal study in New Zealand, involving a sample of 250 secondary-aged students of average ability with eight classes, matched for ability. There were three treatment groups:

They found that there were no significant differences in the quality of writing produced post-test, although the transformational grammar group did improve in their sentence-combining ability.

The Fogel and Ehri study (2000) was different from the previous two studies as it took as its starting point an identified writing problem, the use of Black English Vernacular (BEV) in the writing of low-attaining ethnic minority students in primary school. The study set out to ‘examine how to structure dialect instruction so that it is effective in teaching SE [Standard English] forms to students who use BEV in their writing’ (2000: 215). An experimental design was adopted with three treatment groups:

According to the authors, treatments 1 and 2 mirror typical classroom practice. The results indicated a statistically significant positive effect for group 3, leading the authors to conclude that the guided practice had ‘clarified for students the link between features in their own non-standard writing and features in SE’ (2000: 231). However, because of the small sample size, the EPPI review does not rate this as strong evidence for a positive benefit.

The EPPI review noted the dearth of research in this area, but a much more recent study, conducted in primary schools in the UK by Sheard et al. (2012), would appear to confirm its conclusions. They investigated whether teaching grammar using electronic handsets that provided students with instant feedback would improve their learning of grammatical terminology and whether it would improve their writing. They conducted a 12 week study, drawing on a sample of 42 primary schools, half of which were randomly assigned to use the electronic handsets and the other as a control group. Grammatical knowledge was tested before and after the intervention using discrete grammar questions and a paragraph revision writing task to ascertain whether the grammar knowledge transferred to writing. The results indicated a strongly significant positive impact on grammar knowledge on the experimental group, using the handsets, but no impact on writing.

However, one aspect of grammar teaching that has been repeatedly identified as successful is sentence-combining, the practice of ‘teaching students to construct more complex and sophisticated sentences through exercises in which two or more basic sentences are combined into a single sentence’ (Graham and Perin, 2007: 18). The EPPI review (2004; Andrews et al., 2006) maintained that there was good evidence for the efficacy of sentence-combining, and it was one of the 11 instructional strategies identified by Graham and Perin (2007) as beneficial to writing. Sentence-combining has a long history in the USA with a chain of studies (Daiker et al., 1978; Hake and Williams, 1979; Hillocks and Mavrognes, 1986; Kinneavy, 1979; Marzano, 1976; O’Hare, 1971; Phillips, 1996; Saddler and Graham, 2005; Savage, 1980) all reporting the success of sentence-combining. Most were conducted in high schools or even at university level. The studies were themselves followed by a flurry of similar studies (for an overview of this, see Connors, 2000), most agreeing and confirming the efficacy of the technique. The only UK study to look at this, with secondary school students (Keen, 2004), also found a positive impact. But despite this apparent consistency in results, sentence-combining is problematic in terms of both the studies and the teaching of writing. In general, the studies measure whether teaching students to combine sentences, largely through exercises, improves their ability to create effectively combined sentences, with very few studies looking at whether writing quality is improved. Several critics have pointed out the shortcomings in these studies (Connors, 2000; Faigley, 1980; Witte, 1980), and Crowhurst (1980) conducted a study that demonstrated that increased syntactic complexity, such as that achieved through sentence-combining, was not correlated with writing quality. In addition, critics have suggested that any success deriving from sentence-combining is attributable to discussions about the ‘rhetorical principles characteristic of good writing’ (1980: 298), not to the strategy itself. Significantly, for example, Keen’s study included a considerable amount of contextualized discussion of the effectiveness of different sentences. Summarizing the approach, Andrews et al. argued that ‘sentence-combining suggests a pedagogy of applied knowledge – at its best, applied in situations of contextualized learning; at its worst, drilling’ (2006: 52), and it is probable that Andrews’ notion of applied knowledge rooted in contextualized learning is central to this issue.

Andrews’ observation that the success of sentence-combining might owe more to its contextualization than to the approach per se is important. More recently, research in this area has begun to look at the benefits of teaching grammar that is contextualized or embedded within the teaching of writing. In the USA, Fearn and Farnan (2007) found strong positive effects on students’ writing in high schools when teaching made connections between the grammar being taught and children’s writing, and they claimed that ‘grammar instruction influences writing performance when grammar and writing share one instructional context’ (Fearn and Farnan, 2007: 16). In England, a mixed method study, involving 13–14 year olds, and combining a large randomized controlled trial with a qualitative design has found a strongly significant benefit for students of a contextualized approach (Jones et al., 2013; Myhill et al., 2012). In this study, the instructional focus was the teaching of writing, and the teaching materials were designed around a consideration of what grammatical understanding might support learners in making authorial decisions in the composition of a particular text. It is the first robust study that has provided evidence of a positive impact of grammar teaching on writing, but it is important to note that the learning focus was writing, not grammar. Nonetheless, it offers a coherent and evidenced rationale for developing a purposeful role for grammar in the curriculum.

III Teacher perspectives on the teaching of grammar

Further insights into the role of grammar in the curriculum are provided by research studies or professional publications that attend to teachers’ perspectives on the topic. Professional perspectives on grammar are evident in numerous articles by teachers and teacher-educators, particularly in the special ‘grammar’ editions of the English Journal, namely volume 85 (1996), volume 92 (2003) and volume 95 (2006). These articles tend to fall in to two camps: those which position grammar predominantly as a matter of rule-learning and error-correction (e.g. Benjamin et al., 2006; Brown, 1996; Rose, 1996), and those which position grammar as a tool for promoting metalinguistic awareness and rhetorical choice (e.g. Ehrenworth, 2003; Gold, 2006; Hagemann, 2003; Jayman et al., 2006). The latter position is often informed, in articles from US practitioners, by Weaver’s concept of contextualized grammar (1996, 2006) and Kolln’s concept of rhetorical grammar (2006).

The professional literature reflects the complexity of defining what is meant by the terms ‘grammar’ and ‘grammar teaching’, as Vavra (1996) notes. A handful of studies have explored how teachers define or conceptualize grammar. The largest, a survey of 137 primary and secondary school teachers in the UK, reported that teachers lacked confidence in defining grammar, particularly in understanding ‘the relationship between implicit and explicit knowledge of language’, and noted a strong association of explicit grammar teaching with prescriptivism and old-fashioned teaching methods (QCA, 1998: 26). A similar problem of definition has been reported in the USA from a smaller-scale study by Petruzella which noted confusion between grammatical rules, issues of usage, and spelling and punctuation (1996: 69). This finding was echoed in Cajkler and Hislam’s interviews with trainee primary teachers in the UK, with participants associating grammar with phonics, spelling and punctuation (2002). UK studies by Pomphrey and Moger (1999) and Watson (2012) report tension or inconsistency between prescriptive/descriptive and prescriptive/rhetorical conceptualizations of grammar teaching among trainee and practising secondary English teachers respectively, with the latter finding that teachers tend to immediately conceptualize grammar as prescriptive, rule-bound and focused on correctness, while conversely offering a rhetorical view focused on exploration, choice and effects when asked about how it might be useful in supporting students’ writing development. All of these studies indicate that some degree of conceptual confusion persists within the profession.

Research into teachers’ beliefs about grammar more generally provide evidence that UK teachers tend to associate grammar with negative discourses of ‘old-fashioned’ teaching (QCA, 1998; Watson, 2012), and that these ‘neo-conservative associations’ (Myhill, 2006: 78) can occur even in countries in which grammar is not such a contested subject, such as Flanders and the Netherlands (van Gelderen, 2006). Investigations into teachers’ affective responses to grammar have typically reported ‘anxiety’ (QCA, 1998: 26) and ‘apprehension’ (Watson, 2012), and a general lack of confidence when it comes to dealing with grammar in the classroom, in both the UK (Kelly and Safford, 2009) and USA (Hadjioannou and Hutchinson, 2010). Teachers and trainee teachers have tended to be disparaging and hyper-critical in their self-evaluations (Pomphrey and Moger, 1999; Watson, 2012), and Cajkler and Hislam have found that anxiety remains high even when trainee teachers made improvements in their linguistic subject knowledge (2002). Teachers who feel anxious or insecure about their own declarative grammatical knowledge are also more likely to hold prescriptivist, rule-bound views of grammar (Harper and Rennie, 2009; Kamler, 1995; Macken-Horarik, 2001). However, studies have also reported more positive attitudes. Both Turvey (2000), working with secondary-level trainee teachers, and Cajkler and Hislam with primary-level (2002) found that, regardless of their anxieties about subject knowledge, trainees valued the idea of grammar teaching: Turvey found that her students felt that they had ‘missed out on something’ by not being taught grammar themselves at school, and that this made it ‘all the more important that their pupils should have it’ (2000: 143). Similarly, Watson found that some secondary English teachers felt passionately that the negative discourse around grammar teaching is misleading, and that far from being ‘the pit of doom’, grammar is actually ‘where freedom lies’, thanks to its focus on making knowledge about language and the choices involved in writing explicit (2012: 32). Such reports indicate that there may be willingness to learn about and practise grammar teaching even where teachers or trainees are insecure in their own linguistic or pedagogical knowledge. The most recent research into teachers’ beliefs has found that teachers tend to value the potential that a rhetorical approach to grammar teaching focused on experimentation has to increase students’ metalinguistic awareness and support them in consciously crafting their writing (Watson, 2012). There still persists, however, a tendency to associate grammar with a restrictive discourse of rules and accuracy (Watson, 2012) and to position grammar ‘as antipathetic to freedom and creativity’ (Wilson and Myhill, 2012: 11).

Attitudes to the use of grammatical terminology have been explored in three studies. Two of these indicate a belief that it is more suitable for use with higher rather than lower ability students (Petruzella, 1996; QCA, 1998). The third study examined teachers’ personal epistemologies, their ‘beliefs about what knowledge is relevant and valuable in learning to write’ (Wilson and Myhill, 2012: 4), in relation to literary and linguistic metalanguage and poetry writing. This reported a recurrent epistemological argument that ‘linguistic metalanguage was rule-bound and constraining, limiting young people’s writing in an unconstructive fashion’ (Wilson and Myhill, 2012: 10), which counterpointed a view that literary metalanguage is ‘a valuable part of learning about literature’ (p. 13). This is consistent with Findlay’s findings that UK teachers regard grammar as ‘a chore’ (2010: 4), and that they consistently value the literary aspects of the subject above the linguistic.

The literature, therefore, indicates a continuing trend amongst practitioners to view grammar as reactionary and restrictive, to value it less than literary aspects of the subject, and to be anxious about teaching it. However, there is also evidence of a countervailing belief that grammar taught with an exploratory and rhetorical approach can be valuable in increasing students’ metalinguistic understanding and ability to consciously craft or design their writing.

IV Teaching grammar and the demands on content and pedagogical subject knowledge

While teachers’ knowledge about grammar is not generally raised as a concern in non-Anglophone countries, historical factors mean that it is a significant issue for many first-language English teachers. Shulman (1987) has distinguished between subject content knowledge (knowledge of an academic domain) and pedagogical content knowledge (knowledge of how to teach that academic domain). The demands on teachers’ subject content knowledge about grammar have become increasingly specific in Anglophone countries: in the USA, the Common Core State Standards for Language Arts are heavily focused on grammatical constructions that students are expected to master; in the UK grammar was made a mandatory part of the curriculum in 1988, with the non-statutory but widely adopted National Strategies guidance (DfEE, 1998; DfES, 2001) including detailed teaching objectives for students aged 5–14; Australia is currently developing a new National Curriculum that includes a strand on Knowledge about Language aiming to foster ‘a coherent, dynamic, and evolving body of knowledge about the English language and how it works’ (ACARA, 2009: 1); and in New Zealand, problems have occurred in trying to implement an innovative syllabus with a strong grammar focus due to teachers’ ‘lack of knowledge about language’ (Gordon, 2005: 63).

While the extent to which students need to explicitly know about grammar is still hotly debated (Locke, 2010), there is widespread agreement that teachers’ grammatical knowledge needs to be richer and more substantive than the grammar they may need to teach to students. Teacher subject knowledge requires ‘a higher degree of grammar consciousness than most direct learners are likely to need or want’ (Leech, 1994: 18), with an ability to be ‘conscious analysts of linguistic processes’ (Brumfit, 1997: 163) and possess ‘conscious awareness’ (Armstrong, 2004: 223) of how texts are structured. It has been argued that teachers who understand grammatical forms may be better placed to support developing writers (Andrews, 2005), to identify linguistic development in their students (Gordon, 2005), and to ‘make the analysis explicit’ (Hudson, 2004: 113) when examining texts with their students.

The problem that current English teachers have with attaining the level of grammatical subject knowledge outlined above has arisen for two reasons: first, the fall from favour of grammar teaching in Anglophone countries following the Dartmouth Conference in the USA in 1966, due to the widespread view that the formal teaching of grammar had no beneficial impact on students’ linguistic facility (Hudson and Walmsley, 2005); second, the tendency in the UK for teachers to follow a literature degree route into teaching, along with is a shortage of applicants from a linguistics route (Shortis and Blake, 2010). As a result, many current English teachers were not taught grammar at school or university, a point also noted in the US context by Kolln and Hancock (2005), and by Gordon in New Zealand (2005). Of course, teachers who have literature degrees and are keen readers themselves do have a lot of knowledge about texts that they can draw on in language teaching, and they also have a substantial amount of implicit grammatical knowledge about texts. However, as Andrews (2005) points out, it is likely that ‘a teacher with a rich knowledge of grammatical constructions and a more general awareness of the forms and varieties of the language will be in a better position to help young writers’ (2005: 75).

This subject knowledge problem extends into initial teacher training courses, with Kolln and Hancock (2005) complaining that most pre-service programs for English teachers in the USA do not address grammatical knowledge, and a number of UK studies reporting weaknesses in grammatical knowledge (Andrews, 1994, 1999; Bloor, 1986; Burgess et al., 2000; Chandler et al., 1988; Hislam and Cajkler, 2006; Williamson and Hardman, 1995; Wray, 1993). In Australia, Louden et al. (2005) conducted a survey which indicated that teachers do not feel confident about teaching grammar when they complete their training, and Harper and Rennie’s pre-service teachers (2009) ‘showed limited understandings in their ability to analyse the parts and structure of sentences, and their knowledge of metalinguistic terms did not seem to extend past the basic concepts of “noun”, “verb” and “adjective” ’ (2009: 27). However, studies in the UK and Australia have indicated that some teachers may be over-critical or over-anxious in their self-evaluations, with Cajkler and Hislam (2002) finding that primary trainee teachers in the UK had reasonable knowledge of grammar despite being anxious about their subject knowledge, and Hammond and Macken-Horarik finding that primary teachers in Australia were confident in their knowledge of genres and text types despite expressing a lack of confidence in their own knowledge of ‘rules of traditional grammar’ (2001: 125). At the same time, in England, there has been criticism of the accuracy or appropriacy of curriculum materials prepared to support teaching grammar (Cajkler, 1999, 2002, 2004; Cameron, 1997).

The combination of curricular expectation that students will have explicit knowledge of grammar, and the tendency towards an absence of grammatical knowledge in the academic experiences of English teachers, creates a clear challenge for pedagogical practice and student learning. Research into teachers’ pedagogical content knowledge related to grammar teaching is limited, yet there is evidence from second-language research that strong declarative grammatical knowledge does not necessarily translate simply into effective teaching of language (Andrews, 2001; Borg, 1999, 2001, 2003). Pedagogical content knowledge is dependent upon teachers having a clear understanding of the role that grammatical knowledge can play in the classroom, without which teachers may generate mismatches between their own use of grammatical metalanguage and students’ understanding (Berry, 1997) or unwittingly convey inappropriate messages to learners. Lefstein (2009), for example, illustrates how teachers using policy materials underpinned by a principally rhetorical notion of grammar can use them in meaningless or rule-bound prescriptive ways, while the trainees in Cajkler and Hislam’s study (2002) understood grammatical knowledge as essentially about the naming of grammatical constructions but did not understand that pedagogically ‘grammatical awareness is about making available a range of choices for writers to use for particular purposes in particular contexts’ (2002: 176). In a more recent UK study, Myhill et al. (2013) have identified a range of pedagogical problems faced by teachers who attempt to address grammar. These include:

This study also, however, reported examples of teachers who were able to make meaningful connections between the linguistic features being studied and specific context-relevant effects or purposes, and who responded sensitively to students’ writing by explicitly drawing out effective grammatical choices and challenging students to reflect metalinguistically on their own writing. This focus on the relationship between teachers’ subject knowledge and pedagogical knowledge relating to grammar is mirrored in Australia by Jones and Chen (2012), who have demonstrated how teachers have been able to evolve new understandings of both grammatical knowledge and pedagogical practice through participatory collaboration with a research team. Current research therefore indicates the vital importance of addressing teachers’ pedagogical content knowledge alongside subject knowledge of grammar.

V Discussion

To an extent, what this review signals most clearly is that consideration of a fully theorized role for grammar within the writing curriculum has largely been hijacked by political–professional debates about grammar’s inclusion or not, and that the debate has not substantially developed since it began in the early 1960s. The same arguments are voiced and re-voiced over time, but with little re-theorization or advancement. Silently underpinning this long-standing debate, and perhaps perpetuating it, is the fundamental distinction between prescriptive and descriptive views of grammar. Whilst linguists have long theorized grammar as a description of language which will change and evolve as language changes, public and political views of grammar have tended strongly to prescriptive views of grammar. Hudson and Walmsley, in their analysis of the separate development of academic linguistics and grammar teaching, draw attention to ‘an ever-widening gap in England between the practice of professional grammarians, on the one hand, and the lay public and practice in schools on the other’ (Hudson and Walmsley, 2005: 595). However, this binary division between linguistics and educational perspectives on grammar is not wholly correct. Following the eschewing of grammar teaching in Anglophone countries in the 1960s, an alternative view of grammar emerged that was characterized by a view of grammar as ‘knowledge about language’, a view that was significantly influenced by sociolinguistics (e.g. the work of Peter Trudgill; see, for example, Trudgill, 2000) and was descriptive in theorization. In England, the emphasis in the Bullock Report (DES, 1975) on the importance of developing students’ understanding of language and its rejection of traditional grammar teaching was highly influential, and more fully realized in the late 1980s in the government-commissioned national project, Language in the National Curriculum (LINC), to investigate classroom approaches to developing knowledge about language. The resultant materials, subsequently rejected by the government, drew heavily on the work of Britton and Halliday, focusing on functionally oriented understandings of language in context and on placing meaning at the centre of attention. These materials have had a significant legacy: the dominant professional view of grammar within the writing curriculum is strongly oriented towards a descriptive approach (Watson, 2012; Wyse, 2006). Therefore, it is possible to argue that, despite the prolonged debate, one element of the theorization of grammar in the writing curriculum from both a linguistic and theoretical perspective is that it is founded upon a theory of descriptive, functionally oriented grammar. Given this foundation, the review also points to ways in which this theorization may be developed further.

Within the theoretical perspectives offered by Systemic Functional Linguistics, and rhetorical grammar – and the recent empirical work of Fearn and Farnan (2007), Macken-Horarik et al. (2011), Myhill et al. (2012), and Schleppegrell (2007) – there are signs of an emerging consensus that grammar may be important in developing learners’ understanding of how language works and, specifically, how grammar choices are significant in shaping and constructing meaning. Cameron argued theoretically that ‘knowing grammar is knowing how more than knowing what’ (1997: 236), pointing out that grammatical terminology is simply the tool that facilitates language investigation and analysis. To set this in a classroom context, this means that it is more important to know how a passive construction alters the emphasis in information conveyed than it is to know that it is a passive construction. The recent empirical work, described above, is providing evidence of how this knowledge is realized in practice and how young writers develop understanding that ‘it is through the rhetorical and syntactic forms they choose that the content is constructed and evaluated’ (Schleppegrell, 2007: 122). This consensus is less concerned with grammar as an arbiter of accuracy, and more concerned with ‘insights that go well beyond the minimum needed to write conventionally or correctly’ (Hancock, 2009: 194). Carter and McCarthy (2006: 7) draw attention to the notion of grammar as choice, as well as a grammar of structure, and Myhill et al. (2011) have conceptualized the contextual teaching of grammar within the writing curriculum as one that seeks to open up to young writers ‘a repertoire of infinite possibilities’.

Taking this theorization one step further, fostering young writers’ awareness of the linguistic choices available to them in writing and how those choices differently shape meaning is developing their metalinguistic knowledge of writing. Gombert’s (1992) taxonomy of metalinguistic knowledge, as noted earlier, relates principally to oral language development, but the category of metapragmatic knowledge is concerned with language in use, e.g. language in social contexts. Arguably, adopting descriptive approaches to grammar and nurturing students’ abilities to make choices and decisions in their writing is developing this metapragmatic knowledge, and rendering it available to inform the process of writing. Such knowledge is explicit, what Roehr described as ‘declarative knowledge that can be brought into awareness’ (Roehr, 2008: 179) and goes beyond the use of linguistic metalanguage to label and identify to include explicit knowledge of how linguistic choices subtly shape or alter meanings. It may also be significant that the empirical studies of Keen (2004) on sentence-combining, of Wyse (2006) on vocabulary choices, and Myhill et al. (2011) all seem to be highlighting the importance of talk, or metalinguistic discussion, in enabling this explicit metalinguistic knowledge of writing. Thus a more coherent theorization of a role for grammar in the curriculum might be framed as the teaching of grammar which promotes students’ explicit metalinguistic understanding of how grammar choices shape meaning in texts and of the writing choices available to them, founded upon a descriptive, functionally oriented understanding of grammar.

VI Conclusions

This review of the literature has illustrated clearly that the role of grammar in first language teaching remains a source of dispute, a conclusion underlined by the political/professional debate in England at present about the grammar test at age 11 and the emphasis on grammar in the revised National Curriculum. The review also signals that robust empirical studies in this area are limited and – apart from indicating that simply teaching grammar as the isolated naming and labelling of word classes and syntactical structures is of little obvious benefit – the research is rather impoverished in offering any theoretically-grounded understanding of children’s learning of grammar and how they transfer that learning to their own language use. It offers a way forward in theorizing a role for grammar in the writing curriculum as a functionally oriented endeavour, developing students’ metalinguistic thinking and decision-making in writing. It is also clear, however, that more well designed studies in this area are much needed to provide richer and broader understanding of how children develop metalinguistic understanding, how that learning transfers into their own language use, and the pedagogies that support the development of that understanding.