The Age Factor and Early Language Learning 9783110218282, 9783110218275

Citation preview

I dedicate this book to three outstanding professionals: Professor Mirjana Vilke and Professor Richard Johnstone, the most inspiring researchers of young language learners, and Dr. Jo´zsef Bogna´r, who motivated me to teach children.

An introduction to why context matters

Oleanders and lemon trees are precious plants kept in pots in my hometown, Pe´cs, Hungary. I place mine in the sunniest spots from spring to fall, keep them in the house during winter, and do everything to make them blossom. When I first saw the robust di¤erences of their wild growing relatives abroad I could not help but notice the importance of the local environment. Context matters. Similar to horticulture, the impact of the context is visible everywhere in early language learning. This volume is about how context contributes to early language learning and teaching. The first two chapters prepare the reader for the variety of topics in the rest of the book. The first paper discusses the relevance of the critical period hypothesis, types of curricula for young language learners and provides an overview of recent inquiries. Chapter two, by Peter Edelenbos and Angelika Kubanek, is about principles, research and good practice in Europe. Four chapters focus on assessing young language learners. Helena Curtain gives a detailed account of assessment instruments and proficiency standards used with early start programmes in the USA; Ofra Inbar-Lourie and Elana Shohamy discuss why the construct is the main issue in assessing young learners. Joanna Jalkanen’s chapter shows how teaching and assessment are implemented in harmony at a kindergarten in Finland, whereas Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller give an account of young learners’ oral assessment in Switzerland. The next two chapters focus on how age impacts on language learning. Carmen Mun˜oz overviews the relationship between input and longterm outcomes of early learning in a formal setting, whereas Chise Kasai compares how Japanese children and adults acquire two sounds. Three chapters explore individual di¤erences: Jelena Mihaljevic´ Djigunovic´ overviews the large body of research accumulated on young learners’ motivation, anxiety and other areas. Marina Mattheoudakis and Thomaı¨ Alexiou look into how young learners’ socio-economic status interacts with other factors in Greece; whereas Csilla Kiss gives an account of how an aptitude test was developed for Hungarian learners.

viii

An introduction to why context matters

Innovation is the common theme in the next section. Qiang Wang’s chapter provides insights into a survey documenting Chinese teachers’ enthusiasm towards teaching children English. Another Asian context is the focus of Jayne Moons’ paper in which she shows through case studies how policy decisions are implemented in Vietnam. Classrooms and Chinese learners are focused on in Jing Peng and Lili Zhang’s chapter. Discussions in the last three papers analyze the status of languages. John Harris argues against the traditional division between heritage, foreign and minority programmes based on studies conducted in Ireland, whereas Janet Enever discusses how the status of the target language influences young language learners’ choices in England. Finally, a critical discourse analysis reveals what is involved in early English programmes in the Israeli context. I hope readers will find the richness of the topics and the depth of the discussions in the many local contexts informative and helpful. The chapters document the impressive development in research methods into early language learning and teaching and may serve as a valuable resource for future studies. I am sure readers will find evidence on how context matters in age-related research. Enjoy the book as I enjoy my oleanders and lemon trees! Marianne Nikolov

Contents

Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

An introduction to why context matters . . . . . . . . . . . . . . . . . . . .

vii

1. The age factor in context Marianne Nikolov . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1

2. Early foreign language learning: Published research, good practice and main principles Peter Edelenbos and Angelika Kubanek . . . . . . . . . . . . . . . . . .

39

3. Assessment of early learning of foreign languages in the USA Helena Curtain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

59

4. Assessing young language learners: What is the construct? Ofra Inbar-Lourie and Elana Shohamy . . . . . . . . . . . . . . . . . .

83

5. Language proficiency and age-appropriate methodology at the English Kindergarten of Kuopio Joanne Jalkanen. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

97

6. ‘‘I can you help?’’ Assessing speaking skills and interaction strategies of young learners Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 7. Input and long-term e¤ects of early learning in a formal setting Carmen Mun˜oz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 8. English /l/ and /r/ acquisition by Japanese children and adults Chise Kasai . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 9. Individual di¤erences in early language programmes Jelena Mihaljevic´ Djigunovic´ . . . . . . . . . . . . . . . . . . . . . . . . . . 199 10. Early foreign language instruction in Greece: Socioeconomic factors and their e¤ect on young learners’ language development Marina Mattheoudakis and Thomaı¨ Alexiou . . . . . . . . . . . . . . 227 11. The role of aptitude in young learners’ foreign language learning Csilla Kiss . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253

x

Contents

12. Primary English in China: Policy, curriculum and implementation Qiang Wang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

277

13. The teacher factor in early foreign language learning programmes: The case of Vietnam Jayne Moon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

311

14. Chinese primary school students’ use of communication strategies in EFL classrooms Jing Peng and Lili Zhang . . . . . . . . . . . . . . . . . . . . . . . . . . . .

337

15. Expanding the comparative context for early language learning: From foreign to heritage and minority language programmes John Harris. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

351

16. The status of the target language: Contemporary criteria influencing language choices for early learners in England Janet Enever . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

377

17. English for Young Learners (EYL) in grades 1 and 2 in Israel: A critical discourse analysis Rivi Carmel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

403

Subject index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

423

1.

The age factor in context

Marianne Nikolov Who would ever have predicted such an enormous increase in enthusiasm towards teaching children foreign languages all over the world in the late 1970s during the low ebb of interest due to the Burstall, Jamieson, Cohen, and Hargreaves (1974) report? The world has definitely changed a lot since then. As readers will find out, a smorgasbord of recent studies have inquired into how young language learners develop in a variety of programmes. An early start to learn a new language tends to be seen as conducive to proficiency over time. Evidence used to come from second language contexts; however, in recent years a wide range of studies have been conducted on the teaching and learning of modern foreign languages (FL) at an increasingly early age in contexts where the target language is limited to the classroom. The aim of this chapter is to provide a critical overview of issues and challenges characterizing recent empirical research inquiring into early FL programmes. In the first part of this chapter I discuss the Critical Period Hypothesis, the point of departure usually claimed to provide a theoretical underpinning to early modern FL programmes. Then, I introduce models and their aims and time frames for early language learning (ELL). In the third section I overview recently published studies on various aspects of the teaching and learning of modern foreign languages. As will be shown, a range of large- and smaller-scale studies have been implemented to explore how the age factor works in a variety of educational contexts and they tend to fail to support ‘‘the younger the better’’ assumption. These findings, however, do not mean that ELL is a waste of time. Studies provide insights into complex ways of how young learners develop and o¤er an opportunity to discuss research methodology and areas for further research. As will be argued, the advantages and outcomes of an early start need to be analyzed in the specific contexts where young learners and their teachers interact with one another in classrooms in hugely varying conditions. Therefore, further research is necessary indeed.

2

1. 1.1.

Marianne Nikolov

The age factor and the Critical Period Hypothesis The Critical Period Hypothesis

Young learners are widely perceived to aquire languages in a qualitatively di¤erent way from adolescents and adults. Children before a certain age seem to pick up a new language with ease and success, whereas older learners often fail to do so. Discussions on the age factor tend to focus on the Critical Period Hypothesis (CPH) for language (e.g., Scovel 1988, 2000; Singleton 2001, 2005). The CPH has been a hotly debated research area not only in second language acquisition (SLA) research, but also in linguistic theory and cognitive science (Hernandez, Ping and MacWhinney 2005; Paradis 2004; Pinker 1994; McWhinney 2005). Some researchers argue that di¤erent critical, or sensitive, periods characterize language acquisition in di¤erent linguistic areas, and ‘‘the existence, or not, of one or more sensitive periods for SLA has major implications for the validity of any SLA theory’’ (Long 2005: 311). An important distinction has been confirmed in recent cognitive and neurobiological explanations of SLA reflecting a dual procedural/declarative dimension widely accepted in cognitive science (e.g., McWhinney 2005; Paradis 2004; Ullman 2001). Two systems exist side by side: a rule-based analytic procedural system, and a formulaic, exemplar-based declarative system (Skehan 1998). In the first system, storage and powerful generative rules operate together to compute well-formed sentences; in the second one, a large memory system is responsible for drawing on some rules operating on unanalyzed wholes or chunks. Young language learners use memory-based processes, whereas adult language learners rely on rule-based learning. The CPH concerns implicit linguistic competence (DeKeyser 2003; Long 2005). ‘‘The decline of procedural memory for language forces late second-language learners to rely on explicit learning, which results in the use of a cognitive system di¤erent from that which supports the native language’’ (Paradis 2004: 59). The acquisition of implicit competence is a¤ected by age in two ways: (1) biologically, the plasticity of the procedural memory for language gradually decreases after about age five; and (2) cognitively, reliance on conscious declarative memory increases both for learning in general and for learning a language from about age seven. Learners may apply compensatory mechanisms to counterbalance decline in implicit learning: learning new vocabulary,

The age factor in context

3

for example, relies on declarative memory, thus, it is not susceptible to the CPH (Paradis 2004). This fact explains how exceptionally successful adult learners manage to become proficient despite a late start (for an overview see Moyer 2004; Nikolov and Mihaljevic´ Djigunovic´ 2006: 236–240): most of them possess unusual memory capacity (Skehan 1998) and rely heavily on metalinguistic knowledge and pragmatics (Paradis 2004). Recent neuroimaging techniques have triggered new types of studies on the CPH debate. The overall findings (Stowe and Sabourin 2005; Wattendorf and Festman 2008) show that (1) the same brain areas are used for both first and second language; (2) late learners may draw on other areas as well; and (3) areas responsible for processing language seem to be used less e‰ciently for second languages in the case of very early proficient bilinguals as well as late bilinguals. Discussions on the CPH have focused on the question whether there is an abrupt fall or a continuous decline in SLA. Some researchers analyzing large-scale datasets (Chiswick, Lee, and Miller 2004; Hakuta, Bialystok, and Wiley 2003) used self-assessment data on age on arrival, length of exposure, and language development from immigrants in Australia and the United States and found linear relationships between age on arrival and language proficiency. These findings indicate that success in SLA declines steadily over time and thus, a counter argument against a critical period. A plausible explanation is put forward in MacWhinney’s (2005: 64) Unified Competition Model: older learners become increasingly reliant on connections between sound and orthography and they vary in the constructions they can control or that are missing or incorrectly transferred. They have restricted social contacts and their cognitive abilities also decline. In his view, none of these factors predict a sharp drop at a certain age in SLA, but a gradual decline across the life span. An indepth analysis of maturational constraints is proposed by Hyltenstam and Abrahamsson (2003): they argue that a ‘‘consensus model’’ can integrate the accumulated evidence on empirical facts and the relationships among them. As they claim, maturation can account for the general linear decline in learning potentials with increasing age on arrival for learners in general, ‘‘whereas the variability between exceptionally successful and non-exceptional L2 learners of the same starting age is accounted for best by non-maturational factors’’ (Hyltenstam and Abrahamsson 2003: 574).

4

Marianne Nikolov

Two main perspectives have been examined in research into the CPH: (1) the rate of acquisition and (2) ultimate attainment. Most studies have confirmed the findings of the seminal paper by Krashen, Long, and Scarcella published in 1979: older age is an initial advantage in the rate, but a disadvantage in ultimate attainment. The widely accepted findings state that (1) adults progress at early stages of morphology and syntax faster than children; (2) older children acquire faster than younger children; and (3) child starters outperform adult starters in the long run. Most parents and decision makers in foreign language contexts seem to be aware of the third point, but they mistakenly assume that young learners develop fast. In other words, enthusiasm towards an early start is not supported by empirical studies, as the younger means the slower. Special discussions concern what linguistic areas are mostly a¤ected by a critical or sensitive period. It is widely accepted that accent is at the very heart of the CPH. In Long’s view (2005), native-like accent is impossible to attain unless first exposure occurs before age six or twelve, indicating quite a controversial range; whereas a sensitive period for lexical, collocational abilities and morphology and syntax ends between the ages of six and mid-teens. Although a growing body of studies on successful adult language learners provide evidence that native-like proficiency (including accent) is available to some exceptional adults, these discussions are beyond our scope (for recent overviews see e.g., Moyer 2004; Nikolov and Mihaljevic´ Djigunovic´ 2006; Singleton and Ryan 2004). To summarize: the CPH claims that natural language acquisition is available to young children, whereas older adolescents and adults have limited or no access to it. Although the existence of age e¤ects is widely accepted, many applied linguists disagree on whether they are consistent with a critical period. Overviews on the age factor tend to rely on the same body of empirical evidence; however, some authors interpret studies in favour of the existence of the CPH (e.g., DeKeyser 2003; DeKeyser and Larson-Hall 2005; Hyltenstam and Abrahamsson 2001, 2003; Ioup 2005; Long 2005) or against it (e.g., Bialystok 2001; Birdsong 2005; Marinova-Todd, Marshall, and Snow 2000, 2001; MacWhinney 2005; Moyer 2004, Tomasello 2003), whereas others present a balanced view (e.g., Johnstone 2002; Mun˜oz 2006; Scovel 1988, 2000; Singleton 2001; Singleton and Ryan 2004). No wonder Singleton (2005: 269) calls the CPH ‘‘a coat of many colours’’ and claims that ‘‘it is like a

The age factor in context

5

mythical hydra, whose multiplicity of heads and capacity to produce new heads rendered it impossible to deal with’’ (p. 280).

1.2.

The CPH and early foreign language programmes

One wonders how these scholarly discussions are relevant to early modern foreign language programmes. It is clear from the literature on the CPH that the vast majority of studies rely on data collected in second language contexts and not in educational contexts where the target language tends to be limited to the classroom as a foreign language and is simply one of the school subjects. Even in second language contexts, one needs to bear in mind, there is no guarantee that all children will attain native-like proficiency levels (Long 2005: 289), as there are many other factors contributing to ultimate attainment. To illustrate, a recent longitudinal study found that young children had strong accents after four years of enrollment in English-medium schools in the United States (Flege, Birdsong, Bialystok, Mack, Sung, and Tsukada 2005). Research shows that native accent is not automatically available in second language contexts either. Therefore, we need to take into consideration what these findings imply for ELL in foreign language contexts. Some experts (e.g., Nikolov 2002; Singleton 2001) have cast doubt on the relevance of the CPH to ELL, despite the fact that most programmes for young learners tend to refer to ‘‘the earlier the better’’ slogan. DeKeyser and Larson-Hall (2004: 101) point out that early foreign language and partial immersion programmes do not capitalize on children’s implicit learning skills for two reasons: first, they are formfocused, and second, they are limited in time. The latter is the reason why Larson-Hall (2008) calls ELL ‘‘minimal input situation’’. A third, equally important reason has to be added: in most ELL contexts teachers are non-native speakers of the target language and their proficiency may often be below what enthusiastic advocates of ELL expect to achieve – native-like proficiency (Enever, Moon, and Raman in press; Nikolov 2002, 2008). These points lead us to the aims ELL programmes set and the models they implement in their limited time frames.

6

2.

Marianne Nikolov

ELL programme models: Time and aims in curricula

Research into the CPH has focused on what is possible over many years in bilingual contexts. In foreign language learning contexts young language learners study a new language in di¤erent conditions. 2.1.

Time devoted to language learning in ELL models

Early FL programmes devote a limited amount of time to ELL: between less than an hour a week to short daily sessions (for an overview see Curtain and Dahlberg 2004). A realistic recommendation suggests a ‘‘minimum time allottment . . . should be 30 minutes three times per week’’ (Curtain 2000: 202). A recent overview on European countries found that time devoted to ELL ranges between roughly one hour and several hours in content and language integrated learning (CLIL) programmes (Edelenbos, Kubanek, and Johnstone 2007; Edelenbos and Kubanek present volume; Johnstone in press). This finding shows that in most programmes an extremely limited amount of time is available for ELL. In contrast, bilingual programmes vary in the ratio of the curriculum available to learners in their first (L1) and second language (L2) (Curtain and Dahlberg 2004), but learners tend to study subjects in their L2 in about half or more of their curricular time. As a result, many learners achieve close to native-like levels in listening and reading comprehension, whereas their spoken and written L2 is less good (Cummins no year). The amount of time is only one of many important factors contributing to ELL. Opportunities to learn, the quality and amount of input and interaction available to learners in and outside the classroom also vary to a large extent, similary to the time when ELL starts. In line with research into the CPH, most sources consider learners ‘‘young’’ before age 12, but the typical starting age ranges between 3 and 11 years. This wide age range means enormous di¤erences not only in what children literate in their L1 can do, but also in what tasks and activities may work with them, thus making programmes hardly comparable. Another important point relates to when and how children transfer from ELL programmes to later ones and how their achievements are built on.

The age factor in context

2.2.

7

Achievement targets in ELL

2.2.1. Range between awareness raising and content learning The detailed overview of ELL in Europe analyzed published research, good practice and main principles (Edelenbos et al. 2007) and found enormous variability in the programmes (see Kubanek and Edelenbos in present volume). They range from initiatives aiming for no more than broad awareness in a number of languages to partial immersion. The type of programme has important implications for curricular aims (Julkunen 2000) as well as for assessment (see Inbar-Lourie and Shohamy in present volume) and teacher education. 2.2.2. Achievement targets in the new language Achievement targets vary a lot in di¤erent programmes, but young learners of foreign languages are not expected to achieve native level in school, although sometimes parents and policy makers may voice unrealistic goals and expectations (Curtain 2000; Nikolov 2002). In most contexts achievement targets tend to be modest (for details see e.g., Blondin et al. 1988; Edelenbos et al. 2007; Curtain; Haenni Hoti, Heintzmann, and Mu¨ller; Inbar-Lourie and Shohamy all in present volume) and di¤erent levels may be required in the four skills. Also, in FL contexts the target language is a curricular area in its own right and it is assessed similary to other schools subjects (for a discussion of the construct see: Inbar-Lourie and Shohamy present volume). All in all: achievement targets should reflect developmental stages in ELL and a gradual shift from prefabricated utterances to more analyzed language use (Johnstone in press). 2.2.3. Shaping attitudes and motivation Recent European guidelines for curricula suggest going beyond linguistic aims and setting complex goals for ELL. They promote exposing young learners to an L2 not only for linguistic purposes, but also to allow them to develop favourable attitudes towards languages and language learning, and to help them become proficient users of foreign languages as adults. For example, a European language policy document states that ELL in the kindergarten and primary school is a priority, as in such programmes ‘‘attitudes towards other languages and cultures are

8

Marianne Nikolov

formed, and the foundations for later language learning are laid’’ (Commission of the European Communities 2003: 7). This point coincides with what handbooks for teachers emphasize (Cameron 2001; Curtain and Dahlberg 2004; Moon 2000a; Pinter 2006) and empirical research also underpins (e.g., Johnstone 2002; Nikolov 1999). A further a¤ective aim may concern language learning anxiety in ELL (Mihaljevic´ Djigunovic´ present volume) and learners’ own culture and identity (Prahbu in press; Vickov 2007). Little is known about the implementation of these aims and how they feature in the long run in various contexts. 2.2.4.

Learning to learn languages

A third area besides socio-a¤ective and linguistic aims concerns metacognition and ‘‘learning to learn’’. Research on early bilingualism has consistently found (Bialystok 2001, 2005) that bilingual children process languages more e¤ectively than their monolingual peers and the constant management of two competing languages enhances executive functions and results in a higher level of metalinguistic awareness. Thus, another important aim may involve the development of metacognitive skills and learning strategies (Edelenbos et al. 2007) to support children in learning new languages not only in ELL programmes but also as adolescents and adults. As will be seen some research is available on learning strategies, but more would be desireable. 2.2.5.

Learners’ mother tongue

All young learners in ELL programmes have mastered their mother tongue (L1) to a certain level and the two languages interact with one another. Some ELL curricula may emphasise the role of L1. In recent years English as a lingua franca has been not only the most popular L2 in ELL, but it has been perceived both as a blessing and as a threat in certain contexts (Graddol 2006; Nunan 2003). Prahbu (in press) even voices his concern about English becoming a substitution for other languages in India. In the European context language policy documents promote the learning of two modern foreign languages to counterbalance the spread of English thus promoting the ideal of plurilingual citizens which could be paraphrased as speakers of not only English. Interestingly, no research has been conducted to explore the impact of ELL on L1 development.

The age factor in context

9

2.2.6. Transition and continuity Besides what aims di¤erent models define it is crucial to consider what happens after the first few years of ELL covered in language policy documents and how young learners’ achievements and development in the areas mentioned in the above sections are integrated into later programmes. ELL curricula tend to cover a certain number of years – usually the first years in kindergarten and primary school. Continuity and transfer, however, have represented major challenges since the early 1970s. Continuity may be lacking not only for the particular languages young learners started learning, but also in the area of methodology and integration of what they can do and are good at (Nikolov 2001, 2002). Therefore, it is a must to examine ELL in their wider educational contexts; if teachers in upper grades fail to build on what and how young learners are able to do in their new language, the ELL e¤ort may easily be wasted and lead to frustration. The only published study comparing and contrasting curricula for primary and secondary teaching was conducted in Greece on German as a second foreign language (Papadopoulou 2007). By analyzing objectives, themes and contents of activities, teaching methodology and assessment the author suggests how a common curriculum could be designed and implemented. Such studies would be needed in other countries and for all languages taught in ELL. 2.3.

Teachers

Teachers should be proficient users of both languages and familiar with the general curriculum and the principles of how young children learn. However, this is not always the case. Two main models exist in ELL: the teacher is (1) a generalist homeroom teacher; (2) or a language specialist. In the first case teachers are familiar with the curriculum, can integrate curricular areas and know how to manage young learners with ease, but their proficiency may not be very good. In the second case teachers may be more proficient in the L2 but less skilled in implementing age-appropriate methodology. Therefore, teachers represent a major challenge in ELL (Goto Butler 2004, 2005; Curtain 2000; Enever, Moon, and Raman; Graddoll 2006; Nikolov 2002, 2008; Shohamy and Inbar-Lourie 2006). Research into teacher education and classroom observation studies involving analyses of classroom discourse

10

Marianne Nikolov

and interaction provide insights into these points (see sections on teachers and classrooms). 2.4.

Why ELL is a good idea

Aims in ELL need to be in harmony with how young language learners develop and other components in their general curriculum. In ELL models achievement targets tend to be modest and way below native proficiency matched for the age groups. Given the limited time frame and other typical constrains, such goals are realistic but fall short of why the CPH would be a relevant point for consideration. The main reason why ELL is a great initiative is that by early exposure children may enjoy the potential advantages of starting young, as well as profit from what they experience at later stages in their language learning. In other words, as Johnstone (2002, in press) puts it, those making an early start may benefit from advantages at an early point in their education (relative ease of acquiring the sound system and unanalyzed wholes, higher level of motivation, lower anxiety, more time over years, etc) as well as at a later stage (more background knowledge, literacy and learning skills, strategies and analytical skills). As is widely accepted, ELL may also influence learners’ a¤ective, cognitive and metacognitive development over the years. However, these long-term outcomes need to be researched systematically in individual contexts – an area where more studies are needed.

3.

Themes and research methods in ELL studies

This section provides an overview of the emerging themes, issues and challenges characterizing recent empirical research inquiring into ELL programmes with a primary focus on studies conducted in Europe, where compulsory foreign language education tends to start at an increasingly early stage (Key data on teaching languages at school in Europe 2008). However, the overview goes beyond Europe and attempts to provide insights into trends on other continents as well. The most important criteria for inclusion are that the study (1) was published during the last decade or so; (2) participants are young learners (or older but started young), their teachers, or other stakeholders; (3) the number of hours per week is between one and five, but the aim of the pro-

The age factor in context

11

gramme is neither awarenesss raising towards various languages, nor partial immersion or CLIL; (4) the context is embedded in the state and not the private sector; and (5) the research method is appropriately documented. Chapters in the present volume are also included in the review, but they are not discussed in detail, as the interested reader will find them in the book. The worldwide increase in early foreign language learning in state education has resulted in a growing number of empirical studies. Some of these developments are well documented in country case studies (Enever, Moon, and Raman in press; Nikolov and Curtain 2000), publications of small-scale research projects usually focusing on a particular aspect of ELL (e.g., studies in Moon and Nikolov 2000; Nikolov in press; Nikolov, Mihaljevic´ Djigunovic´, Mattheoudakis, Lundberg, and Flanagan 2007; and see studies in further sections), as well as large-scale longitudinal national projects (e.g., in Spain by Garcı´a Mayo, and Garcı´a Lecumberri 2003; Mun˜oz 2006; in Ireland by Harris and Conway, 2002; Harris et al., 2006; Wang in present volume). Recent stateof-the-art reviews show (e.g., Blondin et al. 1998; Edelenbos et al. 2007; Nikolov and Mihaljevic´ Djigunovic´ 2006) that although many studies are cross-sectional and follow a quantitative or a qualitative tradition, more and more studies apply triangulation: they look into linguistic, a¤ective factors as well as classroom variables, and some of them provide insights into learners’ and teachers’ perspectives; also, some, mostly smaller-scale studies are longitudinal. Overall, as will be seen, reseach on ELL is gradually becoming more sophisticated and complex and the pictures emerging from studies present all shades of various colours. However, most examine either teaching or learning, and few are replicable in other contexts. On the other hand, comparative studies have also emerged aiming to examine issues across borders. 3.1.

Large-scale national projects

3.1.1. Pilot projects versus ELL on a large scale ELL programmes are often launched first as pilot projects to prepare ground for later programmes. In the 2002/03 academic year ten European countries reported a pre-primary or primary pilot project (Key data on teaching languages at school in Europe 2005: 32). As Johnstone (2008) claims, it is well-known that ELL can be a great success when

12

Marianne Nikolov

implemented in individual project schools, where appropriate conditions are provided and where the teaching is of high quality. However, largerscale implementation of ELL across schools is a di¤erent matter. In some countries ELL is introduced without piloting either as a result of strong parental pressure or a governmental decision. In many contexts major di¤erences have been documented between the pilot phase and later large-scale programmmes. For example, in Austria (Jantscher and Landsiedler 2000), Croatia (Mihaljevic´ Djigunovic´ and Vilke 2000), Italy (Gattullo and Pallotti 2000), Ireland (Harris in present volume) and Scotland (Blondin, et al. 1998) enthusiasm was on a high level and good progress was documented in the initial phase; however, attitudes towards ELL changed and professional quality declined as ELL became part of everyone’s daily routine. It has been found in various contexts that experimental and pilot programmes implemented by enthusiastic teachers produce remarkable results (e.g., Edelenbos, Kubanek and Johnstone 2007; Moon and Nikolov 2000; Nikolov 2002); but the spread of ELL often results in less research and funding and teachers do not find daily challenges as exciting and easy to cope with as teachers participating in monitored projects. For example, in Italy, carefully designed in-service teacher education programmes supported the introduction of ELL, but enthusiasm declined as the practice gradually spread (Gattullo and Pallotti 2000). In Croatia, a large-scale national ELL project taught a variety of languages as part of a longitudinal study; however, the ministry withrew support and the project was over (Mihaljevic´ Djigunovic´ and Vilke 2000; Vilke and Vrhovac 1993, 1995). Years later, when the government introduced compulsory ELL from grade 1 for all children in Croatian schools, they failed to intergrate the results. A similarly well-documented ELL project in Scotland ended up di¤erently from the original initiative without specialist teachers (Blondin et al. 1998). The same trend is true not only for foreign languages but also for second languages, making the distinction between these two terms less marked (Harris in present volume). A di¤erent example comes from Hungary and the author’s own experiences. I used to teach young learners English from their age 6 to 14 at an 8-year primary school from the late 1970s for two decades. As there was no tradition or syllabus to rely on, I had to find out how to develop learners’ English with the help of intrinsically motivating and cognitively challenging tasks. As a result of this major challenge, I involved them in negotiations on what they liked and disliked and developed an

The age factor in context

13

eight-year-long story-based syllabus with them (Nikolov 1999, 2002). As a teacher educator in the second part of my professional life I have found out much to my dismay that not all teachers are enthusiastic about ELL. In a number of studies on teachers (2008) and ELL classrooms in nation-wide surveys (2003, in press) I have documented that many teachers of young learners wish they could teach older ones, and lack not only proficiency and pedagogical skills but also a desire to improve their practice in the ELL classroom. 3.1.2. Large-scale projects on start and rate of ELL A number of large-scale projects have been implemented and documented in various publications serving with lessons for ELL. The eightyear-long ELL project mentioned in the previous section (Mihaljevic´ Djigunovic´ and Vilke 2000; Vilke and Vrhovac 1993, 1995) involved over 1,000 first graders learning English, French, German, and Italian. It aimed to find the optimal starting age to launch ELL and to pilot age-appropriate materials and meaning-focused classroom techniques. The project followed learners’ development in the target languages and changes in their attitudes and motivation over time. Control groups starting in grade 4 were involved to make outcomes comparable. At the end of year 8, the earlier starters were slightly better but achievements over time depended on many variables besides length of ELL. Two longitudinal projects carefully designed to examine various aspects of the age factor explored early and later introduction of English as a foreign language into the school curriculum of bilingual (CatalanSpanish; Basque-Spanish) learners. The Barcelona Age Factor (BAF) Project (Mun˜oz 2006 and present volume) started in 1995 and involved 2,068 participants. It aimed to find out whether age has an e¤ect on the rate of ELL, whether older learners surpass younger learners the way they do in natural L2 contexts, and how age a¤ects di¤erent language areas in ELL. Data were collected at three time intervals: after 200, 416 and 726 hours of instruction on several measures: speaking, listening, writing, and reading in English, two tests measuring comprehension in L1 (Spanish and Catalan) and a questionnaire. Some tests were discrete point, others were integrative skills tests, whereas oral interviews and role-plays were meaning-focused. The earliest beginners (age 8) showed the highest rate of learning in the last third of the followed period, the middle young beginners (age 11) progressed fastest in the second third (between 200 and 416 hours

14

Marianne Nikolov

of instruction), whereas adolescent beginners showed fastest initial rate of learning in the first third of the period (after 200 hours). For those starting at ages 8 and 11 the rate of learning became salient at the age of 12. Mun˜oz concludes that younger children need a longer time than young beginners in L2 contexts. After nine years of learning English, the di¤erence in scores on tests implying implicit learning (e.g., listening comprehension) got smaller. As Mun˜oz (2006) points out, di¤erences disappear between younger and older beginners with the same time and exposure once they reach the same stage of cognitive development. A similar project involved Basque and Spanish speaking learners of English from the ages of 4, 8, and 11 (Garcı´a Mayo and Garcı´a Lecumberri 2003). The research design was parallel with the previous longitudinal study and the findings are also similar: on a number of performance measures, including oral and written perception and production tasks, older beginners achieved significantly higher scores than younger learners (Cenoz 2003; Garcia Lecumberri and Gallardo 2003; Garcia Mayo 2003; Langabaster and Doiz 2003; Mun˜oz 2003 and in present volume). Both studies followed a special up-to-date research design allowing for triangulation of the data. Researchers looked into not only linguistic development in several skill areas but also young learners’ motivation and strategies. In both studies, however, the same tests were applied for the di¤erent age groups for the sake of comparability (Mun˜oz 2003: 167). Therefore, it is unclear how learners would have performed on tests more tuned to their cognitive and proficiency levels. Another important complex variable, the quality of teaching including teachers’ proficiency (accent which is at the very core of CPH studies) and classroom methodology were not included in any form. Garcia Lecumberri and Gallardo (2003: 129) point out that all teachers were non-native speakers of English, but ‘‘for obvious reasons, this can be a sensitive issue, which we have not been able to address yet’’. This may mean that younger groups were taught by less proficient teachers than their older peers and it is also possible that the actual classroom activities were more conducive to older learners’ development. Overall findings of these projects support what Krashen et al. (1979) found in their overview of the literature: younger learners develop at a slower pace than older ones. Therefore, it would be desirable to collect more evidence on how ELL is beneficial over a long period of time.

The age factor in context

3.2.

15

Studies on long-term e¤ects of ELL

3.2.1. Studies involving adolescents and adults The best way of testing the impact of ELL programmes and how, if at all, the CPH exerts its influence in the long run, would involve adults who started young. No such studies exist, most probably because research does not go beyond the time span of a few years. The only project relevant here is Urponen’s (2004) study on Finnish women married to native speakers of English. She found age when learning English as a foreign language started to be one of the significant predictors of success over time; however, some later starting women also achieved native proficiency. In a very di¤erent research paradigm Chise Kasai in the present volume uses an experimental design focusing on Japanese speakers’ perception and production of highly problematic sounds in English. She found an age e¤ect for production, but not for perception of /l/ and /r/. In other words, younger learners were better at producing words with these sounds than adults but not at distinguishing them. Yet another research design was applied in a retrospective study involving young adults (Nikolov 2001). In an exploratory qualitative enquiry on 94 adult learners’ school experiences various foreign languages were involved. All participants studied an L2 for 5 to 8 years in primary school, but they all perceived themselves as ‘‘unsuccessful’’: they felt they never benefited from early exposure (see next section on continuity problems). The research design of a study implemented in the US combined a large-scale longitudinal survey and retrospective interviews with a few particpants after ten years (Heining-Boyton and Haitema 2007). The majority viewed their ELL positively and their attitudes towards speakers of foreign languages and cultures were also favourable. Some correlational studies examined the relationship between how long a time learners spend studying the FL and how proficient they are. In one project representative samples of Hungarian learners’ proficiency was assessed in English and German in years 6 and 10 (ages 12 and 16) after studying the L2 for 1 to 6, and 2 to 10 years, respectively. The correlations between achievements and the number of years of language study were very low (between .25 and .29), and the ones between the number of weekly classes ranging from 2 to more than 5 were higher but also modest (.39 to .43) (Nikolov and Jo´zsa 2006).

16

Marianne Nikolov

In a small-scale correlational study Larson-Hall (2008) studied Japanese college students who started studying English between ages 3 and 12. They were examined on a phonemic discrimination and a grammaticality judgement task. Modest relationships were found and age e¤ects seemed to emerge after a substantial amount of input had been gained – a finding in line with what Mun˜oz has repeatedly pointed out. 3.2.2.

Transfer and continuity

The lack of continuity was one of the most important findings of the ‘‘French in the Primary’’ report in Britain (Burstall et al. 1974) and transfer has been a major challenge in ELL in many educational contexts (see country case studies in Nikolov and Curtain 2000). Continuity may be lacking in various areas. (1) Students are not o¤ered an opportunity to study at an appropriate level and are placed in groups of beginners. Such practice may lead to a decline in motivation. (2) They may not be able to study the language they learnt in ELL years. This is acceptable after an awareness raising programme if students are motivated to start a new language. It may, however, result in a demotivating experience if learners are denied an opportunity to continue learning their L2 due to limited access. (3) The lack of continuity may occur in classroom methodology: learners may feel they cannot come up to expectations, as the new course does not build on what they can do. The study on unsuccessful adult learners mentioned in the previous section revealed how ELL opportunities interact in complex ways with other factors in the educational system (Nikolov 2001). Over half of the participants were placed in beginner classes or had to give up the language they had learnt for years. One of the participants claimed, ‘‘I’ve been learning languages for years but I haven’t got anywhere’’. A later large-scale study involving all year 9 (age 15) students in state secondary schools in the same country (Hungary) revealed that twothirds started learning the same language they had learnt for 5 to 8 years in primary school as beginners or false beginners (Va´go´ 2007). This picture is further coloured by the status of the languages they are o¤ered. In summary, in order to find out how ELL programmes are beneficial in the long run, continuity is an area to examine. Unfortunately, no published research is available on how secondary school teachers build on ELL.

The age factor in context

3.3.

17

Studies on young learners

A large number of studies have become available in recent years exploring how young language learners develop and how various factors contribute to their proficiency over time. Among these factors, time was discussed in previous sections. In this part the a¤ective, cognitive and social contribution are examined in interaction with classroom variables. 3.3.1. A¤ective domain Young language learners’ attitudes, motivation and anxiety have been examined in several contexts exploring the arguments for an early start: children’s attitudes are positive, and they are more motivated and less anxious than older students, as research has systematically revealed. The interested reader is directed to an excellent overview by Jelena Mihaljevic´ Djigunovic´ in the present volume on research into individual di¤erences and to the detailed discussions of recent research in Europe (Edelenbos et al. 2007). There is, however, hardly any research into how ELL contributes to openness towards other cultures and speakers of other languages over time, although it is widely assumed that learning a new language at a young age should impact on these areas favourably. The reason must be that most studies inquire into processes and outcomes within realistic time frames, for example, in one type of school. Retrospection would allow researchers to gain insights into these areas, but no such studies are known. A narrower focus, the relationship between language learning motivation and cross-cultural contact was examined among Hungarian eighth-graders by Csize´r and Kormos (2009). Their results indicate that students of English have more positive attitudes towards native speakers and higher level of linguistic self-confidence, invest more energy into language learning and receive more support from their environment than students of German. However, all learners showed low expectancy of success and they did not report investing substantial e¤ort into language learning. The findings also reveal that learners of English experience more frequent written contact than learners of German. Integrativeness, linguistic self-confidence and perceived importance of intercultural contact were found to be closely related to how much e¤ort students are willing to invest into foreign language learning.

18

Marianne Nikolov

Studies providing young learners and their teachers with opportunities to present their emic perspectives allow us to hear children’s and their teachers’ voices and understand what is involved in ELL. Qualitative studies applying interviews, observations, analyses of teachers’ and learners’ diaries and other techniques may o¤er insights not available in other research paradigms. Following the tradition set in a previous study on ELL (Nikolov 1999), Nagy (2009) invited young learners to think about their reasons of learning English, what helped and hindered them, and other questions. Her findings reveal that in many cases ELL is not without its daily problems and even at an early stage children are aware of important points and may provide crucial insights and feedback to teachers. The same is true for many teachers in ELL programmes: they may be familiar with the principles of methodology, but may find it di‰cult to implement meaning-focused tasks on a daily basis (Goto Butler 2005; Kiss 2000; Lugossy 2007, Moon in present volume; Nikolov 2008; Wang in present volume). 3.3.2.

Cognitive domain

Language learning aptitude of young language learners is a generally under-researched area, most probably because children are expected to develop interpersonal communication skills with ease. A few recent projects set out to explore what role cognitive skills play in ELL in the Hungarian and the Greek context. Both educational systems are characterized by a strong streaming tradition – a reason why placement is an issue from an early age. Three studies examined how young Hungarian learners’ abilities contribute to their achievements in learning languages. In a context where streaming has been present for many decades and English tends to be provided for the more able learners, in a study involving over 400 learners from ten schools 22 per cent of the variation in English performances was explained by their aptitude (Kiss & Nikolov, 2005). An even younger age group was involved in a similar project piloting an aptitude test for 8-year-olds (Kiss in present volume). The chapter reveals why such a test was necessary, how it was validated, and what lessons can be learnt from the enquiry. In a large-scale longitudinal study on nationally representative samples of over 40,000 learners of English and German (ages 12, 14, 16 and 18) learners’ scores on an inductive reasoning test predicted a large portion of the variance in their performances on reading tasks (e.g., 14% in

The age factor in context

19

English at age 12). In the younger age groups inductive reasoning skills explained larger portions of the variance than in the older cohorts. The relationships between the cognitive variable, L1 reading, and other factors were stronger than in the case of older learners. Interestingly, relationships between L2 skills became weaker over time (Csapo´ & Nikolov 2009). Greek learners were involved in a project examining five- to nineyear-old Greek learners’ cognitive abilities and their learning of English vocabulary. Significant relationships were found between young children’s aptitude and vocabulary development in English (Alexiu 2009). 3.3.3. Socioeconomic status Although socioeconomic status (SES) is widely accepted to contribute to school success in educational research, it is rarely addressed explicitly in studies on ELL. In contexts where the private sector o¤ers a range of early programmes, socio-economically advantaged children’s parents are more able to a¤ord them (see national case studies in Enever, Moon, and Raman in press). Several nationally representative samples participated in various projects on Hungarian language learners (Csapo´ and Nikolov 2009; Nikolov 2009; Nikolov and Jo´zsa 2006). In all of these large-scale studies showing a larger picture significant relationships were found between learners’ SES and their language proficiency, and language learning plans, as well as motivation. An innovative study explored the relationships between young language learners’ SES and language development in English in a particular town in Greece (Mattheoudakis and Alexiou in present volume). As the results show, all learners, irrespective of their SES, get private tutoring besides learning English at school. Although their families’ SES impacts significantly on young Greek learners’ level and progress in English, no such relationship is found for their language learning motivation – a finding in contrast with the previously discussed studies. 3.3.4. Languages interacting: First language and additional languages In several studies overviewed in previous sections young learners were bilingual, and the additional language was often their L3, but the relationships across languages were not explored. In other contexts attention is paid to young learners’ L1 and L2 as well. For example, in the multilingual Netherlands, Goorhuis-Brouwer and de Bot (submitted)

20

Marianne Nikolov

report on the outcomes of a project evaluating early English language teaching in Dutch primary education where Dutch is not L1 for many children. Mihaljevic´ Djigunovic´ (submitted) reports on di¤erent interactions between performances on listening, reading and writing tasks of the same format in L1 Croatian and L2 English between earlier and later starters, whereas Csapo´ and Nikolov (in press) scrutinize the relationships between reading in L1 Hungarian and L2 English and German, and how the relationships change over time. Overall, interestingly, few studies explore such relationships, most probably because most researchers focus on the target language rather than the interaction between languages. However, if ELL is meant to be accepted widely and to maintain its position in o‰cial curricula, sooner or later the interaction across languages in literacy development needs to be put on the research agenda. In most studies English is the target language in ELL and early start programmes are responsible for the most recent dynamic expansion of English (Graddol 2006), as was pointed out in the previous sections. In some contexts (e.g., Enever in present volume; Harris and O Leary 2009), mostly where English the o‰cial language, the picture is more varied. Discussions related to the role and status of English are beyond the scope of our discussion, but Carmel’s chapter in the present volume o¤ers an excellent example of how English at an early age has recently become part of special discourse on larger issues. 3.3.5.

L2 skills, vocabulary, grammar and strategies

As more fine-grained studies look into various skill areas, a lot of valuable insights are gained into how young language learners’ listening, speaking, reading and writing, vocabulary and grammar develop. These studies provide useful information on what realistic achievement targets include and how teachers can build on developmental stages in ELL. Studies tend to be either cross-sectional (involving more participants) or longitudinal (repeated cross-sectional or following development of a few learners over a period of time. Both types of studies add valuable insights into important areas of processes and outomes. Several large-scale studies followed learners’ development in certain skill areas, lexicon and grammar. The interested reader is directed to the edited volumes discussed in the previous section on projects involving large samples (e.g., Garcı´a Mayo and Garcı´a Lecumberri 2003;

The age factor in context

21

Harris and Conway, 2002; Harris et al. 2006; Mihaljevic´ Djigunovic´ and Vilke 2000; Mun˜oz 2006; Nikolov and Jo´zsa 2006; Vilke and Vrhovac 1993, 1995). Other studies use a much narrower focus and inquire into one area involving fewer students. This type of research provides more detailed information on the actual learners and is often combined with an exploration of strategy use. Some studies look into how young language learners develop in these domains; others focus more on the outcomes. For example, Szpotowicz (2000, 2009) and Orosz (2009) examine young language learners’ vocabulary in two di¤erent contexts: the first one discusses what factors influence the process of vocabulary learning, whereas the latter the growth in size. All four skills were involved and combined with textbook analyses on two target languages in the Ukraine, where Ukrainian is the o‰cial language, whereas English is a foreign language for minority Hungarian learners (Huszti, Fa´bia´n, and Ba´ra´nyne´ Koma´ri 2009). Reading and literacy development are the focal points in several projects: a new literacy programme is adapted from Australian practice in Norway (Drew 2009); and ways of using picture books are examined with Hungarian learners (Lugossy 2007). Young learners’ writing processes are explored by Moon (2000b) in Bhutan and by Griva, Tsakidirdou and Nihoritou 2009) in Greece; whereas Bors (2008) compares how 14-year-olds’ writing performances can be analyzed combining criterion-based and corpus-based methods. 3.3.6. Exploring ELL classrooms Without information on what happens in classrooms it is impossible to evaluate how achievement targets and principles of ELL are implemented on a daily basis. Some recent studies allow us to gain insights into how children interact with their peers and teachers, and what type of tasks and activites characterize ELL. Although the principles of ageappropriate methodology are clearly outlined in teachers’ handbooks, practice varies to a large extent. In this section a few typical examples are chosen to show the variety in learning opportunties (S stands for students, T for teachers). In the first example the teacher is drilling a structure using English only with a large class. Students respond in L2 English exclusively, and this type of practice was found typical in several classrooms in a classroom based study (Peng and Zhang 2009).

22

Marianne Nikolov

As the curriculum is prescribed for teachers the exact same structures were taught in all schools where observations were conducted. Excerpt 1 Large class of 10-year-olds at a Chinese school (Peng and Zhang 2009) T: Where are you from? S: I from Chongqing. T: No. S: (confused) Oh, I from China. T: No, I am from Chongqing. S: I am from Chongqing. T: Good, sit down. In the second extract young learners are busy working on a drawing task. Both students and the teacher use L1 Croatian mostly and insert a few words in L2 English. The excerpt is borrowed from a conference presention on a new international project (Early Language Learning in Europe; ELLiE) aiming to compare classroom practice in seven European countries with the same instruments to allow researchers to compare their findings. Excerpt 2 A class of 22 learners in a first-grade class (Mihaljeviic Djigunovic and Szpotowicz 2008) S: Teacher, sˇta trebamo nacrtati? [what should we draw?] T: Eeee sad gledam [now I look] carefully. Ovdje su [Here are] two trees, na svakom [on each] tree trebate napraviti plodove [you need to make fruits]. Ovdje znacˇi [So here] apples, ovdje [here] pears. A iznad toga, lijepo c´ete ih obojiti [And above that you will colour them nicely], colour them and match them. Ovako. [Like ˇ ekaj, cˇekaj, cˇekaj. [Wait, wait, wait] Who is big? Tko je od this] C njih [Who among them is] big? Ss: Ovaj tu. Ovaj tu. [This one. This one.] Father. Father. T: Ok, who is strong? To c´ete sami odlucˇiti i spojiti. [You will decide for yourselves and connect.] Vidite ko je [See how it is] small, pa povucˇete crte do njega . . . [then you draw lines to it] Another type of study takes a close look at how children do oral tasks in pairs and how they sca¤old one another’s learning. Pinter (2007) analyzes young learners’ discourse patterns and the role of repetition on a

The age factor in context

23

paired oral task involving picture description: children are comparing their pictures with some di¤erences. Excerpt 3 Two 10-year-old boys describe and compare what they see in their pictures (Pinter 2007: 196) S1: Yes. My bedroom in one dog. S2: Yes. In my bedroom there is a spider. S1: No, two spider. S2: /z/ S1: Spiders. In my livingroom on cat. S2: Yes. These three excerpts are meant to illustrate the wide range of learning opportunities characterizing young language learners’ classrooms. Although classroom studies are extremely time consuming they provide valuable insights into what is involved in ELL and they are essential in teacher education as well. A lot more research is necessary on di¤erent task types and work formats in di¤erent contexts. 3.3.7. Assessment in ELL As ELL is becoming more widely spread, stakeholders want to monitor outcomes (see previous section on large-scale projects). This may be the reason why an increasing body of research has been conducted into assessing young learners’ proficiency in their new language, as documented in a special issue of Language Testing in 2000 and other publications (Hasselgreen 2005; McKay 2006). The most important question concerns what the construct to be measured is, as is discussed in detail by Inbar-Laurie and Shohamy in the present volume. It has to be in line with the curriculum. Many assessment projects are limited to a particular context (e.g., Jalkanen in the present book), whereas others are part of institutional evaluation projects (e.g., Henni Hoti, Heinzmann and Mu¨ller in the present volume). Besides assessing what young language learners can do at the end of their ELL, some experts advocate the use of portfolios (Hasselgreen 2005). What is clear from recent developments in the area of assessing young language learners is that as programmes are less an exception and more typical, achievement targets are also more specific, and accountability is a logical next step. In the European context it would

24

Marianne Nikolov

be desireable to develop age-appropriate refined scales in harmony with the Common European Framework of Reference (Council of Europe 2001) and in similar detail to documents illustrated by Curtain in the present volume. It would also be extremely useful to develop an international database of tasks validated in di¤erent contexts at certain levels of performances to allow experts to share expertise in this highly specific area. Some international initiatives are already in place. ELLiE is a longitudinal project on the introduction of ELL in primary school classrooms in seven European countries. It was launched in response to the rapid expansion of provision for ELL in Europe. It aims to ‘‘clarify what can realistically be achieved in European classrooms where relatively limited amounts of curriculum time are allocated to second/foreign language learning’’ (http://www.ellieresearch.eu/). Schools from Croatia, England, Italy, Netherlands, Poland, Spain and Sweden are involved and data are collected along the same lines with the help of the same instruments. The outcomes of this project, now in its second year, will hopefully allow us to see not only the larger picture, but also the dayto-day details in a variety of contexts. A recent international comparative study looked into levels of achievement of year 8 learners of English in Croatia and Hungary (Nikolov, Mihaljevic´ Djigunovic´, and Otto´ 2008) and showed that earlier starters were slightly, but significantly better on several measures (meaning-focused tests on the four language skills and on pragmatics) than later starters. Croatian learners, however, outsmarted their Hungarian peers despite the fact that they tended to start English later, in larger groups and fewer classes – a sobering result for advocates of ELL.

3.4.

Studies on teachers and teacher education

Teachers and teacher education are key areas for explorations in ELL, but what is known about them is not enough. Therefore, this is where more research is definitely necessary. Teachers in ELL classrooms tend to belong to two groups: (1) generalists or homeroom teachers often with modest proficiency but good skills in age-appropriate methodology and familiarity with the overall curriculum, and (2) language specialists with better proficiency in the target language, but not trained to manage young learners and with less insight into what children know in other curricular areas. Teachers

The age factor in context

25

in the first group tend to teach children the whole curriculum and a new language, whereas language teachers usually do the language classes only. 3.4.1. Research on teachers Several studies have examined teachers in ELL. However, even in the most carefully designed complex studies, teachers’ proficiency and its impact on children’s development are seen as a delicate issue. An exception to this rule is the research conducted in Israeli schools by Shohamy and Inbar-Lourie (2006 cited in Inbar-Lourie and Shohamy in present volume). Their study triangulated classroom observation data with proficiency measures for learners and interviews with teachers, thus showing the complexity of how teachers implement ELL curricula and how di¤erent perceptions of what children can do interact with their goals and behaviour. The way teachers of young learners think about their teaching and how they evaluate their own needs is the focus of some recent studies. Challenges are numerous in contexts where English has recently become part of curricula without appropriate teacher education (for various contexts see chapters in Enever, Moon, and Raman in press). On the Asian continent Goto Butler (2005) explored teachers’ views on communicative methodology in Korea, Japan and Taiwan and found that they were concerned about using communicative activities for a number of good reasons. She also asked them to reflect on their own language proficiency and learnt that they felt their productive skills lagged behind their receptive ones and needed to develop their English to be able to implement ELL programmes (Goto Butler 2004). A similar picture emerges from Moon’s chapter in this volume: she worked with Vietnamese teachers and her case studies throw light on teachers’ needs and possibilities of meeting them. Although in European coutries ELL traditions have existed for a long time, several inquiries reveal that there is room for improvement in both pre- and in-service teacher education. Lundberg’s (2007) research shows how Swedish teachers in remote places benefit from a three-year project and how action research allows them to reflect on their own work. Lugossy (2007) looks at a narrower focus: how teachers involved in a picture book project use storybooks with their learners and how their reflection on their own practice changes over time. Kiss (2000) and Lugossy (2006) compared what teachers know and believe, and

26

Marianne Nikolov

what they actually do in their classrooms. Both studies found that there is a discrepancy between the two planes and these points interact in complex ways with teachers’ motivation and how rewarding they perceive their daily work with young language learners. Very young learners may turn out to be a real challenge, as a study on Polish pre-service English teachers documents (Szulc-Kurpaska 2007). Daily practice with young language learners may not be as rewarding as many advocates of ELL think. An obverstation study applying inteviews with teachers reveals that few teachers found satisfaction in teaching young learners, they wished they could stream them and teach the more able only, or teach older learners; they perceived games and storytelling as a waste of time, and looked forward to ‘‘proper teaching’’ in later years (Nikolov 2008). In many educational contexts the introduction of ELL involves major curriculum change (e.g., Moon’s chapter on Vietnam). Wang’s large-scale survey (in this volume) examines how teachers relate to their innovative curriculum involving not only a challenge in teaching English to young learners, but also a major shift towards learner-centered teaching. Chinese teachers’ enthusiasm towards implementing top-down change will hopefully impact on ELL favourably.

4.

Conclusion

The aim of this chapter was to overview some important areas related to the age factor in recent research. As has been shown, the relevance of the critical period hypothesis, widely assumed to underlie the reasons why ELL is a good idea, is not the most important point of departure for discussions on early start programmes. The age factor needs to be viewed in its context; all conditions have to be taken into consideration, as so many other factors contribute to the implementation of ELL programmes that they vary to a large extent. No wonder curricula and models are so di¤erent and achievement targets also vary in line with them. Overall, what young learners are expected to be able to do as a result of ELL tends to be modest. Many fascinating studies have been implemented in recent years and it is di‰cult to overview all research relevant to ELL. This is a welcome move forward and it is hoped that more studies allow us to understand

The age factor in context

27

what is involved in ELL in specific contexts in di¤erent parts of the world. One point is clear from the overview: young children develop at a slower pace than older students and adults – a finding in line with previous research, but rarely emphasised. The research methods in the studies have become a lot more sophisticated and complex. Longitudinal studies are quite frequent, triangulation is applied in many studies and mixed methods are a lot more typical than a decade ago. Some researchers allow us to gain insights into children’s, teachers’ and other stakeholders’ lived experiences, and this new line in research o¤ers a lot of potential. The emergence of classroom research is also a welcome trend in ELL research, but there is a lack of research into what task types work with what age groups at certain levels. Comparative international projects also o¤er new perspectives. It would be good to know, for example, how portfolios work in ELL. These points lead us to a neglected question: to what extent studies are replicable in various contexts. Replication studies are valuable for many reasons; few of the studies overviewed could be replicated, either because the instruments are not made available, or because the contexts are very di¤erent. As for the target language, English is overwhelmingly in the focus of research but other languages are also given attention and hopefully the interaction between learners’ languages will also become a more popular research area. Despite the fact that the wide range of innovative studies provides a lot of information on ELL, more studies are needed. We do not know enough about how young learners actually move along developmental stages in their new language skills, vocabulary and grammar, how the early chunks from classroom discourse, rhymes and games form the basis of rule-based learning, how their ELL contributes to their strategic, pragmatic and intercultural competence over time, how early exposure to a foreign language impacts on the learning of new languages in later school years on in adulthood, and how it helps them to become happier people. A lot more research is needed on teachers and teacher education. What level of proficiency and type of professional knowledge and personal qualities are necessary for teachers? What do they need to know about young learners and their language development to be able to help them achieve goals set in curricula? There is hardly any research into teachers’ motivation and how it changes over time. Working with

28

Marianne Nikolov

young language learners is rewarding but hard work. More studies should involve teachers in their own classroom research – another area for futher inquiries. The overview did not touch upon language policy, but that is also an area worth exploring in depth. Finally, professionals interested in and researching ELL and related topics would benefit from sharing an international forum perhaps in the form of a refereed online journal.

Acknowledgement The author gratefully acknowledges the support she received while working on this chapter from the Research Group on the Development of Competencies, Hungarian Academy of Sciences (MTA-SZTE Ke´pesse´gkutato´ Csoport).

References Alexiou, Tomai 2009 Cognitive skills in young learners and their implications for FL learning. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 46–61. Clevedon, Avon: Multilingual Matters. Bialystok, Ellen 2001 Bilingualism in Development: Language, Literacy, and Cognition. Cambridge: Cambridge University Press. 2005 Consequences of bilingualism for cognitive development. In: J. F. Kroll and A. M. B. De Groot (eds.), Handbook of Bilingualism: Psycholinguistic Approaches, 417–432. Oxford: Oxford University Press. Birdsong, David 2005 Interpreting age e¤ects in second language acquisition. In: J. F. Kroll and A. M. B. De Groot (eds.), Handbook of Bilingualism: Psycholinguistic Approaches, 109–127. Oxford: Oxford University Press. Blondin, C., Candelier, M., Edelenbos, P., Johnstone, R., Kubanek-German, A., and T. Taeschner 1998 Foreign Languages in Primary and Pre-school Education. London: CILT.

The age factor in context

Bors, Lidia 2008

29

Hungarian 8th-graders’ writing skills in English as a foreign language: A criterion- and corpus-based measurement project. Ph. D. Diss. Department of English Applied Linguistics, University of Pe´cs. Burstall, C., Jamieson, M., Cohen, S., and M. Hargreaves 1974 Primary French in the Balance. Windsor: NFER Publishing Company. Cameron, Lynne 2001 Teaching Languages to Young Learners. Cambridge: Cambridge University Press. Cenoz, Jasone 2003 The influence of age on the acquisition of English: General proficiency, attitudes and code mixing. In M.P. Garcia Mayo and M. L. Garcia Lecumberri (eds.), Age and the Acquisition of English as a Foreign Language, 77–93. Clevedon: Multilingual Matters. Chiswick, B.R., Lee, Y.L., and P.W. Miller 2004 Immigrants’ language skills: The Australian Experience in a longitudinal study. International Migration Review 38 (2): 611–654. Commission of the European Communities 2003 Promoting language learning and linguistic diversity: An action plan 2004–2006. www.eu.int/comm/education/doc/o‰cial/keydoc/ actlang/act_lang_en.pdf (retrieved on 11 December 2005) Council of Europe 2001 Common European Framework of Reference. Cambridge: Cambridge University Press. Csapo´, Benő, and Marianne Nikolov 2009 The cognitive contribution to the development of proficiency in a foreign language. Learning and Individual Di¤erences 19: 203–218. Csize´r, Kata and Judit Kormos 2009 An investigation into the relationship of l2 motivation and crosscultural contact among elementary-school students. In M. Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 62–74. Clevedon, Avon: Multilingual Matters. Cummins, Jim no year Immersion education for the millennium: What we have learned from 30 years of research on second language immersion. http:// www.iteachilearn.com/cummins/immersion2000.html (retrieved on 22 December 2005). Curtain, Helena 2000 Early language learning in the USA. In: Marianne Nikolov and Helena Curtain (eds.), An Early Start: Young Learners and Modern Languages in Europe and Beyond, 191–208. Strasbourg: Council of Europe. Curtain, Helena and Carol Ann Dahlberg 2004 Languages and children – Making the match: New languages for young learners, grades K-8. New York: Allyn and Bacon.

30

Marianne Nikolov

DeKeyser, Robert 2003 Implicit and explicit learning. In: C. J. Doughty and M. H. Long (eds.), Handbook of Second Language Acquisition 313–348. London: Blackwell. DeKeyser, Robert and Jenifer Larson-Hall 2005 What does the critical period really mean? In: J. F. Kroll and A. M. B. De Groot (eds.), Handbook of Bilingualism: Psycholinguistic Approaches, 88–108. Oxford: Oxford University Press. Drew, Ion 2009 Using the early years literacy programme in primary EFL Norwegian classrooms: Factors influencing young learners’ vocabulary acquisition. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 108–120. Clevedon, Avon: Multilingual Matters. Edelenbos, Peter, Johnstone, Richard and Angelika Kubanek 2007 Languages for the Children in Europe: Published Research, Good Practice and Main Principles. http://ec.europa.eu/education/policies/ lang/doc/youngsum_en.pdf http://ec.europa.eu/education/policies/ lang/key/studies_en.html Enever, Janet, Moon, Jayne and Uma Raman (eds.) 2009 Young Learner English Language Policy and Implementation: International Perspectives. Reading, UK: Garnet Eduation/IATEFL. Flege, J. A., Birdsong, D., Bialystok, E., Mack, M., Sung, H., and K. Tsukada 2005 Degree of foreign accent in English sentences produced by Korean children and adults. Journal of Phonetics 33 (3): 263–290. Goto Butler, Yuko 2004 What level of English proficiency do elementary teachers need to attain to teach EFL? Case studies from Korea, Taiwan, and Japan. TESOL Quarterly 38 (2): 245–287. 2005 Comparaticative perspectives towards communicative activities among elementary school teachers in South Korea, Japan, and Taiwan. Language Teaching Research 9 (4): 423–446. Griva, Eleni, Tsakiridou, Helen and Ioanna Nihoritou 2009 A study of FL composing process and writing strategies employed by young learners. In M. Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 132–168. Clevedon, Avon: Multilingual Matters. Garcı´a Mayo, M. P., and M. L. Garcı´a Lecumberri (eds.) 2003 Age and the Acquisition of English as a Foreign Language. Clevedon, Avon: Multilingual Matters. Goorhuis-Brouwer, S., and Kees de Bot submitted Impact of early English language teaching on L1 and L2 development in children in Dutch schools. International Journal of Bilingualism. Garcı´a Lecumberri, M. L., and F. Gallardo 2003 English foreign language sounds in school learners of di¤erent ages. In M. P. Garcı´a Mayo and M. L. Garcı´a Lecumberri (eds.), Age

The age factor in context

31

and the Acquisition of English as a Foreign Language, 115–135. Clevedon, Avon: Multilingual Matters. Garcı´a Mayo, M. P. 2003 Age, length of exposure and grammaticality judgements in the acquisition of English as a foreign language. In: M. P. Garcı´a Mayo and M. L. Garcı´a Lecumberri (eds.), Age and the Acquisition of English as a Foreign Language, 94–114. Clevedon, Avon: Multilingual Matters. Gattullo, Francesca and Gabriele Pallotti 2000 Baseline study on FLT to young learners in Italy. In: M. Nikolov and H. Curtain (eds.), An Early Start: Young learners and Modern Languages in Europe and Beyond, 51–58. Strasbourg: Council of Europe. Graddol, David 2006 English Next. London: The British Council. Hakuta, K., Bialystok, E., and E. Wiley 2003 Critical evidence: A test of the Critical Period Hypothesis for second-language acquisition. Psychological Science 14: 31–38. Harris, John and Mary Conway 2002 Modern Languages in Irish Primary Schools: An Evaluation of the National Pilot Project. Dublin: Institiuid Taengeolaiochhta Eireann. Harris, J., Forde, P., Archer, P., Fhearaile, S. N., and M. O’Gorman 2006 Irish in Primary Schools: Long-term National Trends in Achievement. Dublin: Department of Education and Science. Harris, John and Denise O Leary 2009 A third language at primary level in Ireland: An independent evaluation of the Modern Languages in Primary Schools Initiative. In Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 1–14. Clevedon, Avon: Multilingual Matters. Hasselgreen, Angela 2005 Assessing the language of young learners. Language Testing 22 (3): 337–354. Heining-Boynton, Audrey and Thomas Haitema 2007 A ten-year chronicle of student attitudes toward foreign language in the elementary school. The Modern Language Journal 91 (2): 149– 168. Huszti, Ilona, Fa´bia´n, Ma´rta and Erzse´bet Ba´ra´nyne´ Koma´ri 2009 Di¤erences between the processes and outcomes in 3rd graders’ learning English and Ukrainian in Hungarian schools in Beregsza´sz. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 166–180. Clevedon, Avon: Multilingual Matters. Hyltenstam, K., and N. Abrahamsson 2001 Age and L2 learning: the hazards of matching practical ‘‘implications’’ with theoretical ‘‘facts’’. TESOL Quarterly 35: 151–170.

32

Marianne Nikolov

2003

Maturational Constraints in SLA. In: C. J. Doughty and M. H. Long (eds.), Handbook of Second Language Acquisition, 539–588. London: Blackwell. Ioup, Georgette 2005 Age in second language development. In: Eli Hinkel (ed.), Handbook of Research in Second Language Teaching and Learning, 419– 436. Mahwah, NJ: Lawrence Erlbaum. Jantscher, Elisabeth and Isabel Landsiedler 2000 Foreign language education at Austrian primary schools: An overview. In: Marianne Nikolov and Helena Curtain (eds.), An Early Start: Young Learners and Modern Languages in Europe and Beyond, 13–28. Strasbourg: Council of Europe. Johnstone, Richard 2002 Addressing ‘‘The Age Factor’’: Some Implications for Languages Policy. Strasbourg: Council of Europe. in press An early start: What are the key conditions for generalized success? In: Janet Enever, Jayne Moon and Uma Raman (eds.), Young Learner English Language Policy and Inplementation: International Perspectives. Kent, UK: IATEFL Young Learner Special Interest Group. Julkunen, Kyo¨sti 2000 Young learners’ opinions on content and language integrated learning. In: Jayne Moon and Marianne Nikolov (eds.), Research into Teaching English to Young Learners, 374–385. Pe´cs: University of Pe´cs. Key data on teaching languages at school in Europe. 2008 Edition. (2008). Brussels: Eurydice. Kiss, Csilla 2000 ‘‘Words are but wind, but seeing is believing’’: Do the expressed attitudes of primary EFL teachers reflect their classroom practice? In: Jayne Moon and Marianne Nikolov (eds.), Research into Teaching Young Learners, 171–189. Pe´cs: University of Pe´cs Press. Kiss, Csilla and Marianne Nikolov 2005 Preparing, piloting and validating an instrument to measure young learners’ aptitude. Language Learning 55 (1): 99–150. Krashen, Stephen, Long, Michael and Robin Scarcella 1979 Age, rate and eventual attainment in second language acquisition. TESOL Quarterly 13: 573–582. Larson-Hall, Jeni¤er 2008 Weighing the benefits of studying a foreign language at a younger starting age in a minimal input situation. Second Language Research 24 (1): 35–63. Lugossy Re´ka 2006 Shaping teachers’ beliefs through narratives. In: Marianne Nikolov and Jo´zsef Horva´th (eds.), UPRT 2006: Empical Studies in English

The age factor in context

33

Applied Linguistics, 329–352. Pe´cs: Lingua Franca Csoport. Available online: http://www.pte.hu/uprt/ 2007 Authentic picture books in the lives of young EFL learners and their teachers. In: Marianne Nikolov, Jelena Mihaljevic´ Djigunovic´, Marina Mattheoudakis, Gun Lundberg, and Tanya Flanagan, (eds.), Teaching Modern Languages to Young Learners: Teachers, Curricula and Materials, 77–90. Strasbourg: Council of Europe. Lundberg, Gun 2007 Developing teachers of young learners: In-service for educational change and improvement. In: Marianne Nikolov, Jelena Mihaljevic´ Djigunovic´, Marina Mattheoudakis, Gun Lundberg, and Tanya Flanagan, (eds.), Teaching Modern Languages to Young Learners: Teachers, Curricula and Materials, 21–34. Strasbourg: Council of Europe. Langabaster, D., and A. Doiz 2003 Maturational constraints on foreign-language written production. In: M. P. Garcı´a Mayo and M. L. Garcı´a Lecumberri (eds.), Age and the Acquisition of English as a Foreign Language, 136–160. Clevedon, Avon: Multilingual Matters. Long, Michael 2005 Problems with supposed counter-evidence to the Critical Period Hypothesis. IRAL 43 (4): 287–318. MacWhinney, Brian 2005 A unified model of language development. In: J. F. Kroll and A. M. B. De Groot (eds.), Handbook of Bilingualism: Psycholinguistic Approaches, 49–67. Oxford: Oxford University Press. Marinova-Todd, S. H., Marshall, B., and C. E. Snow 2000 Three misconceptions about age and L2 learning. TESOL Quarterly 34: 9–34. 2001 Missing the point: A response to Hyltenstam and Abrhamsson. TESOL Quarterly 35: 171–176. Mihaljevic´ Djigunovic´, Jelena Submitted Starting age and L1 and L2 interaction. International Journal of Bilingualism. Mihaljevic´ Djigunovic´, Jelena, Nikolov, Marianne, and Istva´n Otto´ 2008 A comparative study of Croatian and Hungarian EFL students. Language Teaching Research 12 (3): 433–452. Mihaljevic´ Djigunovic´, Jelena and Magdalena Szpotowicz 2008 Interaction of contextual and individual variables in instructed early SLA. EUROSLA 18. Aix-en-Provence, France: 11–13 September 2008. Mihaljevic´ Djigunovic´, Jelena and Mirjana Vilke 2000 Eight years after: Wishful thinking vs the fact of life. In: Jayne Moon and Marianne Nikolov (eds.), Research into Teaching English to Young Learners, 67–86. Pe´cs: University Press Pe´cs.

34

Marianne Nikolov

Moon, Jayne 2000a Children Learning English. London: Macmillan Heinemann. 2000b Exploring the e¤ects of drafting on Bhutanese children’s story writing in year 3 and 4 of schooling. In: Jayne Moon and Marianne Nikolov (eds.), Research into Teaching Young Learners, 386–416. Pe´cs: University of Pe´cs Press. Moon, Jayne, and Marianne Nikolov (eds.) 2000 Research into Teaching English to Young Learners. Pe´cs: University Press Pe´cs. Pinter, Annamaria 2006 Teaching Young Language Learners. Oxford: Oxford University Press. 2007 Some benefits of peer–peer interaction: 10-year-old children practising with a communication task. Language Teaching Research 11 (2): 189–207. Moyer, Alene 2004 Age, Accent and Experience in Second Language Acquisition. Clevedon, Avon: Multilingual Matters. Mun˜oz, Carmen 2003 Variation in oral skills development and age of onset. In: M. P. Garcı´a Mayo and M. L. Garcı´a Lecumberri (eds.), Age and the Acquisition of English as a Foreign Language, 161–181. Clevedon, Avon: Multilingual Matters. Mun˜oz, Carmen (ed.) 2006 Age and the Rate of Foreign Language Learning. Clevedon, Avon: Multilingual Matters. Nagy, Krisztina 2009 What primary-school pupils think about learning English. In M. Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 229–242. Clevedon, Avon: Multilingual Matters. Nikolov, Marianne 1999 ‘‘Why do you learn English?’’ ‘‘Because the teacher is short.’’ A study of Hungarian children’s foreign language learning motivation. Language Teaching Research 3 (1): 33–56. 2001 A study of unsuccessful language learners. In: Zolta´n Do¨rnyei and Richard Schmidt (eds.), Motivation and Second Language Acquisition (Technical Report #23, 149–169). Honolulu: University of Hawai’i, Second Language Teaching and Curriculum Center. 2002 Issues in English Language Education. Bern: Peter Lang. 2003 Angolul e´s ne´metu¨l tanulo´ dia´kok nyelvtanula´si attitűdje e´s motiva´cio´ja. [Attitudes and motivation of learners of English and German] Iskolakultu´ra 13 (8): 61–73. 2008 ‘‘Az a´ltala´nos iskola, az mo´dszertan!’’ Also´ tagozatos angolo´ra´k empirikus vizsga´lata [‘‘Primary school means methodology!’’ An empirical study of lower-primary EFL classes]. Modern Nyelvoktata´s 10: 1–2, 3–19.

The age factor in context

2009

35

Early modern foreign language programmes and outcomes: Factors contributing to Hungarian learners’ proficiency. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 90–107. Clevedon, Avon: Multilingual Matters. Nikolov, Marianne, and Helena Curtain (eds.) 2000 An early start: Young learners and modern languages in Europe and beyond. Strasbourg: Council of Europe. Nikolov, Marianne, and Jelena Mihaljevic´ Djigunovic´ 2006 Recent research on age, second language acquisition, and early foreign language learning. Annual Review of Applied Linguistics 26: 234–260. Nikolov, Marianne, Mihaljevic´ Djigunovic´, Jelena, Mattheoudakis, Marina, Lundberg, Gun, and Tanya Flanagan (eds.) 2007 Teaching Modern Languages to Young Learners: Teachers, Curricula and Materials. Strasbourg: Council of Europe. Nikolov, Marianne, and Krisztia´n Jo´zsa 2006 Relationships between language achievements in English and German and classroom-related variables. In M. Nikolov, and J. Horva´th (eds.), UPRT 2006: Empirical Studies in English Applied Linguistics, 197–224. Pe´cs: Lingua Franca Csoport, PTE. Available online: www.pte.hu/uprt/ Nunan, David 2003 The Impact of English as a Global Language on Educational Policies and Practices in the Asia-Pacific Region. TESOL Quarterly 37: 589–613. Orosz, Andrea 2009 The growth of young learners’ vocabulary size. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 181–194. Clevedon, Avon: Multilingual Matters. Papadopoulou, Charis-Olga 2007 German as a second foreign language in Greek compulsory education: Curriculum and continuity. In: Marianne Nikolov, Jelena Mihaljevic´ Djigunovic´, Marina Mattheoudakis, Gun Lundberg, and Tanya Flanagan, (eds.), Teaching Modern Languages to Young Learners: Teachers, Curricula and Materials, 91–104. Strasbourg: Council of Europe. Paradis, Michael 2004 A Neurolinguistic Theory of Bilingualism. Amsterdam: John Benjamins. Peng, Jing, and Lily Zhang 2009 An eye on target language use in elementary English classrooms in China. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 212–228. Clevedon, Avon: Multilingual Matters. Pinker, Steven 1994 The Language Instinct. London: Penguin.

36

Marianne Nikolov

Prahbu, N.S. in press Teaching English to young learners: The promise and the threat. In: Janet Enever, Jayne Moon and Uma Raman (eds.), Young Learner English Language Policy and Implementation: International Perspectives. Kent, UK: IATEFL Young Learner Special Interest Group. Scovel, Thomas 1988 A Time to Speak: A Psycholinguistic Inquiry into the Critical Period for Human Speech. Rowley, MA: Newbury House. 2000 A critical review of the critical period research. Annual Review of Applied Linguistics 20: 213–223. Singleton, David 2001 Age and second language acquisition. Annual Review of Applied Linguistics 21: 77–89. 2005 The critical period hypthesis: A coat of many colours. IRAL 43 (4): 259–268. Singleton, David, and Lisa Ryan 2004 Language Acquisition: The Age Factor (2nd edition). Clevedon, Avon: Multilingual Matters. Skehan, Peter 1998 A Cognitive Approach to Language Learning. Oxford: Oxford University Press. Shohamy, Elana and Ofra Inbar-Lourie 2006 The teaching of EFL in the first grade: EFL teachers versus homeroom teachers. The Pedagogical Secretariat, Jerusalem: The Israeli Ministry of Education (in Hebrew). Szpotowicz, Magdalena 2000 Young learners: Do they remember vocabulary? In: Jayne Moon and Marianne Nikolov (eds.), Research into Teaching Young Learners, 360–372. Pe´cs: University of Pe´cs Press. 2009 Factors influencing young learners’ vocabulary acquisition. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes, 195–211. Clevedon, Avon: Multilingual Matters. Szulc-Kurpaska, Malgorzata 2007 Teaching and researching very young learners: ‘‘They are unpredictable’’. In: Marianne Nikolov, Jelena Mihaljevic´ Djigunovic´, Marina Mattheoudakis, Gun Lundberg, and Tanya Flanagan, (eds.), Teaching Modern Languages to Young Learners: Teachers, Curricula and Materials, 35–46. Strasbourg: Council of Europe. Tomasello, Michael 2003 Constructing Language. Cambridge MA: Harvard University Press. Ullman, Michael 2001 The neural basis of lexicon and grammar in first and second language: The declarative/procedural model. Bilingualism: Language and Cognition 4: 105–122.

The age factor in context

37

Urponen, Marja Ingrid 2004 Ultimate attainment in postpuberty second language acquisition. Ph. D. Diss. Boston University. Va´go´, Ire´n 2007 Nyelvtanula´si utak Magyarorsza´gon [Language learning paths in Hungary]. In: Ire´n Va´go´ (ed.), Fo´kuszban a Nyelvtanula´s [Focusing on language leaning], 137–173. Budapest: Oktata´skutato´ e´s Fejlesztő Inte´zet. Vickov, Gloria in press Learners’ own cultural identity in early language learning. In: Marianne Nikolov (ed.), Early Learning of Modern Foreign languages: Processes and Outcomes. Clevedon, Avon: Multilingual Matters. Vilke, Mirjana, and Ivonne Vrhovac (eds.) 1993 Children and Foreign Languages. Zagreb: Faculty of Philosophy, University of Zagreb. 1995 Children and Foreign Languages II. Zagreb: Faculty of Philosophy, University of Zagreb. Wong Fillmore, Lily 1998 Supplemental declaration of Lily Wong Fillmore. http://www.humnet.ucla.edu/humnet/linguistics/people/grads/ macswan/fillmor2.htm (retrieved on 22 December 2005)

2.

Early foreign language learning: Published research, good practice and main principles

Peter Edelenbos and Angelika Kubanek

1.

Introduction

The European Commission, in its publication Promoting Language Learning and Linguistic Diversity: An Action Plan 2004–2006 (2003), had expressed its intent to extend, consolidate and develop the early learning of one or more foreign or additional languages in each of the EU member states. The Commission’s working paper Progress towards the Lisbon Objectives in Education and Training (March, 2005) outlines the strategy and steps to be taken by 2010, in order to make education and training systems in Europe the best in the world, and includes a chapter devoted to foreign language teaching. Two aspects of this working paper were directly relevant to a study o¤ered for tender in April 2005 (EAC 89/04, Lot 1), The main pedagogical principles underlying the teaching of languages to very young learners. Firstly, there is a large target audience for such insight into the main pedagogical principles of early language learning (ELL), as the current teacher age profile suggests that around 1 million teachers are needed in Europe. Secondly, present educational statistics do not give an adequate account of organisation, content or good practice. As stated in the working document, current indicators (i.e. the number of pupils per class) do not deal with the core aspects of early language learning: learning materials, teaching methods and quality. The study focused on the main pedagogical principles underlying the teaching of languages to the very young. The diversity of early language learning throughout Europe and the di¤erences at a conceptual level were not to be underestimated, necessitating the inclusion of a description of good practice and a review of research into early language learning. These two complementary activities aimed to better trace, clarify, compare and describe the pedagogical principles in their relevant sociocultural contexts. In addition, the review of research can be seen as

40

Peter Edelenbos and Angelika Kubanek

a sorely needed update on the Blondin et al. report from 1998. The description of good practice served one major purpose: a homogenous, high-quality description of a range of initiatives leading to good practice can help all practitioners and scholars understand the beauty of teaching languages to very young children. This contribution presents some of the findings of the study. In the Language Portal of the European Union europa.eu/languages the reports can be found in French, English and German.

2.

Method

The structure of the study was clearly dictated by the tender. Because of the European scope of the study close cooperation was sought with internationally operating network partners. For the research update, close collaboration was established with a network of internationally operating partners. These partners operated in di¤erent geographical and linguistic academic settings and were asked to participate in intensive mini-conferences. Two types of procedures were seen as inadequate: first, pure desk-top research and second, the sole use of online-questionnaires or interviews over the telephone. For the research overview, of course, these techniques were among those used, but for the acquisition of data on good practice the international contributors obtained their information, in part, directly from teachers during in-service courses for early language learning. The object of study, according to the explicit wish of the European Commission, needed to be main stream education and not immersion programs. For good practice there should not be a sole focus on those projects that stood out, but rather on the good practice that could be observed in several countries.

3.

Published Research

The study’s first product is a review of the main research in the field, in Europe and elsewhere, since 1999. The review of research is a prerequisite that influences the project by adding to the existing body of knowledge.

Early foreign language learning

41

The year 1999 was chosen so as to complement the 1998 review of research concerning foreign languages in primary and pre-school education by Blondin, Candelier, Edelenbos, Johnstone, Kubanek-German, and Taeschner. In this way, the cohesion and continuity of reviews of research are secured. The present study shifts the focus to the influence that available evidence has on underlying pedagogical principles and the organisation of foreign-language teaching, classroom practice, and teacher training. The full load of available research in this field is reviewed, with a twofold purpose: (1) to collect research on early language learning, especially on ‘good practice’ across member states of the EU and (where appropriate) elsewhere, in respect to modern foreign languages (main category in pre-primary and primary education) and (2) to produce a consolidated report that brings together key findings and, where possible, draws links between good practice and particular outcomes e.g., attainments, proficiency and motivation. A number of activities were undertaken to acquire publications. Some of these were: (1) Identifying high-quality research published in each member state, associated countries or elsewhere via well-established databases such as ERIC and PSYCHLIT; (2) Obtaining research publications not written in English, French or German. Often, these publications are neither easily available nor accessible to an international audience because of the language they are written in. (3) Synthesizing the research publications into practical and scientific overviews, analogous to the annual reviews by Johnstone (2000). The information about research was collected by the project team with strong support from a small group of experienced and expert researchers drawn from various parts of Europe, who met with the project team and also prepared and submitted a large amount of evidence in the form of summaries of research reports considered to be potentially useful. It was found that some areas of potential interest were covered but others less so, and consequently it was decided to draw on research conducted elsewhere in the world if this added some new insight which might be appropriate to European circumstances.

4.

Good Practice

The sub-chapter on the concept of good practice and quality was developed in tandem with a search for examples and their classification. A normative approach, i.e. classifying the examples according to pre-

42

Peter Edelenbos and Angelika Kubanek

determined conditions for success, was one option. However, this would be inappropriate for both epistemological and ethical reasons: the authors would assume the position of arbiters over countries and initiatives they do not and could not possibly know first-hand. Furthermore a normative approach would probably lead to an unbalanced outcome i.e. some positions are overstated and some countries would not be covered or only to a limited extent. Instead, a realistic picture of ELL was sought through the principle of triangulation (Guba 1978), where di¤erent sources and perspectives are combined. The procedure thus comprised several instruments and moments of reflection, validation and renewed data gathering. The first step was to carry out a major literature search, and the information gathered led to the development of two instruments; a) a general questionnaire and b) a description sheet of realisations of good practice. The outcomes of the questionnaire were discussed with experts at a validation meeting. In addition to this there were informal consultations with experts throughout Europe. The questionnaire and the description sheet were sent to each of the national representatives from the Goethe Institut as well as to experts in the networks of the three key researchers. For the general questionnaire quality indicators were taken from o‰cial and academic literature. For the description sheet, a rather detailed explanation of what might constitute good practice preceded the actual description sheet sent out to the informants. The general questionnaire aimed to retrieve information on the status quo regarding provisions, educational principles and assessment, and was sent to one or two experts per country. Their responses were descriptions of what happens in the field. The examples sent in complied with the criterion of relative stability, meaning that the measures or initiative described were sent in with the agreement of the teachers. The initiatives described by the informants often involved large groups of teachers. The authors used available materials to add to the information provided by the respondents, and the remaining references to good practice apply to large numbers of children, teachers and audiences.

5.

Pedagogical Principles

The study’s third product is a description of specific pedagogical principles underlying the teaching of languages to very young learners. The principles must be clearly identifiable, i.e. they must be grounded in empirical research and professional judgements of ‘good practice’ that

Early foreign language learning

43

are meaningful and accessible to classroom teachers. A provisional map of pedagogical principles and any contextual factors which seem to be associated with them was drawn up in consultation with research colleagues (product 1), and the agencies and experts in the field (product 2). The map of pedagogical principles was then extended and refined with more systematic reference to sets of principles from countries such as Germany, Holland, Denmark, Scotland, Italy, Greece, Hungary and Bulgaria. The product is a preliminary set of pedagogical principles. This set consists of political, socio-cultural, psycholinguistic, didactical, methodological and pedagogical principles. The complete set was submitted to two groups for discussion. The first group, made up of five research and agency partners, met during a two-day seminar. The set of principles was discussed in relation to the descriptions of good practice with five representatives. The second group consisted of high-profile educationalists and educational policy makers from various towns, regions (local and border), national governments and supra-national organisations. In all, 56 persons were addressed and asked to cooperate in this project. The results of the survey were presented in tables and overviews during a 2-day seminar attended by five educationalists.

6.

Published research

The research reports and published articles could be divided into five major groups. For each group some examples are provided: (1) Studies on provision: (a) Starting age; (b) Intensive programs; (c) Continuity from primary into secondary education; (d) The professional needs of teachers; (e) Computer assisted learning; and (f ) New national initiatives. (2) Studies on learners’ progression and learning processes: (a) Listening; (b) Speaking; (c) Pronunciation; (d) Reading; (e) Writing; (f ) Self assessment; (g) Teaching and learning strategies; and (h) Social learning. (3) Studies on attitudes, motivation and other a¤ective factors: (a) Attitudes and motivation; (b) Socio-economic status; (c) Gender; and (d) Influence of minority languages spoken in the family. (4) Assessment studies: (a) National assessments; and (b) Extended studies on listening and reading. (5) Studies on language awareness and intercultural learning.

44

Peter Edelenbos and Angelika Kubanek

Three research domains seem to be of particular interest: studies of the starting age, studies of teaching and learning strategies and studies of attitudes, motivation and other a¤ective factors. Studies of the use of portfolios and learning strategies reflect new perceptions on competences taught in early language learning. An example is the following description of research on motivation (Edelenbos, Johnstone, and Kubanek 2007: 57). Nikolov (1999, 2002) followed three cohorts of children in a longitudinal study. They were investigated for eight years and were taught by the same teacher. It was found that learners’ motivation could be maintained by intrinsically interesting and cognitively challenging tasks, and that their attitudes were shaped by what happened in the classroom. In a replication study, Hardi (2004) investigated the characteristics of Hungarian learners’ language-learning motivation, in the case of two cohorts between ages 9–14. The research focused on likes/dislikes in classroom learning and on language choice and language learning. Both cohorts were highly motivated but in those studying English as an optional subject, instrumental motivation was dominant, whereas with those younger learners in compulsory foreign language groups, classroom-related motives were dominant. Mihaljevic´ Djigunovic´ (1998) investigated the attitudes, motivation, language anxiety and attributions of children in Croatia learning a foreign language (English, French, German or Italian). Her longitudinal findings showed that young learners in the Croatian experimental project started out with and maintained positive attitudes towards learning a foreign language over an extended period of time. By grade seven some young learners showed signs of demotivation caused by di‰culties in learning and dissatisfaction with certain aspects of the teaching setting. Learners attributed both success and failure to sources they could control (e¤ort and task enjoyment). In a large-scale cross-sectional study Nikolov (2003) o¤ered a questionnaire to 28,000 Hungarian young learners. She established that communicative activities with a focus on meaning were the least frequent to occur in all year-groups surveyed. What children enjoyed was watching videos, listening to tapes, pair and group work, and acting. Lamb (2004) investigated the motivation of Indonesian children, aged 12–13, learning English. Very high levels of motivation were found, but the traditional concepts of instrumental and integrative motivation tended to be indistinguishable. This was possibly because integrative motivation tended to be seen not as integration into Anglophone

Early foreign language learning

45

culture but rather into global culture. As such, individuals seemed to aspire to a ‘bicultural’ identity involving a global and a local version of the self. Marschollek (2002) found that the classes he studied maintained their motivation throughout primary school. Where there was a decrease of interest, this showed in those children who had not had the opportunity to meet foreigners. The contact with native speakers was found to be the strongest factor as regards motivation. There was evidence for a significant influence of social context and parents on interest to learn a foreign language. The didactic consequences, according to the author, include the need to acknowledge all types of success in learners and to give opportunities for autonomous learning. Kennedy, Nelson, Odell, and Austin (2000) found that students at elementary school in the USA with foreign language programmes showed positive attitudes to school, to perceived di‰culty in language acquisition, to perceived desirability of foreign language study, and had positive cultural views, self-esteem and confidence, in contrast to their peers who were not learning a foreign language at that stage. It was concluded that learning a foreign language at elementary school helps develop motivation to participate, to persist and to succeed in their foreign language studies. As already stated in Blondin et al. (1998), the main gains in early language learning lie in the development of positive attitudes and motivation. At the same time though, research published since then adds some important detail to the picture. None of the research which has been surveyed should be considered as demonstrating universal evidence which can lead directly to particular desirable practices. This is not the fault of the researchers; it has much more to do with the nature of research in our highly diverse, complex, fast-changing and contested field which cannot be as controlled and scientific as is research in various other domains of human investigation. However, there are a few key insights that can be derived from the research publications that have been reviewed. An early start can confer considerable advantages on children by activating natural languages acquisition mechanisms they possess, by a¤ording them more time overall and by providing them with a linguistic and intercultural experience which can have a beneficial formative influence on their cognitive, social, cultural, acoustic, linguistic and personal development (including qualities of persistence and participation) and on their sense of self. An early start by itself however guarantees nothing; it needs to be accompanied minimally by good teaching, by a supportive environment and

46

Peter Edelenbos and Angelika Kubanek

by continuity from one year to the next, taking children smoothly from pre-primary to primary, and from primary into secondary education. Research provides some evidence that intrinsic motivation is not only created by fun and games but also by intellectual challenge and feelings of satisfaction with personal achievement. Children seem to progress naturally through a number of stages of development in their target language, a process which seems to be driven, in part at least, by internal mechanisms which all children possess; it also seems from some research studies that all children progress through these stages in the same sequence but with considerable variation in the rate at which this is achieved. Natural progression of this sort does not seem to be smooth and unidirectional (i.e. upwards) but seems to contain periods of stagnation (plateau-e¤ect or fossilisation) and periods of confusion. It does not sit well with notions of progression built into course-books, national syllabi or transnational frameworks. Two of the individual learner characteristics which seem most strongly associated with proficient performance in the target language are motivation, and aptitude. By the age of twelve or so, aptitude seems to account for significant variation in performance. Therefore, it is all the more important to understand that aptitude is not something which is fixed from birth. It is in fact a quality which can be developed and increased by the process of primary schooling. That is why in the early stages of primary schooling, the evidence suggests that it is helpful not to focus exclusively on fun activities based solely on language-use but to supplement this with activities which help children to internalise meaningful concepts about language (thereby a¤ording them a meta language with which to regulate their language-learning and -use) and also to enhance their sensitivity to sound and to underlying pattern, all of which seem to be associated with language-related aptitude. While confidence is a very important quality to encourage, it is important also to bear in mind a key message of Blondin et al. (1998), namely that many children (even some in immersion settings) do not seem to progress substantially beyond a stage of being able to produce largely learnt-by-heart, pre-fabricated utterances. In order to help them gain a more flexible command of their target languages from an early age, it is helpful to alternate between talk-activities focused on confident, fluent expression and those more focused on accuracy of form and meaning; and also between activities requiring spontaneous performance and those where performance can be planned and prepared.

Early foreign language learning

47

Like all learners, children benefit greatly from appropriate feedback which helps them monitor their progress. The feedback may be positive (in order to o¤er encouragement) or it may be constructively negative (e.g., helping them identify and then repair errors of grammar), provided that this does not undermine confidence or self-esteem. The feedback need not necessarily be given by the teacher through didactic instruction, it may arise from processes whereby children themselves are encouraged to focus on linguistic form as well as on meaning, to notice for themselves particular features of the languages they encounter, and to evaluate their own and their peers’ output. The evidence suggests that young children can benefit in a number of ways if their initial experience is not restricted to ‘listening’, ‘speaking’ and ‘doing’ but includes a gradual and systematic introduction to reading and writing from an early point. The reading and writing ought desirably to fit into a broader school approach to literacy development and equally may focus on local community languages as well as on the particular language the children are learning. Children at all levels (from kindergarten onwards) benefit from being encouraged to reflect on their learning. This includes reflecting on how they might learn from and support each other, as well as on how they might learn from their teacher or others, or how they might learn things for themselves. It is worthwhile for teachers to help their pupils to document, share, evaluate, supplement and refine how they learn, not only in respect of languages but also in respect of their other learning at school. Stories play an important role in children’s language-learning, not only because of children’s natural interest in stories, nor in the undoubted appeal to their imagination, but also because stories embody a narrative structure of discourse which can be useful for learning more generally. The evidence suggests it is worthwhile for teaching to focus not only on the micro-structure of grammatical forms but also on the macro-structure of discourse, including the discourse of narrative. Socio-economic status has been shown in some studies to have a negative influence on the early learning of a modern language at primary school. The term ‘socio-economic’ may in fact subsume a range of background characteristics, not only financial in nature but also possibly reflecting local culture, ethnicity, first language or dialect, employment, peer-group influence, family and other a‰liations. It is extremely important that children should not be penalised for bringing such characteristics with them into their early schooling.

48

Peter Edelenbos and Angelika Kubanek

The language-related outcomes which can reasonably be expected of children by the end of their primary schooling are strongly influenced by the particular models of language curriculum which are adopted. The outcomes will vary considerably as between (for example) 1) teaching a particular target language for roughly one hour per week from various starting points in pupils’ primary school education, and based mainly on working through given textbooks or syllabuses; 2) the same as 1) except that there is a more flexible, learner-centred approach in which attempts are made to link pupils’ language-learning to their personal interests and circumstances, and to their learning of other aspects, such as science or geography; 3) one which initially at least prioritises ‘language and intercultural awareness’ over the teaching of one particular language; and 4) forms of bilingual education, including partial or total immersion in which the factors of ‘time’ and ‘intensity’ are substantially increased. Research tells us clearly that model 4) generates the highest levels of target-language proficiency but is unlikely to be universally applicable. If proficiency in the target language (and the cognitive and intercultural benefits which arise from this) is the main aim, then one or other form of model 4) seems to be the strongest option, since it maximises the key factors of ‘time’ and ‘intensity’ (the greater intensity arising from children having not only to learn the target language but also vital subject matter through the medium of the target language).

7.

Good Practice

An enormous range of invaluable activity has taken place across Europe at various levels, proof of the dedication of large numbers of professionals and of the perceived importance of the theme of early languagelearning. Based on examples which were, by and large, contributed by experts who had first-hand information about particular initiatives, a landscape of primary language-teaching was sketched. The survey is divided into four sections: Creating conditions for good practice, teacher training, teaching the classroom and dissemination. Within the four sections there are sub-classifications. It is suggested that the survey be read on two levels: a) as a collection of examples of good practice, each within a unique language context, and b) on a more abstract transversal level and therefore non-language specific. This second approach can partly compensate for the fact that only a selection of European languages could possibly be mentioned.

Early foreign language learning

49

The two main criteria given to the informants who searched for what to report for their country were stability and acceptance. These criteria were very broad and meant to guide the informants away from unique or very new initiatives. Some recent initiatives were, however, added to the incoming material in order to highlight ongoing dynamics. The summaries which were sent to the authors gave evidence of various underlying ideas about what is considered good. Some contributors sent examples which stand for the expressed policy of the EU (e.g., content and language integrated learning [CLIL], multilingualism), others sent examples of setting up structures, others again more material for direct use in the class which takes some workload o¤ teachers. Some informants reported mainly large-scale initiatives. In Blondin et al. (1998), minimal conditions for success were listed. Similarly, many of the descriptions given are about the creation, in the past years, of conditions for successful teaching and learning. A range of possible methods to collect information on perceived good practice was listed in the introductory chapters of the study. It was necessary to use a desk-research and expert-consultation approach, given time and budget restraints. The advantage of this method is that the examples given as illustrating good practice represent what highly experienced educationalists considered as being worthy of inclusion in a report of this kind meeting the criteria of stability and/or acceptance. What is reported is, therefore, no catalogue of ‘what should be’, of opinions about quality, but gives evidence of ‘what has been’ or is perceived to work in a particular context. There is no automatic way of achieving good practice, be it at the organisational, the teacher-training or classroom level. Therefore, neither the examples nor this part of the conclusions should be read quantitatively or with an if-then attitude. The motivation and learning success of the children cannot be predicted by ticking 10 instead of 7 quality indicators in a checklist. Nor is it possible to predict good practice in a uni-directional way, stating that success is guaranteed if x, y and z conditions are fulfilled. Creating and even improving good practice occurs as an interplay of factors through negotiations and interpretations, as a process over time. Four areas of good practice were distinguished for the classification: Conditions, teacher training, teaching, and dissemination. Conditions include the provision of information, legal provisions and/or transnational or national recommendations or directives, financial management and organisational arrangements, with a predominance of re-

50

Peter Edelenbos and Angelika Kubanek

ported good practice in the last area. In the teacher education section (both pre-service and continuing professional development) examples for internationalisation were presented and a number of types of organisational changes considered to be working well by the experts were consulted. As for teaching, the majority of examples focused on teaching material, but also the intercultural aspect came to the foreground, augmented by a small paragraph about teaching in less favourable conditions. Dissemination of images of ELL and information for a wider public were the topic of the fourth section. The chapter ended with a tentative list of quality indicators. The examples reported by the informants focused to a large extent on the tangible. Good practice at the organisational level was reported in many facets. The second tangible element is teaching material for use by the children. Again, the large variety is documented in the section on teaching. In contrast, little was reported on how a specific research study with its recommendations was felt to contribute to improving practice. It is clear that even if research results are heavily publicised like the national assessment studies, the e¤ect of such findings is not immediate. Still, It seems that the discourses of research and the discourse of practice are perceived as separate occurrences. The need currently voiced to improve diagnostic competence can lead to an intensified dialogue with researchers. Approaches from other disciplines can help refine knowledge and widen the horizon for practice. The e¤ect of story-telling can be understood better if teachers have heard about narrative from the perspective of literary studies. From the sciences, forms of visualising can be learnt from the Children’s Universities. Popular experiments, for example in acoustics, or from computer animations of speech, might even be concretely enlightening when planning to stress the ELL principle ‘training the ear’. Learning from outside of one’s field might seem demanding but the fresh look can be of help. Commercial textbooks were not considered in the study; nevertheless, the role of textbook writers should be mentioned. Views of good teaching in ELL as presented by foreign language educators at universities and colleges are picked up by textbook writers who are, however, more bound by the stated requirements of particular curricula. It could be said that textbooks are a manifestation of an influence of theory of education and research and therefore function as change factors. The European Language Portfolio has been widely recommended,

Early foreign language learning

51

and as a consequence, textbook publishers are adding such material to their books. The educational discussion about scholastic achievements has a¤ected ELL directly in that textbook publishers produce additional material for assessment in countries where such material was not common. While it is acknowledged that publishing houses need commercial success, they have a responsibility, for example, not only in designing teacher handbooks but also in selecting cultural content. It is worth arguing that however good a textbook may be, it is highly unlikely that a learner-centred approach can possibly work with young children if it is based solely on a textbook. Good practice in teaching an additional language to young children seems, from the evidence supplied, to require some flexibility in approach in order to cater for each individual child’s actual environment, interests and needs and to exploit unpredictable opportunities as they arise. This has considerable implications for the role of the teacher in respect of, for example, their confidence in their own capacity to depart from the ‘safety’ of the language items which are set out in sequence in the particular course-book. Evidence exists that bringing in visitors and providing intensive cultural experiences motivate and kindle a will to communicate in the children which is di¤erent from day-to-day lessons. The opportunities for authentic encounters should be increased. Even though agencies exist which provide help, there seem to be barriers: Perhaps due to extra time for preparation, perhaps because crossing the border has become too seemingly normal to arouse interest, perhaps because the border region is considered not attractive compared to a major town in the mainland. Solutions should be sought which make encounters and visits of guests more feasible for teachers. These might include, for example, ad hoc financial support for a cover teacher or a teacher-training student or an ‘animateur’ or actor, when a class goes on a trip for a day. The range of visitors in class can be widened, for example, by a technician from another country temporarily working in a company where a child’s parent is employed, by inviting target-language actors or singers who happen to be in a town. Creating such opportunities should not be left to the classroom teacher alone. In addition, crossing borders may be viewed in a virtual sense, through the use of new technologies which put pupils and their teachers from two or more di¤erent countries in touch with each other for a variety of purposes, including the undertaking of joint projects. Project work is a major characteristic of primary school education, and bi- or

52

Peter Edelenbos and Angelika Kubanek

multilingual project work of a virtual, intercultural sort has potentially much to o¤er. The benefits to a child’s proficiency in their target language as well as their social and intercultural awareness are potentially considerable. The benefits to teachers could be equally great in that their virtual partner teachers could prove to be an invaluable source of advice and assistance in supporting the teacher’s own language and intercultural development. If measures of the sort suggested above were taken, then good practice might well be considered as exemplifying important principles of early language teaching such as ‘respect for otherness’, ‘tolerating cultural dichotomies’ or ‘dealing explicitly with prejudice’. Increasing interculturality also means facilitating short visits of primary teachers to a school in a di¤erent country. Even though o‰cial EU structures exist and are successful, perhaps more might be done at national and other levels to complement these. The majority of primary teachers are women who might have younger children, and ways might be found of supporting their children as part of the travel package.

8.

Pedagogical principles

In order to arrive at a set of validated principles of ELL, a sophisticated iterative process was adopted in collaboration with a range of experts in the field which allowed an initial formulation of principles to be developed and then progressively to be refined. Of the four di¤erent interpretations of the term ‘principle’ which suggested themselves, the two found to be most compatible with the present study were principles in the sense of ‘aims of ’ or ‘reasons for’ and principles in the sense of ‘maxims for action’. 8.1.

Principles as ‘aims of ’ or as ‘reasons for’

The evidence from o‰cial documents identified the following principles as ‘aims of’ or as ‘reasons for’: (1) to develop the hidden multilingual potential of every child which allows a natural acquisition of another language as early as possible; (2) to improve the language competences of students who will be able to take part in European mobility projects and become equal business partners in the global world; (3) to foster

Early foreign language learning

53

positive attitudes towards language-learning; (4) to improve the level of communicative competence reached by students through their educational system; (5) to raise awareness of language-learning at a young age and thus motivate children for learning additional languages; (6) to establish links between language-learning in primary and secondary education; (7) to establish a su‰cient degree of proficiency and awareness in two languages so as to make it easer to begin with a third language and a new challenge; (8) to respond to the demands of parents that their children should learn languages from kindergarten onwards.

8.2.

Principles as maxims for action

In addition, a number of principles were identified which may be viewed as maxims for action which are stated at a more general level than most of the measures identified as being suitable for good practice and which tend to convey some sense of the assumptions which lie behind the actions which are recommended. These assumptions may be of a pedagogical-didactic sort, or may be more oriented to acquisition or learning, or may be more philosophical in nature. A few of these maxims for actions are: (1) stimulate and foster children’s enjoyment to learn an additional language; (2) promote basic skill communication; (3) build on and sustain the initial motivation which children bring with them; (4) provide particular language-activities which are adapted to suit the age- and stage-levels of children; (5) provide meaningful contexts and relevant thematic areas; (6) ensure that comprehension precedes production; (7) make provision for holistic language learning; (8) cater for training of the ear and training of pronunciation; (9) help pupils become aware of the relationship between the sound and written systems of the languages they know and are learning. It should be emphasised that the two sets of principles as set out above are put forward by the authors not as their own final, personal and fully considered opinion but rather as reflecting the substantial iterative process which took place in consultation with eminent experts. What is not at all clear is the extent to which these principles have been explicitly understood and accepted as such by the majority of teachers. It seems highly likely that in many cases they are implicitly rather than explicitly understood.

54

Peter Edelenbos and Angelika Kubanek

9.

Overview of important main principles that are either related to ELL or are unique for ELL

Principles with a clear relationship to the personality of the child and its cognitive functioning – pedagogical cognitive orientation. The main principles are: (1) frequent exposure to the foreign language; (2) taking into account the full range of learner characteristics; (3) encouraging tolerance towards others and providing familiarity with di¤erent sets of values. General principles related to learning – didactical concepts and instruction. The main principles are: (1) taking into account learning strategies and learning styles of children; (2) providing meaningful contexts and relevant thematic areas; and (3) comprehension precedes production. Principles related to language learning – psycholinguistic issues and methodological transformations. The main principles are: (1) holistic language learning; (2) a visual approach and multisensory learning; (3) age-related taking full advantage of the children’s physical predispositions. Principles unique for early language learning are: (1) more comprehension than production; (2) a positive motivation to learning; (3) training of the ear; (4) training of pronunciation; and (5) explanation, and practice in noticing the relationship between phonemes and graphemes of the other language and in comparison to the mother tongue. Although it has been pleasing to report a considerable increase in published research studies since Blondin et al. (1998), in fact the research that has been published provides relatively little information on the extent to which principles such as those set out above are understood, accepted or implemented. There is a strong suspicion that for this to happen, a great deal of discussion and elaboration would be needed, involving all the main stakeholders in early languages learning, such as transnational and national authorities, teachers, teacher educators, advisers, inspectors, school management, not forgetting pupils themselves and their parents.

10.

Issues for future research

Some of the best empirical research on ELL has focused on a limited number of aspects, including language acquisition, motivation and

Early foreign language learning

55

learning strategies. What the field continues to lack are integrated research and development programmes, and a comprehensive view of the main variables that determine success. Such broader research would provide a basis for good practice in a range of key areas that have, so far, been neglected. These are issues of critical importance to the continued success of early language learning: (1) The e¤ects of an early introduction to reading and writing in the child’s mother tongue, and of early foreign language learning; (2) The development of discourses on community/ regional languages and their cultures; (3) The e¤ects of embedding the target language in other subjects during primary school, for example in science, history, geography, sports, art and crafts; (4) Features of classroom instruction in the target language; (5) The e¤ects of corrective feedback by the teacher, and of peer feedback and self-reflection by young learners; (6) Case studies of pupils encouraged to express themselves creatively in the target language, instead of practising standardized utterances provided in class; (7) The e¤ects of o¤ering an additional language to children with learning disabilities or special needs. In addition to expanding the scope of current research, it is necessary to assess the advantages and limits of other fine work that has been done and still needs extra dimensions. Pienemann, Kessler, and Roos (2006) state that all learners of a foreign language go through the same developmental stages in exactly the same sequence, thus variation and learning speed depend on individual and contextual factors. Given the ubiquity of the six developmental stages, Pienemann et al. say education is not a means to skipping one or more of them. Certain approaches, however, such as content-based learning, can increase the speed and success rate with which pupils go through them. The authors conceptualize progression as climbing stairs or a ladder: a series of incremental steps upward. In a critical discussion of the ladder metaphor, Mitchell (2003) has argued that learning a second language is a complex and non-linear process, one with setbacks and multiple points of contact among the different stages. Development may sooner be thought of as the interplay between linguistic ability, accuracy and complexity. A similar point of view is taken by Peltzer-Karpf and Zangl (1997). They argue that children’s utterances develop from short phrases to longer ones before a period of ‘‘Systemturbulenz’’ sets in when grammar control regresses, only to stabilize later on.

56

Peter Edelenbos and Angelika Kubanek

Nonetheless, the ladder metaphor serves as a support to pupils and teachers. It can be employed to obtain a sense of progress or direction, or as a tool for providing feedback and shaping discussions. Researchers and teacher trainers now face a major task: they have to better understand the limitations of the metaphor and the complexities of language learning, for example those uncovered by research into second language acquisition (Verspoor, Lowie, and Van Dijk 2008). This requires conducting several crucial studies to confirm the persistence of the six developmental stages, and to investigate exactly how children can progress through them. Only then can ELL fully benefit from the model. Listening is one of the most basic skills in ELL. The importance of listening as a unique component of ELL has been demonstrated previously (Edelenbos, Johnstone, and Kubanek 2007). However, it is still unclear how listening skills may be improved. What factors contribute to their acquisition and what methods bring this about best? Studies of e¤ects should include direct observations to establish how much teaching time is dedicated to o¤ering English to children. Additionally, these should take into account teachers’ fluency and correct use of the target language, as well as the amount of time spent on discussions and feedback. Here, the first steps to be taken are in the classroom. Another area in great need of research is intercultural learning. In terms of urgency, three aspects stand out. First, how can teachers and teacher trainers best understand and explain possible trajectories for intercultural learning? Second, the manner and extent to which intercultural competence supports linguistic competence must be empirically tested. Finally, what are the results so far? What is the e¤ect on skills among primary school children and even those moving from kindergarten to the primary level? (For more on this see Edelenbos and Kubanek, 2007.) Looking forward, an issue of increasing importance to ELL and education more generally will be the approach termed ‘‘blended learning’’. The main themes for this research will revolve around a fundamental question: How and to what extent is blended learning superior to traditional forms of language learning? Part of this is to explore the different possibilities for integrating out-of-school learning. Finally, questions will arise about flexible learning arrangements that can be optimized based on the type of learner and the specific model of ELL being used.

Early foreign language learning

57

References Blondin, Christiane, Candelier, Michel, Edelenbos, Peter, Johnstone, Richard, Kubanek-German Angelika, and Traute Taeschner 1998 Foreign Languages in Primary and Pre-School Education: Context and Outcomes. A review of Recent Research within the European Union. London: CILT. Edelenbos, Peter, Johnstone, Richard, and Angelika Kubanek 2007 The Main Pedagogical Principles Underlying the Teaching of Young Learners: Languages for the Children of Europe. Bru¨ssel: Schlu¨sselstudie fu¨r die Europa¨ische Kommission. Edelenbos, Peter, and Angelika Kubanek 2007 Bilinguale Kinderga¨rten im Saarland. Ergebnisse einer Vorstudie im Mai 2007. Durchgefu¨hrt im Auftrag des Ministeriums fu¨r Bildung, Kultur und Wissenschaft des Saarlandes. European Commission 2003 Promoting Language Learning and Linguistic Diversity: An Action Plan 2004–2006. Retrieved on 10 December 2008 from http://ec.europa.eu/education/doc/o‰cial/keydoc/actlang/act_ lang_en.pdf March, 2005 Progress towards the Lisbon Objectives in Education and Training. Retrieved on 10 December from http://www.okm.gov.hu/doc/ upload/200507/progress_report_2005.pdf Guba, Egon G. 1978 Evaluation of Educational Programs. New York: Harper Row. Hardi, Judit ´ ltala´nos iskola´s tanulo´k attitűdje e´s motiva´cio´ja az angol mint ide2004 A gen nyelv tanula´sa´ra [The attitude and motivation of primary school children for learning English as a foreign language]. Magyar Pedago´gia 104 (2): 225–242. Johnstone, Richard 2000 Early language learning. In: Michael Byram (ed.), Routledge Encyclopedia of Language Teaching and Learning, 188–193. Routledge: London and New York. Kennedy, Teresa J., Nelson, Jack K., Odell, Michael R. L., and Austin, Laurie K. 2000 The FLES attitudinal inventory. Foreign Language Annals 33 (3): 278–289. Lamb, Martin 2004 Integrative motivation in a globalizing world. System 32 (1): 3– 19. Marschollek, Andreas 2002 Kognitive und A¤ektive Flexibilita¨t durch Fremde Sprachen. Eine empirische Untersuchung in der Primarstufe. Mu¨nster: Lit.

58

Peter Edelenbos and Angelika Kubanek

Mihaljevic´ Djigunovic´, Jelena 1998 Uloga afektivnih faktora u ucenju stranoga jezika [Role of a¤ective factors in FL learning]. Zagreb: Filozofski fakultet Sveucilista u Zagrebu. Mitchell, Rosamond 2003 Rethinking the concept of progression in the national curriculum for modern foreign languages: a research perspective. Language Learning Journal 27: 15–23. Nikolov, Marianne 1999 ‘Why do you learn English?’ ‘Because the teacher is short.’ A study of Hungarian children’s foreign language learning motivation. Language Teaching Research 3 (1): 33–56. 2002 Issues in English Language Education. Bern: Peter Lang AG. 2003 Angolul e´s ne´metu¨l tanulo´ dia´kok nyelvtanula´si attitűdje e´s motiva´cio´ja. [Attitudes and motivation of learners of English and German] Iskolakultu´ra 13 (8): 61–73. Peltzer-Karpf, Annemarie, and Renate Zangl 1997 Vier Jahre Vienna Bilingual Schooling: Eine Langzeitstudie. Vienna; Bundesministerium fu¨r Unterricht und kulturelle Angelegenheiten., Abteilung 1 (1). Pienemann, Manfred, Kessler, Joerg-U. and Eckhard Roos (eds.) 2006 Englischerwerb in der Grundschule. Stuttgart: UTB. Verspoor, Marjolijn H., Lowie, Wander, and Marijn van Dijk 2008 Variability in L2 development from a dynamic systems perspective. The Modern Language Journal 92 (2): 214–231.

3.

Assessment of early learning of foreign languages in the USA

Helena Curtain

1.

Introduction

As early foreign language learning programs have proliferated in the United States, along with a focus on language proficiency, a need has emerged to quantify what students in various types of language programs are able to do in terms of proficiency outcomes measured on a profession-wide scale. The general belief has been that the more time students spend working communicatively with the target language under the guidance of a skilled and fluent teacher, the greater will be the level of language proficiency they acquire. But until the recent advent of several proficiency-based language assessment instruments for early language learners, there has been no manageable, practical way to establish the validity of that belief. There has been a movement toward authentic performance-based assessment in the work done in the USA during the past 25 years in oral proficiency. This work was spearheaded by the American Council on the Teaching of Foreign Languages (ACTFL) that in 1982 made available the first version of the ACTFL Proficiency Guidelines soon to be followed by the Oral Proficiency Interview (OPI) (Liskin-Gasparro 2003). The guidelines and the OPI indicate what students can and cannot do with language in terms of global tasks or functions, context and content, accuracy, and text type. The guidelines are a series of language proficiency descriptors and OPI is a standardized procedure for the global assessment of functional speaking ability conducted in a one-on-one interview lasting from 10 to 30 minutes. The proficiency guidelines and the OPI have served the profession well in providing a common measure for talking about language outcomes. In the intervening time language educators have gained insights regarding the nature of face-to-face communication and the impact of instructional practices on the development of students’ oral proficiency.

60

Helena Curtain

In addition to the OPI and to the guidelines, ACTFL and a consortium of language professional organizations produced in 1996 and revised in 1999 the Standards for Foreign Language Learning the in the 21st Century (ACTFL 1999). Related to the movements for proficiency outcomes, there has been a parallel movement toward performance-based assessment rather than a focus on achievement testing for language classrooms. D’Anglejan, Harley and Shapson (1990: 106) state that ‘‘student evaluations must emphasize communicative language performance in context’’. Smith (1995: 1) indicates that assessment and testing of young learners must reflect the approaches used in teaching this age group, as traditional paper and pencil tests alone ‘‘cannot incorporate the wide range of approaches used in teaching this age group.’’ Performance tasks di¤er significantly from traditional paper-andpencil approaches to assessment and focus on what students can do with language, rather than their ability to manipulate bits of knowledge about language. A performance task is a product or a performance that calls for the student to use the competencies and the vocabulary from a unit of instruction in a meaningful context, preferably a context that replicates the challenges of the ‘‘real world’’ (Curtain and Dahlberg 2004: 164). Linked to assessment of language programs for young learners there are two issues that need to be considered. First, due to the sometimes very small amount of contact time, there are limitations in the variety of second language vocabulary grammar and language functions that students encounter. Since the language learned may be very restricted, assessment instruments must also be able to measure the small amount of language that students have actually learned. Second, although the OPI has also provided an assessment construct that could be applied to classroom testing formats, it has inherent di‰culties at the K-12 level since implementation of the OPI raises cost and feasibility issues in K-12 programs due to the large numbers of students and the constraints on teacher time (Fall, Adair-Hauck, and Glisan 2007).

2.

Language programs for young learners in the USA

In USA elementary schools language programs for young learners there is no consistency in the amount of time available for instruction or in

Assessment of early learning of foreign languages in the USA

61

the time of onset for the program. Programs vary in the grade level at which the language is introduced, the number of times the language is taught per week, and the length of time per language teaching session. Elementary school language classes may meet between two and five times per week, in sessions ranging from 15 to 60 minutes. Some programs o¤er languages once a week for 20 minutes; some o¤er it three times a week for 30 minutes; still others o¤er it every day for 45 minutes. Some begin language instruction at kindergarten, others at grades 2 or 3, still others at grades 4 or 5. Some programs continue through grade 8, while others are o¤ered only in the elementary school with no connections to the secondary levels, so that students must start language study again when they reach the middle school or high school. Some programs are o¤ered in alternating languages to give students a choice and to satisfy parent requests for certain languages. This great variety is due in part to the fact that, within the United States, curriculum is set at the district level. Thousands of individual school districts make thousands of individual decisions regarding the nature of such programs. The plus side of this situation is that school districts have not had to wait for a national or state mandate to start their programs. The negative side is, of course, that there is no consistency, and sometimes no continuity, in such programs. Language programs in the USA can be classified into two major types: immersion programs and FLES (Foreign Language in the Elementary School) programs. FLES has sometimes been used as a general term to describe all foreign language programs at this level. However, it is most often used to describe a particular type of elementary school foreign language program, one that is taught two to five times per week for class periods of 20 to 60 minutes or more. Some FLES classes integrate other areas of the curriculum; others focus on the second language and its culture (Curtain and Dahlberg 2004). In immersion programs the usual curriculum activities are conducted in a second language. This language is the medium as well as the object of instruction. The second language is used for 50–100% of the school day throughout the school experience. The focus is on subject matter, but language learning per se is incorporated as necessary throughout the curriculum. FLES programs have a time allotment of 5–15% of the student day. Time is usually spent learning language per se. The goal of FLES programs is functional proficiency in the second language. The degree of

62

Helena Curtain

proficiency students obtain varies with the program, and is commonly believed to be dependent on the amount of time available for language instruction and the quality of instruction that the students receive. In FLES programs, proficiency goals are most often concentrated on listening and speaking but students are also expected to acquire an understanding of and appreciation for other cultures, and to acquire some proficiency in reading and writing. Some FLES programs are ‘‘content enriched,’’ or ‘‘content-based’’ and integrate topics from the regular curriculum into instruction. The great bulk of language programs fall into the FLES or exploratory and not the immersion category. A survey (Branaman, Rhodes, and Rennie 1998) done in 1997 (and currently being conducted by the Center for Applied Linguistics for 2007) found that from 1987 to 1997 language instruction in elementary schools increased by nearly 10 percent. In 1987, just over one in five (22%) elementary schools reported teaching foreign languages; by 1997 the percentage had risen to 31% (approximately one in three). The percentage of secondary schools teaching foreign language remained fairly stable: 87% in 1987 and 86% in 1997. Among the 31% of programs that o¤ered foreign language study, 79% o¤ered introductory exposure and 21% were programs leading to language proficiency. Only 7% of all elementary schools (increased from 3% in 1987) o¤ered instruction in which the students were likely to attain a high level of fluency, as recommended in the goals of the national standards!

3. 3.1.

Proficiency, standards and performance guidelines ACTFL proficiency guidelines

The ACTFL Proficiency Guidelines, first published provisionally in 1982, marked a pivotal change in the teaching of new languages in the United States and have greatly influenced the language teaching profession. They refocused language instructional goals from what learners know about language to what they can do with the language they have learned, and at the same time they established a common metric for measuring student performance.

Assessment of early learning of foreign languages in the USA

63

The Guidelines describe student performance in listening, speaking, reading, and writing at the Novice, Intermediate, Advanced, and Superior levels. They depict student language performance in terms of function, content, and accuracy of the message being delivered or received. They were adapted from guidelines developed in U.S. government language schools and have made ‘‘proficiency-oriented instruction’’ a part of the vocabulary of every language teacher. The 1982 guidelines were subsequently reevaluated and revised, beginning with Speaking in 1999 and Writing in 2001 (Liskin-Gasparro 2003). The ACTFL guidelines are di‰cult to apply directly to the curriculum of the elementary school program since they were designed to describe the performance of adult-like language users. Many of the functions and much of the content described are not closely related to the interests and needs of children. Yet the principle of purposeful language use is clearly held in common at all levels of communicative language teaching. Emphasis has shifted from grammar and discrete-item testing to a global evaluation of language competencies. Teachers can permit themselves to assess communication of meaning as well as, though not to the total exclusion of, degree of accuracy, and students can be rewarded for a much broader range of language ability and performance. The ACTFL Proficiency Guidelines were an important first step leading to the ACTFL Performance Guidelines for K-12 Learners that are tailored to the language needs of students in elementary and secondary schools. Table 1 shows proficiency levels based on the (ACTFL) Proficiency Scale from Novice Low to Intermediate High (Pre-Advanced). According to these proficiency descriptors, for high school speakers of English beginning a language such as French or Spanish at grade 9, they would be expected to reach the novice mid by the end of the first year of study. By the end of 4 years of high school study in languages such as French and Spanish most of them should reach the intermediate low category. In contrast to the Common European Framework of Reference (Council of Europe 2001) which has six levels of performance, the ACTFL Proficiency scale has ten levels. The superior and advanced levels have been included here since the proficiency of elementary school students who are not in immersion programs is not expected to go beyond the intermediate level. (The complete guidelines can be accessed at www.actfl.org.)

64

Helena Curtain

Table 1. The ACTFL Proficiency Guidelines Intermediate High (Pre-Advanced in the K-12 Performance Guidelines) Intermediate-High speakers are able to converse with ease and confidence when dealing with most routine tasks and social situations of the Intermediate level. They are able to handle successfully many uncomplicated tasks and social situations requiring an exchange of basic information related to work, school, recreation, particular interests and areas of competence, though hesitation and errors may be evident. IntermediateHigh speakers handle the tasks pertaining to the Advanced level, but they are unable to sustain performance at that level over a variety of topics. Intermediate Mid Speakers at the Intermediate-Mid level are able to handle successfully a variety of uncomplicated communicative tasks in straightforward social situations. Conversation is generally limited to those predictable and concrete exchanges necessary for survival in the target culture; these include personal information covering self, family, home, daily activities, interests and personal preferences, as well as physical and social needs, such as food, shopping, travel and lodging. Intermediate-Mid speakers tend to function reactively, for example, by responding to direct questions or requests for information. However, they are capable of asking a variety of questions when necessary to obtain simple information to satisfy basic needs, such as directions, prices and services. Intermediate Low Speakers at the Intermediate-Low level are able to handle successfully a limited number of uncomplicated communicative tasks by creating with the language in straightforward social situations. Conversation is restricted to some of the concrete exchanges and predictable topics necessary for survival in the target language culture. These topics relate to basic personal information covering, for example, self and family, some daily activities and personal preferences, as well as to some immediate needs, such as ordering food and making simple purchases. At the Intermediate-Low level, speakers are primarily reactive and struggle to answer direct questions or requests for information, but they are also able to ask a few appropriate questions. Novice High Speakers at the Novice-High level are able to handle a variety of tasks pertaining to the Intermediate level, but are unable to sustain performance at that level. They are able to manage successfully a number of uncomplicated communicative tasks in straightforward social situations. Conversation is restricted to a few of the predictable topics necessary for survival in the target language culture, such as basic personal information, basic objects and a limited number of activities, preferences and immediate needs. Novice-High speakers respond to simple, direct questions or requests for information; they are able to ask only a very few formulaic questions when asked to do so. Novice Mid Speakers at the Novice-Mid level communicate minimally and with di‰culty by using a number of isolated words and memorized phrases limited by the particular context in which the language has been learned. When responding to direct questions, they may utter only two or three words at a time or an occasional stock answer. They pause frequently as they search for simple vocabulary or attempt to recycle their own and their interlocutor’s words.

Assessment of early learning of foreign languages in the USA

65

Novice Low Speakers at the Novice-Low level have no real functional ability and, because of their pronunciation, they may be unintelligible. Given adequate time and familiar cues, they may be able to exchange greetings, give their identity, and name a number of familiar objects from their immediate environment. They are unable to perform functions or handle topics pertaining to the Intermediate level, and cannot therefore participate in a true conversational exchange.

3.2.

Standards for Foreign Language Learning in the 21st Century

Just as the proficiency movement created a whole new way of thinking about language instruction and assessment in 1996 the Standards for Foreign Language Learning in the 21st Century provided an additional new focus for curriculum and program planning. In contrast to earlier curriculum documents, the Standards created the bold vision of a long sequence of language instruction for all learners, beginning in kindergarten and continuing through grade twelve and beyond. For the first time in the United States, the language teaching profession had made a strong statement about the importance of a long sequence of instruction beginning with the earliest years of schooling. By 1999 the languagespecific professional organizations had developed supporting Standards documents describing student progress from grades K-16, thus focusing the entire instructional sequence on a unified vision of language education. See Table 2. The document contains sample progress indicators for Grades 4, 8, and 12 for each standard. The intent is that these progress indicators can be realistically achieved at some level of performance by all students. The indicators are meant to serve as a basis for curriculum at the district and local level. In keeping with the long-held educational tradition in the United States that education is to be defined at state, district and school levels, the standards do not mandate local curricula. They are also not tied to any particular instructional method, but are instead a broad statement of what foreign language education should prepare students to do. The Standards have come to be known as the ‘‘5 Cs’’ according the five major goal areas: communication, culture, connections, comparisons and communities.

66

Helena Curtain

Table 2. Standards for foreign language learning in the 21st Century (American Council on the Teaching of Foreign Languages 1999: 10) Communication Communicate in Languages Other Than English Standard 1.1: Students engage in conversations, provide and obtain information, express feelings and emotions, and exchange opinions. (Interpersonal) Standard 1.2: Students understand and interpret written and spoken language on a variety of topics. (Interpretive) Standard 1.3: Students present information, concepts, and ideas to an audience of listeners or readers on a variety of topics. (Presentational) Cultures Gain Knowledge and Understanding of Other Cultures Standard 2.1: Students demonstrate an understanding of the relationship between the practices and perspectives of the culture studied. Standard 2.2: Students demonstrate an understanding of the relationship between the products and perspectives of the culture studied. Connections Connect with Other Disciplines and Acquire Information Standard 3.1: Students reinforce and further their knowledge of other disciplines through the foreign language. Standard 3.2: Students acquire information and recognize the distinctive viewpoints that are only available through the foreign language and its cultures. Comparisons Develop Insight into the Nature of Language and Culture Standard 4.1: Students demonstrate understanding of the nature of language through comparisons of the language studied and their own. Standard 4.2: Students demonstrate understanding of the concept of culture through comparisons of the cultures studied and their own. Communities Participate in Multilingual Communities at Home & Around the World Standard 5.1: Students use the language both within and beyond the school setting. Standard 5.2: Students show evidence of becoming life-long learners by using the language for personal enjoyment and enrichment

The communication standard stresses the use of language for communication and approaches the four skills of listening, speaking reading and writing by reorganizing them into interpersonal, interpretive and presentational tasks. – Interpersonal: Two-way communication: Exchange of information through speaking or writing.

Assessment of early learning of foreign languages in the USA

67

– Interpretive: One-way communication: Taking in information through listening or reading – Presentational: One-way communication: Giving out information through speaking and listening

3.3.

ACTFL performance guidelines for K-12 learners Level 2

Since the proficiency guidelines were originally focused on adult-like situations, an ACTFL Young Learner Task Force drew on the Proficiency Guidelines, the Standards for Foreign Language Learning, and classroom experiences to develop a set of performance standards appropriate for K-12 learners. The ACTFL Performance Guidelines for K-12 Learners (1998), focus on students who participate in elementary, middle and high school language programs, taking into account the continuous cognitive development of students at these levels and its influence on their ability to perform language tasks. The Performance Guidelines also take into account the fact that most K-12 language development takes place in a classroom context, with a curriculum that is articulated from one level to the next (Swender and Duncan 1998). The Five Cs make the content (the ‘‘what’’) of language education clear, and the K-12 Performance Guidelines describe the ‘‘how well’’ of language learning based on assessment of student performance on carefully designed activities and tasks. The guidelines identify areas for use as a basis for performance assessment in the second language and have become one of the most important tools for in designing realistic goals and assessments. The Performance Guidelines for K-12 Learners were written primarily to describe the performance of students who begin foreign language study in kindergarten and continue through Grade 12. Parallel to the national standards, they describe student benchmarks at grades 4, 8 and 12. The guidelines also have an element of flexibility in that they describe language outcomes for students who begin instruction at di¤erent entry points. The reason that the guidelines are so important for the future of language education in the USA is that for the first time communication outcomes have been tied to a specific sequence of instruction over a specific sequence of time. The guidelines authors acknowledge (Swender and Duncan 1998) that the present picture of foreign language education in this country does not live up to vision of K-12 programs and they anticipate that outcomes will di¤er for those programs that do not

68

Helena Curtain

live up to the profiles as articulated in the expected outcomes chart in Figure 1 (programs that may have di¤erent entry points or exposure time). They also anticipate that performance outcomes may di¤er for students who are learning a second or third foreign language. The guidelines are organized according to: Three modes of communication – Interpersonal – Interpretive – Presentational Three Benchmark Levels – Novice Learner (K-4, 5–8, 9–10) – Intermediate Learner (K-8, 7–12) – Pre-Advanced Learner (K-12) Six Domains of Performance – Comprehensibility (How well is the student understood?) – Comprehension (How well does the student understand?) – Language Control (How accurate is the student’s language?) – Vocabulary Usage (How extensive and applicable is the student’s language?) – Communication Strategies (How do they maintain communication?) – Cultural Awareness (How is their cultural understanding reflected in their communication?) Table 3 gives a sample of the interpersonal, interpretive and presentational modes of communication in the novice learner range at grades K-4, grades 5–8 or grades 9–10. The descriptions represent what students should be able to do with a foreign language after set amounts of time provided that their instruction is both standards and performance-oriented. Teachers should be able to refer to this document, reflect on their students’ use of language, and gauge if their performance is according to the descriptions in the guidelines. Alternatively, teachers may read the descriptors and determine that their students do not perform at a level consistent with the time and e¤ort spent and, therefore, may want to seek ways of modifying their program to achieve the level of language performance described. The guidelines developers (Swender and Duncan 1998) indicated that the outcomes correlated with time in programs ‘‘required students to be enrolled in elementary programs that meet from 3–5 days per week for

Assessment of early learning of foreign languages in the USA

69

no less than 30–40 minutes per class; middle school programs that meet daily for no less than 40–50 minutes; and high school programs that equal four units of credit’’ (Swender and Duncan 1998: 482). Four units of credit would equal the equivalent of daily instruction for 45 minutes to one hour. Table 3. Novice Learner Range Grade K-4 or Grade 5–8 or Grade 9–10 ACTFL K-12 Performance Guidelines Interpersonal: – Comprehend and produce vocabulary that is related to everyday objects and actions on a limited number of familiar topics; – Use words and phrases primarily as lexical items without awareness of grammatical structure; – Recognize and use vocabulary from a variety of topics including those related to other curricular areas; – May often rely on words and phrases from their native language when attempting to communicate beyond the word and/or gesture level. Interpretive: – Recognize a variety of vocabulary words and expressions related to familiar topics embedded within relevant curricular areas; – Demonstrate increased comprehension of vocabulary in spoken passages when these are enhanced by pantomime, props, and/or visuals; – Demonstrate increased comprehension of written passages when accompanied by illustrations and other contextual clues. Presentational: – Use a limited number of words and phrases for common objects and actions in familiar categories; – Supplement their basic vocabulary with expressions acquired from sources such as the teacher or picture dictionaries; – Rely on native language and phrases when expressing personal meaning in less familiar categories.

The graph in Figure 1 illustrates the anticipated performance outcomes described in the K-12 Guidelines and shows their relationship to the ACTFL Proficiency Guidelines. Note that the Pre-Advanced level in the Performance Guidelines is equivalent to Intermediate High in the Proficiency Guidelines. Once students have attained the Pre-Advanced level, the ACTFL Proficiency Guidelines can be used to describe student progress through the Superior level. (Published and copyrighted by the American Council on the Teaching of Foreign Languages. Reproduced with permission.)

70

Helena Curtain

Figure 1.

Assessment of early learning of foreign languages in the USA

71

These guidelines will help teachers to set realistic goals for students and choose appropriate performance assessments to measure student successes. Thus, for example, the teacher in a third-grade early language program might consult the Vocabulary Use category, for example. If his or her students have been learning the new language since kindergarten, they will be nearing the benchmarks identified in the Novice Learner Range (K-4). The teacher can read the descriptions to see what the students are working toward in the areas of Interpersonal, Interpretive, and Presentational Communication. After examining these guidelines, the teacher realizes that performance tasks will be most useful and realistic if they prompt students to use memorized vocabulary and phrases in familiar contexts. The teacher also realizes that extended writing or speaking in sentences will not be a realistic expectation. It will be important that activities and tasks directed toward vocabulary in the interpretive mode should use familiar contexts and be accompanied by gestures, pantomime, props, and visuals (oral) or illustrations (written) and other contextual clues. Examination of performance indicators in the other five areas should give the teacher a basis for realistic and appropriate lessons and tasks for these third-grade students. Teachers working with beginning language students in grade three would have to recognize that students in their classrooms are at very early stages of development toward these performance levels, but the performance goals are still the same. Teachers working with eighthgrade students in a 5–8 program would also find their students in the Novice Learner Range, but their topics and familiar contexts would likely be di¤erent from those of the third- or fourth-grade child. Taken together, the performance indicators in the six domains provide a snapshot of learner performance at each of three points along the K-12 continuum: Novice, Intermediate, and Pre-Advanced. Teachers can use the Performance Guidelines to help focus their planning and assessment, no matter how many years of previous language experience their students may have had (Curtain and Dahlberg 2004: 188).

4.

Assessment instruments for early language learning level

In the wake of the professional developments mentioned above, several useful measures of student language attainment by young learners

72

Helena Curtain

have been developed. This section will briefly describe each of those instruments.

4.1.

Assessments and Documents Available from the Center for Applied Linguistics: ELLOPA, SOPA, COPE and FLAD

Since the late 1980s, the Center for Applied Linguistics (Thompson, Boyson and Rhodes 2006) has been developing alternative assessments for young language learners. Three of these assessments are the Early Language Listening and Oral Proficiency Assessment (ELLOPA), the Student Oral Proficiency Assessment (SOPA), and the CAL Oral Proficiency Exam (COPE). The purpose of these instruments is to determine the highest proficiency levels in speaking and listening comprehension that students can sustain at a particular point in time. The COPE was developed first. Its interactive format elicits conversational speech between pairs of students, yielding a global proficiency rating for each student in both speaking and listening comprehension. The COPE was originally developed in 1988 for fifth and sixth grade Spanish immersion students (Rhodes and Thompson 1990) and has also been used with immersion students in seventh and eighth grades. In response to requests from many school districts for an alternative language assessment that would be appropriate for immersion students in the lower elementary grades, the SOPA was developed in 1991 for children in Grades 1–4 in a Spanish partial immersion program. Similar to the COPE, the SOPA is also interactive and assesses students in pairs. In 1996, the interview format and rating scale of the SOPA were adapted for non-immersion Foreign Language in the Elementary School (FLES) programs. Over the years, the SOPA has been used in an increasing number of language programs for students in Grades K-8. In 2001, the ELLOPA, an instrument with a format very similar to that of the SOPA, was developedfor the youngest children (Grades K-2) learning languages in FLES programs. Teachers expressed the need for an assessment that provided more opportunity for students to use language at the novice levels. The ELLOPA has been adapted for immersion programs as well, and has also been used with pre-K students. The main di¤erence between the SOPA and ELLOPA is the use of additional visuals, such as puppets, with the ELLOPA. Table 4 lists the programs for which the ELLOPA, SOPA, and COPE have been designed. These assessments have also been developed

Assessment of early learning of foreign languages in the USA

73

for a number of languages, including Chinese, English, French, German, Hebrew, Japanese, Russian, and Spanish. Table 4. Programs and Grade Levels Using ELLOPA, SOPA, and COPE ELLOPA

SOPA

COPE

Grade levels

Pre-K–Grade 2

K–Grade 8

Grades 5–8

Programs

FLES, FLEX, Immersion

FLES, FLEX, Immersion

Immersion

4.1.1. The ELLOPA and SOPA Both the ELLOPA and SOPA use an interview format. Two administrators assess pairs of students in a friendly, non-stressful environment. One administrator primarily conducts the interview and the other primarily takes notes and assigns the students’ ratings in addition to operating the recording equipment for rating verification. This allows the interviewer to focus on eliciting language that will demonstrate the students’ highest proficiency levels. The rater records as much as possible of each student’s speech to facilitate assigning accurate ratings. Interviews are conducted entirely in the target language and take approximately 15–20 minutes to complete. The goal of these assessments is to show what the students can do in the target language rather than focus on what they cannot do. The ELLOPA and SOPA interviews consist of a series of games or tasks with varying levels of di‰culty that elicit both academic and social language (see Table 5). The assessment activities follow the natural development of language skills, focusing first on listening comprehension and then on speaking. This sequence builds the students’ confidence, allowing them to respond successfully, receptive skills being less demanding than productive skills. Students are encouraged to say as much as they can so that adequate speech samples may be obtained for accurate ratings. Students receive ratings in two or four skill areas – oral fluency, grammar, vocabulary, and listening comprehension – depending on the rubric that is used. Ideally, the ELLOPA and SOPA results should not be used as the only indicator of a student’s progress in language development. Rather, they are used in conjunction with teacher observations and other evaluations of the student’s daily work. When these assessments are given annually, students’ ratings are expected to reveal grad-

74

Helena Curtain

ual progress in the target language(s). The amount of progress depends on many factors, such as the type of program, the instructor’s fluency level in the target language, frequency of instruction, and student motivation. Realistic expectations for students in FLES programs can be found in the ACTFL Performance Guidelines for K-12 Learners (1998). A quiet room, free of distraction, is preferred for the ELLOPA and SOPA interviews. Establishing a nonthreatening environment helps ensure that the students feel at ease during the assessment activities, and consequently, attain their highest levels of proficiency.

Table 5. Outline of the Ellopa and SOPA Non Immersion Scripts (Center for Applied Lingusitics) Non-Immersion (FLES) ELLOPA Warm-Up Game 1: The Magic Bag Goal: TPR with people and house objects – Put students at ease with simple listening comprehension first, followed by speaking. Game 2: Come to My House Goal: Identify and name, follow and give commands, describe – Give students opportunities to create at sentence-level. Game 3: Let’s Sing! (Optional) Goal: Sing a song – Give students the opportunity to show what they can do in the language if they are comfortable with singing and are willing to sing. Game 4: Talking With Mrs. Cow Goal: Answer questions – Give students opportunities to create with the language on familiar topics. Game 5: Accidental Mix-Up (Optional) Goal: Sort items that belong to Mr. Cat (U.S.A.) and la Sra. Vaca (Spanish-speaking culture) – Students show that they know cultural di¤erences. [May be conducted in English at a di¤erent time] Wind-Down: Simon Says Goal: TPR – End at students’ comfort level. Non-Immersion (FLES) SOPA Warm-Up Task 1: Fruits & Other Foods Goal: TPR with fruits and other foods – Put students at ease with simple listening comprehension first, followed by speaking. Task 2: All About You Goal: Answer questions – Give students opportunities to create with the language on familiar topics.

Assessment of early learning of foreign languages in the USA

75

Task 3: The Classroom Goal: Identify and name, follow and give commands, describe – Give students opportunities to create at sentence-level. Wind-Down: Simon Says Goal: TPR – End at students’ comfort level.

The students are assessed in pairs so that they will feel more comfortable and to provide opportunities for interaction between them as well as with the administrators. Videotaping the ELLOPA and SOPA interviews is important for subsequent rating verification. Videotaping the students is preferred in order to capture more of the nonverbal communication that occurs during the assessment activities, but if videotaping equipment is not available, the interviews may be audio taped. Ideally, the interviews are both videotaped and audio taped to provide a backup in case of equipment failure. The rater generally operates the recording equipment, records the students’ responses for each activity on the ELLOPA or SOPA rating sheet, and assigns ratings in the appropriate categories for each student. Ratings are assigned in each of the five skill categories on the ELLOPA-RP, in four skill categories on the COPE/SOPA-RS, or in two skill categories on the SOPA-RS. After each interview, the rater and interviewer review the notes and come to an agreement on the ratings. If they require further discussion than is possible between interviews to reach a consensus, or if they want to fine-tune their ratings, the recorded interviews may be viewed later. 4.1.2. The COPE The COPE uses a role-play format and assesses two students at a time. The students interact with each other in the context of realistic situations that elicit both academic and social language interactions are mainly between the students rather than with the examiner. The COPE role plays take approximately 15–20 minutes to complete and are audio and/or videotaped for rating verification. Students are assigned ratings in four skill areas: oral fluency, grammar, vocabulary, and listening comprehension. 4.1.3. Foreign Language Assessment Directory (FLAD) The CAL created a Foreign Language Assessment Directory (FLAD) that is now available online. FLAD is a searchable directory of informa-

76

Helena Curtain

tion on nearly 200 tests in over 80 languages. FLAD users can find information about assessments including the grade and proficiency level for which a test is intended, the skills targeted by a test, information about a test’s development and the publisher’s or developer’s contact information for further inquiries (http://www.cal.org/CALWebDB/ FLAD/).

4.2.

The National Online Early Language Learning Assessment (NOELLA)

The U.S. Department of Education funded a consortium of six states (Georgia, Kentucky, New Jersey, South Carolina, Virginia, and Wyoming) to develop an online assessment of early language that is low in cost, reliable, accessible nationwide in every type of school, time e‰cient, and allows comparisons across all types of elementary school language programs and models of delivery. The consortium contracted with the Center for Applied Second Language Study (CASLS) at the University of Oregon and the Center for Applied Linguistics (CAL) in Washington D.C. to help make this online assessment a reality (Falsgraf and Tollefson 2006). The assessment, the National Online Early Language Learning Assessment (NOELLA), is currently being piloted to assess children’s learning in four languages: French, Japanese, Mandarin Chinese, and Spanish. NOELLA is designed to be developmentally appropriate to measure the language learning of children in grades three through six across program models. The NOELLA gives educators and parents a snapshot of how children are doing in the interpretive mode of communication (reading and listening) as well as in the presentational mode (speaking and writing). Because a computer program is incapable of negotiating meaning with a student, for example, responding to a student’s oral or written communication as would another human being, does not assess the interpersonal mode of communication. Assessments such as the SOPA and ELLOPA are designed to address that mode and can be used in conjunction with the NOELLA to provide a complete picture of proficiency (Falsgraf and Tollefson 2006). The first step in developing NOELLA, therefore, was to develop benchmarks: specific definitions of proficiency levels. Groups of elementary school teachers from the six states in the consortium, joined by

Assessment of early learning of foreign languages in the USA

77

teachers from other states, wrote those benchmarks using the National Standards, the ACTFL Proficiency Guidelines, the ACTFL K-12 Performance Guidelines, and the SOPA scoring guides from CAL (The sample of the benchmarks will be available at http://noella.uoregon. edu. The next step was to develop assessment items (See http://noella. uoregon.edu for sample items). After the test, the computer immediately scores and reports the interpretive tasks. Students’ proficiency levels on the presentational components of the assessment are assigned by human raters and, will be available a few weeks after the test. NOELLA reports allow schools to see how the whole program did in particular skills or item types. There is a similar online assessment currently available for grades 7– 12: The Standards Based Measure of Proficiency (STAMP; see http:// www.avantassessment.com/products/about_stamp.html). 4.3.

Linguafolio Jr.

The United States has developed Linguafolio and Linguafolio Jr, versions of the European Language Portfolio. Five states: Virginia, Kentucky, South Carolina, North Carolina, Georgia, and Nebraska have spearheaded this project with many other states now following their lead. Kentucky led in the Linguafolio Jr. project. LinguaFolio, Jr. is a collection of information about early language learning that is designed for use by educators and younger students. It allows students to record and reflect on their language learning and cultural experiences and can accompany language learning at a young age and is suitable for documenting language abilities for various uses. Like its European counterpart, LinguaFolio Jr. includes three parts: a passport, a language biography, and a dossier (see http://www.doe.virginia.gov/linguafolio/ junior.html). One of the primary purposes of e¤orts with Linguafolio and Linguafolio Jr. in the United States is to connect standards and performance guidelines here to the internationally accepted Common European Framework of Reference (2001). 4.4.

New Jersey Consortium for Assessing Performance Standards

A consortium of educators in the state of New Jersey (Gilbert, Mackenzie, Meulener, Smith, Yetman, and Zeppieri 2008) has established a col-

78

Helena Curtain

lection of thematically organized performance assessments based on the guidelines and frameworks presented in this chapter (see the Foreign Language Educators of New Jersey web site http://flenj.org/CAPS/ ?page=parent). The project includes a database of thematically organized, integrated performance assessment tasks at the benchmark levels of proficiency, novice-mid, intermediate-low and pre-advanced as defined by the ACTFL Performance Guidelines for K-12 Learners. The project also includes a design template for the development of benchmark assessments to be utilized by school districts nationwide. There are 70 assessment tasks with rubrics. Some of the tasks also contain both visual and auditory samples of student work. The thematically organized assessments have been developed around the themes such as – – – –

Art Appreciation Art of Well-Being Celebrations and Traditions Discovering the World around Me.

Each task is presented in a specific language, but is adaptable to other languages. In addition to targeting a proficiency level, the assessments are also intended to measure the full range of student language ability across the three modes of communication: interpretive, interpersonal, and presentational. Each task follows a standard template that includes the following components: – – – – – – –

Task Title Theme Level (Novice-Mid, Intermediate-Low, Pre-Advanced) National Standards Goals Addressed Communicative Mode Time Frame [time needed to complete the task] Description of the Task (always stated in student-friendly language, requiring no adaptation on the part of the using teacher) – Materials Needed – Teacher Notes (any information that would be helpful for the using teacher to have in advance) – Adaptations. Listed here is a sampling of some of the novice level assessments. A significant number of the assessments are geared to elementary school

Assessment of early learning of foreign languages in the USA

79

students. Some of the assessments that are geared to secondary level students could possibly be adapted. – – – – – – – – – – – –

Art Appreciation – Mexican Art Art of Well Being – It’s Hot Out, I’m Thirsty! Celebrations – Carnaval in the Dominican Republic Discovering the World around Me – What’s for Sale Environment – Fire Safety and Prevention Family – People Important to Me How I Spend My Free Time – Elena the Whale Leisure Time – My family People Important To Me – A New Friend The Art of Well Being – I Eat Well The Environment – Let’s Take Care of the Air Work & Career – I Want to be. . .

This collection of performance-based assessments assembled by the Consortium for Assessing Performance Standards is important because it is a template for the nation on how to go about fashioning quality assessments that are based on the content and performance standards found in the ACTFL Performance Guidelines for K-12 Learners and the Standards for Foreign Language Learning in the 21st Century. These sample assessments give the guidance needed by classroom teachers to go ahead with designing authentic performance-based assessments for their own students.

5.

Conclusion

This chapter has reported on the early language learning situation in the United States regarding proficiency outcomes tied to the ACTFL Standards for Foreign Language Learning and the ACTFL Performance Guidelines for K-12 Learners. These tools have enabled the language profession in the USA to begin to make comparisons in the proficiency outcomes of programs that vary along a continuum that ranges from 15–20 minutes once per week to 60-minute classes every day. For the first time in this country parameters have been set for language outcomes that might be accomplished within various time frames. This chapter also reports on various USA-based young learner assessment measures that are based on the above-mentioned tools. It is hoped that

80

Helena Curtain

this information will provide the basis for an exchange of ideas related to the most important task of creating quality language education and assessments for young learners.

References American Council on the Teaching of Foreign Languages 1982 ACTFL Provisional Proficiency Guidelines. Hastings-on-Hudson, NY: American Council on the Teaching of Foreign Languages. 1998 ACTFL Performance Guidelines for K-12 Learners. Yonkers, NY: Author. American Council on the Teaching of Foreign Languages 1999 Standards for Foreign Language Learning in the 21st Century. Yonkers, NY: ACTFL. Branaman, Lucinda, Nancy Rhodes, and Jeannie Rennie 1998 A National Survey of K-12 Foreign Language Education. The ERIC Review 6 (1): 13. Council of Europe 2001 Common European Framework of Reference for Languages http://www.coe.int/t/dg4/linguistic/Source/Framework_EN.pdf Retrieved from Internet March 18, 2008. Curtain, Helena and Carol Ann Dahlberg 2004 Languages and Children: Making the Match, 3rd ed. New York: Pearson Allyn & Bacon. D’Anglejan, Alison, Harley, Birgit, and Shapson, Stan 1990 Student Evaluation in a Multidimensional Core French Curriculum. Canadian Modern Language Review 47 (1): 106–124. Fall, Thekla, Adair-Hauck, Bonnie, and Eileen Glisan 2007 Assessing student’s oral proficiency: A case for online testing. Foreign Language Annals 40 (4): 377–406. Gilbert, Jacqueline, Mary Mackenzie, Carol Meulener, Martin Smith, Beatrice Yetman, and Rosanne Zeppieri 2008 Consortium for Assessing Performance Standards: A New Jersey FLAP Grant Project Retrieved from the internet http://flenj.org/ CAPS/ March 18 2008. Falsgraf, Carl and Ann Tollefson 2006 The national online early language learning assessment (NOELLA) Learning Languages 12 (1): 18–19. Linguafolio Jr 2008 Retrieved from www.doe.virginia.gov/linguafolio/junior.htm March 18, 2008. Liskin-Gasparro, Judith E. 2003 The ACTFL proficiency guidelines and the oral proficiency interview: A brief history and analysis of their survival. Foreign Language Annals 36 (4): 483–491.

Assessment of early learning of foreign languages in the USA

81

National Standards in Foreign Language Education Project 1999 Standards for Foreign Language Learning in the 21st Century. Yonkers, NY: Author. Rhodes, Nancy, and Thompson, Lynn 1990 An oral assessment instrument for immersion students: COPE. In: A. M. Padilla, H. H. Fairchild, and C. M. Valadez (eds.), Foreign Language Education: Issues and Strategies, 75–94. Newbury Park, CA: Sage. Smith, K. 1995 Assessing and testing young learners: Can we? Should we? In D. Allen (ed.), Entry Points: Papers from a Symposium of the Research, Testing, and Young Learners Special Interest Groups. Kent: Cambridge 17th–18th March 1995 Whitstable, Kent: IATEFL. Swender, Elvira and Greg Duncan 1998 ACTFL performance guidelines for K-12 learners. Foreign Language Annals 31 (4): 479–483. Thompson, Lynn, Beverly Boyson, and Nancy Rhodes 2006 Administrator’s Manual for CAL Foreign Language Assessments, Grades K–8 Early Language Listening and Oral Proficiency Assessment (ELLOPA), Student Oral Proficiency Assessment (SOPA), CAL Oral Proficiency Exam (COPE). Washington D.C.: Center for Applied Linguistics.

4.

Assessing young language learners: What is the construct?

Ofra Inbar-Lourie and Elana Shohamy Considerable e¤ort has been invested in recent years in developing language programs for teaching young learners (Curtain and Dahlberg 2004; Nikolov and Mihaljevic Djigunovic 2006). These programs vary in their objectives and formats and in the relative emphases they place on language and content (see Martin 2000). The increasing demand for programs for young language learners (YLL) has led to a growing need for, and an interest in designing suitable assessment tools for determining the quality of the language learned (Bailey 2008; McKay 2006; ReaDickins 2000). This very issue is especially pertinent in the teaching of English as a Lingua Franca (ELF) for its status as an international language has brought about an unprecedented demand for English classes for YLLs worldwide (Graddol 2006). This chapter seeks to explore issues related to the interfaces and connections between various types of teaching programs and assessment practices for YLLs. The structure of the chapter is as follows: we begin by surveying the prevailing types of current language programs for YLL focusing on issues related to ‘language’ and ‘content’. We then examine the congruence between assessment schemes specifically geared for YLLs and these types of programs. Finally, we propose an assessment approach which is based on insights gained from a study that evaluated the teaching of English to first grade students in Israel. This approach, we argue, can be considered as more valid for it better reflects the pedagogical goals and assessment needs of programs for YLLs.

1.

Current approaches for teaching young language learners

A review of the present trends in language programs for YLLs demonstrates extensive variability in terms of objectives, content, teaching modes and expected outcomes. One emerging characteristics is the different foci on language versus content, i.e., an emphasis on language

84

Ofra Inbar-Lourie and Elana Shohamy

components such as grammar, lexicon and on generally viewing language as the goal of instruction rather than as a means, versus programs where the language is utilized to teach certain content areas like mathematics, history, or science. Related to the latter is also the stressing of cognitive competence associated with the specific topics (e.g., mathematical thinking) (Deller and Price 2007; Edelenbos, Johnstone, and Kubanek 2007; Johnstone 2000). We suggest that the di¤erent types of programs can be placed on a continuum ranging from those that emphasize language awareness, on one end, to those which stress mostly content, on the other; such a continuum draws on the scheme proposed by Met (1999).

Figure 1. A language-content continuum for teaching languages to young learners

Figure 1 depicts the continuum where point A represents language awareness programs and point E represents content. Between these two extremes we refer to programs that di¤er in the proportions of these components. Point A then refers to programs which aim to gain explicit knowledge about the language learned rather than to explicitly learn the language as a tool for communication (Hawkins 1984; James 1999). This approach to teaching YLLs implies that learners develop an understanding of and familiarity with one or more language systems as well as development of metacognitive skills to contrast other languages with those of their own. Thus, in these types of programs only little emphasis is placed on developing actual language communicative competence in the target language (Martin 2000). Further along the continuum, in point B, we see programs that we termed as ‘language focused’ as they represent traditional types of language teaching that view competence in the language as a central teaching goal. Typically, such programs are taught by language teaching experts whose teaching focuses on language topics such as lexicon and structures in the target language. The language teaching themes are not intended to draw on or complement the themes in specific content areas

Assessing young language learners

85

concurrently studied by the learners in other classes. In the case of YLLs the topics studied in language focused programs are usually selected from the learners’ environments and include topics such as, the family, food, school, numbers, colors, animals and body parts. Such programs are characterized by a strong emphasis on the lexical items associated with these topics and are aimed at promoting language knowledge (Rixon 2007; Williams 2005). Yet, a more ‘content-based’ teaching approach that integrates topics from other school subjects into the language teaching programs is depicted in point C of the continuum. In this approach language is integrated and embedded with content (Jantscher and Landsiedler 2000; Johnstone, 2000). The content is derived from subjects taught in the general school curriculum. Johnstone (2000: 129) provides a description of this approach: Foreign language activities therefore do not exist in isolation, but are intended to fit into a wider pattern of pupils’ learning activity. For example, a first-language activity in mathematics may lead to foreign-language work on numbers in relation to (say) a geography task. The concepts underlying these numbers will have been established through the first language, and the foreign language activity will have the double purpose of not only enabling the pupils to learn these numbers in the foreign language but also of strengthening their understanding of the underlying concepts by extending their application to a new and di¤erent domain.

The continuing ‘embeddeness in a flow of events’ is perceived to have positive implications for learning as it creates links between the students’ familiar conceptual knowledge and the language learning experience. It is believed that such programs can foster both cognitive development and language knowledge providing that the language teacher is also familiar with and knowledgeable about the various areas of the curriculum. The most suitable teaching model in this case is that of homeroom or generalist teachers, who can integrate the teaching of the language with other meaningful topics and concepts that they teach simultaneously to the same group of learners (Jantscher and Landsiedler 2000). Moving along the continuum yet further to point D, we find approaches that rely more heavily on content, and utilize the target language as a means or medium for acquiring knowledge in a number of school subjects. These models are broadly referred to as ‘languages across the curriculum’ or ‘CLIL’ (Content and Language Integrated Learning), and are implemented mostly in European contexts (see for example Deller and Price 2007; Eurydice 2006; Graddol 2006).

86

Ofra Inbar-Lourie and Elana Shohamy

Point E refers to ‘immersion types programs’ in which the language serves the content exclusively. The distinctions between CLIL and immersion approaches are not always clearly discernible. SeikkulaLeino (2007) di¤erentiates between ‘immersion’ and ‘CLIL’ programs on the basis of a number of parameters. For example, in the CLIL approaches students usually learn to read and write in their first language versus learners in immersion settings who first acquire literacy skills in the target language. Another di¤erentiation relates to the use of the foreign language as the medium of instruction which is implemented proportionally more in immersion type programs versus in CLIL. In terms of language proficiency goals, immersion programs aim at fluency in both the first and foreign language, while CLIL’s objectives often depend on specific settings. Yet, both models are perceived as contentoriented. Our main claim here is that the content-oriented teaching approaches in their various formats, are especially pertinent to English language teaching in view of the current status of English as an international language and a lingua franca, referred to as ELF (Jenkins 2007; Seidlhofer, Breiteneder, and Pitzl 2006). Specifically, we refer to the situation whereby English has permeated the global and local ecology (e.g., media, commerce, academia, the web), is deeply embedded in multiple contexts and intertwined with the local languages (Canagarajah 2007). In such settings learners often encounter English terminology in and out of the school context from a very early age in their use of the internet, computer games, books, and TV. In such settings the traditional language-focused approaches to language teaching become less relevant as English is not confined to the classroom and to the texts and materials introduced by the language teacher. Content-based programs on the other hand, seem to present more suitable approaches for English language teaching in this day and age. Yet, when surveying the commonly used tools for assessing English in YLL programs, we find that the assessment instruments focus by in large on language oriented approaches. In the remaining part of this paper we will demonstrate this phenomenon by selecting two cases where such testing instruments are used for assessing the English proficiency of YLLs. The first case is the widely used Cambridge Young Learners English (YLE) Tests (University of Cambridge ESOL Examinations 2007); the second is instruments we ourselves devised within the context of a study that examined the e¤ectiveness of di¤erent teaching models for YLLs (Shohamy and Inbar 2006). The insights gained from

Assessing young language learners

87

these analyses will lead us to recommend an approach to assessing young EFL learners which we believe is more compatible with the current content-focused modes of English teaching.

2.

Current approaches to the assessment of YLL programs: Two examples

The assessment of young language learners has only recently been recognized as a topic of study which merits specific focus and attention within the general language assessment domain (Bailey 2008; McKay 2006; Pavlou and Smith 2005; Rea-Dickins and Gardner 2000). Most writings on the topic promote the use of formative assessment for evaluating the language knowledge of YLLs in order to inform, improve and diagnose learning progress (Gattullo 2000; Edelenbos and Kubanek-German 2004). In terms of assessment tools, preference is generally voiced for the use of integrative language and content tasks which replicate real life situations and dilemmas and are intellectually challenging (MacKay 2006). However, contrary to such recommendations the assessment tools for young learners are often geared towards language-focused (rather than content-focused) approaches. This implies that the language components are assessed within conventional thematic areas such as family, animals and body parts, and rarely follow embedded or integrated content-based approaches which, as was noted above, are increasingly employed in many contexts. We will now illustrate this point by first referring to the widely used Cambridge Young Learners English (YLE) Tests (University of Cambridge ESOL Examinations 2007). The tests are aimed for three di¤erent levels: Starters, Movers, and Flyers. We will relate to the test intended for the beginning level entitled Starters as per the focus of this paper. 2.1.

Cambridge Young Learners English: Starters

The test is intended for beginning learners and consists of three parts: listening, reading, writing and speaking. It makes use of visual elements in the form of pictures except for the last oral activity which requires the test takers to provide short answers. The topics as described in the test manual (University of Cambridge ESOL Examinations 2007) include areas deemed familiar to YLLs, such as animals, parts of the face and

88

Ofra Inbar-Lourie and Elana Shohamy

body, clothes, colors, family, food, numbers, school, familiar objects, toys and transport (p. 10). In addition to vocabulary the test examines understanding of basic grammar and structures components. In the listening section students are expected to carry out instructions (e.g., position objects in a picture, locate objects, color objects), write down numbers and names, and choose a correct picture in a multiple choice task. In the reading/writing part, test items require marking ‘yes’ or ‘no’ to indicate a correct answer, choosing and copying words and writing single words in gap filling activities or in responses to questions. In the speaking section test takers are expected to carry out instructions, answer questions about object and scene cards, as well as respond to personal questions with short answers (Cambridge young learners’ tests handbook 2007: 5–15). The vocabulary, grammar and structures included in the test are listed in the test handbook. The construction of the test items attempts to comply with the needs and special characteristics of young learners in terms of the stimuli, topics and tasks. The aim is to make the testing conditions as anxiety free as possible and the testing experience a positive one, with each examinee receiving an award for having taken part in the test. Bailey (2005) in a review of the test found it to be practical and superior to other existing tests geared for young EFL learners. The review also notes the possible meaningful positive impact of the test on the teaching of English to young learners in various contexts, especially in terms of the materials produced (Bailey 2005: 251). However, in analyzing the test vis-a`-vis the approaches to teaching YLLs as presented in Figure 1, we found that this test follows the approach associated with ‘language focused’ instruction, point B on the continuum, which views language competence in the English language as central. According to this approach the teaching of English to YLLs is detached from other subject areas and is confined to concrete topics. Furthermore, the test does not build on the learners’ literacy development or previous first language experiences. The level of cognitive processing required is at the word of phrase levels of recognition or production, and there is no targeting of higher order thinking skills. Hence, the underlying approach reflected in the Starters test is not compatible with teaching programs that employ integrated or embedded language and content approaches. One also needs to consider that since the test is widely used it is likely to have meaningful washback e¤ects (Shohamy 2001) reinforcing even further a ‘language-focused’ rather than ‘integrated’ content-focused programs for teaching young learners.

Assessing young language learners

2.2.

89

Tests used as part of a research study

The second example we wish to present to illustrate and accentuate the gap between current approaches to teaching English to YLLs and the assessment procedures employed, is that of listening and speaking tests that we ourselves constructed as part of a research study. We will first describe the research study and the assessment tools used followed by the insights gained as a result of the research findings. The focus of the study was a comparison of the e¤ectiveness of two teacher models implemented for teaching English to young (6–7 year old) students: the first model comprised expert EFL teachers who do not teach the learners any other subject except English, while the second model was that of homeroom or generalist teachers who teach English to their first grade students along with the other subject areas, e.g., literacy, mathematics, science (Shohamy and Inbar 2006). The teaching and learning in these two models were compared on a number of parameters: the students’ achievements, their attitudes, students’ selfassessment and the teaching methods employed. The tools for measuring the students’ achievements were designed to fit the curricula of the teaching programs in the research sites, and included conventional YLL topics such as numbers, colors, family, familiar objects and animals. The assessment tools comprised of a listening comprehension test and an individual oral test. The listening component included a short story presented with a visual stimulus and multiple choice items in the form of pictures or illustrations. The speaking instrument was comprised of two parts. The first part was an individuallyadministered vocabulary recognition and production test using picture stimuli. Test takers were asked to identify objects from five semantic categories (numbers, animals, colors, family and food), and then to name the objects from the same semantic fields. The second part of the speaking test was conducted in the form of an individual oral interaction intended at eliciting short responses to personal questions as well as testing the students’ ability to follow simple instructions. The interactions were carried out in English with the learners’ L1 used whenever needed. Scoring was done on a three level assessment rubric which related to the learners’ ability to respond to questions, greetings and instructions, and to their willingness to engage in and initiate interactions in English. The findings of the study which, as noted, were aimed at finding out which of the teaching models was more e¤ective, showed that students studying with the EFL teachers scored significantly higher on the listening com-

90

Ofra Inbar-Lourie and Elana Shohamy

prehension test than those studying with homeroom teachers (86.46 versus 81.39, p < .05). In the oral test only small di¤erences were found in the word recognition section in favor of students who learned with the EFL teachers. However, we argue here that these results need to be interpreted with reference to the other research findings obtained in the study, specifically the data gathered via classroom observations. Each of the teachers who participated in the research (10 teachers from seven schools) was observed teaching English in the first grade classroom. Teachers were observed for the duration of two or three lessons, with each lesson lasting about 45 minutes. The observation tool focused on the class activities, in particular the opening and closing of the lesson, review of previously learned materials and the introduction of new topics, the activities and materials used, the classroom interactions, and the language used by both teachers and students (English versus the students’ first language, Hebrew). This observation data revealed di¤erences between the two teaching models: while the expert EFL teachers taught English as a subject in a way that was detached from other school subjects studied in first grade, (e.g., following a ‘language-focused’ teaching approach), the homeroom teachers anchored and integrated their teaching within the knowledge of content areas, e.g., L1 literacy, arithmetic, science, arts and crafts and music. In doing so the homeroom teachers followed the notions of ‘the embedding model’ as described by Johnstone (2000). In one instance, for example, a homeroom teacher insisted on teaching the beginning EFL learners how to ask ‘WH’ questions in English despite the students relatively low proficiency in the language. In the interview which followed she explained that she decided to work on WH questions since the children were concurrently studying how to form questions in their L1, Hebrew, in two other subject areas: in L1 literacy and in mathematics. The homeroom teacher felt that presenting an overriding conceptual understanding of what it means to ask such questions in a number of content areas contributed towards establishing and solidifying this concept in the children’s cognition. Our view then is that the aforementioned conventional languagefocused assessment tools that we developed for this research project did not pick up on the very unique teaching and learning experiences and rich interactions that took place in the classroom where content and English were embedded. Our tests focused on lexical items and overlooked these rich interactions where students acquired and discussed conceptual knowledge, formed connections with other subject areas, used high order thinking skills and dialogued with the teacher and with

Assessing young language learners

91

other learners in a mixture of English and their first language. Our claim, therefore, is that the results obtained via usage of the conventional types of listening and speaking tests represent a very narrow dimension of knowledge gained by the language learners, specifically when teaching was conducted by the homeroom teachers. Thus, while the construct which emerged in teaching consisted of language embedded with content, the assessment construct we used was oriented towards the ‘language’ end of the continuum. Reflection on this process ensued in considerations of the interfaces between teaching and assessment and the need to develop a di¤erent more formative assessment construct, one that is closely integrated with the embedding teaching approach observed. We now turn our attention to describing this approach.

3. Proposed assessment format The assessment construct we propose here is grounded in a view of language as a tool for gaining knowledge and meaning making and for developing cognitive processing skills. The language teaching curriculum is closely linked to the general school curriculum, building on knowledge established through the first language. The assessment procedures mirror that approach. They therefore aim to assess not only language proficiency but also newly gained conceptual knowledge in various subject areas. The assessment tasks are integrated with the very learning activities that take place in the classroom and see the learners’ L1 as a vital tool for achieving the above goals. This approach is also grounded in views of assessment as a means to improve (rather than prove) learning (Broadfoot and Black 2004). The role of the assessment activity in this ‘assessment culture’ is to contribute to the teaching-learning process and to provide teachers and students with feedback for sca¤olding and formulating each step in the language learning process (Inbar-Lourie, 2008). Below we present some examples which illustrate this approach. Example 1: Most YLL programs include the topic of colors in their curriculum, the general objective being the ability to recognize and produce color words in the target language. However, since children aged 6–7 are usually familiar with colors in their first language one wonders whether the teaching needs to be confined to this objective, or whether it is possible to build on the existing knowledge, i.e., the ability to notice colors and di¤erentiate among colors, in order to gain novel conceptual

92

Ofra Inbar-Lourie and Elana Shohamy

knowledge. For example, the students could study about the role of colors in relation to customs in di¤erent societies, and to using colors to convey meaning in idioms or similes in di¤erent languages and cultures. Another possible direction is to introduce the di¤erentiation between ‘hot’ and ‘cold’ colors, a concept which may be studied in conjunction with the art class. The teaching and subsequent assessment schemes in this case, will be based on L-1 knowledge, but will focus on a new framework, one which distinguishes among the colors on the basis of the hot/cold categorization. The learning experience can be complemented by a critical discussion as to the legitimacy of the classifications (i.e., is this division always true?) or conventional norms and labeling in general. In the discussion the children will most likely use their L-1 or a mixed code of the two languages due to their very elementary English proficiency. Based on this learning activity the assessment task can, for example, require the learners to di¤erentiate among hot and cold colors in paintings asking them to try and provide the reason for the artists’ choice. Thus the assessment will not only be language-oriented but will also focus on meaningful content, linking or integrating knowledge gained in the domains of art and English. Example 2: In the area of arithmetic, students perform computational functions and process mathematical concepts rather than simply engaging in learning the equivalent word (or term) for the number already known in their L1. Students can then learn about new concepts such as, ‘even’ and ‘odd’ numbers, about counting in fixed intervals (3, 6, 9, 12. . .), and about matching mathematical symbols to functions is in line with L-1 and L-2. The assessment task will focus on the conceptual understanding as well as the learners’ language use which may be in a dual hybrid format (L-1 and L-2). Example 3: When studying about literary genres the new language can serve to demonstrate and solidify concepts and strategies studied in L-1 in the domain of literary analysis, such as analyzing stories and poems, identifying genres (such as fables) and literary conventions. The assessment task will, for example, ask students to di¤erentiate among genres pointing out their typifying features. Example 4: In the music class concepts regarding musical conventions and symbols such as notes, musical keys and letter codes can be used to solidify the letter symbols and infuse some meaningful content into the learning, e, g., f stands for forte, pp stands for pianissimo. Hence in terms of assessment students will not only be requested to identify these letters but also to point out what they represent in terms of the musical conventions studied.

Assessing young language learners

93

Example 5: text messaging. Text messaging utilizes abbreviated language forms to convey a message, and often uses shortened versions of words or expressions in the English language. Teaching text messaging, explaining the rationale for its systematic use, even its grammar, comparing it with other ‘languages’, introducing some common abbreviations and asking the learners to produce some of their own, all of these can constitute a realistic language and content activity which would then be assessed using a similar task type, e.g., write a text message in English to congratulate a friend on his birthday. In all of the above examples the assessment task, like the teaching activity, strives to assess meaningful language use embedded within a relevant content. Students are encouraged to bring to the learning and assessment process the knowledge they have gained from other content areas, to comment, discuss and reflect on their language acquisition process. The children’s first or dominant language is viewed as an important component in the learning and assessment processes, serving as both the basis for furthering one’s knowledge and expressing new realizations which emerge and can not, at this stage of the language learning process, be articulated in the new language (Similar examples are o¤ered in immersion settings for languages other than English in Australia on a site called ‘‘Assessment for Learning’’ Curriculum Corporation http:// cms.curriculum.edu.au/assessment/default.asp). This construct we feel is more valid for the current approaches of content-based language teaching approaches which are of particular uses for YLLs in situations where English is taught at an early age by homeroom teachers and is line with the current view of English as a lingua franca.

4. Conclusions The vast increase in young learner programs requires a systematic debate on the role and format of the assessment of the knowledge and skills gained in these programs. This is pertinent to both classroom and external assessment. Clearly there is a need for further research which would explore the assessment constructs along the language-content program continuum, particularly within integrated or embedded models in young learner classrooms. Such research would provide data on how teachers and learners utilize classroom assessment for promoting learning and whether this use is indeed unique to this learner population and how it complements current teaching approaches. In addition, there is a need to examine the implications of such constructs for defining lan-

94

Ofra Inbar-Lourie and Elana Shohamy

guage knowledge in di¤erent contexts, particularly in the case of English given its present status and use. The process which is reported here, whereby research on language learning and teaching led to re-thinking on the suitability of commonly used assessment tasks, demonstrates the need for language teaching and language assessment experts and researchers to interact and incorporate findings from each other’s sub-disciplines, for as Shohamy (1998:172) says: ‘‘It is clear that researchers who collect language acquisition data and testers who measure language for making decisions must begin to pay attention to one another’s work.’’. This reflective process also illustrates the dynamics of testing construction and validation which constantly need to be re-thought and revised in view of current developments and new understandings.

References Bailey, Alison L. 2005 Test review: Cambridge Young Learners English (YLE) Tests. Language Testing 22 (2): 242–252. 2008 Assessing the language of young learners. In: Elana Shohamy and Nancy H. Hornberger (eds.), Encyclopedia of Language and Education, 2nd ed. Vol. 7. 379–398. Language Testing and Assessment, Springer Science þ Business Media LLC. Broadfoot, Patricia and Paul Black 2004 Redefining assessment? The first ten years of assessment in education. Assessment in Education: Principles, Policy and Practice 11 (1): 7–26. Canagarajah, Suresh. 2007 The ecology of Global English. International Multilingual Research Journal 1 (2): 89–100. Curriculum Corporation Assessment for Learning Melbourne, Australia http://cms.curriculum. edu.au/assessment/about.asp (accessed March, 2, 2008). Curtain, Helena and Dahlberg, Carol Ann 2004 Languages and Children: Making the Match: New Languages for Young Learners 3rd ed. Needham Heights, MA: Pearson Allyn & Bacon. Deller, Sheelagh and Price, Christine 2007 Teaching Other Subjects through English. Oxford: Oxford University Press. Edelenbos, Peter and Anglika Kubanek-German 2004 Teacher assessment: The concept of ‘diagnostic competence’. Language Testing 21 (3): 259–283.

Assessing young language learners

95

Edelenbos, Peter, Richard Johnstone, and Angelika Kubanek 2007 The main pedagogical principles underlying the teaching of languages to very young learners. Languages for the children of Europe. Published Research, Good Practice and Main Principles. Final Report of the EAC 89/04, Lot 1 study European Commission Education and Culture. http://ec.europa.eu/education/policies/lang/doc/young_en.pdf Eurydice European Unit 2006 Content and Language Integrated Learning (CLIL) at School in Europe. Brussels: Eurydice. http://www.eurydice.org/ressources/ eurydice/pdf/0_integral/071EN.pdf Gattullo, Franceska 2000 Formative assessment in ELT primary (elementary) classrooms: an Italian case study. Language Testing 17(2): 278–288. Graddol, David 2006 English Next. Manchester: The British Council. www.britishcouncil. org/learning-research-english-next.pdf Hawkins, Eric 1984 Language Awareness in the Curriculum: An Introduction. Cambridge: Cambridge University Press. Inbar-Lourie, Ofra 2008 Assessment Culture. In: Elana Shohamy (ed.), Language Testing and Assessment, 285–300 (Vol 7). In Hornberger, N. (general editor), Encyclopedia of Language and Education 2nd ed. New York: Springer Science and Business Media, Inc. James, Carl 1999 Language Awareness: Implications for the Language Curriculum. Language, Culture and Curriculum 2 (1): 94–115. Jenkins, Jennifer 2007 English as a Lingua Franca: Attitude and Identity. Oxfords: Oxford University Press. Johnstone, Richard 2000 Context-sensitive assessment of modern languages in primary elementary) and early secondary education: Scotland and the European experience. Language Testing 17 (2): 123–143. Jantscher, Elisabeth and Isabel Landsiedler 2000 Foreign language education at Austrian primary schools: An overview. In: Marianne Nikolov and Helena Curtain (eds.). An Early Start: Young Learners and Modern Languages in Europe and Beyond, 13–27. Strasbourg: Council of Europe. Martin, Cynthia 2000 Modern foreign languages at primary school: A three-pronged approach? Language Learning Journal 22: 5–10. McKay, Penny 2006 Assessing Young Language Learners. Cambridge: Cambridge University Press.

96

Ofra Inbar-Lourie and Elana Shohamy

Met, Miriam 1999 Content-Based Instruction: Defining Terms, Making Decisions. College Park, MD: National Foreign Language Center. Nikolov, Marianne and Jelena Mihaljevic Djigunovic 2006 Recent research on age, second language acquisition, and early foreign language learning. Annual Review of Applied Linguistics 26: 234–260. Cambridge: Cambridge University Press. Pavlou, Pavlos and Kari Smith (eds.). 2005 Serving TEA to Young Learners. Proceedings of the conference on Testing Young Learners, organized by the University of Cyprus, IATEFL and CyTEA. Tivon: Oranim College of Education. Rea-Dickins, Pauline 2000 Current research and professional practice: reports of work in progress into the assessment of young learners. Language Testing 17 (2): 245–249. Rea-Dickins, Pauline and Sheena Gardner 2000 Snares and silver bullets: disentangling the construct of formative assessment. Language Testing 17 (2): 215–243. Rixon, Shelagh 2007 Cambridge ESOL YLE tests and children’s first steps in reading and writing in English’. Cambridge ESOL Research Notes 28: 7–14. Seidlhofer, Barbara, Breiteneder, Angelika. and Pitzl, Marie-Luise. 2006 English as a lingua franca in Europe. Annual Review of Applied Linguistics 26: 3–34. Seikkula-Leino, Jaana 2007 CLIL Learning: achievement levels and a¤ective factors. Language and Education 21 (4): 328–341. Shohamy, Elana 1998 How language testing and SLA benefit from each other? The case of discourse. In: Lyle Bachman and Andrew Cohen (eds.), Interfaces between Second Language Acquisition and Language Testing Research, 156–176. Cambridge: Cambridge University Press. 2001 The Power of Tests. Harlow: Pearson Education Ltd. Shohamy, Elana and Ofra Inbar-Lourie 2006 The teaching of EFL in the first grade: EFL teachers versus homeroom teachers. The Pedagogical Secretariat, Jerusalem: The Israeli Ministry of Education (in Hebrew). University of Cambridge ESOL Examinations. 2007 Cambridge Young Learners English Tests Handbook: Starters, Movers and Flyers. Cambridge: UCLES. Williams, Melanie 2005 Teaching and assessing young learners: balancing language, content and task. In Pavlos Pavlou and Kari Smith (eds.) Serving TEA to Young Learners. Proceedings of the conference on Testing Young Learners, organized by the University of Cyprus, IATEFL and CylTEA, 17–23. Tivon: Oranim College of Education.

5.

Language proficiency and age-appropriate methodology at the English Kindergarten of Kuopio

Joanne Jalkanen

1.

Introduction

This chapter concerns young learners of English in the early total immersion programme at the English Kindergarten of Kuopio (Enkku). A brief history of the organisation is provided, followed by a discussion of the classroom practice and the methodology used. The second part presents the results of a language proficiency study, using an assessment tool designed specifically for this purpose, from the point of view of the Critical Period Hypothesis (CPH), to identify and explain the strengths and weaknesses of these young learners. The general assumption exists that younger learners perform better than older ones, and the CPH tends to support this. The CPH suggests that children are able to attain native-like levels of language proficiency if the second language acquisition takes place before the end of a so-called critical period, which has also been referred to as a ‘window of opportunity’. This sensitive period is said to end at puberty, although some researchers claim that it is over as early as the age of four or five years. It is linked to the idea that the brain begins to lose its plasticity after puberty (see Birdsong 1999 for a discussion on stances). It seems to be generally agreed that very young children do seem to pick up the phonemes of the second language with ease, especially in naturalistic learning situations (Ellis 1997: 484–493). Nevertheless, early second language learners have not yet reached the stage whereby they are able to consider language as a system. Many accepted methods of teaching a second language to adults and older children, for example, grammar-translation, make certain assumptions about the learner, which consequently cause them to be unsuited to the second language instruction of young children, i.e. most children under the age of seven, and all under five (Tough 1977: 80–81). This is because the level of conceptual development attained at these ages does not make such thinking and motivation possible. Children’s metalinguistic

98

Joanne Jalkanen

development has not yet advanced to the stage of their being aware that they are using language. They are still generally unable to consciously commit themselves to learning the language, that is, to accepting that vocabulary must be memorised and grammar rules learned and applied in practice (Tough 1977). For young learners, language is the inherent tool that they use to make sense of and interact with their world. The philosophy of immersion as it is understood here derives from Canada in the 1960s. The radical St Lambert programme was implemented in Quebec in 1965, whereby monolingual English-speaking children received their instruction in French only from kindergarten (age five years) onwards. The programme, conceived in collaboration between parents and experts on bilingualism, was labelled an immersion programme (Johnson and Swain 1997: 2). Johnson and Swain (1997: 6– 8) outline eight core features of a prototypical immersion programme, which describe immersion as it is understood here. These features are that the second language (L2) is the medium of instruction; the curriculum used in the L2 immersion programme parallels that of the monolingual students; there is overt support for the mother tongue (L1); the programme aims for additive bilingualism (meaning that the L2 is learned at no cost to the mother tongue); the exposure to the L2 is largely confined to the classroom; students entering the programme have a similar and usually limited level of language proficiency in the L2; the teachers are bilingual in the L1 and L2; and the classroom culture is that of the local L1 community. Early immersion means that the students are of kindergarten and pre-school age (in Finland from three years to seven years of age). Therefore, in this instance, immersion is relevant to early childhood education. Total immersion means that the second or foreign language (here English) is the working language of the classroom, and the children receive all of their instruction in it. Early total immersion appears to be a successful means of teaching young learners a foreign language. It is a form of bilingual education, the major focus of which is on the development of communication skills in the target language. Metalinguistic aspects of the language, such as grammar are not dealt with directly, but may be addressed when problems occur. Immersion is an alternative option, parallel to the regular curriculum. The programme must be content-based and not languagebased, i.e. the language is the medium of instruction not the object. Here a distinction between second language ‘learning’ and ‘acquisition’ must be made. Language ‘learning’ is concerned with the accumulation of knowledge of the grammar and vocabulary of a language, that is, the

Language proficiency and age-appropriate methodology

99

conscious process of studying the language. Language ‘acquisition’ applies to the gradual development of ability in a language through its use in naturalistic communicative situations (Yule 1997: 191). Children learn the target language in a way that closely resembles how they learn their mother tongue. The goal of the method is additive bilingualism; to enable students to become functionally bilingual, that is, to be able to communicate e¤ectively and with ease in formal and informal situations on a range of topics, live in either speech community without any cultural or linguistic di‰culties in adjusting, and be able to be professionally educated in either language (Helle 1985: 111–112).

2.

The English Kindergarten and Pre-school of Kuopio

A Finnish day-care centre is a Finnish institution; the fruit of a long history of the state supporting parents in the task of raising their child, and its roots go back to the 1800s. After the 1900s, day-care also began to take child welfare issues into account and day-care centres were to help families that could not cope with the task of child rearing. The 1960s saw more Finnish women wanting the right to go out to work and a law was passed to this e¤ect in 1973: thus, the ‘storage’ function became more important. In 1983, day-care was to provide warm and safe relationships and support the physical, social and emotional development of the child. A further aim was to support parents in the task of raising their child. The focus for day-care changed again in the 1990s, when each child was allotted the right to day-care, regardless of whether or not the parents were in work, that is, day-care was no longer just for the socially or economically weak, but was considered as a part of everyday life in Finland (Alasuutari 2003: 24–28). English language early childhood institutions (i.e. kindergartens, nurseries, etc.) also support parents in the task of raising their children, but their role is less of a societal one. These institutions have a linguistic function to fulfil in order to justify their existence. While Finnish daycare centres are seen as a matter of course, foreign language kindergartens are an optional alternative. Therefore, sending a child to learn a language that is not an o‰cial language of the country in which he lives should be the result of a conscious decision taken by the parents, and is a manifestation of the desire for something that is di¤erent to the regular options. Many such institutions remain privately run, as is the case here.

100

Joanne Jalkanen

The English Kindergarten of Kuopio was founded in 1969 and is among the oldest and largest English language kindergartens in Finland. It was founded to provide Finnish children with the opportunity to receive naturalistic early English second language instruction from native speakers. In 1991 it became the English Kindergarten and Preschool of Kuopio, and this was also the year of the inception of the immersion programme. Prior to this, the system was such that each class had a native-speaker teacher and a Finnish assistant whose task was to translate what the teacher said. This system was deficient in two major ways. Primarily, it was not conducive to learning English. As can only be expected, the majority of the children did not listen to the English; they waited for the translation in Finnish. Secondly, it demanded a great deal from the assistant, and could even result in her being regarded as a higher authority than the teacher because of the knowledge and use of the children’s mother tongue. Implementing the early total immersion programme according to the Canadian model provided a solution to this problem by ensuring that the children were exposed to as much English as possible during their time at Enkku, and without undermining the teacher’s authority. It also appeared to dispense with a problem often discussed by teachers at foreign language kindergartens, namely, which should take priority, learning social skills or language learning? In immersion, the social skills are learned in the new language. Concurrently, the entry age was raised to four years. Since most normal indigenous children have established their first language (mother tongue) by the age of 3½ to four years, four-year-olds should be better prepared to cope in the second language environment. Their mother tongue should be developed enough to enable them to process visual/aural input in English using concepts and structures that already exist through their experiences in Finnish. Earlier, Enkku o¤ered only half-day immersion and had morning and afternoon groups because the premises were not suited to full-day day-care. Since August 2002, the pre-school has operated on a di¤erent site to the kindergarten because of a lack of premises big enough to hold both parts. The pre-school has o¤ered full-day immersion since then, and the first experiences of this have been very positive. This has been reinforced by the excellent feedback from the pre-school’s feeder school, Rajala School, which o¤ers an English content and language integrated learning class for former Enkku pupils and bilingual children with English as a home language. The kindergarten moved sites and was able to o¤er full-day immersion from August 2003. There are four

Language proficiency and age-appropriate methodology

101

classes at Enkku, and it can take a maximum of 80 full-day children. It is bound by the same rules and regulations as Finnish day-care centres concerning sta‰ng ratios, safety regulations, and so on. Each class has a teacher and two qualified nurses in accordance with Finnish law, that is, one qualified member of sta¤ per every seven children.

3.

Methodology in practice and the social nature of language acquisition

Young learners’ second language instruction should be holistic, and take account of each child as an individual. It is imperative that the teacher builds a good working relationship with each child, and all situations should be put to use to allow the child to use the language. Often at first, the child goes through an initial silent period, and relies on help (or scaffolding) from the teacher or older peers to learn chunks of English. However, in the first weeks, the children can acquire many such chunks as they become accustomed to the routine. Even small routines such as saying good morning are important to the children, and provide them with real situations in which they can use English. This serves a confidence-building purpose, and motivates the children because they can quickly start to use English themselves, and feel pride when the adult responds to their output. Each classroom is divided into di¤erent play stations, for example, a Lego station, jigsaw table, library corner or home corner (which may be a doctor’s surgery or shop depending on the current topic). Having the play stations makes the classroom more child-centred as it allows the children a certain amount of autonomy. The children are also taught to be responsible for taking care of their classroom and the things in it. Each day, monitors are chosen to help set up tables for meals, keep the books tidy and change the calendar. Independence is very much encouraged: each child is responsible for putting away his worksheets, tidying the classroom, and after each meal, for taking his dirty dishes to the trolley and stacking them in the right places. Everything is done in English. Every day the children have activities with the teacher and the nurses in groups of four or five. These are based on the principle of the ‘integrated day’, meaning that the children have a specific amount of activities, but these can be done at any time throughout the day. The aim is,

102

Joanne Jalkanen

Figure 1. The integrated day system at the English Kindergarten of Kuopio

however, that the majority of activities will be done in the morning, leaving the afternoon free. Almost all of the activities are carried out as small group work. When they are not ‘working’ with one of the adults, the children are free to move around the play stations as they wish, but must ask in English to move from one station to another. The sta¤ make sure that the children have time to play between each activity with a teacher. The number of children who can be at each play station at any given time generally ranges from 3 to 6. Figure 1 shows the components that interact during the child’s integrated day. 3.1.

Introducing new language

New language is presented during activities, either in a small group with an adult, or together with the teacher on the mat, for example, music, singing, story time or drama. Rhymes and singing are a good source for the presentation of prosodic features, for example, intonation and stress. As all the activities revolve around a theme or topic, the same vocabulary is reinforced time after time. Furthermore, the children are exposed to the language in meaningful, authentic situations in which they are free to use their language (or interlanguage) and develop their social/a¤ective strategies. Taken from a psycholinguistic point of view, they receive the same input in a variety of situations, which should lead to intake. In addition, they have natural opportunities to try out their interlanguage hypotheses, and develop their language. Pictures and games that actively involve the children are an excellent way of introducing new vocabulary and structures because the input is

Language proficiency and age-appropriate methodology

103

comprehensible, and they tend to keep the children’s motivation high. This highlights the importance of play and the social nature of immersion, and supports the principle of learning by doing. Through this principle, the children do a great deal of handwork; painting, drawing, modelling, sewing, and so forth. The environment is made as stimulating as possible by the teachers, as this artwork is used to make colourful and educational displays on the walls. In this way, the classroom becomes the children’s own domain; each person contributes something, and the children are very proud of their e¤orts. In many respects the role of the environment is similar to that of the Montessori method: the learning environment (physical environment and the people within it) is considered of particular importance, because it is believed that children learn a great deal from it without significant input from teachers. The function of the environment at Enkku is very similar, although the teachers have a more active role to play because of the immersion method. Theoretically, this is a constructivist way of thinking and is echoed in the theories of Piaget and Vygotsky. The idea is that children construct their own understanding of the world, and the teacher’s role is to provide environments conducive to children being able to discover for themselves how the world works. It is widely accepted in language teaching that, among other factors, negotiation of meaning and social interaction are hypothesised to play a key role in language learning, hence the active role of the immersion teacher. Vygotsky-like pedagogy, in which the child learns in social interaction with others, is relevant to child-centred learning, and this type of pedagogy is one means used in immersion. It is best summarised in the concept of Vygotsky’s Zone of Proximal Development. Basically, this is the di¤erence between what the child can achieve unassisted and what he is able to achieve with assistance. The teacher’s role is one of facilitator, someone to provide sca¤olding. This means that the teacher must choose tasks and activities which challenge the child and require him to take risks, in order to be able to achieve more. Each child has his own strengths, preferences and weaknesses, which must all be taken into account.

3.2.

The child’s mother tongue

As additive bilingualism is the goal, the children are permitted to use their mother tongue in the classroom. Finnish is very much one of the

104

Joanne Jalkanen

classroom languages. This is one of the ‘myths’ concerning the immersion method. Many people assume that the children are not allowed to use Finnish, and do not realise that the children generally do speak Finnish among themselves. The children are always encouraged to produce their own utterances in English, but are never ignored when they use Finnish. The teaching sta¤ always respond, even if it is just to repeat what has been said in English. Thus, it is important that immersion teachers are themselves bilingual in the two languages; although this is not always possible. According to Grammont’s principle, however, the immersion teacher should have only one linguistic role. That is, the teacher should use only the immersion language with the students, whatever the situation, and should not give instruction in the mother tongue, so that the children only associate one language with one person. Experience has shown that the children generally try harder to use their English if they think that the English teacher speaks no Finnish. At times, depending on the individual, the teacher may pretend not to understand the Finnish and try to coax the child to say it in English. This is always done in a gentle and fun manner. In instances when a child is upset, he is taken aside and spoken to in Finnish. The mother tongue is a resource which can quickly and neatly clear up a misunderstanding or establish understanding. In immersion, the use of children with better English proficiency as translators is discouraged because they must all learn to use English actively. However, the use of translation should not be denied, and certainly in some instances is used in the classroom, especially with the younger children. Errors are not overtly corrected, because this could be devastating to the fragile self-confidence of the child, and may result in his withdrawal and avoidance of situations in which he must interact with the teacher. The preferred covert form of ‘correction’ is recasts: this means repeating what the child has said, but in the correct form. As the child’s proficiency increases, error correction also takes the form of clarification requests or elicitation. Humour is one good way of ensuring that the child hears the correct form without any dents to his self-esteem.

3.3.

The curriculum

Principally, Enkku is a place where children come to learn, and the sta¤ aims to provide a secure and stimulating atmosphere and environment for them to do so. As many of the children spend most of the day there, it has also to be as home-like and caring as possible.

Language proficiency and age-appropriate methodology

105

Figure 2. Early childhood education skills

Having a regular routine is important for any child, and especially so for children in an early total immersion programme, because this provides them with stability and security. At the beginning, when the child does not understand the language, just knowing what will happen at a given time on a given day helps him to cope better with the situation. The activities must have working strategies and aims which are clearly planned, and which are known to all the members of sta¤ in the school. The activities and routines also have a pedagogic role in that through them the children must also learn certain elementary social, physical, and motor skills. Enkku has employed an early years curriculum based on the United Kingdom system, but which covers the Finnish pre-school curriculum. Now, however, e¤orts are underway in Finland to develop a new early years curriculum, which will be incorporated. The learning at Enkku is individual and holistic, and each child ‘works’ at his own level in all matters. In addition to learning English, the general goals are to (1) develop the children’s social skills; (2) teach them to respect other people, participate in group work, and co-operation; (3) help them to be good, caring people who help each other; and (4) be independent, selfconfident individuals who can think for themselves and understand themselves as individuals and part of a group. Specific attention is paid to developing self-esteem. A further aim is for the children to develop a good work ethic in relation to their tasks and general readiness for school. All the e¤orts to work towards these goals take place in English.

106

Joanne Jalkanen

Figure 3. Incorporating the UK early years curriculum

4.

Assessing language proficiency

Enkku employs a system of continual assessment concerning the progress of the children in all areas. However, I developed an assessment tool to measure the English language proficiency of the children in the Kuopio situation, to ascertain how well they may be expected to cope with the English-medium programme at primary school. KATE, the Kuopio Assessment Tool for English, was developed in collaboration with colleagues at both Enkku and Rajala School, using an action research approach. It was largely inspired by the Student Oral Proficiency Assessment (SOPA) and the Early Language Listening and Oral Proficiency Assessment (ELLOPA) developed in North America by the Center for Applied Linguistics (CAL) (see Curtain in present volume). 4.1.

The KATE rationale

KATE comprises six tasks; a separate group task to assess knowledge of basic concepts in English; general conversation; a vocabulary task; discussing a picture; synthesis of simple scientific concepts; and disembedded questions. It is an oral assessment tool since very few of the children in question had reading and writing skills in either their mother tongue or English. Language proficiency was seen as a dynamic holistic compe-

Language proficiency and age-appropriate methodology

107

tence involving a range of components (pronunciation, grammar, etc.). The aim was to incorporate Cummins’s (2003) hypothetical sets of skills into the tool, Basic Interpersonal Communication Skills (BICS), Cognitive/Academic Language Proficiency (CALP), to facilitate the measurement of proficiency. As the name suggests, BICS are those skills that are activated when one interlocutor talks to another about everyday things in concrete situations in which the context and meaning can clearly be deduced and understood. Such skills can also be called surface fluency, and according to Skutnabb-Kangas (1981: 111), are measured by tests for accent and native-speaker-likeness, oral fluency, basic vocabulary, and basic syntax. CALP is the ability to use language in academic context reduced situations (Baker 1996: 151), for example, in problem-solving or analysis, where the meaning is not clearly demonstrated in the context. The di¤erence between BICS and CALP is usually portrayed in the image of an iceberg. Conversational proficiency is shown as the tip of the iceberg, visible above the surface, and includes such skills as speaking and comprehension, i.e. the surface features of the language. Cognitive proficiency is underneath the surface, and represents the skills of analysis and synthesis, such as creative composition, i.e. the deeper, subtler language skills. Studies on children’s cognition show that children are better able to understand what is expected of them with tasks that make use of the type of language or situations that make human sense to them, and to which they can relate their already existing knowledge of the world (e.g., Donaldson 1978: 19–25). For this reason KATE principally took a pragmatic approach. It also incorporated elements of the integrative or holistic approach in that the tasks measured more than one skill, and language processing skills were considered as comprising more than one language component at a time (a listening and speaking task can measure both grammar and pronunciation). KATE included only one task that is discrete-point in nature, the vocabulary lexicon task. KATE is a criteria-referenced assessment tool in the understanding of assessment as the evaluation of student performance or behaviour as related to established criteria (in this instance two rating scales, one for general fluency, the other focusing on cognitive skills). The cognitive tasks are to some extent norm-referenced in the sense that they correspond to what we felt could be expected of pre-school children in this programme, taking account of the skills they ought to have in English. It is formative in the sense that it focuses on what the learners can do as

108

Joanne Jalkanen

opposed to what they cannot, but is not used as a basis for further action. KATE is also summative in the sense that it records the overall achievements of each learner in a systematic way. I believe that KATE is a fair and valid tool to assess children’s language proficiency of this age. An additional aim of the original study was to assess the inter-rater reliability of the tool, by comparing the external assessor results with those from the class teacher, based on the continuous assessment of the children, but using the same scoring rubric. KATE scores corresponded surprisingly well with those of teacher assessment: 66% were either identical (45%) or had a half-score di¤erence (21%). A Finnish summary of the KATE was presented to the Educational Committee of the City of Kuopio at a meeting on the 26th of November, 2006. The Educational Committee decided to accept KATE as an o‰cial tool to measure the English language proficiency of children aged six to seven years. KATE comprises two parts: a controlled drawing observation task, which lasts about 30 minutes and was done on one day in a group; and an approximately 20-minute individual interview which took place on another day. All of the tasks were designed to focus on what the child could do or produce through the medium of English, and were presented as a game or something that is fun. The vocabulary lexicon task was time-consuming, but the assessor had to ensure that each of the other four interview tasks lasted approximately five minutes. The controlled drawing observation task concerned knowledge of basic concepts in English, such as shapes, colours, location, and so on, and was done in a group situation. As such, it also assessed how well the children were able to concentrate and work in a group situation in English. This task concentrated entirely on the children’s listening comprehension skills in English, assessed through what they produced on paper. Instances of copying were also noted, and in such cases, the child got the number of points based on his drawings, but mention was also made that he needed support from the others or looked at the others’ work. Four children in the target group did not do this task as they were away on both of the days on which it was organised. The first interview task consisted of general conversation based on a brief background questionnaire sent out prior to the interview, asking about the child (family, hobbies, and so on). Talking about familiar topics should put the children at ease and provide them with the best opportunity to display their fluency in English. The idea was to provide the child with the opportunity to say as much as possible, using, for

Language proficiency and age-appropriate methodology

109

example, open-ended questions, i.e. those that require more than just a ‘yes’ or ‘no’ answer. The presence of the cassette recorder was explained to them and it was switched on in their presence. The lexicon task or discrete-point vocabulary task was added after the results of the pilot test indicated that the general level of proficiency was high. It was done to ascertain whether or not such a task could be used to distinguish between good speakers of English. The task comprised five di¤erent pictures from age-appropriate children’s books: a kitchen, the seaside, a park, a farm, and a shop. Twenty items were identified and numbered in each picture, making the number of lexical items 100 in total. The fourth task involved talking about a picture. Four pictures were chosen and backed on to four di¤erent coloured papers. The child chose one of them, and this was turned face up. The aim was for the child to say as much about the picture as possible in his own words. If a child was shy or had problems, the assessor’s role was to prompt him using suitable questions and elicitation, as would happen in authentic language use. The fifth task was devised to test synthesis, which is included in the CALP skills. Each child had to choose a series of four pictures depicting a simple scientific concept related to the curriculum that they had already covered at pre-school. Four concepts were chosen, the four seasons (a year); the life of a plant, seed; the stages of a frog growing; the lifespan of a butterfly. Each series of pictures was mounted on di¤erent coloured card. At the interview, the piles of cards were face down on the table and the child had to choose his preferred colour. This random choice helped to reduce the child’s expectancy as to the content of the task. The idea was for the child to be able to synthesise the cards into one basic scientific concept. The child had to put the pictures into the correct sequence, thus displaying an understanding of the concept and then describe what the sequence depicted. Academic proficiency was assessed on whether or not the child was able to put the pictures into the correct sequence, and whether or not he could talk about the concept as a whole or was able only to discuss each picture as a separate unit. Notice was also taken of whether or not a child who could not name the concept in English could do so in Finnish, which would suggest only a lack of vocabulary or arbitrary signs in English. The final KATE task took the form of a series of ten disembedded questions, that is questions which occur out of context, and there are

110

Joanne Jalkanen

no clues available from the immediate physical environment as to the nature of the question or its answer. The function of this task was to provide the assessor with an idea of how well the child could think in English in context-reduced situations and provided an indication of the child’s budding cognitive use of English (CALP). The questions focused on everyday concepts or on logical-mathematical thinking. There were four sets of questions, each backed on a di¤erent colour of paper. Again, the child was able to choose one colour. The final question was always a joke, ‘‘What is green and hairy and goes up and down?’’

4.2.

Scoring rubrics and results

Use was made of the Student Oral Proficiency Rating (SOPR) scale’s rationale to rate the children’s BICS. SOPR was developed by Development Associates, Inc., and is a modification of the Student Oral Language Observation Matrix (SOLOM) developed by the San Jose (California) United School District. Only the rating scale was used, not the instrument itself. It consists of five components of proficiency (comprehension, fluency, vocabulary, pronunciation, grammar) and each has five levels or grades. It was felt that the components and levels were well-explained and would serve as a point of departure, to be modified based on experience. In KATE, each level is seen as a percentage, and this also allows the assessors to give a mark on the borderline between two levels. Grades were given using the scale 1–5 and half-scores were also accepted; where one is 20%, 1 to 2 is 30%, two is 40%, five is 100% and so forth. The ratings for comprehension, fluency, general vocabulary, and grammar were given based on the impression gained from all of the KATE tasks. These constitute the Basic Interpersonal Communication Skills. The ratings for general cognition, synthesis and context-reduced situations come from the final two tasks and represent the Cognitive/Academic Language Proficiency. The children’s CALP was rated using a scale that I developed to fit the tasks. Again this was to be a point of departure to be modified after it had been tried during the interviews. The cognition referred to is the children’s cognitive ability in English and should not be confused with their cognitive ability in their mother tongue, which, it is felt, should be higher. The first section is for general cognitive skills, based on the impression gained during the interview. The second and third sections are relevant to the cognitive interview tasks. This rating scale has five

Language proficiency and age-appropriate methodology

111

levels or grades in accordance with the SOPR rating scale, and the assessment grades are given using the same principle. 4.3.

Test results

The results presented here are those given by the KATE assessor to the target group (n ¼ 27). The results of the pilot group (n ¼ 6) are not included because it consisted of a range of ages, from five to almost eight years in order to ascertain if the operationalisation of the instrument had been done correctly and that it was suitable for the target agegroup. Hence the group in question consisted of 27 Finnish children (10 boys and 17 girls) aged six to seven years from two pre-school classes. The children were in either their second or third year in the immersion programme. The observed drawing and discrete-point vocabulary tasks are addressed separately, followed by two figures that concern the assessment of language proficiency components divided into the hypothetical concepts of BICS and CALP. To score the observed drawing task, a rating scale was devised, based on a number of points for each concept; the maximum number of points was 50. Each child received a percentage score, for example, in the case of a student getting 40 points out of 50, the percentage would be 80, which could then be related to the scores of the scoring rubrics (0–9 points was one, 10–19 points was two, and so on). The results of the observed drawing task showed that most of the children who did this task (n ¼ 23) had a fairly good knowledge of basic concepts in English. For the target group, the lowest score was 68% and the highest was 98%. The mean score for this task was 85.9 per cent. The discrete-point vocabulary task was rated separately. An answer rubric was made and synonyms for correct answers listed, for example, spade and shovel. It was agreed that if the word used to describe the item in the picture could conceivably be used by a native speaker, then it was considered as correct. The exception to this was with regard to several birds in the park picture – here the aim was for the children to be able to name the type of bird portrayed. The child only received a point for a correct answer in English. No points were given for knowing the item in Finnish. As the number of lexical items was one hundred, it was easy to obtain a percentage which could then be related to the other rating scales. However, the rating scale did note when the answer was correct (a tick), wrong (a cross), not known (a question mark) or given in Finnish (an ‘F’).

112

Joanne Jalkanen

All of the children showed a drop in score for this activity, but there appeared to be no consistent pattern between the percentage obtained in the general impression of vocabulary and this discrete-point vocabulary task. I feel that the task was not entirely suitable for the children in question, as it was too di‰cult. There were too many items and the children’s attention and concentration wavered from time to time. In instances when a run of wrong answers or not knowing the vocabulary in English occurred, the children became noticeably demoralised. This was a shame because many really tried to please the assessor. In addition, this task was time-consuming and rather laborious for the assessor. Some children were also very concerned about the marks the assessor was making on the paper. Perhaps it would have been better to take a more pragmatic approach to this task instead of a discrete-point one. The same idea of testing the vocabulary in a given situation could have been retained by using a cloze test. By taking a story and removing every seventh word or so, the children would have the opportunity to choose the word they felt best fitted the situation. This avoids the idea of only a right or wrong answer and allows for at least some negotiation of meaning. Furthermore, any children relying on ‘I don’t know’ may be prompted to provide a more imaginative answer.

4.4.

Results for BICS- and CALP-skills

The results showed that the overall level of proficiency was high, and this is immediately apparent from Figure 4. The assessed children’s comprehension of English is excellent, that is, they can understand normal conversation at normal speed without any great di‰culty or that they can understand well, but may need some repetitions. There is a wider range of scores for fluency, but again the level is generally fairly high: over 80 per cent of the children either speak fluently or fairly fluently with only occasional lapses as they search for the correct terms. The same trend can be seen concerning the level of general vocabulary skills, which ranged from having somewhat limited conversation skills and some use of wrong vocabulary, through occasional use of wrong vocabulary to use of vocabulary that approximates that of a native speaker. The results for pronunciation skills indicated that less than 15 per cent of the children in question had pronunciation that requires concentrated listening on the part of the interlocutor. The major-

Language proficiency and age-appropriate methodology

113

Figure 4. Summary of scores concerning BICS-skills as percentages

ity ranged from being always intelligible although with some use of inappropriate intonation patterns to approximating the pronunciation of a native speaker. The range of scores was greatest concerning grammar; some children were graded as two or two to three. This may indicate that grammar is a more di‰cult part of language proficiency. However, I am not surprised at this drop in level since the children in question acquire English naturalistically and there is little or no focus on form. Weaker grammar skills have always been part of the negative side of naturalistic language learning and it is likely that this would improve when the children have English at school. It is nevertheless reassuring to see that so many children have been given a four or above score for this skill, which corresponds to occasional grammatical mistakes or use of grammar that approximates a native speaker. The range of results is from three to five concerning general cognitive skills, so approximately 81 per cent of the children were rated as being able to use English imaginatively and creatively (or approaching this level), and also in context-reduced situations. The results of the synthesis task are perhaps the most interesting, since the assessor used the greatest

114

Joanne Jalkanen

Figure 5. Summary of scores concerning CALP-skills as percentages

range of scores. This was the only skill that generated a grade one. This may be the result of nerves on the part of the child or simply that he did not understand what was required. Despite this, the scores are again generally high, with the majority of the children being able to synthesise the pictures into the whole concept and to name the concept in English or their mother tongue. The assessment results for coping in context-reduced situations appear somewhat varied. It is, nevertheless, encouraging to see over half of the children scored four to five on this task, which I felt was a demanding one. This shows that the majority of the children in question are able to cope well with English in context-reduced situations and suggests that they are well on the way to developing the skills they will need to be able to cope in the English-medium programme at school. I felt that the CALP scores were surprisingly high, as I was expecting more of a drop, which would have been more in line with the lag between BICS and CALP purported by Skutnabb-Kangas (1981: 111– 114). However, its absence here may be explained by the fact that these children are not from a linguistic minority group, but are from the majority Finnish-speaking population. They receive overt support for

Language proficiency and age-appropriate methodology

115

their mother tongue, also within the immersion programme. Therefore, it may be assumed that they have developed their cognitive academic skills in their mother tongue and have learned to apply those skills. This would suggest that they have opened or are in the process of opening up the channel from English to their underlying common proficiency.

5.

Conclusions

So, is younger really better? Although the above case would tend to lend credence to the CPH, I would suggest that it is mainly because the early total immersion programme described above is an appropriate methodology for young language acquirers. Its greatest strength is also its weakness; the naturalistic learning situation does not formally take metalinguistic aspects of language into consideration. Nor can it, given the young age of the learners in question. This is reflected in the results of this study. The results show that the majority of children who were assessed have a fairly good to excellent proficiency in English. They had an excellent command of English when using it in concrete situations. There was a slight lag for some concerning academic skills, but this is held as a natural phenomenon in second language acquisition and concerned only a small number of children. I would have expected the grammar results to be lower than they were. However, it should be borne in mind that the grammar score was given on a general impression of the grammatical correctness of the English used by the children during the interview. Their metalinguistic knowledge of grammatical rules and concepts was not tested. In any case, the grammar scores reflected that some children’s proficiency was less developed in this area. The only discrete-point task, which is more metalinguistic in nature as it tested knowledge of specific vocabulary, generated lower scores. Many of the children were able to explain what the item was or did, but did not know the arbitrary sign in English. This supports the initial discussion about the type of language learners young children are, and what is an age-appropriate methodology for them. Finnish people realise that very few non-Finns will be motivated to learn Finnish, and that they must learn foreign languages, especially English. For this reason, Finns have become interested in their children

116

Joanne Jalkanen

learning foreign languages at an early age through immersion or other bilingual programmes. Parents often opt to send their children to a language kindergarten because of the implications of the Critical Period Hypothesis. Whether or not one chooses to believe in a critical period for or predisposition to language learning, it must still be remembered that even for a young child, learning or acquiring a second language requires time and e¤ort, even when the child is motivated and receives adequate input (Arnberg 1987: 75). I believe that the early total immersion programme as it has been implemented in the above case is an appropriate methodology for young acquirers of English. A marked improvement in the children’s general level of English has been noted since the change from the half-day immersion system to the full-day system. This has been correlated by feedback from Rajala School. Conversely, however, some children do not settle or do well in the immersion programme. Although these children are the exception rather than the rule, it must, nevertheless, be borne in mind that immersion is not a panacea for teaching languages to young children. Finally, KATE was developed to aid teachers in understanding the level of proficiency attained by young learners in English. When we use it, we must ask ourselves what we hope to achieve by doing so. It is my hope that this assessment tool will always be used with the child’s best interests in mind. It is an assessment tool that has been developed for a specific language learning context. It should not be used to test any group of young learners without some initial consideration and adaptation.

References Alasuutari, Maarit 2003 Kuka lasta kasvattaa? Vanhemmuuden ja yhteiskunnallisen kasvatuksen suhde vanhempien puheessa [Who raises the child? The relationship between parenthood and societal upbringing from the parents’ point of view]. Helsinki: Gaudeamus. Arnberg, Lenore 1987 Raising Children Bilingually: The Pre-school Years. Clevedon, Avon: Multilingual Matters. Baker, Colin 1996 Foundations of Bilingual Education and Bilingualism. 2d ed. Clevedon, Avon: Multilingual Matters.

Language proficiency and age-appropriate methodology

117

Birdsong, David 1999 Second Language Acquisition and the Critical Period Hypothesis. London: Lawrence Erlbaum Associates Inc. Cummins, Jim 2003 BICS and CALP. http://www.iteachilearn.com/cummins/bicscalp. html (viewed 13.03.2005). Donaldson, Margaret 1978 Children’s Minds. London: Fontana Press. Ellis, Rod 1997 The Study of Second Language Acquisition. Oxford: Oxford University Press. Helle, Tuija 1985 Bilingual Education: A Study of the French Immersion Program in Canada Considering the Possibilities of Adaptation to Finnish Schools. In: Ha˚kan Ringbom (ed.), Foreign Language Learning and Bilingualism, 88–189. Publications of the Research Institute of the ˚ bo Akademi Foundation. A Johnson, Robert Keith and Merril Swain 1997 Immersion Education: International Perspectives. Cambridge: Cambridge University Press. Skutnabb-Kangas, Tove 1981 Bilingualism or Not. The Education of Minorities. Clevedon, Avon: Multilingual Matters. Tough, Joan 1977 Talking and Learning. London: Schools Councils Publications. Yule, George 1997 The Study of Language. Cambridge: Cambridge University Press.

6.

‘‘I can you help?’’ Assessing speaking skills and interaction strategies of young learners

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller 1.

Educational policy regarding foreign language instruction in Switzerland

In 2004 the Swiss conference of cantonal educational ministers (EDK) issued a new series of guidelines for foreign language (FL) instruction throughout Switzerland (EDK 2004), based on the advice of a group of experts (see EDK 1998). One of the recommended innovations was to introduce instruction of a second FL in primary schools, which comprises grades 1–6 in Switzerland. In addition, a more significant role is to be given to the local language of instruction (particularly the standard language in the German-speaking cantons), as well as the mother tongue of immigrant children. This innovation is in line with the language policy of the European Union, whose professed aim is to foster functional plurilingualism in future European citizens in order to ensure mobility and cohesion in an increasingly globalized and politically, economically and socially interconnected Europe. The implementation of this policy results in an intensification and preponement of FL instruction in primary schools. While one of the foreign languages taught at primary schools should be a national language, the other should be either English or another national language. This specification aims at taking into account both the political and cultural significance of the four national languages (German, French, Italian and Rhaeto-Romanic) on a national level, as well as the increasing importance of English as a lingua franca on an international level. As the linguistic situation is diverse in di¤erent cantons of Switzerland, the decision as to which language to start with is made by the cantonal authorities. This has led to the introduction of di¤erent models in di¤erent parts of Switzerland. In the German-speaking cantons of Central Switzerland, Obwalden, Zug, Schwyz and Lucerne, in which the present study is conducted, the so-called ‘‘3/5 model’’ has been

120

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

established with the primary school children learning English (L2) from 3rd grade onwards and French (L3) from 5th grade onwards. Hence, for the majority of the children who grow up monolingual speaking (Swiss) German, English constitutes the first and French the second foreign language learnt. The di¤erent cantons involved in this study di¤er both in terms of the amount of lessons allotted for the teaching of FL as well as in terms of teaching materials used. In Obwalden, Zug, and Lucerne English instruction starts with three lessons per week for 3rd graders and the pupils work with Young World (Klett-Verlag). In Schwyz 3rd graders have two lessons of English per week and work with Here Comes Super Bus (Macmillan Publishers). A content analysis of the di¤erent course books warrants the conclusion that in terms of its methodological approach English instruction in 3rd grade is still characterized as subject teaching, even though some indications of Content and Language Integrated Learning (CLIL) can be found (see Weskamp 2003: 39). The conceptual innovation of FL teaching in primary schools consists mainly in the explicitness and commitment with which teaching is oriented towards a communicative approach and towards the goal of functional plurilingualism, on the one hand, and in the combination of linguistic, intercultural, and instrumental (communication and learning strategies) educational goals, on the other hand (Bildungsplanung Zentralschweiz 2004: 3–22; Werlen 2006: 11). Erika Werlen characterizes foreign language competence of primary school children as ‘‘[d]ie Haltung, Sprachen lernen zu wollen und sich in der Zielsprache versta¨ndigen zu wollen, und die interaktive Kompetenz, sich in der Zielsprache versta¨ndlich zu machen’’ [the willingness to learn languages and to communicate in the target language and the interactive competence of making oneself understood in the target language] (2006: 11). This idea is reflected in the teaching principles of FL instruction in primary schools in Switzerland insofar as the development of interaction and learning strategies, which are considered central for the build-up of communicative competence, are considered more important than grammatical and phonetic accuracy (Haenni Hoti and Werlen 2007: 141).

2.

Overview of the study

The present study is part of a national research program investigating language diversity and language competence in Switzerland. It is a

Assessing speaking skills and interaction strategies of young learners

121

three-year longitudinal study tracking the development of 1,008 primary school students from three di¤erent cantons from 3rd grade to 5th grade. In the first two years the main focus is on describing the students’ performance in English in three skills: listening, reading and oral interaction. In the third year the main interest will be the possible e¤ects of English language skills on the acquisition of French (L3). Quantitative instruments are used to assess children’s listening, reading and speaking skills in English and in French, as well as their reading skills in German. Additionally, a student and a teacher questionnaire are administered. The student questionnaire assesses the following variables: language learning motivation, achievement-related self-concept, use of cognitive, meta-cognitive and social learning strategies, attitudes towards Englishspeaking countries and people, and feelings of being over- or underchallenged. (The questionnaires and scales can be made available by the authors upon request.) This chapter presents results of the first year of the study (3rd grade). The focus is on children’s oral interaction skills as demonstrated at the end of 3rd grade after one year of English instruction. Results regarding listening and reading comprehension in English will not be featured in this chapter. The following three questions will be addressed: 1. How well can the children communicate in English after one school year of English instruction? 2. Which interaction strategies do they use to overcome communication di‰culties? 3. On which factors do children’s oral interaction skills depend? Oral interaction skill is conceptualized as the ability of functionally communicating with an interlocutor, making oneself understood in a context-adequate manner rather than as grammatical correctness and phonological accuracy (Werlen 2006: 11).

2.1.

Sample and methods

The study consists of a sample of 50 school classes in four di¤erent cantons (n ¼ 1008), including a test group (with English and French instruction) of 30 school classes and a control group (with French instruction only) of 20 school classes. Given the time-consuming nature of conducting and analysing the speaking test, four children (of the test group) were selected for the speaking test in each class (n ¼ 120). Ten

122

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

children had to be excluded from the analyses due to missing values. In order to ensure some variation in terms of the students’ performance, the English teachers were asked to evaluate the speaking skills of their students. In each class one child of low, one child of high and two children of medium ability as classified by their teacher were selected for the speaking test. Fifty percent were girls and fifty percent boys. The average age of the children at the time of the first testing (3rd grade) was 9;5. Thirty percent of the children are bi- or multilingual due to their immigration background. The growing tendency of o¤ering foreign language instruction in primary schools all over Europe has raised the question as to how to assess young learners’ foreign language skills e¤ectively (Zangl 2000: 250). The issue of e¤ective and age-appropriate assessment is of crucial importance for both researchers and teachers. Language assessment of young children no doubt has to take their special needs into account (Hasselgreen 2005: 337–338). Unfortunately, there is not much literature on how language competence of primary school children can be assessed (Johnstone 2000: 125), nor is there much of a tradition of testing established in primary schools (Hasselgreen 2000: 262). There is, consequently, a dearth of appropriate and validated material for assessing young learners’ foreign language skills, in particular their speaking skills. The construction of our oral interaction tasks was inspired by a variety of studies conducted with young learners (Bu¨eler et al. 2001; Hasselgreen 2000; Schaer and Bader 2003; White and Turner 2005), as well as on material developed for and o¤ered by national and international testing institutions (OCR 2006; University of Cambridge ESOL Examinations 2003, 2004). In order to ensure children’s familiarity with the topics of the tasks a content analysis of the set course books was conducted and the topics selected for the di¤erent tasks were those that all children could be expected to have come across in the classroom. An attempt was also made to develop tasks that can be said to map onto the study aims formulated in the English curriculum of Central Switzerland based on the Common European Framework of Reference (CEFR, Council of Europe 2001). The oral interaction test lasted about 25 minutes and involved an interviewer and two school children. It consisted of three parts: in part one the child was requested to answer personal questions (name, age, place of origin, favourite food, etc.). In part two the child was asked to describe a picture and subsequently identify di¤erences between two pic-

Assessing speaking skills and interaction strategies of young learners

123

tures. In parts one and two each child interacted with the interviewer singly. Part three was a paired speaking task in which the two children interacted with each other and carried out a role-play (shopping in a grocery or eating in a restaurant). All the tests were video-recorded. Interacting in a foreign language is a multi-facetted activity and successful interaction requires a variety of skills, such as mastery of the linguistic code (vocabulary, syntax, phonology), strategic skills and social skills. An assessment of children’s interaction skills, therefore, necessarily has to take di¤erent dimensions into account. We analyzed the videotapes according to the following four dimensions: 1) 2) 3) 4)

task fulfillment interaction strategies complexity of utterances vocabulary range.

As far as task fulfillment is concerned, criteria were formulated specifying under what conditions a task is fulfilled, partially fulfilled or not fulfilled. While this method permits a quantification and statistical analysis of the results of the speaking test for a relatively large sample, it does not in itself do justice to the complex nature of spoken interaction. For this reason we also analysed di¤erent interaction strategies used by the children to overcome communication di‰culties, the complexity of their utterances as well as the range of their vocabulary. As a next step, the association between English speaking skills and individual explanatory variables was studied using multiple regression as a statistical method. A significance level of 5% was used in all statistical tests. The results of these analyses will be outlined in the following sections.

3. 3.1.

Results Task fulfillment

This section will outline to what extent the 3rd graders were able to fulfill the tasks set in the oral interaction test and by implication to what extent they can be said to have attained the study aims of the curriculum. The tasks will be assigned to di¤erent study aims and the results will be illustrated with examples (I ¼ Interviewer; S ¼ Student):

124

3.1.1.

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Study aims: ‘‘can use simple greetings and farewells’’; ‘‘can introduce themselves and answer questions about their person’’ (A1.1 Level)

In the first part of the test the children were required to answer a series of questions about themselves. The task was considered fulfilled if the child could provide a comprehensible, situation-adequate answer without requiring any assistance from the interviewer, partially fulfilled if he or she required assistance and not fulfilled if he or she could not provide a satisfactory answer despite assistance from the interviewer. These specifications reflect our conceptualization of learning progress as becoming increasingly independent of contextual and situational cues as well as of direct assistance by the interlocutor (Werlen 2006: 13). Ninety-three percent of the children could return the interviewer’s greeting in comprehensible English (S: Hello.) and all the children (100%) were able to say their name without requiring assistance from the interviewer. Furthermore, the children were asked about their age (I: How old are you?), their place of origin (I: Where are you from?), and their address (I: What’s your address?): Forty-three percent of the children could say how old they are without needing help from the interviewer (S: Nine.). Another 46% needed help from the interviewer in order to be able to answer the question and have, consequently, partially fulfilled the task. They were able to say how old they were after the interviewer had modelled the answer for them (I: I’m thirty-five years old. And you? S: Nine.). About every tenth child (11%) was unable to provide an answer in English despite assistance from the interviewer. That is to say he or she either did not understand the question or answered in German. Fifty-nine percent of the children could say where they come from without requiring assistance (S: I’m from Switzerland.) and 27% could do so with help from the interviewer (I: I’m from Lucerne. And you? S: Er, I am from, er, Sarnen.). The remaining 14% were unable to (partially) fulfill the task. The most demanding task for the children was giving their address, which only 18% (short of every 5th child) could do without help. This is due to the fact that many children misunderstood the question as ‘‘what’s your dress?’’ Modelling of the answer by the interviewer, however, eliminated this misunderstanding and 79% were subsequently able to give a context-adequate answer, thereby partially fulfilling the task (I: My address is Zentralstrasse number ten in Lucerne. S: Er, my address ist Bergstrasse elf in L.). Only 3% of the children did not fulfill the task.

Assessing speaking skills and interaction strategies of young learners

125

These results illustrate that, as might be expected, providing further personal details is more demanding than just greeting and giving their name. Assuming that the above tasks at least partially assess the study aims ‘‘can use simple greetings and farewells’’ and ‘‘can introduce themselves and answer questions about their person’’, which are to be achieved by the end of 4th grade, we can conclude that they have been attained by a portion of the children by the end of 3rd grade. Apart from greeting and giving their name, this portion lies below the benchmark of 65%, however, so that only the first of these study aims ‘‘can use simple greetings and farewells’’ can be considered attained, while the second can be considered partially attained by the end of 3rd grade (Bildungsplanung Zentralschweiz 2004: 12). These aims are part of Level A1.1 in the curriculum, based on the competence levels which are described in the CEFR, (Council of Europe 2001). The curriculum specifies that in the 3rd and 4th grade the children should attain the level A1.1 as far as their oral interaction skills are concerned. In this study a learning aim of the curriculum is interpreted as having been attained when the assigned tasks in the interview have been fulfilled by 65% of the children; this benchmark indicates that the majority of the children are able to complete the task. In order to make more reliable assertions regarding the attainment of study aims, however, a larger sample and a greater number of tasks would be required. Furthermore, no statements can be made as to the number of children who have attained a specific level in accordance with the CEFR because the exact level of the task used for testing cannot be examined. 3.1.2. Study aim: ‘‘can answer simple questions dealing with familiar topics’’ (A1.1 Level) The topics ‘food and drink’ as well as ‘time’ were part of the English curriculum in the 3rd grade and could, consequently, be assumed to constitute familiar topics for the children. Almost half of the children (48%) could name their favourite food (I: What do you like to eat? S: Pizza, spaghetti and lasagne.). Thirty-five percent needed verbal or nonverbal (gesture and mimics) help by the interviewer to do so. The next question by the interviewer ‘‘What do you like to drink?’’ seems to have been predictable for the children, as 87% could provide a direct, satisfactory answer. Next the children were asked at what time they go to bed and get up. More than four out of five children (82%) could say when

126

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

they go to bed without needing help (S: Er a quarter past nine.) and no less than 94% could say what time they get up without help (I: What time do you get up in the morning? S: Er seven. I: At seven? S: Seven o’clock.). It seems, therefore, that the topic of time and the associated vocabulary (numbers) is well mastered by the 3rd graders. The study aim ‘‘can answer simple questions dealing with familiar topics’’, too, seems to be attained by the end of 3rd grade (Bildungsplanung Zentralschweiz 2004: 12). 3.1.3.

Study aim: ‘‘can describe familiar things, animals and people’’ (A1.1 Level)

In the second part of the test the children were asked to describe a picture and subsequently identify four di¤erences between two pictures. The identification of a di¤erence required the correct naming of things or states in both pictures. If a child was able to name a thing or state in one picture but not in the other, he or she was considered to have identified 0.5 di¤erences. The task was fulfilled if the children were able to describe at least three di¤erences, partially fulfilled if they were able to describe at least 1.5 di¤erences and not fulfilled if they could describe less than 1.5 di¤erences. There were two test versions (A and B) that turned out to be unequal in terms of task di‰culty. In the first version the presented pictures featured a grocery. Seventy-two percent of the children fulfilled the task and were able to describe 3–4 di¤erences between the two pictures (S: And here has rain and here not.). The second version featured two pictures of a beach scene. In this setting only 42% of the children fulfilled the task (S: Here look the dog on the left side und here look the dog on the right side.). As this is a longitudinal study, two test versions (A and B) were drawn up and presented to each half of the children. The two versions will be exchanged in conjunction with the second interview at the end of 4th grade. Using the Rasch analysis as an evaluation method, a scale will be computed, enabling us to make statements about the di‰culty of the tasks and the children’s skills (Rasch 1980). While we assume that the above-mentioned task is qualified to measure the study aim ‘‘can describe familiar things, animals and people’’, we also need to point out that the results seem to be heavily dependent on the picture presented and the associated vocabulary. Given that all respondents were able to describe things, animals, persons and to some extent even activities presented in the pictures, the above study aim can

Assessing speaking skills and interaction strategies of young learners

127

be considered at least partially fulfilled after one year of English instruction, bearing in mind the considerable di¤erences regarding the complexity of respondents’ utterances. 3.1.4. Study aims: ‘‘can name food and drinks to buy or order them’’ and ‘‘can say what they like to eat and drink and order something to eat or drink’’ (A1.2 Level) The third part of the oral interaction test consisted of a role-play which the children carried out spontaneously in pairs (no preparation time was given). Again there were two versions: in Version A the children were supposed to go shopping (n ¼ 53) and in Version B they were to eat in a restaurant (n ¼ 57). In the ‘shopping’ version one child acted the role of the shopkeeper and the other the role of the customer. The task was considered fulfilled if all the key elements of the interaction were present. We defined the following four elements as key elements in the given situation: (1) greeting, (2) ordering/bringing all the five items listed on the shopping list, (3) paying and (4) saying goodbye. The children’s utterances had to be comprehensible for an English-speaking person in order to contribute to the fulfillment of the task. About one third (32%) of the children successfully demonstrated all four key elements and, accordingly, fulfilled the task, as illustrated in the example below: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper: Customer: Shopkeeper:

Hello. Hello. Er, err, errr (thinking), a red shirt, a book. (brings T-shirt) Das (‘this’). Thanks. (brings book) Shoes. (brings shoes) Two bars of chocolate. (brings two bars of chocolate) And a bot of milk. (brings milk) Thanks. Thirty, er, thirty-three money. (pays) Thank you. (starts packing) I can you help?

128

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Customer: No, thanks. Shopkeeper: Goodbye. Customer: Goodbye, thanks. Another 49% of the children partially fulfilled the task of going shopping. This means that they successfully accomplished the core of the interaction, namely the exchange of the goods and payment (elements two and three), but did not say hello or goodbye. About a fifth of the children (19%) were unable to accomplish the core of the interaction: they did not order or bring all the items on the shopping list or other items instead or they did not pay. The role-play ‘in the restaurant’ turned out to be more challenging than the ‘shopping’ version which might be due to the fact that this situation is less familiar for children at this age. One child acted the role of the waiter/waitress and the other the role of the customer. Parallel to the other version four key elements were necessary for the interaction to be considered entirely successful and the task fulfilled: (1) greeting, (2) ordering/bringing all the five items on the menu, (3) paying and (4) saying goodbye. Four percent of the children fulfilled the task, incorporating all four key elements in their interaction. The example below is an illustration of a role-play. Customer: Waiter: Customer: Waiter:

Customer: Waiter: Customer: Waiter: Customer: Waiter: Customer: Waiter:

Hello. Hello. I will a salad with bread, a cuk, and spaghetti and ice-cream. Here the salad (brings salad) and the bread (brings bread), cola (brings coke), ice-cream (brings ice-cream) and spaghetti? Yes. (brings spaghetti) (eats) (thinking) Twenty-eight, please. (The children were not expected to calculate correctly.) (hands the waiter 18 francs) Twenty-eight. Yes. (stands up and leaves) Bye. Bye.

More than a third (39%) partially fulfilled the task. They ordered/ brought all the items on the menu and paid, but did not greet or say

Assessing speaking skills and interaction strategies of young learners

129

goodbye. The other children did not fulfill the task because they did not successfully accomplish key elements two and three. The results suggest that the task di‰culty of the role-plays depends on the contrived situation and that it clearly exceeds level A1.1, which is the level targeted in 3rd and 4th grade. The results also suggest that the majority of the children have already partially attained the study aims: ‘‘can name food and drinks to buy or order them’’ and ‘‘can say what they like to eat and drink and order something to eat or drink’’. 3.1.5. Study aim: ‘‘can use everyday words like thank you or sorry’’ (A1.1 Level) In order to assess this study aim in a relatively authentic situation, all the children were given a bar of chocolate and a ruler at the end of the test. Eighty-five percent of the children thanked the interviewer in English (S: Thank you.), which can be cautiously interpreted as attainment of this study aim.

3.2.

Interaction strategies

Interacting in a foreign language often necessitates the use of interaction strategies to overcome communication di‰culties arising from lack of understanding or lack of target language (TL) resources to express intended meanings. It is important to point out that the development of strategic competence is also specified as a study aim in the curriculum (Bildungsplanung Zentralschweiz 2004: 12). Therefore, the interview and role-play parts of the oral test were additionally examined with respect to the interaction strategies employed by the 3rd graders to overcome communication di‰culties (n ¼ 108). In an interaction with an English-speaking person, strategies in English (or mostly in English) can be expected to be most e¤ective (e.g., saying in English that you do not understand or that you do not have the linguistic TL resources for answering available). This type of strategy was rarely used by the 3rd graders, however. Fourteen percent of the children made use of this type of strategy once (I: What’s your address? S: I cannot answer.). Only one child used this type of strategy twice and one child three times. The vast majority of the children (84%) never used this strategy either because they did not require it or because they preferred another type of strategy (see Figure 1).

130

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Figure 1. Frequency of use of di¤erent interaction strategies

Another way of signalling non-understanding or lack of target language resources is employing nonverbal strategies (e.g., shrugging shoulders). This type of strategy was also relatively rarely used. The majority of the children (70%) never applied it. Just under one fifth (19%) of the children employed gesture or mimics once and 9% did so two or three times. One child made use of non-verbal strategies as often as five times (see Figure 1). The 3rd graders more often used strategies in German (or Swiss German) to indicate that they did not understand or that they could not express what they wanted to say (I: What do you like to eat? S: (thinking) To eat. Was heisst das?). About every fifth child (22%) adopted this strategy once and another fifth (22%) twice or three times. Eight percent of the respondents used this strategy four to seven times. Nevertheless, almost half of the children (47%) never used it, either because it was not necessary or because they found some other way to get help (see Figure 1). It needs to be pointed out, however, that the interviewers were native speakers of German and that the children were aware of this. We do not know if the children would have used the same types of strategies with a person not understanding German. In any case, resorting to the use of their mother tongue (German) is better suited for an upkeep of communication than avoidant behaviour (silence, no use of strategies).

Assessing speaking skills and interaction strategies of young learners

131

There are examples of interlanguage (Selinker 1992) which can contain elements of both the students’ first langauge and the target language. The use of code-switching (English and German) can be considered a strategy of overcoming (temporary) shortages in linguistic resources. The following examples illustrate such compensation strategies: Interview: I: What do you like to eat? S: My eat ist, er, Ha¨rdo¨pfelstock (‘mashed potatoes’). Picture description: S: Er, here is a boy who, er, trommels (‘is playing the drums’). I: Mhm. With the drums. And here? S: Is mit the guitar. Code-switching belongs to the types of strategy more commonly employed by the pupils. Almost a quarter of the children made use of this type of strategy once (24%) and another third (33%) twice or three times. Nine percent of the respondents produced four to seven codeswitches. However, one third of the children (33%) did not make use of code-switching (see Figure 1). Slightly more than a third (34%) of the 3rd graders never used entirely German utterances in order to compensate for lack of vocabulary. Nevertheless, the use of entirely German utterances was one of the more frequently used strategy types. Nineteen percent used this type of strategy once and 29% twice or three times. Fifteen percent used it four to six times. One child produced eight, eleven, fifteen, and nineteen German utterances, respectively, in order to be able to express what they wanted to say (Figure 1). These latter interviewees used mostly German in both the interview and the role-play and produced relatively few utterances in English. 3.3.

Utterance complexity

In addition to task fulfillment and use of interaction strategies, the oral tests were analysed with respect to the complexity of the pupils’ utterances. All utterances in comprehensible English were counted. German utterances were excluded, as were code-switched utterances that did not exceed eight words. The utterances were classified as 1- to 9-word and longer utterances. For lack of space only a selection of results concerning the complexity of students’ utterances will be presented here.

132

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Figure 2. Distribution of 1-word utterances

After one year of English instruction all children mostly produced 1-word utterances and in many instances these were perfectly adequate. The number of 1-word utterances produced by a single child ranges from three to thirty-one (I: And where are you from? S: Portugal.) (Figure 2). Twelve percent of the pupils produced 16 and 18 1-word utterances, respectively. As might be expected, the number of utterances produced by the children decreases as complexity increases (see Figures 2 to 5). Across the entire sample the children produced between 0 and 12 3-word utterances. Seven percent of the children did not use any 3-word utterances, while one child used 12 such utterances (I: What is the boy doing? S: He playing guitar.) (see Figure 3). Five-word utterances were produced by the children between never and eight times (child describes a picture: I can see a lake.) (see Figure 4). More than 50% of the pupils produced one or more 5-word utterances in their test. One child used as many as eight 5-word utterances, which can be considered quite complex. Nine-word or longer utterances constitute quite an achievement for 3rd graders after just one year of English instruction, considering the limited amount of exposure to the language, as they have had 2–3

Assessing speaking skills and interaction strategies of young learners

Figure 3. Distribution of 3-word utterances

Figure 4. Distribution of 5-word utterances

133

134

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Figure 5. Distribution of 9( þ )-word utterances

lessons a week. It is not surprising then that such utterances were accordingly rare (child describes di¤erences between two pictures: Here look the dog on the left side und here look the dog on the right side. The following rule was applied in coding the utterances: if an 8-word or longer utterance contained a word in German that was not as semantically important as to impede understanding (e.g. connector, article) the utterance was counted) (see Figure 5). Slightly more than three quarters (77%) of the children produced no utterance of this length, while around every fifth child (21%) was able to produce a 9-word or longer utterance. One child used three and another four utterances of such complexity. 3.4.

Vocabulary range

Successful oral interaction in a foreign language and concomitant feelings of success require, among other things, a basic vocabulary in the target language. The second part of the oral test (picture description and identification of di¤erences between two pictures) was additionally evaluated with respect to the vocabulary range exhibited by the pupils. During this sequence of about four minutes we counted the number of di¤erent, comprehensible English words produced by the children. As Figure 6 clearly shows, considerable di¤erences characterize 3rd graders’

Assessing speaking skills and interaction strategies of young learners

135

Figure 6. Distribution of vocabulary range

range of vocabulary: the number of comprehensible English words produced by the children ranges between 8 and 43. The highest ratio of children exhibited a vocabulary range of 19 or 23 words (8% of the children, respectively). The following examples serve to illustrate the repertoire of target language words participants exhibited: Example 1: a, baby, clock, pear, orange, banana, and, mom, boy, dog, the, it, is, black, red, this, girl, here, apple ¼ 19 words Example 2: I, can, see, a, dog, he, is, out, the, water, and, kite, here, are, two, girls, man, yellow, orange, hat, boat, trees, no, yes, playing, with, his, guitar, rainbow, not, butterfly, have, this, fish ¼ 34 words

Given the large sample it was not possible to conduct a fine-grained linguistic analysis of the students’ language output. A grammatical categorization of the vocabulary produced by all the children, however, revealed that, apart from nouns which constituted the word category most commonly used, the children also mastered verbs, conjunctions, articles, pronouns, prepositions, adjective and adverbs for certain domains.

136

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

3.5.

Factors explaining the children’s oral interaction skills

Having presented the descriptive results regarding the oral interaction skills of the 3rd graders, we will now demonstrate which factors could be shown to contribute to the explanation of these speaking skills. In a first step, a regression analysis was conducted to ascertain if significant e¤ects can be found on the class level. A comparison of the test scores revealed that the variability within school classes was as high as the variability across school classes. School class has neither a significant fixed e¤ect (F ¼ 0.8663; p ¼ 0.66) nor a significant random e¤ect, the intraclass coe‰cient (ICC) being 0. The fact that no class e¤ects could be found might be due to the selective sampling procedure and the small number of children tested per class. The association between English-speaking skills and individual explanatory variables was studied next by using multiple regression. Variable selection was based on the Akaike information criterion (AIC). The stepwise procedure led to a model with five explanatory variables. No interactions were significant. Residual plots showed a few outliers. As they did not have a particularly large influence on the estimates they were not excluded. The first variable, reading skills in German (L1), is the most important predictor for speaking skills in English (L2): The better the students performed in the German reading test, the higher their score in the English speaking test. Another highly significant variable is the use of learning strategies: the more often the children claimed in the student questionnaire to use (meta-)cognitive learning strategies and techniques for self-regulated learning, such as writing down words they want to learn, speaking English with themselves, deciding what to learn, controlling tasks and using a dictionary as well as social learning strategies (asking their teacher or classmates for support, explaining something to their classmates), the higher their performance in the English speaking test (see Table 1). Other important factors are feelings of being overburdened, attitudes towards English-speaking countries and people, and age: the less overburdened children feel during English classes and the less afraid they are of making mistakes, the better their English speaking skills. The result of the English oral interaction test also improved with more positive attitudes towards English-speaking countries (USA, England) and people. Finally, children who were older than the average (10 years or older) turned out to score lower in the English speaking test than their

Assessing speaking skills and interaction strategies of young learners

137

Table 1. Variables explaining achievement in oral skills in English Coe‰cient

Std Error

1.61

0.79

2.41

0.74

0.26

0.17

Use of (meta-)cognitive and social learning strategies

0.62

0.24

Attitudes towards English-speaking countries/people

0.26

0.17

0.40

0.22

(Intercept) Reading skills in German (L1) Feelings of being overburdened and fear of making mistakes

Age: b 10 years old

p

**

*



‘‘**’’ 0.001 < p < 0.01, ‘‘*’’ 0.01 < p < 0.05, ‘‘ ’’ 0.05 < p < .01 (n ¼ 110; R2 ¼ 24%; p < .0001)

younger classmates. It might be that these children were later enrolled into the school system or had to repeat a school year (see Table 1). The following variables were eliminated during the variable selection process because they did not significantly improve the model fit: gender, number of books at home, citizenship (Swiss/binational/other nationality), number of spoken languages in the family, parental support while learning English, motivation to learn English, achievement-related selfconcept. Consequently, they can be regarded as unimportant factors with respect to the explanation of the English-speaking skills of the 3rd graders in our sample. Finally, the school class factor was tested again, together with the individual variables in a multilevel model. The class e¤ect remained insignificant.

4.

Conclusion

Overall, the children’s learning progress regarding their oral interaction skills after one year of English instruction can be considered substantial. Naturally, the rate of progress varies considerably across individuals, as is the case in all school subjects. The vast majority of the 3rd graders can use simple greetings and farewells as well as everyday expressions such as ‘‘thank you’’. Furthermore, they are able to answer simple questions dealing with familiar topics such as favourite food and drinks or

138

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

time. A considerable proportion of the children were also able to introduce themselves and answer questions about their person as well as describe familiar things, animals or people without requiring assistance. Hence, the majority of the study aims formulated for the 3rd and 4th grade (A1.1 Level) which could be assigned to certain tasks in the oral test can be considered attained and a minority of these study aims can be considered partially attained. The role-play task is assumed to reflect a study aim on the A1.2 Level which the children are only expected to attain by 5th grade. Nevertheless, depending on the familiarity with the situation, at least one third of the children in our sample was able to accomplish the core elements of the interaction, namely the exchange of goods and payment. The examples of student utterances provided in this chapter demonstrate that high-performing pupils have acquired a basic vocabulary for certain domains and can produce relatively complex utterances (up to 9-word or longer utterances). Learning progress is also evident in lower-performing students although these still rely more heavily on the German language to bridge communication gaps. As far as the use of interaction strategies is concerned, the analyses showed that the children rarely use strategies in English or non-verbal strategies (gestures, mimics) to overcome communication di‰culties. It seems that these kinds of strategies need to be explicitly taught in the classroom. Achievement di¤erences with regard to the students’ oral interaction skills are clearly identifiable after one year of English instruction. These can probably partly be attributed to the di¤erentially developed vocabularies of the children. Apart from this our analysis showed that the following factors contribute to the explanation of English speaking skills of young learners: reading skills in German, feelings of being overburdened and fear of making mistakes, use of (meta-)cognitive and social learning strategies, attitudes towards English-speaking countries/people, and age. It seems, therefore, that previous language knowledge, specific a¤ective and attitudinal factors as well as learning strategies are crucial for the early development of speaking skills in a foreign language. In conclusion the following pedagogical implications can be drawn: dealing with diversity regarding foreign language skills is certainly a challenge for the primary school teachers, demanding high methodological competence. In order to avoid that the children feel frequently overtaxed or under challenged, they need to di¤erentiate their instruction (Haenni Hoti 2007). Even though fears of communicating in English seem to be minor at this early stage, our results show that, if present,

Assessing speaking skills and interaction strategies of young learners

139

they can negatively a¤ect the children’s speaking skills. To prevent such fears, teachers should keep encouraging their students to speak in the target language and point out that mistakes are part of the learning process. After all, it is one of the main objectives of early foreign language learning to ensure that the children’s contact with the first foreign language at school is positive and enjoyable.

References Bildungsplanung Zentralschweiz 2004 Lehrplan Englisch fu¨r das 3.-9. Schuljahr. Luzern: Bildungsplanung Zentralschweiz. Bu¨eler, Xaver, Sto¨ckli Georg, Stebler Rita, and Stotz Daniel 2001 Schulprojekt 21. Lernen fu¨r das 21. Jahrhundert? Externe wissenschaftliche Evaluation. Schlussbericht zuhanden der Bildungsdirektion des Kantons Zu¨rich. [Electronic Version] Retrieved on 19.01.07 from http://www.paed.unizh.ch/pp1/stoeckli/tsprimgraphs/SP21Teil1.pdf; https://www.unizh.ch/paed/pp1/stoeckli/tsprimgraphs/SP21Teil2.pdf Council of Europe 2001 Common European Framework of Reference for Languages: Learning, Teaching, Assessment. Cambridge: Cambridge University Press. EDK (Schweizerische Konferenz der kantonalen Erziehungsdirektoren) 1998 Sprachenkonzept Schweiz: Welche Sprachen sollen die Schu¨lerinnen und Schu¨ler der Schweiz wa¨hrend der obligatorischen Schulzeit lernen? Bericht der Expertengruppe ‘‘Gesamtsprachenkonzept’’. Bern: Schweizerische Konferenz der kantonalen Erziehungsdirektoren. 2004 Sprachenunterricht in der obligatorischen Schule: Strategie der EDK und Arbeitsplan fu¨r die Gesamtschweizerische Koordination. Beschluss der Plenarversammlung der EDK vom 25. Ma¨rz 2004. Bern: Schweizerische Konferenz der kantonalen Erziehungsdirektoren. Retrieved on 08.01.07 from http://www.edk.ch/PDF_Downloads/ Presse/REF_B_31-03-2004_d.pdf Haenni Hoti, Andrea 2007 Leistungsvielfalt als Herausforderung fu¨r den Englischunterricht in der Primarschule. Beitra¨ge zur Lehrerbildung (BzL). Zeitschrift zu Theorie und Praxis der Grundausbildung 25 (2): 205–213. Haenni Hoti, Andrea and Erika Werlen 2007 Englischunterricht (L2) in den Zentralschweizer Primarschulen: Hat er einen positiven oder negativen Einfluss auf das Leseversta¨ndnis der Schu¨lerInnen in Deutsch (L1)? In: Erika Werlen and Ralf Weskamp (eds.), Kommunikative Kompetenz und Mehrsprachigkeit. Diskussionsgrundlagen und unterrichtspraktische Aspekte, 139–160. Baltmannsweiler: Schneider Verlag Hohengehren.

140

Andrea Haenni Hoti, Sybille Heinzmann, and Marianne Mu¨ller

Hasselgreen, Angela 2000 The assessment of the English ability of young learners in Norwegian schools: an innovative approach. Language Testing 17 (2): 261–277. 2005 Assessing the language of young learners. Language Testing 22 (3): 337–354. Johnstone, Richard 2000 Context-sensitive assessment of modern languages in primary (elementary) and early secondary education: Scotland and the European experience. Language Testing 17 (2): 123–143. Oxford, Cambridge and RSA Examinations (OCR) 2006 Asset Languages. External Assessment Sample Tasks. Cambridge: OCR. Retrieved 01.01.2006 from http://www.assetlanguages.org. uk/about/samples.aspx Rasch, Georg W. 1980 Probabilistic Models for Some Intelligence and Attainment Tests (expanded edition). Chicago: University of Chicago Press. Schaer, Ursula and Ursula Bader 2003 Evaluation Englisch an der Primarschule Projekt 012. Appenzell: Schulamt. Retrieved 30.09.2007 from http://www.ai.ch/dl.php/de/ 20031110083112/Evaluationsbericht+07.11.2003.pdf Selinker, Larry 1992 Rediscovering Interlanguage. London: Longman. University of Cambridge ESOL Examinations 2003 Cambridge Young Learners English Tests. Sample Papers. Cambridge: University of Cambridge. 2004 Cambridge Young Learners English Tests. Sample Papers. Cambridge: University of Cambridge. Werlen, Erika 2006 Neuland Grundschul-Fremdsprachenunterricht. Empirische Ergebnisse der Wissenschaftlichen Begleitung, Fremdsprachen in der Grundschule’ in Baden-Wu¨rttemberg. Babylonia 1: 10–16. Weskamp, Ralf 2003 Fremdsprachenunterricht entwickeln. Grundschule – Sekundarstufe I – Gymnasiale Oberstufe. Hannover: Schroedel Verlag. White, Joanna L. and Carolyn, E. Turner 2005 Comparing Children’s Oral Ability in Two ESL Programs. The Canadian Modern Language Review 61 (4): 491–517. Zangl, Renate 2000 Monitoring language skills in Austrian primary (elementary) schools: A case study. Language Testing 17 (2): 250–260.

7.

Input and long-term e¤ects of early learning in a formal setting

Carmen Mun˜oz 1.

Introduction

The consensus view concerning the e¤ects of age on second language (L2) acquisition is that younger starters reach higher ultimate levels of proficiency than older starters. This view has been related to the Critical Period Hypothesis (CPH) formulated by Lenneberg (1967) primarily in reference to first language acquisition but rapidly extended to second language acquisition as well. According to the CPH there is a fixed span of years during which language acquisition is e¤ortless and fully successful but after which it becomes harder and rarely completely successful. This view has been supported empirically by research that has typically compared the ultimate attainment of immigrants with di¤erent ages of arrival in the target language community. The review of studies in this area has revealed two robust findings (Krashen, Long, and Scarcella 1979). First, that older starters have a faster rate of learning in the initial stages, particularly of morphosyntactic aspects. Second, that in spite of their initial slower rate of learning younger starters outperfom older starters in the long term, that is, younger starters have a longterm advantage. The consensus view has been generalized to formal learning situations as well, setting expectations that an early start will invariably lead to success (see for a discussion Mun˜oz 2008). Age-related research in foreign language learning has provided confirmation for the older starters’ faster rate, but it has failed so far to produce confirmation for the younger starters’ long-term advantage (see collections of recent studies in Garcı´a Mayo and Garcı´a Lecumberri 2003 and Mun˜oz 2006a). Attempts at explaning that lack of evidence have claimed that existing studies do not extend long enough to capture long-term results, and have sometimes considered the number of years that would be required to provide formal learners with an amount of input that is equivalent to the input received by learners in a naturalistic learning situation (Singleton 1989).

142

Carmen Mun˜oz

The aim of this chapter is to present a research project that has extended for a comparatively long period of time in order to capture what could arguably be considered long-term results of an early start in a foreign language learning situation and then to suggest an explanation for the findings that takes into account the input received by formal learners and its interplay with learners’ age. It is the contention of this paper that such an explanation may account generally for the most important age-related di¤erences observed between naturalistic and instructed language learning settings. To that end, the following section discusses the role ascribed to input in age-related studies. Next, a study of foreign language learning is presented that examines short-, mid- and long-term results of di¤erent-aged formal learners. The discussion of the findings suggests that considerations of the input received by L2 learners in both qualitative and quantitative terms are essential for explaining long-term results. The final section argues that the role of input is crucial in explaining the finding that age-related e¤ects are mediated by context.

2.

The role of input in age-related studies

Flege (1987) notes that the absence of a search for potential sources of variance in L2 learning other than age may be explained by the commitment of many researchers over the last decades to the notion that the di¤erential success in L2 acquisition can be explained by the Critical Period Hypothesis. One such potential source of variance is the input received by learners. Obviously, the strength of the commitment to the Critical Period Hypothesis in one or other of its various versions (see Singleton 2005, for a discussion) may be explained by the allegiance to the innatist account of language acquisition characteristic of the dominant linguistic theory in the last fifty years. Relatedly, one of the most important arguments put forward by Chomsky in order to argue against behaviourist accounts of language acquisition focused on the insu‰ciency and inadequacy of input to children (Chomsky 1959, 1965). Innate knowledge of universal formal properties was proposed to explain the ‘‘logical’’ problem of language acquisition, namely how children could acquire language successfully in spite of the insu‰ciency of input (and negative feedback) (see e.g., Hornstein and Lightfoot 1981).

Input and long-term e¤ects of early learning in a formal setting

143

In that perspective, a major issue in SLA has been whether the innate system postulated for child language acquisition continues to function in adults. One possibility is that adult learners lose their innate language acquisition ability with increasing age or, in other words, that adults do not have access to Universal Grammar in the same way as children do. Such a solution fits very well with the notion that there exists a biologically determined period during which language acquisition is e¤ortless, namely the Critical Period Hypothesis. Adults’ failure to learn language successfully (understood as reaching native-like command of the language) can then be explained by the fact that they learn the second language after the end of the critical period, initially set by Lenneberg (1967) at puberty. Bley-Vroman (1989), one of the proponents of that view, claims that child language development and adult foreign language learning are fundamentally di¤erent. In his Fundamental Di¤erence Hypothesis, the nature of the di¤erence is internal, linguistic and qualitative. The first characteristic is of interest in this context: it is internal because it is caused by di¤erences in the internal cognitive state of adults versus children, not by some external factor or factors (such as insu‰cient input) (1989: 50). In his discussion of alternative approaches, that is, views that claim fundamental sameness, Bley-Vroman explicitly rejects Krashen’s Input Hypothesis (1982, 1985), which postulates that adults do not get enough input or do not get the right kind. The Input Hypothesis posits a learner-internal ‘‘filter’’ that prevents the input to which the learner is exposed from getting into the language acquisition device. That filter is weak or inexistent in childhood and strengthens or arises around the age of puberty. In this way Krashen (1982) can explain the di¤erence in second language learning success between children and adults while maintaining that no real change in the language acquisition device occurs at puberty. In contrast, in recent theoretical proposals from cognitive linguistics and usage-based models of acquisition there is no ‘‘logical’’ problem because language input is su‰cient evidence for the general learning processes that give rise to its abstract representation and no innate mechanisms (i.e., Universal Grammar) are needed to supplement these processes. Such models claim that we learn language while processing input and producing language during interaction (e.g., Tomasello 2003). One of the most important recent claims is Ellis’s frequency-based account of language learning. Ellis (2002: 179) remarks that ‘‘the role of frequency has largely been ignored in theoretical and applied linguis-

144

Carmen Mun˜oz

tics for the last 40 years or so’’, and claims that there is now enough evidence to reinstate frequency as an ‘‘all-pervasive causal factor’’. Along those lines it has been suggested that one of the underlying reasons for children’s observed advantage in second language acquisition may be the fact that they are more flexible learners than adults in skilled activities (see Tomasello 2003). Turning now from the influence of theoretical positions to the empirical evidence provided by research, we find that a second, more specific, reason underlying the minor role given to input in age-related studies may be the fact that input di¤erences have not generally appeared as determinant in results. Early research looked at qualitative di¤erences in the input available to younger and older learners, and in particular whether input addressed to younger learners was simplified or more facilitative of learning, as it seemed to be the case in L1 acquisition (see e.g., Snow 1986; Steinberg 1993). Hatch (1977) suggested that child L2 learners receive input that is of a di¤erent kind to that received by adult L2 learners: input that is better tuned and linguistically less complex, providing young children with more and clearer samples from which to learn the target language. Further, Hatch (1978) observed that speech directed at child L2 learners in a naturalistic situation was more concretely oriented and context-embedded than speech directed at adults. This greater support could partly explain the language learning superiority of children relative to adults. However, a counter-argument to that plea came from the field of child language acquisition studies where research showed that the features identified in speech addressed to children in typical western communities were not universally observed (see Foster-Cohen 1999). On the other hand, evidence was also found that speech addressed to L2 learners seems to be adjusted to the addressee’s proficiency level (Krashen 1981, 1985), which means that not only children but also adults receive supportive input. Moreover, it has been suggested that lack of interaction opportunities could be as determinant of age-di¤erences as quality of input (Hatch 1983). A study which provides evidence of di¤erences in both quality of the input and interaction opportunities is reported in Snow (1983). In that study (Snow and Hoefnagel-Ho¨hle 1979) oral exchanges in classrooms were recorded in order to collect information on how nonnative speakers (English-L1 learners of Dutch in The Netherlands) were addressed. It was found that younger children received shorter, simpler utterances and that they had more access to private small group interac-

Input and long-term e¤ects of early learning in a formal setting

145

tions. In contrast, the majority of utterances addressed to high school students were not simplified or adjusted because teacher’s talk was mostly teacher-fronted. However, no strong correlation was found between any input measures and the learners’ subsequent rate of acquisition, although some learners were found to be better than others at generating input for themselves, specifically those with good vocabularies and good accents. In fact, older learners have been claimed to be superior in generating input, which could explain their initial advantage in L2 learning, that is to say, their rate advantage. Scarcella and Higa (1982) conducted an experimental study comparing the conversational management devices used by child and adolescent learners both to repair problems in comprehension and to keep the conversation going. Scarcella and Higa concluded that the reason for the older learners’ headstart in the language learning process could be the fact that they were provided with finelytuned input and with opportunities to stretch their linguistic competence in conversation. Other more recent studies have also found that due to their greater interactional skills older learners’ involvement in conversation is more active than that of young children. These superior interactional skills have been shown in the observation of children dealing with comprehension problems (Mun˜oz 2002) and in the examination of children dealing with breakdowns in interaction arising from production problems (Gran˜ena 2006). Moving on now to the issue of the quantity of input received by L2 learners, it has often been noticed that adult learners (and many children) do not attain target levels in spite of prolonged access to input (Long 1990), which constitutes a defining characteristic of the phenomenon known as fossilization (Han and Odlin 2006; Selinker 1972). In contrast, evidence has been found of exceptionally successful L2 learners who have spent only a limited time abroad (see Nikolov 2000). The most frequently debated considerations relate to learners’ length of residence (LOR), that is, the period of time learners have been immersed in the L2 community. LOR is thus the operationalization of amount of input in a naturalistic learning setting and its e¤ects on language acquisition have often been contrasted to age-e¤ects. For example, Snow (1983) observes that the paradox that children learn foreign languages more slowly than adults, yet the earlier one starts to learn a foreign language the better one will speak it, is basically methodological. Namely, the studies showing that older learners are better than

146

Carmen Mun˜oz

younger ones have tested for L2 ability within the first 2 years of learning, while all the studies showing that child learners are superior to adult learners have tested after at least 5 years of first exposure to the L2. Similarly, in the first comprehensive review of studies concerned with the age factor, Krashen, Long, and Scarcella (1979) already remarked that length of residence only seems to have significant e¤ects during an initial period and suggested that it plays little or no role after the first 5–10 years. Along those lines, DeKeyser (2000: 503) argues that a minimum of 10 years should be required in ultimate attainment studies in order to ensure that it is ultimate attainment that is being picked up; a shorter length of time could reveal rate e¤ects instead. In general, LOR e¤ects reported in research have tended to be small or insignificant. The influence of LOR on ultimate attainment was clearly dismissed in Johnson and Newport’s (1989) study, and it has continued to be dismissed in studies examining long-term e¤ects. For example, DeKeyser and Larson-Hall (2005) argue that input plays a very limited role once variation in age of arrival is controlled statistically. Likewise, Long (2007) remarks that after an initial period LOR ceases to correlate with proficiency scores. Recently, however, a new interest has developed in the area of speech learning for the study of input, an interest spurred particularly by work of Flege and his associates. This line of research claims that LOR may be a good index of second language learning on condition that learners receive a substantial amount of native-speaker input. Consequently, the need for both quality and quantity in input for L2 speech acquisition is emphasized: learners need to receive quality input for speech learning – native-like or almost native-like from a phonological point of view-, and to receive substantial amounts of it. For example, Flege and Liu (2001) found that years of residence in the target language country were predictive of success in L2 speech learning only when learners had received substantial amounts of native-speaker input, in contrast not only to L1 input produced by other immigrants from the same linguistic background but also to L2 accented input produced by non-native speakers. In their study subjects were separated by occupation, and LOR was observed to have an e¤ect on speech outcomes only in the case of students, who were likely to receive native-speaker input, in contrast to non-students, whose exposure to native-speaker input was less intense. Flege (2008) draws a similar conclusion from the study of a Polish boy who arrived in the US at the age of 7 and whose language develop-

Input and long-term e¤ects of early learning in a formal setting

147

ment was followed over a 7-year period by Winitz, Gillespie, and Starcev (1995). This learner showed a strong e¤ect of LOR in contrast to other learners of similar age, and eventually became indistinguishable from native speakers. What was di¤erent in this case was that the young immigrant settled with his family in a small, rural town and attended a school that had very few non-English speaking children and did not o¤er ESL classes. As a consequence, the boy received much more native-speaker input than most other children who immigrate in apparently similar circumstances but are enrolled in ESL classes in schools located in a large city with much more contact with many other immigrants to the US. In conclusion, Flege and his associates claim that L2 proficiency increases with LOR only if the L2 speaker is participating in social settings, such as schools, in which they receive a substantial amount of input from native speakers of the L2. From a methodological point of view Flege (2008) argues that previous research has underestimated the predictive power of L2 input among other reasons because input has not been actually measured. In order to ultimately determine whether input is really less important in L2 than in L1 learning, Flege suggests that input should be rigorously measured rather than estimated via self-report. It must be noted that the findings reviewed so far refer to second language learning in a situation of social immersion. When considering foreign language learning, that is, learning that is constrained to the classroom, the operationalization of amount of input is courses or years of instruction. In such situation input appears often as extremely limited in quantity and of variable quality, both in its interactional traits (e.g., addressee-oriented adjustments) and in its phonological features (see Niz´egordcew 2007: 39 ¤. for a review of di¤erences between L2 naturalistic and classroom discourse). This important asymmetry renders it inadequate to rely on findings obtained in a naturalistic setting and underlines the need for empirical evidence drawn from instructed settings (see for a discussion Mun˜oz 2008). In particular, specific evidence is needed of the length of time required for learners to receive similar amounts of input as those separating, in naturalistic learning settings, rate advantage from ultimate attainment advantage. Or, in other words, evidence is needed of the amount of input that leads instructed learners to show long-term age-related benefits. In the next section a study is presented that explored long- term e¤ects of di¤erent starting ages in a foreign language learning setting, following students along primary and secondary education.

148

3.

Carmen Mun˜oz

Empirical evidence of long-term results in a formal learning setting: The BAF Project

The Barcelona Age Factor (BAF) project extended from 1995 to 2004 with the aim of exploring age-related di¤erences at three di¤erent points in time: short-, mid- and, most importantly, long-term (see Mun˜oz 2006a for a collection of studies from the BAF project). To that end data were collected after the same amount of hours of instruction (200, 416, and 726) from four groups of learners with di¤erent ages of initial learning. Those four groups appear in Table 1 identified in terms of their initial age of learning as younger children (8 years), older children (11 years), adolescents (14 years), and adults (older than 18 years). A longitudinal subsample was formed from the sets of participants who could be followed from one collection time to the next, while new participants were added at each data collection in order to have as large a cross-sectional subsample as possible. The three younger groups were school students drawn from 30 state-funded schools, so that teacher variability and socio-economic background factors were cancelled. The group of adults was composed of students at three di¤erent adult language schools either in the public system or attached to a university. All the participants were Catalan-Spanish bilingual learners of English and the total number neared 2,000. A large battery of tests measuring both productive and receptive skills in written and oral modalities was administered to intact classes. Table 1. Learner groups Younger children AO ¼ 8

Older children AO ¼ 11

Adolescents AO ¼ 14

Adults AO ¼ 18 þ

Time 1 200 hours

AT ¼ 10;9 N ¼ 284 OSE ¼ 164

AT ¼ 12;9 N ¼ 286 OSE ¼ 107

AT ¼ 15,9 N ¼ 40 OSE ¼ 21

AT ¼ 28;9 N ¼ 91 OSE ¼ 67

Time 2 416 hours

AT ¼ 12;9 N ¼ 278 OSE ¼ 140

AT ¼ 14;9 N ¼ 240 OSE ¼ 96

AT ¼ 19,1 N ¼ 11 OSE ¼ 4

AT ¼ 39;4 N ¼ 44 OSE ¼ 21

Time 3 726 hours

AT ¼ 16;9 N ¼ 338 OSE ¼ 71

AT ¼ 17;9 N ¼ 296 OSE ¼ 51

_

_

AO ¼ age of onset; AT ¼ age at testing; N ¼ number of subjects; OSE ¼ only school exposure

Input and long-term e¤ects of early learning in a formal setting

149

The battery included a dictation and a cloze test in the three languages (English, Catalan, and Spanish), a listening comprehension test, a grammar test, a written composition, an oral picture-elicited narrative (or description), an oral interview, a phonetic imitation test, a phonetic discrimination test and a role-play. In addition, learners answered a questionnaire that elicited information about their use of the three languages, a¤ective variables and strategies. On the basis of their answers to the questionnaire participants were classified into those who had only had curricular exposure to English and those with additional exposure (inside or outside school) as well. Only the former learners constituted valid groups for the comparisons aiming at examining the role played by initial age of learning. Because amount of exposure was controlled, the group(s) with higher L2 outcomes could be considered to have a faster rate of learning. The first comparison, after 200 hours of instruction, revealed a superior rate of learning on the part of older starters, both when considering longitudinal and cross-sectional sub-samples (see Mun˜oz 2006b). Each age group had better results than the younger group in the di¤erent measures of proficiency, with the only exception of the listening comprehension test and the phonetic imitation test, where di¤erences were not always significant. The second comparison took place when students in the four groups had had 426 hours of instruction in English. Again, older starters outperformed younger starters extending the expected rate advantage beyond the first years of learning (as long as four years in the case of the youngest participants). By the time of the third comparison, after 726 hours of formal study of English, only the two younger groups provided a large enough sample of valid subjects that could be compared (those who began learning English at 14 had already finished secondary education, and in the group of adults there were very few left that after so many years of study had not spent some time in an English-speaking country, which rendered them invalid for the purposes of this research). The older starters still outperformed the younger starters in most measures, although this long-term comparison revealed that di¤erences generally diminished and were no longer significant in the phonetic discrimination test and on some measures of fluency in the written composition as well as in aural reception skills. In sum, the younger starters after 726 hours of instruction distributed in eight years were only slowly catching up in some measures to the older starters who had had the same number of hours

150

Carmen Mun˜oz

of instruction but distributed in seven years. A follow-up of the early starters after 800 hours, that is nine years, revealed that they had not yet caught up with the older starters, with fewer hours of instruction either in writing skills (Nave´s 2006) or in vocabulary development (Miralpeix 2008). A first, most illuminating, general finding is that this final outcome seems to be strongly influenced by learners’ cognitive development. The influence can be seen, first, in the di¤erential rate of learning of the different age groups: the adults, more mature cognitively, showed the most rapid initial acquisition (up until the first data collection), followed by the group of learners who began at age 14, who showed a more rapid initial rate of learning than the younger groups; the group of learners who began at age 11 showed the fastest rate between the first and the second data collection; and finally, the youngest group showed the slowest rate in the first 416 hours, but a rapid increase from the second to the third data collection. The influence of learner’s cognitive maturity can also be seen in the di¤erent patterns followed by the development of the two younger groups’ morphosyntactic and auditory abilities. On the one hand, morphosyntactic learning appeared to boost at around age 12 (Time 1 for learners who started at age 11, and Time 2 for learners who started at age 8), coinciding with the cognitive growth associated with puberty. On the other hand, auditory skills seemed to be a¤ected by amount of exposure rather than by learners’ age. A regression analysis (Mun˜oz 2001) showed that scores in the listening comprehension tests had a stronger influence of exposure than scores in other tests while L1 development, a measure of cognitive development, had no noticeable influence on the listening comprehension test. This di¤erence is also pointed out by findings in other studies, both in regular programmes (i.e., Burstall et al. 1974; Cenoz 2002) and in immersion programmes (i.e. Lapkin et al. 1980), showing that late starters did not have a clear advantage in listening comprehension measures. Thirdly, learners’ cognitive development seems also to have selectively a¤ected the observed progressive reduction of older L2 learners’ lead over younger learners. This progressive reduction could be interpreted to correspond to the phenomenon observed in naturalistic acquisition, when younger starters first catch up to older starters and then outperform them reaching a higher, sometimes native-like, ultimate attainment. Interestingly, this reduction is noticeable when the age difference between the two younger groups of learners that are being com-

Input and long-term e¤ects of early learning in a formal setting

151

pared has been reduced (16–17 vs 17–18) thanks to a di¤erence in intensity of tuition (from two- to one-year gap, approximately), at a point in which the cognitive di¤erence between the two groups of learners has also been reduced, and in areas in which the influence of learners’ cognitive maturity may be said to play a lesser role (i.e. phonetic discrimination, fluency). This brings us to a reflection on whether age appears to a¤ect foreign language learners in the same way as it is known to a¤ect naturalistic language learners; in other words, whether younger learners in this situation will eventually outperform older learners with the same amount of exposure. The answer seems to be negative. Empirical evidence shows that their superior cognitive development gives older learners a headstart in the first stages. It also shows that when di¤erences in cognitive development between older and younger learners disappear or diminish, the younger starters begin to catch up to older starters. This suggests that, in the same conditions of input and instruction, di¤erences will disappear eventually. But there is no reason to expect that younger starters will surpass older starters. In sum, whatever long-term superiority younger starters may have in a naturalistic setting cannot manifest itself in an instructed setting such as the situation described in the present study. Results from the BAF project, therefore, confirm the older learners’ rate advantage and do not provide evidence of the ultimate attainment advantage attributed to young learners by the CPH and related maturational explanations. One reason for the younger learners’ relative lack of success as well as the older learners’ relative success found in this study may lie in the fact that young children favour and are probable favoured by implicit learning (see DeKeyser’s [2000] interpretation of Bley-Vroman’s [1989] Fundamental Di¤erence). But implicit learning is slow and requires massive exposure to the language, the type of intensive exposure that is available to children learning their first language (DeKeyser and Larson-Hall 2005; Ellis 2002). However, in a typical foreign language situation such as the one studied in the BAF project, learners have very limited access to input. In those circumstances, without massive exposure, without repeated instances of items and constructions, implicit learning cannot take place. Children are then deprived of their potential advantage, an advantage that underlies their eventual long-term superiority (see for an extended discussion Mun˜oz 2006b), which explains their paradoxical disadvantage in a typical foreign language setting. At the same time, the superior cognitive development of older learners also makes them better at

152

Carmen Mun˜oz

explicit learning, the kind of learning that is facilitated in typical classrooms. L2 input is not only limited in quantity and intensity in this type of situation, but it may also be limited in quality. In particular, L2 input may have been poor from a phonological point of view, which can also explain the unexpected results obtained in the phonetic tests by the learners in this study. Both in a phonetic imitation test and in a phonetic discrimination test di¤erences between the learner groups were neither significant or consistent (see Fullana 2005, 2006). This finding runs counter expectations of superior pronunciation skills on the part of young learners, the most robust finding in terms of age e¤ects in a naturalistic language learning situation. In fact, it has been claimed that phonology may be the only linguistic domain a¤ected by the CPH because of its neuromuscular basis (Scovel 1988). However, similar results were obtained in a study of foreign language learning by Garcı´a-Lecumberri and Gallardo (2003), as well as in certain L2 immersion settings where learners receive accented input from their teachers (Flege 1991). A plausible explanation to this finding was given by Flege (1991) in the form of the ‘‘accented L2 input hypothesis’’ which posits that regardless of an early starting age of L2 learning, learners will not be able to perceive and produce L2 sounds accurately if they have received accented input in the L2. In sum, the quantity as well as the quality of the input received by foreign language learners can be seen to play a determinant role in the results obtained by younger and older learners in a typical foreign language setting. The next and final section will bring together the results from the present study with results from other studies in foreign language learning situations and will contrast them to findings from naturalistic learning settings. The latter, as commented above, have shaped the consensus traditional view of age e¤ects on second language learning (see for a discussion Mun˜oz 2008).

4.

General discussion and conclusion

The BAF project extended until the end of secondary education, that is, for a period of time that allowed to obtain results that may be considered to correspond to the ‘‘ultimate attainment’’ that is achievable in the school system. It may be argued that those findings, which are

Input and long-term e¤ects of early learning in a formal setting

153

interpreted to show long-term e¤ects in a school learning setting, have been obtained after a time span that is shorter in number of hours than the periods considered in long-term naturalistic studies (see Singleton and Ryan 2004: 99). A problem for making such comparisons is that given the lack of relevant empirical evidence they must depend on estimations. For example, research by Snow and Hoefnagel-Ho¨hle (1978) comparing di¤erent-age learners in an immersion situation found that older starters outperformed younger starters in the first comparisons but that after approximately one year younger starters were catching up with older learners. Singleton (1989: 236) estimated that 18 years would be necessary in a formal learning situation to provide as much input to younger learners so that they could also catch up to older learners (and eventually outperform them as in a naturalistic learning situation). It is evident that finding a large enough sample of learners who engage in a non-interrupted period of so many years of study of a foreign language without problems of articulation between di¤erent cycles or periods or even problems of tiredness and lack of motivation, and having received no additional input by means of stays in the target language country can be di‰cult. Besides, from an educationally oriented perspective and in view to inform language planning decisions, the most relevant results are arguably those that correspond to the period up to the end of secondary schooling, because that is the end of the language learning process for the majority of school learners. This is, however, an issue that must be empirically investigated and findings would be of the utmost importance for elucidating the influence of age on the long-term of foreign language learning. A di¤erent question that has also been speculated about in naturalistic learning studies that investigate age e¤ects on ultimate attainment is the length of residence required in order to uncover ultimate attainment rather than rate e¤ects. It was seen above that a period of 5 or better 10 years may be methodologically necessary to begin to analyze ultimate attainment results, and that Long (2007) remarked that after that initial period LOR ceases to correlate with proficiency scores (leaving only starting age as the determinant factor). However, an estimation of the number of years that would be needed in order to reach a comparable amount of exposure in a regular foreign language programme as that provided by 5–10 years of social immersion may extend well beyond a life-time. Consequently, if the learning process does not go beyond that ‘‘initial’’ period, it may be argued that the amount of input received by learners will never cease to be a determinant factor in a foreign

154

Carmen Mun˜oz

language setting, and thus it will never cease to correlate with proficiency scores. LOR or number of years or courses of instruction represent a quantitative dimension of input but it is not the only one. The intensity of input or its distribution in time is another quantitative dimension of input that may play a determinant influence on the results obtained in age-related studies. Notwithstanding, estimations of the input that may be equivalent to that received by naturalistic L2 learners have always been based by default on the typical time distribution of a foreign language programme, usually two to three periods of approximately one hour per week. For the purposes of this discussion that type of input distribution could be referred to as ‘‘extended input’’. We saw above that a plausible explanation for the younger learners’ lack of advantage may lie in the fact that they do not have access to massive intensive exposure that can activate their allegedly superior implicit language learning mechanisms. Following this line of thought, what learners – and particularly the younger ones – need is not an extremely, almost impossibly, long period of ‘‘extended input’’, but a substantial period of ‘‘intensive input’’ for implicit learning to take place. The importance of input frequency and intensity in language acquisition is emphasized by a cognitive perspective that explains that retrieval of knowledge from working memory for production and comprehension is facilitated by frequent repetitions in the input. Conversely, if input is distributed in very small doses separated by long stretches of time retrieval and use will be less e‰cient because traces of form-meaning mappings will not be available in working memory and progress in learning will be comparatively deficient. This may also partly account for the superior results obtained by formal learners in intensive courses over learners in regular courses even when the total amount of hours of instruction is the same (Serrano and Mun˜oz 2007; Serrano 2007). Finally, it was noted above that research has shown that in a naturalistic learning setting high-quality input has a determinant role in L2 pronunciation. In the same way, formal learners need high-quality input as well as numerous opportunities for using the language. In that line, Piske (2007) remarks that simply spending many years in a foreign language classroom will not help students to learn a foreign language well and that over the years students will only show accurate pronunciation if they receive a substantial amount of high-quality input (and phonetic training as well).

Input and long-term e¤ects of early learning in a formal setting

155

To summarize, in the present paper it has been claimed that di¤erences in the input L2 learners receive may be relevant for explaining the divergent findings observed in a naturalistic and in a foreign language learning setting. First of all, it has been argued that if input is sparcely distributed it will continue to exert an influence on language learning outcomes beyond the first years in contrast to L2 learning that occurs in a naturalistic setting. Most importantly, di¤erences in input may help explain the divergent results found in terms of the long-term benefis of an early start. In a foreign language setting the lack of massive amounts of input intensively distributed does not facilitate implicit learning on the part of younger learners. This deprives them of the possibility of reaching a proficiency level that is near native-like as in naturalistic language learning, or in any case superior to that of older learners. It is argued as well that even with a substantial amount of contact hours with L2, if input is extended along many years retrieval and use will be less e‰cient and progress in learning will be comparatively deficient. To conclude, the view that emerges from research in foreign language learning settings contrasts with the ‘‘consensus’’ view in which the burden of learning falls almost exclusively on the learner’s innate capabilities, a view which has gone along well with the belief that an early start in a school setting will be automatically successful, without considering the quality, quantity and intensity of L2 input.

Acknowledgement The research described in this chapter was supported by grant HUM2004-05167 of the Spanish Ministry of Education and grant 2005SGR00778 of the Generalitat de Catalunya. I would like to thank Raquel Serrano for her inspiring work.

References Bley-Vroman, Robert 1989 What is the logical problem of foreign language learning? In: Susan Gass and Jackeline Schachter (eds.), Linguistic Perspectives on Second Language Acquisition 41–68. Cambridge, UK: Cambridge University Press.

156

Carmen Mun˜oz

Burstall, Clare, Monika Jamieson, Susan Cohen, and Margaret Hardgreaves 1974 Primary French in the Balance. Windsor: NFER Publishing Company. Cenoz, Jasone 2002 Age di¤erences in foreign language learning. I.T.L. Review of Applied Linguistics 135–136: 125–142. Chomsky, Noam 1959 Review of Skinner. Verbal Behavior, Language 35: 26–58. 1965 Aspects of the Theory of Syntax. Cambridge, Mass.: MIT Press. DeKeyser, Robert 2000 The robustness of critical period e¤ects in second language acquisition. Studies in Second Language Acquisition 22: 499–533. DeKeyser, Robert and Jennifer Larson-Hall 2005 ‘What does the critical period really mean?’ In: J.F. Kroll and A.M.B. de Groot (eds.), Handbook of Bilingualism: Psycholinguistic Approaches, 88–108. Oxford: Oxford University Press. Ellis, Nick C. 2002 Frequency e¤ects in language processing. Studies in Second Language Acquisition 24: 143–188. Flege, James 1987 A critical period for learning to pronounce foreign languages? Applied Linguistics 8: 162–177. 1991 Perception and production: the relevance of phonetic input to L2 phonological learning. In: T. Huebner and C. A. Ferguson (eds.), Crosscurrents in Second Language Acquisition and Linguistic Theories, 249–290. Philadelphia: John Benjamins. 2008 ‘Give Input a Chance!’ In: T. Piske & M. Young-Scholten (eds.), Input matters in SLA, 175–190. Clevedon: Multilingual Matters. Flege, James and Serena Liu 2001 The e¤ect of experience on adults’ acquisition of a second language. Studies in Second Language Acquisition 23 (4): 527–552. Foster-Cohen, Susan H. 1999 An Introduction to Child Language Development. London: Longman. Fullana, Natalia 2005 Age-related e¤ects on the acquisition of a foreign language phonology in a formal setting. Ph. D. diss., University of Barcelona. 2006 The Development of English (FL) Perception and Production Skills: Starting Age and Exposure E¤ects. In: Carmen Mun˜oz (ed.), Age and the Rate of Foreign Language Learning, 41–64. Clevedon: Multilingual Matters. Garcı´a Lecumberri, M.Luisa and Francisco Gallardo 2003 English FL Sounds in School Learners of Di¤erent Ages. In: M. Pilar Garcı´a Mayo and M. LuisaGarcı´a Lecumberri (eds.), Age and the Acquisition of English as a Foreign Language, 115–135. Clevedon: Multilingual Matters.

Input and long-term e¤ects of early learning in a formal setting

157

Garcı´a Mayo, M. Pilar and M. Luisa Garcı´a Lecumberri (eds.) 2003 Age and the Acquisition of English as a Foreign Language. Clevedon: Multilingual Matters. Gran˜ena, Gisela 2006 Age, Proficiency Level and Interactional Skills: Evidence from Breakdowns in Production. In C. Mun˜oz (ed.): Age and the Rate of Foreign Language Learning. Clevedon: Multilingual Matters, pp. 183–207. Han, ZhaoHong and Terence Odlin 2006 Studies of Fossilization in Second Language Acquisition. Clevedon: Multilingual Matters. Hatch, Evelyn 1977 Optimal age or optimal learners? Workpapers in TESL 11: 45–56. 1978 Discourse analysis and second language acquisition. In: Evelyn Hatch (ed.), Second Language Acquisition: A Book of Readings. Rowley, MA: Newbury House. 1983 Simplified input and second language acquisition. In: R. Andersen (ed.), Pidginization and Creolization as Language Acquisition. Rowley, MA: Newbury House. Hornstein, Norbert and David Lightfoot (eds.) 1981 Explanation in Linguistics: The Logical Problem of Language Acquisition. London: Longman. Johnson, Jacqueline and Elissa Newport 1989 Critical period e¤ects in second language learning: The influence of maturational state on the acquisition of ESL. Cognitive Psychology 21: 60–99. Krashen, Stephen D. 1981 Second Language Acquisition and Second Language Learning. London: Pergamon. 1982 Accounting for child.adult di¤erences in second language rate and attainment. In: S. Krashen, R. Scarcella and M. Long (eds.), ChildAdult Di¤erences in Second Language Acquisition, 202–226. Rowley, Mass.: Newbury House. 1985 The Input Hypothesis: Issues and Implications. London: Longman. Krashen, Stephen, D., Michael, H. Long, and Robin C. Scarcella 1979 Age, rate, and eventual attainment in second language acquisition. TESOL Quarterly 13 (4): 573–582. Lapkin, Sharon, Merril Swain, J. Kamin, and G. Hanna 1980 Report on the 1979 evaluation of the Peel County late French immersion program, grades 8, 10, 11 and 12. Unpublished report, University of Toronto, OISE. Lenneberg, Eric. H. 1967 Biological Foundations of Language. New York: Wiley. Long, Michael H. 1990 Maturational constraints on language development. Studies in Second Language Acquisition 12 (3): 251–285.

158

Carmen Mun˜oz

2007 Problems in SLA. Mahwah, NJ: Lawrence Erlbaum Associates. Miralpeix, Immaculada 2008 The Influence of Age on Vocabulary Acquisition in EFL. Ph. D. diss., University of Barcelona. Mun˜oz, Carmen 2001 Factores escolares e individuales en el aprendizaje formal de un idioma extranjero. In S. Pastor and V. Salazar (eds.) Tendencias y lı´neas de investigacio´n en ASL. Estudios de Lingu¨´ıstica, 249–270. University of Alicante. 2002 Codeswitching as an accommodation device in L3 interviews. Paper presented at the 12th EUROSLA Conference, Basilea. 2006a (ed.) Age and the Rate of Foreign Language Learning. Clevedon: Multilingual Matters. 2006b The e¤ects of age on foreign language learning: The BAF Project. In: Carmen Mun˜oz (ed.), Age and the Rate of Foreign Language Learning, 1–40. Clevedon: Multilingual Matters. 2008 Symmetries and Asymmetries of age e¤ects in naturalistic and instructed L2 learning. Applied Linguistics 24 (4): 578–596. Nave´s, Teresa 2006 The long-term e¤ects of an early start on foreign language writing. Ph. D. diss., University of Barcelona. Nikolov, Marianne 2000 The Critical Period Hypothesis reconsidered: Successful adult learners of Hungarian and English. IRAL 38 (2): 109–124. Niz´egorodcew, Anna 2007 Input for Instructed L2 Learners. The Relevance of Relevance. Clevedon: Multilingual Matters. Piske, Thorsten 2007 Implications of James Flege’s research for the foreign language classroom. In: Ocke-Schwen Bohn and Murray J. Munro (eds.), Language Experience in Second Language Speech Learning. In honour of James Emil Flege, 301–314. Amsterdam/Philadelphia: John Benjamins Publishing. Scarcella, Robin and Corrine Higa 1982 Input and age di¤erences in second language acquisition. In: S. Krashen, R. Scarcella and M. Long (eds.), Child-Adult Di¤erences in Second Language Acquisition, 175–200. Rowley, Mass.: Newbury House. Scovel, Thomas 1988

A Time to Speak. A Psycholinguistic Inquiry into the Critical Period for Human Speech. Rowley, MA: Newbury House. Selinker, Larry 1972 Interlanguage. International Review of Applied Linguistics 10 (2): 209–231.

Input and long-term e¤ects of early learning in a formal setting

159

Serrano, Raquel 2007 The E¤ect of the Distribution of Instruction Hours on the Acquisition of English as a Foreign Language. Ph. D. diss., University of Barcelona. Serrano, Raquel and Carmen Mun˜oz 2007 Same hours, di¤erent time distribution: any di¤erence in EFL? System 35 (3): 305–321. Singleton, David 1989 Language Acquisition: The Age Factor. Clevedon: Multilingual Matters. 2005 The Critical Period Hypothesis: A coat of many colours. International Review of Applied Linguistics 43 (4): 269–285. Singleton, David and Lysa Ryan 2004 Language Acquisition: The Age Factor. (2d ed.). Clevedon: Multilingual Matters. Snow, Catherine 1983 Age di¤erences in second language acquisition: Research findins and folk psychology. In: K. Bailey, M. Long and S. Peck (eds.), Second Language Acquisition Studies, 141–150 Rowley. Mass.: Newbury House Publishers. 1986 Conversations with children. In: P. Fletcher and M. Garman (eds.), Language Acquisition: Studies in First Language Development. Cambridge: Cambridge University Press. Snow, Catherine and Marian Hoefnagel-Ho¨hle 1978 The critical period for language acquisition: Evidence from second language learning. Child Development 49: 1114–1128. 1979 The linguistic environment of school-age second language learners. Unpublished manuscript. Harvard University. Steinberg, Danny 1993 An Introduction to Psycholinguistics. London: Longman. Tomasello, Michael 2003 Constructing a Language: A Usage-Based Theory of Language Acquisition. Cambridge, Mass.: Harvard University Press. Winitz, Harris, Brenda Gillespie, and Jennifer Starcev 1995 The development of English speech patterns of a 7-year-old Polishspeaking child. Journal of Psycholinguistic Research 24, 117–143.

8.

English /l/ and /r/ acquisition by Japanese children and adults

Chise Kasai

1.

Introduction

This chapter is based on a larger study (Kasai, 2001) investigating how age plays its role in second language acquisition. Subjects are 53 Japanese children and adults residing in the UK, and their ability is tested to produce and discriminate English /l/ and /r/ sounds, which are said to be the most di‰cult sounds for Japanese to acquire. In the past decades, researchers have investigated the age factor. The conflicting arguments of ‘the younger, the better’ and ‘the older, the better’ still remain controversial. However, one thing which is certain is that when it comes to acquiring pronunciation, children often show much better performance than adults (Asher and Garcia 1969; Ekstrand 1976; Oyama 1975; Patkowski 1990; Snow and Hoefnagel-Hoe¨hle 1978; Tahta, Wood, and Loewenthal 1981a; Thompson 1991), but with regard to the acquisition of other elements such as syntax, morphology and so on, their results often vary. Therefore, the first experiment in this research focuses on the production ability of Japanese children and adults using English words which contain /l/ and /r/ sounds in di¤erent positions. In the second experiment the subjects’ discrimination ability is investigated. The words used in this experiment also contain English /l/ and /r/ sounds in di¤erent positions. In the area of phonetics several studies have investigated Japanese learners’ perceptual ability in conjunction with the age factor. Again the issue is divisive: some researchers support ‘the younger, the better’ argument, whereas others support ‘the older, the better argument’. Children acquire these particular sounds with a great deal less e¤ort than adults. Although it is possible for adult learners to acquire the sounds, intensive training is needed. In both experiments, position of /l/ and /r/ are also investigated to see relationships where these sounds occur in words and di‰culty of acquisition. Namely, /l/ and /r/ in the positions of word initial, word medial intervocalic, and consonantal clusters are examined.

162

Chise Kasai

One of the problems in previous research was subjects’ starting age of second language (L2) learning. For example, comparing a young learners’ group and an older learners’ group was not possible because each group had di¤erent L2 learning histories. Therefore, in this research, the age of entry into Britain was set as a starting age. One may claim that the age of entry to a host country is not exactly the same thing as the starting age of second language learning, since each individual might have a di¤erent L2 learning history prior to entering the host country. This applies to the subjects in this experiment. Those who are categorised as adults had English learning experience ranging from 6 to 10 years before entering Britain, while those who are categorised as children had much less experience. However, the English teaching which the adult group had undergone in Japan was presumably mainly based on grammar and reading, and the amount of tuition to help develop the subjects’ communicative or listening skills were significantly less. More precisely, although such classes are incorporated in the conventional curriculum, it seems that the majority of the teachers are native speakers of Japanese and it does not seem to be the case that the subjects had much exposure in developing native-like fluency. Also it must be noted that the subjects had learned English as a foreign language, not as a second language. This means that their use of English was limited to the classroom and they had few opportunities to use English as a mean of communication outside the classroom. In short, any previous exposure to English that the subjects had cannot be compared to the one which they would experience after entering Britain. Finally, this research focuses on subjects’ pronunciation and listening skills after having a certain amount of exposure to English through living in an Englishspeaking environment. Therefore, in this experiment the subjects’ age of entry is considered as their starting point of second language learning, on the assumption that both groups of children and adults had almost no previous exposure at all to native English.

2. 2.1.

Research design and method Production task and discrimination task

A production task and a discrimination task were used to examine subjects’ ability in acquiring /l/ and /r/ sounds. The production task used a

English /l/ and /r/ acquisition by Japanese children and adults

163

‘recording’ as stimulus. A native English speaker (on tape as a narrator) pronounced 98 English words, including 18 screen words, which in turn had to be repeated by the Japanese subjects, one word after another. The subjects were also given the words written down on paper for visual support. The subjects’ pronunciation was recorded individually and scored on a 5-point scale by three judges. In the discrimination task, the subjects were asked to listen to a stimulus tape which recorded 50 minimal pairs and to decide which word of the minimal pairs they had heard. 36 minimal pairs contained either /l/ or /r/ sounds such as ‘rake-lake’, ‘election-erection’ and the rest did not contain any of the sounds such as ‘hit-fit’, ‘bet-vet’. The sets were mixed randomly so that subjects were prevented from realising which sounds they were being tested on. In both experiments three research questions were asked. 1) Are children better than adults? 2) Which is more di‰cult to acquire /l/ or /r/? 3) Does a position of /l/ and /r/ in words a¤ect acquisition?

2.2.

Subjects

Participants were 27 Japanese children (10 female and 17 male) and 3 male native English children as control subjects. The age range was 6 to 14, and the mean age of entry was 5:5; the mean age at the experiment was 9:6. In the adult group there were 27 Japanese adults (15 female and 12 male) and 2 male and 1 female native English adults as control subjects. Their age ranged between 19 and 35, and their mean age of entry was 21:8, and their mean age at the experiment was 25:10. The average length of exposure to English was 4:1 for children, and 3:3 for adults, therefore, children had 10 months longer exposure to English than adults. For the children, the age of entry into Britain was under 12 and in the case of the adults it was after the age of 13 years. Both groups of Japanese children and adults had lived in the UK for at least two years with similar backgrounds. To classify the backgrounds, in the case of the children, their parents were Japanese businessmen and the adults were Japanese students studying at universities in the UK. Regarding their level of English, both groups were involved with the English education system. Although there is no record of the children’s group having had any sort of tests done to enable them to attend schools in the UK, the adult group had been required to take certain English proficiency tests such as TOEFL or IELTS. On entry into uni-

164

Chise Kasai

versity they were required to score at least 6.0þ on IELTS or 550þ on TOEFL in order to be accepted for an undergraduate course. For the postgraduate course they were required to score at least 6.5þ on IELTS or 600þ on TOEFL. Regarding exposure to English, the children attended local schools with native English children and the adults attended the University of Essex in the UK. Both groups had an equivalent exposure to English. For the production task, the three judges used were all native English RP speakers; a 45-year-old male postgraduate student, a 35-year-old female mature student and a 24-year-old female postgraduate student all at the University of Essex. All of them were reading linguistics and had knowledge in this field. They were instructed how they should rate each subject’s pronunciation on a 5-point scale and were given the tapes with marking sheets. There was slight concern that the judges might have been a¤ected by Estuary English which had spread in the Essex region. However, although all the judges were reading linguistics at the University of Essex, none of them was a native of this region. Also the syllable final /l/ which is pronounced as /w/ in Estuary English is not tested in this research. Therefore, the judges were considered to be suitable for this research. For the discrimination task, no judges were needed since all the answers were written on the answer sheets and it was a matter of right or wrong. Therefore, the scores were marked and calculated using SPSS 10.0. 2.3. 2.3.1.

Data collection instruments Words used in this experiment

There were 98 English words to be pronounced by the subjects; 80 of which contained /l/ or /r/ sounds and 18 screen words which did not. They were mixed randomly on the recording so that the subjects could not tell which sounds were being tested. The 80 words with /l/ or /r/ sounds used in this experiment were selected from Gimson (1989) and are represented in Table 1. As Gimson’s An Introduction to the Pronunciation in English (1989) is a standard authority, it was thought to be the best source of words. Four types of /l/ sounds and three types of /r/ sounds mentioned below were chosen particularly since they were assumed to be the most obvious /l/ and /r/ sounds with which Japanese people have di‰culties.

English /l/ and /r/ acquisition by Japanese children and adults

165

Table 1. Stimulus words chosen from Gimson’s classification (Gimson 1989) Four di¤erent types of /l/, Clear /l/ Word initial alone voiced

leave, let, late, loud, leer

Word medial alone voiced intervocalic

silly, yellow, alloy, foolish, sailor, oily, million, failure, allow, select

voiceless in cluster initial

blow, glad, splice, fly, slow

voiceless in cluster medial

play, clean, butler, hopeless, sprinkler, couplet, antler, simplest, ghastly

Four di¤erent types of /r/ Word initial alone voiced

reed, rag, raw, rude, road, royal, rear

Word medial intervocalic alone voiced in consonantal cluster

mirror, very, arrow, sorry, hurry, furry, arrive, diary, dowry, eery

(following fortis accented plosive ¼ [ä ])

proud, tree, try, expression, surprise, attract, extremely

(following fortis fricative, unaccented fortis plosive, or accented fortis plosive preceded by /s/ in the same syllable ¼ somewhat devoiced [ä])

fry, afraid, throw, thrive, shrink, shrug, apron, cockroach, sprint, street

(following lenis consonant ¼ [ä]-fricative after /d/)

brief, bright, grey, grow, umbrella, address, hungry, general, miserable

Some of the 98 words used in this experiment are also used in the discrimination task. However, in order to build a su‰cient number of minimal pairs, some words were added. Table 1 adapts the classification used by Gimson (1989) illustrated with words from my experiment. 2.3.2. Screen words There were some words which did not contain /l/ or /r/ sounds, (called ‘screen words’ or ‘distracters’), which were mixed randomly with the 80 /l/ and /r/ words. The purpose was to prevent subjects from concentrating on /l/ and /r/ sounds consciously. It was thought that if the subjects realised that they were being tested on the particular sounds, they might have performed them in an unnatural way and that this might a¤ect the result.

166

2.3.3.

Chise Kasai

Frequency of the words

The frequency of each word was checked in the Collins COBUILD English Dictionary (1995). In this dictionary there are five frequency bands shown by black diamonds in an extra column. The most frequent words have five black diamonds, the next four diamonds, and so on. The presence of a word in the top three bands indicates that it is essential in construction, grammar and conversation and these bands account for approximately 75% of all English usage. The remaining bands (which indicate frequencies 2 and 1) are less frequently used, but are still important. In this dictionary there are some other words which do not belong to any of the 5 bands. They are the words which occur less frequently than others, but they are still important and deserve an entry in the dictionary. According to the editor of this dictionary, all the words found in the dictionary were based on the vocabulary which English learners are most likely to meet and thousands of other words were not included in the dictionary. All the words chosen for this experiment were indeed found in the dictionary, indicating that they are commonly used words and therefore, thought to be no problem for the subjects to understand. For reference, those which do not belong to the 5 bands, yet are still in this dictionary are: bleed, crush, brew, clue, globe, glow, glue, grove, lash, loyal, pirate, pray, rash, rot, sin, sprint, vat, vest and vet, which seem to appear commonly in daily life. The reason for checking frequency is that this research aims to test subjects’ ability to produce /l/ and /r/ sounds in a natural setting. So, if words rarely used in daily life had been used in this experiment, it was thought to be unfair for the subjects and that the result might be a¤ected. However, after checking frequency in three di¤erent dictionaries, the conclusion was reached that it does not really matter, or rather it was impossible to argue that those words were guaranteed to be commonly used words. For example, according to BNC Frequency List ‘leave’ is used 12000 times and ‘cockroach’ is used 65 times out of 100 million times. Although the use of ‘cockroach’ is much less frequent than ‘leave’, it was supposed that everyone on this experiment knew the meaning of ‘cockroach’ as well as they knew the meaning of ‘leave’. Thus, as the editor of COBUILD dictionary states that inclusion of a word in the dictionary already indicates common usage, otherwise they would not be included.

English /l/ and /r/ acquisition by Japanese children and adults

167

2.3.4. Stimulus tape For the production task, a stimulus tape was made, which consisted of three sections. The first section recorded an introduction to explain how the experiment was to be carried out. The second section was practice, which allowed the subjects to get used to the style of the experiment: subjects were asked to pronounce some words after a narrator’s voice. (i.e. The narrator’s voice says ‘Now you are going to hear some words. Listen carefully and repeat after me.’ Then the narrator pronounces ‘book’. There is an interval for subjects to pronounce the same word. Three words which will not appear in the actual experiment were given as a practice.) Finally, the actual test section was recorded after these sections. The process was exactly the same as the practice section and it took approximately 19 minutes. As for the discrimination task, a stimulus tape which runs parallel with the answer sheet was made. This tape consisted of an introduction part, practice part and the actual test part. First, the same narrator from the production task explained how the test was going to be carried out in the introduction part; secondly the subjects were instructed to try four examples in the practice part so as to get used to the style of this test. Finally, in the test part 50 words were pronounced. The subjects were required to choose which of the words in the minimal pair had been pronounced. There were approximately 3 seconds between the words, and the subjects were asked to circle on the answer sheet the word which they thought was the one pronounced. 2.3.5. Narrator The narrator who constructed the stimulus tapes was a female native speaker of English aged 25. She was from the Oxford area and was considered to be an RP speaker. The reason for having a native speaker as a narrator was to make subjects hear natural English pronunciation. Also, the stimulus tapes had to be as clear as possible so that the subjects could hear each sound clearly. The narrator was instructed to allow approximately two second intervals in between each word for the production task, and three seconds for the discrimination task. These intervals were decided after pilot tests. Subjects on the pilot tests showed that these intervals were the most comfortable. In both tasks, once the recording was started, the whole process continued without stopping the tape, in order to prevent mechanical stopping noise.

168

2.3.6.

Chise Kasai

Equipment

The stimulus tapes were recorded in a language laboratory and the equipment used was Beyerdynamic MPC 50, which recorded every sound clearly, and a Panasonic BCX CD player was used to play the stimulus tapes. Some blank tapes were prepared to record subjects’ pronunciation. 2.3.7.

Questionnaire

For reference purposes a questionnaire written in Japanese was prepared to ask about subjects’ backgrounds. The first part of the questionnaire asked basic questions such as their names, ages, age at entry to the UK, and length of stay. The second part asked more precise questions regarding their English ability. The subjects were either given a 5-pointscale or spaces in which to freely express their answers. This questionnaire was used to obtain as much information as possible about the subjects. It was interesting to see how each subject had rated their own ability. In the children’s questionnaire the subjects tended to be quite confident in their second language acquisition, while adults largely showed their awareness of failure to learn /l/ and /r/ sounds. Some children even had their own theories on second language acquisition and how to overcome such di‰culties as, ‘‘Listening or comparing my own pronunciation to the others will help’’ and ‘‘You have to like the English people first, then you can become friends and talk to them’’. Some of the questions may not relate directly to the purpose of this research; however, in order to refer to the result later on, it was thought to be useful to obtain a variety of background information. Eventually there were 54 subjects all together. It proved to be a good way of remembering them as individuals. 2.3.8.

Answer sheet

To run parallel with the stimulus tape, an answering sheet was prepared for the discrimination task. At the top of the sheet there were four sets of minimal pairs for practice before proceeding to the actual test part. Then the rest presented 50 minimal pairs for the actual test. The sets were listed from top to bottom from left to right and each set was numbered in order to avoid confusion.

English /l/ and /r/ acquisition by Japanese children and adults

2.4.

169

Procedure of the production task

1) Subjects were admitted into a language laboratory individually for the recording. In the laboratory an examiner operated the equipment. In order to treat all the subjects equally, the examiner remained as silent as possible and the stimulus tape had all the instruction needed. 2) After being seated, the subject was asked to complete a questionnaire. 3) The stimulus tape was switched on. The subjects listened to the first introductory section to understand how the experiment was to be carried out. 4) The tape went onto the second section, which enabled the subjects to practise. 5) Three words were given to them to repeat after the narrator’s voice. 6) After the narrator said ‘Are you ready? Let us begin’, the actual experiment began. The subjects were required to repeat 98 English words after the narrator’s voice. The whole experiment lasted approximately 19 minutes. During recording the subjects were given a word list for visual support. For scoring the judges were instructed to concentrate on subjects’ native-like proficiency of /l/ and /r/ sounds. Each pronunciation was assessed on a 5-point scale, as follows: 1 for absolutely foreign accent – There are many mistakes, mispronunciation of /l/ and /r/ sounds or pronouncing completely di¤erent sounds; 2 for heavy foreign accent – Words are pronounced correctly, but a foreign accent is noticed easily; 3 for slightly foreign accent – Not as bad as 2, but it is easy to tell that the speaker is a foreigner making an e¤ort to sound like a native speaker of English; 4 for near native – Words are pronounced correctly, but still a slight accent is noticed; 5 for absolutely native accent – Without doubt the speaker is a native speaker of English. The point scale method has been used by several linguists in examining learners’ language acquisition, especially in acquiring native accents, and there seem to be two types of scale: 3-point scale and 5-point scale. For instance, Oyama (1975) and Patkowski (1980) used a 5-point scale assessment when examining immigrants’ pronunciation, while Tahta et al. (1981b) adopted the 3-point scale. In this experiment a 5-point scale is adopted because it was thought to give a wider choice of accents to describe subjects’ pronunciation.

2.5.

Procedure of the discrimination task

1) Following a short break after the production task, subjects were instructed to proceed to the discrimination task. As before, the same

170

Chise Kasai

examiner stayed in the lab to operate the equipment. 2) The tape was switched on and subjects were required to listen to the tape to understand the procedure. First, an introductory part was recorded to explain how the experiment was to be carried out. Second, the tape announced a practice part which allowed the subjects to try four examples. Finally, with the narrator’s voice, ‘Are you ready? Let us begin.’ the actual experiment was started. 3) Once the experiment had begun, the stimulus tape went without stopping in order to put all the subjects in an identical situation.

2.6. 2.6.1.

Results from the production task Reliability of judges and words

All the statistical analyses in this research were calculated using SPSS 10.0 version. To verify the results, the figures are checked using Levene’s Test of Equality of Error Variances and Mauchly’s Test of Sphericity. If the value is calculated to be not significant in Levene’s, it proves that the data was suitable for analysis by ANOVA. In the case of Mauchly’s, if the value was calculated to be significant, the further analysis adopts the significance value of Greenhouse-Geisser, and if this was proved to be not significant, the further analysis adopts the significance value of Sphericity Assumed. Regarding the further analysis, Levene’s Test has been passed for all the results. Furthermore, the level of significance is set at p < .05 in all calculations in the following analysis for both the production and discrimination tasks. These tests were adopted because they were commonly used as standard tests in statistical analysis of similar data. 2.6.2.

Agreement in level of judgement

The reliability of the judges was examined in order to check whether they scored the subjects consistently. This was done in two ways. First a comparison was made to verify that the judges were giving the same level of scores to subjects in an absolute sense. Tables 2 and 3 give an overall view of how each judge marked the subjects, presenting the mean, minimum and maximum scores for each judge.

English /l/ and /r/ acquisition by Japanese children and adults

171

Table 2. Scores produced by each judge for child subjects Variables

Mean

Std Dev

Minimum

Maximum (out of 490)

N

Judge 1

386.83

56.82

265

466

30

Judge 2

429.67

74.86

247

490

30

Judge 3

460.73

55.80

188

490

30

At this stage, there was a slight concern over the way Judge 1 marked child subjects. As can be seen in the maximum score, Judge 1 is lower than the other two: Judge 2 and 3 had 490 as the maximum score, whereas Judge 1 had 466. It was found that Judge 1 marked child subject 2 with a noticeably lower score. In most circumstances, omitting a subject from the results of an experiment is not acceptable; however, as all the other subjects were judged similarly by the other two Judges, we decided to omit this special case (i.e. Child subject 2 marked by judge 1). Subject 2 has therefore, been eliminated from all the results, thus reducing the total number of children to 26. It is also apparent that Judge 3 has a greater range for children and a smaller range for adults than the other two judges. Table 3. Scores produced by each judge for adult subjects Variables

Mean

Std Dev

Minimum

Maximum (out of 490)

N

Judge 1

275.97

81.33

173

490

30

Judge 2

359.93

76.07

197

485

30

Judge 3

373.03

67.52

232

488

30

2.6.3. Correlation between judges After adjusting the subjects, the reliability of the judges for the children’s group was calculated, and this time the Pearson values were r ¼ .7575 for Judge 1 against Judge 2, r ¼ .7944 for Judge 1 against Judge 3 and r ¼ .8760 for Judge 2 against Judge 3. In the case of the adults, the values were r ¼ .7944 for Judge 1 against Judge 2, r ¼ .8367 for Judge 1 against Judge 3 and r ¼ .8411 for Judge 2 against Judge 3. These are all close enough to 1.0000 to be acceptable.

172

2.6.4.

Chise Kasai

Internal reliability of subsets of words

The reliabilities of the words used for this experiment were examined. The nine types of /l/ and /r/ words (i.e. in word initial, word medial, and in clusters) were checked to establish whether any words within the sets were inconsistent with the overall results for the set. The following table shows values of alpha from each type. Table 4. Reliability of words within the groups Types /l/ and /r/

Value of alpha

Types /l/ and /r/

Value of alpha

/l/ in initial position

.9786

/r/ in medial position

.9743

/l/ in medial position

.9773

/r/ in cluster following fortis accented plosive

.9657

/l/ in initial cluster

.9714

/r/ in cluster fortis fricative unaccented fortis plosive

.9719

/l/ in medial cluster

.9764

/r/ in cluster following lenis consonant fricative after /d/

.9710

/r/ in initial position

.9758

As can be seen in Table 4, all sets of /l/ and /r/ words proved to be highly reliable. This is attributed in part to the fact that pooling the judgements of 3 judges already increased reliability, in part to the fact that there were at least 9–10 items in each set and in part to the apparent ease of the task itself. 2.7.

General comparison between children and adults

First, general scores of children and adults on the production task are compared. The judges marked the subjects on a 5-pioint scale, as the raw scores show in Tables 2 and 3. However, in order to compare the production task and the discrimination task in later sections, scores from both tasks were calculated into a 4-point scale from this section onwards. This statistical control makes it possible to compare the production task and the discrimination task on the same scale. In Figure 1 the graph shows an upward curve to the right which indicates that there were more subjects who scored higher marks. Thus 15 children out of 26 scored 3.5 to 4.0 that is to say at native levels; 24 out

English /l/ and /r/ acquisition by Japanese children and adults

173

Figure 1. Child overall mean scores for /l/ and /r/ production

Figure 2. Adult overall mean scores for /l/ and /r/ production

of 26 scored better than 2.5, the near native level. Although a few did not perform well, the mean score is 3.34, which is considerably higher than the 2.23 of the adults. As far as the adults are concerned, the graph in Figure 2 is bellshaped with a central node, quite di¤erent from the one for the children. This indicates a greater variation in their scores; few adults achieve higher scores than children. Only 1 adult out of 27 scored 3.50 to 4.00, i.e. native level; only 7 out of 27 scored above 2.50, the heavy to slight accent level. It can be concluded from the above figures that children perform much better, whereas adults’ performance is more widely spread (i.e. Std. Dev of adults was .57 against .51 in the case of chil-

174

Chise Kasai

Figure 3. General comparison between children and adults

dren). In other words, the majority of children scored the highest nativelike score while the majority of adults scored around the average of 2.00 to 2.50 (Figure 3). This visual interpretation of the results is confirmed through the statistical calculations represented in the following tables. In the further analysis, because the significance value is calculated to be significant in Mauchly’s Test of Sphericity, further analysis is carried out which adopt the value of Greenhouse-Geisser, since these values are commonly used in this type of analysis to test significance values. As before, a value of p smaller than .05 is acceptable as significant in all further analyses. Table 5 shows the di¤erence between /l/ and /r/. The value of p was calculated as p ¼ .069; so it was proved to be not significant, meaning that the di¤erence between /l/ and /r/ is small. However, the value could be interpreted as ‘borderline’, as it is close to p ¼ .05. If a few more subjects or words had been added to the experiment, it could have been significant, which suggests that, though not proved here, there could be a di¤erence brought about by the small scale of this experiment, with /l/ slightly easier to produce than /r/. Table 5. A general comparison between /l/ and /r/, and subjects Source Di¤erence between /l/ and /r/ Di¤erence between children and adults Interaction between /l/ and /r/ and children and adults

F

Sig.

Eta Squared

3.461

.069

.064

52.334