INTRODUCTION

This special issue of JFLS focuses on what learners know about French words, on how they use that knowledge and on how it can be investigated and assessed.

In many ways, it is a sequel to the special issue on the Acquisition of French as a Second Language edited by Myles and Towell that appeared in JFLS in 2004. While articles on the L2 acquisition of the French lexicon have appeared in a variety of journals, including JFLS, this special issue (SI) is the first volume which specifically focuses on lexical knowledge and use among learners of French as a second language. The issue is timely, because of the growing importance of vocabulary in the SLA research agenda, but also because research into vocabulary acquisition appears at the top of a list of areas in which teachers of Modern Foreign Languages are most interested (Macaro, Reference Macaro2003: 6).

Over the past few decades interest in research into Second Language Acquisition has grown exponentially, as can be seen in the number of textbooks and handbooks that have appeared in recent years (Ritchie and Bhatia, 1996; Cook, 1996 et seq.; Doughty and Long, Reference Doughty and Long2001; Myles and Mitchell, Reference Myles and Mitchell2004), as well as in the formulation of new theories of SLA, such as Processability Theory (Pienemann, Reference Pienemann1998), Acquisition by Processing Theory (Truscott and Sharwood Smith, Reference Truscott and Sharwood Smith2004) and the Extended Competition Model (MacWhinney, Reference MacWhinney2005), which complement approaches to SLA that are based on Universal Grammar. As Myles and Mitchell (Reference Myles and Mitchell2004: 91) put it, approaches to SLA which are based on Universal Grammar, have mainly focused on describing and explaining morphosyntactic development in learners, and much less on other aspects of the linguistic system. In the Minimalist Program, however, the differences between languages are seen to be mainly lexical in nature. According to Cook (Reference Cook1998), the Minimalist Program is lexically-driven in that the properties of lexical items shape the sentence rather than lexical items being slotted into pre-existent structures. As a result, the task the L1 learner faces is mainly one of learning the lexicon (lexical and functional items), which then triggers the setting of universal grammatical parameters. This approach is reflected in the Lexical Learning Hypothesis (Ellis, Reference Ellis, Schmitt and McCarthy1997) according to which vocabulary knowledge is indispensable to acquire grammar (see also Bates and Goodman, Reference Bates, Goodman and Altmann1997).

There are many reasons why understanding how words are learned and used and how we can measure that knowledge is also crucially important for researchers working in Applied Linguistics, in particular for those with an interest in Education. As Milton (this volume) points out, vocabulary knowledge can be quantified in ways that other aspects of language knowledge cannot, which makes this area particularly attractive for the development of indices and measures that can be used in a variety of educational (and clinical) contexts to assess learners' or patients' language profiles.

The importance of vocabulary has been demonstrated in a wide range of studies. It is widely accepted that lexical knowledge is one of the main prerequisites for academic achievement of monolingual and bilingual children (see Daller, Reference Daller1999; Dickinson and Tabors, Reference Dickinson and Tabors2001). According to Meara and Bell (Reference Meara and Bell2001), teachers' judgements of L2 texts appear to be based to a large extent on the type of vocabulary used by the students. In a similar vein, Malvern and Richards (Reference Malvern and Richards2002) show that teachers' subjective rating of students' range of vocabulary in oral interviews correlated very highly (all values above 0.97) with their judgements about fluency, complexity, content and accuracy.

The language threshold for reading is also largely lexical. Anderson and Freebody (Reference Anderson, Freebody and Guthrie1981) report a high correlation between tests of vocabulary and reading comprehension across a range of studies in first language reading research. Laufer (Reference Laufer1992) points to similar results for second language acquisition. She also provides evidence that for text comprehension a vocabulary large enough to provide coverage of 95% of the words in a text is needed. Hu and Nation (Reference Hu and Nation2000) even put the boundary for unassisted reading at 98%.

A focus on vocabulary can also provide new insights into the distinction between implicit and explicit learning (Ellis, Reference Ellis1994; Hulstijn, Reference Hulstijn, Doughty and Long2003). Recent research has focused on incidental vocabulary, i.e. vocabulary that second language learners develop while they are focused on a task other than on learning new words (Gass, Reference Gass1999). Most scholars agree that except for the first few thousand most common words, L2 vocabulary is predominantly acquired incidentally (cf. Huckin and Coady, Reference Huckin and Coady1999). If this is correct, there are important implications for the ways in which vocabulary is offered in language classes and textbooks. The role of explicit versus implicit (or incidental) vocabulary learning is also taken up in two contributions to this special issue (Houten, Bulté, Pierrard and Van Daele; Graham, Richards and Malvern).

Much less is known with respect to the pragmatic and sociolinguistic knowledge about individual items, i.e. information about the ways in which particular words are used dependent on the formality of the conversation, the characteristics of the interlocutors and the topic of the conversation. How L2 learners acquire pragmatic competence, and how this competence is linked to other levels of linguistic competence is only beginning to be explored (see also Read, Reference Read2000 and Nadasdi, Mougeon and Rehner, this volume).

From a theoretical perspective it is important to study the acquisition of a variety of languages as one cannot discover universal principles or processes of Second Language Acquisition if the focus is on one language (English) only. The recent upsurge in interest in French SLA (Myles and Towell, Reference Myles and Towell2004; Prévost and Paradis, Reference Prévost and Paradis2004; Dewaele, Reference Dewaele2005; Ayoun, Reference Ayoun2007) is therefore particularly welcome. Each of these volumes makes a very important contribution to our understanding of a range of issues in the acquisition of French, and they all demonstrate the important contribution the analysis of French can make to the development of the SLA research agenda.

This special issue is intended to complement the ones mentioned above, in that it focuses specifically on lexical issues. Most of the contributions to this special issue were first presented in the context of the ESRC seminar series entitled Models and measures of vocabulary acquisition, knowledge and use: the interface between theory and applications. This seminar series was jointly organised by researchers of the M4 Applied Linguistics Network (David Malvern and Brian Richards from Reading; Paul Meara and Jim Milton from Swansea and Helmut Daller and Jeanine Treffers-Daller from UWE Bristol) at the three participating institutions between 2006 and 2007.

As Milton (Reference Milton2006) has shown, we know very little about the words that are learnt in French classes or about the vocabulary sizes of learners of French at different levels in the UK or elsewhere, and this volume aims to address this issue. In the first three articles, the focus is on the measurement of lexical richness in speech production. Housen, Bulté, Pierrard and Van Daele make an important contribution to the discussion around the theoretical constructs used in lexical aspects of second language acquisition and how these can be operationalised in empirical studies. In their theoretical framework they distinguish three levels of analysis. At the theoretical level there is the cognitive construct of lexical competence, which consists of different dimensions (width, depth etc.) and a procedural component. As lexical competence cannot be directly observed, they distinguish this from the behavioural construct of lexical proficiency (diversity, sophistication etc.), which then needs to be operationalised in statistical terms in the form of various statistical constructs (TTR, Index of Guiraud etc.). Housen et al. subsequently demonstrate how these concepts and operationalisations can be used to analyse lexical development over time in a longitudinal study among Flemish students of French in the Dutch-speaking schools in Brussels. Using a wide variety of measures, the authors show the students progress significantly in terms of lexical diversity, sophistication and productivity over a period of two years (from age 12 until age 14). The authors also call for a greater conceptual clarity as to what lexical sophistication entails. More specifically they come to the conclusion that it is not sufficient to refer to frequency lists alone in operationalisations of lexical sophistication.

This issue is taken up in the second contribution by Tidball and Treffers-Daller who focus on lexical sophistication in the oral productions of British learners of French in Higher Education. They use different operationalisations of the Advanced Guiraud (Daller, Van Hout and Treffers-Daller, Reference Daller, Van Hout and Treffers-Daller2003) to show that operationalisations based on teacher judgement are superior to those which are solely based on frequency, including a measure based on Vocabprofil, the French version of Laufer and Nation's Lexical Frequency Profile. In addition, they point to the need to address the key role of cognates in vocabulary learning as these often remain outside the scope of measures of vocabulary richness. The study confirms the results of Horst and Collins (Reference Horst and Collins2006) that learners often make better use of words in the higher frequency bands, in that they use a larger variety of the first 1000 words (k1), but do not necessarily improve significantly in their use of infrequent words. The authors therefore call for measures that further differentiate between different frequency layers among the k1 group.

In the third study, David presents a detailed analysis of the lexical richness in the speech of 80 British learners in secondary education (Years 9–13). Using a combination of a controlled active and a free active task she shows that lexical diversity as measured on the basis of D (Malvern, Richard, Chipere and Durán, Reference Malvern, Richards, Chipere and Durán2004) increases significantly from Year 9 until Year 12 (cross-sectional study), with the largest increase between Years 11 and 12. No significant difference was found between Year 12 and Year 13 in the longitudinal part of the study, which could be due to a test repetition effect. David also demonstrates that the proportion of nouns increases between Years 9 and 10 and thereafter it consistently decreases. This confirms the existence of an early noun bias in the early L2 acquisition of French. The proportion of verb types increases over the different year groups. She finally calls for a further investigation of the effect of tasks (free productive versus controlled productive) and lemmatisation on measures of lexical richness, as comparisons of D-figures between different studies are currently hampered by variation in tasks and lemmatisation methods.

In the fourth study the focus shifts from productive to receptive vocabulary. Milton's study is a follow-up on Milton (Reference Milton2006) in which he studied receptive vocabulary growth among 449 learners across all levels in a British school using a French adaptation of X_lex (Meara and Milton, Reference Meara and Milton2003). In the current study, Milton compares the vocabulary of 21 learners who had studied to ‘O’ level French between 20 and 50 years ago study with the 24 best students from an equivalent cohort that took GCSE. The comparison shows that the current GCSE students possess only a half to a third of the vocabulary knowledge of learners who took ‘O’ level in the past. This enormous decline is probably due to pressure to increase the numbers taking and passing the examination, as well as to a reduction in the classroom hours available for learning, and a change in the examination format. Milton also reports about the receptive vocabulary of 66 students taking single and joint honours French at a British university using the same testing methodology. This part of the study shows that students continue to learn words at the same kind of rate that they did at ‘A’ level; about 500 words per year. This means they end up with relatively low levels of vocabulary knowledge that make effective communication in the foreign language very difficult. Clearly there are important implications of this study for everyday practice in MFL in schools and universities in the UK.

The fifth study specifically focuses on vocabulary progress in Year 12 in the context of learner strategy training that targeted listening and writing by 150 pupils in 15 English comprehensive schools. Graham, Richards and Malvern analyse the impact of learner strategy training in eleven schools and compare the results with a comparison group of four schools. On the basis of raters' assessments of written picture descriptions, they show that the students' productive vocabulary increases significantly over the two terms. They also used a French adaptation of X_lex (Meara and Milton, Reference Meara and Milton2003) to assess receptive vocabulary. This tool was shown to be a valid measure to distinguish between students with different GCSE grades, but it was more difficult to demonstrate the existence of measurable progress in receptive vocabulary on the basis of this tool. There was an overall gain in listening, writing and productive vocabulary scores, but no group effect. This means that the additional time used on strategy teaching did not negatively affect vocabulary growth.

In the final contribution, the focus is on a dimension of vocabulary knowledge that receptive or productive measures generally ignore, namely socio-stylistic variation. Nadasdi, Mougeon and Rehner's study of lexical variation in the use of Canadian French expressions for car demonstrates how difficult it is for learners to get to grips with sociolinguistic variation in French. Among native speakers of Canadian French auto is the expression most frequently used variant, with char, machine and automobile being secondary variants whereas voiture is virtually absent from their speech. The 41 Canadian immersion students prefer the form auto, which may be due to the similarity between this form and the Spanish and Italian translation equivalents, as well as to its frequency in the teaching materials. In addition, they use the hyperformal voiture, which is very frequent in the educational input to students, but make no use of the marked informal variants machine or char. The absence of char in the input to the students (in the classroom and the textbooks) is probably a key reason for the students' preferences.

While it is difficult to assess the importance of all six articles for our understanding of lexical dimensions of language proficiency, it is clear that a key aspect of this contribution is a methodological one: the tools and measures used in this study – many of which were developed by members of the M4 Applied Linguistics Network – make it possible to carry out valid and reliable analyses of learner's knowledge across a range of national and educational contexts. The research community will also no doubt find it helpful to see how all tools and measures are evaluated in the different studies.

The studies described here also provide new detailed insights into the development of lexical knowledge among learners in schools and universities. The picture that emerges from these studies is that vocabulary learning is a neglected area in secondary and tertiary education. A further decline in standards can only be prevented if the information provided here is taken on board by all those who are keen to improve MFL teaching, that is teachers, authors of textbooks and policy makers in different countries, but in particular in the UK, i.e. the Qualifications and Curriculum Authority and the Department for Children, Schools and Families.

An area that deserves further attention from the research community is the construct validity of measures of lexical richness. While lexical knowledge is clearly multidimensional, the frequency of lexical items (types or tokens) is being used more often than anything else in the development of measures of lexical richness. This issue is particularly important for measures which claim to tap into lexical sophistication, as frequency is unlikely to the only determinant of lexical sophistication. A related problem is that valid information about the frequency of lexical items in spoken language remains difficult to obtain due to the fact that oral corpora are much smaller than written corpora. Any measures of lexical richness/sophistication that are based on frequency data from written corpora but are being applied to oral data are likely to run into problems. An important question for future research will also be to establish how cognates and units beyond the single word level such as formulaic sequences (Wray, Reference Wray2005) can be accounted for in measures of lexical richness. The aim of a special issue such as this can however not be to answer all questions. We hope that the research community will find this special issue has made a useful contribution to answering some of the questions and to pointing out how follow-up studies can contribute to developing this research agenda further.