Starting age overshadowed: Pfenninger and Singleton (2018)

Here’s a great open access paper in Language Learning on the vexed question of early foreign language (FL) teaching, once again calling into question the “early start” policies favoured by parents and policymakers. Here are the bibliographic details:

Starting Age Overshadowed: The Primacy of Differential Environmental and Family Support Effects on Second Language Attainment in an Instructional Context

Simone E. Pfenninger, University of Salzburg

David Singleton, University of Pannonia/Trinity College Dublin

Photo by Marvin Ronsdorf on Unsplash

Keywords age factor; multilingualism; bilingual advantage; young learners; early foreign language learning

The study looks at EFL learning in primary and secondary schools in a German-speaking region of Switzerland with a view to taking home language and other language experience into account, unlike previous early start research. It compares “monolingual” pupils who spoke German at home with “bilinguals” with a different home language, including recent migrants, and includes age of onset (AO) of English learning as well as home language variables, including family support. Since bilingualism is thought to influence L3 learning differentially for early (simultaneous) and late (sequential) bilinguals, and to include a role for literacy, these variables are also included in the study. (Access the full paper here – no paywall.)

The researchers followed pupils over 5 years in order to investigate the effect of age of onset on L2 English proficiency and the role of family support on bilingual and biliteracy development. Participants were some 600 pupils starting secondary school (age 13): 300 early learners, who began EFL as primary school pupils aged 8, and a similar number of late learners, secondary school pupils just starting English. Within these groups, four subdivisions were made

  1. 200 monolinguals with Swiss German as home language and Standard German at school (literacy); they also had French L2
  2. 144 biliterate simultaneous bilinguals
  3. 107 simultaneous bilinguals literate only in German
  4. 185 sequential bilinguals, literate only in German

Pupils were thus learning EFL as an L3 and L2 of literacy in the case of the monolinguals, and an L4 and L3 of literacy for the second group, L2 of literacy for the third and fourth. The home languages included Spanish, Portuguese, Croatian, Serbian, Albanian, Arabic, and Italian and were balanced across early and late groups. Bilingualism was reported, not tested, and parental biodata confirmed no socio-economic discrepancies between bilingual and monolingual groups.

Sociolinguistic context was measured by questionnaire, covering

  • books in the home
  • parental support (indirect = English proficiency, direct = English homework help)
  • parental attitudes regarding English

Six EFL proficiency measures were used, all tested and piloted in 2008 with some 50 pupils:

  • listening comprehension (B2)
  • receptive vocabulary
  • productive vocabulary
  • written lexical richness, complexity/fluency/accuracy
  • oral lexical richness, complexity/fluency/accuracy
  • grammaticality judgement

Here’s what Pfenninger and Singleton found:

In all four groups, an earlier AO was a significant predictor of 60% of the tested FL skills at the beginning of secondary school […] Specifically, in each group, the early starters outperformed the late starters in receptive vocabulary, written lexical richness, written fluency, oral lexical richness, oral accuracy, and written grammaticality judgments.

The late group performed less well on all those measures at the start of secondary education, but parity on written and oral complexity, written accuracy and oral fluency was reached after only six months, and by the end of obligatory schooling, 5 years later, no difference remained outside the lexical richness measure. These results show that the late learners made much more progress than the early group over the five years of secondary schooling.

Bilingualism had no effect on results – the monolingual group did not perform differently from the various bilingual groups. However, there was an effect for biliteracy: the early starters among this group outperformed their late-starting counterparts in productive and receptive vocabulary, written lexical richness and fluency, and oral complexity. The authors see this pattern as “reminiscent of age effects in naturalistic settings” (p. 221).

Concerning the second research question regarding environmental effects, it turned out that once data concerning parental attitudes and behaviours were included in the analyses, these variables proved more important predictors of learning outcomes than age of onset. Books at home, parental EFL proficiency and attitudes had a strong influence on results, and these variables also interacted with bilingualism

In the authors’ words,

bilingualism was effective only in combination with in/direct parental support and positive parental attitudes

The same result was found for biliteracy:

biliteracy was always better than monoliteracy, and family involvement/encouragement was always better than no involvement, as the main effects establish. However, biliteracy together with environmental support was particularly effective, compared to either bilingualism or environmental support alone.

The biliteracy group had much more parental support from parents with consistently more positive attitudes to FL learning and multilingualism.

[The authors conclude the results section with a caution regarding random effects, of which they found a good deal: tests like ANOVA, without the cluster-randomised design used here, will be vulnerable to Type-I errors and may “suggest age effects which are not really there” (p. 224)]

The study concludes that age effects are less important in the acquisition of a FL at school than proponents of an early start often think. It showed that any advantage which early starters had over late starters was largely erased by the start of secondary school, and completely washed out five years later, by age 17-18. Bilingual and monolingual pupils in the Swiss EFL classes learned English to the same standard. Only bilingual children who were also literate in both languages maintained the advantage of an early start compared with late-starting counterparts at age 18. The authors entertain hypotheses regarding the cognitive advantage of bilingualism and biliteracy, but prefer an explanation based on the home environment, particularly parental support. They suggest intense parental involvement in a child’s FL learning mirrors naturalistic L2 acquisition and provides “a similar sense of this learning meeting family goals and of its being integrated into family life.”

The authors conclude with a call for more multidisciplinary research to accommodate the sociolinguistic dimensions uncovered in this study. They suggest pedagogy, learner motivation, and attitudes of teachers and peers as factors which warrant further research.

Communicative competence in languages for specific purposes

Here’s an open access article I just published in Language Education and Assessment. It has a history of different interpretations of the notion of communicative competence, looks at the concept in relation to the CEFR in terms of proficiency testing and learner profiles, and then makes use of recent work in specific purpose testing to propose an updated definition of communicative competence and a framework for using it in LSP teaching and testing.

Whyte, S. (2019). Revisiting communicative competence in the teaching and assessment of language for specific purposes. Language Education and Assessment.

The conclusion argues for

an expanded view of communicative competence which is more faithful to Hymes’ (1972) original conception and reflects a number of advances in L2 research over the intervening five decades. One is the realisation that native-speaker norms are not the most relevant in LSP: formal linguistic accuracy is of little importance in any real-world context outside the language classroom. Another finding is that indigenous criteria for the assessment of communicative competence in both L1 and L2 contexts can produce categories of language use which offer reliable and valid indicators of speakers’ performances. These criteria are comparable across disciplines and across discourse events and task types yet show little overlap with the linguistic criteria used in traditional EAP or LSP tests. This finding supports the view that our recent interpretations of communicative competence in language testing have failed to take the wider dimensions of this notion into account. I have argued that a middle ground between discrete formal linguistic criteria and broad extralinguistic factors is to be found in Hymes’ original conceptions of both knowledge and ability for use, and that these dimensions can be usefully explored by researching interactional patterns and discursive practices in LSP communication.

And this is how I apply Hymes’ notion of communicative competence in LSP:

Doctoral studies at French universities

Many faculty at French universities receive frequent applications from international students wishing to enroll in a doctoral programme in France. Since different countries have different administrative procedures and also academic customs, here’s a short guide for students in humanities and social sciences at universities in France.

If you want to enrol in doctoral studies at a French university, you need to find an adviser to supervise your dissertation and a research laboratory to accept your application.

  1. You should also look at university websites to find research laboratories and faculty.
    1. You need someone who is professeur or maîtres de conférences HDR (meaning habilités à diriger des recherches). Sometimes maîtres de conférences who are not HDR co-supervise with qualified colleagues.
    2. You should also consider how your research project fits with the lab’s research priorities and with faculty members’ competences and research interests.
  2. Then you need a plan for your doctoral dissertation. This is a 3-5 page document with
    1. a proposed title
    2. a review of research already conducted in the area,
    3. an outline of your proposed contribution and a bibliography.

Your aim should be to show that you have prior experience in your chosen domain and that your proposed topic falls in an area of interest to your potential adviser and their research lab.

If this seems difficult to do immediately, students often consider enrolling in a masters programme first to improve their knowledge and skills in their proposed field, to learn about key terms and current research in the French context, and make contact with labs and professors who might be willing to supervise your doctoral project. It is often easier for professors to work students they already know and who have documented experience gained in local or national masters programmes.