ABSTRACT
While metalinguistic awareness (MLA) has been shown to play a crucial role in second and third language acquisition, a variety of methodological approaches point to the difficulty of operationalising the concept. The present study provides a critical analysis of MLA in relation to its measurement by means of a French-language test of metalinguistic ability. In this article, the results on the Test d’Habiletés Métalinguistiques (THAM-3) completed by 66 francophone college students in Québec are interpreted with reference to the notions of noticing and understanding. Nearly a third of the participants reached the highest level of metalinguistic analysis in the metasemantic section of the test, as opposed to only 5% in the metagrammatical part. An analysis of the coding procedures suggests that this tendency may be related to the fact that explicit reference to grammatical categories was required to attain maximal scores in the metagrammatical section, whereas metalinguistic terminology was not necessary to perform at this level in the metasemantic section. This critical view of the coding procedures used on the THAM-3 points to the need to re-evaluate the role of explicit knowledge of grammatical concepts and metalanguage in the measurement of MLA.
Acknowledgments
The author is indepted to the editors and to the anonymous reviewers who have triggered important reflections on earlier drafts of this article. Very special thanks are owed to Leila Ranta for her insightful comments and her support throughout the crucial phases of the revision process.
Disclosure statement
No potential conflict of interest was reported by the author.
Notes
1 This further mirrors findings on literacy development, namely because literacy skills were shown to be transferable between languages (e.g. Durgunoğlu, Citation2002; Schwartz, Geva, Share, & Leikin, Citation2007).
2 The rationale for using the French version is that test has nearly exclusively been administered in the test takers’ first or dominant language. The exception is El Euch (Citation2010), whose findings do, however, suggest that MLA is language-independent.
3 The test was conceived for three different age groups: MAT-1 (age 4–6 years), MAT-2 (age 9–13 years), and MAT-3 (age ≥16 years).
4 Collège d’enseignement général et professionnel (2–3 year programs prior to university).
5 Some of the students entering cégep had already taken Spanish classes in secondary school, in which case German may have been considered their fourth language. However, in line the TLA literature (e.g. Cenoz, Citation2003), the term L3 was used for any language beyond the second that was currently under observation.
6 Examples of items from each section are provided in the second part of the Results pertaining to The role of explicit knowledge of grammar.
7 Since some participants were under the age of eighteen, it was decided to exclude the possibility of misinterpretations due to lack of knowledge, which could otherwise have been mistaken for a lack of MLA.
8 Two items were excluded due to inconsistencies between the error as it appeared in the text and in the coding procedures, respectively. A third item was considered ambiguous with respect to the possible interpretations that may have served as an appropriate answer.
9 English translations of all examples are provided in Appendix.
10 Since the errors in the text are specific to the French language, an English translation would not have rendered the meaning and scope of the error. In turn, English translations seemed inappropriate.
11 Not only were most of these university students older (mean age = 24.5 years) than the participants who attended a junior college (mean age = 18 years), but their language background also differed slightly in that only 4 out of 150 had knowledge of a fourth language besides French (L1), English (L2) and Spanish (L3). Moreover, all of the participants in the present study were enrolled in a language program, whereas the others were recruited from different disciplines.
12 In the validation of the THAM-3 (Candilera et al., Citation2015, pp. 79–81), the distribution of the results of the French sample was also found to mirror that of the Italian sample (Pinto & Iliceto, Citation2007).
13 In this respect, the results resemble those of the Italian sample who attained 66% of correct L answers in the Acceptability section.
14 As outlined in the Methodology section, the original Italian version of the TAM-3 (Pinto, Citation1999) was tested on 139 senior high school students between the ages of 18 and 19 years, while the large scale validation study of the TAM-3 (Pinto & Iliceto, Citation2007) was performed on 621 Italian speakers from various backgrounds: 281 senior high school students (age: 16–18 years), 281 university students (age: 19–30 years), 67 working adults (age: 31–58 years). Since no significant differences were found between mean scores for each group, only the total mean scores are reported for each subtest. Finally, the Spanish THAM-3 (Lasagabaster, Merino, & Pinto, Citation2015) was administered to 148 university students (age: 18–37 years).
15 A presumed lack of explicit knowledge of grammar may further be related to the fact that the test was taken in the learners’ L1. Even if metalinguistic abilities were shown to transfer between languages (see above), it may be useful to verify whether comparably low ML scores on the metagrammatical part would be reproduced in an instructed L2. However, given the learners’ educational background (post-secondary language studies), no major difference would be expected, especially since explicit instruction of morphosyntactic features is typical in French (L1) courses (Ranta, 2017, personal communication).
Additional information
Funding
Notes on contributors
Nina Woll
Nina Woll is an Assistant Professor at the Department of Modern Languages of the University of Quebec in Trois-Rivières (Canada). Her research interests include psycholinguistic processes in the acquisition of additional languages, specifically with regards to the development of metalinguistic and crosslinguistic awareness in instructed settings.