Age of Acquisition and Language Exposure The Effectiveness of French Language Programs in Canadian Schools

by Kirsten Paula

French Immersion and Core French programs are widespread within Canadian schools and are mandatory to varying degrees across the country. The critiques and scepticism around the programs is unfortunately also quite widespread. This review will attempt to discover linguistic developmental factors to address these concerns. Wariness towards L2 learning can be dismissed through original research to ensure that learning a second language does not impede the acquisition of English literacy, numeracy, or other cognitive development as compared to monolingual peers at school (Barik & Swain 1976). The same factors of acquisitional success have been examined in simultaneous environments, that is, in contexts where both languages are learned while still in the critical period of language learning, near infancy or toddlerhood (the exact age is still debated). The underpinnings of simultaneous acquisition have been explored by Genesee (2007) and Volterra and Taeshner (2007), both of whom agree that, with simultaneous acquisition, the rate of acquisition for each language is driven by language specific input. However, overall acquisition of both languages is not threatened. Children who simultaneously acquire two or more languages were shown to acquire lexical and syntactic knowledge in a predictable integrated process, which has been hypothesized to be the key in bilingual success.

Late life learners must go through two separate acquisition processes instead of a singular integrated one. The effects of this change can be seen in studies that investigate monolingual, native bilingual, early bilingual and late bilingual processing. Native bilinguals were defined as learners who acquired both languages from birth, early bilinguals started monolingually and had the second language introduced later in life usually at least past one year old. Areas where differences in language processing can be explored include: stuttering, non-speech noise distraction, speech comprehension in acoustically difficult environments, and control mechanisms in L3s (Mamdoh & Gomaa, 2015; Martin, Strijkers, Santesteban, Escera, Hartsuiker, & Costa, 2013; Ortiz-Mantilla , Choudhury, Alvarez, & Benasich (2010), Shi (2010)). The former studies support that earlier acquisition and more input will result in higher proficienty. This paper also consider the results of other evaluations within Immersion and Core French programs, which refute the idea that in-school immersion participants inherently acquire better language proficiency overall (Day & Shapson 1988, Lappin-Fortin 2014).

Methods

When selecting articles for this review paper the goal was to collect a diverse and interdisciplinary group of articles. Applied linguistics, pedagogy, and neurolinguistic perspectives were all important in gaining an accurate view of L2 acquisition and Canadian French Language programs. A variety of publication dates ranging from initial research in the field to modern publications were selected, in both electronic and printed formats. The hard copy sources were found through the course textbook and recommended texts from the professor. The online database used to find articles was NovaNet with search terms including: bilingual, early, late, immersion, French, and English. The search was refined to peer-viewed journals with full text availability.

Results

Selective Or Non-Selective Acquisition

Early publications related to second language learning were concerned with the selective or non-selective nature of bilingual language acquisition. Volterra and Taeschner (1978) established their selective view. Genesee (2007) followed this initial research with a paper supporting a non-selective process within bilingual production. The two studies centre on the acquisition of lexical and syntactic information. Volterra and Taeschner (1978) used longitudinal data collected from two subjects between the ages of one and four years old. They developed a theory whereby two languages are acquired simultaneously in three stages: (1) initial unification of both lexical and syntactic substytems; (2) differentiation of the lexicon but continued unification of syntax; (3) finally differentiation of both the lexicon and syntax (2007, p. 304). Their theory predicts that the reduction of language mixing is due to the differentiation of the languages themselves from a shared storage system in their brains.

Genesee (2007), on the other hand, emphasized language mixing and code switching as being something that language learners always have access to, and as tools that are socially acquired along with linguistic knowledge. He highlights the research that has been done on infants, which demonstrates that they have rather developed capacities for language perception based on speech sounds (Jusczyk, 1982). He reasons that this is adequate for learners in order for them to be able to differentiate the languages before they even speak it. Genesee believes that “learners acquire sociolinguistic competence rather than psycholinguistic separation of the language systems” (p. 325).

L2 Relative Proficiency Levels

Further research into how early and late bilingualism could affect a language selection mechanism has been carried out by Martin et al. (2013) in an ERP study with Spanish-Catalan-English trilinguals and Spanish-Catalan bilinguals. Group One had English as an early L2 and Catalan as a late L2. Group Two had both Catalan and English as languages acquired later in life. Finally, Group Three consisted of early Spanish-Catalan bilinguals. These three groups were evaluated for their relative strengths in L2 proficiency with age of acquisition as the dominant variable. They all completed a naming task that required them to identify pictures alternately in Spanish and Catalan. In their statistical analysis, they controlled for mean word frequency, number of letters, number of orthographic and phonological neighbourhoods, number of phonemes and syllables, familiarity, concreteness, and imageability — none of which reached significance (Martin et al., 2013). Early bilinguals from Group One had symmetric switching costs (response latencies were effected equally while switching between all languages) while late bilinguals in group two had asymmetric ones (response latencies were much slower when switching from L3 to L1). They discovered that the N200 ERP, in the frontal region at around 290ms, was mostly affected, which is typically related to cognitive efforts related to “inhibitory processes…conflict monitoring…response selection monitoring and cognitive control.” The other significant results were found with the LPC in the parietal region, which had a hemisphere effect with a higher amplitude in the left hemisphere and a significant Type of Trial × Language of response interaction amplitude at 490ms (Martin et al., 2013). Late bilinguals had higher N200 and LPC amplitude than early bilinguals suggesting that language selection and control was more cognitively taxing. They did not discover a significant difference between high proficiency bilinguals in Group Three and moderately proficient early trilinguals in Group One. Martin and colleagues conclude that this is likely due to some sort of cognitive control function that develops with early bilinguals during early simultaneous acquisition. Once developed early on, the relative strength of the L2 does not make a significant difference.

Another ERP study related to attention investigates responses to complex tones. The subjects were 41 Spanish monolinguals and Spanish-English bilinguals who varied in age of acquisition and years of L2 experience, aged between 20 and 36 years. All participants completed a multiple-deviant oddball paradigm task in which the researchers manipulated four variables including duration, frequency, silent gap, and frequency modulation (Oritz-Mantilla et al 2010.) Mismatch negativity, which indicated detecting variation in a stimulant, had peak amplitudes between 100-250ms and was measured around the frontal, fronto-central, and central regions. The P3a peaked following the MMN, signalling involuntary attention switching, with significant results in the central midline regions. The final measure was Late Negativity following the P3a. LN signalled a reorientation of attention and was measured in the same fronto-central region around 500ms. Cognitive processing energy varied as a function of age of acquisition and length of experience (or input) within a language. All three of the measures displayed more extreme peak amplitudes the later the acquisition, making the MMN more negative, the P3a more positive, and the LN more negative as well.(Ortiz-Mantilla et al 2010). They hypothesized this could be caused by a combination of two things. Firstly, a cognitive auditory attention mechanism in the brain being stronger for early acquisition participants, because they have spent a longer time automatizing the process of detecting speech sound significance in more than one language. Secondly, late learners may still be in an “active learning mode” which makes them more prone to analyzing non-speech sounds for significance because they are less familiar with the typical linguistic cues for their languages.

Shi (2010) investigated auditory comprehension. They also used a more continuum-like scale for rating early to late acquisition of their participants like Oritz-Mantilla and colleges, but still divided participants into larger categories. Forty participants classed as monolinguals, native bilinguals and early, late, or very late bilinguals were tested in comprehension by identifying the last word in a series of acoustically degraded sentences (SPIN test) that varied in degrees of : “noise (+6 vs. 0 dB signal-to-noise ratio), reverberation (1.2 vs. 3.6 s reverberation time), and context ( high vs. low predictability)” (Shi 2010). Reverberation and context were found to be the most influential factors across all groups. Monolinguals performed the best across all conditions, and late bilinguals the worst with a mean in the 40th percentile. The most interesting finding was that early bilinguals and native bilinguals’ use of context was still significantly affected by reverberation and noise, unlike monolinguals who consistently were able to use contextual cues.

In addition to comprehension, fluent production is also an important variable in L2 proficiency. Mamdoh and Gomaa 2015 examined the correlation between stuttering severity and second languages with Arabic-English bilinguals. They label them as late bilinguals with acquisition occurring in kindergarten around age 4. Their study occurred when the subjects were around age 10, with six years of bilingual education. It was found that stuttering severity significantly increased from their native L1 to their L2. Physical or emotional ailments were ruled as confounding variables, leading Mamdoh and Gomaa to conclude that the stuttering increases were due to the late onset of L2 acquisition. These results show production-based deficits. They did not rule out cognitive processing, or a lack of cognitive mechanism strength as a source of production deficits.

Canadian French Immersion Context

The following studies have been selected to illustrate real-world applications of L2 acquisition within Canada and as a way to observe the relative levels of instruction and proficiency students receive. Barik & Swain is a three year review of French Immersion programs in Canada, between the years of 1970-73. Its main concerns were evaluating if there were any detrimental effects to English skills, math skills, or cognitive development and that there were significant positive effects of immersion French over core French. Overall FI students demonstrated normal literacy, numeracy, and cognitive development (comparative to English stream children in school) and performed better in French classes. They described the program as entirely French in kindergarten with spontaneous English speech, entirely French with the introduction of English class in grade one, and slightly more English classes in grade three trending toward a 50/50 split.

Day and Shapson did a 1988 study on 7th graders in three different school districts in BC that offer early and late immersion programs. They differ in age of commencement 5-6 years for early and 12-14 for late and the total number of years spent in the program, 12-13 for late and 6-7 for early. Students were evaluated on grammatical, attitudinal, and auditory comprehension. They discovered that early immersion students performed significantly better across the board for French auditory comprehension, but only district C also performed significantly better with French Language Arts. Attitudinal responses were mixed between the late and early immersion groups about their own French competencies and willingness to use French. The three districts did vary in percentage of teaching time in French, from 50-70%.

In Thordardottir (2014) they demonstrate the importance of input vis-a-vis acquisition with Montreal pre-schoolers. He found a positive non-linear correlation between levels of input from all contexts and scores on grammar and vocabulary tests. A non-word repetition test was used to control for any children with language learning impairments. When learners have divided linguistic environments, for example different languages at home and at school, it limits the total language input they can receive for each language. Thus his subjets fit into three different categories: 40-60% split of French/English, more French with 5-39% English, and more English with 61-94% total (Thordardottir 2014). The 3 year old group of bilingual children had lower levels of vocabulary than monolingual children overall. It was determined that a certain percentage of exposure was needed to reach the maximum or native fluency scores by age 5 (above those thresholds there was little improvement). 40-60% input was needed for receptive vocabulary and 70% was needed for expressive vocabulary. With grammar Thorardottier found that the 3 year old performed within normal monolingual ranges but the 5 year old had grammar scores that correlated positively with input level, revealing poorer grammar in low input cases. French lacks the Optional Infinitive Stage, making it more difficult to accurately test and compare French and English fluency in children. This theory proposed an alternate explanation from Volterra or Genesee for a lengthier acquisition process that can be just as successful. This is a rather non-selective theory that separates competencies and input into distinct categories.

Kerry Lappin-Fortin (2014) completed a more modern study examining 225 core French and French immersion graduates after their first year of university French education. She analyzed a corpus of diagnostic questionnaire writing samples for various indicators of grammatical and syntactic accuracy, which she believes to be the principal weakness of french language proficieny for French Immersion students (Lappin-Fortin 2014). The indicators included evaluations of verb conjugation, word count, phrase subordination and avoir+âge constructions amongst other factors. Three groups were formed from the sample data participants: French immersion (FI) students, Core French (CF) students, and Core French Plus (CF+) students who has also completed a study abroad experience in a francophone location. CF+ students performed the highest in accuracy scores, longest output, and most subordinate/relative clauses. Other accuracy scores failed to show many significant differences in FI’s favour, notably with CF students out-performing FI students with avoir+âge conjugations three times more successfully. FI students did perform significantly better on some tasks, but mostly when CF+ subjects were removed from the CF group. Lappin-Fortin (2014) suggests that this stems from fossilization of Anglicisms and uncorrected grammatical errors made in prolonged French immersion environments.

Discussion

Quantifying “Early”

Age of acquisition was a variable in almost all of the studies cited above and has shown to have many significant impacts on L2 acquisition. Four categories of learners were found in most studies as well: monolinguals who spoke solely one language, bilinguals from birth or simultaneous bilinguals, early bilinguals and late bilinguals. Unfortunately, there was a lot of discrepancy between how “early” and “late” was defined within each study. In Day & Shapson (1988), “early” immersion could start as late as age 6 or 7, but in Mamdoh and Gomaa’s (2015) paper age 4 was labelled a late bilingual environment. There did not seem to be a general age or even milestone event that could consistently predict how each study would divide up their participants. It is more effective to implement a continuous scale for age of acquisition to collect precise and non-bias data.

The high discrepancies between what “early” and “late” bilingualism entails easily causes major problems for policy makers and educators trying to develop teaching methods that better align with linguistic research. More acquisitional research needs to be done to establish these boundaries so they can become more helpful landmarks that can be operationalized in other fields. These adjustments will greatly improve the validity and generalizability of collected data.

With more accurate data researchers can determine the exact age and/or developmental stage in which attention control and language selection mechanisms are still able to be elegantly incorporated into a child’s language processing, as discussed in Genesee (2007) and Martin et al. (2013). Both studies point towards a critical window where this mechanism can be learned and then strengthened through repetition into a bilinguals’ normal processing. But what is not discovered is a certain age, stage of acquisition or stage of cognitive development that would necessarily inhibit this type of attentional control mechanism to manifest itself through later simultaneous language learning. Though people seem to be able to acquire languages at nearly any age, mechanisms like this may be missed if educational systems fail to incorporate the target language early enough into potential bilinguals academic lives. Effectiveness in code switching (Martin et al., 2013) and refined speech sound identification (Ortiz-Mantilla et al., 2010) have been shown to be fragile skills in regards to early acquisition.

Monolinguals as a Standard?

And finally it is worth questioning whether the baseline statistics and performance comparisons should come from a monolingual group of participants, as they do in nearly all studies discussed above. Monolinguals are being used as a control group for “normal” acquisition and competency scores. This is debatable especially considering results like the ones in Shi’s (2010) study that revealed even native bilinguals differed from monolingual speakers when trying to solve the SPIN test under high reverberation and low context conditions. Monolinguals still had a slight advantage at reading minute linguistic and contextual cues. Also, according to Thordardottir (2014) a child needs 70% exposure to a language to gain fluent expressive and receptive vocabulary. Based on simple math it would be impossible for a native bilingual child to receive 140% language exposure each day. Using monolinguals as standard reveals an ideological belief that bilinguals are, or at least should be, able to function as two independent monolingual speakers. It appears in all of the methodological choices of the linguistic and pedagogic researchers alike (excluding Lappin-Fortin (2014) who excluded francophone participants). Despite the devotion to a duel monolingualism ideology, this ideal appears to be increasingly unattainable based on the actual data retrieved. Early acquisition coupled with high input should be able to bring students to near native competencies, but the full acquisition of native performance remains open to debate and further research.

Conclusion

In conclusion, earlier immersion and exposure will likely be beneficial in gaining fine-tuned accuracy and comprehension within a language. Both age of acquisition and overall exposure have been shown to be significant and separate variables predicting overall language acquisition success (Martin et al. 2014, Ortiz-Mantilla et al. 2010, Shi 2010, Thordardttir 2014, Mamdoh & Gomaa 2015). Therefore, by providing earlier immersion it may be possible to increase L1-like performance by introducing a second language earlier on in a child’s cognitive development. This could provide cognitive mechanism involved in bilingualism more time to be strengthened through repetition. These are possibly the underlying skills that produce confident and fluent-sounding bilinguals in oral speech, which is where current programs show the most weakness. Through further linguistic research exact milestones in development can be identified for academic policies to be based around.

Early French programs will also help by allowing the child more opportunities to be exposed to their L2. Simply put, the current programs do not provide adequate input to develop competencies effectively. That being said, the benefits of earlier immersion programs will only be truly achieved if stricter policies around pedagogy styles are employed that avoid the fossilization of “French immersion French” (Lappin-Fortin 2014) where ungrammatical/anglicised constructions are able to persist within the classroom. These ungrammatical artefacts are extremely hard for bilinguals to unlearn later in life. If Canada wants to develop their students to perform closer to simultaneous acquisition status, both cognitive skills to improve oral fluency and grammatical skills to increase accuracy will need to be reinforced.

Bibliography

Barik, H., & Swain, M. (1976). Language -- Bilingual Programs. Three-Year Evaluation of a Large Scale Early Grade French Immersion Program: The Ottawa Study. Journal Of Learning Disabilities, 9(7), 429-430. http://dx.doi.org/10.1177/002221947600900707

Day, E., & Shapson, S. (1988). A Comparison Study of Early and Late French Immersion Programs in British Columbia. Canadian Journal Of Education / Revue Canadienne De L'éducation, 13(2), 290. http://dx.doi.org/10.2307/1494957

Genesee, F. (2007). Early bilingual language development: one language or two? In L. Wei, The Bilingualism Reader (2nd ed., pp. 320-335). New York, NY: Routledge.

Lappin-Fortin, K. (2014). Comparing Written Competency in Core French and French Immersion Graduates. Canadian Journal Of Applied Linguistics, 17(2), 91-112.

Mamdoh, H., & Gomaa, M. (2015). Assessment of Severity of Stuttering in Native Versus Foreign Language in Secondary (Late) Bilingual Children. Indian J Otolaryngol Head Neck Surg, 67(2), 132-134. http://dx.doi.org/10.1007/s12070-015-0850-7

Martin, C., Strijkers, K., Santesteban, M., Escera, C., Hartsuiker, R., & Costa, A. (2013). The impact of early bilingualism on controlling a language learned late: an ERP study. Frontiers In Psychology,4. http://dx.doi.org/10.3389/fpsyg.2013.00815

Ortiz-Mantilla, S., Choudhury, N., Alvarez, B., & Benasich, A. (2010). Involuntary switching of attention mediates differences in event-related responses to complex tones between early and late Spanish–English bilinguals. Brain Research, 1362, 78-92. http://dx.doi.org/10.1016/j.brainres.2010.09.031

Shi, L. (2010). Perception of Acoustically Degraded Sentences in Bilingual Listeners Who Differ in Age of English Acquisition. J Speech Lang Hear Res, 53(4), 821. http://dx.doi.org/10.1044/1092-4388(2010/09-0081)

Thordardottir, E. (2014). The typical development of simultaneous bilinguals: Vocabulary, morphosyntax and language processing in two age groups of Montreal preschoolers. In T. Gruter & J. Paradis, Input and Experience in Bilingual Development (13th ed., pp. 141-160). Philadelphia PA: John Benjamins North America.

Volterra, V., & Taeschner, T. (2007). The acquisition and development of language in bilingual children. In L. Wei, The Bilingualism Reader (2nd ed., pp. 303-319). New York, NY: Routledge.

Made with Adobe Slate

Make your words and images move.

Get Slate

Report Abuse

If you feel that this video content violates the Adobe Terms of Use, you may report this content by filling out this quick form.

To report a Copyright Violation, please follow Section 17 in the Terms of Use.