Learning to Read in Multiple Languages: A Study Exploring Allophone Students’ Reading Development in French Immersion

This article describes a two-year study of the French and English reading development of seven elementary French immersion (FI) students who spoke a home language that is neither English nor French. Given the critical role of literacy in school success and the growing number of third language (L3) learners entering FI, this study focused on L3 learners’ reading experiences. Standardized reading measures were administered in English and in French and think-aloud protocols and interviews were conducted with students. Results suggest that L3 students are similar to, if not stronger than, their bilingual peers with respect to English and French reading ability. They also relied on their knowledge of other languages to support French reading development and evidence of metalinguistic and metacognitive insights is presented. A number of classroom implications for teaching reading in diverse FI classrooms are included.


Introduction
It has been estimated that over 75 % of the world's population speaks more than one language and a large number of this population reads multiple languages. For many students around the world, learning an additional language is done through a languagebased immersion program offered at school. For example, in Finland, Hungary and Australia, immersion programs provide an opportunity for students to 'immerse' themselves in the learning of a minority language such as Swedish and French. In Spain, Wales, and New Zealand, immersion programs are offered in an attempt to revitalize the threatened Catalan, Welsh and Maori languages respectively, while in Hong Kong and Singapore, English immersion enables students to learn English (Johnson & Swain, 1997;Kristmanson & Dicks, 2015).
In Canada, the early French immersion program finds its roots in St. Lambert, Quebec, where the program was first developed in the 1960s to provide Anglophone students with an opportunity to learn French and become bilingual (Rebuffot, 1993). In French immersion (FI), students learn subjects such as sciences, language arts, and mathematics within a rich French second language environment (Genesee, 2007a). Students learn to speak, read, and write in French in their language arts classes and also develop French proficiency through the learning of subjects in French (Lambert & Tucker, 1972). French immersion has continued to grow in popularity, becoming an integral part of Canada's educational landscape. At the time this study took place, 350,000 students were enrolled in the program (Canadian Parents for French, 2010. Moreover, in recent years, the FI program has increased in diversity, with more and more students from various linguistic backgrounds enrolling in the program.
Over the last several years, Canada has continued to see a steady growth in immigration. Among the G8 countries, it has the highest proportion of foreign-born immigrants, who in 2001 made up 20.6% of the total Canadian population (Statistics Canada, 2011). Between 2006 and 2011, Canada welcomed over one million newcomers, most originating from Asian, Middle Eastern, European, African, Caribbean, Central and Southern American countries. With such demographic shifts away from traditionally English-speaking countries of origins, the percentage of allophones (i.e., Canadians reporting a mother tongue other than French or English) now hovers at 20%, up from 17. 6% in 2004(Statistics Canada, 2011. Within the educational context, this increase in immigration means that more and more children are entering Canadian schools speaking a minority language at home and learning English as an additional language. What is more, many allophones enrol in French second language programs, including FI programs (Mady, 2007;Mady & Turnbull, 2007). In such cases, students are in the process of acquiring and developing literacy skills in three languages (their home language, English, and French). In fact, the participants in this study are allophones who are learning French as a third language. Consequently, we will refer to them as "L3 learners".
The positive effects of FI for English-speaking students are well documented (Genesee, 1987(Genesee, , 2007a(Genesee, , 2007bSwain & Lapkin, 1982;Turnbull, Lapkin & Hart, 2001;Wiss, 1987). However, we know very little about how students learning literacy skills in three languages progress in FI. Given the growing number of such learners entering FI and sparse research in this growing field, this study set out to gain a better understanding of L3 learners' reading experiences when enrolled in FI.

Theoretical Constructs
This study adopts an interactive approach in understanding the act of reading, which stipulates that it is the interaction between top-down and bottom-up processes that shapes comprehension (Rumelhart, 2006). Bottom-up reading models focus on lowsensory representations such as letters, syllables, and word recognition as the building blocks for overall text comprehension (Gough, 1972), while top-down reading models view reading as being grounded in what the learner brings to the reading process. In this view, text comprehension starts even before readers start looking at the words and readers draw upon their experiences and personal knowledge to make hypotheses about the content of the text (Cornaire, 1991;Grabe;. The process of identifying letters and syllables is only activated when required (Giasson & Theriault, 1983). Interactive reading models favors bottom-up and top-down processes that work simultaneously and interactively to support one another. Both sets of skills are available to the reader, who is processing and deciphering text (Rumelhart, 2006). Using an interactive model of reading, this study set out to examine students' early literacy skills (i.e. letter naming, word segmentation, phoneme identification) as well as reading accuracy, fluency, and comprehension (Fountas & Pinnell, 2001).
This study was also informed by intersecting theoretical understandings of crosslinguistic transfer, language interdependence, and multiliteracies. The concept of crosslinguistic transfer has long been a key underpinning of second language research and provides this study a necessary framework to examine children learning to read in English and in French. Previously defined narrowly as the "potential influence that one language has on another language" (Kohnert, 2008, p. 77), transfer is now understood as a learner's ability to utilize already acquired resources when learning a new language. Students' prior language learning and experiences can be viewed as "a reservoir of knowledge, skills, and abilities that is available when learning literacy in a new language" (Koda, 2008, p. 71). For many allophones, much prior literacy knowledge will be encoded in either their first-or second-language. Processes involved in cross-linguistic transfer can be understood through the Interdependence Hypothesis (Cummins, 1984;Cummins, 2000). Cummins proposes that, with enough exposure and motivation, transfer will take place from a learner's dominant language to the second language. This is because even though surface features such as grammar and vocabulary of different languages may differ considerably, these languages function as one single integrated process in the brain. This single integrated process, referred to as the Common Underlying Proficiency (CUP), mitigates the need to learn literacy and cognitive skills anew for each new language. Moreover, it supports the claim that simultaneously learning two languages poses no problems for the learner (Cummins, 2000;Edwards, 2009). When applied to reading, this hypothesis suggests that, "learning to read is accomplished only once, and that after learners have matured in their ability to read in their first language, the combined L1 linguistic and cognitive resources transfer to the second language and do not need to be relearned" (Grabe, 2009, p. 146). Explicitly teaching for transfer of skills and concepts from one language to another is supportive of biliteracy development (Cummins, 2001).
Multiliteracies pedagogy re-envisions traditional views of literacy that tend restrict the teaching and learning of reading and writing to "formalized, monolingual, monocultural, and rule-governed forms of language" (New London Group, 1996, p. 61). The scope of literacy pedagogy is broadened under a mutiliteracies framework, not only with respect to promoting multi-modals ways of making meaning and communicating, but accounting for our culturally and linguistically diverse and globally connected societies. This research was particularly concerned with the latter part of this definition. Students bring cultural knowledge and language abilities to new language learning environments. Together, theoretical understandings of cross-linguistic transfer, language interdependence, and mutiliteracies present these knowledges and abilities as important resources in fostering academic engagement.

Reading Strategies
Readers draw upon various sources of information (i.e. meaning, language structures, and phonological and visual information) as well as strategies in order to monitor their reading, search for information, and correct their reading (Fountas & Pinnell, 2001). When readers read increasingly complex texts, they call upon strategies to help them read (Koda, 2005). Students with highly developed metacognitive skills have more strategic control over the reading process and understand that the effectiveness of strategies is dependent on particular reading situations (Koda, 2005;Saint-Laurent, Vésina, & Trépanier, 1995). They use strategies appropriately in combination with one another, and through their use regulate the reading process (Grabe, 2009;Pinnell & Scharer, 1987;Van Grunderbeeck, 1994).

First-and Second-Language Reading Development
Differences among students' first-language reading experiences can impact second-language reading development. Second language learners who can already read in another language have a wide range of reading strategies at their disposal. They can draw from these previously acquired strategies to help them read additional languages (Koda, 2005). As is the case for first-language readers, being strategic plays an important role when learning to read in another language. Geva and Clifton (1994) found that FI students who had highly developed reading skills in one language (i.e. read fluently, accurately, and with comprehension) were also strong readers in another language. Strong second-language readers are able to effectively exploit previously acquired first-language knowledge, skills, and abilities (Chamot & O'Malley, 1994). This accentuates second-language reading development because "prior literacy-learning experience fosters an explicit understanding of what is to be accomplished in the task, and this, in turn, may expedite the process by allowing leaners to be more reflective and strategic" (Koda, 2008, p. 74).

Third-Language Reading Development
L3 learners possess many of the same characteristics as second-language learners. They are able to draw from previously acquired literacy skills. However, because they are already in the process of learning two languages, they have "…more language experience at their disposal" (Cenoz, 2003, p. 71). Consequently, they may present some advantages over their bilingual and monolingual peers, including highly developed learning strategies and metalinguistic awareness (Cenoz, 2003). In a review of the literature on multilingual acquisition, Cenoz and Genesee (1998) found that the process of acquiring a second language facilitates the process of learning a third language (see Bild & Swain, 1989;Cenoz & Valencia, 1994). Additionally, the authors of a French immersion study in Canada found that literacy abilities in one's first language had a positive effect on learning French as a third language in a delayed (Grade 4 entry) French immersion program (Swain, Lapkin, Rowen, & Hart, 1990).

Context and Participants
In most Canadian jurisdictions, like the original St. Lambert model, the early FI program begins in Kindergarten or Grade 1. In New Brunswick, however, where the present study was conducted, the program only began in Grade 3. Within this context, all students received their instruction in English starting in Kindergarten until the end of Grade 2, at which point, they chose to either enrol in French immersion or continue in English. For some language learners, this context means that they are learning French as a third-language. They speak a home language that is neither French nor English, and received formal English literacy instruction from Kindergarten to Grade 2. This unique context provided an opportunity to explore L3 students' French reading development with the understanding that they were entering FI having had three years of formal English reading instruction, in addition to the English they would have acquired in informal settings.
Students from three elementary schools, who were transitioning from Grade 2 English to Grade 3 FI, were invited to participate in a research project. In total, 56 students (20 boys, 36 girls) participated in the study that examined students' English and French reading development, as well as their knowledge and use of reading strategies in both languages (Bourgoin & Dicks, 2013), the role of cross linguistic transfer, and early literacy predictors (Bourgoin, 2014). Within this larger data set, seven students were identified as being allophones, and in particular L3 learners. They were all born in Canada and all spoke a minority language at home as well as English before entering FI. These students became the subject of additional analysis for the purpose of this article. Table 1 outlines the demographic background of the participants as self-identified by the learners and verified by their teachers. The category "Language Spoken in the Home" represents each student's response to the question: "What languages do you speak at home?". Pseudonyms have been used for all participants. How do L3 learners use and transfer reading strategies from one language to another?

Procedures and Measures
Data were collected over a two-year period in order to describe and explain the English and French reading experience of L3 learners as they made the transition into the FI program. The Grade 2 English reading data were of interest in the initial year of data collection. In the subsequent year, the focus shifted to the Grade 3 French reading data in the immersion class.
Grade 2 English reading data. Each student met individually with the researchers twice in Grade 2 (in March and in June). During these sessions, students were administered the Oral Reading Fluency (ORF) and Retell Fluency (RF) measures from the Dynamic Indicators of Basic Early Literacy Skills-DIBELS (Good et al., 2004). These subtests are norm-referenced, standardized English reading assessments used to assess students' reading fluency skills. The ORF measure, designed to assess students' accuracy and fluency rates, required students to read aloud as much of an English reading passage as possible in one minute. Words read correctly were counted to determine accuracy and speed of reading (fluency rate). Individual words read by the student were scored using set benchmarks. The alternate-form reliability of this subtest ranges from 0.89 to 0.94, as reported in the scoring guide (Good & Kaminski, 2003). The RF measure was used as a comprehension check. It required students to do an oral summary of what was read during the ORF measure. Individual words used by the student during the retell were scored using set benchmarks (Good & Kaminski, 2003).
Running Records (Clay, 1993;Fountas and Pinnell, 2001) were also administered during these individualized meetings to capture a more interactive view of reading. This measure documents students' reading abilities in terms of accuracy, fluency, and comprehension. Students were given leveled-reading passages containing 100 to 150 words to read, each one increasing in complexity. The examiner took notes as students read aloud and summarized different-leveled reading passages. Along with selfcorrections, accuracy-related issues such as substitutions, omissions, and repetitions were documented. Accuracy levels were established for each text using set criteria (> 95% independent, 90-94% instructional, < 89% difficult) and, along with fluency and comprehension scales, students' L2 reading levels were determined (Clay, 1993;Fountas & Pinnell, 2001).
Finally, think-aloud protocols (Chamot & El-Dinary, 1999) and interviews were conducted. Throughout the reading task, think-aloud questions were posed to students. Sample questions included were: What are you thinking now? How did you figure that out? What are you doing at this moment? Students were then asked interview questions based on those designed by Saddler and Graham (2007). In the Saddler and Graham study, questions explored students' knowledge of writing and were used with children in Grades 4 through 8. For the purpose of this study, interview questions were redesigned to access Grade 2 and 3 students' knowledge of reading and perceived use of reading strategies (see Appendix A for example of an interview protocol). For each question posed, follow-up questions might have been used to elicit clarification, gain additional information, and/or provide students with the opportunity to expand upon their answers. Think-aloud protocols and interviews were audio-recorded and transcribed.
Grade 3 French reading data. Each student met individually with the researchers four times throughout the year (September, November, March, and June) and were administered French reading measures comparable to those administered in English the year prior. Additional time was spent collecting data in Grade 3 due to the study's emphasis on cross-linguistic transfer of reading strategies and development of reading skills in an additional language.
In grade 3 a French version of the DIBELS test was used; the Indicateurs dynamiques d'habiletés précoces en lecture-IDAPEL (Dufour-Martel & Desrochers, 2011). The same subtests were administered (ORF and RF), this time using French passages. A French version of the Running Records assessment was also administered, the Fiche d'observation individualisée en lecture, to assess students' reading abilities in French. Students were given the option of doing the oral retell in French or in English, and all seven participants chose to do the retell in English. Similarly to data collected in Grade 2 English, think aloud protocols and interviews were conducted in Grade 3. Although data focused on students' French reading experiences, interviews and thinkaloud protocols were conducted in English due to students' limited oral French language proficiency. Additional questions were developed in Grade 3 to gain a better understanding of cross-linguistic transfer (see Appendix B).

Data Analysis
English data from the ORF and RF measure DIBELS were analyzed by comparing them to the Grade 3 French IDAPEL measures. Comparisons with respect to reading levels were also done using the English Running Record scores, administered in Grade 2, and the Fiche d'observation individualisée en lecture, administered in Grade 3. These analyses were useful in exploring students' reading trajectories and students' reading development in both languages.
As was the case in Grade 2, think-aloud protocols and interviews were audiorecorded and transcribed. The analysis of the think-aloud protocols and interviews focused on students' recollection or perceived use of reading strategies. This analysis was conducted in two steps. Firstly, students' answers regarding what they were thinking and doing prior to, during, and after the act of reading were extracted from the transcripts. Secondly, answers that pertained to reading strategies were analyzed.

English and French Reading Development
Our first research question was: How do these L3 students compare to other students with respect to reading in English and reading in French? To answer this question, we analysed a series of measures of reading ability that included oral fluency, retell fluency, and running record results. Below are three tables that present the descriptive statistics derived from these measures. Table 2 provides scores for all Grade 2 students who participated in the study (N = 56) in June (i.e. English reading assessments) and in March of Grade 3 (i.e. French reading) assessments in comparison to allophones' scores. Note. G2 =Grade 2, ORF = oral reading fluency, RF = retell fluency, RR = running record, G3 = Grade 3, ORF = oral reading fluency, RF = retell fluency, RR = running record, a French measures.
With the exception of Scott, all L3 students scored on or above the mean on both the English Fluency (110.56) and Retell measures (45.56). Additional analysis of students' scores revealed that five of the seven placed in the top 25% of all students in the study (N=56) in terms of English fluency skills. Similarly, for comprehension, five of the seven scored in the top 30%. One student, Scott, found himself in the bottom 20% for English fluency and the bottom 10% for comprehension. It should be noted, however, that this student had been diagnosed with an intellectual delay in Kindergarten. Results of French measures, administered in March of Grade 3, revealed that for French reading fluency, six of the seven L3 students performed in the top 25% of the students in the study. Similarly, for comprehension, six of the seven students ranked in the top 20%. Moreover, all L3 students scored above the mean on the IDAPEL French fluency measure (67.89) and six of them scored above the mean on the French Running Record measure, which assesses overall French reading performance (i.e. fluency, accuracy, and comprehension). The results from our initial question led us to believe that L3 students who performed well in English reading also performed well in French reading. It should be noted however that, although differences are noted, there is not enough power to measure statistically significant differences between and among students.

Reading Trajectories
We now turn to our second research question: "Do L3 students follow similar trajectories in their reading development in English and French? If so, in what ways?" In other words, we were interested in whether a student who was reading at a high level in English was also reading at a high level in his or her third language, in this case, French.
To answer this, we compared English Grade 2 and French Grade 3 Running Records scores of the L3 students. This particular measure was used for this analysis because it captured a more holistic understanding of reading development (i.e. fluency, accuracy, and comprehension). The graphs below show the English and French Running Records results.

L3 Students Grade 2 English and Grade 3 French Running Records Scores
The graph indicates that all seven students seem to maintain similar reading trajectories in French and in English. Although students had only been in French immersion for one year, they seem to be moving at comparable rates toward similar levels of reading proficiency in English and in French. These results, described in greater length elsewhere (Bourgoin & Dicks, 2013), support earlier findings indicating that students who read and comprehend well in one language also read fast, accurately, fluently, and with comprehension in another language (Geva & Clifton, 1994;Durgunoglu, 2002;Genesee, Geva, Dressler, & Kamil, 2006). These findings also indicate that L3 students are similar to bilingual learners in this regard. They seem to be at no greater disadvantage as a result of learning to read in multiple languages, and it would appear, have advantages over their unilingual and bilingual peers.

Reading Strategies
Our third and final research question is: "How do L3 students use and transfer reading strategies from one language to another?" In order to gain a better understanding of what allophones did during the act of reading, we analysed data from student interviews, which were conducted following the completion of reading tasks. Students' responses provided a glimpse into the strategic thinking of Grade 3 French immersion allophones. More specifically, we identified two important themes in these data: (1) L3 studnets' knowledge and use of other languages to support French reading development and (2) metalinguistic and metacognitive insights.
L3 students explained that they relied on their knowledge of the English language as well as their knowledge of their home language to help them read in French. According to them, it made French reading "easier". As Danika explained, "to understand [a text], I kind of search for the word in my brain in English. If there's a word I don't know, I kind of see if it might be a little bit the same in English". Anna shared similar insights: "some words in French are spelled similar to English so I think by looking at that and then remembering the words in English helps me". When asked what helps him read in French, Mark explained that, "you can still use your English strategies. It's kind of easy to use the same strategies. They're all words right…French, English, Lebanese, Japanese or something". In this quote, Mark is implying that the processes involved in reading in one language are the same and transfer across different languages. Similarly, Anabelle felt that knowing different languages helps her read a third additional language. As she explained: Because most of the words in English and French can be the same I can really relate to what [the authors are] saying. Probably cause some of the words are like English and the other words like Spanish and since I speak Spanish I kind of understand all of it.
On the other hand, Bruce felt that since his first language, Chinese, was alphabetically unrelated to French, it was not as helpful. "I'd rather learn Korean. It's more related to Chinese…and it would be easier for me than all these guys." This comment relates to the issue of language families and the relative ease or difficulty involved in learning languages that have similar or different grammatical and orthographical systems.
L3 learners also expressed impressive levels of metalinguistic/metacognitive insights. Firstly, they spoke of the importance of learning specific French sounds and vocabulary, and that these two elements were essential to learning to read in French. When asked what he felt he needed to learn from his French immersion teachers, Mark explained, "when you get to know all the letters and sounds, you get the hang of it, so you just read and you get better and better". Moreover, they linked important elements of reading to the self-monitoring process. According to Anabelle "…they're a whole bunch of [French] words. You need to know what the words are so you can keep reading and understand. Even if you can read them, you still need to be able to understand them". Similarly, when asked what she needed to learn next if she was going to progress with her French reading, Anna identified vowels, accents, and French letter combination as crucial elements.
The L3 learners described using many strategies to help them read in French, many of which they had learned when learning to read in English. Table 5 lists the strategies described by these students. Table 3 Reading Strategies Used as Described by Allophones in French Immersion

French reading strategies Sounds (phonological awareness)
-I use the different sounds like "é" and "a" and "o." -It didn't sound right.
-Sounded it out…cause I know what the letters sound like.
-Knowing the silent letters.
-I use the accents I already know.

Vocabulary
-I skip words and see if they make more sense if I skip them.
-I replace the word with other familiar words.
-I thought of the word in English.
-Knowing or recognizing words.
-Find other words in the words that I know.
-I make the words into pieces, like chunking.

Comprehension
-I don't have to just look at the word. I can look at the picture and get a little vision in my head. -I make it stick in my head.
-I just sort of let the information get stored in a file in my brain. To remember, the file opens and all the information comes out.
L3 learners were also flexible and selective in their use of reading strategies. Anabelle's comment illustrates this well when asked whether she used strategies to help her read in French. "I think it depends on what I'm reading. Like if I had really, really hard words I would use a strategy, but sometimes I don't find those hard words so I don't use them." What is more, there were many examples where allophones make use of multiple strategies to orchestrate their reading. For example, Anna would call upon various strategies to problem-solve reading challenges.
Sometimes I just go back and reread, or I ask a teacher if it's right, or I sound it out, or use my strategies like the [French] sounds I know like "oi", "gn". Or, I use the words that I know, or I go to the end of a sentence cause sometimes it's a word in the middle of the sentence so I go to the end of the sentence and then I try to fit in a word that would make sense.
Allophones were able to reflect upon and manipulate language (i.e. metalinguistic awareness). When asked how she knows if she's reading in French correctly, Danika offers this explanation.
When I read French, I always check over the sounds in the words to see if I made any mistake. If I read a sentence and I kind of feel that I didn't say it right or I missed something or I wasn't saying it right, I get a feeling. I get the same feeling when reading in English and in French, it's kind of the same.
Previous research found that, for general language abilities, L3 students are highly capable of reflecting upon language and are able to manipulate it to reach their goal (Cenoz, 2003). According to Anna, "If the word doesn't make sense to me, I try to make it make sense. I use the words I know and replace them with the words I don't know". As the quotes in this section highlight, allophones seem to work with and through their language repertoires and previously acquired knowledge of reading strategies, expanding their skills, abilities, and knowledge of reading in a new language.
Discussion Over a two-year period we examined the English and French reading development of students whose first language was neither French nor English. Our data revealed that young elementary-aged L3 students performed as well or better than their bilingual peers with respect to their English reading development. We reported similar results with respect to students' French reading development once enrolled in the Grade 3 French immersion program. These findings support earlier results showing that students with ethnic minority backgrounds experience no disadvantage being in French immersion in terms of their English language development and overall academic achievement (Genesee, Paradis, & Crago, 2004;Cenoz & Genesee, 1998).
Additionally, our findings indicate that L3 students retained similar reading trajectories with respect to their second-and third-language. In other words, strong readers who read strategically in English also did so in French, supporting the notion of cross-linguistics transfer across multiple languages. As noted previously, first-and second-language reading skills of bilingual minority-language students are closely correlated Stutnabb-Kangas & Toukomaa, 1976). Results from this study, although limited to seven students, seem to indicate a similar relationship between L2 and L3 reading (Cenoz, 2003).
Data from interviews and think-aloud protocols support the claim that L3 learners use a wide range of cognitive and metacognitive resources, including accessing L1 and L2 language literacy skills. Similarly to Cenoz (2003), students in this study reading in multiple languages used insights and skills learned in one language to support the other language being learned. They had more language experiences to draw from and were able to use these experiences as part of their metacognitive processes.
Phonological, syntactic, and functional awareness, along with decoding strategies, good meaning-making strategies, and reading comprehension abilities, have all been found to correlate among languages (Durgunoglu, 2002;Genesee, et al., 2006). Although these studies primarily examined transfer from L1 to L2, the same can be concluded when learning to read in a third language. As van Gelderen et al. (2003) explained, "metacognitive knowledge… is learned in L1 but can be transferred to L2, L3 and theoretically an infinite number of other languages" (p. 11). Our findings with respect to the retention of similar reading profiles across languages and transfer for skills support results found in other third-language acquisition and trilingual education research. Although studies in this area remain limited, the research does indicate that there are potential beneficial effects of third-language acquisition on both first-and secondlanguage development and cognitive/academic development (Cenoz, 2003).

Limitations
There are some limitations related to this study. Firstly, as noted in the results section, because we were dealing with only a few French immersion schools, the number of L3 students was limited. This small number of subjects did not allow us to conduct the kinds of statistical analyses that would allow us to say with certainty that the similarities and differences among students' reading performance are significant. Secondly, we did not consider the issues related to language families and whether or not students coming from Indo-European language backgrounds were advantaged in the learning of French as a third language compared to students of Asian or Arabic language backgrounds. Future research with larger numbers of L3 students from a wide range of backgrounds would allow for a deeper exploration of the key issues.

Conclusion
With one exception, all L3 students in this study were proficient readers in French, their third language, after one year of study in French immersion. Moreover, they were all highly capable of reflecting upon language and manipulating it. They provided insight into how they are learning to read an additional language and what resources and strategies they using to monitor their reading. They spoke of the importance of drawing upon their understanding of French sounds and vocabulary. These elements cannot be overlooked in additional language learning contexts, as they seem to be highly and explicitly used by allophones as a way to monitor their reading.
The results of this study provide insights for the teaching of reading to L3 students in French immersion. Firstly, the role of sound work and word work in French immersion cannot be overlooked when teachings students how to read. Secondly, the majority of students in this study had a fairly strong and established minority language. Moreover, they were very proud of knowing and learning multiple languages and were keen in continuing to do so. By learning about and valuing students' linguistic backgrounds, teachers are able to draw upon students' experiences, discuss similarities and differences, and foster the development of a community of language learners. Finally, it is important to acknowledge and appreciate that students use their knowledge of other languages, including English, as a resource. Many allophones bring with them a substantial amount of linguistic knowledge and strategies that can and should be exploited. It is important to open up conversations with students about how they learn and how they read. This study demonstrates that students not only have the ability to talk about how they learn to read, but also have keen insights into how they do so. They are able to draw upon their experiences, discuss similarities and differences, and help foster the development of a community of language learners (Cummins, 2001).