My many selves are still me: Motivation and multilingualism1

Two concepts of multilingualism that relate to the selves aspect of Dörnyei’s (2009) L2 motivational self system (L2MSS) are highlighted in this article: Thompson’s concept of perceived positive language interaction (PPLI) and Henry’s notion of the ideal multilingual self. With the dynamic model of multilingualism informing both concepts (Herdina & Jessner, 2002; Jessner, 2006, 2008), the intangible advantage that multilingual speakers have over monolingual speakers is clearly articulated in the discussion of this topic. The interconnectivity of language systems is an inherent aspect of the DMM; as such, both Thompson with PPLI and Henry with the ideal multilingual self incorporate the DMM as a framework to indicate the fluid nature of these constructs as additional language learning experiences are added to the system over time. This article further explores the dynamicity of multilingual learners’ language systems and the influences that induce change. Specifically, data from Thompson’s (2017b) study on LOTE learners are reexamined to explore this question. Additionally, excerpts from Natasha Lvovich’s (1997) The Multilingual Self, an autobiography of an L1 Russian speaker, are analyzed to present different possible models of incorporating the multilingual self and PPLI. The article ends with a discussion of an inherently multilingual context, as well as thoughts regarding the possibility of different types of future selves.


Introduction to "selves"
Qui suis-je? ¿Quién soy? Ben kimim? Who am I? This is a question that resonates with all of us, no matter in which language the question is formulated. As Mercer (2014) states, reflecting briefly on the question of 'who you are' leaves you with a multitude of selfdescriptions, incorporating a range of self-related cognitions, beliefs, emotions, motives, roles, relationships, memories, dreams and goals, as well as expressions of who you feel you are not. The self is the hub at the centre of all our lived experiences. (p. 160) The languages we use, and the way that we are exposed to them, are at the epicenter of our conceptualization of self. Multilinguals, operationalized in this article as those who have experience with more than one second/foreign language, have the advantages of multiple world views with every turn of phrase. For an L1 English speaker, pondering "Who am I?" in multiple languages illustrates not only the metaphorical complexities of the question, but also foregrounds the linguistic implications. In French, why is this one of the only instances of first-person singular subject-verb inversion for question formulation? Why does Spanish require a question mark in both opening and closing? Where is the copula represented in the Turkish sentence? Certainly, the conceptualization of self is much more complex than a learner expressing first person singular differently in a variety of languages. Mercer and Williams (2014) dedicated an edited volume to the variety of ways that the self can be conceptualized in SLA. Mercer and Williams (2014) indicate that "there are many different ways to conceptualize, define and thus measure the self" (p. 177). In the same volume, Ushioda (2014) elaborates on the many internal and external self-related cognitions that are related to language learning motivation: Factors internal to the self include the various self-related cognitions briefly mentioned earlier (i.e., cognitions by the self and cognitions about the self), as well as attitudinal and affective factors (e.g., enjoyment, anxiety) and individual characteristics (e.g., gender, developmental age, personality). On the other hand, factors external to the self comprise the broad complex of social, cultural and contextual factors that may influence individual motivation, such as interactions with significant others (e.g., teachers, parents, peers), specific features of the learning environment (e.g., classroom tasks and materials) or less visible aspects of the wider sociocultural context (e.g., educational values and cultural beliefs). (pp. 130-131) This bourgeoning interest in including people's multi-faceted and contextualized language learning experiences has also been especially evidenced in motivation research, as can be seen in the Fall 2017 special issue of the Modern Language Journal, as well as in other recent publications. In particular, May (2014) calls attention to what Ortega (2014) refers to as the monolingual bias, or when language scholars take monolingualism as the default norm in studying language acquisition. Individuals with more than one language learning experience do not experience learning subsequent languages in the same way as those who are starting with the native language as the only linguistic system to which they have access.

Objectives and framework
It is the concept of the internally formulated and externally influenced selves that is the current focus of this piece. Specifically, this article examines two concepts of multilingualism that relate to the selves aspect of Dörnyei's (2009) L2 motivational self system (L2MSS): Thompson's concept of perceived positive language interaction (PPLI, 2016) and Henry's (2017) notion of the ideal multilingual self.
At the core, the conceptualization of self and motivation in language learning was popularized by Dörnyei's (2009) L2MSS. In brief, the L2MSS draws on Markus and Nurius' (1986) concept of possible selves, and Higgins' (1987) selfdiscrepancy theory (SDT). Thompson (2017b) describes the L2MSS as a two-part theory: the conceptualization of selves (ideal and ought-to) and the contextualized learning experience. The ideal self is the self that the learner would like to become in terms of language use (promotion focus) and the ought-to self is the self that the learner feels he or she should become in terms of language use (prevention focus). The process by which the learners acquire the target language and the context in which they find themselves, both of which influence the formation of the selves, comprise the second part of the L2MSS -the learning experience.
Particularly relevant to this article is SDT and the conceptualization of the different selves involved. Higgins discusses an actual self (the current state of the learner) juxtaposed with the ideal and ought selves; the latter two are both what Markus and Nurius would label as future selves (i.e., the imagined future selves of the learner). Higgins also disentangled four main types of self-discrepancies: actual/own with ideal/own; actual/own with ideal/other; actual/own with ought/own; and actual/own with ought/other. As Thompson and Vásquez (2015) indicate: The L2MSS does not strongly articulate the 'I' versus 'other' dimensions of self-discrepancy theory. Thus, the question arises whether there can be an 'other' dimension in the ideal L2 self (the construct with a focus on 'I'), and an 'I' dimension in the ought-to L2 self (the construct with a focus on 'other'), which would be more congruent with Higgins's self-discrepancy theory. (p. 170) As Mercer and Williams (2014) note, there are many other selves in the language learning literature; for example, Thompson and Vásquez's (2015) antiought-to self, which is a self that forms with the desire to reject expectations or that thrives in challenging situations (also see Thompson, 2017a, for elaboration), comes directly from a re-interpretation of the "I"/other" dimension of SDT. Other selves, such as those connected to a specific profession or activity, such as an ideal teacher self, have also been proposed in the literature (i.e., Gao & Xu, 2014), as has a Bildungs-Selbst 'educational self' (Busse, 2017), which is another type of future self.
The following sections will examine the role of context and experiences on multilingualism and self formation. Of particular focus are the internally situated constructs of PPLI and the ideal multilingual self. Throughout the article, theoretical underpinnings and implications of the person-in-context perspective of these two constructs will be portrayed, and the dynamicity of perceptions of multilingualism and self formation will be illuminated. Both quantitative and qualitative analyses are presented in this article, as both types of analyses can illustrate the dynamicity of perceived positive interactions between languages studied and the types of selves developed. For example, as is explained below, a content analysis of short answer questions is used to group participants with those who perceive positive interactions between languages studied and those who do not (the PPLI construct). The participant answers are then used to support and explain the quantitative findings of the group differences. Thus far, the ideal multilingual self has been primarily supported by qualitative data. However, Henry and Thorsen (2017) provided the results of the ideal multilingual self measured quantitatively. Further quantitative measurements of selves have been illustrated by Dörnyei and Chan (2013) and Thompson (2017b); both of these studies use exploratory factor analyses to support the concept of distinct selves in different languages. As such, considering the context and added language learning experiences through time are both necessary when conceptualizing the dynamicity of learning multiple languages, as are the use of both qualitative and quantitative research methodologies.

The internally situated constructs of PPLI and the ideal multilingual self
As Mercer (2014) writes, "It is perhaps stating the obvious to say that the self is complex" (p. 160). The dynamic interchange between more than one second language increases the complexity even more, as it is related to the formation of the self or selves and language learning motivation. Two recent frameworks used in motivation and multilingualism research to illustrate the cognitive representations of learners' language systems are Thompson's theory of perceived positive language interaction (PPLI, 2016) and Henry's (2017) concept of the ideal multilingual self. Both of these constructs incorporate the dynamic model of multilingualism (DMM; Herdina & Jessner, 2002;Jessner, 2006Jessner, , 2008 in their overarching theoretical development. The DMM addresses how linguistic systems of multilingual speakers are interrelated; the so-named M-factor (or a special characteristic that only multilingual individuals have) in this theory is representative of the intangible advantage multilingual speakers have in "language learning, language management, and language maintenance" (Herdina & Jessner, 2002, p. 131). The interconnectivity and dynamicity of a learner's language systems is a fundamental aspect of the DMM, a characteristic that both PPLI and the ideal multilingual self share.
PPLI (e.g., Thompson, 2016) is a learner variable that has been used to explore a variety of individual differences, including motivation (e.g., Thompson, 2017b;Thompson & Erdil-Moody, 2016). As mentioned previously, DMM was one of the frameworks used in the formation of PPLI; several other theories, such as Kellerman's (1979) concept of perceived language distance and Odlin's (1989Odlin's ( , updated in 2008 theory of interlingual identification, were also used when first conceptualizing PPLI. Both perceived language distance and interlingual identification emphasize the importance of the learner's perception of relatedness in terms of language transfer. In other words, the importance of the perception of language relatedness outweighs the actual typological similarities between the languages in question. PPLI also draws from De Angelis' (2007) concept of the relatively small amount of language exposure needed to affect subsequent acquisition.
In terms of data analysis in the PPLI framework, participants answer the question: "If you have studied other languages in the past, do you think that this has helped or hindered your ability to learn subsequent languages? Please provide specific examples where appropriate." The answers are oftentimes used as examples of specific themes, but for a quantitative analysis, learners are placed into PPLI and NPPLI (no perceived positive language interaction) groups based on the answers to these questions. An example of an answer that would place a learner into the PPLI group is as follows: I think learning Spanish before Japanese has helped me with being able to translate from a foreign language to English and the other way around. I think Spanish also really helped me understand pronunciation and phonetic sounds.
An example of an answer that would place a learner in the NPPLI group is as follows: I believe learning Spanish hindered my ability to learn other languages. Thompson (2016, p. 97) provides additional useful tips and examples for coding open-ended comments via the PPLI framework. For example, six coding tips are included: 1. Verify that the participants in question are multilingual (i.e., that they have studied at least two languages beyond the L1). The information should be independently collected in a separate part of the background questionnaire. Bilingual participants (those with only one language beyond the L1) cannot be classified as PPLI. 2. Read the response and identify if a perceived positive language interaction has been stated. 3. Eliminate responses that indicate positive interactions involving the L1 and an L2 (these are interesting, but are outside the scope of the PPLI framework). 4. Remember that responses such as "I'm not sure" or "Neither positive nor negative" do not qualify for PPLI. 5. If a participant states both positive and negative interactions, the PPLI coding can be used if the participant's overall feeling is that the interactions are positive. A NPPLI coding will be used if the participant's overall feeling is that the interactions were negative. 6. PPLI coding is by default subjective. It is always a good idea to have a second or third rater coding the answers for interrater reliability.
As an example of coding tip #5 (indicating both positive and negative interactions with a final PPLI coding), the following quotation was given: Of course, there was an interaction. Thanks to English, I learned other languages more easily because the grammar rules were almost the same. However, I had trouble at first with German. For instance, I would write 'and' instead of 'und.' Except for these it affected positively. (Thompson, 2016, p. 97) Having previously written about motivation and multilingualism in the L2MSS framework, Henry (2017) introduces the ideal multilingual self. Influenced by the DMM and Aronin's (2016) concept of multilinguality, the essence of the multilingual self is that a learner's motivational systems are interrelated to form a multilingual motivational self system. This multilingual motivational self system has various multilingual self guides, two of which are the ideal multilingual self and the contentedly bilingual self: "While the contentedly bilingual self can have the effect of further weakening the power of the ideal Ly self, the ideal multilingual self can have the opposite effect, enhancing the strength of the ideal Ly self" (Henry, 2017, p. 554). The ideal multilingual self and PPLI are both internally-situated constructs, and they both explore the mental representations of the interconnectivity of a learner's language systems. As Herdina and Jessner's DMM is central to both PPLI and the ideal multilingual self, both constructs highlight the dynamicity of the language learning process when multiple languages are involved.

Contextualized selves and PPLI: Implications of the person-in-context perspective
Specific languages studied in particular contexts affect the development of various motivational selves (e.g., Ushioda, 2009). Huang, Hsu, and Chen (2015) in the Taiwanese context, for example, examine learners with English as an L2 and either French, German, Japanese, or Korean as an L3. Using multiple regression analyses with ideal and ought-to selves, learning experience, cultural interest, career opportunity, and role obligation, intended learning effort were predicted for the five languages in question. The ideal self and cultural interest were the strongest predictors for the French sample and the weakest predictors in the English sample. Identification with social role obligations was the strongest predictor for English and the weakest for Japanese. The ought-to self was only a significant predictor of English, Japanese, and German (and not for French and Korean). The predictive strengths of specific languages on the different types of selves is related to the social constructs around the languages in question.
In the US context, re-examining the data from Thompson (2017b) for language specificity, it can be observed that the lowest percentage of PPLI learners can be found with those who study Spanish (see Table 2 and Figure 2); it is also the case that with the Spanish students, there were fewer who had studied an additional language (see Table 1 and Figure 1). Both of these trends are undoubtedly related to the US context and the role of Spanish therein. Spanish is currently the most-taught language in both secondary and post-secondary levels in the US context. If a language other than English is required or preferred for a job in this context, the preference is oftentimes Spanish. Thus, students who study Spanish are likely influenced positively regarding the marketability of the language, while at the same time being negatively influenced by the oftentimes negative discourse in the media about Spanish and Spanish-speaking populations (e.g., García & Mason, 2009). The complex relationship with Spanish in the US context undoubtedly influences students' language learning motivation, and also influences the dynamic interactions of multiple languages that take place in such a complex socio-political setting.  The highest number of PPLI participants (N = 89) are those learners who have a language other than Spanish as the first foreign language studied. For the total number of languages other than English (LOTEs) other than Spanish, 89/144 (61.8%) of the participants have internalized the positive interactions between languages studied into their linguistic systems. An opposite trend is found with the Spanish learners; the lowest number of participants fall into the PPLI Spanish group (N = 31), or 31/78 (39.7%) of the Spanish learners. Within the Spanish group, there are fewer learners in the PPLI group (N = 31) than in the NPPLI group (N = 47). The opposite is true for the group composed of those who studied LOTEs other than Spanish. There are more learners in the PPLI group (N = 89) than in the NPPLI group (N = 55). What can this data tell us? Even though Spanish is the most common language studied in the US context, it does not possess the same utilitarian value as English as a global language. Thus, students may feel obliged to study Spanish because of the relative abundance of Spanish speakers in the US but it is possible that not as many internalize it as those who study languages other than Spanish. This is not the case for all learners of Spanish, however, as 33 out of the 78 Spanish students did perceive positive interactions between languages studied. Perceiving positive language interactions allow learners to formulate a mental representation of the learning experience that is beyond the separate language systems studied. This perceived positive interaction is akin to Henry's idea of the ideal multilingual self, a self that is conceived beyond the selves that are formed for the separate languages in question.
Without at least two foreign language experiences, people are not able to have the M-factor, that is, conceptualize positive interactions between languages. It is possible that in primarily L1 English contexts, another dominant language, such as Spanish in the US, might result in learners having a "contentedly bilingual self" as suggested with Henry's (2017) Swedish learners of English. The situations are not directly comparable, however. The use of English in Sweden is so prevalent, that some have even argued that it should be considered a second, as opposed to a foreign, language, and most Swedes have a high competency in the language (Sylvén & Thompson, 2015). In the US context, it is sometimes the case that the language studied is oftentimes not used frequently outside of the classroom context.

PPLI and the ideal multilingual self in a language learning narrative
Whereas the previous section utilized quantitative data to explore the dynamicity of PPLI, particularly in terms of students of Spanish, this section explores the use of autobiographical narratives to examine PPLI and the ideal multilingual self. The use of language learning narratives as data has been considered a recognized practice only recently. As Pavlenko (2007) indicates, such narratives provide a unique insider perspective to the language learning process, or "people's private worlds" (p. 164). The learner who is interviewed, or as in this case, who writes the autobiography, chooses the specific topics and situations to emphasize. The resulting narratives turned data are the learner's perspective of the truth. If multiple languages are involved, the learner provides evidence for the interactions and dynamicity of multiple language learning processes. As such, Natasha Lvovich's (1997) the multilingual self is an example of another perspective of how to integrate PPLI and the ideal multilingual self, while at the same time theorizing how a multilingual learner conceptualizes her selves in different languages. The text was read multiple times for several rounds of coding.
Structural coding was first employed to find all excerpts of development of selves and perceived interactions between languages studied in the autobiographical narrative. For the second round of coding, focused coding (an adaptation of selective and axial coding) was used to organize the data into salient themes (Saldaña, 2016), and the excerpts presented are examples of the salient themes found. Evident through the text is the consistently shifting concept of self and interactions between languages that Lvovich expresses as she expands her language learning experiences.
Throughout the book, Lvovich primarily expresses language-specific selves (e.g., Dörnyei & Chan, 2013;Thompson, 2017b), primarily with a focus on her French self. Desperate to have an escape from her situation in the Soviet Union, Lvovich created a French self with the encouragement of role models, such as Yulia, her grandmother, who was "a recipient and simultaneously a facilitator of all my French life, information, impressions, and emotions" (p. 8). Natasha was determined to become French, so to speak, visualizing the variety of situations needed to make this possible: I had to learn to do everything a French person does: speak with a Parisian accent, joke about domestic politics, sing children's songs, read and enjoy grotesque detective stories in argot as well as the most sophisticated literature, write in French in any style, curse, gesticulate, give speeches, count mentally, and dip the imagined croissant into coffee. I had to know how the French make their beds, talk on the phone, write business letters, and cook meals from different provinces. (p. 2) This French self was created as an escape or sort of safety net because "A French personality, after all, was much less confusing and safer than being a Jew in Soviet Russia. It was a beautiful Me, the Me that I liked" (p. 8-9). As ideal selves inevitably do, Natasha's French self shifted, based on the context. After eventually immigrating to the US, she no longer needed the safety net of being French: "And then it started to make sense to me. I did not have to continue being French in America" (p. 72). The shift in her perception of self in French did not mean she no longer had an affinity to the language; it just meant that she no longer needed her French self to survive.
Lvovich's English self was a less confident one than her French self, especially at first. Indeed, she had aspects of a feared self in English, something that was not present in her French self: "What will I do when my English is not their English?" (p. 56). Readers can sense the difference in her feelings towards English versus her feelings towards French; Natasha also compares learning the two languages: Something is missing in my way of functioning in English. Something substantial, important, which does not let me enjoy my linguistic performance. It is like I am floating on the surface of the ocean, giving curious glances into its depth. It's like I am fishing for a deeper essence, and sometimes I get some fish, but they are separate fishes, not the overall picture, with everything lying there, on the bottom. Frustrating. Sometimes I felt that way in French, too, but I always ended up sensing the whole structure of everything, in meanings and words, inward and outward, at any level, in any dimension . . . I don't like America. I hate New York. I hate my life. (p. 61) From the excerpt, it is clear that Lvovich's feeling towards English is affected by her feeling towards her living situation (i.e., New York). She reflects on learning English via her experience of learning French and sees the positive effects that French has on her language learning experience (despite the fact that she preferred the process of learning French). In other words, she exhibits the characteristics of other PPLI learners illustrated in previous research: Could I have become more proficient in English as fast as that, without being fluent in French? Could I embrace and feel the depth of the culture? Would I meet so many wonderful people if I had not loved my French friends? Would I have been able to continue to be a linguist, a writer, a scholar, a researcher -a creator, without my French life ? (p. 72) Before arriving in the US, she and her family also had a brief stay in Italy. Interestingly, she did not discuss language interactions between Italian and French; however, there is a brief mention of how Latin helped her with Italian: "Yes, Latin made me admire the language form -maybe that was my first exposure to structuralism?" (pp. 46-47). It is perhaps the case that no real attachment was formed with Italian because of the feeling of non-permanence. They were "stuck in Italy, lost between two worlds and overwhelmingly confused . . . Meanwhile, I am learning Italian" (p. 44). The Italian learning was something to pass the time, but it did not seem to be integrated into her developing self system as much as French was at first and how English became later.
Lvovich, similar to anecdotes of other language learners found in previous research, seemed to conceptualize different selves in different languages: "With each language and each identity, there will be more life, more love, and more growing. Multiplicity is the adjustment" (p. 73). At the same time, however, there is an instance in which she exhibits a sort of ideal multilingual self in the form of being "a linguist:" I am a linguist. What difference does it make for a linguist, which language to learn, which language to teach? I can enjoy the process of constructing the language, mentally building rules and systems, taking notes of irregularities, and admiring -yes, admiring the beauty of linguistic logic. I know I am learning. I am learning very fast, all by myself, like an experienced linguist should do, like I always did. I am learning from people around me, form shopping, form office signs, and from my own teaching. I am learning from trying to become once again what I have always been: a linguist. (pp. 59-60) Using these excerpts from Lvovich's book, we can see the interplay between her French, English, and Italian selves. We can also see evidence of PPLI between French and English, as well as a separate ideal multilingual self. Using the evidence at hand, we can start to conceptualize the interaction of PPLI and the ideal multilingual self, as well as the integration of the ideal multilingual self into the language system. More empirical data from a diverse population of language learners would lead to further development of the hypothesized models. Figure 3 was created to show one possibility in which the ideal multilingual self is the background context with the French, English, and Italian selves situated within it, according to Lvovich's narrative. PPLI is evident for the French and English selves, but not for the Italian self.

Figure 3
The ideal multilingual self as the background self Figure 4 was created to illustrate the same system, conceptualized slightly differently. In Figure 4, the ideal multilingual self is connected to the three language selves and is still a factor in the PPLI conceptualization. As noted above, further research with more empirical data can be used to develop these models further.

Figure 4
The ideal multilingual self directly connected to the language-specific ideal selves How would learners, such as Lvovich, develop their ideal multilingual self? How would this development be related to the PPLI construct? As Dörnyei and Al-Hoorie (2017) state, "From the perspective of the current version of the L2 Motivational Self System, such a 'multilingual self' does not lend itself easily to actual visualization, because the visionary aspect of the theory has mostly been operationalized as imagining oneself using one L2 in a concrete situation" (p. 459). However, Henry (2017) argues that the abstract images conjured for an ideal multilingual self are that of the "essence of the experience" (p. 557, italics in original), and that it focuses the attention "inwards towards the individual's central, intrinsic, and more generally idealistic concerns" (p. 558, italics in original). Nevertheless, the question remains of what those who conceptualize an ideal multilingual self use for visualization; perhaps what is visualized are the language interactions as conceptualized in the PPLI construct, but more research is needed on this matter.
One thing is certain -both PPLI and the multilingual self are dynamic, changing with each language learning experience. For example, Thompson and Vásquez (2015) investigated the developing selves of several advanced language learners turned language teachers of German, Chinese, and Italian. The participant who studied and later taught Italian, Vera, started her journey with an abstract image of wanting to be bilingual: "I had this goal in life -I wanted to be bilingual. I always knew that. I don't know why but I did" (p. 163). Vera's ideal self was originally non-language specific, but as she ended up in Italy, her ideal Italian self overshadowed her more general ideal self. Was the generic ideal self able to be overshadowed because of Vera's status of an L1 English speaker? Perhaps, but in other research, relatively few of the total number of participants seemed capable of conceptualizing a more general multilingual self (e.g., Busse, 2017). Busse suggested that these students might have "an overarching plurilingual ideal Bildungs-Selbst [educational self]" (p. 578), and Lasagabaster (2017) warns that language learners need to be in harmony with individual selves (ideal and ought-to), as well as the multilingual self (p. 592). As several researchers have suggested the existence of a sort of overarching ideal self for learners of multiple languages, the existence and representation of this phenomenon is ripe for future exploration.

Multiple language learning selves in inherently multilingual contexts
One of the few authors who studies inherently multilingual context is Coetzee-Van Rooy (2014), who proposed a multilingual language learning self. Unlike the multilingual self that other researches (e.g., Henry & Ushioda) have proposed, Coetzee-Van Rooy bases this multilingual self primarily on the ought-to self dimension -individuals in these contexts feel that they need to be communicative in a variety of different languages depending on the context in order to fit in to society as a whole.
In her 2014 study, her multilingual southern Sotho and Zulu participants know many languages, using some more than others on a regular basis, and hers is the first mention of a type of multilingual self in the applied linguistics literature: I want to argue that, linked to the 'sociolinguistic language mode' of an environment, the language learning self can be conceptualised as a multilingual language learning self. In the minds of people living in these types of environments there is an expectation that members learn many languages as part of their ordinary behaviour as integrated citizens that belong to the society . . . The 'ought to language self' in a multilingual language mode society directs people to believe that if they are not multilingual in this society, they do not 'fit in', because well-integrated citizens in this society are multilingual. This idea finds support from Bamgbose (1994, p. 34), who argues that a person who speaks several languages is to be regarded as a better integrated citizen than one who is only proficient in one language in African contexts. (Coetzee-Van Rooy, 2014, p. 124) Similarly, Thompson (in press) describes the experiences of a multilingual Senegalese English teacher, Ablaye (pseudonym). His schooling was primarily in French when he was growing up, but he started to seriously study English at the university. During the interview, he compares learning English to learning French, and insists that "English is easier than French." Interestingly, Ablaye did not discuss his native languages, Wolof and Serer, until he was explicitly asked about them. In his answer, he spoke about his daily use of these languages: In this case, Ablaye displays what Coetzee-Van Rooy describes as the multilingual language learning self. Ablaye has developed a strong ideal English teaching self, indicated in his interview as a sense of pride in statements such as: "Yes, [I'm] most interested in English. And whenever I see somebody speaking English, I don't speak French. I speak to him in English. In our country it is what we are doing. We English teachers, we speak English." However, it is clear that he also uses French, Serer, and Wolof as needed in daily interactions. Student Q in Coetzee-Van Rooy describes a similar situation with regard to language use: "And Sotho I learned at home with my grandparents, cos I stay with my grandparents . . . And then Zulu I learned as I visited my mother and my father in Soweto. Cos they stay in Soweto . . . And Xhosa. Cos my mother's family, my mother's side of the family is Xhosa, my father's side is Zulu. So when I visited them [mother and father's family], that's how I learned those different languages [Xhosa and Zulu]" (p. 133). Thompson's Senegalese participant, Ablaye, and Coetzee-Van Rooy's southern Sotho and Zulu participants all have multiple languages that they use daily, which is a distinct situation than those language learners who learn multiple languages primarily by formalized mechanisms, such as in a classroom.

Concluding thoughts
PPLI and the self concepts have primarily been situated in contexts in which multilingualism is a learned phenomenon (i.e., in formal settings, such as a classroom). Of course, there are large parts of the world, which are largely understudied in terms of multilingual and applied linguistics research, where multilingualism is a normal part of everyone's daily life. Additionally, both PPLI and the self concepts have primarily been investigated in contexts where learning a language is a choice, or at least part of the required schooling (i.e., additive bilingualism). The choice of learning a language is quite different than situations where the language being learned was a matter of survival, such as in cases of immigrants (forced or by choice) or refugees. How the different languages interact in the minds of individuals in these contexts, and how their different selves are formed, is an area that needs to be further explored. Additionally, it has been shown that not all multilingual language learners perceive positive interactions between languages learned; some of these multilingual participants envision each language learned as a separate language system. Similarly, not all multilinguals will have a multilingual self identity. Some of these learners might have separate ideal selves (or indeed ought-to, anti-ought-to, and/or feared selves) that are not linked together with any sort of ideal multilingual self. As language learning is inherently dynamic, formation of selves and perceiving positive language interactions are ever-changing and are highly dependent on the context. Qui suis-je? ¿Quién soy? Ben kimim? Who am I? This is a question that language learners might pose when learning one or more languages. The answer to this question at least partially depends on the context and what language is being spoken at the time of the inquiry. How and when positive language interactions are seen, whether an ideal self is conceptualized for each language or if there is a more salient ideal multilingual self, if a language is learned just to prove to others that it can be done or merely out of a sense of obligation -all of these eventualities contribute to the formation of language learning selves, including the ability to see positive interactions between languages. The one certainty is that the answer to "Who am I?" will be ever-evolving.