The Fulbright Foreign Language Teaching Assistant Program (FLTA) and the language policies for teaching Portuguese in the USA / The Fulbright Foreign Language Teaching Assistant Program (FLTA) e as políticas linguísticas para o ensino de português nos EUA

This paper aimed to describe the language policy developed by The Fulbright Foreign Language Teaching Assistant Program (FLTA) and to discuss the agency of its fellows to the spread Brazilian Portuguese in higher education institutions in the USA. Based on the theoretical notion of language policy as a process (RICENTO; HORNBERGER, 1996; JOHNSON, 2013; JOHNSON; JOHNSON, 2015), we analyzed a corpora consisting of documentary data, information gathered through telephone calls and e-mails, and semi-structured interviews with members of the Fulbright Commission Brazil and former Brazilian FLTAs. This is a qualitative-interpretative study (LIN, 2015) of the textsdiscourses that are part of the corpora. The results show the levels and the main agents that are part of this language policy – from the US government to the Brazilian teachers – as illustrated using the onion metaphor with concentric circles that indicate the several layers where these agents act. Results also found there is little formalization of a language policy education for foreign/additional languages by the FLTA Program. This makes the Brazilian FLTAs, who act sometimes collectively and sometimes individually, the main force to form a language policy to spread Brazilian Portuguese in the USA.


Introduction
The Portuguese language is taught in the United States of America (USA) mostly in higher education institutions (MILLERET, 2012). According to the Modern Language Association's latest ISSN: 2317ISSN: -2347ISSN: -v. 9, n. 4 (2020 Todo o conteúdo da RLR está licenciado sob Creative Commons Atribuição 4.0 Internacional http://dx.doi.org/10. 35572/rlr.v9i4.1931 140 report (LOONEY; LUSIN, 2019), Portuguese is ranked 11 th in the list of most studied languages in that country. In the book Ensino de português nos Estados Unidos: história, desenvolvimento e perspectivas, Luna (2012) compiled some historical and contemporary records regarding the teaching of Portuguese in US universities, a subject that has yet to be fully investigated in Brazilian publications. Due to this fact, this study focused on a program that has helped the spread of Portuguese language considerably at college level by sending young Brazilian teachers to the USA: We sought to answer one general question: how does the language policy developed by the FLTA and by its fellows to spread Portuguese in the context of US higher education institutions work?
The research question unfolded into the establishment of two objectives: to describe the aforementioned language policy represented by the FLTA program and to discuss the agency of its fellows. To this end, we analyzed two sets of data based on the theoretical notion of language policy as a process (RICENTO; HORNBERGER, 1996;JOHNSON, 2013;, which places agency as one of the central elements in understanding this process.
The first set of data was obtained from information available on the US and Brazilian websites of the Fulbright Commission, respectively, www.fulbright.org and www.fulbright.org.br, and on the specific website of the FLTA program, https://foreign.fulbrightonline.org/about/fulbright-flta.
The data also came from documentary analysis of the calls for application for the FLTA program. This first set of data was complemented by information obtained from representatives of the Fulbright Commission Brazil via telephone and e-mail and by information obtained when one of the authors of this paper participated in the FLTA program.
The second set of data consists of excerpts from interviews that make up the corpus of the doctoral dissertation under preparation by one of the authors of this paper 1 . The semi-structured interviews were conducted using the Skype video conferencing software, in 2019, with three collaborators -Letícia, Sofia and Juliana 2 -who had completed their participation in the 2018-2019 edition of the FLTA program. These interviews lasted, on average, one hour.
We inserted this investigation into an interpretative methodological paradigm (LIN, 2015), with a view to understanding the meanings of the social actions of these agents. For that, we conducted a qualitative analysis of the corpora's texts-discourses, highlighting the topics that emerged from these texts-discourses 3 .
This study aimed to fill a gap found in a bibliographic survey based on major academic databases (Brazilian Digital Library of Theses and Dissertations, CAPES Database of Theses and Dissertations, Scielo and Google Scholar): the absence of works describing the FLTA language policy conducted by the Fulbright Commission in the process of acquisition of foreign/additional languages 4 in the USA and addressing the political-language performance of FLTAs in the teaching of Portuguese in that country. Therefore, this study was an effort to share information about a program that intervenes in the Brazilian Portuguese language policy in the USA, contributing to the dissemination of where Portuguese teaching in that country stands.
In order to achieve the proposed objectives, we have organized this paper in four parts.
Following this introduction is a theoretical background section that drew a parallel between language policy as a process and the concept of agency. In the third section, we analyzed the data in order to describe the FLTA Program and to describe and to discuss the role of FLTAs as agents of policy language. Lastly, we presented our final remarks.

Theoretical background: language policy and agency
Although focus on the agents became systematized after the notion of language policy as a process came to be (RICENTO; HORNBERGER, 1996;JOHNSON, 2009;JOHNSON, 2012;JOHNSON, 2013;, the idea that language policy incorporates the steps of formulation, implementation and evaluation has already been present in some way in the area of Language Policy and Planning (LPP) since the 1970s (RICENTO; HORNBERGER, 1996).
A highlight from the late 1980s is the notion of acquisition planning -an increase in the number of users of a language -developed by Cooper (1989), wherein language teaching was given center stage. In his description of it, Cooper (1989) exposed different planners ranging from the Minister of Education of a nation-state to teachers in the classroom, emphasizing the role of the latter.
His view is based on Clifford Prator's ideas, to whom language teaching is an object of language planning, an aspect that was not common in the definitions seen at that time. Prator pointed out that [t]he entire process of formulating and implementing language policy is best regarded as a spiral process, beginning at the highest level of authority and, ideally, descending in widening circles through the ranks of practitioners who can support or resist putting the policy into effect" (personal communication). (PRATOR apud COOPER, 1989, p. 160) These steps of "formulation" and "implementation", the "spiral" metaphor and the existence of different "levels" and "professionals" that have power to "support" or "resist" a given language policy sheds light on the perception of language policy as a process and not just a product materialized in texts-discourses. Based on Prator's notion that language planning encompasses "[...] decisions concerning the teaching and use of language [...]" (PRATOR apud COOPER, 1989, p. 31), Cooper described teachers as formal agents in decision-making to influence language behaviors. In this sense, the selection of a textbook and the definition of teaching and learning objectives, for example, materialize the deliberate efforts of these agents (COOPER, 1989).
To describe acquisition planning, Cooper (1989) focused on two variables: the goals of language teaching (acquisition, reacquisition and language maintenance) and the methods used to attain these goals (opportunity to learn, incentive to learn, and both opportunity and incentive to learn a language). Cooper's (1989) ideas about acquisition planning allow us to converge with García and Menken's (2010) view that the role of teachers is "under-theorized". In other words, although Cooper (1989) does mention teachers as important agents in acquisition planning, emphasis is placed first and foremost on the goals of language teaching and the means to attain those goals rather than on teachers themselves.
In the 1990s, Ricento and Hornberger (1996) created the onion metaphor and argued that agents, levels, and processes are layers that together compose the LPP whole. These authors focused on all those who are part of English teaching practices, such as teachers, educational program developers, materials and textbook writers, scholars, among others. Ricento and Hornberger (1996) also pointed out the idea that each of these layers " [...] permeates and is permeated by the others" (RICENTO; HORNBERBER, 1996, p. 408). Thus, agents interpret and implement language policies in different contexts, and these processes do not take place in a neutral manner, as they are located in different institutions that (re)produce discourses loaded with ideologies. Ricento and Hornberger (1996, p. 417) placed the teachers "at the heart of language policy (at the center of the onion)" because "[t]hey [the teachers] are policy transmitters and can become policymakers if they so desire". Although Ricento and Hornberger (1996) used "they" to refer to all the professionals involved in English teaching, they pointed out that teachers' agency happens when they define what, how and why to teach, since all these aspects are issues of language policy. When scholars claim that the policymakers' agency can be understood as a conscious choice, they are not dismissing agency as involuntary either. For Ricento and Hornberger (1996, p. 402), "[...] English language teaching (ELT) professionals are involved in shaping language policy, whether consciously or unwittingly". In other words, a language policy agent does not necessarily need to be aware of being one in order to act as such. Hence, it is possible to conceive two kinds of policies: the ones that develop spontaneously, even by chance, and the ones that are made deliberately.
In 2009, Johnson (2009) resumed and deepened the procedural and agentive view of language policy. According to Johnson (2009), language policy takes place from higher levels of authority (a nation-state, for example) to grassroots organizations (schools, for example). In these different concentric layers, agents are involved with the creation, interpretation and appropriation of different language policies, over which they act in many ways. In 2012, E. Johnson (2012) also added instantiation as one of the steps of language policy. According to Johnson (2013), creation refers to how, why and for what purpose a specific language policy was developed. Therefore, looking at this activity implies observing the broader socio-political processes that led to its creation, the creators' motivations, the intertextual and interdiscursive relationships established with other present and former policies, the language ideologies that they generate or that are generated from them, the circulating or conflicting discourses within them, the reflection of the intention of their authors in the final text, among other factors (JOHNSON, 2013).
Interpretation, on the other hand, is closely related to the appropriation of the language policy, since any appropriation implies a previous interpretation. Although everyone has the right to interpret a given language policy, only a few interpretations are privileged when appropriated . Therefore, interpretation is present from when a language policy is first created. After being publicized, this policy becomes subject to interpretation not only by its creators, but also by all other agents involved in it (JOHNSON, 2013).
Traditionally, the appropriation step of a language policy is considered to be a sequence of the creation step. For this reason, appropriation was investigated basically with the purpose of verifying whether a given policy had been successful or not. In the latter case, ways were sought to ensure that it corresponded to the purpose for which it was created. By introducing the concept of appropriation, Johnson (2013) ratified and theorized the possible actions taken by the agents in this transition from interpretation to appropriation, such as acceptance, rejection and negotiation. Note that Johnson (2013) did not create the idea of appropriation. In the end of the 1990s, Ricento and Hornberger (1996) had already presented a critique of the view of teachers only as implementers of a language policy, and claimed that there are language policies that can be initiated at grassroots level.
In addition, Ricento and Hornberger (1996) emphasized teachers' agency, presenting the possibility that they may or may not reproduce the ideologies that are part of the language policies. For Ricento and Hornberger (1996, p. 417), [...] teachers may internalize normative social attitudes toward speakers of nonofficial languages or nonstandard varieties of official languages, or they may believe that bilingual education programs disadvantage language minority students.
Finally, there is the instantiation step proposed by Eric Johnson (2012)

and incorporated into
David Johnson's definition (2013). Instantiation refers to the transition to language uses that results from the other steps that are part of the language policy proposed by that author (creation, interpretation and appropriation). For E. Johnson (2012, p. 58), instantiation occurs "through the actual instances of language use by individuals within a given policy context".
As we have seen, Johnson (2009) returned to the notion of language policy as a process consisting of the steps of creation, interpretation, and appropriation, and to them, E. Johnson (2012) added the step of instantiation. For Sousa, Pereira and Vilar (2019, p. 201), "[t]hese consequences of the onion metaphor were relevant to the extent that they reinforced the role of agents (the replacement of implementation with appropriation) and language practices (instantiation)" 5 .
In summary, the proposal of Johnson (2013) deepened the view of language policy as a process, which, far from being considered as a product to be consumed by passive recipients, starts to be understood as a dynamic activity filtered by several agents, capable of creating, interpreting, appropriating and instantiating it. For this reason, Johnson (2013) emphasized the role of these agents, empowering them. According to Johnson and Johnson (2015), some of these language policy agents can be seen as "arbiters", when they have more decision-making power with regard to language policy than others. This negates the view that everyone would have the same possibilities of intervention in the process. Thus, some agents assume the role of implementers while others act as arbiters.
In relation to teachers' agency, Shohamy (2009) argued for a "linguistic activism" of these actors, as they cannot be seen only as teachers of specific languages (Portuguese, Spanish, English) or as "soldiers of the system". Therefore, Shohamy (2009) claimed that these professionals need to be involved in the making of language policies. In this perspective, they need to be aware of their role in relation to decisions about languages and their uses, which would be more easily achieved if knowledge of how to influence educational language policies were inserted in teacher education curricula (SHOHAMY, 2009

Portuguese in the USA
The data was analyzed in this section based on the theoretical background discussed above.
Our analysis tried to achieve the following objectives: to describe the language policy represented by the FLTA program and to discuss the agency of its fellows to spread Brazilian Portuguese in US universities and colleges. Due to the objectives and the different nature of the data, we divided the discussion into two subsections: the first follows a more descriptive methodological approach, and the second, a qualitative-interpretative approach.

The FLTA program as a language policy
The analysis in this subsection was guided by the following questions presented by Johnson (2013, p. 224-225): "What were the sociopolitical and historical processes that led to the creation of a language policy?", "Who are the policymakers and what were their intentions?", "What is the goal of the language policy?" and "What language ideologies engender, or are engendered by, this language policy?". We also sought to map the location of the agents involved with the FLTA Program.
The Fulbright Commission was created in 1946, shortly after the end of World War II, by the United States Congress. The Commission was Senator James William Fulbright's (1905Fulbright's ( -1995 proposition and it was thought of as a foreign policy action that would allow exchanges between the Americans and people from different nations of the world. According to Toscano (2014, p. 26), this type of foreign exchange policy is an effective and "long-term strategy for exercising soft power resources" 6 , understood as "the ability of a State to attract and persuade others" 7 .
Funded primarily by the US Department of State, specifically through the Bureau of Educational and Cultural Affairs, and administered by the Institute of International Education (IIE) 8 , 6 In the original: "Sob essa perspectiva, o intercâmbio internacional é uma estratégias (sic) eficaz e de longo prazo de exercício dos recursos do soft power" (TOSCANO, 2014, p. 26). 7 In the original: "O termo soft power (…) está relacionado à habilidade de um Estado em atrair e persuadir outros" (TOSCANO, 2014, p. 26). 8 According to the website https://www.iie.org/, the Institute of International Education (IIE) was created in 1919. It is an organization that aims to establish political, economic and cultural collaborations between students, scholars and the Fulbright Commission operates in more than 160 countries and annually grants 8,000 scholarships to US citizens and non-Americans to study, teach or research, respectively, inside and outside the USA. Governments, corporations and foundations from different places in the world also provide financial support to the Fulbright Commission. In the more than 70 years of the Commission's existence, around 400,000 people have benefited from it with scholarships. The Fulbright Commission stands out for having among its alumni more than 60 Nobel Prize winners.
Thus, we place the Fulbright Commission -an institutional agent and a language policy arbiter -at the outermost layer of the onion (JOHNSON; RICENTO;HORNBERGER, 1996). The FLTA, one of the numerous cultural and educational programs organized by the Fulbright Commission, is a language policy that aims to promote the spread of languages in US higher education. The program began in 1968, more than two decades after the creation of the Fulbright Commission, when a small group of French teachers was sent to the USA to teach their mother language 9 .
The main goal of the FLTA is to increase the number of US higher education students who master at least one foreign/additional language. To achieve this goal, the FLTA brings young teachers from all over the world to the USA to teach their mother languages in higher education institutions. This teaching experience lasts one academic year. It begins in August or September and ends in April or May.
According to the Fulbright Commission's website, in the first years of the 21 st century, the US Department of State increased funding for the FLTA program, allowing for additional grantees.
Currently, more than 300 young teachers from 60 countries around the world go to the USA every year to teach 40 languages that are divided into two groups by the Fulbright program. The first group is formed by 4 languages considered traditional languages: German, Spanish, French and Italian.
The second group has 36 languages that are identified by the FLTA program as the less commonly taught languages -(LCTL): Arabic, Chinese, Russian, Zulu, Portuguese, among others.
Besides the Fulbright Commission, other institutions and, consequently, many agents, are involved in this process both in the USA and in Brazil. On the American side, we have higher institutions worldwide and it is in charge of managing some of the world's most prestigious exchange programs. It has offices in all continents nowadays 9 The Fulbright Commission is also responsible for the English Teaching Assistant Program (ETA) that provides Americans the opportunity to teach English in many countries, including Brazil. education institutions, which welcome young teachers mainly in their foreign language departments.
In these spaces, the FLTAs' activities are supervised by a professor, usually a professional with experience in teaching foreign languages, representing an agent that can affect the performance of Portuguese teachers.
In Brazil, the institution responsible for FLTA program was initially the Instituto Brasil-Estados Unidos (IBEU) of Rio de Janeiro. Since 2007, the selection process for young Portuguese teachers According to the website of the Fulbright Commission Brazil, the requirements to apply for the FLTA are: • being a Brazilian national and not a North American national; • living in Brazil at the time of application and throughout the selection process; • having a bachelor or teacher's degree in Portuguese and/or English earned in the last five years; • having high-level English proficiency as proven by a large-scale exam; and • not being a current or past fellowship recipient for a similar program.
By allowing the participation of graduates with a major in Language and Literature who do not have a teacher's degree, the Commission demonstrates that it adopts an ideology according to which knowledge of linguistics and literature is sufficient to teach Portuguese in higher education.
The selection process begins in the first half of one year and ends in the first half of the following year. This process involves steps such as the submission of documents, letters of recommendations, and oral and/or written texts, which serves to introduce the candidates and their action plans in the event they are granted a fellowship, and interviews by phone or digital platforms.
As stated on the Commission's website, the benefits for the participant are: • monthly stipend that varies based on the city where the activities will be carried out by the grantee; • paid visa fees, airfare and health insurance; • participation in guidance seminars in relation to teaching work and daily life in the USA.
These take place in Brazil pre-departure, and in the USA before the beginning of the teaching activities.
Example 1 below shows the role of the FLTAs according to the program's website. FLTAs spend one academic year at a U.S. college or university. They: • provide a youthful, up-to-date component to foreign language classes. • enhance the teaching by U.S. professors or teach under their supervision.
• make studying foreign languages in the United States more meaningful and real.
• share their culture and values outside of the classroom by joining clubs and becoming involved in the community. • engage in academic study that will enrich their postgraduate studies and enhance their professional development. • return home with a high proficiency in English.
• can speak first-hand about the United States, its culture and its people.
The first sentence in Example 1 is an indication of the target audience of the FLTA program's language policy: students from US colleges or universities. These students are from different majors.
Some students choose to enroll in Portuguese courses because they have an interest in this language while others enroll to fulfill a university requirement to attend a foreign/additional language course. Most of these students have basic knowledge of Spanish and others are proficient in Spanish. For these students, Portuguese will be the second foreign/additional language. Even at elementary-level Portuguese courses, there are students who already master Portuguese either because they used to live in countries where this language has official status, mainly Brazil, or because Portuguese is their heritage language 10 .
Young teachers can take on one of the following roles in teaching the Portuguese language: being a teacher's assistant (TA) or being the main instructor, as pointed out in the list presented in Example 1. FLTAs are seen as ambassadors of their countries in the USA. For this reason, they are supposed to help to promote the culture and values of their homeland, through the organization of clubs and cultural activities, such as exhibitions of films, dances, sports, etc. Therefore, it is clear that the FLTAs need to have a broad view of what it means to learn a language/culture, as they will promote the teaching of Portuguese not only within the classroom, but throughout community in which the university or college is located.
Items five to seven in the list in Example 1 also describe another assignment of the fellow: to attend at least four subjects at undergraduate or graduate level -three of their choice and one related to some aspect of the USA (history, culture and politics, among others). In this process of involvement with the US academic life, many FLTAs participate and organize congresses. Every According to items five, six and seven presented in Example 1, the Fulbright Commission understands that the experience of one academic year of FLTAs in the USA will have an impact on their English language proficiency level, their professional development, and their expansion of cultural knowledge about the USA. In our view, this expectation also implies that agents will qualify to be capable of spreading English language and US culture when they return to Brazil. Therefore, we noted that the FLTA program also encompasses a language policy aimed at the education of English teachers for Brazil.
In the following paragraphs, we will discuss some unpublished data about the FLTA program  Table 1 shows the number of grants awarded, and of US higher education institutions and states that hosted Brazilian FLTAs in 4 of the 12 editions of the FLTA Program in Brazil. As the column "Number of grants awarded" of Table 1 shows, in the two most recent editions of the program, the number of participants, 20, corresponded to less than half of the number of fellows in other calls, such as that of 2013-2014. In all four editions shown in Table 1, 117 Brazilian teachers worked at 58 higher education institutions (universities and colleges), in 30 American states and in the federal capital, Washington D.C. Some of these institutions are on the lists of the best universities in the world.

The reduction in the number of grants is probably related to financial issues that the Fulbright
Commission is going through. According to the Fulbright Commission's website, the maintenance of the same amount allocated to its actions by the US government in the past ten years has already resulted in a 21% decrease in the Commission's financial power.
According to Garcia (2011), the reduction in the amount allocated for the Fulbright Commission by governments of different nations also explains the financial problems that this organization goes through. The analysis of the Fulbright Commission's annual reports show that foreign governments' support increased in 2013 and 2014. But, in general, the Fulbright Commission has faced a decrease in this financial support (J. WILLIAM FULBRIGHT SCHOLARSHIP BOARD, 2011BOARD, , 2012BOARD, , 2013BOARD, , 2014BOARD, , 2015BOARD, , 2016. Besides those reasons, the US Government also proposed budget cuts. In 2017, the Stand for Fulbright campaign, widely publicized on social media, including on the Commission's website, using the #StandForFulbright hashtag, was created to prevent a 47% budget cut in 2018 (FULBRIGHT ASSOCIATION, 2020). The campaign's actions included petitions, rallying support for the cause of US politicians and holding meetings to discuss the impacts of the activities of the Fulbright Commission in the USA and worldwide. In the following years, the campaign was conducted again to prevent budget cuts. According to the petitions available on the Commission's website, the efforts to avoid budget cuts were successful (FULBRIGHT ASSOCIATION, 2020).
In summary, the project proposed by Senator James William Fulbright is facing serious financial problems. This makes it difficult for the Fulbright Commission to comply with the actions that were already developed. In addition, there have been constant threats of budget cuts, which, if implemented, would permanently render the activities of the FLTA and other Commission's programs unfeasible.
Based on the onion metaphor proposed by Ricento and Hornberger (1996) to characterize the layers, levels and agents involved in the language policy process, Figure 1 shows the language policy for the spread of Brazilian Portuguese in higher education institutions in the USA carried out by the FLTA Program. Finally, at the center of the onion, we place the FLTAs, who are young Brazilian teachers in the context analyzed here.

The Brazilian FLTAs have the floor
This subsection analyses the texts-discourses of the agents located at the "epicenter" Example 2: Excerpt from interview with FLTA Letícia -collective agency Letícia: The recommendations and orientation that we had before coming to the United States were related to the country's culture, for example, more or less what to expect of teaching in a university in the USA. So I was talking, of my own free will, I was talking to former FLTAs, I asked them how it was, because I had never taught Portuguese as a foreign language before this experience. It was new to me, you know? So, I spent some time searching about this and asking the former FLTA how it was. There were no guidelines in the university where I was placed either [...] We looked for this type of orientation [on teaching Portuguese language] by ourselves, we were always asking people who had come before us how things worked, what kind of material was good to search for, we looked for it by ourselves and we used to share our findings with the others [...] I talked to former FLTAs. There were FLTAs placed in many universities, but I also talked to the former FLTA 13 in the university where I would be placed 14  . We used to share many things and it was really important to have that basis and that support. Source: Semi-structured interview In her answer, Letícia reflects on the background of teachers selected for the program. She highlighted the feeling of doing something new: like for many fellows, it was her first experience teaching Portuguese to non-native speakers of this language. According to one of the author of this paper that is a former fulbrighter, some FLTAs are either Portuguese or English teachers or both, and few work specifically teaching Portuguese as a foreign/additional language, something which usually changes after they return to Brazil.
It is worth noting that courses related to Portuguese as a foreign/additional language, if any, are usually not mandatory in Brazilian undergraduate courses with majors in Language and Literature, as shown by the study conducted by Dionísio & Costa (2016). Some FLTAs have teaching and research experience in Portuguese as a foreign/additional language, and some have experience teaching Portuguese as a mother language and/or English as an additional language. According to the calls for application, people who have a major in Language and Literature but do not have teaching certificates can also be selected to be an FLTA grantee. Due to this fact, we might find FLTAs who do not have previous experience with language teaching. In spite of not having this experience, in the electronic forms during the selection process, the candidates for an FLTA grant are required to answer discursive questions that focus on language teaching practices, such as, for example, asking them to list previous teaching experiences and possible actions to be developed in the USA as an FLTA.
Even those who already have experience teaching Portuguese as a foreign language also face another unprecedented aspect in their professional lives: teaching in a university in the USA.
Due to the fact that the calls for application of the FLTA require completion of Language and Literature program within the past five years, the Brazilian teachers participating in the program are young professionals who generally do not have extensive teaching experience, although some already hold a master's degree. Therefore, for the majority, the FLTA program will be their first time teaching classes in higher education. In addition, higher education is organized differently in the USA that is, academic practices are different compared to those in Brazil. Moreover, for some FLTAs, going to the USA as fellows is their first experience in a foreign country.
In order to deal with unprecedented experiences in their personal and professional lives in a relatively short period of time, given that the program lasts only one academic year, Letícia recommends that young teachers seek as much information as possible about teaching Portuguese in the USA by contacting participants of the same or previous editions. The lack of professional experience in teaching Portuguese as a foreign/additional language and the lack of guidance on how to work in this field promote the creation of informal networks for sharing practices and materials for teaching Portuguese among FLTAs, as pointed out by Letícia. As Example 2 shows this sharing was "crucial" and "very important" for Letícia's performance as a teacher at the institution where she was placed. She reveals that knowing about the experiences of other FLTAs is a recurring need for the ones selected to the program, given that this process is repeated with each new group of FLTAs sent to higher education institutions in the USA.
When there is no institutional guidance or supervision, it is often up to the FLTAs to decide what and how to teach and what teaching material to adopt or how to prepare it. Due to the construction of mutual support networks among the FLTAs, these shared materials contribute to the expansion and promotion of practices and ideologies about teaching Portuguese. Thus, these teachers become active actors in achieving the objective of offering Portuguese language courses in the institutions where they are placed, going beyond the resources available locally and establishing a language policy for teaching the language based on practices and beliefs shared among them.
Based on the interviewee's report, it is possible to see that the FLTAs' teaching agency overcomes the limitations of the Fulbright Commission institutions and operational procedures to find resources that meet their pedagogical needs for teaching Portuguese at universities and colleges.
In higher education in the USA, the teaching of Portuguese represents a tiny fraction of the total number of students enrolled in language courses, as pointed out by Milleret (2012) and corroborated by data collected by the MLA and reported by Looney and Lusin (2019). However, to obtain a subjective view of the teaching actors on the importance of the language they teach, we asked FLTA Sofia the following question: "In your opinion, what is the space of Portuguese language in universities in the USA?".
Example 3: Excerpt from interview with FLTA Sofia -restrictions on teaching work and linguistic policy arbiters Sofia: [...] talking to the others FLTAs, I realized that this Portuguese language status, it kind of changes depending on the university [...] sometimes, Portuguese language teaching was very limited just to teaching language as a system whereas others [included] activities that would link Portuguese language teaching to others perspectives, you know? Cultural perspectives related to Brazil, they were kind of hindered, you know? By people who were more... who had a certain amount of power over the FLTA's work […]. Source: Semi-structured interview.
Sofia's answer indicates that the size and representativeness of Portuguese language programs may vary from one institution to another. In addition to the immediate answer to the question, Sofia's excerpt indicates there are other aspects to the FLTAs' agency. First, the interviewee reiterates Letícia's discourse about collective agency through productive contact between FLTAs of different editions seeking to share experiences in teaching Portuguese in the USA.
Second, Sofia's answer shows that restrictions can be placed on teachers in terms of how they carry out their teaching activities. Sofia points out that, despite the emphasis on cultural orientation given by Fulbright's FLTA program, the teaching of Portuguese language in some institutions may be oriented towards a traditional approach, based on learning a language as "system" disconnected from the organic relations with the culture of speakers. Using the verb "to hinder", the interviewee reports a restriction that took place at the institutional level in relation to the adoption of cultural approaches to teaching Portuguese language with a focus on Brazilian culture, even though the FLTAs that teach Portuguese in US institutions are all Brazilian, as shown in the program description.
As agents placed at the interpersonal layer of the language policy process, FLTAs are in direct interaction with university students of Portuguese language courses inside and outside the classrooms and in constant interaction with the institutional layer, represented by the programs, departments and centers where they are allocated to. These interactions between agents that are placed in different layers is mediated by the institutional rules and objectives.
Sofia brings up the category of "power" institutionally exercised by others as regulator of the FLTAs' agency, with which these teachers have to deal. The category of power shows that language policy is a process (JOHNSON, 2013) in which several agents located in the layers of creation, interpretation, appropriation and instantiation intervene. Sofia's answer also corroborates Johnson and Johnson's (2015) claim that agency in language policy is hardly shared equally by the various actors, even with arbiters -agents with greater decision power about language policy in certain contexts. Even though Sofia does not identify the institutional degrees of the people who exercise power, she reveals that the role of FLTAs is a constant negotiation with language policy arbiters who can interfere in the work of these young Brazilian teachers.
As exemplified by FLTA Juliana's answer to the question "How did you become a teacher of Portuguese as a foreign language at a university in the USA?", when there is no institutional tradition in teaching the Portuguese language at the university where the FLTA is placed, those young teachers will have less interferences in their work.
Example 4: Excerpt from interview with FLTA Juliana -agency in the creation of a Portuguese language teaching program Juliana: In the beginning it was a bit hard, as soon as I arrived there, right? Because there wasn't a Portuguese language program and I was the first FLTA at that department [...]. So, in the beginning it was a little hard. Why? What is the FLTA program about? People who are placed in a university that had already have an FLTA grantee, this former FLTA can give them a lot of tips, help them with a lot of things, right? I didn't have this former FLTA, so I had to learn a lot of things the hard way and by asking everybody for help. I wasn't shy about asking questions. So, like, I would email my supervisor, my coordinator, the secretaries. So, I was able to get things going [...] And they would often group me with the Spanish department. When the first meeting at the university happened 15 , the members of the department would get together, right? Uh, the Spanish department... Not the department, the language program, you know? The members of the Spanish program, the Italian program, the German program, they would get together with. I would be the only one without a meeting scheduled, you know? So they added me to the Spanish program, I ended up getting used to this group and learning how the Spanish program works, right? They had a professor who was the Spanish program supervisor and I... My supervisor in reality was the head of the whole department, you know? Because there was no professor [...]. Source: Semi-structured interviews Example 4 highlights the importance of an FLTA's role in creating a university program for teaching Portuguese. Given the bureaucratic and budgetary procedures for hiring professors of "less commonly taught languages", Portuguese among them, the temporary nature of the work of the fulbrighters, whose grant does not usually come directly from the institutions in which they are placed, often turns them into "test pilots" for language education programs. In other words, the FLTAs can act as true founders of these programs, which further emphasizes their relevance as actors for the spread of Portuguese language in higher education institutions in the USA and the importance of highlighting the agency assumed by them in the language policy for spreading Brazilian Portuguese abroad. 15 We have preserved the university's anonymity.
According to Juliana's answer, the absence of a former FLTA to convey the ideological and institutional values for teaching Portuguese compelled her take on the responsibility of creating the program herself, taking advantage of the pre-existing institutional structures intended for Spanish and other languages offered by the university. We must also emphasize the presence of other agents who are not necessarily involved in the teaching of Portuguese, but who may have contributed to the work developed by Juliana, since she showed in her text-discourse that she engaged with the group of Spanish teachers.
Juliana's answer also remindful that teachers need to go beyond the understanding of their agency as being related only to the language they teach (in this case, the Portuguese language), and instead see themselves as part of a broader group of language teachers who have a social and political role, as pointed out by Shohamy (2006Shohamy ( , 2009. As argued by this author (2006,2009), this is why it is imperative that language policy courses be included in teacher education. In our view, this need is even more pronounced when we consider that after graduating teachers may find themselves in contexts for which college did not specifically prepared them. By learning about language policy, these teachers may become more capable of reflecting and transforming beliefs into teaching practices and vice versa.
Finally, we understand that, more than interpreting or appropriating a pre-existing institutional language policy for teaching Portuguese at the university where she was placed, according to the steps described by Johnson (2013), the performance of FLTA Juliana reached the heart of creation of this language policy. This can be asserted based on the preparation of a teaching program that is characterized by a given language approach, on the selection of content and language skills to be explored, on the preparation or selection of the teaching materials to be used in the classroom and on the definition of types of assessment.

Final Remarks
In this study, we sought to describe the language policy of The Foreign Language Teaching Assistant Program and to discuss the agency of its fellows to spread Brazilian Portuguese in higher education institutions in the USA. Based on the theoretical understanding that language policy is a process characterized by the existence of layers, levels and processes put in motion by different actors (RICENTO; HORNBERGER, 1996;JOHNSON, 2013;, this study is a first step in analyzing the language policy promoted by the Fulbright Commission, specifically, by FLTA and its fellows, in the teaching of Brazilian Portuguese in US universities and colleges. We outlined the circumstances that led to the creation of the FLTA by the Fulbright Commission and presented data of the program in Brazil, pointing out historical and quantitative aspects of the language policy for the spread of Brazilian Portuguese in the USA. Those aspects are related from the recruitment of fellows to some of the actions undertaken by them in the USA. We identified the levels and the main agents that act in this language policy process, which begins in the US government sphere, permeates the Brazilian government sphere, and takes shape in classroom, through the work of young Brazilian teachers sent every year to the USA to act as cultural ambassadors and agents of the spread of Portuguese. We summarized the understanding that the FLTA's process consists of different layers and different agents -governmental, institutional and individual -through an illustration based on the "onion" metaphor proposed by Ricento and Hornberger (1996).
We also reflected on FLTAs' language policy agency, based on interviews with fellows of the program. We identified that the FLTA program seems to have no explicit language teaching policy established for foreign/additional languages. The institutions, the supervising teachers and, mainly, the fellows are the ones responsible for creating, interpreting and appropriating a language acquisition policy for Portuguese. Teaching activities, despite having a wide scope, can be subject to restrictions at an institutional level, given the generalist guidelines the FLTAs receive at preparatory events organized by the Fulbright Commission. As we have seen, in a given university context, there was orientation to avoid focusing on Brazilian cultural themes coming from language policy arbiters who supervise the FLTAs' work.
We also demonstrated that FLTAs either work collectively or individually. Creating informal networks for sharing knowledge, experiences and teaching materials and promoting events to discuss the teaching of Portuguese are examples of work done collectively. Deciding what and how to teach, based on the beliefs and teaching practices of former FLTAs or other agents who are not necessarily involved in the teaching of Portuguese, as is the case of teachers in the Spanish program, are examples of individual work. The agency in the context analyzed may act as a mechanism to overcome the lack of experience teaching Portuguese as a foreign/additional language, the lack of guidance on the work to be developed and the restrictions sometimes imposed on their work by arbiters from the institutional context of US universities.
Finally, this study has two empirical contributions. First, it unveils the layers that characterize the language acquisition policy of Portuguese in higher education institutions arising from the FLTA program. Second, it focuses on the agency of the fellows who, based on the official objective to act as "cultural ambassadors" of Brazil, are vectors of a diffuse language policy to promote the Portuguese language in the USA. As this study shows, the FLTAs are important agents for language promotion in US universities. In addition, a practical contribution of this study is providing subsidies for FLTAs themselves to reflect on their teaching practice as something that is immersed in a broader socio-political context and that is as endowed with linguistic and political importance, from the pedagogical choices made in the classroom to the activities developed outside it.
Future studies on the different dimensions of the Fulbright FLTA program are needed, given the lack of academic studies about it and the importance of this program.