An Approximation to Inclusive Language in LMOOCs Based on Appraisal Theory

This article explores inclusive language as a form of verbal communication in an open online language course for refugees and migrants. Firstly, the existing evidence of the benefits of using inclusive language in education is analyzed. Secondly, the specific need to use this type of language in the context of online courses for displaced people is discussed. Thirdly, a first approximation towards the identification of linguistic resources that may impact both group inclusion and individual discrimination is attempted, based on principles and categories from Appraisal Theory. Fourthly, the presence and effects of these linguistic resources are analyzed in the materials and forums of a highly successful LMOOC of elementary Spanish for refugees and migrants. Fifthly and finally, conclusions are drawn on the convenience of incorporating inclusive language as a design element in LMOOCs for displaced people, enhance its use in forums, and train facilitators accordingly.


Introduction
The appearance of large numbers of refugees heading toward Europe became evident around 2015, when a combination of political, social and natural upheavals caused numbers to rise drastically. This situation has not greatly improved since then and, according to UNHCR data, over 68.5 million people are forcibly displaced worldwide, of which 25.4 million are refugees. Some of these people managed to reach the destination country of their choice, especially before European borders began to close, while others are stuck in holding camps along the way. For any of these people to integrate in these countries, understanding and speaking the language of the community in question is a must. While education should be a priority for them, given their often precarious economic and social situation, and lack of documents and certificates necessary to demonstrate prior academic achievement, entering formal institutional education is mostly not possible. In some cases, thanks to non-governmental organizations (NGOs), not-for-profits, refugee support groups, etc., they are able to obtain some schooling in relevant languages, culture-specific topics, and administrative procedures. However, the demand is far greater than the supply.
Since the majority of these refugees do have access to some kind of mobile device, a smartphone or tablet, then online learning could appear to be a solution to such imbalance. Although by personal choice, refugees not surprisingly prefer face-to-face (F2F) teaching over online learning (Kamyaba 2017), given their sense of isolation and the far-from-perfect studying conditions in refugee camps. However, where scale, cost, and general effectiveness are concerned, e-Learning can be seen to be a valid option.
Massive Open Online Courses (MOOCs) appeared on academics' radar around 2008 and, since then, have been applied to a wide range of social groups and learning materials (Liyanagunawardena et al. 2013, Spanish, mainly by the teachers and facilitators, but also by the students, as part of an examination to the nature and effects of the use of inclusive language in an open online course, and how it can be explicitly incorporated as a fundamental design element in such courses.

Theoretical background
Refugees and migrants, as well as trying to grapple with the practical realities of life in a new country, such as finding shelter, learning the language, encountering employment, etc., also need to integrate into the society in which they find themselves. Language can be seen to play a key role in this process of social inclusion (Yates 2011). Developing the relevant linguistic skills is, therefore, a goal for all newly arrived refugees and migrants. This is the case not only for communicating one's immediate needs, exchanging basic information with native speakers, and correctly accessing and using administrative and societal resources, but also for establishing interpersonal relationships and moving towards becoming fully participating members of society. After all, language represents a tool for mediating affective relations and social cohesion (Fusaroli & Tylén 2012). However obvious it may appear, the reality is that this tool is not fully potentiated by the host society or by the refugees themselves. This is particularly evident in academic contexts, where language should be carefully sculptured and scaffolded to enhance learning.

Inclusive language in education
The use of inclusive language in education is well documented in the literature, since arguably the educational process is one that changes peoples' knowledge, opinions, and values for the better. Hence, the didactic process is clearly an optimal moment for establishing inclusive language practices.
Considerable progress has been documented in the academic literature about the supportive role of inclusive language regarding personal conditions and self-identity choices, like the ones mentioned in the previous section. In the case of gender issues, Bacon & Finnermann (1992) used discriminant analysis as a statistical tool to demonstrate that there was a significant difference relating to authentic input and exposure variables between men and women and their self-reports of attitudes, beliefs, strategies and experience in language learning. Almost two decades later, He (2010) analysed sexism in English, arguing that it reflects underlying cultural biases and socially unjust attitudes towards women. This author noted that society itself must change for the language used therein to be updated as one of its core elements. Regarding education, Beebe (1996) noted the presence of exclusive (sexist) language in English language teaching for historical and cultural reasons, argued why it was a problem, and concluded that if such language is to be avoided with students, then awareness must be encouraged in the teaching staff. Almost two decades later, Koeser & Sczesny (2014) presented the results of a study that demonstrated that speakers change their language behavior to use more gender-fair (inclusive) language after having been taught arguments supporting the use of such language.
Regarding the difficulties of educating people with disabilities, Wiebe Berry (2006) undertook a study of how teachers create an inclusive classroom for students with disabilities. The author found that teachers use discourse strategically to encourage participation and collective responsibility. However, while this was proven to be effective at a class level, inclusive language was manipulated to reduce inclusion in small-group working sessions. Care is needed to ensure that students with disabilities fit in the educational process in such a way that their conditions do not limit their integration or learning. Tsehelska (2006) argued that teaching politically correct language to English students exposes them to different cultural issues, thereby influencing the language they learn to use. Similarly, Alexander & Banks (2004) undertook an analysis of the literature on computer-assisted writing pedagogies related to issues of sexuality and discrimination. They noted that this approach to teaching and learning can promote critical and rhetorical sensibilities in students about the constructions of sexuality in modern culture. They go on to conclude that networked communication technologies can play a key role in breaking down these prejudices. Ressler & Chase (2009) noted the importance of providing educators with information about the use of inclusive language as part of a holistic approach to what they call "gender variance" in the student community, to support them in their studies and also help other students develop a broader perspective of inclusion. Delargy (2006) noted the importance of the use of inclusive language as part of an array of strategies to teach in multicultural schools. This author focused on "English as an Additional Language" that is taught to all children who do not have English as their native tongue, and presented a series of best practices for facilitating the social inclusion of native and non-native school children (for example, the children were asked to use "object boxes" related to themes such as food or transport. Small culturally-mixed groups of children then played with the contents of the box and the English native-speakers used the language to support the others and facilitate their integration). Alexander (2015) noted the existence of demographic data suggesting that the population in America will have a composition of 50% minorities by 2050. Presumably, similar figures can be expected in European countries then. Lotherington (2010) noted, in the case of Toronto, that 64.2% of children spoke a language other than English at home. The author presented a project that demonstrated how inclusive language use can be scaffolded in the classroom, both inside and outside the curriculum. Collaborative learning was undertaken by students who combine semiotic modes in narrative learning via complex multimodal retellings using, amongst other things, art, music, drama, dance, puppetry, voice-over narration, etc. This approach used different communicative modalities for inclusive, collaborative literacy learning, stacking linguistic modes in bilingual and multilingual "channels".
More recently, Kulatz & Iversen (2019) have presented a similar project that increases multilingual literacy engagement for newly-arrived immigrant students in Norway as part of the development of an inclusive language classroom. Data was collected for 14 adolescent students speaking 15 different home languages by using a language use questionnaire, student logs, students' identity texts, lesson plans, and the teacher's notes and reflections. The results suggest that including all languages in students' repertories can help build inclusive classroom spaces with multilingual identities. As Arber (2010) had noted, teachers need to make efforts both to include their international students and, at the same time, respect their differences. This author emphasized that English language education plays a key role in this process and is more effective when notions of language and culture are simplified.
Displaced people arguably represent a superset of individuals in need of inclusive education, since the abovementioned personal conditions and self-identity choices are not mutually exclusive with the circumstance of being a refugee or a migrant. Therefore, they are open to being supported using inclusive language as per above and, at the same time, are also a collective with its own heterogeneous and complex sociocultural profile which needs to be scaffolded accordingly. This will be discussed in the next section.

The need for inclusive language in MOOCs for refugees
As summarized on the Coursera blog1, an inclusive course (one with an appropriate social, emotional and intellectual climate, in their own words) enhances the effectiveness of learning (cf. Ambrose 2010), since students feel at ease, lower their emotional filters and, therefore, find it psychologically easier to take risks and make mistakes. Wlodkowoski (2008) noted that fear and uncertainty, often common in refugees' mental state, can slow learning and lead to a sense of frustration, aggression and withdrawal. There is a range of factors that can be controlled in an online course like a MOOC to make it more inclusive; for example, providing rules on forum participation to prevent disrespectful comments in content or form (often referred to as a "netiquette") and open up the conversation to all students.
Until now, several projects have focused on different applications of LMOOCs for refugees and migrants. In MOOCs4inclusion, a catalogue of free digital learning initiatives was produced that includes a range of different courses for language learning (http://moocs4inclusion.org). Examples of projects which have led to LMOOCs included there are: Afpa, that has developed an introductory French as a foreign language MOOC for social inclusion; MINGLE (Migrant Language and Social Integration), that offers two Greek language courses for Bulgarian migrants willing to work in Greece or Cyprus; Dutch for Arabic speakers, that offers an A1 language course specifically for asylum seekers of this profile; OEAD4Refugees, that offers courses on German to refugees wanting to subsequently enter higher education. As can be appreciated, the LMOOCs in question here focus on introductory European language training mostly for Arabic natives.
Other initiatives focus on MOOCs for refugees in general and, while they do not only include LMOOCs, they have some of these courses on offer; for example, Coursera for refugees (www.coursera.org/refugees/ for-refugee-learners) and Education without borders (www.universitaperrifugiati.it/en/corsi-di-lingua. aspx). Also, the major MOOC platform providers have LMOOCs (e.g., edX, FutureLearn, MiriadaX) as part of their course portfolio. While these MOOCs could be undertaken by refugees and migrants, and some have been, general contextual problems (cf., Stevenson & Willott 2007, Crea 2016, Crea & Sparnon 2017 and the lack of specific scaffolding for displaced people has made them (and other online courses) less attractive than they are for a general Western audience (Kamyaba 2017, Reinhardt et al. 2018, Witthaus 2018. There are also research projects which, while they have not developed any MOOCs, they have also focused on the pedagogic and sociocultural aspects of designing LMOOCs in general and also for displaced people. Examples include: the LangMOOCs project (www.langmooc.com), that presents the analysis of the state of the art of LMOOCs and a pedagogic handbook on how effective courses can be developed; and the above mentioned MOOCs4inclusion, that also published a report in this sense . Finally, there is also academic literature on the general application of MOOCs to the educational needs of displaced people and their social inclusion, like the work of Zhang (2013), Jemni & Khirbi (2017), and Alonso & Samy (2018), which highlights some of issues related to the use of open education for this collective. In parallel with this collaboration, related research was undertaken by the group members on subjects like digital alphabetization and digital literacy for the typical refugee population who come from the MENA (Middle East and Northern Africa) region (Traxler 2018. What became evident during this process was that teaching people from the MENA region European languages required not just the acquisition of a different vocabulary and grammar, but also a fundamental change in their underlying reasoning. As Kaplan (1966) had noted for English (but is arguable extendible to other European languages), it is characterized by directness and deductive reasoning, while other languages like Arabic favor indirectness and inductive reasoning.
Furthermore, given the difficulties that a lot of people have during the migration from their home countries, and how they are treated along the way, it is necessary for them to overcome the negative psychological states in which they often fall. The role for inclusive language in this process is evident, since as Argyropoulou & Ypsilandis (2017) noted, it can promote a sense of belonging, solidarity and responsibility by both those who use it and receive it. Furthermore, in the findings presented by Jhaveri (2018), exposure and positive reinforcement in academic communication, including open online courses, is one of the most effective ways of helping students adopt the use of inclusive language, which arguably propagates and potentiates, in turn, its use in the refugee community.
While the initiatives presented in this section demonstrate that some effort has been given to the design/application of LMOOCs for refugees and migrants, and some of them do simplify and adapt content and teaching methods accordingly, no focus has been given to the use of inclusive language for promoting integration and learning. The authors of this article defend the hypothesis that the use of inclusive language in LMOOCs can support and facilitate refugees and migrants, directly and indirectly, in the following ways: -By facilitating the participants' social integration as part of their learning process. -By modeling the use of inclusive language when communicating with others, so that it is propagated for future interaction. -By shaping the way that society views their presence in a given country.

Toward a theoretical framework for inclusive language: Appraisal Theory
The authors claim that the use of inclusive language in the teacher-learner and learner-learner interaction within an LMOOC leads to an iterative process of personal (re-)interpretation of the other's contributions and determines a mutual positive emotional reaction. Brown (1991) noted the importance of affective variables in second language learning and, further on, for effective communication in social groups. He went on to argue that the combination of empathy, extroversion and assertion may facilitate the merging of cognition and affect, which vary within individuals and may partly account for the varying degrees of success in learning a second language. In a more general sense, as noted above, Wlodkowoski (2008) related the degree of negative emotions present in a student to the difficulties that s/he had in learning. As Foolen (2015) argued later, there is a clear and direct relation between emotion and language. Following these arguments, it seems likely that it is the appraisal of the affective tone or quality of inclusive language and the underlying conceptual structures it activates, that potentiate a student's learning and helps establish the relationships that build social inclusion.
Appraisal Theory, derived from Systemic Functional Linguistics (Halliday 1985a, Halliday 1985b, Martin 1995, Martin & White 2005, Ananda, Nababan & Santosa 2018, focuses on how language conveys positive or negative assessments, and how attitude and emotion categories its proposals and propositions interpersonally. It provides a framework that enables meaning to be analyzed in terms of three broad dimensions: Attitude (positive/negative assessments, defined in terms of affect, judgement and appreciation, whereby speakers associate emotional responses with objects and other people), Graduation (force and semantic focus, by which speakers raise or lower the interpersonal impact, intensity or volume of the evaluation), and Engagement (mono-and hetero-gloss, as resources of dialogistic positioning that define propositions and proposals conveyed by language). That is to say, this theory is concerned with the linguistic resources by which speakers come to express, negotiate, emphasize, neutralize or naturalize particular (inter-)subjective and ultimately ideological perspectives. The authors claim that Appraisal Theory can offer an adequate appropriate framework that can be extended for modelling inclusive language. Therefore, a set of resources or variables have been identified to play a cognitive and affective role in inclusive open online academic interaction. They have been grouped around a complex axis: elements for favoring social inclusion/avoiding exclusion, elements for positive/negative individual discrimination, and, within them, elements related to Attitude, Engagement and Graduation, as identified in Appraisal Theory. These are presented in table 1 below. Table 1 contains a first approximation to an illustrative list of linguistic resources that influence social inclusion and personal discrimination and has been developed from a synthesis of resources found in the Appraisal Theory literature (Martin & White 2005, Ananda et al. 2018, complementary literature on inclusive MOOCs (Marrone et al. 2013, Sanchez-Gordon & Luján-Mora 2016, and the authors' own research, together with the testimony of the language teachers at refugee learning centers over a series of F2F meetings in the context of the MOONLITE project (Read, Sedano & Barcena 2018). As can be seen, some resources are more conveniently expressed affirmatively (e.g., "Using…") and others negatively ("Avoiding…"). In some cases, the feature would be closely similar either way (e.g., "Using an empathic and conciliatory tone" vs. "Avoiding an aggressive or accusatory tone"). A convenient conventional division into linguistic levels (based on the CEFR's terminology [Council of Europe 2001]) has been adopted to classify the resources for methodological purposes, although the underlying force of inclusion, respect and mutual understanding in the resources often trespasses a single level. The authors have adopted a consistent approach to their classification and assignment although they accept that there could be scope for other interpretations due to the interconnected nature of linguistic levels. Although for the most part the items do not statically correspond to a linguistic level, the level at which a language is best seen to offer most resources to convey the above force is pragmatics. Some of them are relative to social and individual discrimination and are more specific of social groups that are either excluded or at the risk of exclusion, while others refer to general interpersonal communication and are directly influenced by Appraisal Theory dimensions. A few examples of the latter are: passing positive judgments (Attitude), using positive reinforcement (Engagement) and graduating interpersonal impact, force and volume of one's utterances as suitable in the verbal negotiation (Graduation). It should be noted that while some features are cross-linguistic (e.g., using question tags), others are language-specific (e.g., avoiding the use of derogatory words by derivation, for Spanish diminutives have a negative connotation when applied to nouns and used beyond their primary meaning of "smaller in size, age or importance"; using informal and formal second person pronouns, which is of particular importance in conversational politeness in French). Associating emotional/affectual responses with participants and previous performance/processes (e.g., appreciation, empathy, reinforcement) • Making reference to something that the interlocutor has previously said, either to them or to third parties • Using reassurance and positive reinforcement • Using resources to negotiate an interpersonal space for one's own positions within the diversity of view points put at risk by an utterance (e.g., modals of probability, reality phases, attribution, proclamation, expectation/counter-expectation) • Graduating interpersonal impact, force and volume of one's utterances as suitable in the verbal negotiation (e.g., assertive, empathic and encouraging speech, emphasizers/diminishers, imperative/interrogative)

SOCIOLINGUISTIC LEVEL
• Awareness of situational context or culture-driven language • Avoiding vertical analyses of cultural and personal behaviors and perspectives • Avoiding discriminatory comparisons between cultural and personal behaviors and perspectives • Using appropriate levels of formality (e.g., politeness formulae) to convey respect/closeness/intimacy The authors claim that the use and potential effect of the resources for conveying inclusion within a language should be explored in LMOOCs for refugees and migrants, both in the elaboration of course materials (guides, activities, tests, etc.) and in forum participation. The next section analyzes the presence/ absence and instantiation of these resources in a set of two highly successful LMOOCs, and the types of reactions elicited by them.

Methodology
The methodology followed in this research is eminently experimental, or more specifically, quasiexperimental, due to the non-formal nature of the situational context in which it took place (starting, for example, with the actual way in which people signed up to participate in the LMOOCs which, by their very open nature, was not under the research group' control) and also the heterogeneous profile of students in open education and of refugees themselves.
The methodology used followed a mixed approach (Robson 2002), as the evidence was collected with a combination of quantitative and qualitative tools, namely, data obtained from online observation (e.g., user logs, forum and chat interactions, activity results), questionnaires, and interviews, with a thematic analysis of the qualitative data (Kothari 2004). Questionnaires were sent electronically to students before the courses began and after they finished. They included both open-and close-ended questions. Post-questionnaires were followed by selective semi-structured interviews of subjects that manifested specific behaviors. The MOOC platform at UNED, the Spanish distance learning university that hosts the two LMOOCs under study, is Open edX, which has the edX Insights data analytics tool installed. This was used in combination with data files that could be downloaded in CSV format from the platform. It should be noted that the study undertaken here using the above-mentioned data is an exploratory one in the sense that it is a preliminary analysis of the presence of inclusive language features in the LMOOCs under study. In terms of the reliability and validity of the examples presented in the results section, it should be noted that they are illustrative, not representative.
As for the open language learning pedagogy, the approach was essentially heuristic, namely, teachers and facilitators using e-moderating methodologies (Salmon 2000) to structure and scaffold student interaction in order to encourage optimal engagement and maximum learning.
The study undertaken here focuses on the presence of inclusive language in the two LMOOCs created, specifically in terms of the characteristics presented earlier and the use of Appraisal Theory as the supportive theoretical framework for their development. As such, two research questions can be identified here: 1. What evidence is there of the use of inclusive language in successful LMOOCs aimed at displaced people? 2. What type of effects are caused by the use of such inclusive language?
In the ERASMUS+ MOONLITE project (2016-1-ES01-KA203-025731) which contextualizes the development of the two LMOOCs and the research presented in this article, the role of open online courses for supporting social inclusion and employability is being studied, specifically through the development of language competences and entrepreneurial skills. As part of this work, two consecutive online courses of Spanish for the immediate needs of refugees and migrants who arrive in Spain for the first time were developed. The courses were called Open Doors ("Puertas Abiertas" in Spanish) . In a way, Open Doors can be seen as one course divided into two parts, for the initial intention was to design and develop only one. The division of the course was proven necessary due to the amount of content that was identified to be relevant to fulfill its practical goal together with the restriction of the host institution that a MOOC follow a standard six-week structure. Hence, the decision was taken to divide the course into Open Doors I and II. The Spanish level across the two courses went from A1 through to A2+. Neither of the courses obviously claimed to cover the complete levels.
A needs analysis was undertaken before starting the design process of these courses. Twenty different organizations that provide training to displaced people (non-governmental bodies, not-for-profits, foundations, refugee support groups, etc.) in Spain were contacted to participate in this endeavor. The collaboration with them started by an exchange of information, views and opinions on the challenges of teaching and supporting refugees on their arrival to the country (e.g., illiteracy, limited resource, posttraumatic stress, etc.). The research group at UNED that coordinates MOONLITE explained what MOOCs were and how they could be useful to complement F2F classroom teaching to the representatives of the organizations (mostly teachers of Spanish as a foreign language) over a series of F2F formative sessions. The organizations, in turn, explained what the specific educational and sociocultural needs were for the refugee collective and the areas in which they lacked support during the educational process. They were also able to diagnose the linguistic needs of their students for Spanish, as the majority had worked with refugees over a period of years in regular classes. These needs included adapting the didactic rhythm to their abilities, interests, and frame of mind; overcoming the limitations both of Western-centred materials and also of multicultural interactional possibilities in small classroom-based learning; and working in wide realistic communicative situations. In the course of the needs analysis, different types and levels of educational scaffolding were identified to be necessary in order to facilitate both the learning process and improve the social inclusion of the students. Finally, thanks to the outreach of the organizations, it was possible to have several of their refugee students forming part of the LMOOCs' preparation activities, providing their very valuable opinions on what the learning content and format should be like.
By definition MOOCs are open courses so anyone could register to rtake part in the LMOOCs here. However, given the profile of the courses in terms of real-life scenarios that people arriving in Spain would face, and the fact that the publicity for the courses emphasized their appropriateness for refugees and migrants, it is not surprising that the majority of the students on the course were neither Europeans nor North Americans. Furthermore, the organizations that had been collaborating to develop the courses also used their communication networks and channels to reach as many refugees as possible who might find the courses relevant.
Given this demographic, it should be noted that all the students who participated in the courses were adults and had very little in common except for their refugee condition and their lack of acquaintance in terms of previous experience of online courses, use of technology, and level of Spanish. It must be noted, however, that even though the courses were targeting an elementary level and there were not advanced students, some of them had a higher level to start with, as could be seen by their participation in the course. This was probably due to the course focus on real-life scenarios, which could be of great interest to them for practical reasons above and beyond any linguistic considerations.
The LMOOCs presented in this research are hosted on the UNED Abierta (Open UNED) MOOC platform. This institutional initiative was started in 2012 by one of the authors. Today it has a total of more than 200 MOOCs and over 150,000 registered users. Due to the course interface provided by the platform, the two courses developed here have a similar structure to almost all other ones on the platform and combine text, audio, video, with activities, individual tests, and collaborative peer-to-peer evaluations.
Each LMOOC consisted of four independent situation-specific modules and was designed to be undertaken over six weeks, leading to a total study period of twelve weeks. In each course, the first week was designed for students to introduce themselves and become familiar with the MOOC platform. Four weeks were then dedicated to the learning content, and the last week, to complete any unfinished activities. Each LMOOC required an average of 25 study hours (and was accredited with 1 ECTS credit for the students who successfully finished). The situation-specific modules for each course can be seen in table 2: As the courses were being developed, their availability and relevant course information (starting dates, duration, free certification, etc.) was disseminated on refugee-specific web portals and via refugee-support groups. That way, people with the targeted social profile would find out about the existence of these courses, although being MOOCs, they were obviously open to anyone who wanted to undertake them. As well as dedicating effort to attracting students to the courses, teachers from the support groups were contacted to participate in the courses as facilitators. As part of this initiative, specific instructions were given to them for participating in the forums. Given that the facilitators in the course were mostly Spanish nationals, then the contact that the refugees and migrants in the courses had with them in the forums represents a form of implicit social inclusion, since they were actually mixing with people from the country in which they wanted to integrate, which enabled them, to some degree, to experience how the nationals think, talk, and the views they hold. Specifically, they were asked: 1.
"To point out and praise their progress and correct only the main errors progressively, starting with those that make comprehension difficult or cause misunderstandings. As you can appreciate, receiving excessive corrections each time a message is sent can have a negative inhibitory effect. It is also advisable, whenever possible, to explain briefly and simply what the error is to help them not to make it in the future.
Before giving an answer, remember that it is important to say hello, thank them for their question or comment and generate positive feedback. The affective component, as you all know, is fundamental in any learning activities and especially so in this case.

2.
To use 'inclusive language', with common and simple words, respectful of cultural differences. Try and use 'we' instead of 'you', when, for example, you want to encourage everyone to do a certain activity or give a general comment/instruction. When you address a particular student, to answer a comment or question, of course, you can and should use 'you'." (translated fragment of the email sent to the facilitators by the authors before the first course started) Both the facilitators and students that took part in the first edition of Open Doors I and II were encouraged to use Spanish as a means of communication. The reason for this decision is obvious: we wanted to expose the students to native Spanish from the beginning, expand their exposure to this language beyond the didactic activities and initiate them in the receptive and productive use of this language in real communication. This is not common in beginners' courses, so facilitators received online training prior to the beginning of the course so that they could scaffold their input. As for students, as mentioned above, they were encouraged to use Spanish and several statements were made in the guide and the initial email messages about the fact that making mistakes and using very broken Spanish was all part of the learning process and an opportunity for correction and improvement. It must be added, however, that facilitators were also encouraged to use other languages in critical moments if absolutely necessary to overcome misunderstanding and move on in Spanish subsequently.
Starting from the authors' hypothesis (presented in section 2.2) of the value that inclusive language can have in the successful accomplishment of an LMOOC, an analysis of the characteristics of such language was undertaken, with a view to establishing the basis towards a proposal for its incorporation in an LMOOC if positive or promising results were reached.

Results
The courses were successful in several ways: firstly, the number of students who initially registered for the courses was approximately 2200. Of these students, slightly more than 30% of the participants finished the courses, which is arguably the most important sign of student satisfaction. This accomplishment rate was considerably higher than that of previous LMOOCs (e.g., in Professional English, 1st edition, in 2013, it was 7% [Ventura et al. 2014]; and in How to succeed in the B1 Exam, 1 st edition, in 2015, 12% [Martín-Monje & Castrillo 2016]). Secondly, those participants who finished the course undertook between 96% and 98% of the activities, which is remarkable because they only needed to do 80% to obtain the course certificate, and considerably higher than the three above mentioned courses. Thirdly, over 95% of the students declared being neither Europeans nor Americans, which is compatible with a refugee/migrant profile. This was confirmed in the forums, where the login names and specially the interventions of the students were also compatible with a non-Western and refugee profile (e.g., Syria, Cameroon, Venezuela). Fourthly and finally, forum interaction was higher than expected for an A level course, as can be seen by the number of messages in the first LMOOC in table 3. The course had five specific forum threads for activity-related discussions, questions, interactions, etc., with a total of 1647 messages, with the distribution presented. The reduction of messages from the beginning until the end of the course was 38.4%, which is considerably lower than the courses above, where interest in forum participation decreased dramatically as they progressed (e.g., Martín-Monje et al. 2017), because many students found this type of activity too time-consuming and not effective once the course methodology was clear. Therefore, the data shown in table 3 is another indication that many participants felt motivated and engaged to dedicate time to reading and writing when asked to do so in the module instructions. According to the research hypothesis presented in section 2.2, the authors have considered a correlation between the success of this course and the language used therein, not only by the facilitators, who are expert interlocutors with displaced people in academic situations, but also by the students themselves, who also have first-hand interacting experience with other refugees. What follows is an illustration of the evidence found of the language resources used in to enhance inclusiveness in the two LMOOCs developed. The use of inclusive language in these LMOOCs can be seen to be divided into two parts: that made before the course started (by the authors) and what appeared during the course (by the facilitators and the students). The former was designed into the instructions, guides, documents, videos, and activities present in the LMOOCs. Furthermore, the course facilitators, as noted in section 3, were advised to foment its use when engaging with the students in the forums. The latter refers to what was produced by those who participated in the course activities, forums and, to a lesser extent, the courses' Facebook page. It should be emphasized that this research is arguably the first attempt in the literature to identify the theoretical underpinnings and practical effects of inclusive language in an LMOOC. Therefore, the approach followed is all-encompassing and integrates the different types of evidence of this language found in the courses studied. The majority of the examples were originally written in Spanish and have been translated into English so that the reader can understand them. This has been done so with attention to tone and implication, with the intention of preserving the original connotation of the source words and expressions in the transfer process. Mistakes and capitalization and punctuation errors have been kept in translation to be as faithful as possible to the original text. At the phonetic-phonological level, one potential language reference was identified: • Emphasizing prosodic patterns and pronunciation In the video and audio generated by the teachers in this course, care was taken to maintain a reasonable speed together with clear pronunciation and a highly stressed intonation. This is a way not only of scaffolding comprehensibility but also of marking speech, for example, the difference between a question and a statement, and highlighting aspects of grammatical structure. Intonation, the variation of spoken pitch (together with variations of other prosodic systems, like loudness, rhythm and tempo), indicates emotions and attitudes, and helps to regulate conversational flow, causing an involving effect towards the interlocutor. Intonation has a strong connection with the expression of emotion and renders the speech emotionally close to the interlocutor. Since the students on the course did not practice their aural skills, no immediate reciprocity was detected. At the morpho-syntactic level, nine language references were identified: • Avoiding altering speech when addressing certain groups due to prejudices (e.g. further expansion, increased assertion) This feature is related to the modification of linguistic production according to the type of audience, on the assumption of their lower capacity to understand the message, and entails a form of prejudice in itself. This is usually linked to sexist language and examples include giving women more explanations and instructions, or in simpler language, or giving signs of status in the speech. There was no evidence of the presence of this feature in any of the courses.
• Avoiding the use of derogatory words by derivation (e.g., diminutives/augmentatives) In Spanish, it is common to use the diminutive derivational morpheme ("-ito, -ita") with nouns to describe small, young and harmless people, animals or objects, and convey affection towards them. When used metaphorically (e.g., to refer to human nationals or ethnicities), this practice can be considered patronizing and offensive. There was no evidence of that in the course.
• Using 1st person plural (or informal and plural 2 nd person) pronouns and determiners The 1st person plural pronoun (i.e., "nosotros, nosotras" in Spanish, as it is gender-marked) was used consistently throughout the materials: guides, instructions, examples, videos, audios, and evaluation tests. The effect caused was highly integrating, for the far more common use of the 2nd person singular or plural can be perceived as isolating or even accusatory by a sensitive audience. Furthermore, the use of 1st person plural pronouns was explicitly recommended to the facilitators, who used it consistently, producing a highly involving and encouraging effect (although slightly artificial at times). See, for example, one of the titles of the instructional videos: How we take part in this course.
When the use of 2nd person pronouns is unavoidable, in general, the informal form ("tú"/ "vosotros, vosotras") conveys more emotional proximity than the formal one ("usted, ustedes"). Although it did not happen in these courses, in other contexts it might be reasonable to expect a negative reaction on the part of senior students to being addressed with an informal pronoun by a younger speaker. However, since online interaction is rather anonymous and, therefore, hides many personal traits, the use of "tú" is perceived as more inclusive by default. Similarly, when the 1st person pronoun would have been an unnatural sounding option, the generic use of the plural form of the 2nd informal personal pronoun ("vosotros, vosotras") in Spanish) is perceived to be more embracing than the singular counterpart, for the reasons presented in the previous point, that is, to avoid singling out individuals, which may cause distress or embarrassment. What follows is an instance of encouragement which uses this linguistic reference: You (informal, plural) are learning very well. Bravo! Note that similar effects were caused by pronouns and determiners of the same type.

• Using connective adjuncts and discourse markers
The students barely used connectors or discourse markers in the forums, for the majority's level of Spanish was very low. The facilitators, however, used them continuously to aid comprehension for they can have a threading effect in speech. This is compatible with the unusually low number of clarifying questions for an elementary language course. For example: Practicing helps us to learn. Furthermore, it is fun. Therefore, we must use short periods of free time to practice.
As expected, students' speech used basic coordinators, like "and" and "but". (It should be remembered that punctuation and capitalization mistakes have been maintained in translation): Good afternoon I am sissoko. In mali there are public and private hospitals. There is no medical card and we always pay. It is very expensive and the people go to search for traditional Treatments. There are chemist shops but everything is paid for.
My daily routine. I normally wake up at 5:00, I get up at 6. I brush myself, comb myself, shower. At 7:00 I have breakfast. At 8:00 I start doing housework, At noon sharp I have lunch and then at 12:30 I go to work, I come back at 6. On Tuesdays and Thursdays I take my daughter to her ballet classes and on Wednesdays I go to Spanish classes. At the weekend sometimes we dine at the restaurant or go to the club.

• Using stance adjuncts
Stance adjuncts add a personal quality of speech and are, therefore, a valuable feature to be found in language that is to fulfill an inclusive role. There are different types of stance adjuncts. The ones found in the courses are evaluative and, therefore, attitudinal in nature, and reflect the subjective (or objective) attitude of the speaker towards the content of the message (or the addressee). This is reflected in the following example: Unfortunately, there is no census certification in my country.
There are also instances of epistemic stance adjuncts, which express the speaker's opinion regarding the validity of the content, commenting on the certainty or possibility of the proposition. See the following example: Thanks for sharing the routine in your country, Nfaly! In Spain we don't have the habit of watching football matches at the cinema. It is probably great fun.
• Using non-assertive constructions (e.g., interrogative clauses) When possible, non-assertive constructions, e.g., interrogative clauses, are used instead of assertive ones, like imperatives and exclamatives. Similarly, modality is used to diminish the impact of the message and prevent potential confrontation. Note the use of negative questions in the following example that favors one of the two possible answers (which, in turn, evidences their illocutionary force as statements): What is your name, "Cameroun"? How interesting that it [census certification] doesn't exist there. Isn't there any type of registration either? Don't you [informal, plural] have to present any document? Here in Spain, yes, the DNI (National Identity Card) and an invoice or a rental contract.
Interrogatives are also the main type of clause for eliciting information, and starting a conversation or keeping it alive: My name is Serhii. I am Ukrainian, and I am 43. I live in Madrid. Hello everybody! Today too cold in Madrid, What is the weather like in your place?
We don't have a health card in Côte d'Ivoire. Medicines are not free but expensive. What is the health system in your country like?
A type of speech act commonly related to imperative clauses is error correction. The following example shows how it is typically done in the courses: • Using constructions with an emotive quality in a controlled fashion (e.g., exclamative clauses) "Emotional" type of constructions, like exclamatives, are only present in these courses in messages that convey a positive attitude towards the interlocutor, such as encouragement, gratitude or reinforcement and, occasionally, enthusiasm, but not for self-assertion, as shown by the following examples: Thanks a lot for participating in the forum! I'm very happy you like the course! Hello to everybody, teachers, facilitators, and student friends from all over the world!

• Using question tags
The use of question tags is one of the most highly interactive resources in a language. While in English they can be grammaticalized, in Spanish they are constant. Although question tags may have different pragmatic functions like expressing doubt, seeking confirmation or agreement, etc., their use in these courses is mainly related to turn taking and stimulating participation. There is an extensive use of question tags in the forums, for example: Thanks, Ismail. You're from Ghana, aren't you? ("¿no?", in Spanish) How interesting! You're from Argelia, right?
• Using repetitions, parallelisms, paraphrases, and circumlocutions A fundamental distinction must be made between repetitions, etc., of messages uttered by a previous speaker and those that refer to something previous said by the same speaker. In the first case, this may be done with a didactic purpose, e.g., as a subtle way of correcting a mistake present in the first message or re-expressing it in a way that is stylistically superior, less offensive, etc. By doing this, the course facilitators did not only avoid bringing out the implicit failure in rendering a correct message: it was done without interrupting the conversational flow. The second case when facilitators express their own message more than once in the same intervention can be a way of scaffolding potential intelligibility and keeping the audience engaged and potentially active. This is illustrated in the following examples: Facilitator2: Hello, Abdel. So, they need your full name and your date of birth; very interesting :) Do they also ask you, for example, for your place of residence?
Thank you very much! In Spain it is also very important when you want to elect a president, when you want to vote.
At the semantic level, five potential language references were identified: • Avoiding the use of marked language (e.g., gender-specific words, gender in professions, predictable cultural and linguistic mismatches and biases, culture-specific idioms, imagery, symbolic and metaphoric language) There is no evidence of marked language in a potentially conflictive way on the part of any of the agents that interacted in the courses. Most Spanish nouns (and their satellite elements, like adjectives and determiners) are morphologically marked for gender (the most common of which is the "-o" ending for masculine and "-a" for feminine). In one of the modules, the students learn about this, and how the default gender and number in Spanish is masculine singular, so that an audience with a female majority is traditionally addressed to using the masculine pronoun, a practice currently being questioned. Hence, in the courses there are instances of some alternatives that have been gaining popularity recently: Hello, bienvenid@ (welcome).
• Avoiding clichés and stereotypes (e.g., belief, political opinion, origin, social and economic background, race and ethnicity, age, disability, gender, sexual preferences) Attention was paid throughout the courses not to include any stereotypes or biases in the course activities or forum interactions. For example, no assumption was made about beliefs, occupation, or life style, after noticing that a student was female and from a certain developing region.

• Using words with a positive emotional quality
There are words that have a positive primary meaning and others can gain this quality when used in context. There are many examples of both types in the forums of the courses, like "friends", "happy", "pleasure", and "honour", as can be seen here: Hello to all my friends all over the world! I come from Egypt and I am very happy to learn Spanish in this course. I want you to know that it is a pleasure for me, an honour, to be able to interact with you and doing it makes me very happy.

• Using the interlocutor's name in direct interaction and reported speech
Using the interlocutor's name either to address him/her directly or to refer to that person to third parties was common practice, especially in the first forum, as part of introductions. Almost all the replies to initial greetings (of the type: "Hello, my name is Francisco / Maxmos", repeated the name of the previous speaker; for example: Hello Francisco, thanks for participating in this debate. Hello, Maxmos. Thank you very much for participating :) Similarly, students used the same element, without prompting by the teachers or facilitators. Finally, little or no use of the third-party names were made by anyone in the course.

• Using emoticons
There are many instances of emoticons in the LMOOCs, although little variety. Their presence shows the effort on the part of the speaker to project a relaxed attitude and a friendly atmosphere. Only positive emotions are expressed, mainly "smiles". They also convey a pleasant mood and the existence of friendly and relaxed interpersonal relationships that have been established between the facilitators and the participatory students; for example: As can be seen in this example, the atmosphere enhanced by the use of smiling emoticons propitiates the intervention of a third individual in the conversation, which is subsequently gently appreciated/celebrated by the facilitator. At the pragmatic level there were sixteen language references identified: • Avoiding offensive and derogatory statements No offensive language was used in the courses by the teachers or students towards other participants. The course materials were expectedly void of this type of statement. The only messages in the forum that contained negative criticism were uttered by the students and referred to some aspect in their home countries. Interestingly, there was no criticism about any other country or collective, including the host community. See the following message posted by a student called Issam about the health system in his country: The health system in my country is the worst system in the world like if you go to a slaughterhouse when you enter the door of the center is when you get sickest all the people are waiting for the doctor some shouting others fainted some bleeding and the doctor is not there, we do not have health cards, medicines are usually given free at the hospital but when you see the doctor s/he makes you a prescription and tells you to go buy this and that's all at most if it is something serious such as a deep cut or a very strong pain you are injected to remove the pain and that's it at best you see some doctor and that's it but if you have contacts it is different and you will be helped by three doctor with free medicines and put you in some room without noise or screams oh well we very often call the health centres in Morocco slaughterhouses People die waiting for the doctor at the entrance of the hospital.
The atmosphere of respect and empathy in the courses is such that interventions like this receive no direct reply.
The majority of opinions about oneself and one's country are more balanced. See, for example: Look, I like many beaches in Brazil, especially those located in the state of Rio de Janeiro, they are in fact very beautiful and very visited by tourists from all over the world. As for the Price of tickets, I can assure you that it is not a very easy task I come from the Central African Republic There is registration at the town council too in my country, it consists of identification especially when the time comes to elect a President and the necessary documents are the birth certificate and the national identity card. They ask me for the following information: name(s), surname(s), date and place of birth, district, postal code if there is one.
• Using a positive tone (e.g., affectionate, empathic, conciliatory) The empathic tone used throughout the courses rendered a relaxed and friendly atmosphere as a whole. This general tone was only interrupted when some aspect of the home country was described, which contained an implicit or overtly negative criticism, as seen above. There is a high number of instances, although the range of means to convey that tone is limited. Examples include the following: Hello there! How are you today?" // "How are you?" // "It is a pleasure to help you" // "A warm farewell for now" // "Don't worry" // "I have the same problem" // "It is not important" // "Hello dear friend Tomás!

• Avoiding hyperbole in expression or tone
Some extrovert students use hyperbolic language to express gratitude and satisfaction with the courses: Greetings from Brazil. I want to thank you, immensely, for the correction of my mistakes. Making a note of them is absolutely crucial to my learning process.
Regarding negative judgments, the majority are moderate, like the following: The health system in my country is a little bad. It is not easy for people to access the services. There are hospitals and health centres but they are expensive for the people up until now.
However, as seen above, several severe negative criticisms and exaggerations about certain aspects of students' countries were observed in the forums, sometimes expressed in figures of speech, which contained an evident negative emotional charge. An example is the use of the metaphor "slaughterhouse" above, in reference to a hospital. The following message is of the same type: In India we have hospitals in private and gubernamental sectors. Hospitals well established and excellent doctors. But the treatment is a nightmare for the poor. One can receive treatment from any hospital, but the payment is a burden. We don't have a common (public) insurance system. There are private insurance companies and they are trying not to give any service to clients.
The simile below is also hyperbolic: The registration office is like madhouse.
Although there were instances of exaggerated negative criticism in the courses, the responses to these comments were non-reactive, and, therefore, no conflicts arose.
• Avoiding sensitive topics (e.g., historical, political, intimate) While the teachers and facilitators made no direct reference to any topics related to war, terrorism, etc., or alluded to the personal circumstances of the participants (for example, when they claimed lack of time to dedicate to the study of the language in hand), given the background of the refugees who participated in the course, it was not surprising that some of them did occasionally; for example: In my country it is not possible to register at the town hall because of the war.
Conflict never arose because, when a potentially sensitive topic was introduced by a student in conversation, there was no continuity by the rest of the members. Regarding topic management in conversation, there were several occurrences where students started to answer a question in an activity and then drifted off-topic. For example, when asked about the health system in their countries, a student started to talk about it and then moved on to talk about the doctors' skills, the availability of medicines, and finally, how impatient people get while waiting to be attended. This type of behavior in conversation can be caused by personal implication on the part of the speaker and is compatible with an open and participatory frame of mind.
• Avoiding references to characters or situations, events, etc. that are not of general domain (e.g., advanced scientific knowledge, knowledge related to a particular community) References to characters or situations, events, etc. that are not common knowledge (e.g., advanced scientific knowledge, expertise related to a particular community) and may, therefore, provoke feelings of insecurity or inferiority have been generally avoided in the LMOOCs. Care was taken not to use references or examples that only represented Spanish, European or Western names, situations, etc., since most students might not be familiar with them. This was a design criterion of the course from the beginning and partly the reason why refugees were involved in the preparation of the course materials and the dialogues that are at the core of each module.

• Selecting suitable topics for the interlocutor's profile
The topics used is the discussion activities in the forums of the courses were generally related to the students themselves (e.g., personal likes and dislikes) and their country of origin (e.g., daily life, travelling, healthcare), in order to empower their original identities and previous existences and, therefore, their selfesteem, and prevent any feeling of detachment and alienation.
Topics of conversation and debate in the forum, therefore, dealt both with people and activities expected in countries with strong traditional roots, and also with transversal issues of general interest. When messages required specialized knowledge, the participants in the courses showed a high level of consideration, for they generally offered enough background knowledge to enable their interlocutor to understand and interpret their words adequately. See the following examples whose comprehension requires different levels of specialized knowledge: Hello, in my country there are many interesting place and different cultures, for example, in the Sahara, there is a region called Hogar where a tribe, the Targuis, live, who go dressed like in the old times and men, unlike women, hide their mouths. In order to go to Hogar you must take a plane because it is 2500 km away from Algiers the capital.
I come from Touba, Senegal. Touba is the city of Serigne Touba, Marabou and spiritual guide of the mourides.

•
Providing external multimodal scaffolding (e.g., interlingua, glossary, subtitles, images, videos) The teachers gave the facilitators indications to use all strategies available in an online course to provide linguistic scaffolding to the students of both courses, foreseeing that difficulties in comprehension due to their low level of Spanish may demotivate them. Their potentially frail state of mind and the difficult circumstances that they were going through were also taken into account, all of which might easily lead to course abandonment. Facilitators, who were expert language teachers from the organizations who collaborated with the research group in the design and development of the LMOOCs, followed the instruction given by the teaching team closely, which consisted of accurately judging when they needed to use another language momentarily (after repeatedly failing to be understood) and turned back to Spanish afterwards without interrupting the general flow of interaction in this language. Below is one of the first messages in the forum of Open Doors I: Badajoz, in the Padre Cristóbal Cáritas reception centre. I don't know the role of town hall registration explain it to me. Facilitator4: Hello Abdoulaye. When you live in a city or town you need a census certificate. You will be asked to show it to go to the doctor or enrol in a course. Bonjour Abdoulaye, quand tu vis dans une ville ou un village, tu as besoin un document d'enregistrement. On va demander si tu veux aller chez le médecin ou d'assister à un cours de formation.
As for the students, interestingly, as can be seen in the example in figure 1, due to the continuous encouragement to use Spanish only in the course forums, hardly any other language was used. Nearly 97% of Spanish messages in the forums were expressed in the language object of study, in spite of the low level that most of the students had. This was not the case, however, in the Facebook page associated with the course, where students felt more relaxed to use it. This was, in fact, one of the reasons for providing a Facebook page for the LMOOCs. Including a social network played a highly inclusive role for students with the lowest level of Spanish who were most disoriented about the methodology and the technology (the platform and the tools therein) of the courses. As predicted by the facilitators, they used English, French and Arabic on the Facebook page.
All the instructions in the LMOOCs were formulated using simple and direct language and supported by icons to indicate when the students were expected to read a text ("lee"), watch a video ("mira"), repeat or respond to an audio ("repite/responde"), download a file ("descarga"), select something ("escoge/elige"), or listen to an audio ("escucha"), as can be seen in figure 2. Subtitles in Spanish and in the most common languages among the refugees and migrants interested in Spain and Spanish (namely, English, French and Arabic) were created. All the videos in the courses (e.g., about the structure, how to navigate around the platform, how to carry out activities, etc.) had subtitles in those languages (as well as transcriptions in PDFs ready for download).

• Projecting a low self-image
Projecting a humble image of oneself is (and even self-deprecating humour) is a doubtful but common social strategy to be accepted, liked and taken as an unpretentious and nice individual. Both facilitators and students in the courses used this resource in several occasions, as shown in the following examples: Hello Lucciano! Don't worry. The same thing happens to me in English and my replies are sometimes chaotic. What is important is practising and learning, just like you do. I really like the connecting words you use in sentences.
Hello to everybody! My name is Carmen and I live in Caceres. I am a Spanish teacher, and I am just starting to receive training in teaching migrants and refugees.
Accept these greetings from this humble friend from Brazil!

• Passing positive judgments about others
This is a bidirectional constant between facilitators and students with an evident inclusive and empowering or boosting effect; for example: Many thanks. You write very well in Spanish. I like this course very much. The teachers are very helpful. I don't have any doubts about the course that is why I can only say thank you vey much to my teachers who are organized to help us so I am very happy to learn with you because the way you prepare the courses is very agreeable and it is the way that can help a person from far away who doesn't know any Spanish so that they can participate in this course. Thank you very much for your time.
• Associating emotional/affectual responses with participants and previous performance/processes (e.g., appreciation, empathy, concern) One of the most salient pragmatic resources present in the courses is the emotional and affectual responses triggered by previous interventions. In the following sequence of fragments, for example, firstly, a student gives an opinion about census certificates; secondly, he sends another message when he realizes that he has confused two terms ("empoderamiento" and "empadronamiento") and presents his apologies twice; thirdly, the facilitator sends a comforting message to the student about his mistake and deviates his attention towards his participation, for which she greets him twice. As can be seen, both participants are very reactive about each other's messages:

Student6:
Empowering is fast and easy in my country if you have money to pay you do the paperwork in one day […].
Student6: I must tell you that maybe I have not interpreted the term properly. I have confused 'empadronamiento' for 'empoderamiento' (I present my apologies for it), for this is not a common thing in our country […]. I apologize once more for not having understood the nature of the question and, for that reason, having made a mistake. Thanks to everyone! Facilitator4: Good morning, Lucciano. Thank you very much for writing in the forum and explaining about census certification in your country. You don't have to apologize because in order to learn you need to make mistakes and 'empoderamiento' looks similar to 'empadronamiento' but the meaning is very different […]. Thank you very much again and lots of luck with the course.
Apart from gratitude and apology, other very common reactions to previous speech between teachers and facilitators on the courses are agreement and support, which has an engaging effect. Common phrases found in the forums include: "I agree with you", "You are right", and "It is true". The reaction on the part of the interlocutor is typically to produce a message with a similar emotional tone; for example: Student7: Yes, it is true, breakfast in Spain has a lot of sweet things. I love toast with jam.
Facilitator6: You are right, we have so much in common.
• Making reference to something that the interlocutor has previously said, either to them or to third parties Very little evidence was found for this resource on the part of the students, possibly due to their low level of Spanish. One of the few examples produced is the following: Thanks Beatriz for your help, as you said, I will wait to see if it [my Spanish] gets better.
• Using reassurance and positive reinforcement Apart from the summary of contents at the beginning of each module and the checkup list at the end, where the student can see the acts that they will be able to/can already perform in Spanish thanks to their studies, there is considerable evidence of positive reinforcement, strictly speaking, and mutual reassuring statements on the part of facilitators and students. Examples of facilitators' reassuring comments are: You write very well // I really like that // Well done for having finished the module // That is very good, try to continue that way As for the students: Thank you teacher, you speak very well, I would like to do so // I don't have any doubts about the module because it is well explained // Thanks for your attention Occasionally, the facilitator may ask for reassurance: Facilitator7: I hope I have clarified your doubts and helped you with your curiosity. Student8: For sure!!! • Using resources to negotiate an interpersonal space for one's own positions within the diversity of viewpoints put at risk by an utterance (e.g., modals of probability, reality phases, attribution, proclamation, expectation/counter-expectation) There are a few messages in the forums of the courses where the facilitators and the students negotiate the difficulty and the contents of the course. The interaction in both cases is very mildly at risk and, in any case, saved after the intervention. In the first example, the facilitator uses an empathic strategy: she recalls a similar situation when she experienced difficulty and how it was solved, thus, giving the students the hint of what needs to be done. In the second example, the facilitator starts her reply by agreeing with the student, only to implicitly show him that he is mistaken: Student9: I can understand many words in Spanish but I can never use them well Facilitator8: That happened to me in German and I thought it was impossible. But it isn't! We need to practise very much and don't worry about mistakes: only understanding is important.
Student10: I don't understand why we study so much about paperwork. The words are difficult, and I am not sure it is so useful.
Facilitator9: You are right that some of the words are long but you will be happy that they are familiar to you if you come to Spain.
• Graduating interpersonal impact, force and volume of one's utterances as suitable in the verbal negotiation (e.g., assertive, empathic and encouraging speech, emphasizers/diminishers, imperative/interrogative) There are many examples of speech in the courses where the facilitator's or the student's utterance is graduated, particularly to diminish or increase the effect of the speech, which prevents any tension or negative feeling from building between speaker and interlocutor. The range of resources used, however, is greatly limited by the students' low command of the language. Let's see the following example: I think writing is perhaps a little more difficult than speaking. That is only at the beginning so don't worry.
Finally, at the sociolinguistic level, four potential language references were identified: • Awareness of situational context or culture-driven language In order to undertake an online conversation between people who do not know each other and have little in common, an effort must be made. In the LMOOCs, politeness formulae on the part of facilitators and students were far more common and emphatic than in standard F2F interaction, but they were interpreted in the key of the specific setting in which they were uttered. This realization helped the (virtual) integration of the agents involved. Another example occurred when the students were asked to go to the forum if they had any query about a particular activity: Please comment any problems you have with this activity here.
Instead of simply refraining from visiting the forum, those who did not have any queries (around 10% of the students) went to the forum to insert messages like the following: I have no problems // I have no problems thank you very much // There is nothing I don't understand // I understand everything The only way that this behavior could be interpreted correctly would be by being aware of an underlying cultural difference for this type of situations.
In other occasions, the students were aware that the interlocutor may not be acquainted with certain cultural elements, and explained them, like in the following text: The price of tickets depends on elements like the region and also what we call 'high season' (important dates), like Christmas, New Year, and now, with the arrival of the Carnival, the price figures keep going up and up, from 50 dollars a day to incredible 1000 dollars a day. Ufaa!!!!

• Avoiding vertical analyses of cultural and personal behaviors and perspectives
It is sometimes the case that prolonged debates around a particular subject may unveil controversial conclusions. There is no evidence of this happening in the course.

• Avoiding discriminatory comparisons between cultural and personal behaviors and perspectives
There were no comparisons between social or individual elements in the courses. For example, facilitators were specifically requested to avoid negative comparisons, and when they appeared, move the focus of conversation onto a more descriptive aspect instead. Student11: The medical system in my country is much worse than here.
Facilitator10: What do you do to get hold of medicines in your country?
• Using appropriate levels of formality (e.g., politeness formulae) to convey respect/closeness/intimacy As seen above, informal pronouns were used multidirectionally between the different course participants. Despite this evidence, however, there was a continuous, almost excessive, use of politeness formulae in what appeared to be a persistent attempt to maintain an inclusive and empathic interaction. We have seen above that greetings, the expression of gratitude, agreement, and concern are constants in the courses. In fact, as seen in table 4, "thank you" (and its variants) was the most common expression in the forums of Open Doors I, followed by positive remarks, expressing welcome, etc. Respect and recognition were feelings which were constant throughout the courses and explain the presence of so many politeness formulae, as seen repeatedly above. It is to be noted that, as expected in an LMOOC, there was not any expression of intimacy, endearment, or personal closeness.

Discussion
For refugees and migrants, learning the language of a country where they are trying to integrate, or even for later on, if they haven't managed to enter yet, can be seen to be of great importance as a step forward. In this article, inclusive language has been described as a tool that can not only be used to help such students improve their communication in the target language and understand its formal and functional aspects efficiently, but also play a significant role in improving their social inclusion. In table 1, in section 2.3, a first draft proposal of the resources that play a cognitive and affective role in inclusive language for open online communication is presented. They are grouped into five levels: phonetic-phonological, morphosyntactic, semantic, pragmatic and sociolinguistic levels, following a conventional linguistic division, although pragmatics permeates all of them. While their selection could have been undertaken exclusively based upon the previous experience of the authors and, particularly, of the support groups that help and scaffold this collective from their F2F support centres, Appraisal Theory was used because it provides a theoretical basis on how speakers use language to evaluate, adopt stances, construct textual personas and manage interpersonal positionings and relationships, yet it is versatile enough to be applied and adapted to different genres and linguistic variants. The authors claimed that Appraisal Theory provided a grounded key for interpreting the fragile and complex interpersonal communication between most displaced people and their teachers and peers in the challenging scenario of massive open online courses. From here onward, the authors identified specific language resources that favor social inclusion and avoid exclusion, and lead to positive/negative individual discrimination, together with instantiations of resources related to standard Attitude, Engagement and Graduation, as per Appraisal Theory.
The context for the study of the linguistic characteristics presented in this article, as noted above, was a course of elementary Spanish (ranging from A1 to A2+ levels), that was divided into two contiguous LMOOCs (Open Doors, parts 1 and 2). As noted previously, inclusive language was used both before the course (in the supporting documentation and activity descriptions) and also during it (by the relevant actors). Advice was also given to the course facilitators to potentiate its use where possible in the forums.
When addressing the first research question, about the evidence for inclusive language learning criteria in the LMOOCs, as section 4 has demonstrated, there is ample evidence of its use both by the facilitators, and arguably more importantly and spontaneously, by the refugees and migrants that make up the student community. The results in that section take the form, firstly, of the "presence" of inclusive language corresponding to some of the characteristics from table 1, which represent how the students relate verbally to the facilitators and to each other. Secondly, the results also show the "absence" of exclusive language, in terms of other characteristics from the same table that would only alienate and demotivate a collective like refugees and migrants, and eventually lead to course abandonment and worse. Not surprisingly, there was no evidence for the expression of prejudices, stereotypes, clichés and discriminatory judgments in the courses. There was an extraordinary level of verbalized politeness, empathy and gratitude between the participants which, together with the above, are claimed by the authors to be associated with the success of the courses (in terms of measurable accomplishment, engagement, interaction, etc.). The inclusive language not only played a pedagogic role but also supported the students emotionally. This was shown when they talked negatively about aspects of their home countries and other potentially delicate subjects (such as war and terrorism), which received no follow on either from their peers, facilitators or themselves in subsequent messages.
When the general inclusiveness of the language used by the students is compared with that of students on other standard LMOOCs run by the authors, considerable differences are found, since the latter do not present evidence of the majority of the resources presented in table 1 (cf., the Professional English LMOOC described in Barcena et al. 2015, where there was even a generic "netiquette" included in the course guide with specific recommendations for verbal interaction). Therefore, the following question arises about the use of inclusive language in the LMOOCs: does it happen spontaneously by students or as a result of how others speak? This is a difficult question to answer. There is some evidence for both in the results presented in section 4. On the one hand, little advice was given to the facilitators and none to the students about how to manage their interaction. Furthermore, there was evidence of spontaneous examples of inclusive language. On the other hand, some of the produced language could be due to mere imitation, in a similar way to what happens with non-verbal communication in F2F conversations (cf. Jokinen & Pärkson 2011).
Moving onto the second research question, that of the effects caused by using inclusive language in an LMOOC, particularly one aimed at people in disadvantaged situations, as well as its presence in the student production in the forums and activities, there are also other quantitative factors that suggest a positive effect on student performance. Firstly, the high degree of interaction in the forum and the number of messages that went beyond single question-answer interactions. Secondly, the number of students who finished the course. It was recognized in section 4 that MOOCs have high levels of abandonment; typically, figures are around 90% of enrolled students. In this course, more than 700 students finished the first LMOOC, representing a completion rate of over 30%, well above average. Thirdly, the number of messages in the forums, which dropped slightly but did not fall off massively, as is usually the case in MOOCs. While there may be other reasons for these numbers, the students consistently demonstrated throughout the course, and subsequently in the post-questionnaire, their high level of satisfaction overall. As well as the quality of the content and structure of the course, this was arguably thanks to the inclusive language used for communication.
Further evidence of the effect of inclusive language in these LMOOCs can be detected in the reactions to its use, i.e., the motivation and engagement shown by the students when addressing the topics in the forums show how its use helps the students maintain their interest and proactiveness in the course as it progresses. This can be identified in terms of the following types of responses, which are exemplified in the two courses run: 1. Longer responses. As the courses progressed, conversations about any topic tended to have more messages which, in turn, also tended to be lengthier as each forum thread progressed. A schematic example of one piece of conversation in the third module of the first course is presented here in terms of the number of lines of text used by the interlocutors: Student11: 3 lines Facilitator11: 3 lines Student12: 6 lines Student11: 1 line Student13: 2 lines Student12: 13 lines 2. More elaborate responses. As the courses progressed, and the students got to know each other in the forums, richer interaction took place, which led to more complex messages. Although it could be argued that knowing more Spanish could account for the increase in the number and verbosity of the messages, as mentioned above, literature on online courses makes reference to a certain tendency on the part of students to become less participative towards the end of courses, once they control the methodology and the steps to be taken to finish.
3. Responses with a marked positive emotional tone. Emotion tend to increase in the responses of the course participants and their scope as the courses progressed. They also diversified, ranging from hope and gratitude to happiness and affection. For example, students were complementary about the support and assistance received: Student14: In my country Mali, most people like to drink tea in the shade of a tree. During my free times I like to go for a walk.
Facilitator12: That sounds like fun. Remember that we say in my free time and not in my free times.

Student14:
Thank you for correcting my writing. A big hug! 4. Responses which use inclusive language. As courses progressed, the presence of inclusive language, in the multiple forms identified in this article, became more evident, probably due to a combination of an imitation process of the linguistic model provided by the facilitators and a verbal representation of the emotional tone of the course. Some evidence has been presented on this in the results section, to which the following eloquent example can be added: Many thanks for the course, I love that you are concerned that we all mix and get along.
5. The absence of interpersonal conflicts. There was no evidence of any confrontation in the forums, which is not usually the case in these courses.
6. Responses that reflect a cognitive gain. Although this effect is difficult to prove without a proper evaluation process, there are signs that there was significant progress on the part of those students that accomplished the courses and contributed most actively in the course activities and the forums: In module 0: In module 4: Student16: In my country there are health centres and also hospitals But the health system is very bad, we don't have a health card and we have to pay everything.
Facilitator14: Very good. Are there doctors of all health specialties or just one general practitioner?
The analysis undertaken in this article has presented illustrative evidence that reveals the affordances of using inclusive language in LMOOCs for displaced people, which have been identified by the authors as a first step toward a fully developed account of this language type in the context of the descriptive and analytical framework provided by Appraisal Theory. The answers to the research questions presented in this section illustrate the effectiveness of this language as a tool for improving the social inclusion of refugees and migrants, arguably while they are undertaking the courses and, subsequently, if and when they continue to interact with peers and native speakers of the target language, using the same linguistic variant.

Conclusions
A lot of refugees and migrants, who have been forced to flee their home countries, appreciate the value of continuing with their education, as a way of repositioning themselves for new roles in the European countries in which they find themselves, or toward which they are traveling. Their preferred option for learning, F2F classes, is often not available due to the distributed geographical locations of displaced people and the lack of sufficient teachers and educational resources for such large numbers. This article started by noting that MOOCs, as free online courses, represent an important educational option for the many refugees and migrants who need the corresponding knowledge and skills. However, even if the refugees accept that such courses are the best or only option that they have, and try to undertake them, quite often they abandon them, since the courses are not designed specifically for the very people they are aimed at, i.e., people outside the Western civilization with different learning, social and digital culture. The goal of the research presented in this article was to collect and extend previous work on inclusive language to explore its use for refugees and migrants in an online learning context like that of MOOCs. In order to identify and develop the resources that underlie inclusive language for this collective, Appraisal Theory was used. This preliminary exploratory study is by no means exhaustive and future work will be needed both to quantify the inclusive language features presented here, to explore the domains in which they can be applied, and to analyze the cognitive gain and course performance of those students who present features of non-inclusive language. Also, a finer-grained analysis could be undertaken to differentiate the effects of inclusive language generated by the course design, development, materials, tuition, facilitation, and learning processes. The results gathered from the two LMOOCs that make up the introductory Spanish course have provided evidence in support of the value of inclusive language for refugees and migrants in online courses, and as such, have offered insight into answers to the research questions presented here. It has been noted that, while results are positive, the study is preliminary, based upon exploratory data gathering and, as such, need to be further explored and examined in future research.
However, above and beyond the data that support inclusive language in LMOOCs, there are other general quantitative data that arguably do the same for this type of language in the case of refugees and migrants, namely the lower abandonment figure here (almost three times higher than on standard MOOCs), the higher motivation present in the students (shown by the number of activities undertaken and the higher degree of forum interaction) and, therefore, the improved resulting practice in the target language overall.
LMOOCs were arguably born with the objective of democratizing language learning online. However, up until now, this has not been the case since the profile of students on these courses show the majority to be European or North American. The case for inclusive language presented in this article represents a way to effectively open up LMOOCs for wider, disadvantaged and vulnerable social groups, like refugees and migrants. It also lays the foundation for online language learning for and towards this collective, which not only enhances target language knowledge and communication, but also increases social inclusion and our opportunities for having a more inclusive and heterogeneous modern society.
I just want to say thank you… Hello, I want to say thank you for the excellent course, you are angels. I want to carry on studying this language, because I like Spanish very much. Thank you for the pieces of advice at the end of the course, I want to look them up and try to do other courses. My dream is to visit Spain, but unfortunately it is not very easy for those who live in Brazil, the trip and the accommodation are very expensive. I do not get tired of dreaming, who knows if one day I will show up in that beautiful country. Thank you