Taking a Closer Look at the Speaking of Ideal Self of Spanish Undergraduates

Over the past fifteen years, research on second language (or L2) motivation has been dominated by Dörnyei’s influential paradigm, the L2 Motivational Self System. Students’ imagined visualisations are key components in this theory, as those students who have a clear ideal selfimage with an L2 component will probably be more motivated to learn a language than others that have not established a desired future state goal for themselves. This article reports the qualitative findings of a mixedmethod study that explored the effects of a multimodal intervention with influential speakers on changing the students’ attitudes in public speaking. Semi-structured interviews and open-ended questionnaires were conducted with 11 engineering undergraduates, who volunteered to take part in the present study. Qualitative data showed that the multimodal intervention accompanied by goal setting (i.e., students’ classroom oral presentations) triggered an increase in some students’ future speaking selves. Six of the eleven students demonstrated a slight development in their levels of linguistic self-confidence, which made their vision of their ideal L2 speaking selves more realistic and clearer. The article discusses the implications of these findings and calls for a pedagogical shift that embraces more opportunities to assess the multimodal skills and strategies students need to become fluent L2 speakers.


INTRODUCTION
Lately, research on second language (or L2) motivation has been largely influenced by Dörnyei's (2005Dörnyei's ( , 2009 comprehensive theoretical construct, the L2 Motivational Self System (L2MSS). This paradigm encourages a viewpoint of L2 motivation through the perspective of students as future L2 users. Dörnyei (2005, pp. 105-106) establishes three dimensions in this theoretical model: 'Ideal L2 Self', which refers to a learner's goals and aspirations as a language learner, 'Ought-to L2 Self', which refers to the personal characteristics a language learner believes he or she ought to avoid for negative results, and 'L2 Learning Experience', which refers to specific aspects related to the language learning environment (i.e., methodology, language lecturer, materials, and classmates).
The roles that vision and imagination play in Dörnyei's (2005Dörnyei's ( , 2009) L2 Self perspective are quite relevant. An increasing body of research has brought into focus the relationship between L2 motivation and vision (Dörnyei et al., 2016;Dörnyei & Kubanyiova, 2014;Safdari, 2019) and L2 motivation and imagery (Dörnyei & Chan, 2013;Munezane, 2013;Shen, 2010). These studies offer robust empirical evidence about the facilitating role that self-enhancement activities, including visuals and guided imagery, have on students' involvement and commitment in learning a second language.
Designing successful motivational intervention programmes, or ideal-self generating activities that focus on influential role models or successful L2 learning achievers may allow students to enhance their possible language selves, following Dörnyei's (2005Dörnyei's ( , 2009) L2 Motivational Self System. Role models, as Muir et al. (2019, p. 1) argue, "can exert considerable influence in shaping our values, attitudes, and beliefs" and might be "a highly influential component of the psychological context of second-language acquisition".
The present article aims to contribute to filling the gap in the literature produced by the increasing importance of a multimodal approach in the learning of foreign languages and the absence of works focused on the impact of multimodality on learners' motivation in general and in the Spanish ESP undergraduate context in particular. Accordingly, the overriding objective of this article is to investigate whether undergraduate students' L2 speaking selves could be increased with the regular use of TED (Technology, Entertainment and Design) Talks.
This article reports the qualitative findings of a mixed-method study (García-Pinar, 2019) that explored the effects of a multimodal intervention with the influential TED speakers on changing the students' attitudes in public speaking. For that purpose, a one-group pretest-posttest research design as a part of a mixed method study was conducted in order to analyse the positive, negative, or neutral motivational effects of carrying out a multimodal intervention followed by the visualisation of different technological TED Talks in the classroom during a semester. One important quantitative finding was that the multimodal intervention impacted positively on students' L2 motivation, as there were statistically relevant differences between the second questionnaire and the first (i.e., first and fifth phase of the mixed-method study).
Meanwhile, the study reported in this article only presented the qualitative results (i.e., phases 3 and 4 of the mixed-method study) and can be regarded as a follow-up to the quantitative study by García-Pinar (2019). Qualitative data, as the following sections shall describe, allowed the researcher to obtain a further understanding of the development of students' ideal L2 selves and their stance concerning variables such as linguistic self-confidence in using English when giving their classroom oral presentations. The following research questions were formulated: 1. To what extent is there a development of students' ideal L2 selves as a result of a multimodal intervention? 2. How does a multimodal approach to public speaking influence learners' linguistic self-confidence in their oral presentations?

Possible and Ideal Selves
The construct of 'possible selves' was introduced by Markus and Nurius (1986) and encapsulates an influential motivational self-mechanism that represents (p. 954) "individuals' ideas of what they might become, what they would like to become, and what they are afraid of becoming". In academic contexts, ideal and positive visions students generate about themselves for their future might lead them to academic engagement and eventual achievement. Individuals have a desire to reduce the gap between their actual and ideal selves (Higgins, 1987) that may initiate specific actions to soothe the feeling of dissatisfaction, guilt, fear, or disappointment.
Possible selves are future-oriented self-conceptions (Markus & Nurius, 1986), and as such, cannot be conceived as static constructs (Henry, 2015). Individuals' awareness of the distance between the actual and the ideals selves might trigger changes in their possible selves (Henry, 2015). Possible selves, therefore, may undergo up-and downward revisions whenever they interact with other self-concepts. In L2 settings, the lecturer has the challenging role of providing students with engaging tasks, relevant teaching materials, and positive learning experiences that are personally and academically meaningful and of value to L2 students, allowing them to reassess and enhance their ideal L2 selves. The focus, therefore, should be on the necessary and specific actions involved in the process of achieving students' desired L2 selves, since as Safdari (2019) contends, future self-guides can be trained, manipulated, and made more vivid and real.
A few classroom motivational interventions (Chan, 2014;Magid, 2014;Munezane, 2013;Safdari, 2019) have focused on vision and imagery to develop students' L2 ideal self-images and to increase their L2 motivation. These interventions included extensive tasks and activities: oral communication projects, ample exposure to role models, and the elaboration of future-oriented autobiographies. To determine the impact of these tasks and activities, mixed-method approaches (Chan, 2014;Magid, 2014;Munezane, 2013) were used to collect qualitative and quantitative data. Quantitative and qualitative findings showed that participants' L2 motivation, their attitudes toward the learning experience, and their imagery capacity improved.

TED Speakers as Role Models for L2 Students
Designing successful motivational intervention programmes, or ideal-self generating activities that focus on influential role models can be optimal opportunities A. García-Pinar, Taking a closer look at the speaking of ideal self of Spanish undergraduates | 904 to allow students to enhance their possible language selves (Dörnyei, 2020). This process, referred to as observational or vicarious learning, has its roots in Bandura's social learning theory (Bandura, 1997). Widely researched in educational psychology (Cox et al., 1999;Lockwood & Kunda, 1997;Murphey & Arao, 2001), observational learning assumes that individuals who have achieved remarkable success can raise the observers' expectations for their future and motivate them to excel in their pursuits. The important aspect of observational learning for the present study has to do with the influential role that TED Talks and their speakers' multimodal way of communicating may have in changing students' attitudes to speaking in public. These short, carefully crafted talks offered by leading experts in a broad range of fields, have been shown to provide a set of guidelines for compelling communication (Anderson, 2017(Anderson, , 2020. During 18 minutes, speakers at TED provide targeted enlightenment on a wide variety of topics and ideas that are deemed 'worth spreading'. Common subjects in these talks are business, technology, global issues, and politics. The researcher assumed that students' regular visualisation of these talks could allow them to observe and analyse speakers' multimodal way of communicating in relation to their purposeful use of gestures and facial expressions, the use of wellthought-out visuals, and their precise use of space on stage. The regular visualisation of TED Talks in the English course, therefore, could create a suitable context where concrete action plans (i.e., those afforded using these talks) helped learners envision tangible pathways to create their desired future selves. The construction of communication from an accurate combination of verbal and non-verbal modes might have a strong influence on how students receive and interpret the message being conveyed. It follows, therefore, that providing students with information which helps them understand and interpret the co-occurrence of different modes in complex multimodal talks is a prerequisite in today's L2 educational settings (Jewitt et al., 2016;Kress, 2009;Rowsell, 2013).
TED Talks could be deconstructed in the L2 classroom so that the students observe that the way TED speakers communicate with a salient interweaving of verbal and non-verbal modes is attainable and relevant (Lockwood & Kunda, 1997). This multimodal way of speaking may raise the students' hopes for their future, as they are likely to assess and contrast their speaking capabilities and multimodal skills to TED speakers' or salient others (Markus & Nurius, 1986). Once students realise what the speaking task is doable, the L2 lecturer could establish some concrete plans of action that lead to real progress (Dörnyei & Kubanyiova, 2014).

Setting and Participants
The present study was conducted at a well-established Spanish polytechnic university founded in 1998, and presently has over 7.200 graduate and postgraduate students. The university offers students a wide range of undergraduate degree programs in the fields of Engineering, Architecture, Tourism, as well as Economic and Business Sciences. Engineering undergraduates at this university should attend the course of Technical English during one semester of their four-year degree. The course covers different topics that are common to different engineering fields and includes a task-based approach contextualised in daily engineering contexts to practice oral and written skills. Group in-class presentations are one of the tasks students prepare throughout the course and are generally regarded by many students as the most challenging.
The participants for this study were 11 Engineering undergraduates, all of whom had also participated in the quantitative phases of the mixed-method study (García-Pinar, 2019). They were on average 19 years of age. Concerning the Cambridge Assessment Test, which assessed the grammar and listening skills, their English proficiency oscillated between B1 and B2. To respect the participants' privacy, confidentiality and anonymity were ensured. Accordingly, the participants' original names were replaced by pseudonyms.

Data Collection and Procedure
The 11 students were interviewed after having attended the multimodal intervention at the beginning of the semester and before they had to give their classroom oral presentations. The interview protocol included nine different questions and the average length of the interviews was about 30 minutes. In addition, all of these interviews were audio-recorded and conducted in Spanish. The main aim in conducting the individual, semi-structured interviews was to allow learners to ascertain different aspects related to the feeling of linguistic self-confidence whenever they had to speak in the classroom in front of their classmates and their language lecturer.
Another objective was to study the beliefs in relation to their regular implementation of verbal and non-verbal modes in students' in-class oral presentations. Finally, the researcher aimed to study the effects that the L2 multimodal intervention could have on developing learners' future identities as proficient speakers of English. Students received some coaching for the classroom oral presentations they were about to give. They were given some guidance about the different verbal and non-verbal modes they had learned during the multimodal intervention. In the next phase, the six different questions included in the post-intervention open-ended questionnaires were aimed to gather specific information of the students' own impressions of their performance in their oral presentations. Students completed these questionnaires upon completion of their in-class presentations so they could give accurate information of how the implementation of different modes had made them feel.
Following thematic content analysis (Dörnyei, 2007), the interview transcripts were thoroughly read through three times. During the first reading, the transcripts were marked to obtain closer familiarity with the overall organisation. Significant and meaning-carrying details were highlighted, such as nouns, adjectives, and adverbs that conveyed students' motivational intensity. The second reading was aimed at highlighting the expressions and utterances that expressed goal/vision orientation. The third reading involved singling out the expressions that conveyed participants' linguistic self-confidence. Participants' responses that are closely related were grouped (Loewen et al., 2009). Additionally, the researcher identified positive and negative descriptors to better understand students' common beliefs about their L2 learning experiences and their ideal L2 speaking selves.
A. García-Pinar, Taking a closer look at the speaking of ideal self of Spanish undergraduates | 906

The Multimodal Intervention
The multimodal intervention was specially designed to help learners create a more vivid representation of themselves as proficient L2 speakers. The key assumption underlying this focused intervention was that by addressing the different verbal and non-verbal modes that contributed to meaning-making in these multimodal talks, L2 students could recognise how speakers at TED engaged and persuaded their audience. If they were able to visualise and notice how different modes (i.e., language, visuals, gestures, facial expression, and proxemics) were arranged in these talks, students might be encouraged to voice their 'engineering ideas worth spreading' in a TED Talklike style.
During the first part of the intervention, students were initiated into different nonverbal modes in communication to make them realise that language was no longer the only mode to construct meaning. Thus, the intervention provided students with information that might help them understand and interpret the co-occurrence of verbal and non-verbal modes in complex multimodal talks. Students were guided on how to proceed with this implementation of modes, and how this implementation could extensively enhance their oral performances and heighten their levels of linguistic selfconfidence. Additionally, the students also learned about the important role that beat and deictic gestures (McNeill, 1992) play in communication. These types of gestures might help speakers highlight important stretches of their discourses, coherently integrating the overall discourse. Other non-verbal modes that students were introduced to were proxemics, facial expression, and visuals. Furthermore, the students learned important multimodal concepts, such as visual salience (Kress & Van Leeuwen, 1996) (i.e., the specific ways different elements in a visual layout such as colour, size, and contrast appear to capture the viewer's attention), modal aptness and modal affordance (Kress, 2009). The researcher assumed that these were notions that students should be aware of when designing their Power Point files, as these are likely to have a determining role in their overall performance. The second part of the intervention showed students the most relevant minutes of six different TED Talks. The main purpose of this part was to make students notice the ways different modes deployed in these multimodal talks helped speakers to better convey their messages. TED speakers' postures, gestures, facial expressions, the way they used space, and their well-crafted power points had clearly relevant roles in the process of communication. Students actively participated during this part by naming all the nonverbal modes TED speakers used, and by justifying speakers' intended purpose in using those modes.

Development of Students' Ideal L2 Selves as a Result of a Multimodal Intervention
The qualitative results reported in the following sections offer relevant insights into the motivational learning behaviour of the 11 participants. Data, as detailed earlier, was collected at two different times during the semester, and through different instruments. The first qualitative tool was individual, semi-structured interviews (time 1) with 11 students, that took place before students' in-class oral presentations. The second qualitative tool was an open-ended questionnaire (time 2) that these 11 students filled in and sent by e-mail after having finished their oral presentations. One important focus of the analysis was to trace whether students' L2 self-visions could develop, from time 1 to time 2.
The first research question of the present study asked about to what extent there is a development of students' ideal L2 selves as a result of a multimodal intervention. The findings from the qualitative data revealed that, by visualising multimodal TED Talks, students were able to realise that public speaking also involved nonverbal communication conveyed through diverse behaviours such as gestures, facial expressions, proxemics, and prosody. Students' visualisation of TED speakers' way of communicating allowed them to observe and analyse their multimodal way of communicating, and the multimodal intervention that focused on different influential role models may have enhanced their possible selves.
Most students acknowledged that multimodal public speaking was not a skill innate to them. Yet, they also highlighted that this was a skill that they could master over time and through extensive training. Questions 1 and 2 from the interview (time 1) ("Can you imagine yourself as someone who will be able to speak fluently and correctly in English in the future? and, as someone who can give a talk in a TED style incorporating the modes you learned?") plus question 1 from the post-intervention open questionnaire (time 2) ("Do you think that with rehearsal and with the incorporation of different modes, you might give a talk as a TED speaker does?") enquired into students' capacity to envision themselves as highly proficient users of English in the future. The analysis of the students' answers to these three questions allowed the researcher to trace the evolution of the students' ideal L2 selves, and to observe how even in a short time (from the day of the interviews to the moment they completed the post-intervention questionnaires), students' comparisons with proficient and successful TED speakers triggered a certain reformulation and revision of their ideal selves. This fact, in turn, made these future selves more realistic and tangible.
This study used three coding categories to measure the development of students' L2 speaking selves between the time the interviews took place (time 1) and the time students completed the open questionnaires after having finished the in-class oral presentations (time 2). These categories were previously used by Chan (2014): "(a) emergent", "(b) fading", and "(c) stable". The coding categories this study used to include the multimodal facet were emergent ideal multimodal L2 self, fading ideal multimodal L2 self, and stable ideal multimodal L2 self. The emergent ideal multimodal L2 self-referred to five students that during the interview described the possibility of including non-verbal modes in their in-class presentations as very unrealistic and showed themselves quite suspicious of the positive influence that the regular use of modes could have on their performances. Yet, during the postintervention open-ended questionnaire, these five students realised the positive influence that the regular use of modes had had on their speeches. Additionally, these students acknowledged that despite the multimodal public speaking was not a skill innate to them, this was a skill they could command in time and through extensive training.
The fading ideal multimodal L2 self-referred to one student who acknowledged the relevance of using different non-verbal modes beyond the verbal one and conceived the possibility of using these quite plausible during the semi-structured A. García-Pinar, Taking a closer look at the speaking of ideal self of Spanish undergraduates | 908 interviews. During the post-intervention open-ended questionnaires, the ideal multimodal L2 self in this student faded because he realised that the implementation of these modes had not improved his speech. Finally, the stable ideal multimodal L2 self-related to five students whose multimodal facet kept stable from the time they were interviewed until the time they completed the post-intervention open-ended questionnaires.
The five students with emergent ideal multimodal L2 selves showed a more specific multimodal L2 self at time 2 and reflected on what they had learnt about multimodal presentation techniques. The language they used (i.e., vocabulary and verbal tenses) when they explained their capacity to envision themselves as proficient speakers able to incorporate an array of modes, shows some positive development from time 1 to time 2: At time 1, Student 1 embraced the possibility of being a competent L2 speaker who was able to use a talk with diverse modes. He claimed that this possibility "might be a goal". By time 2, he unhesitatingly stated that he could give a multimodal talk; time and training being the two necessary prerequisites: "yes, it is definitely a question of time and lots of practising to include all the modes in a natural way". Student 2, in the same way as Student 1, viewed the probability of becoming a fluent L2 speaker. Yet, he was not comfortable with giving a presentation in English at time 1 ("I'm not sure..."). By time 2, he was more confident about giving a presentation to the whole class ("I'm completely confident…"). Although Student 3 appeared to be quite determined to become a competent L2 speaker at time 1, he demonstrated a more solid ideal L2 multimodal self at time 2. For example, he visualised the goal of giving a multimodal talk at time 1. Yet, he admitted that this faith in his possibilities might be due to being overly optimistic ("I don't know whether I am being too optimistic"). By time 2, he kept his optimistic view, but he was able to give a practical form to his multimodal skill and acknowledged that he needed to practice several modes (gestures, intonation) in order to give a good talk: "I feel able to give a talk at a good level, in a TED-style. I would need to practice".
Unlike Student 3 and Student 1, Student 4's future vision was vague at time 1. She knew that she needed to become fluent in English, but giving a multimodal talk was a challenge that was unattainable ("I don't know", "I am not sure"). At time 2, Student 4's language reflected more determination ("right now I could"). She clearly stated being able to give a talk in a TED style in Spanish, and one in English in a few years.
Student 5 also considered the possibility of giving a multimodal talk in the future (i.e., five or six years), but this possibility was placed at time 1 and not at time 2, as in Student 4's case ("well, in the distant future, I could view myself as speaking good English"). By time 2, she was able to be more specific and claimed that with adequate training she would be able to give a talk in a TED style "in five, six years".
One of the students reported a fading ideal L2 multimodal L2 self (present at time 1 but not at time 2). At time 1, Student 6 was doubtful about whether she could give a multimodal presentation, even in Spanish. Her language demonstrated these insecurities ("I am not sure", "I don't even know if I could"). Yet, she implied that having enough time might be the key prerequisite she needed to give a good talk ("it might be a question of time"). At time 2, she ruled out any possibility of giving a multimodal talk, even with the time issue ("I don't think I will ever achieve…").
In contrast to the previous cases, five students reported stable ideal L2 multimodal selves (present at time 1 and at time 2). Student 9 could envision himself as a competent user of English. When asked about the possibilities of giving a talk in a TED style at time 1, he relied on his skills, but made this ability conditional on time and training issues ("it is a matter of time, and practice"). By time 2, he showed the same self-confidence and he again mentioned the time and training conditions ("I am not sure. I should practice a lot (…). It is a question of time"). Student 7 could also imagine himself as a fluent speaker of English. Giving a multimodal talk like in TED was a goal for him ("a tall order") and he verbalised the possibility of delivering a good multimodal talk using the conditional both in time 1 ("I guess I could, but I would need a lot of time") and in time 2 ("this could be a good speaking goal").
Student 8, in the same vein as Student 9 and Student 7, exhibited a similarly positive position towards giving a multimodal talk. He showed a decisive and determined attitude at time 1 and time 2. He acknowledged his capability to give a multimodal talk in Spanish ("I can visualise myself giving a talk like those in TED. I could give one in Spanish") and did not appear so confident about giving one in English at time 1. At time 2, he verbalised this confidence again ("I would be able to give a good talk in Spanish"). He seemed to be hesitant about giving a multimodal talk in English at time 1 and stated that being able to give a talk in a TED style would keep him motivated to learn English. With a quite remarkable difference compared to the rest of the students, Student 11 fully maintained confidence in his language ability. Giving a talk in a TED style was an achievable possibility both at time 1 ("I think I could, yes") and at time 2. At time 2 he was even able to name what he needed in order to obtain his goal ("with lots of practice, adequate gestures, intonation, and visuals, with a topic I have control of and about which I feel passionate, I might give a talk in a 'TED style'"). Unlike the previous cases, Student 10's perception of his public speaking skills was quite negative. He maintained a stable ideal multimodal self between time 1 and time 2. The least pronounced development between time 1 and time 2 was the case of this student. In the interview, he openly admitted his lack of confidence in public speaking ("not at all", "not in my dreams"). In the open questionnaire, he maintained this lack of confidence ("I don't feel confident at all").
This qualitative analysis proved to be relevant to corroborating the dynamic nature of future self-guides (Henry, 2009(Henry, , 2011(Henry, , 2015Markus & Nurius, 1986). Most students' ideal selves were reformulated and revised during the post-intervention openended questionnaires. Higgins ' (1987) self-discrepancy theory can help explain, to a certain extent, why some students may have been able to envision themselves as proficient L2 speakers more easily than others. During the interviews, students were asked whether they liked how speakers at TED communicated, some students' ideal selves were reformulated and revised.
Raising consciousness about their multimodal skills and the strategies that might complement the verbal mode in their presentations may have played an important role in reducing the gap between these students' actual and their ideal L2 speaking selves. It might have been the comparison with these proficient speakers that triggered a process whereby their ideal L2 selves would be progressively constructed. Other students may have realised that the gap between their desired and their actual speaking selves was too big to even consider the implementation of different modes in their oral presentations.
A. García-Pinar, Taking a closer look at the speaking of ideal self of Spanish undergraduates | 910 The findings of the present study are in line with the previous research by Chan (2014), Mackay (2014), Magid (2014), and Safdari (2019), who carried out selfintervention studies that focused on vision and imagery. The self-intervention studies that these authors conducted were aimed at developing the participants' L2 ideal selves, especially through the implementation of different activities. In the four studies, participants' ideal L2 selves were increasingly more robust, their goals for learning English became more clearly delineated, and participants' imaginations improved. In Chan's (2014) and Magid's (2014) studies, the paired sample t-tests demonstrated that students' ideal L2 selves improved at the end of the intervention. Qualitatively, diverse aspects of L2 selves were also modified. In Mackay's (2014) study, the analysis of the results revealed that the combination of visualisations and strategy training had the potential to motivate EFL students. Some students in the intervention group proved to be positively influenced as a result of the intervention process. Students showed clearer goals, recognised enjoyment in the L2 experience, and acknowledged their selfefficacy beliefs. Safdari's (2019) implementation of a vision-based motivational intervention program also proved to be effective in strengthening L2 learners' vision, imagery capacity, and motivation.

Influence of a Multimodal Approach on Students' Linguistic Self-Confidence in Their Oral Presentations
The second research question of the present study investigated whether multimodal public speaking affected the students' linguistic confidence. Questions 5 and 6 from the interview or time 1 (respectively, "Do you feel self-confident when you speak English?" and, "Would the incorporation of the modes in the oral presentation help you feel more confident?") together with question 3 from the post-intervention open-ended questionnaire or time 2 ("Do you think that with rehearsal and with the incorporation of different modes, you might give a talk as a TED speaker does?") addressed the construct of linguistic self-confidence.
The researcher assumed that upon completion of their oral presentations (after the intervention), the students would be in the position to establish whether they perceived that the implementation of modes had had a positive effect on their classroom presentation performances. These three questions contributed, therefore, to trace the development of students' levels of linguistic self-confidence from time 1 to time 2. Six of the eleven students widely acknowledged that the use of different modes contributed to building a high sense of their self-efficacy: Student 9, who declared that he was linguistically self-confident, and who had played down the importance of nonverbal modes before his classroom performance at time 1, stated that the use of gestures, in particular, had enhanced his speech by time 2. Student 1 admitted his capability for public speaking. Before his oral presentation, the only aspect he feared was not being able to explain the technical processes, and he conferred a facilitative role to modes in a hesitant way ("they might help me feel more confident"). By time 2, he admitted that the use of well-thought-out visuals had helped him convey his verbal part quite fluently.
The use of good visuals, as in the case of Student 1, might have contributed to raising Student 7's level of confidence at time 2. At time 1, he admitted not having enough confidence about public speaking and thought modes could give him the extra confidence he lacked ("the use of some modes might help"). By time 2, he indicated that good visuals seemed to have captivated his audience, which helped him lower his initial feelings of anxiety ("I was nervous most of the time, but I think I did well. My classmates seemed to listen. Our visuals were quite good").
The facilitating role of modes was also highlighted by Student 2, who acknowledged being quite anxious about speaking in public. At time 1 he stated that modes might help him lower his anxiety ("I might feel more confident if I use these modes"). By time 2, he convincingly declared that modes (i.e., gestures, visuals, and facial expression) extensively helped him convey the verbal part better. He "felt good".
Student 3 showed a wary attitude. He openly declared not having any objection to speaking in public. He also stated that the implementation of modes could have an overall positive effect. Still, he stated that these should be used with caution. At time 2, he indicated that the use of different modes contributed to the increased level of selfefficacy he already showed at time 1 ("gestures helped me to be focused all the time").
The last student in this group is Student 4. She declared that she felt fear of negative social evaluation when having to speak in public ("I feel really nervous and start trembling. I don't like to speak in front of many people"). Yet, she informed the researcher that modes, including the use of gestures, could help her reduce the fear of being negatively evaluated. By time 2, she showed a higher level of self-efficacy ("I think I did a good presentation") and stated that the use of good visuals and gestures seemed to keep their classmates' attention. Two students (Student 8 and Student 5) acknowledged that the use of modes had enhanced their presentations. However, these had not had a facilitative role in controlling their anxiety. Student 8 declared some communication apprehension originated from his lack of confidence in being able to express complex engineering processes. He showed some uncertainty about the enabling role of nonverbal modes at time 1. At time 2, he stated that the use of modes did not help him feel at ease. Yet, he thought he did "OK".
In the same sceptical way, Student 5 openly declared for having fears of being assessed by others when she speaks in public. She also appeared quite sceptical about the facilitative and enhancing role of modes in the classroom oral presentation. By time 2, she declared that she had felt quite anxious throughout, and stated that training and rehearsing were the only strategies needed to succeed.
Three students (Student 10, Student 6, and Student 11) widely acknowledged that the implementation of modes had not contributed at all to enhance their speeches. The use of gestures, in particular, had made them deviate their attention. Therefore, gestures had been a hindrance in their oral performance. Student 10 is the student who felt the most anxious and least self-confident about his speaking skills during the interview ("I don't feel confident at all") and did not think that any technique in public speaking could control his anxiety. By time 2, he stated that his initial intentions of implementing modes had made him feel even more anxious ("I felt even more nervous"). Student 6 was another student who stated that she had foreign language anxiety related to the fear of being evaluated by her classmates. She seemed doubtful when she was asked about the possibility of using any of the modes, she had learned at time 1 ("I need to practise a lot if I finally decide to incorporate them"). By time 2, she stated that the use of modes had made her too concerned about these, which had had a negative influence on her verbal part ("I was too concerned about the use of gestures and this fact made me feel anxious").
Student 11's opinion in terms of the facilitative role of modes was quite remarkable as one might suppose that the high levels of self-confidence, he reported A. García-Pinar, Taking a closer look at the speaking of ideal self of Spanish undergraduates | 912 during the interview might have brought about a positive comment in relation to the facilitative role of modes in communication. He was, indeed, the student who acknowledged having the greatest self-confidence of the 11 students ("I feel confident about public speaking. I am sure I will be able to make a good presentation"), and who did not show any of the two components kinds of anxieties his classmates described (i.e., communication apprehension and fear of negative evaluation). Despite these positive comments, in the post-intervention open-ended questionnaire, he thought that the use of gestures had not helped him to upgrade the content of his performance ("I focused too much on when I should incorporate the gestures"). Qualitative data allows one to ascertain that, on the whole, raising students' awareness about the facilitative role of different modes used in their oral performances could improve their self-efficacy. Six of the students acknowledged that the implementation of modes had enriched the content of their presentations. They explained that their classmates seemed to follow their presentation and this fact helped them visualise themselves as proficient L2 speakers, and even reduced their levels of anxiety in some cases. However, it should be borne in mind that, according to the findings of the present study, the use of modes can be counterproductive, especially with those students who did not have a high level of English.
In the post-intervention open-ended questionnaire, the students had to name the modes they had used in their talks. All of the students marked the visual mode. In response to Question 3 ("Did the use of these modes help you feel more confident?"), students reported how the use of appealing visuals seemed to have captured their audience's attention, and how this generalised involvement had made them feel at ease and more self-confident. Ten students reported having implemented either beat deictic or iconic gestures during their performances. Three students acknowledged having kept eye contact, seven students used some type of facial expression (i.e., smiling and nodding), three students resorted to varied intonation, and finally, four students were especially aware of word stress during their performances (see Tables 1 and 2).
It is particularly noteworthy that none of the students used proxemics in their presentations. During the multimodal intervention, the students were introduced to the ways TED speakers arranged and utilised their space to enhance meaning making. Students could observe how TED speakers perfectly knew how to move on stage with the flow of their speeches, and when they chose to stop in order to emphasise significant parts in their speeches. One of the reasons that might justify the students' reluctance of using the classroom space was the layout of the classroom, which was traditional and far from anything resembling the stage design at TED. A further possible reason why students did not make use of the class space was that this mode entailed considerable practice and rehearsal to achieve its emphatic purpose, a prerequisite that might have deterred many students. The following table describes the types of modes signalled by the students in the interviews to be potentially used in their oral presentations.
As detailed in Table 1, most students stated that they would use some of the modes they had learnt during the multimodal intervention. In particular, they showed themselves determined to use the gestural (i.e., beat and deictic gesture) and the visual modes. Table 1. Modes to be potentially used according to the students in their oral presentations.
On the whole, the analysis of qualitative data manifested that the ideal L2 self is a construct that cannot be conceptualised as static (Henry, 2015). In the present study, most participants' ideal selves were reformulated and revised during the postintervention open-ended questionnaires. The visualisation of speakers' multimodal way of communicating contributed to manipulating and making more vivid and real students' L2 speaking selves. In line with the above remarks, which provides the ideal L2 self with a dynamic, changing nature, and extensive practical potential, the L2 lecturers have a challenging role of providing students with engaging tasks that can fuel them with the motivation which further gives direction to their actions and allows them to envision their future as proficient L2 users.

CONCLUSION
The main assumption underlying the present study is the belief that an intervention with a focus on the multimodal skills of experienced speakers could influence students' attitudes to speak in public and develop their ideal speaking L2 selves. These results are extensively consistent with the quantitative results (García-Pinar, 2019), which showed that the intervention positively influenced the students' possible L2 selves as there were statistical differences between the pre-and the postintervention questionnaires.
The multimodal skills of influential speakers may have contributed to making students realise that public speaking not only involved the verbal mode but that this equally entailed nonverbal communication conveyed through diverse behaviours, such as gestures, facial expressions, proxemics, and prosody. The use of gestures, more specifically beat and iconic gesture, and of well-thought-out visuals during participants' in-class oral presentations, made them realise the engaging role these had. Six of the eleven interviewed students realised that the implementation of modes had enriched the content of their presentations. This fact led to heightening their levels of linguistic self-confidence, which further facilitated their performance and, in some cases, helped them visualise themselves as competent L2 speakers, as later shown in the analysis of the post-intervention open-ended questionnaires.
This study suffers from one important limitation that has to do with the noninvolvement of the researcher as a lecturer of the students under analysis. This fact might be considered both advantageous and disadvantageous. The decision of having conducted the study with students whose English lecturer was not the researcher herself, allowed her to analyse data completely objectively. Additionally, students might have felt themselves less inhibited to speak and give their personal opinions about specific issues such as their learning experiences. Yet, the researcher esteems that it could have been quite relevant to lecture these participants herself in order to follow their motivational behaviours throughout the semester. Despite this limitation, the present study can serve as a springboard for future research by ascertaining the role that individual variables such as personality, play in the construct of motivation as mediated by multimodal interventions in order to attain a finer-grained picture of the effect of such interventions. Research in this area might help understand why different students react differently to the same learning conditions.