Theory and Practice of Second Language Acquisition vol. 8 (1), 2022, pp. 35–52 https://doi.org/10.31261/TAPSLA.10482 Xavier Martin-Rubió https://orcid.org/0000-0002-9643-9530 Universitat de Lleida, Spain Irati Diert-Boté https://orcid.org/0000-0002-6964-5996 Universitat de Lleida, Spain Catalan Law and Business Students in Italy: The Impact of a Stay Abroad on Fluency and Accuracy A b s t r a c t The growing number of English-Medium Instruction courses offered across Europe is increasing the opportunities for student exchanges. This study follows the progress of three students from Universitat de Lleida after their Erasmus experience at three different European universities, two in Milano and one in Macerata. The students took a monological English oral test before and after their stay abroad, and f luency and accuracy measures have been calculated from it. The students were also interviewed and participated in focus-group discus- sions. The measures from the two students who went to Milano show an improvement in their English level, whereas the student who went to Macerata performs even worse on his return. However, the experience from the two students in Milano was substantially different from the one who went to Macerata. Using the ethnographic information and the qualitative data available, we bring forward arguments that can help to account for these different outcomes. Keywords: f luency, accuracy, study abroad, additional language learning, Englishization, ELF Student Mobility in Europe A growing number of universities across Europe offer English-medium bachelor degrees. According to Maiworm and Wächter (2014), the number of English-taught programs in Europe drastically increased between 2007 and 2014 in percentages that range between 112% for Central West Europe (the lowest increase) and 866% for South West Europe (the highest increase). https://creativecommons.org/licenses/by-sa/4.0/deed.en https://doi.org/10.31261/TAPSLA.10482 https://orcid.org/0000-0002-6964-5996 Xavier Martin-Rubió, Irati Diert-Boté36 The advantage of this process of Englishization of Higher Education is that it re- moves linguistic barriers that made student exchanges more difficult in the past. The disadvantages include a domain loss for the local languages (Salö, 2014) and the obstacles for integration for incoming students and staff who might be considering a longer stay (Martin-Rubió & Cots, 2016). In this study, we follow the progress of three students from Universitat de Lleida (UdL hence- forth), a higher education institution (HEI henceforth) situated in Catalonia, who spent one or two semesters at another European HEI. The three host HEIs are situated in Italy (Milano and Macerata). The students took an English gram- mar test and carried out two tasks, one written and one oral, before and after their stay abroad. They also participated in focus group discussions before and after their stay, and were interviewed during their exchange. Thus, the study focuses, on the one hand, on the analysis of the measures of accuracy and of the temporal measures of fluency of the oral task; and on the other, on the contextual circumstances of their stay, as self-reported in the group discussions and interviews. Fluency and Accuracy Measurements There is a long tradition in Second Language Acquisition to examine the evolution of fluency, accuracy and complexity in learners’ production as a way to keep track of their learning process. For the present article, only fluency and accuracy are considered. Fluency has been defined as “the ability to talk with normal levels of continuity, rate and effort” (Starkweather, 1987, p. 12). When speech is constantly interrupted by pauses, continuity is affected; when only a small number of syllables are uttered in a given second, the rate is low; and when the speaker struggles to find the suitable word, an unusual amount of ef- fort is required. Accuracy, on the other hand, has been described as “the ability to produce error-free language” (Czwenar, 2014, p. 82), although what can count as an error is rather debatable and a subject of controversy. Determining the levels of fluency and accuracy depends on, at least, two elements: the task to be developed and the level of language proficiency. Apart from these, the level of accuracy might be influenced by the socio-educational context of the speaker, and the level of fluency by the speaker’s personality, such as self-confidence and extroversion (MacIntyre, Dörnyei, Clément, & Noels, 1998; Arnold, 1999), and by personal traits such as speech impairments, like stuttering (Fortunato- Tavares, Howell, Schwartz, & Furquim de Andrade, 2017). Nevertheless, for the present study, only the task at hand and the language proficiency are the foci of interest. Catalan Law and Business Students in Italy… 37 Different tasks present different demands on the people that engage in them. This can be easily gathered from a task comparison. By way of illustration, we could take the three following tasks: reading a text, describing a comic strip, and participating in a debate. While reading a text, the reader must only process the language written on the page; there is no need to improvise, even though the reader must still make sense of the words in the page. The person reading might produce pronunciation errors, but they would not be responsible for the lexical and grammatical elements in the text; the author of the text would be. When asked to describe a comic strip, one needs to choose what to say, and although the task itself narrows the vocabulary to be employed, one has a larger range of lexical choices than while reading. Furthermore, in a monological comic-strip description, the participant has to solve the task in a self-sufficient way. In debates with other people, conversely, one can pick and reuse lexical and grammatical elements from the other participants. Additionally, one needs to process the information conveyed by the other participants and decide what to say and how to say it, while simultaneously dealing with issues of taking, holding and yielding the floor. Any person, thus, will speak with different levels of fluency and accuracy across these different tasks. Language learners at the early stages of their learning processes will evi- dently struggle to formulate long meaningful chunks, subsequently producing a higher number of pauses in their speech, a portion of which will be situ- ated in the middle of phrases, thus giving the impression to the listeners that they are not fluent (Wennerstrom, 2000). As Chambers (1997, p. 540) points out, “becoming fluent therefore is […] about pausing less often and pausing at the appropriate junctures in an utterance.” At a phonological level, Hieke (1985) argued that fluent speech equals connected speech. When the number of uninterrupted syllables is measured, what is actually being measured is how capable the speaker is of connecting syllables without pausing; this is a great indicator of the progress in the language. This is further exacerbated in English, considering it is a stress-timed language; therefore, the learner needs to learn to connect syllables with an alternation of stressed and unstressed syllables that is different in syllable-timed languages like Spanish (Leal, 1995). Accuracy is intimately-related to the notion of error and mistake. According to Brown (2004, p. 216), “a mistake refers to a performance error in that it is a failure to utilize a known system correctly; while an error is a noticeable deviation from the adult grammar of a native speaker, reflecting the interlan- guage competence of the learner.” In this case, the adult grammar of a native speaker is taken as the norm, and noticeable deviations are considered er- rors, whereas performance errors are described as mistakes. In the same vein, Canagarajah (2015) argues that errors, unlike mistakes, are systematic and in- dicate the personal “grammar” of the language user. Errors are an integral part in the learning of an additional language, and one good indicator of progress Xavier Martin-Rubió, Irati Diert-Boté38 is precisely the learner’s decrease in the number of errors. Errors can be clas- sified as lexical (word choice), phonological (pronunciation), semantic (mean- ing), syntactic (grammar), and pragmatic (content) errors (Jiménez Arias, 2004, p. 177). A special focus is placed in this article on pronunciation errors. Figure 1 shows one of the two comic strips that our participants had to describe for their oral task. One of the participants used the expression “sequence” three times in the description, but pronounced it as / s̍e.kwens/ rather than the standard / s̍i .ːkwəns/ in all three occasions. It is, thus, not a mistake but a pronuncia- tion error in the sense that the participant probably is unaware of the standard pronunciation of this word. Figure 1. Comic 1: Picnic task (Heaton, 1966). While describing the third picture of the strip, the participant says “the boys search the ball/ (0.6) /in.a:̍ prɒk.si̩ meɪt.li̩ es̍ neɪk/.” In the picture, we can see a boy looking for the ball inside the hole, and a snake approaching the hole, so we can imagine he wants to say something like “and a snake is approach- ing,” but the fact is that the sentence he produces is ungrammatical and that we can only make sense of it all with the help of the picture. Focusing on pro- nunciation accuracy, if we decide he was producing the expressions “approxi- mately” and “snake,” we could identify two pronunciation mistakes. Rather than /ə̍ prɒk·sɪ·mət·li/, he generates a secondary stress in the fourth syllables (/a:̍ prɒk. Catalan Law and Business Students in Italy… 39 si̩ meɪt.li/); he also utters “snake” in three rather than two syllables. Moreover, he makes a lexical mistake, since he uses the adverb “approximately” rather that the verb “(to) approach,” and a grammar mistake, because he is using the wrong order and structure. Needless to say, significant differences are bound to be found amongst different raters, which is why the two authors of this paper rated the cases independently and then discussed the different positions until a common ground was found. Additional Language Learning at Home and Abroad The status of English in international communication has progressively gained momentum and its impact upon many societies (including the Catalan and the Spanish) is much more far-reaching nowadays than a few years ago. The necessity to learn the English language is thus evident, being English one of the central subjects in the Catalan curriculum of both primary and secondary education. As an example, Catalan students are gradually obtaining higher results in the assessments of basic competences at the end of primary and secondary education (Generalitat de Catalunya, 2017); nevertheless, those evaluations do not assess oral production, which students believe to be the most problematic area for them (Diert-Boté, 2016). In order to improve the competence in the English language during com- pulsory education, it is not uncommon for English learners to attend private language schools, to receive content subjects in English in many high schools— referred to as Content and Language Integrated Learning (CLIL)—and to go abroad, particularly during summer breaks. Spending some time abroad is frequently regarded as one of the most effective ways of learning the language of the country where the stay takes place. Although not as intensely as in other countries (Park, 1997), sending one’s child(ren) to countries like the UK or Ireland for two to five weeks during the summer break is generally perceived in Catalonia as an effective way to boost the chances to learn English (Tragant et al., 2017). When students start a university degree, HEIs normally offer similar options to help students to keep progressing in the English language. On the one hand, HEIs usually run language courses through their language centers addressed to two collectives: local students, who study additional languages (mostly—but not only—English); and international students, who can study the host univer- sity’s local language(s). For instance, UdL offers Catalan and Spanish courses to international students, and there are several courses of English and other languages at different levels. Xavier Martin-Rubió, Irati Diert-Boté40 On the other hand, the equivalent to CLIL subjects at university are English for Specific Purposes subjects (ESP) and content subjects taught in English, commonly referred to as English Medium Instruction (EMI). Maiworm and Wächter (2014, p. 48) analyze the evolution of the number of programs taught in English (ETPs henceforth) in HEIs across Europe and argue that the Nordic Region and Central West Europe are “the pioneers in the implementations and running of ETPs”; the authors also state that although the growth rates in these regions are below the average, this is only because “further growth becomes more and more difficult” given their leading role in this process. In turn, “the growth rates were highest in South West Europe (866%),” although “most South European countries are still at the bottom of the ranking list—de- spite the impressive growth rates.” Spain and Italy belong to this South West Europe region. Several studies (Cots, 2013; Dafouz & Camacho-Miñano, 2016; Doiz et al., 2014; Mancho-Barés & Arnó-Macià, 2017; Salaberri-Ramiro & Sánchez-Pérez, 2015) testify to the intensification of the Englishization process in Spanish HEIs. Particularly at UdL, the teaching in English in the academic year 2012–2013 accounted for 4.6%, whereas in the year 2016–2017 it rose to 6.1%. Notwithstanding the increase, the percentages of programs in English, and more specifically of students enrolled in such programs, are still very low. Ultimately, the third option that HEIs offer is study abroad programs, most notably the Erasmus exchange program, which provides some students with the opportunity to spend one or two semesters at a foreign HEI. Unlike the summer-break stays discussed above, which normally take place in the UK or Ireland, university exchanges cover many different countries. The top destina- tion for UdL students, for instance, is Italy (UdL, 2014). Research has shown that study abroad (SA henceforth) benefits second/foreign language acquisi- tion, especially oral proficiency (Freed, 1990; Ginsberg & Miller, 2000). SA programs are naturalistic settings in which students can meet native speakers and thus participate in different communicative situations from those provided in educational settings (i.e., formal classroom learning) (Manchón & Murphy, 2002); yet, many of these students also enroll in language courses, so the combination of both settings seems to be the most appropriate one to learn a foreign language (Llanes & Serrano, 2011). Although Manchón and Murphy (2002) mention contact with native speak- ers, on many occasions, and specially for the present study, native speakers of English are not what students (expect to) encounter in their SA experiences. Lecturers, students, and staff generally coexist in an environment in which several languages are employed, but where English is often used as the only available lingua franca. When UdL students decide to spend one or two se- mesters abroad, one of their goals is to improve their English in one of these English as a Lingua Franca (ELF henceforth) settings, which is a feasible goal according to research. Kaypak and Ortaçtepe (2014) found out that students Catalan Law and Business Students in Italy… 41 in an SA ELF setting showed a shift in focus from accuracy to intelligibility and a higher level of self-efficacy, which led them to engage more frequently in interaction in the target language. In turn, Borghetti and Beaven (2015) point out that in an ELF setting students not only experience lower levels of embarrassment, fear of being judged, and concern about participating in inter- action; they also perceive a greater level of accommodation, negotiation, and cooperation strategies on the part of non-native speakers of English. Therefore, these non-Anglophone countries appear to be suitable places to practice English. Koylu (2016) confirms these positive findings, adding the analysis of oral and written fluency, accuracy, and syntactic and lexical complexity to the picture. The participants in the study are Turkish students in three different contexts: at home internationalization (AH), stays in countries where the tar- get language is the national language, in this case England (SA) and stays in countries where the target language is not the national language (ELFSA), in this case universities in Austria, Czech Republic, Denmark, Finland, Germany, Greece, the Netherlands, Italy, Poland, and Portugal. Koylu (2016, p. 173) identi- fies different accounts from students studying in countries such as Denmark, the Netherlands, Germany, and Finland in relation to those from students in Italy or Greece, with the different English level of the local residents as an important factor. Figures from the Eurobarometer (European Commission, 2012, p. 23) in relation to the percentage of people in the different EU countries who claim they can hold a conversation in English show a very stark contrast between countries like the Netherlands (90%) or Denmark (86%), and countries like Italy (34%) or Spain (22%). There are, however, many other factors, such as the length of stay, the teaching practices of the host HEI, the place of residence in the host city, or the network of friends sojourners establish, that also play a big role in the learning opportunities of the exchange students. Research Questions and Methodological Aspects The data used for this project come from a mixed methods research project which aimed to analyze the impact of a stay abroad on students from UdL. The study consisted of a quantitative part that targeted all the students from UdL selected to participate in the Erasmus program in the 2013–2014 academic year, and a qualitative part that focused only on those travelling to Denmark, Italy, and the United Kingdom. These three countries were selected because English plays very different roles in them. The 109 students who had been selected to participate in an Erasmus exchange during the 2013–2014 academic year were asked to complete a questionnaire, a grammar test, a written task, and an oral Xavier Martin-Rubió, Irati Diert-Boté42 task. Twenty-five of those 109 students completed all of it, and three of those 25 students had chosen an HEI in Italy for the SA. All the students provided their written consent to participate in the research project and the name of the three participants of this study has been replaced to preserve anonymity. Beatriu, a Business student, went to Milano (HEI code IMILANO16) for ten months. She spent a month in Venice taking an Italian course before travelling to Milano. She was placed in a hall of residence on the outskirts of Milano, where a great number of Spanish Erasmus students resided, although she shared a room with a Slovakian student until February. Alma, a Law student, went to a different university in Milano (IMILANO01) and stayed there for five months, from February 2014. She travelled to Milano with a friend from Lleida (that we will call Gertrudis), another Law student with whom she shared a room. They stayed in a different hall of residence, this one situated in the center of Milano. At first, they mingled with Italian students, but ended up hanging out with other mostly Spanish Erasmus students. They argued it was just easier to do that because they went to the same parties and had the same schedules, whereas many Italian students worked and partied less often. They had many German, Dutch, and Romanian students in the hall of residence with whom they often talked. Finally, Josep Maria, another Law student, went to Macerata (IMACERATA01) for six months, although in this case from September 2013. Macerata is a much smaller place than Milano. His best friends in Macerata were two Italians, a girl from Córdoba (Spain) and a Greek student who had spent several years in Italy and with whom he always spoke in Italian. The oral task mentioned above consisted of a comic strip of six different frames that the participants had to describe in less than five minutes after spending up to a minute preparing what to say. Two different comics were employed, comic 1 in the pre-tests (see Figure 1) and comic 2 in the post-tests (see Figure 2). A mistake while implementing the test to Josep Maria meant that he actually described comic 1 in both the pre- and post-tests. Given Josep Maria’s performance in the post-test, one can only wonder what the result would have been if comic 2 had been used as originally planned. The six oral stories produced were subjected to a two-phase manual analysis. The first phase analyzed temporal measurements of fluency and measurements of quantity, and the second included disfluency and accuracy measurements. The first phase started with the identification of three types of chunks: silent pauses, filled pauses, and between-pauses units (henceforth bp-units). Silent pauses are pauses of 0.25 seconds or longer, and filled pauses include hesita- tions, laughter or coughing. Bp-units are stretches of talk found between silent/ filled pauses and are measured in syllables. A spreadsheet was used to indicate the type of chunk, the length in milliseconds for all chunks, the number of syllables, and the rates of each bp-unit. Catalan Law and Business Students in Italy… 43 Figure 2. Comic 2: Football task (Heaton, 1966). Five measures of quantity have been employed. Speech Time (ST) re- sults from adding the time of all the bp-units. Ginther et al. (2010, p. 387) define speech time as “speaking time, excluding silent and filled pauses.” In turn, Pause Time (PT) is calculated by adding the time of all pauses (both silent and filled). Total Response Time (TRT) is the number of seconds the participant takes to complete the task and results from adding ST and PT. Speech Time Ratio (STR) gives us the percentage of time that the speaker spent speaking in relation to the TRT. The last measure of production is the number of meaningful syllables uttered. As for fluency, three temporal measures have been used, two devised for this project and the third (MSR) adopted from Ginther et al. (2010). The first two measure the speech rate, that is, the speed at which the syllables are delivered: the Rate of Speech Time (ROST henceforth) results from dividing the total number of Xavier Martin-Rubió, Irati Diert-Boté44 syllables uttered by the ST; the Average Rate per bp-unit (ARbpu) results from cal- culating the speech rate of every bp-unit, and then establishing the average rate of all bp-units. The third measure is the Mean Syllables per Run (MSR), which results from dividing the total number of syllables by the number of runs (or bp-units). For the second phase of the analysis, the bp-units in this spreadsheet were copied and pasted into another tab, and four new columns were added. Two disfluency and two accuracy measures were chosen: on the one hand, rep- etitions and false-starts/self-corrections (FS/SC) on the one hand (disfluency); and pronunciation, and lexical and grammatical errors on the other (accuracy). These measures were determined separately by the two authors of this paper, who then compared and discussed their results. We started from a number of rules and guiding criteria, but even so several minor issues emerged during the meetings. For example, a rule to measure repetitions was that all repeated consecutive syllables would be counted. In the case of the utterance “in the_ in the house,” for instance, two repeated syllables were counted because the participant was undoubtedly repeating these syllables. However, in “ha_ had,” the two syllables are different so it was agreed that this would be considered a false-start, as we do not surely know what the speaker intended to say in the first syllable. Pronunciation errors were particularly complex to determine. On several occasions, isolating the specific expression was what allowed us to come to a decision. The next two sections deal with the two different kinds of findings. An analysis of the results of the fluency and accuracy measures will be presented in Results. Results Table 1 contains the figures for the six stories produced by the participants before and after their stays in Italy. Josep Maria and Beatriu start from very low MSRs (3.88 and 3.15 respectively), but whereas Beatriu increases to 5.26 MSR (a 67% increase), Josep Maria actually decreases to 3.09 MSR (a 20% decrease). Alma’s MSR increases 120%, from 5.45 to 12.18. Alma and Beatriu produce many more syllables in their post-stories (90% and 59% increases, respectively), whereas Josep Maria produces just one more syllable. Also remarkably different are the pausing behaviors. STR for Alma and Beatriu remain similar (small in- creases of 4.2% and 8.9%), whereas Josep Maria’s STR decreases 33.2%. Josep Maria pauses more than he produces syllables, although we must indicate that there is a pause of more than eight seconds (chunk 62) that is partly responsible for this very unusual measure. Speech rates also increase for Alma and Beatriu, whereas Josep Maria’s rate is slower in his post-story. Catalan Law and Business Students in Italy… 45 Table 1. Quantity, temporal fluency, disfluency and accuracy measures Josep Maria Pre-story Post-story Alma Pre-story Post-story Beatriu Pre-story Post-story Syllables 101 102 109 207 63 100 ST 37.08 40.68 39.43 54.75 24.70 37.54 PT 23.72 59.13 14.32 16.88 14.90 17.76 TRT 60.80 99.81 53.74 71.63 39.60 55.30 STR 60.98% 40.75% 73.36% 76.43% 62.37% 67.89% ROST 2.72 2.51 2.76 3.78 2.55 2.66 ARbpu 2.58 2.46 2.63 3.66 2.44 2.65 MSR 3.88 3.09 5.45 12.18 3.15 5.26 Rep 9.90 6.86 0 0 1.59 7.00 FS/SC 2.97 3.92 0.92 0.97 3.17 6.00 Pr E 6.93 0.98 1.83 0 3.17 7.00 L&G E 10.89 17.65 6.42 5.80 14.29 10.00 Disfluency measures indicate very different patterns: Alma never repeats syllables and produces very few FS/SC; Beatriu clearly increases her rep- etitions and FS/SC in the post-story; and Josep Maria slightly decreases the number of repeated syllables per 100 syllables and slightly increases in FS/SCs. Combined with the fluency measures, we see that Alma improves a great deal, Josep Maria performs worse, and Beatriu is the one who improves in fluency measures but with more disfluency, which may be indicating that she is more willing to take risks. In pronunciation accuracy terms, Josep Maria only produces one error in his post-story, but this is due to the very limited number of words he uses. He spends 43 of the 102 syllables he produces in his post-story with the words “and,” “the,” “boy(s),” and “(foot)ball.” He even utters at least one word in Italian (acqua), and possibly three more ( per, te and i). We are not sure about these last three because they are also words in Catalan. In his pre-story, he makes seven pronunciation errors, but three of them correspond to the way he pronounces “sequence,” a word he does not use in the post-story. In Josep Maria’s case, and due to a mistake in the protocol, he described the same comic on both occasions, so that a very unusual situation arises: bp09 of his pre-story and bp12 of his post-story are almost identical. In the first he says “in approximately snake” and in the second “and approximately the snake.” We have counted two pronunciation errors in the pre-story (wrong stress in “ap- proximately” and wrong vocalic sound before “s” in “snake”) and one in the post-story (still wrong stress, but this time the vowel has a reason to be there because of the definite article “the” before “snake”). Xavier Martin-Rubió, Irati Diert-Boté46 Beatriu produced more pronunciation errors in her post-story, but this is mostly because of the vocabulary in the two stories. Four of the seven pronun- ciation mistakes have to do with voiced final consonants, which she pronounces with total voicelessness (“dog”—three times and “bag”—once); two errors correspond to “mountain,” which she pronounces without the diphthong in the first syllable; and the last corresponds to the wrong presence of an “l” sound in “walking.” In the pre-story, in which she only uttered 63 syllables, there were no words used ending in final “g.” In the pre-story she mispronounced “hole” in two different ways, in fact (first without the diphthong and then without final /l/). So we can actually see some slight improvement in general, and the increase in pronunciation errors could be due to the lack of chances to make the errors. Finally, Alma makes no pronunciation errors in her post-story, and only two in the pre-story. Wrong pronunciation of “hole” (no final /l/) and wrong vocalic sound in “third.” Alma produces seven lexical & grammatical errors in her pre-story and 12 in her post-story, but since her post-story is almost twice as long, there is actually an improvement in accuracy (6.42 to 5.80 errors per 100 syllables). Beatriu also improves (nine errors in her pre-story and ten in her post-story mean a 30% decrease in the number of errors per 100 syllables). Josep Maria produces 18 errors in his post-story, coming from 11 in the pre- story. Given both stories are just one syllable apart in length, accuracy worsens. In the next section, details from the contextual circumstances of the three students’ stay, as self-reported in the group discussions and interviews, will complement the fluency and accuracy analysis. Insights from the Qualitative Data of the Study The qualitative data for the Italy data sub-set are, on the one hand, focus group discussions, and, on the other, interviews. Table 2 provides details about the three group discussions. The reason why there are two post-focus group discussions is that the students who went to Macerata did so during the first semester, which means they were back from Italy in January already, so the fo- cus group in June would have taken place too long after their return. The three students in the other focus group had just returned from Italy, although whereas Beatriu had spent a whole academic year in Milano, Alma and Gertrudis had been there during the second semester. The interviews were conducted while the students were in the host HEIs. Two researchers travelled to Milano and Macerata, and conducted three inter- views in Milano and one in Macerata. Alma and Gertrudis were interviewed Catalan Law and Business Students in Italy… 47 together, and the audio recording lasted 40’ 47’’ (INT_MI_1). Beatriu was interviewed separately, and it lasted 53’ 39’’ (INT_MI_2). Josep Maria was interviewed in Macerata, and the audio recording lasts 56’ 58’’ (INT_MA_1). Table 2 Group discussions CODE Date Participants Time pre-FG 18/06/2013 4 students (including Beatriu) + 2 researchers 123 minutes post-FG1S 20/03/2014 3 students (including Josep Maria) + 2 researchers 53 minutes post-FG2S 17/06/2014 Alma, Beatriu, and Gertrudis + 2 researchers 44 minutes Three factors emerge from these groups discussions and interviews. A first factor concerns the starting self-confidence and proficiency level of the stu- dents. Beatriu could not apply for an Erasmus grant in the North of France, as she wanted to, because she did not have the required B1 level of English. She passed the test in June 2013, right before going to Italy. However, throughout the interview she expresses her lack of confidence when speaking English. Josep Maria also starts the stay without the B1 level, whereas Alma had a strong B1 and she is the one who benefits the most from the stay. The second element is where they lived and the people they interacted with the most (and the languages in which they did that). Beatriu was accom- modated in a hall of residence located on the outskirts of Milano, which was full of Spaniards (something she complains about). However, she shared a room with a girl from Slovakia, with whom she had to use English. Beatriu saw the need to move to another room to practice more Italian and English. In fact, the three participants complained about a great number of Spanish students in the program, and they realized that Italy was not the best place to practice English, at least for two reasons: the low English level of Italians and an ex- cessive number of Spaniards. Gertrudis even said she practiced Spanish (her L1 is Catalan, after all). However, Alma (and Beatriu to a lesser extent) did improve their English, which might be due to the fact that they end up using more English than at home: Beatriu with her Slovakian room-mate, and Alma (and Gertrudis) with the German, Dutch, and Romanian people in their hall of residence. The third element is the nature of the expectations the three students had in relation to their stay. The three students in post-FG1S indicate they expected a higher presence of English in general, and admit they would change their destination if they could go back. Josep Maria mentions the Czech Republic and Finland as places he would now choose, since at least there he would have an opportunity to practice English and to experience a truly different culture. However, learning Italian was also part of the plan. The lessons at the three Xavier Martin-Rubió, Irati Diert-Boté48 universities were conducted in Italian, and they had to take oral exams in that language, whereas non-Spanish Erasmus students took the lessons in English, as most did not have learning Italian amongst their goals. Alma, for example, improved her English but also learned Italian. In the post-focus group she claimed that she would even write “a bit of Italian” in her résumé. From Josep Maria’s post-story, it is obvious he also learnt Italian. Discussion and Concluding Remarks Most research conducted in SA contexts seem to indicate that oral fluency is benefited from SA, whereas this does not seem to be the case for accuracy, where findings are less conclusive (Juan-Garau & Pérez-Vidal, 2007). Freed et al. (2004) compared the progress of 28 American students of French studying in three different learning contexts. The authors discovered that determining the amount of time students actually spent using the target language was es- sential in order to accurately interpret the findings. Apparently, the students in the immersion program, who were the ones with the greatest improvements, reported “devoting significantly more time to using French in out-of-class activities compared not only to the students in AH [at home] context but also to those in the SA context” (Freed et al., 2004, p. 294). In this line, Llanes et al. (2018) suggest that certain individual differences such as motivation and foreign language use in free time (among others) play a role in the language gains students will experience in the different contexts to the point that they can account for the different outcomes. In our case, the two students who went to Milano learnt some Italian and simultaneously improved their English proficiency. Although their immersion in the Italian language was high, they also mixed enough with students from different countries and linguistic backgrounds, which may have made an impact on their progress in English. In contrast, the student who went to Macerata does not show an improvement in the language. The stay has certainly done little for his English, and his feeling at the end is that his destination was a place similar to Lleida (similar weather, food, culture, etc.), and although he has learnt some Italian, he would have probably benefited more from a stay in a country further North within Europe. The results highlight the many factors contributing to the context of stay abroad (which in turn affect the process of additional language learning) and resonate with findings from other studies like Borghetti and Beaven (2015), Kaypak and Ortaçtepe (2014), Kalocsai (2009) and Koylu (2016). These factors can be the role that English plays in the specific campus and in the town/city/country, the language repertoires of the network Catalan Law and Business Students in Italy… 49 of friends the students establish, and the attitudes which the different subjects adopt towards the different local languages. Some limitations of the study must be pointed out. The oral tests were monological productions in which students had a minute to prepare themselves and then they improvised a short descriptive text. However, what they prac- ticed the most during their stay were dialogical improvisations. In this sense, it would have been interesting to analyze the students’ progress in dialogical productions in English. Another limitation of the study is that Josep Maria was given the same comic on his return, and even though several months had elapsed, this makes the comparison with the other two participants problem- atic. Certainly, it would have been more suitable to follow the same strategy for all the participants. Another issue is the small number of participants; this entails that the results must be taken with caution. A final limitation has to do with the monolingual bias of this study. The project as a whole aimed at identifying the impact of the stay on three main areas: language level, intercultural awareness, and feeling of Europeanness. Although the students participating in the qualitative part of the study were asked about their rela- tionship with the local languages, it would have been very enriching to include tests in the local languages of their host HEIs to see their progress in those languages. I believe this paper points towards at least two different lines for further research. The first would be new studies comparing pre- and post-oral produc- tions of not just monological but also dialogical tasks from students engaged in different kinds of methods to improve their English. The second would be a more ethnographic approach to the many factors influencing additional lan- guage learning while studying abroad. References Arnold, J. (Ed.) (1999). Affect in language learning. Cambridge University Press. Borghetti, C., & Beaven, A. (2015). Lingua francas and learning mobility: Ref lections on stu- dents’ attitudes and beliefs towards language learning and use. International Journal of Applied Linguistics, 27(1), 221–241. Brown, H. D. (2004). Principles of language learning and teaching (4th ed.). Longman. Canagarajah, S. (2015). Clarifying the relationship between translingual practice and L2 writing: Addressing learner identities. Applied Linguistics Review, 6(4), 415–440. Chambers, F. (1997). What do we mean by f luency? System, 25(4), 535–544 . Cots, J. M. (2013). Introducing English-medium instruction at the University of Lleida, Spain: Intervention, beliefs and practices. In A. Doiz, D. Lasagabaster, & J. M. Sierra (Eds.), English-medium instruction at universities: Global challenges (pp. 106–130). Multilingual Matters. Xavier Martin-Rubió, Irati Diert-Boté50 Czwenar, I. (2014). Analysing spoken language for complexity, accuracy and f luency: Some methodological considerations. In W. Szubko-Sitarek, Ł. Salski, & P. Stalmaszczyk (Eds.), Language learning, discourse and communication (pp. 81–92). Springer. Dafouz, E., & Camacho-Miñano, M. M. (2016). Exploring the impact of English-medium instruction on university student academic achievement: The case of accounting. English for Specific Purposes, 44, 57–67. Diert-Boté, I. (2016). “I didn’t learn English at school”: An exploration of English language learners’ beliefs and emotions.” Unpublished Master’s Thesis, Universitat de Lleida. Doiz, A., Lasagabaster, D., & Sierra, J. M. (2014). Language friction and multilingual policies in higher education: The stakeholders’ view. Journal of Multilingual and Multicultural Development, 35(4), 345–360. European Commission (2012). Europeans and their languages. Special Eurobarometer, 386. http://ec.europa.eu/commfrontoffice/publicopinion/archives/ebs/ebs_243_en.pdf Fortunato-Tavares, T., Howell, P., Schwartz, R. G., & Furquim de Andrade, C. R. (2017). Children who stutter exchange linguistic accuracy for processing speed in sentence com- prehension. Applied Psycholinguistics, 38(2), 263–287. Freed, B. F. (1990). Current realities and future prospects in foreign language acquisition research. In B. F. Freed (Ed.), Foreign language acquisition research and the classroom (pp. 3–27). DC Heath. Freed, B. F., Segalowitz, M., & Dewey, D. P. (2004). Context of learning and second language f luency in French: Comparing regular classroom, study abroad, and intensive domestic immersion programs. Studies in Second Language Acquisition, 26(2), 275–301. Generalitat de Catalunya (2017). Avaluacions censals de competències bàsiques. Consell Superior d’Avaluació del Sistema Educatiu, July 6. Accessed 5 March 2018, http://premsa. gencat.cat/pres_fsvp/docs/2017/07/06/17/17/5365f81b-cb55-45ba-b015-a070b412178e.pdf Ginsberg, R. B., & Miller, L. (2000). What do they do? Activities of students during study abroad. In R. D. Lambert & E. Shohamy (Eds.), Language policy and pedagogy: Essays in honor of A. Ronald Walton (pp. 237–260). John Benjamins. Ginther, A., Dimova, S., & Yang, R. (2010). Conceptual and empirical relationships between temporal measures of f luency and oral English proficiency with implications for automated scoring. Language Testing, 27(3), 379–399. Heaton, J. B. (1966). Composition through pictures. Longman. Hieke, A. E. (1985). A componential approach to oral f luency evaluation. Modern Language Journal, 69(2), 135–142. Jiménez Arias, I. (2004). Treating students’ errors in oral production. Letras, 36, 175–188. Juan-Garau, M., & Pérez-Vidal, C. (2007). The effect of context and contact on oral performance in students who go on a stay abroad. Vigo International Journal of Applied Linguistics, 4, 117–134. Kalocsai, K. (2009). Erasmus exchange students: A behind-the-scenes view into an ELF community of practice. Apples – Journal of Applied Language Studies, 3(1), 25–49. Kaypak, E., & Ortaçtepe, D. (2014). Language learner belief and study abroad: A study on English as a lingua franca (ELF). System, 42, 355–367. Koylu, Z. (2016). The influence of context on L2 Development: The case of Turkish under- grtaduates at home and abroad. (Unpublished doctoral dissertation). University of South Florida, USA. Leal, C. F. (1995). A pause as a cause of change. ERIC document ED 379935. Educational Resources Information Center. Llanes Baró, À., & Serrano Serrano, R. (2011). Length of stay and study abroad: Language gains in two versus three months abroad. Revista Española de Lingüística Aplicada, 24, 95–110. http://ec.europa.eu/commfrontoffice/publicopinion/archives/ebs/ebs_243_en.pdf Catalan Law and Business Students in Italy… 51 Llanes, À., Tragant, E., & Serrano, R. (2018). Examining the role of learning context and individual differences in gains in L2 writing performance: The case of teenagers on an intensive study-abroad programme. The Learning Language Journal, 46(2), 201–216. MacIntyre, P. D., Dörnyei, Z., Clément, R., & Noels, K. A. (1998). Conceptualizing willingness to communicate in a L2: A situational model of L2 confidence and affiliation. The Modern Language Journal, 82(4), 545–562. Maiworm, F., & Wächter, B. (2014). The big picture. In B. Wächter & F. Maiworm (Eds.), English-taught programmes in European higher education the state of play in 2014 (pp. 25–62). Lemmens. Mancho-Barés, G., & Arnó-Macià, E. (2017). EMI lecturer training programmes and academic literacies: A critical insight from ESP. ESP Today, 5(2), 266–290. Manchón, R., & Murphy, L. (2002). An introduction to second language acquisition research. UMU/Diego Marín. Martin-Rubió, X., & Cots, J. M. (2016). Englishisation at a global space: Students and staff mak- ing sense of language choices. Language and Intercultural Communication, 16(3), 402–417. Park, J.-K. (1997). ‘English-fever’ in South Korea: Its history and symptoms. English Today, 25(1), 50–57. Salaberri-Ramiro, M. S., & Sánchez-Pérez, M. M. (2015). Analyzing writing in English-medium instruction at university. Linguarum Arena. Revista de Estudos em Didática de Línguas da Universidade do Porto, 6, 45–58. Salö, L. (2014). Language ideology and shifting representations of linguistic threats: A Bourdieusian re-reading of the conceptual history of domain loss in Sweden’s field of language planning. In A. K. Hultgren, F. Gregersen, & J. Thøgersen (Eds.), English in Nordic universities: Ideologies and practices (pp. 83–110). John Benjamins. Starkweather, C. W. (1987). Fluency and stuttering. Prentice Hall. Tragant, E., Serrano, R. & Llanes, A. (2017). Learning English during the summer: A com- parison of two domestic programs for pre-adolescents. Language Teaching Research, 21(5), 546–567. Universitat de Lleida, 2014. Memòria Acadèmica del curs 2013–2014. Lleida: Edicions de la Universitat de Lleida. http://www.udl.cat/export/sites/universitat-lleida/ca/organs/secretaria/. galleries/docs/Memoria-Academica/Memoria_general_de_la_UdL_13_14.pdf Wennerstrom, A. (2000). The role of intonation in second language f luency. In H. Riggenbach (Ed.), Perspectives on fluency (pp. 102–127). University of Michigan Press. Xavier Martin-Rubió, Irati Diert-Boté Katalanische Jura- und Wirtschaftsstudenten in Italien: Die Auswirkung eines Auslandsaufenthalts auf Sprachkompetenzen und Ausdrucksfähigkeit Z u s a m m e n f a s s u n g Die wachsende Zahl der in ganz Europa angebotenen englischsprachigen Lehrveranstaltungen verbessert die Möglichkeiten des Studentenaustauschs. In der vorliegen- den Studie werden die Fortschritte von drei Studenten der Universitat de Lleida nach ihrem Erasmus-Aufenthalt an drei verschiedenen europäischen Universitäten – zwei in Mailand und Xavier Martin-Rubió, Irati Diert-Boté52 einer in Macerata – analysiert. Vor und nach ihrem Auslandsaufenthalt nahmen die Studierenden an einer monologischen mündlichen Englischprüfung teil, bei der ihre Sprachkompetenzen und Ausdrucksfähigkeit beurteilt wurden. Darüber hinaus wurden sie interviewt und be- teiligten sich an Fokusgruppendiskussionen. Die Ergebnisse der beiden Studenten, die nach Mailand gegangen sind, zeigten eine Verbesserung ihrer Englischkenntnisse, während der Student, der nach Macerata gegangen ist, nach seiner Rückkehr noch schlechter abschnitt. Allerdings unterschieden sich die Erfahrungen der beiden Studenten aus Mailand erheblich von denen des Studenten aus Macerata. Anhand von ethnographischen Informationen und ver- fügbaren qualitativen Daten wird ein Versuch unternommen, die voneinander abweichenden Ergebnisse nachzuvollziehen. Schlüsselwörter: Sprachkompetenzen, Ausdrucksfähigkeit, Auslandsstudium, Fremdsprachen- lernen, Engischisierung, ELF