Kari Tanner and Dr. Jeffrey Turley, Spanish Department
When it comes to learning a foreign language, study abroad contexts have generally been assumed to produce the best results. However, with rising program costs it becomes necessary to evaluate alternate options as well. The three main language learning contexts are classroom, foreign language housing (domestic immersion) and study abroad. While many studies have analyzed language gains in these settings independently of each other, little research has actually compared the gains across these three environments. I chose to conduct this research with learners of Italian, so I started by finding participants in each of the three contexts. The classroom learners consisted of university students currently enrolled in Italian classes of various levels. The domestic immersion group were students currently living in BYU’s foreign language student residence (FLSR), with varying levels of Italian. Lastly, the study abroad group were students with varying levels of Italian participating in BYU’s study abroad program in Siena, Italy. Since the majority of previous studies have analyzed language improvements over several months or even years, I decided to examine a much shorter period: 7 weeks.
I interviewed all the participants at the beginning of the 7-week period and asked them to respond (in Italian) to two free-response questions (‘What do you like to do in your free time?’ and ‘Describe your perfect vacation.’), and read one Italian passage out loud (Il vento e il sole’), while audio-recording their responses. At the end of the 7-week period, I conducted the exact same interview for the post-test. I also found a group of native Italians willing to answer the same questions, and used their recordings as a reference point to compare with the non-native speakers. After obtaining the recordings, I spliced the sound files and uploaded them to the online survey program Qualtrics to be rated by native Italians. The native Italian speakers listened to each sound clip and rated them based on a 7-point Likert Scale (e.g., where 1 = no foreign accent; sounds like a native Italian, to 7 = very heavily accented). In addition to the native speaker ratings, I also used the software program ‘PRAAT’ to conduct an acoustic analysis on the recordings.
After tabulating the results and subtracting each student’s post-test score from their pre-test scores, the findings surprised me. Based on the native Italian speaker judgements, classroom learners improved their accent by .064 on the Likert Scale, but actually received negative scores for the other two categories: -1.05 for fluency and -.146 for comprehensibility. Classroom learners received scores of .236 improvement for accent, .320 for fluency and .195 for comprehensibility. Finally, the study abroad students were scored at .481 improvement for accent, .561 for fluency, and .473 for comprehensibility. One surprising factor was the negative numbers in the data, meaning that the post-test scores for certain students were actually worse than their pre-test scores. This outcome may be attributed to the randomization of my survey. (The native Italian speakers would hear the sound files in a random order, not knowing who the student was or if they were listening to the pre- or post-test recording). If I were to do this project over, I would still randomize the pre- and post-test order, but I would keep each student’s recordings grouped together. The native Italian speakers rated the study abroad students as superior in all three categories, but what was surprising to me was that the classroom learners ranked higher than the FLSR students in all three categories as well.
The acoustic analysis tests I performed reflected the native speaker judgements. In almost every category the study abroad students were superior, followed by the classroom learners, while the FLSR students had the lowest scores. Using the ‘PRAAT’ program, I measured the vowel quality based primarily on the first two formants–F1 which measures the height of the tongue and F2 which measures the placement of the tongue. The voice onset time (VOT) was measured by the length of time that passed between the release of the stop consonants /p/, /t/, and /k/ and the start of vocal cord vibration (seen in the sonogram as darker black lines, as compared to the lighter fuzzy sections when the vocal cords are not vibrating). I also analyzed geminate consonants–the double consonants in Italian which give the language its famous melodious tone. I previously believed that geminate consonants were produced by lengthening the consonant, however, my analysis showed that the more important factor at play was the shortening of the preceding vowel, which I found fascinating!
From this data we can conclude that if an Italian learner’s priority is to improve accent, fluency, and comprehensibility over a 7-week period, a study abroad context will provide the best results. Surprisingly, a domestic immersion setting such as BYU’s FLSR did not produce better results than a regular classroom setting, so if students are unable to study abroad, a classroom learning experience will still yield significant improvements. It would be important to note however, in future studies, the starting Italian level of each student, their knowledge of other languages, motivation, and time actually spent reading, writing, and speaking Italian, for these factors also play a large part in second language acquisition. With that being said, one of the most exciting and encouraging findings from comparing the acoustic data with the Italian judgements was that even small phonetic improvements made big differences in the Italian perceptions–every little bit counts!