TL;DR
This study investigates how native English speakers adjust their speech (hyperarticulation) when interacting with non-native English speakers in a conversational exam setting. They found that speakers tend to hyperarticulate more when addressing less proficient learners, suggesting a potential didactic intent to aid comprehension.
What Problem Does This Paper Address?
The paper tackles the phenomenon of "foreigner-directed speech" (FDS), which refers to the systematic changes native speakers make when addressing non-native speakers. While FDS is thought to aid comprehension, its exact function and the factors influencing it remain unclear, especially in naturalistic conversational settings. Previous research has primarily focused on infant-directed speech (IDS) and has yielded mixed results regarding FDS in conversational contexts. This study aims to model FDS within a controlled, yet ecologically valid, EFL conversational exam to understand how native speakers hyperarticulate and whether learner proficiency influences these adjustments.
Methods
The researchers designed a triadic conversational exam involving native English-speaking examiners and Mandarin-speaking EFL students in Taiwan. They collected video recordings of these interactions. The core methodology involved:
- Data Collection: 31 triadic conversations were recorded, each lasting approximately 12 minutes, focusing on themes of adventure with five guiding questions.
- Participant Grouping: Students were categorized into "high-communicative" (outspoken) and "low-communicative" (less outspoken) groups based on their propensity to engage verbally in English. Triads were treated homogeneously if they contained one or two outspoken students, with control groups consisting of only non-outspoken students.
- Speech Act Coding: Utterances were coded based on speech act categories: "Directives/Facilitating" (positive), "Defending/protesting" (negative), and neutral.
- Acoustic Analysis: The study focused on vowel space, specifically the first and second formants (F1 and F2) of the corner vowels /a/, /i/, and /u/. This involved analyzing acoustic features to quantify hyperarticulation.
- Statistical Analysis:
- Hierarchical Log-linear (HILOG) method: Used to describe discourse patterns and identify exchanges where speech act categories differed across groups.
- Three-way ANOVA: Employed to analyze vowel formant data, considering factors like group membership (high/low communicative), type of event (positive-coded initial utterances, introduction of new questions), and vowel formant (F1/F2).
Datasets
The study utilized a newly created dataset of 31 triadic EFL conversational exam interactions. This dataset consists of video recordings of native English-speaking examiners interacting with Mandarin-speaking EFL university students.
Key Insights and Ideas
- Learner Proficiency Matters: The outspokenness of EFL learners significantly influences how native speakers adjust their speech. Less proficient (low-communicative) learners elicit more hyperarticulation from examiners.
- Didactic Intent Hypothesis: The observed hyperarticulation, particularly towards less proficient learners, suggests a potential didactic purpose – an attempt by the examiner to make speech clearer and aid comprehension, aligning with theories of FDS as a teaching tool.
- Context-Specific Hyperarticulation: Hyperarticulation is not uniform across all interactions. It appears more pronounced during specific speech acts, such as introducing new questions, and differs between vowel formants (F1 vs. F2).
- Vowel Space Differences: The study observed distinct patterns in vowel space adjustments. In F2, more vowels underwent distinct changes across group memberships in positive-coded utterances. In F1, the high-communicative group showed the greatest vowel variation during the introduction of new questions.
- FDS vs. IDS Parallels: The findings echo observations in infant-directed speech (IDS), where hyperarticulation is also thought to aid learning, suggesting a common underlying mechanism for facilitating communication with less experienced listeners.
Key Results
- Examiner Hyperarticulation: Native English-speaking examiners exhibited hyperarticulation in their speech directed at EFL learners.
- Group Differences: The low-communicative group consistently showed more hyperarticulated vowels compared to the high-communicative group, particularly in F2.
- Introduction of New Questions: This specific event type was a significant trigger for hyperarticulation, especially for the low-communicative group.
- Vowel Formant Specificity: F1 and F2 formants showed different patterns of variation. F2 was more sensitive to group membership in positive-coded utterances, while F1 showed greater variation in the high-communicative group during new question introductions.
- Statistical Significance: Several interactions between group, event, and formant were statistically significant (p < .05), supporting the observed differences in hyperarticulation.
Limitations and Future Work
- Limitations: The paper acknowledges that the study was conducted in a specific, albeit ecologically valid, exam setting. The generalizability of findings to other conversational contexts might be limited. The study also notes that the range of F1 formants was narrower than F2, potentially influencing the observed variations.
- Future Work: The authors suggest future research should focus on detailed exploration of sequential interactions between examiners and learner groups, analyzing interlocutor attitudes and expressions. They also propose modeling conversational hyperarticulation during speech acts or problem-solving sequences in cross-cultural settings. Further work is needed to confirm the didactic intent and its effectiveness.
Summary updated