TL;DR
This study investigates how well Japanese learners of French can produce and understand French rising intonation contours, which are crucial for conveying meaning and structure. The findings suggest that while learners can mimic rising contours, they struggle to correctly produce or perceive them when they are violated or tied to specific syntactic boundaries, indicating a lack of phonological awareness of these French prosodic features.
What Problem Does This Paper Address?
Understanding the prosody of a language – its rhythm, stress, and intonation – is key to mastering it. In French, rising intonation contours play a significant role, signaling everything from questions to syntactic boundaries within phrases. However, Japanese, a pitch-accent language, marks prosodic boundaries differently, often with a default low tone. This fundamental difference raises a question: how well do Japanese learners of French acquire the nuanced use of French rising contours, particularly when these contours are tied to specific syntactic structures? The paper aims to bridge this gap by examining Japanese learners' ability to both produce and perceive these French intonational patterns, hypothesizing that their proficiency is directly linked to their phonological awareness of these specific French prosodic features.
Methods
The researchers employed two main experimental paradigms:
- Sentence Reproduction Task: Participants listened to French sentences where the natural rising contours at the end of certain phrases (Accentual Phrases - APs, and Intonational Phrases - IPs) were either violated or maintained. They were then asked to reproduce these sentences. The goal was to see if learners could correct unnatural (non-rising) contours to the expected rising ones, and if their ability to do so improved with proficiency.
- Sentence Completion Task: Participants heard the beginning of a French sentence and had to choose the correct word sequence to complete it from two options. This task was designed to test their ability to perceive the subtle differences in rising contours that signal different syntactic boundaries (specifically, the distinction between AP and IP boundaries).
Both experiments involved Japanese learners at different proficiency levels (beginner, intermediate, advanced) and a control group of native French speakers. Stimuli were carefully manipulated using Praat, and data was analyzed using Generalized Linear Mixed Models (GLMM).
Datasets
The study did not introduce new datasets but rather created specific experimental stimuli based on French sentence structures. These stimuli were recorded by native French speakers and then manipulated to create grammatical and ungrammatical prosodic variations. The participants were 38 Japanese learners of French and 12 native French speakers.
Key Insights and Ideas
- Mimicry vs. Awareness: Japanese learners can often reproduce rising contours when they hear them, but this doesn't necessarily mean they understand why they are there or can correct violations. This highlights a distinction between surface-level mimicry and deeper phonological awareness.
- L1 Influence: The default low-tone marking of prosodic boundaries in Japanese likely makes it challenging for learners to internalize the French system of rising contours for continuation and syntactic marking.
- Proficiency Matters, But Slowly: While advanced learners showed some improvement in correcting contours compared to beginners, the gains were not always statistically significant, suggesting that acquiring this aspect of French prosody is a slow process.
- Perception is Particularly Challenging: The sentence completion task revealed that Japanese learners, regardless of proficiency, struggled to differentiate between the subtle rising contour cues that signal different syntactic phrase boundaries (AP vs. IP). This suggests a significant gap in their perceptual understanding of French intonation.
Key Results
- Reproduction: In the sentence reproduction task, Japanese learners were generally able to reproduce rising contours when they were present in the model. However, they were less successful at correcting non-rising contours to the expected rising ones, especially in conditions where the model had a falling contour.
- Perception: In the sentence completion task, Japanese learners performed at chance level, indicating they could not reliably distinguish between the prosodic cues for AP and IP boundaries. Native French speakers, however, performed significantly better, correctly identifying the boundary types.
- Proficiency Trends: While not always statistically significant, there was a general trend for accuracy to increase with proficiency in the reproduction task, particularly in correcting certain types of contour violations.
Limitations and Future Work
The authors acknowledge that the difficulty of the tasks themselves might have contributed to the poor accuracy rates, especially for advanced learners. They also note that the stimuli containing two APs were particularly challenging.
Future research directions suggested include:
- Immersion's Impact: Investigating whether learning French in an immersion environment, as opposed to solely in a classroom setting in Japan, could accelerate the acquisition of phonological awareness of French prosody.
- Exploring Prosodic Explanations: Further exploring the prosodic reasons behind native speakers' perception of different boundary types, potentially linking it to phrasing and syntactic structure.
- Developing Intervention Strategies: Designing and testing pedagogical approaches to specifically target and improve Japanese learners' phonological awareness of French rising contours.
Summary updated