TL;DR
This paper investigates how Swiss-German speakers' native accentual patterns transfer to their French speech, a variety known as "Français Fédéral" (FF). The study reveals that FF speakers exhibit distinct prosodic features compared to native French speakers, suggesting a stronger influence of their lexical accentuation system from Swiss-German.
What Problem Does This Paper Address?
The paper addresses the under-researched area of prosodic transfer in L2 French, specifically focusing on "Français Fédéral" (FF), a variety spoken by Swiss-German native speakers. While it's known that French and German coexist in Switzerland, little is understood about how the prosodic systems of these languages interact when speakers learn French as a second language. The core question is whether FF speakers' French prosody is closer to the lexical accentuation system of Swiss-German dialects or the supra-lexical accentuation system of native French. This distinction is crucial for understanding language contact phenomena and the acquisition of prosodic features in L2 speech.
Methods
The researchers employed a semi-automatic approach to analyze recorded speech data. Key prosodic features examined include:
- Prominence Distribution and Metrical Weight: They analyzed the ratio of prominent syllables and the average number of syllables per Phonological Phrase (PP).
- Phonological Phrase Formation Constraints: They specifically tested adherence to two constraints:
- Align-XP: A rule dictating that a PP should contain a lexical head and all dependent elements to its left.
- No-Clash: A constraint that restructures PPs to avoid accentual clashes between adjacent words.
- Sandhis Phenomena: They investigated the realization of liaisons and resyllabifications, both within and across PP boundaries.
Statistical tests (chi-squared tests and generalized linear models) were used to assess significant differences between groups, and Multi-Dimensional Scaling (MDS) was employed to visualize the overall similarity and differences among speakers.
Datasets
The study utilized speech data from 16 participants:
- Native French Speakers: 4 from Neuchâtel (NE) and 4 from Paris (PA).
- "Français Fédéral" Speakers: 4 Swiss-German speakers from the Emmental-Oberaargau area (Bern, BE) and 4 from Zürich (ZH).
All participants were adults within a specific age range (55-78 years old). The speech material consisted of a journalistic text read aloud by each participant.
Key Insights and Ideas
- Lexical vs. Supra-lexical Influence: The findings strongly suggest that FF speakers' French prosody is more influenced by the lexical accentuation system of Swiss-German than by the supra-lexical system of native French.
- Smaller Prosodic Units: FF speakers tend to parse their speech into smaller prosodic groups (Phonological Phrases) compared to native French speakers. This aligns with the idea that lexical accentuation often leads to more frequent, shorter prosodic units.
- Challenges with Native French Constraints: FF speakers show less adherence to native French phonological constraints like Align-XP and No-Clash, indicating they haven't fully acquired these supra-lexical rules.
- Sandhis Patterns: FF speakers realize fewer sandhis phenomena (liaisons and resyllabifications), particularly across PP boundaries, which is consistent with a more lexically constrained prosodic system.
- Regional Variation within FF: While both BE and ZH speakers showed FF characteristics, there were some subtle differences, with BE speakers producing even shorter PPs, potentially reflecting regional variations within Swiss-German dialects themselves.
Key Results
- Prominence: FF speakers (BE and ZH combined) produced significantly more prominent syllables than native speakers (PA and NE combined).
- PP Metrical Weight: FF speakers generally produced shorter PPs (fewer syllables per PP) than native speakers. Specifically, BE speakers had the shortest PPs (3.0 syll/PP) compared to PA (3.5), NE (3.3), and ZH (3.2).
- Align-XP: Native speakers (PA and NE) respected the Align-XP constraint in 45% and 40% of cases, respectively. In contrast, FF speakers (ZH and BE) adhered to it in only 10% and 2.5% of cases, showing significantly less respect for this rule.
- No-Clash: Native speakers avoided accentual clashes in about 80% of cases, while FF speakers did so in only about 30% of cases, indicating a significant difference in their ability to restructure PPs.
- Sandhis: Native speakers produced a higher proportion of both internal (PA: 100%, NE: 95%) and external sandhis (PA: 80%, NE: 54%) compared to FF speakers (ZH: 78% internal, 27% external; BE: 74% internal, 21% external).
The MDS analysis visually confirmed a clear separation between the native (PA, NE) and non-native (ZH, BE) speaker groups based on their prosodic characteristics.
Limitations and Future Work
- Spontaneous Speech: The study acknowledges that the findings are based on read speech. Future work is needed to assess if these observed patterns hold true for spontaneous speech.
- Subtleties of Swiss-German: While regional differences within Swiss-German were hinted at (e.g., BE speakers' shorter PPs), a deeper analysis of the specific accentual systems of Bernese and Zurich German could provide further insights.
- Markedness Differential Hypothesis: The authors suggest their findings support the Markedness Differential Hypothesis, which posits that typologically marked systems (like French lacking a strong lexical-prosody system) are less likely to be maintained when in contact with unmarked systems (like Swiss-German's lexical pitch-accent system). Further research could explore this hypothesis more broadly.
Summary updated