A Comparison of the Syllable-initial /n/ and /l/ in Standard Mandarin: Articulatory, Acoustical, and Aerodynamic Characteristics

Introduction. The syllable-initial consonants /n/ and /l/ are common in languages. While /n/ being a nasal is articulated with complete oral closure but lowered velum for nasal airflow, /l/ is a lateral involving central tongue-palate contact with lowered sides of the tongue for lateral airflow. The said articulatory characteristics consequently contribute to the differences in acoustics and aerodynamics between /n/ and /l/. In Standard Mandarin, /n/ and /l/ are clearly distinguished as two separate phonemes, while the merger between /n/ and /l/ is observed in many other Chinese dialects. This study investigates the articulatory, acoustical, and aerodynamic characteristics of the syllable-initial /n/ and /l/ in Standard Mandarin. It compares between /n/ and /l/ in different prevocalic environments.
Participants and Materials. Eight young adult speakers of Standard Mandarin participated in the study to utter a list of CV monosyllabic words. In the words, the syllable-initial consonants were /n/ and /l/, followed by the vowels /i/, /y/, /u/ and /a/.
Ultrasound Imaging. The ultrasound speech system was used to obtain the mid-sagittal view of the speakers’ tongue configuration during the test words. The images of the tongue contours of the syllable-initial /n/ and /l/ and the following vowels were subsequently analyzed using the generalized additive mixed models (GAMMs). The results show that /n/ and /l/ generally in alveolar articulation share similar tongue shape and tongue position, and their major difference lies in the tongue body posture. Compared to /n/, /l/ shows a larger variation in tongue body posture across different vowel contexts, due to anticipation of articulatory gesture of the following vowel. This suggests that in the articulation of /l/, the lowering of the lateral rims provides more flexibility for the tongue body to move in anticipation of the following vowels. As for /n/, it achieves complete anterior and lateral closure, which appears to inhibit the movement of the tongue body to coarticulate with the following vowels.
Acoustic Analysis. During the ultrasound imaging, synchronized audio data were recorded for acoustic analysis of the formant frequencies and energy distribution of the prevocalic /n/ and /l/ in the test words. Statistical analyses, including linear discriminant analysis (LDA) and linear mixed-effects models, were performed to determine the differences in acoustic variables between /n/ and /l/. It is found that the primary difference between the two consonants lies in their spectral energy distribution. While /l/ shows enhanced high-frequency energy, /n/ exhibits a spectral energy distribution skewed toward the lower frequency region. Regarding the spectral characteristics, /n/ and /l/ generally share similar formant structures. However, the formant frequencies of /l/ are more sensitive to vowel context, in contrast to /n/’s formant patterns which are relatively stable across the vowel contexts. This finding is in parallel with the greater coarticulatory variation for /l/ than /n/ in different vowel contexts.
Aerodynamic Analysis. The DualView airflow system was also used to collect the synchronized oral and nasal airflow data from the speakers in order to compare the difference in nasalance between /n/ and /l/. The nasalance is quantified as a nasal ratio (NR) which indicates the proportion of nasal airflow to the total (nasal + oral) airflow. As the NR value between 0 (low nasalance) and 1 (high nasalance), the zero-or-one-inflated beta regression models were used for statistical analysis. The results show that as expected, /n/ exhibits a higher NR than /l/ in any vowel context. However, the difference in NR between /n/ and /l/ is reduced in the /i/ context, due to a notable increase in NR for /l/ before /i/. The finding suggests that the perceptual distinction in nasality between /n/ and /l/ is less apparent in the prevocalic /i/ context. Comparing between the vowels following /n/ and /l/, the NR is considerably high for the vowels following /n/ due to the carryover effect. For the vowels following /l/, the NR is noticeably raised for the low /a/. This consequently causes in a reduction in NR difference between /a/ in the /n/ and /l/ contexts. The finding suggests that the prevocalic /a/ context has minimal help for the perceptual distinction in nasality between /n/ and /l/.
Conclusion. Overall, the experimental data obtained in this study reveal the differences in articulation, acoustics, and aerodynamics between /n/ and /l/ in different prevocalic contexts, which contribute to the perceptual distinctiveness between the two coronal consonants in Standard Mandarin. The findings of this study are useful information for studies investigating the merger of the syllable-initial /n/ and /l/ as reported in many Chinese dialects.