The three syllable-final stop consonants (-p, -t, -k) in Middle Chinese syllables have been lost or altered in many modern Chinese dialects. In comparison, southern dialects such as Hong Kong Cantonese preserve the three final stop consonants. However, in recent years, the final stop consonants in Hong Kong Cantonese have shown signs of instability, such as cases of -t and -k merging. This study analyzes the pronunciation and phonation features of final stop consonants in checked syllables in Hong Kong Cantonese through experimental phonetic methods. Based on the analysis results, it identifies (1) changes in place of articulation and glottalization phenomena of -p, -t, -k in contemporary Hong Kong Cantonese, (2) explores differences in change patterns among different finals and their correlation with individual phonetic properties, and (3) investigates whether the sound changes of final stop consonants are influenced by preceding vowel types and tone types. Based on the obtained results, it summarizes the phonetic properties and changes of final stop consonants in Hong Kong Cantonese, and establishes models and theoretical discussions for the changes of stop consonant finals in Chinese dialects as a whole.

The merger of nasal finals -n and -ŋ in Chinese dialects is a long-standing phonetic evolution phenomenon. Research on the production of nasal finals by Putonghua speakers in dialect regions is already quite abundant, but the underlying cognitive mechanisms, particularly processing mechanisms at the levels of lexical representation and articulatory planning, remain to be revealed. This study focuses on the increasingly numerous "receptive bidialectals" (i.e., individuals who use Putonghua as their daily communication language, can understand the dialect but rarely use it), exploring the activation patterns of nasal final features in Chinese character recognition and production, as well as the influence of different dialect backgrounds. The study adopts the Stroop interference paradigm from cognitive experiments to examine the implicit role of multiple phonological representations in lexical access.

Rhotacization (erhua) is an important phonological phenomenon in Putonghua and Chinese dialects. The r sound in rhotacization can appear at the end of a syllable through morphophonological changes, and interacts with the preceding vowel in complex phonological and phonetic ways. This is particularly special in Putonghua, which only allows nasals as finals, and provides an important window for examining final changes and syllable structures in the Chinese phonological system. Existing research has shown that the acquisition of final r sounds is one of the difficult points in children's phonological development in many languages. This study focuses on the acquisition process of rhotacized finals in Chinese children, using acoustic analysis methods to examine children's production characteristics of rhotacized finals at different developmental stages, providing new evidence for understanding the acquisition mechanisms of Putonghua children's phonetics and phonological development.

Based on the Functional Load (FL) index, this paper conducts quantitative statistics on Cantonese rhyme contrasts, focusing on analyzing functional differences among different finals. Results show that Cantonese rhyme contrasts form a clear hierarchy: vowel contrasts without finals carry the highest load, nasal finals overall outperform checked stop finals; within nasal finals, there is a functional gradient of -ŋ > -n > -m, while checked finals show -k > -p > -t, with -t final weakening being the most significant. Homotypic finals are prone to contrast neutralization, and weak finals significantly reduce the overall semantic discrimination ability of combinations. Combining the Functional Load Hypothesis, it can be seen that Cantonese finals exhibit a differentiated decline pattern: the vowel and -ŋ final systems remain stable, while -m nasal finals and -p/-t checked finals continue to disappear, representing a quantitative manifestation of diachronic sound changes at the synchronic level.

Previously, Luo Changpei and others in "Tang and Five Dynasties Northwestern Dialects" used Sino-Tibetan transcription, Tibetan transcription materials, etc., to discuss the characteristics of 8th-10th century northwestern dialects. Regarding nasal finals, Luo Changpei believed that the nasal finals of the Dang and Geng rhyme groups had partially begun to change, while -m and -n finals were relatively well preserved. Regarding checked finals, due to the nature of Tibetan script itself, Luo Changpei believed that Tibetan corresponded to -b, -d(r), -g, etc.; although there was no evidence of complete disappearance, -r and -g finals showed signs of disappearance from the Five Dynasties period.

Since then, apart from using Dunhuang excavated materials, there have been almost no other materials available for discussing northwestern dialects. Fortunately, the large-scale excavation of Western Xia documents in the 20th century has provided us with materials on 10th-13th century northwestern dialects, allowing us to use Western Xia's Sino-Western Xia and Sanskrit-Western Xia transcription materials to explore the characteristics of northwestern dialects. From current academic research, checked finals in the northwestern region during the 10th-13th centuries had almost completely disappeared, and the Dang and Geng rhyme groups with nasal finals had also nearly been lost. However, since the Western Xia language itself had no nasal finals, the characteristics of nasal finals other than the Dang and Geng groups should bear "foreign-style" features, that is, features appearing when people in ethnic regions learned Chinese.

At the same time, the northern Khitan regime established the Liao dynasty. Academic research through Khitan small script and other materials suggests that checked finals also disappeared in northern dialects, with main vowels becoming diphthongs, which is very likely the earliest source of the checked tone vernacular readings in modern Beijing dialect. The yang-tone finals -m, -n, -ŋ remain completely preserved.

This study focuses on the speech perception mechanisms of nasal finals in Putonghua and adaptation effects in second language acquisition. Based on Enhancement Theory, through Gating Test, synchronized ultrasound tongue imaging, and Generalized Additive Mixed Model (GAMM) analysis, it was found that under conditions where the closure segment of nasal finals is weakened or disappears in continuous speech, the formant dynamic shifts produced by tongue body coarticulation of the low vowel /a/ can serve as distinctive phonological feature cues to assist native speakers (L1) in completing /n/-/ŋ/ contrast recognition before the final appears; while second language speakers (L2) universally exhibit phonetic-phonological cue integration deficits, unable to use vowel transition cues, and highly rely on the auditory presentation of the final closure segment.

Further analysis based on HSK proficiency levels shows that L2 speakers' perception levels vary with Chinese proficiency: HSK Level 4 learners have no stable perception pattern, while HSK Level 6 learners' dynamic perception trajectories approach those of native speakers. Research results indicate that L1 phonological representations contain fine phonetic enhancement details, while the bottleneck in L2 perception may precisely lie in the insufficient acquisition of such contextual variants.