Backness Agreement in Consonant + Glide Onsets: Evidence from Mandarin

This paper extends Duanmu’s (2000, 2007) analysis of Mandarin consonant–glide (CG) onsets by arguing that articulator dissimilation alone cannot account for the full range of grammatical and ungrammatical onset clusters. While Duanmu’s Articulator Dissimilation (2007: 28-30) principle explains the ban on Labial + Labial sequences (e.g., *[fw]), it fails to account for illicit clusters like *[fj], where no articulator overlap exists. This study proposes that both *CG-LabLab and backness agreement constraints are required. Specifically, constraints such as Agree[back]:DorG, Agree[back]:C[-son, +cont]j, and Agree[back]:C[-son]ɥ better explain the distribution of CG onsets. Mandarin phonotactics reflect both OCP effects and anti-OCP agreement patterns for ease of articulation and phonetic distinctiveness. I also investigate the classification of the front glides as Coronal versus Dorsal and demonstrate that this distinction does not alter the requirement for backness agreement, suggesting that both major articulator features and dependent features like [back] are essential to a unified analysis of CG onset constraints in Mandarin. The present analysis takes a novel approach by combining the Optimality Theory with the sonority scale to better account for the missing CG combinations and the asymmetry of obstruent patterns in Mandarin onsets. Additionally, glide asymmetry is observed: [ɥ] is distributionally scarce and more restricted. Such markedness scale and sonority scale within complex onsets are also found in other Chinese varieties, including Lanzhou, Fuyang and Putian. These findings suggest broader cross-linguistic applicability and shed light on the phonological principles governing onsets across languages.