Featural Faithfulness in Consonantal Repairs of Arabic Loanwords in Sundanese

This paper provides an Optimality-Theoretic account of the consonantal repairs employed by Standard Sundanese to adapt seven illicit Arabic consonants in Arabic loanwords and thus bring them into conformity with its phonemic system. The study supports Best's Perceptual Assimilation Model (PAM), which posits that the unfamiliar consonants are perceived to their closest native counterparts in terms of articulatory features. It has been found that all seven Arabic consonants are replaced with native Sundanese counterparts with, first and foremost, the same major articulator (i.e., labial or coronal: IDENT (ARTICULATOR)) and voicing quality (i.e., [±voice]: IDENT (VOICE)), followed by the same manner of articulation (i.e., IDENT (MANNER)), which is preserved in the adaptation of five Arabic consonants only. In addition, while the source place (i.e., IDENT (PLACE)) and stridency (i.e., IDENT (STRIDENCY)) features are matched in only three adaptations each, IDENT (C) (i.e., no change of features) is never reflected in the data. Moreover, the adaptation of Arabic consonants satisfies markedness restrictions (represented by *EMPHATIC, *LABIODENTAL, *DENTAL, */f/, */z/, and */ʃ/) in Sundanese phonology against the importation of Arabic consonants. This is captured by the following OT constraint ranking: MARKEDNESS, IDENT (ARTICULATOR), IDENT (VOICE)>> IDENT (MANNER)>> IDENT (PLACE)>> IDENT (C). In conclusion, this paper contributes to the insufficiently researched area of Sundanese loanword phonology.