An OT Analysis of Nasal-obstruent Clusters in English

This paper argues in favor of an Optimality Theory(OT)constraint hierarchy governing the phonotactics of nasal+obstruent clusters in English.The paper shows that in careful speech English has a strong tendency favoring voiceless obstruents after nasals both in medial and coda positions,and retains a contrast between voiced and voiceless obstruents,as supported by statistical data.Thus the paper proposes that,in English,the Optimality constraint IDENT-IO(ObsVce)dominates*NO.We also show that in English voiced coronal stops are retained post-nasally,whereas voiced peripheral(labial and velar)stops are deleted(or,rather,not permitted).To explain this phenomenon,the paper proposes that the markedness constraint VOICED-CODA ranks between MAX-COR and MAX-PER,and the constraints IDENT-IO(ObsVce)and*VOICED-CODA work in tandem to result in the deletion of peripheral voiced stops.It shows that the same constraint hierarchy also applies to nasal+non-stop obstruent clusters,i.e.clusters with fricatives and affricates,both medially and in final position.Finally,It is pointed out that as language is changing,this is reflected in a changing constraint hierarchy as well.If*VOICEDCODA and*NO become higher ranked,this would cause either devoicing or deletion of voiced obstruents in the coda,or it would cause obstruents after nasals to become voiced.Both processes are found in faster,informal speech styles and may reflect change in progress.