Cost and benefit in loanword adaptation

This paper examines the model of loanword adaptation developed within the Theory of Constraints and Repair Strategies in the light of Marshallese borrowings from English. The claim that structure preservation is subject only to a cost threshold principle is shown to be untenable. A modification to the model involving the introduction of a benefit threshold is suggested, together with an alternative strategy within the framework of Optimality Theory. 1. Structural problems Any language which prohibits consonant clusters in syllable initial and syllable final positions or perhaps- even worse- completely bans syllable final consonants, faces difficulties when borrowing words from a language in which the structural constraints on onsets and codas are less severe. The borrowing of English words by a wide variety of Oceanic languages provides many examples of the range of problems encountered and of the types of solution adopted in order to integrate alien forms. 2. Structural solutions In order to focus on syllable structure problems without distraction, let us ignore for a moment the possibility that some or even all of the segments of an inadmissible syllabic structure might individually need to be adapted before they are able to surface as acceptable segments in the borrowing language. With that proviso, there are clearly two ways in which the unacceptability of a consonant sequence or an illicit single coda consonant might be dealt with: A. By epenthesis Consonants in a cluster may be separated by inserting an (epenthetic) vowel-or if necessary vowels- between them. Alternatively a vowel may be inserted before a cluster word-initially or after a cluster (word-finally). Single final consonants, when codas are completely excluded, may be similarly modified by the addition of a following vowel. All of these tactics resolve the structural problem by the creation of additional syllables within which the problem consonants may be relocated as single onsets or codas