Early creole syllable structure: evidence from early varieties

Early creole syllable structure 9(17) Engl./Port. ESAnasty > nâsibastard > bassrasoftly > safri, sapri(18) English ESRdoctor > datranasty > nasisoftly > safrisister > sisaIt is not quite clear what determines the choice of the deleted consonant. According toWilson (2001:148), phonological processes of cluster reduction by deletion canonicallytarget the first consonant. This means that the data in (17) and (18) above (with theexception of datra < doctor) would call for a refinement (or even abandonment) of thisputative typological generalization. Taking into account data such as those in (19) rathersuggests an analysis on the basis of sonority. It is the least sonorous consonant that isdeleted:(19) a. English ESAgoodmorrow > gumarragoodnight > gunêtib. English > ESRgoodmorrow > kumaragoodnight > kunetiOnly in those cases where sonority cannot decide (such as doctor > datra), it is the firstconsonant that has to be deleted, in accordance with Wilson’s generalization. The sonoritybasedargumentation is, however, only convincing if we accept the idea that plosives areless sonorant than fricatives (cf. nasty > nasi), and that the /d/ in expressions such as goodmorrow and good night was still present in the input to the restructuring process.ESK and EJAM show partly an identical, and partly a different pattern. They both allowinternal nasal-obstruent(-sonorant) clusters under the appropriate syllabification, asevidenced in (20):(20) English ESKsomebody > somebodysomething > suntinghungry > (h)ungry(21) English EJAMmpempem (Twi) > pempeny ‘plenty’something > suntingmaster-thank you > massa-tenky