| Summary: | <p> This dissertation identifies three previously unexplained typological asymmetries between creoles, other types of language contact, and `normal' sound change. (1) The merger gap deals with phoneme loss. French /y/ merges with /i/ in all creoles worldwide, whereas merger with /u/ is also well-attested in other forms of language contact. The rarity of /u/ reflexes in French creoles is unexplained, especially because they are well attested in French varieties spoken in West Africa. (2) The assimilation gap focuses on stress-conditioned vowel assimilation. In creoles the quality of the stressed vowel often spreads to unstressed vowels, e.g. English <i>potato</i> > Krio /&rgr;ϵ&rgr;&tgr;ϵ&tgr;ϵ/. Strikingly, we do not find the opposite in creoles, but it is well attested among non-creoles, e.g. German umlaut and Romance metaphony. (3) The epenthesis gap is about repairs of word-final consonants.These are often preserved in language contact by means of vowel insertion (epenthesis), e.g. English <i> big</i> > Sranan <i>bigi</i>, but in normal language transmission this sound change is said not to occur in word-final position.</p><p> These case studies make it possible to test various theories of sound change on new data, by relating language contact outcomes to the phonetics of non-native perception and L2 speech production. I also explore the implications of social interactions and historical developments unique to creolisation, with comparisons to other language contact situations.</p><p> Based on the typological gaps identified here, I propose that sociohistorical context, e.g. age of learner or nature of input, is critical in determining linguistic outcomes. Like phonetic variation, it can be biased in ways which produce asymmetries in sound change. Specifically, in language contact dominated by adult second language acquisition, we find transmission biases towards phonological rather than perceptual matching, overcompensation for perceptual weakness, and overgeneralisation of phrase-final prominence.</p>
|
|---|
