Discussion Details
| Title: | Optimality Theory and trends in children's speech |
| Submitter: | Aubrey Nunes |
| Description: | While OT has enjoyed great descriptive success in relation to 'end-state'
or competent phonology, it seems to me less capable of explaining trends in children's speech. Because these trends have not been taught, it seems to me that they have to fall out, either from the theory of acquisition, or from the properties of the Articulatory/Perceptual interface - in terms of ‘ease of articulation’ or analogical or some other sort of mishearing. Some of these trends are asymmetric. Two asymmetries, noted by Cruttenden (1978) are: Fronting with Coronal context-freely replacing Dorsal in stops as overwhelmingly the commonest such process; Dorsal harmony in 'doggy' as [gogi] overwhelmingly commoner than coronal harmony. There is also Smith's (1973) 'puddle puzzle', with Dorsal again replacing Coronal, but this time in a way which seems to me disharmonic, in 'puddle' as [pugal]. In Nunes (2002) from two experiments with a total of 120 normally-developing children up to the age of 8;6 I also found further asymmetries, such as: Metathesis in 'hospital' as [hostipal] the only articulator in any of the 120, very rarely labial harmony (in one child with a serious speech disorder), and coronal harmony, not at all, as far as I know; Disharmony in ‘monopoly’ as [manokali], as the only one step process I have ever observed involving the articulators in this word. Migration of the /s/ in 'spaghetti' to a position in the onset of the stressed syllable with the surface form almost always as [basketi], [psketi], or [sketi], with the mere reduction of the initial cluster as [sageti] or [pageti] seemingly very rare (I have never heard it); In 'soldier' the splitting of the affricate and the floating of the non-anterior property of its fricative edge into the onset of the stressed syllable to make the realisation a homonym of 'shoulder' rather than the reduction of the affricate, giving 'solder' (not in any of the 120, and otherwise rare); Typically coronal harmony in a set of cases including 'cardigan', 'hippopotamus' and 'calculator' as [ka:didan], [hitapotamus] and [kaltalayta] where the domain contains two oral stops. matching in all respects including ambisyllabicity, differing only in their articulator. In all of these cases, there is at least one possible alternative, seemingly at least as easy to say, which children either don’t say or say only rarely. These asymmetries can be seen in normally developing children's speech from 2;0 to 8;6, as well as in disordered phonologies. On independent grounds, it is surprising to find asymmetries in errors. Not speculating on whether these errors are at the point of lexicalisation or in the speech production system, the asymmetries as a set are not easily or plausibly attributable to A/P interface effects. This leaves only the theory of acquisition. In every case, there may a description of the process in terms of Descriptive Faithfulness or some other constraint. But this does not explain why the process occurs where it does. For OT to account for asymmetries which have no reflex in competent phonology, it would seem necessary to postulate constraints which are universally outranked, violating Pinker's 1984 Continuity criterion and the similar idea from Atkinson (1982). Atkinson, M. (1982) Explanations in the study of child language. Cambridge: Cambridge University Press. Cruttenden, A. (1978) Assimilation in child language and elsewhere. Journal of Child Language 5: 373 – 378. Nunes, A. (2002) The Price of a Perfect System: Learnability and the Distribution of Errors in the Speech of Children Learning English as a First Language. PhD Thesis, University of Durham. Pinker, S. (1984) Language learnability and language development. Harvard: University Press. Smith, N. (1973) The Acquisiiton of phonology. Cambridge: University Press. |
| Date Posted: | 26-Nov-2004 |
| Linguistic Field(s): | Phonology |
| LL Issue: | 15.3307 |
| Posted: | 26-Nov-2004 |
