Intelligibility Of Certain Kinds Of Non-native Cluster Simplification

By Author: Barbara Carter
Total Articles: 9
Comment this article

This is not to suggest that vowels must necessarily be realised in native-like ways; it is not inconceivable that future varieties of English might distinguish between word pairs such as pen and pan somewhat differently. But there will still be a contrast between them, barring major changes to the lexicon. In addition, there may be ways in which Hong Kong listeners are better able to distinguish sound contrasts made by Hong Kong speakers, so that theoretical merging may not always mean actual merging. This was suggested by the data in Hung (2000:126), where Hong Kong subjects were better able to differentiate words containing the [u, u:] contrast, when read by a Hong Kong speaker, than native-speaker listeners were. However, if international intelligibility is accepted as being important and it certainly is in Hong Kong, where English is principally a tool for international communication then accepting the merger of [ae] and [e] would seem to be problematic. The merger of and caused Five Fingers Shoes intelligibility problems in Jenkins's data. Although this is not discussed by ...
... Jenkins herself, the importance of this contrast may be related to the relatively large number of minimal pairs involving this contrast (471, according to Higgins, 2008). The importance of such 'functional load' calculations is controversial, but the merger of [ae] and [e] 302 pairs (Higgins, ibid.) might also be expected to be problematic. Brown (1991) assigns top priority to this contrast, partly because of his experience of miscommunication involving near-minimal pairs such as bag and peg in Singapore.

I myself have experienced communication breakdown as a result of the merger in Hong Kong, as when a local artist told me that he was going to send the painting. 'Send it where', I asked, before his puzzled expression and other contextual clues made me realise he was going to sand the painting. Some ELF researchers might claim this native-speaker observation to be irrelevant, but if it causes problems for me it is likely to cause problems for others, assuming that auditory discrimination ability is by and large species-specific. It is not therefore entirely clear why Jenkins (2007:27) appears to propose 'the removal from the syllabus of nineteen of the twenty RP vowel and diphthong qualities'; it is the contrasts between the qualities that are important, and most of these will still have to be worked on.

Turning to the features of Hong Kong English that do not pass the test of international intelligibility, there is the conflation of [n] and [1] ('rare' in Deterding et al., but present in 37% of tokens in Hung) and the tendency to subsitute [w] for [v] (also more prevalent in Hung's data). The simplification of initial clusters is proscribed by Jenkins's core, so that pronouncing crowded as [klaudid] would be expected to lead to intelligibility problems. The speakers in Deterding et al. deleted final consonants in 54% of clusters occurring before a pause or before a word beginning with a vowel, environments where deletion would not normally be expected in NS varieties of English (Schreier, 2005:30, in Deterding et al., 2008:15 7). This is therefore a Five Fingers KSO feature that might lead to intelligibility problems according to Jenkins's criteria, in which final clusters can only be simplified according to LI rules of elision.

But on further examination, some of the simplifications noted by Deterding et al. seem unlikely to be problematic, such as the common Hong Kong pronunciation of think without the final [k] consonant. The relatively lower importance of final clusters is entirely predictable, from a psycholinguistic standpoint: 'The deletion of cluster-final segments at the end of lexical items results in a lower degree of information loss than reduction in the beginning of words' (Schreier, 2003:12). There is a need for more research into the consequences for intelligibility of certain kinds of non-native cluster simplification. For example, in Hong Kong one can often hear multisyllabic words such as government or consultant produced without the final consonant; while native speakers would not normally do this it seems unlikely to affect intelligibility. Other features of Hong Kong English may be found to actually contribute to intelligibility, rather than detract from it, especially in the suprasegmental realm. The tendency towards syllable-timed rhythm noted by Deterding et al. has been argued by Kirkpatrick (2007b) to enhance intelligibility, although this remains controversial and relatively unsupported by empirical evidence.