The sound patterns of Englishes: representing phonetic similarity

Abstract

Linguists are able to describe, transcribe, and classify the differences and similarities between accents formally and precisely, but there has until very recently been no reliable and objective way of measuring degrees of difference. It is one thing to say how varieties are similar, but quite another to assess how similar they are. On the other hand, there has recently been a strong focus in historical linguistics on the development of quantitative methods for comparing and classifying languages; but these have tended to be applied to problems of language family membership, at rather high levels in the family tree, not down at the level of individual accents. In this article, we outline our attempts to address the question of relative similarity of accents using quantitative methods. We illustrate our method for measuring phonetic similarity in a sample of cognate words for a number of (mainly British) varieties of English, and show how these results can be displayed using newer and more innovative network diagrams, rather than trees. We consider some applications of these methods in tracking ongoing changes in English and beyond, and discuss future prospects.1 How are accents different; and how different are accents?In November 2004, the BBC commissioned an online poll on attitudes to accents, as part of the 'Voices' project (which more generally explored 'how we speak in the UK now'). The fact that over 5,000 responses were received shows how interested speakers of English are in each others' accents; and the results and comments make fascinating reading (http://www.bbc.co.uk/voices/yourvoice/poll_results.shtml, accessed 13 September 2006). Part of the fascination, however, lies in the opacity of many of the comments, which are difficult to interpret in phonetic terms, as shown in (1).(1) (a) I don't really like the Birmingham accent that much (even though I've got one), but I do like the Black Country Accent . . . It sounds singy and old-fashioned. (S. Murphy, Birmingham) (b) Yorksher accent and more specifically 'ull accent rules!. (red badger, 'ull) (c) I never notice my accent until someone points it out and the way we shorten words and the rs we sound here in Bristol. (Debbie Smith, Bristol) Many of these comments are straightforwardly attitudinal, and are expressed in terms of liking one accent and not liking another; and typically respondents are positive about their own varieties (though with distressing frequency this is not the case for the denizens of Birmingham). More interesting are the cases where respondents attempt to ground their comments with reference to particular phonetic or phonological characteristics of varieties, as shown in (1) by 'red badger', who encodes [h]-dropping in his/her response, and in Debbie Smith's comment on Bristol 'r'. These mentions, however, are hardly very specific, and are consequently hard to interpret. What is the issue with 'r' in Bristol here? For instance, is it the phonetic quality that counts, or its distribution? And does it matter whether someone is co...

Issues in the genetic classification of contact languages

“…Where such similarities are straightforwardly divergent, then a network program will display an output indistinguishable from a tree; where there are interrelationships among units, suggesting parallel innovation or contact, then a more weblike structure will emerge. For most biological populations, and as it turns out for most cases explored so far involving languages and dialects (Forster and Renfrew 2007;McMahon et al 2007), the truth lies somewhere in between and there will be both tree-like and non-tree-like signals. Networks, in other words, are flexible; they do not require us to abandon the insights of the family tree, and they do not rule out the possibility that tree-like divergence will be the sole or the dominant factor in the histories of some language groups; but they do not limit us to the tree model either.…”

The past, present, and future of English dialects: Quantifying convergence, divergence, and dynamic equilibrium

“…In Heggarty, McMahon, and McMahon (2005), McMahon and McMahon (2005), and McMahon, Heggarty, McMahon, and Maguire (2007), a new method for comparing the phonetic similarity of different varieties of the same language and of different languages was introduced. This method shares with other dialectometric methods a concern with quantifying the similarities and differences between varieties in an objective way, but it differs from methods such as that described in Nerbonne, Heeringa, and Kleiweg (1999) and Nerbonne and Heeringa (2001) in that it prioritizes linguistic accountability over computational simplicity.…”

“…4. Representation of the resultant distance matrix using trees and networks, in particular the phylogenetic software program NeighborNet (see Huson & Bryant, 2006; for applications to linguistic data, see Heggarty et al, 2005;McMahon & McMahon, 2005;McMahon et al, 2007; and Investigating multidimensional relationships below).…”

“…2. As it is our intention to develop a method that is applicable beyond varieties of English-so that, for example, (varieties of) English can be compared to (varieties of) German-these cognates must be shared by English and its Germanic relatives (see McMahon &McMahon, 2005, andMcMahon et al, 2007, for further details).…”

Measuring Foreign Accent Strength in English

“…Our work differs in using only unedited speech, in examining accents from a range of languages, and in not restricting attention to a small number of features. McMahon et al (2007) explicitly aim to measure the degree of accentedness in various forms of English world-wide, but they use an algorithm that is not fully specified and therefore not easily applicable to other datasets:…”