July 28, 2004

The beauty of Brummie

Following up on recent posts about accent evaluation, here's a clear explanation by Steve Thorne of an experiment comparing how different English accents are perceived by native and non-native speakers.

In May 2002, I recorded short samples of 20 different accents of English... In order to limit the influence of extraneous variables, the speakers chosen were all male, white, aged between 35 and 40, and upper-working to lower-middle class. These recordings were played to 96 native and 109 non-native English speakers who were then asked to briefly describe each accent and rate each one on a scale of 1-10 (1 = very unpleasant, 5 = neutral, 10 = very pleasant).

According to Thorne:

... the native speakers reacted predictably. The French, Southern Irish, Edinburgh Scottish and Geordie (Newcastle-upon-Tyne) accents received the most favourable responses (none, incidentally, described the very nasal French accent as 'nasal'), the American and rural accents such as Cornish and Norfolk also did well, but Welsh, RP (Received Pronunciation), Northern Irish and accents associated with large urban conurbations such as London (Cockney) and Liverpool (Scouse) fared badly. No prizes for guessing which accent came bottom. Black Country.

Black Country is "an accent associated with the South Staffordshire area of the English Midlands". I'll freely confess that I had never heard of it, and so would not have guessed that it would come in last (or "come bottom", as Thorne puts it in that charmingly quaint UK-ish patois of his :-).

In the cited page, Thorne doesn't explain in detail who the "native speakers" were and where they came from -- that surely would make a big difference, since Americans' ability to distinguish among UK accents is usually rather poor, and their evaluative associations with UK accents are quite limited. I expect that the same is true in reverse for the British speakers with respect to American accents, though perhaps American television and movies have conveyed some clues about the stereotypes involved.

Anyhow, one of Thorne's original goals is to defend the speech of his native Birmingham, which seems to be something like the New Jersey or Brooklyn of England:

Ask a British person what their least favourite accent is, and they will more than likely say 'Brummie' - the variety of English spoken in the West Midlands city of Birmingham. Ask them why, and they will more than likely use adjectives such as 'nasal', 'monotonous', 'miserable' and/or 'ugly' to justify their responses. Such views are based on the belief that all other accents are higher in aesthetic value than Brummie, and even those who are prepared to accept that Brummie is not 'wrong' (and many aren't) seem fundamentally opposed to the idea that other accents are not more aesthetically pleasing. But is Brummie really ugly?

In support of his claim of anti-Brummie prejudice, Thorne quotes from a shockingly smug and nasty BBC page on "How to speak Brummie", which he accuses of speading untruths:

[A] common misconception about Birmingham intonation... is that it 'falls' at the end of sentences, and this leads to criticisms that it is 'dull', 'miserable', 'depressing' and/or 'downbeat': "In Brummie, the lowering suggests despondency and makes it less attractive to the listener . . . the lack of aural variation quickly begins to grate".

The results of Thorne's experiment support his intuition:

The responses of non-native speakers, on the other hand, were inconsistent - ranging from 'harsh' (for Brummie), through 'nice', to 'melodic', 'lilting' and 'musical', and from 'clear' (for Southern Irish), through 'boring', to 'disgusting'. Although there was no significant difference between the overall scores for each accent, many appeared to prefer the characteristically Brummie 'rising' and 'high tone at the end of sentences', criticising instead the 'cold and unemotional' character of Edinburgh Scottish - one respondent even going so far as to describe the Scottish speaker as 'untrustworthy'. Scouse was also praised on many occasions for its intonational distinctiveness - its clarity, 'pleasant tonality', and dynamic 'rolling of the r', but reactions on the whole were generally mixed, and there was little evidence to suggest that foreign speakers were dipping into the same adjective cluster as their British counterparts - no high occurrence, for example, of the words 'nasal', 'common', 'whingey', or 'wrong' to describe the Birmingham accent.

Thorne's conclusion:

These findings demonstrate that non-native speakers work to a totally different set of criteria when evaluating English accents, and do not discriminate on the same grounds as native English speakers. Judgements of the perceived beauty or ugliness of accents are based almost entirely upon a knowledge of the social connotations which they possess for those familiar with them.

Returning to that BBC Brummie page again, I observe that its author confidently asserts that

The Birmingham accent hits one note - usually a low one - and sticks to it no matter what. It is this lack of aural variation that is the principle cause of irritation for others. It is also the source of the stereotype of the unimaginative Brummie. The accent stays the same and never varies, and so subconsciously people assume the same must be true of the speaker.

although I suspect that this assertion has no factual foundation whatsoever.

It seems much more likely that the individual who wrote this BBC page has a standard old-fashioned snobbish distaste for Brummies, finds their accent to be irritating and unpleasant for the usual social-psychological reasons, and has decided to rationalize these competely irrational prejudices by offering a pseudo-scientific explanation in terms of hypothesized properties of Brummie intonation.

As Thorne points out, what this BBC page says about Brummie intonation is internally contradictory -- it is described as having "a downward intonation at the end of most sentences" which "suggests despondency and makes it less attractive to the listener", but it is also described as "[hitting] one note - usually a low one - and [sticking] to it no matter what". Thorne also points out that the different aspects of the description are in any case factually incorrect, and especially that "[a]n extensive use of rising rather than falling tones ... is typical in Birmingham speech". His non-native listeners bear him out, describing Brummie as "'melodic', 'lilting' and 'musical'".

One could also study this question using the tools of instrumental phonetics rather than native vs. non-native perception. Given an appropriate sample of speech, one could evaluate quantitative properties of pitch contours, such as percentage of final rises, falls and levels, or distributions of rates of change over final syllables or final post-stress regions. One could look at distributions of pitch values and their first couple of derivatives at different time scales, or the distribution of pitch shapes over syllables and words.

As far as I know, no one has ever done this, in a systematic way, in comparing different accents of English, though Pierre Delattre did something similar across languages, 50 years ago, in comparing the pitch contours of lectures by Margaret Mead and Simone de Beauvoir.


Posted by Mark Liberman at July 28, 2004 10:44 AM