Monday 28 February 2011

On talking among(st) yourselves

A correspondent writes from the USA to say he's noticed British English words and phrases increasingly entering public and written discourse. He gives as examples He was sacked instead of He was fired, gone missing, and such slang words as snarky. In particular, he says he's beginning to hear amongst rather than among, and wonders whether there's a difference between British and American English in this respect.

There certainly is. A table in the Quirk Grammar (9.21) shows amongst occurring ten times more frequently in British English, and this is confirmed by later corpus studies. In the huge COCA corpus (Corpus of Contemporary American English) we find 2405 instances of amongst compared with 144,461 instances of among - 1.7 per cent. This compares with 4449 instances in the British National Corpus compared with 22,385 of among - 20 per cent. On the other hand, those 2405 US instances spread pretty evenly over the past decade, so there's no evidence of any kind of very recent dramatic increase. I'd be interested to hear what US readers of this post think.

In the UK, my impression is that all the -st words are reducing in frequency. They began as a development of an ending attached to the base form: among + an -es genitive. We see that ending still in besides. Then in the 16th century, people evidently felt this was related to the -est superlative form, as gradually we find the -st ending used. We see it also in against, where it's the standard form, and in amidst (vs amid) and whilst (vs while), where usage varies. Fowler thought that these differences might be explained with reference to pronunciation - the -st forms would be used when the following word began with a vowel - but this isn't supported by the large corpus collections. The variation in standard English seems to be primarily stylistically conditioned: some people like the sound of whilst; others hate it. There's also a chronological factor: the -st forms tend to be found in older texts and among older people. And there's a great deal of regional dialect variation too.

The OED makes an interesting point about amongst, suggesting a semantic nuance not found in among: 'generally implying dispersion, intermixture, or shifting position'. So, I walked amongst the crowd would suggest a rather more active moving about than I walked among the crowd. If this is so, then I'd expect to see an increase in the proportion of amongst usages in contexts where these notions are dominant; and a quick dip into Google suggests that this is the case. With among(st) a group the proportion of amongst usage rises to 12.6 per cent; with among(st) the waves it is 45 per cent. With talk among(st) yourselves, the usage actually reverses, with amongst being four times as frequent. If there is a trend in US English to use amongst, semantics may well be an influential factor.

Sunday 13 February 2011

On pronouncing Purcell

A correspondent writes to ask about the pronunciation of two 17th-century names: Henry Purcell and Andrew Marvell. He says: 'In preparatory school I was taught to place the stress on the first syllable of Purcell and the second syllable of Marvell, always assuming that to be correct. However, I frequently hear the former pronounced with the stress on the second syllable and the latter with the stress placed on the first. Was my English instructor correct? And do American and British usage conform or differ? Has the stress shifted historically?'

When establishing an earlier pronunciation, as seen in earlier posts in this blog on the Shakespearean sound system, there are several kinds of evidence to look for - rhymes, puns, metre, spelling, and explicit comments by contemporaries. In the case of Purcell, we find clear evidence of the stress falling on the first syllable from contemporary spellings. Before spelling standardized, the vowel in an unstressed syllable would be spelled in different ways. So when we find such spellings as Pursal, Purcel, Persill, and Pursall in the 17th century, an initial syllable stress is clearly suggested. It is reinforced by the ode John Dryden wrote on the death of his friend, in which the metre requires the stress to be on the first syllable:

Now live secure and linger out your days,
The Gods are pleas'd alone with Purcell's Lays,

The same stress pattern is found in a rhyme in a poem by Gerard Manley Hopkins, who rhymes Purcell and reversal. That was 1918. So there doesn't seem to have been any historical change.

Nor is this just a British pronunciation, as American dictionaries say the same thing. W Cabell Greet's World Words, compiled in association with the Columbia Broadcasting System in 1948, gives initial-syllable stress for both Purcell and Marvell, but adds, after Purcell, 'As an American family name the last syllable is often accented'. The Random House Dictionary confirms this. It lists three Purcells: Edward Mills (the US physicist), Henry (the composer), and a town in Oklahoma. The first and third, it says, have the stress on the second syllable; for Henry, the stress is on the first. And the very next entry is for the Purcell Mountains in British Columbia and Montana, with, once again, the stress on the second syllable. American intuitions are thus split down the middle, with Henry apparently in the minority, so it's hardly surprising that people assume he is like everyone else.

However, intuitions in Britain are split too. It's never possible to anticipate the crazy ways in which the English like to pronounce their surnames or placenames, as famous cases such as Cholmondley ('chumley') and Happisburgh ('haysbruh') illustrate. So, when we encounter a surname ending in -ell, there's no way of predicting the stress pattern. There are several examples of surnames ending in -ell which have the stress on the second syllable, such as the Irish politician Charles Stewart Parnell. This end-stress is a typical feature of polysyllabic words in Irish English. But even here there are problems, because in Parnell Square the stress usually reverts to the first syllable (a similar alternation to what we find with he's sixteen and sixteen people). And Parnell himself preferred to say his name with the stress on the first syllable.

So we get the result we see in, for example, the Cambridge English Pronouncing Dictionary, where we find both pronunciations given, in both British and American English. And because linguistic uncertainty is always contagious, it's not surprising to find other surnames vacillating. No dictionary I've looked at gives any other pronunciation for Marvell than one with the stress on the first syllable, and similarly for the members of the famous Durrell family, but we do hear the alternatives, especially from American speakers, from time to time.