I just finished William Labov’s Principles of Linguistic Change: Social Factors, which is a detective story. No, really. You don’t expect a linguistic tome to have the literary quality of suspense, but this book does. It’s organized around the central puzzler of historical linguistics: why does language change? Why do people bother with sound changes, especially when everyone agrees that they’re destructive if not positively evil? It takes the whole book to create a framework to answer the question.
This is mostly because Labov details his methods, his data, and what he does in the dark with statistics. He mostly works with surveys he and his students have done at the University of Pennsylvania, though he references similar work that’s been done all over the world: New York, Detroit, Montreal, Cairo, London, Belfast, Seoul, Buenos Aires, Mexico City, Hong Kong, and more.
One hypothesis we can immediately reject: that people imitate the leaders of society. Bluntly, people don’t come to talk like the king, or Congress. All of the ongoing sound changes that have been identified are divergences from the standard. Labov calls this change from below; other sociolinguists speak of covert prestige. One obvious example is AAVE, the speech of urban American blacks, which more or less completely ignores both standard General American and the local dialects of northern whites. Blacks and whites in the US don’t want to sound like each other. (This isn’t a universal– Jamaican Londoners talk like everyone else, which Labov confirmed by playing recordings of them to white folks; they couldn’t tell that the speakers were black.)
In order to address how sounds change, Labov focussed on whose speech is changing. The community doesn’t advance uniformly. His findings:
- The leaders of sound change are almost always women; they’re often a generation ahead of the men.
- Women keep advancing a sound change in a linear fashion; men’s advance is stepwise. The obvious interpretation is that men don’t pick up the change from their contemporaries, but from their mothers.
- There’s a typical curvilinear function of class: neither the lower class nor the upper class are in the forefront of change, but those in the middle– even more specifically, the upper working class.
- Nonstandard variants often peak in adolescence. So older speakers may retreat from a change.
- There’s only a very small contribution from ethnicity or neighborhood (except to the degree that these correlate with class).
- A phoneme doesn’t change all at once; some words are leaders, some laggards. For some reason, the tensing of short a in Philadelphia strongly affected the word planet, while Janet remained lax. (This is reminiscent of the effect of Trojan horse words in gender change.)
Beyond this, Labov was able to identify individuals who were in the forefront of sound changes in Philadelphia. Interestingly, they shared several characteristics. They were upper working class women, with a strong nonconformist streak. Perhaps most interestingly, they were what Malcolm Gladwell calls Connectors, people who were not only intensely involved with their neighborhoods, but had strong connections to other areas as well– the perfect people to spread ideas.
This tends to falsify notions that sound change is due to ignorance or laziness; the leaders are bright and upwardly mobile. Sound changes are also not due to isolation; they’re centered on the most social people. The paradox is that these women are just rebellious enough to fight social norms, but not enough to be dissipated or burned out.
So what happens, exactly? Labov outlines the steps like this:
- Some phoneme P has asymmetrical neighbors in phonetic space: there’s a farther gap between its near neighbor N and a farther neighbor M. Phonemes are realized with a certain amount of spread; as there’s more room in the direction of M, outliers in that direction are heard as valid instances of P.
- New languages learners thus move the phoneme in the direction of M– in effect, they mishear the outliers as normal tokens.
- The change is taken as characteristic of younger speakers and less formal speech. It’s preferentiallyl taken up by nonconforming young women.
- Upwardly mobile women spread the change to higher and lower social classes.
- Men catch up to women in the next generation, as they pick up the now advanced sound change from their mothers.
Now, all this is unconscious. These are not overt markers like a regional dialect– people are generally unaware of these changes, and if they’re pointed out the speakers are typically apologetic. If a change does reach public awareness, it’s stigmatized. It may continue to advance (it still has appeal as a marker of nonconformity), or it may just be retained as a long-term class marker. (E.g. there’s some evidence that the pronunciation of -ing as -in’ goes back for centuries.) If a local language variety is losing ground (generally to the standard language; this seems to be common in Europe), the leaders in this process also tend to be women.
A corollary is that people are lousy self-reporters. Labov played people recordings of words showing different stages in different sound changes; invariably people reported themselves as much closer to the standard than they were, and even claimed that “no one talked like” the more extreme variations. This should be a note of caution for linguists who rely on people’s evaluations of grammatical correctness!
Another curious fact: the closest analogue to sound change may be fashion, which is also driven by the preferences of middle class, highly social women.