Social networks

Labov-type surveys rely on collecting data from a random sample of individuals. Their speech is analysed for various key characteristics,

–114–

which are then correlated with their socio-economic background. The result, perhaps not surprisingly, suggests that human society is some­what like a layer-cake, with different socio-economic layers stacked up on top of one another. In one respect, this is a useful insight into the way societies function. But, as with many surveys, the result is oversimplified. In practice, people do not normally live in such clear-cut layers: someone from the so-called working class might well have middle class friends and neighbors.

In fact, human beings tend to cluster into social networks, groups of people who regularly interact with one another. A detailed study of the social networks within one particular speech community can pro­vide a useful corrective to Labov-type studies, which tend to suggest humans are rigidly stratified. Network studies can provide a more realistic picture of the way people interact in real life. Furthermore, if a sociolinguist manages to be introduced into a network, its members are more likely to chat in a natural way than in a Labov-type survey in which it is sometimes difficult to observe people speaking 'normally'.

The British linguists Jim and Leslie Milroy pioneered the linguistic study of social networks with a study of three communities in Belfast. Leslie was introduced into each group as 'a friend of a friend'. This ensured that she was accepted, and that people would talk relatively normally in front of her: when one youth tried to show off by talking in a somewhat affected way,, his friend punched him and shouted: 'Come on, you're not on television now, you know'.

Networks can be of high density, when the same people tend to work, play, and live together. On the other hand, they can be of low density, when people only have a small amount of contact with any one network, in that they may live in one area, work in another, and travel elsewhere for their social life (Figure 50).

When the Milroys examined their data, they discovered a number of things which would not have been detectable in a Labov-type sur­vey. For example, Labov's work suggested that men and women's

–115–

speech tended to differ, with women on the whole being closer to the prestige norm. The Milroys found this pattern also in Belfast overall, but with some interesting subtleties. When Leslie examined the three communities separately, and charted the occurrence of the way [aj was pronounced by people of different ages and both sexes, she found some modification of the overall pattern. In the oldest, most tightly-knit, and most traditional community, she found that the predicted pattern of male-female difference was most prominent. But in the other two, which were newer, and fairly loose-knit, this pattern was less evident, and was even reversed among the younger women of one community. This suggests that a blurring of sex differentiation in language may be linked with the break-up of close-knit networks. Findings such as this indicate that linguistic variation needs to be con­sidered from at least two angles: from the point of view of a broad Labov-type survey based on a random sample of people, but also from a close-up view of a number of social networks.