Introduction: Sociophonetic perspectives on language variation 5
In this approach, the authors appear to owe much to the view according to
which the balance between storage and computation in language processing can-
not be defined once and for all. The cost of storage is not necessarily more than
the cost of processing (e.g., Baayen et al. 2002); in investigating corpora of actual
uses, storage may be found to cover much of the labour necessary for processing
specific phenomena, provided that a fine-grained analysis of frequency distribu-
tions is realized.
The proposed generalizations are found to hold true also for subgroups of data
as defined by different types of liaison consonants (/n/, /z/, /t/) and the speakers’
age and educational level. Age was chosen because it is known to be a relevant
sociolinguistic factor in liaison variation (Durand et al. 2011); on the contrary,
educational level was chosen because it has always been disregarded in previous
analyses of French liaison. The results show that variation in liaison production
as a function of educational level is present in the “tail” of the distribution, that
is, in that variegated sample of low- and very-low-frequency items which clearly
turns out to be “the most likely repository of lexical environments differentially
selected by different groups of speakers” (p. 47). In the authors’ opinions, such
a result confirms the importance of adopting a corpus perspective for the study
of sociolinguistic variation and suggests that some unexpected forms of socially
structured variation may emerge if the analysis focuses on the basin of those rare
productions that only very large databases may include.
- Patterns of sociophonetic variation
The paper “Derhoticisation in Scottish English: A sociophonetic journey” by Jane
Stuart-Smith, Eleanor Lawson, and James M. Scobbie presents sociophonetic data
from the cities of the Central Belt of Scotland, Edinburgh and Glasgow, whose
varieties show evidence of derhoticisation. According to Wells’ (1982) taxon-
omy, Scottish English is usually thought to be a classic rhotic variety of English.
Nevertheless, historical sources and several sociolinguistic inquiries have estab-
lished that a derhoticisation process is present in selected varieties of Scottish
English, though with variations related to the speakers’ gender, social class, and
speech style. The richness of points of view adopted in this study, ranging from
auditory to acoustic and articulatory analyses, as well as from the discussion of
different transcription methods to an investigation of the speakers’ perceptual
responses, forces the reader to reflect on what should be the best way of represent-
ing the complexity of sociophonetic data in our explanation of the speech process-
ing mechanisms adopted by the speakers/hearers in normal linguistic interaction.