32 Bernard Laks, Basilio Calderone and Chiara Celata
the first doctoral thesis applying the framework and methods of SPE was defended.
Schane (1965) thus devoted a 45 page chapter to the analysis of French schwa,
consonantal liaison, h aspiré, liaisons within the verbal group, inversions and final
fixed consonants. Notwithstanding the complexity of the phenomena addressed,
which had hitherto produced thousands of pages of analysis, Schane’s very modest
chapter based treatment of those six issues in French phonology on a total of 73
examples. In all, the thesis put forward a total of 41 ordered rules. It was up to the
reader to assess such a ratio between rules and examples.^1
During the 20th century the surge in the power of linguistic corpora and
databases, coupled with the development of efficient sampling tools in the frame-
work of computational linguistics, has facilitated the availability of an ever-larger
and more varied range of linguistic data. Sociolinguistics (especially ‘Labovian’
variationism) had a great role in developing the paradigm of datum in linguistics
(Labov 2004). The problem of structured heterogeneity and inherent variation in
the grammar of any language system has been the focus of the scientific approach
since Weinreich et al. (1968); sociolinguistics presents itself as a corpus linguistics
that takes seriously the internal social organization of the linguistic community
being examined. However, with the contemporary quantitative approaches and the
development of corpus linguistics proper the use of the datum is no longer limited
to the description of social variation; on the contrary, it reaches to the very heart
of linguistic explanations in several domains such as acquisition, lexical storage,
community patterns of variation, heterogeneous linguistic competence (such as
pidginization, creolization, dialect formation or loss). The scientific study of lin-
guistic usage (Langacker 2000; Bybee 2006) and corpus linguistics converge in
these contemporary approaches, giving rise to quantitative and variable models
of language acquisition in psycholinguistics (e.g., Tomasello 2003, 2008), proba-
bilistic and stochastic models of language variation in formal descriptions (e.g.,
Boersma & Hamann 2009), and contemporary sociophonology, that very strongly
shows the traces of this new empiricism in linguistic studies. At the crossroads of
traditional sociolinguistics and experimental phonology, contemporary sociopho-
nology investigates the phenomena of speech use and comprehension from a cog-
nitivist perspective and explicitly borrows themes and concepts from experimental
psychology, particularly from exemplar theory and the idea that multi-sensorial
mnestic traces are stored in the mind of the speaker according to their distribution
(frequency) and recoverability (recency) (Goldinger 1998; Foulkes & Docherty
2006; Johnson 2006). In recent sociophonological studies, the emergence of a lin-
guistic category (be it phonological or socio-indexical) is predicted on the basis- Twenty-six examples for truncation and elision, 14 for fixed final consonants and numeral
adjectives, 3 for aspirated h, 2 for hiatus, 20 for junctures in verbal groups and for inversions, 8
for postpositioned pronouns. Twenty-two additional examples are quoted in the footnotes.