Chapter 6. Sound archives and linguistic variation 173
to the individual speaker: not only different speakers produce different values of
the same phenomenon, but there exists a range of different realizations for one
linguistic variable within the speech of the individual speaker. Therefore, in order
to study the range of variation of a phenomenon, it is necessary to start from the
inspection of the concrete realizations of a linguistic variable. According to this
view, individualism is not equivalent to atomism or fragmentation, inasmuch as
there are clear limitations to the range of variation.
The groundbreaking paper by Weinreich et al. (1968) represents a fracture
with respect to the contemporary framework of linguistic studies; at the same
time, it develops some central tenets of European linguistics at the beginning of
the twentieth century. In fact, the authors devote much attention to the European
traditions concerning the study of variation, and to Mathesius himself. More spe-
cifically, the programme of contemporary European linguists in the framework of
individualism is revised according to typically American beliefs, by introducing an
emphasis on the regularities inherent in variation and on the community pattern,
rather than on the variability of the individual.
In European individualism, the relationship between variation and the limits
of variation, as well as between heterogeneity and invariant structure, was dis-
cussed by making an appeal to the behavior of the individual speaker; the rela-
tionship between the individual and the community was problematic in itself.
In particular, Mathesius was very clear about the dialectics between oscillation
and limits of the oscillation. On the other hand, Weinreich, Labov and Herzog
criticize Mathesius precisely for not having sought structured heterogeneity: the
individual’s behaviors were seen as referring to a class of speakers, also because
of the influence of macro-sociological models. The specificity of the individual as
historically determined source is reduced or nullified also because of the massive
use of statistics, which brought to light regularities. In this respect, the dominant
trend of Anglo-Saxon sociolinguistics (the so-called “first wave of social studies
of variation”, in Penelope Eckert’s words; Eckert 2012) achieved interesting results
on the analysis of diastratic variables, but these results did not appear to be revo-
lutionary in the domain of stylistic variation.
Models of multidimensional variation such as those obtained, for example,
for (r) in the North-American context have rarely been applied in the European
context (apart from the English area): this result cannot be caused only by the dif-
ferent research areas of European scholars. The call for regularity and structured
heterogeneity is consistent with the correlational research method, but above all it
is related to the particular socio-cultural conditions of North-America. Therefore,
a philological, historical study of spoken texts (where by ‘historical’ we mean
related to the conditions of the text and of the speaker producing it) is desirable,