184 Rosanna Sornicola and Silvia Calamai
which were extremely illuminating in this respect: despite being associated with
the traditional questionnaire technique, they tried to register all the complexity of
the speaker’s behavior through multiple answers. Spontaneous speech is actually
the place of multiple answers, very fertile ground for the analysis of phonetic varia-
tion patterns: usage data allow the inspection of the variants, followed by attempts
of generalization which take into account such factors as prosodic organization of
the utterance, organization of turns, sociolinguistic functions. A similar model,
mutatis mutandis, is employed by Temple in this volume: phonetic detail, however
minute it may be, can convey multiple meanings on different levels.
In addition to representing a precious resource for stylistic analysis, sound
archives also allow us to introduce the analysis of linguistic phenomena to a truly
historical perspective. The growing attention to the Intangible Cultural Heritage
which is spreading in different parts of the world may offer the scientific com-
munity a huge quantity of audio recordings by anthropologists, dialectologists,
ethnographers from the entire twentieth century (Ginouvès 2011). It thus becomes
possible to envisage a historical experimental sociophonetics, and therefore to
“repeat the past” by means of the analysis of old recordings, which represent
“invaluable data for the study of change in the community, and for the studies
of change or the absence of change in individual systems” (Labov 1994: 77). The
limits of these real-time comparisons are evident, but certainly counterbalanced
by the information on sound change which can be obtained through an intelligent
exploitation of these sound resources, which have only rarely been used so far.- Conclusion
By disposing of the large quantity of spontaneous speech gathered in sound
archives, we may rely on a solid empirical basis for the analysis and description of
phenomena that do not appear to be related to classical sociolinguistic variables.
This possibility is particularly important for those phenomena that pertain to the
dialectal substratum / dialectal competence of the speakers, that is, to varieties that
are hardly evocable by many techniques of speech elicitation, including the most
sophisticated ones (such as map-tasks), and certainly impossible to be captured
through traditional questionnaires. The so-called spontaneous diphthongization
attested in the Phlegrean varieties as well as in several Romance areas is one of
these problematic phenomena, being diphthongization captured only with diffi-
culty by traditional dialectological and sociolinguistic analyses. The many sound
archives collected throughout the Peninsula turn out to be a goldmine for socio-
phonetic variation, and for the most part they are still unexplored.