Talking stones. Phonology in Latin inscriptions?

In spite of the long-standing debate on the value of epigraphic data, especially in the studies aiming at reconstructing the sociolinguistic framework of Latin, scholars still disagree on the value to be assigned to abnormal (i.e. non classical) spellings occurring in inscriptions. Are they clues suggesting pronunciations re- flecting the social class of the reader/speaker? Are they simple mistakes in writ ing? Are they a sign of the archaizing style typical of the epigraphic register? The paper focuses on the graphemic alternations ˂I˃ ~ ˂E˃ and ˂U˃ ~ ˂O˃ occurring within CLaSSES I, a corpus of inscriptions of the Archaic and Early periods of the Latin language. The distribution of vowel alternations in spelling is not casual, but rather suggests a plausible correspondence in phonological variation. The fine-grained comparison carried out on the lemmata occurring in CLaSSES I reveals a quite complex orthographic picture, where vowel alternations cannot be as- cribable to archaism pure and simple but rather may be interpreted as evidence for a sociophonetic process sensitive to both lexical and prosodic constraints.