Ipsa Latinitas et Regionibus Cotidie Mutetur et Tempore : Some Methodological Considerations on the Use of Herman’s Quantitative Method

This article discusses whether statistical analysis (as opposed to quantitative analysis) can be successfully applied to the study of diatopic variation in Latin (when using inscriptions). Section 1 introduces the methodological problem. First, previous research on the topic is reviewed, with particular attention to the quantitative method proposed by Jozsef Herman in the year 2000. Second, the limitations of this method (and of inscriptions in general) are addressed in the light of the criticism made by James Noel Adams in 2007. In section 2, the findings of Herman (2000) on the existence of regional differences in the phonological system of the Italian peninsula are validated by means of various statistical tests. The results in this section suggest that the quantitative method proposed by Herman is reliable. However, they also suggest that this same method should be supplemented by the application of objective statistics. More precisely, one should verify that the observed figures differ significantly from the figures expected based on a random distribution. In section 3, the discussion turns to the use of "negative" evidence, i.e., the count of correct forms alongside the corresponding deviations from the "Classical Latin" norm. This kind of evidence is distrusted by Herman, but the statistical study in this paper indicates that even such evidence can be useful, at least in limited circumstances. Finally, section 4 draws some preliminary conclusions and clarifies the importance of the epigraphic sources for the study of diatopic variation in Latin.