Related to the summary of Adams, 2013, the present study first of all solves the problem how the absence of a linguistic change, in this case of the merger of /b/ and /w/, from inscriptional corpora can be evidenced. For this purpose, two separate yet interconnected methods, both suggested by J. Herman, will be used and presented. By involving both methods in the analysis of data sets recorded from a selection of Roman provinces already examined in the Computerized Historical Linguistic Database of Latin Inscriptions of the Imperial Age (
http://lldb.elte.hu/), not only will we demonstrate that inscriptional data are adequate and reliable for the examination of the B/V confusion, but we will also offer a picture of the frequency of the merger of /b/ and /w/ in Latin over the centuries and in various regions which is more realistic than any explanation before. During this process, a correlation between this merger and the loss of the intervocalic /w/ evidencing the change from [w] to [β] will also be detected and presented for discussion.
