Lenition as a Celtic substrat

I managed to get my hands on Maiden and Parry's Dialects of Italy and am enjoying some of the book's contents. I got recently to the part where consonantal weakening was theorized to be due to a Celtic substratum. "This [lenition] has been attributed to the influence of the Celtic substratum since at least the time of Schuchardt (1866: 88), who pointed out modern insular Celtic languages display historical weakening similar to that of French, Spanish and Portuguese." (p. 39) Iberia, Gaul, and northern Italy had intervocalic voicing...unsurprisingly all three areas had a significant Celtic populations prior to their conquest by Rome. Outside of these areas, Romania, southern Italy, and southern Spain (Mozarabic) had retained voiceless consonants intervocalically.

As Pannonia since Roman times had a significant Celtic population (the region was "thoroughly Celticized" according to Pannonia and Upper Moesia, a History of the Middle Danubian Provinces), I've thought that it would likely have lenited its intervocalic voiceless consonants. Though when I was speaking to Dewrad he insinuated that many of these changes arose in Gaul and may not have reached as far east as Pannonia. I still think there is a possibility of intervocalic voicing during the timespan when Pannonia was apart of the Carolingian empire's tributaries, a reconnection with Latinum may have brought "Gaulish" features into the language. Alternatively, lenition may have arisen independently in this language. Lenition is not unknown to this area — Slovak with /g/ > /ɦ/, Hungarian with /t/ > /z/. A compromise I've thought up might include lenition of intervocalic /p/ > /v/ but /t/ and /k/ is kept as is.

So I am curious to what the peanut gallery thinks of the idea of lenition being a Celtic substrate as a basis to include it in a romlang.