An initial assay of Wu and Min lexical tone sandhi opens inquiry into a possible source : Austronesian pitch accent, and into a sandhi origin for the "third" tone ( qu/departing ). One important element in such an assay is the feature Right/Left : focus = preservation, stress, retention, etc, towards the Right or towards the Left. Northern Wu appears to be focus Left, and destressing seems to be spreading in the area. Southern Wu and Min are focus Right. Southern Wu focus does not prevent some Right mergers, and often it is the last two syllables acting together that is the focus. Northern Min shows similarities to Southern Wu, but Southern Min can be said to have no tone sandhi at all : The Amoy et al tone circles appear to be artifacts of changes in isolation values, since they are virtual reconstructions of the probable prototone values. The one Hakka dialect examined appears to be like Northern Min/Southern Wu. On the basis of this assay, I would hazard the guess that in the study of the origin of lexical tone sandhi, Southern Min should be classified with the Cantonese/Thai type, Northern Wu as a separate type heavily influenced by Mandarin, and Southern Wu/Northern Min as the preservation of the oldest, most Austronesianoid type of sandhi. Further speculation would be foolhardly until more information is available and more detailed comparisons and histories are drawn up.
Homophony and Contrast Neutralization in Southern Min Tone Sandhi Circle