Hi,
http://www.pkucn.com/viewthread.php?tid ... a=page%3D1
I think I may have finally found the right explanation for Hokkien's stereotony. I am so excited.
In short, pre-sinicized Hokkien almost certainly was a tone-less language, and used topic markers much like Japanese or Korean. If you know a bit of Japanese, you know about the topic markers わ(wa)、が(ga), etc. Hokkien must have had similar topic markers or final particles. When Hokkien was importing large amount of vocabulary from Chinese, which was a tonal language, it still retained these toneless particles. At a later stage, these final particles coalesced with the final syllable of tonal phrases, giving rise to the standing tones.
If the story stops there, it would all be just pure conjecture. But the most striking part was, after I put forth this conjecture, I thought, hmm..., maybe I should check the case of Hmong, too. Luckily someone in Thailand already did an exhaustive review of Hmong particles:
http://mulinet9.li.mahidol.ac.th/e-thesis/4237996.pdf
and voila! I hit a gold mine. Topic markers are used ubiquitously in Hmong! If you read section 7.3, you will find out that (1) particles are used everywhere, almost in every sentence, (2) many of them carry several tonal values, which indicates to me their historically toneless background.
So, to me, Hokkien's steoretony is not a mystery anymore. It has a solid theoretical explanation.
This also opens a gate of possibilities into more understanding of tonal features of Wu and Mindong languages. All this is good news. Now we have some support to the claim that in a wide area (today's Jiangsu, Zhejiang and Fujian) of ancient Southeast China, the native languages there were toneless and rich in suffix particles. I never even knew that ancient Hmong/Thai shared some structural similarities with Korean/Japanese.