You may be right that people wouldn't honour the list for fear that they'd already inadvertently received the taint.
Perhaps it would only work if there was only a small percentage fee on taint - which was cranked up every now and then.
Even a zero-penalty taint blacklist might be useful - the community could more easily see where the stolen coins were coming from.
I an not fully understanding what you are saying I think. Here's the situation which concerns me:
Say that at a certain point in time, some threat requiring activation of an 'internet kill switch' pops up. Suddenly, for all intents and purposes, all networks are off-line. Fairly soon critical infrastructure would be wired back in, but that would be large corporations, and there would be significant packet level filtering. Consumer grade connectivity may be out for a long time. Months. And as it came back on-line it would also probably be carefully analyzed and regulated. I could even imagine a situation where encryption was only allowed between authorized endpoints (like people and their banks.)
Obviously such a scenario would spur interest in mesh networks and the like, but the technologies are extremely immature and untested, and anyone deploying them would likely be labeled a 'terrorist' which would complicate development.
Bitcoin seems to be developed under the assumption that peer-to-peer communications are available in a wide-spread and relatively reliable manner. I fear that Bitcoin could fail if that happens to not be the case, particularly if it were actively exploited by competent attackers during such a time of limited connectivity.
So, I would envision the 'block level blacklist' capability of Bitcoin being useful for a situation where there was a very obvious and very clear need to simply black out the entire economy for as long as needed in order to save it. Actually, in such dire (and probably far-fetched) circumstances, it might be fairly tenable and make the most sense to simple upgrade the client software in order to get going again.