Preface: I completely agree with the idea of having a server where players can be killed off, but I absolutely feel that IF there will only be one server, then the server can't have province deaths and still be a good business move for Reverie.
I make no claims of understanding your playstyle, but I believe that you aren't a griefer.
I do know that nice, reasonable people can drift into griefer mindsets when they've had a bad day, so when a game permits that kind of conduct, those kinds of people can tear apart a couple weaker provinces - I also know that griefers flock to games that allow them to grief others. If a small alliance decides they want to "go out in a blaze", they can tear apart dozens of provinces (this happened in Aegis a couple times while I was playing). I worry that we have a small community right now and that it may be harder to grow that community with province-death and reset as a common issue. I could be wrong, though.
If Reverie has the resources to support multiple servers, the obvious solution is to have province death on a small portion of those servers. I'd be interested in seeing how the populations vary. If it holds true to current WoW populations, it'll be about 30-35% in the more hardcore style, 65-70% on the no-player-death servers. I still think in a single-server approach, it's better to go with the no-player-death option, but if they opt for 2 servers, I agree that one should be player-death positive.
I think you're right about the greater alliances and communications aspect for the player-death positive approach, but again in my experience that can lead very quickly to a bully faction and victim province landscape.
Consider this anecdote - I had a friend who was pre-med in college at a school where the bottom 10% were guaranteed an F. After the first test, the people in the bottom 10% dropped the class. That meant the next group would get Fs...so they would drop the class too. After the second test, the class rank was more solid, so the next couple groups would drop. By the end of the semester, classes only had around 30-40% of their initial enrollment.
In a college setting, this be economic - they pay for the class again in the spring. For a game, it isn't - those who drop just move on to something more rewarding for them. It might guarantee that the people who are left love the game, but it won't help word-of-mouth advertisement, and it will probably hurt business significantly.