[silvarilon]

I'll agree with you on that. Permanent death as a feature (i.e. my PC can get in a fight and permanently kill another PC) is very straightforward. And it has straightforward implications, some of which are good, some of which are bad.

Circumstantial permanent death is less clear precisely because it depends on the circumstances.

The purpose of those game mechanics would depend on what you are trying to achieve with your game.

What it adds is the ability to strongly differentiate your clans. Join clan X and you have power over sentences of executions. Join clan Y and you have power over resurrections. Join clan Z and you have better resources to buffy you in combat. Join clan A, B, or C and you get various social advantages.

The way you would differentiate your clans would depend on the game you're making. If you're making a game primarily about combat, then having one clan able to resurrect would be hugely unbalancing. If you're making a game where it's easy to avoid combat, then having one clan able to resurrect has a smaller game impact.

If implemented badly, yes.
If implemented well, every clan or character has the same potential - their choices would lead them towards different strengths.
In my implementation, the clans can only implement permadeath on other characters when certain conditions are met. I.E. the character that is going to die has taken some action to lead towards that event. So, that inherently makes it a rare event, and not relevant if we're talking games where "every character can kill off each other" - in games with more plentiful death, having one clan holding the keys to permadeath does seem very unbalanced.

The game is intended for politics and social roleplay, so the dynamics it creates are primarily political. Want someone removed, you've got to catch them out in certain ways. Drive them into doing something that would get them executed. Want to save a character from the gallows? You'll need to get the right characters on-side. Your clan might not have the right abilities, but your clan will have something else that you can barter with.

It shifts the emphasis. In a "pure" permadeath game, if you have an enemy you always have the option of walking up to them and killing them. That will often be the easiest option, and thus the one taken most often. It leaves little reason for players to come up with Machiavellian schemes to manipulate their enemies into a misstep. That is perfectly fine, since most games aren't trying to push for scheming politics. For games that are pushing for something different, the careful control of permadeath can provide an achievement. Being able to protect yourself from permadeath, or being able to permadeath an enemy is something that motivates players. So if you create the gameplay mechanics to achieve that based around the "core gameplay" of your game, you'll be encouraging more players to interact with your core gameplay.

Regardless of whether that core gameplay is the combat system, politics, fighting monsters, buying treasures, or whatever.

But yes, limiting permadeath does "weaken" the impact. I'd still argue that a good design in limited permadeath has the potential to strongly reduce the negative impacts while only mildly reducing the positive impacts. But, in the final tally, permadeath is always going to be an issue of personal preferences. And players will self-select the games based on their preferences. So there will always be a demand for both types of games.