1) I was just playing a game where the Korath Clan declared war on me. Several turns later, the Altarians contacted me, and said something like "Your war against the Korath Clan has no merit. We declare war." Wtf, they attacked me first! Similarly, I've seen cases where the AI would say that "X is the victim of a war of aggression by Y, so we're helping X", but I suspect that actually it was the other way around because X had a much larger military compared to Y. Basically, I'm saying the AI should keep proper track of who starts wars, and view those who start wars (including other AI players, not just the human player!) with increased suspicion.
I believe this happens only happens when they have an "alliance" with the civilization that attacked you. The gameplay mechanic is fine, since their treaty requires them to go to war, but I completely agree that it should be worded differently. I was annoyed by that a few times as well.
2) A common occurrence in GalCiv2 is that when one civilization is becoming very dominant, another will contact you and say "Have you been keeping an eye on X? We are concerned about their growing dominance" etc. But they never actually proposed a specific action. If they are going to reach out to you, they should put something on the table, like offer you money or a tech to attack the strong civilization.
Yeah, that entire conversation always seemed pointless. I too would like to see them suggest going to war together now "before they become too powerful," or they could even suggest making a "Research Treaty" after learning that two powerful civilizations have recently made a treaty together. Perhaps offer some attack ships if you agree to "keep an eye on them."
3) Finally, an extremely common occurrence in GalCiv2 is get messages like "We have decided to go to war against X. A third party with a great deal of influence persuaded us to do so". Ok, fine. I take it that they were bribed to go to war. But the AI never tries to bribe the human player to go to war. In general, I feel it's very important that the AI players treat the human and other AIs the same way.
This is a great point. Nobody ever tries to get you to go to war against another AI. I always felt that was a big let down. The AI should be willing to offer you money and technologies to attack, just like they do with other AI. It doesn't make sense that you can make them go to war against each other, and they can make each other go to war, but they can't make an offer to you. The option "Never ask me again! They are my ally!" or something similar could be included so that it doesn't become annoying, but I would love to see this in GalCiv3.
Another one that you didn't mention is the "Your warmongering cannot continue!" or something to that effect. This never happened if you played as a "good" civilization. You could attack anybody in sight, and the other "good" civilizations wouldn't accuse you of war-mongering. It would always happen when you played as an evil civilization, but even if EVERYONE attacked you first, the other civilizations would form a coalition to destroy you, because you were "evil." I would like to see that addressed a little too. It would be nice to see who attacked who being taken into account, since even "evil" civilizations should have a right to defend themselves, without having their attacker calling other "good" civilizations to their aid when they attacked you.
Also: It would have been nice if there was a notification that civilizations were ready to start research and/or economic treaties. I would try and see if "it is too soon" every 10 turns or so, but it was WAY too easy to get into the game for bit (maybe 10-15 min), go to diplomacy, and find that 7/10 races already have research and economic treaties. You then have to pick one of the weak races, which even sucks more, because you had been building a fleet to destroy them. It would have been nice if there was a notification that civilizations were ready to start research and/or economic treaties.