Since this suggestion is less popular than I suspected, I thought I should elaborate on my reasoning behind my objection to the current rules. Sorry for the delay in a follow up post; I had a busy few days and could only budget time to read the boards and not post...
As Gdj's post indicated, declaring war on a nation while receiving military access is a treacherous act that violates the norms of warfare at the time. Fair enough, it is (though one that I think that players only make on accident rather than as intentional treachery). Still, it is treachery... but more severely punished than other violations of the norms of warfare in the game:
- Declaring war despite a Royal Marriage (which was historically equivalent to an alliance in the game's early time frame), only costs 1 stability. The current mechanics treat a DOW on a country granting Military Access 5 times as egregious of a violation of diplomatic norms. At the same time, by every other metric in the game, Royal Marriages are more serious: signing one improves mutual relations by +25, they can't be canceled on a whim, and relations suffer if you cancel one. By contrast, countries don't care if you cancel a military access agreement through them and signing one doesn't increase their opinion of you.
- Declaring war on a vassal costs 3 stability. The current mechanics treat a DOW of war on a country granting Military Access more serious than a DOW on a nation with an agreement of formal protection that the player's country has extended to a junior partner. That seems odd.
- Declaring war without just cause (a Causus Belli) and forcing your nobility and peasantry to fight a war of aggression only costs 2 stability. The current mechanics treat this violation of the norms of the era's warfare as less severe than DOWing someone who grants military access.
- Likewise, initiating Westernization - that is, overturning the social, political, and sometimes military norms of a nation - only costs 3 stability. Currently, the nobility, peasantry, etc., of a nation get more upset over violation of one norm of warfare (don't declare on someone who said you could move through their territory) than overturning some of the fundamental social contracts (forgive the anachronistic terminology) and privileges of the country/its estates.
In short, even if we consider that a declaration of war on a country granting military access is a violation of the norms of warfare and international diplomacy of the time, it seems that the stability cost is excessive compared to other stability hits for similar diplomatic actions. On that basis, I think that the mechanic should be revised. The computer never makes the mistake (as far as I can tell), so it really serves the function only as a punishment to careless players.
If that situation really is an opportunity cost necessary to prevent a worse problem (as Gdj speculated) and we have no better alternatives, then it's fine. If it isn't, then I suggest addressing the problem for a minor "quality of life" improvement for players.
If the current mechanics are in place to prevent exploits (which is entirely possible), then I would suggest (as I noted in the OP) making a DOW impossible on any country granting Military Access. In other words, we could prevent exploits and accidental stability hits by treating treaties of Military Access like other treaties. We currently cannot declare war on allies, countries granting us trade power, or countries we have guaranteed. The mechanics could follow that pattern without any meaningful downside that I can see. Basically, the inability to declare war on a country granting military access would just be a reminder to cancel access and declare war in a month with no stability cost.
I don't see that either casual or hardcore gamers would lose out if we tried either revision.
Since I got more downvotes than upvotes (and only one reply with an explanation), I encourage those who disagree to vocalize their reasoning. Do you, for instance, think that violating military access (which really wasn't a thing as the game depicts during that time period) so much more shocking and appalling of an act than the diplomatic actions that cost us 1-3 stability? If you see an exploit that the stability hit is supposed to prevent, do you think that the current mechanic helps more than it hurts? Is the current set of mechanics really the best and easiest way to avoid whatever exploit might be there? I'm sincerely curious on these points since, as I noted above, this issue isn't my current crusade or anything; I was just reminded of the issue by seeing it brought up 4 times in 3 pages on the "
Stupidest things you have every done in EU4" thread.