A handy command would be something like 'Speaking'. When you typed it, it would tell you what language your speaking currently.
Modifying @stats to show it is easier than making your speech tell you (and I wouldn't want to see it every time I speak, that'd be spammy I think).
Maybe an @speech command that includes what your speaking, in what voice, who your addressing and who your watching is the right way to go about it.
Or maybe it would be easier to just make 'speak <whatever>' not clear your voice history if its the same language you were speaking. However, the speak command is the easiest way to remind to everyone what language you are speaking.
Decisions, decisions.
I reckon that's clearly the best way to go about it, although it would involve some more work. Hey, the best things don't come with the least effort right?
Anyway, I remember having problems numerous times where I thought my char was addressing or watching someone, later to find out that he really wasn't. It gets tedious having to manually check all the time to make sure the conditions that one is assuming are in fact in place, when one should be concentrating on the RP.. the content of conversations, not how they're being technically delivered.
Correct me if I'm wrong, but when ones character follows another, they don't stay addressed when entering a new room right? They're treated just like anybody else, the following isn't taken into consideration.. same with watching. Surely when people are walking together, they communicate in transit.. having to readress etc is a pain and unrealistic in my view. Same with watching. If one is shadowing, fair enough.. one probably shouldn't be able to 'watch' another, to eavesdrop.. but following is explicit.. the people would be close by, so I reckon watching should stay intact for someone following another through mulitiple rooms.
Erm, that was about 10 times as verbose as it had to be, but yeah.. yay for @speech, and yay for watch and addressing staying intact from room to room when followed (unless it's not like that anymore).
I think it's a good idea to be able to keep track of all these communication factors easily considering how complex it has now gotten, so an @speech is more an essential than just a handy little feature for the lazy bums.
etc.
So, since pose/emote can't be relied on to know who your addressing, try seperating your speech from your pose at least the first time you want someone to hear you. Start off with the 'to' and go immediately into your following pose. Or even use a social, if you don't mean to say anything. Surely there are enough subtle socials like nod, brow, eye, etc that -will- add them to your addressing list just as 'to' would.
Pose/Emote/Spoof are too broad and general for us to try and put more logic in them that we've done already. I do agree with the 'follow' thing, file a bug, I consider it one.
I don't think there are any holes in the help surrounding our speech system. If so, point them out and we'll see what we can do.