I think it's more to do with certain sounds, we've all played first person shooters where you try to avoid using a certain machine gun because it has a very loud repetitive sound. In Perfect Dark Zero when Joanna takes a hit she sometimes lets out a "EEEEE!", which isn't fun, but get hit in Gears of War and you'll get a lesser sounding "ugh" which you can barely hear, and won't be annoyed by.
From the developers point of view there is also memory to consider, you could have 50 different variations of "ugh" "oof" "ahh" "oowhahoo" for 1 character, but that could be anything up to 25MB depending on sound quality, multiplied by the number of characters who need similar treatment like squad members plus other sound effects. These voices and sound effects get played on headsets and surround sound speakers so they need to be high quality.
You may well say "25MB is nothing, you could compress it even smaller", more compression requires longer load times to decompress everything, to have a very large number of sounds loaded and ready to play in time with instant events would require more powerful hardware or it will directly impact how smoothly the game plays.
You've probably noticed in some games with big conversations, the disc spinning or the hard drive ticking away whenever someone is about to speak, followed by a brief pause, and then the audio coming out, sometimes the character will even proceed to talk before the words come out, seen often in Oblivion and Mass Effect, these audio files are not pre-loaded as they don't need to be, gamers can afford to wait in these slower moments of gameplay.
So, for now, we will have to put up with our 4 or 5 "barks" per character, and developers will have to carry on making them more discreet to avoid repetition.