Just to qualify noahbickart's suggestion, the L/R -> M/S -> L/R approach works best if a coincident miking setup was originally used, and less well (sometimes quite poorly) if the microphones were spaced apart.
In general the notion of "filtering" or other post-processing to decrease reverberation is a "holy grail" type of quest that can only be achieved to some given degree, never absolutely (unless the reverberation came from separate microphones and was recorded onto separate tracks in the first place, and you can go back to the original multi-track recording and remix--obviously not what we're discussing here).
Apart from that type of situation, you'll always have to give up some degree of the naturalness and believability of your recorded sound quality--quite possibly a lot of it. In the end it may not be worth the cost. The universe doesn't guarantee a technical fix for every possible kind of mishap.
For example, even with M/S (in its conventional, two-microphone or two-signal form), when you decrease the reverberance by lowering "S", you are also cross-blending the left and right channels. So you can get a recording with less reverberation, but it will also have less stereo separation. And as I said, that whole approach really only works well when the original recording was X/Y (or was M/S itself); any kind of spaced recording will have comb filter effects when you go to derive the "M" and "S" signals from your left and right channels.
If you suspect in advance that you're going to need this type of post-processing, you might consider double M/S or even Soundfield recording, if those are possible options for you.
--best regards