Well, it's not a regexp and it's not a perl one liner, but this will do what I think you actually want to do using any awk in any shell on every Unix box:
$ awk '{c=$0; $1=$1} k==$0{print p ORS c ORS} {k=$0; p=c}' file Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. Elvis was the greatest singer of all time. He has 18 number one in billboard charts. Elvis was the greatest singer of all time. He has 18 number one in billboard charts. Mary has a little lamb. The live near the mountain. Mary has a little lamb. The live near the mountain. This has been the coldest winter in the last 20 years. This has been the coldest winter in the last 20 years. I'm assuming you favor brevity since you asked for a regexp or a perl one-liner but if you prefer legibility, here it is again with more meaningful variable names and better white space:
awk ' { currRec=$0; $1=$1; currKey=$0 } prevKey==currKey prevKey == currKey { print prevRec ORS currRec ORS } { prevKey=currKey; prevRec=currRec } ' file