$/ is the input record seperator (default is "\n"), and if you change it to something else (such as "a" or "\n\n"), perl changes where it seperates "lines" on a file. For instance, here is a script that will read a file on paragraphs (defined as "\n\n") rather than newlines, and print !!!!!PARAGRAPH!!!!! before each one, to let you know that its working:

#!/usr/bin/perl -w
use strict;
open (FILE, "input.txt");
while (<FILE>)
{
     print "\n!!!!!PARAGRAPH!!!!!\n", $_;
}
[download]

Here is a sample text file to try it on: (because im nice :))

dstgs dahfgvsd kjghkls djgklsdjgl ksdjglk;j sdkl;ghsd jkljg lkds;j glk
+;dsjglk; sdjgkldsjgl ;ksdjgkl;d sjglk;s djgkjdskgjdsjg lkdsjglk;sdjg 
+l;ksdjglkd sjglkdsjlg sd glsd 

dstgs dahfgvsd kjghkls djgklsdjgl ksdjglk;j sdkl;ghsd jkljg lkds;j glk
+;dsjglk; sdjgkldsjgl ;ksdjgkl;d sjglk;s djgkjdskgjdsjg lkdsjglk;sdjg 
+l;ksdjglkd sjglkdsjlg sd glsd 

dstgs dahfgvsd kjghkls djgklsdjgl ksdjglk;j sdkl;ghsd jkljg lkds;j glk
+;dsjglk; sdjgkldsjgl ;ksdjgkl;d sjglk;s djgkjdskgjdsjg lkdsjglk;sdjg 
+l;ksdjglkd sjglkdsjlg sd glsd
[download]