It's harder to describe without code than with.

First define what constitutes a word. Upper and lower case letters obviously, but what about the quote in "don't"?

Then define what comes in between 'words'. Spaces for sure, but what about punctuation like ,.?!- etc.

Once you've defined your words and separators, you want to capture n words and separators before the match and n separators and words after.

$word = qr[[a-zA-Z']+];;
$sep = qr[[ \t.,!?]+];;
$n=10;
$match = 'match';
$re = qr[((?:$word$sep){$n}\Q$match\E(?:$sep$word){$n})];;

print $para =~ $re;;
want to extract n words on either side of the match. I am sure modules
+ such as Plucene or KinoSearch already
[download]

So far, so good, but what happens if the word is less that n from the start or end? It won't match, so the quantifier has to become {0,$n}:

$re = qr[((?:$word$sep){0,$n}\Q$match\E(?:$sep$word){0,$n})];;
$match = 'CPAN';;

print $para =~ m[$re];;
There probably is already a CPAN module for this, but I am stuck tryin
+g to figure
[download]

Better, but it will still fail if the definition of words and separators isn't good enough

For example, if there are numbers in there somewhere. Or brackets. Or quotes. And that's before you get to trying to deal with unicode.

At best the above is a starting point. How far you need to go beyond that will depend upon your application.