OK, now that's solved, let's look at the definition of 'word' (yes, things are going to get hairy...). Take this sentence, for example:

"No, he said."

The 'words' that your current code would extract are:

"No,
he
said."
[download]

Is that really what you want? Or would you prefer:

No # or, better, 'no'
he
said
[download]

?

during the foreach loop I remove punctuation and set to lowercase (such that No and no are the same word.

while (<FILE>) 
{ 
   my @these_words = split(' ', $_);
   foreach my $this_word (@these_words)
   {
      
      $this_word =~ s/[[:punct:]]//g;
      $this_word = lc($this_word);
      push @all_words, $this_word;
     

   }
}
[download]

$this_word =~ s/[[:punct:]]//g;

The only problem with that approach is that it removes internal punctuation (ie apostrophes) as well, so that I'll becomes ill, she'd becomes shed, etc. ('Why was Virgina Woolf so obsessed with sheds?' I hear someone ask.)

I'd use this instead:

$this_word =~ s/^[[:punct:]]+//; # Remove leading punct.
$this_word =~ s/[[:punct:]]+$//; # Remove trailing punct.
[download]

Update: added Virginia Woolf sentence.