You can reduce the code quite a bit by playing to Perl's strengths. The quick way to load the story into an array (once you've opened the file) is

  my @story = <FILE>;
[download]

You can reduce the amount of work you're doing to pull out bracketed words by doing

  foreach my $line ( @story ) {
      $line =~ s/\[([A-Za-z-]+)\]/Replace($1)/eg;
  }
[download]

This replaces each bracketed work with whatever Replace() returns for that word. The /eg modifiers on the regular expression say to replace whatever is found with the results of executing (/e) the expression on the right-hand side, and to repeat this match "globally" (/g) across the target string. And, because $line is "aliased" to each row in @story, you don't need to make a second pass through @story.

Then, to print the story

  print @story;
[download]

Now, you've localized all of the interesting work to

  sub Replace {
      my $word = shift;
      ...
  }
[download]

which I'll leave as an exercise.