I have used antiword successfully in the past for reading the text of Word files at the command line. It doesn't seem to be actively maintained any more, though.
I also notice that AbiWord has a command line option for converting Word to other formats. You could of course use the full GUI version of AbiWord, or indeed OpenOffice.
(Update) I realise of course that none of my answer directly answers the question of reading these files in Perl, but in practice the command line possibilities mentioned are often a practical way to go.