Here's one way of using Perl to read Microsoft word documents. I don't have Windows or Word to test—and the post cleverly mentions that solutions like this go out of date rapidly, but doesn't provide any date for the post itself, so I have no idea how old it is. Anyway, this might be enough to get you started. (There's a Win32::Word::Writer module, but, of course, that's the opposite of what you want.)

The simple and glib answer to “Can you re-construct the original Word file?” is “No”. Is it acceptable to convert to some other format instead of plain text? Word itself can save as RTF or HTML, I think; the HTML it generates is pretty awful, but there are programs to tidy up for you. An HTML document will be much more faithful to the original than plain (unformatted) text would be.