"#Antiword" is a command line program that can read the proprietary "doc" format (not docx) used by old MS Word versions. Pandoc and antiword can be combined via #DocBook XML; this offers a way to convert old doc files:

antiword -x db input.doc | pandoc -f docbook …

@toxision I would try #antiword first, and if things get despirate I would use the “strings file.doc”, which will simply extract all printable characters from a binary file of any kind.