It would've been 2003 or - much less likely - 2008 or VC6. I don't remember the specifics; I would think it would be to enable the define (FS2_SPEECH, IIRC, but it might be SPEECH_SDK or USE_SPEECH or something) and install the SDK needed. Then enable the custom build step for the xml file and build it.
Again IIRC, I think I upconverted the original speech-enabled project files to 2003, and that's why I remember the custom build step, because I had to redo it. So those 2003 files should be based off of speech-enabled ones; however, it's been a long time, and a lot of people were having trouble with 200*, but there's been a lot of confusion as to which project files you're supposed to use, and which are supposed to be up-to-date or not.
So as a starting point if someone wanted to try to enable it, I would suggest opening the 2003 project files (either with 2003 itself, or with one of the express editions you can get online for free). Check and see if it has an XML file in the speech folder, and if so, whether it has a custom build step associated with it, and if it does, check the project defines and make sure speech is in there. If all of those are done right, you've done everything I remember being necessary for speech.
Finally, I'm not sure whether you're talking about speech input or speech output, but I would start with the above steps for either one. The two are considered to be separate systems.