Depending on your needs, yes. You can't do speech-to-text, but you could do text-to-speech. Basically, the easiest way to do this is with VOIP (a phone that runs on your computer) - Skype is the best known, but there are others. I use Gizmo; Vonage is also popular.
So, it's running on your computer. Then you use a text-to-speech (TTS) program. What you use is going to be dependent on your operating system; I don't know what you'd use on Windows, but Mac OS X has TTS built in (in terminal, it's the "say" command). On linux, I forget the exact name of the program I played with - either FreeTTS or Festival. Whichever it is might run on Windows as well, but I'm not sure.
You have to connect the output of your TTS program to the input of your VOIP program. On Linux, I think you could do this with esd, but I'm not sure on the details. On OS X, there used to be a program called ... SoundFlower? That was free, but I don't think it works on the new (Intel) Macs, and it's not being developed anymore. Still, I'm sure there's a way to do it. With Windows, I'm not sure what you'd use.
This is all fairly doable if you're fairly geeky, or it'd probably be a modest-size project for a decent programmer to put it all together in an "install and go" package that anyone would be comfortable with.
Hope that helps.