Thanks Tim, Ian, Beatriz and everyone for your replies and discussion of this idea.
The SpeechAce people have been in touch with me and have created an opportunity for me to try SpeechAce. It is an exploration at this point.
From what I understand, SpeechAce does speech analysis at the phoneme level aiming it at language learners more than what typical voice recognition does, say at Google. I play to explore it some and will report back.
Of course, costs remain a reality. However, they seem open to innovations at that level.
Again, my idea here has been to make use of existing fill-in-the-blank question technology where a layer would get added replacing the keyboard with the microphone. The interaction would compare spoken student statements converted to text to anticipated correct answers programmed into the question by the question creator. It strikes me that having a computer seeking to understand what a student is saying greatly increases the number of oral practice and interactions a language teacher can offer to his or her students.
Mel