0

While now many smartphones feature the "assistant" of one or another kind, and even have dedicated buttons to invoke them, the experience with these assistants, at least for me, is invariably frustrating. Even short, trivial phrases must be repeated multiple times to be understood. I have experimented with various languages talking with my phone and finally thought, if it cannot do with human language, why cannot we have a specialized one computer understands?

I thought what could make the life of the smartphone easier. Words should have recoverable meaning even if pronounced somewhat wrongly. Minor errors should never result in a valid word with different meaning. Written and spoken forms must have simple and consistent rules for translation in between. Phonemes should be restricted to the subset common in many widespread languages, do not expecting something I have now clue how to pronounce.

Have this ever been attempted?

Improve this question
2
  • phonemes should be restricted? Every language has it own phonemes to differentiate meaning. if you do not represent the [i] in English for example, you cannot express: The bit of the ship that was missing was the bit that was hit. So the common subset idea just does not work. Sorry.
    – Lambie
    Commented Jan 17 at 15:41
  • Rhis depends on speech recognition software, Whisper is quite well
    – Anixx
    Commented Jan 17 at 16:03

1 Answer 1

Reset to default
5

Yes, there was, and the project went a quite a long way to implement it. It was named Robot Interaction Language (Roila) and their website https://www.roila.org is still up (10 years after the completion of the research project).

Improve this answer
6

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.