I know Apple and Google have predictive text, but I believe it is purely statistically generated, without analysis of a sentence's grammar. I know various computer parsing systems of human language exist (https://en.wikipedia.org/wiki/Parsing#Human_languages).

Is there any system which uses the actual grammar rules of a language (for example, the Cambridge Grammar of English) and the actual lexical information of the words (for example, "the" must take a noun as its complement), and a dictionary, to suggest, in an incomplete sentence, which words could come next which would be grammatically correct?

  • 3
    This was the norm before statistical models. Note that suggesting based merely on grammar would produce infinite options. – Luke Sawczak Jan 21 '20 at 14:11
  • Few languages have an infinite vocabulary from which to choose the next word. – hippietrail Jan 21 '20 at 19:08
  • What do you mean by "actual"? Grammatical rules as stated by Cambridge or anyone are a mere human attempt at capturing the workings of a living language, an these attempts are very useful but neither complete nor perfectly accurate. They do their best to describe how speakers of a language use it and when they perceive it to be used effectively. – LjL Jan 23 '20 at 23:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.