2

I am trying to apply Cornell Semantic Parsing framework (implementation of Combinatory Categorial Grammars CCG) to the German language. This framework takes natural language texts, learns grammar and translates those texts into logical expressions (lambda calculus) that are useful for natural-language processing.

My question is: Are there some developments of CCG for German language? Google gives almost nothing fundamental, only some examples are considered. But I searched only resources in English language. Maybe there are good developments and efforts that are available only in German language? I would be happy to hear about them.

What is general consensus about applicability of CCG to German language? From the one side German language has strict word order and therefore CCG seems to be applicable. From the other side, German has article and word inflections and this can make things more complex.

Any ideas and hints are welcome. It would be nice to continue already existing efforts and not to reinvent everything from scratch.

  • 2
    You are wrong in your assumption German had strict word order. Apart from V2, which in addition is only applicable on declarative sentences, and some prefixing rules, word order in German is horribly complex and not easily put into rules. – Janka Jun 30 '17 at 8:35
  • 1
    As Janka already said: German has NOT a strict word oder. German word order is very flexible. Learn more here: german.stackexchange.com/a/37200/1487 – Hubert Schölnast Jun 30 '17 at 9:10
  • 1
    I absolutely doubt you could easily adapt an algorithm on word order that was originally designed for the English language to German without doing some baseline work first and fundamentally change it. Except for entirely trivial sentences, German word order cannot only be horribly complex, it also tends to carry notion and emphasis, which is much more important for language processing. – tofro Jun 30 '17 at 9:45
  • 1
    On the other hand, German doesn't have very free word order either. Changes in word order generally come with changes in information structure and there is a pretty clear basic word order that speakers usually adhere too, all things being equal. There is also nothing in this post to suggest that the algorithm was designed (only?) with English in mind. – user9315 Jun 30 '17 at 11:00
  • 1

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.