1

Besides Frege's principle of compositionality are there any other philosophical frameworks of language to form distributional semantics in the English language?

Edit - Although I'm still yet to complete in full, my doctoral studies involved modelling free association strengths between words. This heavily relied on free association networks collected manually from human subjects. As a means to overcome the extreme cost (both finally and temporally) of creating such networks I sought to form representations / approximations of said networks using geometric representations of words derived through distributional semantics. At the time, Frege's principle was the basis upon which representations were formed. I have since returned to this field of inquiry an in inspection of the standard journals and research outlets I am yet to find an alternative framework upon which distributional semantics can be employed in creating representations for words to be used in linguistic tasks.

I'm unsure whether this is a result of my deficiency in the philosophy of language, so to repeat, are there any other philosophical frameworks to form representations of words using distributional semantics?

  • 1
    Just ask you question on Linguistics. You're more likely to get suitable answers there. – BillJ Sep 30 '19 at 14:13
  • 5
    Have you looked into Wordnet and Framenet? They go quite a ways beyond what you're asking for. – jlawler Sep 30 '19 at 15:03
  • 2
    I might have attempted to answer it, but realized I'll have to write a lot. Instead, I'll refer you to the canonical textbook, Speech and Language Processing, specifically, Ch. 18, 19 and Appendix C. – prash Oct 2 '19 at 16:47
  • 1
    @curiousdannii - There is a huge field of study into this area. Generally the mathematical representations employed are High Dimensional Real Vectors, Discrete Probability Distributions, Weighted Di-Graphs, and Deep Neural Networks. – David Galea Oct 3 '19 at 4:18
  • 2
    @curiousdannii See word2vec (or word embeddings). In short, map (somehow) a word to a high dimensional vector such that vector operations (usually just distance) capture something about meaning of the words. Also see ontologies for a logical framework (which I consider mathematical). – Mitch Oct 3 '19 at 13:17

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.