2

Where can I find machine readable datasets for use in phylogenetic estimation? I am looking for anything---lexical cognate data, phonological data, morphosyntactic data---in any language family.

  • Everywhere. Online, in books, journals... you might rein it in and ask e.g. "where can I get datasets for this, involving Luhya languages?". I would not even go as high as "Bantu". – user6726 Jun 14 '18 at 1:25
  • Unsupervised approaches to this are going to have hidden subjectivity. Vocabulary, orthography and even grammar bleed so much across language families and subfamilies, and today's ML does not even attempt the kind of reasoning needed to work that out from unstructured data. – Adam Bittlingmayer Jun 17 '18 at 13:42
  • So you could go with a supervised approach, whereby you decide on some initial relations in a way that you do not need to give falsely precise numbers, like just the assertion that distance(pt, es) < distance(ro, es). There could still be subjectivity especially if you do more than the very safe ones, but at least it is explicit subjectivity. – Adam Bittlingmayer Jun 17 '18 at 13:45
  • And then you can build representations of the language from anything, including not just data in the language but things like the vector of the English Wikipedia article about the language. – Adam Bittlingmayer Jun 17 '18 at 13:47
1

For a quick and dirty approach to phylogenetics, you can look for Swadesh lists for many languages.

For doing something better, I am afraid you will have to vet your data carefully, use the best etymological dictionaries available, and prepare it on your own.

| improve this answer | |

Your Answer

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

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