I was looking (for a statistics project) to the Lexical Distance between languages and I came across this post

Worldwide map or data for linguistic distance?

I was wondering if there is any "numerical table" of these distances, in order to do myself some analysis!

  • You would probably have to contact Prof. Tyshchenko about that. It may be in the original Ukrainian article. But it doesn't seem to represent anything except shared vocabulary, which is not the same as "distance between" languages in some theoretical Language Space.
    – jlawler
    May 23, 2016 at 13:39

3 Answers 3


It's not in a convenient format at all, but if you look at the SI (supporting information) tab for the paper Support for linguistic macrofamilies from weighted sequence alignment there are .svg files with very large phylogenetic trees with edges labelled with similarity between languages. You could contact the author for something a bit more directly useful!


Except of the prof. Tyshchenko lexical distance map there are lexical similarity tables of Indoeuropean languages families by the outstanding American linguist Isidore Dyen https://books.google.com.ua/books?id=AoGmexxHvJQC&hl=uk&source=gbs_book_other_versions Here are similarities of The Romance languages (page 102), Germanic languages (page 107), Slavic languages (page 112). In the lexical distances there were included loanwords and partial cognates with so-called irregular sound change, as it is shown for example on the pages 95-96. Probably, ASJP scientists tried to calculate all language families of the world with using of some (not thoroughly clear yet) kind of Normalized Levenshtein distance. It seems that the latter used for complete cognates, partial cognates and non-cognates. Here is one of the examples of the results of their researches (but they are shown graphically and not numerically)

https://www.researchgate.net/publication/273347143_ASJP_World_Language_Tree_of_Lexical_Similarity_Version_2_April_2009 These "trees", which were made by a computer, may be often with mistakes as it was noticed by a linguists V.Solovyev and T.Galeev here in page 899 - http://journals.uran.ua/visnyknakkkim/article/view/171877/171594

So, it's necessary to follow the nrwer versions of the ASJP lexicostatistics.

Still, these and the other kinds of the lexical distances were generally related to the Swadesh list and to the aim concerning the calculation of the exact time of a language emergence (glottochronology) or to putting it in some traditional or not traditional subgroup (so called - genealogy like Isidore Dyen put Ukrainian and Belarusian in Central-Eastern Slavic subgroup).

Intelligibility studies of two or more languages relied mostly on experiments with understanding defined spoken (more) or written (less) texts. In the texts, there the average lexical distance (the words with the different roots or meaning) is calculated The cognates and the partial cognates are mostly calculated for the prediction of the possibility of their mutual intelligibility with the variants of the Normalized Levenshtein distance. Here is a good example in the article "How easy is it for speakers of Dutch to understand Frisian and Afrikaans and why?" by the Dutch linguists Renee van Bezooijen and Charlotte Gooskens



Maybe you already have finished your statistics project, but anyway here is a matrix of the available lexical distances according to Tishchenko. It includes no lexical distances that I have calculated.

I hope you can share anything you come up with.

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.