Inferring diachronic changes of typological variables from phylogenetic trees

Disclaimer: Please note that this is a work in progress. Neither the data nor the results are final.

From Natalia Levshina, I got a csv file with a collection of five typological variables pertaining to word order and case from several hundred languages.

She works on a project on causal inference over typological variables.

The languages are non-independent due to common descent, I did a little study inferring a continuous latent variables for each observed variable. The fitted models provide a posterior distribution over diachronic changes of these five variables along the branches of phylogenetic trees.

Here are the steps I carried out

• prepare_data

• Bayesian phylogenetic inference

• create_posterior_trees

• fit_wordorder

• fit_case

• fit_entropy

• evaluation