Chaining and the Growth of Linguistic Categories

This repository includes code and data for analyses in the following work:

Habibi, A.A., Kemp, C., and Xu, Y. (to appear) Chaining and the growth of linguistic categories. Cognition.

Python Files

• main.py runs the whole process, using the other files. You can follow the flow of the code from this file, if you need to.
• paths.py indicates the paths to the data and directories used throughout the code.
• model_saver.py creates the necessary format for the word embedding input file.
• super_words_builder.py creates the file that contains the mapping of the classifiers and their nouns for each year with the right format for the input of the program.
• the_predictor.py has the SuperPredictor class that does the prediction for a set of classifier nouns for a number of steps and saves a prediction file.
• predict.py includes the function for running the prediction processes and kernel-width optimization using the SuperPredictor class.
• the_analyser.py contains the SuperPredictionAnalyser class that has the necessary methods to analyse the accuracy of a prediction file.
• analyse.py includes functions that use the SuperPredictionAnalyser class to do the evaluation and create plots.
• simulation_analysis.py creates a simulation of the important chaining mechanisms on a randomly generated data and plots the results.
• utilitarian.py contains some utility functions and classes.

Data Files

• data/super_words-chi-Luis-YbyY(w2v-en).pkl is the file containing the classifier-noun data over time with the specific format that is used in the code.

• data/w2v-chi-en-yby.pkl is the word embedding file that maps Chinese nouns to the embedding of their English translations.

• data/gwc2016_classifiers/ includes the data on classifier-noun pairs provided by: Morgado da Costa, L., Bond, F., & Gao, H. (2016). Mapping and generating classifiers using an open chineseontology. InProceedings of the 8th Global WordNet Conference.

• English_translation_survey_results.zip includes three xlsx files for surveying 100 sampled Chinese nouns in English translations.