Source Code Accompanying the Paper "More on Network Approaches in Historical Chinese Phonology"

This repository contains source code and data for the following paper:

• List, Johann-Mattis (2018): More on network approaches in Historical Chinese Phonology (音韻學). Paper prepared for the LFK Society Young Scholars Symposium. Taibei: Li Fang-Kuei Society ofr Chinese Linguistics. URL: https://hal.archives-ouvertes.fr/hal-01706927.

@InProceedings{List2018a,
  author      = {List, Johann-Mattis},
  title       = {{More on Network Approaches in Historical Chinese Phonology (音韻學)}},
  booktitle   = {{LFK Society Young Scholars Symposium}},
  year        = {2018},
  publisher   = {Li Fang-Kuei Society for Chinese Linguistics},
  pdf        = {https://hal.archives-ouvertes.fr/hal-01706927/file/main.pdf},
  url        = {https://hal.archives-ouvertes.fr/hal-01706927},
  address     = {Taipei},
  hal_id      = {hal-01706927},
}

Requirements

• lingpy (http://lingpy.org)
• networkx (https://github.com/networkx/networkx)
• cytoscape (http://cytoscape.org, for exploring output graphs)
• SplitsTree (http://splitstree.org) to render the data-display networks

Xíeshēng Networks

The folder xiesheng offers the following code, data, and additional files:

• sinopy.py: a little library offering a script needed for the computation
• `karlgren1957.tsv': Karlgren's data on Grammata Serica Recensa, taken from the WikiBooks (see paper for details)
• network.py: main script which computes the network
• xinsheng.tsv: data from CJKVI-IDS on the phonetic structure of characters
• `ab-groups.tsv': output produced by the algorithm, all groups in which third and first division are distinguished
• graph-full.pdf: full graph in PDF
• graph-all.gml: full graph in GML, can be read into cytoscape
• web-session.zip: Unzip the directory and open the file index.html in your browser to inspect the network interactively (exported thanks to Cytoscape)

To run the code, just make sure to have all dependencies installed (and that you use python3), and run:

$ python3 network.py

This will re-create the data (network, groups).

Fǎnqiè Networks

The folder guangyun offers all data and files you need to replicate the study on fǎnqiè spellings mentioned in the paper.

• guangyun.py: main script to run the analysis
• guangyun.tsv: main data needed for the analysis
• sinopy.py: helper library
• tls.json: helper data for MC readings
• guangyun.gml: basic output (the network)
• web-session.zip: unpack and open index.html in browser for interactive inspection

To run this code, cd into the folder, and type:

$ python3 guangyun.py

Evaluation

Evaluation scripts can be found in the folder evaluation.

Comparing Reconstruction Systems

The folder evaluation contains the following four files related to the comparison of reconstruction systems:

• matrix.dst: output of the algorithm, convenient to use in SplitsTree
• reconstruction.py: the code used to compute the network
• reconstructions.tsv: similar to reconstructions.json below, but different reconstructions
• reconstructions.json: eight different reconstructions for Old Chinese in json format (for a selected number of characters)

To run this code, cd into the folder, and type:

$ python3 reconstruction.py

Comparing Rhyme Analyses

The folder evaluation also contains the data and code for the rhyme analyisis comparison.

• Baxter1992.tsv: rhyme annotation by Baxter (1992)
• Wang1980.tsv: rhyme annotation by Wáng (1980)
• rhymes.py: script that runs the comparison

To run this code, cd into the folder, and type:

$ python3 reconstruction.py