Stefan Th. Gries and Martin Hilpert: Variability-based Neighbor Clustering: A bottom-up approach to periodization in historical linguistics

This companion website to our chapter provides additional online material for users to explore. More specifically, we provide

• an R workspace that contains the function vnc.individual; this function also contains an explanation of the required file format and pointers to papers discussing VNC analyses;
• a tab-separated two-column text file articledata.txt that contains the data on the get-passive discussed in section 4 of the chapter;
• a tab-separated two-column text file newdata.txt that contains a subset of the data on the get-passive discussed in section 4.

With the above R workspace, interested readers can perform VNC analyses on data where each year / time period is characterized by a single numeric value. The following briefly outlines the relevant steps:

1. download and install R from CRAN;
2. start R and enter load(file.choose())¶ and, when prompted to choose/enter a file name, choose/enter the path to vnc.individual.RData
   1. to just repeat the analysis of the get-passive in section 4, enter vnc.individual() and follow the directions (i.e., press ENTER) or enter vnc.individual(file.choose()) and, when prompted to choose/enter a file name, choose/enter the path to articledata.txt on your harddrive;
   2. to perform an analysis on different data, enter vnc.individual(file.choose()) and, when prompted to choose/enter a file name, choose/enter the path to the file with the data to be analyzed on the harddrive (e.g. newdata.txt).

The script will then output two graphs in separate windows: a dendrogram of the type of Figure 1:

Figure 1. VNC dendrogram

And a scree plot of the type of Figure 2:

Figure 2. Scree plot

This can then be saved into a user-specified graphics file (we recommend using .png files).