Manuscript Relationship Calculator

If you don't know why you might want a manuscript relationship calculator then check out New Testament Textual Criticism. This tool computes the similarity matrix that is the first stage in the traditional method of clustering.

Clicking on any of the hyperlinks below will bring up a more detailed explanation in a separate window.

The variants need to be added into the box below ;one variant per line with a // separating each variation. The format is supposed to be somewhat in the style of the UBS format.


		p10 B 81 // p26 01 A E G 044 6 33 104 256 263 424 436 459 1175 
			1241 1319 1506 1573 1739 1852 1881 1962 2127 2200 2464 Byz [K L P] 
			Lect

		B 0172 6 424c 1739 1881 L596 // 01 A // C 33 81 1506 // D G 1852 
			// 1912 L593 // P // 044 256 263 365 424* 436 1175 1241 1319 1573 
			1962 2127 2200 2464 Byz [L] Lect // 104 459 // K

The program has an internal definition of what a valid manuscript is. Firstly note that only Ascii (Roman) characters are valid. Greek and Hebrew need to be converted. Parenthesis and most other punctuation is ignored.

The text of the manuscript must then be one of these

A single upper case character
A 'P' or 'F' followed by a number
A number
'Byz' or 'Lect'
'L' followed by a number is valid only if you check this box:

Any manuscripts entered that are not in the list will simply be ignored. The first part of the output shows which manuscripts have been counted for which variants and shows which if any, have been ignored.

If you have other manuscripts that you want to be counted too then they have to be entered into this box separated by spaces. Note that CASE is important.

Please note too that variants with only two variations one of which is a singleton as defined by those manuscripts which are accepted will not be used as part of the correlation tables. However those singletons are counted for in the singleton count tables.

By default all of the manuscripts that are accepted take part in the output grid or similarity table. Obviously that can be a lot of MSS and you may not be interested in all of them. If you wish you may list MSS in the box below and then only those MSS listed will appear in the grid. The MSS will appear in the grid in the order you specify them.

This program can compute Colwell Clusters. The traditional manner requires you to enter the MSS that you wish to act as the center of the cluster. You should do that on the text line as a space separated list:

If you do not enter any starting points then the program will try all possible points and tell you which ones work.

Colwell is famous for the 70/10 rule which states that the MSS must be within 70% agreement and with a 10% gap to the next cluster. This is the default the program uses but if you wish to change that you may do so here:

Agreement: Gap:

K-clustering is a more modern method of grouping MSS and this program will do that. Presently the output of this process shows which manuscripts join together in sequence. Unless you are familiar with these things it can be a little intimidating. However I believe that a study of the process will produce valuable insight.

The program has three different methods of measuring the distance between two clusters:

The program also supports three different methods of terminating the clustering process, any or all of them may be used at the same time:

Number of clusters to leave: Minimum agreement to allow: When these appear in one cluster: