Load cognate sets

[LinguList]

This requires to aspects: normal cognates, and cognates which are partial.

[LinguList]

Cognates should best be loaded independently of loading a normal wordlist.

[LinguList]

A first solution for "normal" cognate sets has been added. We need new solutions for

• partial cognates (where the reference is to a part of a form, not the entire form, by segment slice
• multiple cognate sets in the same dataset

[LinguList]

Note that the situation of multiple cognates has still not reached a standardized practice. We were discussing to use a contribution table for this purpose, but there is no exemplary dataset for CLDF that would illustrate this for the time being.

[LinguList]

I we are to illustrate multiple cognate sets in one dataset with a contribution table, it may be useful to use the liusinitic dataset, as we describe four variants in this preprint to compare partial cognates to "normal cognates". The major new function has now also been added and tested in lingrex. So this may be a good fit. By now, we do not offer the different cognate sets in CLDF.