Quote:I still don't see how this really explains it. It seems like a good theory should account not only for which words exist, but also for how common they are.
@Sam G:
First of all a theory is good if they can explain at least one thing better then all competitive theories. So far I have only given three simple rules to describe the relations between similar words in the VMS. This three simple rules are not a theory and they did not explain the VMS. They only allow to estimate how common the words are compared to other words in the VMS.
Now to your question. The important point is that there is a network of relations between words in the VMS. Even the three simple rules result in a huge number of possibilities. Let us for instance check three possible changes for the word 'qokaiin'. This three changes are:
1) the gallow glyph replaced with another gallow glyph
2) the glyph 'in' replaced with 'iin'
3) groups starting with and without 'q'
With this three changes you have already a grid with 16 possibilities:
qokain (279) | qokaiin (262) | okain (144) | okaiin (212)
qotain ( 64) | qotaiin ( 79) | otain ( 96) | otaiin (154)
qopain ( -) | qopaiin ( 6) | opain ( 2) | opaiin ( 13)
qofain ( 1) | qofaiin ( 1) | ofain ( -) | ofaiin ( 5)
With this grid alone it is also not possible to tell why 'okaiin' is more frequent then 'okain'. The reason is that each word in this grid is also connected to other similar words not shown here. 'okaiin' is for instance also related to 'daiin' and since 'daiin' (836) is four times as frequent as 'dain' (211) it makes sense that 'okaiin' is more frequent then 'okain' even if 'qokaiin' is less frequent then 'qokain'.