Another statistic that can be explained by the VBM
Background: I came across the following because I wanted to verify whether the aiin family actually encodes endings like "en" in German.
Hypothesis: If the “aiin” ending corresponds to endings in the German language, such as “en,” then theoretically it can also be formed using the “E” of the vowel bridge ="y[space]qo" followed by a coded consonant cluster.
Assumption: Since the “aiin” family appears with varying frequency across the individual sections and in Currier A / B, the question arises as to whether this alters the frequency of "y[space]qo" = "E" which would be an indication supporting the hypothesis that the “aiin” family encodes German endings like En.
In short:
A low frequency of “aiin” should lead to an excessive frequency of "y[space]qo", and vice versa.
To get straight to the point: This is exactly what the data show.
The Measurement
[
attachment=16670]
The sum is not universal but is grouped into two text-type levels: The sections Herbal A, Herbal B, Pharma average around ~15; the sections Balneo, Stars average around ~23,6. (This two-level division is interesting in itself, and there are reasons for it, but that would take us too far afield here.)
More importantly: Within a given level, the two groups swap places.
The clearest example is Balneo versus Stars, since both have nearly identical sums (23,85 vs. 23,62) but a very opposite distribution:
Balneo routes the material predominantly through the y[space]qo bridge (18,05) and only slightly through the aiin ending (5,81); Stars does exactly the opposite (11,92 / 11,70).
Same total amount, different distribution!
Herbal-A versus the rest shows a slightly weaker version of this same
opposite relationship within the 15th level.
This gives us another statistical peculiarity that can be explained relatively easily using the VBM. Since no one has apparently tested "aiin" against "y[space]qo" yet, I’ll go out on a limb and say that this statistical anomaly is new. (but i checked it: Emma May Smith has discovered something similar, but of course without the vowel bridge
.)
If "aiin" and "y[space]qo" should encode different things, one has to ask why their sum should remain constant while the respective parts fluctuate so greatly - difficult...
Jost
(Note: And since I've been writing for weeks that y[space]qo = e, and aiin is probably a German ending like en, this isn't some neatly defined finding that came up later either.)
