22-07-2018, 06:26 PM
(This post was last modified: 22-07-2018, 06:32 PM by Torsten.
Edit Reason: small change
)
(22-07-2018, 05:07 PM)Emma May Smith wrote: You are not allowed to view links. Register or Login to view.Your observation is valid but your hypothesis isn't. The problem you face is that the word statistics for the line starts are often different from the main text (especially in Currier B). It's unsafe to use an observation for a single glyph when there many other differences. Any theory should attempt to account not only for the use of gallow glyphs in this position, but also d, y, s.
It seems that you are talking about something I didn't say in this thread. In this thread I only said something about frequencies for the word types in the VMS. Similar word types occur with similar frequencies. There is no exception for this observation. Even some chained words for which no similar words exists behave as expected. They occur only once. Therefore the conclusion that similar words depend in some way on each other is inescapable.
The usage of t and k is not more then an interesting detail. It is interesting since at first sight we would not expect that qotar and qotal are less frequent then qokar and qokal. I have tried to explain why this word types behave this way. But this didn't mean that I have to explain every detail. The frequencies for the word types differ for Currier A and B. Therefore, if you want to go into all the details you have to distinguish between Currier A and B, between different line positions and between labels and normal text anyway. The important point is that all the details didn't change the overall picture for the VMS. It is possible to describe a mountain as a mountain even if he has some bumps and slacks, since without bumps and slacks it would be a cone.