Quote:Have you produced any text that has this property? I don't see why auto-copying should produce it.
I have written an app for doing so [see You are not allowed to view links.
Register or
Login to view.]. The sourcecode for this app is available via Github [see You are not allowed to view links.
Register or
Login to view.].
The entropy values for the text generated with this app are comparable with entropy values for the VMS:
Currier A Currier B App
H0 4.46 4.46 4.39
H1 3.82 3.88 3.81
H2 2.11 2.01 2.21
I only say that it is possible to explain the VMS with the autocopy hypotheses.
Quote:1) Asserting, without any evidence, that it cannot be a property of a natural language text
First of all, that it a text using natural language is your hypothesis. Therefore it is on you to demonstrate evidence in favor of your hypothesis. Sorry, but this is the way science works. It is a common mistake to assume a starting hypothesis while trying to interpret an undeciphered script. The danger this way is that every characteristic of the script will be interpreted with this starting hypothesis in mind. Because of this reason you should search for evidence for your starting hypotheses.
Secondly, the weak word order for the VMS alone is evidence against a natural language [see You are not allowed to view links.
Register or
Login to view.].
"Only a few repetitive phrases can be found. There are only 35 word sequences which use at least three words and appear at least three times. Only for five of these sequences is the word order unchanged for the whole manuscript, whereas for 30 out of 35 phrases the word order does change." [You are not allowed to view links.
Register or
Login to view.].
Quote:2) Asserting, without any evidence, that auto-copying would produce that property
"In 66 out of 72 cases (91%) at least one similar word was found within a maximum distance of three lines and a maximum edit distance of three. Furthermore, if near is defined as both glyph groups must be used one after another or in two consecutive lines one above the other, the result is still interesting. In 62 out of 140 cases (44%) and in 25 out of 72 cases (35%) a similar glyph group can be found for both samples. In other words, similar glyph groups can be found above each other twice as often as they can be found side by side." [You are not allowed to view links.
Register or
Login to view.].
"The words in the VMS build a network of words similar to each other. Therefore it is no surprise that a larger number of similar words exist for each word in the VMS. However, they occur near to each other and they occur with comparable frequencies. ... A better explanation is that since similar words do co-occur throughout the text, the spelling variations for a frequently used word also occur more often. In other words, for the VMS, the observed word frequencies are a result of the fact that similar words do co-occur throughout the text" [You are not allowed to view links.
Register or
Login to view.].
Moreover, it is possible to order the types by their similarities to build a multidimensional grid containing all word types which occur at least four times [You are not allowed to view links.
Register or
Login to view.]. See the network graph for the VMS compared to that of the arabic You are not allowed to view links.
Register or
Login to view.:
[
attachment=1117] [
attachment=1118]
An enlarged part of the graph for the Quran reveals that typical for the arabic language are multiple smaller networks with two, three or up to 20 words:
[
attachment=1119]