Zodiac labels

[takrobat]

I find it interesting that there are some indications of seasonal change, but since we also know that there are changes within the text itself (line to line -> paragraphs etc), I am not sure if this is seasonal or just an artifact of the text changing from page to page (like a rolling ruleset that changes). 
Would be happy to take any ideas on how to proceed.

[ReneZ]

I ignore all that have "*" or "!" in them, as I am not completely sure what those mean?

The "*" represents an unrecognised or unreadable symbol. It is sensible to ignore words with these, but it is also an option to look at other transliterations that may have made a different decision.

The "!" should only appear in the LSI 'interlinear' file and can be ignored. It means that one of the other transliterations (in the same file) has a symbol here.

Ficticious example:  ols!ey should be read as olsey

[takrobat]

Thank you very much for the clarification. Made som quick changes : 

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

You are not allowed to view links. Register or Login to view.

[takrobat]

Midix patterns:
Midfixes used:


Are they after prefix, alone or with "non known prefix (other)":


other+ means for example words like "olfsheoral", where it recognizes "she" as midfix, but olf is not a valid prefix. Other examples:

  1. "ot" + midfix:
    - otchodals = "ot" + "ch" + "odals"
    - otshshdy = "ot" + "sh" + "shdy"
    - otchody = "ot" + "ch" + "ody"
  2. "of" + midfix:
    - ofchdady = "of" + "ch" + "dady"
    - ofsholdy = "of" + "sh" + "oldy"
    - ofchdysd = "of" + "ch" + "dysd"
  3. "chc" + midfix (hard before hard!):
    - chckhhy = "chc" + "k" + "hhy"
    - chcthhy = "chc" + "t" + "hhy"
  4. Other unusual patterns:
    - olf + "she" → olfsheoral
    - ole + "t" → oletal
    - eeoee + "t" → eeoeety


Some of these are probably some transcription errors, I don't have time to look too much into it right now.  But what we again can see is that there are differences between zodiac in how they are built. I will try to summarize this better in a later post, right now I am just dumping my finds here so other people also can take a look.

[takrobat]

Some zodiacs seems to have a lot of midfixes (Aries always have them), but some have a lot of "soft" words, cancer for example have 43% of words without midfix. Examples:
  - Cancer: or, oral, ory, os, oeees (13 total)
  - Leo: daiin, loly, ogeom, oreeey (9 total)
  - Virgo: oeedey, oeeo, oraiin (8 total)
  - Scorpius: Only ar (1 total!)

[takrobat]

I think it is totally clear that there is a real difference between the zodiacs. Is it because of the zodaics themselves, or because it is an artifact of other things in the encoding, I don't know. But I will continue to look at this, as I find it very interesting.

[nablator]

I think it is totally clear that there is a real difference between the zodiacs.

You will find such statistical anomalies on most pages. No section is homogeneous.

[takrobat]

I think it is totally clear that there is a real difference between the zodiacs.

You will find such statistical anomalies on most pages. No section is homogeneous.

Thank you for the feedback.
I know about this phenomen, which is why I say I am not sure if this is an artifact of the "different" pages, or an artifact of the content of the different pages.  I could compare it to other pages, but since it is labels - mainly one word labels - I think it will not be a good comparision. But it's on my list

[takrobat]

I have been thinking if it could be some sort of counting system or something that would reveal something increasing or decreasing, so I decided to check different steps through the "ring" of labels (which isn't even "one" ring). My idea is that measuring levenshtein distance with different steps would either show some sort of "cyclic" underlying knowledge, or it would show no or little difference - which would mean that the text has no obvious "order" or it would show a lot of difference which means that there is some kind of order to the labels. I am not 100% sure that my thinking here is correct, but my experiment yielded the result I expected : No real difference with increased steps.



To me the conclusion is this : 
The order does not reveal anything special, the words does not seem to be get more and more different. Then again, I am not sure levenshtein distance is the correct way to measure this - levenshtein distance measures the number of transformations needed to go from word A to word B, but we can be pretty sure that these are not "words" as we normally think of words.

[takrobat]

My next chart is not the best, for a lot of reasons, but I still find it a little interesting. What I am trying to achieve is finding phases within the labels, which means a switch from "o"->"y" prefix, for example. I did this very simple, and just counted "y"-phases, "o"-phases and other phases. So whenever the labels switch we increase number of phases. 
This has a lot of problems, first of all where we start in the ring will automatically influence the number of phases (if you start in the middle of a phase you will count it twice), and second not everything is _one_ ring, but multiple. I haven't tried to do anything to counter these obvious error sources, but instead use the numbers as a very simple overview of the number of times a "phase" changes. Do not take these numbers as a 100% correct representation, but instead as an indication. We also know that "o" and "y" in itself is not necessarily separate phases, as we have more complex prefixes (o followed by another letter). If someone finds this interesting, I am of course willing to look further into it. But for now, this is what I have.