(14-10-2016, 07:27 PM)ThomasCoon Wrote: You are not allowed to view links. Register or Login to view.Hi Emma - recently I've also been wondering about the relation between o and y - the most common vords beginning <ok> often have <yk> parallels (same with <ot> probably). And in Marco's work, they are in the same group...
The
yk parallels, and indeed most words beginning with
y, have a tendency to be line initial. If characters represent sounds then
y and
o are the same kind of sound and the beginnings of lines induce sound change. It's all very wonderful to think about and I wish I understood what it means.
(14-10-2016, 06:44 PM)julian Wrote: You are not allowed to view links. Register or Login to view.Marco, was your analysis made with the text from the whole manuscript, or a subset of the folios? I wonder about conflating Language A and B ...
Hello Julian,
the analysis was made on the whole manuscript - Takeshi Takahashi's transcription (so it doesn't include the "Rosettes" page).
I will come back to this thread in a few days, in order to comment more. I want to understand these results better
(14-10-2016, 11:10 PM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.It's all very wonderful to think about and I wish I understood what it means.
I agree with you there. At this point I am thoroughly confused. Especially now that I've seen Marco's work.
In many words (but certainly not all), it seems:
<
Sh> can substitute for <
ch>
<
l> can substitute for <
r> (or at least changing <r> for <l> still creates a valid word)
<
y> can substitute for <
o>
<
o> can substitute for <
a>
And the 4 gallows can often substitute for each other.
These 21 strings are all valid (disregarding spaces):
otchord
otchold
ytchold
ytchord
okchord
okchold
ykchold
ykchord
okchald
okchard
ykchald
otShold
okShord
okShold
ykShold
opchord
opchold
opchald
ypchord
ypchold
ofShold
EDIT:
opchard does not appear, as I previously claimed.
Many of those strings are not valid. Indeed, the string rd is not at all common.
(16-10-2016, 05:45 PM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.Many of those strings are not valid. Indeed, the string rd is not at all common.
My bad, I should've clarified: by "string" I meant "sequence of letters (i.e. disregarding spaces)" - I'll edit my post. (Otherwise I would've said vord / token.)
Either way, I'm still trying to make sense of these substitutions...
It's still not true. Even disregarding spaces, I'm not sure some of these occur once.
(16-10-2016, 07:54 PM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.It's still not true. Even disregarding spaces, I'm not sure some of these occur once.
I'd be happy to double-check the Takahashi transcription (which is where I got this information in the first place):
otchord - 7 occurrences
otchold - 8 occurrences
ytchold - 7 occurrences
ytchord - 3 occurrences
okchord - 9 occurrences
okchold - 9 occurrences
ykchold - 5 occurrences
ykchord - 4 occurrences
okchald - 2 occurrences
okchard - 2 occurrences
ykchald - 1 occurrence
otShold - 2 occurrences
okShord - 1 occurrence
okShold - 1 occurrence
ykShold - 1 occurrence
opchord - 3 occurrences
opchold - 4 occurrences
opchald - 1 occurrence
ypchord - 1 occurrence
ypchold - 3 occurrences
ofShold - 1 occurrence
Somehow I must have mistyped 1 of them:
opchard does not appear.
I try to be as objective and factual as possible when I post things like this
If finally put together something that I hope might help in the analysis of the previously posted data. The attached image presents "radar charts" computed on the basis of the clustering input (the csv you can find in You are not allowed to view links.
Register or
Login to view.).
I have manually added in red the cluster boundaries produced by the two algorithms discussed You are not allowed to view links.
Register or
Login to view.. Brighter red lines correspond to the "macro cluster" the were identified combining the outputs of the two algorithms. Fainter red lines correspond to some of the dubious cases (eg EVA:k
k could belong with F K P T or with EVA:t
t).
[edit: I had to redraw the d chart using a different Y-axis scale... it looked weird because of some LibreOffice bug]
You meant obviously not "
macro cluster" but "marco cluster" 
Just to be sure: the clusters are for any position, but relatively for 1 word only ?
h radar showing high for bf.y and af.t and af.y
q radar showing high for bf.o and af.0 and af.l
Could you give 1 word as an example for these please?
(22-10-2016, 09:41 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.Just to be sure: the clusters are for any position, but relatively for 1 word only ?
h radar showing high for bf.y and af.t and af.y
q radar showing high for bf.o and af.0 and af.l
Could you give 1 word as an example for these please?
Hi David,
thank you for your interest!
The clusters are based on the data in the radar charts (same data as the csv attached You are not allowed to view links.
Register or
Login to view.). For each character, the frequency with which it appears before or after each other character are considered. The statistics are based on single unique words in isolation (words are
not weighted on their number of occurrences).
The q radar only shows high for "before o". The graphing software plots a larger blue area because the value for bf.o greatly exceeds 0.2, the limit I set for the radar diameter (in the You are not allowed to view links.
Register or
Login to view. you can see the actual value for "q before o" is 0.88). Since the diagram is so small, the effect is that the blue are comes close to the confining lines for af.l and af.o: I understand that this can be confusing but I cannot think of a different way to plot all these numbers in a single image, I am sorry.
I attach a slightly enlarged version of the q radar-chart.