I am looking for a data base list of glyph counts for whole manuscript. I know there is debate about the exact identity of some glyphs and difficulty in reading some but I'm guessing all the analysis I have seen has been using some kind of close as set of data fro this. No problem finding word counts and single folios but need preferably list that covers most common 30.
thanks,
Bunny
Here's the info you required via the EVA explanation on Rene's site:
You are not allowed to view links.
Register or
Login to view.
Thanks for the links. I was looking for an actual list but can do a search from complete transliteration files. I speak EVA fluently but i'm no computer analyst so I was looking for something simple like a predone list.
While taking a dose of my own medicine so to speak, and trying other tokens on the text, I noticed with the best will in the world (and I know nothing of sheet music, playing instruments, singing or composing) that music/glyph substitution is producing some "bum" notes that just cant be fitted harmoniously. I expect that if assumptions that exist about some of the glyphs are changed and reordering done, then a more melodious sound may follow. Very in the realm of "monkeys on typewriters" but I look forward to any results. I did hear a brief soundclip someone had done in past along these lines, but was pretty much an ear jamming squeal. From what little results I have seen so far I expect sounds to be able to factually pick out "incorrect" glyphs, now that is a new concept? I know the transliteration types have differences and will look at how the sounds produced compare and if I can get a recorded clip of some of the offerings I will pop it on.
Maths is not my thing but I would be very interested to see some results from an exercise in that area.
Bunny Monkey - typewriter
You can find some frequency counts in the form of graphs at the start of this page:
You are not allowed to view links.
Register or
Login to view.
Note sure if this is what you need. I have the numbers on which these graphs are based, and can include them here too.
This kind of thing ? ( i dont vouch for its accuracy but it's not too far out ).
Takahashi Transliteration ( No Rosette data in Takahashi; words contaiining question marks excluded )
#=IVTFF Eva- 1.5
# Extracted from LSI_ivtff_0d.txt
# Version v0a of 26/08/2017
file loaded 5447 lines.
Total word count: 37759
Total char No. 191009
How many individual chars: 22
letts=['o','e','h','y','a','c','d','i','k','l','r','s','t','n','q','p','m','f','g','x','v','z']
(25368, 'o') (20007, 'e') (17811, 'h') (17611, 'y') (14251, 'a')
(13276, 'c') (12940, 'd') (11707, 'i') (10906, 'k') (10492, 'l')
(7429, 'r') (7364, 's') (6921, 't' (6127, 'n') (5414, 'q')
(1627, 'p') (1114, 'm') (503, 'f') (95, 'g') (35, 'x')
(9, 'v') (2, 'z')
Thanks for all the replies.
It is the sort of thing I'm looking for, but apart from looking at individual letters also need to look at combinations in prallel, c, h, ch. s, h, sh, t, cth etc. I think I can do this now from the transliterations but disentangling the glyph combos from each other and reconing each alone from the numbers will take some time (and from the intros, line numbering and stats on the transliterations that all need to be acurately eliminated). It doesnt matter so much where there are large gaps in frequency but some are closer and want to be accurate. I will look at the individual letters as sounds in sequence and the combos as single sounds also, whichever gives a "bum note" will be presumed to be proven wrong.
I have faaar too much time on my hands atm!
Monkey - typewriter.
Still struggling a bit in disentangling extraneous glyphs of the transliteration from main text eg. V101 carries 9, in both parts. I know EVA and V101for instance come up with different counts based on how the original was read but is there a source that shows just body of text rather than all the workings or is it assumed that the extaraneous glyphs are not significant in number? Project Muse and others have graphs of frequency are any of these dne on the pure text alone?
thanks again,
Bunny
I'm not sure what you mean about "extraneous" text. Do you mean uncommon or rare glyphs? They are not necessarily extraneous.
There is no definitive or 'correct' transcript because we don't know how to parse the VMS glyphs. Is a gallows character one glyph or two (a ligature)? Is a benched gallows one glyph, two glyphs, or three?
Is ch a ligature (two glyphs) or is it one?
Is the cap on EVA-sh sh an abbreviation symbol (like an apostrophe) that indicates missing letters?
Is a analogous to Latin "a" or is it a combination of e + i ?
If there are ligatures, it would significantly increase the overall glyph count.
[quote="-JKP-" pid='37907' dateline='1591790143']
I'm not sure what you mean about "extraneous" text. Do you mean uncommon or rare glyphs? They are not necessarily extraneous.
There is no definitive or 'correct' transcript because we don't know how to parse the VMS glyphs. Is a gallows character one glyph or two (a ligature)? Is a benched gallows one glyph, two glyphs, or three?
Is ch a ligature (two glyphs) or is it one?
Is the cap on EVA-sh sh an abbreviation symbol (like an apostrophe) that indicates missing letters?
Is a analogous to Latin "a" or is it a combination of e + i ?
If there are ligatures, it would significantly increase the overall glyph count.
Hi,
I meant these kind of details on the files, letter count counts all examples.
<f1r> <! $Q=A $P=A $F=a $B=1 $I=T $L=A $H=1 $C=1 $X=V>
<f1r.4,+P0>
Also found with V101 that could not distinguish C and c in word counts. Anyway doing it the long way round to be sure to remove all out of text examples, seems ok, just a long way around.
Bunny
They are the annotations, comments, page headers etc of the transliteration file.
If you want to extract only the pure voynich text from whatever trasliteration file you have.
Use IVTT. Or roll your own, or if you want to do it by hand, thats cool too.
You are not allowed to view links.
Register or
Login to view.