- Options
- Login
- Desktop Style
Options
A key to understand the VMS
- A key to understand the VMS
-
RE: A key to understand the VMS
stellar > 05-02-2017, 09:23 PM
Sorry about the last post as it may cause confusion.
What we are seeing with the VMS and at last a person like Torsten who maybe resting on the shoulders of giants is that he sees prefix's are changing in huge % differently from suffix's or just the opposite.
Now does a natural language have these properties? Or does a natural language retain prefix's and suffix's more evenly placed across a language?
Can the prefix's and suffix's be the window for unlocking the VMS? I believe so and I'm will to beT that and invented language using numbers would follow this properties of suffix out weighing prefix or the other way around. It just depends on how the Author would write. But the most obvious choice would be to edit the prefix more often then the suffix to change the word number. WAKE UP VOYNICHEROS! -
RE: A key to understand the VMS
nickpelling > 05-02-2017, 09:36 PM
(05-02-2017, 07:49 PM)Torsten Wrote: You are not allowed to view links. Register or Login to view.
(05-02-2017, 05:35 PM)nickpelling Wrote: You are not allowed to view links. Register or Login to view.Torsten: so what you are saying is that even though Voynichese is autocopied from itself, the "itself" from which it is autocopied (i.e. that seeds the copy) contains near-universal adjacency rules which are then preserved in the autocopying?
No, adjacency rules are not preserved from the seeding part they are a result of the autocopying process. If a word is copied and some glyphs are replaced with similar ones the order of the glyphs stays unchanged.
Well... if your starting point is akin to that of a verbose cipher (e.g. where al ol ar or qo are all irreducible tokens, as per Curse 2006 etc), you're starting with (what to nearly everybody else's eyes looks like) a set of adjacency rules within those irreducible tokens.
And then you run into problems parsing a and o, because they appear both inside and outside your core set of irreducible tokens. What is someone to make of ok, ot, of, op? And - even more problematically - how about aiin and aiir family words? Shouldn't they all be irreducible tokens too?
For me, the answer seems more likely to be yes than no: but having lots of carefully chosen irreducible token blocks formed of what look like sets of letters is something that is as close to verbose cipher as it is far away from everything else, linguistically.
Which would seem to lead you to the position that what you're describing is autocopying where the seeding text rules are indistinguishable from verbose cipher.
Which would rather reduce the need for autocopying at all, surely? -
RE: A key to understand the VMS
Emma May Smith > 05-02-2017, 09:47 PM
..... -
RE: A key to understand the VMS
Torsten > 05-02-2017, 10:34 PM
(05-02-2017, 05:23 PM)nickpelling Wrote: You are not allowed to view links. Register or Login to view.It's also interesting that we see all four gallows inserted in the middle of ch, but never anything else in the middle of ch. That's a rule all of its own as well, etc etc.
See for instance words similar to 'chcKhy' as grid:
chkchy ( 6) chckhy (140) checkhy (47) chckhey (30) checkhey (10) chockhy (21) chockhey ( 5) chckhdy (13) chckhedy (11)
chtchy ( 2) chcthy ( 79) checthy (28) chcthey ( 7) checthey ( 4) chocthy (18) chocthey ( 6) chcthdy ( 7) chcthedy ( 7)
chpchy ( 4) chcphy ( 11) checphy (--) chcphey ( 3) checphey ( 2) chocphy ( 3) chocphey (--) chcphdy ( 2) chcphedy ( 3)
chfchy (--) chcfhy ( 2) checfhy ( 1) chcfhey ( 1) checfhey (--) chocfhy ( 1) chocfhey (--) chcfhdy ( 1) chcfhedy (--)
shkchy ( 2) shckhy ( 60) sheckhy (30) shckhey (12) sheckhey ( 4) shockhy ( 5) shockhey ( 3) shckhdy ( 2) shckhedy ( 6)
shtchy ( 3) shcthy ( 31) shecthy (20) shcthey ( 7) shecthey ( 1) shocthy (12) shocthey ( 1) shcthdy ( 1) shcthedy ( 1)
See 'chcKhy' at the network graph of similar words for the VMS:
See for the usage of words similar to 'chckhy' on page f76r:
Why words more similar to 'chcKhy' are more frequent within the VMS and on page You are not allowed to view links. Register or Login to view. then words which are less similar? For instance 'checkhy' and 'chckhey' are more frequent then 'checkhey'. Why so many spelling variants for 'chcKhy' exists within the VMS and on page f76r?
With the autocopy hypotheses it is possible to answer this questions. Did you have an alternative hypothesis to explain this observations? -
RE: A key to understand the VMS
Torsten > 05-02-2017, 11:14 PM
(05-02-2017, 09:36 PM)nickpelling Wrote: You are not allowed to view links. Register or Login to view.
(05-02-2017, 07:49 PM)Torsten Wrote: You are not allowed to view links. Register or Login to view.No, adjacency rules are not preserved from the seeding part they are a result of the autocopying process. If a word is copied and some glyphs are replaced with similar ones the order of the glyphs stays unchanged.
Well... if your starting point is akin to that of a verbose cipher (e.g. where al ol ar or qo are all irreducible tokens, as per Curse 2006 etc), you're starting with (what to nearly everybody else's eyes looks like) a set of adjacency rules within those irreducible tokens.
And then you run into problems parsing a and o, because they appear both inside and outside your core set of irreducible tokens. What is someone to make of ok, ot, of, op?
I didn't say that they are irreducible tokens. I only say that 'al', 'ol', 'ar' and 'or' can replace each other. But this doesn't mean that they are irreducible. After the network graph of similar words for the VMS it is possible to write 'chal' without 'l'. But since most times 'a' is written as 'y' at the end of a word the result is most likely a 'chy':
cthaiin ( 13) ckhaiin ( 3) chaiin ( 45) chdaiin ( 16) chodaiin (44)
cthain ( 4) ckhain ( 1) chain ( 18) chdain ( 9) chodain ( 9)
cthan ( 2) ckhan (--) chan ( 11) chdan (---) chodan (--)
cthair (---) ckhair (--) chair ( 1) chdair ( 2) chodair ( 2)
cthar ( 20) ckhar ( 3) char ( 72) chdar ( 20) chodar (14)
cthal ( 7) ckhal ( 4) chal ( 48) chdal ( 19) chodal ( 7)
ctham ( 1) ckham ( 3) cham ( 20) chdam ( 10) chodam ( 1)
cthos ( 1) ckhos ( 3) chos ( 38) chdos ( 1) chodos (--)
cthor ( 45) ckhor ( 9) chor (219) chdor ( 8) chodor (--)
cthol ( 60) ckhol (22) chol (396) chdol ( 2) chodol ( 2)
cthy (111) ckhy (39) chy (155) chdy (150) chody (94)
cthey ( 50) ckhey (32) chey (311) chedy (501) cheody (89)
ctheey ( 13) ckheey (11) cheey (174) cheedy ( 59) cheeody (12)
Quote:And - even more problematically - how about aiin and aiir family words? Shouldn't they all be irreducible tokens too?
No. The network of similar words for the VMS shows that 'aiin' and 'aiir' are just intermediate forms between 'daiin' and 'ar':
aiiin( 41) daiin ( 17) odaiiin( 4) okaiiin( 4)
aiin (469) daiin (863) odaiin (60) okaiin (212)
ain ( 89) dain (211) odain (18) okain (144)
an ( 7) dan ( 20) odan ( 2) okan ( 5)
aiir ( 23) daiir ( 23) odaiir ( 2) okaiir ( 6)
air ( 74) dair (106) odair ( 5) okair ( 22)
ar (350) dar (318) odar (24) okar (129)
ail ( 5) dail ( 2) odail (--) okail ( 1)
al (260) dal (253) odal (13) okal (138)
am ( 88) dam ( 98) odam ( 6) okam ( 26)
os ( 29) dos ( 1) odos (--) okos ( 8)
or (363) dor ( 73) odor ( 8) okor ( 34)
ol (537) dol (117) odol ( 2) okol ( 82)
y (151) dy (270) ody (46) oky (102)
Quote:For me, the answer seems more likely to be yes than no: but having lots of carefully chosen irreducible token blocks formed of what look like sets of letters is something that is as close to verbose cipher as it is far away from everything else, linguistically.
Which would seem to lead you to the position that what you're describing is autocopying where the seeding text rules are indistinguishable from verbose cipher.
Which would rather reduce the need for autocopying at all, surely?
No, I didn't see any other way how it was possible to someone in the 15th century to write a text with a network of 6837 similar word types out 8026 word types [see You are not allowed to view links. Register or Login to view.].
-
RE: A key to understand the VMS
stellar > 05-02-2017, 11:28 PM
Torsten I have been reading your paper online. You should read my request, because it would add evidence to your study. What's missing is the affix comparison. In languages it is known that both prefix's and suffix's are uniformly distributed through out their individual corpus's.
You are not allowed to view links. Register or Login to view.
Maybe I'm missing something here or I have not found a comparison online. Or is it self evident in the Voynich community that the affix study is known so could anyone online here point to the study where it shows that affix are uniform non-uniform within the VMS? -
RE: A key to understand the VMS
Torsten > 06-02-2017, 01:12 PM
(05-02-2017, 09:47 PM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view......
I think the problem here is that we discuss about the results of the autocopy hypotheses instead of discussing the observations the autocopy is based on. Let me explain what I mean.
The autocopy hypotheses is bases on certain observations. One observation is that similar words exists for the VMS. The next observation is that similar words do co-occur. The last statement is true for Currier A and B, for the sections of the VMS, for a single page and in my eyes it is even possible to say that similar words typically occur side by side or one above the other. My next observation is that similar words are related in some way. There are certain rules how it is possible to replace one glyph or glyph sequence with a similar one. Moreover I say that it is possible to predict how frequent a word using for instance 'sh' instead of 'ch' is. This results in a more general statement that there is a tendency that a word is more frequent as more it is similar to one of the three most frequent words of the VMS 'daiin', 'ol' and 'chedy' . My next observation is that it is possible to build a grid or network of similar words in which the words are ordered by similarity and frequency. Based on the observation that it is possible to generate another word type from the word pool by replacing a glyph by a another one, or by adding or deleting a glyph I have build network graph for the VMS. In this graph for 6948 out of 8026 (86,5 %) word types at least one similar type exists (using the transcription of Takeshi Takahashi). Moreover the main network of similar word types is connecting 6837 out of 8026 (85 %) types with each other. At the end I come to the conclusion that the autocopying hypotheses is a possible explanation for this observations. (The list of steps did not contain every detail and is therefore not complete. Sorry.)
If we discuss now the autocopy hypotheses it is unclear if you refuse only the autocopy hypotheses or also the observations the hypotheses is based on. Therefore I would suggest to discuss the observations first. Only if we know where we agree or disagree for the observations we should try to discuss if the autocopy is a possible explanation for this observations. -
RE: A key to understand the VMS
nickpelling > 06-02-2017, 02:07 PM
Torsten: my hypothesis (that Voynichese combines scribal abbreviation with strong elements of verbose cipher, along with other cipher tricks) does explain in a properly forward sense pretty much all the properties that you are focused on. So your by-now-much-repeated claim that autocopying is the only explanation for these behaviours is simply not true.
I think the reason your explanation makes Emma May Smith want to ragequit is that by focusing so heavily on the connectivity of the network of Voynichese words, you seem to have lost sight of the frequency counts of nodes. You are not talking about a network where all paths have equal weights: you are talking about a network where there are some often spectacularly divergent values.
I'm sorry, but your attempted post-rationalization of why eol/ole/leo are sensible in terms of autocopying (but elo/oel/loe are not) makes no sense to me whatsover. -
RE: A key to understand the VMS
Torsten > 06-02-2017, 04:14 PM
(06-02-2017, 02:07 PM)nickpelling Wrote: You are not allowed to view links. Register or Login to view.Torsten: my hypothesis (that Voynichese combines scribal abbreviation with strong elements of verbose cipher, along with other cipher tricks) does explain in a properly forward sense pretty much all the properties that you are focused on. So your by-now-much-repeated claim that autocopying is the only explanation for these behaviours is simply not true.
First, to say only that you can explain everything is not an explanation for anything. Secondly, that the autocopying hypotheses is the only possible explanation are your words. I only say that from the observations I know I came to the conclusion that the autocopying hypotheses is a possible explanation for the VMS. BTW: Where have you published the[font=Trebuchet MS] observation that similar word types for the VMS build a network and your explanation for it?[/font]
Quote:I think the reason your explanation makes Emma May Smith want to ragequit is that by focusing so heavily on the connectivity of the network of Voynichese words, you seem to have lost sight of the frequency counts of nodes. You are not talking about a network where all paths have equal weights: you are talking about a network where there are some often spectacularly divergent values.
[font=Trebuchet MS]The frequency count of the words is a central observation in my paper [You are not allowed to view links. Register or Login to view.]. This observation was the reason for me to build the grid and the network graph. [/font]
You expect that it should be possible to calculate the exact frequencies for each word. Why did you expect that? In my eyes this would indicate a level of order in the VMS I would expect for a cipher. Therefore such a level of order would be a counter argument for the autocopy hypotheses. The autocopying hypotheses on the other hand suggests that the VMS was written manually. For someone writing the script by copying words already written there was no need to count the number of times he was writing this or that word. Therefore it is expected for the auto-copying hypothesis that the frequencies for the word types vary.
To argue with a network of equal weights means that you expect that each change has the same chance and didn't see that a human mind would have his own preferences and that even if multiple words are possible you can still only write one word at a time. A balanced multi dimensional network in all dimensions is therefore impossible. Even for a computer it would require some iterations to calculate such a network. A google search for "balanced networks" show that this is a topic on its own. Therefore you ask for something impossible to reach manually if you didn't write a text[font=Trebuchet MS] using some simple palindromes only.[/font]
Quote:I'm sorry, but your attempted post-rationalization of why eol/ole/leo are sensible in terms of autocopying (but elo/oel/loe are not) makes no sense to me whatsover.
First, please don't ask a question and then call the answer post-rationalization.
Second, the script has elements like the four gallow glyphs where for each element multiple invertible replacement exists [You are not allowed to view links. Register or Login to view.]. In this way also 'ol', 'al', 'ar', 'or' build such a cycle of four elements which can replace each other. But it is a fact that beside 'al' also 'ail' exists and beside 'ar' also 'air'. Therefore this observation alone is a counter argument to your idea of some irreducible tokens. Since multiple changes are possible the network is multidimensional. Therefore it is not easy to draw it in two-dimensions. Sometimes it helps if you reorder the nodes:
aiiin ( 41) daiiin ( 17)
aiil ( 1) daiil ( 1) aiir ( 23) daiir ( 23) aiin (469) daiin (863)
ail ( 5) dail ( 2) air ( 74) dair (106) ain ( 89) dain (211)
al (260) dal (253) ar (350) dar (318) an ( 7) dan ( 20)
ol (537) dol (117) or (363) dor ( 73) on (---) don (---)
oil (---) doil (---) oir ( 2) doir (---) oin ( 4) doin ( 4)
oiil (---) doiil (---) oiir ( 4) doiir ( 8) oiin ( 33) doiin ( 19)
oiiin (---) doiiin ( 3)
'ch' is interchangeable with 'ee' and 'ee' is interchangeable with 'eee' [You are not allowed to view links. Register or Login to view.]. Therefore 'ch' is not to far from 'che'. Since only 'chol' exists in the VMS but not the words 'ochl', 'chlo' or 'loch' the sequences 'oel', 'elo' and 'loe' are less likely for the VMS. The word 'lcho' only exists three times. Therefore a sequence 'leo' is also less likely.
The most frequent word containing 'eol' is 'cheol' (172 times). For 'eor' this most frequent word is 'cheor' (100 times). In the case of 'eal' it is 'cheal' (30 times) and in the case of 'ear' it is 'chear' (51 times). As more frequent a word is as more similar words exist for it. Moreover since 'cheol' (172 times) is frequent also the word 'Sheol' (114 times) is frequently used:
char ( 72) chear ( 51) cheear ( 1) okar (129) okear ( 7) okeear ( 2) qokar (152) qokear ( 6) qokeear ( 2)
chal ( 48) cheal ( 30) cheeal ( 2) okal (138) okeal (12) okeeal ( 1) qokal (191) qokeal ( 4) qokeeal ( 1)
cham ( 20) cheam ( 5) cheeam (--) okam ( 26) okeam ( 2) okeeam ( 2) qokam ( 25) qokeam (--) qokeeam ( 1)
chos ( 38) cheos ( 33) cheeos ( 7) okos ( 8) okeos (14) okeeos ( 6) qokos ( 1) qokeos ( 5) qokeeos ( 4)
chor (219) cheor (100) cheeor (14) okor ( 34) okeor (22) okeeor (14) qokor ( 36) qokeor (21) qokeeor (10)
chol (396) cheol (172) cheeol ( 9) okol ( 82) okeol (66) okeeol (18) qokol (104) qokeol (52) qokeeol (11)
[font=Courier New][font=Courier New]shar ( 34) shear ( 21) sheear ( 2) [/font] otar (141) otear ( 4) oteear (--) qotar ( 63) qotear ( 2) qoteear (--)[/font]
[font=Courier New]shar ( 34) shear ( 21) sheear ( 2) otal (143) oteal ( 6) oteeal ( 2) qotal ( 59) qoteal ( 2) qoteeal ( 1)[/font]
[font=Courier New][font=Courier New]sham ( 7) sheam ( 2) sheeam (--)[/font] otam ( 47) oteam (--) oteeam ( 2) qotam ( 12) qoteam (--) qoteeam (--)[/font]
[font=Courier New][font=Courier New]shos ( 10) sheos ( 17) sheeos ( 1)[/font] otos ( 4) oteos (29) oteeos (10) qotos ( 1) qoteos ( 1) qoteeos ( 3)[/font]
[font=Courier New][font=Courier New]shor ( 97) sheor ( 51) sheeor ( 9) [/font] otor ( 46) oteor (12) oteeor ( 4) qotor ( 29) qoteor ( 5) qoteeor (--)[/font]
[font=Courier New][font=Courier New]shol (186) sheol (114) sheeol (14)[/font] otol ( 86) oteol (42) oteeol ( 9) qotol ( 47) qoteol (12) qoteeol ( 5)[/font]
In this way the result fits with the autocopy hypotheses:
[font=Courier New]eol 961 eor 464 eal 140 ear 171[/font]
[font=Courier New]ole 39 ore 11 ale 10 are 7[/font]
Beside 'ol' there are 1116 words starting with 'ol-' and 3092 words ending with '-ol'. With other words 'ol' is more frequently used at the end then at the beginning. Therefore it is no surprise that also more words with 'ol' after 'e' exists then words with 'e' after 'ol'. Because 'ol' occurs more frequently at the end of a word then at it beginning the glyph sequence 'eol' is more frequent then the glyph sequence 'ole'.
Why exactly this explanation makes no sense to you? -
RE: A key to understand the VMS
nickpelling > 07-02-2017, 12:52 PM
Your explanations makes no sense to me precisely because you see the presence of "al" (3002 occurrencees) and "ail" (29 occurrences) in your network as having equal argumentative value. The former occurs 100x more often the latter: and yet you give not a single practical argument as to why there should be disparities of the order of 100x within such small edit distances of common groups.
Your autocopying argument requires the presence of a proto-network from which the full network is somehow an extended manifestation (via autocopying): yet you are silent as to what mechanism drove that proto-network. You spend your time talking about your overfeathered chicken but the matter of the missing egg from whence that chicken grew is a debate you prefer to explain away.