Why and how the text could be Bavarian

[ololololo]

A skilled cryptologist would probably be able to decrypt the text quickly.

Were there many skilled cryptologists in the Middle Ages? What is the probability of coincidence that they met together, agreed that one would write a book and the other would translate it? What is the probability that the "recipient" will actually be able to decipher it, given that he knows the algorithm?

[JoJo_Jost]

Were there many skilled cryptologists in the Middle Ages? 

A skilled cryptologist would probably be able to decrypt the text quickly.

Sorry, that sentence wasn't referring to the VMS, but to my simple encryption of the English text using vowel bigrams—and thus to the “now” time.

[ololololo]

Were there many skilled cryptologists in the Middle Ages? 

A skilled cryptologist would probably be able to decrypt the text quickly.

Sorry, that sentence wasn't referring to the VMS, but to my simple encryption of the English text using vowel bigrams—and thus to the “now” time.

Yes, I understand, but the question is about the algorithm as a whole, not this example. How easy is it to decipher, knowing the algorithm?

[JoJo_Jost]

Yes, I understand, but the question is about the algorithm as a whole, not this example. How easy is it to decipher, knowing the algorithm?

So far, it's pretty simple: Just find a space, look at the two letters to its right and left, find that bigram in the list below the text, and then replace it with the vowel next to it.

It gets more complicated later on—there are also different lists for the consonants, but I'll get to that soon. Overall, though, it’s not that easy if you don’t know the cipherer’s writing habits and dialect. That’s why I also think it’s a bad cipher. But more on that later.

[ololololo]

Yes, I understand, but the question is about the algorithm as a whole, not this example. How easy is it to decipher, knowing the algorithm?

So far, it's pretty simple: Just find a space, look at the two letters to its right and left, find that bigram in the list below the text, and then replace it with the vowel next to it.

It gets more complicated later on—there are also different lists for the consonants, but I'll get to that soon. Overall, though, it’s not that easy if you don’t know the cipherer’s writing habits and dialect. That’s why I also think it’s a bad cipher. But more on that later.

Unfortunately, handwriting and dialect are problems for us, but not for the "recipient".
And at what point does the cipher become more complicated (the same consonant lists)? Is this observed in VMS?

[JoJo_Jost]

Unfortunately, handwriting and dialect are problems for us, but not for the "recipient".
And at what point does the cipher become more complicated (the same consonant lists)? Is this observed in VMS?

But this is also true for the recipient, because with this VBM, even slight changes in spelling conventions can disrupt the flow.

It gets more complicated because of compression. The edy family and the other families are compressions of spelling conventions. Thus, nasal sounds and spellings can be—and likely are—combined. So probably n nn and m mm. Or the p and b are treated as a single consonant because they were “interchangeable” at the time. The v and w—the same goes for t and dt.

The consonants groups: “s,” “z,” “ss,” “sz,” “cz”—all are apparently used “indiscriminately,” and to make things really complicated, so are “ts,” “z,” “cz,” and “tz.” Also “k,” “c,” “ck,”. The final sounds: “t,” “d,” “dt” in the same text. 

And of course there are also problems with the vowels. The “j” and “y” are often written as an “i,” the “u” as “v,” (and the "f" as "v"

And it’s not just consonants that are like this. In Bavarian, the “a” and “o” are sometimes indistinguishable. The “uo,” “ue,” “u,” etc., and so on.

On top of that, there are oddities like: “und” (and)
This is written as: und, vng, unde, vnde, vnnd, vnn, vn, vndd, vndt, unnd, unn, vnnde, un, and others.
Sometimes a writer even uses multiple versions in a single text.

This doesn’t make things any easier—it makes them extremely complicated. Especially with a cipher based on the flow of vowels and consonants.

So why do I consider a cipher of this kind to be weak? Because the “inventor” probably didn’t think that through, but maybe it was also only intended for a very small circle of persons... who knows.

[DG97EEB]

I just came across this.. JoJo: Do you know this project? Worth emailing them?

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

[JoJo_Jost]

Hi Ed

If I understand correctly, they haven't even written the manual yet. Do they refer to the VMS in their studies? Since it hasn't been cracked, probably just as a side note. 

[JoJo_Jost]

The Matrix 1.1 Part 1

Continuation of the post: You are not allowed to view links. Register or Login to view. 

Well, the creator of the cipher has now made the vowels “unrecognizable.” The compromising “e” is spread across two vowel bridges and is thus quite well concealed. Now he has to conceal the consonants—especially “d,” the most common consonant in German, and “n.”

But let’s first take a look at what’s left when we remove the vowels.

Step 1: Remove all spaces
Step 2: Replace all vowels with a space.

We'll do this now using a German text, because everything that follows in the next posts doesn't work in other languages. I've chosen a passage from the “Breslauer Arzneibuch” (standardized (v/u)) for this purpose:

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

b nd k ld l ftw ls h  ns g tv rd ns  b nh st ns lm nl z n nd r d rd d h  z tb s l  t rr m nd sn chtt ns s lm nscr pf n nd r t td nn nd v cht r nn td d rh st  n std rn chs lm n mg b nz tr nk n  ns r p  n s p  nd  n    b s nd s l mg b ng rst nw zz rm tm nd l l  nd m tp nn ln nd    lnh n g  nd m tz ck r d rg p ms r p  nm rgr m d rk t nk rn  nd p pp ls fm tz ck r nd m tp nn ln nd g p m  ss lm tg rst n  cl  ng p mz  zz n p pp lp z n nd c rb zg s t n nw zz rm tm nd l l d z g p m  ng h n r bh s w ls lb  md br  st z n m tw chs  nd m t  ln l d z l z ns n nd r  n nd r nd m sch d rz  rst nm l nd p rz lns f nd c rb zs fd rs  h m c  chw l zz nw ch    r  nd v rb s chm tc ss  flst l g p m  n l ct  r  d m ch  ls n mg mm  r b c m nd w z ntr g nt sl chsz h ndr gm n mw z nm  h n nd p n ln sl chszw nz cdr gm n md zz m nd lk rn d nd h t b g zc ns c h ndr gm n m m d n d k t ns m n nd p rz lns m n sl chsv nfdr gm n mc rb zs m n nd  rd pf ls m n nd pf d m ns m n nd k mb rs m n nd l t ch ns m n sl chss bcndr gm d zst z ll z nd r  n nd r nd gr zdr ns r p  n  ln ls l  zs nb d rfz  n rl ct  r  n nd m sch  z nd r  n nd rh rt s r sw rd nh st nh t  nd rbr  sts chtd d  st  nk ld  nd  nv cht d sw w  stn chtgr z nd t n mw rm d ncw ld  ns lt  mg b nr s nh n cm tw zz rd  nch l nd  pp  nn g s t ns  nd d rn chg s g nd rch  nt chd s lt m  chg b n  n  ss ld z nk  rn g s t ns m tm nd l l g p mz  zz n g nz pf l k rn  nd  g ng p m  ss l  nc  rnm tw nb rnd d d  rr s nt nd g p mf n gr c mg s t nm t  g n nd m tc t lns lb  md br  stm tw chs  nd m t l  ns mb c d  sc r d sspr ch td z  g nm t s p g s t ng ts n  rd nk ld nh st nd v rs n  chg tm r s m n rs d t ls gr z ls  nb n s d rh st  st  nd rv cht d d r nn t zd m h  b t  d r  n nd rnl d nd  nd br  stg ns g pd m p ch n  ns r p d nm ch  ls n mm m b  mzv nz cdr gm n m pp  nd v nch lw rz lc l m nt  nd  s p  nd c p ll m  n r s sl chsv nfdr gm n mf n gr c m nd l n  t sl chsdr dr gm n md  rr w nb rd n chtk rn  nn s h nd rtdr gm n m  g nzv nz cdr gm n mp nn lns chs nc n nd h n c nd m ch d  n  ns r pd gr z c r  c  st  chg tv rd nh st n nd p l  r  nw r  ch m th n g  nd  nch ls fm tm nd l l g p mz  zz n  s n nd r ph nr m t  m p l  r  nk m l  nd  npf fl rm tr t n nd pf rr n nd h nt g nm nd l l d zh rc  t r t tw nn v nd rv cht d s chg s m n th t nd nb lg lnd  mm d zh rc g nd rs cht mk m td g s nd n ls g h s n nd  chd s  b ns d v cht  l z t nd zh rc  nd  std v cht g m sch t  nbl t  nd  nc l r s c t r td zh rc  nd  std m m nn h  z nd d rst t n nd z h td n d m f  nd rt f  st b rd v cht g m sch t  ngr z nbl t  nd  nd rm l nc l  ns  t r td zh rc  nd  std rm ntr r c nd tr g  nd  n d rstd m s  b nh rc ns lm nb c t z h lf k m nw nn  z std zb st l td z nd m  b  sts d rh rc w w  nd mbl t  d r  nd rc l r s l  nd r d rd d h  z tb s l c  bd c t nd d  g ntd z ld r nd d  mm  htn cht n rr tg pd m s  b n  n ntr nc  n  ss  flst h  n    b spr n s x f n    nd  n  ln nd  nb z  s f nd  nm rgr ms fg p mz  zz n g  z n l  s h nd h n r zm rgr ms f d rh s l  chn m nr s nw zz rm tk nf rs d rs cht m  nd rm l nc l  ns v rb d ns  b nm td r rc n  d d h  z t  r g l  n f rt ss m g p md z l ct  r  md  m rg r t n nd h  z  n b rb d n nb d d zn chtz h  zs n chz k ltd rn chg p md gr z nt r  c m nd s lb  md br  stm t l  nt ll  d rm t  m  nd rn l d zw lsm ck  nd w rms  tw nn g sch ts cht m  nd mg br st nd rd  st nd r nc t ng d d l b rd m h rc ns nd td  ns lm nd s rst nd l b rb zz mm t rc n  lsw rh rn chs g n nd m c p t l  nd rl b r  d chw nn d rs lb g br st  ld ck  rg t nd rl b r  nd sm g n nt g ntd  ns lt d nm g nv rb nm tsp  nm t rc n  mm chtg sch t  nd m h r  ns  zd nd nstd n z n mh tn cht zm cg tr b nd z b r zd sn chtm cg t nd z rg t  ngr z rv    d r  ngr z rl r  d r  nw nd l ng d g sch t  nk ld  d r  nh tz d v    st ntw d r  nd rv cht d d w n t n ld m  b  d r ll  n  nd m m g n d r nd nd rm n bd v ll s  n ld m l b d zm rk b d rm  nd b d rsv r d d rl ph t  b r l st b rd v llc nd mm g nd zm rk b d m r pz n nd b d rsp  nd b d mw nb r nd m h  b t gr z l r g sch t  n mm zl ch rtr  b  nd  n mm zl ch rsp    nd  nd mbl t d zd  l z td rchd n s  nd  n mm z g m sv  z  nd  nd nr nn nd n  cbl t m nd d nw b n  r rn t rl ch ns cht mv nsv lch m nd rd n ll nd zg sch td zb sv r td nl p nd kr nk td l t nd  rw nd l td  rw  nd  rl sch tn t rl ch h tz d  nm z  chd zh rc  mm cht cw rd nd w nd l ng d g sch t  nh tz  nk ld  rst pf td sv  zl ch nd  rsp rr t nn rh lp nd m l b d nd nstd  nw rtd zh rc  mm cht cg w nd l ng

You’ll notice that there are mainly one / two consonants left here, occasionally three, and rarely four/five

Interestingly, the top 160 most frequent consonant clusters in the Breslauer account for 92% of the entire text.

"Breslauer Arzneibuch":
Total number of German clusters:  approx. 134,317
Occurrences in the top 160:          approx. 124,077
Coverage by the top 160:              approx. 92.38%
Remaining occurrences:                approx. 7.62%

So, in order to cover these consonant clusters with the VMS kernels (without the vowel bridges), the VMS kernels would need to contain sufficient information.

To do this, let's take a look at the E-family.

In the VMS, there is the “e” family which has roughly 10 stems:
e
ee
eee
eo
eeo
ed
eed
eod
eeod
eeed

There are many VMS glyphs before these "e family"; let's take for example the 16 most common ones:

k
t
ch
sh
kch
pch
lk
tch
lch
ckh
lsh
ksh
cth
tsh
fch
psh

That are 160 possible combinations. 

An theoretical exampel:

   

Fill the cells with German clusters and you will get a Matrix

(But the actual VMS cells are not uniformly filled. In the hard bridge matrix, 111 of the 160 possible cells are occupied and 49 are empty. And, of course, it’s important to note that in Currier A/B, the bigrams “od” and “ed” are almost completely interchanged, and other effects —from this perspective, the “e” family shrinks even further. - but it's only an example)

Then there are all these other families—the aiin family, the air family, etc.—that can take on additional roles. 

And, of course, there’s also the rest of the more unstructured glyph distributions in the VMS (I’m currently assuming that these remaining glyphs are simply substituted. And that the labels, for example, are just consonants without vowels—which explains why they tend to be a bit short).

So generally speaking, there is enough information to encode these leftover consonant clusters. That is an important insight.


More on that later....

[JoJo_Jost]

Okay, we’re dealing here with one of the strangest statistical anomalies in the text:

   

The graph shows how often parts of the E-family in the left column force a “qo” when preceded by a t, k, ch, or sh.

After “shedy,” for example, “qo” follows in 49% of cases, while after “tedy,” “qo” follows in only 25% of cases.

This shows that the letter preceding the “E-family” influences the frequency of the following “qo.” The glyphs preceding the “E-family” seem to have a sort of long-range effect.

The Vowel Bridge Model offers a simple explanation for this oddity

1. If the “edy” family resolves consonant clusters and “y[space]qo” is an “e,” then it’s actually quite simple. There are certain consonant clusters that simply attract an “e” more often, and others that do not attract an “e.”

An example from the qokedy Crib suggests that “ked” is an “nd.” “nd” is one of the most common consonant clusters in the VMS.

"ked = nd / y[space]qo = e  = nde" 

In german the consonant cluster “nd” is often followed by an “e”

But there are letters in German that are much more likely to be followed by an “e”—for example, the letter “g.” So if “shed” were a “g,” it would only be logical that “shed” would be followed much more frequently by “y[space]qo,” i.e., an “e.”
k   edy = 27%
sh edy = 49%

(That's just one example; unfortunately, the actual frequencies depend heavily on the spelling conventions of the time, so all we can really say here is that “ge” is much more common (in most texts) than “nde.”)

This strange “remote effect” thus boils down to a direct interaction between consonants and vowels. And that makes it very easy to explain—it’s even a typical phenomenon in a normal language.

And with that, the VBM can easily explain yet another statistical phenomenon of the VMS.  

But it also explains, almost as an aside, what happens before “ked”—that is, why “qo” is so drawn to “ked.”

Treating “(y) qo” as part of the “e” vowel bridge would turn “ked” = “nd” into “(y) qoked” = “end.” And “end” is also a very common glyph cluster in German.

We have now been able to explain a number of VBM phenomena simply by revealing that the spaces in the VMS are part of the vowel cipher (y[space]qo) -

- The uniform, almost static word distribution
- The limited word lengths
- The almost repetitive occurrences of similar words
- Words that differ by only one letter
- Repetitive phrases
- The text’s highly structured nature.
- And the many hapaxes, which contradict any normal linguistic structure

Further investigations suggest that the kerns, without the vowel bridge segments of the VMS, could encode consonant clusters, thereby quite simply explaining additional statistical peculiarities.

- Strange long-range effect of shedy / tedy
- qo -> ked

That’s quite a lot that suddenly makes sense if you simply make two (!) basic assumptions that are also related to each other: The glyphs before and after a space are vowels, so the remaining cores (especially the “E” family) encode consonants.

According to Ockham’s Razor, for that reason alone, this VBM must be considered a very strong candidate that may be capable of solving the VMS.