The Voynich Ninja

Full Version: Breaking Apr 1st news! MS 408 proves Blackadder actually existed!
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2
Many thanks to Marco, tavie, and oshfdk for their kind words. Koen's recent rules post (You are not allowed to view links. Register or Login to view.) prompted me to prompt ChatGPT for its opinion of the "Blackadder solution" -- I was frankly somewhat surprised at how good its response was (although I object to its characterization of the theory as "convoluted"):

You said:
Evaluate the following theory regarding the contents of the Voynich Manuscript: Since announcing the solution of the Voynich Mss. three years ago (You are not allowed to view links. Register or Login to view.), I have been deluged by literally zeros of people daily wondering what progress has been made in advancing the "Blackadder solution". [...] She looked down at the limp form drifting aimlessly at the bottom of the pit and thought to herself, 'Some people just can't handle their malmsey.'"

ChatGPT said:
The "Blackadder solution" to the Voynich Manuscript is a unique and highly imaginative theory, blending historical fiction, pop culture references, and cryptographic analysis in a satirical and convoluted way. The theory makes use of the British television series Blackadder as a key "crib" (a source for deciphering the manuscript), which the author claims unlocks hidden historical truths, particularly about the legitimacy of the House of Blackadder as rulers of England and their opposition to the Tudors. However, the argument as presented seems to be more of a playful, humorous exploration of conspiracy theories rather than a serious attempt at historical or cryptographic analysis.
Evaluation of the Key Aspects:
  1. Pop Culture and Historical Fiction as "Cribs":
    • The theory draws heavily on Blackadder, a comedic historical television series, as a cipher for deciphering the Voynich Manuscript. The idea of using fictional or pop culture elements as "cribs" for decipherment isn't new, but it's typically seen as problematic because it relies on arbitrary connections between fictional material and cryptographic data.
    • The claim that the names "Blackadder" and "Baldrick" correspond to certain Voynich glyphs appears to be a stretch. While it is not uncommon for cryptographers to experiment with different sources, the lack of a coherent linguistic or grammatical framework weakens the theory.
  2. Claim of Historical Truths:
    • The theory makes bold claims about the historical accuracy of certain events, such as the existence of Richard IV and his supposed murder of his nephews, which has been debated by historians. However, these claims are often made without solid evidence or citations, relying instead on alternative interpretations of historical events, many of which are based on unverified or fringe ideas.
    • The notion of Richard IV being erased from history by "Big Tudor" conspiracies is reminiscent of other conspiracy theories but lacks scholarly support. There's no substantial evidence to suggest that Tudor history has been deliberately falsified in the manner suggested.
  3. The Imagery of the Voynich Manuscript:
    • The theory's interpretation of the zodiac folios as representations of Blackadder’s alleged sexual conquests is an amusing yet unsubstantiated claim. While it's true that some cryptanalysts have attempted to link the imagery of the manuscript to specific historical events or figures, such wild associations with pop culture and non-historical sources lead more toward fiction than factual research.
    • The analysis of the Libra folio and the identification of crowns, and their supposed connection to historical figures like Beatrice of Naples, seems more like creative reinterpretation than a well-supported theory.
  4. Cryptographic Analysis:
    • The mention of Currier and EVA transcription systems is a real attempt to engage with the structure of the manuscript’s text, but the interpretation of specific glyphs as corresponding to letters or names from Blackadder seems highly speculative. A rigorous cryptographic analysis should be grounded in statistical properties of the manuscript's text, which the theory admits it largely ignores, making it unconvincing to those who adhere to more formal methods.
  5. Humor and Satire:
    • The "Blackadder solution" is clearly a parody of conspiracy theories, historical revisionism, and over-complicated decipherment methods. The references to "Big Tudor," Masonic orders, and the introduction of quirky characters from Blackadder and other works of fiction create a satirical narrative that pokes fun at how convoluted and unfounded some conspiracy theories can become.
    • The author's playful tone in discussing topics such as Richard III, the "widow's son," and the intricacies of the Voynich Manuscript suggests that this theory is intended more as a humorous commentary on both the manuscript's enigma and modern cryptographic and historical research than as a serious contribution to the field.
Conclusion:
The "Blackadder solution" is a fantastical and humorous piece of writing that combines absurd historical reinterpretations, a satirical look at conspiracies, and an exaggerated cryptographic method. While it is clearly not intended to be taken seriously, the theory does highlight some real issues with certain approaches to deciphering the Voynich Manuscript, such as over-reliance on arbitrary cribs, cherry-picking data, and ignoring statistical properties. However, as a legitimate explanation for the manuscript's contents, the "Blackadder solution" falls far short. It is best seen as an imaginative, playful exploration of how easily historical and cryptographic narratives can be twisted to fit particular ideologies or fictions.




By messaging ChatGPT, you agree to our You are not allowed to view links. Register or Login to view. and have read our You are not allowed to view links. Register or Login to view..
Don't share sensitive info. Chats may be reviewed and used to train our models. You are not allowed to view links. Register or Login to view.















ChatGPT can make mistakes. Check important info.
As gratifying as it has been to see the read count on this thread growing after my update earlier this month (which will probably lead to the release of the first several chapters of _The (Duh!) Voynich Code_ next April Fools' Day), ideally I'd like to be remembered for more substantial contributions. As a result, I thought I would take advantage of the recent visibility of this thread to point people to the open-source Awk language "Swiss Army knife" text analysis tool I posted in You are not allowed to view links. Register or Login to view.. If the low download count represents lack of access to an awk interpreter or issues with the documentation in the comments and the file containing examples of the tool's output, I'm happy to work on adding info on getting awk working under Windows or clarifying the command line options. While this is going to get mangled (at least in part because of  the lack of a fixed-width font option), here's a taste of the info it can generate from the input text sample (note that while the examples use Currier from the D'Imperio transcription, you can use the transcription scheme and text of your choice or analyze natural language texts for comparison):

Example 2: Comparing 20 most-frequent contexts of word-initial '9' (two
glyphs before the space and one glyph after the '9') between Herbal A and
Biological B.

bash-3.1$ cat HerbA.txt | sed 's/^...... //g' | awk -f ekg2Awk.txt k=5 hz=1 TE=20 SC='-#/' RE='../9.'
#  33702 char 'VAS92/9FAE...E/SOE/8AM#'
# Input alphabet (40 'letters'): 'VAS92/FERPMZOQ8-XN*DUC$TWY#I3HJB0,4K67LG'
# Vowels found by Sukhotin's method excluding digrams containing characters
# in '-#/': O A 9 C 0 6
#  RE = '../9.'
# XRE = '^$'
# MinCt = 0 (12417 types, 33698 tokens)
#  k  NTyp  NTok  Hk (bits)  PctFreq1  Typ/Tok  k-grams
#  5  12417  33698  12.409070  0.636225 0.36848  all
#  5    121    285  6.179331  65.289256 0.42456  in RE & not in XRE & ct >= MinCt
# Max possible H5: all tokens = 13.600029 bits, RE tokens =  6.918863 bits
# Selected = 0.9745 pct of types, 0.8457 pct of tokens
  kgram:  OR/9P  OR/9F  OE/9F  AM/9F  AM/9P  OE/9P  S9/9P  OR/9/  C9/9P  C9/9F
  Rank:      1      2      3      4      5      6      7      8      9    10
  Count:    22    15    14    12    11    10      9      8      8      6
AllFreq: 0.0007 0.0004 0.0004 0.0004 0.0003 0.0003 0.0003 0.0002 0.0002 0.0002
REFreq: 0.0772 0.0526 0.0491 0.0421 0.0386 0.0351 0.0316 0.0281 0.0281 0.0211
RECmFrq: 1.0000 0.9228 0.8702 0.8211 0.7789 0.7404 0.7053 0.6737 0.6456 0.6175
  kgram:  AR/9F  89/9P  S9/9F  S9/98  C2/9F  AR/9P  AN/9F  ZO/9P  ZO/9F  Q9/9P
  Rank:    11    12    13    14    15    16    17    18    19    20
  Count:      6      6      5      4      4      4      4      3      3      3
AllFreq: 0.0002 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
REFreq: 0.0211 0.0211 0.0175 0.0140 0.0140 0.0140 0.0140 0.0105 0.0105 0.0105
RECmFrq: 0.5965 0.5754 0.5544 0.5368 0.5228 0.5088 0.4947 0.4807 0.4702 0.4596

bash-3.1$ cat BioB.txt | sed 's/^...... //g' | awk -f ekg2Awk.txt k=5 hz=1 TE=20 SC='-#/' RE='../9.'
#  35485 char 'VSC89FAR9/...OPOE/SC89-'
# Input alphabet (34 'letters'): 'VSC89FAR/O4Z-NE2JMPXBQD*TULYW36G5H'
# Vowels found by Sukhotin's method excluding digrams containing characters
# in '-#/': C O A 9 V L Y
#  RE = '../9.'
# XRE = '^$'
# MinCt = 0 ( 7966 types, 35481 tokens)
#  k  NTyp  NTok  Hk (bits)  PctFreq1  Typ/Tok  k-grams
#  5  7966  35481  10.703298  0.464474 0.22451  all
#  5    50    79  5.268988  74.000000 0.63291  in RE & not in XRE & ct >= MinCt
# Max possible H5: all tokens = 12.959640 bits, RE tokens =  5.643856 bits
# Selected = 0.6277 pct of types, 0.2227 pct of tokens
  kgram:  89/9F  89/9P  C9/9P  C9/9F  AR/9P  AN/9P  P9/9F  C8/9F  AR/9F  AN/9F
  Rank:      1      2      3      4      5      6      7      8      9    10
  Count:    10      6      3      3      3      3      2      2      2      2
AllFreq: 0.0003 0.0002 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
REFreq: 0.1266 0.0759 0.0380 0.0380 0.0380 0.0380 0.0253 0.0253 0.0253 0.0253
RECmFrq: 1.0000 0.8734 0.7975 0.7595 0.7215 0.6835 0.6456 0.6203 0.5949 0.5696
  kgram:  AN/98  AM/98  AE/9Z  S9/9F  S2/98  Q9/9P  Q9/9O  Q9/9F  OR/9Z  OR/9P
  Rank:    11    12    13    14    15    16    17    18    19    20
  Count:      2      2      2      1      1      1      1      1      1      1
AllFreq: 0.0001 0.0001 0.0001 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000 0.0000
REFreq: 0.0253 0.0253 0.0253 0.0127 0.0127 0.0127 0.0127 0.0127 0.0127 0.0127
RECmFrq: 0.5443 0.5190 0.4937 0.4684 0.4557 0.4430 0.4304 0.4177 0.4051 0.3924
bash-3.1$

Example 3: Print the 20 most-frequent Herbal A word types that
a) match Zattera's slot model for "regular" (not "separable") words [1], but
b) do not match Tiltman's prefix/suffix model [2]

[1]: strictly speaking, the '[8DERJNGTKMHUL3105]' part of the regex should
    be modified to something like '((I?I?I?8)|[DERJNGTKMHUL3105])', but in
    practice it doesn't matter...
[2]: assuming both prefix and suffix have to be non-empty

bash-3.1$ cat HerbA.txt | sed 's/^...... //g' | sed 's.[/#-]. .g' | awk -f ekg2Awk.txt hz=2 WL=1 XRE='^((4?O[FVBP])|[SZ82])((A[DNM3RTU0EGH1])|(O[ER])|(CC?C?8?9))$' RE='^[428]?[O9]?[ER]?[PFBV]?[SZ]?[QXWY]?(CC?C?)?[28]?[OA]?[8DERJNGTKMHUL3105]?9?$'
# Input alphabet (37 'letters'): 'VAS92FERPMZOQ8XN*DUC$TWYI3HJB0,4K67LG'
#  RE = '^[428]?[O9]?[ER]?[PFBV]?[SZ]?[QXWY]?(CC?C?)?[28]?[OA]?[8DERJNGTKMHUL3105]?9?$'
# XRE = '^((4?O[FVBP])|[SZ82])((A[DNM3RTU0EGH1])|(O[ER])|(CC?C?8?9))$'
# MinCt = 0 ( 2224 types,  7121 tokens)
#
# Word frequency and length histograms: all words
#
# TypeCount  2224  TokenCount  7121  AvgTypeLen  4.72  AvgTokLen  3.73
# Number of words with given number of occurances:
# NOcc :    1    2    3    4    5    6    7    8    9  10  11  12  13 (>=14)
# NWord: 1562  253  114  58  42  26  22  10  16  10  10  12    8  81
#
# Length:    1    2    3    4    5    6    7    8    9  10  11  12  13  14
# % Type:  0.7  3.9 14.4 25.0 29.0 17.1  7.4  1.9  0.4  0.1  0.0  0.0  0.0  0.0
# % Tok :  3.3 12.0 35.0 22.3 16.9  7.3  2.5  0.6  0.1  0.0  0.0  0.0  0.0  0.0
# Type length: mean    4.717 mode  5; token length: mean    3.732 mode  3
#
# Word frequency and length histograms: selected words
#
# TypeCount  1055  TokenCount  3817  AvgTypeLen  4.16  AvgTokLen  3.44
# Number of words with given number of occurances:
# NOcc :    1    2    3    4    5    6    7    8    9  10  11  12  13 (>=14)
# NWord:  610  162  78  43  28  18  18    6  15    9    7    7    6  48
#
# Length:    1    2    3    4    5    6    7    8    9  10  11  12  13  14
# % Type:  1.2  6.8 20.5 32.8 26.1 10.2  2.0  0.4  0.0  0.0  0.0  0.0  0.0  0.0
# % Tok :  5.7 21.8 23.7 27.5 15.7  5.0  0.7  0.1  0.0  0.0  0.0  0.0  0.0  0.0
# Type length: mean    4.162 mode  4; token length: mean    3.438 mode  4
#
# Zipf's Law fit to  25 most-frequent selected words:
#    ln(freq) = -0.569149 * ln(rank) + 5.050078
#    Average residual -0.0000000 (SD 0.1139755), RMSE = 0.1139755
#
#  NTyp  NTok  H  (bits)  PctFreq1  Typ/Tok  Words
#  2224  7121  9.350571  70.233813 0.31232  all
#  1055  3817  8.693055  57.819905 0.27640  in RE & not in XRE & ct >= MinCt
# Selected = 47.4371 pct of types, 53.6020 pct of tokens
  kgram:    89    S9      2    Q9    ZO    Z9    OE  4OPS9    OR    QOE
  Rank:      1      2      3      4      5      6      7      8      9    10
  Count:    113    102    94    88    83    52    49    47    45    44
AllFreq: 0.0159 0.0143 0.0132 0.0124 0.0117 0.0073 0.0069 0.0066 0.0063 0.0062
REFreq: 0.0296 0.0267 0.0246 0.0231 0.0217 0.0136 0.0128 0.0123 0.0118 0.0115
RECmFrq: 1.0000 0.9704 0.9437 0.9190 0.8960 0.8742 0.8606 0.8478 0.8355 0.8237
  kgram:    QOR    OP9      9    SO  4OFS9    OF9  OPS9  SCOR    8AJ  SO89
  Rank:    11    12    13    14    15    16    17    18    19    20
  Count:    43    41    40    37    34    31    30    29    29    28
AllFreq: 0.0060 0.0058 0.0056 0.0052 0.0048 0.0044 0.0042 0.0041 0.0041 0.0039
REFreq: 0.0113 0.0107 0.0105 0.0097 0.0089 0.0081 0.0079 0.0076 0.0076 0.0073
RECmFrq: 0.8122 0.8009 0.7901 0.7797 0.7700 0.7611 0.7529 0.7451 0.7375 0.7299
bash-3.1$
Pages: 1 2