• Options
• Login
• Desktop Style

Calgary engineer believes he's cracked the mysterious Voynich Manuscript

• Index

  • The Voynich Ninja
  • Voynich Research
  • Analysis of the text
  • Calgary engineer believes he's cracked the mysterious Voynich Manuscript
• Thread Closed
• Calgary engineer believes he's cracked the mysterious Voynich Manuscript

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  joben > 13-06-2024, 08:14 PM

  (13-06-2024, 05:42 PM)Ahmet Ardıç Wrote: You are not allowed to view links. Register or Login to view.1- In my study, the machine only looks at the data at hand and classifies them in accordance with the command, summarizes them, and compares them with each other in different details.

  An LLM can hallucinate and give false answers. It happens all the time. Maybe in the future an AI will be more useful for VMS related tasks.
  

  (13-06-2024, 05:42 PM)Ahmet Ardıç Wrote: You are not allowed to view links. Register or Login to view.2- Machines do not become dishonest, prejudiced, hypocritical, double-standarded, biased, nationalist, or racist.
  For this reason, in the near future, the works performed by machines will be able to interpret some of the old academic works in a new and more accurate way, which will surprise humanity. Because, AI is more consistent in distinguishing between "works rich in personal opinion but inconsistent in evidence and facts" presented in many "scientific" articles and studies, especially on history and linguistics.

  This is false. At least the current LLMs are not that impressive when it comes to giving non-biased answers. It is very easy to see that ChatGPT for example has many filters for certain questions that you ask. ChatGPT is also biased if you ask it to compare political leaders or extreme ideologies, or make jokes about them. ChatGPT currently has a noticable left wing bias. It has gotten better, but it is still noticable. You also have to take into account that ChatGPT has been trained on texts created by humans. Garbage in, garbage out. What looks good in theory about a rational machine that only looks at hard facts is simply not what we are seeing today.
  
  

  (13-06-2024, 05:42 PM)Ahmet Ardıç Wrote: You are not allowed to view links. Register or Login to view.For this reason, I felt the need to make a new addition to a post I made here before about the founders of artificial intelligence and understanding its capacity. That's why I wrote this note. Scientists have been oppressed in various ways by the ignorant and the ruling power in every period of history. The founders of artificial intelligence also experienced difficulties due to their lifestyles or choices. I think it would be disrespectful not to note situations such as discrimination or racism against them while talking about this issue. We humans cannot move forward by sweeping problems under the carpet or ignoring them. When necessary, it is necessary to remind the racist that he is a racist and the discriminator that he is a discriminator, and I do this from time to time and I can do this on every platform. If we cannot write about racism and discrimination here when appropriate, the admin should close my page and terminate my membership/following. But do not try to lecture me or guide me on matters that are not your responsibility. This is none of your business.

  You are free to express yourself as long as the moderators allow it. I am also free to express myself that I think what you just wrote is ridiculous and off-topic. It is also my belief that you specifically attack historic UK bigotry while ignoring the motherlode of even worse and currently existing bigotry originating from other places in the world, mainly outside of Europe. "Strain at a gnat and swallow a camel", as the proverb goes. Sorry, I just don't find this genuine at all. When someone resorts to singing "Kumbaya" instead of discussing the actual topic, I lose my interest. Good luck with your theory.

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Koen G > 13-06-2024, 08:24 PM

  I have noticed before that discussing politics on forums leads to noting but unpleasantness. And besides, it is off topic in this thread. Let's keep it on the subject of Ahmet's theory, please.

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Ahmet Ardıç > 13-06-2024, 11:53 PM

  Dear Joben,
  
  I will not respond to your comment, which has a part that may have a political aspect.
  
  I read your comments about ChatGPT. I partially agree with these. But let me give a clear example here without further ado. I have been talking about some structural and phonetic features of Turkish here for months (in fact, the first time was more than a year ago). I wrote these features here by showing evidence, academic articles, and examples. Later, I also showed that 1/1 of the same structural and phonetic structure exists in VM texts. I presented this at an international symposium, where the scientific committee consisted almost entirely of Old Turkish expert professors, and it was published as an article.
  
  For months, I have been asking researchers and linguists here how these structural overlaps should be evaluated from a scientific perspective and what these findings and overlaps actually mean. None of the answers from people to this question were answered with scientific understanding. I do not prefer to communicate with the dialogue jargon in the bar or on the street. The question I am asking is a very linguistics-oriented question and is based on very clear and concrete findings. People's unscientific and vague answers or asking new questions that have nothing to do with the questions I asked are forms of communication that I do not prefer. I have stated this many times.
  
  I asked ChatGP the same question. I asked GPT to read my VM-related articles and summarize them. Then I asked the machine to look at the sources for the evidence I presented in my articles too. In this sense, I only asked the GPT to examine academic articles and academic criticisms of academic articles. Then I asked him to list all the other articles that said VM-texts are natural languages and read them. Then, I ordered the GPT to find itself the related pieces of evidence presented in those articles to read all of them too. Then, I asked GPT to compare all the articles claiming that VM-texts are natural languages, in terms of the rationality of the evidence, consistency, and whether the overlaps are concrete or not. Then, I asked GPT to express its judgment with answering the question "The VM-texts are more likely to be which language?"
  
  I also wrote the GPT answers to all these here & share. Many people are reading the posts I make on these pages. They are at liberty not to answer my question about what the two structural and phonetic overlaps I have presented mean. People may have different reasons for not answering. I don't know the reasons. For example, they may say; "I cannot answer your question because I do not know Turkish or I am not a linguist." There may be reasons for not answering, or trying to answer a question with a question or trying to divert the discussion from its axis by writing nonsense and irrelevant things to the main question and topic.
  
  So, This is what the GPT machine does not do. When you ask the machine a question, it answers in a particular frame. The machine that answers my simple question does not depend on any of the elements you have listed here as "the current weaknesses of ChatGPT". No excuses or weaknesses play a role in his answers to a simple question. My questions are simple reading and comparing content questions for ChatGPT. There is no detail in my questions that would confuse the ChatGPT.
  
  In fact, while there are many scientists and academicians reading these pages; If the situation of not being able to say "what it actually means to understand that two clear structural and phonetic matches are compatible with a single language" continues, I can say that I will close my membership account on this page and leave soon. People with an academic identity should be able to open and read the evidence presented and express their opinions within the scientific framework, even if they do not know Turkish. I don't think it would make any sense for me to be in an environment where this cannot be done. Because I can say that continuing like this is quite annoying, apart from being a waste of time on my part. I indicated two structural overlaps and asked some simple questions. I will wait for the answers for a while and then can conclude that not be interesting to be in a group where researchers cannot interpret this simple data. At that moment, I could leave this group myself without waiting for the admin, and I think I am very close to this point.

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Ahmet Ardıç > 14-06-2024, 07:52 PM

  Dear Researchers,
  
  If some readers/researchers perceive that VM content is more likely to be in Turkish and want to support our research, I can ask them for help on some issues. I would like to thank in advance everyone who may wish to contribute to our research voluntarily.
  
  I would like to explain to potential volunteers how they can contribute to our research even if they do not know Turkish. In addition, researchers who may volunteer in this research can also write and publish their articles based on their findings or overlaps.
  
  Researchers who may volunteer should note the following information:
  
  1- To compare the words in VM content, they can find academic articles or dictionaries containing information such as Old Turkish plant names, star names, animal names, and drug names.
  
  Here, I would like to share with you the link to an article that I have just seen for the first time, as an example. I also share a dictionary page. If anyone wants to compare the plant names in this article using VM texts and ATA alphabet transcription, they will contribute to this research:
  
  Article name: THE PROBLEM OF PLANT NAMES’ LATIN SCIENTIFIC EQUIVALENTS IN OLD ANATOLIAN TURKISH MEDICAL MANUSCRIPT STUDIES
  Author: Anıl ÇELİK
  You are not allowed to view links. Register or Login to view.
  &
  Historical Dictionary of Ottoman Turkish Terms For Gardens and Gardening:
  You are not allowed to view links. Register or Login to view.
  
  Additionally; People who know old nomenclatures (plant names, toponyms such as old names of castles and cities, etc.) in European languages, especially Old Latin and Old Greek, can also search for these words on the VM plant pages and map pages.
  
  2- You can take into account the information about the sound differences between the most common dialects in Turkish in your word comparisons. For example, if the plant name is PIRASA and you know that the "I" and "U" sound changes frequently between dialects, you can also search for this word as PURASA.
  
  For this reason, you should know that the most common sound changes among dialects in Turkish are as follows:
  I >  U / İ > Ü
  B > P / P > F
  O > A / Ö > E
  G > K / G > Ğ
  D > T
  S > Z
  C > Ç / C > J
  T > K / D > G
  L > N
  B >M
  F > V
  
  You may also be interested in the following pages where you can get information about the phonetic and phonological structure of Turkish:
  
  Vowel Harmony is a Basic Phonetic Rule of the Turkic Languages: You are not allowed to view links. Register or Login to view.
  
  DEVIATIONS FROM THE GENERALIZED PHONOLOGICAL RULES IN TURKIC LANGUAGES: > You are not allowed to view links. Register or Login to view.
  
  Dialects on the Eastern Turkish Coast of the Black Sea:
  You are not allowed to view links. Register or Login to view.
  
  Turkic sound changes:
  You are not allowed to view links. Register or Login to view.
  
  THE SOUND STRUCTURE OF THE TURKIC LANGUAGES IN CONNECTION WITH THAT OF THE FINNO-UGRIC ONES: (An instrumental-phonetic and phonologic investigation):
  You are not allowed to view links. Register or Login to view.
  
  COMPARATIVE PHONOLOGY OF HISTORICAL KIPCHAK TURKISH: You are not allowed to view links. Register or Login to view.
  
  The Influence of Turkish Sound System on English Pronunciation:  You are not allowed to view links. Register or Login to view.
  
  ASSIMILATION OF SOUNDS IN THE TURKISH LANGUAGE: A CASE STUDY OF HIGH SCHOOL STUDENTS AT KARABUK REGION (TURKEY):
  You are not allowed to view links. Register or Login to view.
  
  Turkish phonology: You are not allowed to view links. Register or Login to view.
  
  Thank you,

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Ahmet Ardıç > 30-06-2024, 07:25 PM

  If anyone wants to answer the questions below, I'm curious about your opinions. 
  
  I also partially asked these questions to artificial intelligence;
  
  How many different languages are there in the world?
  If we were to write human sounds in modern and ancient European languages with Latin letters, which language would need the largest number of letters? How many letters should or are there in this language?
  Which written language requires the largest number of letters in the entire Indo-European language family? How many letters are there in this language? I would like the same questions to be answered for the Semitic family languages.
  
  You can see the AI answers below. Artificial intelligence says that the Indo-European language, which needs the most letters in Europe, can be written with approximately 28 (or maybe 32?) letters. (I would like to discuss the exact question and the answer given by the machine here.) Do you think the answer to the machine is correct? Which European language needs the largest number of letters? Considering the ancient sound structures of these languages, including today's European languages, which European language do you think needs the most letters?
  
     
  
  Chat GPT4's answer was more detailed than GPT3.5.
  
  Here is the answer:
  Addressing your questions involves delving into phonology (the study of the sounds used in speech) and orthography (the method of writing a language). To analyze the complexity in representing human sounds (phonemes) in written form across different languages, we'll need to consider both modern and historical linguistic scenarios within the Indo-European and Semitic language families.
  Indo-European Language Family
  Modern and Ancient European Languages:

  • Across all European languages, both modern and ancient, the distinctions are nuanced. Ancient languages like Classical Latin, Ancient Greek, and Sanskrit had complex phonetic systems, but each was efficiently represented with an alphabet adapted to its sounds. However, if we consider languages in terms of needing adaptations like diacritics (marks added to letters), the picture changes.
    
  • Old Church Slavonic and its derivatives in the Cyrillic script might be a contender among ancient European languages because of its range of sounds represented by a combination of basic letters and diacritics.
    

  Language Requiring the Largest Number of Letters:

  • Among Indo-European languages, Sanskrit stands out for its complex phonetic system, especially when represented in the Devanagari script, which isn't based on Latin letters but can be transcribed with them. Sanskrit phonology includes a wide range of vowels (both short and long), diphthongs, and consonants, including voiceless, voiced, aspirated, and nasal sounds, leading to a large number of distinct phonemes.
    
  • When Romanized (transcribed with Latin letters), Sanskrit requires a significant number of additional diacritic marks to represent its phonetic nuances, arguably making it the language that would need the largest number of Latin letters for accurate transcription. The exact number can vary based on the transcription system, but it would easily surpass the basic 26 letters, incorporating diacritics to represent different phonetic qualities.
    

  Semitic Language Family
  Largest Number of Letters Needed:

  • Semitic languages, including Arabic and Hebrew, use abjads—a type of writing system where each symbol typically stands for a consonant, leaving the reader to supply the appropriate vowel. This makes the phonetic representation seem simpler but can complicate precise transliteration into Latin script.
    
  • Arabic could be considered the Semitic language that requires a large set of Latin letters for transcription due to its phonetic range, including sounds not found in many other languages, such as the emphatic consonants. However, the core Arabic alphabet contains 28 letters, and additional letters are used in some languages that use the Arabic script to represent sounds not found in Classical Arabic.
    
  • Proper representation in Latin script would require the use of diacritics and additional letters to accurately capture these sounds, especially if we're including the variations present in modern Arabic dialects.
    

  Conclusion:
  In conclusion, within the Indo-European language family, Sanskrit (when its sounds are transcribed into Latin script) likely requires the largest number of letters due to its rich phonetic inventory. For the Semitic family, Arabic, when transcribed into Latin script, potentially requires a considerable array of letters and diacritic marks to represent its full range of phonemes accurately. The exact number of letters or symbols needed can vary significantly based on the transcription system used, but both languages would necessitate more than the standard Latin alphabet to represent their phonetic complexities accurately.
  
     
  
     
  
     

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Koen G > 30-06-2024, 07:55 PM

  I just don't see the benefit of using ChatGPT for this. If we have to do research to fact check it, wouldn't it be better to do and present the actual research? Why does a deeply flawed large language model have to be involved?

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  asteckley > 30-06-2024, 08:18 PM

  (02-06-2024, 07:56 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.Current generation AI is not smart. It is simply good at producing convincing language. Do not trust it for complex analysis! By using ChatGPT for this kind of task, all you are doing is signaling your lack of understanding of its limitations.

  
  True!
  ChatGPT is actually extremely powerful and useful, but only if one understands why it works at all, and therefore what its limitations are.
  
  But, unfortunately, not understanding its limitations is something that even some of the "top" experts in the field are guilty of. In fact, it's regularly argued about in social media forums.

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  AmaliaRose > 30-06-2024, 11:02 PM

  For anyone following this thread, who is unaware of the limitations of LLMs, please follow the work of You are not allowed to view links. Register or Login to view.. For those leaning to academic reads, You are not allowed to view links. Register or Login to view." is the paper to read, freely available on ACM. AI can bring a lot to this research, however, we do need to understand promise and peril.

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Ruby Novacna > 30-06-2024, 11:35 PM

  That's exciting, Ahmet, but what's next? Are you going to test the 7 thousand languages?

• RE: Calgary engineer believes he's cracked the mysterious Voynich Manuscript

  Aga Tentakulus > 01-07-2024, 12:01 AM

  I don't know what to make of the AI's statement.
  The German alphabet alone has 30 characters.
  Sometimes you count it as 26, then 29, just as you want it.
  The long "S" (sch) and "C" (tsch) are no longer used.
  Interesting: We speak "tschau" and write "tschau" in Italy we speak "tschao" but write (ciao). Say goodbye.
  See Wiki.
  
  You are not allowed to view links. Register or Login to view.
  You are not allowed to view links. Register or Login to view.
• Next Oldest Next Newest