The Voynich Ninja

Hello to everyone who reads my articles and comments,

I would like to extend my gratitude to everyone who has been following my linguistic research through various articles, paying attention to the linguistic findings and details therein, and sending me private messages to congratulate me. Some experts working in the field of the Turkish language also joined in, offering their congratulations due to one of our articles. I will share information and the link to one of our articles, which was the reason for some of these congratulations, here.

A book containing symposium papers, with evaluation committees consisting of experts in the Turkish language field, was published by the "Türk Kültürünü Araştırma Enstitüsü" (Turkish Culture Research Institute) at the end of last year. However, I just have received this publication. One of the supporters of this symposium was the Turkish Language Association (TDK), which is the authoritative institution of the Republic of Turkey regarding the Turkish language.

Thus, I would like to inform everyone who has not yet heard about this topic that our article titled "READING OF THE PAGE 33v OF VOYNICH MANUSCRIPT" submitted for the 1st International Turkish Culture Symposium, has been published. Our article can be found starting from page 40 in this booklet.

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

file:///C:/Users/user/Downloads/1.ULUSLARARASI%20T%C3%9CRK%20K%C3%9CLT%C3%9CR%C3%9C%20SEMPOZYUMU%20(2)_compressed%20(1).pdf

The scientific committee of the international symposium I mentioned consists entirely of academics working in the field of Turkology. In fact, the evaluators and members of the scientific committee of this peer-reviewed publication are professors, associate professors, and doctors who hold titles in the field of science. As far as I know, all members of the scientific committee of this symposium are linguistic experts with valuable contributions in the field of the Turkish language (including Old Turkish).

When you open the link, you may observe that the images of our paper shared in this proceedings booklet appear colorless due to black-and-white printing. Therefore, as of today, I am also sharing a colored version of the same paper on our page. The text is in Turkish, but interested individuals can utilize artificial intelligence to translate it into their respective languages for reading.

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

We have demonstrated how we utilized comparative linguistics methods to read a full page (page 33v) of the Voynich Manuscript (VM) texts and translated it into modern Turkish. Throughout this process, we provided examples illustrating the evolution of old word spellings over time and identified these words by referencing known dictionaries. It's crucial to address this issue, like any other unknown, through scientific methodologies rather than unscientific approaches. We have scrutinized the techniques employed by linguists and have strictly adhered to these methods. However, it's essential to note that words in agglutinative languages such as Turkish should not be analyzed in the same manner as those in Indo-European languages. Therefore, when examining the details, strive to acquire fundamental information specific to the language under study.

We are pleased to share this positive development with you. Moving forward, we anticipate receiving constructive criticism from experts engaged in academic research in the field of Turkish language studies. While we have yet to receive any criticisms (in fact, we have only received congratulatory remarks from experts who have reviewed our work), we hope that subject matter experts will begin contributing to our efforts in deciphering the VM with their valuable critiques and suggestions in the future.

Thank you for reading.

Kind regards,

Ahmet Ardıç

Thank you once again for the congratulatory messages by e-mail. The fact that we received congratulatory messages from linguists, especially those working in Old Turkish, indicates that there is a certain level of increased awareness regarding the outcomes of our work.

So far, the majority of individuals still fail to recognize the presence of Turkish sentences within the VM content. Of course, this is understandable and quite normal, and at this stage, it is not surprising.

Here, I would like to offer some advice to scholars and researchers and also draw attention to certain issues. This is because researchers who are unfamiliar with the Turkish language and have no knowledge of Old Turkish often misinterpret our work or fail to examine it carefully. Therefore, I would like to highlight the following points.

1-) It is not possible to reach a solution with approaches outside linguistic methods.

2-) It is not possible to read the VM content with incorrectly made alphabet transcriptions.

3-) Every researcher who believes that there are natural languages in the content must clearly present the correspondences with an old-language they claim to exist in the content. For example, they must present the forms of readings from 600 years ago (based on dictionaries, academic articles, and similar examples shown in other old manuscripts). Also, they must show the correspondences between the sentence structure, phonetic features, and word structure of the old form and the current form of that languages, and it must be shown that the presented correspondences are also found in other old manuscripts. For example, if the letters A and E are represented by the same symbol, it must be shown that similar examples were also seen in other old manuscripts. Thus, the proposed solution cannot be specific only to Voynich. Therefore, by referring to historical records and linguists' studies, it must be shown that similar examples have been recorded elsewhere.

4-) When matching sounds corresponding to VM-writing characters, the proposed alphabet transcription values must be used consistently every time. With our ATA transcription, we actually limit ourselves in creating multiple words. For example, if you know that linguists sometimes match 5 or 6 sound values for one sign when reading the writings of ancient civilizations such as Hittite, Sumerian, and Akkadian, and that a word can often be read in different phonetic forms (to the extent that some words have been accepted as correct readings in 5 different phonetic forms), it will be easier to understand that such a study is more inconsistent compared to our ATA transcription readings. Many researchers, even if they are not aware, have provided reading suggestions within a much narrower range based on our ATA alphabet transcription, considering the flexibility criteria accepted by linguistics.

5-) Our transcription also includes familiar vowels in Turkish. For abbreviations, usually only a single reading is often provided. For example, we are mentioned and reading that the word "ÇOCUK" (child) is written by the author as "ÇCK," (such as similar as CHLD in English) in VM and we read this word with the same meaning in every time. We have not read this word as "child" on one page and in another meaning on another page. Not all words written in abbreviation form have been read yet. However, as sentence analyses/translations are made, it must be ensured that each abbreviation will always be read as the same word and meaning. If there are exceptional cases, why this occurs must be explained again with scientific methods and historical similar examples.

6-) Many words we read have been written (by containing with vowels) without any abbreviation. It has also been the case that many drawings-word overlaps have been shown. For example, on the page where the drawing of a certain plant is made, the name of that plant has been read, and it has been shown that a plant name written approximately 600 years ago by the author is still written in the same form today. In other words, even the phonetic value of a single letter can be read unchanged. Such overlaps have been seen throughout the VM manuscript in numerous instances. The fact that this situation has occurred is actually invaluable findings in terms of linguistics. In terms of mathematical probabilities, it can be put forward with formulas that such 1/1 overlaps are not random. In our readings based on ATA transcription, we have presented many overlaps with 1/1 identical phonetic values (including vowels) between the old and new writing forms compared so far.


7-) We have proposed the most realistic alphabet transcription to date. Our ATA alphabet transcription, when compared to all other transcriptions made so far, offers the highest number of matches to known symbols of the Latin alphabet. Even this occurrence has value and significance for accurate reading (especially in terms of scientific and realism). Because this situation should make (as many previous researchers have claimed) researchers realize that the "VM symbols are not unprecedented symbols," but rather the majority (almost all) of them are already found in known alphabets.

8-) Turkish language is quite rich in terms of word repetitions such as doubling, tripling, etc., in all known historical texts. This richness perfectly matches the VM texts in terms of word variety and frequency of occurrence in the texts. So much so that the same word written three times, four times, or five times in a row has been demonstrated by linguists to be shown in other historical (ancient) manuscripts. We also demonstrated the same structural condition in VM texts. Especially word triples, quadruples, and quintuples together are specific to the Turkish language only. In other words, this phenomenon has not been recorded in ancient texts of any other language in the world. These types of matches provide structural and language-specific matches among compared texts.

9-) It is important to demonstrate a large number of word and drawing matches. In addition to showing these matches, matches should also be demonstrated in terms of sentence integrity and sentence structure. This should be done by reading numerous sentences. In fact, this is not enough. Examples of reading full pages must also be shown. When doing this, the old phonetic forms of the words must be found and demonstrated in the known dictionaries of linguistics in the same phonetic form and meaning. Our work encompasses all of these.

10-) Proposed solution suggestions should also be compatible with the known structures of linguistics. Anagrams should not be suggested. Using methode of outside of recognized known scientific ways never be useful in any time. If artificial intelligence translation pools like Google are to be utilized in solutions, the proposed words must be read in the same way phonetically without manipulating their phonetic value. Moreover, this is not enough. It must also be demonstrated that the words read are found in other ancient texts in the same meaning and phonetic form. Our work presents this as well. While following the methods accepted by linguistics, if we cannot demonstrate that even a single word is found in ancient texts and dictionaries, we do not claim that the page in question has been read completely.

11-) Words written in abbreviation form are almost present in every line. Some of these must be plant names or compound plant names. So much so that examples have been identified in the content where the author has created a new name by grafting two different plants together and using the names of both plants (in the unabbreviated form). Moreover, it has been shown that these examples coincide with plant drawings made on the same pages. In this case, it would be realistic to consider or think that the author may have been wrote the names of two plants in Old Turkish or Latin with in abbreviation form too, and combined them as into a single word. In this case, botanists should participate in the translations. When botanists consider the ATA transcription while looking at the VM content, it will be very beneficial for the translation of the entire 240 pages into modern language. So, it should not be expected that all 240 pages will be translated into modern languages by a single family working group. These studies should be conducted with the collaboration of experts from different disciplines. But it is now necessary to abandon alphabet transcription variants such as EVA that do not work. EVA and other transcription variants have not led to a conclusion. We, on the other hand, clearly present the most realistic results so far with the ATA transcription.


I'm not claiming that all the results of our work are perfect. There may be some mistakes in some details. However, all errors will be resolved through the involvement of experts working in various fields/disciplines using scientific methods. It is not possible to achieve accurate results with inconsistent approaches outside of scientific or logical realms. Illogical and unscientific approaches will not lead to solutions in the future too.

None of the solutions we have proposed have been outside the realms of scientific or logical domains. We have not received any realistic academic criticism that could demonstrate inconsistencies, even in a single point in our VM study pages. A few years ago, an English article we submitted to a foreign journal was evaluated quite poorly, and we do not know the experts who made this evaluation, but none of them were Old Turkic experts, that's for sure. Because we have shown so many sentences and all the words on some full pages in old dictionaries, any researcher who evaluates what we have written correctly will also clearly see the realistic results obtained.

We now only write our books and articles on VM in Turkish. Because our native language is Turkish. Especially, we submit our articles to symposiums attended by expert lecturers in Old Turkic in the departments of Turkology at universities. Everyone who claims that there is a natural language in the VM content must do the same. For example, if you say that the content is in Latin or Hebrew or another language, you must share the evidence with the experts in those languages. We are doing exactly that. So far, when experts with scientific data have made this evaluation, they only conveyed congratulatory messages. However, there is still no widespread awareness internationally that the content is in Turkish. However, there is no need to rush because sooner or later it will be widely accepted that the content is in Old Turkish. My reason for commenting here and providing information to readers is to explain that they (may) need to assist in the scanning of words and proper nouns in European languages in VM texts using the ATA transcription.

Thank you for reading.

Best regards,



See The Base ATA Alphabet Transcriptions here > 
You are not allowed to view links. Register or Login to view.
[attachment=8204]

See The Concatenation logic of simple (base) alphabet-characters here > 
You are not allowed to view links. Register or Login to view.
[attachment=8205]

Dear Ahmet,

there are a few issues with some of your arguments.
I just want to highlight the first one:

" 1-) It is not possible to reach a solution with approaches outside linguistic methods. "

This is not generally correct. It could be (partially) correct *IF* the text is meaningful, in some language.

The two main problems are:
- It is not yet certain, at all, that the text is meaningful.
- If it IS meaningful, it may have been encrypted, and in that case, it would require cryptological methods in addition to linguistic methods.
So, yes, liguisitic methods would be needed, but by themselves they might not be sufficient.
Note also that statistics are nowadays an important tool in linguistic methods, and they can be used to either confirm, or disprove, certain hypotheses.

There are two very different situations:

A) Someone presents a solution that translates the vast majority of the text, and it largely makes sense.
In this case, there is:
- much less doubt that the text could be meaningless, 
- not much doubt about the language of the text
- much less need for statistical tools to confirm or deny the result

B) Someone presents the beginning of a solution
(This is what we get in almost all cases. The few exceptions could be discussed separately).
The beginning of a solution means that one presents translations of individual words, or some short fragments.
This is an incomplete solution and in this case nothing is confirmed. Instead:
- there is still doubt that the text could be meaningless
- the language has not yet been confirmed
- statistical methods can be used to estimate the likelihood that this method can translate the rest of the text.

For me, there is no doubt that you are in this second situation (B).
If not, you would be able to translate the most frequent words (as I proposed in another thread), and these words would also be frequent words in old Turkish. This is just one example.

To judge whether any particular solution of type B has a chance to be correct, requires two things:

- Linguistic knowledge of the proposed source language
AND:
- Knowledge of the special features of the text in the Voynich MS.

The second point is of critical importance. Now I have to assume that you have some understanding of these special features, but you do not seem to address them.
I am very doubtful that the linguistic experts who are positive about your work have this knowledge, so it will be impossible for them to properly estimate the likelihood that your solution is correct.

Right now, there are more than half a dozen proposed solutions that are based on individual words or short fragments, and I cannot see any reason to prefer one above the other.
In fact, all of them have serious issues which make all of them equally unlikely to be correct.

It is a perfect example of an undecided match:

Person A argues that it is language X, and you need to know language X to judge the result.
Person B argues that it is language Y, and you need to know language Y to judge the result.

Now person A does not know language Y and person B does not know language X.
In that situation there is nothing to decide who is more likely to be correct.

Person C might not know either language X or Y, but understands statistics, and can see that the methods proposed by both A and B cannot reproduce these statistics.

Have you ever tried translating a fragment of a Turkic text into Voynichese?

(27-02-2024, 09:03 AM)Koen G Wrote: You are not allowed to view links. Register or Login to view.Have you ever tried translating a fragment of a Turkic text into Voynichese?

Too bad that the only test that would easily convince anyone who knows anything about the properties of Voynichese isn't possible, because the dialect isn't identified:

Quote:Our current conclusion is that the author’s dialect is a mixture of Anatolian, Azerbaijani, and Uzbek as it is congruent to all three. Another possibility is that the Turkish dialect used by the author is one of the dialects that are no longer employed / forgotten today.

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

Let me just point out that this hypothesis is non-falsifiable: there is no corpus to compare Voynichese with and any word that is not found in old dictionaries could be an abbreviation or a variant spelling or belong to a forgotten dialect.

Gerard Cheshire did the same thing with his proto-romance mixture.

Koen and nablator: I agree with both of you, but the argument of Ahmet seems to be, that one has to know old Turkish. If one does not know old Turkish, then one is not qualified to judge his solution.
This would mean that the criticism of the three of us has little value.
The support from other Turkish linguists would appear to have more weight ....

The point I was trying to make is, that that could perhaps be true if he had a complete solution.
However, that is not yet the case.
It is very partial.

In order to go from this partial solution to a complete one will be very difficult, and that is caused by the particularities of the Voynich MS text.

According to Ahmet's logic, only Turkish linguists can judge the VM and claim or disclaim whether the language is actually Old Turkic. The Voynich solution is not depended on confirmation only, but also on disproving other possible theories. As long as any other theory is floating around, the Turkish team and the experts defending Turkic theory would have to convince non-Turkish Voynich experts why Turkish theory is better and why it can exclude all possible future proposals.  As I understand, they are not yet determined which dialect the VM is written, or if it written in some extinct Turkish dialect. This brings me to one extinguished dialect, spoken in today's Turkey, before the Turks came from Mongolia. That was the Slavic dialect in Thessaloniki, which was so close to the Slavic language spoken in the 9th century in Pannonia and other parts of Great Moravia, that the missionaries from Thessaloniki were invited to Great Moravia to instruct the Slavs in their native language. The language subsequently called the Old Church Slavonic has been best preserved in Slovenian dialect of Prekmurje.

Since by 15th century, Turkish grammar and vocabulary was already explained in the Codex Cumanicus in the words written in Italian and German writing conventions, it would be relatively easy for the Turkish team to find some words that are spelled exactly the same in the Voynich Manuscript and in the Codex Cumanicus, possible in the same grammatical form.
They would also have to address all those grammatical and structural oddities of the VM language which were pointed out by various VM researchers.

The situation cannot be resolved by disputation, rather by demonstration. Whoever has a proposed solution, let them show what is needed to read a VMs page in a way that makes sense. And if the proposed solution works well in one place, then try it on the text in other sections. If it works - it works.

The fundamental determination is: does the VMs text make sense? - and what does it say?

If the solution is so close at hand, it should be possible to read something.

Dear Rene,

If a researcher has introduced a new scientific method to the methods applied by linguists and if this method is considered explanatory, then the method, which is thought to be explanatory, cannot be outside the scientific field. But in any case it is not possible to reach a solution with approaches outside scientific & linguistic methods.

Linguists cannot use the statistical approach as a verification tool in unknown dialects. However, we also partially use statistical methods as an additional evidence.

The text is meaningful with using ATA transcription in Turkish shown.
- It is certainly that the text is meaningful.
In our articles that have garnered praise from linguists, we precisely demonstrated meaningful sentences and full-page readings in VM texts. In doing so, we relied on evidence and ancient dictionaries to illustrate them.

There is no such thing as encrypted texts spanning 240 pages. It's entirely written in a natural language. The Voynich Manuscript texts are written in Old Turkish (in a dialect that we haven't fully explained yet). Determining the dialect precisely will take some more time. However, current results indicate that this dialect is more likely from the Black Sea region, Istanbul, and Balkan accents. The Voynich Manuscript texts are partly written in word spelling style, dividing words into syllables, and sometimes combining a few words and often written in abbreviated form. (Moreover, our author may be one of the minorities whose native language is not Turkish, but he/she wrote in Turkish.)

However, it's shown that there are also numerous sentences that are read very clearly. In fact, it can be said that there has been no manipulation in any writing style in these sections.

As you know, linguistic methods certainly involve statistical approaches, but they are not limited to that alone.

The language of the texts has been understood. Turkish words have been seen on every line throughout the 240 pages. When the unidentified words are identified, the entire 240 pages can be translated into modern languages.

Our work papers never reject by linguists who works in Turkish. This may be relevant for those who do not speak Turkish and have not yet thoroughly examined our work.

We as ATA team never presents the beginning of a solution only, but more than this. Firstly, we created an alphabet transcription that allows us to determine the language in the content, and this key has enabled us to read the entire 240 pages. As I mentioned, we see Turkish sentences and words clearly in every line of this book (approximately throughout 240 pages). Some sentences are read without difficulty, while others, due to a few words, cannot yet be translated into today's language, but eventually, all will be read. As I said, our study not only helped us understand which language is in the texts but also facilitated the understanding of the texts.

Turkish is an agglutinative language. For example, when writing scientific texts and using the present tense, it's common to see the same suffixes (phonetic elements) in words. However, this doesn't mean that every Turkish text should follow the same pattern. This is precisely where you have fallen into the fundamental misconception. In order to use statistical approach in comparative linguistics, you first need to find two texts to compare.

If one of these texts is the Voynich Manuscript (VM) and the other is selected when we clearly identify the dialect, Turkish's different dialects can be match another with statistical way. In fact, if you compare 2 different dialects in Turkish, those dialects may not show statistical matches with each other. Sentence structures may even differ. While in one dialect the subject pronoun may be at the beginning and the predicate at the end, in another dialect this structure may not be followed. Moreover, word suffixes may also vary. Furthermore, in some Turkish dialects, there may not be enough written or ancient texts to compare statistically. Many texts are in this situation and they are read as Turkish without resorting to statistical data. Thus, a solution can be reached without using statistical methods. If this were not the case, Sumerian inscriptions would have remained unread. Linguists write that the Sumerian language is not related to any known language. So, what statistical comparison method could linguists who first managed to read this language have used? Therefore, statistical methods you mention will not be verification tools. If so, the fact that the Sumerian language has been read would have to be explained as unverified information. However, linguists say they can read the Sumerian language and translate meaningful sentences into modern languages, which is true. Moreover, tens of thousands of Sumerian inscriptions remain unread. Nevertheless, linguists claim that this language has been deciphered, and they are correct. Linguists cannot use the statistical approach as a verification tool in unknown dialects. However, we also partially use statistical methods. For example, we have shown with this method that the writing patterns of word triples, quadruples, and quintuples overlap between the Voynich Manuscript and Turkish.

We already be able to translate the most frequent words as I proposed in another thread. But never think that these all of the words would also be frequent words in old Turkish. Because this detail is clearly a dialect oriented point.

With our study in Turkish, linguistic knowledge of the proposed source language already matched. With our study, knowledge of the special features of the text in the Voynich-MS/Turkish are already matched too. We have clearly demonstrated these with specific examples in our previous posts here and in our own articles.

Do not be very doubtful that the linguistic experts. Your sentence about related linguistic experts that like "it will be impossible for them to properly estimate the likelihood that your solution is correct." you uttered seems to convey a rather haughty tone. Please try to understand our articles in a consistent and detailed manner, recognizing that you may not have any knowledge of Turkish and Old Turkish, rather than assuming that experts in the field lack the linguistic competence to evaluate our work.

Yes, there are more than half a dozen proposed solutions that are based on individual words or short fragments, and may be you cannot see any reason to prefer one above the other because your linguistic knowledge.

Your lack of knowledge of Turkish in making such a distinction is quite understandable. However, to be able to compare the evidence we present with similar claims, you need to be able to examine the evidence presented in comparative linguistics. Our claim that "VM contains Turkish" has now been proven. It is not for me to determine at what stage you will understand or whether you will understand at all. Even if no one on this page is convinced, the conviction of experts in Old Turkish is sufficient. The content of VM is Turkish. Whether you believe it or not, this has been proven beyond doubt and in a very clear manner.

In fact, if person A says it's language X or I say it's Turkish, you can still see whether the results and evidence are scientific or realistic without knowing those languages in detail. If you can't do this or think you won't be able to see it, then of course, in this case, you can put all the claims in the same bag.

We understand that those who will approve the outcome are professionals in the field of that language. In other words, if Person A claims that it is Language X, then those who will approve or reject the outcome should be experts in Language X. By the way, whether you have understood or not whether Language X is a real result or not for you or myself or any other person opinion has no value for a linguist in X language area. However, if experts in Language X begin to congratulate the claimant, carefully re-reading what the claimant has written may perhaps show you the true result as well.

The statistical approach is valuable and is also used by us. Many ancient texts that were previously unreadable have been deciphered without the statistical approach. Because what matters is to use the methods within linguistics. We have used both and demonstrated the overlaps.

Your thoughts on the confirmatory value of statistical methods are not consistent with scientific practices. In other words, I can say that you make many incorrect inferences in many details, and you try to proceed assuming that your incorrect inferences are correct. However, as I mentioned, statistical comparisons are not always made. Evidently, ancient texts have been deciphered without statistical comparisons. Statistical methods are useful for providing additional evidence, but they are not indispensable. Linguistics has read almost all ancient inscriptions without them, and you should be aware of this or take it into account. What adds value to a study is not the opinions of critics but the information and evidence in the content, and whether it is scientific or not.

I believe you have the maturity to accept my criticisms with understanding. My intention is not to offend you, but I wanted to point out that it would be erroneous to view an incorrect inference as a means of verification.

Thank you for reading.

Kind regards

A. Ardıç

(27-02-2024, 09:03 AM)Koen G Wrote: You are not allowed to view links. Register or Login to view.[Have you ever tried translating a fragment of a Turkic text into Voynichese?]

"Voynichese" has never accurately represented compound syllable characters/signs with their correct sound values. Therefore, this application is not useful to me. However, sometimes it is helpful for finding out which pages certain words are on. If we want to automatically convert texts containing VM-writing signs into the Latin alphabet, ATA transcription will need to be applied to these softwares, but I have never tried this before.