04-04-2022, 06:28 PM
Dear Davis,
Thank you for writing. Yes, we were able to stay busy with our work without losing our health during difficult times. Thank you. I hope you are well too.
Thank you for the details and reminders you drew attention to on the subject.
I perceive and understand the detail you have stated. Moreover, I had read the VM statistics studies you referenced in the past. Yes, it is possible to examine different aspects of texts using statistics. For example, Ratio of vowels and consonants in the compared texts can be checked. Or, Character entropyis (another metric by which we can compare languages to each other and to Voynichese) can be checked. And some other comparisons can be made through statistics too.
Remember, Bennett (1976) indicate the unusual nature of the Voynich script in his discussion of conditional character entropy (also known as second order character entropy). Which conditional character entropy can also be thought of as the overall predictability of a letter given the pre-ceding letter. In that case, the overall conditional entropy can be calculated from "the conditional probabilities of each bigram" weighted by their overall occurrence in the text for sure.
As you know, based on many past study indicate that character sequences within the words in Voynich text are unusually predictable compared to European languages.
We already have full-page reading of VM as like author wrote (in undefined dialect) with enough words form many all VM pages and phrases and thats all available reads can be use in lexicostatistics measurements and comparisons. For example, the folio 33v we worked on before was read in full. (We are still waiting for Turcologists to evaluate our reading on this page.)
The words of the Turkish language have a phonetic vowel harmony within themselves. In other words, it is predictable and what kind of sound structure will come from which sound is more predictable. Moreover, in Turkish, some letters/sounds are never at the beginning of words, and likewise, some letters/sounds are never at the end of words. I have mentioned these in my previous comments on this page and stated that the same structure coincides with corresponding to the same sounds for Voynich texts. I stated for that structure as "it should be defined as a structural overlap between today's Turkish and Voynich texts".
The vowel harmony rule always works in the language, and most of the time, even if you do not write the vowels partially or completely, a natural speaker of this language will continue to read the unwritten letter at the sound value it should be.
In addition, there is an almost complete overlap with Turkish in terms of repeated side by side word structures such as reduplication, trinity, and quatrain etc. I also mentioned this with examples and past evidence.
ATA transcription enabled us to read these articles, and structurally, such overlaps seem very clear. These linguistic findings and overlaps and the readings that have been made have already shown that the content is Turkish without using the statistical method.
On the other hand, statistical studies for VM have not been able to obtain such a result or identify the language in which the work was written.
The study of linguistic statistics for VM texts is approximately similar to "reverse engineering" in technology and industry. First, the inner parts and sections are examined in a way that can be understood, and their relations and numerical connections are reproduced again, resulting in a whole that will serve to compare the original old parts to other knowns.
So, Linguistic statistics is one of many methods, but it is never the only method that can be applied. We have explained our method and if the VM texts are not rendered correctly with the correct volume value, no statistical comparison will give correct results. This method has already been applied many times, and the positions shown for the VM in the plotted graphs are incorrect.
Because those who do this work create texts by reading the original letters incorrectly and compare unrealistic texts with today's languages in statistical calculations. All of those works are not work yet just for this reason. But I'm not saying that they were applied as good experiments and it is not possible to reach a result using statistics.
However, regardless of the fact that we have enough raw vocabulary words to do Lexicostatistics studies, these words and full-page texts will be corrected by linguists. If these studies are done after specifying the Turkish dialect to be compared, they will give correct results.
Almost every 2 words written by the author were either shortened or written as if they were separate words by separating their syllables or phonemes, or the vowels were written incompletely or sometimes they were written with only consonants.
Moreover, some words were combined even though they should not have been combined, and some sounds were written by lengthening certain sound expressions, almost like a spoken song.
Thus, half or more of the texts were manipulated by the author to make the texts difficult to read. Therefore, if these will corrected more accurately by linguists for Lexicostatistics than correct results can be obtained.
We think that the use of texts in a Lexicostatistics study before these are completed will give misleading results. But there is another reason why we did not do a study for comparison in Lexicostatistics in detail.
We first made the ATA alphabet transcription for the Voynich texts by examining the old alphabets in terms of the Turkish language and blending them with the familiar characters of the Latin alphabet. This transcription enabled us to read the texts in Turkish to a great extent which no transcription has been able to achieve until today, despite the fact that the texts were written by multi-directional manipulation.
Therefore, after the texts are read in Turkish, we never thought to prove it again with using Lexicostatistics comparison to see whether it overlaps with Turkish or not. This can of course be done as a second validation after a while.
Moreover, none of the alphabet transcriptions that enable the reading of unknown historical alphabets and texts have been revealed by lexicostatistics studies. So there is no such example in history. The primary requirement for being able to read ancient texts is to make an alphabet transcription that works as a reading key. None of the transcriptions made to date, including EVA transcription, showed a voice equivalent to more than 100 characters of the Voynich alphabet.
They never predicted true multi-syllable characters and matched them with correct sound values. Moreover, they misread even simple alphabet characters. For example, they were read 600-year-old different VM letters with same phonetic value.
In our study, some letter signs can be read with a double voice, as like other known Old Turkish texts. While this is the case, no matter what method we compare the sound indexes without correctly predicting the compared sound values, it will not give correct results.
In summary, if we are going to do the study you mentioned, it will mean that we try to confirm the findings, which we have already confirmed with our own method, with the proposed method for the second time. This can of course be done. But first, let's get help from linguists to correct the texts, then this suggestion is tried if necessary and it shows overlap.
About the "High number of occurrences of EVA [p] and [f] in the toplines of paragraphs as opposed to elsewhere in a paragraph" as for the subject you mentioned that the signs perceived as P and F are actually a large number of very different syllable character signs.
In other words, if you/we look at them carefully, it will be seen that their number is more than two. So that the ones read as P are different from each other and the same situation is valid for those read as F.
EVA P and F are multiple syllable characters, each of which is different from each other.
No matter which structure in the text you/we want to examine through linguistic statistics, you/we must first find an alphabet transcription whose sound values must be correctly matched. This transcription work is not something that can be done by using eye and analogy or based on personal opinion only. Voynich alphabets are not located of the outside area of the historic alphabet evolution process. These VM letter signs not can be read with personal analogies far from science. EVA is just created with baseless personal analogies except for the well known Latin alphabet characters we use today. It has almost no scientific value that can be used in a scientific study.
Sorry my post was a bit long.
Thank you for reading
Thank you for writing. Yes, we were able to stay busy with our work without losing our health during difficult times. Thank you. I hope you are well too.
Thank you for the details and reminders you drew attention to on the subject.
I perceive and understand the detail you have stated. Moreover, I had read the VM statistics studies you referenced in the past. Yes, it is possible to examine different aspects of texts using statistics. For example, Ratio of vowels and consonants in the compared texts can be checked. Or, Character entropyis (another metric by which we can compare languages to each other and to Voynichese) can be checked. And some other comparisons can be made through statistics too.
Remember, Bennett (1976) indicate the unusual nature of the Voynich script in his discussion of conditional character entropy (also known as second order character entropy). Which conditional character entropy can also be thought of as the overall predictability of a letter given the pre-ceding letter. In that case, the overall conditional entropy can be calculated from "the conditional probabilities of each bigram" weighted by their overall occurrence in the text for sure.
As you know, based on many past study indicate that character sequences within the words in Voynich text are unusually predictable compared to European languages.
We already have full-page reading of VM as like author wrote (in undefined dialect) with enough words form many all VM pages and phrases and thats all available reads can be use in lexicostatistics measurements and comparisons. For example, the folio 33v we worked on before was read in full. (We are still waiting for Turcologists to evaluate our reading on this page.)
The words of the Turkish language have a phonetic vowel harmony within themselves. In other words, it is predictable and what kind of sound structure will come from which sound is more predictable. Moreover, in Turkish, some letters/sounds are never at the beginning of words, and likewise, some letters/sounds are never at the end of words. I have mentioned these in my previous comments on this page and stated that the same structure coincides with corresponding to the same sounds for Voynich texts. I stated for that structure as "it should be defined as a structural overlap between today's Turkish and Voynich texts".
The vowel harmony rule always works in the language, and most of the time, even if you do not write the vowels partially or completely, a natural speaker of this language will continue to read the unwritten letter at the sound value it should be.
In addition, there is an almost complete overlap with Turkish in terms of repeated side by side word structures such as reduplication, trinity, and quatrain etc. I also mentioned this with examples and past evidence.
ATA transcription enabled us to read these articles, and structurally, such overlaps seem very clear. These linguistic findings and overlaps and the readings that have been made have already shown that the content is Turkish without using the statistical method.
On the other hand, statistical studies for VM have not been able to obtain such a result or identify the language in which the work was written.
The study of linguistic statistics for VM texts is approximately similar to "reverse engineering" in technology and industry. First, the inner parts and sections are examined in a way that can be understood, and their relations and numerical connections are reproduced again, resulting in a whole that will serve to compare the original old parts to other knowns.
So, Linguistic statistics is one of many methods, but it is never the only method that can be applied. We have explained our method and if the VM texts are not rendered correctly with the correct volume value, no statistical comparison will give correct results. This method has already been applied many times, and the positions shown for the VM in the plotted graphs are incorrect.
Because those who do this work create texts by reading the original letters incorrectly and compare unrealistic texts with today's languages in statistical calculations. All of those works are not work yet just for this reason. But I'm not saying that they were applied as good experiments and it is not possible to reach a result using statistics.
However, regardless of the fact that we have enough raw vocabulary words to do Lexicostatistics studies, these words and full-page texts will be corrected by linguists. If these studies are done after specifying the Turkish dialect to be compared, they will give correct results.
Almost every 2 words written by the author were either shortened or written as if they were separate words by separating their syllables or phonemes, or the vowels were written incompletely or sometimes they were written with only consonants.
Moreover, some words were combined even though they should not have been combined, and some sounds were written by lengthening certain sound expressions, almost like a spoken song.
Thus, half or more of the texts were manipulated by the author to make the texts difficult to read. Therefore, if these will corrected more accurately by linguists for Lexicostatistics than correct results can be obtained.
We think that the use of texts in a Lexicostatistics study before these are completed will give misleading results. But there is another reason why we did not do a study for comparison in Lexicostatistics in detail.
We first made the ATA alphabet transcription for the Voynich texts by examining the old alphabets in terms of the Turkish language and blending them with the familiar characters of the Latin alphabet. This transcription enabled us to read the texts in Turkish to a great extent which no transcription has been able to achieve until today, despite the fact that the texts were written by multi-directional manipulation.
Therefore, after the texts are read in Turkish, we never thought to prove it again with using Lexicostatistics comparison to see whether it overlaps with Turkish or not. This can of course be done as a second validation after a while.
Moreover, none of the alphabet transcriptions that enable the reading of unknown historical alphabets and texts have been revealed by lexicostatistics studies. So there is no such example in history. The primary requirement for being able to read ancient texts is to make an alphabet transcription that works as a reading key. None of the transcriptions made to date, including EVA transcription, showed a voice equivalent to more than 100 characters of the Voynich alphabet.
They never predicted true multi-syllable characters and matched them with correct sound values. Moreover, they misread even simple alphabet characters. For example, they were read 600-year-old different VM letters with same phonetic value.
In our study, some letter signs can be read with a double voice, as like other known Old Turkish texts. While this is the case, no matter what method we compare the sound indexes without correctly predicting the compared sound values, it will not give correct results.
In summary, if we are going to do the study you mentioned, it will mean that we try to confirm the findings, which we have already confirmed with our own method, with the proposed method for the second time. This can of course be done. But first, let's get help from linguists to correct the texts, then this suggestion is tried if necessary and it shows overlap.
About the "High number of occurrences of EVA [p] and [f] in the toplines of paragraphs as opposed to elsewhere in a paragraph" as for the subject you mentioned that the signs perceived as P and F are actually a large number of very different syllable character signs.
In other words, if you/we look at them carefully, it will be seen that their number is more than two. So that the ones read as P are different from each other and the same situation is valid for those read as F.
EVA P and F are multiple syllable characters, each of which is different from each other.
No matter which structure in the text you/we want to examine through linguistic statistics, you/we must first find an alphabet transcription whose sound values must be correctly matched. This transcription work is not something that can be done by using eye and analogy or based on personal opinion only. Voynich alphabets are not located of the outside area of the historic alphabet evolution process. These VM letter signs not can be read with personal analogies far from science. EVA is just created with baseless personal analogies except for the well known Latin alphabet characters we use today. It has almost no scientific value that can be used in a scientific study.
Sorry my post was a bit long.
Thank you for reading