Marco, I appreciate all of your observations. I see them not as refutations of my theory, but as good questions that need to be further investigated and explained. I do not deny that much further investigation and explanation is certainly necessary. If I ever suggested or claimed otherwise, then that was a mistake and a misstatement, and I apologize if I gave anyone that impression.
My previous statement about "all of the statistical analytical properties" was clearly a misstatement on my part. The word "all" was of course inaccurate. I apologize for this misstatement. I did address a number of such statistical properties in my previous post, so I should have written a phrase along those lines in my summary description.
I very much appreciate nablator's research on the frequency of Eva [d] in lines without [p] or [f] vs. lines with [p] or [f]. I certainly did not mean to ignore it. I must have temporarily set the issue aside in the course of other research, and I inadvertently failed to get back to it to follow up. I apologize for this oversight as well.
Again, I view all of these questions Marco raises as important issues that need to be further investigated and explained. I believe that further investigation may well lead to further insights that are capable of explaining these issues, within the broad context of my hypothesis as I have presented it. I get the feeling that some researchers here may believe that the issues Marco raises mean that my theory should be rejected or abandoned. I on the other hand view the post as a very useful summary of difficult issues, which I can use as a basis and a guide for further research and investigation to identify possible explanations for these issues. Such research and investigation may take months, not a day or two.
I cannot reply to all of Marco's observations right now in this comment. But I can give a few brief comments and observations, while recognizing that much further research is clearly necessary.
1) First, I want to briefly address nablator's research on the frequency of [d] in lines with and without [p] or [f]. This is not intended as a full or complete answer or explanation, just as a brief initial comment on the issue. nablator showed that [d] is not more frequent in lines without [p] or [f] than it is in lines with [p] or [f]. I understand that this analysis
prima facie fails to confirm my hypothesis of [d] as a substitute for [p] and [f]. I will have to consider other or additional explanations for what is going on, in order to explain and make sense of this statistical data. For example, in general in the MS text, I have read and also observed myself that similar word forms tend to occur in more or less close proximity to each other; that is, one often finds variants of a word form with a letter or two altered or with other small changes, within the surrounding lines above or below the word form. I would describe this as a form of "character alliteration" in the Voynich MS text. We may not know for certain which actual letters or sounds each character represents, but whatever letters or sounds they are, they tend to occur in a repeated manner within a paragraph or passage or block of lines of text. I suggest that this may represent a frequent usage of alliteration and assonance on the part of the author in the underlying language of the text. If this is the case, then it would mean that lines with [p] or [f] may contain alliteration or assonance of the sounds that these characters represent. My theory proposes that these sounds are Greek /p/, /f/ ("ph"), /v/ ("b"); others may have other ideas. In this case, the use of alliteration or assonance would mean that the combined frequency of [p], [f], and [d] as a substitute for the same letters and sounds, would naturally be greater in lines containing this alliteration / assonance of labial sounds, and the combined frequency would be smaller in lines without such alliteration or assonance of labials. This would explain why [d], even as a substitute for [p] and [f], is still no less frequent in lines with [p] or [f] than it is in lines without [p] or [f].
Now the follow-up question would be, how then does one explain the particular greater combined frequency of [p], [f], and [d] in the first lines of paragraphs, as opposed to their smaller combined frequency in other lines of the text? All I can suggest for now is the following idea: Perhaps the author chose to use alliteration or assonance of labial sounds more often in the first lines of paragraphs in particular. One may then ask, is there any particular plausible reason one can think of to explain why the author would do this? I can suggest this idea: The elaborate and ornate characters [p] and [f] simply look better aesthetically in first lines, with more room to draw their tall shapes above the line, so the author preferred to employ alliteration and assonance of their type of sounds in those lines. I do acknowledge and accept Emma's point in the other thread that the author was capable of adjusting the size and height of glyphs, and so
could have used these characters and these sounds just as well and just as often in other lines if he or she had wished to do so. Still, for aesthetic reasons the author may have decided that these particular characters [p] and [f] look best when there is more room to draw them to their full height, and thus he or she was inspired to employ alliteration and assonance of their labial sounds more often in these first lines.
Still, just because such labial sounds would thus be less frequent in other lines, does not mean of course that the author would be able to avoid these sounds entirely elsewhere. For that purpose of representing these sounds in other lines, I propose that the author there usually chose to use [d] as a substitute for [p] and [f].
2) Concerning other issues of word structure and the line as a functional unit, I believe that my table at least provides the basis for a possible plausible explanation for many of these issues. Not only can [d] be used as a substitute for [p] and [f], but also likewise [s] is essentially a substitute for [k], and [q] is a substitute for [t]. (Also, although I have not added it to the table yet, I suspect that [g] and [m] are essentially most often substitutes for [y].) The possibility of all of these substitutions of alternate characters, without affecting the underlying language of the text at all, may possibly explain the existence of some of the more unusual features of Voynich word structure and line structure. Basically, having all of these optional alternate characters at his or her disposal as substitutions wherever and whenever desired, the author would have been free to create many unusual character distribution patterns, and again it would not have affected or changed the underlying language of the text at all.
Consider for example the virtual impossibility of double gallows (non-occurrence of any two characters of the set [k, t, p, f, ckh, cth, cph, cfh] next to each other, with only 6 isolated exceptions in the entire MS text), which Emma raised and discussed in a long-running thread on this forum. My theory and table offer a simple explanation: the author was easily able to avoid double gallows by always using an alternate character as a substitute for one of them instead. Now in Greek, many of these combinations will not often naturally occur anyway, but one such type of combination will occur indeed: "-pt-" and "-kt-", as in "hepta" and "okto". The author did not have to use double gallows for such combinations, because he or she could use [ps] and [ts] instead (usually in the form of [psh] and [tsh]). The combination [ks] (usually as [ksh]) for "-tt-" is also possible. Alternately, the author could write [dk] to represent "-pt-" or "-bd-" for example, which in fact I propose occurs in the word [chedkaly] that I identified as "hebdom[e]s" ("sevenths" or I suggested "sabbaths").
On the other hand, in Greek such combinations with "k" rather than "t" as the
second letter in the cluster will be much, much rarer. This is why we don't have words such as "hepcagon" or "otcagon" in our languages today. And here we find a possible explanation for the positional restriction of [q] in the Voynich MS text, which issue Marco raises in his post. Unlike [s]/[sh], the author didn't need to use the alternate character [q] in this cluster-final position, because unlike Greek "t", Greek "k" rarely occurs there.
As for the almost exclusive restriction of [q] to word-initial position, and the high frequency of [qok-] (by far the most frequent [q]-sequence) and [qot-] (by far the second most frequent one), I can also offer plausible explanations in line with my theory. It would be plausible that the author preferred to use the most prominent gallows character for the first consonant of the root of a word. So for example when the ubiquitous simple Greek conjunction "kai" ("and") is prefixed to a following word, the author chose to use [qo-], representing "ka'-", rather than [to-], so that the first consonant of the following root word would remain the most visually prominent character in the word, as the first and usually the only gallows character in the word. This way the little "ka'-" conjunction prefix would not "steal the thunder" of the much more important and significant first consonant of the main root word: the pride of place as the prominent first gallows character of the word was reserved for that first root consonant, and the smaller, simpler, less prominent [q] was used for the "k" in "ka'-" instead.
However, when the first consonant of the root word was not a gallows character anyway, this aesthetic issue was not relevant, so there the author did not necessarily need to use [q] as a less prominent alternate character for [t] in the prefixed conjunction "ka'-" ("kai"). Thus, the primary purpose and usage of [q] throughout the MS text is mainly to represent the "k" in "ka'-" before "gallows-initial" root words, in order to avoid the use of the "competing" prominent gallows character [t] as "k" in such positions. This explains the frequency of the occurrence of [qo-Gallows] that Marco raises the question about.
3) About word frequency, and the correspondence of the most frequent Voynichese words with the most frequent Greek words, of course I am highly focused on this issue. This is and always has been one of my main concerns when I even begin to investigate any possible hypothesis about the underlying language of the Voynich MS text. I do not recall that I have ever stated or implied that my theory does not allow for the representation of some prefixes and especially suffixes as separate words that appear as distinct in the manuscript. For example, let us go back and consider the final line of the four-line poem at the top of You are not allowed to view links.
Register or
Login to view. 1, which lines were in fact the very first passage of text that I posted on this forum and in this thread with a quasi-Judaeo-Greek reading and interpretation. If we go back and review my reading and interpretation of that final line, we will see that I read and interpreted the two separate, distinct Voynichese words "[soar cheey]" as the one single Greek word
δολιας ("dolias"), the genitive singular feminine form of the adjective
δολιος, meaning "deceitful" or "treacherous". To review all the steps, I read [soar cheey] as "doAl--iies", which I interpreted as "dol-ies", which I consider a plausible possible pronunciation of
δολιας. (In fact, if you listen carefully to a modern Greek pronunciation of this word, you will hear that the "a" is not pronounced as a European "a" sound, but rather much closer to an American English "short a" sound as in "cat". This latter sound is in fact phonetically closer to short "e" than it is to a pure "a".) Thus, it is fully in line with my theory, and with my readings and interpretations in practice to date, to interpret some of the most frequent Voynichese words as Greek suffixes / prefixes.
Entropy is of course also a vital issue to consider, discuss, and address. It is such a big issue that I will have to return to it and treat it separately in another post. There I will also include a discussion of the relevance of the "9-phoneme (11-letter) Greek text generation".
In the meantime, I am also going to continue to respond to requests such as Koen's, who asked about how my system handles labels such as those on You are not allowed to view links.
Register or
Login to view. , You are not allowed to view links.
Register or
Login to view. , You are not allowed to view links.
Register or
Login to view. , and You are not allowed to view links.
Register or
Login to view. (mainly labels of roots and leaves of plants). I have been working on such readings and interpretations in response to Koen's request, and I hope to be able to post something about this topic relatively soon.
Geoffrey