geoffreycaveney's Judaeo-Greek theory

[geoffreycaveney]

(continued)

3. [d] as a substitute for [p] and [f]

From You are not allowed to view links. Register or Login to view., it seems that you propose d as a possible (but not purely) substitute for p and f, because of:

• d may be representing /u~v/, while
  
• p and f may be representing /b~f~p/, and
  
• Betacism may caused the writer to write /v/ for /b/, and from where it was extended to writing /v/ for /f~p/
  

The last line does not seem likely to happen to me, but I cannot say it’s impossible. However, Emma then posted these statistical facts:

• [d] is common on the first line of a paragraph, which would need to be explained.
  
• [p, f] are shorter when written away from the first line of paragraphs and [k, t] are taller when written on the first line. The scribe was capable of adjusting the glyph height.
  
• [d, l, r, s] all lack significant number of [e] following them.
  
• [p, f] take relatively good numbers of [ch, sh] following them, which [d] does not, especially away from the start of the line.
  

So, it seems that the d cannot replace p and f freely, as they have distinct stats.

Again, the question arises in the first line of You are not allowed to view links. Register or Login to view., where the same Greek word είπαν (eípan) appear twice, but the same Greek letter π (p) are written in Voynichese as d and p, respectively.

Since the stats are different for d, p and f, it would be great if we could have a general rule about when to use d. I guess the final answer must be based on very large amount of texts, so that the rule could be general enough to cover most cases. Therefore, I’m not expecting to see an answer soon, but when you have finished translating the whole manuscript, there should have already been an answer to this question.

(to be continued…)

Dear Chen Zhe:

Thank you again for raising this question, which is clearly a critical issue, as Emma, Marco, nablator, and others have pointed to as well.

Your observation of the contrast of the same word in the same line written once as [teeodaiin] and once as [epaiin] prompted me to consider and test the following hypothesis:

[d] is written as a substitute for [p] or [f] in the following cases:

(1) anywhere other than the first line of a paragraph

(2) in any position in which one of the gallows characters [k], [t], [p], [f] has already been written earlier in the word

To test this second and latest hypothesis, I performed the following statistical analysis: 

Using the data available at voynichese.com, I calculated the total number of Voynich words in which each significant Voynich character [a, ch, d, e, f, k, l, o, p, r, s, sh, t, y] occurs in medial position, using the string [*x*]. (Medial position is relevant because [p] and [f] are rare in final position in any case, and for this hypothesis we are only considering a position somewhere after a gallows character in the word.) 

Then I calculated the amount of such words in which each character occurs in medial position somewhere after, but not immediately after, a gallows character in the word, by using the strings [k*x*], [t*x*], [p*x*], [f*x*], and adding the four values together. ("Not immediately after" is relevant, since we are investigating substitution for [p] and [f], and we know from the work of Emma and others that two gallows characters almost never occur adjacent to each other in a word in any case in the MS.)

Then for each character I divided the latter value by the former value, to produce a percentage of medial occurrences of each character that occur somewhere after, but not immediately after, a gallows character in the same Voynich word.

Here are all the data on all 14 of these Voynich characters for each of these three calculations:

medial [o] : 14441
medial [e] : 14209
medial [a] : 11749
medial [k] : 9633
medial [d] : 8606
medial [t] : 5864
medial [ch] : 4933
medial [l] : 3198
medial [r] : 1299
medial [sh] : 1243
medial [p] : 1076
medial [y] : 471
medial [s] : 445
medial [f] : 360

-------

5236 [gallows*e*]
4705 [gallows*d*]
3053 [gallows*o*]
1641 [gallows*a*]
899 [gallows*l*]
465 [gallows*r*]
451 [gallows*ch*]
248 [gallows*k*]
167 [gallows*sh*]
139 [gallows*s*]
127 [gallows*t*]
111 [gallows*y*]
45 [gallows*p*]
41 [gallows*f*]

-------

gallows*[x]* / all medial *[x]* for each character:

[d] : 54.7%
[e] : 36.8%
[r] : 35.8%
[s] : 31.2%
[l] : 28.1%
[y] : 23.6%
all glyphs : 22.3%
[o] : 21.1%
[a] : 14.0%
[sh] : 13.4%
[f] : 11.4%
[ch] : 9.1%
[p] : 4.2%
[k] : 2.6%
[t] : 2.2%

-------

We see from the data that medial occurrences of [d] after, but not immediately after, a gallows character in the same word, constitute a much higher proportion of all medial occurrences of [d], more than half of them, than is the case for any other Voynich character.

We also see that [p] in particular is extremely rare in medial position after, but not immediately after, a gallows character in the same word, with only 45 of its 1076 total medial occurrences being in such a position.

The data is so extreme, in fact, that the question becomes not whether [d] is a substitute for [p] and [f] in this position, but rather, whether [d] is not also a substitute for all gallows characters in this position!?

However, I hope that is not the case, because that would be taking the ambiguity of the value of [d] a bit too far, to the point where we might never be able to hope to read and correctly interpret the text.

Rather, we also see that [s] is quite frequent in the same such position as well, and my theory already proposes that [s] is an alternate character for [k].

Accounting for the rarity of [t] in this position is more challenging, but I may propose the following: 

First of all, in Greek the letter kappa is not nearly so frequent in medial position, or as the second consonant of a root, as tau is. Words and roots like "kata" and "okto" are common; words and roots like *"taka" and *"otko" are not. That's why we have Greek borrowings such as "category" and "octagon" in our languages today, but not words like *"tacagory" or *"otcagon". (Note: in my analysis of Voynich word breaks, the second parts of words such as these could be written as separate Voynich words, so that "-gory" and "-gon" would be distinct Voynich words, and the Greek gamma would still occur as the first consonant of the Voynich word, not as a medial letter somewhere after another Voynich gallows character consonant in the same word.)

Second, in some cases we may have Greek gamma pronounced as a y-glide, as in "gyro", in which case it could be written in the Voynich script as [e], as I have already accounted for in my letter correspondence table. In this case, the relatively high frequency of medial [e] somewhere after, but not immediately after, a gallows character in the same word, as my data presented above clearly show, could also be the result of the substitution of [e] for [t] to represent Greek gamma.

The main point of this study, in any case, is to highlight the environment of medial position after a gallows character in the same word, in which position [p] and [f] are very rare, and in which position [d] is extremely common, much more so than any other character in the Voynich MS, in comparison to all of its medial occurrences. The substitution of [d] for [p] and [f] in this position would explain this data, and it would also explain additional occurrences of [d] in the first lines of paragraphs, where we would otherwise expect more [p] and [f] and fewer [d], but where we in fact still find plenty of occurrences of [d]. 

Geoffrey