The Voynich Ninja

(13-03-2019, 03:10 AM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.On the first point only: can we statistically show that [p, f] + [d] on the first line of paragraphs is the same as [d] in other lines? Assuming that your proposed values are evenly spread throughout the text, we should still see lower rates of [d] where [p, f] occur.

Emma, to answer this question requires a complete and detailed statistical analysis which will take some time and effort to perform. It is a simple matter to note that [p, f] tend to occur in the first lines of paragraphs but not elsewhere: clearly the first lines only constitute a limited minority of the total text, and it is equally clear that a substantial majority of [p, f] occur in these lines. But to analyze the relative frequency of [d], in these first lines vs. elsewhere, is a more complicated undertaking. I will work on it, but it will take some time.

For now, however, I will note a couple other observations I have made, just in my initial investigation into this question:

I see a tendency in the initial pages of the ms for [e] and [ch] to occur less frequently on the first line of paragraphs than they do elsewhere. Conversely, [sh] seems more likely to occur on the first line of paragraphs than it does elsewhere.

Now I hasten to add that this is just an initial provisional observation based on limited data. Nevertheless, I think it is worth mentioning.

How about just adding up all [p], [f], and [d] for each line and creating one average for all lines with some [p] or [f] and another average for those lines with none? If the former is significantly higher than the latter then we have our answer.

quote: I propose that [d] represents a distinct letter in its own right

I fully agree, and confirm that is correct. 

(14-03-2019, 03:18 AM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.How about just adding up all [p], [f], and [d] for each line and creating one average for all lines with some [p] or [f] and another average for those lines with none? If the former is significantly higher than the latter then we have our answer.

I have just observed a couple additional possible positional preferences of [d] that may significantly impact the statistics:

Looking at You are not allowed to view links. Register or Login to view. , 2nd paragraph (bottom of the page), 1st line, will illustrate the potential tendencies well:

1) [d] may tend to occur (as a substitute for [p, f] ?) within a word in which a gallows character has already appeared in the first part of the word. 

On this line of You are not allowed to view links. Register or Login to view. , we see two examples: the first word [kydain] and the sixth word [fodan] (the last word before the plant illustration breaks up the line).

2) [d] may tend to occur (as a substitute for [p, f] ?) toward the end of first lines more than toward the beginning of first lines.

On the same line, we see [d] occur twice in the very last word [daiildy].

These same two tendencies occur in You are not allowed to view links. Register or Login to view. , first line: 
1) In the first word [kshody], following a gallows character in the first part of the word.
2) [d] occurs in each of the last three words on the line.

So to be more precise in this analysis, we would need to separate the beginnings and ends of the first lines of paragraphs, and we would need to distinguish between occurrences of [d] in words with a gallows character vs. [d] in words without one. Only the occurrences of [d] in the beginning parts of first lines, in words without any gallows characters, would rule out the three possible environments in which I conjecture that [d] may substitute for [p, f] in the ms text.

I do note as well that there may also be a phonological explanation of tendency (1) above in the underlying language. If my conjecture is correct that [d] may represent the /v/~/w/~/u/ phonemes, and [p, f] may represent phonemes such as /p/, /b/, /f/, then there would be a natural tendency for [p, f] to occur more frequently in word-initial position, and less of a tendency for [d] to do so. The statistics about each consonant phoneme's frequency in initial vs. non-initial position in the global 50-language study in Carsten Peust's "On Consonant Frequency in Egyptian and Other Languages" ( You are not allowed to view links. Register or Login to view. ) confirm the word-initial tendency for /p/, /b/, /f/ vs. the lack of such a strong tendency for /v/. It is even possible that a sound change or allophonic alternation process in the underlying language, such as betacism (merging of the phonemes /b/ and /v/), caused the scribe to more often substitute /v/ for /b/ in non-initial position.

(14-03-2019, 12:03 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.quote: I propose that [d] represents a distinct letter in its own right

I fully agree, and confirm that is correct. 

I propose that d is more frequent on lines that contain (or start with) p or f.

(14-03-2019, 03:20 PM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view.So to be more precise in this analysis, we would need to separate the beginnings and ends of the first lines of paragraphs, and we would need to distinguish between occurrences of [d] in words with a gallows character vs. [d] in words without one. Only the occurrences of [d] in the beginning parts of first lines, in words without any gallows characters, would rule out the three possible environments in which I conjecture that [d] may substitute for [p, f] in the ms text.

Excellent! We await your results.

(14-03-2019, 05:13 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.

(14-03-2019, 12:03 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.quote: I propose that [d] represents a distinct letter in its own right

I fully agree, and confirm that is correct. 

I propose that d is more frequent on lines that contain (or start with) p or f.

If that is true, no conclusion or implication would reveal us anything. If you have a suggestion I would love to hear that.

As written before, somewhere, I made an analysis of all initial Page letters and words + initial Paragraph letters, in the VMS as well as in Culpeper's herbal.
Based on that I could not form any valid hypothesis.

On the [f and p] I tried to publish and start a discussion here a year ago, but some people were furious that I wanted to start a hypothetical discussion on letters.
So, then I immediately removed the thread. Anyway, it is useless to discuss letters here, because there are so many different ways to look at them (as you can already see)
In my opinion, every participant in the discussion is obligated to read and understand the field of all other participants. Otherwise everybody is preaching to the choir.

To me:  [d] is not "dependent". 

(In my "letter dependency research" -unpublished because still ongoing- most angles, such as, letters, words, positions, etc. were investigated, based on that I formed this conclusion.)

(15-03-2019, 02:39 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.If that is true, no conclusion or implication would reveal us anything. If you have a suggestion I would love to hear that.

A quick-and-dirty count gives the following results:

To remove formatting from TT_ivtff_v0a.txt in paragraphs:
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep "^[pf]" > ttpfstart.txt
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep "[pf]" > ttpf.txt
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep -v "[pf]" > ttnopf.txt

To compute the frequency (as a percentage) of "d" among EVA characters including "." in paragraphs:
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttnopf.txt
5.81984
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttpf.txt
6.21357
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttpfstart.txt
6.24086

(length-1 because I am using these UNIX tools on Windows: every line has a cr beside lf.)

These small frequency variations (5.8% in lines without p or f vs 6.2% in lines with p or f) do not tell us much, but they certainly do not suggest that d is a good candidate for a substitution of p and f.

A more plausible hypothesis IMHO than a substitution would be that p and f are inserted (no idea why, and reading discussions about "Neal keys" did not help). The (weak) evidence for this is that removing p and f very often yields words that exist elsewhere in the VMs.

After removal of p and f in ttpf.txt the frequency of d is 6.45%.

(13-03-2019, 02:47 AM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view....

* If [d] represents /u/~/v/, which would also be the labial glide /w/, and [ch] represents the palatal glide /j/, then it would be understandable that the two distinctly different glides cannot naturally be written in combination together.

Hi Geoffrey,

Thank you for your discovery and detailed analyze of texts.

I am a bit confused by this line of statement.

What do you mean by something like /u~v~w/ cannot co-exist with /i~j/? I could easily remember two very basic English words which are made from those sounds: we and you. The former is /u~v~w/ plus /i~j/, and the latter is /i~j/ plus /u~v~w/.

(15-03-2019, 03:34 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.

(15-03-2019, 02:39 PM)Davidsch Wrote: You are not allowed to view links. Register or Login to view.If that is true, no conclusion or implication would reveal us anything. If you have a suggestion I would love to hear that.

A quick-and-dirty count gives the following results:

To remove formatting from TT_ivtff_v0a.txt in paragraphs:
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep "^[pf]" > ttpfstart.txt
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep "[pf]" > ttpf.txt
grep ",.P" TT_ivtff_v0a.txt | cut -c18- | sed "s/<->/./g;s/<$>//g;s/ //g;s/\.$//g"| grep -v "[pf]" > ttnopf.txt

To compute the frequency (as a percentage) of "d" among EVA characters including "." in paragraphs:
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttnopf.txt
5.81984
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttpf.txt
6.21357
awk -F "d" "{ll += length-1; cc += NF-1} END {print 100*cc/ll}" < ttpfstart.txt
6.24086

(length-1 because I am using these UNIX tools on Windows: every line has a cr beside lf.)

These small frequency variations (5.8% in lines without p or f vs 6.2% in lines with p or f) do not tell us much, but they certainly do not suggest that d is a good candidate for a substitution of p and f.

A more plausible hypothesis IMHO than a substitution would be that p and f are inserted (no idea why, and reading discussions about "Neal keys" did not help). The (weak) evidence for this is that removing p and f very often yields words that exist elsewhere in the VMs.

After removal of p and f in ttpf.txt the frequency of d is 6.45%.

As I wrote yesterday in the other thread about my theory, I want to express my belated thanks and appreciation to nablator for performing this very helpful statistical analysis. It clearly shows that [d] is not more frequent in lines without [p] or [f].

However, I still believe [d] may be a substitute for [p] and [f]. I believe the explanation may lie in the author's use of alliteration and assonance. The sounds that these letters represent may simply tend to appear very frequently clustered together on the same lines, as an expression of this alliteration and assonance. In this case, [d] may be a substitute for [p] and [f], and still [d] may appear just as frequently or even perhaps more frequently on lines with [p] or [f] than on lines without [p] or [f].

This idea also suggests that the author had a preference for employing alliteration and assonance with these particular sounds in the first lines of paragraphs. A possible reason for this would be that the author simply preferred the aesthetic appearance of the tall, elaborate, and ornate characters [p] and [f] on the prominent first lines, and so was inspired to choose their sounds for alliteration and assonance on these lines more so than elsewhere.

ChenZheChina, I am not trying to ignore your comment either, thank you for the comment and the question, I will do my best to get back to you and answer as soon as I can.

Geoffrey