The Voynich Ninja

Uncalibrated years are nothing else than a different way of referring to the percentage of C-14 contents.
The relation between the two is linear in the time frame of interest.

Averaging the uncalibrated years is the same as averaging the C-14 percentages.

Furthermore, the uncertainty in a date of any sample is far greater than the typical duration of creating a book. Of course, there are exceptional cases where this could have taken decades, but there is no particular reason to believe that this is the case for the Voynich MS.

What most people don't realise is that we have been very lucky. The best guess for the origin of the book, at the time that the experiment was prepared, was 1460-1470-ish. Had this been true, the result would have looked very different indeed, because the calibration curve is much flatter here.

When I have more time, tomorrow, I can show an example of such a case, and the rather horrible date estimate that may result from it.

The fact that the calibrated uncertainties are much smaller than the uncalibrated ones is exactly a consequence of the very favourable slope at the start of the 15th century.

(16-05-2021, 06:07 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.Now, while I myself used to like the idea that we deal with a "single object", I have to admit that there are arguments towards the opposite, namely that this is not a single object, but rather a collection of different objects which may have been produced at different time, and bound only later.

That brings the whole thing to the point. This could fundamentally influence my view of the VMS. Until now, I have also always assumed a "single object".

(16-05-2021, 08:09 PM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Furthermore, the uncertainty in a date of any sample is far greater than the typical duration of creating a book.

Rene, the point that Rich raises in this thread is that it's not 100% certain that the whole "project" was to create a single book. This is in the vein of what Wladimir suggested recently that the VMS was not designed to be a single, let's say, "opus", but rather a set of more or less disjoint manuscripts collated by a later owner into the thing which we now consider a uniform "book".

Thanks for the clarifications, insights and explanations. Interesting and informative discussion.

(16-05-2021, 08:09 PM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.Furthermore, the uncertainty in a date of any sample is far greater than the typical duration of creating a book. Of course, there are exceptional cases where this could have taken decades, but there is no particular reason to believe that this is the case for the Voynich MS.

As Anton points out above, and as I have repeatedly, this idea of limiting the interpretation of these results only in light of "one book, one duration" excludes many other possible reasons for the individual dating being what we see here. Not just for my case, but for untold others. Possible useful evidence is removed by this assumption, this averaging. Even if the Voynich is from, say, 1450, and genuine, the use of a wider date range of vellum for its construction may be a useful clue. Average those dates, and the clue is gone.

(16-05-2021, 08:09 PM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.The best guess for the origin of the book, at the time that the experiment was prepared, was 1460-1470-ish.

I understand what you mean about that, that some date ranges could have been problematic. But as for the dating range you cite, I don't believe this is the case. If one looks at the mass of pre-C14 opinions, they were all over the calendar. I'd made a couple of lists of them, and remember well our past discussions over those opinions. What happened after the testing put the vellum in the early 15th century was a "weening" of all previous expert and amateur opinions which did not fit that dating. Eventually, the reality of so many experts getting it wrong... the majority did get it wrong... morphed into the present claim that the "experts got it right".

The list I compiled from D'Imperio shows this well:

Wilfrid Voynich- late 13th century
Romaine Newbold- late 13th century
Theodore C. Peterson- 13th to 14th centuries
John H. Tiltman- 13th century
Hugh O’Neill- late 15th century (on)
Helmut Lehmann-Haupt- Early 15th century
Erwin Panofsky- Late 15th century, changed to early 16th century (1510-1520)
Elizebeth Friedman- late 15th, early 16th century (1480-1520)
A.H. Carter- “far later” than 13th or 14th century
Dr. Charles Singer- early 16th century (1520 or later)
Leonell Strong- 16th century (1525 or later)
Robert Brumbaugh- 16th century (1500)
Professor Giles Constable- 16th century
Professor Rodney Dennis- 16th century
Dr. Franklin aLudden- late 15th to mid 16th  century (1475-1550)
Sergio Toresella- late 15th century (1460-1480)

The more modern... post D'Imperio... trend from the middle 16th to the early 17th century. So when the claim is made that the expert opinions, who whomever, were in the early 15th century, and therefore the dating shows the experts were correct; and conversely, then, the dating backs up expert opinion... no, that was not the case at all. The C14 dating was a surprise, and well out of the mass of expert opinion. In fact, that was how it was first reported when the dating was released. The pruning of the experts to fit the results happened soon after that, though, so the story changed.

(16-05-2021, 08:09 PM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.The fact that the calibrated uncertainties are much smaller than the uncalibrated ones

If we use calendar dates so that sigma is 15 rather than 35 that surely changes my perception of the consistency among the results. Except that the range of the calendar dates for the individual samples also narrows, to 1410 for the oldest and 1426 for the newest!

Why do we calculate the calendar dates and then seem to ignore them? Is there an a priori reason for this?

I believe the problem with those expert datings may have been the same as we are discussing here in relation to carbon dating: the experts dating the VMS from the "single object" perspective, while in fact they'd rather date various portions thereof.

I say "may have been" because I'm not sure if it's really been thus in all the cases. In fact, it is not uncommon that a single-bound manuscript contains writings even from different centuries, so I would be surprised if all aforementioned experts tried to impose a "single" date on the whole VMS.

(16-05-2021, 09:30 PM)DONJCH Wrote: You are not allowed to view links. Register or Login to view.f we use calendar dates so that sigma is 15 rather than 35 that surely changes my perception of the consistency among the results.

The "15" figure of the calibrated PDF is not its sigma (standard deviation) anymore. It's just half the interval of 68,2% probability, and nothing more. The calibrated curve is no longer Gaussian, instead it's some weird stuff, two-modal, to make it worse.

(16-05-2021, 09:42 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.The calibrated curve is no longer Gaussian

Yeah but it's close enough. What I did was a hack to get a rough estimate. Look at the rest of the post above though, the means for the individual samples become closer after calibration too.

I am approaching this from the POV of a lab person to see if the report was justified and I agree with the author and Rene that there are no outliers in the results.

Not that I disagree with your proposal, but these results are not good evidence for that.

Every measurement of a single page is a process that has its inherent inaccuracies. Note that the measurement for a single page is derived from several shots based on small parts of the individual sample. The spread of these shots contributes to the uncertainty of the measurement.
This single measurement is the ratio of two ratios:
(C-14 / C-12) as measured , divided by: (C-14 / C12) for 1950.
This single value is of the order of 0.94 for the various sheets in the Voynich MS.
This single value has an uncertainty or error distribution which can be assumed to be Gaussian in the area of interest.

When computing the calibrated dates, one computes a convolution with the 2-dimension probability density function of the calibration curve. The result is no longer Gaussian at all.

It is the calibrated date that is of interest, and that is used at the end result. I don't quite understand the question from DONJCH in post #25.

Even if the MS is several parts, that we not all produced not at the same time, what would be the time difference? One year? Five years? 25 years? 

In the first two cases, combining the results is completely valid.
The book looks consistent enough, from beginning to end, to make a much longer duration rather unlikely.

Now keep in mind that the end result still allows 5% probability that the date is outside the published range. This 5% is considerably greater, in my humble opinion, than the probability that it took well over 5 years to produce this MS.
Furthermore, the results show (to the extent that they can) that the parchment seems to have been acquired within a short time span.

This is all about statistics, and one should not mix in things that have a very small likelihood of being correct.

The manuscripts too, do not allow for a greater time gap. (Handwriting)

And if.....even 50 years would not change much. There is still nothing where one would have to seriously rethink.