RE: Storehouses of parchment
R. Sale > 31-08-2020, 12:34 AM
ReneZ,
Thanks again. Discussion helps to develop an improved perspective. As I understand it, the early 1400s is a relative 'sweet spot' for C-14 data with the values involved for the margins of error being notable smaller than during other periods of time. That certainly helps.
The real question is, from the four samples analysed, how many different production batches of parchment are there? The answer is unknown. Even if the range for a single batch of production is extended from annual to a five-year period, there would still be six different batches in a 30 year span. The C-14 test is not going to be sufficiently accurate to tell one end from the other, or anywhere in between. The test simply lacks resolution and cannot discriminate[font=Tahoma, Verdana, Arial, sans-serif] between individual batches with such proximity. Yet the actual difference could really be more than just a couple of years. The mathematical results can be subjected to several interpretations.[/font]
With the four VMs samples, the dates are close enough, and the degree of imprecision is great enough to permit a valid statistical analysis *as if* all samples came from a single source. All is well and good. However, the data is such that the most recent sample is something of an outlier, and as such it is the primary candidate for production by an alternate source, and in this case an obviously more recent parchment source. Calculation based on a single-source assumption gives a good answer. However the C-14 inability to discriminate between batches made more than a decade apart means that the dual-source interpretation is also a valid possibility. And this version of dual-source interpretation is going to have the most significant effect on the potential chronology.
While initial VMs production may have occurred over weeks, months, or years, production cannot be completed before the final parchment date. Hence the effect of this dual-source possibility depends significantly on the date of the most recent sample. And this sample has the potential capacity to move the earliest date of VMs final production to a year that is somewhat later than the single source average indicates (1438) - to a date that is closer to 1450, and perhaps a bit beyond.
As you said, "[font=Tahoma, Verdana, Arial, sans-serif]The probability that any single sample is after 1450 is negligible." That would mainly be the one that is the most recent of the four samples. And negligible is not zero. So I can agree with that. There's more than one way to interpret the data, and current evidence is insufficient to confer validity on a particular method.[/font]
[font=Tahoma, Verdana, Arial, sans-serif]What do you say to 1434-1445+ as a date for the last sheet of parchment?[/font]