Posted by Steve Price on February 17, 1999 at 10:20:40:
In Reply to: Re: How We Know What We Know? posted by Alan Nagel on February 17, 1999 at 10:05:38:
Dear Friends,
Just to add a point to Alan Nagel's: virtually every attribution - date or geography or tribal - is ultimately probabilistic.
To illustrate, here are some trivial examples:
1. A rug has kepse guls as its major motif.
a. It is, therefore, very likely, almost surely, Turkmen.
b. It is quite likely, but not quite as surely, Yomud.
c. If, in addition, it has symmetric knots it's likelihood of being Yomud is increased. If it has assymetric knots open to the right, it is more or elss equally likely to be Yomud or Ersari (in the absence of additional information).
2. A Caucasian rug has what appear to be all natural dyes.
a. The likelihood that all the dyes really are natural depends on the experience off the person who is judging it, but is always less than 100% certain.
b. The likelihood that all the dyes are natural is higher if chemical tests confirm that, but even those tests are fallible, so we are still left with a probability rather than a certainty.
c. To the extent that we can trust the conclusion that the dyes are natural, we can make some age estimates, but those will include margins of uncertainty well beyond those of the dye identifications.
And so forth. What all of this means (I think) is that we routinely deal with uncertainties, sometimes we will be wrong, and it's fairly important that we recognize that fact.
Regards,
Steve Price