Bayesian analysis (Re: Qayafa chronology )
- --- In ANEemail@example.com, Graham Hagens <rgrahamh@...> wrote:
>[The URL of the Manning article is
> One of the points of divergence involves the utilization of
> the technique known as 'sequence analysis' which Manning and
> co-authors have used extensively in the past.
> The weakness of this approach is that it sometimes requires the
> author to make assumption about the duration of particular
> archaeological strata, or sequences, 'for which there may little
> external evidence.
> I would argue that utilization of 'single point' data is less
> prone to interpretive errors.
> Graham Hagens
> --- On Thu, 4/30/09, BjÃ¶rn Lindborg <bjorn07se@...> wrote:
http://dendro.cornell.edu/articles/manning2007b.pdf (no spaces!).]
> It's interesting to note that the high and low schools inHi Graham (and All),
> Palestine Iron Age chronology (Mazar and Bronk Ramsey vs. Sharon
> and Gilboa) arrive at their resp. positions by analysis of
> basically the same raw 14C dates.
> [I think the Radiocarbon article by Mazar and Bronk Ramsey (Vol.
> 50:2 pp. 159-180; August 2008) is not yet available on the web.]
> Björn Lindborg
Manning and co-authors are not the only ones that use 'sequence
analysis' (Bayesian analysis). It seems to be the usual method when
for some reasons, single 14C measurements don't give decisive
results, e.g. when the calibration curve is 'wiggly' because of
fluctuating atmospheric 14-CO2 levels in the past. This is the case
at ca. 1250-1120 BC and was probably one main reason for Manning to
write his article. We see this in his Figs. 8 and 9, "Calibrated
calendar age ranges for the average of the 20 simulated radiocarbon
ages for 1200 BCE (2961±7 BP)". Even an absurdly tiny error of ±1,
thus 2961±1 BP above, still leads to a similar calibration outcome:
12511240, or 12121190, or 11771160, or 11431131 BCE at 1sigma.
The spokespersons for the 'Low' Palestinian Iron Age chronology
(Boaretto et al., "Dating the Iron Age I/II Transition in Israel"
in Radiocarbon, Vol 47:1) also state:
<< In the present study, we employ the standard Bayesian inference
approach (Bronk Ramsey 1994, 1995). >>
Finkelstein uses a similar method in his criticism of the 'High'
Tel Rehov chronology.
In such 14C catastrophy areas, single measurements may be 'less
prone to interpretive errors' as you write, but also less reliable
for absolute dating purposes because of the wide ranges obtained.
- The material assemblage (primarily pottery) that I have seen from Kh. Qayafa has types that continue late Iron I types (but not all types and relatively few types typical of "Philistia" in the late Iron I), but at the same time, has various types that seem to indicate an early Iron IIA dating (but many typical types of the slightly later Iron IIA are missing). From what I have seen, the assemblage is different from contemporary assemblages in Philistia (e.g., Tell es-Safi/Gath) AND from the far too little we know from late Iron I/early Iron Age II Judah (such sites as Jerusalem/City of David [yes, there is this phase in Jerusalem, despite what is often written], Kh. Dewara, etc.).
Until more is published from the site, we can all talk ourselves blue in the face, each sticking to his/her own presuppositions.
And yes, to get good 14C dating (one can that "prove" or "disprove" this or that theory), one would need a series of stratigraphically-sequenced clusters of single year cultigens, all from extremely well-defined and secure contexts. Statistical manipulation is only secondary to this.
At times one get the feeling that we all suffer from: too little secure data - too much talk...
"Telling it at Gath" ...