Phase models

[Chris Chinnock (Chris)]

Hi all, 

I have something niggling at the back of my mind with a model Im using for a site we're working towards publication on and I want some advice on whether the niggle is a problem.

Im working with a series of dates from a group of mesolithic pits which a loosely scattered across an area. They are all broadly comparable in terms of dimensions and fill sequence and so it seems reasonable to treat them as a group. I have built a phase model for all of the dates we have from the pits whilst also accounting for some stratigraphic relationships and occasional residuality. The estimated span for the phases is approx 2-2.5k years at 68%, which is a much longer span than I was expecting though apparently not unprecedented. 

My understanding of the phase model is that it assumes, in this case, that the pits were excavated/backfilled at a relatively constant rate before their use ended. My concern with such a long span of use that I cannot be sure that either the excavation or the deposition sequences were at all constant. Does this fundamentally undermine the validity of the phase model?

Thanks in advance

Chris 

[Thomas S. Dye]

Aloha Chris,

"'Chris Chinnock (Chris)' via OxCal" <ox...@googlegroups.com>
writes:

> My understanding of the phase model is that it assumes, in this
> case, that the pits were excavated/backfilled at a
> relatively constant rate before their use ended. My concern with
> such a long span of use that I cannot be sure
> that either the excavation or the deposition sequences were at
> all constant. Does this fundamentally undermine
> the validity of the phase model?

If I understand correctly, the phase model per se is not a
problem. Nevertheless, you have a choice of adding Boundary()
commands, or alternatively First() and Last() commands to
determine span of use. Boundary() commands will likely yield a
shorter span, but at the cost of the assumption that niggles.
First() and Last() don't make that assumption, but will likely
yield a longer span.

Note, too, that Boundary() commands will likely alter the
calibrated ages of some of the pits, but First() and Last() will
not. So, if the calibrated age of an individual pit is important
in your analysis, then you will likely want to use First() and
Last() commands rather than Boundary() commands.

My colleagues and I addressed this issue in Journal of
Archaeological Science vol. 153, "Bayesian chronology construction
and substance time". Our example was Anglo-Saxon female graves,
not Mesolithic pits, but I suspect you'll find stratigraphic
similarities.

Let me know if you have questions.

All the best,
Tom
--
Thomas S. Dye
https://tsdye.online/tsdye

[Chris Chinnock (Chris)]

Thanks Thomas, that's very helpful. I'll take a look at that paper and come back to you if I have any questions.

Cheers

Chris