effect of site-level phases on cultural transition boundaries
norbu
Hello everyone, I am new to OxCal, and apologise for a basic question.
I am currently modelling a cultural transition and would appreciate clarification on how OxCal estimates transition boundaries in hierarchical models.
My main question is: how does the incorporation of site-level phase boundaries influence the estimated transition boundaries between two cultural occupations? Specifically, are the transition boundaries between Culture 1 and Culture 2 derived solely from the combined distribution of all dates within each cultural phase, or are they indirectly computed by the start and end boundaries of individual sites nested within those phases? If so, how does this hierarchical structure affect the posterior estimates of the overall cultural transition interval?
I would like to understand how this differs from a simpler model in which all radiocarbon dates are placed directly within cultural phases without explicit site-level structuring.
Thank you so much!
Here is a sample code:
Sequence()
{
Boundary("Culture 1 start");
Phase("Culture 1")
{
Sequence()
{
Boundary("Site 1 start");
Phase("Site 1")
{
R_Date(...);
};
Boundary("Site 1 end");
};
Sequence()
{
Boundary("Site 2 start");
Phase("Site 2")
{
R_Date(...);
};
Boundary("Site 2 end");
};
};
Boundary("Culture 1 end");
Boundary("Culture 2 start");
Phase("Culture 2")
{
Sequence()
{
Boundary("Site 3 start");
Phase("Site 3")
{
R_Date(...);
};
Boundary("Site 3 end");
};
};
Boundary("Culture 2 end");
};
Christopher Ramsey
- the top level phases (or phase) contain all of the un-nested events and the boundaries of any subsidiary groupings. So in your phase 'Culture 1' for example, there are effectively just four events the start and end of the two site phases. At this top level the subsidiary boundaries behave just like events although always orders.
- the subsidiary phases are essentially independent and just constrained to lie within the outer groupings.
The effect of this can be quite marked because the subsidiary phases might typically have many more dated events. For example if Site 1 and Site 2 each have 20 dates and you treated this as a simple un-nested model, phase "Culture 1" would have 40 dates associated with it - whereas in the nested model it only has 4. This is likely to mean the the precision of the boundaries for the outer cultures is likely to be lower. However, assuming you only have two example sites - each of which might be different - this is actually more realistic than assuming that the 40 dates are all independently sampled from "Culture 1" and are truly representative of the phase as a whole.
Best wishes
Christopher
> You received this message because you are subscribed to the Google Groups "OxCal" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to oxcal+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/oxcal/7be31217-b254-46cd-8c7e-911609ffb4ccn%40googlegroups.com.