Questions about the utilization of the Oxcal program

[은채]

I have a question regarding the command utilization of the Oxcal program.

Hello, I am an archaeology major student using the Oxcal program in South Korea. I would appreciate it if you could understand with a generous heart even if I am not good at English.

I'd like to use Oxcal's R_combine command, is there any problem if I use the carbon dating value from various residences of the ruins as one? If there is a problem, I'd like to know why.
(For example, if 10 residences were excavated from ruins A and one carbon solidarity value was acquired for each dwelling, I wonder if there is any problem in considering it as the representative carbon solidarity value of ruins A by combining the carbon solidarity values from 10 dwellings.)

I wonder if the R_Combine command simply combines several carbon dating values statistically or can only be used for the same carbon dating sample value.

In addition, if an outlier exists, the outlier command is used, but there seems to be a problem that the subjective part of the researcher is included when setting the value for removing the outlier. Will the existence of a residence be considered when using the command? If the part of existence is considered, I ask because I think it can be used.

I'm posting a question here because it seems like an open group because I have a problem to understand it even after reading several papers! This may be a humble question, but I'd appreciate it if you could answer it. Thank you.

[Christopher Ramsey]

The R_Combine function really only makes sense, in modelling terms, for combing dates on material which has the same radiocarbon composition. Most obviously this could be several measurements on the same sample - but it could also be, for example measurements of different seeds in the same storage jar all thought to be from the same harvest.

The measurements are then combined BEFORE calibration ensuring that the uncertainty from calibration is maintained. This is why you should not use Combine in such situations which would result in a spuriously high precision. Combine can be used where you have for example one radiocarbon date and one date using a different independent method.

In most other circumstances the samples dates are likely to span more than one growing season and so, although you may not be able to see it within the precision of the measurements, will have different radiocarbon compositions. They ultimately relate to events spanning time.

If you have events spanning time then you are really dealing with a phase of events and treating them as a single event does not make sense for age modelling.

You might have the situation where you think they are all related to one major event - eg a destruction event which relates to multiple buildings. In this case the event you are interested in is the end of the phase and will not be well represented by an average of all the dates in the context. How you treat such a situation will depend on how you think the samples spread in time. If you think they are evenly distributed through a period up to the final event you would use a simple uniform phase model:

Sequence()
{
Boundary("Start");
Phase("1")
{
R_Date("A",...);
R_Date("B",...);
R_Date("C",...);
...
};
Boundary("End");
};

with the End boundary being the event of interest. If as is often the case you expect the events to be clustered at the end with a long tail of earlier residual material then an exponential distribution might be more appropriate:

Sequence()
{
Tau_Boundary();
Phase("1")
{
R_Date("A",...);
R_Date("B",...);
R_Date("C",...);
...
};
Boundary("End");
};

These approaches do need a significant number of dates to work well.

The other way in which R_Combine is used is just to check the statistical test (Ward and Wilson 1978) to see if you can say at 95% confidence that they are NOT all the same age. If it passes the test it does not mean that they are all the same age though - just that they could be. In other words it does not mean they should be included in an age model in this form. The evidence that they are all the same age needs to come from the context and not from the measurements.

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 on the web visit https://groups.google.com/d/msgid/oxcal/9c0d003f-ad85-4d74-a85c-eb3f8dbf97ben%40googlegroups.com.