Dating a burn event?

[Erik Marsh]

Hi all – 

I want to date a burn event. I have 13 dates with a clear stratigraphic set up: the first 7 are from lower levels dating to the site's construction and the second 6 are from an overlaying 2-m ash layer from when the site was burned down.

However, the dates do not follow the stratigraphy, nor are the 6 dates from the ash layer very similar, which I was expecting (they fail a chi-square test). They are all samples are from building materials (grass reeds and timbers), which could have been harvested well before the burn event. The one very early date (1450±90) could be old wood; the agreement is low but it is not an outlier. (Outlier models have not helped much. I am trying to avoid removing dates arbitrarily). However, this does not explain the two late dates – the only thing I can think of it contamination from later use of the site.

Anyway... what's the best way to model the burn event? For now I have a single-phase model that ignores the stratigraphy and assumes all the dates are from when building materials were harvested. The ending boundary is the burn event (there are no dates from after the burn).

Thanks for any suggestions!

Erik

  Sequence()

      {

       Boundary("Start");

       Phase()

       {

       //7 dates from the construction of sector 2

        R_Date("1", 1370, 60);

        R_Date("2", 1450, 90);

        R_Date("3", 1170, 80);

        R_Date("4", 1310, 60);

        R_Date("5", 1200, 60);

        R_Date("6", 1080, 50);

        R_Date("7", 1350, 60);

       //6 dates from overlaying ash layer  

        R_Date("8", 1110, 30);

        R_Date("9", 1120, 30);

        R_Date("10", 1270, 30);

        R_Date("11", 1170, 30);

        R_Date("12", 1250, 30);

        R_Date("13", 1240, 30);

       };

     Boundary("Burn Event");

    };

[Bayliss, Alex]

Erik,

 

I don’t think this is a modelling problem; this is an archaeological problem. What was dated? Only annual samples from absolutely not redeposited contexts will date the burning event. If you date a timber from a burnt house, you are dating the time the ring(s) you sent to the laboratory were growing. If you date a burnt piece of wattle, that is a better indication of the date when the house was constructed, not when it burnt down. You are looking for charred grain in a storage bin, a leg of mutton on the kitchen table. Once you have had a hard stare at the dated material, you may be able to say that samples relating to the construction of the house are earlier than samples that were in it when it was burnt down.

 

If you have a mixed bag of samples, then you may have to manage their defects in the modelling process. Try a Charcoal outlier model for anything that may have an old-wood offset, and a tau boundary for the burning event (ie things in the house a more likely to be close in age to the burning event than a long way away from it).

 

Hope this helps,

 

Alex

 

 

 

From: ox...@googlegroups.com [mailto:ox...@googlegroups.com] On Behalf Of Erik Marsh
Sent: 18 February 2021 16:59
To: OxCal
Subject: Dating a burn event?

 

THIS IS AN EXTERNAL EMAIL:  do not click any links or open any attachments unless you trust the sender and were expecting the content to be sent to you

--
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/af2c21d9-d4e0-4f1b-8b15-09559eda257dn%40googlegroups.com.

We are the public body that helps people care for, enjoy and celebrate England's spectacular historic environment, from beaches and battlefields to parks and pie shops.
Follow us:  Facebook  |  Twitter  |  Instagram     Sign up to our newsletter     

This e-mail (and any attachments) is confidential and may contain personal views which are not the views of Historic England unless specifically stated. If you have received it in error, please delete it from your system and notify the sender immediately. Do not use, copy or disclose the information in any way nor act in reliance on it. Any information sent to Historic England may become publicly available. We respect your privacy and the use of your information. Please read our full privacy policy for more information.

[Thomas S. Dye]

Aloha Erik,

IIUC, dates "3" and "6" are sealed by the thick burn layer, so
they must pre-date the burn event.

If so, remove Boundary("Start") and change Boundary("Burn Event")
to Event("Burn Event"). This will remove the uniform distribution
assumption introduced by Boundary(), which might be superfluous
here.

In-built age won't contradict the logic of the model because the
deposition event for each dated sample is older than the burn
event. The search for outliers based on the uniform distribution
assumption is likely not worth the effort.

HTH,
Tom

--
Thomas S. Dye
https://tsdye.online/tsdye