Dating a Destruction Layer with Four Disagreeing Clocks
A regional museum is preparing catalogue labels for material from a destruction layer at Tell Harran, excavated in 1998 and never fully published. The archive holds four conflicting items. (1) A radiocarbon determination on charcoal from the destruction debris, 2-sigma calibrated to 1210–1180 BCE; no replicate, context recorded only as 'debris.' (2) A floor laid directly over the debris, containing a faience bead of a type dated 1150–1100 BCE at three comparanda sites. (3) A tablet from within the debris naming a king whom the Assyrian king list places at 1233–1197 BCE, though the list's absolute dates depend on a single eclipse record and transmitted reign lengths. (4) A scarab of a 19th-dynasty pharaoh recorded in a field notebook as 'in the destruction level,' with no locus number. The registrar must choose between a label reading 'c. 1200 BCE,' one reading 'c. 1180 BCE,' and an open range. Publication closes in three weeks. The director wants one number, because 'visitors do not read ranges.'
1. Rank the four items as dating evidence for the destruction, and state precisely what each one does and does not establish.
Read the worked answer
Ranking, strongest first: bead, radiocarbon, tablet, scarab.
- Faience bead in the overlying floor. Because the floor seals the debris, a bead dated 1150–1100 BCE gives a terminus ante quem: the destruction cannot be later than the floor's construction. Two caveats: the bead could be residual (older object still in use or redeposited), and the comparanda dates are typological, carrying uncertainty of decades.
- Charcoal radiocarbon. It dates the death of the wood, not the fire. Structural timber can be decades old, so 1210–1180 BCE is a terminus post quem only, at 95% confidence, from a single unreplicated run on vaguely provenanced debris.
- The named king. Weakest usable item. It requires that the king-list dates are correct (they rest on a chain anchored to the 763 BCE eclipse and assumed reign lengths), that the tablet is contemporary with the destruction rather than curated or residual, and that the name is not a homonym. It offers a synchronism, not a date.
- The scarab. No locus, no stratigraphic control; scarabs circulate for generations as heirlooms. Discard as dating evidence.
Combined bracket: destruction after c. 1180 BCE and before the floor, i.e. c. 1180–1150 BCE.
2. Write the dating statement you would put in the catalogue entry, and explain why each word of it is hedged the way it is.
Read the worked answer
Statement: 'Occupation at the site ended in a destruction dated between c. 1180 and c. 1150 BCE. The layer is sealed by a floor of c. 1150–1100 BCE and post-dates a charcoal sample of 1210–1180 BCE. The associated tablet names a late-thirteenth-century BCE king; this may reflect curation or residuality rather than contemporaneity.'
Why each hedge:
- 'between c. 1180 and c. 1150' - the bracket, not a point, because the two usable bounds are independent and each is soft. Using one number would silently discard the other bound.
- 'sealed by a floor of c. 1150–1100' - the bead's typological range is quoted rather than converted into an event date, so the reader can see the source of the upper bound.
- 'post-dates a charcoal sample' - makes the terminus post-quem logic explicit; a reader who knows about old timber will understand why 1210–1180 is not offered as the destruction date.
- 'may reflect curation or residuality' - flags the synchronism problem without asserting either.
What the entry must not do: present the king's name as the cause or date of the destruction, or hide the scarab's evidential weakness by omitting it silently. A short footnote should record that the scarab lacks a locus.
3. The director demands a single number, and a colleague wants to cite the tablet as proof that an Assyrian campaign destroyed the site in 1197 BCE. Address both.
Read the worked answer
On the single number: give 'destroyed c. 1170 BCE' on the label, with a companion panel or catalogue note showing the full bracket 1180–1150 BCE and the reasoning. 'c.' plus a round decade signals approximation conventionally; a label is a summary, not the argument. The compromise costs nothing evidentially as long as the bracket remains accessible and the label does not claim precision the evidence lacks. What would be illegitimate is writing '1170 BCE' without the 'c.', which invites citation as a fixed point.
On the Assyrian campaign claim: refuse it, for three reasons.
- Chronological dependency. The king-list dates for the late thirteenth century BCE are themselves reconstructions; a 1197 BCE campaign date is not an independent anchor.
- Association. A tablet in destruction debris need not be contemporary with the destruction. Without an internal regnal or eponym formula, the name gives at best a synchronism.
- Circularity. Using the destruction to date the tablet and the tablet to date the destruction closes the argument on itself. If the colleague wants a campaign, they need independent evidence - a royal inscription naming the site, or a destruction horizon attested at several sites with the same bracket.
Offer instead: 'a destruction in the later twelfth century BCE, broadly contemporary with late Assyrian expansion.'