How Archaeologists Date Ruins
The Problem of Dating
Every date in archaeology is an approximation, and every approximation rests on one or more lines of evidence with their own margin of error, their own assumptions, and their own potential for misapplication. The date '2500 BC' attached to a site in a museum label is the end result of a chain of reasoning that may involve stratigraphy, radiocarbon analysis, the typological comparison of artefacts with dated parallels, and historical documents — each method contributing to a range of probability that the label reduces to a single figure.
Understanding how dates are arrived at matters for anyone who reads about ancient sites seriously, because the precision and confidence of a date varies enormously with the method used. A building can sometimes be dated to within a specific year by a datable inscription. A burial can be dated to within roughly a century by the ceramic assemblage around it, or to within a decade or two by radiocarbon analysis combined with Bayesian statistical modelling. A peat deposit can be dated to within plus or minus 300 years by radiocarbon. Knowing which method applies, and what its error margins are, changes how any claimed date should be understood.
Stratigraphy: Reading the Layers
The fundamental principle of stratigraphy is among the oldest in archaeology: layers of sediment, construction debris, occupation deposits, and destruction events accumulate over time, with later deposits sitting above earlier ones. An archaeologist excavating a site encounters this sequence in reverse, peeling back the most recent layers to reach the earliest. The sequence of layers — the stratigraphic profile — records the site's history in compressed form.
Stratigraphy alone gives relative dates: layer A is older than layer B, which is older than layer C. Converting relative dates into calendar dates requires either material within the layers that can be independently dated (datable coins, inscriptions, radiocarbon samples, tree-ring sequences from timber), or comparison of the layer's artefact assemblage with assemblages from other sites where calendar dates have been established.
Edward Harris developed the Harris Matrix in the early 1970s, a systematic diagrammatic method for recording and interpreting stratigraphic sequences that has become standard in excavation practice worldwide. The matrix represents each deposit, interface, and cut as a node, connected by lines indicating stratigraphic relationships. It allows complex sites with hundreds of stratigraphic units to be analysed consistently and communicated unambiguously.
Radiocarbon Dating: Physics and Precision
Radiocarbon dating, developed by Willard Libby at the University of Chicago in the late 1940s, depends on the fact that living organisms continuously incorporate carbon from the atmosphere, including a small proportion of the radioactive isotope carbon-14. When the organism dies, it stops absorbing carbon and the carbon-14 decays at a known rate — its half-life is approximately 5,730 years. By measuring the ratio of carbon-14 to stable carbon-12 in an organic sample (charcoal, bone, seed, shell), the time since the organism's death can be calculated.
Raw radiocarbon dates must be calibrated against the IntCal calibration curve, which accounts for historical variations in atmospheric carbon-14 concentration. The calibration curve is built from samples of known age — tree rings, corals, lake varves — which have been independently dated by other means. Calibration converts the raw radiocarbon measurement into a calendar date range, typically expressed as a probability distribution: 'the sample has a 95% probability of dating to between 1400 and 1250 BC.'
Accelerator Mass Spectrometry (AMS) radiocarbon dating, which became practical in the 1980s, requires only milligrams of carbon rather than the grams needed for the older decay-counting method, allowing the dating of small or precious samples such as a single seed, a fragment of rope, or a charred grain from a sealed deposit.
Bayesian statistical modelling, applied to sets of radiocarbon dates in combination with stratigraphic information, can significantly tighten date estimates. At Çatalhöyük in Turkey, Bayesian modelling of several hundred AMS dates reduced the uncertainty on the occupation sequence from individual date ranges of around 400 years to estimates of roughly 50 years for specific phases, allowing the pace of the Neolithic settlement's development to be understood in meaningful human timescales.
Dendrochronology: Tree Rings and Absolute Dates
Dendrochronology — tree-ring dating — is, in favourable circumstances, the most precise dating method in archaeology, capable of identifying the exact calendar year in which a timber was felled. It depends on the fact that trees in temperate climates form one growth ring per year, and that the width of each ring reflects the growing conditions of that year: wide rings in good years, narrow rings in poor ones. The result is a unique pattern that can be matched against a master chronology for the region.
Regional master chronologies have been built from overlapping sequences in living trees, historic timbers in dated buildings, and archaeological timbers, extending back thousands of years in some areas. The central European oak chronology now extends continuously back over 12,000 years. The key requirement is that the timber must be from the same region as the chronology used for matching, because climate patterns are geographically specific.
When a structural timber is found in an archaeological context with its sapwood preserved, the outermost ring gives the year of felling very precisely. Without sapwood, the date is a terminus ante quem — the timber was felled no earlier than the last ring present, but the missing sapwood (typically 10–50 rings for oak) means the actual felling date was somewhat later. In Roman Britain, dendrochronology has been applied to waterfront structures at London, to the pile foundations of bridges, and to the timber buildings of forts, producing precise construction dates that form anchors for the regional archaeological chronology.
Coin Evidence and Historical Documents
Where coinage was in use, coins provide highly useful dating evidence because minting dates and rulers' names are often known from historical records. A coin found within a sealed deposit provides a terminus post quem — the deposit cannot predate the coin's minting. But coin evidence is imprecise in one important respect: coins circulate for long periods after minting. A coin of Augustus found in a 3rd-century AD context is not evidence that the deposit dates to Augustus's reign; it is evidence that the deposit dates from no earlier than his reign.
Inscriptions naming datable rulers or consuls can provide precise construction dates for major public buildings. The triumphal arch of Titus in Rome names the Senate and People of Rome as dedicators, and the dedication to a deified emperor places it after Titus's death in AD 81. The Vindolanda writing tablets from Hadrian's Wall include dates written explicitly in Roman calendar form, giving year-specific anchors for the fort's occupation sequence.
Luminescence Dating: Fired Materials
Thermoluminescence (TL) and its derivatives — optically stimulated luminescence (OSL) — date the last time a mineral or ceramic material was exposed to heat or sunlight. When clay is fired to make pottery or brick, the heat resets the luminescence clock; when sediment is buried and removed from sunlight, the burial date can be calculated from the accumulated luminescence signal. The method is particularly valuable for sites where organic material is not preserved and radiocarbon cannot be applied, and for dating fired mud-brick architecture or kilns.
OSL dating of sediment samples has proved especially valuable for Palaeolithic and early Neolithic sites where stratigraphic integrity can be verified and where the age range lies beyond the practical limit of radiocarbon (around 50,000 years). At Göbekli Tepe in southeastern Turkey, a combination of radiocarbon dating of associated organic material and OSL dating of backfill sediments has confirmed the site's extraordinary early date — construction activity beginning around 10,000 BC, placing it at the very beginning of the Neolithic period, before any evidence of agriculture or animal domestication in the region.
How Multiple Methods Work Together
No single dating method is infallible, and the most reliable chronologies use multiple independent lines of evidence that converge on consistent results. Inconsistency between methods is itself informative: it may signal a disturbance of the stratigraphic sequence, residual older material incorporated into a later context, or an error in the calibration applied. Open the map to explore ancient sites and see how their assigned dates connect to the archaeological evidence that supports them.