The short answer
Three things to know before you read on
- 1 establishes relative order — older layers generally lie beneath younger ones — before any laboratory work begins.
- 2 measures when a once-living organism stopped exchanging carbon with the atmosphere; it dates the organism, not the artifact it accompanied.
- 3A raw radiocarbon measurement is not a calendar year. It becomes a date range only after against an internationally agreed atmospheric curve, and only when read alongside its archaeological context.
01
The Ground Keeps a Record
Every excavation begins with observation, not instruments. As soil accumulates over centuries, it does so in layers — strata — and the principle that lower layers were deposited before upper ones is one of the most reliable rules in all of historical science. Archaeologists document these layers in profile, recording their colour, texture, and boundaries before a single sample is sent to a laboratory. This sequence of superposition gives the site its internal calendar: earlier, later, and contemporary are established by position alone.
Stratigraphy is powerful precisely because it makes no assumptions about chemistry or style. A hearth sealed beneath an undisturbed floor is demonstrably older than that floor. A pit cut through earlier deposits is demonstrably younger than what it cuts. These relationships are direct observations, not inferences, and they anchor everything that follows. When laboratory dates later arrive, they are tested against this stratigraphic logic — and any result that contradicts a well-documented sequence demands explanation before it is accepted.
02
Shape and Style as Relative Clocks
Once relative order is established, archaeologists compare what they find to what is already known. Typology — the systematic study of how artifact forms change over time — allows a distinctive pottery rim or a particular flint-knapping technique to serve as a chronological marker across a region. When the same vessel type appears at a site with a known date and at an undated site, the comparison carries real inferential weight, though it must be used carefully: styles spread unevenly, and local traditions can persist long after fashions change elsewhere.
Associated finds sharpen the picture further. A coin, a glass bead of a known production period, or an imported ceramic whose origin is independently dated can anchor an otherwise undated assemblage. Archaeologists call these index fossils by analogy with geology — objects whose temporal range is well enough constrained to narrow the possibilities for everything found alongside them. The inference is probabilistic, not certain, and its strength depends on how securely the comparison objects are themselves dated and how tightly the association can be demonstrated.
Stratigraphic Profile
Excavators record the vertical sequence of deposits, establishing which layers are earlier and which are later through direct observation of superposition and cutting relationships.
Relative order — no calendar years yetTypological and Associative Comparison
Artifact forms and associated finds of known date are compared to the assemblage, narrowing the probable period through inference from regional parallels and index objects.
Probable period — decades to centuries of resolutionRaw Radiocarbon Measurement
A biological sample is measured in an accelerator mass spectrometry laboratory. The result is a conventional radiocarbon age with a statistical uncertainty — not yet a calendar date.
Conventional radiocarbon age ± laboratory errorCalibration and Contextual Integration
The raw measurement is mapped onto the IntCal20 calibration curve and combined with stratigraphic priors using Bayesian modelling, yielding a probability distribution across calendar years.
Calendar date range at 95% confidence03
What Carbon-14 Actually Measures
Radiocarbon dating rests on a well-understood physical process. While an organism is alive, it continuously exchanges carbon with the atmosphere, maintaining a ratio of the radioactive isotope carbon-14 to stable carbon-12 that reflects the atmospheric concentration at that moment. When the organism dies, exchange stops and carbon-14 begins to decay at a known rate — its half-life is approximately 5,730 years. Measuring how much carbon-14 remains in a sample therefore indicates how long ago that exchange ceased: in other words, when the organism died.
This is the measurement's most important boundary condition, and it is frequently misunderstood. Radiocarbon dates the death of a biological organism — a tree, an animal, a cereal grain — not the manufacture of any object made from or associated with it. A wooden beam cut from an ancient tree and reused centuries later will yield a date reflecting the tree's death, not the building's construction. Charcoal from a long-lived species may be centuries older than the fire that produced it. Recognising these gaps between the dated material and the archaeological event of interest is a core skill of the discipline.
04
From Raw Measurement to Calibrated Range
A laboratory returns a radiocarbon result as a conventional radiocarbon age expressed in years before present, accompanied by a statistical uncertainty. This number is not a calendar year. The atmospheric concentration of carbon-14 has not been constant through time — it has fluctuated due to changes in solar activity, ocean circulation, and other factors. To convert a raw measurement into a calendar date range, researchers apply a calibration curve: a record of past atmospheric carbon-14 levels reconstructed from tree rings, corals, speleothems, and other archives with independent age control.
The internationally adopted standard for the Northern Hemisphere is IntCal20, published by Reimer and colleagues and maintained as a community resource. The calibration process, formalised by Bronk Ramsey and others at the Oxford Radiocarbon Accelerator Unit, maps the raw measurement and its uncertainty onto the calibration curve and returns a probability distribution across calendar years — typically expressed as a range at 95 percent confidence. Because the curve is not a straight line, a single radiocarbon measurement can sometimes map onto multiple calendar intervals, and the resulting distribution may be broad or multimodal. The final reported date range is therefore a model output, not a direct reading.
05
Combining Evidence into a Coherent Chronology
No single method stands alone. The strength of an archaeological chronology comes from the convergence of independent lines of evidence. A radiocarbon date that agrees with the stratigraphic position of a sample, matches the typological period suggested by associated ceramics, and falls within the known production range of an imported object is far more persuasive than any one of those observations in isolation. Bayesian statistical modelling, now standard practice in the field, allows researchers to incorporate stratigraphic constraints directly into the calibration process, tightening date ranges considerably when the prior information is well documented.
This integration also makes the reasoning transparent and testable. When a new date contradicts the existing model, the contradiction is informative: it may indicate a residual sample, a disturbed deposit, or a genuine revision of the site's history. Archaeologists are therefore not simply collecting dates — they are building and testing chronological arguments. The published date range for a site or phase is the output of that argument, and its credibility depends on how well the underlying evidence is documented and how honestly the uncertainties are reported.
Stratigraphy establishes relative order — older layers generally lie beneath younger ones — before any laboratory work begins.
The most important limitation in archaeological dating is the inferential distance between the dated material and the human event of interest. Radiocarbon measures when an organism died; calibration converts that into a calendar range; stratigraphy and typology link that range to a cultural episode. Each step introduces assumptions. When those assumptions are well-founded and the evidence converges, the resulting chronology is robust. When they are not — when samples are residual, deposits disturbed, or calibration curves flat — the reported range may be wider, or less securely tied to the event, than a headline date implies. Responsible reporting requires stating these conditions explicitly.
06
The Limits Honest Archaeologists Acknowledge
Even a well-executed chronology carries irreducible uncertainties. Organic material is not always preserved; some periods of the calibration curve are flatter than others, producing wider date ranges regardless of measurement precision; and the relationship between a dated sample and the human activity of interest is always an inference that must be argued, not assumed. Sites in the Southern Hemisphere, marine environments, or regions with locally anomalous carbon reservoirs require additional corrections that introduce further uncertainty.
Public communication of archaeological dates often strips away this complexity, presenting a single year as though it were a direct measurement. It is not. A responsibly reported date is a range, derived from a model, applied to a sample whose relationship to the event of interest has been carefully assessed. The discipline has developed rigorous tools for quantifying and communicating that uncertainty — and the most important advance in recent decades may be the growing expectation that those tools are used, and their outputs reported in full.
07
Sources and references
This article draws on the following primary and reference sources. All claims have been checked against the supplied evidence package.
- 01Oxford Radiocarbon Accelerator UnitLaboratory reference ↗
The Oxford Radiocarbon Accelerator Unit at the University of Oxford operates one of the leading accelerator mass spectrometry facilities for archaeological dating and provides methodological documentation on radiocarbon measurement procedures.
- 02Reimer et al. · IntCal20 Northern Hemisphere calibration curvePrimary standard ↗
Reimer and colleagues published the IntCal20 calibration curve, the internationally adopted standard for converting Northern Hemisphere radiocarbon measurements into calendar date ranges, covering the period from the present to 55,000 calibrated years before present.
- 03Bronk Ramsey · Radiocarbon Calibration and Age EstimationAcademic reference ↗
Bronk Ramsey's work on radiocarbon calibration and age estimation, produced at Oxford, provides the formal statistical framework — including Bayesian modelling approaches — used to integrate radiocarbon results with stratigraphic and other archaeological evidence.