3 Dating Limitations and Chronological Interpretation

What a date can establish

A date helps place events in time, but it answers a specific question and does not automatically date an archaeological event. Relative chronology establishes which events came before or after others. For example, a layer below another layer is usually earlier if the deposits have not been disturbed.

Radiocarbon dating estimates when carbon-bearing material stopped exchanging carbon with its environment. It measures a sample, not a layer, building, or human activity. A sample date can represent an event only when the sample is securely associated with that event and was deposited at about the same time.

Sample association and depositional history

A short-lived seed recovered from a sealed hearth may closely approximate the hearth activity if the seed was burned during its use and the deposit remained undisturbed. By contrast, charcoal from a long-lived tree in the same hearth may be older than the burning event: the tree's inner wood could have formed decades or centuries before it was cut or burned. This difference is the old-wood effect.

Older material can also be redeposited in a later layer as residual material. Younger material may move downward into an older layer as intrusive material, for example through roots or burrowing animals. These possibilities mean that stratigraphy, the sample's context, and its taphonomy—the history of deposition and subsequent movement or change—must be considered alongside the laboratory result.

Contamination and reservoir effects

A sample can acquire carbon after its original formation. Modern carbon, such as rootlets or some conservation materials, can make an old sample appear younger. Radiocarbon-depleted carbon, such as geological carbonate, can make it appear older.

Laboratories use physical and chemical pretreatment to remove many contaminants, but pretreatment cannot correct a sample that is the wrong age for the event being investigated. Good sample selection and clear information about handling and preservation remain essential.

Reservoir effects can also affect radiocarbon ages. Carbon from marine or some freshwater environments may contain less radiocarbon than contemporary atmospheric carbon, making organisms that used it appear older. Samples from these environments, and samples from people or animals whose diets mixed terrestrial and aquatic foods, may require an appropriate calibration curve or reservoir correction. The correction can be uncertain, especially where local reservoir conditions varied.

Calibration and calendar-date uncertainty

A laboratory reports a conventional radiocarbon age, usually in years BP, meaning before AD 1950, together with measurement uncertainty. This is not yet a calendar date because atmospheric radiocarbon levels have varied over time. Calibration compares the measurement with dated records of past radiocarbon levels, using a curve appropriate to the sample. IntCal is used for Northern Hemisphere terrestrial samples, while Marine20 is an example for marine samples. Calibration curves are revised as evidence improves.

The calibrated result is a probability distribution, commonly reported as one or more calendar-date ranges with associated probabilities. A result can have several separate ranges rather than one continuous interval because the calibration curve has bends or relatively flat sections called plateaus. On a plateau, even a precise laboratory measurement may correspond to a broad span of calendar time. Reporting only a midpoint or “best date” hides this uncertainty, so the ranges and their probabilities should be reported and interpreted.

Combining and evaluating evidence

A chronology is stronger when it is supported by multiple lines of evidence. Several carefully selected radiocarbon samples can test whether a sequence is consistent; stratigraphic relationships can constrain their order; and dendrochronology or luminescence dating can provide evidence based on different materials and processes. For example, dates on short-lived seeds from successive sealed layers, combined with the known order of those layers, can help estimate when a phase began and ended.

Agreement among methods increases confidence only when their assumptions and contexts are sound. Several samples from the same disturbed deposit do not necessarily provide independent confirmation, and multiple dates can share reservoir or calibration-related uncertainties. Statistical chronological models can combine radiocarbon probability distributions with well-supported stratigraphic information, but they cannot make an incorrectly associated sample reliable.

Disagreement is useful evidence. It can prompt checks for contamination, redeposition, reservoir effects, or mistaken assumptions about the sequence. The most defensible chronology combines carefully chosen samples with stratigraphy and genuinely complementary evidence while preserving and explaining uncertainty.