Table of Contents
A fragment of burnt seed, pulled from a hearth in a South African cave. A potsherd with a thumbprint still pressed into the clay. A bone tool worn smooth at one end. Before any of it tells us anything, someone has to answer archaeology’s simplest and most stubborn question: how old is it?
Dating an ancient artifact is almost never a single measurement. It is a chain of arguments, cross-checks and educated doubts — and the honest answer often comes with a margin of error wide enough to drive a century through.
Relative Dating: Getting the Order Right Before the Numbers
Long before anyone could produce a calendar date, archaeologists could produce a sequence. Relative dating doesn’t tell you when something happened; it tells you what happened first.
Stratigraphy
The basic principle is older than archaeology itself. Layers of soil and debris pile up over time, so in an undisturbed site, deeper usually means older. At Near Eastern tells — mounds built up from millennia of rebuilding on the same spot — excavators have cut through more than 20 metres of occupation debris, layer stacked on layer like pages in a book. The catch is that pits, burrows and later digging can flip the pages, which is why context matters more than depth alone.
Typology and seriation
Objects change style over time, and those changes tend to follow patterns. In 1899, Egyptologist Flinders Petrie noticed that pottery from hundreds of predynastic graves could be arranged into a gradual sequence based on handle shape and decoration, even without any absolute dates. Seriation takes that idea and scales it up: line up enough assemblages and the fashionable styles rise and fall in a curve, much like car colours today. It gave scholars their first real grip on the timelines of ancient civilizations whose lost worlds we are still piecing together.
Radiocarbon: The Method That Rewrote Prehistory
In 1949, Willard Libby and his team at the University of Chicago published a technique that turned archaeology upside down. Living things absorb carbon-14, a radioactive isotope produced in the upper atmosphere. When an organism dies, uptake stops and the isotope begins to decay at a fixed rate, losing half its quantity roughly every 5,730 years.
Measure the remaining carbon-14 in bone, charcoal, wood, textile or shell, and you can calculate when that organism stopped breathing. Libby won the Nobel Prize in Chemistry in 1960; within a decade, radiocarbon had pushed human prehistory back thousands of years. It sits comfortably among the discoveries that genuinely reshaped how we understand our own past.
Two practical details matter. First, radiocarbon tops out around 50,000 years — beyond that, too little isotope remains to measure reliably. Second, raw radiocarbon ages are not calendar years. Atmospheric carbon-14 levels have wobbled over time, so labs calibrate their results against tree-ring sequences and other archives. The current standard curve, IntCal20, was built from thousands of measurements and pushes calibration back to 55,000 years. That is why a sample might come back as “4,120 ± 30 BP” in the lab report and “2,850–2,760 BC” in the publication.
When Carbon Won’t Cooperate
Some materials have no carbon to measure: volcanic rock, fired pottery, flint. Others are simply too old. For those, scientists reach for a different clock.
- Potassium-argon and argon-argon dating — volcanic minerals trap argon as they cool, and the accumulation gives an age. This is how the 1.8-million-year-old beds at Olduvai Gorge were dated, and how Lucy’s species was placed at about 3.2 million years.
- Luminescence (TL and OSL) — quartz and feldspar grains store energy from natural radiation; heat or sunlight empties that store. A burnt flint from a hearth or a quartz grain buried in sediment can therefore reveal when it was last heated or exposed to daylight, covering a range from a few decades to a few hundred thousand years.
- Uranium-thorium — flowstone and cave carbonate can be dated back half a million years. It was this method that showed a hand stencil in a Spanish cave must be older than 64,000 years, long before modern humans reached Iberia.
- Dendrochronology — tree-ring sequences, matched ring by ring, give year-by-year precision. European oak chronologies now stretch back more than 12,000 years, and bog oaks preserved in Irish peat have anchored whole prehistoric timelines.
- Tephra layers — a distinct volcanic ash bed acts as a timestamp that can be traced across a continent and dated elsewhere.
Why the Answer Is Rarely a Single Number
Competent dating projects use several methods on the same site and see whether the results agree. Bayesian statistical models then combine radiocarbon results with stratigraphic order — if layer B sits above layer A, the algorithm will not let B’s date come out older.
A neat demonstration came from L’Anse aux Meadows in Newfoundland. Researchers found that a wooden artifact had been cut in a year when a sharp spike in atmospheric carbon-14 occurred, traceable to a solar storm in 993 CE. Combined with tree-ring evidence, they pinned the Norse presence there to 1021 CE — a single year, from a lump of wood.
Where results conflict, the disagreements are instructive. The Shroud of Turin is a well-known example: radiocarbon testing in 1988 returned a medieval date, while later researchers argued the sampled corner may have been repaired or contaminated. Such disputes are not proof that dating fails. They show how much rides on where a sample came from.
Contamination, Old Wood and Other Traps
A date is only as good as the sample. A few grams of intrusive rootlet, museum preservative or even handling oils can skew a radiocarbon result by centuries. Labs now run chemical pre-treatment steps — acid-base-acid washes, ultrafiltration on bone collagen — specifically to strip out modern carbon.
Other problems are built into the material itself:
- The old wood problem — a roof beam carved in 1200 BC might come from a tree that started growing 300 years earlier. Dating the beam dates the tree, not the building.
- The reservoir effect — marine organisms take up older carbon from deep water, so a seal bone or a fish hook can appear centuries older than it is.
- Calibration plateaus — around 800–400 BC, atmospheric carbon-14 barely changed, so a wide span of calendar dates flattens into nearly identical radiocarbon values. Even a perfect measurement cannot break that knot.
These stubborn gaps sit alongside other puzzles, including several ancient mysteries that still baffle scientists and historians today — cases where the object is well dated but the explanation is not.
What a Date Still Cannot Tell You
Knowing that a bronze pin is 2,700 years old says nothing about who wore it, what it meant, or whether it was treasured or discarded. A radiocarbon result is a number with a probability distribution attached, and the interpretation is a separate act of reasoning.
That is why dating now travels alongside residue analysis, use-wear microscopy and isotopic studies of diet and migration. A pot’s charred crust can reveal what was cooked in it; a bone’s strontium signature can show where a person grew up. Newer tools keep adding layers — portable lab kits, high-precision mass spectrometry, and even protein sequencing that can identify species from a splinter too small to see clearly. Researchers chasing the scientific mysteries that still baffle researchers in 2025 often start with exactly this kind of evidence.
The artifact on the lab bench is never just a date. It is a material with a history of use, a context of deposition, and a chain of custody from trench to storeroom. Push the dating methods as far as they will go, and the most valuable thing they produce is not a number on a certificate but a sharper question about the people who left the object behind.
