Key takeaways
- Dating methods split into two families: relative methods that establish order without numbers, and absolute methods that attach a figure with a stated uncertainty.
- Radiocarbon does not produce a calendar date directly. A raw measurement has to be calibrated against a reference curve, and the published result is a probability range rather than a single year.
- Every scientific method dates the material, not the event — charcoal gives the death of the tree, not the night of the fire, which is why the gap between the two has to be argued rather than assumed.
- Tree-ring dating can reach single-year precision where the wood and the regional chronology allow it, which is why it is used to anchor and check other methods.
- Dates get revised when calibration curves improve, when better samples are found, or when a chain of inference is re-examined. Revision is the method working, not the method failing.
"How do they even know that?" is the right question to ask of any date attached to something that happened before living memory, and it has real answers. There are perhaps five families of method doing the work, each with a domain where it is strong and a domain where it produces confident nonsense. This is a tour of all five, including the failure mode of each.
The question behind every date on a timeline
A timeline gives you a number and moves on. What it does not show is the chain of reasoning that produced the number — and that chain varies enormously in strength. Some dates are recorded in a document written at the time by someone present. Some are inferred from a chemical measurement on a fragment of charcoal, calibrated against a reference curve, and published as a range. Some are conventions that scholars agreed on in order to have something to point at.
All three appear on the same timeline in the same typeface. Knowing which kind you are looking at is most of the skill.
Two families of method
Dating methods divide into two broad families, and understanding the split makes everything else easier.
Relative dating establishes order without producing numbers. If one layer of soil sits above another and neither has been disturbed, the lower one was deposited first. That is a real, defensible finding, and it carries no dates at all.
Absolute dating attaches an actual figure, almost always with a stated uncertainty. Radiocarbon measurement, tree-ring counting, and a document that names its own date all belong here.
In practice they are used together. Relative methods build the skeleton — this before that, this contemporary with that — and a comparatively small number of absolute dates pin the skeleton to the calendar. A site with a hundred stratified layers and four radiocarbon dates is not four-hundredths dated; the four dates constrain the whole sequence, because the layers between them are bracketed.
Stratigraphy: the sequence in the ground
The oldest principle in the toolkit is also the simplest. In an undisturbed deposit, material accumulates from the bottom up, so lower is earlier. Archaeological excavation, as the National Park Service describes the discipline, records the position of everything it removes precisely because that position is itself evidence.
Two qualifications matter more than the principle:
- "Undisturbed" is a claim, not a default. Burrowing animals, later construction, ploughing, flooding and looting all move material between layers. Establishing that a deposit is intact is part of the work, not an assumption preceding it.
- A layer dates the deposition, not the objects. A coin in a layer tells you the layer formed no earlier than the coin was minted. It does not tell you the layer formed when the coin was minted — the coin could have circulated for decades first. This is why archaeologists speak of a terminus post quem: a date after which something must have happened.
That second point is a general one, and it recurs in every method below.
Radiocarbon: what it measures, and what it does not
Radiocarbon dating is the method most people have heard of and the one most often described inaccurately.
The physical basis: living organisms continuously exchange carbon with their environment, including a small proportion of the unstable isotope carbon-14. When the organism dies, that exchange stops and the carbon-14 present begins to decay at a known rate. Measuring how much remains indicates how long ago the exchange stopped. Willard F. Libby was awarded the 1960 Nobel Prize in Chemistry, in the Nobel Foundation's words, "for his method to use carbon-14 for age determination in archaeology, geology, geophysics, and other branches of science."
Three things about the method are routinely lost in summary.
A raw measurement is not a calendar date. The amount of carbon-14 in the atmosphere has not been constant through time, so a measured radiocarbon age has to be converted using a calibration curve built from independently dated material. The Oxford Radiocarbon Accelerator Unit distributes OxCal, the widely used calibration and chronological modelling program; its current release works with the IntCal20 curve, alongside SHCal20 for the Southern Hemisphere and Marine20 for marine samples. A date reported without saying whether it is calibrated is ambiguous, which is why publications mark the distinction explicitly.
The output is a probability distribution. Because the calibration relationship is wiggly rather than straight, one measurement can map onto more than one stretch of calendar time. Results are therefore published as ranges at a stated confidence level. When a popular article converts that range into a single crisp year, information has been discarded.
It dates the death of the organism, not the event of interest. A charcoal fragment dates the tree, and the tree may have been dead for a long time before it was burned — the "old wood" problem. Short-lived material such as a seed, a twig or a nutshell sits much closer to the event, which is why excavators prefer it. Bridging the gap between what was measured and what you want to know is an argument that has to be made, not a step that can be skipped.
The method also has a practical ceiling. Beyond roughly fifty thousand years there is too little carbon-14 remaining to measure reliably, so anything older requires an entirely different technique.
Tree rings: the method that reaches single years
Dendrochronology is the highest-precision dating method available for the periods it covers. Trees in seasonal climates add one growth ring per year, and the pattern of wide and narrow rings — driven by shared climate — repeats across trees in a region. Match the pattern in an unknown sample against a regional master chronology built from overlapping known samples, and the sample's rings can be assigned to specific calendar years.
The Laboratory of Tree-Ring Research at the University of Arizona, founded in 1937 by A.E. Douglass, who originated the science of tree-ring dating, remains a central institution in the field.
The constraints are specific:
- You need enough rings to produce a distinctive pattern; a short sample can match many places by chance.
- You need a master chronology covering the right region and species.
- You need the outermost ring — ideally with bark still attached — to date the felling year rather than some earlier point in the tree's life.
- The tree must grow in a climate with a genuine annual cycle, which limits the method geographically.
Where all of that holds, dendrochronology can date a timber to the year and sometimes the season. That precision is also why tree-ring sequences are used to build and test radiocarbon calibration curves: one method with absolute annual resolution anchoring another with statistical resolution.
Written records: the most direct and the most deceptive
A document that states its own date looks like the end of the problem. Frequently it is the beginning of one.
Records are dated within a system, and systems differ. Regnal years count from a ruler's accession, which requires knowing when the accession was and how the local convention handled partial years. Calendar reform shifted dates by days and moved the start of the year. Different regions adopted the same reform decades or centuries apart. None of that makes the records unreliable; it makes them require translation. Archives generally state which system a document uses — The National Archives in the United Kingdom and the U.S. National Archives both publish research guidance to their own holdings — and the translation is a solved problem once you know you have to do it.
We cover the calendar side of this in more depth in how to actually learn something from a timeline, including why there is no year zero and why two sources can give different years for the same event without either being wrong.
The subtler problem with written records is that a text tells you what someone wrote, which is not automatically what happened. A chronicle compiled two centuries after the events it describes is a source about the compiler's period as much as the one it narrates. Historians handle this by asking when the text was produced, from what, and for whom — and by looking for independent corroboration.
Synchronisms are the practical tool here. If a record in one tradition mentions an event that appears in another tradition dated by a different system, the two chronologies can be locked together. Chains of these links are how ancient chronologies get built, and their weakness is that a single bad link propagates through everything downstream.
Astronomical anchors
Some ancient records describe events in the sky — eclipses, planetary conjunctions, heliacal risings. Because celestial mechanics can be computed backwards, an unambiguous description can sometimes be matched to a specific computable date. The U.S. Naval Observatory's Astronomical Applications Department publishes the kind of positional data this depends on.
The catch is "unambiguous". Ancient descriptions are often brief, sometimes figurative, and frequently do not record the observing location — and eclipses recur. A proposed match is a hypothesis whose strength depends on how many candidate dates fit and how specific the description is. When it works, though, an astronomical tie-point is exceptionally strong, because the calculation does not depend on any human record being accurate.
Physical remains and their limits
Beyond radiocarbon there are other physical approaches — luminescence methods for when mineral grains were last exposed to light, techniques based on other decay systems for geological timescales, and typological sequences built from how object styles change. Each has its own domain, its own error behaviour, and its own version of the same warning: the method dates a material property, and connecting that property to a human event is a separate inferential step.
Sites on the UNESCO World Heritage List frequently carry dates derived from several of these methods at once, cross-checked against each other. Agreement between independent methods is the strongest evidence available in the field, and disagreement is what drives the next round of work.
Why dates change
Historical and archaeological dates get revised, and the revisions are sometimes large. Three ordinary reasons account for most of it:
- Calibration improves. Radiocarbon calibration curves are periodically revised as more independently dated material is incorporated. A date published against an older curve may shift when recalculated.
- Better samples appear. A date resting on charcoal from long-lived timber can be replaced by one resting on a seed from the same context, cutting out the old-wood gap.
- An inference is re-examined. Many dates depend on chains — this layer relates to that structure, which relates to that record. Break or improve a link in the chain and everything attached to it moves.
A field that never revised its dates would not be more reliable. It would be one that had stopped checking.
What this means when you read a timeline
Three habits are worth carrying:
- Notice the phrasing. "Circa", "conventionally dated to", "traditionally attributed to" and "no earlier than" all mean different things, and each is a signal about how much weight the number can take.
- Ask what was actually measured. For any scientific date, the useful question is what material was tested and how close that material sits to the event.
- Prefer converging evidence. A date supported by a document, a stratigraphic sequence and a radiocarbon range is in a different category from one supported by any single strand.
Using a timeline app for this
World History Timeline Sim covers more than 2,000 events and lets you browse by year, era, region, country and category, alongside a world history map, quizzes, flashcards and timeline-ordering exercises. That format is genuinely useful for the thing it does: building the relative structure — what was happening alongside what — which is the scaffolding that dating methods hang on.
It is an orientation tool, not a citation apparatus. When a date matters for something you are writing or arguing, take it to an archive, a museum catalogue entry or a publication that states its evidence. The archives listed in our sources below are free and open.
More apps in this area are collected under education and brain training, and our other writing on the subject sits under education and brain training articles. The world history timeline guide covers how to use a timeline as a study tool rather than as a reference work.
Good to know
Frequently asked questions
What is the difference between relative and absolute dating?
How does radiocarbon dating work?
Why is a radiocarbon date given as a range instead of a year?
How precise can tree-ring dating be?
Why do historians write 'circa' or 'conventionally dated to'?
Does World History Timeline Sim show how its dates were established?
Sources
- The Nobel Prize in Chemistry 1960: Willard F. Libby (opens in a new tab)Nobel Prize Outreach — accessed
- OxCal: Radiocarbon Calibration (opens in a new tab)Oxford Radiocarbon Accelerator Unit, University of Oxford — accessed
- Oxford Radiocarbon Accelerator Unit (opens in a new tab)Research Laboratory for Archaeology and the History of Art, University of Oxford — accessed
- Laboratory of Tree-Ring Research (opens in a new tab)University of Arizona — accessed
- Archeology (opens in a new tab)U.S. National Park Service — accessed
- Astronomical Applications Department (opens in a new tab)U.S. Naval Observatory — accessed
- The National Archives (opens in a new tab)The National Archives, United Kingdom — accessed
- Founding Documents (opens in a new tab)U.S. National Archives and Records Administration — accessed
- World Heritage List (opens in a new tab)UNESCO World Heritage Centre — accessed