Harris Matrix
Also known as: Stratigraphic Sequence Diagram, Harris-Winchester Matrix, Single-Context Recording, Context Sequence Diagram
The Harris matrix is a method for recording and diagramming the stratigraphic sequence of an archaeological site as a partial-order diagram of individually defined contexts. Devised by Edward C. Harris at the Winchester excavations in 1973 and codified in his Principles of Archaeological Stratigraphy, it treats every deposit, cut, and interface as a separate stratigraphic unit and reduces the tangle of physical relationships among them to a minimal directed acyclic graph that expresses only relative temporal order. By distinguishing physical superposition from temporal sequence and stripping away redundant relationships through transitive reduction, the matrix turns the three-dimensional complexity of a dig into a single, auditable diagram. It is the structural backbone of single-context recording and the standard interface between excavation and chronological modeling.
Key highlights
- Reduces the full three-dimensional stratigraphy of a site to a single, minimal, unambiguous diagram of relative time via transitive reduction.
- Provides a built-in consistency check, since any cycle in the diagram exposes a recording or interpretation error.
- Underpins single-context recording, giving each unit an independent record that supports re-analysis and digital data management.
- Supplies exactly the partial-order constraints that Bayesian chronological modeling needs to combine with radiocarbon dates.
Intuition
This section is available to Pro members. Upgrade to Pro
How it works
This section is available to Pro members. Upgrade to Pro
When to use it
Use the Harris matrix whenever an excavation produces multiple superimposed or intercutting stratigraphic units and you need an explicit, auditable record of their relative sequence. It is the default recording framework for complex multi-period sites — urban tells, deeply stratified settlements, cemeteries with intercutting graves — where verbal or purely drawn section records cannot keep the sequence straight. It is also the natural front end to absolute dating, because it supplies the ordering constraints that Bayesian chronological models require. The method is less informative on shallow, single-phase sites with little superposition, and it records only relative order, so it must be combined with artifactual, radiometric, or documentary dating to attach calendar ages and with formation-process analysis to interpret how each deposit came to be.
Strengths & limitations
- Reduces the full three-dimensional stratigraphy of a site to a single, minimal, unambiguous diagram of relative time via transitive reduction.
- Provides a built-in consistency check, since any cycle in the diagram exposes a recording or interpretation error.
- Underpins single-context recording, giving each unit an independent record that supports re-analysis and digital data management.
- Supplies exactly the partial-order constraints that Bayesian chronological modeling needs to combine with radiocarbon dates.
- Records only relative order, not duration or absolute date, so it cannot stand alone as a chronology.
- Its accuracy depends entirely on correct context definition and relationship recording in the field, which is irreversible once a deposit is removed.
- Large sites generate very large matrices that can be hard to lay out and read without dedicated software.
- The diagram captures temporal sequence but not the depositional processes behind each contact, which must be interpreted separately.
Common pitfalls
This section is available to Pro members. Upgrade to Pro
Applications
This section is available to Pro members. Upgrade to Pro
Frequently asked
How is the Harris matrix different from a section drawing?
A section drawing is a scale picture of the deposits exposed in one vertical face; it preserves geometry and appearance but mixes spatial and temporal information and is limited to what a single plane shows. The Harris matrix abstracts from all sections and plans across the whole site to express only one thing: the relative temporal order of every context, as a minimal diagram. Harris argued that the stratigraphic sequence is independent of the section and must be recorded as its own entity. In practice the two are complementary — sections and plans document each context, and the matrix records how they relate in time.
Why is the diagram drawn with redundant relationships removed?
Because the relationships form a partial order, many links are logically implied by others: if context A overlies B and B overlies C, then A is necessarily later than C without drawing that arrow. Keeping such implied links would clutter the diagram and obscure the genuine immediate contacts. The Harris matrix is therefore the transitive reduction of the order — the unique smallest set of arrows that preserves the complete sequence for an acyclic graph. This is what gives the matrix its clean appearance and lets very large sequences remain legible.
What does a cycle in the matrix mean?
A cycle — where a context is both earlier and later than another, directly or through a chain — is stratigraphically impossible, since nothing can predate itself. When one appears it always signals an error: a relationship recorded the wrong way round, two units that were actually the same context, or a misidentified cut. Detecting cycles is therefore a valuable internal audit. The acyclicity requirement makes the directed acyclic graph the correct mathematical model and turns the matrix into a self-checking record rather than a passive drawing.
Sources
- 1.Harris, E. C. (1989). Principles of Archaeological Stratigraphy (2nd ed.). Academic Press.ISBN 9780123266514
- 2.Renfrew, C., & Bahn, P. (2016). Archaeology: Theories, Methods, and Practice (7th ed.). Thames & Hudson.ISBN 9780500292105
You have read it. What now?
Cite this page
ScholarGate. (2026, June 23). Harris Matrix. ScholarGate. https://scholargate.app/archaeology/harris-matrix