ISAD(G) explained: the UK standard for archival description

Inconsistent catalogues are almost always a training problem disguised as a standards problem. A new volunteer describes a box of letters at item level; a colleague describes the same fonds at series level; a third person adds a folder and copies the fonds title into the series record. The result is a finding aid that confuses researchers, exports badly to AtoM or ArchivesSpace, and fails an NLHF grant review.

The fix is ISAD(G) — the General International Standard Archival Description published by the International Council on Archives (ICA). Not as a bureaucratic checklist, but as a shared logic for how archival description works across levels. Once that logic is clear, the standard is straightforward to apply.

What ISAD(G) actually is

ISAD(G) is an international standard — its second edition was published in 2000 by the ICA and remains current — that defines how archivists describe archival materials in a way that is consistent, hierarchical, and exchangeable between institutions. It sits alongside ISAAR(CPF), the companion standard for creator authority records, and underpins most modern archive management systems and export formats including EAD, EAD3, and Dublin Core.

The International Council on Archives designed ISAD(G) to work across every institution type — national archives, local authority record offices, university special collections, museum documentation departments — which is why it is framed as a general international standard rather than a sector-specific template. In UK practice, ISAD(G) is the backbone of archival cataloguing standards and the de facto requirement for any project seeking public funding.

The areas of description

ISAD(G) groups its 26 elements of description into seven areas. You do not need to complete all 26 for every unit of description, but knowing the areas helps you decide what belongs where.

1. Identity statement area — the who, what, when, how much. Reference code (your unique identifier), title, date(s), level of description, and extent and medium. These five elements are mandatory for every record, at every level of the hierarchy.

2. Context area — name of creator, administrative or biographical history, archival history, and immediate source of acquisition or transfer. This area explains why the archive exists and how it arrived at your institution. Leaving it incomplete loses provenance evidence that cannot be recovered later.

3. Content and structure area — scope and content, appraisal and destruction and scheduling decisions, accruals, and system of arrangement. This is where you describe what is in the unit and how it is organised — including the classification of the unit within the overall hierarchy.

4. Conditions of access and use area — conditions governing access, conditions governing reproduction, language and scripts, and archival finding aids. Researchers and rights holders need this information before they can use the collection.

5. Allied materials area — existence and location of originals, existence and location of copies, related units of description, and publication notes. This area links your holding to related material elsewhere and supports cross-institutional discovery.

6. Notes area — a general notes field for anything the other five areas cannot accommodate.

The seventh area — description control — records the archivist, date, and rules applied. It is often omitted in practice, but including it supports audit trails and is good professional hygiene.

The four rules of multilevel description

The most misunderstood part of ISAD(G) is not the elements — it is the multilevel description rules. These govern how archival description works across a hierarchy, from the highest level of description (the fonds) down through series, sub-series, dossier (file), and item.

ISAD(G) sets out four rules of multilevel description. Getting them right prevents most of the problems in practice.

Description from general to specific. Start at the highest level — the fonds. Each lower level inherits context from above, so you do not re-explain the creator at item level.

Information relevant to the level. Scope and content at fonds level describes the whole collection. Scope and content at file level describes that file. Don’t blur them.

Non-repetition of information. If the creator, date range, or custodial history is already recorded at fonds level, don’t copy it into every series and file below. Repetition creates inconsistency and breaks interoperability when systems try to link parent and child records.

Relationship between descriptions. Every unit of description, except the fonds itself, must be explicitly linked to its parent. This hierarchy is what makes archival finding aids navigable and EAD XML valid. A missing link between levels is one of the most common reasons EAD exports fail to import cleanly into AtoM or ArchivesSpace.

Together, these four rules mean that a well-formed ISAD(G) catalogue is a tree, not a flat list — every unit uniquely identified and explicitly linked to its parent.

The essential elements — the minimum you must record

ISAD(G) treats a small core of its 26 elements as essential. Five sit in the identity statement area and apply at every level of description:

• Reference code — the unique identifier linking the record to its institution and its classification within the hierarchy
• Title — formal or supplied, clearly marked if devised
• Date(s) — of creation, not accumulation
• Level of description — fonds, series, sub-series, file, item
• Extent and medium — how much material, in what format

To these five, ISAD(G) adds name of creator as essential for international data exchange — six essential elements in total — recorded at the level where it applies. Everything else is conditional on the level and the nature of the material. A fonds-level record for a business archive will almost always need creator, administrative history, and system of arrangement. A single photograph at item level may need only the five mandatory fields plus conditions governing reproduction.

The discipline is knowing what is enough at each level — and resisting the urge to either over-describe items or under-describe the fonds.

ISAD(G) and authority records

ISAD(G) records the name of the creator but stores the full creator description separately, in an ISAAR(CPF) authority record. This separation is deliberate: one creator may be responsible for multiple fonds, and maintaining a single authority record avoids duplicated or divergent biographical histories across the collection.

In UK practice, institutions link ISAD(G) descriptions to creator authority records through their catalogue system. AtoM does this natively; ArchivesSpace has its own agent records that serve the same function. The link between the unit of description and the authority record is what makes aggregated discovery possible across institutions — and what ISO standards for archival information exchange depend on. For NLHF-funded projects, clean authority records are increasingly scrutinised at grant review; see our ISAD(G), EAD3 and Dublin Core for NLHF projects post for what funders expect.

Common mistakes and how to avoid them

Titling the series with the fonds name. If the fonds is “Hartley Family Papers” and you have a series of correspondence, the series title is “Correspondence”, not “Hartley Family Papers: Correspondence”. The hierarchy provides the context; the series title should not repeat it.

Putting scope and content at the wrong level. A fonds-level scope note that describes individual letters is working too hard. A file-level scope note that summarises the whole collection is simply wrong.

Omitting archival history and source of acquisition. The chain of custody — who held the material between creation and transfer, any gaps, any known interventions, and the immediate source of acquisition or transfer — belongs in the context area. This is not optional information: it is the provenance record.

Forgetting the identifier. Every unit needs a unique reference code. Systems that export to EAD or import into AtoM will reject or mangle records without one.

Conflating collection and fonds. In ISAD(G), the fonds is the primary highest-level unit of description: material created or accumulated by a single creator in the course of their activities. “Collection” is sometimes used loosely to mean any gathered group of records, but the two terms are not interchangeable. Keeping them distinct matters when you are describing artificial collections alongside organic fonds in the same catalogue.

Where ISAD(G) is heading: Records in Contexts

The ICA is developing a successor standard called Records in Contexts (RiC), which moves from a hierarchical tree model to a graph model. A single record can belong to multiple contexts simultaneously, representing the full complexity of provenance, functional history, and institutional relationships. RiC is designed as an ontology compatible with Linked Data and ISO standards.

UK adoption of RiC will be gradual. Most institutions are running ISAD(G)-based systems and will not migrate overnight. The practical implication for current projects is: apply ISAD(G) rigorously now, because well-structured ISAD(G) data — with clean identifiers, explicit hierarchy, and complete context area entries — maps more cleanly to RiC than loosely described material does.

The metadata investment you make in a current project has a longer shelf life than it may appear.

How Archivers.ai supports ISAD(G)-aligned cataloguing

Applying ISAD(G) consistently across a large accession — especially with a mixed team of professional staff and volunteers — is where archival description most often breaks down. The Archiver assists by proposing description at the appropriate level of description for each unit, flagging where mandatory elements are absent, and suggesting scope and content text that the archivist reviews and confirms before anything is written to the catalogue.

The output is standards-ready metadata that exports in EAD3, Dublin Core, and AtoM CSV — formats that import cleanly into the systems UK institutions already run. Read more about how we use AI to support, rather than replace, professional judgement. The archivist stays in control of every description; the tool compresses the time spent on first-pass drafting so that expertise can go where it matters most.

If you are starting a cataloguing project and want to see how this works in practice, join the early access list for collections currently in scoping.