Mare Nostrum LAB Thesaurus

Research teams work with many kinds of terminology: pottery types, vessel forms, production places, chronological periods and the characteristics of archaeological contexts. Keeping these terms consistent helps make records easier to find, compare and interpret across a project.

The Thesaurus organises these vocabularies into thematic dictionaries. They cover pottery classification and description, provenance and chronology, stratigraphic relationships, trench and context characteristics, architectural details, and metadata for visual and textual resources. Each entry represents a defined concept with its own identifier, so researchers can refer to the same concept across different tools, even when its name has alternative spellings or is expressed in another language. Descriptions and relationships provide the context needed to distinguish similar terms.

The dictionaries are hierarchical: entries are organised under broader categories, with more specific concepts nested beneath them. Researchers can follow these relationships to understand where a term belongs and navigate from a general category to a particular type.

For sharing vocabularies, the export uses SKOS - the Simple Knowledge Organization System, a W3C standard for representing thesauri and other controlled vocabularies on the web. It expresses concepts, their names and their broader–narrower relationships in a form that other systems can interpret.

A concept can have labels, short descriptions and alternative names in several languages. The entry forms currently support English, Polish, Greek, German and French, allowing teams to develop multilingual terminology within a shared structure. This multilingual structure lets teams add translations and terminology as their work develops; the languages available in the forms describe the system's capabilities, while the translations present depend on the content of each record.

Alternative names can record synonyms, abbreviations or spelling variants without creating a separate entry for every name. The concept keeps the same identifier, helping researchers recognise that different names refer to the same thing.

Chronology entries combine period names with structured date fields. Four boundaries allow contributors to express uncertainty about when a period starts and ends:

• Earliest Date: the earliest possible beginning of the period.
• Latest Start Date: the latest possible beginning of the period.
• Earliest End Date: the earliest possible end of the period.
• Latest Date: the latest possible end of the period.

Together, these fields describe a range for the beginning and a range for the end, allowing uncertain transitions to be recorded without forcing a single exact date. Values can be supplied where known; not every entry has all four boundaries filled in. For example, the "2nd c. CE" entry records an earliest date of 101 CE and a latest date of 200 CE. An application can therefore resolve the selected term to its recorded date range, or retrieve the latest possible start and earliest possible end where these have been supplied.

These date values are available to other applications through SPARQL, alongside the concept identifier and label. Applications can use them to translate chronological vocabulary selections into date ranges for filtering, comparison or timeline displays, while retaining the original period term and any recorded boundary uncertainty.

The Thesaurus connects archaeological terminology with the tools researchers use to document and study their material. In GeoNode, concepts can be assigned to dataset metadata. In Recogito, they can be attached to annotations on texts and images. Selected vocabularies can also be exported for use in Arches, supporting consistent descriptions of archaeological records.

The GeoNode and Recogito integrations retrieve concepts through the Thesaurus's SPARQL service. SPARQL is a standard query language for linked data: it allows software to request concepts and follow their relationships, rather than search only for matching words. For example, an application can request entries belonging to a particular dictionary branch, together with their identifiers and labels. Researchers use these connections through the applications' search and selection interfaces-writing a SPARQL query is not required for everyday vocabulary selection, though a separate query interface is available for users who want to explore the data directly.

Mare Nostrum LAB extends the underlying software with a dedicated dictionary view. Expandable lists display the hierarchy directly, so researchers can browse related entries, open a concept's record and return to the dictionary. Checkboxes allow users to select individual entries or whole branches; selecting a parent also selects its descendants, making it easier to assemble the vocabulary needed for a particular project.

The "Download selected as Arches format" action exports the selection as SKOS in RDF/XML, saved in a .rdf file. RDF, the Resource Description Framework, represents information as identifiable resources and their relationships; RDF/XML is its XML-based file syntax. The export includes concept identifiers, English labels, available English alternative names and descriptions, and the selected hierarchy, with ancestor concepts included automatically to retain the selection's place within the dictionary. The file is prepared for import into Arches, and using an established vocabulary standard also provides a basis for reuse in other systems that support SKOS, subject to their import requirements.

Dedicated forms help contributors add records to the correct dictionary and place them within its hierarchy. Open the form for the relevant vocabulary, enter the label and description, add any alternative names or translations, and select the parent concept. The forms are based on Cradle, adapted by Mare Nostrum LAB for the platform's dictionaries. Required fields and dictionary-specific choices guide data entry, and the parent can be the dictionary's main category or an existing entry, allowing a new record to be added at the appropriate level.

The adapted forms also create the corresponding inverse relationship on the parent record, supporting navigation in both directions and helping contributors maintain a connected vocabulary as it grows. Browsing is available without signing in, while adding and editing content requires an account with the relevant permissions; saving through the forms also requires authorising the form tool to make changes on the contributor's behalf.