Becoming Operational - Our Own Vocabularies, Linked to the AAT
Part 3 of 4 in Becoming Operational
- Becoming Operational - Why I Switched from Mistral to Claude
- Becoming Operational - The Conventions
- Becoming Operational - Our Own Vocabularies, Linked to the AAT
- Becoming Operational - Arslantepe Is Online
So, back in The Vocabularies, I chose the Getty AAT for object and feature types, and Pleiades and PeriodO for sites and periods. I still stand by that choice. But the moment we started working with real data, I realised I needed something in between: local controlled vocabularies that say exactly what the sources say—and that are linked to the big vocabularies instead of being replaced by them.
Why not just use the AAT?
Here’s the thing: the AAT and an archaeometallurgical typology answer different questions. The AAT knows awls and chisels. It doesn’t know a Maykop fork or a single-bladed shaft-hole axe. If I simply replaced every type with the nearest AAT concept, I’d throw away exactly the information the typology was made for.
It’s the knife-or-scissors problem from my first data example all over again: the words we use for objects carry the judgement of the people who classified them. A vocabulary that’s too coarse silently overwrites that judgement.
The typology: built on SAM
For the object types, I didn’t start from scratch. The LOD-AM typology is based on the typology of the old SAM project (Studien zu den Anfängen der Metallurgie), as published by Krause in 2003, plus a few additions of my own:
Rüdiger Krause, Studien zur kupfer- und frühbronzezeitlichen Metallurgie zwischen Karpatenbecken und Ostsee (Rahden/Westf. 2003).
Why SAM? Because the SAM database will become one of the most important sources for LOD-AM. Its thousands of analyses of early metal objects are classified with exactly this typology. If LOD-AM speaks the same typological language, that data can come in without being forced into categories it was never meant for.
One might say, that I am not free of bias in regards to Rüdiger Krause. That's correct. I studied Vor- und Frühgeschichte in the Goethe University Frankfurt and I have been a student of his for many years. He did a lot to put me on the path, which let me to this point and I am deeply thankful for that and for many other things. So he more than earned this little citation.
The typology is a SKOS concept scheme with up to four levels. At the top are broad classes like daggers and other blades, axes and adzes or elements of dress and jewellery. Below them come type groups (daggers) and the types themselves (flat dagger or dagger, unspecified). Every concept has an English and a German label.
Connecting it to the AAT
Instead of replacing our types with AAT concepts, every concept that has a counterpart in the AAT is linked to it, with one of two SKOS relations:
skos:closeMatchwhen both mean (nearly) the same: our awl is a close match of AAT awls, our bowl of AAT bowls (vessels).skos:broadMatchwhen the AAT concept is broader: our single-bladed shaft-hole axe has the broad match AAT axes (tools), our Maykop fork AAT forks (object genre).
Of the roughly 230 concepts in the typology, about a quarter are close matches, most others have a broad match, and only a handful (like unknown) have none. And yes—every single AAT identifier was checked against the Getty SPARQL endpoint. No invented IDs.
The best part? You can ask LOD-AM for "all daggers" in AAT terms and still get every object with the precise type the publication gave it. Global and local at the same time.
Site and feature types
The same idea works for sites and features. LOD-AM has its own small vocabulary of site and feature types—settlement, hoard, grave, kurgan, layer and a few more—each linked to the AAT where a fitting concept exists. Arslantepe is a settlement (close match: AAT settlements), the Palace Hoard a hoard (AAT hoards (groupings)), the Royal Tomb a grave (AAT graves).
Periods
For periods, I followed the way PeriodO organises them—and full transparency, after missing PeriodO’s hierarchies at first, I was glad to finally get this right. Every period belongs to the chronology system of its source. "Arslantepe VIA" always means Arslantepe VIA as defined there, with its own time-span (3350–3000 BC). The phases are grouped into cultures: Arslantepe VIA is a phase of the Northern Mesopotamian Late Chalcolithic, Arslantepe VIB1 of the Northern Mesopotamian Early Bronze Age. A context is dated to a phase; the culture follows from it.
What’s next: Pleiades and PeriodO
These local vocabularies aren’t the end point—they’re a bridge. Next, I’m going to connect them to the gazetteers I chose at the very beginning:
- Sites will be linked to their Pleiades places. Where a site is still missing in Pleiades, I’ll add it.
- Periods will go to PeriodO. Since our chronology systems are already organised the way PeriodO organises its collections, each of them can become one PeriodO collection.
Until then, the local vocabularies do their job: precise enough for the sources, and linked enough to be found.
Curious for the itty-bitty details?

View LOD-AM/knowledge-graph on GitHub →
As always, if you have thoughts, suggestions, or just want to talk vocabularies, you know where to find me: info@lod-am.net or (if you have a GitHub account) in a comment below.
Comments
Comments are provided by giscus and stored as GitHub Discussions. Loading them connects your browser to giscus.app and GitHub, Inc. in the USA and transfers your IP address and browser data there. The USA does not have a level of data protection equivalent to the EU's; US authorities may access the data. By clicking the button, you consent to this transfer (Art. 49(1)(a) GDPR). Details: privacy policy, section 5.