Ancient Artifacts in MANTO

Written by Greta Hawes

We’re moving into a new phase.

Up to now, all our data has been drawn from ancient texts. But these are just one kind of evidence for ancient storytelling, and to give a more complete picture of the tradition, we need MANTO to also encompass the extra-ordinary amount of information contained on ancient artifacts that depict mythic events.

1887 drawing of the François Vase, a large volute krater ca. 570 BCE. It was found in 1766 in an Etruscan tomb in the necropolis of Fonte Rotella near Chiusi and named after its discoverer Alessandro François. The drawing depicts the frieze showing the procession of gods and goddesses at the wedding of Peleus and Thetis. This file is made available by Wikimedia under the Creative Commons CC0 1.0 Universal Public Domain Dedication.

Since the middle of the year, Ewan Coopey and I have been working out how to do this.

We’ve had to think a lot about what MANTO can (and does) do well, and what we should not take on ourselves. Decisions that we made right back at the beginning of the project mean that there are technical constraints to what we can do now. Equally, we have needed to confront the practical difficulties of dealing with a corpus that is so large and complex with our limited resources.

I’ve just spent a wonderful week at the Canellopoulos Museum in Athens working my way through their collection and thinking about these technical and practical constraints. They will be the topic of future blog posts. Meanwhile, back in Sydney, Ewan has been spearheading our first pilot scheme, which focuses on finding relevant artifacts in Australian and New Zealand collections and adding them into MANTO.

Our first major challenge was working out how we were even going to identify artifacts within our dataset. Ancient texts are really pretty straightforward: we simply give the author, the title, and the passage or line numbers and link to an online repository using CTS-URN where possible (see here). Of course, there can be disputes and disagreements over exactly which author wrote which text, or which numbering system should be used, but nothing insurmountable.

By contrast, ancient artifacts exist in separate collections with their own systems for identification. Larger datasets like Beazley, LIMC, and Wikidata aggregate these collections, but there is no single system for identifying and classifying all artifacts.

This means that, practically speaking, we are investing a lot more time and energy creating a corpus of artifacts for MANTO than we ever did for texts. Our concern has been to find a way of doing this as efficiently as possible while also ensuring that we are able to maintain accuracy across the data collection process.

This process is starting to bear fruit. You will now see in some of the filecards a list of “depictions”. The artifacts are designated with a little paintbrush icon. (Try here, here and here.)

We have ended up with a schema for artifacts that focusses on two things: firstly, the identification of the artifact, and secondly its classification. So, for the François Vase:

MANTO’s schema for identifyng ancient artifacts illustrated by the François Vase

(The image above does not capture the entirety of the information: you can see it properly in MANTO here.)

Essentially, we have identified the François Vase by recording both its inventory number in its home collection - the Museo archeologico nazionale di Firenze - and where it appears in aggregator repositories: so, in this instance, catalogue references for Beazley and LIMC as well as URLs for the Beazley Archive Pottery Database, Digital LIMC, and Wikidata.

We also wanted to classify our artifacts in a very lightweight way. The metadata you could collect in a situation like this are almost infinite: shape and size, technique, material, composition of the material, provenance etc. We have decided just to use a few basic categories: type (e.g. coin, vessel, freestanding statue), period (e.g. early classical), place of creation and recovery, and then whether the artifact features any writing in Greek or Latin. We realise that these categorisations will not suit everyone’s purposes, but they are a workable solution for us which we think we can deploy with adequate accuracy and efficiency. Further, we hope they will be useful to others in an LOD environment in which others can reuse and enrich our data as needed.

These classifications have a purpose too, in that Nodegoat’s developers are currently working on new functionality that will allow us to filter our filecards by the source of the information, and so these broad categories will provide the basis for much more targeted discovery.

I thank core faculty and my fellow participants at the recent Alexandria Archive NADAC workshop in Berkeley, for their extraordinary support and feedback on this part of the project, for their third-hand accounts of appendectomies on remote digs done with nothing more than some local moonshine, a properly sharpened trowel and a vague understanding of human anatomy, and for not being too impolite about my controversial contention that all vessels are simply vessels.