Artifacts in MANTO: technical (im)possibilities

Written by Greta Hawes

In the midst of our recent efforts to add data from ancient artifacts into MANTO, I was delighted to take up an invitation to spend some time at the Canellopoulos Museum in Athens. CAMU has an extraordinary collection of Greek art, which you can explore on-line here (or of course in its beautiful premises tucked in under the Acropolis). Because CAMU holds a good number of ancient objects that depict myths, my visit was a great opportunity to think through some of the issues we are facing in working with this material. I’ll be documenting these in my next few blog posts.

MANTO now has a very mature ontology and data collection methodology. In general, this is a very good thing: we’ve worked long and hard to develop it. But this also means that decisions we made five years ago when working on (say) Apollodoros or Pausanias now control how we have to capture images on artifacts. There are, in short, technical limitations to what we can do now without radically redesigning MANTO.

Redesigning MANTO at this point is not really an option, but nor is it necessary: what we are facing is nothing unusual since all database structures afford certain advantages and disadvantages in different contexts. In this post I look at the way these technical design elements shape how we capture images on ancient artifacts.

As a relational dataset, MANTO requires that there are recognisable mythic entities (people, places, objects, etc) to draw connections between. The first technical limitation I struck was the impossibility of capturing recognisably “mythic” scenes which depict no recognisable entities. So, take this black-figure kylix as an example:

Black-figure kylix (ca. 530 BCE).

Paul and Alexandra Canellopoulos Museum, Athens, no. Δ 436. Photograph: Nikos Stournaras. Image used with permission.

The scene of men and women in various poses is obviously intended to invoke an heroic atmosphere. But whereas we might easily say that is depicts the mythic storyworld (which is of course the key concern of MANTO), we cannot recognise here any particular mythic episode or characters. And so, because data collection MANTO requires us to identify particular people, places and objects, there is no way that we can include this particular object in our dataset.

A second technical limitation became clear as I began to understand how our method of categorising stories cuts out other possible ways of categorising them. From the beginning, our focus has been on identifying clear connections between identified entities. In narrative terms, we are most interested in the outcomes of actions. So, we end up with lists of all the people Diomedes kills, or all the men that Helen is said to have married, or all the mythic relics dedicated at the Sanctuary at Lindos. This is a kind of factoid-heavy approach. What it misses are the repeated motifs that appear and reappear through Greek myth. We might have rather based our approach on something like the folk-tale indexes that try to categorise and name plot points (“a young man appears in disguise”; “a monster threatens the city”) but did not.

When it comes to looking at depictions of Greek myth, our concern with concrete outcomes means that we cannot easily capture the kinds of iconographic patterns that others might in fact be interested in. So, whereas there are recognisable scene types in art that show (e.g.) warriors arming, or bidding farewell to their parents, in MANTO we cannot capture such nuances. This means that scenes which are formally different from one another are captured using the same (or very similar) ties in MANTO. Consider these three images, all showing Heracles with the Nemean Lion:

Fragmentary metope from the Temple of Zeus at Olympia showing Heracles standing over the body of the Nemean Lion.

Image from Wikimedia.

Captured in MANTO here

Attic black-figure amphora attributed to the Painter of Tarquinia RC 3984 with scene showing Heracles fighting the Nemean Lion. Dated to last quarter of the 6th c BCE.

Paul and Alexandra Canellopoulos Museum, Athens, no. Δ 2606. Photograph: Nikos Stournaras. Image used with permission.

Captured in MANTO here.

Three views of an Attic black-figure lekythos attributed to the Emporion Paintern with a scene showing Heracles luring the Nemean Lion from its cave. ca. 480-460 BCE.

Paul and Alexandra Canellopoulos Museum, Athens, no. Δ 646. Photograph: Nikos Stournaras. Image used with permission.

Captured in MANTO here.

The first image shows Heracles resting over the body of the slain lion; the second shows the struggle; and the third, a prelude to the fight: Heracles lures the lion from its cave using a deer as bait. Despite the fact that each scene represents a different point in the episode, we capture all using a tie that expresses the ultimate outcome: “HERACLES kills THE NEMEAN LION”.

We do this in part because the killing of the Nemean Lion is such a stable part of Heracles’ mythology: there is no version of the story in which the lion escapes, or turns around and mauls the hero. The outcome is essentially guaranteed, and so although only one of these images shows a dead lion, any ancient viewer would understand what is about to happen in the other two. By focusing on the outcome like this, we can continue to build up our factoid-heavy dataset, i.e. in this instance a reversal will make sure that these images are captured as evidence for the Lion appearing on Heracles’ “kill-list”.

I mentioned that database structures have advantages as well as disadvantages. The third image above - the lekythos - demonstrates this clearly. MANTO is designed to capture mythic places, and so the very fact that the lion is here depicted in a cave is of great interest to us. Indeed, this cave is already a landmark in MANTO: it’s here!

By investigating the data that we have already linked to this cave, you will discover that it played two quite different roles in stories of how Heracles killed the creature. The image on the lekythos (which was perhaps inspired by a tragedy on the theme) has Heracles lure it from its bolt-hole. Apollodoros 2.5.1 has a different account, in which Heracles blocks his means of escape and ventures into the cave itself:

Heracles tracked the lion. He first shot an arrow at him, but when he perceived that the beast was invulnerable, he heaved up his club and made after him. And when the lion took refuge in a cave with two mouths, Hercules built up the one entrance and came in upon the beast through the other, and putting his arm round its neck held it tight till he had choked it.

It’s no surprise to find the same event narrated several different ways — that is, after all, the nature of Greek myth. One storyteller does not need to agree with the next. What is intriguing is that the very idea of the lion lurking in his cave had such a hold on the Greek imagination that Pausanias (2.15.2) pointedd out its location in the 2nd century CE. Those kinds of connections — the implications of Greek myths on the physical landscape — are what MANTO’s structure is well-placed to reveal.

My warmest thanks to Nikolas Papadimitriou, director of the Canellopoulos Museum, for hosting me in Athens and for providing the images of objects in the collection that appear in this blog series.

This blog captures some material I presented at a workshop at the Centre for Digital Humanities at the University of Uppsala: I thank also Anna Foka and the team there for a fika-licious visit.