Semantic Geo-Annotation for Ancient History and Beyond

Sometime in the second century CE, Pausanias of Magnesia wrote the Periegesis Hellados (Description of Greece). Representing a unique deep dive into ancient Greece’s built environment to the level of individual statues and paintings, this text projects a tour of the Greek mainland in ten books, from Attica (I) to Phocis (X), in a clockwise circuit around the Peloponnese (Figure 1). However, as Pausanias reveals in a rare methodological statement (1.39.3), he records not only the sights (theōrēmata) but also the stories (logoi) about them, whether because something of interest had happened in this or that place or because the person to whom a statue, say, had been dedicated had done something of interest. Cutting across and rerouting the narrative’s ‘relentless linearity’ are Pausanias’s movements in and through time, from the contemporary period of living under Rome to an era when heroes walked the earth, establishing institutions and rituals still practised in Pausanias’s day.

Despite appearances, tracing his itinerary around mainland Greece is far from straightforward and presents many challenges to digital mapping. Conventional software like GIS, whose capability of aggregating data by location has made it indispensable for historical geography, is less well equipped to deal with narrative space. In this blog post, I want to sketch out two ways in which the Digital Periegesis project is bringing to the fore spatial patterns underpinning Pausanias’s text, using Book 1 as a case study, while also developing digital tools and methods for research into textual geographies more broadly. In both cases, we draw on the work of Pelagios, the humanities Linked Data initiative that is helping domain experts produce resources that can be connected to each other by references to place.

First, to capture the spatial information in Pausanias’s narrative, we use Pelagios’s free open-source tool, Recogito. (For a comprehensive guide to Recogito’s functionality see the GitHub Tutorial.) For our purposes, it enables us to identify spatial features mentioned in the text – places, objects in space, peoples – and align those references to a global authority file. For Pausanias, such an authority record is provided by the Pleiades gazetteer, which assigns unique stable identifiers (URIs) for the places of the ancient world. We also use Recogito’s tagging feature to provide more information about the place or object, especially its type – whether it is part of a built or physical environment, for example, or whether it is a temple, statue, or painting, etc. In addition, we use tags to classify the moments in his narrative when Pausanias pauses his itinerary to relate an account about a particular place, person, or event. Rather than using our own modern dating system of BCE/CE to label these shifts in time, we use a broad-brush temporal anchor based on critical moments in Greek history, which would have been recognisable to his readership, such as the Persian or Peloponnesian Wars.

What does the semantic geo-annotation and tagging of Pausanias allow us to do? Well, first, because of their alignment to gazetteer records, we can map the places in Pausanias in a far more accurate, comprehensive, and granular way than was possible before. Figure 2, for example, shows all the places mentioned in the first book of the Periegesis. If we recall that Book I is devoted to a ‘tour’ of Attica (the region around and including Athens), the spatial footprint of the narrative is striking. The book in fact refers to places and peoples from across the Greco-Roman world and beyond, to Britain in the far west to India in the far east.

Because of our temporal tagging system, we can visualise the same information in different ways. Consider, for example, Figure 3a, which is filtered using the label ‘period of the Macedonians’, i.e., when various Macedonian-based powers held hegemony over the Mediterranean – from around ‘338 BCE’, when Philip of Macedon defeated the combined Greek forces at Chaeronea, to ‘146 BCE’, when Rome sacked Corinth. Paying particular attention to the ‘timeline’ of the narrative at the bottom right, it’s clear the critical role this period of history plays in Pausanias’s account of Attica, particularly in his description of the space of the Athenian agora (highlighted by the red circle).

We may compare this to the map now filtered by the ‘time of the heroes’ – i.e., all those places/objects that Pausanias describes in relation to the heroic age. The heroes are a constant presence, but they especially crowd in as the book nears its end, as Attica’s borders with Megara come under particular scrutiny.

These maps have barely scratched the surface of Pausanias’s description, but I want to move on to my second point, which is about the new Peripleo*, the visualisation tool used to create these figures, again developed by Pelagios’s Rainer Simon. To a certain extent, Figures 2 and 3 could have been created in a GIS – after all GIS specialises in showing spatial breadth and density. But being able to feature the text within the mapping interface already provides important additional functionality. This aspect really comes into its own when switching from the ‘map all places’ mode to ‘places in view’ (see Figure 4). As well as seeing all places mentioned in the book, you can read through the narrative and see the places mentioned section-by-section. Such a view is particularly powerful for a text like Pausanias’s, which is, for a large part, ‘itinerary’ based. So, you can ‘follow’ Pausanias’s movements through the civic space of Athens.

But that’s not all. We may recall that marking up the place references in Pausanias using Recogito involved not only identifying their mention in the text but also their alignment to a global authority – specifically a record of a place in a gazetteer with a unique identifier or URI. By virtue of this two-step process of semantic annotation, we can publish our work as Linked Data. Peripleo can take advantage of the research possibilities this format provides in two ways. First, you can link from Pausanias’s text to other online resources, which refer to the same places mentioned and which have also been marked up as Linked Data (i.e., have used the same gazetteer URI for the places mentioned).

Figure 5 provides a very simple example. It shows a building mentioned by Pausanias – the Erechtheion – which was located on the Acropolis (Figure 5a). By virtue of having aligned this reference to the appropriate record in Pleiades means that not only can we locate it on a map. We can also link to its record in Pleiades, and not only that, but also access the information found there, including links to further resources produced and curated by other online resources  (Figure 5b). In this way, we can better contextualise the reference in Pausanias and start to bring together other information that could have a bearing on our understanding of its description and function within Pausanias’s narrative.

But that’s also not all. By clicking on the Pelagios Network logo in the bottom left of the map interface (highlighted in Figure 6), you can reveal other places not mentioned in Pausanias, but which are found in the vicinity of those he does describe. (They show up because these resources have also been produced as Linked Data, using Pleaides URIs.) For example, Figure 6 shows the ‘other places’ located in the vicinity of Pausanias’s description of the famous Temple of Olympian Zeus in Athens. Notable among these is the equally famous Arch of Hadrian, which Pausanias fails to mention, though he would have had to walk right past it to get to the Temple of Zeus. As the classicist Tim Whitmarsh has argued, Hadrian’s arch was critical in demarcating Athens both spatially and temporally, drawing a distinction in the very topography of Athens between its ancient western part (or ‘city of Theseus’) and its Roman present in the east (the ‘city of Hadrian’). By erasing this arch from (his) record, Pausanias implicitly deconstructs and reconfigures the imperial space of Athens.

With this last example, we are getting a little closer to better exploiting the power of linked data, where we’ll be able to compare Pausanias’s description of objects to other resources about them and start to note and explore the gaps and the silences more systematically.

To sum up: both Recogito and Peripleo are open source tools, which are free to use, don’t require any specialist software to be downloaded, and work in your browser. Moreover, both are really easy to use. (I can say that with some confidence, since I can use them!) While Recogito is designed to help you semantically annotate place information in a range of digital documents (not only texts, but also maps and databases), you can also use it to markup all kinds of information, notably (but not limited to) people and events. Peripleo can help you ‘trace’ the historical representations of places or objects. In fact, the Digital Periegesis example in Peripleo is one of the first implementations of the emerging ‘Linked Traces’ format, developed by Karl Grosser of the World Historical Gazetteer in collaboration with Pelagios’s Rainer Simon, to enable the tracing not only of textual geographies but also the histories of artefacts, peoples, and events in space. (The idea will be to enable you to go seamlessly from annotating in Recogito to using Peripleo to visualise your annotations and to link to any other resources produced using the same Linked Traces format, which you want to compare and explore.)

One final take-home point. It’s the nature of digital resources that applications don’t stand still but continually evolve. This is one reason why development is no longer planned or run by a small core-team relying on external funding (“Pelagios the project”) but is now in the hands of the community (the Partners of the “The Pelagios Network”). The new Peripleo has been made possible due to a small grant (£5K) from my institution (The Open University), which was used to fund the creation of a lightweight visualisation environment for the annotations that we made on the Digital Periegesis, as one of the Partner projects of the Pelagios Network. Work on Recogito is proceeding in a similar way (i.e., it’s now funded by individual Pelagios Partners), and is already leading to the exciting new development of embeddable annotation interfaces Recogito, in which you won’t take your data to Recogito so much as take Recogito to your data. For annotating images, this work is going under the name of Annotorious, while for texts it’s RecogitoJS. In what ways and directions these tools and methods for producing and using Linked (geo)Data will develop further will be up to you.

*A note on Peripleo. The old Peripleo (developed in 2017) suffered from being too closely identified with Pelagios, as if it were somehow the ‘telos’ of our initiative. But Peripleo was only ever (meant to be) a demonstrator of what you could do if data were linked. As a result, the datasets ‘in’ Peripleo are random. It’s true that there’s a preponderance of ancient world resources, because they’re the people with whom we worked most closely at the beginning. But essentially, we worked with anyone who approached us. And many people did approach us, and continue to do so, because Peripleo is a nice visualisation of linked datasets (as well as providing the spatial footprint of collections). Yet the randomness of the resources meant that apples were being linked (/compared) to pears. Moreover, there was a more pressing problem. Pelagios was never meant to be a portal or an aggregator – “the one ring to rule them all” – where people went to explore interconnected resources. This for us is the very antithesis to the idea of linked data as decentred, non-hierarchical, community directed. Hence the transformation of Pelagios ‘the project’ into the community owned and directed PelagiosNetwork – since in our view linked data is as much about linking people as it is about linking resources. Which brings me back to the new Peripleo. Our view is that there won’t be ‘one’ Peripleo. Rather, it’ll be a robust but flexible (web browser) app that can be used to bring together the linked data resources that any one person is interested in exploring further.

This article was first published in the fall of 2021 on the Pelagios medium.com site, under the title: “Early steps in digitally mapping Pausanias‘s Description of Greece”.