Meet a Historian: James Baillie on Digital Humanities and the Medieval Caucasus

Note from the Editor: I’m excited to feature another guest post with you all! This week we have James Baillie discussing how digital humanities and prosopographic methods can be used to better understand the history of the medieval Caucasus. Prosopography is a historical tool-set that is about charting the networks, connections and commonalities of people, often people about whom we otherwise know very little. It’s a tremendously important methodological approach, especially in pre-modern history where everything we know about even important figures may only add up to a few lines of text. In the days of yore, this all had to be done by hand (slowly and often missing things), but this is a field where the potential of information technology – that’s the digital humanities here – is very great indeed. That said, this is a form of historical work I generally do not do, so I’m very excited to have someone who does talk through some of the methods and how they can help us to understand the past.

Baillie is in the final stages of his Ph.D. at the University of Vienna and is also a Project Assistant with the Caucasus Digital Repository in the Austrian Academy of Sciences. His research demands not only that he be a historian but also experience with coding and other information technologies, an impressive dual skill-set to be sure!

And with that, over to James!

I’m James Baillie – a historian specialising on digital approaches to the history of the medieval Caucasus region. I’m based at both the University of Vienna and the Austrian Academy of Sciences, and am coming towards the end of my doctoral work. In this post I’ll start with a look at the Caucasus and why I take my particular approach to it, and then go on to the underlying method – prosopography, the study of collected historical person data – before looking at some of the conclusions we can draw from it, in order to give you a bit more insight into the particular processes of research I go through. Finally, I’ll wrap up with some thoughts on why this matters for the Caucasus and why all of this is important in a wider context too.

So, without further ado, welcome to…

Fig. 1: Some notable places and settlements in the 12th century Caucasus

High Medieval Georgia

The Caucasus – the mountainous region that sits between Anatolia, Iran, and the southern Russian steppes, between the Black and Caspian seas – has throughout recorded history been a diverse and contested part of the world. It was, in the High Middle Ages as now, a culturally varied region with a range of overlapping ethnic and political entities: the overlaps, as we shall see, very important for understanding the medieval period. As a quick aside on what we mean by medieval – I tend to use the European-style terms and rough chronological definitions of what is medieval here, and my high medieval focus stretches from the 11th to the very early 13th centuries. Scholars in the region, however, often have a different and much broader understanding of ‘medieval’ which has roots in Soviet schools of history that suggested that the longer periods of ‘serfdom’ in the region meant the ‘medieval’ or ‘feudal’ period lasted until around 1800. Within all this I focus on the history of Georgia in particular, or as it is natively known Sakartvelo (საქართველო), the land of the Kartvelian peoples.1 In traditional narratives, the medieval history of Georgia is the precursor history of a Georgian state. This arises from the needs of nineteenth and twentieth century Georgian historians both before and under Soviet rule: for those opposed to Russian domination, the country’s twelfth century rulers, who were able to project power far further, could help present Georgia rhetorically in the same frame as European states and nations. Both Georgian liberalism and nationalism have typically been West-facing projects over the last two centuries, in reaction to Russian imperial claims.2 Meanwhile, for the USSR in particular, the history of medieval statehood, suitably framed, was an acceptable form of history to be writing, portraying the state as protagonist versus the story’s antagonists, the nobility and church who might undermine state authority. The history of the medieval state as such was both important to Georgian liberals and nationalists and acceptable to Soviet authority, and flourished as a result.

The history that emerged from this is one of a “golden age” triumph of a series of increasingly powerful state rulers, most notably David IV “the Builder” in the first quarter of the 12th century and Tamar in the last quarter, who are portrayed as having successfully won two zero-sum contests. The first of these was an external one, between Christian Georgia and its Muslim rivals such as the Turco-Persian Eldiguzid dynasty. The second was the internal struggle, in which the great feudal lords as a class struggled to maximise their independence from the central power of the state, with the aim of becoming as independent as possible. Finally, however, the Mongol invasions crushed the power of this Georgian state. Much of this school of thought dates back to the grandfather of modern Georgian history, Ivane Javakhishvili: his work is not available in English but books like Mariam Lordkipanidze’s ‘Georgia in the 11th-12th Centuries’ and Roin Met’reveli’s more recent ‘Golden Age’ follow similar lines.  For those new to the field, Stephen Rapp’s overview paper ‘Georgia before the Mongols’ is one of the best available introductory counterpoints to show the issues with how the traditional narratives are put together.

The traditional view of Georgia’s medieval history is not entirely wrong – but it is some way, too, from being entirely right. Two main critiques can be levelled: first, that it emphasises grand narratives based on classes of person, so the ‘nobility’, or the ‘church’, or indeed the Bagrationid dynasty’s monarchs across several generations, are often assumed to consistently have the same collective ideas and aims. This is often too simplistic a model for any real political system, especially when stretched over many decades. The second critique is that traditional histories tend to be very explicitly focused on the state and office-holding, and to use these elements to explain what’s going on: much ink tends to be spilt on topics like whether it was important that the creation of the formal title of Atabeg in around 1210 created a rival role of similar prestige to the Mtsignobart’ukhutsesi-Chq’ondideli, a senior religious advisor and vizier responsible for the royal chancery. I would certainly not attempt to claim that this was entirely unimportant, but I would suggest that comparatively insufficient attention has been given to the social context of power: who knew who, how ethnicity and gender affected people’s connection to power, and so on. It is in this context that a methodology that helps us examine those latter problems, that of prosopography, enters the chat.


Fig. 2: A very small subset of interconnected prosopgraphy data. Different colours indicate different types of ‘node’ representing people, places, sources, events and identities.

Prosopography is the systematic study of groups of past people: it requires us to create formal, structured sets of data about people, but in ways that do not assume that what we aim for is a statistical data-set to run calculations on. Medieval chronicle sources are too selective and patchy to allow for that sort of statistical analysis. Indeed, the fact that premodern data are patchy is precisely why building prosopography collections matters. We might not be able to say much from individually looking at the small number of elite women recorded at the medieval Georgian court, many of whom get one or two source mentions, and we don’t have a consistently sampled data-set on all or indeed any of them. However, if we can collect all of those references together in a structured data system and look at them as a group, we can start identifying shared characteristics and build up a picture that we couldn’t see if we left them sitting individually in their respective bits of source material. Being able to find patterns as possible explanations or lines of research can open up new windows on the past by, in effect, rearranging the things our sources already said into new configurations and by connecting in bits of external data such as geolocating the places that sources discuss.

Another thing we get out of prosopography besides this change of perspective is a process of learning about our information by putting it into a structure to start with. One finds quickly that there are certainly limits on any such system – there are some things about the past that cannot be captured by systematising them, because they are too subjective or contextual. When I am at home in Vienna, the fact that I am a European is unremarkable, and the fact that I am British stands out because almost nobody else is asking for milk in their tea: conversely, when I return to my native island, my being British fades contextually into the background, but my firm sense of European identity is a much more contested, political part of who I am. It is relatively easy to capture the fact that I hold multiple identities, but not why and when they matter.

Structured data tends to lose that information – but despite that, we gain an awful lot by using it is a more rigorous assessment of our starting assumptions. On what basis are we saying someone was a “Georgian” or an “Armenian” in a historical context? Is it a specific designation given to them in a text, an assumption based on their name and family ties, or a contextual reading based on their religion and allegiances? Forcing ourselves to go through historical sources, person by person and place by place and event by event, and asking ourselves what claims we feel comfortable making is valuable – sometimes vital – work.

It might be useful here to give a more concrete example. Enter the Orbeli, or Orbelian, family. The Orbelis were at the heart of Georgian politics through the middle decades of the 12th century, holding some of the highest offices of the court. Intrigue was rarely far behind: they were deeply tied up in the events around the death of the short-lived Mepe (King) David V, and may have been responsible.3 They raised David’s son Demna through the reign of his younger brother Giorgi, eventually rebelling in an attempt to replace Giorgi with their protégé. Their defeat opened the way for Giorgi’s elder daughter, Tamar, to become the first woman to rule Georgia.

The English-language Wikipedia page for the Orbelis confidently states that they were an Armenian family. If we look at Georgian-language Wikipedia, it equally bluntly opens by stating that they were a family of eastern Georgian nobility. Which is correct? Well, when we look at our sources, both. The 12th century Orbelis were religiously Orthodox rather than Armenian apostolic, but clearly had roots in both communities – their fortress of Lori was in the Georgian-Armenian marchlands. Their family chronicle, written some decades later, further adds interest to the puzzle by claiming that they took names from both Georgian and Armenian, but were in fact a Chinese family originally that had turned up some thousand years earlier: this is almost certainly a family myth or heroic ancestry tale, but shows the complexity of how people in the 12th century wanted to present their own ethnic backgrounds. Even if the rough data-level solution (I tag them as both Georgian and Armenian) is a huge simplification, the process of gathering that data on an individual level can teach us a huge amount about how we try to fit people into categories. It can certainly be a quick lesson in the risks of accepting the defaults from one source tradition or of taking the assumptions of scholars who, even if doing very good work, simply weren’t interested in this part of the problem, used a word that seemed about right, and then moved on.

You can now appreciate, hopefully, what prosopography is – a structured system, but fundamentally a humanities method that remains open to multiple approaches and interpretations. But what does it tell us?

Fig. 3: Forms like this, combining explanatory notes and structured data, are where most time is spent on this project. Fortunately for this piece, we get to skip filing in nearly a thousand of these and go straight to the interesting bits.

More Prosopography!

Heading back to the Caucasus, and skipping over several years of data entry and structuring (a project I may be continuing to add to for years to come), we certainly can reach some interesting conclusions about how medieval Georgian society worked that change our picture of the dominant, centralising state.

One thing is the importance of connectivity – which was often linked to ethnic identity. People often start looking at the past by assuming that being outside the dominant ethnic group was a disadvantage in historical societies – and that is often true, but the picture can be more complex. The Georgian commander of the armed forces (amirsp’asalar) was rarely a Georgian with no other ethnic background in the 12th century, with two mixed Georgian-Armenian families (the Mkhargrdzeli/Zakarian and Orbeli/Orbelian) dominating the post, plus figures like a Kipchak general called Kubasar holding the role. One likely reason for this is that these people had better connections to local nobility or sources of troops – Armenian-connected families may have helped legitimise Georgian claims on Armenian-majority regions, and Kubasar was appointed in the middle of a rebellion, probably to ensure the loyalty of large forces of Kipchaks who had been invited into Georgia a generation earlier in exchange for military service.

This wasn’t something I expected to find, at least in this form. One of my starting points was to look at regional rulers (eristavis) and evidence for decentralisation and local/regionalist sentiment within Georgia: but there’s actually little evidence that this was a motivating political factor. On the contrary, Georgia’s elite culture seems to have been quite centralised on the court. This was not because the state had effectively centralised its infrastructure and brought the nobility under its heel into serving the state, though: Georgian monarchs seem to have often aimed to increase, not decrease, the power of their vassals, on the grounds that this reflected significantly greater prestige onto them.4 The court could then become the social focal point for the south Caucasus, with the courtiers’ connections beyond it being vital for then turning that social focus into raising military manpower, governing cities, raising religious buildings, and other such functions.

Connectivity was probably also vital for women’s power. Queen Tamar is sometimes held up as the singular, sole example of feminine power in the 12th century Caucasus, but her aunt Rusudan was the widow of a Sultan who negotiated a key peace deal between the Seljuks and her brother Giorgi III. When Tamar faced down a coup at the start of her reign, her negotiators with the plotters were senior women from noble houses, too. Elite women tended to move between households and regions, and so their ability to form information connections – what in network science might be termed brokerage – gave them access to particular styles of power-as-diplomacy that were not necessarily as accessible to their male counterparts.

Prosopography can also help us redraw some of the maps of the region in the period. By systematically mapping out who was where, when, we get a better picture of where conflict zones actually were and the stability (or not) of control of different regions. The ancient Armenian capital of Ani was a particular political ping-pong ball, for example, but despite this, few maps of the period even show it as being near a ‘border’, preferring to map out the outer extent of places that were nominally captured at some point rather than taking a descriptive approach. The approach of doing political maps at “high points”, and showing them as consisting of matched borders in the manner of a modern country, can sometimes obscure as much as it tells us about the norms of a particular period.

Fig. 4: A standard map (CC-by-4.0, user Ercwlff on Wikimedia Commons) versus a map of conflict sites in Tamar’s reign (author’s work). Note how most of those blue dots are well inside the kingdom’s “border” on the left.

Finally, prosopography can help us sort and cross-check our own ideas: structured data created from chronicles isn’t an ‘accurate model of the past’, but it can be an accurate model of our best assessments and estimates, and so we can check it for internal consistency. By connecting together a set of dates and other bits of time-span data (e.g. X happened 1 year after Y), we can actually check mechanically what the things we said about one event imply about the dating of every other event in the data, reducing a problem that could take days of manual calculation into a button-press. This, in turn, can identify problems or places where certain ideas don’t fit together – for example, the Orbelis’ aforementioned rebellion is presented in the family chronicle with some passages about a member of the family going to see Atabeg Eldiguz to ask for his support against Giorgi. The slight problem here is that Eldiguz is generally held to have died three years before the conventional date of the rebellion, which is the sort of thing that a computer is very good at pulling out of a set of rather tangled data and this can then lead onto some interesting reassessments of who was actually dealing with who, and when, and what that might mean for the politics of a particular situation more widely.

Models of the Past

In many ways, I hope not too many people reading this will be asking why this all matters. As part of the ACOUP readership, you probably find the back and forth of medieval politics interesting for its own sake. Nonetheless, I think it does matter much more widely, and I’d like to finish by talking a bit about why.

For the Caucasus, medieval prosopography matters because people understand who they are and who they ought to be firmly in relation to their past. Georgian politicians evoke David the Builder in speeches to claim him for their political brand, Tamar is brought up in political debates over gender and gendered violence in the country, and the Georgian government is funding a literal computer game about the battle of Didgori in an attempt to inspire national sentiment among the young.5 The strong connection between these figures and national and ethnic status makes having serious, at times difficult discussions about their very complex realities more important. A simplistic understanding of the past can mean a simplistic understanding of the present that is assumed to flow from it – and that, all too easily, can turn into a simplistic case for violence when ownership of the past is applied to ownership of culture and land.

Even if you’re not at all interested in the Caucasus region, though, you should care about prosopography for more or less the same reasons. How we fit the world into boxes matters, because those boxes then affect our thinking and decisions on a wide range of issues. In your own view of the world, who is part of what group? How do you assign them to that, and what can you or can you not imply about them on that basis? How do you take account of the overlaps, layers, and complications? How you handle all of these things is partly determined by the mental structures you use to understand particular aspects of the world: those sets of ideas about what information about people can look like then create particular ways that these pieces of information can fit together and can be processed.

What we should realise from this is that a data structure or a block of code are things that make implicit and subjective arguments about how to see the world. This is possibly the single most important basic insight that Digital Humanities as a field needs to impart, because it affects so much of the world around us. We live in an age where people are structuring data about each other all the time, and often treating both the information and structures as statements of absolute fact. Taking a humanities eye to those structures can show code for what it is: it is not for nothing that we refer to programming languages: whilst a computer may not be able to discern subjectivity, that simply means that it takes the programmer’s position as factual, not that it gains some imagined freedom from the doubt of humanistic positions and enquiry.

We also live in an age where people are modelling worlds all the time – and I’ll finish there on a lighter note, which is that I think this sort of data and modelling work can offer a tremendous amount when it comes to areas like game development and digital storytelling. I’ve been developing mods and computer games as a hobbyist for some years – indeed, that’s where I learned the coding skills that I’ve more recently used to put together my research database and the custom-built editing tools I use for it. This is an area where, by thinking about and playing with the data structures that tell us who characters are and how they relate to one another, we can create characters and social-political game elements that utilise those connections, and provide the depth that comes from people having complex, tense, overlapping identities and networks that change over time.

This has been a whistle-stop tour through two-and-a-bit sub-fields of history, and is definitely here only to give a flavour rather than the full drink. I hope, though, that this has taken you through some of the processes of digital humanities scholarship, given you a few glimpses into the lives of some fascinating people, and made a little of the case for this work and what it can bring us. Thank you for reading!

James Baillie can be found on Twitter (@JubalBarca), Mastodon (@JubalBarca@Scholar.Social), and the web forums for geeky-creative projects which I help run and where I often write for the articles section. I pile up side projects at a disconcerting rate: these include in-development computer and tabletop RPG projects, convening the Coding Medieval Worlds workshops on games and medieval history, miscellaneous songwriting and storytelling pieces, and maintaining a blog on British liberal politics and policy.

  1. “Sa… o” is a common circumfix, also found in e.g. mepe for king becoming samepo for kingdom or mzareuli for cook becoming samzareulo for kitchen.
  2. By ‘imperial’ I don’t just mean ‘during the period of the formal Russian Empire’: the approach of the USSR to places like Georgia was still fundamentally imperial in practice, even if some parts of the ideology and rhetoric had nominally changed.
  3. The Georgian term mepe is gender-neutral for a reigning monarch (although Georgians at times inconsistently use dedopali, a term for queen consort, when talking about any other country’s queens, their own queens regnant are always mepe just as is the case for their male rulers and there is no contemporary evidence that this came with any assumption of masculinity or masculine traits).
  4. This would be a really interesting system to model in certain Crusader Kings 3 government forms.
  5. There is a trailer! It does not look entirely promising.

63 thoughts on “Meet a Historian: James Baillie on Digital Humanities and the Medieval Caucasus

  1. Very interesting stuff, and thanks for introducing me to a concept I’d been unaware of before, the circumfix!

    1. Yes, they’re very rare in English but much commoner in Georgian! Having just had a quick nose around online, apparently one of the few English ones is en-en for increases of quantity/intensity (enliven, enlighten) but there’s only a couple of uses of that.

    1. You can strip off the “?resize=1024%2C509&ssl=1” part of the image URL to get the unresized image.

  2. The “Georgian empire with tributaries” map seems to have been contested in the past on Wikimedia Commons. According to the revision history, someone claimed that Trebizond and Shah-Armens were not in fact vassals / tributary states, and attempted to de-color them (though the map was restored). Any opinions on this? (Setting aside how much Tamar controlled its putative internal territory, for a moment…)

  3. Both are pretty complicated situations, especially because “tributary” in the sense of “sent regular tribute according to a pre-agreed formula” probably often isn’t what we’re dealing with here and unless you have actual treaty documents surviving (which we usually don’t) it’s very very hard to know what you can extrapolate from an individual case of tribute mentioned in a narrative source. I’ll try and give a run-down below, anyhow, though this is done a bit from memory so advance apologies if I end up misremembering anything:

    I’ll do Shah-Armen first as the one I know less well (I’d like to do a lot more work on it, but at this point that’ll have to wait until post-PhD and there’s harder-to-access Arabic material that would be important to do it justice). In any case, I don’t think one can very easily argue that Khlat/Shah-Armen is a vassal state for more than perhaps very occasional parts of Tamar’s reign: it starts falling into the sphere of what the Mkhargrdzelis can raid by very late in her reign, and it , but by then it’s rarely referred to by the term Shah-Armen, which tends to be used to refer to the Sokmenid dynasty which has died out by 1207, and it’s not clear that Khlat necessarily actually dominates the amount of territory shown on the map above at the point where it’s actually being brought into the Mkhargrdzeli sphere (and it really is the Mkhargrdzelis spearheading that front, we don’t get much indication that the Georgian monarchs had a heavy involvement there). So part of the problem with that overall map is that it’s not really clear when in Tamar’s reign it’s trying to show, and that makes a difference.

    As for Trebizond – I don’t think we have any evidence that would suggests any tribute flowed from Trebizond to Georgia in this period. The one bit we have good evidence of here is that the army that captured Trebizond for the Megakomnenoi was Georgian, and Trebizond is among places that a Georgian chronicle says Tamar gave their freedom to: so this was very much a Georgian operation, and it’s likely there was some ongoing influence. But it’s unlikely that constituted vassalage, or tribute, in part because neither of those things were remotely the point of capturing Trebizond – indeed, capturing Trebizond probably wasn’t the point of capturing Trebizond. What’s much more likely (at least in my view) is that this was very much an attempt to install a friendly ruler in Constantinople – the Georgian force under David Komnenos marched down the Pontus pretty much at speed until he was stalled by the Nicaeans, leading to a stalemate that really ended up producing an accidental Trapezuntine state for the next quarter of a millennium. The Georgian monarchs could claim some success, in that they’d very well rewarded their relatives, but it’s really unlikely that the outcome was what they intended and it’s not clear that there was any ongoing vassalage relationship (not least because the Georgians probably still hoped for some years for things to swing in David and Alexios’ favour, and it might’ve been awkward to seat either of them on the Byzantine throne if they’d already accepted Georgian vassal status, so presenting them as The Real Empire might well have been more important at least for the first few years).

    So I’m not sure if that’s entirely helpful – sorry – but hopefully it shows some of the issues in trying to fit these sorts of relationships into neat “this guy’s a tributary, this guy’s a vassal” type boxes. It’s tricky, maps are so useful historically, but they can conceal as much as they reveal sometimes and I think it’d be good if we had a wider range of types of map and approaches rather than tending to revert to states-with-borders being the key standard for historical mapping purposes.

    1. Thanks for the reply! I am absolutely out of my depth here, but based on what you’ve said, it sounds like denoting Shah-Armens as a “tributary” is at best optimistic, while Trebizond was more like an “ally”, although perhaps that’s what the map maker meant by “Dependency.” And yeah, if the ideal goal was to hope the leaders of Trebizond took over leadership of the Byzantine Empire, it wouldn’t look good if they were too obviously subservient to Georgia.

  4. By ‘imperial’ I don’t just mean ‘during the period of the formal Russian Empire’: the approach of the USSR to places like Georgia was still fundamentally imperial in practice, even if some parts of the ideology and rhetoric had nominally changed.

    And so it’s not left implied: The same is true of the Russian Federation under Putin, cough Ukraine cough. (Possibly also Yeltsin? Not that familiar with Russian history.)

    Their family chronicle, written some decades later, further adds interest to the puzzle by claiming that they took names from both Georgian and Armenian, but were in fact a Chinese family originally that had turned up some thousand years earlier: this is almost certainly a family myth or heroic ancestry tale, but shows the complexity of how people in the 12th century wanted to present their own ethnic backgrounds.

    It always strikes me how often premodern people presented themselves (or their retinues/empires/culture/whatever) as being as ethnically diverse as possible. Not universally, but often enough that it emphasizes how arbitrary modern ideas about ethnic purity are. (In particular, the idea that homogeneous nation-states are somehow more cohesive or resilient than heterogeneous ones.)

    1. Nation-states didn’t exist before the 18th century, since nationalism hadn’t been invented yet. And depending on what you mean by “heterogenous”, you’re probably looking more at an empire anyway.

      And I can totally see how a ruler that got his (/her) power non-democratically (which was then the normal situation) would have benefited (depending on the context) from having family ties from many places (whether real or made up).

      In this case I’d guess that it might have helped to present yourself as a foreigner with not *too* strong ties/claims to one side or another, to be seen as someone impartial, to be able to unite Georgians and Armenians ?


      Otherwise, speaking of these “Chinese” claims –
      (and we know how inaccurate that might have been, considering the two kinds of “indians” mishap in the modern era)
      – they might as well have origins from Parthians, Scythians, or other Asian nomads ?

      1. I never said anything about premodern states being nation-states? I very deliberately only used that term in the specified context of “modern ideas about ethnic purity”.

        And I can totally see how a ruler that got his (/her) power non-democratically (which was then the normal situation) would have benefited (depending on the context) from having family ties from many places (whether real or made up).

        I don’t know what your point is. There are reasons that modern people who support ethnostates think they’re a good idea, and ways that they can benefit the people in power.

        Otherwise, speaking of these “Chinese” claims –

        (and we know how inaccurate that might have been, considering the two kinds of “indians” mishap in the modern era)
        they might as well have origins from Parthians, Scythians, or other Asian nomads ?

        I don’t know what your point is here, either. First off, it’s blatantly unfounded speculation, assuming that Georgians would have conflated Parthians and Chinese because…Europeans mistook indigenous Americans for Indians and the name stuck? Second, it doesn’t really matter to the point I have, so even if you were right, all you’d accomplish is wasting everyone’s time.

      2. The idea that nation-states didn’t exist prior to the 19th century is wrong. There were polities with strong national identity: England had a very clear identity tied to common language and system of government. The Norwegians had a rather clear concept of what “Norway” was even in the 13th century, and Sweden of the 16th century was a nation-state which defined itself by its nationalist opposition to Danes, Russians and the Hanseatic League.

    2. “It always strikes me how often premodern people presented themselves (or their retinues/empires/culture/whatever) as being as ethnically diverse as possible. ”

      IIRC, to be a citizen of a Greek polis you had to be descended entirely from citizens of that polis, on both sides of your family. I doubt they claimed to be ethnically diverse.

      But it seems a bit like the series on Crusader kings we had a few weeks ago. In that case, everyone wanted a lot of important people as their vassals, with a lot of other vassals serving them. Naturally, you would want to boast about how big and wide-ranging your network was.

      Bur for a polis, the first question would be about who was a citizen of which polis. Only then can you give them a say in the rule of that polis. The last thing you want is to give complete strangers who care nothing for you a say in how your own polis is run. So citizenship has to be more exclusive than vassalage.

      Giving someone else power over you is a much more fraught business than agreeing to be given power over him.

      The same logic applies to modern states. You can’t grant people citizenship and authority as easily as you could grant people vassalage and subordination.

      1. That was required in Athens, but Aristotle described it as the last of the restrictions put in — he was well aware of places where paternity was sufficient.

      2. IIRC, to be a citizen of a Greek polis you had to be descended entirely from citizens of that polis, on both sides of your family. I doubt they claimed to be ethnically diverse.

        The very next sentence after the one you quoted addresses it. I never claimed it was a universal thing.

        But it seems a bit like the series on Crusader kings we had a few weeks ago. In that case, everyone wanted a lot of important people as their vassals, with a lot of other vassals serving them. Naturally, you would want to boast about how big and wide-ranging your network was.

        Which explains Georgian nobles claiming they were themselves Georgian, Armenian, and Chinese…how?

        Giving someone else power over you is a much more fraught business than agreeing to be given power over him.
        The same logic applies to modern states. You can’t grant people citizenship and authority as easily as you could grant people vassalage and subordination.

        You realize that this argument boils down to re-asserting the thing I was saying about arbitrary ideas of ethnic purity, right?

        It’s also demonstrably wrong. Irish and Italian immigrants in the USA were considered non-white; when freed black people became a significant demographic, they were accepted as white (to discourage them from finding solidarity with freedmen). This didn’t have any effect on the cohesion or whatever or the white majority or the USA as a whole. It just kinda happened.

        1. “You realize that this argument boils down to re-asserting the thing I was saying about arbitrary ideas of ethnic purity, right?

          It’s also demonstrably wrong.”

          On the face of it, you are saying both that you think I agree with you, and also that you think I am wrong. You may imagine that I find that a little confusing.

          What do you mean by “arbitrary”, exactly? On the face of it, an arbitrary decision is one where there is no reason to prefer the favoured choice over the unfavoured one.

          People deciding who to be loyal (or claim to be loyal to) and whose loyalty to trust do not generally make arbitrary decisions. For example, consider those polis described above in which citizenship was inherited from the father. Family membership was also inherited from the father. So their rule meant that a member of a (for example) Roman family would himself be a Roman citizen. Change the rule so that citizenship was inherited from the mother, and that would no longer be true. Change the rule so that citizenship was a matter of being born in Rome, and again that would no longer be true.

          So their citizenship rule was different to many possible rules, including the modern American rule, but not arbitrary.

          “Varied” and “different” do not mean “arbitrary”.

          As for systems of vassalage, they might have had little need to decide who to class as a citizen. But they did have a lot of people looking for ancestors from whom they could claim to have inherited claims to other people’s loyalty. Or at least ancestors who seemed important.

          Thus the claim that Rome was founded by Aeneas, Prince of Troy. Or that Wales was founded by Brutus, son of Aeneas, Prince of Troy. Or that Alexander the Great was a descendent of Heracles. I admit these claims, and many like them, were made-up drivel. But not *arbitrary* made-up drivel.

          They were drivel made-up to support Roman and Welsh and Alexandrian claims to greatness.

          Likewise, people trying to extract loyalty from group X might well be happy to tell them of their own ancestor from group X, like American politicians claiming an ancestor among whatever ethnic group of voters they happen to be appealing to today. The claims may be true, or not true, or complete fantasy. But they are not *arbitrary*.

          Switching to the example you gave at the end, I confess myself sceptical of the idea that Scarlett O’Hara was considered Black until 1865, and White thereafter. Indeed, I am told that Catholic Irish Americans at the time had a reputation of being especially enthusiastic racists (1). And you can see why. They were Catholics in a Protestant country, and the country from which they claimed descent did not rank too highly. Being White was a claim to rank which they could make in spite of all that. As the poorest French aristocrats were precisely the ones who made the biggest thing about being aristocrats, so the Whites with the fewest claims to status could be expected to make the biggest deal out of race. You have to emphasise the things that are going for you.

          It’s not arbitrary at all. It’s all about trying to get allies, subordinates, and status. Mean Girls writ large.

          (1) Battle Hymn of the Republic is the reference that comes to mind.

        2. “Irish and Italian immigrants in the USA were considered non-white”–That is a canard. What would it mean? Did Irish and Italian immigrants face prejudice and discrimination? Sure. But were they denied the right to vote anywhere in ante-bellum America? No. Were they legally discriminated against in their ability to become citizens? To claim homesteads? Were they forced to ride in separate sections of public conveyances? Or to attend separate schools? No, no, no, no, no. To equate the prejudice encountered by various white ethnic groups with the injustices visited on blacks is an unjustifiable denial of the uniqueness of the black experience in America.

          1. A miscegenation conviction was overturned on the grounds that you had to prove that the woman was white when she was an Italian immigrant.

          2. The Southern establishment thought that miscegenation laws were stupid–heck, half of them slept with their black housemaids–but the masses enacted them. Judges fought back by overturning convictions on technicalities, such as the claim that an Italian person could be black (obviously, so could an English person for that matter) and that therefore the prosecution had failed to produce evidence of an essential predicate for conviction. No one made Italian immigrants ride in the colored car.

  5. >By ‘imperial’ I don’t just mean ‘during the period of the formal Russian Empire’: the approach of the USSR to places like Georgia was still fundamentally imperial in practice, even if some parts of the ideology and rhetoric had nominally changed.

    I’m interested in learning more about this. Could you recommend some books on this topic?

    1. That’s a really good question: unfortunately as I’m very much a pre-modern specialist (so I’ve dealt a lot with but never actually taught modern Georgian history) I’m not sure I’m the right person to give you a confident answer in terms of good intro-level texts, especially if you want a good analysis using imperialism as a framework. There is quite a bit written on this, most of what I’ve picked up has been from a rather mixed range of papers and sources, but I don’t think I have a good recommendation for a general English-language primer. It’s something I should have a better answer to though, will look into it for future – thanks for asking.

  6. The story of how Tamar became queen would make an amazing novel or even a movie. I’d never heard of any of this so it was really enlightening. Thanks!

    1. Yes, it’d be interesting to see more creative responses to the period. In some ways the dramatisation of Tamar’s life really even started during her reign – the opening of The Knight In Panther Skin, generally considered one of the greatest Georgian epic poems and written during Tamar’s time, is set in “Arabia” but with characters in a very analogous setup to the Georgian situation, with an elderly king being about to die without an heir, planning to enthrone his daughter, and so on. There’s an early discourse in it between the king and his viziers where they spend a lot of time making the case for women as being fit to rule, which is super interesting.

  7. Really neat to see the intersection of tech and humanities.

    Sometimes I feel as though there is a lot of missed low-hanging fruit in research simply because the “sort of people who go into engineering” and “the sort of people who go into humanities doctorates” barely talk to each other, let alone collaborate.

    Oddly, I’ve met a ton of other engineers who love history or philosophy — but very few of them actually did anything with that interest at university. So they end up reading pop-history stuff, but almost never talk to an actual historian.

    I wonder what other kinds of opportunities there are for the rare folks who have skills from both worlds.

    1. The lack of intersection or interconnectivity between disciplines isn’t something limited to humanities and engineering, or “hard” and “soft” sciences; even within the humanities people could benefit from talking to others in a slightly different discipline, and it would help clear out some of the most out-there works.

      My go-to example of this is Florian Curta, specifically his work titled “the Making of the Slavs.” It’s fundamentally a thesis based on archeological evidence – or the lack thereof, as his core argument revolves around there being no material culture that would evidence a Slavic migration – and used to argue Slavic peoples emerge as a compound people and pigdin language of subjects of the Avar khaganate. Which is the poster-child example of a work that stands on its own within their own discipline and falls apart when introduced to something different, in this case comparative linguistics.

      If people worked on cross-contact between fields of study and disciplines, there’s definitely a lot of progress that could be made. New ideas, new subdisciplines, and maybe putting to rest some of the weirder lingering theories and pieces of junk history.

      1. My favorite example is how a historian of science has to understand both the chemical reaction involved and the amount of soot in the 19th century air (and how it would affect the reaction) to understand certain 19th chemistry results.

    2. “I’ve met a ton of other engineers who love history or philosophy”

      But to import techniques from Domain X into the humanities, you need the reverse: A humanities scholar with a deep knowledge of this technique from Domain X. You are not going to introduce anything to a discipline without becoming a student of that discipline.

      (Francis Crick started out as a physicist, but I think it fair to say that by the time he was invited to Stockholm, he had become a biologist.)

  8. The argument of this post is interesting, but out of my depht. However, there is this (tangential) thing that irks me: when I was in high school (in Italy), I studied that the middle ages start from 476 (Odoacres deposes the last western roman emperor) and end in 1492 (Columbus reaches the Americas). The middle ages are then split in two: up to the year 1000 there are the high middle ages (the middle ages as a dark period of barbarism and illiteracy), and from AD 1000 onwards the low middle ages, with an upward swing culturally and economically, that end in renaissance/age of the exploration/early modern age. This classification makes a lot of sense, certainly it does for italian history.

    But the OP calls the year 1300 “high” middle ages, whereas for me it would be close to the end of the middle ages and certainly in the “low” middle ages. It also says that for the soviets the middle ages coirresponded to feudalism so went on to 1800 (I read somewhere that the french historian Le Goff similarly saud that the long middle ages ended with the french revolution).

    Now I’m not going to say that one terminology is better than the other, it probably depends on what part of the world you’re looking at and what definitions you are using, but this is confusing: if for one “high middle ages” goes from ad 500 to 1000, and another means from 1000 to 1500, evidently a lot of confusion can arise.

    So if you look at wikipedia nowadays you’ll see that “historians” don’t like the concept of “middle ages” because it is misleading. Well it certainly is if everyone uses it to mean different time periods!
    So we end whit that thing that happens in japanese mangas (I’m an avid scanlation reader, I’m a sinner I know) where there are guys that look like germans from 1750 and someone says “thios is a world that is similar to european middle ages”. This clearly is going to create a lot of confusion!

    1. This is probably due to an Italian connotation of “high” as “intense”, while in English the connotation is more as “great” ?


      And that’s nothing compared to the use of the word “modern” !

      A lot of people still use it to mean “contemporary”, even though I’m pretty sure that future historians will have the “modern” era (or whatever other name they will pick) end up in the 20th century (my bet is 1945, with 1905 and 2001 as extremes) : we don’t have much in common with Christopher Columbus / the Ottomans any more, *especially* in terms of how warfare between great powers is conducted ! (=> not directly)

      1. If the Middle Ages could last for a thousand years, and the classical period for a similar time, I feel the modern era can be with us for quite a while.

        If there is a break period before the end of this century, I would expect it to be due to something exceptionally dramatic, such as global thermonuclear war, or AI.

      2. Things get interesting when some version of “modern” becomes associated with a specific past era. The Hypermodern school of chess is associated with a set of ideas, openings, and masters from roughly the 1920s-1940s, which makes one wonder what superlatives are left to describe more contemporary developments.

    2. Prof Devereaux wrote a whole bunch of posts about Game Of Thrones where one of his main points was that it was much more like the early modern era (1500s-1600s) than the medieval era people claim it resembles.

      1. I usually see about 1000-1300 or so as the “high middle ages”, with 1300-1500 or so as “late middle ages” (usually the Black Death is used as a boundary marker)

        Scandinavian chronology is of course further different, in that we don’t generally start the “middle ages” until about 1000 or so, with various local subdivisions prior to that.

        So: Roughly 500-1000: Early middle ages (overlapping with late antiquity in complex ways) 1000-1300: high middle ages, 1300-1500: Late middle ages, 1500-1800: Early-modern. (with various more specific dates sometimes used, but these are roughly the ones)

  9. Are these prosopography projects visible online? Can we try them out?

    It seems like, much like with scientific visualization, figuring out how to present these datasets and diagrams in understandable and engaging ways to non-experts will be helpful. (And has overlap with video games.)

    1. My project isn’t publicly available yet, I’m hoping to get there in the next year or two. But you can browse a lot of more index-style prosopographies, such as the Prosopography of the Byzantine World or People of Medieval Scotland databases which should be easily google-able. Their interfaces are often a little bit clunky, but there’s tons of interesting stuff in there.

      And yes, engagement is something we need to think about more – especially as knowing what you’re looking at and what the database is claiming about its own entries isn’t always obvious. Hopefully things like this to improve awareness and interest in the field help a little!

  10. Very interesting post – kudos to the writer and the editor for featuring this on the blog!

    I’m a person who works in IT, but with an interest in the humanities – actually my undergraduate degree was in the humanities, although it’s never been directly relevant for my work. I think it’s great to see projects like this that bring digital toolkits to the study of history – very interesting to see the types of inferences and theories you can come up with just by having structured data to play with.

    It would be nice to see these sorts of data sets become open and shared, perhaps that would allow people who wouldn’t normally care about medieval history to take an interest

    PS: Are you using ontotext graph db for your data? Is it in RDF format? When your PhD thesis is complete and published, I’d love to read it for more tech details 🙂

    1. Thanks! On the tech side, I’m using a neo4j graph database with a PHP front-end that I built myself. Happy to send some notes over sometime if you’re interested, just drop me an email (firstname dot lastname at

  11. Great post! Absolutely love this! My data science masters capstone was on graph theory, albeit in a PR context. (Looks like you maybe even used the same tool, Neo4j?) Graphs are so much fun. It is amazing how this type of analysis makes hidden communities and influencers pop in a way they don’t through unstructured text or traditional flat files, at least not without a lot of time-consuming digging. How widespread is this in the history profession? I’ve been surprised in the communications field how slowly it’s spread to more mainstream tools that put this in the hands of non-technical users since marketing departments in academia and communicators with a more technical background proved the value some time ago. Still, I think we’re on the way to having a basic understanding of networks become a core competency in many fields.

    1. That’s neo4j, yes!

      I think it’s becoming more common to use this sort of analysis in history: it’s something one needs a lot of caution on, especially as there are lots of network approaches that assume much more robust data than I’m dealing with so I’m often using network concepts (like brokerage) as a tool of understanding more than I’m necessarily using the calculation-and-algorithm side of social network analysis as a methods toolkit (working out what out of what you’re seeing is an artefact of how the sources are written versus what’s an artefact of the society you’re examining is a long hole to dive down!)

      The adoption of network approaches is sometimes happening slowly, though, in part because history as a whole field is really really struggling for resources and in part because this sort of analysis is only as good as what you can feed into it and the understandings of context you can build around it. Whilst I think there’s tons to be gained here, in many ways some of what I’ve done has ended up opening ideas and avenues that need someone with more specific methodological/theory specialisms to pursue them. Good close-reading histories of gender or race/ethnicity in the Caucasus in this period would be invaluable: so in terms of how widespread things are, there’s often a multiplier effect on understanding by having a diversity of approaches.

  12. “A simplistic understanding of the past can mean a simplistic understanding of the present that is assumed to flow from it – and that, all too easily, can turn into a simplistic case for violence when ownership of the past is applied to ownership of culture and land.”
    It is a common conceit among left/liberal intellectuals that sophisticated or nuanced understandings will lead to the political conclusions they prefer, and that their political opponents are untutored rubes. It really doesn’t work that way. That said, I agree that historical understanding is important; I just think it leads to political conclusions different from those of the average academic.

      1. Occam’s razor would suggest that, all else being equal, simple models should always be preferred to complex ones. Indeed, a simple theory that roughly matches the facts is more likely to be right than a complex model that exactly fits the facts.

        As Francis Crick remarked, “Some of the facts are sure to be wrong.”

  13. Thank you James, for this fascinating look at your research program! I have met a few people in the past who self-identify as prosopographers, but never had a close-up view of one at work. I have just one unmitigatedly pedantic question: You refer to medieval chronicles, but do you also draw on non-narrative sources, such as church-dedication inscriptions? I imagine these would be especially useful, since they match a name (e.g. “Liparit, Eristavi of eristavis”) to a particular place and time. There must be hundreds of dedications of this kind throughout Georgia, but I do not know if they have all been transcribed and published.

    1. Hi Kevin, nice to hear from you here!

      The answer is basically “not yet” – there are some catalogues of these things, but they’re not easily accessible and the chronicles are much lower hanging fruit for the time being. I absolutely do want to make use of inscriptions more in future developments of the project, though, it’s a major aim of mine.

  14. “A simplistic understanding of the past can mean a simplistic understanding of the present that is assumed to flow from it – and that, all too easily, can turn into a simplistic case for violence when ownership of the past is applied to ownership of culture and land.”

    This argument would seem to imply that a complex understanding of the past can lead to a complex case for violence in the present. In fact, since a complex model of the world will have more degrees of freedom than a simple one, it would probably be easier to find some way to tweak it to produce a desired justification for violence.

    It is better for us to have a detailed understanding of the past than a less-detailed understanding. Truth is a goal in its own right. And in general, a better understanding of the world should lead to better judgements about it. But there can be no guarantee that any given truth will be convenient for the cause of peace, or for any other cause.

  15. Not on topic, but I don’t know where else to contact you. Anyway, I see you’re still on Twitter, despite what Elon’s been doing to the place. Can I convince you to come to Mastodon? If you’re trying to pick an instance, would probably do quite well for you.

  16. How much of an effort is there to produce these kinds of graphs communally as part of the research process?

    I was just thinking… in some ways it’s kind of horrible that the best way to go research something is to read half a dozen expensive books (and then track down their citations) to end up with a synthesis only in your own head, instead of having that information already indexed and cross-referenced in a database like this.

  17. Of course medieval kingdom never can be equated with modern-period nation states. But attempts at denying that the medieval Georgian legacy is modern Georgian’s heritage first and foremost will go futile.
    This medieval unity and its remembrance (which existed through the whole time after it and was not created in the 19th century) was the only reason the Russian colonization failed at dividing Georgians into several pity “nations”. It is why Mingrelians rejected Russian attempts at making Mingrelian a literary language with cyrillic script and introducing it to church services.

Leave a Reply