Episode 66. Rune Rattenborg, Seraina Nett, Gustav Ryberg Smidt: Geomapping Cuneiform: transcript

00:13 JT
Hello, and welcome to the Thin End of the Wedge, the podcast where experts from around the world share new and interesting stories about life in the ancient Middle East. My name is Jon. Each episode I talk to friends and colleagues and get them to explain their work in a way we can all understand.

00:32 JT
We’re used to the idea that clay tablets survive very well in Iraq’s soil. And we take for granted the astonishingly long life of the cuneiform writing system: three and a half thousand years. This all means that the cuneiform corpus is huge. But just how big is it actually? What are those tablets? Where are they now? And where did they come from in the first place? These kinds of basic information can tell us all sorts of interesting things about ancient Iraq, and the wider cuneiform world.

01:09 JT
Our guests have spent the last few years systematically gathering and analysing these data about tablets around the world. They tell us how they did it, and what they learned. So get yourself a cup of tea, make yourself comfortable. And let’s meet today’s guests.

01:35 JT
Hello, and welcome to Thin End of the Wedge. Thank you for joining us.

01:39 RR
Thank you very much for having us.

01:41 SN
Thank you for having us.

01:42 GRS

01:43 JT
Could you tell us please: who are you, and what do you do?

01:46 RR
My name is Rune Rattenborg. I’m an assyriologist from the University of Copenhagen. And also an archaeologist through my doctoral degree from Durham University. And I’ve previously been involved with a lot of archaeological fieldwork in Jordan, Syria, and Iran with various projects. And my interests are quite wide, both across assyriology and archaeology. But my primary focus is on digital assyriology, digital humanities research designs, the materiality of writing. Mainly in a large-scale and comparative data-driven perspective. I like data a lot.

02:24 SN
My name is Seraina Nett. I am also an assyriologist. I’m currently employed as a part-time lecturer in assyriology, at University of Copenhagen. And for the past few years, I’ve also been a researcher affiliated with the project GLoW — Geomapping Landscapes of Writing — at Uppsala University in Sweden. I originally have a background in near eastern archaeology and linguistics, and have a Master’s from the University of Bern in Switzerland, and then a PhD in Assyriology from the University of Copenhagen, in Denmark. By training, I’m originally a sumerologist. And so I work mainly with Sumerian texts and languages. So I’m mostly interested in the earlier parts of Mesopotamian history — the period up to roughly 1500 BCE. And in particular I work with the time of the Third Dynasty of Ur — so the 21st century BC. But I’m also more generally interested in social and economic history. So lives of normal people, gender, social status, and so forth. And also the social context of language, and how we can address these questions, again using digital methods, including Big Data, larger datasets, and so forth.

03:38 GRS
My name is Gustav Ryberg Smidt. I have worked on the GLoW project. And now I’m currently working on a project at Ghent University. I have done my master’s degree in assyriology at the University of Copenhagen, where I also know Seraina and Rune from. And I’ve mainly focused around digital and computational humanities in assyriology. And I’ve also done a bit of computer science in Copenhagen. I’m interested in the Old Babylonian period and how we apply different computational and digital humanities approaches to that corpus.

03:40 SN
Seraina, you mentioned the wonderfully named GLoW project. Can you tell us more about that?

04:17 SN
Yes. So the GLoW project, GLoW stands for Geomapping Landscapes of Writing. And this is a project that Rune and I and several research assistants had at Uppsala University. It actually just ended; so this is very recently over. And this is a project where we’re essentially looking into three main avenues. We wanted to create an index of the geographical locations of where cuneiform texts were found. But also where they’re currently located. So museums, libraries, private collections around the world. And to do this to provide accurate, correct, usable geographical data, because a lot of the geographical data that you could find, when you look these sites up online and, yeah, was not particularly accurate. It was often crowdsourced, which is great. But we wanted something that was curated, that would provide a good context for these archaeological sites where the texts were from. And it was a particular research assistant–Carolin Johansson–who created these datasets of geographical data: latitude, longitude, ancient name of the site, modern name of the site, etc, both for the ancient sites and for modern museum collections.

05:37 SN
And then we also wanted to make more texts available electronically, by adding more information to the available online databases. So in particular, the CDLI — the Cuneiform Digital Library Initiative; the largest existing database within assyriology — because the CDLI as a database was and is by no means complete. So not every single cuneiform inscribed object is in there. This is due to the history of the database, and it being a collaborative project. The content in particular was quite skewed towards the earlier periods of Mesopotamian history. And so another aim of our project was to contribute as much as possible to this database. So to add more content into the CDLI database. And also thus making it a more reliable resource. And to actually find out how much was missing from this database. We had Gustav, who was a research assistant for our project at the time, actually start by just looking through the literature and try to find out how many texts are there? How many tablets do we have? And what percentage of those tablets is actually represented in the database?

06:51 GRS
I was working with a lot of the data collection for not so much the geographical aspects of it, but the textual aspects. So I worked with collecting the “estimate dataset”, as we call it. So the estimates of how many texts were found at a given site. And then I also work with trying to take some of the bigger corpora. So some of the sites with a bigger core, and try to figure out how much of the data is correct. And then basically go through each text and correcting things that might be wrong. So that could be go through all of texts found in Ur, for example, and go text-by-text, and simply looking okay, is this an administrative text as it says here? Or should it be corrected? And that process was, of course, not always going from one text to another, but you figure out which texts, which publications, which datasets tend to be more corrupt than others.

07:44 JT
Hmm. How do you deal with that on a practical level, then? Because if you have a dataset of half a million points, and even 5% of it is something where you have to check, that must multiply to a huge amount of time and effort. How do you actually manage that level of detail in such a large dataset?

8:02 GRS
Well, you begin from the point where you can find the most information that seems somewhat reliable. And that’s of course, a bit of an iffy answer, because “somewhat reliable” is already difficult to assess. But here, of course, this is what the normal research skills come in you learn throughout your studies about figuring out what seems to be a good source. Always, of course, try and find something that’s related to the archaeological excavations, if that exists for a place. And then simply try to see if somebody has in that connection published a catalogue. If not, then you find someone who might have done a catalogue afterwards. And if that’s not possible, it’s going through each publication per site, and finding the section that’s called “textual finds”, and then noting which ones they have found.

08:51 GRS
And then you get a picture of how many texts are there. Plot it into your Excel sheet or whatever program you’re using to help you look at the data. And then you go through each element and see, can I verify this information? Does it say anything in the publication about it? Or can I actually see the text and try to verify it? Or do I have to leave it as unknown? Or change it to something else that seems more appropriate or is probably correct. But yeah, when you’re talking about so many texts, it’s of course, a huge undertaking. That’s also why we didn’t manage to do that with every single site. But we simply said, let’s start from the south of Iraq then work our way up north. I think going a bit north and then west, I managed to get up to Nippur and finish Nippur.

09:38 EB
Sounds like a lot of work.

09:41 GRS
A lot of hours, sitting with stacks and stacks of books, flipping through them and trying to figure out is this a good publication of economic texts, for example? And then this seems okay, and then just go through the catalogue and say, this doesn’t correspond to my other information. Is this one then correct and then go and double check it in the text. And say then OK, this is correct, then on to the next one.

10:02 SN
So this is the starting point for our project. We looked at what was there supposed to be in terms of total numbers? And what are we missing? Where should we start in contributing additional data to this database? And then of course, once we have actually gathered all of this data … this is what we are currently very much in the process with … is to use this dataset that was created. So the geographical data, but also the more detailed complex dataset to actually look at, where do texts come from? What kinds of texts do we have? When do they date? What material are they made out of? What languages are they written in? And so forth. So we kind of did a lot of groundwork to create more reliable data and then started to actually look into what does this data tell us on a larger scale?

10:51 JT
I wonder if we could start digging a little deeper at what should be the most simple question — how many tablets are there? Cuneiform famously has this very large corpus. People throw around numbers, like half a million, a million. But it’s a very complex problem, isn’t it? Because not only do you have the size, you have the complexity in that people talk about tablets either as museum numbers or things that come from excavations or through their publication reference. I guess from a museum perspective, the very basic question of how many tablets are there in a museum differs a lot. It depends on how you count them. So how many tablets are there actually in the world? And how do we measure that?

11:31 RR
Yeah, that is a very simple question. It’s also perhaps not surprisingly far from the simple answer that you have to it really, Jon. I think the first thing you should ask is, of course, as you say, how do we count tablets? What are we talking about? Are we talking about fragments? Are we talking about complete tablets? Are we talking about the hypothetical number of complete tablets that may be derived from any number of fragments? Are we talking about something else?

12:00 RR
Already there, you’re running into some challenges in different studies, different databases, different repositories, different indices, take different attitudes toward this very simple exercise of counting, if you will. From our perspective, we initially thought, well, just as anybody else, this is quite a simple question, you know, you just have to go out and count them. But once you dive into the various databases and the general literature in the field of archaeology, you’ll find out that it’s actually far from always so straightforward just to pick up a number. Studies we managed to conclude so far, based on the data collected, that the GLoW project amassed, came up with an estimated 430,000 tablets and fragments that could be related to locatable archaeological proveniences. But if you go back to the seminal, important, and actually also the first thorough study of the overall size of the cuneiform corpus back in 2010, published by Michael Streck, called “Grosses Fach Altorientalistik” — a fantastic article that every assyriologist should read — you will see the overall number there landing at around 430, oh, no sorry, 530-540,000 tablets and fragments worldwide.

13:16 RR
And then, of course, you can also then scale up that number, if you will, by talking about all of the inscriptions that have not been counted as such. It is, theoretically speaking at least possible to calculate how many inscribed bricks you might find in one certain ziggurat, for example. So the calculation can sort of go on and on and on. But I think it’s generally fair to assume at the present time that the studies that have actually dealt with this question, you would land at somewhere between 500,000 and 600,000 tablets and fragments known worldwide.

13:50 JT
You were very careful there to specify tablets that were in some way relatable to a location. Many tablets were purchased in the 19th early 20th century, so don’t really have that kind of a provenance. Do you have a sense of how many other tablets there are? Or as far as the project was concerned, were you prepared to accept “assumed Babylon”, say, as an origin, a geographic location?

14:17 RR
Yeah, that’s another problem that we ran into often, of course, because you might also ask, of course, what you actually mean by archaeological proveniences? The provenance of a tablet, how secure is that? Cuneiform is a very special corpus in that way. Because given the size of the entire corpus and the size of individual assemblages that you excavate from archaeological sites, we often can say with quite some certainty, that a tablet, for example, from Drehem is from Drehem, even though it’s not ever actually been recorded as being excavated there. We seem to just assume so on the basis of information that is internal to the inscription in the tablet. But given the number of inscriptions that share common traits coming from the same archive or from the same dossier of texts, you can often rely on that information to an extent that, at least at the face of it, seems much more reliable than what you’d find for many other cultures of ancient writings that are being studied.

15:17 RR
So in that sense, when saying that we the GLoW project, were able to allocate somewhere around an estimated 430,000 inscriptions to archaeological proveniences, that, of course, includes a huge number of texts that are not actually recorded in a proper archaeological sense provenience-wise, but simply the tablets that can be with the general certainty that’s assumed within the discipline of assyriology, associated with this archaeological site.

15:47 EB
I just have a question about letters. What do you do when you have a letter that is sent from one place–like point A to point B–is found in point B? How do you deal with a letter like that?

15:59 RR
Yeah, that is another issue. And one, which you can sometimes find a lot of confusion as to what the actual provenience, of a tablet or a letter in this case will actually be. There are several online databases now that have begun to operate with distinguishing between information that’s related to where it’s happened was produced, as to where it was actually archaeologically found, which is the kind of qualification that should certainly lie on the mind of every assyriologist going forward. But it’s not something that’s been fully implemented yet. In those cases, often, we would have to spend a lot of time diving into the exact nature of a provenience that was given. But it really is a question of examining and differentiating between different types of information to relate to a tablet. So in the future, hopefully, we would have more well-curated information that would allow us to look at the archaeological provenience, and the place of production of cuneiform tablets at the same time, which would no doubt reveal a lot of interesting patterns also.

17:06 EB
Can you give some examples of some of the things that you’re wanting to do? The kinds of research questions that you’re interested in looking for with these to really show why it matters to know where these tablets were from?

17:18 RR
I think that the general question as to why the archaeological provenience or the geographical origin of a cuneiform tablet really matters, is something you should consider against the history of assyriology as a discipline. By tradition assyriology is an epigraphical, philological discipline. So the primary focus is quite naturally really trained on the inscription themselves, and not so much, at least not in the first instance, on their archaeological context. But it does matter a great deal where an inscription is found and what its archaeological context is, and how it compares to other tablets of a similar genre, format or, or time period.

18:00 RR
I think the cuneiform corpus is particularly interesting when considering geographical distribution, because it is, as we’ve already mentioned, one of the largest, if not actually the largest discrete corpus of writing from the ancient world. And if you’re only talking about a body of writing that you’re studying of perhaps a couple of hundred artifacts from a couple of different archaeological locations, of course, spatial distribution or geographical distribution is not going to tell you very much. But when you’re looking at hundreds of thousands of records, and trying to understand their distribution over wider areas, some quite interesting things will necessarily reveal themselves.

18:37 RR
In the GLoW project, we’ve been very interested in trying to understand, for example, where cuneiform can be said to have been a prominent feature in the material record of ancient society, and where it seems more of a peripheral occurrence, if you will. The study we published in the autumn of 2023, we were looking at, or trying to compare, the number of cuneiform inscriptions that were found in any given area in archaeological excavations with the density of locations in which they were found. So basically, you’re looking at how many texts do you have from a given area, versus how many places where you’ve found texts do you have from a given area? And what is interesting when you’re looking at it that way, so trying to qualify not the number of texts, but the regularity with which texts turn up at archaeological excavations, you see very different patterns.

18:50 RR
And my impression is that because we don’t have a firm grasp of the full geographical distribution of the cuneiform script, or cuneiform inscriptions, we’re also not ideally equipped to talk about how common cuneiform texts or cuneiform writing really was in the ancient world from a quantitative perspective. So, for example, we pointed out as part of our research that it’s interesting to see that in Israel, for example, and in the Palestinian territories, you have a very high concentration of the number of sites with finds of cuneiform inscriptions, although the overall number of inscriptions is very low. But if you compare the density of the number of archaeological locations with finds of cuneiform inscriptions in Israel and the Palestinian territories to the density of archeological locations with cuneiform inscriptions found in southern Iraq, it’s actually almost the same. Which is interesting in the sense that if you’re looking at a simply from the perspective of the distribution or the regularity with which you find cuneiform inscriptions on archaeological sites without thinking about the numbers, from that angle, it seems as if cuneiform is actually just as common in a lot of areas we tend to think of as peripheral, as it is in the areas we tend to think of as the core areas of where you use cuneiform writing in the ancient world.

20:50 JT
Yeah, that study was quite surprising to me, actually, because I tend to think of tablets as group objects for a number of reasons, I guess. Cuneiform is very heavily administratively focused; you have these large organisations producing hundreds or thousands of documents. And you imagine that when you find one, you probably find hundreds or thousands, of others. But what was quite striking was how many sites produced just one or a handful of tablets. I was wondering if you could say something about what you think that tells us?

21:21 JT
To continue in my train of argument from earlier on, it would be interesting to think about what the cuneiform corpus actually is in this perspective. We did some comparative examinations and workshops during the period of the GLoW project at Uppsala University, with researchers from other philological and epigraphical disciplines. And the cuneiform corpus is quite unique in that the predominance of certain genres is very different from, say, if you’re looking at the corpus of Latin inscriptions from the Roman Empire, or Runic inscriptions from Scandinavia, for example. If you’re looking at strictly from the perspective of geographical distribution, you will see that for the Roman and Runic epigraphic corpora, which are mainly concerned with monumental inscriptions, you know, Runic stones that are standing out in the open commemorating someone’s grandfather or brother or mother, or architectural inscriptions from the Roman world, it produces a very, very large number of archaeological proveniences, so different locations where you find a stone inscription. But the overall number of inscriptions that you find in one place is comparatively low. Whereas if you go to a corpus where the genres that are represented are predominantly administrative or economic texts, letters, so everyday types of writing, it’s actually quite common for discoveries of texts to contain these very large lots that Jon is talking about.

22:18 SN
If I might add to that, what’s of course also worth keeping in mind here, if we talk about the cuneiform corpus as we discuss it in the context of the GLoW project, is that, we are talking about 3000 years, 11 different languages. So there are also differences if we look at it in a more kind of fine-grained. If you look at different geographical regions, if you look at different languages. For example, if we look at the Kingdom of Urartu in the first millennium, there’s lots of one inscription sites, but those are all rock inscriptions; they’re all royal inscription. They maybe tell us something else than all the sites in the Mesopotamian periphery that only yielded one cuneiform tablet. But yeah, this is really something we should look into further and we will look into further. So these individuals finds, what are they? How did they get there? What genres are they? What context do they belong to? And which time periods did they belong to ultimately?

22:49 RR
So if that’s the kind of genres that you predominantly find in a corpus, you tend to find a lot of them in the same place. But I think a general argument or the interesting thing about how many places there were, where you would only find one or just a handful of inscriptions within the cuneiform world, was that if you qualify for the serendipity of archaeological discovery, you could say, it will actually suggest that there’s a lot more and like really a lot more, cuneiform inscriptions to be discovered in many places.

24:20 EB
This all seems like somewhat basic groundwork when you’re wanting to look into an ancient society of like, I’m looking at a text and I want to know exactly where it’s from. So I was wondering, why has this work not been done before? This seems like it’s strangely difficult to do. Do you have any inkling or comment on why this has been not done before? Has it been difficult to work out where a tablet is from historically in assyriology?

24:48 SN
Yes, it has been difficult. It still is difficult. {LAUGHS} There’s many reasons why it is difficult. So on one hand, of course, again, as Jon already pointed out, there’s a large percentage of the tablets that we are dealing with today are from the art market. They have no nice, good archaeological provenance. Often the text finds are also quite badly published, especially older text finds. You have things like “I was given this tablet by someone in a village 10 kilometers south of Mosul”. So linking that to an archaeological site is very difficult. This is also one of the reasons why we have worked with ranges of probabilities. So how likely is it that this text is in fact from this particular site? But also, sometimes it’s … the museum collections are not very accessible. It’s difficult to figure out if the text that is in the collection of a museum is the same text that was excavated at a certain archaeological site. Texts are being transferred from different collections, especially without a clear ID or reference or reliable photographs of the tablets. It’s sometimes very difficult to figure out if this tablet we are talking about in publication A is in fact, the same tablet that is also available in publication B. And so that has been an ongoing struggle, I would say, in this project.

26:10 JT
I can only imagine.

26:11 RR
I would add that there is also the thing to consider now that Ellie’s asking, saying that this seems basic work, why it hasn’t been done before. I mean, of course, it has. Assyriologists do that all the time. But I think we often tend to forget even as assyriologists, that the cuneiform corpus is extremely large, by any standard. I mean, there are excellent databases for Latin inscriptions, for example. But these are also the results of many decades of work by a much, much larger cohort of scholars than assyriology could ever aspire to gather. So I think really just reflecting that there is still loads of very basic cataloguing, recording and ordering of data to be carried out in the field of assyriology. Not just because this information is hard to find. It’s also just simply because there are so many tablets to try to keep track of.

27:06 JT
I wonder if we could pursue that a little bit more. So you mentioned at the start of our conversation that you’d pull together your own corpus using the various databases that are available — various projects, institutional repositories. Could you say something about what that was like? How did you integrate so many diverse datasets? And what kind of recommendations would you have for how they could curate the data?

27:29 RR
Yeah, that’s the kind of problem that I’m still struggling with even now, after three and a half years of GLoW. I think there are tons of things that we would have done very differently, have we known just how complicated that would be by now. I think the most important part of the question as to how you integrate large datasets is really to put the emphasis on integration and not on the datasets themselves. I mean, of course, the rise of digital humanities computing and humanities and assyriological research has meant that there’s an astonishing body of digital data available out there that you can go and use many ways. But the challenge is that we’re still at a stage where not much of this is actually thoroughly mapped to the same standards. That doesn’t mean necessarily that one standard is better than another. It’s just to say that even within such a relatively constrained academic discipline as assyriology, there are many different ways of structuring, ordering, categorising, and labeling your data. And that inevitably produces challenges and some quite difficult choices in terms of how you’re going to order things back and forth.

28:40 EB
This sounds like you have to make a lot of decisions about what kind of information you’re including in your database. How big a discussion was that, creating the right kind of framework for the project?

28:54 GRS
We put a lot of thought behind what were we going to look for. As mentioned, we had CDLI in mind for our work. So we of course, had to work along the lines of them, so that it was compatible. There are some things we can’t necessarily get, and some things just takes too long. So this is why I think it’s important to consider what information is useful for the user. To determine which texts to look at and which texts not to look at. What information can we gather quickly that can give that for a user? And then of course, it was a long debate sometimes, I think the famous one we came across a lot was a genre called “school texts”. Is that something we will include, because a school text is a context. It’s not a genre of texts. What if this text has two different types of school texts on it? It has maybe a lexical list on one side and literature text on the other side.

29:52 GRS
So those kinds of questions prompted a lot of issues which we then had to deal with. Are we then working with multiple values per cell or are we then taking it up another hierarchical level and saying, okay, can we find another overarching category? Is it an academic text instead? We managed to solve to some degree. But of course, there are also still holes, where it’s not just us, but the whole field, as such that needs to get some sort of consistency level where we can work together and not just in small silos of field or specific for, let’s say, one period. You have a specific set of genres or specific set of languages, for example. But we need to be able to broaden that out.

30:33 GRS
So when we’re working with so many texts, we have to of course, consider a huge amount of different databases as well, or datasets and repositories, that are often very specific to a specific period or a specific genre. And we want to be able to integrate that into one place, CDLI in this case. But the important part is more that there is some sort of common place, so that when you get the data from somewhere else, that the values, the way they describe the data is consistent in a way that you can either say there’s a one to one relationship with whatever you have in your own dataset, or that it can be fairly easily transferred into a set of values you have in your dataset.

31:20 RR
Part of the research design for GLoW, which I think was also a major factor in us getting the grant from the Riksbankens Jubileums Foundation in Sweden, and then we got it, was that we did not want to create another database as such, but rather, we wanted to build a large data collection that was tailored to integrate with the existing digital ecosystem in assyriology. So we basically built the global database around the data model of the Cuneiform Digital Library Initiative. Because in that way, all the data that was collected during the GLoW project was, ideally speaking at least, tailored to be transferred directly to the CDLI upon the conclusion of the project, without too much hassle.

32:05 RR
This would mean that the data will be more readily available. And we would avoid the need to maintain an actual database for many years for which we did not have the institutional backing, and certainly not the money. And I think that is a problem that many assyriological research projects will be facing. And I would strongly encourage people to think about how to integrate their data within the existing array of data collections that are available on the World Wide Web. But it also meant that we spent a lot of time trying to getting all of these different data collections to play and integrate.

32:42 RR
To go back to the more specifics of the question you asked, Jon, the way we started out was simply working mainly with the focus on the archaeological proveniences. So that was our benchmark. Working through all of the archaeological proveniences we could find for cuneiform inscriptions, and then building a data collection of cuneiform inscriptions from there. And then, of course, cross referencing that back and forth with the datasets that we could acquire, from downloads online, from all of the other amazing projects that are out there, like the Cuneiform Digital Library Initiative, the Database of Neo-Sumerian Texts, all of the various Oracc projects, and so on, and so on. There’s a huge amount of data that assyriologists can make use of now, for which we should all be very, very grateful. And then from there on, then building on to trying to attain as comprehensive a picture of the overall corpus as is possible. But that sounds a lot easier than it actually was. So I think even now, looking back in this, as I said, a lot of things that I would have done differently. And also, I came to have much more respect for the amount of knowledge and the amount of specialisation and training that goes into understanding each of these large data collections on their own terms.

33:56 EB
It sounds like you definitely have a bit of a wish list, which is maybe a bit too technical for us to go into today. But do you have a wish list for what you would love to do next with this dataset? And how to develop it further?

34:11 SN
Well, maybe one thing to point out here, in general is that all of our data, either is or will soon be publicly available. So yes, we are working further with this. But it’s actually also available for other people to work with. Our geographical data set is already available for download and has been for a few years now. Other datasets are and will be integrated into the CDLI database. So this is all something that other people can work with, if they so choose. So far, I think most of our analysis have been focusing largely on the bigger picture because we were doing data collection until more or less the last minute. So we have been working based off of the estimated datasets that Gustav created, but now that we have actually our more in-depth data being incorporated into the CDLI, some of the things we’re also looking into is again, looking at distributions of chronological distributions of text genres, language use, and so forth.

35:12 SN
Myself together with Gustav, I’m looking at plotting grammatical variation, dialect variation, orthographic variation on maps. Because now that we have this reliable geodata, the geographical dataset that was created for GLoW, this is actually something that we then can combine with the data that we have available for, for example, Sumerian, where we can actually plot spelling variants onto a map and look at geographical variation there. Another thing that I have been playing with, it’s maybe a little bit left field, is looking into unseen species models. So this is something that comes from biology actually. It’s a model that lets you predict what is missing. And I’ve been trying to play with our datasets to see if I can get that model to predict what texts we don’t have. This has actually been applied in the humanities. It’s not something I’ve invented. It’s just been used for medieval manuscripts and other things. So far, I haven’t really managed to get very useful results. But this is something I’m still trying to tweak and work with as well to actually, now that we’ve collected everything that we have, do we have any way of predicting what we are lacking? What we don’t have based on the data that we have for the ancient Near East?

36:25 JT
Wow, wow, wow.

36:26 RR
I think I would also add that I hope that the work that we’ve been doing is also something will contribute to broader acceptance of the geographical extent of the cuneiform corpus. How big and extensive that body of writing actually is. And I think that’s the kind of information that is of interest to a lot more people than just assyriologists, and those archaeologists who take a particular interest in cuneiform. I’ve been recently working with a bio-archaeologist who takes an interest in understanding what data from administrative cuneiform texts, for example, can tell them about the flora and fauna of the ancient world. And in that kind of engagement, it’s actually worth something to have a more comprehensive understanding of the distribution of all or nearly all of the written sources that are available. And then using that as a roadmap to be able to single out, find, the kind of information that the bio-archaeologist would be looking for.

36:31 RR
So I think that or I hope at least that the work which happened during GLoW is something that will help motivate more cross-disciplinary research, more multidisciplinary research that is actually making use of cuneiform sources to meet their ends, or their research questions in another way. And I think that being able to clearly show the geographical distribution, being able to more clearly and more comprehensively index cuneiform sources is an incredibly important part of that exercise.

37:56 JT
I can think of lots of uses for this kind of information. It’s fantastic that you’ve been able to pull it together now in such a comprehensive way. In such a useful way. And it’s out there and can be used by everybody else. This is going to be a really important component of future digital assyriology. It’s going to have lots of uses for experts, people in other fields, but also in terms of public engagement, as well. So thanks very much.

38:21 RR
Thank you. Thank you very much.

38:23 SN
Thank you so much.

38:24 GRS
Thank you, Jon, and Ellie.

38:25 RR
It’s been a pleasure.

38:26 RR
I’d also like to thank our patrons: Enrique Jiménez, Jana Matuszak, Nancy Highcock, Jay C, Rune Rattenborg, Woodthrush, Elisa Rossberger, Mark Weeden, Jordi Mon Companys, Thomas Bolin, Joan Porter MacIver, John MacGinnis, Andrew George, Yelena Rakic, Zach Rubin, Sabina Franke, Shai Gordin, Aaron Macks, Maarja Seire, Jaafar Jotheri, Morgan Hite, Chikako Watanabe, Mark McElwaine, Jonathan Blanchard Smith, Kliment Ohr, Christina Tsouparopoulou, TT, Melanie Gross, Claire Weir, Marc Veldman, Bruno Biermann, Faimon Roberts, Jason Moser, Pavla Rosenstein, Müge Durusu-Tanrıöver, Tate Paulette, Willis Monroe, Toby Wickenden, Emmert Clevenstine, Barbara Porter, Cheryl Morgan, Kevin Roy Jackson, Susannah Paulus, Eric Whitacre, Jakob Flygare, Jon Ganuza, as well as those who prefer to remain anonymous.

39:47 JT
I really appreciate your support. It makes a big difference. Every penny received has contributed towards translations. Thanks of course to the lovely people who have worked on the translations on a voluntary basis or for well below the market rate. For Arabic, thanks in particular to Zainab Mizyidawi, as well as Lina Meerchyad and May Al-Aseel. For Turkish, thank you to Pinar Durgun and Nesrin Akan. TEW is still young, but I want to reach a sustainable level, where translators are given proper compensation for their hard work.

40:25 JT
And thank you for listening to Thin End of the Wedge. If you enjoy what we do, and you would like to help make these podcasts available in Middle Eastern languages, please consider joining our Patreon family. You can find us at patreon.com/wedgepod. You can also support us in other ways: simply subscribe to the podcast; leave us a five star review on Apple Music or your favourite podcatcher; recommend us to your friends; follow us on Twitter: @wedge_pod. If you want the latest podcast news, you can sign up for our newsletter. You can find all the links in the show notes and on our website at wedgepod.org. Thanks, and I hope you’ll join us next time.