 Hello everyone and thanks a lot for tuning in. My name is Isera Manowska and I'm a fellow at the Orchids Institute of Advanced Studies. I worked at the intersection between computer science and archeology to very different disciplines, but we can bring them together to do cool stuff such as my current topic of research which is modeling ancient demography. You may ask, why would you even do that? I mean, do we really care how many people there were in the past? Well, the thing is, even if we are not that interested in that particular question, in order to reconstruct, to learn more about any aspect of how human societies work or worked in the past, we need to know how many people were there. If you're interested in economy and trade, you need to know how many people traded with each other and what was the supply and demand like. If you're interested in how societies acquired their resources and whether that was sustainable, again, you need to know how many people were there and how many people did you have to feed. All the stuff about cultural change, about resilience of groups, all of those topics have one constant basis and that basis is knowing how big the group we're looking at is. That's why I focus on the baseline of any further research that can be done within those topics. You may think, OK, well, this is pretty easy nowadays, right? You want to know how many people live in Denmark? You just look it up on Wikipedia, right? Unfortunately, and we do regret all of that, people in the past did not keep records and that is a bit sad because that would be very, very helpful if they recorded every single birth and every single death. It would be fantastic, but that only happened way later on. So what we need to do is we need to use archaeological data as a proxy to basically try to figure out how many people there were in the past and how that number changed over time. The three types of proxies that I use in this project are very different from each other. And the idea is to use them all in combination so that where there's a gap in one line of evidence, we can fill it up with another evidence. The type of data I use are data such as funerary data. For example, in the city of Palmyra, wealthy elites, they made portraits of themselves that they located in their tombs. And you can also look at the capacity of those tombs and therefore kind of try to extrapolate as to how many people died. And if you're to die, then you probably were born as well. And that brings us to the burials, which is the most straightforward way of estimating people's births and deaths. To even approach this kind of data, you do have to use pretty robust quantitative methods in order to count the things, but also model them because it's not that you go into any archaeological site and you say, okay, this is dated to year 197 before Christ. That just doesn't happen. So there's a lot of uncertainty related to that data. And we have to kind of regard it in a probabilistic manner. There is a group of methods that is used for that. They're called heuristic methods. And this is what I applied to this kind of data. But that's not it, that's not the whole story. If you only looked at those data, then probably you will find some patterns which may or may not reflect the past. So what I'm trying to do is contrast them with other types of data that are completely independent. The next type is the city data. And that's related to a simple observation that big cities, there's usually more people. Small cities, there's usually less people. And it's not a very linear relationship. It has to be said. But in general, the size of a settlement and the changes to that size and the size of the houses, et cetera, you can try to extrapolate it to come at some kind of range of population that lived in that place. And how did that range change over time? And finally, well, you can put as many people in the city as you want, but the question remains, how are you gonna feed them, right? So the first type of data I look at is the hinterland data. This is related to the fact that the environment around us has certain carrying capacity. And carrying capacity means that there's just as much stuff that you can get out of it as there is. You cannot really extract more energy that is available in that environment. So you cannot feed the city the size of New York from three fields of wheat. That's just not gonna happen. In my case studies, this is all related to the land and water and the settlements around that would produce the food and feed that food into the city. By combining those three types of data, we're hoping that we'll be able to catch a glimpse of how the population of two different regions during the Roman Empire evolved over time. But there is one thing about looking for patterns in the data, but there's another thing about learning why those pattern happens. What has happened? Why did the population grow? Why did it go down? Why did it expand? What has happened to those people that their population trajectories took this turn and not another one? And the method that is my absolutely favorite method in the whole of science and the one that I use the most is agent-based modeling. So we can ask, how can we leverage agent-based modeling to learn more about the past? And this is one of the best methods to investigate different hypotheses, different theories about the past. This is something that we use quite a lot in archeology actually. So agent-based modeling is a simulation techniques. That means we have little agents and they live within a certain environment and they interact with each other and they interact with their environment as well. And through those interactions, we gain those population level patterns that we can then compare to, yeah, the three types of data that I just told you about. So what can we do? We can model those people, those little agents as they're creating their households, as they're having their babies, getting married, et cetera. So we can create this agent-based model of age structure population dynamics to create this kind of in silico, artificial society and see how their trajectories change. And that allows us to kind of test a whole range, a whole phase space of scenarios of how the societies could have survived in the areas where they did. And that allows us to integrate all of the environmental data because obviously those little agents need to eat. But also, and this is the best bit, right? We can drop some catastrophic events on them, which means you wanna drought, you have a drought and you see what happens. Nobody's hurt because they're all in a computer. So without doing any dubious experiments of people, we can actually test what would happen to a society under different circumstances, including some pretty extreme ones. And that also allows us to kind of measure how the cultural interactions might have shaped the way those people lived and how did they go around with their business. So I use several different frameworks such as NetLogo and Pandora, as well as General Programming Language Python to create the simulation in my computer, to create a little population of agents that are the virtual, the artificial past society from the times of the Roman Empire. And seeing their trajectories and comparing them with those that I see in the data allows me to basically say what might have happened in the past. And you may ask, why would we even do that? I mean, who cares, right? And here I wanted to bring this quote from Winston Churchill who said very famously, history is simply one down thing after another. And this is how we very often perceive history. This is how we learned at school. And this is frankly how we very often approach it when we study it. First thing, this happens. Second thing, this happened. Third thing, this happened. Perhaps this thing caused another thing. But it doesn't have to be like that. What history provides us with is the whole kind of set of experiments that human societies have done in the past. Some of those experiments were successful and those societies thrived. And some of those scenarios, they were not that great and those societies are not among us anymore. So I can say history does not need to be one down thing after another. It can be a source of information about how societies develop, how they respond to challenges, to change. What are the signs that things are going well and what are the signs that things really aren't going that well? And I'm really hoping that by doing this very synergetic type of research, crossing through several different disciplines of science, we can really bring information from the past to the present and learn from it. As an IS fellow, I am funded by the European Union Horizon 2020 Research and Innovation Program under the Marie Skodowska-Kiri Current Agreement and also Arhus University Research Foundation. And I'm really grateful for this support because when it comes to this really risky and not typical research, it is this kind of funding that allows us to basically explore the new borders of science. Thank you very much.