 I'm Scott Ortman. I'm with the University of Colorado in Boulder and what I wanted to do with my time here is present sort of a thought experiment about one possible way of expanding the contemporary relevance of archaeology. It's something that a lot of us are talking about and I have some thoughts about how we might go about doing it and so although there's a study here I think it's probably better thought about for this audience as a stylized example perhaps of how one way we might go about doing this. So the first thing I want to mention is that you know contemporary archaeology is already relevant in all sorts of wonderful ways whether it's supporting marginalized groups by getting the facts of history right, by helping people to imagine alternative worlds, to inspiring people. I mean archaeology already does all of that so I'm not trying to say that we shouldn't keep doing all of those things but what I want us to think about is what would an archaeology that contributes directly to discussions around contemporary issues look like. And from that point of view I think you know one of the things I think we could do is just think about the archaeological record a little differently than we have. I think a traditional approach in archaeology has been to say our goal is to reconstruct the past and so when we do our work you know what we're thinking about is the way it was back then. I think most people who aren't archaeologists if you were wanting to say the things we learned about the past are relevant to today the first thing they would say is but the past is different. We know it was different and there's a million ways in which the past was different than the present. So why should I believe that what an archaeologist is telling me should have any impact on my view of what's going on now? So I think the shift we have to make is from thinking about the archaeological record as a record of the past to thinking about as a record of as a compendium of human behavior and as a source for investigating social phenomena regardless of when or where they occurred. And if you start thinking about it that way and let those thoughts roll downstream I think what you realize is that there are some kinds of problems that people that we confront today that are much more easily addressed by other means than archaeology. So there's lots and there's lots of things we would like to know about the past that are hard to learn about the past for which the results are more difficult to link to the present whereas there are some things we can learn about in the past that are more are easier to answer and perhaps more directly relevant to the present and that we need to just accept that and focus on those sorts of things. There are some problems that archaeological data are well suited for and some that they're not. You know one of the other things I think we have to think about is what archaeological data really come into their own at the big scale at long time frames along the long time horizons. This is something ethnography can't do this is something an opinion survey can't do and so forth. This is where our data really shine and where they stand out from other kinds of evidence that we have about social phenomena. And of course the way most of the data that's being generated to address these sorts of things today is coming out of cultural heritage management. So figuring out how to leverage the massive amount of data coming out of that process I think is critical. One of the things we have to do is continue to work on developing the linking arguments that connect the things that are easy to observe archaeologically to the dimensions of social phenomena that people care about today. We have to turn them into proxies for the things we want to know more about and contribute to and my example here will be focusing on that. And then finally I think we need to again in a way be more specific reconstructing the past is too general. What we need to do is say what is a specific social phenomenon that we want to learn about and what does the archaeological data have to say about it. You know these need to be specific problems that matter beyond our field and they'll often be problems that people in other fields are grappling with not the traditional problems that archaeologists have grappled with. So the example I'm going to talk about here is a very simple one. It's the accumulation of pot broken pottery or the consumption of pottery in archaeological sites as a potential economic indicator. I think we all know as archaeologists you know one thing there's lots of in the archaeological record is pottery. It's pretty systematically preserved. Of course there's ways in which you know it's not 100% preserved but it's preserved pretty well compared to most things. There's a tradition of research on pottery accumulations that suggests it's connected in a pretty deep way to demographic signals. So what I'm going to explore here is the possibility that pottery consumption also tracks economic development and can be used as a proxy for thinking about social economic development. And I'm going to use two just terrific data sources that I've come across that have been are available through the Archaeology Data Service in in Britain. One of them is the results of the Rural Settlement of Rome in Britain project. So this is a project involving 3,800 excavations at non-urban sites in the UK since 1990. For every one of those excavations there's a database that tells us the area of excavation, the number of structures encountered, the occupation span of the remains, the total pottery found in that area, and also the total number of coins found in those excavations with the coins subdivided by rain according to the scheme created by Richard Rees who's a numismatist that works in the area. And the second big data set I'll look at is a current compilation of chance finds of Roman coins in the portable antiquity scheme that you mentioned earlier. Many of you probably already know the links to the sources but they're there at the bottom. This is a map of the database for the Rural Settlement of Roman Britain project. So from this database there's information for about 2,200 excavated sites if you put it all together. And of course there's a jillion reasons why the specific measures that we have for individual excavations would not be a precise estimate of the properties of that specific settlement if we could excavate each one systematically using systematic random samples and so forth. So most of the arguments here are going to be based on the idea that we're not looking for the properties of individual settlements, we're looking for average properties of all of the settlements and major ways in which they vary. So there's a couple of measures I've plotted here. The kilograms of pottery per year of occupation per hectare of excavation is in green. The number of coins per year of occupation per hectare is red and the number of structures per hectare is in blue. And I've log transformed all of these before making the plots because as a way of saying well let's say that these simple measures are really reflections of socioeconomic rates. Well one of the first properties of most social and economic rates in the contemporary world is they have a log normal distribution. Actually you think about a wealth distribution for example it's always log normal. So that what that means is if you take the logarithm of the measure and then plot the distribution it should look Gaussian. And so here's the distributions here from these databases and all three of them they're pretty good. They look pretty pretty like you know Gaussian. So even though I'm sure there's incredible lots of unknown and unquantifiable error in the individual data points in these distributions you know the logic of large data synthesis is that all of those what ifs sort of balance each other out and that the signal still comes through. And so that's this let's say is a little bit of evidence that that's that that is perhaps the case. A second way of thinking about it is in the contemporary world social and economic rates tend to be faster in larger settlements than in smaller ones. So a next check would be to say well how do these measures what is you know if these are rates per year of occupation how do they change with site size. So in the database each of these sites has been categorized by area here's four categories of the sites that are the data that are available. So you can see that all three of them they get the rates go up. So the densities of the settlements go up the rate at which pottery was consumed and the rate at which coins were lost all increase with settlement size. There's a third way of investigating this stuff that connects to the the ideas that Jack was talking about earlier. These are ideas that are involved in something we call the social reactors project that Jack and I are both a part of. This is the typical pattern of urban indicators today. On the left you see measures of social infrastructure in contemporary cities. If cities needed infrastructure at the same at a needed to add infrastructure at consistent rates as the population of the city grew you would see all these dots line up along the black line but in fact they line up along the red line which has a shallower slope than one. So what it means is that larger cities get by with less street per person. There's an economy of scale there. On the on the right you have a measure of outputs GDP versus population. Again a GDP group apportionately to people the data would line up along the black line but they actually line up along the red. What shows you is that as cities get bigger the people in them produce more per person. There's increasing returns to scale in the way contemporary cities function. Empirically the ranges of the slopes of these relationships are shown at the bottom of the slide here and there's we have a theoretical framework that you know suggests why that is the case that we're investigating here. So what we're going to be looking at now is the degree to which the data from Roman Britain follow the same pattern and they do. On the left is the relationship between the number of structures or the found in an area. So structures over area versus the area is the structure density. What you see is that the densities of Romano-British sites increase with population following the same slope as contemporary cities do. And on the right we have two different measures of outputs the rate at which pottery was consumed and the rate at which coins were lost. They both show increasing returns pretty darn close to the rates that you see in contemporary cities. So far so good. There's actually you know the possibility that some of these simple measures might actually be useful for this. But there's a little bit of a wrinkle here and I know I'm trying to walk through this quickly. This is what's going on in contemporary cities. This is data from the People's Republic of China. Each city is a dot. There's 20 years worth of data here. There's a fit line for each year and the yellow dots are the centers or the mean coordinates of the data for all the cities in China for each year. So if you look at what happens year on year and year out is that the average sizes of cities is increasing a little bit. But the end the slope of the relationship between a city size and GDP is the same. You can see the slopes of those fit lines are basically the same. But the height of the fit of the relationship is changing. It's growing up. What it means is that the baseline rate of production in China is expanding per person. That's what's typically what you see in contemporary data. In archaeological data so far we do not see that. This is data from the US Southwest to give an example. It's the population of Pueblos in the Southwest versus the pottery consumption rate that we have for them. And I have the centers there are the yellow dots again. And what you see is that for a variety for time periods in a time series the centers just meander along the same fit line. The fit line is not growing or increasing as it is in the modern world. And the same data are represented as a time series on the right. So you have the average populations of cities growing. You have the average consumption rates of cities growing. But the black line at the bottom shows you the intercept of the fit line for each time period. You can see it's just meandering about a fixed value. So what's going on here? Is it because we don't have good data? Is it because the methods we use to generate these proxies are washing out the possibility of what we might call intensive growth? Is it because our measures aren't sensitive enough? Is it because pre-industrial economies didn't do that? Perhaps economies of the past didn't generate this kind of process that we see in the modern world. Well I would say so if there was ever an ancient society that experienced intensive growth it was probably Roman society. So let's see if there's evidence of this in the data for Roman Britain. Well the answer is complicated. The pottery measure, this is a time series where what you see is the intercept of the scaling relation. So the intercept of the fit line putting all these data together. The series for pottery in red meanders about just like it does in other cases. But yet the intercept for coin loss shows consistent growth over the century, century on century. So what's going on here? It seems as though the baseline rate of coin loss is growing even though the baseline rate of coin or sorry pottery consumption is not. Well you know folks you know I'm no expert on Roman coins so you know I'm looking forward to talking to more experts about this but I do know from reading that inflation was an issue in the Roman Empire and that the values of coins got less over time. The purchasing power you know at least by their face value. So one thing to investigate is the role inflation in that pre-factor there perhaps the economy didn't actually change. Perhaps it was static it's just that people had to you know had coins that were less valuable and so there was more small change because it was worth less. So that's where the portable antiquity scheme comes into play. What I tried to do in this slide is in a sense convert the coins found in the PAS to the silver standard. So what I did was I used an index from Numismatis about the silver content of coins over time and combined that with the you know the face values and the conversions from different denominations of coins and the end result of that process at the bottom here is the graph to look at. So the silver line is the the silver the grams of silver in the average coin the blue line is the average face value and the point here is that even after you control for the reduction in the silver content you still see that the in a sense the silver standard real value of coins seems to have increased a little bit. So at least by this way of thinking about it it seems like we could say that perhaps the Roman economy did experience this intensive growth over time but at rates of maybe a fraction of a percent a year up to maybe two percent when it when the area was first brought into the empire but it means that our pottery consumption index that I was sort of hoping would work doesn't show the intensive growth it only shows the agglomeration effects. So it's a partly disappointing answer but it seems to be the reality in this of the data here. Perhaps what's going on here is as the Roman economy expanded people shifted from pottery containers to perhaps containers made of other materials like metals and other things like that so perhaps pottery consumption in this environment in this context just doesn't show the intensive growth but it does show agglomeration effects. So this is sort of a progress report of ongoing research and thought processes that we're going through trying to think about ways of developing arguments that connect these common observables to properties that we care about today. Welcome your feedback on that and here's some links to some of the websites of the groups involved in doing this work. Thank you. Thanks.