 First of all, since we're back to the second part, I'll quickly pause in case anybody had some desperately urgent thing they thought of over the tea break and wanted to ask about what we've discussed already. Total silence, okay? I assume not. Oh, Seth, that's something. It might be a really stupid question, and I was just trying to follow the whole thing you were saying about, I mean, the quantification of coins is kind of interesting. I imagine it would be an easy thing to do, but actually it seems way more difficult than counting pottery. But the thing that you were trying, using the counting the number of ddies per year, I can see that that gets you quite close to the information that you want. But then there's the next step beyond that, which I don't think you did show on the table, which is you've got the number of ddies per year. But then there's another inference that we've made about how many coins were struck from each dye. And you said that there is some minor variation, but like how big, I would have thought there's probably quite big variation there, isn't there? Which could then also... That's, I think, there's a big debate a few years ago in Numismatics, where there was one school saying, this tells you nothing because a dye can break after five strikes, or it can go on to strike 20 or 30,000. And the other people were saying, well, on average, we know from written records from many of Europe, how many dyes were struck. So it's, it is a... You'd want to quantify that as well, wouldn't you? In that table you'd want to know what could be somewhere between 5,000 coins or 10,000 coins, or is it... Actually, partly as a result of that debate, we want very hard not to quantify that. What happened is, is that Numismatics has no confidence in its ability to tell you how many coins a dye struck. But it has quite a bit of confidence that given similar procedures, metals, shape, size, etc. of coin, that on average the dyes made about the same number of coins. So what's happened is, the closest to a consensus opinion that Numismatics has developed is essentially a cautious one to say, we can tell you that they were making two or three times as many coins in Samartata as they were in Sind, but we don't want to say how many coins that was. So we can give relative values, but not absolute. One of the caveats... So if you give the qualification in, because you put in the table just the figures of how many dyes, but would you also try and put in there what kind of metal they're... Yes, so in that particular case, all of those coins are of the same weight, they're all gold of about a 50% purity, they all have about the same relief, they're all derived from similar procedures that existed. So across those three coinagers, we feel quite confident. I'd feel even more confident if we were looking at the Silver Central Provinces types, where we think that it's literally the same institution making the coins all the way through. But for example, if I had a dye study that told me how many... We mentioned Vishnu Kundins, who make these copper coins with quite high relief on them. If I had a dye study for them, and I had a dye study for the silver coins of one of the Alcon rulers in the north, which are relatively low relief, quite flat space ones, I wouldn't want to compare those at all. I wouldn't have any confidence that their relative number of dyes told me anything about their relative production. So we're only able to do this in a relatively narrow way when we can say that these things are very similar in their particular distribution. And that's partly a result of this debate, essentially a consensus, very cautious opinion, that we can establish relative values for things that are similar, but we can't establish absolute values and therefore can't compare dissimilar things, which does cause us a problem, which I'll come to in a bit, which is of course sometimes we're comparing against monetary instruments that are not made by dye-striking. And then we have a problem of comparison at all. So for example, cowrie shells at a later stage are absolutely definitely testified as a currency. Were they a currency in the 6th century? Since I can't tell you how many coins we're actually talking about being made, I can't compare them with cowrie shells. If a coin is cast rather than struck, which is the case for coinage in Orissa at this period, and is the case for coinage some coins in Sri Lanka, I think, at this period, then I can't compare the dye counts against any information I have for those, because again the dye counts are providing only a relative value against other dye counts. The caution numismatics exhibits on this is a limitation on our ability to do the counting and measurement. It's quite interesting when you compare the dye counts with real-world counts of, like if you've got an area where there are lots of coin hordes where you have actually got, you can count numbers of coins. We have comparisons of this sort that can be made. To throw the information out there for you to make what you will of it yourself, we have an ancient set of accounts from the 4th century BC for a group of coins we also have a dye study for, and they would give a figure somewhere in the region of 15,000 per obverse dye. It's Greek coins though. We do as Joe pointed out, have a set of medieval mint records which we can match against dye numbers, and they give averages between about 2,000 and 3,000 per dye and about 80,000 per dye, with again the range broadly falling somewhere in the 10,000 to 30,000 area. We have some experimental archaeology in that largely bears out a 10,000 to 20,000 number. Essentially we think it might be possible to do that, but personally I think that numismatics is not far enough forward with that sort of study to have any confidence in its results. Whereas these kinds of purely relative ones we make of, these look very similar in all the respects we can reasonably measure, therefore their dye counts are probably indicative of their relative production. I feel a certain amount of confidence in, but I wouldn't want to go to actual numbers of coins, and there's a historiographic reason for that as well as a genuine theoretical concern. I am probably more cautious because of this debate that took place than it was vicious. Like all fields you are cautious about things that have been heavily criticised, even if you're not necessarily justified in your caution, but that's why I didn't go there. It's a good question. Talking about other things numismatics do, I said that we've therefore got a measurement of how much is made, but how much is made is not how much is in circulation. You don't know whether or not it's being used. We had that envy for the money supply. If the coins get given out to Royal Duchess and then they're buried in the ground, and they're not exchanged, they're not actually having any effect on the money supply. So Sharman might be right, they might make a lot just as many coins, but they might not use them in the period, in the way that they did in the earlier periods and therefore his assertion that the amount of coinage being used has gone down, which of course remember he came out having looked at other sources, so he's not drawing that from the coins themselves, might be valid. I'll just very quickly, numismatists can use techniques based on hordes to tell you whether or not the coins in the hordes have changed hands. These are graphs showing how the weight of coins alters after they've been initially made based on different factors, and I'm going to skip over them beyond telling you that, and this is an example of distributions, no, this is an example of distributions in a particular horde. My point is we have ways of measuring the velocity and circulation in numismatics. If anybody wants to at the end I can come back and explain a bit more about that, but believe me for the moment they're very dull, but they are effective. On to the next one, so this is what I did want to say a little bit about, which is this is an example of a Gandhara copper piece, a long anonymous series issued in the region of a site called Kashmir-Smast in the northwest in modern day Pakistan. It's on the Pakistan side of the border isn't it? Oh yeah, it's the northern Gandhara. Yeah, northern Gandhara. So we have here an assemblage from the site. Now it's mostly been acquired through trade and people revealing collections they've made from the site, but it's a single site. We have a large publication with close to a thousand coins in it, so a good sample. There is an argument has been put forward by one Roman specialist, which essentially says when it comes to base metal coinage, your dye studies are pretty much worse thus because they tell you how much was made but not how much people were using. And he says much more useful is what people drop, because when they exchange, use coins to change hands, they lose some of them. So the amount you find in site assemblages is a better measure of coin use than the amount that's actually produced. Name of the Roman specialist I've forgotten, so somebody will just have to ask him. No, it's... It will come back to me or somebody can ask me at the end and I can look through my references. Is he willing for the pub quiz? Yeah. So what I did was I dug out one of two sites where we have this kind of information, which is cashmere smashed, and then I ordered the information in two different ways. The charts on the top left here is the information Ashama ordered in, so by its class, by the type of coin. And the gigantic spike in the middle is the anonymous Gandaran copper. Shailendra's point that the anonymity, the inability to attribute it changes the way you look at it. The chart below distributes the coins over their period of production. So what it does is if the coins are issued for a hundred-year period, and there are 200 of them, then it's assigned two coins to each year, evenly distributing them over the period of production, and then breaks those down into 50-year baskets to give the graph at the bottom. And you can see cashmere smashed on an escalating scale up to the sixth century. The ones marked in red at the back, cashmere smashed is definitely not active as a site before 200 AD. So the coins in red are old coins still in circulation. It's probably not active before 300 AD. So the coins in the reddish blue, the purple, are probably coins that were still in circulation. Now I'm raising that because it raises one of the other things about the amount of money that's available. A significant proportion of the money that's available is not money produced at the time. It's old money that's still hanging around. Now this is true of the coins in your pocket, where if you look, most of the coins are going to be a couple of years old. Some of them might be 10 or 15 years old. It's even more true of ancient coins. The bulk of coins that you're using, that you have in a horde, are probably 20, 30, 50 years. The best example we have are the western satraps, where the coins are individually dated. And the age profiles there are... We have four to each other, and one to each other, and western satraps. Yeah, so centuries between the two groups of coins. And that's quite important. It's important for two reasons. One, it gives an important caveat on the dice studies. Two, it's important for Sharma's argument. A sudden dramatic and short gap in the production of coinage need not have had any significant effect on the amount of coinage that was actually available. If a particular dynasty collapses and is a result of mint stock producing coins, there's probably a lot of coinage still around. People are probably used to using old coinage. There's no particular reason to assume that the amount of coinage available to them will drop significantly unless the production remains dormant for a relatively long period. So, for example, at the end of the Gupta period, the reason the Sametata coins, the very first of them, look like Cushan coins. And they're made at the end of the Gupta Empire or after its collapse. So, the coins are being made based on a prototype that's 200 or 300 years old at the time, which implies that that prototype is still familiar. So, next one. We can also count in a slightly different way. Hord's give us this opportunity. The only place in the sixth century we can do this is West India, which Shailendra alluded to earlier. This is a posthumous issue of Kamaragukta. So, it's a Kamaragukta coin, but Kamaragukta was dead when this was made. And these carry on being made a significant period after his death in West India. There are Scandagupters from West India, but no Budagupters. So, just a little interjection is that you're going to ask, why this is so and how do we know that these were made posthumously? Because the inscription is pretty clear. It reads well, there is no sort of corruption. There have been another set of methodologies where people have mapped the typological degradation, the way the constant copying of the designs leads itself to degradation in terms of the detail, how that sort of detail then maps onto the debasement in the silver content. And then you sort of map those two things, we get an approximation of how long after the issue, the first issue is coins. And the giveaway feature on this particular one is that his collar around his neck is made of dots, which isn't on his lifetime issues. And that's a feature it shares in common with the Mitrakussen, which places it into the same period in the 6th century. So, there are Kamaragukta, Scandagukta and Mitrakka, all three coins circulating in parallel. And possibly some others. So, again, there's this large silver coinage of the western satraps, these small silver coins. This is very well known, there's a tight catalogue of it, it's well studied. There are then lots of successor coinages. You could take the approach that was taken by Dale and recall the number of hordes in which these things appear. I was able to pull together about 20 or 30 hordes relatively quickly. But what I did was take an alternative way of counting it. I took the latest coin in each horde, assumed that the horde was deposited at the time the latest coin was issued or close to it. And therefore all of the coins present were in circulation at that point in time. And therefore came up with a chart that shows the rate of deposition. How many coins per year are being lost rather than how many are being made. And what you'll notice is there's a massive peak between 350 and 380 AD. Shalindra has written a whole article on this, so he can explain to you what the important political event that was happening at this period was. 350 to 380 AD, the western satra hordes. 350 to 380 AD. These are the very late western satra hordes. Yes, and well there is, the guptas are coming in around that time. So they are threatening western shatraps and then around early 400s the ompwist actually happens. But there is a sort of a window period that you see from coins that the guptas are pushing in and the satraps are pushing back and then ultimately they go. And that's around early for the first decade of the fifth century. So that's what is happening. Some hordes created by a crisis. Yes. Peak basically means that people in western-satra territory are burying their money. Because the more money guptas are coming. I mean there is political uncertainty of various kinds. So I mean that is, there is bigger transition, you know, so anxiety obviously. A hordd yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch yn ysgolwch. And the important point of course is that what the child also then shows you is that losses are still extremely significant in the sixth century. That actually compared to the period pre third fourth century, the coins that are being deposited are in very large numbers. We have to be careful about this as Shalindra says, political uncertainty that either the possibility that your coin might not be worth anything in a few years or that somebody might seize it. Or of course the possibility that you bury your coins in a routine way but are unable to return and collect them can create hoarding. But also these probably also just reflect that coinage is there in significant numbers. And since, you know, the best that I'm offering is an opinion on this, I think Shalindra and I broadly agree that actually there is a cross regional trend for South Asia and it's that the amount of money and coinage and circulation just keeps going up. I think that the country that South Asia just becomes gradually more and more monetised, its monetary economy becomes gradually more sophisticated as monetisation spreads. And any retractions are extremely local events. West India seems to be the best case for that. So there's a very clear path of most of the evidence when you look at it in isolation in West India seems to indicate ever increasing amounts of money and direct contravention to the evidence deployed by Shalindra and others. So again we've got, so what we've done is we've looked at all those different ways of looking at it and the very last one of these by the very last step is you've established how much is in circulation either because you've done dye studies and then applied all sorts of modelling to it or because you've found good assemblages or hoard data you can work from. And therefore you've established something about the money. The opposite, therefore more coins were in circulation, therefore more coins were in circulation, therefore there was more money. Yes. So it's exactly upside down. Yeah. But here's the problem, commodity money, not everything that's money is a coin. There are other forms of money in operation. This is, we had a whole talk on the Bactrian documents. Here's a Bactrian document from the year 2056. So this is in the 6th century, right in the period we've done. The loan, it's a loan agreement. Tet is borrowing from Muzd Pasbarun. Now here's the critical thing. He's borrowing some grain. This is presumably because he hasn't brought his grain in yet and he needs to cook some bread. This is probably a food. It's probably facilitated by various social relationships. But it's monetised without having any coins involved. Because along with the grain he borrows some wine, 21 jars of wine. And he has to return the grain in the contract, same amount of grain, but he has to return 24 jars of wine. The three extra jars of wine is his interest. So the wine isn't being drunk. It isn't that he drinks 21 at a party and promises that he'll supply 24 later. It's that it would be silly to borrow 10 bushels of grain and then give 12 bushels of grain back because the other person doesn't want 12 bushels of grain, but everybody wants wine. So he gives him the grain and he gives him some wine with it. And then later he has to give more wine back. The wine becomes a way of attaching a numerical value to the loan. We know this happens with cowardly shells. We know it happens with silk. We know it happens with other forms of commodity that the commodities are turned into money. A numismatist referred to this as commodity money, standardised units of a commodity that can be used in similar ways to the ways we use coins, but are not coins. This is just a commodity's contract. It's like a port bell. The futures contract is created that way rather than money. You can, yes. It's the only example, so it's a bit hard to be confident about what it says. You're probably missing something here, but you're saying that he doesn't want to drink the wine. If he doesn't, why does he need to borrow 21 jars of wine? Why don't he just return the grain plus three jars of wine? If we assume that what's going on is a commodity money transaction, it's very likely he never collects the wine. Why did he mention 21 jars of wine? Because it gives you a one-seventh rate of interest. 21 goes to 24. What you're doing is you're turning a social exchange into a way of measuring money. We don't know for certain that's what's happening here, but we have other more complicated situations in places like Egypt or the Taran Basin, where much larger amounts of documentation survive and we can trace people doing this kind of thing. We know that things like this can operate this way. I give you some service and I lend you along with it two rolls of silk. You promised to return with three rolls of silk. At the end, I've created a loan contract using the silk like I would use money. It's not immediately clear that the silk actually changes hands, but it's a way of attaching value to something. We don't have a deep insight into the details of what the person is doing, because there's only two bacheloring documents where people do this. Normally, they do everything from coin. They've written the document in a very similar way to the way they would write it with coins, but they've done this funny transaction involved handing over the wine and the grain and getting wine back. Normally, everything's in coin. The chart next shows the coin values and even where the document doesn't contain the exact number of coins, we can see how much somebody's valued it by, because the penalties have a fixed relation with the value of the contract. Even where things are exchanged in the kind, the economy thinks of them as having monetary values. Unfortunately, the bacheloring documents are the only set of sort of minorial records that survive from central Asia in this period. Presumably, every small kingdom and principality maintain an equivalent set of records, and some of them might use them to facilitate monetary exchanges, creating an effect, sort of money's account that operate alongside actual physical money. We know this happens in the Taran Basin, that there's a central pot of money, and sometimes it's given out, but a lot of time people just have an amount of money in their account, and the amount is transferred between accounts. So it's, again, it's potentially increasing the money supply and the velocity of circulation without any actual coins changing hands. Which brings us to how do we measure any of that and the extent of monetisation or market activity if we don't have, if we can't use the coins, has to measure. And Gethin was actually got, do you want to come down to the front to this? Gethin actually discussed this with me last night, and he'd come across a Roman example of somebody doing this archaeologically. I think we're all very impressed with Robins on a meticulous approach and his deep knowledge of this data set. Now what I'm going to talk about now exhibits none of that. We, thinking about synergy, I think this is probably about synergy, and Nathan was quite keen for us to exhibit some synergy today here, and I think in order to do that, you have to get about your comfort zone. So with that in mind, I'm going to talk about economics, the Roman Empire, and computational modelling. That's basically on the basis of a paper I was reading that I just mentioned, I mentioned to Robert yesterday when we were at a research associate's dinner and he was interested in it, so I'm just going to present a little bit about that paper. Now, Roberts used a few macroeconomic concepts, and two that are kind of underlying what some of what he talked about are rationality and knowledge. Now, agents in economies do not always have full knowledge, so when he's talking about price or the amount of money in circulation in the economy, we don't know how many coins might be in circulation in the economy then, but at that time, they wouldn't have known that either. So when we're applying modern macroeconomic principles to ancient economies, we have to be quite careful. So I'm going to talk basically what I'm talking about is an article by Tom Brumans, who's a friend of mine, and Jerome Pobleau, and it's called Roman Bizarre or Market Economy, exploiting tableware distributions through computational modelling. So they talk about two theories of the Roman economy. The first is Peter Bang's Roman Bizarre. In this theory, the integration of markets around the empire was weak, meaning traders had a poor knowledge of prices and availability of goods. To quote Peter Bang, such a scenario is distinguished by high uncertainty of information and the relative unpredictability of supply and demand. This makes the prices of commodities in the Bizarre fairly volatile. As a consequence, the integration of markets is often low and fragile. It is simply difficult for traders to obtain sufficiently reliable and stable information on which effectively to respond to developments in other markets. Considerable fragmentation of markets prevails. Now, in this model, the results were market integration. Merchants would aim to benefit from opportunism and speculation in individual markets. And secondly, the social network of personal trusted relations and strong communal ties was maintained between markets. So, unpredictable supply and demand meant that prices were volatile in Peter Bang's theory, and markets functioned very differently to present day, where well informed specialist trade facilitated by extensive and efficient networks. Now, there might be some similarity between this and, yeah, so there's the article. I must say, everything I'm talking about basically is coming from this article. So, this is not by any means why we research it. So markets functioned very differently to the present day, where well informed specialist trade facilitates, is facilitated by extensive networks. Now, the second theory is Peter Tenon's Roman market economy. In this theory, markets were strongly integrated. A quote from him, the economy of the early Roman empire was primarily a market economy. The parts of this economy were created far from each other, were not tied together as tight-tight as markets are often today, but they still function as part of a comprehensive Mediterranean market. So in this model, commercial information in one market was easily accessible in the others and knowledge was widespread. So those are the two theories. Now, the article takes up a social network perspective to examine this and they're suggesting that integration of markets is high if the potential to share commercial information and goods directed between markets is high and low if it is limited. The degree of market integration can therefore be represented as the proportion of all possible ways that connect traders on different sites. So, yeah, the bizarre argument there's fewer connections on this side with weak market integration and on this side, we have a market economy with greater number of connections. So, in the article, they invoke a computational model where the Roman east is considered to have functioned as a complex system where the small-scale actions and interactions of agents with only limited access to information gave rise to large-scale patterns. Now, I'm not going to get into the details of this model partly because I don't fully understand it myself, but it's an interesting approach that's being used in a lot of different archaeological projects at the moment. Now, a pair of traders to give a couple of points about how the model functions, a pair of traders connected in the social network are able to share commercial information, supply, demand and price estimates and to trade table wires, the model models, the movement of Teresidioata. Now, production centres on the... in a few places in a few locations in the economy and based on the knowledge of prices, supply and demand available to the traders, they can choose to buy, sell, store or store goods and pass them to other... or with other traders or local consumers and the model therefore gives rise to different patterns of distribution for the four table wires based on the two variables that I mentioned, the number of links between traders on different sites and the amount of information shared between connected traders and so those two variables are the two key aspects and they also underpin these two models and so they can be explored. Now, Brumans and probably apply their model to the Teresidioata distribution in eastern Mediterranean. Now, it's a large area and a well-defined database of shards with clear patterns in the overall distribution so they have a large proportion of this type called ESA and it becomes more even between four different types as time passes. So, their conclusion based on the model is that the high integration of markets have the potential to give rise to the archaeological observed differences and the structure of the social network within individual markets is less important than a high degree of integration between markets and the potential for one production centre to produce more than others. So, of course they give a lot of caveats but they end up suggesting that this model is more likely for the distribution of the Teresidioata eastern Mediterranean. So, I would say that this has a couple of implications for Robert's presentation. The first is the possibility of applying a formal model like the one used to the coin distributions. Now, this we're a long way from that, I would say. We require a pattern in the coinage to be well-defined and then because it has to be tested given different variables. Models generated could then be compared but the specifications of models need to be explored using greater amounts of data and the models themselves need to be tailored to certain data so that they can be falsified or not. That's a big problem with models like Charmas as well. But it can give us an idea of where we can further work to carry out further research that could be into factors that enabled and structured and maintained communication between faraway communities. So, we can look outside the coins to look at the extent to which markets may be integrated and other forms of data. The second implication is that the macroeconomic concepts that Robert was talking about in the first part of his presentation are highly dependent on an understanding of the market and we also saw that when he was talking about the commodity because those can completely slew the picture as can prices and market integration. So, Robert has mentioned that production and deposition of coins can be examined and highlighted the problem I mean, you talked in your paper anyway that he's circling you talked about the production and the deposition but it's this exchange in the middle that we don't understand so well and perhaps broader archaeological evidence can help fill in gaps here. Two examples that could be what that would be might be the distribution of red polished wear around South Asia. So, I thought that it was interesting what gethnic race with me at dinner and I wanted to just kind of insert it there because it made clear that there are alternative ways to measure how economies are working. It's not solely a function of coins just like the coins themselves are not solely. So, I suppose I've only thought on how easy it would be to count and categorise pottery in South Asia in a way that was comparable with those kinds of models of the Roman Empire. There'd be even more the approach is quite innovative anyway isn't it? I just think within obviously the data set that's available for the eastern Mediterranean is so much more complete than probably anywhere else in the world isn't it in terms of the volume of archaeological data the knowledge of the pottery, the dating the precision of dating of Teresa Jalata all those things mean that you have actually got a huge body of data to play around with and apply statistical models to by India it would be really hard wouldn't it because we don't even have the basic chronology of classification of the the pottery to start with I think probably most areas of India I've imagined a lot more about it than me but we're dealing with four, five hundred year time periods that the material can be slotted into so there's huge huge amounts of work that just needs to be done with very basic stuff of doing large scale excavations of going through big assemblages of categorising things drawing things getting radiocarbon dating and doing the basic sequencing before you can do anything in terms of fancy computer models but in some ways you're violating the first slide in your presentation which was you weren't going to get to new theory you weren't going to get to an explanation of feudalism and you weren't going to get to well the number of things but Gettin's paper has opened those questions because what you're the pottery distribution is going to show you is something that the coins can never show you the first thing is people don't go around picking up pottery they go around picking up coins you know and like this great horde of pottery shirts never sort of comes on the market and it's just a map and it's just rubbish lying out there in the fields so it is possible to do a survey and collection in a systematic fashion not in an idiosyncratic fashion but in a systematic fashion across the landscape on these sites and that seems to me that coupled with what you can do in anismatics it's got to be sort of tied with the ceramics because the ceramics the ceramics are a proxy for people you don't have a broken bit of pottery there by magic you have it because somebody dropped it apart so the typologies and the quantities of ceramics are a proxy for human activity and scale of activity scale of settlement and from there you can move into what they are but pots mean lots of different things at the same time some pots are used for cooking some pots are used as commodities that are bought and sold some pots are high status I mean these are very basic categorisation of ceramic sandwiches but we're a long way from being able to identify what certain pots were used for there's a lot more of them around than there are coins yes, absolutely absolutely, but we know less about them than we do coins yes, they have a heavy attention that they deserve because they're even less attractive than the ugly coins there was in Sharma's model of not just feudalism but also the urbanisation there was a connection between the availability and the lack of certain kinds of pottery and he uses as much as he uses paucity coinages, bolster he also uses pottery and Derek Kennett had began to assemble data to demolish it as it was saying again the same thing has happened with coins there are lots of methodological problems I mean I'm not a ceramicist so I don't want exactly the methodological problems that he was referring to but he said that there are particular I think the pottery the characteristic pottery that he was associated with is sort of red-slipedware red polycystic that's right and the kind of occurrence and disappearance of RPW was deployed in the same model to say that there was the urbanisation but Derek's data was also a certain amount of settlement happened issues were discussed in the same paper by Derek I don't think he's published it he has and one of the reasons why collecting pottery is again for the problems is because the site settlement has sort of moved around and you can't really take where people are still living so all you have to base your data or get your data from is all the peripheral things and they're not really the sites that were inhabited in as intensively as they were so there is that kind of bias as well in this paper this junction is that when you have such a large amount of pottery taken to such a wide area that those biases then are less you're not looking at sites individually you're looking at a large complex where that's the case or not I actually doubtful I would say that one thing I was actually going to say when I was up there is that I think often the case for these digital tools or digital techniques they there's not enough account taken of the problems of the data itself and you know they could do with perhaps taking Robert's approach to really unpack them I think that's a really important point which is there is a difference between systematic and stochastic errors it doesn't matter how many collections you look at the pretty coins are still going to outnumber the ugly ones because the error is systematic it's built into the method but increasing the number of small silver dramas you have access to is going to remove random elements that occur because a particular hoarder is large or a particular hoarder is small and so is going to eliminate those kinds of errors so there are certain types of errors that are eliminated by going to large numbers and there are certain types that are not one of the problems we have in numismatics is that we face a very large number of errors that are systemic they're built into the analysis and so it can't be eliminated simply by increasing the sizes of our samples the assessment is good it's kind of built on what you're saying but also what Michael's point was interesting but the pottery is an indication of actual activity and use and you made the same point earlier didn't you about referring to this numismatist who'd argued that actually coin loss was a sort of more interesting indicator of coin use than rates of minting so it's the same argument isn't it and I suppose when you were talking about that systematic error that's built into coin collecting this focus on the pretty examples I've had exactly the same problem with ceramic studies for many years being dominated by other fields of art history and collecting pretty bits of glaze pottery but once we start to look at the whole assemblage that's when it becomes In India the thing is that as Charmaine was saying the interest that is that very seldom when you're looking at a big archaeological site the current settlement is usually on the highest ground for all these reasons don't get washed away in the rain and so people are living up there so you don't excavate there and then this is kind of compounded the problems compounded by the way archaeological practice goes on in India so the government of India is old and the historians have to buy some land here to have an archaeological reserve and it's always on the side because the people are up on the top so the archaeological reserve is on the side of the big settlement mountain and so if there's money to do an excavation they excavate on the side and as you've said it's not representative of where people lived historically it's just that it's in the reserve or it's near the reserve and they can excavate there so that if all the data across all your archaeological excavation that's coming from the side of the settlement you're going to have a very skewed the thing, I'm not saying it's always happening but that sort of seems to be the way it's going and then the other thing about the other sociological and historiographical thing about it is that Sharma was as you've just read from the first slide he's got a really compelling style he writes really well and he's kind of humorous he's engaging and who follows him as the professor in Delhi the pre-eminent historical post in India is that sort of weakling Chatharpavili and so the result is that there's been no significant intellectual challenge really to his rather strong theories which on the face of it are extremely compelling and so added to by his personal charm everything else and so you end up with a kind of mindset of feudalism and decline and so that you know as Einstein said you can only see what the theory allows you to see you can't see anything otherwise so because there's going to be no city city sites out in the countryside they're not there they don't exist the theory doesn't allow you to see huge urban settlements in archaeological terms they've just vanished and some foreigner like me walks along and you kind of stumble on what says gigantic mountains it's never even mentioned it's a gigantic city site because it doesn't exist in the literature it's just not there because the theory that Sharma has propounded which he recognises to his credit is just the theory let's see how it works there cannot be cities because there's only villages so they don't exist what are the things of massive distortion this method of counting points and drawing lots of historical conclusions was subsequently taken up by lots of other people in and sort of drawn there is a whole a book called Imperial Monetary System and there are sort of essays long essays written about how certain mines function and what their outputs are because moogle points are dated so you know that they were issued in a certain date unlike these ones and then there was sort of drawing conclusions saying that because in between 1650 to 1655 the main gasura extant number of specimens 100 55 to 63,000 65 to whatever back again in that period the production went up so the GDP went up and there's this whole argument about what was India's GDP in moogle terms which is all based on this sort of methodology of counting coins which is I think perlicious it's not because the British Museum has so many numbers of rupees stuck at surat in so many years that doesn't mean that the surat meant was actually functioning better in those years as compared to other years so it's a really flawed method which was sort of I would say imagine and discharged by Shalman but then taken up by lots of his students and followers and then applied to different areas of economic history and I have to say maybe it's a way of conclusion you're perpetuating this because whenever I come back from India with a handful of Indian coins you always pick the ones that the museum doesn't have and so there's these little five rupee coins whatever they are and they end up in the collection and so we've got all the times because you're collecting them so someone will come along in 300 years and the British Museum said gosh the amount of coins has increased so there you are because museums are interested in typologies that's what the collection is for either to display them to show you here is an example of something that existed or to have a record of here is an example of something that existed and of course additional to your list of distorting this typology of what represents delivery time yes absolutely and of course what happens is if you have a period in time when for some reason you get fewer types because in mobilisation this is the practice of simply issuing the same type of coin over and over again or anonymity which means that of course the bits of the coin that you are usually collecting doesn't change when this happens you have fewer coins in the museum it's the natural money Joe wrote very interesting articles on the coin's tradition of India one of the things that he pulls out which is is that Indian coinage seems to show a higher degree of tendency towards immobilisation and anonymity in its production but actually that's not a universal feature of Indian coinage it's not evenly spread across the whole of the two and a half thousand years of Indian it's good a huge amount of it is concentrated right here in this period just that were not immobilised like the western satraps where the name and the date changes and they're being replaced by these coins that have comaragoptis and legend just repeated over and over again replaced by the coins the Sasanian coinage an import of Sasanian coinage into the north north of India in presumably about 500 AD then that coin design continued in use till about 30 in western India and some of them you can distinguish there are other small features out in which one names one to distinguish them but the majority of them all you can say is this isn't early and that's about it and so it's possible that that change in the way coinage is visualising itself has something important to say but one of its effects is to make it harder and less interesting to study it and to reduce the number of examples that you will find in collections so but immobilisation repeating designs is usually associated with the coins being used it's usually associated with a need for you to have confidence that a coin looks the same way coins have always looked more confidence in that than you're placing in the political authority is backing it so it's often a function of large production so the irony in a sense is that this is especially true of Gupta coins and Kushan coins there are moments in the Gupta and Kushan coinage where the coin design is incredibly variable so every museum needs 20 or 30 Chandragupta's minimum no matter how small your museum is you need a lot of Chandragupta coins you need a lot of Huvishka coins because they all have different designs you have to have one of each every private collector does but in fact Huvishka coins are incredibly small production there's not very many of them at all but almost every single one that's extant is published it's very it's very unlikely that there's more than a handful of Huvishka coins in the world that are not published in some form you can pull together a little over a thousand of them published but that's probably all the coins that's right I think Jason almost I'm in trouble everything everything you've just said makes perfect sense to me it's almost charmesque in it do not trust it I suppose I'm just talking back to what Gathin was saying it's at that level speaking as an archaeologist it's at precisely that level of okay we can see this going on in the coins therefore that means this must be in the case it's at that level that I start to question and start to wonder to what extent are we applying wider theory macroeconomic theory borrowing from proven examples where the nemismatic data has been matched with other archaeological data from other contexts and other parts of the world at other periods or are there truisms that really can be applied across the board I just sort of wonder what would happen at what point what do you need as a nemismatist for there to be archaeologically, texturally, historical to base your chains of inference on surrounding data as opposed to on the sort of starting point in nemismatics is that coinage is an official device largely an official device and it follows particular patterns of behaviour in order for it to work and therefore one can make some deductions about it when you find that Gory Hymns redoing re-examling on the first Crescian kin, there's a huge increase in the number of coins cutting off some might also pick in the same colours and quarters and so one can look at what economic phenomenon might induce that to happen and we have other parallels in elsewhere so one can draw some conclusions about what's happening to the monetary system about time so because coins are money and money is part of so money works in the same way but it doesn't work in the same way but there are patterns of behaviours that one can see so when you find one coin in excavation so when you find a whole rather than coins then you can draw some conclusions about the monetary system but not the monetary system about the coin system which isn't always the same as the monetary system I think I think it does but I in terms of what we would need just to make absolutely clear I would say that for each of the 20 or so series of coins that I kind of looked at in the survey which you saw I would say you would need a solid type catalogue syllogies of most of the major collections a die study 5 or 6 archaeological sites with a large number of coins from well stratified contexts and in the region of 50 to 100 hordes and then given that data the system is reconstructable we've got that and we've got that and in South Asia before 1100 we have that for punch mark coins and early cushion and that's it what I'm also getting at then is and this is an honest question I remember that is it that you're only interested in making conclusions about the monetary economy the flip side to which is is it that you absolutely have to have those type series and all of that quantification to make any conclusions is it that you need that because you don't have because there is an absence of information about everything else that was going on without the classifications typologies and contextualising information that sites and hordes provide the coins aren't really telling the coins are part of anything the coins are part of systems they are part of the system that hammered them at the mint they're part of the political system in fact their authority they're part of the economic system that moved them they're part of ritual and social systems that resulted in their deposit they are affected by military systems like the example of the western satraps they can tell you something about any of those systems but because they only tell you something because they're trying to tell you something about the system they can only tell you that when the coin data has been embedded within this system within our system of classification and the data and the contextualising without that without that contextualisation of the coins can't yes that's part of the story of newismatics how do we how is the data assembled is part of the story if you ignore that then you're going to miss it I mean the literature on coins is absolutely grim full of examples of people taking coins out of context and writing things about them which are not true well the literature on Sanskrit text people pick up a text of some author of the period we're working on and they think they can open that book and just step right into the age of Kallidasa without any understanding that how does Kallidasa land in his room he's been subject to copying and copying and copying over 700 or 800 years and the whole point of text critical study is to understand the layers and try and recover the oldest part of this text you can actually make a historical conclusion about what the shape of this text was now the fidelity is quite strong so it's not so bad but the point is that it's the same kind of methodology that's exactly what you're doing trying to get the chronology of the typologies people that are archaeologists think he's broken bits of pottery don't tell me anything but the thing is that getting the sequence of the pottery in order so that when you walk around beside an India that you're not going to excavate you can pick up a few pieces you can say right it's inhabited from the 6th to the 11th century and there's no sign of anything that's a big conclusion to be able to make in 10 minutes walking across the site which you can do in the eastern Mediterranean I think Seth wanted to raise a point that's its synagogus well about when we don't find coins I mean this seems to be a big problem as well because we're trying to infer economic social systems and all the rest on the basis of the coin finds and the main fact that you talked about is kind of collecting behaviour and the impact that's had but there's also surely lots of the actual occurrence of coins is high even as well and doesn't also necessarily reflect the usage and there must be so many factors there and most of the sort of archaeological situations that I've worked and it's been in that like the Indian Ocean where we have there's big debate about what extent were the Sasanians involved in Indian Ocean trade there are almost no Sasanian coins in circulation in the Indian Ocean does that mean they were or they weren't involved in Indian Ocean trade in that room you could argue it both ways couldn't you really and I think the same in India there's been I'm sure everyone knows a lot more about the me but the whole thing of coin hoards in Roman coin hoards and whether that indicates or not the absence of data is the question of absence and that's really interesting and we had the one example of this document with the jugs of wine and those bactrian documents have seals on them play seals and bits of string and they were sealed up and of course there's just this small cache of them that are complete a few others around but if you go to a place like Sunig in Punjab where the British Museum is about 20 or 30 just the seals and apparently somebody told me they were reading some report where 40,000 seals were found in Sunig I don't know if they're typo in the report but anyway there's a vast number of them there so this is like a banking centre with thousands and thousands of seals with all the documents gone the bait seal left so every single one of those had some kind of transaction on them like this and non-monitised probably so there's a huge amount of economic just the presence of those things in the case a huge amount of economic activity of the type by analogy you've got to be bold and say it's by analogy that's the scale of it and also the fact that the coins are un-attributable to us which means they're less collective but in fact we're looking at a very complex monetary system people are using it they've got to know what the values are how the values are determined so the fact that there's a huge number of un-attributable ugly coins actually refers to a very complex money use so there must be money changes involved so when 100 of these dramas were deposited against you would get interested in what sort of coin so there is a lot of complexity there which is sort of obliterated by the fact if you just look at the coins and say well these are ugly and unattributable then all that context is lost thinking about your context that is the main that is the same word but the distribution psychologically is influenced by so many important factors isn't it and so precious metal coinage it depends what kind of coinage precious metals are going to go out of circulation and so you might have huge areas that are totally monetised where you find almost zero coins psychologically and so and is there a bias in where the coins end up are we only looking at a small perhaps bias sample of a particular use of coins by looking only at the ones that have appeared to us it's very important to remember that the only two things that we know or that we can establish we don't in this case know because we don't have the information the only things we can establish are the moment when a coin was born the moment when it was made and the moment when it effectively died when it went into the ground not to be recovered until we got it we have those two moments but everything we want to know about the coin really happens in between it's the bit in between we want to know but we can only with numismatics in most cases tell what happens at the very beginning and the very end of that coin's activity and that's and there's enormous numbers of biases you're right there's systematic distortions in the way we find material the systematic distortions in the way the material is reported and something I picked up and was just looking up here is that one of the cases made is that the de monetisation is also associated with the de urbanisation phenomena sites go out of use in this period and they come back into use later claim that some of the people involved and there are a number of sites of significant sites which report large numbers of coins of the Cushans right which I could use for good stratigraphy there are no sites and people will reference this that there are no sites that would give me the same information for the Cushans successes in the north west the killerites, the Alcons et cetera except that there's a single reference in the literature about Ropar which says a hoard of 600 copper coins found in an earthen pot in sub period 4 de contains issues from 200 BC to 8600 these include Chandergupta local coins Sasanian fire altar coins which are all also found regularly in layers in the site itself but Ropar's total coin information is that one reference written in the encyclopedia of Indian archaeology by the excavator the material from Ropar has never been published whereas Achahtra published Cushan coins Taxila published twice Cushan coins so there's a systematic bias in terms of the reporting that takes place as well as the digging and of course in the sites themselves but the critical thing is that the critique of Sharma is not that Sharma's approach of attempting to count was wrong it's that when Sharma attempted to count he did so by excluding all the possible factors that might have distorted his count you could make that approach and say where can I count where I have some confidence that I'm not this is the whole point of this group of central provinces coins you're deliberately picking out a group where you say actually the various factors are not operating to the same degree on here here I might have confidence he's dealing with a damaged data set he's also carrying through certain mistakes that were made in the generation of those data sets but he is also doing something himself which is that he is making that neat equation of an absence of evidence is evidence of absence which we all know is an affixional you can't make that equation when Sharma himself is applying to his interpretation also one aspect of the value of coins is that as well as being archeological items they also have pictures of them which can actually tell you quite a lot so there are in most branches of the Univismatics two schools the people who are interested in what's on the coin and they rarely speak to each other but also the coins and how to quickly attribute a label that is the bias to actually identify and put a label as quickly as possible is the bane of it all because then you tend to sort of lose the sites of it you have to have a big fix that this is aching and here he is mentioning an inscription it must be hidden so we can put them in the trade with a nice label King for every coin King for every coin We've overrun our time and we've kept the video a tendon longer than he should be here so we are having dinner at the usual place everyone's invited because many people couldn't make it because of the weather so please come along for dinner if you want to follow the crowd usual place