Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 1 Introduction According to the Icelandic Sagas, Eric the Red founded the Greenlandic Norse settlement in the year AD 986 (Jones 1986). This first settlement (the Eastern Settlement) was augmented with another settlement (the Western Settlement), situated about 400 km to the north of the first one. The land was uninhabited, as the Thule Eskimos had not yet spread to the southwestern part of Greenland. A measure of the prosperity and the obvious steady contact with Iceland, and thereby Norway, was the ordination of a bishop to Gardar in the Eastern settlement in AD 1124 (Arneborg 1991). It seems, however, that contact with Iceland and Norway gradually faded, although the settlements were subjected to the Norwegian crown in AD 1261. According to written sources, the Western Settlement had become deserted by approximately AD 1360, and the last known written testimony from the Eastern Settlement dates to AD 1408 (Grønlands Historiske Mindesmærker III:145–150). It has been assumed that the Eastern settlement was finally depopulated a century later (Meldgaard 1965). The Big Picture Various explanations on the demise of the Norse settlements very much hinge upon how big the population was. If it was small, then even slight perturbations may rapidly have brought the population below sustainable levels, whereas a large population would have a better “buffering” effect. Estimating the population size may be done by, for example, analyzing the number and size of those churchyard sites that have been thoroughly excavated, and then calculating an average burial density. It is then possible to extrapolate from the known churchyard areas to a total inhumed population, which indicates an average population of some 1500–2000 people (Lynnerup 1998). Another way to evaluate population size would be to look at the biological framework for such a population. Was an increase from a starting population of 400–500 individuals to at least 2000 individuals possible within a 300-year period, without having to assume extreme values for life span, mortality, and fertility? And, consequently, could a population of this size pass into extinction? I arbitrarily set a starting population at 500 individuals based on historical research, drawing on information in the accounts in the Grønlændingesaga and Eric the Red’s Saga that “32 ships sailed for Greenland ... but only 14 made it there” (GHM I: 179, 207). It has been estimated that some of the larger Viking cargo ships of the period 1000 AD had a cargo capacity of 40 tons (Crumlin-Pedersen et al. 1992), allowing for a capacity of about 30 individuals per ship. Consequently, about 300–400 people could have settled in Greenland in the first wave (Meldgaard 1965). Keller has mentioned that the capacity of the ships may have been smaller, but that there was more regular immigration, leading to a starting population of about 300–800 people (Keller 1986). A minimum starting population of some 500 people would fit with the generally accepted minimum levels for sustainable populations of about 400–500 (Dyke 1984, Geist 1978). Using an exponential model, a rate of increase of 0.62% would have to be assumed for the population level to increase to about 2000 within about 200 years. Endperiod demographics of the Greenland Norse Niels Lynnerup* Abstract - The Norse who settled and lived in Greenland AD 1000–1500 provide a unique biological anthropological material for the investigation of human and environmental interaction. They were a relatively isolated population, constrained in both space and time. This study reports on demographic modelling of (1) marriage and fertility patterns and (2) total population level numbers with crude death rates and birth rates, including immigration and emigration scenarios. In addition to linear models, I also used stochastic models, which may be especially important when dealing with small populations. My results indicate that a small population such as the Norse inevitably will be close to a minimum sustainable level, and that simple emigration may explain the demise of the settlements. Furthermore, when the societal and economical implications of the population dynamics are considered, certain activities of the Norse in the latter phase of the settlement, such as the building of new churches, may acquire a new significance. To illustrate this, a so-called sunk-cost economic model was applied. Viking Settlers of the North Atlantic: An Isotopic Approach Journal of the North Atlantic *Laboratory of Biological Anthropology, The Panum Institute, Blegdamsvej 3, DK-2200 N Copenhagen, Denmark; n.lynnerup@antrolab.ku.dk. 2014 Special Volume X:XX–XX Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 2 This rate does in fact parallel a calculated population increase in Iceland for the period 970–1095 AD, using population figures from Thorarinsson (1961), and corresponds also to rates calculated for many other populations reported in ethnographic studies (between 0.05% and 0.59%; Hassan 1981:234). Indeed, recently, the peopling of Iceland has been proposed to take place over just 20 years (although with continuing “reinforcements”), basically going from none to more 20,000 (Vesteinsson and McGovern 2012). Although different numbers and circumstances apply, the pattern would probably be identical for Iceland and Greenland: a first phase with rapid peopling, followed by a longer-term, continuous but less-numerous immigration. For Greenland, the immigration rate probably dropped to zero (or rather the net rate of immigration and emigration was zero) over the years 1100–1200 AD. This reduction, along with a slowing of population increase (as the population approached carrying capacity and thus began to strain resources), could be consistent with a levelling-out by 1200 AD. It would be difficult to precisely set a definite carrying capacity, as the carrying capacity depends on general climatic environmental factors as well as the energy-extraction technology. However, it is generally assumed that the climate was favorable in the beginning of the settlement period, but then deteriorated throughout the period with the so-called Little Ice Age, as first demonstrated by Dansgaard et al. (1975). In fact, it has been possible to also demonstrate a changing climate directly in the dental enamel of the Norse and Thule Culture Eskimos (Fricke et al. 1995). Recent comprehensive stable isotope analyses of the Norse point to an increasing reliance on marine foodstuffs, reflecting the terrestrial carrying capacity was being reached (Arneborg et al. 2012a). But can the Greenland Norse be considered to have gone extinct? Given a positive growth rate, emigration remains a possible explanation of a decreasing population. Emigration may occur when a population exceeds its optimum level in relation to living conditions (Hassan 1981). This scenario could have occurred because the Norse population reached the carrying capacity of their habitat, which may in fact itself have been decreasing, at least as concerns farming. Cultural identity probably also played a strong role: the Norse saw themselves as farmers, and having livestock was important in this respect, as was also probably a sense of wanting to be part of the wider Norse North Atlantic “sphere” with regular contact and trade (Arneborg et al. 2012a). Allowing for a decreasing rate of growth (as especially young people tend to emigrate, thereby not only counting themselves out of the population but also removing their potential offspring), the emigration rate would have had to have been about 8 individuals per year. For a small population like the Norse, such a level of emigration would certainly have had a dramatic effect. The question is whether such rates are realistic, keeping in mind that these rates represent averages. In other words, emigration would probably have taken place in waves of, for example, 100 people (i.e., three boats) every ten years. The size difference between the Western and Eastern Settlements must also be taken into account. Judging by the numbers of farmhouses (West: 80, East: 250; Gad 1984) and by the number of churches (West: 3, East: 11), there seems to be a 1:3 ratio between the settlement population sizes. Viewed in isolation, the Western Settlement would rapidly approach the minimum population size of 500. Indeed, given the above emigration rates and lowered fertility rates, the population would have fallen below this level after just 20 years. It is thus possible that a decline began at around AD 1300, and that approximately 50 years later, most people would have had emigrated from the Western Settlement, leaving perhaps only a few, mainly elderly, settlers. The Western Settlement could thus have been completely depopulated during the 14th century. Perhaps the Eastern Settlement then experienced some immigration, which for some years offset the rate of decline, but then decline proceeded at the same rate as in the Western Settlement. This scenario would leave the Eastern Settlement depopulated in the mid-15th century. A population may also become extinct due to exceptionally high mortality rates. High mortality rates are usually linked with war and epidemics. However, even rates of up to 10% of young adults killed in warfare would not substantially decrease birth rates. Thus, highly lethal epidemics are the only probable cause of a dramatic reduction due to mortality in the population within a short time span. Plague struck both Iceland and Norway in the 14th and 15th centuries, and total mortality rates of between 30% and 50% have been suggested (Benedictow 1992). Clearly, a halving of the Norse population in just one or two years would be disastrous for such a small population, and on the basis of the numbers alone could quite plausibly explain the population reduction. There is no historical evidence of plague in Greenland, and plague is a disease that does not leave any signs in the skeleton. While a few multiple synchronic graves have been found, most recently at Ø64 (J. Arneborg, Danish National Museum, Copenhagen, Denmark, pers. comm.), such burials do not Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 3 point to plague as such, and certainly not anything resembling the plague pits of European cemeteries at the time of the great plague epidemics. Based on the above, a bigger picture emerges, allowing for an accumulated population size of 26,500 for the total period of Norse presence in Greenland, equivalent to an average population size Nc of 1377 at any one time. These figures are very close to the figures arrived at in the preceding section, based on the number of interments in cemeteries. These figures highlight that the Norse population could very well have reached reasonable levels and conversely pass into extinction within the 500-year span of Norse settlement, without assuming undue biological parameters either for immigration and fertility, or emigration and mortality. Such a hypothetical population “profile” is shown in Figure. 1. The Smaller Picture The above demographic model is based on linear models. Thus, the population model is at best a grand average. While the overall model may reflect real changes, no real population behaves as “smoothly” as linear models. Real-life populations have a much more “rugged” pattern, i.e., small fluctuations due to deaths and births not occurring at regular intervals (so-called stochastic variation). This discrepancy is not a problem for such general models as the fluctuations will tend to even out. Yet, it may be of great importance for small populations; once the population is sufficiently small it becomes vulnerable to fluctuations in fertility and mortality (Weiss and Smouse 1976). The Norse population was small, and therefore the occurrence of births and deaths may not even out. In terms of sustaining the population level, the number of girls born is especially important. Assuming, as is often done for medieval populations (Hassan 1981), a gross child mortality rate (i.e., up to adulthood) of 50%, then each woman would have to give birth to two girls in order for one of them to reach fertile age and have children (thus keeping the population level steady). In a small population such as the Western Settlement, there might only be 15 to 20 births a year, corresponding to 7–10 girls, of which maybe 3–5 survived childhood. If for some reason even more of the small children and newborns die due to, e.g., an infectious disease, food poisoning, etc., this may mean that the entire replacement stock of girls that year are removed. The population as a whole would in that case be severely challenged to keep up its number. There would simply not be enough females to give birth to enough children for the population to be sustainable. A bigger population may buffer this, but for small populations this can have severe downstream demographic implications. Figure 2 shows a simple stochastic modelling, where a computer randomly selects a birth or a Figure 1. Population profile of the Norse settlements in Greenland. Upper curve is for the total population (East and West Settlements); middle curve East Settlement only; lower curve West settlement only. The curves chart a course of population development from ~500 settlers, reaching a peak level of ~2000 individuals, and then, inside 200 years, a decline to 0. Note that the West settlement is the smallest settlement, thus the one most vulnerable to stochastic changes and the first to collapse (see text and Fig. 2). Figure 2. Simple stochastic modelling, where a computer randomly selects a birth or a death. Purely by chance such a “run” may be rather steady (“I”: deaths more or less equal births) or show a decrease (“II”: more deaths than births). Such random developments may have big implications for small populations. Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 4 death. Purely by chance, such a “run” may be rather steady (deaths more or less equal births) or show a decrease (more deaths than births) or show an increase (more births than deaths). Further mathematical analysis concerning the West Settlement suggests that a decline in the sex ratio (fewer females as to males) may very well have been the case, with the total population number falling below 50 well within the attested timeframe of Norse settlement (Di Bacco et al. 2006). These calculations were based on finds at the churchyard at W51 Sandnes, where the skeletal material dates to around the 12th century (Arneborg et al. 2012b), and is thus not a reflection of, e.g., an earlier “founder” population, where one might perhaps suppose there was a skew towards males. Societal Changes If the Norse were facing an ever more arduous and difficult life in Greenland, then why are the historical and archaeological sources silent about what happened? If people left, why is there no incontrovertible evidence for this, e.g., in terms of evidence such as the Sagas telling about the Landnam period? Intriguingly, there is a complete absence of rich finds from Greenlandic Norse settlements. No church inventory, weapons, objects of fine metals such as jewellery has been found among the Norse settlements (Berglund 1986). Indeed, this might point to the very fact of a gradual, depopulation. When you leave, you take along your valuables. However, other archaeologists have pointed to churches, which seemingly were built almost up to the demise of the settlements. If the Norse were leaving, then why invest labor in building new churches? Would this not rather point to a sudden and disastrous demise? The latter question touches upon a very fundamental question of how past populations “fail”, and how this is perceived in the archaeological record. I have tentatively applied the “sunk-cost” model, as proposed by Janssen et al. (2003). In brief, this model first states that past societies (or recent for that matter) will balance between “recovery” and “collapse” periods. Depending on the size, ecology, economy, and other parameters, each society may possess a high or low buffering capacity or resilience. Thus, a society where surplus is generated may weather a period of, for instance, crop failure for a certain period of time. The Norse society, being small, and probably under climatic stress, would have a small buffering capacity and low resilience. That is, maybe just a few years of poor or failed crops, disease among the livestock, etc., may have driven the society past a demographic and social tipping point, so that it collapses without being able to recover. Furthermore, the model states that human investment in societal matters is not always rational. For example, the Norse in the Eastern Settlement established water canals to sustain grazing land. This is an example of a clearly rational investment. On the other hand, more irrational investments may be made (or, and this must be emphasized, irrational to us), such as building a new church. But it is not so much a question of whether the Norse or modern researchers studying them define an investment as rational or irrational, nor whether such activity may be deemed “wasteful”, following Dunnell (1999), i.e., activity not directly related to immediate welfare or reproduction. The central issue in the “sunk cost” model is that once an investment is made, then it is very difficult to change it. Indeed, the investment might “take a life of its own”, with ever increasing investments, because the object of investment may become a symbol for the society, something which binds the society together or shows defiance in the face of adversity. The investment seemingly offsets collapse; it may even in the archaeological record look as if the society is capable of, e.g., constructing huge and magnificent buildings, but in an economic reality, the society is collapsing. And when it collapses, it looks all the more sudden (Janssen et al. 2003). To put this model in a Norse perspective, one could envision the following scenario. Adversity hits first the Western Settlement, the smallest and most challenged ecologically/economically. Some leave, but others show defiance, perhaps hoping that things will reverse. As a (for them) rational investment, a new church is erected; perhaps a new church showing the devoutness of the people, which will appease God. Building a new church is a big investment, and to perhaps abandon it half-finished would be a sign of weakness or lack of faith. Thus the church gets built, even though more people leave, and the general situation is ever more bleak. Finally the church is finished, but only few years later, the settlement collapses completely, leaving perhaps only a few elderly people on their farmsteads. In other words, collapse started early, and the overall societal activity was in steady decline. Yet for the archaeologists finding a new church, it may seem that decline only set in late (Fig. 3). In other words, the Norse settlements in Greenland experienced a steady decline over 100– 200 years, not a sudden collapse, even though investments such as a new church, were being made right to the end. I feel that the “sunk cost” model, coupled with archaeological and anthropological data holds potential for future investigations. New buildings, big investments, are not necessarily evidence of a society far from collapse. Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 5 Conclusion Ultimately, the demise of the Norse settlements was not something out of the ordinary. In fact, there seems to have been a massive depopulation in most European countries at that same time. As mentioned earlier, a 60% decrease in population has been projected for northern Norway, and at least a 30% decrease for Iceland. This massive depopulation, usually ascribed to the great plague epidemics, had enormous demographic, economic, and social repercussions throughout the wider social and demographic networks of Europe and the North Atlantic region. In the wake of the plague, there was largescale population resettlement, where inhabitants of the more unproductive areas left for the better, “vacant” zones. Whether Norse Greenland was directly affected by plague or not, it would, as a highly peripheral area, most certainly have been affected by the indirect effects of plague. For instance, the export prices of several Icelandic commodities fell dramatically (Keller 1986), and this may well have had economic consequences for Greenland (Arneborg et al. 2012a). Since it seems that the decline in population levels in fact had already started before the plague epidemics reached the northern European countries, and we can assume from palaeoclimatic and archaeozoological evidence that there was climatic change in the years after 1300 AD, this shift likely “stressed” the population (Brothwell 1998), probably resulting in a trend towards ever harsher living conditions. Adaptive responses would include increasing marine foodstuff reliance and seeking to increase land yield; the former evidenced by the stable isotope analyses, and the latter perhaps reflected by irrigation systems and the buffering capacity of the local community, headed by large farmsteads (Arneborg et al. 2012a). However, perhaps after some internal resettlement, emigration accelerated in the 15th century. Better land became available in a larger community, e.g., Iceland, and it is even possible that old family claims could be invoked (e.g., Halstead and O’Shea 1989). The marginal land of Greenland no longer held the same attraction. If correct, the scenario sketched out here would also serve to eliminate some of the “spectacularity” of the demise of the settlement. The Norsemen moved to Greenland because of a perceived gain and the possibility of owning land, perhaps pushed to some extent by population pressure and the rapid exploitation of Iceland (Anthony 1990, 1997). They Figure 3. Model of the Norse collapse. The minimum level of societal sustenance is the (not well-defined) level at which a society irreversibly breaks down (dashed line). The black line indicates that the society was over many years in continuing decline (“what really happened”). The dotted line shows that the societal activity may have seemed to be not in declince, e.g., due to building of a new church (“what the archaeologists see”). Yet the building of a new church did nothing to change the fundamental decline, it only offset the rate of final collapse; thus, the demise of the Norse settlements may seem sudden, but in reality the decline was ongoing. Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 6 Benedictow, O.J. 1992. Plague in the Late Medieval Nordic Countries. Middelalderforlaget, Oslo, Norway. Berglund, J. 1986. The decline of the Norse settlements in Greenland. Arctic Anthropology 23:109–135. Brothwell, D. 1998. Stress as an aspect of environmental studies. Environmental Archaeology 2:7–13. Crumlin-Pedersen, O., M. Schou Jørgensen, and T. Edgren. 1992. Skibe og Samfærdsel. Pp. 42–51, In E. Roesdahl (Ed.). Viking og Hvidekrist. The National Museum, Copenhagen, Denmark. Dansgaard, W., S.J. Johnsen, N. Reeh, N. Gundestrup, H.B. Clausen, and C.U. Hammer. 1975. Climatic changes, Norsemen ,and modern man. Nature 317:806–809. Di Bacco M., L. Del Panta, P.F.G. D’Amore, and N. Lynnerup. 2006. The effect of an unbalanced demographic structure on marriage and fertility patterns in isolated populations: The case of Norse settlements in Greenland. Genus LXII(1):97–119. Dunnell, R.C. 1999. The concept of waste in an evolutionary archaeology. Journal of Anthropological Archaeology 18:243–250. Dyke, B. 1984. Migration and the structure of small populations. In A.J. Boyce (Ed.). Migration and Mobility. Taylor and Francis, London, UK. Fricke, H.C., J.R. O’Neil, and N. Lynnerup. 1995. Oxygen isotope composition of human tooth enamel from medieval Greenland: Linking climate and society. Geology 23:869–872. Gad, F. 1984. History of colonial Greenland. Pp. 556–576, In W. Sturtevant (Series Ed.). Handbook of North American Indians, Volume 5 Arctic (D. Damas [Volume Ed.]). - Smithsonian Insitution Press, Washington, DC, USA. Geist, V. 1978. Life Strategies, Human Evolution, Environmental Design. Springer-Verlag, New York, NY, USA. Grønlands Historiske Mindesmærker (GHM) I–III. 1838–1845. Det Brünnichske Forlag (reprinted 1976 by Rosenkilde and Bagger), Copenhagen, Denmark. Halstead, P., and J. O’Shea. 1989. Introduction: Cultural responses to risk and uncertainty. Pp. 1–7, In P. Halstead and J. O’Shea (Eds.). Bad Year Economics. Cultural Responses to Risk and Uncertainty. Cambridge University Press, Cambridge, UK. Hassan, F.A. 1981. Demographic Archaeology. Academic Press, New York, NY, USA. Holopainen, J,. and S. Helama. 2009. Little Ice Age farming in Finland: Preindustrial agriculture on the edge of the grim reaper’s scythe. Human Ecology 37(2):213–225. Janssen, M.A., T.A. Kohler, and M. Scheffer. 2003. Sunkcost effects and vulnerability to collapse in ancient societies. Current Anthropology 44(5):722–28. Jones, G. 1986. The Norse Atlantic Saga (2nd Edition). Oxford University Press, New York, NY, USA. Keller, C. 1986. Nordboerne på Grønland 985–1350. Bidrag til en demografisk økologisk diskusjon. Universitetets Oldsaksamling Årbok 1984/1985:145–157. moved back when this possibility arose again in Iceland. Indeed, it would be surprising, in the light of the almost universal demographic changes and overall depopulation in Norway, Iceland, England, etc. (see, for example, Holopainen and Helama 2009, Lagerås 2007), if a remote and already economically vulnerable settlement like the Norse settlement in Greenland did not decline. Based on our exhaustive stable isotope study, concerning the Norse, their domesticates, and game (Arneborg et al. 2012a, b), we hope to be able to carry out much more detailed analyses of the Norse farming strategies and their use of marine foodstuffs, in order to better calculate the carrying capacity of the Norse settlements. These efforts in turn may then allow for a higher-resolution demographic modelling. As such, the Norse case study may apply to other small and isolated populations in the past. It has been suggested that regional extinction and re-colonization can explain many patterns in the archaeological record of the Pleistocene (Premo and Kuhn 2010, Riede 2009), but perhaps also Neolithic settlements should be seen in this light, if even the relatively advanced agricultural populations of the medieval period were prone to stochastically caused demographic collapse following the relatively minor climatic changes of that period. Author’s Note This paper is based on my previous publication: Lynnerup, N. 2011. When populations decline: Endperiod demographics and economics of the Greenland Norse. Pp. 335–345, In T. Meier and P. Tillesen (Eds.). Ûber die Grenzen und zwischen den Disziplinen. Archaeolingua Alapitvany, Budapest, Hungary. Literature Cited Anthony, D.W. 1990. Migration in archaeology: The baby and the bathwater. American Anthropologist 92(4):895–914. Anthony, D.W. 1997. Prehistoric migration as social process. Pp. 21–32, In J.C. Chapman and H. Hamerow (Eds.). Migration and Invasions in Archaeological Explanation, 664. Oxbow Books, Oxford, UK. Arneborg, J. 1991. The Roman church in Norse Greenland. Acta Archaeologica 61:142–150. Arneborg J., N. Lynnerup, and J. Heinemeier. 2012a. Human diet and subsistence patterns in Norse Greenland A.D. c. 980–A.D. c. 1450: Archaeological interpretations. Journal of the North Atlantic Special Volume 3:119–133. Arneborg J., N. Lynnerup, J. Heinemeier, J. Møhl, N. Rud, and A.E. Sveinbjörnsdóttir. 2012b. Norse Greenland dietary economy ca. A.D. 980–ca. A.D. 1450: Introduction. Journal of the North Atlantic Special Volume 3:1–39 Journal of the North Atlantic N. Lynnerup 2014 Special Volume X 7 Lagerås, P. 2007. The Ecology of Expansion and Abandonment: Medieval and Post-medieval Agriculture and Settlement in a Landscape Perspective. Riksantikvarieämbetet, Stockholm, Sweden. Lynnerup, N. 1998. The Greenland Norse. Meddelelser om Grønland, Man and Society Series 24. Meldgaard, J. 1965. Nordboerne i Grønland. Munksgård, Copenhagen, Denmark. Premo, L.S., and S.L. Kuhn. 2010. Modeling effects of local extinctions on culture change and diversity in the Paleolithic. PLoS ONE 5(12):e15582. Riede, F. 2009. Climate change, demography, and social relations: An alternative view of the Late Palaeolithic pioneer colonization of southern Scandinavia. Pp. 3–10, In S. McCartan, P.C. Woodman, R.J. Schulting, and G. Warren (Eds.). Mesolithic Horizons. Oxbow Books, UK. Thorarinsson, S. 1961. Population changes in Iceland. Geographical Review 51:519–533. Vésteinsson, O., and T. McGovern. 2012. The peopling of Iceland and comments. Norwegian Archaeological Review 45(2):206–235. Weiss, K.M., and P.E. Smouse. 1976. The demographic stability of small human populations. Journal of Human Evolution 5:59–74.