Climate and social change at the start of the Late Antique Little Ice Age

Abstract

The Late Antique Little Ice Age, spanning the period from 536 CE to roughly 560 CE, saw temperatures in the Northern Hemisphere drop by a degree C in less than a decade. This rapid cooling is thought to have caused widespread famine, epidemic disease, and social disruption. The relationship between cooling and social disruption is examined here using a set of high-resolution climate and historical data. A significant link between cooling and social disruption is demonstrated, but it is also demonstrated that the link is highly variable, with some societies experiencing dramatic cooling changing very little, and others experiencing only slight cooling changing dramatically. This points to variation in vulnerability, and serves to establish the Late Antique Little Ice Age as a context within which naturalistic quasi-experiments on vulnerability to climate change might be conducted.

Keywords

archaeology climate change history Late Antique Little Ice Age social resilience vulnerability

Introduction

Climate scientists are in agreement that a period of global cooling, commonly referred to as the Late Antique Little Ice Age (LALIA), began in the mid-6th century, probably the result of a series of volcanic eruptions in 536, 540, and 547 CE (Büntgen et al., 2016; Helama et al., 2016; Toohey et al., 2016). Climate scientists, historians, and archaeologists are also in general agreement that the LALIA provoked widespread social disruption, famine, and episodes of epidemic disease across the Northern Hemisphere (Gunn, 2000; Helama et al., 2018; but cf. Moreland, 2018). Despite good evidence for both the LALIA and temporally associated social disruptions, no one to date has attempted to systematically determine if the two are empirically related to one another at both local and global scales. Here a set of high-quality historical data on twenty Northern Hemisphere societies along with high-resolution paleoclimate data are used to conduct a cross-cultural study of the relationship between the LALIA and social change. This study finds that the LALIA is significantly associated with social change across the Northern Hemisphere and that there is a statistically significant relationship between the intensity of cooling and the degree of social change in this sample.

The beginning of the LALIA is here assumed to be 536 CE, when the first of three large volcanic eruptions forced dramatic cooling across the Northern Hemisphere (Sigl et al., 2015; Toohey et al., 2016). While the exact locations of these eruptions are debated (Nooren et al., 2017), it is clear that they produced the largest and longest atmospheric loading event in recorded history (Dull et al., 2019). This extraordinary atmospheric catastrophe took place during a period where both good archaeological and historical data are available, and thus may provide a unique a naturalistic quasi-experiment through which hypotheses about social vulnerability and resilience in the face of massive climatic change might be tested (deMenocal, 2001). A naturalistic quasi-experiment refers to a condition where a sample of societies are differentially impacted by a treatment (in this case, the onset of the LALIA) and the differential effects of that treatment can be systematically measured and evaluated (Dunning, 2008, 2012; Leatherdale, 2019). If such a condition in fact existed it would offer a context for quasi-experimental research that is of vital importance today as the world’s societies face the possibility of a similar, if not greater, period of climatic change by the end of the century (Collins et al., 2013; Riede, 2014).

Two basic questions about climate and social change at the start of the LALIA are explored here. The first is whether social change can be empirically shown to have increased among societies in the Northern Hemisphere following the onset of the LALIA. In terms of a quasi-experiment, this question can be understood as asking whether or not cooling at the start of the LALIA can be taken as a “treatment” of the cases. The second is whether a predictable association exists between the degree of social change and the degree of cooling. In the context of a quasi-experiment, this question can be understood as asking whether or not a “dosage effect” is present. If either or both questions are answered in the affirmative then there may be a context where naturalistic quasi-experiments are possible (Diamond and Robinson, 2010; Leatherdale, 2019).

Methods

The cross-cultural method is used to explore these two questions. The cross-cultural method differs from the more common case study analysis in that a systematic sample of cases are examined that are expected to display a range of variation on variables of interest rather than cases that exemplify particular variants. A systematic sample allows the potential to answer questions probabilistically rather than simply illustrating answers anecdotally. The logic of the cross-cultural method is that if an hypothesis about cultural stability or change has merit then the hypothesized causes and effects should be strongly associated across a wide range of cultural and ecological variation (Ember and Ember, 2000; Smith and Peregrine, 2012). Samples for cross-cultural research are selected in order to represent this range of variation and to avoid spatial or network (through ancestral links) autocorrelation (“Galton’s problem”) which might increase apparent relationships between dependent and independent variables (Dow, 2007).

The sample used in this study was selected in order to minimize autocorrelation while simultaneously providing cases with adequate historical and archaeological data to allow for coding. Fifteen of the 20 cases were selected from pre-existing (and partially pre-coded) samples. Twelve cases were selected from the Seshat Databank (seshatdatabank.info), which assembles historical and archaeological information (e.g. archaeological site reports, primary and secondary historical sources, demographic data, and the like) to provide an archive of both raw and pre-coded data to test hypotheses about long-term cultural stability and change (Turchin et al, 2018, 2020). The Seshat Databank “World Sample 30” provides data for cultural sequences at 30 locations across the globe, selected to minimize autocorrelation and maximize cultural and environmental diversity (http://seshatdatabank.info/methods/world-sample-30/). Twelve of the cases in the “World Sample 30” were in the Northern Hemisphere and had societies that existed during the LALIA. These 12 cases were selected for coding and analysis. It must be noted that while there has been some controversy surrounding the way data have been coded for the Seshat Databank (e.g. Slingerland et al., 2019), none of the coded variables in question were used here, and coding itself followed a standard protocol used extensively in cross-cultural research (see Ember and Ember, 2000)

Second, three cases previously coded for a project on social resilience to climate-related disasters overlapped the LALIA and were also selected for coding and analysis (Peregrine, 2018a, 2018b). The dependent variable used in this study had already been coded for these three cases, had already been demonstrated to be valid and reliable, and thus provided a clear protocol for extending coding to the new cases. Finally, five cases were added to ensure that there would be some coverage of all major regions of the Northern Hemisphere. These five additional cases were selected (as all the others) from areas that had well-studied archaeological and historical records. The locations of the cases are shown in Figure 1 and listed in Table 1.

Figure 1.

Locations of the 20 sample cases, showing social and temperature change for each case in the sample. The left bar indicates the percent of total change in the SCI. The right bar indicates the percent of total change in mean annual temperature. Under the assumption that greater temperature change leads to greater social change (i.e. that there is a “dosage effect”) one would expect the right bar to be equal to or higher than the left.

Table 1.

Cases used in the analyses and associated information. The first column gives the full case name followed by the abbreviated name in parentheses. The second column gives the full time period from which data were collected (although data were focused on the 20 year period before and after 536 CE). The third column gives the mean temperate change between the period 525–535 CE and the period 536–546 CE. Values with an asterisk indicate a temperature decline that is statistically significant at the 0.05 level (1-tailed) based on a dependent-samples t-test comparing mean annual temperatures 525–535 CE with 536–546 CE. The final column gives the case’s score on the SCI. Values with an asterisk indicate that change is statistically significant at the 0.05 level (1-tailed) based on Z-tests comparing the SCI score with the distribution of expected change.

Case name (short name)	Time frame	Temperature change (degrees C)	Social change index
North America
Point Peninsula Complex (Ontario)	300 BCE–700 CE	−0.74*	6
Mund Phase (Cahokia)	450 CE–600 CE	−0.70*	8*
Pioneer/Formative Phase (Gila)	1 CE–750 CE	−0.74	9*
Monte Alban IIIB and IV (Oaxaca)	500 CE–900 CE	−0.15	15*
Early Classic (Tikal)	250 CE–600 CE	−0.15	12*
Europe
Early Merovingian (Paris)	486 CE–543 CE	−0.79*	6
Ostrogothic Kingdom (Latium)	489 CE–554 CE	−0.69*	14*
Migration Period (Jutland)	500 CE–700 CE	−1.00*	10*
Brega (Ireland)	Reign of Túathal Máelgarb 533 CE–544 CE	−0.51*	6
Toledo (Spain)	Reign of Theudis 531 CE–548 CE	−0.58*	15*
Eastern Asia
Rouran Khaganate (Orkhon)	300 CE–555 CE	−0.73*	14*
Hephthalites (Sogdiana)	408 CE–561 CE	−0.35	14*
Kofun (Kansai)	250 CE–710 CE	−0.33*	10*
Early Imperial Period (Luoyang)	200 BCE–900 CE	−0.60*	9*
Liang (Yangtze)	Reign of Emperor Wu, 502–549 CE	−0.60*	7*
Southern Asia
Sasania Period (Susiana)	224 CE–642 CE	−0.40	6
Kadamba Empire (Deccan)	354 CE–540 CE	−0.26	11*
Gupta Empire (Ganges)	Reign of Kumaragupta III 530–540 CE	−0.36	10*
Northern Africa
Byzantine Empire (Egypt)	395 CE–631 CE	−0.24	14*
Jenne-jeno III (Niger)	400 CE–900 CE	−0.26	10*

The dependent variable in the analysis, the Social Change Index (SCI), was created by summing the six measures of social change that are listed in Table 2. Coding protocols have been published elsewhere and both the six variables comprising the SCI and the SCI itself have been demonstrated to be valid and reliable (Peregrine, 2018a, 2018b). With the data employed here the SCI has an alpha of 0.776. The six dependent variables were coded by contrasting the conditions for the roughly 20-year period prior to 536 CE and those for the roughly 20-year period following. Data of fine enough resolution to keep within those 20-year ranges was not always available, and in those cases data with the best temporal resolution available for the periods both before 536 CE and after were employed. In all cases the values coded were within a 100-year range of the 536 CE date. Data collection was focused, to the extent possible, on a single community or region within the larger case. This is standard practice in cross-cultural research and is done as a way to control for the range of social diversity found in different geographical locations within any given society (Ember and Ember, 2000).

Table 2.

Variables comprising the Social Change Index (SCI). All are coded on a three-point (1) none, (2) some, (3) much scale.

Variable	Question coded for social change index
DV 1	Was there a change in population size or migration?
DV 2	Was there a change in the frequency or intensity of famine or disease?
DV 3	Was there a change in the frequency or intensity of conflict?
DV 4.1	Was there a change in the organization of communities?
DV 4.2	Was there a change in the regional organization of the society?
DV 4.3	Was there a change in the religious or ritual organization of the society?

It is important to emphasize that there is diversity in all societies, and there is diversity in the inferences made about past societies. But knowledge of the past changes as more information is uncovered, and interpretations of the past change as more is learned. This is the reality of research in the historical sciences. The data presented here represent the best approximation of reality based on the available information and interpretations of that information, but they do not in any way represent the “truth” about the past. By providing as much supporting information as possible in the Supplementary Materials and the raw data posted on-line, it is hoped that current and future scholars might return to these data, make corrections or provide new interpretations, recode variables based on their own protocols, and either replicate of falsify the results presented here.

Coding itself was done on the Dacura platform which allowed both numeric codes and all supporting documentation to be input in a Linked Data format (Peregrine et al., 2018). Dacura uses RDF-triplestore to create semantic links between both textual and pre-coded data that allows for both data harvesting and data sharing on the semantic web. Coding began by collecting quotes from textual sources (specifically from archaeological site reports and secondary archaeological and historical sources—very few primary historical documents were employed) and data from images (primarily archaeological) and inputting them directly into Dacura along with bibliographical information. After the source materials were input they were read independently by the author and a researcher assistant who then worked together to agree on initial codings for each variable. Initial codings were re-evaluated after all the cases were completed, and experts from Seshat’s group of contributors (seshatdatabank.info/seshat-about-us/contributor-database) were invited to review the codings and underlying source materials. Revisions were made based on expert’s responses and suggested additional source materials, and final codes were decided upon by the author and the researcher assistant. It is important to note that expert input varied, with numerous suggestions made for some cases, and none for others. As noted above, readers are encouraged to examine the Supplementary Materials and on-line data to revisit and refine codings based on new material or interpretations.

Measures of expected annual social change were calculated from two different datasets. The first employed Murdock and Provost’s (1973) 10-item Index of Cultural Complexity to examine general patterns of cultural evolution over the last 14,000 years (Peregrine, 2003). The second employed a set of variables from the Seshat Databank created to explore factors underlying the evolution of cultural complexity over the past 12,000 years (Turchin et al., 2018). For each dataset the slope of the least-squares regression line for the association of time and cultural complexity was used to estimate the annual amount of change (Supplementary Figures 1 and 2 available online). This estimate was then transformed into a Z-score and applied to the SCI (M = 10.3, σ = 3.21, min. = 6, max. = 18), so that an estimate of change specific to the SCI could be produced. The estimates from both data sets were very close: that change over the 20 year period from 536 CE to 556 CE was estimated to be 1.14e-6 Z-scores or 1.87e-6 Z-scores, respectively, implying that change in SCI scores would be expected to increase by 3.6e-6 or 6.0e-6 scale points, respectively. A distribution of expected change was then created for each estimate using a mean of 6 plus 3 times the expected increase in scale points, and a standard deviation equivalent to the expected increase in scale points. These distributions were used to perform the Wilconxon tests reported. They were also used to define the prior probability distribution for the reported Bayes factors. SPSS script and output for these calculations and analyses are provided in the Supplementary Materials.

Temperature data were derived from Raphael Neukom and colleagues’ (2019) analogue method five-degree annual temperature reconstructions, which were themselves based on the PAGES 2k global temperature-sensitive proxy collection (2017). Neukom and his colleagues ran one hundred reconstructions on these data for each year in each five-degree unit. The mean of these 100 reconstructions is used in the analyses here (Figure 2). The R script used to export data from the raw NetCDF datafile and to calculate mean annual temperatures for each of the 20 cases is provided in the Supplementary Material. It is important to note that some of the proxies used in the reconstructions covered large geographical areas, and therefore the resolution of these data are variable and some degree of autocorrelation is anticipated (Raphael Neukom, 2019, personal communication).

Figure 2.

Mean annual temperature variation for each of the 20 cases. Data are derived from Neukom et al. (2019).

Data on growing season length and sunlight constraints were derived from the GIS shapefile values at the focal community for each case. The GIS shapefile for length of growing season was produced by the Food and Agriculture Organization of the United Nations based on the Global Agro-Ecological Zones 3.0 database (Supplemental Figure 3 available online) (IIASA-FAO, 2012). The sunlight constraints shapefile was produced by the Conservation Biology Institute (consbio.org) based on data provided by the Numerical Terradynamics Simulation Group (www.ntsg.umt.edu) (Supplemental Figure 4 available online) (Churkina and Running, 1998).

Results

Previous studies have demonstrated that global and regional temperatures dropped during the LALIA, and it appears that the 20 sample societies experienced a dramatic period of cooling beginning at 536 CE and continuing until roughly 558 CE (Figure 2 and Supplementary Figures 5–24 available online). Comparison of mean annual temperatures for the period 486 to 526 CE with those of the period from 536 to 556 CE yields a statistically significant difference (t = 9.410; df = 19; p < 0.000). However, comparison of temperature trends for 525 to 535 CE with those for 536 to 546CE for each individual case (using Z-tests) indicates that only 11 of the 20 cases show a statistically significant cooling across the 536CE threshold (Figure 1 and Table 1). While the average temperature across these 20 cases cooled after 536CE, the cooling was not uniform and was not statistically significant at almost half of the locations. Thus, what has been suggested as a global cooling event was in fact highly localized (Neukom et al., 2019). This localization appears to be true for social change as well.

The mean of the Social Change Index (SCI) is 10.3, which is significantly different from the mean expected (6.0) if there was of no change during this time period (Wilcoxon Z = −3.529, p < 0.000; Wilcoxon tests were used because the SCI is not normally distributed). However, social change is something that is always taking place, so an assumption of no change over a 20-year period is not entirely warranted. As discussed above, two sets of data measuring cultural evolution over the last 12,000 years (Peregrine, 2003; Turchin et al., 2018) were used to estimate the amount of social change that might be expected during the period from 536 to 556 CE. Comparison of the expected change based on these data with the actual change give results that are very similar to those under the assumption of no change (Wilcoxon Z = −3.547, p < 0.000 and Bayes Factor 0.000 for both estimates of expected social change; a Bayesian analysis is included here because the shape of the prior probability distribution is clearly defined). Thus the first question is answered in the affirmative—the start of the LALIA clearly does mark the beginning of a period of dramatic social change, supporting, in a systematic and generalizable way, the conclusions reached by many scholars based on case studies (Gunn, 2000; Helama et al., 2018).

However, social change did not occur to the same degree across all 20 cases; indeed, four cases experienced no change (Figure 1 and Table 1). While social change did occur in dramatic fashion at the beginning of the LALIA, social change, like temperature change, appears to have been highly localized. The second question asks whether or not there is a predictable association between cooling and social change such that societies that experienced greater cooling also experienced greater social change (i.e. is there a dosage effect). Because both temperature and social change were varied and localized, there should be enough variation even in this small sample to address this question. Correlation of mean temperature difference for the periods from 486 to 526 CE and from 536 to 556 CE with the SCI indicates a statistically significant relationship (Spearman’s ρ = −0.459, p < 0.021 (1-tailed)). However, such a relationship is not indicated when only those cases that experienced significant cooling are analyzed (Spearman’s ρ = −0.157, p > 0.323 (1-tailed)). Thus while the degree of cooling in the LALIA does appear to be generally associated with social change, it does not appear to be predictably associated with social change among those societies that experienced the most significant cooling. This is puzzling.

The lack of a predictable association between the local degree of cooling and local social change seems odd given that dramatic cooling did take place during the 536 to 556 CE period and that cooling is predictably associated with social change when considering all 20 cases. It is apparent from historical records and historian’s interpretations of those records that there seems to be a direct link between the 536/540/547CE eruptions, dramatic climatic disruption, and social change in specific cases (Gunn, 2000; Keys, 1999; McCormick et al., 2012). What then are the factors that might mediate the direct effect of cooling on social change in some of the cases examined here? One possibility is local variation in the density of atmospheric aerosol particles released by the 536/540/547CE eruptions (Helama et al., 2018; McCormick et al., 2012). To test this idea the correlation between cooling and social change was performed controlling for the modern length of the growing season (IIASA-FAO, 2012) and modern sunlight constraints (Churkina and Running, 1998) at each of the 20 case locations (Supplemental Figures 3 and 4 available online). These partial correlations do affect the uncontrolled correlations and produce statistically significant results (Spearman’s ρ = −0.862, p < 0.006 (1-tailed) controlling for length of growing period and Spearman’s ρ = −0.743, p < 0.035 (1-tailed) controlling for sunlight constraints). While there are obvious concerns comparing modern conditions with those 1500 years in the past, it is reasonable to assume that there would be some continuity, and controlling for these conditions clearly do affect the correlations. With these controls it appears that a predictable association between the degree of cooling and the degree of social change can be established.

Conclusion

It seems clear that the start of the LALIA is associated with a significant increase in social change and the intensity of local cooling is generally associated with the intensity of local social change. But within these general relationships there is significant local variation, variation that can be understood by controlling for mediating variables. The mediating factors point to an important element that must be addressed in any consideration of climate-related social change—vulnerability (Wisner et al., 2004; Tierney, 2014). In the context of a naturalistic quasi-experiment this is because the cases are not matched on key variables. Those key variables can all be understood as aspects of vulnerability (Blong et al., 2018). It makes sense that atmospheric dust would be one aspect of differential vulnerability, as all 20 cases were selected in part to be located in unique environmental contexts. But there other are important environmental, technological, and socio-political vulnerabilities that also mediate social stability and change in the context of climate-related hazards (Peregrine 2018a, 2018b). Understanding these vulnerabilities is key to developing social and political structures resilient to the growing climate crisis (Hegmon and Peeples, 2018).

The results presented here indicate that the start of the LALIA marks a time where dramatic climatic change is empirically linked to dramatic social change. It is therefore a context rich for studies that could prove essential to our future. The societies that experienced the LALIA had to struggle through a period of dramatic climatic change just as societies are beginning to struggle today. Some of those societies persisted and even flourished, while some collapsed. The central question we need to address is what conditions—social, political, technological, or other—provided for survival (Comfort et al., 2010; Moreland, 2018). The LALIA provides a naturalistic quasi-experiment through which we can empirically identify causal factors in social stability and change under the pressures of a dramatic temperature shift. Such research is of vital importance today as climate change threatens to impact the world as dramatically as the 536/540/547CE volcanic eruptions did, and for a much longer period of time (Collins et al., 2013). Understanding how societies adjusted to the onset of the LALIA provides the opportunity to empirically identify both effective and detrimental strategies of adapting to such change.

Supplemental Material

Supplementary_Materials – Supplemental material for Climate and social change at the start of the Late Antique Little Ice Age

Supplemental material, Supplementary_Materials for Climate and social change at the start of the Late Antique Little Ice Age by Peter N Peregrine in The Holocene

Footnotes

Acknowledgements

The author wishes to thank Raphael Neukom, Joel Gunn, Felix Reide, and Payson Sheets for their many stimulating discussions, suggestions, and insights about the LALIA and its aftermath. Research assistant Joe Kortenhof worked tirelessly with the author coding cases and checking sources and has earned the author’s respect and gratitude. The author also thanks Carol Ember, Marilyn Hentz, and the entire staff at HRAF for their ongoing support. Finally, the author thanks the staff at DataChemist, and especially Kevin Feeney, for their development of the Dacura interface and their ongoing assistance with data input and management.

Data Availability Statement

The project codebook, an Excel file containing all codes and supporting evidence, the full coded dataset in csv and SPSS format, the extracted and analyzed temperature dataset in csv and SPSS format, and an ArcMap package containing data for all the maps presented in this paper are archived at the HRAF Advanced Research Center at Yale University ().

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by the National Science Foundation (Award # SMA-1416651) and the Army Research Office (Contract Number W911NF-17-1-0441). The views and conclusions contained in this document are those of the author and should not be interpreted as representing the official policies, either expressed or implied, of the Army Research Office or the U.S. Government. The U.S. Government is authorized to reproduce and distribute reprints for Government purposes notwithstanding any copyright notation herein. Both awards were administered through the HRAF Advanced Research Center (hrafARC) at Yale University.

ORCID iD

Peter N Peregrine

Supplemental material

Supplemental material for this article is available online.

References

Blong

Riede

Chen

(2018) A fuzzy logic methodology for assessing the resilience of past communities to tephra fall: A Laacker See eruption 13,000 year BP case. Volcanica 1(1): 63–84.

Büntgen

Myglan

Ljungqvist

, et al. (2016) Cooling and societal change during the Late Antique Little Ice Age from 536 to around 660 AD. Nature Geoscience 9(3): 231–236.

Churkina

Running

(1998) Contrasting climate controls on the estimated productivity of global terrestrial biomes. Ecosystems 1(2): 206–215.

Collins

Knutti

Arblaster

, et al. (2013) Long-term climate change: Projections, commitments, and irreversibility. In: Stocker

Qin

Plattner

G.-K

, et al. (eds) Climate Change 2013: The Physical Science Basis. Cambridge: Cambridge University Press.

Comfort

Boin

Demchack

(2010) Designing Resilience: Preparing for Extreme Events. Pittsburgh, PA: University of Pittsburgh Press.

deMenocal

(2001). Cultural responses to climate change during the Late-Holocene. Science 292(5517): 667–673.

Diamond

Robinson

(2010) Natural Experiments of History. Cambridge, MA: Belknap Press.

Dow

(2007) Galton’s Problem as multiple network autocorrelation effects: Cultural trait transmission and ecological constraint. Cross-Cultural Research 41(4): 336–363.

Dunning

(2008) Improving causal inference: Strengths and limitations of natural experiments. Political Research Quarterly 61(2): 282–293.

10.

Dunning

(2012) Natural Experiments in the Social Sciences. Cambridge: Cambridge University Press.

11.

Dull

Southon

Kutterolf

, et al. (2019) Radiocarbon and geologic evidence reveal Ilopango volcanic eruption as source of the colossal ‘mystery’ eruption of 539/540 CE. Quaternary Science Reviews 222: 105855.

12.

Ember

(2000) Cross-Cultural Research Methods. Lanham, MD: AltaMira.

13.

Gunn

(ed) (2000) The Years Without Summer: Tracing A.D. 536 and its Aftermath. Oxford: BAR International Series 872.

14.

Hegmon

Peeples

(2018) The human experience of social transformation: Insights from comparative archaeology. PLoS One 13(11): e0208060.

15.

Helama

Arppe

Uusitalo

, et al. (2018) Volcanic dust veils from sixth century tree-ring isotopes linked to reduced irradiance, primary production and human health. Nature Scientific Reports 8(1):1339.

16.

Helama

Jones

Briffa

(2016) Dark ages cold period: A literature review and directions for future research. The Holocene 27(10): 1600–1606.

17.

IIASA-FAO (2012) Global Agro-Ecological Zones (GAEZ v3.0). Available at: http://www.fao.org/nr/gaez/about-data-portal/en/

18.

Keys

(1999) Catastrophe. New York: Ballantine.

19.

Leatherdale

(2019) Natural experiment methodology for research: A review of how different methods can support real-world research. International Journal of Social Research Methodology 22(1): 19–35.

20.

McCormick

Büntgen

Cane

, et al. (2012) Climate change during and after the Roman Empire: Reconstructing the past from scientific and historical evidence. Journal of Interdisciplinary History 43(2): 169–220.

21.

Moreland

(2018) AD536—Back to nature? Acta Archaeologica 89(1): 91–111.

22.

Murdock

Provost

(1973) The measurement of cultural complexity. Ethnology 12(4): 379–392.

23.

Neukom

Steiger

Gomez-Navarro

, et al. (2019) No evidence for globally coherent warm and cold periods over the preindustrial Common Era. Nature 571(7766): 550–554.

24.

Nooren

Hoek

van der Plicht

, et al. (2017) Explosive eruption of El Chichćon volcano (Mexico) disrupted 6th century Maya civilization and contributed to global cooling. Geology 45(2): 175–178.

25.

PAGES 2k Consortium (2017) A global multiproxy database for temperature reconstructions of the Common Era. Nature Scientific Data 4(1): 170088.

26.

Peregrine

(2003) Atlas of cultural evolution. World Cultures 14(1): 2–88.

27.

Peregrine

(2018a) Social resilience to climate-related disasters in ancient societies: A test of two hypotheses. Weather, Climate, and Society 10(1): 145–161.

28.

Peregrine

(2018b) Reducing post-disaster conflict: A cross-cultural test of four hypotheses using archaeological data. Environmental Hazards 18(2): 93–110.

29.

Peregrine

Brennan

Currie

, et al. (2018) Dacura: A new solution to data harvesting and knowledge extraction for the historical sciences. Historical Methods 51(3): 165–174.

30.

Riede

(2014) Towards a science of past disasters. Natural Hazards 71(1): 335–362.

31.

Sigl

Winstrup

McConnell

, et al. (2015) Timing and climate forcing of volcanic eruptions for the past 2500 years. Nature 523(7562): 543–549.

32.

Slingerland

Monroe

Sullivan

, et al. (2019) Historians respond to Whitehouse et al. (2019), “Complex societies precede moralizing gods throughout world history.” Journal of Cognitive Historiography 5(1–2) [preprint accessed at PsyArXiv: /doi.org/10.31234/osf.io/2amjz]

33.

Smith

Peregrine

(2012) Approaches to comparative analysis in archaeology. In: Smith

(ed.) The Comparative Archaeology of Complex Societies. Cambridge: Cambridge University Press, pp. 4–20.

34.

Tierney

(2014) The Social Roots of Risk: Producing Disasters, Promoting Resilience. Stanford: Stanford University Press.

35.

Toohey

Krüger

Sigl

, et al. (2016) Climatic and societal impacts of a volcanic double event at the dawn of the Middle Ages. Climatic Change 136(3–4): 401–412.

36.

Turchin

Currie

Whitehouse

, et al. (2018) Quantitative historical analysis uncovers a single dimension of complexity that structures global variation in human social organization. Proceedings of the National Academy of Sciences 115(2): E144–E151.

37.

Turchin

Whitehouse

François

, et al. (2020) An introduction to Seshat: Global history databank. Journal of Cognitive Historiography. DOI: 10.1558/jch.39395.

38.

Wisner

Blaikie

Cannon

, et al. (2004) At Risk: Natural Hazards, People’s Vulnerability and Disasters. London: Routledge.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.64 MB