Overall, the charts provided in chapter 1 serve as a means to a very specific end, which is to establish and compare long-term patterns in imperial state formation in different parts of the Old World. For them to be substantively misleading, they would have to be wrong in ways that seem incompatible with the historical record. Nonetheless, I advise readers not to use these tentative reconstructions in ways that go too far beyond this limited objective: they are based on controlled conjecture, and are not to be mistaken for historical “facts.”8
NOTES
INTRODUCTION
1. The simile of an “escape” has been employed before, most notably by Fogel 2004, Deaton 2013, and Vries 2013. I use it here in part to continue the cinematic theme of this book’s motto; cf. also Deaton 2013: 2–3 for a reference to the 1963 movie The Great Escape.
2. For notable premodern growth phases, see, e.g., Goldstone 2002 (cross-cultural); Saller 2002 (ancient Rome); Scheidel, Morris, and Saller 2007 (ancient Greco-Roman world); Jursa 2010 (ancient Babylonia); W. Liu 2015a (Song China); Ober 2015 (ancient Greece). For the underlying methodology of the Maddison Project Database 2018, see most recently Bolt et al. 2018.
3. In terms of the world’s economic center of gravity, the West’s preeminence peaked around 1950, see https://www.economist.com/graphic-detail/2018/10/27/the-chinese-century-is-well-under-way for a visualization.
4. See Bourguignon and Morrisson 2002.
5. I. Morris 2013b, expanding on I. Morris 2010. Eastern Eurasia shows a similar upturn after a time lag. Morris’s scoring (not shown here in full) reaches back to the beginning of the Holocene.
6. I. Morris 2013b: 239–48 defends the broad validity even of earlier variations by discussing margins of error. Quote: Landes 2003: 5.
7. I. Morris 2013b: 53–143 (energy capture), 249, table 7.3 (role of energy capture), 63–65 (eighteenth and nineteenth centuries); see 66, fig. 3.4 for a visual. Energy capture refers to energy flows to humans from food (consumed by humans or by animals used or consumed by humans), fuel (including wind and water power) and raw materials (I. Morris 2013b: 53). Lower estimates of premodern energy capture in other studies are to some extent the result of narrower definitions (76–80). Even though Morris appears to underestimate the contribution of coal to the nineteenth-century expansion (compare I. Morris 2013b: 63 to Warde 2007: 117–20 and Warde 2013: 133, table 5.1), his totals are sufficiently similar to other estimates to support their inclusion in his development index.
8. Warde 2007: 127–28, accepted by Wrigley 2016: 34. See also Warde 2013: 134, fig. 5.2 for an analogous tripling of per capita energy consumption in a larger sample of Western and Southern European countries. For England’s pioneering role in the fossil fuel transition, see, e.g., Warde 2013: 131–41 and chapters 10 and 12 in this volume. For the transformative shift from organic and fossil fuel economies, see most recently Wrigley 2016: 1–3.
9. On life expectancy, see Norberg 2017: 43; Pinker 2018: 54. See Roser 2019 for more detail. See Deaton 2013: 29–41 for the link between GDP and life expectancy. On global per capita GDP growth, see Maddison 2010. On poverty and malnourishment, see Norberg 2017: 9, 20, 65, 76.
10. Floud et al. 2011: 69, table 2.5 (heights), 364 (quote).
11. On literacy, see Norberg 2017: 131–33; Roser 2018. On freedom, see Pinker 2018: 202–3; Roser 2018. Fifty-six percent of the world’s population currently lives in democracies. On happiness and “life evaluation,” see Deaton 2013: 18, 21, 53; Pinker 2018: 269. On global economic inequality trends, see Milanovic 2016: 118–32. On resilience of within-society economic inequality, see Scheidel 2017: 405–23.
12. For a taste of the growing literature on the risks the future holds, see Rees 2003, 2018, and, more specifically, http://www.ipcc.ch/, https://www.globalchange.gov/, and http://www.lancetcountdown.org/ for the fossil-fuel economy’s impact on the climate and human welfare.
13. McCloskey 2010: 125–384 (with 2016: 83–146) and Vries 2013: 153–438 are the most comprehensive (and strident) surveys-cum-critiques. Daly 2015 offers a scrupulously neutral survey of the main approaches, primarily aimed at students.
14. When empire once again emanated from Europe, from the sixteenth century onward, it did so by exporting it to an increasingly global periphery while coexisting with and arguably even reinforcing fragmentation in Europe itself. Russia’s initial lack of demographic heft long offset its massive spatial expansion. Thus, neither colonial empire nor Russia (which Lieven 2000 and Burbank and Cooper 2010: 185–99 put into comparative context) represent deviations from this pattern.
15. This observation cannot be undone simply by calling ancient Rome a “Mediterranean” rather than a “European” empire (e.g., Gat 2006: 392–93): Rome ruled most Europeans at the time, well more than half of the empire’s population was located in Europe, and the empire controlled almost all of Western Europe, which is at the center of this book’s attention.
16. For the nomenclature, see chapter 7 in this volume.
17. I refer to empire within—above all, Western—Europe. Colonial empires overseas were extensions of competing European polities.
18. See chapter 10 in this volume for “long-term” and “short-term” explanations of modern economic development and for the “Little Divergence,” and see chapter 7 for definitions of various “divergences.”
19. With no claim to completeness: among the chief proponents are J. Hall 1985: 111–44 and J. Hall 1988: 33–38; Landes 1998: 37–39; E. Jones 2003: 104–26, 225, 233; Mokyr 2007: 23–26 and Mokyr 2017: 165–78; Cosandey 2008: 175–316; van Zanden 2009a: 32–68, 295; Rosenthal and Wong 2011, e.g., 124–25, 145, 198–99, 208–10, 225–30; Vries 2013: 379–82, 434–35; Vries 2015, e.g., 363, 430. Among the many others, see, for various aspects, Wallerstein 1974: 60, 63; Baechler 1975: 74–77; Mann 1986: 376; Macfarlane 1988: 191; Diamond 1997: 416; Wong 1997: 148; Abernethy 2000: 206–24; Crone 2003: 161–63; Chaudhry and Garner 2006; Mann 2006: 383; Arrighi 2007: 315; Ringmar 2007: 289; Karayalcin 2008: 977; Chu 2010; Voigtländer and Voth 2013a, 2013b; Hoffman 2015, esp. 213; McCloskey 2016: 396–400; Cox 2017: 746–47; Roeck 2017: 22; Dincecco and Onorato 2018; Ko, Koyama, and Sng 2018: 310–11; and cf. also Mielants 2007: 154–62 and Stasavage 2011: 47–109 for the role of city-states and small polities.
20. I develop this approach in chapter 1.
21. Bang, Bayly, and Scheidel (in press) offer a more systematic survey than the very substantial collection of Rollinger and Gehler 2014. For the nature of empire and the debate, see the introductory chapter by Bang (in press), and notably also Haldon (in press), with references. For partial syntheses and/or theory, see Eisenstadt 1963; Kautsky 1982; Doyle 1986; Lieven 2000; Motyl 2001; Wood 2003; Chua 2007; Münkler 2007; Darwin 2008; Burbank and Cooper 2010; Leitner 2011. See also, among collaborative works, Tracy 1990, 1991 (early modern merchant empires); Alcock et al. 2001 (eclectic); Morris and Scheidel 2009 (antiquity); Scheidel 2009d, 2015c (on the two largest ancient empires, Rome and Han China).
22. Not on its own: Mokyr 2009: 83–84. Quote: Landes 2003: 12. See also Vries 2013: 22–27, for the breadth of the process but also the key role of industrialization.
23. As Wickham rightly observes regarding the survival of states (and not just large empires, but state structures as such), “It is survival that is the norm, failure that is the deviation” (1994: 74), and that this poses a challenge to scholars of post-Roman Western Europe. I hope to have reduced this challenge.
24. Quotes: Johann Wolfgang von Goethe in Italienische Reise (ch. 24, December 3, 1786, on the city of Rome); Gibbon 1788b: 645.
25. Paired by Campbell 2004: 167. The inscription appears as no. 138 in Sartre 1993, and Virgil’s quote is from the Aeneid 1.279.
26. “Getting to Denmark” is a metaphor for establishing political and economic institutions that are highly conducive to human welfare, a concept that goes back to Pritchett and Woolcock (2002: 4) and that has since been popularized especially by Fukuyama 2011: 14. Metaphorically, the
escape from Rome was also an escape from China, and indeed from rule by any large agrarian empire.
27. This imbalance was already noted more casually by Klein 2017: 302. My sample, with due apologies for unconscionable omissions and idiosyncrasies: Economists (19): Daron Acemoglu, Robert Allen, Stephen Broadberry, Kent Deng, Ronald Findlay, Bishnupriya Gupta, Eric Jones, Cem Karayalcin, Mark Koyama, Timur Kuran, Deirdre McCloskey, Joel Mokyr, Patrick O’Brien, Kevin O’Rourke, Jean-Laurent Rosenthal, Jared Rubin, Jan Luiten van Zanden, Nico Voigtländer, and Hans-Joachim Voth. Sociologists (9): Jean Baechler, Joseph Bryant, Ricardo Duchesne, Jack Goldstone, John Hall, Toby Huff, Eric Mielants, Immanuel Wallerstein, and Dingxin Zhao. Historians (8): Philip Huang, Margaret Jacob, David Landes, Michael Mitterauer, Prasannan Parthasarathi, Kenneth Pomeranz, Peer Vries, and Bin Wong. Political science is only marginally represented by Gary Cox, Erik Ringmar, and David Stasavage, and the list is rounded off by the physicist David Cosandey, and Ian Morris, a classical archaeologist turned world historian. One could have added Joseph Needham, trained as a biochemist. If one wanted—even more idiosyncratically and thus anonymously—to single out the dozen most influential scholars in this sample, ratios would change only a little, with two-thirds of the putative top spots held (mostly) by economists and a couple of sociologists, and the remainder by historians. What is even more striking, of course, is that 95 percent of the full sample is (now) male: this is not by any means an inclusive domain of research.
28. Fowden 2011: 172 (quote). See 168 for the “first millennium” (CE) as a meaningful unit of analysis, a notion that is borne out in my book: all of the critical events—the fall of the Roman empire and the inexorable early medieval dismantling of its institutions, as well as the revival of empire in China (Sui-Tang and Song) and elsewhere in the Old World—took place in precisely that period.
29. Quote: Kocka 2009: 15. For comparison in history and cognate social sciences more generally, see Bonnell 1980; Skocpol and Somers 1980; Tilly 1984; Ragin 1987; Goldstone 1991: 50–62; Haupt and Kocka 1996a (esp. Haupt and Kocka 1996b; Osterhammel 1996); Mahoney and Rueschemeyer 2003 (esp. Mahoney 2003; Rueschemeyer 2003); Kocka 2009. There does not seem to be an introduction specifically to comparative history: for now, Lange 2013 offers the closest approximation. Specifically for comparative ancient history, see Mutschler and Scheidel 2017; Scheidel 2018.
30. Quotes: Lloyd and Sivin 2002: 8; Rueschemeyer 2003: 332.
31. Quote: Kocka 2009: 15.
32. Quote: Kocka 2009: 15.
33. Quote: Kocka 2009: 17.
34. Weber 1920–1921.
35. For analytical comparisons, see Bonnell 1980: 164–65.
36. For these concepts, see Skocpol and Somers 1980: 175–78; Ragin 1987: 17.
37. Tetlock and Parker 2006: 15 (quotes) and 17–28 for a forceful statement of this position. See also Demandt 2011: 51 for the same point. For counterfactual history, see Hawthorn 1991; Tetlock and Belkin 1996; Ferguson 1997; Cowley 1999, 2001; Tetlock, Lebow, and Parker 2006b; Demandt 2011.
38. Tetlock and Parker 2006: 18–19 for these two approaches; Demandt 2011: 28–31 for causality.
39. Tetlock and Parker 2006: 33–36 on procedure, esp. 34 (quote). See also Tetlock and Belkin 1996: 1–31 on criteria for judging the quality of counterfactuals, esp. 23–27 on consistency with historical facts and theoretical laws.
40. Imperative: Tetlock and Belkin 1996: 19–23. Problems: Tetlock and Belkin 1996: 21; Tetlock, Lebow, and Parker 2006a: 10; Tetlock and Parker 2006: 34–35.
41. Junctures: Tetlock and Belkin 1996: 7–8; Tetlock and Parker 2006: 33–34.
42. Tetlock and Parker 2006: 35–36 consider this risk the greatest challenge of counterfactual reasoning.
43. Tetlock, Lebow, and Parker 2006a: 10–11 (change over time); Tetlock and Parker 2006: 21 (question).
44. “The reader should remember that to describe a phenomenon is not to praise it” (Strayer 1970: xxviii, on the creation and strengthening of European states). The same applies to interstate competition (a euphemism for pointless warfare and consumer-unfriendly protectionism), religious strife, colonization, plantation slavery, and so much more.
Not least because of its wide scope, this is not designed to be a book of overt controversy. I note disagreements rather than seek to arbitrate them as long as they do not materially affect my argument, and I focus on the plausible and on positive contributions as I try to build my case. On rare occasions when I do engage, I tend to do so in dialogue with fellow Princeton University Press authors, in what I hope will be understood as a gesture of collegial respect: e.g., Beckwith 2009; Pines 2012; Hoffman 2015.
CHAPTER ONE
1. I draw here on Scheidel in press-b.
2. Taagepera 1978a, 1978b, 1979, 1997; Turchin 2009; Scheidel in press-b.
3. Translated from Braudel 1966: 326. For an analysis of cost constraints on the functioning of the Roman empire, see Scheidel 2014, based on http://stanford.orbis. edu.
4. For the “shadow” empires of the steppe, see Barfield 2001 and chapter 8 in this volume.
5. The “most populous polity” of a given region is defined as the polity with the largest number of people who were located within that region, rather than as the polity with the largest total population that extended into that region. In practice there is hardly any difference between these two criteria. A few exceptions occurred in the Middle East and North Africa region, where the Roman empire would have been the most populous power overall in 150 and 100 BCE and the Mongol empire in 1250; in Europe, where the Mongols were the most populous power overall in 1250; and in the “Roman empire region,” where the most populous powers overall were the Seleucid empire in 300, 250, and 200 BCE and the Abbasid caliphate in 800 and 850 CE.
6. For Chinese censuses, see Bielenstein 1987. I have firsthand experience with ancient Roman population numbers: see esp. Scheidel 2001: 181–250 (Roman Egypt) and 2008b (Italy). Uncertainties are particularly extreme when it comes to the Pre-Columbian and early colonial Americas: see, for example, Henige 1998; McCaa 2000; and the technical note to this chapter.
7. I include Afghanistan in the MENA region because from the perspective of historical state formation that is where it (most) belongs. The other regions are defined in keeping with standard conventions.
8. According to the figures in McEvedy and Jones 1978, these regions accounted for approximately 92 percent of the global population in 1 CE, 83 percent in 1000, and 81 percent in 1500. Their share in 2012–2013 amounted to 62 percent (or 4.45 billion out of 7.1–7.2 billion).
9. The former Spanish Sahara that is currently claimed by Morocco is omitted from these tallies. Omission of the Algerian and Libyan deserts reduces the total to roughly 9.5 million square kilometers, and further adjustments for outlying portions of western Egypt and the Rub’ al Khali in Saudi Arabia would lower it to 8–8.5 million square kilometers.
10. The rather severe fluctuations in the early second millennium CE are in part a function of McEvedy and Jones’s specific assumptions (see the technical note) and need not be taken at face value. Once again it is only the overall pattern that matters.
11. Here and in the following, the combination of a polity followed by another polity in parentheses refers to a scenario in which the leading empire was closely followed by the runner-up; de facto they might be considered roughly equivalent. If two polities listed for the same year are divided by a slash (/), the demographic estimates do not allow even tentative ranking. For the data, see the technical note. I consider the Roman empire in 400 CE to be united even though it effectively (though not formally) split up five years earlier, and I treat the medieval German empire (eventually known as the Holy Roman Empire) as a single polity until the death of Frederick II, a somewhat arbitrary cut-off point but necessary to account for the progressive internal segmentation of this peculiar entity. To be sure, both the German empire and the kingdom of France (the runner-up) were strongly decentralized during this period.
12. Even this may well be a slight underestimate. A Roman share of around 85 percent at its peak in the second century CE is perfectly possible and indeed plausible once we adjust for the underestimate of the Roman imperial population found in McEvedy and Jones 1978 (see the technical note) in keeping with my (fairly conservative) estimates in Scheidel 2007a: 48, table 3.1. Even if we upgrade the non-Roman European population to the levels of 800 CE given by McEvedy and Jones 1978 in order to compensate for some of this adjustment, the Roman share still reaches or exceeds 80 percent in the mid-second-century CE.
13. See the technical note.
Escape From Rome Page 58