Escape From Rome

by Walter Scheidel

  14. Jackson 2005: 69 (Esztergom); Pow 2012: 70 (Silesia and Moravia); Jackson 2005: 65 (Croatia); Pow 2012: 75 (Serbia and Bulgaria).

  15. Chambers 1979: 164–65; Pow 2012: 76–77.

  16. Pow 2012: 46. France’s 1999: 234 verdict that the Mongol retreat owed nothing to European military prowess is correct as far as battles are concerned but neglects the role of fortifications, which he himself discusses at length (77–106).

  17. Sinor 1972: 181–82; Sinor 1999: 19–20; and, in general, Lindner 1981. Criticism in Pow 2012: 24–34. For 1918, see Sinor 1972: 181. Total Mongol forces probably numbered a little over 100,000 in the early thirteenth century, which is consistent with these figures: Sverdrup 2017: 110.

  18. Estimates vary. See Lindner 1981: 14–15 for 42,000 km2 of grassland in Hungary and a maximum of 320,000 horses, or closer to half that in real life, and a 1:10 man:horse ratio for Mongols (i.e., 16,000–32,000 men). Sinor 1999: 19–20 allows for a theoretical maximum of 415,000 horses and 83,000 men at a 1:5 ratio (superseding 1972: 181–82, for 205,000 horses and 68,000 men). Man 2014: 144 estimates that 24,000 square kilometers of grassland could have supported 150,000 men with 600,000 horses.

  19. Lindner 1981, for this adaptation among Huns, Avars, and Magyars, a thesis that has stood up reasonably well to later criticism. Dynamics: see chapter 8 in this volume.

  20. Mongols in unsuitable terrain: Pow 2012: 33–34. For comparative analysis of how Mongols dealt with fortifications elsewhere, see Pow 2012: 79–121, emphasizing their use of artillery against mud and brick walls and the role of collaboration and defection (120–21). Russian craftsmen: Hartog 1996: 163. See Allsen 2002: 268–71, 278–80 for the use of Chinese gunpowder weaponry (rockets) during the conquest of Central Asia and Iran (1220s–1250s) and the subsequent transfer of counterweighted catapults from Syria to China to assist in the final stages of the conquest of Song China in the 1270s.

  21. May 2012: 67–80 gives a summary of fragmentation and hostilities between the khanates in the late thirteenth century.

  22. For the importance of alternative targets, see, e.g., Jackson 2005: 74; Pow 2012: 121.

  23. See Jackson 2005: 113–34 for the halting of Mongol advances.

  24. For these later attacks, see Chambers 1979: 164–65; Jackson 2005: 123, 203–6, 210, 212.

  25. Pow 2012: 122 and Sverdrup 2017: 327 share this view.

  26. See chapter 8 in this volume.

  27. For Otto I, see chapter 5 in this volume.

  28. Allsen 1987: 207–9; Hartog 1996: 54–56 (census in Russia), 55–56 (tax), 162–67 (influence); Ostrowski 1998: 133–248 (ideology).

  29. Holland 1999: 102 (quotes), 104. See also Parker and Tetlock 2006: 381 for a consequent halt to “Western” development. Cf. Hartog 1996: 166 for the Mongol invasion’s initially disastrous effect on the Russian economy, which was later partly offset by the expansion of long-distance trade.

  30. Blame: cf. Pow 2012: 127–33. See Allsen 1987: 190–94 for mobilization practices, esp. 191 on drafting for sieges. It is not explicitly recorded that Russians provided troops, though it is highly likely (209n73). Pow 2012: 133 employs this scenario as a counterfactual for Latin Europe.

  31. See chapter 8 in this volume.

  32. Watts 2009 offers a recent survey.

  33. Blockmans 2002: 36.

  34. Pagden 2001: 43–44; also Blockmans 2002: 25–26 for the intensity of fragmentation.

  35. Empire: König 2002: 200–209; Pagden 2001: 43–44. Christendom: König 2002: 199–200, 209–12; Luttenberger 2002.

  36. P. Wilson 2016: 172–73, 422–82, esp. 423, 436, 448.

  37. Reinhard 2002.

  38. Babel 2002; Durchhardt 2002; Tracy 2002a; Kleinschmidt 2004. For the contribution of Ottoman pressure on the success of the Reformation, see Iyigun 2008.

  39. Pietschmann 2002: 537–38; H. Thomas 2010: 315–36.

  40. Brendle 2002, on the opposition of the German estates to Charles V.

  41. Brendle 2002: 702–3; Kleinschmidt 2004: 208–10. See Blockmans 2002: 80–99, 108–13 on Charles V and the Reformation.

  42. Kleinschmidt 2007: 225.

  43. Tracy 2002b: 242 (Metz). This estimate is based on the following values: 0.2 grams of silver per liter of barley in Augsburg in 1552 (nine-year average) ~0.3grams of silver per liter of wheat; 1 ducat (3.5gram of gold at 13:1 gold:silver) ~45 grams of silver; 3.25 million ducats ~150 tons of silver ~490 million liters of wheat. For Roman military compensation, see chapter 3 in this volume.

  44. Tracy 2002b: 50–90 (sources of revenue), 91–108 (credit), 109–248 (campaign funding). 248; 2002a: 159 (rising costs). See also Blockmans 2002: 139–68, esp. 154–60 on financial pressures.

  45. Tracy 2002b: 182, 247; cf. 2002a: 158. Kleinschmidt 2004: 146.

  46. Tracy 2002b: 229, 247.

  47. Blockmans 2002: 25–46 gives a pithy overview.

  48. Eire 2006b: 150–56.

  49. Ibid. 159–61, 164–65.

  50. Ibid., 162–65.

  51. Parker 1998, esp. 143, 181, 202, 273–75, and 280 for unfavorable outcomes overall. Dutch Revolt: Black 2002: 107–17. Philip as king of Spain: Kelsey 2012.

  52. For these three counterfactuals, see Parker 1998: 292–94 and 2014: 372–73, and Parker 1999: 144–53 specifically on the events of 1588.

  53. Murphey 1999: 36–49. For the nature of Ottoman rule, see, e.g., Lowry 2003; Barkey 2008.

  54. Ibid., 30–32 on the limits of Ottoman state power and control; Karaman and Pamuk 2010: 603–9 (flat tax take), 609–12 (comparison with Europeans). Population: McEvedy and Jones 1978: 18, 79, 97, 111, 137.

  55. A truce with Persia—envisioned by Parker and Tetlock 2006: 374–75 as a prerequisite for taking Vienna and the downfall of Habsburg power—would hardly have been enough: the more the Ottomans had been involved in Latin Europe, the stronger the incentive for attacks in their rear would have become.

  56. Troop numbers: Karaman and Pamuk 2010: 610, fig. 4 and 612, table 1.

  57. Goldstone 2006: 172–73.

  58. England: Pestana 2006: 199.

  59. Goldstone 2006: 173–78, 184–85 for this scenario’s negative effects on science and technology. Mokyr 2006 disagrees. For these factors, see Part V.

  60. Horne 1999 considers this counterfactual.

  61. For the contrast with East Asia, see chapter 10 in this volume.

  62. Horne 1999: 208, 210–12, 214–16; Roberts 2014: 445, 625.

  CHAPTER SEVEN

  1. See chapter 1 in this volume. For the term “First Great Divergence,” see note 17 below.

  2. Visualized in chapter 1 of this volume.

  3. For parallels, see Gizewski 1994; Scheidel 2009a: 13–15. For the comparative history of Rome and China, see Scheidel 2009d, 2015c; Auyang 2014; and my website, https://web.stanford.edu/~scheidel/acme.htm.

  4. In contrast to Rome’s experience, Qin’s fairly rapid conquest of its main competitors had been preceded by close involvement in several centuries of intensifying competition within an integrated state system not unlike that of post-Roman Europe: see Hui 2005 for a comparative study of why the Chinese Warring States eventually succumbed to a hegemon whereas early modern Europe did not.

  5. By then, just four empires—Rome, Parthia, Kushan, and Han—laid claim to at least two-thirds of the entire world population: the global heyday of empire. For the numbers and later parallels, see Scheidel in press-b.

  6. For convergent evolution and lasting differences, see Scheidel 2009a: 15–20.

  7. M. Lewis 1990 is fundamental. Fukuyama 2011: 110–24 provides a clear outline.

  8. Scheidel 2009a: 16–17 (general); Tan 2017 (taxes). For warfare, see Rosenstein 2009. Asymmetries: see chapter 3 in this volume.

  9. Scheidel 2009a: 17–19. For a comparison of administrative structures, see Eich 2015 and Zhao 2015b.

  10. Bureaucracies: see previous note, and for budgeting see Scheidel 2015d. Urban development: M. Lewis 2015; Noreña 2015; Scheidel 2015b: 9–10; and Noreñ
a forthcoming.

  11. See chapter 1 in this volume. Dincecco and Onorato 2018: 3, 5, 20–21 emphasize the role of the fall of the Carolingian empire.

  12. Updated from Hui 2015: 13, for the area controlled by the Qin at the height of their power in 214 BCE, dubbed the “interior.”

  13. Kaufman, Little, and Wohlforth 2007: 231, table 10.1 calculate that East Asia was characterized by a unipolar/hegemonic system for 68 percent of the years between 220 BCE and 1875 (adjusted from their own virtually even split for the longer period from 1025 BCE to 1875), and South Asia for half of the time between 400 BCE and 1810, compared to 2 percent for Europe from 1500 to 2000.

  14. According to Lorge 2005: 181, “The period between 907 and 979 was the last moment of a real multi-state system before the chaotic warlord period in the early twentieth century.”

  15. Sinocentric order: M. Lewis 2009b: 146, 153.

  16. The quote is taken from the title of Rossabi 1983, whose editor claims that “a true multi-state system operated during Sung times” (11). This notion is common in modern scholarship on East Asia from the Tang period onward. Contrast Ringmar 2007: 289, who sensibly notes the lack of a balancing state system in China: imperial hegemony was the only game in town. Cf. also Lang 1997: 26, for the lack of serious state-level competitors of China-based empires. Rosenthal and Wong 2011: 22 stress, contra Rossabi 1983, the differences between the Song era and European multistate systems. The same reservations apply to the strawman position taken by Goldstone 2009: 100 that “it is a misleading oversimplification to say that from 1500 onward Europe had a system of competing states whereas Asia had empires without competition” and “it is wrong to see the major empires as always dominating their neighbors and being free of military competition.” These statements are both correct and irrelevant in that they misrepresent most existing scholarship. Cf. Hoffman 2015: 175–78 for the counterfactual of lasting divisions in China had the Mongols failed to take over the Southern Song, and the possible developmental consequences of a resultant genuine state system.

  17. I initially used the term “First Great Divergence” in Scheidel 2009a, echoing the title of Pomeranz 2000 (which contrary to claims on the internet is not indebted to Huntington 1996). For the concept of the “Little Divergence,” see van Zanden 2009a, referred to as a “First Great Divergence” by Karayalcin 2016. See also Davids 2013, an entire book devoted to “great and little divergences.” Moore 2009: 577 introduced the notion of an eleventh/twelfth century CE “First Great Divergence,” and elaborates on this in Moore 2015 (with a gracious reference (20) to my own use of this term). For the deemphasizing of extended kinship networks in northwestern Europe, cf. chapter 12 in this volume.

  18. Cicero, In Defense of the Manilian Law 7.17 (66 BCE); Schumpeter 1954 [1918]: 6 (from Goldscheid 1917); All the President’s Men (Warner Bros., 1976); Tilly 1992: 96–97.

  19. For the history of taxation in the long term, see Ardant 1971–1972; Bonney 1995, 1999; Cavaciocchi 2008; Yun-Casallilla and O’Brien 2012; Monson and Scheidel 2015a. For evolutionary taxonomy, see esp. Bonney and Ormrod 1999; for theory, Levi 1988; for a model of variation in contractual forms of revenue collection, Cosgel and Miceli 2009. Articulation: Motyl 2001: 13–20, for a spoke model of empire.

  20. For repeated empirical demonstration of these basic trends, see the contributions in Monson and Scheidel 2015a, and esp. Monson and Scheidel 2015b: 19–20. See also Wickham 1994: 43–75 and Haldon in press. Haldon 1993 provides the most extensive discussion.

  21. See chapters 2, 5, and 6 in this volume.

  22. For conditions during the reconsolidation phase, see chapter 10 in this volume.

  23. For a comparison of the fiscal regimes of the Roman empire and Han China, see Scheidel 2015d. For the late Roman tightening of the earlier more-relaxed regime (Scheidel 2015a), see Bransbourg 2015.

  24. Institutional decentralization was a function of the paramount importance of the self-governing city as the focal point not just of taxation but of society in general, and of the involvement of rent-seeking local elites in state revenue collection: Wickham 1994: 50, 73–74. Tax remissions: see, e.g., Bransbourg 2015: 275–76.

  25. Land distributions: Wickham 2005: 84–86. Initial tax shares: Goffart 1980 (228–29 on the eventual conversion to land assignments); Goffart 2006: 119–86 (quotes: 183). Cf. also Halsall 2003: 42–43. Tedesco 2013 references recent scholarship.

  26. Halsall 2003: 64; Wickham 2005: 92–93, 115–20; and Tedesco 2015: 104–12, 166–87 (Italy); Wickham 2005: 87–93; and Modéran 2014: 167 (Vandals).

  27. Wickham 2005: 93–102; and see chapter 5 in this volume.

  28. Halsall 2003: 54; Wickham 2005: 102–15.

  29. Wickham 2009: 553–54. See also Wickham 2005: 146–48. For late imperial China, see chapter 10 in this volume.

  30. Wickham 2005: 84 (state size and military status), 149 (localization), and also 830 for the “crucial importance of the end of Roman imperial unity.”

  In this context, the cultural dimension of the observed shift toward decentralized support for followers that was based on direct access to land yields is also worth considering: we must ask to what extent it might reflect compromises between rulers whose position had grown out of leadership of diverse and mobile warrior confederations and the expectations of their members. Here one must tread carefully: this approach does not merely entail conjecture but also invites charges of essentializing traits of the successor regimes by an academic tradition that has increasingly sought to deconstruct the very notion of “Germanic.” Even so, while acknowledging the considerable diversity and fluidity of these conquest groups, we should be open to the possibility that they prioritized certain types of rewards: at the very least, this notion deserves an outsized endnote. (Wickham 2005: 82–83 rejects the notion that cultural differences between old and new ruling classes were responsible for the observed fiscal changes: but more deeply rooted factors might have played a role. Wickham 2016: 29 notes that senior Germanic officers preferred land to pay.)

  The tradition that ancient Germans (or “Germans”) sought farmland goes back a long way. When Caesar, the highly indebted and wildly ambitious governor of Provence in 58 BCE, was in need of an excuse to intervene in free Gaul, he found one (of several) in the encroachment of Germans from across the Rhine on the Gallic Sequani in what is now eastern France: “Ariovistus, the king of the Germans, had settled in their territory and seized one-third of their land” (Caesar, War in Gaul 1.31).

  Taken on its own, it would surely be unwise to put much weight on his self-serving account that cast Romans as defenders of soon-to-be-conquered Gauls. But this was part of a much larger pattern. More than half a millennium later, in the dying days of what was left of the western Roman empire, unruly Germanic mercenaries put pressure on the renegade general Orestes, who had just placed his teenage son Romulus (nicknamed Augustulus) on the throne: they “demanded that they [the Romans] should divide with them the entire land of Italy; and indeed they commanded Orestes to give them the third part of this, and when he would be no means agree to do so, they killed him immediately. Now there was a certain man among the Romans named Odoacer … and he at that agreed to carry out their commands … giving the third part of the land to the barbarians” (Procopius, Gothic Wars 5.1.3). While we do not know whether this scheme was actually implemented, the fact that Odoacer managed to hold on to Italy for the next thirteen years until the Ostrogoths invaded and overthrew him and his followers suggests that he did something right: some form of land grants might well have been part of it. Goffart 1980: 62–70 questions Procopius’s credibility. However, the Eugippius quote in the main text, referring to Odoacer’s great gifts to many, might reflect an actual land distribution.

  This would fit well with what had repeatedly happened during the intervening centuries: on dozens of occasions, “barbarians”—mostly of Germanic origin—had been settled within the confines of the Roman empire, beginning with the resettlement of the Ubii in Gaul in
38 BCE. Sometimes they were accommodated once defeated and at other times simply received as settlers. Land features as an explicit objective: in the 160s CE, Germanic groups supposedly threatened war “unless they were accepted” (Historia Augusta, Life of Marcus 14.1). Self-governing settler communities are recorded from the 380s onward and became the norm in the early fifth century, thereby precipitating the demise of the western empire. Ste. Croix 1981: 509–18 lists cases from 38 BCE to the 590s CE.

  The inhabitants of free Germany were primarily farmers, and many groups that later arrived in former Roman provinces were linked to populations that had long led a settled lifestyle. In view of this, purely functionalist explanations of the later abatement of taxation might be too narrow: the characteristics of the conquest groups and of consensus-building between leaders and followers may also have played a role. The Germanic experience invites comparison with the contrasting preferences of invaders from less settled areas north of China or Arabia, who for the most part did not seem to share this interest in taking over land directly but preferred to be compensated by means of a centralized tax system, keeping primary producers at greater remove.