Empires and Barbarians

by Peter Heather

  11 See Green (1998), chapter 7; cf. Wolfram (1997), chapter 1; Pohl (2000), 66ff. ‘They chose kings’: Tacitus, Germania 7 (‘reges ex nobilitate duces ex virtute sumunt’).

  12 See Chapter 6 below on the rise of Clovis. Clovis operated on Roman soil, however, which meant that he could support a much larger retinue, whereas a Germanic economic context (see below) would have imposed tighter economic constraints and perhaps made this impossible.

  13 On Chnodomarius’ armour, see Ammianus 16.12.25; we will return to these swords on p. 78.

  14 On Odry, see Kmiecinski (1968). In these eastern areas of Germania, the cemeteries were much more permanent than any settlements in the first two centuries AD, and are marked by large stone circles which contained few if any burials. It has been plausibly suggested that this reflects the fact that cemeteries rather than settlements provided the locus for social gatherings.

  15 The fullest discussion is Haarnagel (1979).

  16 On Wijster, see Van Es (1967). See more generally the relevant studies in Krüger (1976–83): compare vol. 1, chapter 11 with vol. 2, chapter 5; Myhre (1978); Steuer (1982), 258ff.; Hedeager (1988), (1992), 193ff.; Todd (1992), chapter 4. There is a useful discussion of the Roman side of the frontier in Carroll (2001), chapter 4.

  17 Goffart (2006), 26–32 objects to old-style assumptions, based on the famous Jordanes, Getica 4.25, that Scandinavia in particular and Germania in general was a womb of nations, endlessly producing future invaders of the Roman Empire until it was overwhelmed. As a comment on old-fashioned historiography, this is fair enough, though his work does not engage with the detailed archaeological evidence.

  18 See Urbanczyk (1997b).

  19 On the Pietroasa treasure, see Harhoiu (1977). On fibula production at the Runder Berg (see note 24 below), see Christlein (1978), 43–7, 171. On pottery, see Heather and Matthews (1991), chapter 3 (Cernjachov); Drinkwater (2007), 89–93; cf., more generally, Krüger (1976–83), vol. 2, 123ff.

  20 On glass, see Rau (1972). On combs, see Palade (1966).

  21 The groundwork was laid by Steuer (1982).

  22 For an introduction to the historiography, see Thompson (1965). I strongly suspect that measuring social status via artefacts will tend to place the basic erosion of human equality (to the extent that it ever existed) at far too late a date in the history of Homo sapiens sapiens.

  23 For useful surveys, see Thompson (1965), chapters 1–2; Todd (1992), chapter 2; for more detailed discussions, see Gebuhr (1974); Hedeager (1987), (1988), (1992), chapters 2–3; Hedeager and Kristiansen (1981); Steuer (1982), 212ff.; Pearson (1989). For Odry, see note 14 above.

  24 On Runder Berg, see Christlein (1978); Siegmund (1998); and cf. Brachmann (1993), 29–42; Drinkwater (2007), 93–106, which point out that there must have been other lowland Alamannic elite sites, none of which has yet been identified. On Feddersen Wierde, see Haarnagel (1979). On Gothic areas, see Heather (1996), 70ff. (with references). For more general discussion, see Krüger (1976–83), vol. 2, 81–90; Hedeager (1988), (1992), chapter 4; Todd (1992), chapter 6; Pohl (2000).

  25 The two classic and highly influential general accounts are the solidly Marxist interpretation of Fried (1967), and the more optimistic line adopted by Service (1975). These studies set the agenda for more detailed subsequent studies of intermediate societies (between the very small and the more modern). The four areas I identify represent a distillation from the helpful collections of papers in Claessen and Skalnik (1978), (1981); Claessen and van de Velde (1987); Skalnik (1989); Earle (1991); Claessen and Oosten (1996).

  26 This is true whether (see previous note) one adopts Service’s view of the process (by which a wider range of functions is more efficiently fulfilled) or Fried’s less optimistic Marxist view (whereby the growth of the bureaucracy entails the further rigidification of power structures).

  27 The key term here is ‘reciprocity’, meaning that ruler and ruled exchange something that is of mutual value. This probably won’t be (and certainly doesn’t have to be) an equal exchange, but even the act of exchanging makes the interaction honourable. If it is one-sided, then it is demeaning.

  28 Alamanni: Ammianus 16.12. Tervingi: Heather (1991), 109 (on pre-376 AD, based on Ammianus 20.8.1, 23.2.7, and 26.10.3), 146. Drinkwater (2007), 142–4 proposes that there were 15,000 Alamanni and allies at Strasbourg. He consistently downplays Alamannic numbers on the basis of his prior assumption that they posed no real threat to Roman frontier security, which is in my view a circular and unconvincing approach: see Heather (2008a). The evidence strongly suggests that these societies possessed slaves and that slaves were not normally liable for military service. We do not know the proportion of slaves, but they are likely enough to have been a significant portion of the population, so that merely to number fighting men will be to underestimate the total of young adult males in these societies.

  29 For waterborne summits, see Ammianus 27.5.9 (cf. Themistius, Orations 10), 30.3.4–6. For Burgundian/Alamannic boundaries, see Ammianus 28.5.11.

  30 For an introduction to the evidence, see Heather & Matthews (1991), chapter 5.

  31 On the Gothic contingents, see note 28 above, with Heather (1991), 107ff. for the crucial link that military service was something imposed on the Goths by the Romans when they held the diplomatic upper hand. On the Alamannic contingents, see Heather (2001). On the loan word, see Green (1998), chapter 11.

  32 Vannius: Tacitus, Annals 12.25. On Roman imports on elite Gothic sites, see Heather (1996), 70–2. On trade and diplomacy, see Heather (1991), 109. Of course, Chnodomarius may possibly just have been offering a share of war booty rather than cash up front.

  33 On the ‘wall’ of Athanaric, see Ammianus 31.3.8, with Heather (1996), 100 for the identification. On Runder Berg and other sites, see note 24 above.

  34 Based on a trawl through the literature cited in note 25 above. Not even the famously inert Irish kings of the Middle Ages – so wonderfully caricatured by the late Patrick Wormald as a ‘priestly vegetable’ – failed to exercise powers over dispute settlement. In the famous tract on Irish kingship, Crith Gablach, one day was reserved for this function: see Binchy (1970b); cf. Wormald (1986).

  35 For an introduction to early Anglo-Saxon tax systems, see Campbell (2000); Blair (1994). These kinds of arrangement have also been found in areas of Britain that never fell under Roman rule: see Barrow (1973).

  36 The mobility of Alamannic kings is suggested by the difficulty the Romans faced in trying to kidnap one of their number: see Ammianus 29.4.2ff. For an excellent introduction to the immense bibliography on itineration, see Charles-Edwards (1989).

  37 See Thompson (1966); cf. Heather (1991), 177ff. (with full references). For Gundomadus, see note 3 above. Even if one accepts the hypothesis of Drink-water (2007), 142–4 that there were 24 Alamannic canton kings, they would have produced no more than 4,800 retinue warriors between them. On the range of material in burials, see e.g. Steuer (1982); Weski (1982); Harke (1992). On burials entirely empty of goods, see e.g. Heather and Matthews (1991), 62, for some examples from Gothic-dominated territories.

  38 A quick read of the relevant law collections from the Visigothic, Frankish, Lombard, Burgundian, and Anglo-Saxon kingdoms brings out the importance of this group, who also feature in materials from the ‘smaller’ political entities, such from Thuringia, Bavaria, and Alamannia.

  39 On the proportion of freemen to slaves, see Heather (1996), 324–5, after Procopius, Wars 3.8.12 (1 elite to 4 subordinates in one Gothic force); 8.26.12 (close to 1:1 in a Lombard force). On this warfare, see Heather (1996), especially Appendix I (collecting the evidence for two classes of warrior being mentioned in Roman narrative sources). For charter evidence, see Wickham (1992); (2005), part 3. Post-Roman society did not immediately fall under the sway of the much smaller landowning elite, who can be seen to be dominant from the Carolingian period of c.800 and beyond: see for example Chapter 6 above on the growth of landed estates, which was the basis of aristocratic/gent
ry domination in Anglo-Saxon England and northern Francia; and, for more general comment, Wickham (2005), part 2.

  40 The village community at least attempted to protect the Christians in their midst: see Passion of St Saba 4.4; Heather and Matthews (1991), chapter 4.

  41 See Ammianus 31.3.8.

  42 The law codes again show that social value varied substantially with age, with women’s value famously being highest during child-bearing years. But age was clearly important to men too: older men were buried with spurs but not weapons, for instance, suggesting that there was an age limit to military obligation: see Hedeager (1988). Children were likewise sometimes not buried in cemeteries: see e.g. Siegmund (1998), 179ff.

  43 On the general importance of feasting as part of ‘reciprocity’ (see note 27 above), see Earle (1984), (1991). The first-century evidence is discussed in Thompson (1965). On Anglo-Saxon ideologies and realities, see Charles-Edwards (1989); Campbell (2000), chapter 8.

  44 For the early Roman period, see Thompson (1965), 37ff. On Roman control of assemblies, see Dio 72.19.2; 73.2.1–4. On fourth-century village assemblies, see Passion of St Saba; cf. Heather and Matthews (1991), chapter 4. On the decision to cross into the Empire, see Ammianus 31.3.8: ‘diuque deliberans’ (see Chapter 4 below). Thompson (1965), (1966) emphasizes the absence of reference in fourth-century sources (which basically means Ammianus) to regular councils among the Goths and other Germani. While a correct observation, it does not mean they weren’t happening.

  45 The literature on sacral kingship is huge, but see e.g. Wenskus (1961) and Wolfram (1994). The terminology and concept of heilag is nevertheless clear: see Green (1998), chapter 7 for the linguistic evidence; and cf. Pohl (2000) and Moisl (1981) for a practical application. On the actual (as opposed to invented) history of the Amal dynasty, see Heather (1991), chapters 1–2, and part 3; Heather (1996), chapters 6, 8 and 9.

  46 See Gregory of Tours, Histories 2.9; the Chatti are also mentioned at Ammianus 20.10. Salii: Ammianus 17.8; cf., amongst a huge range of possible secondary literature, James (1988), chapter 1, and the relevant papers in Wieczorek et al. (1997). The political processes behind the generation of the Alamanni may not have been totally dissimilar. No old names survived into the fourth century, but the confederation does seem to have built up gradually over time. In the third century, for instance, the Iuthingi (itself a new name) seem to have been a separate grouping, but by the fourth were operating as an integral part of the broader confederation: see Drinkwater (2007), 63ff.

  47 For Gargilius’ cow, see Boeles (1951), 130, plate 16, cited in Geary (1988), 3; the calculation of legionaries’ demands is from Elton (1996).

  48 Julian’s treaties are discussed in more detail in Heather (2001). On the frontier and its operations, see generally Whittaker (1994); Elton (1996); Wells (1999), chapter 6; Carroll (2001), of which the two latter focus greater attention on the Roman side of the Rhine.

  49 On loan words and trade, see Green (1998), 186f. and chapter 12. On iron production, see Urbanczyk (1997b); cf. more generally Krüger (1976–83), vol. 2, 157ff.

  50 On the forced drafts of recruits, see Heather (2001).

  51 See Green (1998), chapter 12.

  52 Caesar, Gallic War 4.2; Tacitus, Germania 5 (who notes, however, that interior Germanic groups still did not value Roman silver coins); cf. Green (1998), chapter 12. On fourth-century coin concentrations, see Drinkwater (2007), 128–35; Heather and Matthews (1991), 91–3.

  53 On the Tervingi and trade, see Themistius, Orations 10, with the commentary of Heather (1991), 107ff. For general orientation on Roman imports and their patterns, see Eggers (1951); Hedeager (1988); von Schnurbein (1995); Wells (1999), chapter 10; Drinkwater (2007), 34ff.

  54 For Roman goods and social status, see Steuer (1982). For the amber causeways, see Urbanczyk (1997b). For tolls, see Green (1998).

  55 See Caesar, Gallic War 6.17; Tacitus, Annals 13.57. For the bog deposits, see Orsnes (1963), (1968); Ilkjaer and Lonstrup (1983); Ilkjaer (1995); for more general comment, see e.g. Hedeager (1987); Steuer (1998); Muller-Wille (1999), 41–63.

  56 For a thoughtful critique of the importance of trade, see Fulford (1985). On ninth- and tenth-century beneficiaries, see Chapter 10 above. For an introduction to ‘agency’, and its more particular problems, see Wilson (2008).

  57 For a detailed report of the find, see Kunzl (1993); for an English summary, see Painter (1994).

  58 For a more detailed account, with full references, see Heather (2001).

  59 Ammianus 17.12–13, with Heather (2001). On the removal of potentially dangerous leaders, see Ammianus 21.4.1–5; 27.10.3; 29. 4.2 ff.; 29.6.5; 30.1.18–21.

  60 On the rationale of hostage-taking, see Braund (1984). On subsidies, Klose (1934) collects the evidence from the early period, Heather (2001) for that of the later Empire.

  61 ‘So eagerly did our forces’: Ammianus 19.11. For further comment on the balance between resettlement and exclusion, see Chapter 3 above; and cf. e.g. Heather (1991), chapter 4, on standard Roman immigration policies; and Carroll (2001), 29ff., on the amount of organized restructuring of adjacent populations that went on as Rome created its German frontier.

  62 Valentinian’s reduction of gifts: Ammianus 26.5; 27.1. For commentary, see Heather (2001); and Drinkwater (2007), chapter 8 (who seeks, in my view damagingly, to demonstrate that the Alamanni could never have represented any kind of threat).

  63 On the Rhine–Weser, see Drinkwater (2007), 38–9. On fifth-century economic expansion in Alamannia, see ibid., 355–44.

  64 See Wells (1999), chapters 10–11, following von Schnurbein (1995), who stresses the increase in imports of Roman weaponry into Germanic contexts after the mid-second century.

  65 Athanaric: Ammianus 27.5; Macrianus: Ammianus 30.3. In both cases, though, the relevant emperor was being pressed by problems elsewhere – Valentinian in the Middle Danube, and Valens in Persia: see Heather and Matthews (1991), chapter 2.

  3. ALL ROADS LEAD TO ROME?

  1 ‘They were expecting . . .’: Dio 32. 8–10.

  2 For a good introduction, see Birley (1966), chapters 6–8, with Appendix III; see also Böhme (1975).

  3 See Dio 72.20.1–2 (on the stationing of troops); 72.11–12, 72.20.2, 72.21 (on the movements of the Asdingi, Quadi, and Naristi respectively); 72.15, 72.16.1–2, 72.19.2, 73.3.1–2 (on trading privileges and neutral zones); 72.19.2, 73.2.1–4 (on assemblies).

  4 ‘Not only were . . .’: Historia Augusta: Marcus Aurelius 14.1; for an introduction to the trickeries of this text, see Syme (1968), (1971a), (1971b). For Roman aggression, see Drinkwater (2007), 28–32, who adds further thoughts on the possible impact of the plague, and Marcus’ sense of duty, to the argument.

  5 On Rhine frontier damage, see Carroll (2001), 138; and cf., on the legions and Marcus’ self-monumentalization, Birley (as note 2 above). See also Chapter 2, note 28 above.

  6 For the first-century homeland of the Langobardi, see Tacitus, Germania 40. That group of 6,000 clearly did not represent more than a subgroup, and they would be followed south by more Langobardi in the fifth century (see Chapter 5 below). These later Langobardi invaded the Middle Danube proper from intermediate settlements in Bohemia, but it is unknown whether this was true also of the second-century group. For references to permanent displacements, see note 3 above.

  7 See Dio 72.3.1a.

  8 See e.g. Barford (2001), introduction and chapter 1.

  9 Of fundamental importance here is the work of the late Kazimierz Godlowski, especially his general treatment of north-central Europe in the Roman period: Godlowski (1970). Shchukin (1990) supplies a good general survey, building on Godlowski’s pioneering work. The argument continues over details, and many more ‘cultures’, and phases within ‘cultures’, have acquired much more precise and absolute dates. In pioneering days, only Roman coins provided any indication of absolute chronology. Since 1945, the chronological development of Roman wheel-turned pottery became better understood, both
for fine wares (dinner services) and amphorae (storage jars for olive oil and wine). Two later techniques supply still more precise dates: carbon-14 (which produces a date-range) and dendrochronology, based on tree rings (which tells you precisely when a given tree was cut down). Combined with Godlowski’s general method, these technical advances have generated a wealth of knowledge that would have astonished previous generations of scholarship.

  10 In technical dating terms, the expansion occurred in Roman Iron Age periods B2, B2/C1a. These paragraphs distil information in two important collections of papers: Peregrinatio Gothica 1 and 2; and cf. Shchukin (2005).

  11 For a fuller discussion, see Heather (1996), 35–8. There is a range of fragmentary references in classical sources indicating that Gothic groups were moving south and east: see Batty (2007), 384–7.

  12 The relevant literature is huge. For a brief introduction, with full references, see Heather (2005), chapter 2.

  13 For a recent comprehensive treatment, with full refererences, see Drinkwater (2007), chapter 2. (Note his important argument on pp. 43–5 that a group called the Alamanni clearly existed already in the 210s, a point to which we shall return.) On the brutal violence, see ibid., 78f. (with further examples); Carroll (2001), chapter 9.

  14 On Alamannic origins, see Drinkwater (2007), 48f., 108–16 (with full references).

  15 Argaith and Guntheric: Jordanes, Getica 16.91 (cf. Historia Augusta: Gordian 31.1 on ‘Argunt’, which probably represents a conflation of the two). Cniva: Zosimus 1.23; Jordanes, Getica 18.101–3; Zonaras, Chronicle 12.20.

  16 The principal source is Zosimus 1.31–5. For additional sources and commentary, see Paschoud (1971–89), vol. 1, pp. 152ff., n. 59ff.

  17 Zosimus 1.42–3, 46, with Paschoud (1971–1989), vol. 1, pp. 159ff. n. 70ff.

  18 Historia Augusta: Aurelian 22.2. There is no evidence that he was related to the Cniva who had been operating in the same region a generation before: see note 15 above. On all these third-century attacks, see Batty (2007), 387–95.