Empires and Barbarians

by Peter Heather

  41 Both points – i.e. the Middle Danubian origins of the crisis, and the subsequent appearance there of the Huns – were first argued by Heather (1995a), and are now been generally accepted: see for example Goffart (2006), chapter 5; Halsall (2007), 206ff. The crucial passage of Claudian which has been misunderstood to refer to Huns on the Danube is Against Rufinus ii.26ff. (especially 36ff.)

  42 Heather (1995a).

  43 Goffart (2006), chapter 5, especially 75–8 (Huns appear in Middle Danube shortly after the crisis); 78–80 (Radagaisus); 94–5 (summarizing the knock-on effect among the expectations of other groups of the fact that the Goths had survived their arrival on Roman soil with their coherence more or less intact).

  44 Halsall (2007), 195–212; cf. Halsall (2005), particularly on the disruptive effects of ending subsidies.

  45 On Tribigild, see Heather (1988); Synesius, De Regno 19–21.

  46 The first practical help from the east consisted of 4,000 soldiers who arrived in Ravenna in 409/10: Zosimus 6.8.

  47 For slaves, see Orosius 7.37.13ff.

  48 Either Constantine III or Flavius Constantius has usually been considered responsible for the transfer: see Chastagnol (1973); cf. Kulikowski (2000a). Halsall (2007), 209 raises doubts, but offers no specific evidence in their support.

  49 On Constantine III, see Zosimus 6.1, which specifically identifies British, Gallic and Spanish military forces as sufficiently united behind him to drive the Vandals, Alans and Suevi into Spain, and to take the usurper to the brink of Empire: see Matthews (1975), 312ff. On the general role of subsidies in Roman diplomacy, see Heather (2001).

  50 The relevant sources are, above all, Ammianus 17.12–13; Anonymous Valesianus 6.31–2. For a recent discussion of the Vandals in the fourth century, see Goffart (2006), 82–7, who convincingly concludes that the evidence places them in Silesia and on the Upper Tisza.

  51 For Vandals in Raetia, see note 31 above. For their fourth-century placement, see previous note.

  52 For fourth-century Goths, see for example Heather (1991), chapter 3. For Alans, see Goffart (2006), 89–90, with Ammianus 31.3.1, who records that the western-most group of Alans in c.375 were called ‘The Don People’ (Tanaites).

  53 For the Alans in 377, see Ammianus 31.8.4ff., with Heather (1991), 144–5 and Appendix B; and in 378, see Ammianus 31.11.16. For their drafting into the Roman army, see Zosimus 4.35.2.

  54 The identity of Uldin’s followers emerges from Sozomen, Historia Ecclesiastica 9.25.1–7 and Codex Theodosianus 5.6.3.

  55 Ammianus 17.12–13 (Constantius’ arrangements in 358); 19.11.1–3 (the return of the Limigantes in 359).

  56 On the differences between the cyclical movements inherent in a nomad economy and ‘real’ migration, see pp. 208–12).

  57 Ammianus 31.4.13; I take it these are the Sarmatians defeated by Theodosius prior to his elevation: Theodoret, Historia Ecclesiastica 5.5; Panegyrici Latini 12(2).12.9–10.

  58 The nomadic character of the Alans’ economy makes one expect a priori that they had a different social structure from agricultural Germani such as the Vandals or Goths, and this is strongly implied, if in a rather general way, by Ammianus 31.2.25.

  59 On the Sciri, see Codex Theodosianus 5.6.3. The contrasting fates of the ‘better’ among Radagaisus’ following who were drafted into the Roman army, versus the many others sold into slavery, might suggest that the latter had had little choice over whether to participate in the action.

  60 On Uldin’s force, see Sozomen, Historia Ecclesiastica 9.25.1–7; and Codex Theodosianus 5.6.3. Radagaisus: Olympiodorus fr. 9 for the ‘best’, as against Orosius 7.37.13ff., who records the miserable fate of the mass of the rank and file sold into slavery. ‘Best’ translates Olympiodorus’ optimates, which has sometimes been translated as ‘nobles’, but to reckon so many nobles is absurd, so the word can only make sense as a reference to the higher-grade caste of warriors: see Chapter 2 above. The elites of both the Rhine crossers and the Burgundians, neither of whom of course were faced with as powerful and immediate a Roman counterattack as Radagaisus or Uldin, showed no obviously similar propensity to abandon the migrant mass.

  61 The only group for whom any case can be made for an invitation is the Tervingi, by Valens, in 376, but in my view even here Valens had no real choice: see p. 169 above.

  62 Cf. Heather (1991), chapter 5, and Appendix B. I don’t believe that the Emperor Gratian made a separate peace agreement with the Greuthungi in the summer of 380. That this adjustment in traditional Roman policy affected only these particular Goths is well understood: see Stallknecht (1969). Kulikowski (2002) and Halsall (2007), 180ff. have recently tried to argue that nothing out of the ordinary was granted in 382, but the case does not stand up to scrutiny: see Heather (forthcoming).

  63 It may well be, then, that local Roman landowners cut a deal with the invaders to prevent less organized and hence inherently more damaging assaults upon their property. Cf. Hydatius, Chronicle 41[49]: the settlement saw particular groups of invaders settle in particular provinces, so it is possible that the Spanish provincial councils were responsible for the Roman provincial side of the negotiations.

  64 See for example Kulikowski (2002); Halsall (2007), chapters 7–8.

  65 Claudian, De Bell. Get. 166ff., 610ff. (dating to 402); Synesius, De Regno 19–21 (dating to 399), with Heather (1988). Neither Kulikowski (2002) nor Halsall (2007), 189–94 offers any explanation of the fundamental distortion they suppose these authors to be incorporating.

  66 Zosimus 5.5.4. Privileging the short, non-contemporary and confused Zosimus over the more contemporary sources is the basic approach adopted (even if leading to slightly different interpretations of Alaric’s career) in Liebeschuetz (1992); Kulikowski (2002); Halsall (2007), 191–4. Amongst other problems, Zosimus conflates Stilicho’s two campaigns against Alaric (in 395 and 397) and wipes out ten years of the history of Alaric’s Goths in making the join between his two main sources here: Eunapius and Olympiodorus (at Zosimus 5.26.1: see Heather (1991), 210). To say that Zosimus had no real grasp of Alaric’s career, therefore, is an understatement.

  67 The activities of Gainas are well covered, if certainly with hostility, in Synesius, De Providentia; cf. Cameron and Long (1993).

  68 To my mind, this is why Liebeschuetz (1992) cannot be correct in viewing Alaric as leading no more than a regiment or two of Gothic auxiliaries in 395. Halsall (2007), 192–3 tries to wriggle round this problem by continuing to deny the overlap with the Goths of 382 while accepting that Alaric’s armed following must have been large, mostly Gothic and from the Balkans. Having accepted these points, he is in fact most of the way to the conclusion that Alaric led the 382 Goths in revolt. He resists this conclusion because he doesn’t believe there was a peace deal in 382 which licensed Gothic autonomy, but see the next note.

  69 Themistius, Orations 16.211. Continuing Gothic autonomy up to and beyond c.390 is signalled, beyond Themistius, in sources both sympathetic to Theodosius and his treaty such as Pacatus, Panegryici Latini 12.(2).22.3–5 (where the Goths are one of a series of foreign peoples serving Theodosius), and hostile to them: Synesius, De Regno 19–21, with the commentary of Heather (1988). Halsall (2007), 180–4 oddly argues that there is no evidence that any continued Gothic autonomy was licensed in 382; he appears not to have read the closing words of Themistius’ speech closely enough. Cf. Kulikowski (2002).

  70 For Roman policies towards leaders, see Heather (2001). Neither the original leaderships (Athanaric, the dynasty of Ermenaric) nor their immediate successors (Fritigern, Alatheus and Saphrax) survived the struggles of 376–82: for more detail, see Heather (1991), 188–92.

  71 Fritigern: Ammianus 31.12.8–9, with Heather (1991), 175–6, 179–80. The best example of the post-382 jockeying is provided by the quarrel between Fravitta and Eriulph. Both led factions and both held different views over the proper ordering of Gotho-Roman relations: see Eunapius fr. 59, dated by the summary of it at Zosimus 4.56. For further discu
ssion, see Heather (1991), 190–1. That Theodosius should have held such a banquet undermines the contention of Halsall (2007), 188–9 that Alaric couldn’t have been the leader of the Goths of 382 in revolt, because there is no evidence that their sociopolitical hierarchies had continued in place after that date (Halsall does not discuss the incident). For Sarus and Sergeric in more detail, see Heather (1991), 197–8. Neither Kulikowski (2002) nor Halsall (2007) bothers to discuss its potential significance. In my view, extracting Roman recognition of his leadership was also precisely the significance of the generalship that Alaric periodically demanded of compliant Roman regimes – probably along with the financial package for his followers that came with it. But note that the generalship was an optional extra that he was willing to drop to make a deal: see Heather (1991), chapter 6.

  72 Kulikowski (2002) – largely followed by Halsall (2007), 187–9 – denies large-scale Gothic military service in the years between the 382 treaty and Alaric’s revolt in 395, but this involves too much special pleading to be convincing. Panegryrici Latini 12.(2)32.3–5 strongly implies that the main Gothic contingent was recruited only for the campaign against Maximus (especially Pacatus’ explicit comment that it would have been dangerous to leave the Goths behind), while Eunapius fr. 55 and Zosimus 4.45.3 note Maximus’ attempts to undermine the recruited Goths’ loyalties, which again implies that this was something unusual. A range of sources note the participation of large numbers of Goths in the campaign against Eugenius (Zosimus 4.58; John of Antioch fr. 187; Orosius 7.37.19), and Theodosius’ banquet for the Gothic leaders (see note 71 above) was held precisely when Theodosius was mulling over his answer to Eugenius’ envoys (Zosimus 4.56). In my view, the banquet was probably a first move towards securing Gothic participation.

  73 For Maximus’ revolt, see Eunapius fr. 55; Zosimus 4.45.3, 48–9. Alaric of course led the revolt after the Eugenius campaign. The arguments of Kulikowski (2002) and Halsall (2007), 187–93 comment neither on the Maximus revolt nor on the significance suggested by the exact chronology of the banquet quarrel.

  74 Orosius 7.35.19 (casualties confirmed at Zosimus 4.58). Neither Kulikowski (2002) nor Halsall (2007), 187–93 discusses this backdrop to the Gothic revolt.

  75 In recent times, we have seen one successful example of this kind of diplomatic strategy in the Good Friday Agreement in Northern Ireland, and one so far unsuccessful example in the Oslo Accords on the Middle East.

  76 Zosimus 5.5.5ff.

  77 Themistius, Orations 16.211.c–d.

  78 Whether this made any practical difference to Alaric’s position in Illyricum in the short term is unclear; having been commanding general there, he was perhaps in a position to retain control of the levers of power.

  79 For this argument in more detail, see Heather (forthcoming).

  80 See in more detail Heather (1991), chapter 6.

  81 Fravittas, Sarus and Modares: PLRE 1, 605; 372–3. On the 402 battle, see Claudian, VI cons. Hon. 229ff.; cf. Cameron (1970), 186–7.

  82 It disappeared to the extent that of the sources that discuss the events of 376, only Ammianus knew there were originally two separate groups of Goths. In my view, both Greuthungi and Tervingi were settled under the treaty of 382 and Alaric’s revolt in 395 involved and definitively united both. An alternative view sees the unification happening when Alaric summoned his brother-in-law Athaulf from Pannonia in 408: Zosimus 5.37.1ff.

  83 Zosimus 5.35.5–6. Pace Kulikowski (2002), it is hard to see who this large body of barbarian soldiery in Roman service was, if not mainly the 12,000 followers of Radagaisus drafted by Stilicho: Olympiodorus fr. 9.

  84 Heather (1991), 151ff. looks to unravel Zosimus’ confusions.

  85 Gothic subgroups were destroyed by Frigeidus (Ammianus 31.9), Sebastianus (Ammianus 31.11) and Modares (Zosimus 4.25), and there is no reason to think this a comprehensive list. On this process in general, see Heather (1991), 213–14, 223–4, 314ff.

  86 Zosimus 5.45.3; cf. Liebeschuetz (1990), 75ff.; Kulikowski (2002).

  87 Exactly how much larger this Gothic force was involves a huge amount of guesswork, but if it is right to calculate the military manpower of fourth-century Gothic units at around 10,000, then the Visigoths who formed around Alaric could certainly field at least twice this number of soldiers, and possibly between three and four times as many.

  88 Victor of Vita, History of the Persecutions 1.2.

  89 Hydatius, Chronicle 77 [86].

  90 On the mid-410s: Hydatius, Chronicle 59–60 [67–8]; on the 420s: ibid. 69 [77]; on the 440s and 460s: Heather (2005), 289ff. and 390ff.

  91 Suevi: Hydatius, Chronicle 63 [71]. Alans: see the convenient listings of Bachrach (1973).

  5. HUNS ON THE RUN

  1 Jordanes, Getica 50.261–2.

  2 Uldin’s Huns and Sciri: Sozomen, Historia Ecclesiastica 9.5; Codex Theodosianus 5.6.3. Huns’ Gothic subjects in 427: Theophanes AM 5931; cf. Procopius, Wars 3.2.39–40, with Croke (1977). The date could be either 421 or 427. The best general survey of Attila’s subject peoples is Pohl (1980).

  3 See Maenchen-Helfen (1973), chapters 8–9, who also notes that leaders like Attila could easily have had ‘proper’ Hunnic names as well as Germanic nicknames; Attila means ‘little father’ in Germanic.

  4 Ammianus 31.2.1–2; Zosimus 4.20.3–5 (cf. Eunapius fr. 42); Jordanes, Getica 24.121–2.

  5 Ammianus 31.2.3ff.

  6 The Alanic digression: Ammianus 31.2; the Saracen digression: Ammianus 14.4. In treating this material, the approach of Maenchen-Helfen (1945) was much more critical than that of Thompson (1995), even though it was Maenchen-Helfen who noticed the meat being placed under saddles. For further comment, and recent bibliography, see G. Kelly (2008), chapter 2.

  7 For some orientation on nomadism, particularly of the Eurasian-steppe variety, see Cribb (1991); Khazanov (1984); Krader (1963); Sinor (1977), (1990).

  8 Bury (1928).

  9 Avars: Pohl (1988). Magyars: Bakony (1999).

  10 For general accounts, see Thompson (1995); Maenchen-Helfen (1973); cf. Heather (1995a) on relations with Aetius.

  11 Attila’s more or less complete indifference to additional territorial gains emerges with striking clarity from the surviving fragments of Priscus’ history.

  12 Huns up to 376: Ammianus 31.3. Huns and Alans in 377: Ammianus 31.8.4ff. Huns and Carpo-Dacians: Zosimus 4.35.6.

  13 ‘Improvised leaders’: Ammianus 31.2.7. Jordanes does place a Hunnic king called Balamber in this era, but these are really events of c.450 and Balamber is in fact the Gothic king Valamer: see Heather (1989), and p. 234 above.

  14 Uldin: Sozomen, Historia Ecclesiastica 9.5, with further comment and full references in Heather (1995a). Analogous phenomena occurred in the Viking era, when leaders thrown up in the first generation of small-scale expansion were quickly subdued as larger numbers under more important leaders joined in the flow: see Chapter 9.

  15 Olympiodorus fr. 19; cf. Priscus fr. 11.2, p. 259 (on the Akatziri).

  16 On the Huns’ bow, see Heather (2005), 154–8, with further references.

  17 For some calculations based on grazing room on the Great Hungarian Plain, see Lindner (1981). On the great raid of 395, see Maenchen-Helfen (1973), 51–9.

  18 On nomad devolution in general, see the literature cited in note 7 above.

  19 See Heather (2005), 325ff., with full references.

  20 Procopius, Wars 8.5 seems to preserve by an indirect route the story originally told by the contemporary Eunapius, which was cast after Herodotus 5.9 (on the Sigynnae).

  21 On the third-century Heruli, see Chapter 3. Sciri: Zosimus 4.35.6. Rugi: Tacitus, Germania 43. On likely placements within the Middle Danubian region, see Pohl (1980).

  22 We will return to the Amal-led contingent in more detail. Bigelis: Jordanes, Romana 336. For the third group under Dengizich’s control, see Priscus fr. 49.

  23 Dengizich: PLRE 2, 354–5. Hernac: Jordanes, Getica 50.266, with PLRE 2, 400–1. Hormidac: PLRE
2, 571. Bigelis: see previous note.

  24 Jordanes, Getica 50.264. Pohl (1980) suggests – in a compromise – that the Amal-led Goths may have moved at this point only from Transylvania. Much ink has been spilled on the relationship between the surviving Gothic history of Jordanes and the Gothic history of one Cassiodorus, which we know to have been written down at Theoderic’s court in Italy. In my view, the textual evidence indeed suggests that Jordanes worked using Cassiodorus’ text (as he claims) and I find the various conspiracy theories that have been offered against this unconvincing: see Heather (1991), chapter 2; Heather (1993). The archaeological evidence for such a late Gothic move to the Middle Danube is indecisive. Kazanski (1991) has placed the end of the Cernjachov culture as late as c.450, but this is not the usual view, and the argument is essentially circular since it is based on Jordanes’ report that there were still Goths east of the Carpathians at this date.

  25 Odovacar: PLRE 2, 791–3. On the Balkan adventures of the Amal-led Goths, see Heather (1991), part 3.

  26 For full references, see PLRE 2, 457 and 484–5.

  27 The Lombards should strictly be called Langobards. The narrative source is Paul the Deacon, History of the Lombards 1.19. For modern commentary, see e.g. Christie (1995); Jarnut (2003); Pohl and Erhart (2005).

  28 Procopius, Wars 6.14–15. Cf. Pohl (1980): the archaeological evidence suggests that the Gepids were slowly expanding south into Transylvania at this point.

  29 An earlier exception would be the Goths rescued from Hunnic domination in 427: see note 2 above. These may also be the same as the Thracian Goths, as we shall shortly see. The Gepids too engaged in expansion within the region: see previous note.