Drinkwater J.F. The Alamanni and Rome 213-496 (Caracalla to Clovis)
Подождите немного. Документ загружается.
are rhythmic and memorable. The kings were part of a social hierarchy.
This is seen most clearly in Ammianus’ account of the Alamannic host
before the battle of Strasbourg in 357 and of Hortarius’ entertainment
of Alamannic leaders facing Julian in 359.7 Combiningthese narratives,
we Wnd superior kings (excelsiores ante alios reges), kings (proximi
reges), lesser kings (reguli) and princes (regales).8 The same passages
tell us that beneath the kings were nobles (optimates) and beneath the
nobles were warriors (armati). This hierarchy has attracted a good deal
of attention; but we should note Elton’s point that it is very similar to
whatthe sourcesrelateofmost other northernbarbariansocieties,9 and
that Ammianus uses much the same terms when describing the
Sarmatians.10 Ammianus probably envisaged all northern societies in
this way, and we should not expect his account to contain much in the
way of anthropological subtlety. Indeed, in most cases it is probably
best not to worry too much about precise distinctions and think simply
of ‘kings’, ‘nobles’ (a number of whom were probably related to the
kings11) and ‘warriors’.
I have so far employed the term ‘king’ since this is the usual
translation of rex/basileus. However, since Alamannic leaders were
clearly not kings in the same sense as, say, Philip of Macedon or Herod
the Great, it may be better to think of them as ‘chieftains’. Yet even
this term needs to be treated with care. Were the reges established
‘chiefs’ (succeeding ‘. . . to a hereditary position in a context of social
hierarchy’, their power ‘ascribed’ and coinciding ‘with privileged
control of wealth’) or competing ‘big men’ (manipulating wealth to
achieve their positions of power)?12 A related possibility is that they
were anyway only temporary, wartime, leaders.13
The crucial question is whether Alamannic ‘royal’ power was
hereditary and permanent; and the communis opinio is Wrmly that it
was. Geuenich, for example, is conWdent that dynastic kingship was a
7 AM 16.12.23, 26; 18.2.13.
8 For regales as men of royal blood see Werner (1965: 75); Geuenich (1997b: 77);
Zotz (1998: 402–3). Cf., in a context which suggests Alamanni, Historia Augusta,
V. Probi 14.1: reguli.
9 Elton (1996a: 15–19, 32–7).
10 AM 17.12.1: subreguli, optimates; 18.2.13: reges omnes et regales et regulos.
11 Zotz (1998: 403).
12 Curta (2002b).
13 Geuenich (1997a: 49).
118 Society
normal part of Alamannic government, noting the cases of
Chnodomarius and Serapio (uncle and nephew), Gundomadus and
Vadomarius (brothers), Macrianus and Hariobaudes (brothers), and
Vadomarius and Vithicabius (father and son).14 One has to be careful:
some of these links, especially those between brothers, may have been
contemporary, the result of related challengers’ cooperating to top-
ple a dominant ‘big man’. However, dynastic succession is the most
likely convention, at least for those communities living close to the
Empire. Rome would have preferred it because it oVered more stabil-
ity and the chance of inXuencing future generations.15 Roman depos-
ition of an Alamannic king must have been a policy of last resort.16
Very telling, in this respect, is the case of Serapio’s father, Mederichus.
He was taken by Rome as a hostage for what seems to have been a
considerable period of time, before returning to the Rhineland where
he probably preceded his son as a local king.17 If there had been a
signiWcant danger of his absence damaging his status in Alamannic
society, that is if his position had not been assured by birth and could
have been open to challenge by an ambitious big man, it would not
have beneWted the Empire to detain him. We may envisage the
Alamannic kings as chiefs, and interpret Hortarius’ feasting of border
Alamannic leaders in 359 in these terms: not as a big man creating a
power base but as an established chieftain and fellow chieftains
demonstrating (feigned) unity in diYcult circumstances, after a fash-
ion that was as much Roman as Germanic.18 However, Rome also
made sure that no such chieftain became a ‘Great King’.19
According to Ammianus, the reges ruled over discrete territories:
regna (‘kingdoms’) or pagi (‘districts’/‘cantons’). Pagi were probably
discrete geographical areas, such as major valleys.20 It appears that a
14 Geuenich (1997a: 45, 49–50); cf. Zotz (1998: 402). Chnodomarius and Serapio:
AM 16.12.25. Gundomadus and Vadomarius: AM 14.10.1, 10; 16.12.17; 21.3.4.
Macrianus and Hariobaudes: AM 18.2.15. Vadomarius and Vithicabius:
AM 27.10.3; 30.7.7.
15 Cf. Burns (2003: 242–3).
16 Cf. the case of Macrianus and Fraomarius, below 150–1, 284, 353.
17 Below 155.
18 AM 18.2.13. Cf. below 251.
19 Below 124.
20 Elton (1996a: 32).
Society 119
kingdom might comprise more than one pagus.21 The inheritance of
constituent pagi would explain how brothers might come to share
power, with the oYce of king being seen as divisible rather than the
kingdom.22 From Ammianus’ accounts of Roman campaigns it
appears that the kingdoms were small—much smaller than, say, the
average Gallic civitas—from which one may infer that they generally
comprised no more than a couple of pagi.
Subordinate to the kings werethe nobles. Thesecouldprove trouble-
some subjects, but from them the kings must have drawn their most
trusted helpers, their comites, ‘companions’. In 357, Chnodomarius
surrendered with ‘three very close friends’.23 In 358, Julian detained
Hortarius’ four comites, ‘onwhose aid and loyalty he chieXy relied’.24 In
361, Julian’s oYcials arrested Vadomarius but released his comites.25
Ammianus does not specify the number of Vadomarius’ attendants
and clouds our conception of royal companions in generalby reporting
that Chnodomarius surrendered not only with his close friends but
also with 200 comites. I propose that the kings’ intimates were few in
number, and that Chnodomarius’ ‘200’ was the garrison of his
‘Herrschaftszentrum’, under himself and three ‘captains’. The 800 or
so warriors that I suggest formed the remainder of a pagus-army of
c.1,00026 will have been dependants of optimates living elsewhere. If
each, like Chnodomarius and his cronies, provided around 50 armati,
they will have been about 16 in number. Kings would therefore have
depended on optimates for their main military strength.
Optimate allegiance may have been Xexible, with successful kings
winning new supporters at the margins of their realms.27 It appears
that in times of great need kings could strengthen their power and
authority by calling in help from outside their territories;28 but this
must have been unusual, and the more so because, by disturbing the
current order, it was potentially dangerous for kings and optimates
21 Zotz (1998: 400–1) referring to AM 18.2.8.
22 Elton (1996a: 33); Zotz (1998: 402). Both cite the example of Gundomadus and
Vadomarius.
23 AM 16.12.60: amici iunctissimi.
24 AM 17.10.8: quorum ope et Wde maxime nitebatur.
25 AM 21.4.5.
26 Below 238.
27 Elton (1996a: 31–2).
28 Below 238.
120 Society
alike. Optimates, as well as generally supporting their chiefs, were
expected to Wght alongside them in battle—even, when necessary, as
berserkers.29 They were also used as envoys in dealings with the
Empire (and, presumably, with other regna).30 On the other hand,
it appears that any king who was perceived to be failing would face
trouble from his optimates. In 357 Gundomadus was murdered and
Vadomarius forced to accept anti-Roman policies. Ammianus puts
this down to their respective populus and plebs, but we are probably
justiWed in ascribing the ‘plots’ involved to leading optimates.31 This
raises questions concerning the mechanism of succession. It seems
likely that, on Gundomadus’ death, Vadomarius was able to join his
brother’s kingdom to his own.32 However, if Vadomarius and his
family had been killed, from whom and how would a new king have
been chosen? One can only guess at some meeting of warriors, its
decisions reXecting the political scheming of leading optimates.
This brings us to the little we know of the character of Alamannic
society: that it was violent. This is in line with the martial nature of
barbarian society in general, and the habitual violence and feuding of
Germanic society in particular.33 The latter is visible in early medi-
eval Alamannic laws, which reveal a highly competitive warrior-
society, based on family groups.34 The cause of such violence was
probably social insecurity resulting from chronic poverty rather than
population increase or natural catastrophe.35 Under such circum-
stances, warriors would Wght each other and the Roman Empire not
out of malice or long-term political strategy but simply because they
felt they had to. And it seems that brutal practices, such as head-
hunting, were encouraged by Rome for imperial ends.36 It would not
be surprising if the urge to raid remained a feature of even settled
Alamannic society. This would help explain excessive Alamannic
29 AM 16.12.49.
30 AM 14.10.9–10 (357); 26.5.7 (365); 28.2.6 (369); 28.5.11–12 (370).
31 AM 16.12.17: [Gundobadus] per insidias interemptus. Cf. Zotz (1998: 404).
32 Cf. above 119. I take Ammianus’ later (21.2.3: 360) a pago [singular] Vadomarii
to be just a generalization, with pagus as equivalent to regnum.
33 Hedeager (1993: 122); Elton (1996a: 46–7); cf. Heather (2001: 41).
34 Steuer (1997b: 275–8); Pohl (1998a: 641), (2002a: 65).
35 Gutmann (1991: 14); cf. above 49.
36 Libanius, Orat. 18.45; cf. Zosimus 3.7.2, which Paschoud (1979: 79) relates to
Historia Augusta, V. Probi 14.2.
Society 121
enthusiasm to ‘aid’ Constantius II against Magnentius in the 350s
and, generally, spasmodic trouble indicating that even friendly chiefs
had diYculties in controlling their own warriors.37
Though the most able and ambitious optimates may have been
removed from local politics by entering imperial serv ice,38 one sus-
pects that it was never easy to be an Alamannic chief. This explains
the attractiveness of Roman support39 and is also likely to have
aVected Roman terminology.
The derivation and meaning of Alamanni and Alamannia have been
dealt with. Here, it is suYcient to repeat that we do not know what
the Alamanni called themselves. However, it is reasonable to suppose
that people identiWed themselves and where they lived according to
the name of their current chief: ‘the people of n’ or inhabitants of ‘the
pagus of n’. 40 From a Roman viewpoint this would have been unsat-
isfactory because kings were bound to change. For its purposes, the
Empire needed more permanent names, which brings us to the
question of the ‘Teilsta
¨
mme’.
Striking in Ammianus’ accounts of how the Alamanni organized
themselves for war are his references to two regional groupings.
These comprise the Lentienses, just north of Lake Constance,
whose name is plausibly associated with the medieval ‘Linzgau’;41
and the ‘Bucinobantes’, in the area of the lower Main, approached by
way of Wiesbaden.42 To Ammianus’ groups modern historians have
added two more, found in the Notitia Dignitatum alongside mention
of the Bucinobantes, which appear to reXect the same sort of organ-
ization. These are the ‘Brisigavi’, associated with Breisach/Brisiacum
in the Breisgau on the upper Rhine; and the ‘Raetovarii’, understood
as ‘men of the (sc.) lost territories of Raetia’, that is north of the upper
Danube.43 All are now generally referred to as ‘constituent’- or
37 Cf. Stroheker (1975: 30); Gutmann (1991: 14); Elton (1996a: 74). Below 203,
222, 262 and 286.
38 Below 154.
39 Cf. above 82.
40 Wenskus (1961: 498); Geuenich (1997a: 29–30, 46), (1997b: 74–5); Zotz (1998:
401). Cf. Ammianus 16.12.17; 18.2.8.
41 AM 15.4; 31.10.1. Zotz (1998: 393).
42 AM 29.4.3, 7. de Jonge (1939: 97).
43 Not. Dig. Or. 5.58; Occ. 5.201/7.25, 5.202/7.128. Zotz (1998: 401); Steuer (1998:
276); Geuenich (1997a: 29).
122 Society
‘sub’-tribes (‘Teilsta
¨
mme’), not in a past (‘this is how the Alamanni
arrived in Alamannia’) or even contemporary (‘this is how the fourth-
century Alamanni were organized’) sense, but as emerging aggregates
of pagi which would eventually coalesce to form ‘the Alamanni’.44
It has been suggested that the ‘Teilsta
¨
mme’ were originally Roman
constructs. ‘Lentienses’, ‘Raetovarii’ and ‘Brisigavi’ seem Celto-
Roman, not Germanic, in form.45 They are locational, indicating
where people lived, not necessarily what they called themselves.
They therefore could have been applied to Germanic newcomers by
outsiders already acquainted with the areas in which these people had
settled. They have the air of having been given to provide permanent
descriptions of groupings irrespective of who led them. It has been
further suggested that such outsiders could have been Roman
recruiting oYcers, raising Alamannic units for the imperial army.46
I consider these issues below, and propose modiWcation of current
thinking. In brief, ‘Lentienses’, ‘Brisigavi’ and ‘Raetovarii’ were even
more Roman than has been accepted so far: they should not be seen
as permanent indigenous geopolitical entities in any state of devel-
opment. An exception is provided by the ‘Bucinobantes’, the only
group we know with a Germanic name. These may have exploited the
identity given to them by Rome to create a temporary miniature
empire on the Main.47
Another controversy concerns the existence of a Great King. This is
important for events during the Wfth century,48 but debate begins with
the fourth. In describing the Alamannic alliance of 357, Ammianus
calls Chnodomarius and Serapio ‘kings hig her than the rest in
authority’. Further, his narrative has Chnodomarius clearly in charge of
operations.49 The alliance was destroyed by Julian, but the question
arises as to whether Chnodomarius’ position indicates that, despite
political division, Alamannic communities already had the potential
to unite under a single leader. In shor t, are we in the presence of
the Wrst stage of development of a single Alamannic monarchy, like
44 Geuenich (1997a: 29–31).
45 Wenskus (1961: 498).
46 Geuenich (1997a: 29–31), (1997b: 74–5); Steuer (1998: 276); Zotz (1998: 401).
47 Below 174, 250, 304, 324.
48 Below 335–7.
49 Below 236.
Society 123
that of the Salian Franks under the Merovingians? Some express
doubt by pointing out that, though the confrontation of 357 was
dominated by Chnodomarius, nominally at least he shared power
with Serapio; that not all Alamannic communities were directly
involved; and that the alliance was shattered following military
defeat. The enterprise should therefore be regarded as partial and
temporary.50 Others are more positive, interpreting the collapse as
reXecting the extent of Chnodomarius’ inXuence and proving that
there were means by which an individual might gain prominence in
Alamannic society.51
We should not overlook the obvious. Though there is a case for
suspecting that Ammianus exaggerated the extent of Chnodomarius’
standing,52 he was clearly a key player and he was followed by
other Alamannic leaders with more than local inXuence, such as
Macrianus.53 We must also take into account the militaristic and
competitive nature of Alamannic society, likely to produce warriors
willing to challenge the leaders of their own community or, as kings
themselves, other kings. History has produced enough examples of
such men creating great empires for us to be familiar with the type.
Why not among the Alamanni? The problem for such people in the
fourth century was that the Alamanni were in a cultural and strategic
cul-de-sac. For a leader to emerge, he would have to be victorious in
battle, and the sole worthy opponent was Rome. But Alamannic
leaders relied on Rome for pay and goods; and Rome was anyway
all-powerful: able and willing to nip any such aspirations in the bud
and, if this failed through imperial inadvertence, distraction or
incompetence, to crush anyone who managed to emerge. 54 There
was potential, but this could be realized only after the removal of
imperial power. Without such leadership, in the fourth century there
could be no clear Alamanni and indeed no clear Alamannia.
Like the ‘Teilsta
¨
mme’, in the Roman period Alamannia was a
geographical expression, not a geopolitical entity. There is no reason
to assume that the old limes formed a frontier. The traditional case
50 Geuenich (1997a: 49), (1997b: 77); cf. Stroheker (1975: 36).
51 Zotz (1998: 395–6, 402).
52 Below 236.
53 Below 304. Zotz (1998: 402).
54 Cf. Heather (2001: 42–6); below 358.
124 Society
for this as the Burgundian ‘border’ is not convincing; and anyway
hill-sites extend far beyond its line.55 In addition, at Gauko
¨
nigshofen,
further up the Main, recent research has shown Elbe-Germani, who
may have been Alamanni or Iuthungi, rubbing shoulders with
residual Rhine–Weser Germani and probably also Burgundians.56 The
region was not administratively and politically integrated until the
Merovingian ducate of the sixth century.57 Allemagne/Alemania
entered modern European languages from Alamannia’s status as a
Frankish dependency, not from its original designation as an area just
outside the Roman Empire. In fact, relative to its four th-century
position, Frankish Alamannia appears to have ‘slipped’ signiWcantly
southwards and westwards.58 Medieval Alamannia came into exist-
ence as a result of the perception and needs of outsiders, and the
same may be said of its earlier relationship with Rome.
Rome can have been interested only in people near the borders of
its Empire. This gave Roman Alamannia a deWnite shape in the
south-west, where it abutted the Empire, but in the north-east its
boundaries are unknown and were probably never clear. It is likely
that people became or ceased to be Alamanni in Roman eyes as they
moved in and out of the region.59 So, again, as in the case of the
‘Teilsta
¨
mme’, such shape as Alamannia had was Roman-imposed, like
so much else in developing Germanic society.60 It is hardly surprising
that in the fourth century, as in the third, we Wnd no Alamannic unit y
or even aspiration to unity.61 There had been progress. The settled
kingdoms were more sophisticated and, presumably, more in contact
with each other on a regular basis than itinerant war-bands; and, as
with the Franks, Roman-inspired ethnogenesis will have fostered
fellow-feeling. However, integration was still weak and liable to be
disrupted by fresh raiders from the east,62 and so was not expressed
in political or military terms. There was still no more than a mosaic
55 Above 94, 110–11.
56 Steidl (1997: 76), (2000: 108–10); cf. above 100.
57 Geuenich (1997a: 92); cf. below 347.
58 Geuenich (1997d: 204–7).
59 Keller (1993: 100–1); cf. Zotz (1998: 401). Cf. above 48.
60 Geary (1988: vi); cf. Pohl (1998b: 2).
61 Wenskus (1961: 503); Geuenich (1997b: 77–8); Pohl (2000: 30–1).
62 Below 313.
Society 125
of kings and pagi (although, unlike the Franks, there is curiously little
direct evidence for Alamanni Wghting each other), incapable of act-
ing together on a grand scale. ‘The Alamanni’ posed no direct threat
to Roman territorial integrity. Alamannic raiding could be distress-
ing, but despite the impression given by the sources, especially
Ammianus, there were no really dangerous attacks on the Empire.
Against this background, it is hardly surprising that the groupings
we know about, or think we know about, form no tidy pattern. Thus,
while the Bucinobantes and Brisigavi fall neatly into the coin-using
areas, the Lentienses and the Raetovarii do not. And while the hill-
sites along the Swabian Alp might take in the Raetovarii, they lie
north of the Lentienses.
The question as to whether Alamannic settlements comprised a
signiWcant proportion of the original imperial population and its
descendants is much debated. Some scholars reserve judgement;
others are hotly opposed to the idea.63 However, before dealing
with this directly, we need to consider the proposal that, under
Gallienus or the ‘Gallic’ emperors, the people of the region were
formally evacuated and settled in north-eastern Gaul.64 This is based
on two observations. The Wrst is that an operation of this nature was
later put in train for the inhabitants of Dacia. These were settled
south of the Danube in two ‘new’ provinces of Dacia Ripensis and
Dacia Mediterranea. Second, the part of Upper Germany that was
lost to Alamannia was known as the Agri Decumates, ‘The Ten
Canton Lands’.65 In the late Roman period, we Wnd a similar name,
Decem Pagi, ‘The Ten Cantons’, referring to a place but perhaps
originally designating a district, to the west of Strasbourg. The
evacuation hypothesis has won support,66 but I reject it here.
The evacuation of Dacia was put in hand by Aurelian, a strong-
minded ruler under no direct threat, w ith the conWdence to take a
painful but necessary decision and the wit to obscure imperial
contraction in a confusing new nomenclature.67 Any evacuation of
the Agri Decumates would have fallen in the early stages of protracted
63 Schach-Do
¨
rges (1997: 101). Cf. the observations of Sommer (1999: 190–1).
64 Hind (1984).
65 Cf. above 32–3.
66 e.g. Matthews (1989: 308); Jehne (1996: 204); Witschel (1999: 211–12).
67 See Watson (1999: 143, 155–7).
126 Society
civil war, in an area which had the attention of at least two interested
parties. Its eVective loss did not need to be so carefully disguised. The
abandonment of Dacia was the abandonment of a huge province,
conquered by the mighty Trajan. The abandonment of the Agri
Decumates could be projected as the partial withdrawal from a
marginal area of Upper Germany which had been annexed only
after it had been colonized unoYcially, under the reviled Domitian,
by the ‘riV-raV’ of Gaul.68 The similarity between Agri Decumates and
Decem Pagi is fort uitous.
What happened to the Roman population of what was to become
Alamannia? It is likely that, at the same time as the military aban-
donment of the region, if not earlier, those provincials who could
moved to territory that was still under imperial protection.69
Alamannic raiding and settlement, however gradual, cannot have
been pleasant. One thinks again of signs of torture and murder at
Harting.70 But others must have chosen, or felt compelled, to remain
behind. Choice would have been easier for the poor: the reluctance of
peasants to leave their land is notorious.71 As for compulsion, one
may assume, for example, that Germanic warriors detained and
supported a number of provincial females and their joint oVspring.
Generally, chances of demographic continuity would have been
much higher if Germanic settlement was a piecemeal process, not a
sudden, violent conquest. There may even have been a phase when
indigenous peasants and warrior newcomers lived alongside each
other.
Early upper-class Xight is, it has been suggested, revealed in
inscriptions showing the presence of civitas-elites of the area in
Mainz.72 Alternatively, these might simply reXect the propensity of
‘country’ decurions to spend their money in the regional capital.73 As
far as lower-class reluctance to move is concerned, the picture is even
more obscure. There is no sign of the continued use of Roman-style
goods in rural settlements. However, this may be because a Roman
68 Tacitus, Germania 29.4 (cf. below 140 n.146).
69 See, e.g. Trumm (2002), for the area around Schleitheim.
70 Above 78–9.
71 Drinkwater (1992: 214).
72 Nuber (1997: 64); Carroll (2001: 138 and Pl. 8).
73 Witschel (1999: 345–6).
Society 127