Drinkwater J.F. The Alamanni and Rome 213-496 (Caracalla to Clovis)

Подождите немного. Документ загружается.

communities composed of Alamannic settlers in the Empire is the

Notitia Dignitatum’s indication of the presence of laeti gentiles Suevi

in Gaul.156 But this too is questionable. Later Roman writers used

‘Suebi’ as a synonym for Alamanni, but the usage tended to be

poetical, with a view to style and metre.157 Soldiers and adminis-

trators dealing with Germanic settlers would have been subject to no

such constraints and would, like the legal draftsman who composed

CT 7.20.12, probably have preferred the prosaic Alamannus. ‘Suebi’

are more likely to have been other Danubian Germani.158 This Wnds

some conWrmation in the fact that the Notitia Dignitatum’s reference

to the (few) Suebi settled in Gaul is followed by that of the (many

Danubian) Sarmatians settled in Italy.

There is also the question of the number of dediticii/laeti available

for service. HoVmann proposed that the later Roman army in Gaul

consisted for the most part of western Germani. During the fourth

century, the bulk of these came from laetic settlements. Unit names

which appear to reXect regional (e.g. Gallicani)orcivitas-based (e.g.

Ambianenses, Bituriges, Picatavenses) recruitment of native Gauls are

deceptive. They indicate only where laetic units were originally sta-

tioned.159 Gauls could still be recruited into legions and, occasionally,

into other units; but where the latter occurred it was only of second-

ary importance.160 This is controversial, and HoVmann’s views

have been disputed.161 It returns us to the question of numbers. To

resolve the issue as to the importance of the laeti we need to know far

more than we do about: the number of laetic settlements; the size of

their populations; the distribution of such populations by gender and

age; the quality of their land; the level of their agricultural expertise;

the housing, tools, seed crops, work animals etc. made available to

them; and the proportion of working males they were obliged to give

up to the army, and for how long. Such questions can be answered

only by means of archaeology, and currently archaeology is not in a

156 Not. Dig. Occ. 42.34 (Baiocasses/Bayeux), 35 (Cenomanni(?)/Le Mans),

44 (Arverni(?)/Clermont-Ferrand); Stroheker (1975: 39).

157 Green (1991: 535); Heinen (1995: 82). Cf. above 46.

158 Castritius (2005: 195–6); cf. above 48, below 324, 335.

159 HoVmann (1969: 142, 145–54, 160, 164, 168, 180, 196, 200).

160 HoVmann (1969: 177, 154 (citing ILS 9215: below 171 n.177)).

161 Cf. Elton (1996a: 150–1).

168 Service

position to do so. However, what we now know about Gaul and

Germania tells against HoVmann’s reconstruction.

The population of the Free Germanic territories was not massive:

there were no many thousands of pairs of hands ready to be put to

the service of Rome.162 Some dediticii and laeti, like those settled in

the Po valley, succeeded.163 For others, however, life must have been

hard. Living near the frontier, many must have been exposed to

raiding by their free cousins. And the lands they worked, though

previously farmed, are likely to have gone to waste. It would have

required enormous amounts of skill and energy to re-instate even

basic subsistence farming—the best that Alamannic farming exper-

tise could oVer. But most laeti and dediticii will have been young men

with limited experience of and interest in wresting a bare living from

the soil. The acquisition of wives and children may have increased

their commitment but, like Romulus’ Romans of old, they would

have faced diYculties in obtaining womenfolk. And one can hardly

imagine the Roman authorities taking a close interest in their prob-

lems. Laeti appear to have survived in some fashion, since the term is

used, somewhat enigmatically, of troops employed by Merovingian

kings.164 However, during the fourth centur y, most laetic communi-

ties must have found it a struggle to survive; and overall they could

hardly have been numerous or strong enough to meet the regular

manpower needs of the western army.165

Models I–IV envisage Alamannic warriors coming into Roman

service and being used, dismissed or absorbed as the occasion

demanded. Now, however, we have to consider a situation in which

Rome went to Alamanni and hired them under longer-term com-

mitments. This is dealt with as Model V, involving units which we

appear to know most about and which therefore should be the easiest

to understand: the ‘regionally ethnic’ regiments of the Notitia

162 Above 39.

163 Above 167.

164 Wood (1994: 64).

165 MacMullen (1963: 9–13, 19–21) has argued vigorously for farming as the part-

time occupation of full-time soldiers and soldiering as the part-time occupation of

full-time farmers in the later Roman Empire. However, despite his direct comparison

(14), this was hardly the same as the use of barbarians, who would have needed

constant surveillance and support.

Service 169

Dignitatum, Bucinobantes, Raetovarii and Brisigavi. These have

attracted a great deal of attention in the past because they appear

to give the names of constituent elements of the Alamanni—their

‘Teilvo

lker’ or ‘Teilsta

mme’.166 Attempts have been made to connect

this putative ‘tribal’ history with the political history of Ammianus’

narrative. For example, Vadomarius is frequently termed king of the

Brisigavi.167 But this reconstruction is not as straightforward as it

might seem.

These units were signiWcant components of the imperial military

establishment. The formation of permanent regiments was a serious

business because of theimplied cost. It was not something that can have

been taken lightly. HoVmannhas show n that suchincreases in imperial

strength happened in waves, when, presumably, the normal sources of

troop recruitment were unavailable or inappropriate. HoVmann sug-

gests that this happened in circumstances of exceptional need, when

just a few western commanders-in-chief—Julian, Valentinian I and

Stilicho (for Honorius)—rapidly raised new regiments to pursue par-

ticular objectives.168 This is clearest in the case of the Bucinobantes and

Raetovarii.HoVmannsees these asa linkedpair of units,and dates their

formation from the career of king Fraomarius, sent to Britain to take

command of the Bucinobantes in 372/3.169 Both the Bucinobantes and

the Raetovarii had been raised by Valentinian I in the tense period after

his arrival inGaul in 365. In 368, they were transferred to Britain to help

deal with the ‘Barbarian Conspiracy’ of 367.170

Similarly, if we accept HoVmann’s dating, the formation of the

twin units of Brisigavi occurred during the acute political and mili-

tary unease at the beginning of Honorius’ reign. At this time, his

regent, Stilicho, prudently increased his strength by recruiting from

western Germani. He formed at least 18 new regiments, two of which

were Alamannic: the Brisigavi Seniores and Brisigavi Juniores.171 In

this instance, however, the Alamannic connection is less strong. The

166 e.g. Stroheker (1961: 40); HoVmann (1969: 167). Above 122–3.

167 e.g. Stroheker (1961: 46: ‘zweifellos’); Lorenz, S. (1997: 78: Vadomarius’ son,

Vithicabius, also as ‘Ko

nig der Brisigavi’).

168 HoVmann (1969: 141–3, 159, 165–8).

169 For the date, see below 285.

170 HoVmann (1969: 165–6). On the troubles of 367 and later, see below 237.

171 HoVmann (1969: 143, 168); cf. Bury (1923: 1.118), for Stilicho’s visit to the

Rhine in 397. Below 320, 322.

170 Service

Brisigavi are accepted by historians as Alamannic because of

the similarity between their name and the modern name of the

area on the right bank of the Rhine, immediately west of the town

of Breisach/Brisiacum: the Breisgau.172 But it should be remembered

that late Roman Brisiacum was a military base guarding a Rhine

crossing, and lay on—or, at least, attached to—the river’s left

bank.173 In principle any regiment taking its name must have been

raised in its immediate, left-bank neighbourhood, and so need have

been little diVerent from the many others recruited in or from Gallic

communities, from dediticii/laeti or native Gauls, or both. The associ-

ation of Brisigavi in the Notitia Dignitatum with regiments bearing the

‘tribal’ names Atecotti, Marcomanni and Mauri is not helpful. Though

the Wrst of these is likely to have indicated an authentic tribal aYlia-

tion,174 the other two are suspiciously vague, and may be fanciful.

If the Brisigavi are still to be regarded as in some way Alamannic, a

supposition supported by archaeological Wnds suggesting a strong

Germanic presence in and around Breisach in the later fourth cen-

tury, similar interpretation ought to be extended to the several

household units of ‘Mattiaci’ found in the Notitia Dignitatum.175

These must have been raised around Wiesbaden/Aquae Mattia-

corum, on the right bank of the Rhine in an area of Romano-

Germanic activity. HoVmann, indeed, locates their formation in

the late third or early fourth century, when such activ ity was gaining

strength.176 He suggests that they were recruited from prisoners of

war or recruits from the high imperial tribe of the Mattiaci which was

somehow, despite the turmoil of the third century, still in existence;

but this is implausible. More likely is that they were raised from a

mixed transrhenish population, consisting of provincials and incom-

ing Alamanni, and perhaps Burgundians.177

172 Matthews (1989: 306); Bu

cker (1999: 22).

173 Bu

cker (1999: 23).

174 HoVmann (1969: 154).

175 In the east, Mattiaci seniores, Not. Dig. Or. 5.53 and Mattiaci juniores, Or. 6.53;

in the west, Mattiaci seniores, Occ. 5.164/7.15 (Italy), Mattiaci juniores, Occ. 5.165/

7.64 (Gaul) and Mattiaci juniores Gallicani, Occ. 5.209/7.77 (Gaul). Above 101.

176 HoVmann (1969: 156–7); above 130.

177 Cf. above 168 n.160. ILS 9215 suggests Gallic recruitment into the Mattiaci.

However, this could have taken place long after the original formation of the units.

Service 171

The Notitia Dignitatum does not mention the one Alamannic

grouping which we know provided recruits, the Lentienses.178

‘Urgency’, ‘rarity’, mixed populations and awkward omissions raise

other questions about the areas from which units derived their

names. We return to the extent to which the names we have denote

‘Teilvo

lker’ or ‘Teilsta

mme’. Current thinking is that most of these

names were not coined by the people they referred to. ‘Men of

the Linzgau’, ‘men of [the old province of] Raetia’ (I omit ‘Brisigavi’

and ‘Mattiaci’ for reasons obvious from my remarks above, and also

‘Bucinobantes’ to which I will return below) is not what Germanic

newcomers are likely to have called themselves but what they were

called by outsiders already acquainted with the areas in which they

had settled and their old, Celto-Roman, names. Such outsiders have

been tentatively identiWed as Roman recruiting oYcers, raising

Alamannic regiments but unwilling to heed Alamannic self-labelling

based on the names of kings.179 This interpretation requires some

adjustment.

First, it must have been the case that Roman regional names for

the Alamanni existed independently of and probably long before the

raising of regular regiments from these regions. As in the case of

‘Alamannia’,180 Roman soldiers and administrators would have

needed ways of describing the people they found themselves dealing

with which avoided the names of individual rulers. So they hit upon

regional names. This explains the existence of ‘Lentienses’.181 Second,

when it came to regimental recruitment, the question of the imper-

manency of par ticular Alamannic rulers cannot have been an issue.

The Late Empire did not name its regiments after anyone outside the

imperial family.182 In addition, it is unlikely that a full regiment

could have been raised from a single pagus. The removal of around

1,000 of his best warriors, even if he had that number at his disposal,

178 AM 31.10.17.

179 Above 122.

180 Above 68–9.

181 AM 15.4.1, listing Constantius’ Alamannic opponents in 355 ( . . . et Lentiensi-

bus, Alamannicis pagis : below 207), appears to suggest that the Lentienses were a

single pagus. But he obscures the sense of the word by using it in the plural; and at

31.10.2–3 he represents the Lentienses as a populus and then a natio.

182 Cf. above 164.

172 Service

would have seriously weakened any pagus chief. Thirdly, these regi-

ments were apparently raised in a hurry, outside the regular system of

recruitment. Rome would have wanted as many suitable recruits as

possible in the shortest time. Overall, such activity, though it would

need to have been agreed with individual kings and princes, cannot

have been organized by them and would probably have had to cover

several pagi. It is these wider areas, already identiWed, which were

targeted by Roman recruiting oYcers, and it is these areas which

are reXected in the names of the regiments formed in this way.

They were indeed Roman constructs, not indigenous geopolitical

entities.183 On the other hand, such regional groupings may have

aVected Alamannic political and social structures. Ambitious local

leaders, whose help Rome would have welcomed in raising regiments,

may have attempted to exploit the boundaries and names given by

Romans to increase their power. This takes us to the Bucinobantes.

Of Brisigavi, Bucinobantes, Lentienses and Raetovarii only

‘Bucinobantes’ appears to be basically Germanic in form. It has

been interpreted as ‘the people of the area of the defensive quick-

set hedge’.184 If this is correct, the hedge in question will have been a

real and considerable feature of the landscape, creating some form of

redoubt. This suggests an interesting alternative to the ‘Ho

hensied-

lung’.185 A hedge also Wts well with the little we know of the middle

Main, where the careful deWning and defending of areas seems to

have been an important issue in a region in which Romans,

Alamanni and Burgundians confronted each other.186 ‘Bucinobantes’

therefore appears, uniquely, as more than a neutral imperial label. Is

it an indication of a ‘Teilstamm’? Probably not: ‘Bucinobantes’ is

another, albeit diVerent, example of a spurious ‘constituent tribe’.

We Wrst hear of a king of the Bucinobantes, Macrianus, not long

after the likely date of the formation of the unit of this name.187

Though Ammianus makes out that Macrianus was highly inimical to

Valentinian I, this is probably overstated. Macrianus appears gener-

ally to have been well-disposed to Rome, using the relationship to

183 Cf. van Driel-Murray (2003: 204) on the Batavi, as interpreted by comparison

with the modern Gurkhas: ‘Tribal boundaries were established by Roman adminis-

trators, tribal identities registered by Roman census oYcials.’

184 Neumann (1981): ‘Knick, Verhau, verXochtene Heckenschutzwehr’.

185 Above 98. 186 Above 112. 187 AM 28.5.8.

Service 173

expand his inXuence down the Main valley.188 A decade or so before,

Macrianus had been cultivated by Julian,189 and it takes little to

envisage him creating a Lilliputian empire on the middle Main,

useful to Rome as a buVer against neighbouring Burgundians. Like-

wise, we may imagine the people of Macrianus’ empire taking their

name from the nature of their chief stronghold. As Macrianus took

over other Alamannic pagi, the name was retained: the Bucinobantes

drove westwards. Valentinian’s need to raise a new regiment,

recruited along the middle and lower Main and called, accordingly,

‘Bucinobantes’, caused him to continue to cooperate with Macrianus,

allowing the latter to come closer to the Rhine and so increase his

power and prestige. However, once Valentinian’s position was secure,

he decided to break Macrianus, now seen as a threat to the Mainz

crossing.190 This was eventually only partially achieved.191 Macrianus’

empire continued, but did not survive the death of its founder

at the hands of the Frank, Mallobaudes.192 This helps explain the

Burgundian presence at Mainz by the early Wfth centur y.193 Overall,

the Bucinobantes were a temporary political construct, not a

constituent tribe.

Ammianus’ report of Gratian’s diVusion of Lentiensian recruits

and the Historia Augusta’s approbatory account of Probus having

done the same thing in the later third century hint that by the end of

the four th century there was concern in educated circles about the

recruitment of external Germani into their ‘own’ regiments.194 Yet

recruits into such units would, in the longer term, have been even

more subject to the integrating pressures outlined in respect of

Models I and II. Immediately enrolled on the permanent strength,

they could have soon been transferred far from home, as happened to

the Bucinobantes.

A Wnal model, Model VI, may be used to postulate a situation in

which, in changed circumstances, the Empire handed over the guard-

ing of the Rhine to Germanic kings and princes. Roman allies thus

became Roman agents, serving in Roman installations on Roman

soil. Though this is more a feature of the Wfth century than the

188 Below 304–5. 189 Below 248–9. 190 Below 286–7, 304–5.

191 Below 304–9. 192 Below 318. 193 Below 324.

194 Above 160–1.

174 Service

fourth,195 the process may have begun before 400. This is suggested

by W nds of artefacts dated to the later fourth century in the military

cemetery at Sponeck, near Breisach. These are indistinguishable from

those of neighbouring Alamannic settlements, and may indicate that

the men of the garrison were recruited from Alamanni or had ver y

close links with local villages.196 However, there are problems in this

interpretation. The presence of Alamannic artefacts could result from

the garrison’s having been supplied by the local community, in the

process adopting something of its lifestyle (including permanent

liaisons with its womenfolk). Further, even if the garrison were

Alamannic, our current inability to localize Alamannic artefacts

means that we cannot tell if it was recruited locally. Finally, thanks

to Roman views on the recruitment of barbarians during the fourth

century, it is likely that any local recruits, if there were any, were fed

into what was still at least nominally a Roman unit.

Alamannic warriors served Rome in signiWcant numbers from the

late third to the early Wfth century, from almost as soon as both were

in close contact to the beginning of the dissolution of the western

Empire. Unlike that of Alamannic oYcers, the pattern of recruitment

of other ranks continued unchanged throughout this period. What

was the proportion of Alamannic recruits in the Roman army as a

whole? This is impossible to establish with any precision, but the

impression is that it was not overwhelming. In the sole area for which

one can perform simple calculations, the Notitia Dignitatum’s listing

of guards-regiments and comitatenses, Alamannic representation

is not high. In A. H. M. Jones’ tables, the overall proportion of

Alamannic units is 3 per cent including the Mattiaci and 1 per

cent without them.197 This, together with the fact that known Ala-

mannic Wrst-class regiments (the auxilia palatina) were infantry

regiments, provides no comfort for Okamura’s suggestion that

Aurelius Victor’s description of the third-century Alamanni as a

gens populosa ex equo miriWce pugnans was an anachronism, based

on his knowledge of their contemporary army service.198 The absence

195 Below 327.

196 Fingerlin (1990: 123), (1997a: 104); Steuer (1997a: 152–3); Bu

cker (1999: 160,

217). Cf. below 328.

197 Jones (1964: 364 Tables V, VI).

198 Aurelius Victor, Caesares 21.2; Okamura (1984: 90).

Service 175

of archaeological evidence for the Alamanni numerous or keen

horsemen likewise tells against the idea that Victor’s statement

reXected fourth-century experience of them.199 In Ammianus’

account of Constantius II’s 355 campaign, there is passing mention

of Lentiensian cavalry, but the associated Wghting does not seem to

have amounted to a cavalry battle.200 Victor was probably referring

cryptically to Julian’s recent victory at Strasbourg.201 Finally, it

should now be clear that we must resist the temptation to treat

Roman regimental names as tribal names.

Appreciation of the extent and nature of Alamannic service as a

whole helps in understanding the events of the following chapters.

For example, as we have seen, Latinus, Agilo and Scudilo were,

though nominally Alamanni, probably raised in the Roman Empire.

They pursued careers in the Roman army, probably Constans’ army

of Illyricum, which came to Constantius II through Vetranio, but

possibly even Constantius’ own army of the east. In culture, experi-

ence and outlook they were a long way from Alamannia. However, if

their origins lay in the relatively few proximate pagi that Rome knew

best, they may have been able to maintain blood ties with their home

communities. One can envisage how, when Constantius II reached

Italy, the idea of using these connections to turn the Alamanni on

Decentius and Magnentius came from ‘cour t’ Alamanni like Agilo;

and a key intermediary in this could have been Vadomarius.202

Ammianus’ report that in 354 Latinus, Agilo and Scudilo ‘betrayed’

Constantius’ invasion plans to ‘their own people’ may contain a grain

of truth: 203 that they originated from the region over the Rhine from

Kaiseraugst, the pagi of Gundobadus and Vadomarius. However,

their ‘treason’ was no more than their involvement in negotiation,

aimed at securing peace without bloodshed on the High Rhine.204

199 Keller (1993: 90). Cf. Whittaker (1994: 214), on a similar lack of evidence for

the Huns as horse-riding nomads; below 340.

200 AM 15.4.12. Below 210–11.

201 Above 141–2.

202 Above 155–6, below 207.

203 AM 14.10.8: populares suos.

204 Below 204, 207.

176 Service

ConXict 285–355

Fighting—by Rome against Alamanni and by Alamanni against

Rome—features most prominently in the literature, ancient and

modern. We can now examine this against the background of

Alamannic development established in preceding chapters.

Precisely how one constructs the overall picture is determined by

the extent to which one believes that the Alamanni posed a real

danger to the security of the Empire. I propose that, whatever either

side may have felt or claimed, there was no substantial ‘Germanic

threat’ on the Rhine from the later third to the early Wfth centuries.

Alamanni and Franks were never a menace to Rome. The Empire

exploited them much more than they could ever have damaged it.1

A more sceptical view of the Germanic menace derives from a

more questioning attitude to the sources, in particular from a more

critical assessment of Ammianus Marcellinus. Since the publication

of Edward Thompson’s pioneering study in 1947, many scholars

have contributed to the identiWcation and examination of the vagar-

ies of Ammianus’ narrative, and of what he does not say quite as

much as what he does.2 Stroheker, for example, noted Ammianus’

intense dislike of Alamanni, and his constant projection of a negative

image of these people, especially in respect of their duplicity.3

Stroheker explained this by reference to Ammianus’ feelings of

resentment against Constantius II for his rough treatment of Roman

1 Drinkwater (1996: 26–8); cf. above 125, 133–5, below 360.

2 See Drinkwater (1997: 1 n.5). Fundamental is now Sabbah (1978). Among

English-language studies, par ticular mention should be made of Elliott (1983) and

Barnes (1998). Cf. most recently Rollinger (2001).

3 Stroheker (1961: 32–3, 48–9).