Erdkamp POLYBIUSLIVYALLIES 2007
Erdkamp POLYBIUSLIVYALLIES 2007
Erdkamp POLYBIUSLIVYALLIES 2007
Chapter Title: POLYBIUS AND LIVY ON THE ALLIES IN THE ROMAN ARMY
Chapter Author(s): Paul Erdkamp
Book Title: The Impact of the Roman Army (200 B.C. – A.D. 476)
Book Subtitle: Economic, Social, Political, Religious and Cultural Aspects
Book Editor(s): Lukas de Blois, Elio Lo Cascio
Published by: Brill. (2007)
Stable URL: https://www.jstor.org/stable/10.1163/j.ctv2gjwsqg.9
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
Brill is collaborating with JSTOR to digitize, preserve and extend access to The Impact of the
Roman Army (200 B.C. – A.D. 476)
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
POLYBIUS AND LIVY ON THE ALLIES
IN THE ROMAN ARMY
Paul Erdkamp*
From the fourth or third century until the beginning of the rst century bc,
Rome’s armies were also the armies of her allies. The socii and nomen
Latinum raised at least half of the soldiers that fought wars for Rome.
The Italic allies were clearly distinguished from the non-Italic troops,
such as Cretan archers or Numidian horsemen, by the fact that they
were governed by the formula togatorum. This can be concluded from their
‘denition’ in the lex agraria from 111 bc: socii nominisve Latini quibus ex
formula togatorum milites in terra Italia imperare solent. The formula togatorum
is seen as a dening element, distinguishing the Latin and Italic peoples
from Rome’s overseas allies. Although in the second century bc a con-
sciousness of Italy as a political and cultural unity gradually emerged,
it was still referred to as a military alliance of Roman citizens and allies
at the end of that century.1 The beginnings of this system remain in
the dark, due to the inadequacies of our sources. The foedus Cassianum
between Rome and the Latin League (traditionally dated to 493 bc)
supposedly established a federal army under Roman command, but
next to nothing is known about its functioning. The participation of the
allied peoples was based on the treaties between their communities and
Rome. The position of the Latin colonies was slightly different, because
their obligations were probably based on the lex coloniae governing each
Latin colony.2 We may assume that the role of the allies was re-dened
* I wish to thank John Rich, Luuk de Ligt and Simon Northwood for their many
valuable comments.
1
Thus, T. Hantos, Das römische Bundesgenossensystem in Italien (München 1983), 185;
P.A. Brunt, ‘Italian aims at the time of the Social War,’ in The fall of the Roman republic
and related essays (Oxford 1988), 113. On the term socii Italici, E. Gabba, ‘Rome and
Italy in the second century bc’, Cambridge Ancient History VIII (Cambridge 1989), 209
observes, “it is only in relation to the predominant partner, that is to say Rome, that
they are seen as a group and thus bear this title”.
2
W.V. Harris, Rome in Etruria and Umbria (Oxford 1971), 85ff.; H. Galsterer, Herrschaft
und Verwaltung im republikanischen Italien. Die Beziehungen Roms zu den italischen Gemeinden vom
Latinerfrieden 338 v.Chr. bis zum Bundesgenossenkrieg 91 v.Chr. (München 1976), 84ff., 101ff.;
Hantos 1983, op. cit. (n. 1), 150ff.; D.W. Baronowksi, ‘Roman treaties with communities
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
48 paul erdkamp
after the Latin War, but we know little about this or how the system
worked and developed during the next century.
In the absence of documentary evidence, we have to rely largely
on the literary sources. Dionysius of Halicarnassus (20.1.5) may offer
a rare glimpse of the structure of the Roman army at an early date,
when he tells us that at the battle of Ausculum against Pyrrhus (279
bc), Latins, Campanians, Sabines, Umbrians, Volsci and Frentani were
divided into units that were posted between the legions. This is the
more remarkable, as the Campanians were cives sine suffragio, who at
least at the end of the century served in the legions. However, there is
no evidence that corroborates Dionysius, and so it remains likely that
we are dealing with “eine reine Phantasieschilderung”.3 Livy is missing
until the Hannibalic War, while Polybius’ narrative of Roman wars
in books 1 and 2 is nearly silent on the role of the allies, except for
the famous survey of peoples joining Rome against the Gauls in 225
bc. His account of the Gallic threat leading to mass mobilisation of
citizens and allies in 225 bc sheds sudden light, mentioning the formula
togatorum and manpower gures, but the passage is notoriously fraught
with problems.
When the sources start to offer information on the allies, all seems
well-established. Livy gives us no general account of the Roman army,
comparable to book six in Polybius’ Histories, since he is not interested
in analysing the Roman state and its institutions. However, his narrative
of Roman war in books 21–45 seems to give a good view of the allied
system as it functioned during the half-century from 218 to 167 bc.
In these books, Livy mentions units consisting of allied peoples (the
Paeligni, Marrucini, Vestini and Marsi) and Latin colonies (Firmum,
Fregellae, Aesernia, Cremona and Placentia). Moreover, we also meet
cohorts from Gaul and Picenum,4 and cavalry units of Latin, Lucanian
and Etruscan horsemen. In 213 bc, for example, Vibius Accaus from
Paeligna, praefectus of a cohors Paeligna, is rewarded for bravery during a
battle near Beneventum against Hanno (Livy 25.14.13). In 204, four
cohortes sociorum Latini nominis are stationed as garrsion in Locri in Brut-
tium (29.19.9). Twice Livy mentions a praefectus socii operating inde-
of citizens,’ Classical Quarterly 38 (1988), 172–178; id., ‘Sub umbra foederis aequi’,
Phoenix 44 (1990), 345–369; D. Timpe, ‘Erwägungen zur jüngeren Annalistik’, Antike
und Abendland 25 (1979), 379ff.
3
Galsterer 1976, op. cit. (n. 2), 106.
4
Mobilised during an emergency levy. Livy 23.14.2.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 49
pendently against the enemy, losing the battle and his life.5 At Pydna,
Marrucinian, Paelignian and Vestinian cohorts fought side-by-side with
Samnite or Latin turmae (44.40.4ff.).
Polybius
5
Livy 25.1.3f.; 31.2.5ff.
6
Rejected by W.V. Harris, ‘The Italians and the empire’, in W.V. Harris, ed., The
imperialism of mid-republican Rome (Rome 1984), 89ff.
7
Gabba 1989, op. cit. (n. 1), 210.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
50 paul erdkamp
those who were overtaken to right or left along the intervals between the
lines. (Polybius 15.9.6–10)
Attention to the contribution of Rome’s Italian allies is limited to the
cavalry: the Italian cavalry is posted on the left wing. The posting of
the Numidian horse on the right wing, however, and the absence of
any Roman cavalry makes clear that the Italian cavalry in this case
in fact includes that of Roman citizens. Typically Polybius, one might
say: ‘Roman’ or ‘Italian’ is almost synonymous. This is also indicated
by his use of such a phrase as “the allied horse of the Romans” in his
account of the battle of Cannae (3.116.6). In his account of the battle
of Zama, the Roman army in general or parts of it are referred to in
such general terms as ‘the Romans’ or ‘the infantry’, and in technical
terms, such as hastati, principes and triarii. However, not once do we see
an explicit reference to the allied infantry.
The picture is the same in all Polybian battle accounts. Apart from the
non-Italic allies of Rome, such as the Aetolians or Spanish peoples, the
only part of the allied contribution that is ever referred to is the Italian
horse (sometimes in the sense of the Roman and allied Italian horse),
and that not even in every battle. Concerning the infantry, Polybius
mainly uses technical terms for units or differently armed types of
soldiers that do not distinguish between Romans or allies. The only
‘ethnic’ term he often uses is ‘the Romans’, but this obviously in most
cases means all those on the Roman side. In short, in his narrative
of military actions, Polybius seems to ignore the distinction between
Romans and allies as far as the infantry is concerned.
One obvious way to distinguish Romans from allies, it might be
argued, is by the usage of the term ‘legion’, since legions are manned
by Roman citizens. However, the term used by Polybius to indicate a
legion—stratopedon—has also much wider meanings.8 This is shown for
instance in 10.16.4, in which Polybius has the general observation that a
consular army consists of “two Roman stratopeda and two of the allies”.
Clearly, also the allies were organized in what Polybius calls a stratopedon.
The term can have at least three meanings in the histories of Polybius:
1. that of ‘army’ in general; 2. that of a large unit of heavy infantry in
Roman or other armies and 3. as the equivalent of ‘legion’. The Loeb
translation by W.R. Paton tends to translate the term as ‘legion,’ but
8
Cf. M. Gelzer, ‘Die Glaubwürdigkeit der bei Livius überlieferten Senatsbeschlüsse
über römische Truppenaufgebote’, Hermes 70 (1935), 283.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 51
that is misleading, since many times the term clearly refers to citizen
as well as allied troops.9
The same applies to Polybius’ terminology for the sub-units of the
stratopedon. The main term he uses is ‘semaia’, which is derived from the
word for the standard of a unit; the direct Latin equivalent is signum, not
only meaning standard, but also unit (cf. vexillum). Basically, the semaiai
are the tactical units grouped around a military standard. Polybius
mainly uses the term semaia as the Greek equivalent of manipulus. This
is most clearly shown in 6.24, where Polybius describes the structure
of the Roman legion. In battle, he writes, the three main lines are:
the hastati, the principes and the triarii. Each of these is divided in 10
companies, the maniples. “These companies are known as tagma, speira
or semaia.”10 The rest of the passage clearly shows that the three terms
are used as equivalent in meaning. It is only for the sake of variation
that Polybius sometimes uses semaia and speira in the same passage for
the same thing. (See for instance the rst lines of the battle array at
Zama quoted above.) In short, the main term for manipulus, which is
the primary tactical unit of the Roman legions in Polybian times, is
semaia, less often speira. Now, the point is that these terms might also
refer to allied troops. This is for instance clearly shown in 6.30.4, where
Polybius mentions the semaiai of the allied infantry. A wider meaning
of the term is also indicated in 15.4.4, where it is used as “ten units of
Roman horse and foot”. In Latin, one cannot use manipulus as referring
to horse. In sum, even such a basic term, which is used as the Greek
equivalent of the Latin manipulus, and which occurs numerous times
in Polybius’ narrative of military operations, refers to allies as well as
Romans.
One last item may be allowed to conrm this point: Polybius informs
us in book six that a Roman legion had six military tribunes (chilíarchoi),
its main ofcers. The allied equivalent of the military tribune, he writes,
is the prefect of the allies, in Latin the praefectus sociorum.11 It may be
noted that the prefects of the allies were Roman citizens, just as the
military tribunes were. Hence, there is little difference between them in
this regard. The interesting part is that Polybius very often—in almost
9
Polybius, The histories, transl. W.R. Paton (Cambridge MA 1922).
10
Polybius 6.24.5.
11
Polybius 6.26.5; 6.37.8.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
52 paul erdkamp
12
Some scholars have pointed out that Polybius’ account of the Roman army in book
six “is written entirely from the point of view of a set of important but subordinate
ofcers, the military tribunes”. E. Rawson, ‘The literary sources for the pre-Marian
army’, Papers of the British School at Rome 39 (1971), 13–31. Reprinted in Roman culture and
society. Collected papers (Oxford 1991), 36 (with older literature). They therefore conclude
that Polybius used some kind of instruction to military tribunes as a starting point.
The existence of such documents is unproven. Moreover, reliance on written instruc-
tions would seem unlikely in an era in which young nobles had seen many campaigns
before their term as ofcers and in which many posts of military tribunes were taken
by ex-praetors and ex-consuls. Finally, military tribunes are not only prominent in book
six, they also play a large role in Polybius’ account of campaigns.
13
Polybius 6.26.7; 6.30.2. According to P.A. Brunt, Italian manpower (Oxford 1971),
678, the ratio varied, pointing to Appian and Livy.
14
Polybius 6.26.7–9.
15
Polybius 6.31.2–6.32.6.
16
Polybius 6.39.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 53
17
On this passage, see A.M. Eckstein, ‘Physis and nomos. Polybius, the Romans, and
Cato the Elder’, in P. Cartledge et al., eds., Hellenistic constructs. Essays in culture, history
and historiography (Berkeley 1997), 85.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
54 paul erdkamp
At the third signal, the leading troops must advance and set the whole
camp in motion. As a rule the extraordinarii are placed at the head of the
column; after them come the right wing of the allies and behind them
their pack animals. Next in the order is the rst of the Roman legions
with its baggage behind it, after which comes the second followed by its
pack animals, together with the baggage train of the allies, who bring up
the rear, the left wing of the allies providing the rearguard. (6.40.3–5)
A different marching order is used when the army expects to meet the
enemy:
In this case the army advances in three parallel columns, consisting of the
hastati, principes and triarii. The baggage trains of the leading semaiai are
placed in front, those of the second immediately behind, and so on, the
baggage trains being interspersed between the bodies of ghting troops.
With this information, if the column should be threatened, the troops
face to the right or left, according to the direction from which the attack
comes, and can then quickly get clear of the baggage and confront the
enemy. (6.40.11)
While the rst passage explicitly distinguishes between the units of
allies and Romans, the battle formation of the second passage implies
a homogenously organized army consisting of three columns. It is clear
that the allies were part of the columns of hastati, principes and triarii.
In his description of the formation of the consular army, Polybius
tells us explicitly that the allies are organized by Roman ofcers. The
magistrates of the allied communities were rst instructed to send a
certain number of troops. Polybius (6.21.5) adds that the magistrates
of the allies used the same procedure as the Romans to choose their
recruits, each community appointing a commander and a paymaster.
At the nal stage, the recruits selected from among the citizens and
those from the Latin and other allies came together and were divided
into units. About this nal stage, Polybius writes:
The allies having now assembled also at the same places as the Romans,
their organization and command are undertaken by the ofcers appointed
by the consuls known as praefecti sociorum and twelve in number. (Polybius
6.26.5)
Moreover, the ttest horsemen and foot soldiers among the allied troops
were selected by the Romans in order to form the extraordinarii (6.26.6).
Ethnic units of varying size would not t the army as he describes it,
not only in book six, but also in his account of the Roman army on
campaign. There is no contradiction between book six and his accounts
of Roman campaigns, which at the least does not support the wide-
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 55
spread supposition that book six describes the Roman army as it had
once functioned, but not anymore in Polybius’ time.18
One more important point should be made. The Roman light armed
soldiers (velites) were men of little property. The three lines of hastati,
principes and triarii represented different age classes. The hastati were
the youngest recruits, the principes men in their prime, the triarii were
seasoned veterans.19 If the allied units were organised in a similar man-
ner, their troops had to be distinguished along similar lines. In other
words, also allied troops had to be distinguished according to age and
property. This is conrmed by the brief remark that we have seen above
concerning the mobilisation of allied troops: the allied magistrates used
the same procedure to select their recruits.
We may conclude that Polybius did not ignore or underestimate the
role of the allies, on the contrary. However, he chose to depict them
as an integral part of the Roman army. It is clear now why he invari-
ably refers to the Roman side as the Romaioi and why he uses general
terms like the semaiai, or hastati, principes and triarii to refer to parts of
the Roman army. Since the allied units fought in exactly the same way
as the Roman maniples, there was little point in distinguishing between
them. The Roman armies that Polybius witnessed and described fought
as a single body, in which distinctions between citizens and allies were
simply not important.
18
Thus, Brunt 1971, op. cit. (n. 13), 625ff. Otherwise, Rawson 1991, op. cit.
(n. 12).
19
Polybius 6.21.7–10. On the interpretation, see recently N. Rosenstein, Rome at war.
Farms, families and death in the Middle Republic (Chapel Hill 2004), 141.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
56 paul erdkamp
of the battle near the Metaurus in 207 bc, for example, no reference
whatsoever is given to allied infantry or cavalry, this despite the fact
that the left and right wing (cornu) are regularly mentioned, while we
know that the allied alae are usually stationed on the wings of the
Roman battle formation. In short, in the majority of full-scale battle
narratives, including all the major battles of the Hannibalic War, no
mention is made of specic allied units. From this point of view, the
majority of Livy’s battle scenes closely resemble those of Polybius and
paint the same picture of a uniform Roman army. This surely reects
the absence of details on allied units in the sources that Livy in these
cases used.
Details concerning the allied contribution are conned to a limited
number of battle narratives, all of which have in common other char-
acteristic features, such as numbered legions and the regular mention
of legates. Remarkable about this type of battle narratives, which are
late-annalistic and can be securely ascribed to Valerius Antias, is the
pattern of its occurrence in Livy’s third, fourth and fth decades. Full-
scale narratives of this type are conned, rst, to the Italian theatre of
war during the central period of the Hannibalic War (213–207 bc) in
books 25–27, and, second, to the wars in Spain and northern Italy from
205 bc onwards in book 29–41. The last case in which ethnic allied
units and numbered legions occur in the extant books of Livy—the
battle of Pydna (168 bc)—may be exceptional, since, on the one hand,
there is clear Valerian material in Livy’s battle narrative of Pydna, but,
on the other hand, there is second-century evidence on specic allied
units playing a prominent role at Pydna.20
The following table summarises the occurrences of specic allied
units and/or numbered legions in Livy’s full-scale battle narratives.
In other words, these are the battles designated in this article as late-
annalistic.
20
A more detailed discussion of Livy’s late-annalistic battle scenes and the reliance
on Valerius Antias in these cases can be found in P. Erdkamp, ‘Late-annalistic battle-
scenes in Livy (books 21–44)’, Mnemosyne 59 (2006[a]), 525–563; id., ‘Valerius Antias
and Livy’s casualty statistics’, Studies in Latin literature and Roman history 13 (2006[b]),
166–182. In these publications it is shown that (1) Livy used the terminology of his
sources on army units; (2) late-annalistic battle scenes are limited to particular periods
and theatres of war; (3) characteristic features, such as numbered legions, allied units
and details on legates, tribunes and centurions are common features of casualty reports
and late-annalistic battle scenes; (4) battle scene and casualty report are often integrated;
(5) nearly all casualty reports can be securely ascribed to Valerius Antias.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 57
Livy 213 attack 3rd legion cohors Paeligna 6000 enemies killed,
25.13.11– on Punic cohors 7000 captured,
25.14.11 camp near sociorum et wagons and cattle
Beneventum civium legio
Livy 212 battle near 1st legion ala sinistra of 18,000 men, only
25.21.5– Herdonea 2,000 escaped
25.21.10
Livy 26.5.7– 211 battle near ‘auctores’ 6th legion equites 8000 men from
26.6.13 Capua others: legionum, Hannibal’s army and
no sig- equites 3000 Campanians
nicant sociorum killed, 15 signa from
battle Hannibal, 18 from
the Campanians
captured
Livy 27.1.6– 210 battle near 5th + 6th sinistra ala 11 tribuni militum
27.1.13 Herdonea legion 13,000 or 7000
Romanorum
sociorumque killed
Livy 209 battle in 18th legion sinistra ala 8000 enemies killed
27.14.2– Apulia (II) and 5 elephants, 1700
27.14.14 legionaries and 1300
sociorum killed, many
civium sociorumque
wounded
Livy 207 battle near 1st + 3rd equites legionis 8000 enemies killed,
27.41.1– Grumentum legion dextra ala 700 captured,
27.42.8 (Lucania) 9 signa captured,
4 elephants killed,
2 elephants captured,
500 Romanorum
sociorumque victores
killed
Livy 29.2.4– 205 battle 12th +13th Romani equites 13,000 enemies
29.2.18 against legion killed, 1800 captured,
Ilergetes e.a. 200 Romanorum
(Spain) sociorumque
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
58 paul erdkamp
Livy 194 attack on a Valerius 2nd legion equites 40,000 enemies killed
34.14.1– city (Spain) Antias Romanorum
34.15.9 Cato dextrae alae
only says
‘many’
(on casu-
alties
etc.)
Livy 35.4.6– 193 battle 2nd legion sinistra ala 14,000 enemies
35.5.14 against Boii extraordinarii killed, 1092 captured
(Gallia Cis.) dextra ala (also 721 horsemen
equites and 3 leaders),
legionum 212 signa and 102
equites alarum wagons captured,
5000 Romani and
socii killed, including
praefecti socium and
tribuni militum of the
2nd legion
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 59
Livy 41.1.5– 178 attack by 2nd + 3rd cohors 8000 enemies killed,
41.4.8 Histri (Italy) legion Placentina none captured,
equitatus 237 victores killed
legionis
The important point is that all the information that can be found on
the tactical functioning of allied units is conned to a limited number
of battles narratives, almost all deriving from the same late-annalistic
source.
In most cases, allies are referred to as the extraordinarii or the left or
right ala. See, for instance, Livy’s account of a battle against the Boii
in 193 bc:
The second legion was sent forward and the extraordinarii relieved. Then
the battle was restored, since fresh troops, a legion with full ranks, had
entered the ght. The ala sinistra was withdrawn from the battle and the
ala dextra took its place in the battle-line. (35.5.6)
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
60 paul erdkamp
As usual, Livy ends his account of the battle with a detailed list of
casualties and booty, the latter including military standards and wagons.
He notes that among the Roman dead, there were many allied prefects
and military tribunes from the 2nd legion.
There are similar cases when Livy briey mentions a battle without
an actual narrative. Two cases are especially interesting: in book 39,
Livy mentions an attack by Ligurians, as a result of which not only
4000 Roman soldiers were killed, but 3 military standards from the 2nd
legion and 19 standards of the allies were lost.21 Similarly, in book 42
Livy says that some authors mentioned a battle against Perseus:
There are those who declare that a great battle was fought on that day,
that 8,000 of the enemy were killed, among them Sopater and Antipater,
ofcers of the king, that there were taken alive about 2,800 men, and
that 27 military standards had been taken. Nor was the victory bloodless,
they say; more than 4,300 of the consul’s army fell, and ve standards
of the sinistra ala were lost. (Livy 42.66.9–10)
These instances clearly show that the numbered legions, the details
concerning allied contingents and the loss of military standards are
closely integrated. We can take this relation between numbered legions
and allied contingents even further, since all the battle scenes having
details on allied units also have numbered legions. Livy’s battle accounts
with numbered legions, details concerning allied units and military
standards in the lists of casualties and booty can be related to Valerius
Antias. We may be certain that Valerius Antias was Livy’s only source
for battle accounts of the late-annalistic type.22
Livy clearly distrusted some of the battles he found in the work of
Valerius Antias and the gures provided by this late-republican annal-
ist. Sometimes he dismisses outright the obvious falsehoods in his
predecessor’s account, but on occasion he summarises or copies the
information he found, telling the reader of the origin of this informa-
tion, and leaving it to the readers’ judgement whether to trust it or
not.23 The damaging depiction by Livy of Valerius Antias does not
inspire condence in the details on specic allied units that we nd in
this context.
21
Livy 39.20.6–8.
22
Erdkamp 2006[a+b], op. cit. (n. 20).
23
Livy 30.3.6; 30.29.7; 32.6.5ff.; 37.34.5; 38.48.1ff.; 42.11.1; 44.13.12ff.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 61
Livy’s casualty gures for the battle of Pydna are cited as evidence
by W.V. Harris for his statement that “the allies suffered more heavily in
combat than the citizen legions”.24 Besides Pydna, two other instances
support this hypothesis, both dealing with battles fought against the
Celtiberians in Spain (in 181 and 180 bc). In 40.32.7, Livy writes
that 200 Roman soldiers were killed, 830 allies and Latins, and 2400
‘external auxiliaries’. The number of casualties reects the previous
account: allied troops and 6000 Spanish auxiliaries played a special
role during the battle. In 40.40.13, 472 Roman soldiers, 1019 allies
and Latins and 3,000 auxiliary troops were killed. It is only rarely that
Livy gives separate numbers of casualties for citizen and allied troops. I
can only add one further instance, which Harris, unsurprisingly, fails to
mention: 1700 legionaries and 1300 allies fell during the second battle
of Marcellus in Apulia in 209 bc (Livy 27.14.14). If these gures are
somehow based on ofcial data, it is surprising that only four out of
approx. 80 cases distinguish between citizens and allies. Many cases say
‘Romans and allies,’ but give no separate gures. It is hard to explain
why gures specifying citizens and allies were only available for Pydna
and three obscure battles in Spain and Italy. Such detail is worrying.
All of these four instances include military standards among the booty.
In view of what we have seen on the reliability of late-annalistic battle
statistics, we should conclude that Harris’ statement is founded on
shaky evidence.
In 168 bc, the nal year of the Third Macedonian War, the armies of
the two powers camped near Pydna, the one under the command of
L. Aemilius Paullus, the other led by the Macedonian king Perseus.
There was a stream of no great size nearer the camp of the enemy,
from which both the Macedonians and the Romans were drawing water
after posting guards on either bank in order to accomplish this mission
safely. There were two cohorts on the Roman side, a Marrucinian and
a Paelignian, and two troops of Samnite cavalry under the command
of the legate Marcus Sergius Silus. Another xed outpost was stationed
before the camp under the legate Gaius Cluvius, composed of three
24
Harris 1984, op. cit. (n. 6), 97. Likewise, A. Keaveney, Rome and the unication of
Italy (London 1987), 15.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
62 paul erdkamp
cohorts, from Firmum, the Vestini and Cremona respectively, and two
troops of cavalry from Placentia and Aesernia. While there was quiet at
the river, since neither side took the offensive, about the ninth hour a
baggage-animal shied from the hands of his grooms and escaped towards
the other bank. While three soldiers were chasing him through the water,
which was about knee-deep, two Thracians dragged the animal from mid-
stream to their bank. The soldiers pursued them, killed one, recaptured
the animal, and retired to their post. There was a guard of eight hundred
Thracians on the Macedonian bank At rst a few of these, angry at the
killing of their fellow-countryman before their eyes, crossed the river in
pursuit of the killers, then more went, and nally the whole force, and
the guard . . . (Livy 44.40.4ff.)
There is a big lacuna in our text, since two pages of the manuscript
are missing, but the further description makes clear that the allied
troops came to blows with the forces on the Macedonian side of the
river, leading to a erce battle. The cohorts of Paeligni and Marrucini
fearlessly attacked the Macedonian phalanx. The Roman citizen-forces
consisted of the 1st and 2nd legion. When Livy’s narrative resumes, the
2nd legion attacked the centre of the Macedonian force. On the right
wing, Roman elephants and allied troops put the Macedonian wing to
ight, which was the rst step to a brilliant Roman victory.
The importance of this passage lies in the survey of allied contingents
on the Roman side, consisting of infantry cohorts from the Marrucini,
Paeligni, Firmum, the Vestini and Cremona, and cavalry units of Sam-
nium, Placentia and Aesernia.
Livy’s battle narrative of Pydna contains elements that are clearly
late-annalistic, such as the numbered legions. The presence of num-
bered legions in Livy’s account of Pydna shows that Livy did not simply
paraphrase Polybius’ narrative of the battle, as he did for instance on
Cynoscephalae (197 bc) and Thermopylae (191 bc).25 There are two
reasons for this assumption: Polybius never has numbered legions in
his battle accounts, and numbered legions make only one appearance
in battles in the East, viz. at Pydna. As we have seen above, numbered
legions are closely connected to Valerius Antias. In fact, they may be
seen as characteristic of Antias. Hence, Livy’s account partly stems
from Antias.
On the other hand, we may not simply assume that the allied con-
tingents originated in the work of Antias. The parallel account of the
25
Livy 33.7.4; 36.18.1.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 63
26
Cf. H. Nissen, Kritische Untersuchungen über die Quellen der vierten und fünften Dekade des
Livius (Berlin 1863), 300ff.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
64 paul erdkamp
not have this story, but it may have been in the lacuna. Noteworthy is
that Plutarch has no earlier mention of the Paeligni and Marrucini.
(4) The analysis of the Roman victory in both accounts is similar,
but with signicant differences. Livy assigns a crucial role to allied units
and elephants. Plutarch does not mention them. Plutarch has a detailed
tactical analysis, claiming that the rough terrain caused the formation of
the Macedonian phalanx to break down, which then fell victim to the
superior weaponry of the Romans in close combat. This analysis closely
resembles Polybius’ comparison of the strengths and weaknesses of the
Roman legions and Macedonian phalanx in battle.27 The niceties of
Plutarch’s tactical analysis are missing in Livy’s account. It is impossible
to say whether this is the result of his lack of interest in such matters,
or because the source Livy used did not have these elements.
(5) Livy notes that the Romans killed 20,000 Macedonians and lost
about 100 men, most of whom were Paelignians. Plutarch differs from
Livy in the number of Macedonians killed (25,000). The number of
Roman dead was 100 according to Poseidonius and 80 according to
Nasica. No gure for Polybius is given. We may observe, however, that
in his extant battle accounts, Polybius never species particular groups
among the Roman casualties.
All in all, there is little to support the hypothesis that Livy’s account
is largely Polybian, apart from the fact that normally Livy relies on
his Greek predecessor for Greek matters. The Paeligni and Marrucini
clearly are part of a tradition that was already present in the second
century bc and it is possible that they occurred in Polybius’ narrative.
However, it is impossible to say which elements were already pres-
ent in the second-century sources, and which elements derived from
Valerius Antias. Cohorts from Firmum, the Vestini and Cremona, and
cavalry units from Samnium, Aesernia and Placentia play no role in
what remains of Livy’s battle narrative and do not occur in Plutarch’s
account. A rm conclusion is not possible, but we may point out that
allied units are not normally a feature of Polybius, while they are of
Valerius Antias.
27
Polybius 18.31.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 65
Secure evidence for ethnic allied units does not emerge before the time
of Marius, but it is sparse: in the Jugurthine War, Marius commanded
a cohors Paeligna (Bell. Jug. 105.2), while two cohorts from Camerinum
fought at his side during the battle of Vercellae against the Cimbri
(Cicero, Balb. 46; Valerius Maximus 5.2.8). In the latter case, we are not
informed about their composition or role, since it is only said that these
men were rewarded Roman citizenship by Marius for their bravery.28
We learn more from Sallust’s Jugurthine War, which contains ve
references that are relevant to our discussion.
28
Plutarch, Marius 28.2 does not mention cohorts, but talks about 1,000 men from
Camerinum.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
66 paul erdkamp
29
I hope to have shown that Livy’s regular use of cohorts regarding the wars in
Spain is of little consequence. Contra M.J.V. Bell, ‘Tactical reform in the Roman
republican army’, Historia 14 (1965), 405.
30
Plutarchus, C. Gracchus 5.1. Recently, P. Erdkamp, ‘The transformation of the
Roman army in the second century bc’, in T. Ñaco del Hoyo et al., eds., War and ter-
ritory in the Roman world (Oxford 2006), 41–51.
31
L. Keppie, The making of the Roman army. From republic to empire (London 1998),
57ff.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 67
32
Timpe 1979, op. cit. (n. 2), 103ff.; S.P. Oakley, A commentary on Livy I (Oxford
1997), 75f.; U. Walter, ‘Opfer ihrer Ungleichzeitigkeit. Die Gesamtgeschichten im ersten
Jahrhundert v.Chr. und die fortdauernde Attraktivität des annalistischen Schemas’,
In U. Eigler et al., eds., Formen römischer Geschichtsschreibung von den Anfängen bis Livius
(Darmstadt 2003), 149ff.
33
Livy 3.5.13. Cf. Brunt 1971, op. cit. (n. 13), 695; R.A. Laroche, ‘Valerius Antias
and his numerical totals. A reappraisal’, Historia 26 (1977), 359; Id. ‘Valerius Antias as
Livy’s source for the number of military standards captured in battle in books I–X’,
Classica et mediaevalia 35 (1984), 95; G. Forsythe, ‘Dating and arranging the Roman
history of Valerius Antias’, in V.B. Gorman and E.W. Robinson, eds., Oikistes. Studies in
constitutions, colonies and military power in the ancient world (Leiden 2002), 110f.
34
Laroche 1984, op. cit. (n. 33), 102.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
68 paul erdkamp
A second part of the answer must have been the increased integra-
tion of Italic families in the political and cultural elite of Rome.35
Valerius Antias, just like Livy, did not come from Rome. There may
be little evidence to support the supposition of Dieter Timpe that
Valerius Antias and Claudius Quadrigarius were members of the Italic
municipal aristocracy, but we may agree with him that the rst-century
historians wrote for a widening audience that increasingly included
leading families of the former allies.36 Not only in politics, but also in
literature and art, we see individuals from the municipal aristocracies
coming to the fore. The losses amongst the elite of Rome in the age
of Sulla as a result of political prosecutions and civil wars were partly
compensated by the inux of leading Italic families.37 These families
had an outlook on Roman history that was not dominated by the city
of Rome. While Polybius and the early Roman historians emphasized
the homogeneous nature of the Roman army, in which Romans and
allies were fully integrated, it suited the interests of many rst-century
readers to spell out the role that allies had served in Roman history.
Later authors celebrated the Italian peoples even in their role as oppo-
nents of Rome. Take, for example, the following lines in which Ovid
(Am. 3.15.7–10) stresses his Paelignian background:
Mantua rejoices in her Vergil, Verona in her Catullus.
It is of the Paelignian race that I shall be said to be the glory,
the race whose desire for freedom had compelled it to take up honourable
arms,
when anxious Rome feared the allied bands.
Rome may have feared the Paeligni, but their cause was freedom and
their struggle thus honourable.38 An interesting parallel is offered by
35
Brunt 1988, op. cit. (n. 1), 117ff.; Gabba 1989, op. cit. (n. 1), 210ff. However,
H. Mouritsen, Italian unication. A study in ancient and modern historiography (London 1998),
59ff. is sceptic about the degree of integration or romanisation: “When looking at
Augustan Italy one is immediately struck by the remarkable homogeneity achieved in
most areas: language, urbanisation, and monumentalisation, institutions, cult, artistic
production, burial practice, epigraphic habits, etc. The obvious romanisation encoun-
tered in this period contrasts sharply with the barely traceable Roman inuence prior
to the Social War” (p. 81).
36
Timpe 1979, op. cit. (n. 2), 113f. Cf. T.P. Wiseman, Clio’s cosmetics. Three studies in
Greco-Roman literature (Leicester 1979), 23; K.-E. Petzold, ‘Zur Geschichte der römischen
Annalistik’, in W. Schuller, ed., Livius (Konstanz 1993), 170f.; Walter 2003, op. cit.
(n. 32), 141.
37
Galsterer 1976, op. cit. (n. 2), 145.
38
The example was taken and quoted from M. Pobjoy, ‘The rst Italia’, in E. Her-
ring and K. Lomas, eds., The emergence of state identities in Italy in the rst millennium BC
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 69
the historian Velleius Paterculus, who declared that the cause of the
rebellious allies in the Social War had been just.39
The fortune of the Italians was as cruel as their cause was just. For they
were seeking citizenship in the state whose power they were defending
by their arms. Every year and in every war they were furnishing a dou-
ble number of men, both of cavalry and of infantry, and yet were not
admitted to the rights of citizens in the state which, through their efforts,
had reached so high a position that it could look down upon men of the
same race and blood as foreigners and aliens. (Vell. Pat. 2.15.2.)40
To strengthen his point, he claimed that the allies had provided two
soldiers for every citizen soldier. This statement is certainly wrong,
but it served his purpose.41 It agreed with the Italic identity of the
historians and that of many of their readers to pay attention to the
non-Roman contribution to Roman history. The relationship between
identity and historiography was, of course, no new phenomenon: one
of the functions of Roman historiography was to dene and character-
ise Roman identity. Virtues and moral superiority were central to the
Roman understanding of themselves and their role in world history.
This understanding is reected in the emphasis on moral exempla in
Roman historiography.42 It was only a natural development that, when
the word ‘Roman’ expanded in meaning and steadily began to incor-
porate Italic peoples, historiography had to incorporate the identity of
these new ‘Romans’ as well.
The growing tension between Romans and allies at the end of the
second century bc, culminating in the Social War, surely increased the
interest in the role of the allies in Roman history. As E. Gabba stated:
(London 2000), 197. See also G. Bradley, ‘Iguvines, Umbrians and Romans. Ethnic
identity in central Italy’, in T.J. Cornell and K. Lomas, eds., Gender and ethnicity in ancient
Italy (London 1997), 62 on the “strong sense of regional heritage in the work of the
poet Propertius, who was a native of Asisium [Umbria]”.
39
This notwithstanding the fact that his greatgrandfather, Minatius Magius from
Aeculanum, had remained loyal to Rome, had raised a legion (!) among the Hirpini and
fought successfully against the insurgents. He was rewarded Roman citizenship. Velleius
Paterculus 2.16.2. M.H. Crawford, ‘Italy and Rome from Sulla to Augustus’, Cambridge
Ancient History X (Cambridge 1996), 417f. observes: “Velleius was well aware that the
Italian cause was just, but that loyalty to Rome was an overriding obligation”.
40
Pobjoy 2000, op. cit. (n. 38), 190: “Velleius’ presentation of the Italians as effec-
tively Roman in all but formal status.”
41
Brunt 1988, op. cit. (n. 1), 126: “This sounds like a relic of allied propaganda.”
Cf. Harris 1984, op. cit. (n. 6), 96, 99f.; Bradley 1997, op. cit. (n. 38), 60.
42
B. Feichtinger, ‘Ad maiorem gloriam Romae. Ideologie und Fiktion in der Histo-
riographie des Livius’, Latomus 51 (1992), 16ff.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
70 paul erdkamp
“The allies became increasingly aware that they had helped create an
empire in which they enjoyed only part of the fruits.”43 I suggest that
we see some of the emphasis on the allied role in Rome’ glorious past
in Livy’s account of Roman battles and of the Roman war-effort in
general. Two noteworthy examples may be highlighted: the Latin colony
of Fregellae and the allied people of the Paeligni.
Every educated rst-century Roman will have been aware of the
fact that Fregellae, founded as a Latin colony in 328 bc, was taken and
destroyed by a Roman army in 125 bc, after part of its citizens had
expressed serious discontent with Rome and possibly even had advocated
revolt.44 A contrasting picture of Fregellan loyalty arises from a few
events during the Hannibalic and later wars as described by Livy. In
the year 211 bc, Hannibal tried to deect some of the legions besieg-
ing Capua and marched towards Rome. In an attempt to slow down
Hannibal’s approach, and thus offering precious time to the Roman
authorities, the citizens of Fregellae demolished the bridge across the
Liris. Furthermore, a messenger from Fregellae, travelling day and night
and apparently bringing the rst news of Hannibal’s approach, caused
great panic in Rome.45 A few years later, a great crisis emerged in Rome
when 12 Latin colonies refused to contribute any more troops, saying
that their manpower and resources were exhausted by the past years
of ghting. The consuls summoned the envoys from the remaining 18
colonies and asked them to elucidate their position.
On behalf of the eighteen colonies Marcus Sextilius of Fregellae replied
that they had soldiers in readiness according to the treaty, and would give
more if more were needed, and would exert themselves to do whatever
else the Roman people might command and desire. (Livy 27.10.3.)
43
Gabba 1989, op. cit. (n. 1), 223.
44
On the causes of the rebellion of Fregellae, Keaveney 1987, op. cit. (n. 24), 64ff.;
Mouritsen 1998, op. cit. (n. 35), 118f. Galsterer 1976, op. cit. (n. 2), 179ff. argues that
many of its citizens were settled in the citizen-colony Fabrateria Nova, which replaced
Fregellae. In other words, the punishment was not as harsh at it seems, since the
destruction of the ‘political’ unity of Fregellae does not imply the destruction of its
entire population. Similar, E. Gabba, ‘Rome and Italy: the Social War’, Cambridge Ancient
History IX (Cambridge 1994), 105. However, Brunt 1988, op. cit. (n. 1), 96f. disagrees,
pointing out “of this there is no evidence” (p. 97). For the image of the peoples of the
Central Apennines, see E. Dench, From barbarians to new men. Greek, Roman and modern
perceptions of peoples of the Central Apennines (Oxford 1995).
45
Livy 26.9.3–6. Oakley 1997, op. cit. (n. 32), 83: these events are probably
unhistorical.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 71
The prominence of the city und its unwavering support of Rome could
not have been expressed more clearly. Their next appearance is during
a famous disaster: when Marcellus was killed by a troop of Numidians,
he was accompanied by 180 Etruscan and 40 Fregellan horsemen. The
negative outcome of this event actually emphasizes the courage and loy-
alty of the Fregellani, as Livy carefully distinguishes between the conduct
of the Etruscans and that of the Fregellans. While the Etruscans all ed
when the enemy appeared, the Fregellans defended Marcellus until the
end. Seeing that Marcellus was mortally wounded, the few who survived
turned to ight with Marcellus’s wounded colleague and his son.46 No
reproach is possible here. Finally, Livy mentions a particular version
on the capture of the son of P. Scipio from an unnamed source: he
had been captured by the enemy during the War against Antiochus III,
while being accompanied by a cavalry contingent from Fregellae. He
had fallen from his horse and had been captured together with a few
horsemen, so it does not appear that the fault lie with the horsemen
from Fregellae.47 It was the wealthier people among the allies that served
in the cavalry. Hence, it is the upper class of Fregellae that is presented
in a very favourable light in these anecdotes and events.
A second example is provided by the Paeligni, who were famous
in the rst century for the fact that their central town Cornium had
been renamed Italica at the start of the Social War and had served
as the capital of the revolt. Their rst great appearance in the third
decade is during the Roman attack on a Punic camp near Beneventum
in 213 bc. Interestingly, this battle, occurring early in book 25, is the
rst of the late-annalistic type in the third decade, characterised as it
is by numbered legions, legates, prefects—and, of course, an ethnic
allied unit. The latter plays a prominent role when the Roman consul
Fulvius, depicted as weak and indecisive, orders his troops, who are
eager to ght, to retreat.
Nearest to the enemy happened to be a Paelignian cohort, whose prefect
Vibius Accaus seized the banner and threw it over the enemy’s earthwork.
46
Livy 27.27.6f. Cf. D.-A. Kukofka, Süditalien im Zweiten Punischen Krieg (Frankfurt
am Main 1990), 116. To interpret these events as evidence of the anti-Roman atti-
tude of the Etruscan horsemen really goes too far. Harris 1971, op. cit. (n. 2), 137f.
contra D.W.L. van Son, ‘The disturbances in Etruria during the Second Punic War’,
Mnemosyne 16 (1963), 268; A.J. Pfg, ‘Die Haltung Etruriens im 2. punischen Krieg’,
Historia 15 (1966), 201f.
47
Livy 37.34.6.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
72 paul erdkamp
Then, with a curse upon himself and the cohort, if the enemy should
get possession of that banner, he was himself the rst to dash over the
trench and wall into the camp. And already the Paelignians were ghting
inside the wall, when from the other side of the camp, while Valerius
Flaccus, tribune of the soldiers of the third legion, was reproaching the
Romans for their cowardice in yielding to allies the honour of capturing
the camp. . . . (Livy 25.14.4–5.)
The centurion of the rst maniple follows the Paelignian’s example,
takes his unit’s standard (signum) and leads his troops into the ght. After
the victory, both ofcers are rewarded for their valour. Can it really
be coincidence that the military tribune—whose role, by the way, is
perfectly superuous—is named Valerius?48 In any case, the numbered
legions, ethnic unit and prominent role of military standards all point
in the same direction. We have already seen that the Paeligni play a
crucial role in Livy’s account of the battle of Pydna, which is reected
by the fact that most of the casualties on the Roman side were Paelig-
nians. Finally, we may note that in 205 bc, many Paelignians, besides
Marsians and Marrucini, volunteer for Scipio’s eet.49
It seems remarkable at least that peoples, who had been bitter
enemies just a few generations before, are depicted in Livy’s work in
such a positive way. One may also add the heroic action of Oblacus
Volsinius, leader of the Frentani, who in the battle of Heracleia cut
his way through the ranks until he was face to face with Pyrrhus and
nearly managed to kill him (Dionysius 19.12; Plutarch, Pyrrh. 16.8–10).
All these peoples and communities revolted against Rome at the start
of the rst century.50 Of course, the former foes of the Social War are
not the only ones that are depicted positively. Placentian knights and
units from Praeneste, Firmum or Cremona are also mentioned, for
48
On the tendency of Valerius Antias to introduce Valerii in major roles throughout
early Roman history, T.P. Wiseman, Roman drama and Roman history (Exeter 1998), 77ff.
Undoubtedly he also included Valerii in minor roles in later events.
49
Livy 28.45.19. One may add that, in 196 bc, a cohors Marsorum bears the brunt
of a battle against the Insubres (Livy 33.36.10), and that, according to one tradition, a
Samnite-lead force saved the army of Minucius from destruction by Hannibal’s troops
in 217 bc (22.24.10ff.). Note that in the rst case, Valerius Antias is mentioned as the
source of the casualty gures. Regarding the same passage, A. Ziolkowski, ‘Credibility
of numbers of battle captives in Livy, books XXI–XLV’, La Parola del Passato 45 (1990),
27 concludes that ‘no battle was fought at all’.
50
In Per. 72, a list is given of seven peoples that revolted against Rome, which included
the Picentes, Vestini, Marsi, Paeligni, Marrucini, Samnites and Lucani. Cf. Appian,
Bell. civ. 1.39. E.T. Salmon, ‘Notes on the Social War’, Transactions and Proceedings of the
American Philological Association 89 (1958), 159ff.; Galsterer 1976, op. cit. (n. 2), 189f.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
polybius and livy on the allies in the roman army 73
instance, but it would be hard in these cases to show the same deliber-
ate effort to depict the allies in a favourable light. It is impossible to
say whether the stories in favour of the rebellious allies were invented
when the conict between Rome and some of her allies emerged, or
possibly at a later date, or whether there were actually local traditions
that were transmitted somehow to Roman historiography. In the case
of Fregellae and the Paeligni, I want to suggest that the above pas-
sages are meant as a vindication of these prominent former enemies.
The existence of such traditions in the wake of the Social War may
be seen as the third element that partly explains the prominent role of
the allies in Livy’s late-annalistic battles.
Conclusion
In Polybius’ view, the allies fought side by side with citizen troops in
a uniformly organised army. In his battle narratives, he hardly ever
mentions the allies explicitly, let alone ethnic or local units. While the
homogeneous nature of the Roman army suited his conception of “the
solidity of the Roman state”, he did not need to adapt the rst to t
the latter. Polybius described the Roman army as he saw it. The pic-
ture is no different in many other battle narratives in the work of Livy,
where we may suppose that he relied on such other Roman historians
as Fabius Pictor or Coelius Antipater. We may observe that the Roman
army as depicted by Sallust is as uniform as that described by Polybius.
While the second-century Greek author could regard the role of the
allies as inherent in Rome’s harmonious unication of Italy and his
Roman colleagues saw no reason for special emphasis, the position
was much different for Roman historians of the rst century bc, who
felt the need for a more detailed and explicit role for the allies. The
disturbed relationship with the allies, and the Social War in particular,
played a role in this tendency, as witnessed by the deliberate effort to
paint a favourable picture of the role played by soldiers from Fregellae
and the Paeligni in the wars against Rome’s enemies.
The campaigns that Livy (solely or largely) based on Valerius Antias
are notoriously unreliable. Livy’s late-annalistic battles in Italy during the
years 213–207 bc include some dubious encounters between Marcellus
and Hannibal, whereas Appian notes that Marcellus achieved little.
The battles fought against the Ligurians, Gauls and Spaniards in the
early second century bc are equally untrustworthy. Some of the battles
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms
74 paul erdkamp
that Livy took from Valerius Antias never occurred at all. Clearly, the
passages in Livy’s battle accounts on allied units offer a very insecure
base for any theories on their role. The study of allied units should be
based on Polybius and the more trustworthy parts of Livy. While the
existence of ethnic or ‘local’ cohorts and cavalry units is beyond doubt,
our conclusion leaves us with little secure evidence for their functioning
in the period 218–168 bc.
This content downloaded from 94.122.147.210 on Tue, 22 Aug 2023 20:46:08 +00:00
All use subject to https://about.jstor.org/terms