pire’s twenty-eight legions under the command of the future emperor
Tiberius. A few years later, after Arminius’s revolt, the same com-
mander invaded Germany with eight legions—the entire army of the
Rhine, some 40,000 men, plus an auxiliary army of unknown strength
Counterinsurgency 167
but probably equal or greater in number. The Jewish revolt of 66 tied
up four legions and a total of about 50,000 troops for several years.16
But this analysis oversimplifies. Insurgency under the Roman Em-
pire was not a series of discrete events and responses; insurgency is
attested in all periods of Roman history and in many locations. Armed
revolt and conventional warfare were only two of its aspects. How and
with what permanent institutions did the Romans prevent, manage,
and respond to resistance day by day?
Some insurgents used terror as a tactic. The example most histori-
ans point to is a group that Josephus calls the sicarii. According to him,
they arose in Jerusalem in the 50s CE; they assassinated their targets in
daytime, often under cover of a crowded festival; and they took their
name from the type of sickled dagger they used. Like some modern
terrorists, the sicarii chose symbolic targets; their first victim was the
high priest Jonathan, “symbol of the sacerdotal aristocracy’s collabora-
tion with the alien Roman rulers and its exploitation of the people.”17
They also attacked wealthy landowners in the countryside and de-
stroyed their property, again apparently as a warning and deterrent to
collaboration with the Romans. According to Josephus, the sicarii were
ideologically motivated adherents of the “fourth philosophy,” which
advocated rebellion from the Romans on religious grounds.18
Josephus calls the sicarii “bandits” ( lestai), and he uses the same de-
rogatory term to refer to other insurgents besides the sicarii. Banditry
was a very widespread phenomenon in the empire, and even when it
lacked ideological aspects it can fairly be described as insurgency be-
cause of the Roman government’s oft-stated interest in eliminating it.
Although some generals and emperors claimed in their propaganda
that they had eradicated banditry from the territory under their rule, in
fact references to banditry pervade literary and documentary evidence
from all periods of the Roman Empire and from every provenance, in-
cluding Rome’s most ancient provinces and especially including Italy.19
The Greek and Roman terms for banditry usually signified predatory
rural violence, which might include raiding, rustling, kidnap, extor-
tion, highway robbery, and murder; because of banditry, travel was
very dangerous in the Roman world, even over short distances. Bandits
often came from the margins of society; they might be slave shepherds,
168 Mattern
pastoralists who lived on the margins of civilization (this is especially
well attested in Sicily and southern Italy), retired soldiers, or deserters
from the army.
Large groups living within the Roman Empire, including certain
tribes and certain ethnic units, were also classified as bandits by ancient
sources. Among the most notable of the latter were the Boukoloi of the
swamps around the Nile Delta. In Cilicia, in southeast Asia Minor, the
Isaurians of the highlands never were incorporated into the Roman
Empire but maintained their own language, tribal organization around
strongmen, and predatory conflict with the more urbanized lowlands
throughout the Roman period and throughout history; the Romans
negotiated and fought small-scale wars with them as against a foreign
enemy.
In Judaea, which is the only area for which a large body of literary
evidence exists over several centuries, banditry was endemic in all peri-
ods of Roman rule. Much banditry in that province had an ideological
element: locals perceived bandits as champions of Jewish freedom from
Rome. The distinction between banditry and guerrilla warfare in this
region is difficult to draw. Networks of rock-cut caves in some settle-
ments in Judaea could be headquarters for bandits or hiding places for
guerrilla rebels, perhaps in connection with the Bar-Kokhba revolt, or
these populations might have overlapped substantially. They are diffi-
cult to date and may have functioned over decades or centuries.20
The difference between a bandit, a tribal chief, a petty king, or the
leader of a rebellion could be open to interpretation; many individuals
are located in more than one of these categories by the ancient sources.
Thus, large geographic areas within the Roman Empire were indepen-
dent of Roman authority, mostly highlands with mobile populations
and inaccessible terrain. There were pockets of Rome’s empire where
its writ did not run.
The analogy between ancient banditry and modern terrorism is
loose. Ideology might or might not figure in ancient banditry, which
was largely economic in nature, and even where resistance was ethnic
or ideological, terror (in the sense of random, unpredictable violence
designed to create instability and fear) is poorly attested as a tactic, ex-
cept for the sicarii. Again with the exception of the sicarii, those labeled
Counterinsurgency 169
bandits in antiquity operated in the countryside, often based in inacces-
sible highlands, and not in the crowded cities that are the preferred tar-
gets for modern terrorists. However, there are also significant parallels.
Bandits were not perceived as common criminals; they were enemies
of the state, against whom the Romans waged war. This was not, as
they thought, war in its truest sense, as against a legitimate state; in-
stead, they conceived the war against bandits as guerrilla, bush, or (as
we now say) “asymmetrical” warfare, though they did not use those
terms. Bandits were not imagined as working alone. They commanded
the loyalty and resources of a local community that would aid and abet
them, or else they enjoyed the protection of powerful landowners,
who employed them for their own purposes: to rob and rustle from,
kidnap, and bully their neighbors, and as shock troops in the continual
competition for land and power in which they were all engaged. Some
landlords amassed what amounted to private armies of bandits, more
than a match for anything the Roman state could muster locally.
Bandits with connections to the local community or to a landlord
were best apprehended by stealth, information, and betrayal. One
caught them by “hunting” them; our sources describe posses of sol-
diers, hired hit men, and local vigilantes.21 Professional or semiprofes-
sional experts in bandit hunting are attested; some of these were not
easily distinguished from bandits themselves.22 The law encouraged or
required communities and individuals to hunt and surrender bandits,
and individuals could kill bandits with legal impunity.23 Roman law also
took aim at those who abetted, protected, or received stolen goods
from bandits, probably with little effect.24 Some of the empire’s wealth-
; iest and most powerful individuals benefited the most from banditry,
perhaps including many members of the Roman senatorial class. The
simplistic rhetoric that opposed banditry to legitimate power masked a
situation in which a rudimentary Roman state operated in the shadow
of, or as part of, a much more complex and highly developed system of
personal power that included bandits and their protectors.
The army policed for bandits. Augustus and Tiberius maintained
military detachments or stationes throughout Italy to control the
banditry, which had escalated during the civil wars that ended the re-
public.25 In some areas roads were militarized for protection against
170 Mattern
banditry, and some structures perceived as frontier systems were con-
structed and staffed to control banditry. In Cilicia, the Romans eventu-
ally (in the third and fourth centuries) fortified an inner frontier against
the Isaurian bandits of the highlands.26
Against large groups of bandits or outlaw peoples, Roman gover-
nors and their subordinates waged small wars. Cicero led brutal puni-
tive expeditions against bandits during his term as governor of Cilicia,
during which he razed villages and exterminated their inhabitants, and
took hostages from one settlement after a long siege, but without long-
term success. Tacitus describes further campaigns in the region under
delegates of the governor of Syria in the 30s CE and in 51.27
Roman governors and emperors sometimes tried to neutralize ban-
dit gangs by hiring them to enforce order, or by recruiting bandits into
the army individually or en masse.28 More often, Roman commanders
negotiated diplomatically with bandits. Cicero established a tie of hos-
pitium or hospitality with one Isaurian strongman (Cicero calls him a
“tyrant”); this and other types of “ritualized friendship” were the main
instruments of Roman foreign and internal relations in the late repub-
lic and throughout much of Roman history generally. Cicero, Pompey,
and Mark Antony all successively recognized another Isaurian leader,
Tarkontidmotos, as a “friend” of Rome or of themselves.29
In the tiny mountainous province of Mauretania Tingitana in ex-
treme northwestern Morocco, Roman provincial governors negotiated
peace with highland chiefs; records of their ritualized agreements, in-
scribed on stone, are almost the only written evidence of Rome’s expe-
rience in the area. Outside the heavily militarized zone in the lowlands,
Roman cultural influence did not extend, though this was a region
surrounded by Roman provinces of long standing. While no surviving
sources refer to the highland population of Mauretania as bandits, the
analogy with Isauria is very striking.30
As with banditry, so with insurgency in general: the military factor
is important to the equation, but the army operated parallel to, and
to some extent within, a wider set of social relationships. The Roman
army was an army of occupation as well as of external aggression and
defense. This was especially true of the provinces of Spain, Britain,
Mauretania (modern Morocco), Syria, Palestine after the Jewish revolt
Counterinsurgency 171
of 66, and Egypt. In these provinces the army was stationed in urban
centers or dispersed throughout, rather than heavily concentrated on
frontiers (the province of Spain had no frontier, and Britain’s frontier
was very short).31 A few areas of the empire were intensively occupied,
notably Judaea, a very small territory that housed perhaps 20,000 troops
in the early second century CE, and Mauretania Tingitana was essen-
tially an armed camp occupied by 10,000 troops, with little evidence of
Roman influence outside the military zone.32 In both cases, intensive
occupation proved ineffective. Judaea’s garrison failed to prevent the
Bar-Kokhba revolt or to suppress endemic banditry in the province, and
Mauretania Tingitana was abandoned in the third century.
A complicating factor is that the army was not deployed from center
to periphery, as one people dominating many others, even though the
emperor claimed ultimate authority over the whole apparatus. Dur-
ing the first century CE the Roman army rapidly changed from a force
of Italian citizen-soldiers to one that was recruited from all over the
empire, not mainly from Italy. Legionaries mainly came from citizen
populations in the provinces, notably veteran colonies. But veteran col-
onies were not isolated, ethnically distinct communities; their citizen-
inhabitants might be veteran settlers or their freedmen, or remote
descendants of those settlers or of their freedmen, having mixed for
many generations with the local population. The descendants of retired
auxiliary soldiers were probably another important source of legion-
ary recruits. The auxiliary army, recruited entirely from noncitizens,
was much larger than the legionary army. Discharged after decades of
service, these soldiers acquired Roman citizenship on retirement and
usually settled in the regions in which they had been stationed. Thus
Rome’s army was recruited from among its subjects.33
Consider the situation in Judaea before the revolt of 66: when Herod
the Great ruled with Roman support (until his death in 4 BCE), he com-
manded a typical Hellenistic army of troops from local military settle-
ments.34 These troops were mainly Sarmatians, Idumaeans, Babylonian
Jewish archers, and ethnic Palestinian Jews, although Herod also had a
famous bodyguard of Germans, Thracians, and Gauls. The indigenous
settlements continued to supply the army that supported all of Herod’s
successors and formed the garrison of Judaea after 6 CE. Much of this
172 Mattern
army remained loyal to the Romans and fought with them to suppress
the revolt of 66 CE, under the command of Herod’s great-grandson,
King Agrippa II. Conversely, Herod probably modeled his army on the
Roman army and used some ethnic Roman officers, and it is likely that
many Jews were recruited into the Roman legionary or auxiliary army
both in his reign and later. But Mel Gibson’s portrait of Latin-speaking
soldiers in The Passion is inaccurate. The soldiers in the “Roman” gar-
rison of Judaea spoke Aramaic.
The army that the Romans initially sent against the rebels in 66,
under the command of Cestius Gallus, governor of Syria, included le-
gionary (i.e., Roman citizen) soldiers, auxiliary (i.e., noncitizen) sol-
diers, contingents from the armies of two local allied (“client”) kings,
and local Syrian militia. At the same time, the royal Jewish forces loyal
to Agrippa II moved against the rebels in three separate locations. Later,
the larger Roman army under the command of Vespasian numbered
55,000–60,000 legionary and auxiliary troops and 15,000 allied troops,
and included at least one Jewish high officer, Tiberius Julius Alexander
of Egypt.
In this example and in others that could be discussed from around
the empire, it is difficult to distinguish the Romans from their subjects,
a procedure made eve
n more complicated because natives, when en-
franchised, took Roman names, and many cannot be distinguished in
the historical record from ethnic Italians. For example, while several of
Herod’s military officers had Roman names, we do not know if they
were enfranchised Jews or soldiers imported or borrowed from Rome’s
legionary armies. Should we count all Roman citizens as “Roman,” in-
cluding King Herod and his successors and Josephus himself, and all
retired auxiliary soldiers, their freedmen, their descendants, and the de-
scendants of their freedmen? Should Paul of Tarsus, whose family had
acquired Roman citizenship from an unknown patron, therefore count
as Roman, along with any freedmen in his family? Should we rather
count only the staff sent from Italy, the procurator and his entourage?
Should we count only legionary soldiers—but there were no legions
deployed in Judaea before the revolt of 66? One scholar, writing on
Gaul, has described the term “Roman” as a social status, not an ethnic-
ity, and as such it was fluid: one could be more or less Roman, more or
Counterinsurgency 173
less enmeshed in the web of Roman culture and influence, and there
was no sharp line between ruler and subject.35
What most historians refer to simplistically as the conquest of Ju-
daea by Pompey in 66 BCE and its subjection to Rome through a series
of puppet kings, Josephus describes as a bafflingly complex process en-
twining Jewish dynastic intrigue, Roman civil war, a triangular set of
international relations and conflicts among Romans and Parthians, Ro-
mans and Jews, and Jews and Parthians, as well as a more local sphere of
relationships, especially between Judaea and Arabia. Every withdrawal
of Roman troops from the region saw a new uprising under a new can-
didate for leadership and a reassertion of local power bases until finally
a period of relative stability followed Herod’s defeat of Aristobulus in
37 BCE. Rome’s incorporation of the Greek East or Caesar’s conquest
of Gaul could be described in a similar way: only close attention to
tensions and conflicts indigenous to the area can adequately explain
Roman intervention and describe its results.36 The story of Roman im-
perialism is not the story of an invincible army deployed from the em-
pire’s center against a surrounding ring of hapless, less sophisticated
Makers of Ancient Strategy: From the Persian Wars to the Fall of Rome Page 27
