BMCR 2007.02.04

Hostages and Hostage-Taking in the Roman Empire

, Hostages and hostage-taking in the Roman Empire. Cambridge: Cambridge University Press, 2006. xiv, 291 pages : illustrations, map ; 24 cm. ISBN 0521861837. $80.00.

In the current age of terrorism the theme of hostages immediately captivates. As an instrument of international relations, hostages vanished in the eighteenth century, only to see extra-legal hostage-taking re-emerge in the last decades. Hostage-taking by common criminals is an historical constant, as is the implicit use of hostages in the various chess games of international, domestic, and even corporate or familial politics, whenever coercion comes into play. The precise condition of individuals used extra-legally as pawns to guarantee bona fides or to coerce behavior is subject to chameleonic change from different perspectives and can be as ambiguous as discerning freedom-fighters from terrorists — a problem that plagues this study.

Joel Allen (hereafter “A”) attempts a “history of the history of hostage-taking” (p. 26) 200 BCE-CE 200.1 The book, whose readership will be devotees of cultural studies, family history, gender studies, and literary-theory-inspired historiography, should have been entitled “The Rhetoric of. . .” Students of international law and political or military history will find only supplements to previous studies.2 For A, the theme of hostages, a murky jumble of unclear definitions and conflicting sources (pp. 16-17, 23-25, 35), finds clarification in his perceived common set of motifs and metaphors in Roman writers and artists. This book, a considerably revised Yale dissertation (1999), belongs to the new Yale school of Roman military history, which already features the works of Kimberly Kagan and Susan Mattern-Parkes,3 who served both as an outsider reader for the dissertation and apparently an advisor for the revisions. Her influence is evident (pp. xiii, 27-28).

Defining the work’s thesis proved difficult — always an ominous bellwether — and requires thirty-seven pages (= “Introduction”). Indeed, the reader receives no hint of the book’s intent or its relationship to earlier work for the initial eleven pages. It then emerges that A wants to examine Roman perceptions of hostages in order to understand Roman expectations about non-Romans and how peoples on the periphery could be controlled. For A, hostages were at the center of Roman thinking about empire: noble or royal males, twelve to forty years old and detained at or near the center of power, could be forged into a new “overclass” (jargon borrowed from recent discourse on globalization) to collude with the Roman exercise of power. Hence Romanization of hostages and subsequent romanization of the hostages’ native populations was a strategic aim: hostages represented an inexpensive, low-risk method of winning territory or extending Roman influence (p. 224). The discussion, largely based on a small number of cases (the Antigonid Demetrius, son of Philip V; the Seleucid Demetrius I Soter; Polybius; Juba II; and Phraates IV’s sons) is primarily concerned with the hostage as a “type”: the “hybrid” character of a young elite male transported into a foreign environment, where he experienced inferiority in various power relationships. These relationships are examined in individual chapters: “Creditor-Collateral” (chapter two), “Host-Guest” (chapter three), “Conqueror-Trophy” (chapter four), “Father-Son” (chapter five), “Teacher-Student” (chapter six), and “Masculine-Feminine” (chapter seven).

The curious ground rules of this study must preface a review of contents. A’s pursuit of the hostage as a “type” renders irrelevant whether an individual is legally a hostage, i.e., delivered by treaty or held as surety to another form of agreement, or voluntarily in Rome for an education. For A, the fact that Roman political superiority fostered a desire to learn Latin and Roman customs constituted a form of compulsion. A rejects Elbern’s alleged classification (n. 2) of hostages according to types of treaties (such as foedus aequum or foedus iniquum), because Rome’s inconsistent treatment of hostages did not correspond to specific types of treaties. Elbern, however, made no such elaborate claims about treatment, but only provided lists of references classified by types of treaties. Similarly, use or non-use of ὅμηρος / Latin obses does not define a hostage, although A foregoes a systematic presentation of philological evidence and eschews distinguishing metaphoric from technical use of these terms or euphemisms. These Greek and Latin terms do not correspond to A’s notion of a hostage “type.” A maximizes the elasticity of hostage as a concept: some prisoners of war (at A’s discretion) are “hostages,” as are political detainees (not legally hostages) like the 1,000 Achaeans after Pydna. Criminal use of hostages is excluded (pp. 17-18), although A does not observe his own rule. All of A’s “hostages” faced a stick/carrot (violence/release) dilemma during their “detention,” which produced a shift (variously real or feigned) to a pro-Roman attitude (pp. 32-33), despite A’s awareness (pp. 53-54) that the Roman government did not execute hostages 200 BCE-CE 200. Granted the concept of hostage is slippery, but A’s hybrid type strikes this reader as somewhat contrived.

Further, A gives equal worth in assessing Roman perceptions to all sources 200 BCE-CE 200, whether Greek or Latin, historians or non-historians. Exempla from the regal or early Republican period (for instance Livy or Dionysius of Halicarnassus) are assumed to reflect Augustan values rather than to have any relevance for their own time. Likewise exempla in Greek of Greek events before 200 BCE (in Plutarch, for instance) are considered valid for Roman views. For example (pp. 149-51; cf. 175), the coincidence of schools and hostages at Plut. Sert. 14, Tac. Agr. 20.2-21.2, and Suet. Calig. 45.2 is taken to demonstrate a common Roman belief c. 100 CE in the re-education of whole populations through hostages. Only Suetonius’ passage shows formal hostages in a school; Tacitus’ hostages at 20.3 become the students of 21.2 through special pleading (p. 150 n. 4); and Sertorius’ school, as Plutarch indicates, was only a pretext for collecting hostages from the Spanish tribes. Yet A believes Sertorius was interested in romanization and omits his position as a rebel. Schools established to educate illiterate provincials (for example those of Agricola and at Augustodunum, pp. 163-65) certainly aimed at romanization, but need have nothing to do with hostages.

In general, cases are often accepted at face value with little or no analysis of their own historical context or Quellenforschung, although supposed “tropes” are dismissed automatically. The exemplum of Frontin. Strat. 2.11.6 is summarily dismissed because of its appearance in a stratagem collection (p. 61 with n. 72), but here A ignores both the purpose of such collections to provide examples for imitation and his own principle that all sources 200 BCE-CE 200 are of equal value. Even conceding the penchant of Roman historians to re-write and invent Roman history, A’s methodology is more literary than historical. A has also been seduced by Mattern-Parkes’ dubious notion of the common attitudes and opinions of a “decision-making elite.” Hence A’s view that all authors 200 BCE-CE 200 think alike.4 The view that one socio-economic class produces only one set of opinions flies in the face of the heterogeneity of views evident in legislative bodies like the U.S. Congress, generally populated by a single socio-economic class of politicians.

The influence of the recent fascination with “spectacle studies” also strikes the reader. A wants to know what the Roman “man/woman on the street” thinks about hostages and presents himself as a keen student of the average Roman’s psyche. Hence a frequency of such phrases as “would have been,” “must have,” and “may have been” in recounting Roman reactions to hostages. But how often would the average Roman on the street have encountered a hostage of the elite “overclass”? Were there no foreigners in Rome except hostages? Was Rome even in the Augustan age still a town of country bumpkins rather than a cosmopolitan city? When a Roman saw a hostage, would he/she immediately think of him as a victim from a powerless community and the object of a lusty Roman official’s sexual gratification (p. 179)? Much is assumed and much exaggerated.

These caveats aside, chapters two to four address various hostage scenarios (discussed in earlier works) under the rubrics of “Creditor-Collateral,” “Host-Guest,” and “Conqueror-Trophy.” An effort to tie the origins of Roman hostage-taking to debt-slavery is not argued in detail and ignores that hostages in international relations were already a standard practice. A distinct Roman origin of the practice is unlikely. A stresses Roman inconsistency in the practice of taking and detaining hostages, but insists on the existence in theory of a veiled threat to the hostages’ lives despite no governmental executions of official hostages after 212 BCE. For A, Romans always combined the role of a gracious host with ruthless application of blackmail. Romans preferred hostages to formal treaties and could make, as in the prelude to the Third Punic War, additional demands even after the delivery of hostages. As often in this work, A’s narrow focus on hostages misses the bigger picture. Before the third century CE Rome rarely contracted formal treaties and always aimed at maximum flexibility in international relations. Legal chicanery in international affairs was typically Roman, although Rome scrupulously observed fides and the letter of the law. Rome in 149 BCE, as in 219-218 BCE, had already decided on war with Carthage and needed only suitable pretexts.

A denies (pp. 93-94) the sacrosanctity of hostages, which he (predictably) makes relative to an author’s choice in relating a narrative. But Elbern is surely correct (n. 2: pp. 109-110): hostages held by the Romans under treaty would be sacrosanct, if the treaty’s stipulations were observed; any sacrosanctity of hostages after a deditio fell to the victor’s generosity, since the surrendered were legally prisoners of war and the victor’s property. A’s misunderstanding of this point later leads (pp. 190-91) to a dubious argument that female hostages (even those delivered under an agreement) would be raped, but his cases mix capture of another state’s hostages (Scipio Africanus’ protection of Spanish females, who were Carthaginian hostages and became Roman POWs, 209 BCE, deception and possible violation of philia (Cleonymus of Sparta at Metapontum, c. 303 BCE, and an individual’s criminal activity unrelated to military operations (a centurion’s kidnapping and ransom of the Galatian Chiomara, 189 BCE).

A correctly stresses how hostages (though many of the examples are really POWs) enhanced a general’s prestige and demonstrates a fondness of the Cornelii Scipiones for association with hostage-taking, a trait dating from Scipio Barbatus’ activities in the Third Samnite War. A’s assertion, however, that hostages were a requirement for a triumph is not demonstrated and ignores the rule that the number of the enemy slain was the deciding criterion, although scholars debate specifics. A cites Domitian’s contrivance of fake German POWs (Tac. Agr. 39.1; “hostages” according to A) three times in the same chapter and makes hostages the basis of Domitian’s envy of Agricola. Yet this is hardly what Tacitus says ( Agr. 38.3-39.1).

Paternalism, the theme of chapter five (“Father-Son,”), is found in examples of paternal or filial vocabulary used in describing the relationship of hostages to individual Romans or the emperor. Although formal adoptions of hostages in Roman families are unknown and marriages between peregrini/-ae and members of the imperial family were strictly forbidden, A imagines, on the analogy of the combination of the Aemilii, Fabii, and the Cornelii in the early second century BCE, the Roman construction of an international network with the emperor as an “imperial paterfamilias.” Juba II (originally a POW and not a hostage, as A concedes at p. 21 but soon forgets), Herod the Great’s various sons, and Phraates IV’s four sons (like Herod’s sons, wards of Augustus and not legally hostages) provide much of the evidence. A perhaps lends too much credence to the ideological picture of an international family on the Ara Pacis as the true backdrop of Augustan foreign policy.5 A’s emphasis (pp. 105-107, 134-35) on the Ara Pacis‘s “foreign children” as confirmation of his hostages as a “type” (specific identifications are denied) contrasts with Elbern’s view (p. 108 n. 2) that Romans did not want small children as hostages.

A’s thesis that the re-education of hostages served the strategic aim of romanization of peripheral non-Roman populations occupies chapter six, “Teacher-Student,” where such re-education is equated with “cultural imperialism.” The cultural effects of a prolonged residence in Italy is cleverly shown in A’s connection of Plautus’ Poenulus with the 300 Carthaginian hostages after Zama (pp. 52, 161-63), although A seems unaware of Dubuisson’s study of just how romanized Polybius became — a work with damaging repercussions for A’s view of Polybius (see below).6 For A, this strategy of re-education demonstrates a Roman effort to create a subtle long-term policy for managing the empire rather than a reliance on brute force (p. 150), but A overstates Rome’s emphasis on force. Tiberius preferred consilium to vis (Tac. Ann. 2.26.3; 6.32.1). Earlier works have already discussed in detail the theme of the education at Rome of foreign princes (cited p. 151 n. 6), though not all of them were hostages, even by A’s definition. A does not address Elbern’s denial (pp. 118-24 n. 2) of romanization of hostages as a goal of Roman foreign policy and exaggerates the “schools for hostages” of Sertorius and Agricola into attempts at instant romanization.

Nevertheless, much of this chapter stresses the failure of this alleged Roman policy and draws the obvious conclusion that the romanization of hostages did not produce romanization of peripheral populations. Proof is found in the (quite limited) excavations of Syrian Antioch, where Livy’s claim that Antiochus IV romanized the city is not confirmed, and in the sparse evidence for Juba II’s additions of Roman elements at Mauretanian Caesarea. The futile efforts to enthrone in Parthia an Arsacid prince raised at Rome (never attempted without the solicitation of a disgruntled Parthian faction) are well documented and already much discussed.7 For Vonones I (r. c. 8-12 CE) A seeks confirmation of Tacitus’ account of his Roman ways on a drachma depicting Vonones with a shorter hair-style than was typical for Arsacid kings. The drachma is also notable for its legend explicitly declaring his defeat of Artabanus II — a rare contemporary reference on a Parthian coin. The suggestion would be more convincing if this drachma (like Vonones’ tetradrachms) had been minted at Seleucia-on-the-Tigris, a Greek city. But the peculiar drachma was minted at Median Ecbatana, a less likely site for romanized presentation of an Arsacid monarch.8

A completes his survey of hostage themes with gender issues, “Masculine-Feminine” (chapter seven) — no doubt the weakest chapter. Discussions of female hostages, brides, and marriage alliances have appeared elsewhere. A seeks, however, to sexualize the hostage experience: the hostage’s inferiority in power implies assumption of a female role to the hostage-taker’s masculine position. For A, authors are free to characterize hostage recipients as rapists, donors as pimps, and hostages as sluts, and Roman (and Greek) authors expected sexual exploitation of hostages. The Roman tradition could even whitewash “heroic” rapes (like the Sabine women). A’s cute (but ungrammatical) translation of Juv. 2.166-70 concludes the discussion with Zalaces, an Armenian hostage in Rome, whose penchant for pederasty illustrates Rome’s corrupting influence. For unknown reasons A doubts Zalaces’ historicity, but an Armenian hostage under Domitian is not unlikely.

Two chapters featuring Polybius and Tacitus and an epilogue conclude the work. A’s discussion of Polybius, his most interesting chapter, argues that Polybius reversed the expected roles of a hostage (teacher/father to Scipio Aemilianus: Plb. 31.24.9-25.1) and duped his Roman hosts into believing him subservient to Roman interests. Unfortunately, Polybius’ relationship with Scipio Aemilianus coincides with an influx of Hellenism into the Roman ruling class and need have nothing to do with a reversal of a hostage’s roles. Not all Greek intellectuals coming to Italy at that time were hostages or detainees. Yet A correctly discerns Polybius’ advocacy of a middle course between subservience and resistance. Polybius, always an Achaean patriot, by no means approved of all Roman conduct after Pydna. Like Thucydides, Polybius prized intelligence and was a realist about Roman power. A assumes, however, that the bulk of the Histories was written, free from Roman influence, back home in Megalopolis after 150 BCE. Although opinions vary, only books one to fifteen seem to have been composed before 146 BCE and the publication of books three and four dates to 145 or 144 BCE. Thus (contrary to A) Polybius’ self-justifications for his role in the events of 149 BCE (Plb. 36.11) would not have been contemporary with the events, or, if they were, they were not published at that time. Further, Polybius’ occasional switches to first-person narrative and emphasis on his personal role in some events (such as the escape of the hostage Demetrius I to Syria) may not be Polybius’ assertions of his independence from Roman views, but rather declarations of an eyewitness’s accuracy or simply self-promotion, not unlike the “celebrity status” of some political exiles on the payrolls of American universities. More significantly, Dubuisson (n. 6) has shown from Polybius’ Greek his conversion to Roman modes of thought and has emphasized that Polybius’ audience was not exclusively Greek. A’s chapter adds fuel to the fire of the debate on Polybius’ attitude toward Rome.

In contrast to A’s stimulating discussion of Polybius, chapter nine, “Tacitus on Hostage-Taking and Heroism,” is superficial and poorly researched. A returns to the topic of Parthian hostages to show how Tacitus exposes the failed policy of romanized hostages, for which the successes of Tacitus’ heroes, Germanicus and Corbulo, compensated. Throughout the work, despite consultation of some standard handbooks, A does not display any real understanding of Armenia and Parthia, the larger political contexts of receipt of hostages from these states, or (in some cases) proper prosopographical investigation. All inadequacies cannot be exposed here. For Armenians, A seems not to know (pp. 90 with n. 73, 117) who Tigranes II the Great was, nor the significance of Rome’s detention (though not as a hostage) of his ambitious son Tigranes (married to the Parthian Phraates III’s daughter). Zeno/Artaxias III (pp. 230-31) does not provide a parallel to Vonones I. Zeno, never a hostage in Armenia (as A thinks), knew Armenian customs (Tac. Ann. 2.56.2) from exposure to Armenians west of the Euphrates in Cappadocia and the later Armenia Minor, part of the realm of his step-father, Archelaus of Cappadocia, once married to an Armenian princess. Zeno’s placement on the Armenian throne represented Roman desperation in finding a new candidate with at least a pretension of Artaxiad blood in his veins. Remarkably, A’s discussion of Zeno, Germanicus, and the Syrian governor Calpurnius Piso (pp. 229-32) follows Tacitus’ account blindly without even a hint (perhaps gratefully) at the new SC de Cn. Pisone, already discussed ad nauseam.9

The reader does not fare better with A’s discussion of Parthian hostages. The famous Parthian settlement (not a treaty) of 20 BCE, which A correctly recognizes as a sham, was not about hostages (pp. 84-86, 124). A’s confused discussion misses the wider context of the settlement and he is unaware of a fuller exposure of this propagandistic event already in print.10 A’s skepticism (pp. 145-47) about Phraates IV’s motive for delivering unwanted sons to Augustus in 10/9 BCE is unjustified, as his son by the slave-girl Musa/Theamusa later (c. 2 BCE-4 CE) ruled as Phraates V/Phraataces (not Augustus’ ward Phraates, whom A erroneously calls Phraates V: pp. 176 n. 100, 232-33). Despite occasional citation of Kahrstedt’s book11, the whole discussion of Roman-backed Parthian pretenders from Vonones I to Meherdates (Vonones I’s son, but unknown to A) suffers from a lack of appreciation of Parthian dynastic politics and civil wars, caused in part by resistance to Artabanus II’s problematic Arsacid credentials. A’s treatment of hostages in Corbulo’s Armenian campaigns continues the application of his metaphors and motifs, which, tiresome by this point, carry (for an historian) little conviction for his interpretation of events. The reviewer, however, cannot help but notice that his article on the chronology of those campaigns is twice favorably cited (pp. 73 n. 15, 152 n. 12), although A persists (pp. 118, 119, 124, 241) in retaining Tacitus’ erroneous date (owed to his annalistic presentation) of 54 for Vologaeses I’s delivery of hostages to Corbulo and Ummidius Quadratus, rather than the actual date of (probably late) 55.12

The work concludes with an epilogue, “The Altar and the Column,” in which A argues that a difference in the presentation of hostages on the Ara Pacis and Trajan’s Column signals a change in Roman attitudes about the periphery from “welcoming condescension to that of stampeding military ambition” (p. 252). That the Antonine age was no longer the age of Augustus will surprise few Roman historians. Like others, A assumes that the young Dacian males at Cichorius’ scene CXLVI on the column are Decebalus’ sons, whom A thinks became “hostages” (really POWs). We do not know that. Since Dacia was annexed as a province, survival of Decebalus’ house served no purpose and the youths, if indeed Decebalus’ sons, were probably executed, just as the Romans obliterated Sarmizegethusa Regia and all traces of native Dacian religion.

A’s book will raise consciousness about hostages. The volume is nicely produced and almost flawlessly edited with few typographical errors (p. 176: misspelling of a coin legend in Greek; pp. 57 nn. 62-63, 237-38 nn. 54-55: transposition of notes?). One wishes, however, that the press’s readers had done a better job of earning their fees and guiding the author. Much here is problematic, especially the “hybrid” view of hostages, trying too hard to make a thesis work, and too narrowly focusing on hostages without considering the wider political and strategic contexts of events. Apparently it did not occur to A to define what is “Roman” about the Roman use of hostages by distinguishing Roman practice from that in the Greek world and the Near East, or — even for rhetoric — what distinguishes Latin presentation from Greek. Of course A assumes no differences between Greek and Latin views 200 BCE-CE 200: all is Roman. Students of ideology and rhetoric may find more value here than historians.


1. The reviewer is obliged to confess that he once instructed the author (second semester, freshman year), but had no subsequent role in his undergraduate or graduate education.

2. C. L. Walker, Hostages in the Roman Republic (diss. University of North Carolina/Chapel Hill, 1980); various papers of M. J. Moscovich, e.g., “Hostage Princes and Roman Imperialism in the Second Century BC,” Echos du monde classique 27 (1983) 297-309; and S. Elbern, “Geiseln in Rom,” Athenaeum 68 (1990) 97-140, a convenient overview. C. Phillipson’s The International Law and Custom of Ancient Greece and Rome, 2 vols. (London 1911) remains useful. A is unaware of D. J. Bederman, International Law in Antiquity (Cambridge 2001).

3. S. Mattern, Rome and the Enemy: Imperial Strategy in the Principate (Berkeley/Los Angeles 1999); K. Kagan, The Eye of Command (Ann Arbor 2005).

4. A, pp. 27-28, citing Mattern (above n. 3) 2-4. Mattern’s book miraculously escaped critical assessment in anglophone reviews.

5. See B. Severy, Augustus and the Family at the Birth of the Roman Empire (London/New York 2003) 110.

6. M. Dubuisson, Le Latin de Polybe: Les implications historiques d’un cas de bilinguisme (Paris 1985).

7. A is unaware of E. Dabrowa, “Les premiers ‘otages’ parthes à Rome,” Folia Orientalia 24 (1987) 63-71, as well as other works with much the same approach as A in attempting to calculate the average Roman’s reaction to Parthian princes at Rome: J. Wiesehofer, “Die ‘Sklaven des Kaisers’ und der Kopf des Crassus. Römische Bilder des Osten und parthische Bilder des Westens in augusteischer Zeit,” in P. Freeman et al. (eds.), Limes XVIII: Proceedings of the XVIIIth International Congress of Roman Frontier Studies held in Amman, Jordan (September 2000) (Oxford 2002) 293-300; R. M. Schneider, “Die Faszination des Feindes: Bilder der Parther und des Orients in Rom,” in J. Wiesehofer (ed.), Das Partherreich und seine Zeugnisse/The Arsacid Empire: Sources and Documentation, Historia Einzelschriften 122 (Stuttgart 1998) 95-146; and H. Sonnabend, Fremdenbild und Politik: Vorstellungen der Rmer von Ägypten und den Partherreich in der späten Republik und frühen Kaiserzeit (Frankfurt a.M. 1986).

8. D. Sellwood, An Introduction to the Coinage of Parthia (London 1980) 195 no. 60.5.

9. On the Piso text see W. Eck, “Cheating the Public, or: Tacitus Vindicated,” SCI 21 (2002) 149-64; on Tacitus’ misrepresentation of the military activities of Piso and his wife Plancina, see E. L. Wheeler, “The Laxity of Syrian Legions,” in D. Kennedy (ed.), The Roman Army in the East, JRA Suppl.18 (Ann Arbor 1996) 262-64.

10. See E. L Wheeler, “Roman Treaties with Parthia: Vo+lkerrecht or Power Politics? in Freeman et al. (above n. 7) 287-92; on the connections of the settlement of 20 BCE with the failed Arabian expedition of Aelius Gallus, see C. Marek, “Die Expedition des Aelius Gallus nach Arabien im Jahre 25 v. Chr.,” Chiron 23 (1993) 121-56; A. Luther, ” Medo nectis catenas ? Die Expedition des Aelius Gallus in Rahmen der augusteischen Partherpolitik,” OTerr 5 (1999) 157-82.

11. U. Kahrstedt, Artabanos III und seine Erben (Bern 1950). A is unaware of M. Schottky’s response to Kahrstedt: “Parther, Meder und Hyrkanier. Eine Untersuchung der dynastischen und geographischen Verflectungen im Iran des 1. Jhs. n. Chr.” AMI N.F. 24 (1991) 61-134.

12. E. L. Wheeler, “The Chronology of Corbulo in Armenia,” Klio 79 (1997) 383-97.