Cheryl Walker, Hostages in Republican Rome
List of Abbreviations
1: Meaning and Purpose of Hostageship in the Greco-Roman World
2: The System of Hostage Regulations in Rome and the Greco-Roman World
3: Roman Conduct toward Foreign Hostages
4: The Termination of Hostageship
Appendix II A
- Cheryl Walker, Hostages in Republican Rome, Chapter 3: Roman Conduct toward Foreign Hostages [94-132], http://chs.harvard.edu/publications.sec/online_print_books.ssp. Center for Hellenic Studies, Washington, DC. 2005
Chapter 3: Roman Conduct toward Foreign Hostages [94-132]
A meticulous examination of the evidence concerning Roman conduct toward and treatment of foreign hostages is crucial to an understanding of Roman policy on hostages and its probable intended results. The difficulties involved in such an examination are threefold. First, material relating to the period of detention, when Roman authorities actually held the hostages, is scanty, and in many cases this material is based upon a single source. Thus we can neither assess the accuracy of the information nor compensate adequately for authorial bias.  Second, the evidence which we have concerning the period of detention has been interpreted, we must assume, in the light of the hostage’s later fate, and we can undoubtedly assume that the ancient writers frequently supplied suitable elaborations, emendations, suppressions, and inventions of fact. Last, there is no method for determining conclusively the degree to which the few individual cases about which something is known are typical of Roman practice in general.  This is due partly to the paucity of apposite data and partly to the nature of historical narration: ordinarily, it is the unusual rather than the commonplace which is described. We must therefore proceed with caution in a most careful analysis of the surviving evidence.
Nevertheless, since the ancient literary sources are all we have, we must also utilize them to the fullest. It is particularly unfortunate that, in addition to the problems mentioned above, information on only  two subjects emerges: the places of detention, and the privileges which governments allowed the representatives of foreign nations who were held as hostages. We are given almost no explicit information concerning domestic details, friendships among the hostages themselves or between hostages and Romans, or Roman attitudes toward hostages, although some scholars have speculated about individual hostages.  In these matters, therefore, we can only make inferences. First, however, we must concentrate on what can be stated with some confidence about the physical disposition of hostages after their delivery to the recipient government.
The place of detention to which the donor government sent hostages was reported more frequently than any other detail concerning their treatment. There is no obvious reason why this should be so, although it may suggest that the place of detention was of some interest and not entirely predictable. For cases in which the escape or recovery of hostages figured or in which the hostages’ whereabouts were otherwise important to the subsequent events (e.g., the display of hostages before the walls of Same, Livy 38.28.9), the place would have been a necessary part of the episode, but such cases form only a small portion (ten cases, or 14.5%) of the total.  Whatever the reason for the inclusion of this type of information, we must be grateful. 
TABLE IV: Distribution of Places of Detention According To Nationality Of Recipient State
|Nationality||In a Military Camp||In Subject or Allied Territory||In Own Territory||Other||Total|
|Greek||3 (13.6%)*||2 (9.1%)||11 (50.0%)||6 (27.3%)||22|
|Roman||17 (42.5%)||3 (7.5%)||20 (50.0%)**||0 (0.0%)||40|
|Other||1 (11.1%)***||3 (33.3%)||4 (44.5%)||1 (11.1%)||9|
* Of these three instances, all of which belong to the campaigns of Alexander, two (Frontinus Strategems 2.11.3; Polyaenus 4.3.30) attest the physical presence of hostages in Alexander’s camp, but the third, describing how the Spartans had so declined from their military excellence that they had sent hostages to Alexander (Aeschines Against Ctesiphon. 133), may be meant only figuratively, as Alexander’s regent Antipater appears to have conducted the negotiations with Sparta (Plutarch Moralia 235B). The testimony of Aeneas Tacticus (1.10.23) is not relevant here, because the hostages in the enemy camp would probably have been brought expressly for the siege; see pp.177-83 below for an analysis of this passage and broken agreements in general.
**This single incident is the legendary case of Cloelia and her escape from Porsenna. Its value is highly questionable, but some elements of the story are probably historical; see below, pp.263-274.
*** Included in this category is an unusual episode only fleetingly mentioned by Josephus (Jewish Antiquities 14.290-2). In 43 B.C. Malichus attempted to recover his son, who was a hostage in Tyre, and flee to Judaea. Since at this period Judaea was governed as an annex of the province of Syria, it is understandable that Judaean hostages were kept in the territory of Syria rather than at Rome. Also included is the school for Iberian boys established by Sertorius in his capital and last bastion, Osca (Plutarch Sertorius14.2).
Even a cursory survey of Table IV reveals some significant differences between Greek and Roman practices in the matter of place of detention. The greatest dissimilarity, that of the marked Roman preference for accepting hostages in their own camp, can be easily traced to a difference between Greek and Roman procedures in negotiation. A  Roman magistrate usually was empowered with both military and diplomatic authority and could represent, subject to ratification, the entire Roman government. The Greek governments more often communicated between capitals by means of embassies, without intermediate negotiations carried out by the Greek commander in the field.  The relatively short distances between many warring Greek nations may also account for the scarcity of hostages held in Greek military camps, for while lengthy travel time was ordinarily required between Rome and her enemies, at least from the third century B.C. on (precisely the period when our sources are fullest), no such difficulty existed for the neighboring states of Greece.
One might well assume that a Roman magistrate would have kept foreign hostages in his camp only as a temporary measure, since providing shelter, supplies, and guards for potentially hostile aliens, particularly if they were numerous,  would have been distinctly inconvenient. This would have been especially true in the event of protracted campaigns against several peoples; the presence of one tribe’s hostages would not necessarily have prevented an attack by that tribe’s neighbors. Moreover, in the absence of a standing army or garrison, some other provisions for the hostages’ disposition would have been necessary before the army disbanded and the magistrate’s term of office expired. And some hostages are indeed attested as being elsewhere after their initial delivery to the camp (e.g., the hostages sent by Antiochus III the Great of Syria to Ephesus, who later appeared in Rome, Polybius 21.17.11; Livy 37.45.20; Summaries 41; 42.6.9., etc.). How long would this temporary detention in a camp have lasted? Although we do not possess any absolute dates for their detention in a military environment, none  seem to have remained long. In general, it is reasonable to postulate that hostages would have been kept in the camp itself or close at hand in the newly conquered territory until the entire campaign conducted by that magistrate had ended.  Some historical instances can be adduced in support of this hypothesis: in 189 B.C. the Roman commander displayed the Samean hostages before the walls of Same to induce the city’s second surrender to him (Livy 38.28.9), as Alexander had paraded hostages before an Indian city as proof of his honorable treatment toward surrendered people (Polyaenus 4.3.30);  in 61 B.C. Albanian and Iberian hostages who had first been dispatched to Pompey in Asia later graced his triumphal procession in Rome (Plutarch Pompey 45.5; Dio 37.2.7). The behavior of Caesar in Gaul indicates the essential accuracy of this hypothesis. In addition to the four cases of hostages sent directly to Caesar (Gallic War 3.27.1; 4.18.3; 8.20.2; 8.46.2), and those sent to the town of Samarobriva, which served as a temporary headquarters for the Roman army (Gallic War 5.47.2), Noviodunum, a town of the Aedui, is mentioned as a repository for Gallic hostages (Gallic War 7.55.2; Dio 40.38.2).  Under the protection of a Roman garrison in allied territory, these hostages would have been securely in Roman power, had not the Aedui defected to join Vercingetorix. Although no source declares that Gallic hostages swelled the procession walking before Caesar’s triumphal chariot, it seems very likely that such was the case, to judge from the precedents of Flamininus in 193 B.C. (Eutropius 4.2.3; Orosius 4.20.2; Livy 34.52.9) and Pompey in 61 B.C. (Plutarch Pompey 45.4; Appian Mithridatic Wars 117). 
Why would a Roman commander wait until the conclusion of his entire campaign to send his hostages to Rome? I surmise that the general would have preferred to keep hostages close at hand until he was sure  that all resistance had ended and that the agreement had been ratified in good faith by both governments; their presence could guarantee swift restoration or retaliation, if either should prove necessary. Further, the frequent detachment of sufficient troops to guide and safeguard the hostages on the way to Rome might easily have hampered later military operations. Nor is the greater dramatic effect of the total number of hostages displayed at once likely to have been overlooked by the ambitious general eager for popular acclaim. Thus it seems fair to conclude that hostages may have spent up to a year or perhaps more in the “temporary” confines of the camp or garrisoned town.
It is obvious that, for both the Greeks and the Romans, it was unusual to detain hostages within the newly subjected territory. Two of the three Roman examples of this procedure belong to the Caesarean conquest of Gaul and have already been examined; the third, the removal of Penestini and Parthini hostages to Appollonia and Epidamnus respectively (Livy 43.31.2-3) may be likewise a “temporary” measure or simply the coastal towns from which they immediately took ship for Rome.  The Greek incidents are puzzling. Agathocles, the tyrant of Syracuse, left the Bruttian hostages under a garrison on the Italian mainland; shortly thereafter, the Bruttians rose up, defeated the garrison, and recovered the hostages (Diodorus 21.8). Why Agathocles failed to consider this possibility and to act accordingly is rather mysterious, for one must conclude that either Diodorus’ account is erroneous or that Agathocles was quite foolishly trusting in the integrity of the Bruttians, in the strength of his mercenary garrison or in the restraining power of the Bruttians’ fear for the hostages. In the only other Greek instance, Demetrius I Soter of Syria allowed  the Judaean hostages to live in the citadel of Jerusalem under guard (Josephus Jewish Antiquities 13.17); it is possible, although it cannot be proved, that Demetrius took into account Jewish dietary restrictions, for Demetrius, unlike his uncle Antiochus IV Epiphanes, seems to have avoided giving religious offense where he could.
Although detention within enemy territory is unusual in the Greco-Roman world, it does occur in other Mediterranean cultures. Despite the paucity of examples from such cultures, which so limits the basis of comparison that valid generalizations are impossible, it is curious that the Carthaginians are three times mentioned in ancient sources as maintaining hostages for a long time in the vicinity of their native tribes. The earliest reference is found in Diodorus Siculus (23.5) for the year 263/2 B.C., when the Carthaginians, fearing that the Tyndarians might betray their city to the Romans, exacted hostages from Tyndaris and removed them to Lilybaeum. While the Punic stronghold in the western part of the island was certainly a safe repository and indeed remained so until the end of the First Punic War, nevertheless it is interesting and perhaps significant that no non-African hostages are attested in Carthage itself.  Even more interesting is the Punic treatment of Iberian hostages. Since the historical validity of the Abilyx episode has been disputed and it has been considered by some scholars  a mere duplicate of a later incident, it is appropriate to examine first the historically genuine episode. In 209 B.C., Cartagena, the Punic capital of the central part of the peninsula, was seized in a brilliantly conceived and directed raid by Roman troops under P. Cornelius Scipio, later surnamed Africanus. Besides the money, booty, supplies, and other goods which Scipio thereby acquired for the  Romans or at least kept his Punic foes from using, Scipio gained control of the Iberian hostages surrendered to the Punic commanders during their subjugation of Spain. Scipio then restored these hostages to their families in return for their tribes’ alliance to the Roman cause.  It is incontrovertible that in this case a Punic garrison in a major Punic outpost was intended to guard the hostages from tribes within striking distance of Cartagena and, since there is no ground for supposing their eventual transfer to Carthage, that this place of detention was permanent. 
An earlier incident is said to have occurred in 217 B.C., when the Iberian Abilyx deceived Bostar, the Punic commander at Saguntum, into entrusting him with the restoration of the hostage children there, an act designed to promote the popularity of the Carthaginian cause in the face of the Roman invasion. Instead of delivering the children and the presents which Bostar had given him to the parents, however, Abilyx brought both hostages and gifts to the Roman commanders Publius and Gnaeus Cornelius Scipio, who then received full credit for their subsequent restoration (Polybius 3.98-99; Livy 22.22.4-18; Zonaras 9.1). Critics of the story have observed that the similarity between these incidents--the restoration of hostages through the clemency of the Scipios--is clear, that the fullness of detail in the Abilyx story is itself suspicious, and that the location of the hostages in Saguntum is unlikely in view of the superior suitability of Cartagena. 
Nevertheless, none of these objections is entirely convincing.  Certainly a basic thread is common to both incidents; but it is significant that the means by which the hostages were secured for the Romans are totally disparate, as are the emphases of the two anecdotes. The  character and activity of Scipio are the pivotal points of the Cartagena episode, whereas the older Scipios are essentially incidental figures in the story of Abilyx’s treachery.  Nor should the difference in conclusion escape notice; the elder Scipios gave away the advantage which their physical control of the hostages bestowed, but Scipio bartered them in return for alliances. To the charge of suspicious fullness no direct counter can be made, although critics must eventually answer why the authority of Polybius ought to be ignored in this case; the refutation of explicit information found in Polybius should not be undertaken lightly and without substantial cause. The final objection appears to be the most tenuous. Given the absence of examples other than the seizure of Cartagena and the Tyndarians detained at Lilybaeum,  it is very difficult to assert anything about Carthaginian practice or theory; as mentioned earlier, Agathocles’ policy toward the Bruttians appears to be neither logical nor customary, and yet there is no reason to disavow the incident.  Until the Abilyx episode can be convincingly proven or refuted, it seems better to accept it tentatively as at least partially historical.
From Table IV it emerges that another possibility, in addition to military encampments and subject territory, was available, although the Romans seem never to have employed it. In six cases, recipient Greek states entrusted hostages to a third party otherwise independent of the contracted agreement. The authority for the two earliest cases is quite late, and the story of the Samian hostages detained at Lemnos on behalf of Athens in 441/0 B.C. (Diodorus 12.27.2-4; Plutarch Pericles 25.1) could be fictitious.  But no such doubt can be easily cast on the remaining four incidents, for they belong to a period whose historicity  is well illuminated, and all three are attested in authorities relatively close in time to the events they described.
In 418 B.C., Arcadian hostages were kept in Arcadian Orchomenus by the Lacedaemonians (Thucydides 5.61.5); in 219 B.C., the Achaean League agreed to an alliance with Messene on condition that the children of the Messenian ambassadors to the League be hostages in Sparta (Polybius 4.9.5); in 173 B.C., to quell the internal dissension of the Aetolian League, both creditors and debtors agreed to surrender hostages, who were then detained in Corinth (Livy 42.5.12); finally, in 168 B.C., Perseus chose Cnossus, in Crete, as the place of detention for the hostages Eumenes had agreed to provide as sureties for his efforts to negotiate a peace between Perseus and the Romans in 168 B.C. (Polybius 29.8.6; Livy 44.25.8). The factor which determined the need for a third government also explains why the Romans did not use this alternative: the relative parity of the two contracting governments. If the Achaean League accepted the alliance with Messene, Messene would have become a member of the League, and the League would have been placed in the equivocal position of simultaneously dominating Messene through control of the hostages and granting equal powers to Messene as a member state.  In view of the traditional enmity between Sparta and Messene, Messene could have expected no favors from Sparta, and it appears that, in selecting Sparta as the place of detention, the Achaean League appointed a guardian potentially more hostile to Messene than the League itself.  Certainly the Aetolian hostages of 173 B.C. adhere even more clearly to the premise that a third party’s control maintained the equal balance of the two disputants. The negotiations between Perseus and Eumenes may even have suggested the necessity of a  third party which could have acted as arbiter, for the contract which was proposed seems to the modern reader indefinite and open to rather broad interpretation. Perseus was to pay a stipulated sum to Eumenes in return for Eumenes’ efforts to reconcile the Macedonian king to the Romans; Eumenes was to guarantee his efforts by means of hostages, whose number, place of detention, and degree of care were chosen by Perseus. It is probably all for the best that the agreement fell through, since the issue hinged on Eumenes’ good faith, and in the event of Eumenes’ failure the contract was too vague concerning both the money and the hostages. Nevertheless, the resemblance between this unratified agreement and the others is obvious: a third party prevented any clear dominance being established between the two contracting parties, and the degree to which the third party was biased in favor of the recipient state probably reflected the actual imbalance in the power of the two contracting states. 
The most common place of detention, somewhere in the recipient state’s own territory, requires little discussion, for the reasons for its frequency are manifest: ready availability for retaliation, difficulty in effecting escape or recovery, the psychological encouragement and sense of superiority which the hostages’ physical dependency fostered in the recipient state’s citizens. More nebulous was the possible advantage of bonding hostages to their keepers through specific personal ties or general feelings of gratitude. Much argument has arisen on this aspect of exacting hostages, usually in terms of the degree to which conscious Romanization was a factor or even a goal in Roman policy; an analysis of Roman behavior toward hostages throughout  the Republican period may provide some slight evidence on this important question. 
The earliest date at which the question of conscious Romanization emerges as an issue with reference to hostages is the Second Punic War. The two episodes from this war furnish the fullest information on the processing of these foreign representatives who were also captives.  In 215 B.C., after Roman troops had suppressed a revolt in Sardinia and had exacted tribute, grain, and hostages, the commander on his return to Rome delivered the tribute to the quaestors, the grain to the aediles, and the hostages, evidently now called captivi, to the praetor urbanus, Q. Fulvius (Livy 23.41.6-7). Three years later, in 212 B.C., the Tarentine Phileas approached the Tarentine and Thurian hostages who were guarded in the Atrium Libertatis and persuaded them to escape (Livy 25.7.10-14).  Both incidents suggest that the responsibility for the hostages was public and magisterial, rather than private. The parallel construction which allotted tribute to the quaestors and grain to the aediles is a strong indication that the captivi were the concern of the praetor urbanus ex officio, just as the maintenance of hostages in a public building like the Atrium Libertatis implies that it was a public charge; it seems improbable that a private citizen could have assumed the authority necessary to regulate a public place. 
Later incidents demonstrate an altogether different pattern. With the conclusion of the Hannibalic War, allied Italy became secure from foreign invasion and in the following century grew increasingly Roman. Foreign hostages could safely be stationed outside the city itself, and as early as 199 and 198 B.C. Carthaginian hostages are found in the Latin colonies of Fregellae (Nepos Hannibal 7.2), Norba,  Ferentinum, Signia, and Setia (Livy 32.2.3; 26.5). Significantly, the Roman Senate, not a magistrate, assigned hostages to these colonies, presumably because the governmental structure of Latin towns permitted strong officials to whom the responsibility for the hostages could easily be delegated.  The Thracian hostages exacted by Perseus and captured by L. Aemilius Paulus in 167 B.C. were in a like manner quartered in Carseoli (Livy 45.42.5-11), the Carthaginian hostages after 149 B.C. in various places in Italy (Zonaras 9.30), and Oxynta, the son of Jugurtha, in Venusia in 90 B.C. (Appian Civil Wars 1.42). 
The only documented exception to the Senate’s assignment of hostages to Latin towns is Antiochus, later King Antiochus IV Epiphanes of Syria. For the second son of Antiochus III the Great, a house was built at public expense which, according to Atticus and others, later belonged to the poet Lucilius (Asconius Commentary on Cicero’s Speech Against Piso p.12K). Inasmuch as Lucilius, undoubtedly a wealthy man,  was content to own it half a century later, the house must have befitted Antiochus’ royal station. Unfortunately, the quartering of the other nineteen hostages tendered to Rome by Antiochus III the Great is unknown,  but it is reasonable to assume that the distinction displayed by the Senate in permitting his residence in the capital was at least in part due to his high birth and position as the second son of the Syrian monarch. Nor is it apt to be coincidental that Demetrius, son of Seleucus IV and Antiochus’ replacement as hostage to Rome, resided at least briefly in the capital; we cannot tell whether his quarters during his appearances before the Senate were permanent or just temporary (Polybius 31.2; 31.11).  Although the evidence is very scanty and quite circumstantial, one might justly suspect that the place chosen by the  Senate for the detention of hostages was neither determined mechanically nor entirely without significance; the seemingly preferential treatment accorded Antiochus and perhaps the other Seleucid prince Demetrius as well argues for a conscious and fluid policy which can be asserted all the more confidently in light of other privileges granted them during their stay in Rome.
Only a handful of cases describe Greek treatment of hostages, but all suggest that Greek hostages were usually accorded care commensurate with their station in life.  That liberated hostages became proxenoi or received such honors as tax exemptions from the states which had detained them  argues for the mutual regard of hostage and recipient, an incredible situation had the recipient government infringed on the privileges normal to the hostages’ rank. Philip II of Macedon wished to reward Philo, his benefactor and guest-friend while he was a hostage (Plutarch Moralia 178C); a ridiculous story if hostages were normally mistreated. It is unfortunate that, in the three incidents which provide direct testimony concerning the care which Greek hostages enjoyed, the disposition of the hostages was one of the causes of the resulting situation, so that the value of the testimony may be impaired because of the unusual and perhaps exaggerated importance placed on the treatment of the hostages. A story told of Alexander by a late and not very reliable source is particularly instructive. To counter a report of his brutality, we are told, Alexander displayed hostages from the city he had most recently captured to a city which hesitated to surrender; the good physical condition  of these hostages mitigated his reputation as a savage, and the besieged city and others as well capitulated at this proof of his honorable conduct (Polyaenus 4.3.30). Even if this episode is historically accurate, it reveals very little about Alexander’s treatment of other hostages, since reasons of propaganda might have prompted his decency in this case. 
It is equally difficult to deduce much from the brief reference to Cratesicleia’ s retinue in Alexandria while she and her grandsons prevented a separate peace between Cleomenes and Antigonus Doson (Plutarch Cleomenes 38) or from the statement about the Judaean hostages held under guard in the citadel of Jerusalem during the reign of Demetrius I Soter (Josephus Jewish Antiquities 13.17). In the former case, the retinue appears to have been mentioned primarily to present the disaster in the worst possible light, since the members of the retinue were also murdered; besides the impossible problem of determining the retinue’s size, is the retinue perhaps a rhetorical means of showing that the massacre of the Spartans in Alexandria was complete? The mere fact of the presence of attendant women and the mere fact that the Judaean hostages were guarded can do little more than indicate that hostages in the Greek world could expect at least the trappings of respect and the reality of custody.
The major exceptions to the usually considerate treatment of hostages are attested, predictably, in hostile sources: Duris charged Cleonymus the Spartan with exacting two hundred lovely women from Metapontum for the satisfaction of his own lust (Diodorus 20.104.3; Athenaeus 13. 605D-E); the Roman historians and Polybius contrasted the chivalry of Scipio with the abuses of the Carthaginians in New Carthage;  Plutarch accused Philip of Macedon of poisoning the younger Aratus (Plutarch Aratus 54.1);  and Caesar composed a speech for Ambiorix in which Ambiorix expressed his gratitude for Caesar’s rescue of his relatives who had suffered servitude as hostages to the Aduatuci (Gallic War 4.27.2). The possibility of abuse, especially sexual assault,  is a frequent theme even when no actual harm occurs, and it appears in sources as divergent as Livy’s account of Cloelia (2.13.10), where it explains why she chose to release impubes when Porsenna allowed her to free some of her fellow hostages, and Polyaenus’ anecdote of Agathocles’ deception of Ophelas (5.3.4), which appears to assume that Agathocles’ son could delay the consummation of Ophelas’ desire for him until Agathocles could rescue him: less a rape than a determined seduction.
It should be apparent, however, that even hostile sources did not usually report atrocities against the persons of hostages without another purpose in addition to blackening an enemy’s reputation. Although the fragmentary nature of Duris’ history precludes any certainty concerning his reason for mentioning Cleonymus’ mistreatment of the Metapontine women, Diodorus used the incident to illustrate Cleonymus’ betrayal of the Tarentine cause (20.104.4) and probably his eventual failure in South Italy. The clearest example of such a use of an anecdote concerning hostages to make a point is the episode which followed Scipio’s capture of Cartagena in 209 B.C. The disgraceful behavior of the Carthaginians toward their female Iberian hostages, so delicately implied by Mandonius’ wife (Polybius 10.18.7-15; Livy 26.49.11-15), differs sharply from the gallant conduct of Scipio, who was so morally superior to his Punic foes that at first he did not comprehend why Mandonius’ wife would request trustworthy guards. The juxtaposition of Punic brutality and Scipio’s charming naiveté  furnishes a contrast both flattering to a Roman audience and of exemplary morality, a dual purpose which has been, not surprisingly, embroidered and amplified with appropriate speeches. Nor is this contrast the sole reason for the incident’s narrative prominence. In accordance with his view of pragmatic history, it is the disaffection caused by the exaction of hostages in general (10.35.6) and their subsequent mistreatment (10.38.1-2,4) which Polybius chose to emphasize. The validity of Polybius’ attribution of Iberian discontent to the Iberians in his account of this episode has been accepted without comment by modern scholars.  Similarly, Caesar inserted very artistically his restoration of Ambiorix’ s relations from what he portrayed as slavery to demonstrate his own generosity and the enemy’s cruelty and to depict Ambiorix’s later revolt from Roman allegiance in the most flagrant colors of rank ingratitude. 
It is completely predictable that such brutality does not overtly occur in Roman examples until the last half century of the Republic, when the losing sides in the civil wars stand accused of this and other atrocities. An analysis of individual episodes in chronological order, however, is perhaps the easiest method of deriving without prejudice what can be known or inferred about the care which Romans devoted to foreign hostages. The earliest information concerns the Tarentine and Thurian hostages in the Atrium Libertatis who, corrupted by Phileas, attempted to escape (Livy 25.7.10-14).  Their actual custody is described as minore cura (25.7.12), and the seeming negligence is dismissed because, says Livy, escape was inexpedient for both the hostages and their governments. The security measures which permitted the hostages to escape not only from the Atrium Libertatis  but from the city itself to Tarracina, some fifty miles away, can scarcely have been more than formalities, and this in spite of the Punic campaigns around Tarentum and Thurii which might have been expected to prompt the Romans to greater watchfulness. A second point is somewhat more subtle. Phileas, who seems to have been a representative of Tarentum,  was evidently allowed such free access to the Tarentine hostages  that he could persuade them to become involved in his conspiracy, a fact which demonstrates rather conclusively that the Romans did not monitor these conversations or at least failed to do so effectively. It would also appear that these hostages maintained contact with their government despite the restrictions imposed on communications by the Hannibalic War.
The group of hostages about whom our information is fullest is the young Carthaginians delivered to Rome after the Second Punic War.  In 199 B.C., a Carthaginian embassy, in Rome on official business, requested that the Roman Senate permit some of the hostages to change their place of residence, and the Senate concurred (Nepos Hannibal 7.2; Livy 32.3.4). Again, we observe that communication between the hostages and their government appears indisputable and that, since the Senate granted the request, suggestions concerning the hostages’ care formed an appropriate part of the diplomatic exchange between the two governments. A very interesting point is the reason proposed for their removal from Norba: ab Norba . . . ubi parum commode (sc. obsides) essent (Livy 32.2.4). From Livy’s statement it is impossible to determine whether the objection pertained to the town, to the general circumstances, or to the hostages themselves; the specific problem or problems which led the embassy to seek alternate quarters lie open to  speculation.  Nevertheless, it is the obvious right to appeal which is of significance here. Whatever the reason for the dissatisfaction of the Punic hostages, they were permitted to complain to their own government, which could then petition the Roman government to ameliorate conditions, and the Roman Senate was willing to acknowledge such petitions.
For the following year, 198 B.C., a remarkable and indeed unique episode is attested by the Roman annalistic tradition. A plot to touch off a slave uprising centered around Setia, with the alleged object of freeing the Carthaginian captives and hostages detained there (Livy Summaries 32; 32.26.5-8). Although the historical accuracy of the plot is basically irrelevant for the purposes of this study,  the skepticism which has attached itself to the narrative of the Setia revolt has spread to the details of the arrangements which followed the suppression of the plot and has cast doubt on the orders related in conjunction with similar unrest in Praeneste (Livy 32.26.15-18). Since these orders reflect upon the actual freedoms normally allowed to hostages, it is important to rescue them from the suspicions which encompass the framework in which they are embedded.
There are essentially two arguments made against accepting Livy’s account of the uprisings in Setia and Praeneste: first, the implausibility of a servile conspiracy to liberate Punic captives and hostages; and second, the absence of these conspiracies from Polybius’ history, which indicates that Livy’s source was an unspecified annalist.  It is no longer judged sufficient to discount an episode’s historicity merely because its source is an annalist. Many of the annalists, it is true, bowdlerized, exaggerated, or invented Roman  history, but for the greater glory of their families, their factions, or their idealized conception of a moral and patriotic past; it is not at all obvious what purposes the invention of abortive servile uprisings in Setia and Praeneste would serve. The major magistrate, L. Cornelius Merula,  has no elaborate role aggrandizing his personal career or the distinction of his family, nor does the suppression of a slave revolt merit special notice of acclaim, according to Roman standards of military valor. This police action possessed no political or military significance, and that might well explain why Polybius did not refer to it. Thus its annalistic origin should not prevent acknowledgement of its possible historicity.
The plausibility of the uprising’s alleged purpose is a different matter. On the most superficial level, it is certainly unlikely that slaves, even those of Punic origin, would have risked their lives for the altruistic liberation of the Carthaginians held in Setia.  This objection, however, is not a serious one, if two circumstances are considered. First, only the Livian epitome (Summaries 32) asserts that the revolt had this intention; a lacuna in the full narrative (32.26.8) has been supplied on the basis of the epitome,  although the third source for the episode (Zonaras 9.16) has an entirely different scenario in which the hostages themselves rebel, and in that case there would have been no need of a slave revolt intended to liberate the hostages. Since the accuracy of the epitomes can be challenged in some places where they can be compared to the full text,  it is perhaps unwise to rely too heavily on this testimony, especially when a careless reading of the full account which has survived could very easily give rise to just such an impression as the one conveyed by the  epitome and by Zonaras: the Carthaginian hostages played a primary role in the unrest, as its objects or its instigators. Moreover, to infer from the epitome that the hostages’ release was the ultimate goal of the uprising is unrealistically rigid. If the slaves actually intended to free the Carthaginian hostages and captives they may have had more practical and cynical reasons than altruism or patriotism. Possession of the persons of these noble Punic youths could have materially advanced the bargaining position of the slaves, for the return of these nobles to Roman authorities could have purchased  clemency or safe conduct to territory outside Roman dominion, or - more dangerous to the Roman government - the release of these nobles might have enabled Carthage to abrogate their treaty with Rome and to resume hostilities in Italy. In view of the Gallic unrest, reputedly under Punic stimulus (Zonaras 9.16), it is possible that the slaves had planned to seize the Carthaginian hostages as a means to exact Carthaginian support. That the Romans feared another Punic war or even the possibility of such a war would account for the prominence of the hostages in events in which they were not personally involved.  And this prominence of the hostages has the curious feature of being entirely passive in nature, which is inexplicable if Livy believed them guilty of conspiracy. Characterized throughout (32.26.4; 6; 8; 13; 15) as servile, the uprisings threatened Roman control over the hostages, and it is this threat to Roman security from foreign pressure  which condensed the report of the slaves’ intentions to the only one significant to the Romans: the seizure of Punic hostages and captives. 
If, as argued above, there is no compelling reason to dismiss from history the Punic slave disturbance of 198 B.C., on grounds of  either its annalistic origin or its “implausible” intention, it now becomes feasible to examine the details pertaining to the hostages’ care as probably representative of the usual treatment accorded them in their Italian residences. Livy wrote: cum iis (sc. obsidibus) ut principum liberis magna vis servorum erat; augebant eorum numerum . . . (32.26.5-6). The large number of slaves was the fitting and seemingly unremarkable consequence of the hostages’ socioeconomic status; one may reasonably postulate that the hostages enjoyed the comforts and luxury (at least those available to them in these small Italian towns) to which their position entitled them, despite their detention. Further, the Punic nobles could perhaps add (augebant) to their retinues while in Italy, which suggests that they had some financial independence as well.  In a second piece of testimony, which describes the uneasiness at Praeneste, the curtailment of certain activities provides evidence for normal privileges. As one of a series of security measures designed to reassure the state that the slave uprisings had been completely suppressed and that the Carthaginians could not foment further trouble, L. Cornelius Merula, the praetor urbanus,  issued orders that hostages were to be kept in privato, with no opportunity of appearing in publicum (32.26.18). This prohibition is meaningless unless the hostages had previously had some freedom of movement, at least the minimal mobility offered by occasional forays into the public places of the towns in which they lived. There is an additional indication that the Punic hostages either stayed in Italian households or maintained their own houses;  since jails and temples were public property, from which Merula’s decree debarred them (unless this prohibition is a condensed report of a complete change in quarters), their normal  domiciles were private. Punic captives, however, were to be lodged only in public prisons and to be bound with shackles of not less than ten pounds; the clear distinction between the two groups of Carthaginians furnishes another, albeit negative, proof that hostages generally received considerate treatment more consonant with that of guests than that of victims.
For the next century, the lives of other hostages for the most part substantiate the assumption that Rome treated these foreign guests with the respect which was their birthright. Only one hostage is said to have died while in Roman custody, which, if accurate, is certainly a good record. The dubious honor goes to Armenas, son of the Spartan tyrant Nabis, who was not restored with the four other Lacedaemonian hostages in 191 B.C. and who soon sickened and died (Polybius 21.3.4). Although his death may have been convenient for the Romans, who could have had little desire to see Nabis’ son and possible successor free to continue Nabis’ revolutionary policies in the Peloponnese, there is no reason to suspect that the Romans in any way aided his ailment, for his detention as a hostage would have served Rome as well as his death did.
As a general rule, then, hostages enjoyed some freedom with regard to physical mobility, in addition to generous treatment with respect to their privacy and their physical needs; in both Greek and Roman society, hostages were accorded such visible proof of their elevated status as retinues. Nevertheless, these liberal policies could be cancelled without notice by the recipient state, and hostages could suffer not merely curtailment of privileges, but overt brutality, particularly sexual abuse, evidently without recourse or appeal. 
Hostages of the Second Century B.C.
The histories of the three royal hostages, Demetrius of Macedon, Antiochus of Syria, and Demetrius of Syria, offer different problems for the modern analyst. Demetrius of Macedon and Antiochus belonged to a group of hostages (e.g., Polybius 18.39.5-6; 21.42.22), and it is likely that Demetrius of Syria also had colleagues in his detention.  But only the royal hostages figure in accounts subsequent to the initial delivery of the hostage group, and it is thus impossible to know whether these princes were separated from the rest for preferential treatment or whether their experiences were typical of the group. The question is further complicated by the factors of nationality and degree of civilization: is it reasonable to assume that the treatment of Macedonian or Syrian hostages is characteristic also of the treatment of hostages from less powerful or less sophisticated peoples? No evidence remains to attest the privileges allowed the hostages whom Scipio Nasica exacted from the Boii in 191 B.C. (Livy 36.39.3), and it is perhaps too democratic to believe that the rather barbaric Cisalpine Gauls were granted the same license as Greek courtiers and princes. On a more pragmatic level, Philip V and Antiochus III were capable of dangerous retaliation if they were offended by injury or if insult were offered their hostage sons; in the second century B.C., the Boii never generated an equivalent feeling of respect in their Roman adversaries, and their hostages may have suffered some loss of privilege in comparison. A more disheartening problem for the modern scholar is the suspicion which arises when we find that the extant testimony about the time they spent in Rome was included in our sources specifically in order to illuminate or explain the later careers of these hostages.  The difficulties in analyzing such testimony are enormous, and for the sake of convenience, I have examined the accounts of hostages’ careers after their departure from Italy separately from their earlier detention.  In this chapter I shall concentrate only on what seems fairly certain about their residence in Italy.
Concerning Demetrius of Macedon, few details are available. He graced the triumph of T. Quinctius Flamininus in 194 B.C. (Livy 34.52.9; Eutropius 4.2.3; Orosius 4.20.2-3).  According to Plutarch (Flamininus 14.2) Flamininus was influential in securing Demetrius’ restoration.  Since the Scipios were the field commanders who benefited from the aid which Philip V of Macedon furnished and for which Demetrius was released, and since Flamininus and the Scipionic party frequently supported each other’s policies,  Plutarch’s attribution of influence to Flamininus is quite plausible. In 183 B.C. the Roman Senate chose to give a signal honor to Demetrius’ embassy, allegedly because of the favor which his conduct as a hostage had evoked (Polybius 22.14.9-11; Livy 39.35.2-3; Justin 32.2.3). The exact nature of this impressive conduct is unclear; Polybius’ account suggests that the young prince was expected to have made useful contacts among the Roman ruling class, and Livy and Justin mention his kingly nature (specimen indolis regiae) and his decency (uerecundia) respectively as the motivating element in the Senate’s decision. 
Information on Antiochus’ stay in Rome is even scantier. The sole evidence of value is that of Asconius (Commentary on Cicero’s Speech Against Piso p.12K), who says that a house was built for him at public expense.  Since it later became the house of the wealthy poet Lucilius, it was probably an expensive edifice. No other financial details, such as the furnishings  or routine domestic expenses, are mentioned; whether the Roman government or Antiochus III covered the other costs cannot be determined. Nevertheless, this is the first concrete proof that the Roman government assumed any financial responsibility for the foreign hostages whom it detained.
As king of Syria, Antiochus IV Epiphanes always maintained a cordial relationship with Rome, even after a Roman official’s arrogant conduct toward him,  and this fact is obviously the basis for a remark which Livy wrote for Antiochus’ ambassador after his accession (42.6.9). The ambassador’s speech, which expressed Antiochus’ gratitude that he had been treated by all classes, not as a hostage, but as a king, is less a reliable historical statement than it is Livy’s own self-congratulatory comment on the Roman sense of propriety, but it may yet be true. However Antiochus had been treated while at Rome, Roman customs had impressed the Syrian prince,  and - whether from gratitude or from a less lofty motive - he never openly opposed Roman policy.
The obstacles which Demetrius of Syria met throughout the course of his relationship with the Roman government and the measures which he was forced to employ in countering them were evidently unusual and so feature largely in the narratives which mention him. We are fortunate that the best historian of that generation, the Achaean Polybius, was not merely present during Demetrius’ later years of residence, but also a personal friend of both the young prince and of some important Roman senators. Moreover, Polybius actively abetted Demetrius’ escape, and his account of the prince’s earlier actions and of the escape itself has survived the mutilation from which other portions of his history suffer. The earliest incident, however, is an anecdote concerning  Demetrius and his cousin, Ptolemy VI Philometor. In 164/3 B.C., the Egyptian king was going to Rome to plead his case before the Senate, the self-appointed arbiter of the quarrel between the king and his brother Ptolemy Physcon. Despite the king’s lack of insignia and his servile mode of conveyance (he was afoot), Demetrius recognized him and sought to equip him as befitted a king, a gesture which Ptolemy appreciated but refused (Diodorus 31.18). Obviously, Demetrius’ financial situation must have been comfortable to enable him to make such a princely offer, and he must have been afforded some mobility outside the city in order for such a chance meeting to occur. One may assume that Demetrius’ own conduct was compatible with this lofty conception of what was appropriate for royalty, and this anecdote has been interpreted as a foreshadowing of the attitude which culminated in the later difficulties between the prince and the Senate. 
After the death of Antiochus IV Epiphanes, Demetrius appeared before the Senate to present his plea for release and his claim to the Syrian throne (Polybius 31.2; Zonaras 9.25; Appian Syrian Wars 46). His case seems to have been legally and morally sound. Demetrius had replaced Antiochus as a hostage at least in part because Seleucus IV Philopator would have valued his son Demetrius more than his brother Antiochus; in turn, Antiochus IV Epiphanes’ son, rather than his nephew Demetrius, should have become the hostage for his fidelity, if there were need of a guarantee. On Seleucus IV Philopator’s death, Antiochus had succeeded because of the minority of Seleucus’ sons;  Demetrius possessed the same right to the monarchy as had Antiochus, since Antiochus V was a minor. The cynical practicality which Polybius described as the Senate’s motive for detaining Demetrius (it also  appears in Appian’s summary) is typical of both Polybius’ realistic pragmatism and of the policies of that era. This emphatic practicality and the seeming absence of other Syrian hostages in Rome strongly indicate that Demetrius’ continued detention was without legal justification.  It is impossible to believe that the Senate would not have stridently insisted on replacement after Demetrius’ escape, had there been the slightest legal pretense for doing so. Further, Demetrius’ eagerness for Roman recognition of his authority gave the Senate a diplomatic advantage which could have been traded for another hostage, but nothing of the sort happened. It is obvious that the Senate’s refusal to release Demetrius was the result of political, not legal, considerations. 
The murder of the Roman legate Cn. Octavius, in 162 B.C, prompted a second appeal to the Senate, but again the majority of the senators denied Demetrius permission to leave Rome. The Syrian prince, however, had no intention of spending his entire life as a hostage and with several of his friends, including Polybius, planned and executed his escape to Syria and his throne (Polybius 31.11-15.10; Livy Summaries 46; Appian Syrian Wars 47; Zonaras 9.25). The escape is fully discussed elsewhere,  but the following points are apposite here. The entourage which attended Demetrius was numerous (πλει?νων, Polybius 31.13.2), and, since it was his custom to invite all of them to dine with him when he ate at home (13.5), it may be inferred that they were men of personal distinction.  Further, a retinue of slaves attended his person (14.11). He was permitted, on very short notice, to go hunting as far from Rome as Anagnia and Circeii, without question (14.2; 15.1-2). In short, his social life, so far as friends, dinner  engagements, and diversions were concerned, was unrestricted. For Demetrius enlisted in his plan not only his own personal associates but also an interned Achaean whose legal position was more circumscribed than his own. This fact argues powerfully for a remarkable degree of freedom for both Polybius and Demetrius. The success of the plan indicates that there was no elaborate security system designed to prevent just such events; news of the conspiracy did not reach anyone interested in Demetrius’ continued detention in time to block his departure from Roman territory.  As in the case of the Tarentine and Thurian hostages in 212 B.C., the Romans evidently did not anticipate active resistance to their will.
From these incidents, what can we conclude about Roman treatment of Demetrius? He seems to have had a rather generous fortune at his disposal, although the source of this wealth is unknown.  He twice addressed the Senate, on his own initiative and without any ambassadorial mediators, unlike the Punic hostages of 199 B.C. Perhaps most importantly, he was apparently subject to few restraints in his acquaintances and to little monitoring of his conversations, correspondence, or movements.  Indeed, the only negative aspect of his residence is his illegal detention, which of course made even the most distinguished treatment insulting.
The three hundred Carthaginian hostages exacted before the outbreak of the Third Punic War merit attention for several reasons. They and the most prominent Carthaginian captives spent their lives in Italy on parole, so to speak (?ν φρουρα?ς ?δ?σμοις, Zonaras 9.30). The difficulty in ascertaining whether the Romans were ethically justified in retaining these hostages after the Carthaginian government refused  to accept evacuation of the city as a condition for peace has already been mentioned;  whether there is any validity in examining the ethical propriety of historical decisions is itself questionable.  The practicality of keeping the hostages throughout the duration of the war needs no explanation, but the position of the hostages altered after the total annihilation of Carthage in 146 B.C. The Carthaginians could no longer be the physical guarantees of a treaty or a means of blackmailing politically prominent parents into a disadvantageous peace, for there was no Carthaginian state; nor could the Romans formally restore them, because there was no government or parents to receive them. With the destruction of Carthage the Punic hostages became politically meaningless. Under these anomalous circumstances it is not surprising that the young Carthaginians remained in Roman custody, in all likelihood in the same conditions as had pertained during the war. It is ironic that the group which in 149B.C. might easily have suffered the penalty for Carthage’s defiance became the nucleus of the noble Punic survivors of that war. 
The royal hostages of the second century B.C. appear to have been well treated by the Romans; there is no indication of any restriction besides that of hostageship itself. They lived well; they do not seem to have been subjected to abusive or insolent behavior; they were permitted freedom of movement and freedom of association. Although the same privileges cannot be proven to have been available to the Punic hostages of 149 B.C. and later, the fact that these noble Carthaginians remained in honorable custody after the belated resistance of Carthage in the Third Punic War suggests that the Romans accepted some responsibility for their continued maintenance in respectable  surroundings.
Hostages of the First Century B.C.
All the remaining incidents which illuminate the care provided to hostages by the Romans belong to a later period and to a different political situation. Despite the great wars of conquest and, in the first century B.C., the expansion of Roman territory to the borders of tribes previously unknown to Rome, these incidents describe the behavior of a civil faction from the viewpoint of the hostile and ultimately victorious opposing faction. Thus the brutality of Sertorius toward the sons of Iberian chiefs when the chiefs defected or otherwise failed him (Plutarch Sertorius 25.4); the outrage of Caesar when he learned, or was said to have learned, that the Pompeians in Africa snatched the children of distinguished local citizens into slavery under the guise of holding them hostage (African War 26.5); the arrogance of M. Antonius in gathering so many notables at Alexandria (Dio 51.16.1-2).
The author of the African War provided no more information than an allegation of Pompeian abuse of African hostages, but two sets of parallel cases are brought to mind. Several examples of hostages being delivered or demanded from a subject people by one faction in a civil war demonstrate that this measure was a standard safeguard; Carbo’s demand for hostages from Placentia in 85 B.C. (Valerius Maximus 6.2.10) and from Italian colonies and towns in 84 B.C. (Livy Summaries 84); the Pompeians Afranius and Petreius in Spain in 49 B.C. (Caesar Civil War 1.74.5); and the Pompeian Staberius in Apollonia in 48 B.C. (Caesar Civil War 1.84.2).  In view of the Pompeian precedents in Spain and Apollonia, a Pompeian  demand for African hostages seems quite probable, but subsequent mistreatment of them is neither substantiated nor likely, since the Pompeians had no desire to alienate possible African supporters. The equation of hostage with slave would seem less a reflection of actual abusive treatment of the African hostages by the Pompeians than an assessment of the condition of any hostage; a comparison of this comment with the Parthian equation of hostageship and slavery (Josephus Jewish Antiquities 18.47) is irresistible. 
The cases of Sertorius and Antony are more complicated. The romantic figure of Sertorius has exerted great appeal upon modern scholars, and his innovative methods have aroused controversy as to his intentions regarding the Iberian boys whose education he subsidized.  Although an evaluation of his goals lies for the most part beyond the scope of this inquiry,  the details of this anecdote nevertheless furnish some suggestive points which have interesting parallels in later Roman history. The status of the Iberian boys is peculiar. They were not, apparently, given over as hostages; Sertorius’ own personality and military ability won the Iberian chiefs’ devotion, and Sertorius lacked the military force necessary to demand children from reluctant parents. Instead, the allure of Sertorius’ promises of future honors induced them to surrender their children to his school at Osca.  The school at Osca is another odd feature. To my knowledge, the only other episode in which education and hostages are associated did not occur until A.D. 39 (Suetonius Caligula 45), but common sense strongly urges that this rarity is coincidental. It seems probable that some effort was made to teach hostages, especially those from the western provinces, at least the Latin language and some Roman practices.   Apart from the question of whether the school was a conscious attempt to Romanize the noble Iberian youths, Sertorius’ educational policy appears to have been both popular and generous. The boys wore the garment appropriate for senators’ sons, the toga praetexta, and might obtain golden bullae for good work; Sertorius paid their fees and awarded prizes, which suggests that the school was not entirely a ploy designed to lure young Iberians into Sertorius’ control. And so long as the chiefs remained loyal, their sons were safe and treated honorably. 
The compression of Dio’s account of foreign notables at Alexandria (51.16.1-2) after the death of Antony prevents any clear understanding of either Antony’s foreign policy or the changes which Octavian effected. Of the many (συχνο?) children of sovereigns held as hostages or “through arrogance” (?φ' ?βρει), Octavian restored some, retained others, and created marriage alliances among them, but only two specific cases are discussed: he restored Iotape to her father, the Median king, but he refused to permit the brothers of Artaxes, king of Armenia, to return to their homeland because Artaxes had executed the Romans left in Armenia. Artaxes’ brothers remained with Octavian until 20 B.C., when Artaxes was deposed and Tigranes, his brother, was installed as king with the aid of Tiberius (Dio 54.9.4-5). The fate of the three other brothers is not known. In the absence of proper identification of the other hostages and involuntary guests detained by Antony, little else can be ascertained about their case. It is, however, logical to assume that the nameless noble children who were not released became part of the unofficial court which was Octavian’s household.  
The four Parthian princes who, together with two wives, four children and (probably) retinue, were sent to Rome around 10/9 B.C. by their father, Phraates IV, are the last historical group whose treatment will be considered. Their status as hostages in any formal agreement is certainly dubious, despite the ancient sources. Several authorities describe the princes as hostages: Tacitus (obses Augusto datus a Phraate, Annals 2.1); Suetonius (Parthorum obsides, Augustus 43.4); Velleius Paterculus (obsides,2.94.4); Orosius (regiis obsidibus traditis, 6.21. 27); Eutropius (obsides, 7.9); Justin (obsides Augusto dati, 42.5.6-11); Strabo (?ξομηρευσ?μενος, 6.4.2C288; ?μηρα α?τ?, 16.1. 28C749); and Josephus (?φ' ?μηρε??, Jewish Antiquities 18.42). The language of the Res Gestae is much more restrained: filios suos nepot[esque] misit in Italiam non bello superatus, sed amicitiam nostram per [libe]ror[um] suorum pignora petens . . . (32). The motivation which Josephus attributes to Phraates’ action, the removal of his legitimate heirs in order to clear the succession for an illegitimate son by an Italian slave, is supported by the fact that the bastard son did in fact succeed. Further, domestic revolt invariably centered on an Arsacid prince, and the additional presence in Parthia of four sons and four grandsons could have imperiled the bastard Phraataces or even Phraates IV himself.  Only the late source Orosius mentioned a treaty (foedus); the φιλ?α or amicitia to which Strabo and the Res Gestae refer need not mean anything more than a private agreement between Caesar and Phraates to leave off active hostilities.  Moreover, it is difficult to discover why Phraates would have surrendered legal claim to the princes unless he was compelled to do so; since the two nations seem to have negotiated as equals at this time, it is likely that the  princes were not formally hostages but guests whose presence in Rome is explained by the danger of their position in Parthia, and the embarrassment it might have caused. 
Although there is good reason to suspect that the Parthian princes were not hostages for an international treaty, their physical presence in the Italian capital argues powerfully that they were in fact hostages in the modern sense of the word, and it is quite probable that the legal distinction which escaped the ancient sources was reflected in their treatment. The princes lived royally at Roman expense (δημοσ? βασιλικ?ς τημελο?νται) Strabo (16.1.28C149); two even died and were buried in Rome (CIL VI 1199 = ILS 842). One day at the games, Augustus seated his distinguished visitors in the second row of seats above his own (Suetonius Augustus 43.4). This gesture of honor had the effect of displaying to the audience the first Parthians dependent on Augustus, as Suetonius remarked, but it probably also established the princes as worthy of attention. The most obvious proof of favorable Roman treatment is the release of Vonones and Phraates, sons of Phraates IV, and of Tiridates and Meherdates, grandsons of Phraates IV, all of whom then became kings or contenders for the throne of Parthia; that four of eight heirs could contend for the rule of Parthia so many years after their delivery to Roman custody speaks well of the Romans. 
Although the connection is extremely tenuous, material from the Irish Leabhar na g-Ceart (Book of Rights) casts some light on the generous behavior usually accorded hostages in the Greco-Roman world. In this poetic document describing the privileges and restrictions of the kings and clans of Ireland, the Oirghialla supplied hostages to the king of Tlachtgha, who kept them without fetters and was to provide  “befitting attire to them, a steed, a sword with studs of gold, secret confidence, elegant apartments . . .” The price for these gifts was the hostage’s oath not to escape, on pain of withering.  Not all hostages were so privileged as to bear arms, however; the Fomorian sureties could only use “bites, kicks, and blows” in the king’s house, lest, armed, they might do a misdeed.  Yet hostages who were true to their oath held honorable positions in the king’s retinue, sitting second on the king’s right side; those specially pledged for the good behavior sat next to the king’s bodyguard on the southern couches, while forfeited hostages, in chains, were at the extreme left. 
Thus it is clear that the Irish Celts employed a kind of parole system to eliminate the need for close supervision of hostages, a system based ultimately upon a religious contract binding on both hostage and king.  In view of this oath, the seemingly unrestricted movements of Demetrius of Syria and the generosity afforded the Parthian princes take on a new significance. Interestingly, it is possible that the Romans too enjoined an oath upon hostages, and a passage from Polybius (followed by Appian Syrian Wars 47) may provide support for such a hypothesis. After the assassination of Gnaeus Octavius, Demetrius of Syria again appeared before the Senate, seeking release from the necessity of serving as hostage for his cousin Antiochus V (παρεκ?λει τ?ς γε κατ? τ?ν ?μηρε?αν ?ν?γκης α?τ?ν ?πολ?ειν, Polybius 31.11.9). Demetrius’ eagerness for official termination of his status as hostage, even without any Roman recognition of his claim to the throne, may have its roots in a personal obligation besides the one originally undertaken by Antiochus III the Great and renewed by Seleucus IV Philopator: a pledge not to attempt escape. Such a promise would  explain the lack of Roman custody and Demetrius’ remarkable freedom of movement and association. Such a hypothesis of parole is of course speculative, but it is also attractive.
Finally, three legendary tales deserve some study because they reflect the storytellers’ attitudes toward the possibilities inherent in the treatment of hostages. The first two stories are found solely in Dionysius of Halicarnassus and concern Latinus, the eponymous king of Latium. According to Dionysius, Latinus was the son of Heracles and a Hyperborean girl given as a hostage by her father; at first untouched, she later became pregnant on the journey to Italy, where Heracles left her with Faunus (1.43.1). The second legend is one of the many variants of the Romulus and Remus myth: Aeneas’ grandsons, part of an exchange of children who thus safeguarded the peace between the Trojans and the Latins (1.59.1-2), were not merely treated well by Latinus, but were adopted by him since he had no male issue (1.73.2). Although the account of Latinus’ lineage is a fiction created to link Rome with Greece and with the archetypical Greek hero, it does clearly delineate a possible consequence of exacting female hostages. The adoption of the twins, no less fictitious, neatly resolves the problem of legitimizing the Trojan Romulus’ claim to Latin territory at no cost to aboriginal prowess or dignity. Despite its worthlessness as historical fact, it suggests that, to the storyteller, decent treatment of hostages was not inconceivable, and that an intimate bond might arise between hostage and keeper. The third case, that of Cloelia and Porsenna, is fully discussed elsewhere;  nevertheless,  it is interesting to recall that in most versions the generosity of Porsenna figures prominently.
What, finally, can be concluded about the care of hostages? Phillipson’s generalization appears correct: they resembled guests more than prisoners, but there were ways in which they resembled captivi,  and potential abuse was always a disquieting element in their detention.  Cultures like the Greek and Carthaginian, which allowed overt homosexuality and/or the exaction of female hostages, were, predictably, open to charges that hostages in their custody suffered sexual molestation. To the Roman government, which encouraged neither practice, accrued the additional advantage of favorably biased historians. Even in partisan sources, however, the Roman government occasionally emerges as having indulged in abuses, as the case of Demetrius of Syria proves, and it is quite probable that other examples of mistreatment have been glossed over, suppressed, or lost. The attitudes toward hostages which prevailed in the hostage’s own society may have contributed to their comfort or discomfort as much as Roman prejudices did, but evidence for these cultural attitudes is unfortunately sparse. 
Specifically, hostages to Rome were at least occasionally and probably regularly granted considerable freedom, they had access to their governmental representatives and to the Roman Senate;  they were allowed privacy in conversation (and even in conspiracy);  and some, although perhaps limited, physical mobility;  they received monetary support (again, perhaps limited) and marks of deference.  No single recorded incident illustrates all of these points, and the  chronological distance which these cases span produces considerable difficulty if one tries to discern a pattern. Yet it seems reasonable that this benevolence toward hostages was not restricted to the incidents discussed above, and that generous treatment was the standard intention, even if it was not always attained in practice.
This work is licensed under a Creative Commons License.