To refer to this please cite it in this way :

  • Cheryl Walker, Hostages in Republican Rome, Chapter 1: Meaning and Purpose of Hostageship in the Greco-Roman World [1-20], http://chs.harvard.edu/publications. Center for Hellenic Studies, Washington, DC. 2005

Chapter 1: Meaning and Purpose of Hostageship in the Greco-Roman World [1-20]

The circumstances in which persons acted as sureties in the Greco-Roman world may be classified under four basic categories: exchange, unilateral exaction by formal national agreement, private contract, and extralegal seizure. The words which describe such persons are ??σια, ?μηρος, ?ν?δοχοι, ?ν?χυρα, and obses in all these categories, although obses seems to have had a technical application to just the first three arrangements, which were bilateral agreements. [1]   Three of the Greek words, however, occur in situations only vaguely similar to those in which ?μηρος was used; ??σια, of persons held to compel restitution or restoration of property (Dionysius of Halicarnassus 5.33.3; 7.2.3); ?ν?δοχοι of persons who stand bail (Dionysius of Halicarnassus 6.84.2; Plutarch Dion 18.2 and 20.1; cf. Latin vades); and ?ν?χυρα of persons acting as pledges (Dionysius of Halicarnassus 6.84.2; Aeneas Tacticus 5.1; Appian Civil Wars 1.44; cf. Latin pignus). Because only the situation outlined in Appian bears a strong resemblance to the cases in which ?μηρος and obses are employed and because the other words are rare in this type of context (six incidents out of two hundred fifty-seven, or 2.3%), an analysis of the use of the words ?ν?δοχοι, ?ν?χυρα, and ??σια would have little effect on the bulk of the data collected here, for these variations occur too infrequently to permit any meaningful analysis. Let us now consider each of the basic categories enumerated above in [2] order to formulate the ancient concept of hostageship (i.e. ?μηρε?α and obsidatus) and its purpose, as explicitly noted in the ancient sources.

Mutual exchanges of hostages occurred when governments of approximately equal power wished to secure an agreement. [2]   Because bilateral exchange is a recognition of political equality, the history of Rome’s successful expansion provides few examples of such exchanges; Rome rarely recognized the political parity of another nation, and it is a striking fact that there is no surviving evidence for even a single instance of Rome’s involvement as an equal partner in a historical exchange. There are, however, some examples of Rome’ s opponents attempting to increase their power by associating themselves with other nations hostile to Rome. In 168 B.C. Perseus, king of Macedon, concluded an alliance (συμμαχ?α) with Genthius, king of Illyria, to gain assistance in his conflict with Rome. Polybius (29.3-4, followed by Livy 44.23.2-9) has given us an extremely detailed and precise account of the procedure of this transaction: first, the agreement of Perseus to the preliminary terms negotiated by Hippias and Perseus’ instructions to his envoy; second, Genthius’ oath of allegiance, followed by the dispatch of the Illyrian hostages and Genthius’ envoy to Perseus; next, Perseus’ oath of allegiance and acceptance of the Illyrian hostages, and the subsequent dispatch of Macedonian hostages to Genthius; finally, the release of the stipulated sum of money to Genthius’ envoy. This is the fullest extant description of the mechanics of hostage exchange. Since nothing from the other extant sources appears to contradict it, we may assume, with due caution, that it [3] describes a probable, but by no means certain, pattern of procedure for such exchanges.

The Romanized Italian cities which composed the Italic League in 91 B.C. determined upon hostage exchange as a means of guaranteeing the fidelity of the individual cities to their joint cause (?μηρα δι?πεμπον ?ς π?στιν ?λλ?λοις, Appian Civil Wars 1.38), thereby indicating the federal nature of the League. The importance of hostages to the alliance must not be underestimated, as if they were merely tokens; when the Roman magistrate Servilius learned of the exchange taking place at Asculum, his interference led directly to the violence which precipitated the Social War. Nor should we forget that the agreement between the Sequani and the Helvetii in 58 B.C. became the pretext for Caesar’s involvement in Gallic affairs outside his province; the Helvetii provided hostages to guarantee that they would not harm the territory of the Sequani, while the Sequani furnished hostages for their promise to permit peaceful passage to the Helvetii (Caesar Gallic War 1.9.4; 19.1). The suspicion with which Caesar regarded the man responsible for this agreement (Caesar Gallic War 1.19.1) suggests that such coalitions, however innocent their stated intentions, were considered by the Romans as potential threats to their national security. Caesar encountered three other such alliances, of various towns of the Belgae in 57 B.C. (Gallic War 2.1.1), of the Aquitani in 56 B.C. (Gallic War 3.23.2), and of the Carnutes in 52 B.C. (who abstained from exchanging hostages in order to prevent a premature revelation of their plans for revolt, Gallic War 7.2.2).

Two other cases deserve brief mention. According to legend, the Trojans and the Italian aborigines agreed to exchange hostages in [4] order to secure their treaty; Latinus kept Aeneas’ offspring Romulus and Remus, and upon Latinus’ death without heirs of his body, the erstwhile hostages succeeded him (Dionysius of Halicarnassus 1.59.1-2). The second case is of interest because the exchange was not between governments or tribes but between the economic classes of a single national unit. In 173 B.C., in an attempt to settle internal unrest, Aetolian creditors and debtors established hostages from both groups at Corinth (Livy 42.5.12). [3]   Both stories indicate that hostage exchanges were expected to bring together peacefully the disparate elements of a single regional or political entity.

By far the greatest number of incidents involving hostages belong to the second category, unilateral exaction by the terms of a formal agreement between two nations. The vocabulary which may be used to describe such arrangements is quite broad and for the most part not employed technically. A perusal of the “terminology” column in Appendix I.A indicates its range. [4]   Some of the words are general, meaning almost any sort of negotiated settlement: ?μολογ?α, προτειν?μενα, σπονδα?, συγκε?μενα, συνθ?και, συντ?θημι, pactio, and derivative forms from these words. Others are more precise. The Greek ?νοχα? and the Latin indutiae appear to refer only to truces, and they are not used to signify a termination of the war. The word “peace” (ε?ρ?νη, pax) is interchangeable with almost any permanent settlement: φιλ?α, διαλ?σεις, συμμαχ?α, συναλλ?γη, foedus, deditio, dicio, amicitia, and imperium. Even terms which might be expected to connote political equality of the contracting parties (e. g., συμμαχ?α) could be coupled with another word which clearly signifies [5] the inferiority of one party (e.g., ?γεμον?α); [5]   we may here compare the socii, who were most frequently not the equals of the Romans, but a people more or less subject to Roman domination.

With regard to vocabulary, Livy’s discussion of the peace settlement at the Caudine Forks is very illuminating. Livy repeatedly calls Pontius’ proposal before its acceptance a foedus (Livy 9.4.4; 5; 7; 5.1; 2), one in which the terms of peace are, with the exception of the yoke, equal to victor and vanquished (Livy 9.4.3-4). But, according to Livy, the final agreement between the Roman officers and the Samnites was a sponsio, which differs from a foedus in that a foedus requires ratification by the Roman people and the official fetial ceremonies (Livy 9.5.1-2). Most importantly, he continues his argument for a sponsio by asking, quite rhetorically, what need of sponsores or of hostages a foedus would have, since the fetial oath makes the contracting nation as a whole responsible for fidelity to the oath and for their actions (Livy 9.5.3). We should be able to conclude from this passage that states which agreed upon a foedus did not exact hostages. Unfortunately, Livy himself employed foedus to describe the contract by which hostages were exacted unilaterally: the treaties with Porsenna in 508 B.C. (Livy 2.13.8,9), with the Lucanians in 298 B.C. (Livy 10.11.3), with the Aetolian League in 189 B.C. (Livy 38.11.6), and with Antiochus III in 188 B.C. (Livy 38.38.1,9). I do not believe that, in Roman historical writing, foedus can be restricted to the narrow technical sense which Livy assigned it in this idiosyncratic passage. [6]  

There are in practice few truly technical words applied to situations in which unilateral exaction occurred. The Latin deditio [6] seems to have a fixed and precise meaning and generally signifies that the dediticii gave hostages. Indeed, Scipio Nasica argued in 19B.C. that he merited a triumph over the Boii, and as proof of his military subjection he cited the exaction of hostages from them (Livy 36.39.3; 40.3). Yet hostage exaction was perhaps not an unfailing part of deditio. Consider the argument concerning the surrender of Ligurian villages in 197 B.C.: the tribunes maintained that deditio without any pignera was fictitious (Livy 33.22.9). Since the Roman commander accepted the surrender without demanding hostages and still asserted the validity of his claim for a triumph, we cannot state unequivocably that every deditio invariably required submission of hostages. Still, it is clear that deditio normally did involve such a submission.

One further point on vocabulary should be mentioned. Moscovich, observing the frequency with which pignus occurs in conjunction with obses, correctly described the term’s application to hostageship: the donor, who had surrendered immediate possession of the object/person pledged as pignus, received the pignus back upon fulfillment of his obligation; in case of default, the pignus passed into the full possession of the recipient. [7]   Yet another word, arrabo, was also employed in connection with obses. Aulus Gellius (Attic Nights 17.2.21) remarked that Claudius Quadrigarius used arrabo of the Roman equites whom the Samnites took in 321 B.C. for the Caudine Forks agreement because arrabo is gravior acriorque. Rather than signifying that the equites were, as it were, the collateral which pignus indicated, the word arrabo suggests that the equites were a non-returnable part of the obligation itself, a kind of down payment. [8]   It is regrettable that only this single reference to persons as arrabo has survived from the [7] Republican period, for the Augustan prose of Livy does not necessarily reflect Republican terminology. [9]  

What did the hostages exacted by a treaty guarantee? The answer to this question may seem obvious, but it is in fact disputed. Täubler, followed by Gsell and de Sanctis, believed that the hostages secured the financial clauses of the agreement; [10]   Aymard, however, argued that the hostages from the Treaty of Zama guaranteed, not merely the indemnity, but the entire treaty. [11]   The easiest method of resolving the question is to examine those instances where an explicit purpose is assigned to the exaction by the ancient sources. In direct support of Täubler’s hypothesis, we may cite the arrangement of the Samians to give hostages as security for their indemnity payments (Plutarch Pericles 88.1), the proposal of Nicias that the Syracusans keep Athenian troops to guarantee the Athenian repayment of Syracusan war expenses (Thucydides 7.83.2; Plutarch Nicias 27.2), and Caesar’s demand for hostages for the promised subsidies (Caesar Gallic War 6.2.2). But other reasons are better attested: as an alternative to one of the recipient’s conditions (garrison, Livy 2.13.4; Josephus Jewish Antiquities 13.247-48; a forced move, Livy 40.38.5); [12]   a complete set of conditions (?π? το?τοις, Zonaras 9.26; Plutarch Pelopidas 27.3; Poplicola 18.2; Moralia 250B; ?π? ταις συνθ?καις, Appian Illyrian Wars 28; περ? το?των, Diodorus 19.75.1-2; Polybius 15.8.7; τ?ς συνθ?κης, Polyaenus 1.47.2); and to prevent further hostilities and/or disobedience (Livy 9.5.5; 22.22.5; 27.24.1; Dionysius of Halicarnassus 2.55.5; Josephus Jewish Antiquities 13.204-206; Appian Samnite History 4; Punic Wars 54; Illyrian Wars 23; Suetonius Augustus 21.2; Polybius 36.11.3). Obviously, hostages could secure more than the financial clauses of an agreement. [13]   [8]

Less obvious is what our sources meant by the frequent reference to hostages as π?στις or fides (Polybius 15.18.8; 21.17.8; 31.2; Diodorus 20.104.3; Dionysius of Halicarnassus 9.17.3; Livy 21.34.3; 37.45.16; Caesar Gallic War 8.27.1; cf. Dionysius of Halicarnassus 6.30.1; Livy 34.35.11). Fides appears to be the verbal expression of an agreement (Livy 44.25.8), for which the hostages themselves are the physical expression. Thus the description of hostages as π?στεις or as fides seems to suggest that the hostages guaranteed the entire contract to which the donor had agreed, since the donor must abide by all conditions of the agreement if he is to maintain it.

Hostages might also be exacted according to a less public agreement, in which the contracting parties did not officially represent their governments and the hostages did not guarantee a formal written treaty. Instead, such hostages appear to have acted as sureties against treachery in a number of different situations. Enemy commanders sometimes wished to open preliminary negotiations with their opponents, in both international and civil warfare. Because (for self-evident reasons) the weaker general usually requested a conference as a means for obtaining a peaceful solution, the stronger general, who was less concerned with the prospect of peaceful negotiations, demanded some pledge in return for the favor of a conference (Livy 42.39.6-7; cf. Livy 28.35.4; Caesar Civil War 1.84.2; [14]   Plutarch Eumenes10.2-3; Cleomenes 17.1 and Aratus 39.2; Moralia 197A; Appian Civil Wars 1.85). [15]   Similarly, when a mercenary or an alleged traitor offered his services, a wise leader took sensible precautions (Diodorus 17.23.5; Livy 10.10.3; 43.10.3; [16]   Dionysius of Halicarnassus 7.10.6; Plutarch Pelopidas 9.5-7 and Moralia 595A-E; Appian Hannibalic War 47). [17]   Perhaps the most frequent exaction of this sort [9] was that which intended to secure the dominance or safety for one faction during civil war (Livy Summaries 84; Caesar African War 26.5, [18]   Civil War 3.12.1-2; Valerius Maximus 6.2.10; Frontinus Stratagem 2.11.1; Appian Civil Wars 1.44; [19]   3.91; Plutarch Cicero 18.1; Cicero Letters to Friends 10.17.3; and the most famous of these, the sons of Antony and Lepidus who became hostages for the safety of Caesar’s assassins, Livy Summaries 116; Velleius Paterculus 2.58.3; Cicero Philippics 1.13.31; 2.36.90; Dio 44.34.6; Appian Civil Wars 2.142; 3.15). Finally, there was the apparently Greek custom of offering hostages (usually oneself) for the veracity of one’s statements (Herodotus 8.94; 9.90; Nepos Themistocles 7.2; Phocion 2,4) or one’s promises, particularly if they involved money (Livy 44.25.7-8; 26.11; Velleius Paterculus 2.42.1; Polyaenus 4.6.17). For the hostage, the major differences between these types of circumstances and the formal treaty exactions were in the length and purpose of the exaction; as these informal circumstances tended to generate very specific conditions (e.g., that the donor observe a strict truce during the conference or that the traitor not lead the recipient into an ambush), the hostages seem to have been released upon the bona-fide conclusion of the conference, successful surprise attack, cessation of hostilities, or payment of the promised sum. Thus, such persons might serve for some few hours or weeks instead of for years, as did formal hostages. [20]  

Inevitably, there are incidents which, because of our ignorance of their details, lie between the second and third categories. Sometimes it is uncertain whether an exaction occurred between two generals of opposing armies as individuals or as official representatives of their government. Another problem is posed by exactions from peoples already allied, in addition to the previous agreement; this was evidently [10] standard Roman policy when there was fear that the allied people might conspire and revolt. [21]   It is difficult to classify these as private contracts, because the exaction had the approval of the Senate, but at the same time it is difficult to prove that such an exaction had the assent of the donor government. In the absence of better information, I have left these ambiguous cases uncategorized.

The last category to be considered here [22]   is that which parallels the contemporary use of the word “hostage” today, i.e., an unwilling person, usually captured or forcibly detained. Although Greek authors regularly called persons so detained ?μηροι. (e.g. Polybius 1.68.3; Dionysius of Halicarnassus 4.83.3; 5.31.2; 6.53.1; 6.62.5; 8.43.4; Diodorus 33.15.1; Plutarch Cicero 18.1; Sertorius 14.2; Appian Civil Wars 3.91; 5.137), Latin authors employed obses with some circumlocution (e.g. pro obsidibus, Livy 33.14.2-3; 36.11.11; obsidum loco, Livy 31.25.8; Caesar Gallic War 5.5.4; Civil War 1.74.5). From this we may with caution infer that obses was a technical term which appears properly to have described one exacted by a bilateral voluntary agreement. Relatively few extralegal seizures are attested, but the evidence for many such cases is undoubtedly hidden under other words. For example, native levies serving with the Roman army are very occasionally referred to as hostages (Livy 23.4.8; 7.2; 31.10; 40.47.10 et al.) [23]   as are prisoners of war (Dionysius of Halicarnassus 3.57.1). [24]   Since no agreement restricted their selection, treatment, or release, their fate remained entirely at the discretion of their captor; unfortunately, the subsequent activities of such extralegal hostages are for the most part unknown, and it is this failure of our sources which makes this category a historical dead end. [11]

In general, then, a hostage in the Greco-Roman world guaranteed in his person the fulfillment of obligations voluntarily undertaken by his party. [25]   Although the circumstances of his exaction and the obligations he secured varied enormously, the terms ?μηρος and obses are consistently applied. Let us consider the attitudes which the Greek and Roman authors have transmitted to us and the religious and legal positions of hostages in order that we may then try to formulate a definition for the concepts of ?μηρε?α and obsidatus.

Cultural Attitudes Toward Hostageship

It should be clear from the circumstances under which unilateral exaction occurred that such a situation could easily become a means of forcing the donor nation to acknowledge, at least implicitly, inferiority; [26]   one might reasonably expect the recipient government to express condescension or insolence and the donor government to resent the institution of hostageship and perhaps the hostages themselves, as the visible tokens of their inferiority. The picture presented by our evidence is in truth rather less predictable. Let us first consider the attitudes found in Greek cultures.

That the Spartan kings Agesipolis (Plutarch Moralia 215B-C) and Cleomenes (Plutarch Cleomenes 22) felt the delivery of fellow citizens (or at least, in Cleomenes’ case, his mother) to a foreign power to be shameful cannot be seriously disputed, despite the lateness of the attestations, for a similarly negative response is apparent elsewhere. [27]   The comic poet Antiphanes, whose floruit is the first half of the fourth century (Athenaeus 15.681C), and the geographer Strabo (11.14.15 C532) both employed the paired antonyms sovereignty-hostageship, as if these two [12] concepts were the opposite ends of a single continuum; the former uses hostageship to demonstrate how far the mighty Spartan empire had fallen, and the latter describes the extremes of fortune experienced by Tigranes the Great of Armenia, who had been a hostage in Parthia and then become a powerful monarch. Nor was offering a guarantee for a personal agreement [28]   less distasteful than hostageship which secured a treaty. Evidently the exaction of a hostage for a private contract suggested that the donor’s promise, if unsupported by a pledge, was insufficient to compel his cooperation. When Charon brought forth his young son as an object of vengeance, should he betray the Theban conspirators to the Spartans in 379 B.C., Pelopidas and the others repudiated his proposal because of the insult that acceptance would have offered to Charon’s loyalty (Plutarch Pelopidas 9.5; Moralia 595A-E). [29]  

There seems little difference between Greek and Roman attitudes in this regard. The same distaste for surrendering hostages is obvious - indeed, intensified - in Latin sources, as an examination of Roman law [30]  and such accounts as the aetiologies of the festival of the Capratine Nones (Plutarch Moralia 313A; Romulus 29.4; Camillus 33.2-7; Macrobius Saturnalia 1.11.35-40), the descriptions of the lamentations on the report of the submission of hostages at the Caudine Forks (e.g., Gellius Attic Nights 2.19.8 and Appian Samnite History 6), and the Pompeian brutality toward the Africans in 46 B.C. (liberos eorum obsidum nomine in servitutem abripi, African War 26.5) demonstrates.

Yet the Roman awareness that hostages taken from other states were visible tokens of Roman supremacy did not necessarily give rise to patronizing treatment or arrogance. [31]  Philip V of Macedon decided to send his son Demetrius as his ambassador to Rome because of the good [13] will [32]   with which the senators had previously acted toward Demetrius during his residence in Rome as guarantee of his father’s obedience to the Senate (Polybius 22.14.9-11; Livy 39.35.2-3; Justin 32.2.4). Antiochus IV Epiphanes’ envoy to the Senate after his accession declared that Antiochus had been treated, not as a hostage, but as a king (Livy 42.6.9). [33]   Dionysius of Halicarnassus could relate, and apparently believe, that Latinus had adopted the hostages Romulus and Remus in the absence of any heirs of his own (Dionysius of Halicarnassus 1.73.2). Further, we may infer that, by the time of the early Principate, there was some revulsion at the idea that innocent hostages could be used to expiate the guilt of their countrymen (Livy 28.34.7-10). [34]   Nevertheless, it is important to distinguish between the gallantry which the Romans could display toward vanquished foreigners and the repugnance shown toward the concept of Roman hostages.

But there was another reason besides the overt loss of sovereignty and fear for the hostage’s safety which could bring discomfort upon the donor government: the reaction known today as “Stockholm syndrome.” Certainly recognized, although not named, in antiquity, the Stockholm syndrome is the phenomenon whereby persons physically dominated in stressful circumstances identify themselves with their captors. Phillipson has observed that former hostages served after restoration as proxenoi to the citizens who had detained them. [35]   The attested histories of the three hostage princes of the early second century B.C., Demetrius of Macedon, Antiochus of Syria, and Demetrius of Syria, refer to their assimilation to Roman culture as if it were a rhetorical commonplace (which it may be): Perseus’ accusation that Demetrius’ soul is Roman (Livy 40.5.12, supported by 39.47.10), Antiochus’ later [14] indulgence in inappropriate Roman customs such as canvassing for office and wearing magisterial garb (Polybius 26.1.5-6; Diodorus 29.32; Athenaeus. 5.193D-196A; 10.438D-439D), and the assertion of the Seleucid Demetrius that he considered the senators his fathers and their sons his brothers (Polybius 31.2.5) suggest the strong influence which their residence in Rome was believed, rightly or wrongly, to have upon them. [36]  

Cultures contemporary with Republican Rome appear even less enamored of hostageship and hostages, although my data are by no means exhaustive. [37]  Gsell accounted for Punic unpopularity in Spain during the Hannibalic War as resulting from the Carthaginian dependence on hostageship. [38]   The military pretensions of the Helvetii in 58 B.C., as recorded by their opponent Caesar (Caesar Gallic War 1.14.7; Dio 38.33.1), equal in haughtiness even the arrogant Roman claim to sovereignty; in the campaigns of 56 B.C., Caesar maintained that a revolt was due to resentment over the exaction of children as hostages (Caesar Gallic War 2.5). [39]   The Parthians extended the disgrace of hostageship to the persons who experienced it: τ?ν ?μηρε?αν ?ντ? δουλε?ας ?νομ?ςοντες κα? τ?ς ?πικλ?σεως τ?ν ?δοξ?αν (Josephus Jewish Antiquities 18.47). [40]   Again, hostageship is the opposite of sovereignty; it is slavery. This, despite the kingly manner in which Augustus maintained the Parthian princes (Strabo 16.1.28C749) and the fact that they were not actual hostages. [41]  

Thus the donor government had cause to fear, not only the confession of weakness which surrender of hostages signified, but also the more subtle threat of the hostages themselves, whose patriotism could be successfully undermined. The recipient government, however, had much to gain and little to lose by treating hostages in a dignified, [15] honorable fashion. [42]   The ambiguity of the hostage’s position, distrusted by his countrymen and courted by his country’s enemies, could not be clearer.

Religious Standing

Westington has remarked that hostages enjoyed sacrosanctity in the absence of any disobedience by the donor state; he based this opinion on the sole testimony of a passage in Dionysius of Halicarnassus. [43]   After the escape of Cloelia and the other women, these hostages and some ambassadors were returning to Porsenna when Tarquin’s party ambushed them. Porsenna, outraged at this conduct, terminated his association with the Tarquins and allied himself to the infant Republic (Dionysius of Halicarnassus 5.34.1-2). What outraged Porsenna was, I submit, not the violation of the hostages’ supposed sacrosanctity, but that Tarquin had ambushed an official Roman party during a truce to which Tarquin himself had agreed. It is the truce which made the persons of the ambassadors and hostages ?ερ?, not any inherent quality in hostageship itself.

Although accredited ambassadors were certainly granted diplomatic immunity in accordance with ius gentium, it is difficult to understand how hostages could have shared this benefit. Even in so flagrant a case of ambassadorial misconduct as tampering with delivered hostages and instigating their escape, the guilty diplomat, captured with the fleeing hostages, may have tried to claim immunity (Livy 25.7.11); [44]   but no hostages other than Dionysius’ Romans of legend invoked any such religious right. Indeed, the Romans do not seem to have observed extraordinary religious scruples upon capturing the [16] hostages of any other nation. [45]   Moreover, conditional sacrosanctity (i.e., sacrosanctity so long as the donor state preserved the agreement for which the hostages were exacted) does not seem to have been an element in the ius gentium, for interpretations of the violation or non-violation of an international agreement are particularly influenced by partisan evaluations of expediency. [46]  

Nevertheless, hostages in all likelihood did enjoy a de facto protection on religious (or quasi-religious) grounds: unprovoked hostility toward them violated the agreement under which they were exacted, an agreement which often sought divine vengeance upon the violators. [47]   Insofar as oaths were a binding force upon those who swore them, religion was a safeguard. Even in a relatively early period, however, religion was sometimes insufficient against the demands of expediency or the promptings of anger or indignation. [48]   In short, hostages might generally have invoked an honored and protected status on religious grounds, based on the terms of the treaty by which they were exacted, but this was not sacrosanctity, and the surviving evidence does not permit us to assume that any particular privileged status was granted to hostages solely on the basis of their hostageship.

Legal Status

The legal status of hostages at Rome is somewhat more distinct, although by no means clear. [49]   Our evidence for testamentary procedure belongs to the legal digests of the Empire, but Republican practice probably differed very little. Dediticii could not make wills as Roman citizens, since they were foreigners; but neither could they claim [17] testamentary rights as peregrini, for they were not citizens of an independent (i.e., non-subject) civitas (Gaius 1.25; Ulpian Digest 20.14). Hostages whose states had surrendered unconditionally in this manner appear to belong to this category. (Hostages from politically autonomous nations presumably retained rights as peregrini, although this is a surmise on my part.) As a legally distinct group, hostages might bequeath their possessions only by special permission, and in the absence of such permission, the Imperial fiscus confiscated the goods (Ulpian Digest 28.1.11); perhaps the treasury did so during the Republic. In some cases hostages might even inherit legacies from Roman citizens, provided that they were regarded as citizens rather than as hostages (Marcianus Digest 49.14.21-32). [50]   Such a law may, however, have been intended to guarantee the rights of a Roman citizen to complete testamentary freedom and only incidentally to grant the hostage a legal privilege. It is possible that Roman law recognized other rights as well, particularly such private rights as marriage and parental authority, but no evidence concerning these private rights has survived. [51]  

The rare cases of Roman citizens who became hostages - there are three Republican episodes, the Cloelia legend, [52]   the Caudine Forks disaster, [53]   and the defeat of Cassius Longinus - are more fully covered by the surviving law codes. Roman citizens captured and held by the enemy lost legal control over other Roman citizens de facto and de jure; children in patria potestas were forthwith released until potestas was regained by ius postliminii, in the event of the parent’s safe return from the enemy (Gaius 1.129; Ulpian Digest 10.4; Paulus 2.25.1; Theophilus 1.12.5). [54]   It is possible that Roman citizens surrendered [18] as hostages would have suffered similar loss of status and enjoyed similar rights of recovery unless these were specifically excluded by treaty conditions. [55]   The only historical incident of the Republican period which involved Roman hostages is infuriatingly incomplete; we know nothing of the hostages’ identities, their treatment, or their release. After the decisive defeat of Cassius Longinus’ army by the Tigurini in 107 B.C., the senior surviving officer, C. Popillius, was compelled to deliver hostages and half the army’s baggage in order to secure a safe withdrawal for the remaining troops (Livy Summaries 65; Orosius 5.15.24; Appian Samnite History 1.3; Caesar Gallic War 1.7.12: Rhetoric to Herennius 1.15.25). [56]   He was then tried and convicted, probably on a charge of maiestas. Whatever the specific law on which he was arraigned, handing over a civis to the enemy was a capital crime (the Twelve Tables, cited in Marcianus Digest 48.4.3). This law does much to explain the obstinate and sometimes suicidal resistance displayed by Roman commanders in desperate straits; this law made serious attempts at negotiations almost impossible, since the commander’s life was forfeit even if he did preserve some of the army by giving hostages.

Nor is this attitude toward delivering hostages surprising. The surrender of one’s or a fellow-citizen’s person, however limited and conditional in theory, could in practice be viewed only with the gravest suspicion; there was little or no appeal or recourse for the helpless hostage. The hostage’s rights, such as they were, lay entirely in the hands of his keepers and could be denied or refused at any time. The two appeals made by Demetrius of Syria for his release indicate the cynical ease with which the Roman Senate could ignore just claims (Polybius 31.2; 11-15), [57]   and at least two Greek incidents [19] exhibit the inability of a donor government to enforce a proper observance of an agreement upon the stronger recipient nation (Polyaenus 1.47.2; Appian Mithridatic Wars 46-47). Indeed, the very distaste which flavors Roman response to the concept of submitting Romans as hostages strongly suggests that such status was open to great abuses and that the unenviably vulnerable position of hostages was in all likelihood the normal ancient attitude rather than the exception. [58]  

With this understanding of the vocabulary, purposes, and cultural responses to hostageship, we are ready to formulate our definitions of ?μηρος and obses. The major distinction between the Greek ?μηρε?α and the Latin obsidatus is the exclusion of extralegal seizures from the Roman concept; thus the Roman terminology differs from both the Greek and the English usage of the respective words ?μηρος and “hostage” as generally applied. For the Greeks, the ?μηρος was a person in the control of the recipient, one who acted as a bond for the fidelity of the friends and relatives of the ?μηρος, whose safety and whose rights were at the recipient’s discretion. For the Romans, the obses was a person, voluntarily submitted, who guaranteed the performance of a bilateral agreement and whose treatment, as a consequence, was theoretically limited by the restrictive clauses, if any, which were included in the agreement. [59]  

From these definitions, we can better understand the cultural attitudes toward hostageship and the anomalous position which hostages held in ancient societies. Because hostages secured an agreement, their religious and legal status had to permit the recipient appropriate [20] recourse against the hostages if the agreement was violated; because unilateral exaction signified the donor’s inferiority and because the hostages were in practice at the mercy of the recipient, hostages were, in a sense, in bondage and not free agents, a cause for social stigma. Further, we may anticipate from the distinction between Greek and Roman hostageship that Roman hostages perhaps more rigidly adhered to selection criteria and to predetermined treatment and restoration. Let us now employ these definitions to investigate the conditions and careers of the people designated as ?μηροι and obsides.

 1. For the restricted use of obses, see below, p. 10.    back
 2. Aymard, “Les otages barbares,” 137; Coleman Phillipson, The International Law and Custom of Ancient Greece and Rome, 2 vols. (London: Macmillan & Co., 1911), 1: 399; pace George Rawlinson, The Sixth Great Oriental Monarchy: or, The Geography, History, and Antiquities of Parthia (New York: Dodd, Mead, & Co., 1872), p.131.    back
 3. For the use of unilateral hostages in similar cases of factional dispute, see below, 8-10.    back
 4. See Appendix I.A. below, 212-244.    back
 5. Although such explicit couplings are rare, it is often manifest from the context of the exaction that one party is inferior.    back
 6. There is at least one case in which Livy's distinction is maintained. Lamenting the pitiable state of Gaul after Caesar's conquest, Orosius remarks ...profiteri sponsionem servitutis aeternae, avulsis insuper obsidibus, cogeretur... (Orosius 6.12.4). The addition of insuper certainly suggests that hostages were not an expected part of the sponsio's burden.    back
 7. Moscovich, “Obsidibus Traditis”, 126. The list in footnote 35 errs twice; it should read Livy 36.40.3 and Justin, Epitome of the Philippic History of Pompeius Trogus, 7.3.2; it appears to be used in place of obses only at Livy 33.22.9.    back
 8. Lewis and Short, A Latin Dictionary (Oxford: 1969), 165, s.v. "arrabo."    back
 9. The coupling of obses and pignus in Cicero's Pro Caelio 32.78 refers not to people, but to prosecutions; for this rhetorical usage, see below, 261. The alliterative power of pignus pacis may explain its popularity with the stylist Livy (Livy 2.13.9; 9.15.7; 28. 34.9; 36.40.3; 40.15.8; cf. Silius Italicus, Punica 10.494; Justin Epitome of Philippic History of Pompeius Troges, 7.3.2).    back
 10. Eugen Täubler, Imperium Romanum: Studien zur Entwicklungsgeschichte des Römischen Reichs, (Rome: Bretschneider, 1964), p. 1 n. 3. For acceptance of this thesis without comment, see Stéphane Gsell, Histoire Ancienne de L'Afrique du Nord, 6 vols. (Paris: Librairie Hachette et Cie., 1918-1924), 3: 294 n. 7; Gaetano de Sanctis, Storia dei Romani, 4 vols. (Florence: La Nuova Italia, 1953), 3: 2: 605 n. 15; cf. P.C. Sands, Client Princes of the Roman Empire (Cambridge: The University Press, 1908), pp. 51-52.    back
 11. Aymard, “Les otages carthaginois,” pp. 436, 449-450; for individual cases in which hostages did not guarantee money, see F.W. Walbank, A Historical Commentary on Polybius, 3 vols. (Oxford: Clarendon Press, 1957-1979), 1: 604, 3: 134-35.    back
 12. Although this alternative proposal was rejected, it still indicates a possible purpose for hostage exaction.    back
 13. Nevertheless, we must not dismiss as pure coincidence the frequent proximity of the financial and hostage clauses and the correspondence of the periods over which both the indemnity and the hostages were required. The Aetolian Treaty of 189 B.C. furnishes the clearest correlation of the two clauses (Livy 38.11; Polybius 21. 42); cf. Aymard, “Les otages carthaginois,” 436-37; Walbank, Polybius, 3: 134-35; de Sanctis, Storia dei Romani, 4: 1: 258-62; and the correlation of hostages and tribute in Irish legend (e.g., Whitley Stokes, "The Death of Crimthann son of Fidach, and the Adventures of the Sons of Eochaid Muigmedon," Revue Celtique 24 (1903): 181.    back
 14. The slight circumlocutions found in Livy 28.35.4 (pro obsidibus) and Caesar B. C. 1. 84. 2 (obsidis loco) may suggest that perhaps obses originally did not refer to this kind of exaction and was only later applied, as an extension of formal hostageship.    back
 15. Note the similarity of expression in obsides dare nihil fraudis fore...pignus fidei, Livy 42.39.607 and pignus afore fraudem, Livy 43.10.3.    back
 16. There is a case of a third party asking for hostages from two other groups. Antigonus Gonatas and Pyrrhus were fighting in the neighborhood of Argos, and Argos sought to resolve the quarrel by establishing a truce. Both Antigonus and Pyrrhus were asked to give hostages, but only Antigonus did so (Plutarch Pyrrhus 31.2).    back
 17. For a discussion of mercenaries supplying hostages to guarantee a private contract, see below, pp. 31-33.    back
 18. Again, a sort of circumlocution: liberos eo rum obsidum nomine in servitutem abripi.    back
 19. Appian does not use ?μηρος here, but ?ν?χυρον see above p.1.    back
 20. See below, pp. 58-72, for the duration of the hostageship for formal agreements.    back
 21. I have omitted a full-length discussion of the rhetorical and non-personal uses of ?μηρος and obses because neither is helpful in understanding hostageship. See Appendix 1. C. below, p. 261.    back
 22. A.H. McDonald, "Rome and the Italian Confederation (200-186 B. C.)," JRS 34 (1944): 13; cf. Diodorus 23.5 for a Punic example of this sort of exaction.    back
 23. It is not unusual that these military units are called hostages, for their repatriation might well have depended upon their own and their country's faithful obedience. Such levies, simultaneously hostages and soldiers, may be the basis for the seemingly inflated numbers of hostages, e.g., Diodorus 24.10.2; cf. Ulrich Kahrstedt, Geschichte der Karthager von 218-146, 3 vols. (New York: Arno Press, 1975), 3:105, 110, 495-498. Cf. Pluto Moralia 865 and Norman J. De Witt, Urbanization and the Franchise in Roman Gaul (Lancaster, PA: Lancaster Press, 1940), pp. 7-8.    back
 24. Even a Greek source preserves a circumlocution: ?ν ομ?ρων...λ?γ?.    back
 25. The Irish legends mention another cause for exaction not found in the Greco-Roman material. According to these legends, hostage exaction was an integral part of sovereignty, so integral that a lack of hostages nullified one's claim to kingship. Cf. Myles Dillon, The Cycles of the Kings (London: Geoffrey Cumberlege, 1946) pp. 37, 82-83; Eugene O'Curry, On the Manners and Customs of the Ancient Irish, 3 vols. (Williams & Norgate, 1873), l: ccxxxviii; for hostages as a surety for one's promise, Whitley Stokes, "Prose Tales in the Rennes Dindsenchas ," Revue Celtique 15 (1894): 313; Cross and Slover, Ancient Irish Tales (New York: Henry Holt and Co., 1936), pp. 30, 33, 109 (of wolves!), 116, 256-258. Hostages may also have figured in arbitration (Dillon, Cycles, pp. 58-59), although the Greeks appear to have avoided hostage exaction in seeking judgments; cf. Adcock and Mosley, Diplomacy in Ancient Greece (New York: St. Martin's Press, 1975), p. 214.    back
 26. See Aymard, “Les otages barbares,” p. 137, and above, pp. 4-5.    back
 27. Whether the anecdote of Agesipolis' retort to a heckler belongs to the fourth or the third century seems to me uncertain; see below, p. 81 n. 10.    back
 28. For the distinction between national and private agreements, see above, pp. 8-10.    back
 29. Plutarch is a more reliable authority for this anecdote because his major source for the Pelopidas was evidently a contemporary fourth-century historian. H. D. Westlake, "The Sources of Plutarch's Pelopidas,CQ 33 (1939): 11-22.    back
 30. See below, pp. 16-19.    back
 31. Since all the sources for this paragraph display a marked pro-Roman bias, we should perhaps allow for some exaggeration of Rome's magnanimity. Still, there is no evidence that this magnanimity is pure invention.    back
 32. According to Livy, it was Demetrius' kingly disposition which gained their good will (obses specimen indolis regiae dedisset...,Livy 39.35.3). This grouping of obses/rex, like the remark of Antiochus's envoy, may be rhetorical elaboration, but it probably refers to a real dichotomy of status.    back
 33. Note again the contrasting antonyms: …ut pro rege, non pro obside…    back
 34. Livy has here used innoxii and saeviturum with reference to hostages and their punishment. Such loaded words clearly demonstrate the attitude that Livy thought appropriate for Scipio Africanus, but this passage is unusual. Except for proper adjectives which describe ethnic    back
 35. International Law and Custom, 1: 151. Phillipson cites the decree which approved the treaty between Athens and Selymbria (Russell Meiggs and David Lewis, A Selection of Greek Historical Inscriptions to the End of the Fifth Century B.C. (Oxford: Clarendon Press, 1969), #87, pp. 267-269), which names the Selymbrian Apollodorus as a restored hostage and confers the proxenia upon him. The other incident on which Phillipson based his remark is an inscription which in seemingly conventional phraseology confers proxenia for the Achaean League upon Phocians and Boeotians (CIG 1542). Cf. Hatto H. Schmitt, Die Staatsverträge des Altertums, vol. 3: Die Verträge der griechisch-römischen Welt von 338 bis 200 v. Chr. (Munich: C.H. Beck, 1969), #507, p. 215.    back
 36. Cf. also the dependence of Demetrius of Macedon and Vonones I of Parthia on Roman support. Aymard (“Les otages barbares,” pp. 139-142) eloquently discusses this issue with regard to the practice of Augustus as described by Suetonius (Aug. 21.2) and its possible effects upon the barbarian chiefs who furnished hostages to Rome.    back
 37. Since I have concentrated my research upon Republican Rome, I have not searched systematically for information regarding Carthaginian, Palestinian, Gallic and other non-Greco-Roman cultures.    back
 38. L’Afrique du Nord, 3: 166.    back
 39. One might have thought a Gallic revolt the last thing desired by the parents of the hostages, but see below, pp. 177-182.    back
 40. It is most unfortunate that no extant source attests the Greek or Roman attitude in such unequivocal language.    back
 41. See above, pp. 1-11 and below, pp. 127-128. Later Indo-European societies, like the Irish Celts and the Scandinavians in Iceland, seem to have treated hostages rather as Augustus did these Parthians or even better. They are described as foster relatives or as royally supported guests within the king's household. Dillon, Cycles, p. 75; John O'Donovan, Leabhar na g-Ceart or the Book of Rights, (Dublin: Printed for the Celtic Society, 1847), p. 135; Stith Thompson, Motif-Index of Folk Literature (Bloomington, Ind.: Indiana University Press, 1957), vols. 5 and 6.    back
 42. See Chapter III below, pp. 94-143.    back
 43. Mars McClelland Westington, Atrocities in Roman Warfare to 133 B.C. (Chicago: Private edition distributed by the University of Chicago Libraries, 1938), p. 31; Moscovich ("Obsidibus Traditis," pp. 126-127) cites Dionysius without comment. Phillipson (International Law and Custom, 1: 406) denies absolute immunity but indicates the conditional inviolability of hostages.    back
 44. Although Livy has certainly scorned Phileas' claim to status (per speciem legationis, Livy 25.7.11), it seems likely Phileas must have been a genuine envoy from Tarentum; the most plausible reason for the pointed denigration of Phileas’ accredition is that Phileas tried to claim immunity.    back
 45. For Roman treatment of captured hostages who had been delivered to a nation at war with Rome, see below, pp. 182-190.    back
 46. Treaties rarely specify every contingency that may arise, and therefore the exact point at which a treaty may be judged broken is usually a matter of interpretation, much influenced by patriotic considerations.    back
 47. Certainly foedera which were struck in accordance with fetial ceremonies incorporated such oaths (Livy 1.24.7-9). Although in one passage (9.5.1-2) Livy asserts that such foedera do not require hostages, he is inconsistent on this point of technical vocabulary; see above, p. 5. In Greek practice, of course, these oaths are well attested (e.g., Polybius 29.3-4).    back
 48. Cf. the summary execution of Volscian hostages in the early years of the Republic (Livy 2.16.9; Dionysius of Halicarnassus 6.30.1). In the event that the recipient government violated the agreement, the donor seems to have had few options. In none of the surviving accounts of such cases are the subsequent careers of the hostages recorded. The cities of Byzantium (Polyaenus 1.47.2), Cauca (Appian Hispanica 52), and Chios (Appian Mithridatica 46-47) were seized and looted in disregard of their conditional surrender; Tryphon killed his prisoner Jonathan Maccabaeus despite the delivery of Jonathan's two children as hostages for Jonathan's release (Joseph. Jewish Antiquities 13.204-206); Oropus' attempts to pressure Athens into providing redress led to some of the unrest which prompted Roman intervention and destruction in 149 B.C. (Pausanias 7.11.4-6). Only a single episode may indicate that the recipient felt some misgivings about keeping hostages while at war with the donor. In 388 B.C. Dionysius I of Syracuse restored Rhegian hostages before the beginning of open hostilities (Diodorus 14.108.1-3). For a discussion of this peculiar story, see below, pp. 145-146.    back
 49. I have not attempted to examine the subject of the legal status of hostages in other ancient cultures. I know of no survey accounts that might deal with this issue in regard to specific cultures    back
 50. This is a very curious edict; it seems to prove that foreign hostages might be honored with Roman citizenship in the imperial period.    back
 51. We know that at least two of the Parthian princes delivered by Phraates IV brought their wives and children (Strabo 16.1.28 C749), but the princes were not formal hostages. The Carthaginians exacted under the Treaty of Zama maintained comfortable establishments (Livy 32.26), as did Demetrius of Syria (Polybius 31.11-15); they were perhaps allowed wives, though they may not have chosen to avail themselves of the right.    back
 52. See below, pp. 263-274.    back
 53. See below, pp. 221 and 249, case no. 54, for further details.    back
 54. As late as 137 B. C., the case of Hostilius Mancinus revealed that the application of ius postliminii had not been fully developed in Roman law. The Imperial law codes, which are our major source for this question, may have established legal interpretations which had not yet gained ascendancy during the Republic and may therefore oversimplify the fluidity of the Republican legal interpretations.    back
 55. Roman soldiers who had surrendered could lose standing, and commanders who had surrendered were also penalized. Watson (The Law of Persons in the Later Roman Republic [Oxford: Clarendon Press, 1967], pp. 237-255) has discussed the restrictions on ius postliminii and the reasons for them.    back
 56. The year of the trial and the law under which he was indicted can be debated at fair length (Last, CAH 9: 159; H. Volkmann, RE 22 (1953): 58-59, s.v. "Popillius," no. 19; Aymard, “Les otages barbares," p. 137); for our purposes here the trial is irrelevant.    back
 57. For the treatment and mistreatment of hostages, see below, pp. 107-132.    back
 58. For the explicit attitudes of ancient cultures toward the institution of hostageship, see above, pp. 11-15.    back
 59. Aegidio Forcellini et al. in Lexicon Totius Latinitatis (Padua, 1940), s.v. obses, provided a useful point of departure for the definition of obses proposed here.    back

This work is licensed under a Creative Commons License.