We have seen that hostageship was an integral part of Roman foreign policy, an institution rooted deeply in both the legendary past (Dionysius of Halicarnassus 1.43.1; 1.59.2) and the contemporary accounts of Rome’s incredibly swift expansion during the second and first centuries B.C. (e.g., Polybius and Caesar). Only one hypothesis could explain the long duration of hostageship over so many centuries and involving such diverse peoples: it worked. No mere custom, however well entrenched, is likely to have survived solely because of the sentiment engendered by longstanding tradition. This is not to say that hostageship never failed, but rather that the potential gains of hostage exaction outweighed both the potential disadvantages and the effort which the recipient government expended upon the process.

So much we can argue from common sense. Where is the proof? In the majority of cases where our sources refer to the exaction of hostages, we are not explicitly told whether concern for the hostages restrained the donor from violating the agreement, and it is such explicit testimony upon which our evaluation of efficacy must rest. For many of these cases, we know that the agreement was kept, but we do not know precisely why. Philip V of Macedon remained a loyal, though perhaps not an enthusiastic, ally of Rome during the Greek campaign of Antiochus III of Syria, but we cannot prove that the {205|206} detention of his son Demetrius was solely—or even partially—responsible for Philip’s fidelity. Similarly, Carthage’s patience in the face of Masinissa’s depredations while Punic hostages resided in Italy may have been due purely to the exaction of hostages, or to her military weakness, or to her reluctance to engage in another African war. [1] Although it is probably correct to suspect in such instances that hostage exaction played an important role in forcing donor governments to observe their agreements, let us instead consider evidence which cannot be readily disputed.

First, the failures. There are some thirty-five revolts attested after the delivery of hostages, out of a total of two hundred eighteen historical exactions, or 16.1%. [2] For Caesar’s campaigns in Gaul, the greater wealth of detailed information allows us to compute a higher percentage of failures, about 27.1%. Even if this higher rate of failure is accepted as a more accurate reflection of the efficacy of hostageship, the reverse of the failure rate is a surprising 70%—the rate of success of hostageship in preventing revolt.

Yet the explicitly acknowledged successes of hostagcship, though fewer than the obvious failures, are perhaps more revealing. The attempts to rescue hostages before beginning overt resistance (Diodorus 12.27.1-2; Plutarch Pericles 25.1-3, Thucydides 1.115.3, 5; Diodorus 19.75.1-2; Josephus Jewish Antiquities 14.290) must assuredly demonstrate the efficacy of the restraint which exaction entailed. [3] Ancient authorities expressly attributed to hostageship no fewer than seven cases of good faith in the face of danger to hostages: in 396 B.C., the Campanians around {206|207} Aetna declined Himilco’s offer of aid in revolting because of their hostages in Syracuse (Diodorus 14.61.4-6); the Iberians in 217 B.C. and 209 B.C. did not rise against the Carthaginians until their hostages were no longer in Punic custody (Livy 22.22,5; 26.43.4; 27.17.1-3; Polybius 3.99.3; 10.34; etc.); [4] the execution of their hostages provoked the defection of Tarentum and Thurii in 212 B.C., (Livy 25.8.1-2; 15,7); Cotys of Thrace tried to explain away his support of Perseus against Rome on the grounds that his son Bithys was a hostage at Perseus’ court (Livy 45.42.7); threatened with the torture and/or death of their kin, the Aedui in 58 B.C. employed a spokesman not bound by hostages to present their complaints against the Germans and Sequani to Caesar (Caesar Gallic War 1.31.12,15), and in 52 B.C. those Gauls whose hostages had fallen into the power of the Aedui joined Vercingetorix’s party (Caesar Gallic War 7.63.3). [5] This explicit testimony, coupled with the reluctance with which states submitted hostages, proves definitively the constraint which hostageship could wield. [6]

Thus it is clear that, with regard to its primary purpose of compelling fulfillment of a contract, hostageship was an effective weapon. Was there another purpose? We have observed in the previous discussion that Romanization was a possible factor in selection criteria, treatment, and termination of hostage status. Although this question of the Romanization of hostages has frequently risen among modern scholars, it cannot be absolutely answered; no ancient source categorically states such a motive. Opinion is divided on the important issue of whether Romanization was in fact a conscious and deliberate {207|208} policy and on the subsidiary matter of when this policy, if such there was, came into being. [7]

There is, however, some circumstantial evidence. Roman practice differed markedly from Greek in the selection of males, although female hostages appear to have worked as well. But only men managed the business of government. The selection of youths of impressionable age seems a regular feature of the treaties with Carthage in 201 B.C., with Philip V of Macedon in 196 B.C., with the Aetolian Treaty in 189 B.C., and with Antiochus III of Syria in 188 B.C. The Romans maintained a procedure of substitution (mutatio obsidum), which meant that more hostages were exposed to Roman culture; whether it is better policy to expose many for a short time or a few for a long time is debatable. [8] Finally, hostages of Rome were as a rule very well treated.

These facts, together with the obvious effect of Roman flattery and ideas found in the subsequent careers of Demetrius of Macedon and Antiochus IV of Syria, suggest that the Roman Senate was fully cognizant of the potential value of Romanization. [9] Certainly Sertorius’ use of a Roman school at Osca becomes more intelligible in view of these precedents, as does the failure of Vonones and most of the other Parthian and Armenian princes to remain secure on their thrones; their long residence at Rome and their unfamiliarity with their own culture caused them to become estranged from their subjects, to become incapable of governing an alien people. Romanization had progressed too successfully.

As hostages assimilated Roman customs and Roman ideas on {208|209] politics, they must also have acquired an understanding of the role which the Romans expected them to play in future dealings. Just as dediticii in general became the clients of the commanding field officer, so must have hostages, often from these same surrendered peoples, become dependents of the man who had directed their exaction. Indeed, when the hostages were taken to Rome or its environs, there was undoubtedly a deepening of this relationship, because of the greater contact which the patron and client had. Moreover, the additional social contact with the ruling elite of Rome might well have brought about the development of similar bonds of obligation to other distinguished Romans, although probably to a lesser extent.

In conclusion, I submit that during the Republican period the Romans employed hostage exaction first and chiefly as a means of insuring that treaty obligations be met and secondly as a method of cultivating politically valuable allies among the hostages themselves. Except for the effort and expense incurred by the selection and maintenance of the hostages, there was little that Rome could lose by such a process, and there was a great deal to be won.