The Fate of the Population after the Withdrawal

Sources are consistent regarding the evacuation and surrender of Dacia. In Eutropius' representative account, 'the province of Dacia, established by Trajan on the far side of the Danube, was evacuated and abandoned by Aurelianus after the devastation of Illyria and Moesia. The Romans were resettled from the towns and land of Dacia to the middle of Moesia, which he renamed Dacia. It separates the two Moesiae, and after standing on the left of the Danube, it now stands on the right'.[63]63. Eutropius, Breviarium... IX, 15. Is it conceivable that despite Eutropius' testimony, a sizeable number of Latin-speaking citizens remained in the province abandoned by Rome, survived the era of great migrations, and became the ancestors of a neo-Latin people?

This hypothesis would be plausible if historical and demographic conditions at the end of the 3rd century made the resettlement impossible or unverifiable; and if it could be established that after 271, there remained in the original province a large population of Latin-speakers, of citizens who had reached the highest level of Romanization. The size of this hypothetical population is critical, for 130–150 years later, after the collapse of Roman administration and military defence in their respective provinces, the Latin-speaking masses of the middle Danubian region were overwhelmed by the waves of the great migration; this happened sooner in the case of northeastern Pannonia, and later in western and southern {1-123.} Pannonia, Raetia, and parts of Noricum. Those who failed to flee in time eventually blended in with the successive newcomers or were swept away. No neo-Latin people emerged from the maelstrom, and this in spite of the fact that historical circumstances (including a society made homogeneous by four centuries of Romanization) and the natural setting were much more favourable than in the case of a Transylvania encircled by Barbarians. That region enjoyed no peace after the Romans' early withdrawal, for it continued to be harassed by the Carpi, Sarmatians, Vandals, and Gepids.

In considering the possibility that a Latin-speaking people survived in Dacia, one must examine the circumstances that might have facilitated the emergence of neo-Latin peoples in the other, more westerly provinces. It is important to note that Dacia was surrendered in deliberate and organized fashion, at a time when the empire still had solid foundations, and not in the midst a great crisis. The neo-Latin peoples that eventually emerged in Rome's onetime possessions did so 130–150 years later, after the fall of the empire. Even in these lands, the degree to which a Latin culture was preserved varied widely, notably between the middle Danubian provinces, and the western provinces or the coastal strip on the Adriatic. In the western border provinces of the empire, the Barbarians drove away most of the inhabitants in the early years of the 5th century and settled down. The new conquerors became either the ancestors of nationalities that exist to this day (in the case of Germania), or blended over time with the local population to form new nations: such was the case with the Ostrogoths and Lombards in Italy, the Visigoths in Spain, and the Franks in Gaul. There was little mixing at first between followers of the Nicene Creed and the Germans, who were either pagan or attached to Arianism, but once the conquerors converted or took baptism, there remained little hindrance to social integration. Thus these various tribal and ethnic groups evolved with little difficulty into territorially-organized feudal societies.

{1-124.} In the former empire's territories in western and southern Europe, a certain continuity prevailed once the conquerors had settled down. The situation in the Danubian provinces and the Carpathian Basin was radically different. There, the same peoples who later established states in the west or in Africa made only brief stays; they progressively eliminated the remnants of Roman culture, but neither settled nor founded states. Whether of not some Romanized inhabitants stayed behind in Dacia, this destructive process had begun in the province some 130–150 years earlier.

There was another factor that assisted in the survival of provincial populations in the west European and middle Danubian regions. With the collapse of Roman administration, Christianity — both its institutions and its creed — remained as the only source of social cohesion. With its promise of salvation in the afterlife, the Church brought no small comfort to people suffering through the vicissitudes of the 5th–7th centuries. The Church's task was facilitated by an organizational structure that was fully developed by the 4th century and which quickly spread to the provinces. The integration of church and state administration was already under way when the administration of the Roman empire began to lose its grip and its ability to defend the towns, which were the fountainheads of Romanization. The Church stepped into the breach as bishoprics supplanted Rome's urban administration; the universality of the Christian creed compensated for territorial fragmentation. The bishops thus assumed a leading role, organizing defence and negotiating with the conquerors. However, the bishoprics had been established and granted privileges within the empire as it existed in the 4th century, and thus Dacia, which had been surrendered in 271, could not benefit from these defenders of Roman culture.

To recapitulate, many circumstances conspired to impede the survival of Latin culture in Dacia, even in the eventuality that fragments of the provincial society stayed in place. This border province jutted deep into Barbaricum and was connected by only a {1-125.} narrow strip to the Latin-speaking part of the empire. It was exposed to the storm of the great migrations at least 130 years earlier than Rome's other possessions. As the distance increased from the center, less and less of Latin culture survived. Dacia lay on the easternmost fringe of this sphere, or, indeed, beyond it, separated as it was from the Romanized Dalmatian coast by territories where Greek, then Slavic language predominated. And it had no organized Church to rally and defend its people.

Rome effected an organized withdrawal from the province at a time when she had brought the Goths' long advance to a halt in the northern Balkans; there a safe haven awaited Dacia's people. The curtailment of anarchy at the end of the 3rd century and greater domestic stability may not have impressed the peoples of the far-flung empire, but the end of Goth attacks on Moesia was a palpable success and indicated the possibility of effective defence. Battles south of the Danube had decimated the population, creating ample room for large-scale resettlement. Eutropius recorded that Dacia was evacuated not only because it was no longer possible to defend it, but also because Illyria and Moesia had been devastated. At least in Illyria, war was not the sole cause of depopulation in the 250s; Zosimus[64]64. Zosimus, I, 37. records that 'a terrible epidemic of pestilence broke out in the town, such as never before witnessed: it surpassed the devastation wrought by the Barbarians, to the point that towns occupied and sacked by them felt fortunate to have escaped the fate of those infected by the plague.' The population losses were so great that in neighbouring Thrace, the Romans were actively promoting new settlement as late as the 4th century. Thus the relocation of Dacia's reduced population was not only feasible but also coincided with the need to repopulate the Balkans. Rome must have been pleased that for once it could resettle its own citizens, and not Barbarians.

To test the credibility of Eutropius, one can also ask whether Dacia's population welcomed the prospect of resettlement. To be {1-126.} sure, they were in no position to block execution of the policy. Soldiers and their families, who made up a significant part of the population, had no choice; the former left with their units, and the latter followed. The empire, suffering from depopulation, was hungry for people, but the evacuation of Dacia was not accomplished by force. The province's inhabitants, and particularly its Latin-speaking Roman citizens, had no interest in remaining in an undefended territory, exposed to the depredations of the Barbarians and doomed to sink to the level of their conquerors. At the beginning of the 5th century, when the empire was already disintegrating, people in the other provinces spontaneously fled southward, although they had little expectation of finding security and new homes. For Dacia's Roman citizens, who spoke Latin, Greek, and perhaps Syrian, staying put was hardly an attractive alternative when they could move a short distance across the Danube to a secure and civilized new home. The sources indicate a pattern of behaviour that is the very opposite. The free Dacians asked to settle in the empire on two occasions at the end of the 2nd century, and following the Marcomanni wars, two groups, one of unknown number, the other 12,000-strong, received the necessary permission. Although they could not have been Romanized in the mere 50–60 years that had elapsed since their settlement in the empire, now, with the Goths nearby, they were not likely to insist on remaining in the abandoned province. Ever since the 2nd century, the peoples on the periphery had regarded the empire as a prosperous and well-guarded haven, and one tribe after the other begged for permission to settle there. Nor was the empire's attractiveness diminished by wars and anarchy in the middle of the 3rd century; the Marcomanni were resettled under Gallienus, and so were the Dacianized Carpi in 295, despite the fact that they had earlier raided Dacia.[65]65. Aurelius Victor, Epitomae de Caesribus 39, 43.

The narrative sources give a uniform account of Dacia's surrender and of its population's resettlement. The evacuation and resettlement of Roman citizens was well within the capacities of {1-127.} Rome's efficient administration. Although many civilians fled earlier, the evacuation could not have been accomplished overnight, and while there is no evidence to show this, it is just possible that some of the inhabitants stayed behind. For all the reasons noted, their numbers could not have been significant.

From the 250s onward, epigraphs raised by citizens grow scarce, and remain so for a long time. There is, however, one indication that resettlement could be traced through epigraphic finds. The new Dacian capital, Serdica, was in a previously Greek-language region, yet a large proportion of the epigraphs found there, and dated from the 4th century, are in Latin. A plausible explanation is the influence of Latin-speaking people who had been resettled from Dacia.

Cemeteries are a good guide to the history of a region, to change and continuity in its settlements. When a settlement's life was interrupted, whether for reasons of extermination, emigration, or flight, there would be no more burials. The cessation of burials — signifying the extinction of a settlement — can be dated with some precision in the Roman era on the basis of grave goods, such as apparel, household objects, and coins. If, despite the written evidence, provincials stayed behind in Dacia, then finds in the cemeteries of Roman towns, rural settlements, and military camps should indicate burials beyond the 270s. Setting aside individual graves and small portions of cemeteries as being unsuited to such analysis, one finds only one wholly excavated town cemetery, at Romula; and there, burials evidently ceased in the second third of the 3rd century. The cemeteries at Napoca, Apulum, and Potaissa, and those at military camps have yet to be excavated.

Roman coins, minted after 271, are commonly cited as evidence of the provincial population's continued presence, as are finds of clear or supposed Christian character. However, none of these finds demonstrate the validity of the hypothesis.

{1-128.} The finds in Dacia of Roman coins identified with the second half of the 3rd century are exceedingly few; the incidence of coins rises only in the 4th century. This type of find offers no proof of the survival of the earlier population, either in Transylvania or in the Banat. Such money circulated outside the empire, in Barbaricum, in amounts that varied according to historical circumstance. Roman coins circulated in the Sarmatian-settled areas of the Great Hungarian Plain and the Banat, and finds of such coins do not prove that the grandsons of onetime Dacian provincials engaged in commerce, nor that the Banat was part of the Dacian province and therefore of the empire. The traffic of coins in Barbaricum reflects a different historical process. The coin finds in the former province of Dacia and in the Banat are distributed unevenly, the dividing line being an earthwork, raised in the 4th century, that ran north-south across the Great Hungarian Plain. The explanation is found in the history of the Sarmatians. The coins found in those parts of the Banat and Transylvania west of the earthwork are evenly distributed; their traffic, as on the Great Hungarian Plain, reached a peak in the 360s and 370s. The traffic of coins in the rest of Transylvania presents a different picture, though it is similar to that in other regions beyond the borders of the empire, for example in Slovakia: it reached a peak in the middle of the 4th century, then another during the reign of Valentinian I. This is may be considered the normal pattern in the Carpathian Basin during the 4th century. The coins found in Transylvania prove only one thing: that the region's population, depleted by the Roman withdrawal, eventually started to grown again, if only at a modest rate.

Most of the fifteen Christian finds are of uncertain provenance or are incorrectly identified; only a biased analysis can attribute them unequivocally to surviving provincials. To be sure, after the empire collapsed, Christianity contributed greatly to the cultural survival of Romans in the western and southern parts of Rome's former domain, where the decisive factor was not religion itself but {1-129.} the Church's institutions, the bishoprics that were constituted in the 4th century. No such bishoprics were found based at Dacia's towns in the 3rd century. The finds of a Christian character therefore do not serve to prove the existence of Christian communities or bishoprics; they indicate, at best, the presence of individual Christian believers, who may well have been Christian Goths. These finds could have reached Transylvania thanks to trade or plunder by the Goths (Christian or not) who resided there in the 4th century; or, indeed, thanks to later collectors of Roman objects. The fortified posts found along the left bank of the Danube in the 4th century facilitated the transfer of such relics. Apart from the Berethalom finds, which can be plausibly attributed to the Goths, these objects are not linked to liturgical ceremony. In the final analysis, there is no proof that their original owners were Christians. The same conclusion can be drawn in the case of the finds at Tápiógyörgye (Pest county, Hungary) and Luciu (Ialomiţa county, Romania), and of the early Byzantine bronze lucernae found in the Barbaricum. Similarly, it cannot be assumed that the bronze hand of Jupiter Dolichenus, plundered in Dacia and subsequently unearthed in Ukraine, or the bronze statue of the goddess Victoria found at Akasztó (Bács-Kiskun county, Hungary) once belonged to Roman citizens or devotees of Roman cults.

There is, apart from the ancient chronicles and archaeological finds, another source that can throw some light on changes in ethnic patterns, and that is toponymy. The names of settlements and rivers harbour clues to the geographical and ethnic distribution of the population. Such names reflect local culture, and their modification or alteration can be analyzed to reveal changes in the nature of a population. Toponymic changes generally do not coincide with ethnic change, for only rarely does an ethnic group disappear in its entirety. Where newcomers settle alongside the remnants of older-established ethnic groups, they will commonly preserve some of the earlier toponyms. Thus toponyms can be traced even where the {1-130.} people who gave rise to them have become assimilated or have disappeared. On the other hand, if the population of a region remains in place after the arrival of new settlers, at least some of the earlier toponyms will survive.

Toponymic change indicates the arrival and settlement of new inhabitants. The scale of ethnic change will influence the degree to which toponyms are modified, as will differences in the culture and social organization of the ethnic groups. In larger territories, there seldom occurs a wholesale change in geographical names. Parts of the banks of longer rivers have been inhabited without interruption, and thus the names of the waterways — such as the Elbe, Rhine, Danube, Tisza, Drava, Sava, and Maros — are preserved through the centuries. The scrutiny of toponyms is of particular utility when one examines the period following the breakup of the Roman empire, which suffers from a dearth of alternative sources. In regions where there remained a sizeable Roman population — that is, in what became the neo-Latin countries — many Roman place- and river-names survive to this day, subject of course to linguistic adaptation. In other regions, where the Roman population was depleted and joined by new settlers, there was greater change in toponyms. The change was even greater in places where the earlier population largely or totally disappeared. Prolonged cohabitation served to preserve some old toponyms, while recurrent ethnic transformation brought with it more rapid toponymic modification and replacement.

In the 5th century, after the breakup of the empire, the Roman population of Raetia, Noricum, and Pannonia soon changed in number and character due to emigration and assimilation, and thus relatively few of the earlier toponyms were preserved; yet in some regions where the provincials disappeared early, a few Roman toponyms survive to this day. In northeastern Pannonia, where there were several wholesale changes in population, no Roman place-names survive. On the other hand, between the Drava and the Sava {1-131.} and in western Pannonia, numerous river names (e.g. Rába < Arrabo, Marcal < Mursella, Zala < Sala, Zöbernbach < Sevira < Savaria, Mura < Mura) as well as place-names (Vienna < Vindobona, Ptuj/Pettau < Poetovio, Sisak/Sziszek < Siscia) bear a Roman imprint; the Szerémség (the triangle between the Drava and the Sava rivers, northwest of Belgrade) carried over the Roman toponym Sirmium, and the town of Mitrovica bears the name of that settlement's Roman defender, Demetrius. The Roman name of Sopron (Hungary), Scarbantia, can be found in documents until at least the end of the 6th century. The name of Savaria is documented in 860 and in the Middle Ages, and as late as the 19th century it was in official use along with the Hungarian name of the locality, Szombathely. These toponyms survive in regions of Pannonia where there is no neo-Latin population, and where most of the Roman population had disappeared by the time of the Avars. Yet an early period of cohabitation and assimilation had facilitated the transmission and preservation of some place-names, and when the Hungarians arrived, they, too, adopted these names. It follows that if Roman names had survived in that other Roman province of the Carpathian basin, Transylvania, the Hungarians would have preserved them as well. Thus, if part or all of a Romanian ethnic group that spoke neo-Latin had lived in Transylvania, it could be expected that there — as in other neo-Latin countries — many Roman toponyms would have survived.

To be sure, the names of certain major rivers betray a Roman origin. In the case of the Szamos (only the place-name of Samum is known from Roman times) and the Maros (Marisus, Marisia), only a stretch of the rivers lies within Transylvania, and thus their names were not necessarily given by Dacia's inhabitants. Two northern tributaries of the Danube, the Olt (Alutus) and the Cserna (uncertain, probably from Dierna) preserve their Roman appellation; this probably owes to the fact that after the Romans evacuated Dacia, they and the Eastern Roman empire for several centuries maintained {1-132.} a foothold on the left bank of the Danube. Beyond this, and apart from the debatable origin of a few river names in central Dacia, no Roman toponyms survive in Transylvania. In no other Roman province did Roman toponyms suffer a similar fate, but their disappearance is consistent with the historical record concerning the evacuation of the province. (The Roman place names found in the Dacian region of today's Romania were attributed in modern times, in evocation of the Roman period.)

Apart from one exception, the isolated, 4th century cemetery at Baráthely, there is no evidence that any of the provincial population might have stayed behind after Aurelianus evacuated Dacia, and thus no reason to question the written records of the period. The extent of toponymic change confirms this wholesale withdrawal. Transylvania, Oltenia, and the Banat were touched by the Great Migrations some 130–150 years earlier than was Pannonia, and 150–200 years before the first settlements of Franks and Goths in the neo-Latin territories. The chances of survival for the former provincials during this long period of turbulence were slight. Relinquished in 271 by the empire, Dacia would remain excluded from the Romanized world.