Barbarians and Romans, A.D. 418-584

ITALY is the model case of barbarian settlement. Only there can we know more or less precisely who conducted the allotments, what resources were allocated to this purpose, how the allocation was administered, how much each Goth was given, and even what kind of oppression a Goth could exercise upon the Roman taxpayers assigned to him. The Italian example offers modern imaginations a firm guide to the course that a fifth-century settlement in Roman territory might take; and more, it documents an instance where allotments could be made without expropriation of private property because tax assessments were awarded instead.

It is tempting to assert flatly that the procedures used in Italy had been adopted at an earlier date by the Roman authorities when they established the Visigoths in Aquitaine and the Burgundians in Sapaudia. Such a view would have the advantage of eliminating the various difficulties that the currently accepted interpretation of the settlements in Gaul presents. The hard part is to tailor the evidence of the Visigothic and Burgundian laws to the Italian pattern. It is not as though these sources are more favorable to the traditional interpretation; for example, the “one-third” of Roman hospitalitas bears little relation to the “two-thirds” of the land and “one-half” of the woods, house, and gardens stipulated in some texts from Gaul. The fundamental problem is that the Visigothic and Burgundian laws speak at a great distance in time from the original allotments and that they say very little. There is, in fact, only one substantial text for Gaul, namely, title fifty-four of the Burgundian code, itself bristling with difficulties. All the rest are odds and ends, of little independent value. The kind of conclusions attainable with the help of Cassiodorus’s Variae are beyond reach in Gaul. The best that can be hoped for is an approximation, compatible both with the evidence that there is and with the lessons of Ostrogothic Italy.

A closer look at the sources for Gaul illustrates their shortcomings. The chroniclers tell us no more about the original Visigothic settlement in 418 than that it was initiated by the patrician Constantius, who assigned to King Wallia a territory for his people to inhabit.¹ The church historian Philostorgius adds details drawn—we cannot tell how accurately—from Olympiodorus: the Visigoths were granted both rations and land for tillage.² There are also the personal reminiscences of Paulinus of Pella, largely confined, however, as regards this topic, to the period when the Visigoths were temporarily in Bordeaux (413-414).³ In short, we are denied the slightest glimpse at the distribution of allotments and cannot even tell precisely when it took place. The first text documenting the existence of Visigothic sortes occurs in the Code of Euric (hereafter CE), whose date is about 475, and, at any rate, in Euric’s reign, 466484/5.⁴ Two extracts bear on the subject, and the more circumstantial of them lacks the opening lines that are indispensable to its interpretation. For further information, we must turn to the mid-seventh-century collection of Reccesvinth (hereafter LVisig). Under the rubric Antiqua, it reproduces accurate extracts from the second Visigothic codification, the Codex revisus of Leovigild, drawn up in Spain between 568 and 586. In keeping with the role of a reviser, Leovigild modernized whatever laws of Euric he incorporated into his code.⁵ Four or five Antiquae contain unmistakable references to the allotments, but they can hardly be taken to document fifth-century conditions.⁶ With the exception of CE 276 and 277, therefore, the Visigothic information available to us is more indicative of how the old settlement looked, or was understood, at the close of the sixth century in Spain, than of how it had been in Gaul under Euric, let alone at its inception some fifty years earlier.

The information that the chronicles supply about the Burgundians is no more detailed than that about the Visigoths. It is noted under 443 that the “remnants” of the Burgundians—after the destruction in 436 of their kingdom in northern Gaul—were given Sapaudia “to divide with the natives”; the patrician Aëtius promoted their relocation. Then, in 457, the Burgundians enlarged their territory: with the assent and support of the Visigoths, King Gundioc “entered Gaul with his people and all their goods in order to reside.”⁷ There are also later accounts of these events. At the close of the sixth century, Marius of Avenches portrayed Gundioc’s move as involving the Burgundians’ “dividing the lands with the Gallic senators.” Finally, in the mid-seventh century, Fredegar tells us that, much earlier, under Valentinian I (363-375), the Gallo-Romans invited the Burgundians to immigrate so that their sponsors might be able to stop paying taxes to the imperial government.⁸ Although Marius and Fredegar valuably illustrate how the epoch of each one wished to remember the original settlements, they are without standing as witnesses to fifth-century conditions.⁹ The contemporary chronicles, for their part, establish that there was settlement and “division” in 443 and 457. “Division” might signify a fifty-fifty partition (a ratio also suggested by the laws, as will be seen), but of what? On the precise terms of allotment and the date when sortes were distributed, the chronicles are silent.

The Burgundian legal sources, however, markedly surpass the Visigothic ones in length and intelligibility. Inspired by Euric’s Visigothic code, King Gundobad (ca. 480516) caused the Burgundian laws to be collected, and his code has been more fully preserved than Euric’s. Its condition, however, is by no means ideal: Gundobad’s collection survives only in the enlarged edition issued by Sigismund (517-523) and in ninth-century manuscripts at that. In Sigismund’s edition, the Burgundian code is formed of an original core (LBurg 1-36), attributed to Gundobad’s initiative, that includes several articles pertinent to the status of barbarian sortes. The core was then augmented by a series of Gundobad’s new laws, of which some are dated, but none, apparently, is earlier than 501; these novellae are of considerably greater interest to allotments than are the original series. Moreover, one Burgundian law transmitted outside the code (lex extravagans), and dated 524, contains a clause relevant to our subject.¹⁰ The information on allotments that the Burgundian laws supply is thus comparatively abundant but difficult to date and, in any case, it is far removed from the events of 443 and 457. The crucial text, LBurg 54, frequently and extensively commented upon by modern historians, is an added law of Gundobad, almost certainly issued later than 501. In sum, a few Burgundian laws permit inferences to be made about the original conditions on which allotments were made, but, since such inferences are possible only after lengthy textual in-terpretation, their validity is much less certain than if the sense of the laws were self-evident.¹¹

Early or late, none of the texts just described provides a narrative account of the Visigothic or Burgundian settle-ments, not even one comparable to Procopius’s tale of Odoacer’s land grants. The versions in our manuals are modern reconstructions, based chiefly on three laws. Of these, CE 277 juxtaposes Gothic sortes to the “third of the Romans”; CE 276 mentions some Goths entering the “place of the hosts (locus hospitum)” in order to supervise a boundary delimitation; and LBurg 54 records, among other things, a recent ordinance that allowed a Burgundian to claim “one-third of the bondsmen [and] two parts of the lands from that place in which hospitality had been assigned to him.”¹²

Two features of these extracts—the terminology of hospitalitas and the explicit or implicit fraction two-thirds— occupy center stage in modern syntheses. A typical version is as follows (though limited, in the event, to the Visigoths, it might easily be extended to the Burgundians):

As allies of Rome, the Goths based themselves primarily on the Roman regime of military quartering (hospitalitas). According to it, the military hospes received a tertia (tertia hospiti deputata) of the country estate in which he took up residence ... At some time after the Gothic occupation of Aquitania II and Novempopulonia ... a new system of sharing, more favorable to the Goths, was put into practice; by virtue of it, only one-third of the cultivated land went to the Roman proprietor, and the two other parts to the Gothic partner (consors).¹³

An admirable effort has been made here to expand scanty data; yet there are serious problems. We are asked to believe that the barbarian allotments entailed expropriation on a large scale of Roman property, without its being shown, however modestly, that any Romans were in fact dispossessed. Moreover, the legal basis for these expropriations is derived from Roman hospitalitas, even though the military quartering of the Theodosian Code had nothing to do with agricultural assets or the income from land. These flaws are decisive. Our chances of reconstructing precisely how the settlements in Gaul were organized may be remote, but we may be close to certain that they happened in some other way than is currently supposed.

At a distance of a half century or more from 418 and 443, the Visigothic and Burgundian laws had no reason to feature the Roman refinements that entered into the original distribution of allotments. A barbarian sors, regardless of how it was initially constituted, was property from the standpoint of its recipient, and landed property at that; it was bound, as time went on, to look increasingly like an ordinary private estate. It is only to be expected, therefore, that the two codes should seem to be dealing with mere property when they regulate allotments. Be that as it may, we have now found that, in Italy, tax assessments were the assets distributed to the Ostrogoths; we have examined two hagiographies in which Visigoths behave like their Italian counterparts, and probably for the same reason; and we also know that, in Gaul, a group of Alans was assigned a category of assessments to share out among themselves. It remains to be seen whether the Visigothic and Burgundian laws, though ostensibly alien to such schemes, contain passages and provisions suggesting that the original awards consisted of a state-derived “superiority” over land rather than of (expropriated) private property. Only then will it be possible to consider what relationship the rules of Roman billeting may have had to the installation of the barbarians.

The tax machinery of Roman Gaul was in working order when the Visigoths and Burgundians were settled; Salvian bears witness to its ravages, and even Sidonius felt its impact sufficiently to ask for a little relief in 459.²⁰ What became of this lucrative resource when whole districts were made available to barbarians “for residence”? Little is known about the forms of government in the earliest barbarian kingdoms of Gaul. Nevertheless, it is unlikely that the apparatus of Roman taxation was simply dismantled. What evidence there is suggests that, however altered taxation may have been, it certainly was not suppressed. As far as the Roman tax officials were concerned, the hypothesis that they became answerable to the barbarian kings, Visigothic or Burgundian, is preferable to the alternative that they continued to render account, even for a while, to the prefecture at Arles (that justice over the Roman population followed the latter course for a time is at least conceivable).²¹ In Italy, the usurpation of Odoacer involved a reorganization of public finance; in effect, the proceeds of tax assessments were divided between the new king and his followers: the former retained two-thirds, and the latter were allocated the balance, as the fund from which their sortes were awarded.²² If something of the same kind took place in Gaul with the advent of the Visigoths and Burgundians, it would partly explain the fate of the Roman tax machinery under new management.

Although expressed in terms of Romans and barbarians, rather than king and followers, a division is basic to our knowledge of the settlements in Gaul. The chronicles mention it, and, by the terms of CE 277 and LBurg 54 (leaving details aside), the shares appear to be two-thirds for the barbarians and one-third for the Romans. The phraseology of the texts is vague or ambiguous in regard to the main object of division. Modern commentators have no choice but to be more specific than their sources. The usual surmise has been that the division applied to private property—the estates of the Roman indigenae.²³ These estates, however, had more than one dimension. Where private property in land, expressed by the word possessio, coexisted with taxes assessed on landowners’ declarations of their resources, called professiones, the ownership of land and the rents arising from it had to be distinct from the property assessments that brought annual returns to the state. Since taxes assessed in this way were levied in fifth-century Gaul, the productive lands of the districts occupied by the Visigoths and Burgundians must have yielded two forms of revenue: the private rents of landlords and the taxes collected by the state. Both stemmed from “the land” and, for all one knows, they were much the same in composition. Yet they were separate entities.

As soon as this distinction is recognized, the award of barbarian allotments is bound to look like a more complex process than the mere partition of ownership between Romans and newcomers. One is forced to confront and answer the question whether the division embraced only the private category of revenue (rents), or the public category (taxes), or both. One apparent certainty is that neither Visigoths nor Burgundians paid taxes to their kings; another is that Roman landlords kept at least a fraction of their property, together with whatever tax liability it used to bear.²⁴ Let us suppose, then, for the sake of argument, that all taxable Roman property in the district of settlement was subject to division, and that the distribution was made on such a basis that the tax-exempt barbarians acquired ownership of two-thirds of these properties, whereas the balance was left to the Romans, whose taxes now went to the king. On this hypothesis, the troops would seem to have been decidedly overadvantaged: as recipients of allotments consisting of expropriated estates, they would have pocketed both the private and the public revenues of their double share, whereas the barbarian king was limited to collecting only the tax revenues of his portion—the properties left in Ro-man hands. Once the fact that Roman landowners regularly paid taxes is injected into the story, it becomes apparent, first, that a division between king and followers was necessarily implied by any sharing out of “lands”; and, second, that an interpretation of the texts regarding Visigothic and Burgundian divisions cannot disregard the tax value of property. The understandable forgetfulness or vagueness of the Visigothic and Burgundian codes concerning such bygone conditions has to be compensated for by our efforts.

The texts themselves, to which we now turn, present arresting oddities. Let us begin by examining CE 277:

[1] Gothic allotments (sortes) and the third (tertia) of the Romans that have not been recovered within fifty years are on no account to be reclaimed. [2] The same holds for fugitives: it is not allowed to recall to service those who were not found within fifty years. [3] We order, indeed, that ancient boundaries shall be firm, just as our father of good memory ordered in another law. [4] And all other cases—regardless of whether good or bad . . . and regardless of whether they are slaves placed in dispute or debts that have not been collected—shall on no account be pursued if they have not been terminated within thirty years. [5] If anyone, after this number of thirty years, attempts to plead a case, let that very number oppose him, and let him be forced to pay a pound of gold to that person whom the king will order.²⁵

The first sentence is of most immediate interest to us but deserves to be seen in context. In LVisig 10. 1. 8—from Leovigild’s Codex revisus—one finds a heading “about the division of lands made between a Goth and a Roman” followed by a provision on dividing the “two parts of the Goth” from the “third of the Roman”; hence, an easy gloss for CE 277. 1: that the parts consisted of land.²⁶ But if this assistance from the late sixth century is dispensed with, the meaning of sortes and tertia is anything but obvious. The main problem is this: according to the standard interpretation (that private property was the object of division), sortes plus tertia (that is, the two-thirds allotted to the Goths plus the one-third left to the Romans) had to have embraced the totality of landed estates in the Visigothic kingdom; if so, then the fifty-year prescription edicted by CE 277. 1 necessarily applied to the totality of lawsuits over the recovery of property. But a prescription of this length— ostentatiously exceeding the, by then, consecrated thirty years—is altogether extravagant.²⁷ Why would Euric have applied it to something so commonplace as the recovery of all real estate? It is true that the same is done with regard to fugitives. But here, too, a prescription of such inordinate length, safeguarding an owner’s title for almost two generations as then counted, calls for explanation. As it happens, fugitives subject to servitium were a complex category. They included not only private law slaves (comparable to private law property), but also bound tenants— coloni—whose servitium was a matter of fiscal or public law, rather than of private ownership; even slaves were partly in the public sphere, provided they were registered in tax declarations.²⁸ We would be mistaken to think that the prescription laid down in CE 277. 2 was meant to defend only a private right.

These observations suggest a way out of the difficulties of CE 277. 1. A prescription of fifty years reeks of special privilege; it reminds one of the hundred-year prescription that Justinian—briefly—lavished upon ecclesiastical property.²⁹ That Euric casually granted a fifty-year prescription to all owners of lands and bondsmen is too unnecessarily generous to be believable. On the other hand, by instituting a time limit of this length with regard to something fiscal, he may have met a genuine and urgent need to safeguard the title of those persons having enjoyment of originally governmental rights (or revenues) or, alternatively, to defend the public from abuses on the part of such persons. Euric’s intent, as the phraseology suggests, may indeed have been to offer protection to those against whom claims were being pressed, rather than to assist claimants. And, if this is so, then the sortes and tertia of CE 277. 1, without ceasing to be property or land, would be essentially a fiscal entity, as fiscal as the servitium of the assessed slaves and tenants of CE 277. 2.³⁰ The terms would refer, not to possessiones, but to something whose proceeds were comparably lucrative, namely, the professiones of lands subject to taxation.

The argument for identifying the original Visigothic sortes and Roman tertia with tax assessments states a possibility rather than a firm conclusion; yet it does not lack at least partial confirmation. Another oddity encountered in the Visigothic texts is that two laws from Leovigild’s Codex revisus portray the Roman “third” in the guise of direct royal domain:

[LVisig 10. 1. 8] [After formal partition has taken place, a Goth must usurp nothing] from the tertia of the Roman, except that which perchance has been given him by our [royal] generosity.³¹

[LVisig 10. 1. 16] Let the officials of each city, the reeves and provosts, take away the tertiae of the Romans from those who hold them by occupation and return them by their efforts, without delay, to the Romans, so that nothing might be lost to the fisc—provided [however, that they are not excluded by the fifty-year prescription].³²

To repeat a point made in regard to CE 277. 1: if tertia(e) signifies the private lands left in Roman hands after expropriation of two-thirds in favor of the Goths, then the term, at least by Euric’s time and thereafter, had to mean the totality of Roman estates. It would appear, therefore, that the Visigothic king of the 570s could give away any Roman property he wished to a fellow Goth. Furthermore, there would have been a standing order to royal officials to safeguard all these Roman properties from encroachment, presumably on the part of non-Romans; the owners could rely on the state, rather than on their own efforts, to maintain them in secure possession. A recent commentator has suggested that the royal prerogative expressed in LVisig 10. 1. 8 was derived from an overriding right of the Roman emperor to provincial land, but this will not do at all.³³ The existence of such a prerogative in LVisig 10. 1. 8 is the best reason for deciding that the Roman tertia in the Visigothic kingdom was not ordinary private law property but a more restricted and strictly regulated category of ownership: the king could give it away because it was his. The same reasoning applies to LVisig 10. 1. 16: that which public officials can be expected to safeguard is royal domain, and not the farms of any Roman subject. These laws should not be understood to mean that the "third” had always been royal property; they provide a hint about what became of the "third” in the sixth century, rather than information about what it was in the fifth. Nevertheless, they confirm the fiscal dimension of the original tertia Romanorum that was detected in CE 277. The tertia, though rightly associated with Romans, was the king’s share of a division between himself and the other Visigoths.

Although these analyses of Visigothic texts fall short of establishing certainties, they permit at least a hypothetical account, along new lines, of what occurred in 418. The "system” or "principle” or "regime” of Roman military hospitalitas had little relevance to the installation of Wallia and his people in southwestern Gaul. The officials who presided over the arrangements of 418 may have availed themselves of the ancient munus hospitalitatis as the legal pretext for requisitioning houses from the population in order to shelter the Goths, but this was a subsidiary (and undocumented) matter. The key economic provision of the settlement—the one that assured the Goths of sustenance —consisted in the surrender to them of all the public revenues in the district. These resources of the Roman state, rather than private Roman property, composed the total, or basic, barbarian "land grant.” From this grant, by appropriate adjustments over the years, there would devolve the king’s domain as well as the Gothic sortes.

The initial award of all Roman taxpayers’ assessments and payments was subjected to an immediate division between the Visigothic king and his people. We do not know at whose behest, Roman or Visigothic, or by virtue of what principle this largely theoretical apportionment took place. According to our rather late information, it called for one-third to fall to the king and two-thirds to the troops. If the tax machinery was concentrated in royal hands and kept in working order, the troops’ share could have been dispensed to them as salaries, in whatever portions were set down in the muster rolls. But the occurrence of the term “allotment,” and the proprietary orientation of Euric’s Code, imply that, sooner or later after 418, the soldiers’ two-thirds of tax proceeds were converted to individual awards of tax assessment whose yield each beneficiary collected at its source.³⁴ If hospitalitas played a part in this process of individual apportionment, it must have been in a subtler way than is suggested by the terms of Arcadius’s law of 398. Neither a house nor a set fraction had anything to do with the matter. Instead, a Goth received as his allotment the full tax payments of one or more Roman tax-payers, together with the detailed inventory of assessed property (professio) on whose basis the tax was paid. The useful function that hospitalitas might serve in these arrangements was to establish a legally definable relationship between the two parties who had been assigned to one another, the Roman (singular or plural) as “giver,” the Goth as “receiver”; the relationship between them was not of patron to client or of victim to thief, but of host to guest.³⁵ Thus, quarters in a Roman house became coupled with the receipt of revenues from a definite quota of land, probably of uniform quantity per Visigoth, and not necessarily belonging to the same host as the one providing shelter.³⁶

Additional aspects of these, admittedly hypothetical, events are uncovered by considering the situation from the perspective of the Gallo-Romans. If all that the Visigoths obtained (aside from shelter) was the tax value of the district, there obviously was no need to overturn proprietary relationships by expropriation: ownership was undisturbed. But the division of tax proceeds, if combined with allotments, implied that a certain group of Romans paid taxes hereafter to the Visigothic king, whereas all other Roman landowners were mediatized, apportioned individually or in small groups to pay the tax on their properties directly to a Visigoth. It would not be surprising if an element of status entered into this distinction; one recalls the story of the nobiles of Saintes in the Vita Bibiani, whom the Goths had a harder time dispossessing than the mediocres.³⁷ Might not the obstacle to the Goths’ ambitions have been that the nobiles were the king’s own taxpayers? By virtue of contributing directly to the monarch (in gold?), the Romans who underwrote the tertia look as though they formed a minority comparable in status to those other direct contributors to royal needs, namely, the Visigoths whose military service was compensated by the (tax) proceeds of their sortes; together, they would have formed a privileged category by comparison with ordinary Roman possessores.³⁸ The process of allotment located the Visigothic recipients at an honorable social level, much as would later happen to the barbarians in Italy.

An account of the Visigothic case is necessarily limited by the dearth of source material. The various laws suggesting that the barbarian sortes included woods and other resources not registered for taxation are best reserved for the next chapter, where the richer Burgundian evidence permits a more comprehensive exposition. As for CE 276— the only passage of Euric’s Code in which the Romans are called “hosts”—its fragmentary state and uncertain interpretation make it suitable only for discussion in an appen-dix.³⁹ Nevertheless, the few texts examined here justify our concluding that state resources, rather than those of the provincials, paid the cost of the Visigothic settlement. The tax burden, as it weighed on the payers, need not have been reduced; but, with two-thirds of the proceeds of the land tax passing out of central control and into the hands of individual Visigoths, the apparatus of fiscality was more drastically dismantled than it would be in Italy. From its beginnings, the Visigothic kingdom was fated to be a different kind of state from that which it more or less hesitantly replaced.

¹ Prosper Chron. a. 419, “Constantius patricius pacem firmat cum Wallia data ei ad inhabitandum secunda Aquitanica et quibusdam civitatibus confinium provinciarum” (MGH AA, vol. 9, p. 469); Hydatius Chron. a. 418, “Gothi intermisso certamine quod agebant per Constantium ad Gallias revocati sedes in Aquitanica a Tolosa usque ad Oceanum acceperunt” (MGH AA, vol. 11, p. 19). The Chronicle of 452 has “Aquitania Gothis tradita” under 413 (MGH AA, vol. 9, p. 654), where the others note Athaulf’s seizure of Narbonne. Jordanes Getica 160-165 and Isidore of Seville Historia Gothorum (MGH AA, vol. 11) add nothing. See also Thompson, “Settlement,” pp. 65-67; J. F. Matthews, Senatorial Aristocracy and Imperial Court, A.D. 364-425 (Oxford, 1975), pp. 318-319.

² Philostorgius Historia ecclesiastica 12. 4. On relations to Olympiodorus, J. F. Matthews, “Olympiodorus of Thebes and the History of the West (A.D. 407-425),” JRS 60 (1970): 81, 91. Olympiodorus is discernible at two removes: Philostorgius’s adaptation and Photius’s ninth-century summary of Philostorgius. The chronology of Philostorgius’s account (as we have it) is completely confused, and the grant of rations was probably unrelated to the settlement in Aquitaine. All in all, the reference to land “for cultivation” offers little guarantee of accuracy. Jones, LRE, pp. 202 and 1109 n. 65, may have taken Philostorgius more seriously than he deserves.

³ Paulinus Eucharisticus lines 282-290 (ed. H. G. Evelyn-White, Loeb, Ausonius, vol. 2): his favorite dwelling, the only one of his houses without a Gothic hospes, was plundered; in several cases, Gothic guests saved the houses they were in from damage. For the date, Matthews, Aristocracy, p. 317 (who, however, confuses guests with “settlement”). Paulinus later lost his extensive property, but for political and family reasons (Eucharisticus fines 503-515). Nothing more in his autobiographical poem can be definitely related to the special circumstances of Visigothic settlement, including lines 575579, which record the voluntary purchase by a Goth of one of his estates (it is tempting, though, to explain why the price was, in Paulinus’s estimation, “scarcely just” by conjecturing that the Goth already had rights to the estate by virtue of allotment, and that what he paid represented only Paulinus’s share; but Paulinus’s words permit only speculations). See also Chapter VI n. 27.

⁴ Rudolf Buchner, Die Rechtsquellen, Wattenbach-Levison, Deutschlands Geschichtsquellen im Mittelalter. Vorzeit und Karolinger, Beiheft (Weimar, 1953), p. 7 with n. 16.

⁵ Ibid., with n. 18.

There is a tendency, in discussions of hospitalitas, to equate Antiquae with laws of Euric (e.g., Lot, “Hospitalité,” p. 988 n. 4, on p. 989); as a comparison of CE 276 with LVisig 10. 3. 5 shows (see Appendix B), the revision was considerable, and in most cases there is no way to establish its extent.

⁶ LVisig 10. 1. 8, 9, 16, 10. 3. 5 (opening lines); moreover, 10. 1. 6 (an Antiqua emendata, i.e., revised a second time for inclusion in Reccesvinth’s collection) seems relevant because of its resemblance to LBurg 31. Antiquae mentioning consortes (e.g., LVisig 8. 5. 5, where, exceptionally, one reads consortes vel ospites) might be pertinent, since the revisers may have substituted consors for hospes (10. 1. 6); but partnership was too common an institution to be exclusively between Romans and Visigoths (e.g., 10. 1. 3, 7).

⁷ Chron. of 452, a. 443, “Sapaudia Burgundionum reliquiis datur cum indigenis dividenda” (MGH AA, vol. 9, p. 660). On the localization of Sapaudia—not identical to modern French Savoy—see Pierre Duparc, “La Sapaudia,” Académie des inscriptions et belles-lettres, Paris. Comptes rendus (1958), pp. 371-383 (with map). Auctarium Havniensis Prosperi a. 457, after the death of the Suevic king Rechiarius (Dec. 456), “Gundiocus rex Burgundionum cum gente et omni praesidio annuente sibi Theudorico ac Gothis intra Galliam ad habitandum ingressus societate et amicitia Gothorum functus” (MGH AA, vol. 9, p. 305). (An entry for 455 records that the Burgundians repulsed a Gepid attack on Gaul; after 457 there is a large hiatus until substantive entries resume.) Although the Auctarium Havniensis was assembled in early seventh-century Italy, its fifth-century entries are now believed to stem from contemporary or near-contemporary sources, notably the Ravenna Annals: Wes, Ende des Kaisertums, pp. 57-58. But critical analysis of the Auctarium Havniensis is still in flux.

For the circumstances of Gundioc’s move, which was hardly so final as the chronicle suggests, see Schmidt, Ostgermanen, pp. 140141.

⁸ Marius Chron. a. 456, “Eo anno Burgundiones partem Galliae occupaverunt terrasque cum Gallis senatoribus diviserunt” (MGH AA, vol. 11, p. 232); Fredegar Chron. 2. 46, “Et cum ibidem [that is, ad Renum] duobus annis resedissent, per legatis invitati a Romanis vel Gallis, qui Lugdunensium provinciam et Gallea comata, Gallea domata et Gallea Cesalpinae manebant, ut tributa rei publice potuissent rennuere, ibi cum uxoris et liberes visi sunt consedisse” (MGH Script. rer. Merov., vol. 2, p. 68).

⁹ Marius (d. 581) was a contemporary of Gregory of Tours; on the date and authorship of the Fredegar chronicle (ca. 660), W. Goffart, “The Fredegar Problem Reconsidered,” Speculum 38 (1963): 206241. Marius’s account has been accepted without question by most authors (e.g., Thompson, “Settlement,” p. 65; Lot, “Hospitalite,” pp. 989-990, with small reservations); it constitutes the chief prop for the idea that barbarian settlements were at the expense of rich men only—even though Marius leaves unclear whether the terrae that the senatores divided were their own. Sparsely informed about early Burgundian history, Marius must have drawn upon a source for this, his unique entry for the Burgundians in the fifth century. But he seems to have rewritten the information in his fashion; note, for stylistic comparison, partem Galliae under the year 509, and senatores under 538. His entry is best understood as a late sixth-century interpretation of distant events. Not surprisingly, its spirit anticipates that of Fredegar’s migration legend.

As regards Fredegar, Thompson, “Settlement,” p. 66: “although this is absurdly dated . . . the tradition itself seems genuine”; in the same way, Schmidt, Ostgermanen, pp. 140-141, believed he could transpose a conjectured authentic core of the passage to the events of 457; literal acceptance by Boehm, Geschichte Burgunds, p. 58; Lot, “Hospitalité,” pp. 990-991 n. 3, was more prudent. The historicity of the passage was argued by Gabriel Monod, “Sur un texte de la compilation de Frédégaire relatif à l’éstablissement des Burgondions dans l’Empire romain,” Mélanges publiés par la section historique et philologique de l'École des Hautes-Études pour le Xe anniversaire de sa fondation, BEHE, vol. 35 (Paris, 1878), pp. 229-239; he was followed by Léonzon le Duc, “Le régime de Thospitalité chez les Burgondes,” NRHD, 3d ser., 12 (1888): 232-247. Monod’s astonishing premise—that Fredegar was incapable of invention—is unacceptable. Fredegar's account partly depends on Cassiodorus-Epiphanius Historia tripartita 12. 4. 11-14 (cited above, Chapter III n. 53) and was deliberately interpolated into Jerome Chron. a. 373; it has obvious affinities to the “origin” legends then being developed by Fredegar himself, the Liber historiae Francorum 2 (where Valentinian also occurs), the Passio s. Sigismundi 1-2, and others. The reference to taxation is a characteristically Fredegarian touch (cf. Goffart, “Three Notes,” p. 379 nn. 140-141), and so is the verb form visi sunt consedisse.

For an idea of how Romans perceived the Burgundian settlement in the decades after it occurred, Vita patrum lurensium 94-95 (cited above, Chapter III n. 78): a mysterious (and presumably unwelcome) decree of Providence.

¹⁰ Buchner, Rechtsquellen, pp. 11-12.

¹¹ Lot, “Hospitalité,” p. 977, disregarded the discrepancy between the date of Burgundian settlement and that of Lex Burgundionum; LBurg 54 is treated as though it applied to 443. So also Stein, Bas-Empire , vol. 1, p. 331.

¹² CE 277 is dealt with at length below, and LBurg 54 in Chapter v. For CE 276, see Appendix B.

¹³ Alvaro d’Ors, Estudios Visigoticos, vol. 2, El Codigo de Eurico. Institute Juridico Español, Cuadernos 12 (Rome-Madrid, 1960), p. 173; the word I translate as “country estate” is finca (cf. Gaudemet, Institutions, p. 724 n. 5, “domaine”). Along the same lines, Gaupp, Ansiedlungen, pp. 394-403; Schmidt, Ostgermanen, pp. 505-506; Levy, Law of Property, p. 84; Jones, LRE, pp. 249, 251-252.

¹⁴ Chron. of 452, a. 442, “Alani, quibus terrae Galliae ulterioris cum incolis dividendae a patricio Aetio traditae fuerant, resistentes armis subigunt et expulsis dominis terrae possessionem vi adipiscuntur” (MGH AA, vol. 9, p. 660).

¹⁵ Gallia ulterior was the region that withdrew from Roman societas under the leadership of Tibatto in 435 (Chron. of 452, MGH AA, vol. 9, p. 660); for earlier unrest in the region, Thompson, “Settlement,” pp. 70-71 and Wallace-Hadrill, “Gothia,” pp. 27-28. It might be conjectured that the Alan settlement of 442 was Aëtius’s solution to the rising of 435, recovering something for the central government from a region that was lost to it anyway.

¹⁶ Chron. of 452, a. 440, “Deserta Valentinae urbis rura Alanis, quibus Sambiba praeerat, partienda traduntur” (MGH AA, vol. 9, p. 660).

¹⁷ Musset, Vagues germaniques, p. 286 n. 1. His interpretation of deserta conforms to the common modern practice; e.g., Lot, “Hospitalité,” p. 1010, "parties dévastées,” and the discussion of agri deserti in Jones, LRE, pp. 812-823, 1039-1040 (note that the index entry “deserted land” also includes pp. 774-775, presumably as an explanation of how desertion took place).

¹⁸ The important study of C. R. Whittaker, “Agri deserti” in M. I. Finley, ed., Studies in Roman Property (Cambridge, 1976), pp. 137175, 193-200 (notes), is an effective attack on the widely held modern view that “agri deserti are to be regarded as a malignant growth of the later Roman Empire, with a datable origin into the bargain.” Whittaker’s well-documented demonstration of the continued vitality of agriculture offers a strong incentive for a reassessment of the legal texts, on which he prudently commented (p. 138), “How one should interpret those regulations is a problem which I must confess I regard with uneasiness.” For provisional remarks on these texts, Goffart, Caput, pp. 67 n. 4, 137 n. 4. Note the sense of “desertion” in the phrase “ne sub praetextu militiae privatum seu publicum debitum deseratur” (NTheod II 7. 1. 2 [439]). The best indicator of the tendency of these laws occurs in those that facilitate the substitution of solvent owners for bankrupts: CTh 12. 1. 161 (399), CJ 11. 59. 11 (405s), CTh 13. 11. 13 (412), 16 (417). Where assessment registry is intermittent and less than meticulous, there will be a discrepancy between the static records and the actual vicissitudes of taxable owners and cultivable lands. This discrepancy, rather than agriculture, is what the laws on agri deserti are chiefly about They bear witness to the dynamics of the tax system, rather than to the realities of rural life.

¹⁹ A severe burden of being a late Roman curialis was to make up for deficits in tax payments, and, in the longer run, the deficit was to be shared with the other taxpayers (CJ 11. 59. 1 [312-337]).

²⁰ Salvian, as above, Chapter III at n. 79; Sidonius Carmen 13 (459); ibid., 5. 446-448 (Panegyric of Majorian, 458): though wearied by continual tributes, Gaul gladly endures a tax for building a fleet against the Vandals. Also Sidonius Epistolae 5. 13 (460s), and 3. 6. 3: “It is a common saying with provincials that a good year depends less on ample crops than on a good administration” (trans. O. M. Dalton [Oxford, 1915], 1: 73). Operations of tax machinery in the Visigothic kingdom: Cassiodorus Variae 5. 39, and the Testamentum s. Caesarii Arelatensis, cited and discussed in Goffart, Caput, p. 131 n. 26. LRB 40, on the transfer of tax liability, reiterates an old fiscal rule. The very categorical and circumstantial statement on the continuity of Roman taxation in the fifth-century kingdoms by Musset, Vagues germaniques, p. 282, is based, tacitly, on Merovingian evidence (below, Chapter VIII n. 33).

²¹ Stein, Bas-Empire, vol. 1, pp. 382-386 (too categorical in view of the evidence); A. Loyen, “Les débuts du royaume wisigoth de Toulouse,” Revue des études latines 12 (1934): 406-415; Heinrich Mitteis, Der Staat des hohen Mittelalters, 7th ed. (Weimar, 1962), pp. 31-39 (imaginative rather than reliable); Jones, LRE, p. 257 (questionable handling of the evidence); Musset, Vagues germaniques, pp. 281-284; Herwig Wolfram, Intitulatio, vol. 1, Lateinische Königsund Fürstentitel bis zum Ende des 8. Jahrhunderts, Mitteilungen des Instituts für österreichische Geschichtsforschung, Ergänzungsband 21 (Graz-Vienna, 1967), pp. 32-89 and, esp., 44-56. One of the most explicit statements about royal rights in Roman lands ceded to barbarians occurs in Claudian De bello Gothico line 539, about Alaric in Illyricum: “oppida legitimo iussu Romana coegi.” This provides an excellent gloss to the leges Theudoricianae mentioned by Sidonius Epistolae 2. 1. 3 (see Appendix D n. 10).

Claude, Westgoten, pp. 43-44, stated that the Romans remained subject to Arles but surmised that the links of the financial personnel to Arles must have been severed by about 440, when Salvian noted flight from Roman fiscality to Gothic territory. The real question, however, is whether any exercise of authority by Arles in Visigothic territory after 418 (even by way of appeal) may be positively documented. To that extent, G. Kaufmann, “Über das Foederatverhältniss des tolosanischen Reichs zu Rom,” Forschungen zur deutschen Geschichte 6 (1866): 443, 458, was right to insist on the internal autonomy of the Visigothic kingdom.

²² As above, Chapter III at nn. 81-84, where this aspect of the Italian settlement is not mentioned. It is true, however, that whatever part of the tertiae was not distributed as sortes paid its taxes to the royal government, not to the Ostrogoths, and there is no indication that the proceeds of the illatio tertiarum were treated otherwise than as general royal revenues.

²³ Since the division of private estates is asserted without reservations in the literature (e.g., Gaupp, Ansiedlungen, p. 394), it is worth observing that no source applicable to Gaul specifies that praedia, fundi, or the like were divided; the closest is LVisig 10. 1. 8 (text in n. 26 below)—a late law and still not explicit. Cf. Chapter III n. 31.

²⁴ The idea that the barbarians were taxed has a much weaker basis in the sources for Gaul than in those for Italy. Although the Variae permit a plausible argument that only closer study proves to be mistaken (above, Chapter III at nn. 65-66), nothing comparable exists in Gaul. Several modern authors have maintained that, sooner or later, the Visigoths and Burgundians were subject to taxation: Gaupp, Ansiedlungen, pp. 403-406; Dahn, Könige, vol. 6 (Würzburg, 1871), pp. 261-262; Fabien Thibault, “Les impôts directs chez les Visigoths et les Burgondes,” NRHD, 3d ser., 26 (1902): 35-38; Stein, Bas-Empire, vol. 1, p. 383. But none of the evidence they cite sustains such a conclusion. As regards the taxation of Roman property, LVisig 10. 1. 16 and, perhaps, LRB 40.

²⁵ CE 277: “[1] Sortes Gothicas et tertiam Romanorum quae intra L annis non fuerint revocate nullo modo repetantur. [2] Similiter de fugitivis qui intra L annis inventi non fuerint, non liceat eos ad servitium revocare. [3] Antiquos vero terminos sic stare iubemus sicut et bonae memoriae pater noster in alia lege praecepit. [4] Et alias omnes causas, seu bonas seu malas aut etiam criminales quae intra XXX annis definitae non fuerint, vel mancipia quae in contemptione posita fuerant, sibe debita quae exacta non fuerint, nullo modo repetantur. [5] Et si quis post hunc XXX annorum numerum causam movere temptaverit, iste numerus ei resistat, et libram auri cui rex iusserit coactus exsolvat” (ed. d’Ors, p. 21). See also next note.

²⁶ LVisig 10. 1. 8: “De divisione terrarum facta inter Gotum adque Romanum. Divisio inter Gotum et Romanum facta de portione terrarum sive silvarum nulla ratione turbetur, si tamen probatur celebrata divisio, ne de duabus partibus Goti aliquid sibi Romanus presumat aut vindicet, aut de tertia Romani Gotus sibi aliquid audeat usurpare aut vindicare, nisi quod a nostra forsitan ei fuerit largitate donatum. Sed quod a parentibus vel a vicinis divisum est, posteritas inmutare non temtet.”

Lot, “Hospitalite,” p. 998, and d’Ors, Codigo, pp. 200-201, were mistaken in stating that CE 277. 1 sets a time limit on disputing the division of sortes from tertia. They disregarded the obvious parallel with CE 277. 2 and failed to observe that Leovigild’s revised version of CE 277. 1 (= LVisig 10. 2. 1-2) is entered under the title “de quinquagenarii et tricennalis temporis intentione,” and not under “de divisionibus” (LVisig 10. 1). CE 277. 1 is about the “recovery” of both sortes and tertia—whatever that may mean—and not about their delimitation.

²⁷ On time prescription, Levy, Law of Property, pp. 184-190. See also LBurg 79. 2-5, where Euric’s example was not followed.

²⁸ Counting of generations, Levy, Law of Property, pp. 186-187. The western laws on fugitives descend from a law of Constantine (CJ 6. 1. 4, fugitive slaves) and, less clearly, from one of Valentinian III (NValent III 31, esp., parag. 5-6); but there are significant novelties, especially in the provision for nonslave fugitives and the requirement to declare strangers, which are unparalleled in East Roman legislation. See LVisig 9. 1. 1, 3, 6, 8; LRB 6. 1-2; LBurg 6, 39. 1-3; Edictum Theodorici 80, 84; Fragmenta Gaudenziana 18 (MGH Leges, vol. 1, pp. 471-472), where tributarius, rather than colonus, is coupled with servus and the latter term is often used to mean both. See also Sidonius Epistolae 5. 19. (The substitution of tributarius for colonus is of exceptional interest, since it may indicate the wholesale assimilation of taxpaying Roman possessores to the level of coloni, thus explaining the absence of coloni from the later Visigothic kingdom, as mentioned below, Chapter v n. 66.) On the public dimension of late Roman servitude, Goffart, Caput, pp. 66-90 (on slaves in particular, p. 79) and “Three Notes,” pp. 182-187.

²⁹ Nicolaas van der Wal, Manuale novellarum Iustiniani (Groningen, 1964), nos. 668-669, about Justinian Novel. 9 (535) and 111 (541).

³⁰ The tradition of special privilege in favor of the imperial fisc was firmly anchored (see the titles “de iure fisci,” Dig. 49. 14, CJ 10. 1-7, CTh 10. 1); e.g., it had priority over private creditors in the collection of debt. When Reccesvinth abolished Euric’s fifty-year prescription, in favor of thirty years, he excepted royal slaves: they could be recovered absque temporum preiudicio—fiscal privilege again (LVisig 10. 2. 4). A good illustration of the privilege extended to a “private” dominus occurs in Fragmenta Guadenziana 16: a slave or tributarius indebted to a third party must fully acquit tributa de labore suo to his master before satisfying his other debt.

³¹ Above, n. 26.

³² LVisig 10. 1. 16: “Ut, si Goti de Romanorum tertiam quippiam tulerint, iudice insistente Romanis cuncta reforment. Iudices singularum civitatum, vilici adque prepositi tertias Romanorum ab illis, qui occupatas tenent, auferant et Romanis sua exactione sine aliqua dilatione restituant, ut nihil fisco debeat deperire; si tamen eos quinquaginta annorum numerus aut tempus non excluserit.”

³³ D’Ors, Codigo, p. 176, on the basis of the famous statement (Gaius Institutes 2. 7, cf. 2. 21) that dominium over provincial land pertained solely to the populus Romanns or the emperor. But on the inapplicability of this dictum, A. H. M. Jones, “In eo solo dominium populi Romani est vel Caesaris,” JRS 31 (1941): 26-31.

³⁴ The date of the allotments is uncertain. Brunner, DRG, vol. 1, p. 74; Alfonso Garcia Gallo, “Notas sobre el reparto de tierras entre Visigodos y Romanos," Hispania 1, no. 4 (1941): 40-41; and others have maintained that the prescription of CE 217. 1 documents the antiquity of the assignment of sortes: fifty years before Euric’s accession or the issuance of his code, thus, between 419 and 431. The deduction may be correct, but, as shown here, other reasons may be imagined for the choice of a fifty-year prescription.

³⁵ Fuller discussion of this aspect, below, Chapter VI.

³⁶ There is no Visigothic evidence bearing on the size of allotments or on whether the allotment-hosts were identical to the house-hosts; for discussion of the Burgundian case, below, Chapter v at nn. 35-37.

³⁷ Above, Chapter III at n. 76.

³⁸ Sidonius wrote to two distinguished Romans in Visigothic territory who were serving the king in military capacities (one of them participated on the Gothic side in the siege of Sidonius’s Clermont) : Epistolae 5. 12, 8. 6. If the lands of these men had originally been reckoned among the tertia Romanorum, it is at least doubtful, in view of the owners' military service in the 470s, that they were still tax-payers.

³⁹ Woods and wastes, below, Chapter v at n. 26. For CE 276, Appendix B.