I INTRODUCTION: A LEGAL ENIGMA

Women had a remarkably strong position in Roman law. Especially in the imperial period, they could own substantial properties and use it fairly independently, even if they were married — something which is quite exceptional in European historical perspective. Yet for a very long time Roman law required adult women to have a male guardian (tutor). The guardian did not administer her property, but he had to endorse her major legal and financial transactions. Thus, she needed his authorisation (auctoritas) if she was drawing up a will, manumitting a slave, contracting an obligation or taking part in any of the old-style formal legal acts which the Romans were so fond of.Footnote ¹ Notably, the guardian's consent was required if a woman wished to alienate real estate in Italy, and the human and animal workforce used to cultivate it. These items (called res mancipi) had been regarded the most vital in the old agrarian society. All the other types of property, like real estate in the provinces, gold or money, were res nec mancipi and could thus be alienated without a guardian.Footnote ²

In many respects, the force of the guardianship (tutela mulierum) was already weakening in the early imperial period, as will be described in Section V. However, it was only much later that the institution of tutela was officially abolished, though the exact law has not been preserved — or so it has been believed. This lack of evidence has prompted scholars to advance widely diverging hypotheses on the date when the guardianship finally disappeared from Roman law:

The assumptions thus range from the late third century to the fifth. Still others have suggested that tutela was perhaps never formally abrogated, but might just have been allowed to fall into desuetude.Footnote ⁷

It is the aim of this paper to show that we have all been wrong, or at least too cautious. We can in fact with reasonable certainty determine the narrow time frame when tutela mulierum, the guardianship of women, disappeared from Roman law, perhaps even identify the law in question. I shall first briefly review the legal sources in Section II, before turning in Section III to the papyrological evidence, which will prove vitally important for the argument. Section IV will discuss the relevant laws within the proposed time frame, while the concluding Sections V and VI will place the legislation in the broader historical context of guardianship and the legal position of women.

II LEGAL EVIDENCE

The most notable breach in the system of tutela mulierum had been made already at the beginning of the Principate. The Augustan marriage laws, which tried to encourage citizens to have more children, had created various disadvantages for unmarried and childless people, and privileges for those who had fulfilled their marital duties. Most of these rules concerned inheritance and public obligations, but for women there was a special incentive: if they had given three live births they received the ius liberorum, ‘right of children’, and were exempted from tutela altogether. For freedwomen the requirement was four live births.Footnote ⁸ The demographic realities of the ancient family meant that in practice a very large proportion of freeborn women could attain the ius liberorum: we shall return to this in Sections III and V. However, not all did, so tutela remained for centuries a fact of life for many women, before the Roman lawgivers finally decided that the guardianship of women deserved to be abolished — but when?

We shall start by examining the limits of the study: the terminus post quem and terminus ante quem. The former is provided by the Vatican Fragments (hereafter Fr. Vat.), an enormous, though only partially preserved, collection of existing legal principles taken from earlier imperial enactments and juristic writings. The surviving part contains almost 350 such excerpts. The guardianship of women figures in so many places that it clearly was a relevant institution at the time the compilation was made. Two examples shall stand for the others:Footnote ⁹

The existing text of the Vatican Fragments derives from a palimpsest of the fifth century. However, the original redaction must have taken place at Rome in the first decades of the fourth century.Footnote ¹⁰ The collection of material may have started already in the last years of Diocletian (as he is in the earlier part presented as living, later as divus), but the work contains many rescripts up to the year 318, where the original compilation seems to have ended.Footnote ¹¹ It is clear that the collection then underwent revisions up to the end of the fourth century. In that process, at least a couple of further laws, some juristic commentaries and a few dozen scholia were added.Footnote ¹² Moreover, the controversial legitimacy of the Tetrarchic emperors after Diocletian's abdication caused insurmountable difficulties for the original compilator and for the subsequent editors as they attempted to produce and reproduce the politically correct (retrospective) composition and titles of the imperial coalition at any given time. For example, after the damnatio memoriae of Licinius in 324, his name and title were inconsistently removed from the headings and consular dates of individual laws.

All this makes it difficult to define when the compilation was ‘finished’. It remains possible that the first technical revisions and perhaps additions were made already before Licinius’ fall, i.e. between 320 and 324, as some preserved imperial titulatures may reflect the way they would have been edited in the west during that time.Footnote ¹³ However, this does not change the fact that almost all the incorporated laws date from before 320, implying that at the very least the core substance of the original compendium had been finished by that year and not much before.Footnote ¹⁴ We may thus conclude that tutela had not yet been abolished by around 318. This remains the secure terminus post quem irrespective of the uncertain subsequent editorial history.

As to a terminus ante quem, it has proved elusive. By 390 at the latest, mothers could assume the guardianship over their own children (CTh 3.17.4). It might be natural to conclude that they would then not have been under tutela themselves. However, the guardianship over children and that over adult women were legally very different constructs. While the latter had become often a formality, the former entailed real administrative trouble and power. Thus, it is not impossible to think that tutela mulierum could still survive, especially as so many women were in practice freed from it.Footnote ¹⁵

In 410 the ius liberorum was made universal (CTh 8.17.2–3):

As can be seen, the text of the law did not concern the guardianship. It refers only to the restrictions placed on testamentary succession between man and wife, and it gives the impression that at this time the ius liberorum no longer had any other effects. However, this impression is not quite accurate, as the ius liberorum still determined the succession rights of mothers up to the sixth century.Footnote ¹⁶ Hence, there remains a slim possibility that it might have played a role in the question of tutela as well — theoretically even after 410.

Thus, legal sources do not provide a secure terminus ante quem, which is of course precisely the reason why scholars have disagreed so much on the date. And that is why we must consider a different kind of evidence: the papyri from Egypt.

III PAPYROLOGICAL EVIDENCE

In Egypt, the Greek legal system of the Ptolemaic period remained in force after the Roman conquest. A type of guardianship was an integral part of it. Thus, in many legal transactions an adult woman had to be assisted by a guardian (kyrios), who was usually her husband or close relative. Such guardians are very common in the papyri of the late second and early third centuries. Yet, there was a possibility of confusion, as the same Greek word kyrios could in the documents refer either to this traditional guardian of the Greek type or to a Roman tutor (if the woman was a Roman citizen) or even to a local interpretation or mixture of these. That is why the guardians appearing in the papyri are, especially after 212, often styled ‘guardian according to Roman custom’ (κύριος κατὰ τὰ Ῥωμαίων ἔθη).Footnote ¹⁷

Subsequently, we encounter more and more documents where the concept of the Roman tutela mulierum is evident. A prime example comes from the year 261 (P.Oxy. XXXIV 2710):

Another relevant papyrus is roughly contemporaneous, from the year 263 (P.Oxy. XII 1467 = FIRA III 27):

Here a woman claims that she has given birth to the required number of children and thus has the ius liberorum and is therefore exempt from guardianship. It shows that women in Egypt were well aware of the system of guardianship and also the main loophole in it. This petition is the only one of its kind, probably because there was no requirement to record the ius liberorum in any official archive. The actual purpose of Aurelia Thaisous’ petition thus remains a matter of speculation.Footnote ¹⁸

For the whole of the third century, we have over one hundred documents where individual women are conducting financial transactions or approaching authorities and either assert that they are assisted by a guardian (μετὰ κυρίου), or are entitled to act without him (χωρὶς κυρίου) on the strength of the ius liberorum.Footnote ¹⁹ It should be noted that even after 212 not all of these documents concern cases where Roman law actually required the guardian's consent. Quite obviously women tended to record the ius liberorum even when it would not have been needed — just to be on the safe side, or because it had status value.Footnote ²⁰

Fig. 1 shows the number of papyri where either the presence or absence of a guardian is recorded, by decade. It is not an exhaustive display of all surviving transactions in which a woman has taken part. Sometimes the text is too fragmentary to record the guardian's presence or absence, sometimes the matter did not need to be confirmed by a guardian, and sometimes the information may just have been omitted.Footnote ²¹ For example, from the same period there are dozens of papyri where a woman is said to be assisted (συνεστῶτος or συμπαρόντος) by a man who is not a guardian, though in over thirty of these the ius liberorum is not adduced either. On the other hand, there are two dozen cases where such assistance is explicitly connected with the lack of guardian because of the ius liberorum.Footnote ²²

Fig. 1. Numbers of documents mentioning guardians or their absence (in black; scale on the right) compared to the total number of dateable published papyri (in grey; scale on the left), by decade. See text for exclusions.

This curve is compared with the number of all documentary papyri published from each decade. I have adjusted the curve of all papyri between 241 and 270 by omitting three large homogeneous groups of administrative documents which have no relevance for this question but inflate the statistics. They are the Heroninos correspondence, the dossier of the Hermoupolis city council (CPR XXXV) and the Decian libelli. Other comparable dossiers are not as large and do not affect the curve's general shape. The anomalous increase in the number of published papyri in the decade 211–220 may partly be explained by the shorter reigns of Caracalla, Geta and Macrinus, which make it easier to date documents by decade, while the preceding and succeeding reigns of Septimius Severus and Alexander Severus were split between decades. I am not aware of any study which has attempted to analyse this effect.

As can be seen, the number of documents mentioning guardians or their absence roughly follows the overall curve up to the early fourth century, when it starts to fall somewhat more rapidly. However, within this group, those where the woman is assisted with a guardian and those where she is acting without him develop in totally different ways. Table 1 gives the numbers of these two types between the years 180 and 360. After around 240, the documents where a guardian is present become very rare and almost disappear in the late third century. At the same time, the number of documents where a woman has the ius liberorum is increasing and remains significant up to the beginning of the fourth century. Fig. 2 shows this more clearly in a graph by decade.

TABLE 1 Women with and without guardian.

Fig. 2. Numbers of documents mentioning presence of guardian (grey) and absence of guardian (black), by decade.

These statistics suggest that after the Constitutio Antoniniana in 212, it took almost thirty years for the people in Egypt to come to terms with the Roman type of guardianship and the way out of it. In practice, it meant that subsequently much fewer women were under guardianship, as most of them could boast the ius liberorum and the legal exemption. Of course, this development does not show that tutela as such had been abolished — on the contrary, it shows that it remained in force and women still wanted to mention the ius liberorum to escape it. And now we come to the final phase, and the crucial period where we can detect a change.

Table 2 shows the last published papyri where the ius liberorum is mentioned. Between 300 and 325, there is on average almost one document per year. The short gaps in 310–313 and 315–317 can simply be explained by random variation: no other reason is plausible. Clearly women were still keen to claim that they did not need a guardian. But after 325, there is a sharp stop, and then a few stray cases separated by several years. As time went on, these outliers tended to mention only the ius liberorum but no longer the guardianship. They may just reflect a conservative practice of a few scribes, using antiquated pattern books, or some uncertainty in a transitional period after the guardianship had disappeared from imperial law. I shall return to this transition in the property regime in Section VI.

TABLE 2 Last papyri with the ius liberorum.

* Note: This unpublished papyrus is cited in P.Mich. XV p. 167.

We may thus conclude that around 325 or a little after (allowing for some random variation here as well) Egyptian women, or their scribes, rather abruptly ceased to plead the ius liberorum to escape the guardianship. If we combine this discovery with the evidence of the legal sources, we can now narrow the search for a law which abolished the guardianship to the period between around 320 and 325. And there are in fact two candidates for such a law.

IV THE ABOLITION OF TUTELA

On 31 January 320, Constantine issued an edict which, after three centuries, relaxed the Augustan penalties affecting people who were not married or did not have children (CTh 8.16.1). This law is very well known, and was extolled already in antiquity by the church historians Eusebius and Sozomen, who underlined its benefits for Christian celibates. How far it really was motivated by Christian ideals is less clear, as the Augustan regulations had always been very unpopular.Footnote ²³

Whichever reasons had prompted Constantine or his advisors to promulgate his edict at the end of January 320, his wording implies that at least the most onerous inconveniences for the unmarried and childless people were going to be removed. However, the extant text does not give any detailed list of the cancelled rules. Only the capacity to receive testamentary bequests, which was perhaps the most important issue for males, is mentioned (‘aequa condicio capessendi’). Further, as Constantine's favour did not extend to the bequests between husband and wife, that was separately cited (‘huius beneficii maritis et uxoribus inter se usurpatio non patebit’).Footnote ²⁴ Otherwise the text, as we have it, passes over all the other disadvantages. Nor are the privileges mentioned which had been granted to those who were blessed with a large family — although the lack of them might also be seen as a kind of penalty. Because of this general silence, the law has always been taken to concern only the rights of inheritance, the topic where caelibes and orbi were particularly disfavoured (e.g. Gai., Inst. 2.286). According to this interpretation, all the rest of the Augustan regulations and their later modifications were left to stand as they still were around 320.Footnote ²⁵

However, this is not necessarily the whole truth. There is an odd sentence (bolded in the quote) referring particularly to women and citing ‘commands of the law which were placed on their necks like yokes’. It is striking how emphatically the lawgiver underlines the effects on women. It would have been quite a prolix way to express a principle which would have been self-evident for any educated jurist: the same rules normally applied to both sexes, unless otherwise indicated.Footnote ²⁶ If the same yokes had been placed on both male and female necks, we would expect a different formulation. Instead, it seems that some very specific legal burdens, which had been imposed on females, were now also relieved.

What could this burden be which applied just to women? The sole inheritance restriction which affected females alone was connected with a mother's right to inherit from her children (SC Tertullianum), a right which depended on the ius liberorum. But as in that particular context the ius liberorum remained in force up to the reign of Justinian, it could not be the burden which was removed in 320.Footnote ²⁷ A second-century administrative handbook from Roman Egypt, the Gnomon of the Idios Logos, seems to codify some of the Augustan regulations, though it is unclear if they include also details which were specific to Egypt. There were gender-specific limits of age and property value, but these would hardly have justified a separate mention in Constantine's law.Footnote ²⁸ The Gnomon also mentions an extra property tax of one per cent on unmarried women, not attested in other sources. However, if this was meant in CTh 8.16.1 it could not have been covered by the extant text because the tax was in no way linked with inheritances and thus would definitely have required a separate lost clause anyway.Footnote ²⁹

An obvious explanation is that the constitution here referred to tutela mulierum, which was thus abolished together with the testamentary restrictions. To be able to read the law in this way, we have to understand the textual history of the Theodosian Code. It does not preserve any constitutions exactly in their original form. The laws were often split up and the pertinent segments were relocated where their subject matter in the Code belonged. And we indeed know that the text of CTh 8.16.1 was just one excerpt from a much longer edict which covered various subjects, including testamentary processes, but also other matters.Footnote ³⁰

Moreover, even the Theodosian Code itself has not survived in the form in which it was once compiled. The first five books were covered by the manuscript T (lost in a fire in 1904), but it was so fragmentary that only around one-third of its text has been preserved.Footnote ³¹ A part of the losses can be supplemented from the Breviarium Alaricianum, which is a sixth-century abridgment of the Code. A further part can be reconstructed from the constitutions which were later incorporated in the Justininan Code. The extent of omissions within a single title can be seen e.g. in title 2.8, where the Breviarium preserves only three constitutions from the original twenty-six.

Thus, we may surmise that some relevant sections of Constantine's original extensive law were included in those titles of the Code that are now deficient or missing in its fragmentary manuscript tradition. This is actually demonstrated by the three excerpts from the same constitution which have been preserved in the Justinianic but not in the Theodosian Code, although they cannot have ended up in the former in any other way than through the latter.Footnote ³² Hence, a passage specifying more distinctly the fate of tutela may subsequently have been inserted in some other title of the Theodosian Code, where it was thought to fit better, because it did not concern inheritance but guardianship.Footnote ³³

It is impossible to know exactly how the original text might have run before the constitution was split up. The compilers of the Code could adjust the wording of a law to make the text flow better after they had removed one or more sentences. This practice can be seen in some rare cases where an original constitution has been preserved independently outside the Code. Our most valuable source in this respect is a separate collection of statutes, the so-called Sirmondian Constitutions. It includes many extensive laws which were later abridged, slightly edited, cut into pieces, and then inserted in a truncated form under different titles of the Theodosian Code.Footnote ³⁴ The crucial extract on tutela may have followed the last sentence of the quote, or even preceded it, depending on how much the compilers had edited the section. The passage certainly appears to us more ambiguous than to the contemporaries, who could read the intact wording.

Assuming that we have correctly identified this constitution and determined 31 January 320 as the date when tutela mulierum was officially eliminated, it might be asked why the change was reflected in the papyri only five years later. We could of course suspect that laws enacted in Rome often reached Egypt, or at least became known to scribes, only after some lapse of time. But there is an even simpler explanation. Between the years 320 and 324 Egypt was not ruled by Constantine, but by his rival Licinius. We are not well informed about the legal situation in Egypt during those years. However, the two rulers’ relationship was already quite strained at that time.Footnote ³⁵ This would have made it even less likely that Constantine's enactments affected Egyptian scribal practice. Constantine finally defeated Licinius and captured Egypt in late 324. At the end of 324, Constantine took care to annul Licinius’ legislation and confirm his own, and the church historian Eusebius in his religious and political zeal indeed contrasted Constantine's beneficial legislation with the barbarous laws of Licinius.Footnote ³⁶

True, we have to admit that the connection of CTh 8.16.1 with the abolishment of tutela cannot be made watertight. That is why it is worth noting that there is another possibility within the same narrow timeframe. A law of 324 (CTh 2.17.1) decreed that young men and women who had not yet attained the age of legal majority (twenty-five years), but had behaved properly, could be granted a personal majority (venia aetatis) a few years earlier (age of twenty for males, eighteen for females).

The law's focus is on the age of majority, and not on guardianship, which at this time may not have been regarded as a very important element in everyday transactions anyway (see next Section). That is why this law has not been taken as conclusive proof of the disappearance of tutela, especially as we do not have the full text preserved.Footnote ³⁷ However, the wording strongly suggests that women had exactly the same rights as men, and the guardianship was no longer in play. Whether it had been removed a few years before, or only by this very same constitution, is impossible to say because there is again a fair chance that important parts of the text have been omitted and possibly placed somewhere else in the Code.

In sum, there are two alternative ways to understand Constantine's constitution of 320. We may assume that a vital part has disappeared, and the law removed the tutela mulierum together with the restrictions on inheritance. Alternatively, we may believe that the law was aimed at the Augustan penalties only in a narrow sense, leaving the ius liberorum and tutela untouched. This would require us to believe that whoever drafted the law just chose for some unknown reason to emphasise its effect on females. However, even the second alternative would not much change the general picture. There must have been a law which abolished the guardianship, and it almost certainly must have been enacted between 320 and 325.

V GUARDIANSHIP, LAW AND SOCIAL REALITIES

To put the Constantinian legal change into a historical and social perspective, we have to assess the impact which tutela might have had on Roman women's lives. This is not an easy task, as it affected them very differently depending on their personal circumstances, such as age, domicile, financial position, free or slave birth, number of offspring and even the personalities of the woman and her guardian. Moreover, changes in Roman society and legal institutions over the centuries meant that the effects of tutela did not remain constant.

Already in the late Republic wealthy women seem to have administered their property quite freely, not really hindered by their guardians. It became possible to change the guardian and to overturn his veto in a court. For some time, a certain category of guardians (tutores legitimi), typically the nearest relatives in the paternal line, exercised more effective power, but this type was abolished by Claudius. After this, in practice only the patronus of a freedwoman preserved the strong position which guardians in the earlier stages of Roman history had held. We may assume that in the case of freeborn women it was above all individual character which determined how much influence the guardian really had.Footnote ³⁸

The result of these developments is described in the Institutes of Gaius, a legal textbook from the late second century. His remarks on the nature of guardianship are one of the most famous contemporary statements on Roman women, cited in almost every study written on their legal and social position.Footnote ³⁹

Clearly Gaius did not consider the guardianship a very useful, or meaningful, institution, at least for many women. There were exceptions which will be discussed shortly. And of course the whole idea behind the Augustan laws — that exemption from the guardianship was a privilege — shows that tutela was not regarded by the contemporaries as a general blessing to women.

The significance of the Augustan innovation, the ius trium liberorum, cannot be overestimated. It was not only a conspicuous symbolic gesture. From comparative population studies we know that in a pre-industrial society with high infant mortality, the average life expectancy at birth must have been between twenty and thirty years. Consequently, female fertility had to be very high, unless the Roman population was in drastic decline, which does not seem to have been the case. Few ancient sources provide enough statistical material to support this theoretical model, nor did the actual fertility need to remain exactly uniform across time, region or social class. However, the Empire as a whole could not possibly have escaped these demographic realities. This means that on average each woman who survived to her menopause must have borne four to six children, half of whom would have died before adulthood.Footnote ⁴⁰ It follows that a woman who reached, say, her late twenties would usually have produced three live births.Footnote ⁴¹

Of course, there were still quite a few women who were under guardianship. For freedwomen it was much more difficult to achieve the required family size, especially because they were usually not freed at an early age. We shall return to this. But even among freeborn women there was an obvious large group who had not yet acquired the ius liberorum: young women. Although Mediterranean girls married quite early by modern standards, not a few in their mid-teens, and probably almost all by the age of twenty, it is unlikely that many had given three births before their early twenties. The Roman guardianship of orphan girls, tutela minorum, ended at the age of twelve, so for at least ten years, and often more, they were supervised by a tutor mulieris, before they had produced the required number of offspring. We may imagine that for such teenagers the assistance of a guardian could often seem beneficial. All this of course concerns only girls who had lost their father — otherwise they would have been under patria potestas, paternal power, and would not even have owned anything.Footnote ⁴²

For a long time, these younger women provided one logical argument for keeping the system of guardianship working. But from the late second century this argument started to crumble. A new type of guardianship was formalised, cura, the curatorship.Footnote ⁴³ Now curators supervised the affairs of young men and women after they had attained puberty but before they had reached the legal age of twenty-five. And a law from the reign of Gordian (Cod. Iust. 5.37.12) confirms that in the third century ius liberorum did not exempt women from curatorship:

As a result, since cura now covered exactly the age when tutela mulierum had been the most useful, there would certainly have been reason to question the existence of two overlapping systems.

Moreover, the universal grant of Roman citizenship in 212 made an additional curiosity evident:

The vexed juridic question in this precise case need not interest us further here.Footnote ⁴⁴ The passage reveals a more important general anomaly. As was noted in the beginning, a woman needed her guardian to alienate res mancipi, the most important of which was land in Italy. However, after 212 the land that most citizens owned was not located in Italy but in the provinces, so it was not res mancipi. This meant that most citizen women could now freely sell or donate all their landed property without asking the guardian, but still needed him to transfer a donkey or cow properly. If this did not seem outright silly, at least it must have undermined the logic of guardianship even further.

The developments described so far would suggest that the guardianship had gradually lost much of its previous importance and thus could be removed by Constantine without further consequences for the legal, or social, order. However, it remains to consider a sizable further group affected by tutela: freedwomen (libertae). As they were mostly freed only after the age of thirty, it would have taken them longer to achieve the required number of four freeborn births, if they reached it at all.Footnote ⁴⁵

Although many rules of the guardianship affected freeborn and freed women in the same way, the existence of tutela legitima was a notable exception. It had otherwise been removed already in the first century, but the patronus of a freedwoman (and even his male heir) still became her tutor legitimus, a fact which gave him much more power than the other types of guardians had.Footnote ⁴⁶ He could not be easily compelled by the praetor to give his authorisation. And perhaps most important of all, he could prevent the woman from writing a will, thus remaining her primary heir on intestacy.Footnote ⁴⁷ This meant that, in the early Principate, the patron could exclude even the freedwoman's own children from inheritance because in the Roman family system they did not count as her nearest relatives. Gaius was well aware that the guardianship in this case was directly meant to benefit the patroni, and it remained an important aspect of tutela in the late second century (Inst. 1.192, 3.43). Of course, we do not know how often the patrons really used their authority. And in any case, this prerogative soon lost part of its force, as the SC Orphitianum in 178 decreed that a woman's children would be her primary heirs on intestacy before anyone else (Tit. Ulp. 26.7; Dig. 38.17.1).

In sum, it can be argued that in the early fourth century libertae remained the largest group of women who still to some extent suffered ‘under the yoke’ of guardianship.Footnote ⁴⁸ And as the Romans certainly wished to protect the interests of slaveholders, this may have served as one rationale for keeping tutela in force. Nevertheless, I suspect that Roman legal conservatism still played the decisive role. After all, if freedwomen had been the primary reason behind the perpetuation of tutela and if the Romans had been inclined the adjust their law to social realities, it would have been simple to determine that the guardianship applied only to libertae. At least it is difficult to believe that Constantine particularly wanted to favour freedwomen with the abolishment of tutela, though it remains remotely possible that his words in CTh 8.16.1 (‘promiscue omnibus’) might have referred to freeborn and freed women together. Finally, one additional reason why tutela survived so long might have been that it formed part of the official population policy based on the Augustan marriage laws.Footnote ⁴⁹

The history of tutela mulierum in the imperial period demonstrates how difficult it is to use Roman law as a source for social history. The decisive change in the social and economic role of women had taken place already by the beginning of the Principate, but it did not suffice to put an end to the institution of guardianship. Nor were the subsequent legal developments directly linked with the status of women. They were determined by the Augustan population policy, the weakening of agnatic ties, the care of fatherless minors and the universal grant of Roman citizenship. During the three centuries of the Principate, the Romans did not see a need to remove tutela from their law, even if step by step it lost most of its original meaning. And as we cannot establish the various motives behind Constantine's comprehensive but imperfectly preserved legislation in the 320s we do not know if his court was interested in the gender roles in any particular way.

Still, Constantine's law can be seen essentially as the final removal of an antiquated and, for freeborn women, largely obsolete legal relic, while it had the more or less unintentional practical consequence of benefitting above all many freedwomen. Whichever we regard as more important, the abolition of tutela was nonetheless a milestone in the long history of Roman women's legal status. It took place at a time when their social and economic position was still at the same level as in the early Principate, if not better. In the 290s, one third of all imperial rescripts were sent to female petitioners, and the proportion of women had indeed risen from the early third century.Footnote ⁵⁰

On the other hand, we must admit that the history did not end here. The guardianship definitely vanished from imperial law, once and for all. There is no trace of it in Justinian's compilation or in later western legislation. However, the papyri in particular indicate that after this the story of women's supervision continued in another form. We shall conclude with a brief review of this sequel.

VI FEMALE SUPERVISION AFTER CONSTANTINE

Unlike most other legal systems, Roman family law in the Principate had been based on a strict separation of the spouses’ properties. This was partly due to the paramount position and economic power of the Roman father (patria potestas). Hence, women were not normally placed under the supervision of their husbands — and as the fathers were likely to die sooner rather than later, the result was a wide practical independence of adult women. Husbands do not seem to have been generally banned from being their wives’ guardians, but it was certainly not an automatic choice in Roman society. And since the ius liberorum usually exempted mature freeborn wives from tutela this was rarely an issue. Perhaps more importantly, husbands were forbidden to serve as curators for their young wives.Footnote ⁵¹

A totally different system is found in the papyri of the sixth century, though the roots may be visible much before. Already in the third century Egyptian husbands had often informally functioned as ‘assistants’ for their wives who had the ius liberorum. By the sixth century, the old Roman tutela had been replaced by a semi-official arrangement where husbands acted as de facto guardians for their wives, sometimes (though rarely) even called kyrioi. At the same time, widows (and perhaps nuns) were free to conduct their own affairs, often using the phrase ‘acting without husband as guardian’.Footnote ⁵² This later system may be called semi-official because the phrases appear regularly in Byzantine papyri, although there was nothing to support the practice in contemporary imperial law. Similar tendences can be perceived in the limited sources of post-Roman western Europe, in Gaul, Italy and North Africa: married couples in the fifth and sixth centuries conducted their affairs together, while widows acted alone.Footnote ⁵³

Such developments after Constantine may help us to understand better the history of the guardianship and its connection with the whole Roman social and legal system. As said, that system emphasised to an exceptional degree the paternal family line at the cost of other human relationships, such as the maternal line or the marital bond. That is why tutela was so closely connected with the agnatic family and, unlike in most other societies, was separated from the conjugal relationship. The agnatic principle and its corollaries were already weakening in the early Principate, and the spread of Roman citizenship to all the Empire's inhabitants certainly reinforced this process. When the guardian was no longer needed to protect the interests of the agnatic family, he became dispensable in Roman law.

On the other hand, there was less need for a nominated guardian in the new (and more universal) system of family finances, which prevailed by the early medieval period. The spouses’ properties were thought to be held in common, and the husband assumed his role in the family affairs by nature. Besides, the wife also had rights to his assets, so that she had to consent, too, if anything was going to be alienated from the common estate.Footnote ⁵⁴ A western constitution of 444 quite rightly remarked that, even if the properties remained separate in theory, the everyday costs of a household were difficult to divide afterwards, when the marriage ended. A contemporary eastern law suggested that a wife might do well to leave property decisions to her husband, but it still stressed that the choice was entirely hers.Footnote ⁵⁵

As widows were not placed under male supervision even where husbands more or less informally controlled their wives, we can conclude that in the centuries after Constantine there was no general female incapacity in the Mediterranean world. True, local customs in different regions may have varied considerably. But by and large the juridic sources indicate that in the eyes of the law Roman women in Late Antiquity had not lost the legal and financial independence which they had achieved by the early Empire, at least in the urban upper and middle classes. The position of women was as strong as ever in the legislation of Justinian. It is more difficult to follow the developments in the west, but even there it seems that at least widowed women preserved their legal autonomy until the traces of the old Roman populations fade away in the Early Middle Ages.Footnote ⁵⁶