Kelly ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 2017 Benjamin Kelly Proving the ius liberorum: P.Oxy. XII 1467 Reconsidered Benjamin Kelly Department of Manuscripts, British Museum, Inv. 2458 Oxyrhynchus A.D. 263 ̣ ̣] ̣α̣[ ̣ ̣ ̣] ̣ ̣[ ̣] ̣ ̣ ̣[ ̣ ̣ ̣,] δ[ιαση]- µότατε ἡγεµών, οἵτινες ἐξουσίαν διδόασιν ταῖς γυναι- ξὶν ταῖς τῶν τριῶν τέκνων 5 δικαίῳ κεκοσµηµένα[ι]ς ἑαυ- τῶν κυριεύειν καὶ χωρ[ὶς] κυ- ρίου χρηµατίζειν ἐν αἷς ποι- οῦν[τ]αι οἰκονοµίαις, πο[λλ]ῷ δὲ πλέον ταῖς γρά[µ]µατα 10 ἐπισταµέναις. καὶ αὐτὴ τοί- νυν τῷ µὲν κόσµῳ τῆς εὐ- παιδείας εὐτυχήσασα, ἐνγράµµατος δὲ κ̣α[ὶ ἐ]ς̣ τὰ µάλιστα γράφειν εὐ⟦β⟧ κ̣̀ ό̣́π̣ως 15 δυναµένη, ὑπὸ περισσῆς ἀσφαλείας διὰ τούτων µου τῶ[ν] βιβλ{ε}ιδίων προσ̣φ̣ω<νῶ> τῷ σῷ µεγέθ<ε>ι πρὸς τὸ δύνα- σθαι ἀνεµποδίστως ἃς ἐν- 20 τεῦθεν ποιοῦµαι οἰκ[ον]οµία[ς] διαπράσσεσθαι. ἀξιῶ ἔχε[ιν] αὐτὰ ἀπροκρίτως τ̣ο̣[ῖς δι-] καίοις µ[ο]υ ἐν τῇ σῇ τοῦ [δια-] σηµοτάτου τ[ά]ξ<ε>ι, ἵνʼ ὦ β[εβο]- 25 ηθ[η]µέν̣η̣ κ̣[α]ὶ̣ ε̣ἰ̣[σ]αεί σ[οι] 106 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 χάριτας ὁµολογήσω. διευτ[ύ]χ[ει.] Αὐρηλία Θαϊσ[ο]ῦς ἡ καὶ Λολλ[ι]- ανὴ διεπεµψάµην πρὸς ἐ- π̣ίδοσιν. ἔτους ι Ἐπεὶφ κ[ ̣]. 30 ἔσται σο[ῦ] τὰ̣̣ βιβλία ἐν τῇ̣ [τάξει.] (Laws exist), most distinguished prefect, which give to women possessing the ius trium liberorum the right to be independent and to act without a guardian in whatever business they transact, especially those who know how to write. Accordingly, as I too enjoy the happy honor of being blessed with children and as I am a literate woman able to write with a high degree of ease, it is with great assurance that I appeal to your highness by this my petition with the object of being enabled to accomplish without hindrance whatever business I henceforth transact, and beg you to keep it without my rights having been judged in your eminence’s office, in order that I may obtain your support and acknowledge my unfailing gratitude. Farewell. I, Aurelia Thai- sous, alias Lolliane, have sent this petition for submission. Year 10, Epeiph 2[.]. Your petition will be kept in the office. 1. The text and its interpretations This papyrus, initially published in 1916 as P.Oxy. XII 1467 (cf. BL VIII 246), has often been reprinted in collections of papyrological documents1 and translated in sourcebooks.2 It owes its celebrity to the fact that—at least on its face—it would appear to give precious information about two legal require- ments that female Roman citizens had to satisfy in order to obtain the ius liberorum and thereby be exempted from tutela— requirements evidenced in no other text. The document sug- 1 Jur.Pap. 14; Sel.Pap. II 305; FIRA III 27; New Docs. II p.30; Pestman, Prim.2 65. 2 J. Rowlandson (ed.), Women and Society in Greek and Roman Egypt: A Source- book (Cambridge 1998) 193, no. 142; T. G. Parkin and A. J. Pomeroy, Roman Social History: A Sourcebook (London/New York 2007) 307, no. 8.11; A. Arjava, “The Romanization of Family Law,” in J. G. Keenan et al. (eds.), Law and Legal Practice in Egypt from Alexander to the Arab Conquest (Cambridge 2014) 186–187, no. 4.3.3. BENJAMIN KELLY 107 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 gests: (a) that women who wished to obtain this right had to be literate; and (b) that to qualify for or exercise the right, women had to petition the governor of their province (or, potentially, some other official, in the case of women in Italy). The mention of literacy is, however, a red herring. The papyri contain many cases of women acting without a guardian pursuant to the ius liberorum. Many of these women are also said to be illiterate in the document—too many for literacy to have been a prerequisite.3 But there has been rather more support for the suggestion that women had to petition to obtain the right. Grenfell and Hunt entitled the editio princeps “Petition for ius trium liberorum,” a title echoed in many of the republications of the text.4 They described it as “the first papyrus to illustrate the process by which the right was secured” (p.196). This in- terpretation of the text has found weighty support ever since.5 3 S. Solazzi, “ ‘Ius liberorum’ e alfabestismo: a proposito di P.Oxy. 12. 1467,” Scritti di diritto romano II (Naples 1957) 229–237, at 231–232 (= Rend IstLomb 51 [1918] 586–597). Solazzi’s position was reinforced by Peter Sijpesteijn’s updated collections of the relevant evidence: “Die χωρὶς κυρίου χρηµατίζουσαι δικαίῳ τέκνων in den Papyri,” Aegyptus 45 (1965) 171–189, at 180–187, and P.Mich. XV pp.158–171. 4 For example, Hunt and Edgar, Sel.Pap. II p.321; Pestman, Prim.2 p.243; Parkin and Pomeroy, Roman Social History 307. 5 See for example P. Meyer, Jur.Pap. p.34: “Eine ingenua (s. Z. 4) be- antragt auf Grund der augustischen lex Iulia et Papia Poppaea und ihrer Ausführungsbestimmungen für Ägypten (Z. 1 ff.) Befreiung von der Vor- mundschaft durch das ius liberorum.” Rowlandson, Women and Society 193: “Aurelia Thaisous alias Lolliane applies to the prefect of Egypt for per- mission to conduct her affairs without a guardian…” For statements along similar lines see E. Turner, “Roman Oxyrhynchus,” in A. K. Bowman et al. (eds.), Oxyrhynchus: A City and its Texts (London 2007) 141–154, at 146 n.11 (= JEA 38 [1952] 78–93, at 84 n.1); H. C. Youtie, “ἀγράµµατος: An Aspect of Greek Society in Egypt,” in his Scriptiunculae II (Amsterdam 1973) 611–627, at 616–617 (= HSCP 75 [1971] 161–176, at 166–167), and “ὑπογραφεύς: The Social Impact of Illiteracy in Graeco-Roman Egypt,” in Scriptiunculae posteriores I (Bonn 1981) 179–199, at 199 n.62 (= ZPE 17 [1975] 201–221, at 221 n.62); S. Pomeroy, “Women in Roman Egypt: A Preliminary Study Based on Papyri,” ANRW II 10.1 (1988) 708–723, at 721; J. A. Sheridan, 108 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 For example, in her survey of women and the law in Roman society, Jane Gardner interprets the document as evidence for a compulsory application regime: “One third-century docu- ment survives from which we gather that women claiming the right to be exempt from tutelage had to apply to the prefect’s office (or, presumably, that of the appropriate official else- where) to have their claim placed on record.”6 In her im- portant sourcebook on women and the law in the Roman Empire, Judith Evans Grubbs goes further, conjecturing that additional evidence had to accompany the application:7 Presumably women who qualified for the ius liberorum had to submit proof of their child-bearing to the authorities so that their right to act without a tutor would be officially registered. However, only one such application for the ius liberorum actually survives. If this interpretation of the text is correct, then it has ramifications for how we understand the operation of the ius liberorum in the societies of the Roman Empire. Although not necessarily prohibitively expensive, petitioning had its costs, in terms of both time and money, so not every woman would have been in a position to create and deliver the necessary application. It is also conceivable that some men would have prevented women in their families from applying, so as to maintain control over family property.8 ___ “Women without Guardians: An Updated List,” BASP 33 (1996) 117–131, at 130–131, and “Not at a Loss for Words: The Economic Power of Literate Women in Late Antique Egypt,” TAPA 128 (1998) 189–203, at 198; K. Vandorpe and S. Waebens, “Women and Gender in Roman Egypt: The Impact of Roman Rule,” in K. Lembke et al. (eds.), Tradition and Transformation: Egypt under Roman Rule (Leiden/Boston 2010) 415–435, at 420–421. 6 J. Gardner, Women in Roman Law and Society (London 1986) 20. 7 J. Evans Grubbs, Women and the Law in the Roman Empire: A Sourcebook on Marriage, Divorce, and Widowhood (Oxford/New York 2002) 38. A similar sug- gestion is made by Sheridan, BASP 33 (1996) 130–131. 8 Sheridan, BASP 33 (1996) 130 n.11; Vandorpe and Waebens, in Tra- dition and Transformation 420. BENJAMIN KELLY 109 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 As it happens, what has become the mainstream interpre- tation of P.Oxy. XII 1467 was challenged within a decade of its initial publication by Solazzi, Burger, and Roos.9 Burger and Roos pointed out that the petition does not actually ask for the grant of the ius liberorum, but only for the lodgment of the document in the office of the prefect. Moreover, the prefect’s subscription is simply a statement that the petition has indeed been so lodged. Burger went on to speculate that, instead of being a request for the grant of a right, the petition was an at- tempt by Aurelia Thaisous to create a piece of lasting testimony to the existence of her children. This would be especially useful in the event that one or more of them died or was absent and their mother’s entitlement was challenged.10 These suggestions then sank almost without a trace, and have been ignored by most Roman social historians, especially in the English-speaking world.11 This paper, therefore, is partly an attempt to save this earlier position from the overwhelming condescension of posterity. This does not imply any criticism of scholars who have seen Aurelia Thaisous as adhering to a compulsory application regime: it is a reasonable default assumption that sane people will avoid doing administrative paperwork unless compelled. Moreover, Solazzi, Burger, and Roos had no access to the parallel documents that they needed 9 Solazzi, Scritti II 230–231; A. Burger, review of Meyer, Jur.Pap., Zeit- schrift für vergleichende Rechtswissenschaft 39 (1921) 296–311, at 303–306; A. G. Roos, “Animadversiones in papyros nonnullas graecas,” Mnemosyne 51 (1923) 415–420, at 418–420. 10 Burger, Zeitschrift für vergleichende Rechtswissenschaft 39 (1921) 304–305; cf. Solazzi, Scritti II 231. 11 There are a few cases in which scholars have noted with approval the suggestions of Solazzi, Burger, and/or Roos: Sijpesteijn, Aegyptus 45 (1965) 176–177; R. Haensch, “Die Bearbeitungsweisen von Petitionen in der Provinz Aegyptus,” ZPE 100 (1994) 487–546, at 506; C. Fayer, La familia romana: aspetti giuridici ed antiquari II Sponsalia, matrimonio, dote (Rome 2005) 593 n.1038. Since these scholars were focused on other issues, they under- standably did not seek to provide further support for these early suggestions against the interpretation that now prevails. 110 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 to make their position truly cogent. But in this paper I contend that enough parallel documents have now been published to show that Aurelia Thaisous was exploiting an existing petitioning procedure that aimed to create a public record of forensically relevant facts and was used by people who foresaw that sooner or later they could possibly become involved in litigation. Furthermore, the publication of papyri over the last century has made manifest another problem that they could not see: why are there no parallels for women using this petitioning procedure to evidence the ius liberorum? This paper will suggest a reason for this lack of a precise parallel—and at the same time provide a possible explanation for the apparently gratuitous references to literacy in the text. 2. Legislative architecture and the ius liberorum The ius liberorum was established by the Augustan marriage laws, the Lex Iulia de maritandis ordinibus and the Lex Papia Pop- paea. At the outset, it is worth asking whether it is intrinsically plausible, in view of what we know generally about that legis- lation and its impact, that there was a requirement for women to apply to a provincial governor or some other authority be- fore qualifying for or exercising the right. The answer would seem to be in the negative for three reasons. First, there is no trace in Roman juristic writings of such an application procedure.12 If there was an intermediate step be- tween the birth of the requisite number of children and the enjoyment of the privilege, we would expect the legal texts to mention it. Because the guardianship of adult women had been abolished at some point in late antiquity—it is not clear when13 —few pertinent discussions of exemption from guardianship in earlier juristic texts and imperial constitutions have been pre- served in the Justinianic collections. However, there is an unambiguous statement in the Sententiae Pauli: “In order to be 12 This point was first noted by Solazzi, Scritti II 230. 13 On this issue see A. Arjava, Women and the Law in Late Antiquity (Oxford 1996) 116–118. BENJAMIN KELLY 111 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 regarded as having acquired the ius liberorum, it is sufficient for freeborn mothers as well as freedwomen who are Roman citi- zens to have given birth three or four times respectively, so long as they give birth to live children and at full term.”14 Fur- thermore, in discussing the cases of freedwomen and freeborn Latins, the Sententiae go on to state plainly that the ius liberorum crystalizes upon the birth of the requisite number of children.15 Exemption from guardianship was not the only legal benefit accruing from fertility.16 Both women and men qualified for other privileges under the Augustan marriage legislation if they produced children. In none of the discussions of these other privileges in juristic writings or imperial constitutions is there a sign that the lex Iulia de maritandis ordinibus or the lex Papia Poppaea imposed an application requirement. Since it was not subjected to post-classical tampering, Gaius’ Institutes is the most important text in this connection. Gaius discusses the testamentary rights that freedmen obtained through siring children (3.42), and rights of heirs or legatees in wills who had children to take caduca (i.e. legacies that had, for whatever reason, not been taken by their intended recipients: 2.206– 14 Paulus Sent. 4.9.1 (ed. Krueger): matres tam ingenuae quam libertinae cives Romanae, ut ius liberorum consecutae videantur, ter et quater peperisse sufficit, dummodo vivos et pleni temporis pariant. Cf. Burger, Zeitschrift für vergleichende Rechtswissen- schaft 39 (1921) 305. 15 Paulus Sent. 4.9.7–8: libertina ut ius liberorum consequi possit, quater eam peperisse ut ingenuam sufficit. Latina ingenua ius Quiritium consecuta si ter peperit, ad legitimam filii hereditatem admittitur: non est enim manumissa. On the textual diffi- culties in these sections see B. Kübler, “Über das Ius liberorum der Frauen und die Vormundschaft der Mutter: Ein Beitrag zur Geschichte der Rezep- tion des römischen Rechts in Ägypten,” ZSav 30 (1909) 154–183, at 161– 164. 16 Here I focus on privileges unequivocally attributed to the Augustan laws, and which are discussed in juristic sources. For full discussions of all the privileges connected with the ius liberorum see M. Zabłocka, “Il ‘ius trium liberorum’ nel diritto romano,” BIDR 30 (1988) 361–390; S. Treggiari, Roman Marriage: Iusti Coniuges from the Time of Cicero to the Time of Ulpian (Ox- ford 1991) 66–77. 112 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 207). He nowhere hints that the ius liberorum in these cases was contingent on an application process: the privilege is said to emerge simply once the children have been born. He would surely have mentioned an application to the authorities if it were legally necessary.17 This adds plausibility to the impres- sion given by the Sententiae Pauli that release from guardianship depended purely on the number of children that a woman pro- duced. This privilege was, after all, part of the same Augustan legislative framework. A second consideration makes it unlikely that women were required by law to petition to record the fact that they had qualified for the ius liberorum before they could exercise the right. Two Augustan statutes, the lex Aelia Sentia and the lex Papia Poppaea, established a system for the registration of citizen children.18 The child’s father, mother, or grandfather (or their agent) had to go in person to the templum Saturni in Rome or the provincial tabularium publicum, and make a declaration (professio), which was entered into the public records.19 By virtue of the fact that a child’s mother was mentioned in the professiones, the 17 This supposition is strengthened by the fact that Gaius is careful to mention an application regime created by another Augustan law, the lex Aelia Sentia: Junian Latin freedmen would obtain Roman citizenship if they had a child who lived to the age of one, but they had first to appear before the praetor or a provincial governor to prove the key facts (Inst. 1.29, 31; cf. [Ulp.] Regularum liber singularis 3.3). 18 For an outline of the different schools of opinion and earlier literature on birth registrations see F. Schulz, “Roman Registers of Births and Birth Certificates,” JRS 32 (1942) 78–91 and 33 (1943) 55–64; R. Haensch, “Das Statthalterarchiv,” ZSav 109 (1992) 209–317, at 283–290; C. Sánchez- Moreno Ellart, Professio Liberorum: Las declaraciones y los registros de nacimientos en derecho romano (Madrid 2001), and “Notes on Some New Issues Concerning the Birth Certificates of Roman Citizens,” JJurP 34 (2004) 109–119; G. Geraci, “Le dichiarazioni di nascita e di morte a Roma e nelle province,” MEFRA 113 (2001) 675–711, at 675–696. 19 On the procedure for illegitimate citizen children, which seems to have been slightly different until the time of Marcus Aurelius, see Sánchez- Moreno Ellart, Professio Liberorum 167–171, and JJurP 34 (2004) 108–110, citing earlier literature. BENJAMIN KELLY 113 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 procedure created a public record of how many children a woman had borne. Evidencing the ius liberorum has therefore plausibly been seen as one of the goals of the Augustan birth registration regime.20 This creates a problem for the theory that mothers were required to apply for the ius liberorum, a problem that has apparently never been noticed: why would women have been required to apply for the ius liberorum if the fact that they had produced the requisite number of children was already a matter of public record? It should be noted that it is not entirely clear whether this process was compulsory. The case for its being compulsory rests on a passage in the Historia Augusta that is clearly faulty in other respects, since it has Marcus Aurelius rather than Au- gustus initiating the whole system.21 This does not inspire con- fidence. The question does not have to be decided here. The critical point for us is this: the legislator would have built point- less redundancy into his scheme if women who had already opted to register their children were also compelled to petition to record the fact that they qualified for the ius liberorum. One would surely have to hypothesize that applications like P.Oxy. XII 1467 were not part of a compulsory process, but a fallback measure for women whose children had not been properly registered at birth. Of course, if the Augustan legislation made birth registration compulsory for legitimate citizen children, then even this hypothesis is untenable: the lawgiver is hardly likely to have provided an alternative procedure to assist people who had broken the law. A final consideration suggests that a compulsory petitioning regime did not exist. If women were obliged to petition before they qualified for the ius liberorum or exercised it, it is strange 20 J. Gardner, “Proofs of Status in the Roman World,” BICS 33 (1986) 1– 14, at 3; Zabłocka, BIDR 30 (1988) 375; Treggiari, Roman Marriage 76; Sánchez-Moreno Ellart, Professio Liberorum 21. 21 HA Marc. 9.7–8. On this passage see Schulz, JRS 32 (1942) 80–82; Haensch, ZSav 109 (1992) 283–284; Sánchez-Moreno Ellart, Professio Li- berorum 30–39. 114 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 that only one example of such petition is known. This is an ar- gument from silence, but an unusually weighty one. Petitions and applications have survived in impressive quantities on papyri; a search of the Synallagma database for the period between A.D. 212 (when most women in Egypt received Roman citizenship and therefore became eligible for the ius liberorum) and the end of the fourth century (by which time the right was clearly no longer relevant) produces records for 556 different petitions and applications.22 In this period, very significant numbers of women must have qualified for the right. From the census returns it is evident that in Roman Egypt over half of women had married by their late teens, and marriage for women was nearly universal.23 The Total Fertility Rate has been estimated at 5.979 children of both sexes on average for women who survived to 50.24 Saying that such women on average had almost six children does not, of course, tell us what percentage of women had three or more children. A very rough minimum percentage can be obtained from the census declarations in Bagnall and Frier’s catalogues25 and the census register P.Oxy.Cens.26 Combined, these attest 159 freeborn women whose ages are recorded as or can be inferred to be 25 or over,27 and who appear in a reasonably 22 Synallagma.tau.ac.il/?project=glrt&username=guest&password=guest (accessed 8 August 2016). 23 R. S. Bagnall and B. W. Frier, The Demography of Roman Egypt 2 (Cam- bridge 2006) 111–116. 24 Bagnall and Frier, Demography 138–139. 25 Bagnall and Frier, Demography 179–323. 26 R. S. Bagnall, “Households,” in R. S. Bagnall, B. W. Frier, and I. C. Rutherford, The Census Register P.Oxy. 984: The Reverse of Pindar’s Paeans (Brus- sels 1997) 57–88. 27 Persons who were sui iuris and under 25 required a curator. Women under 25 with three children would therefore still have needed a curator to oversee their legal transactions, even if the ius liberorum technically freed them from needing a tutor. For this reason, if an application was required, there would have been little point in women applying for the ius liberorum until they reached 25. BENJAMIN KELLY 115 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 complete return in which the family relationships are clear. Of these women, 48 (i.e. just over 30%) are attested as having three or more children. It must be stressed that this figure is much too low: children, especially young children, are under- reported in census returns;28 many women would have pro- duced children who then died before the next census; many adult children would have lived separately from their mothers; and some divorcées (and perhaps also widows) lived separately from their young children.29 Let us however take the figure of 30% as an absolute minimum that serves to illustrate the lower limits of the possible. To illustrate rough orders of magnitude, let us use Bagnall and Frier’s reconstruction of the age structure of the Egyptian population, which (conservatively) assumes a total population of 4.5 million.30 This is very much a hypothetical model; as its creators stress, it has some problematic aspects, such as the heavy skew in the sex ratio in favor of men. Nevertheless, it will serve for illustrative purposes. This model suggests a population of 2,065,312 females, of whom 934,174 were 25 or over. This number needs to be reduced to take into account the fact that a proportion of these women would have been slaves, and hence not eligible for the ius liberorum. Firm data on slave numbers is scarce, but extrapolating from the census returns, a reduction of 11% might be in order, giving a notional free female popu- lation of 831,414.31 As we have seen, a bare minimum of 30% 28 Bagnall and Frier, Demography 41–42, 81, 97–99, 152–153. 29 Bagnall and Frier, Demography 124–125; S. Huebner, The Family in Roman Egypt: A Comparative Approach to Intergenerational Solidarity and Conflict (Cambridge 2013) 101–105, 166. 30 Bagnall and Frier, Demography 104. For literature on the overall size of the Egyptian population see W. Scheidel, “Progress and Problems in Roman Demography,” in Debating Roman Demography (Leiden 2001) 1–81, at 57–59. 31 This percentage is extrapolated from Bagnall and Frier, Demography Tables A and D, which respectively show that 161 females in the census re- turns are identifiably 25 or over, of whom 17 (10.6%) are slaves. 116 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 of these can be assumed to have had three or more children, which means that any given time between 212 and the early fourth century, at the very least around a quarter of a million women would have qualified for the ius liberorum. If it was compulsory to petition to qualify for or exercise this right, it is astonishing that only one such petition survives. Nor can one plausibly hypothesize that women who qualified for the ius liberorum generally declined to submit the paperwork necessary to exercise it. In the third decade after the Constitutio Antoniniana, women in the papyri claiming the right to act without a guardian by virtue of the ius liberorum started to out- number heavily those still acting with a Roman-law guardian: Arjava has recently calculated that in the period 236 to 310 something over fifty women are found claiming exemption from Roman-law guardianship thanks to the ius liberorum;32 on the other hand, only twelve cases of women acting with Roman-law guardians are attested in documents from the same period.33 Qualified women clearly did often avail themselves of the right.34 Thus, given the lack of other known petitions for the ius liberorum from women in Egypt, it is very unlikely that a petition 32 It should be mentioned that nobody would seriously claim that the survival of around 50 documents involving women with the ius liberorum should mean that we would expect around 50 applications to have survived, if it were compulsory to apply for the right. Different genres of papyrus documents have different rates of survival. Moreover, a woman would (hypothetically) make only one application for the ius liberorum during her lifetime, but might potentially act without a guardian in multiple trans- actions. However, there is simply no statistical data of the sort that would allow us to say that the average woman acted independently in, for example, two, four, eight, or more such transactions. 33 Arjava, in Law and Legal Practice 178–179. 34 The scale of the disparity between the number of women with the ius liberorum and the number with guardians also underlines the point that 30% is an unrealistically low estimate: the percentage of women aged 25 and over who had produced three or more children must have been very much higher than this. BENJAMIN KELLY 117 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 was required to qualify for or exercise the right. This con- clusion makes it unlikely that the Augustan legislation provided for a compulsory application process; likewise, it rules out any attempt to hypothesize that a local application regime was in- troduced in Egypt at some point, for example in the mid-third century A.D. In a recently published sourcebook, Arjava was apparently prompted by the lack of known parallels to wonder whether women were legally required to apply for the ius trium liberorum. Nevertheless, he still entitled P.Oxy. XII 1467 “Re- quest for the ius liberorum.”35 3. Registration petitions If P.Oxy. XII 1467 is not the result of a legal requirement to petition to qualify for or exercise the ius liberorum, what is it? It is my contention that sufficient documents have now been pub- lished to give compelling support to the suggestion made by Burger almost 100 years ago, that the goal of the petition of Aurelia Thaisous was to evidence the fact that she had borne the required number of children. In Roman Egypt there was a practice of submitting petitions not in order to commence litigation immediately or obtain some administrative decision, but to create a timely record of facts that would potentially be of later forensic use.36 Over thirty of these petitions addressed to the strategos have been published to date. Typical of these requests to the strategos to register the petition (i.e. deposit the document in some kind of archive) is BGU I 35, a complaint from A.D. 222 that the 35 Arjava, in Law and Legal Practice 186: “It is not likely … that such an official registration was legally required from women if they wished to ap- pear without guardians. At least this is the only specimen of its kind.” 36 This paragraph and the following three present in summary form the findings of my study “Petitions with Requests for Registration from Roman Egypt,” in R. Haensch (ed.), Recht haben und Recht bekommen im Imperium Romanum (Warsaw 2016) 407–456. Readers will find there a very detailed discussion of the ‘registration petitions’ and an analysis of earlier attempts to understand the function of these documents. A full list of registration peti- tions is in the Appendix to that article. 118 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 petitioner’s cow had been slaughtered illicitly by unknown attackers, which ends with the request “wherefore I submit this petition and ask that it be registered so that my right [or ‘account’] may remain against the culprits when they come to light.”37 One finds similar requests for the registration of the document in petitions directed to a nomarch,38 a centurion,39 a logistes,40 a katholikos,41 a prefect,42 and a board of officials that John Rea has plausibly suggested were the nyktostrategoi of Her- mopolis Magna.43 The petitions that request registration con- cern a wide variety of subjects, including disputes over liturgies and the ownership of slaves, and complaints about thefts, as- saults, and acts of property damage. It is clear that this pro- cedure was not limited to any particular kind of case. The purpose of this process of registration is made evident by three features of the documents. First, in making their requests the writers of some of the petitions say that they want the document registered for the purpose of testimony, or say that the document itself is evidence. Thus in P.Ant. II 88.11–12 (A.D. 221) one finds ἀξιῶ αὐτ̣ὸ̣ τοῦτο µαρτυρόµενος [καὶ ἐπιδιδ]οὺς τόδ[ε τ]ὸ βιβλίδιον εἶναι α[ὐ]τὸ ἐν καταχωρισµῷ πρὸς µ[αρ]- τύριον (“Testifying to this very fact and submitting this petition 37 BGU I 35.10–15: ὅθεν ἐπιδίδωµι τάδε τὰ βιβλίδια καὶ ἀξιῶ ἐν κατα- χωρισµῷ γενέσθαι πρὸς τὸ µέν<ε>ιν µοι τὸν λόγον πρὸς τ[ο]ὺς φανησο- µένους αἰτίους. 38 P.Fam.Tebt. 38 = SB IV 7363 (A.D. 168). 39 BGU II 651 = Chrest.Mitt. 111 (A.D. 192). On petitions to soldiers generally see M. Peachin, “Petition to a Centurion from the NYU Papyrus Collection and the Question of Informal Adjudication performed by Soldiers,” in A. J. B. Sirks and K. A. Worp (eds.), Papyri in Memory of P. J. Sijpesteijn (P.Sijp.) (Oakville 2007) 79–97. 40 P.Oxy. LXI 4122 (A.D. 305). 41 P.Ammon II 30.1–17 = SB XIV 11929 = P.Ammon I 6 (A.D. 348). 42 SB XXII 15608 = P.Stras. VI 560 (A.D. 324 or after). 43 SB XX 15036 = CPR I 232 (second half of the third century); cf. J. R. Rea, “On κηρυκίνη: P.Heid. IV 334, P.Köln VI 279, and CPR I 232,” ZPE 79 (1989) 201–206, at 203, 205. BENJAMIN KELLY 119 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 I ask that it be registered for the purpose of testimony”). The phrase πρὸς µαρτυρίαν (or µαρτύριον) (“for the purpose of tes- timony”) also appears in a similar context in two other petitions from this period, and has been plausibly reconstructed in a third.44 Four other cases lack the phrase πρὸς µαρτυρίαν (or µαρτύριον), but express the same idea: a petition of 218 to the strategos uses a participal, [δι]ὸ̣ τοῦτο µαρτυρό[µενος] (“there- fore bearing witness to this”: PSI III 249.17); another to the strategos, of 224, calls itself an ἐνµάρτυρα (“a piece of testi- mony”: P.Fouad 29.13); the registration of the petition in 305 addressed to the logistes is said to be desired µαρτυρίας … ἕνε̣κεν (“for the sake of testimony”: P.Oxy. LXI 4122.16–17); and the petition from the second half of the third century, perhaps addressed to the nyktostrategoi of Hermopolis, is called a διαµαρτυρία (“a written testimony”: SB XX 15036.6). The sense in which registration petitions were useful evi- dence is clarified by a second feature: as far as we can tell, they were invariably submitted very quickly after the wrong of which they complain. Sometimes, indeed, they even appear to be submitted merely in anticipation of a wrong. The speed of submission can be determined in cases where both the date of a wrong and the date of submission appear, or when the text says that a wrong occurred “yesterday” or the like. In all the cases in which the interval can be determined with precision, that interval is very short: never more than nine days, and usually between one and three.45 A third feature is relevant as well: it is quite often evident from the narrative that the petitioner an- ticipated some kind of delay in bringing the case to court. Most commonly, the problem is that the identity of the petitioner’s opponent is unknown, as was the case in the complaint about the slaughtered cow (BGU I 35). Other factors could cause peti- 44 BGU VII 1577.4 πρὸς µ ̣[α]ρ[τ]υ[ρία]ν (A.D. 199–209; addressee lost); P.Fam.Tebt. 38.15 = SB IV 7363 (A.D. 168; nomarch); P.Ryl. II 116.18 = Jur. Pap. 92 = Sel.Pap. II 287 = C.Pap.Hengstl 50 (A.D. 194; strategos). 45 For full references see Kelly, in Recht haben und Recht bekommen 420 n.49. 120 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 tioners to anticipate delay, however. For instance, a wrongdoer whose identity was known, but who had disappeared, could cause such an apprehension on the part of a petitioner. The procedure was also used by people who feared that they might potentially become defendants in a lawsuit—an eventuality whose timing they obviously could not predict. The petitioners in the Hermopolite document discussed by Rea are clearly preparing for such a possibility (SB XX 15036). In SB XXII 15608 = P.Stras. VI 560, a man petitions the prefect complain- ing that a soldier has extorted from him a false written ac- knowledgement of a debt. He asks for a court order for the return of the extorted document. But he also requests the regi- stration of the petition, presumably fearing a scenario in which the soldier fails to attend court but then later tries to enforce the acknowledgement of the debt. The purpose of petitions with requests for registration sub- mitted in Aurelia Thaisous’ era is therefore clear. They were attempts by people who anticipated litigation in the future to put their version of the facts on record. The evidentiary value of this, as far as the courts were concerned, would have been that this was done at an early opportunity. This would have allowed plaintiffs to escape the accusation that they had been tardy in airing their version of the facts, and were probably therefore lying. This kind of argument based on an opponent’s delay was a stock feature of Attic oratory, visible in Demos- thenes and Lysias;46 it also surfaces in Cicero.47 It then made its way into the legal culture of Roman Egypt, as one would expect, given the strong influence of classical rhetoric in the courts of the province.48 Thus, for example, a prefectural edict of A.D. 142 decrees that debtors, if they have allegations of 46 Lys. 3.39, cf. 19–20; 7.42, cf. 17; 13.83–84; Dem. 18.15; 36.53–54. 47 Cic. Cael. 19. Note too Dig. 47.10.11.1. 48 See especially M. Heath, “Practical Advocacy in Roman Egypt,” in M. Edwards and C. Reid (eds.), Oratory in Action (Manchester 2004) 62–82, and Menander: A Rhetor in Context (Oxford 2004) 21, 311–331. BENJAMIN KELLY 121 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 fraud or forgery to make against their creditors, must raise these as soon as their creditors demand repayment. If they delay, they will lose their right to bring criminal charges al- together, or at least have to wait until the action for debt has been decided.49 This provision is clearly provoked by the suspicion that delay is evidence of bad faith and dishonesty in a litigant. In another papyrus, a report of proceedings, we see a defense advocate in a debt case attacking the credibility of the plaintiff on the grounds that he had waited until the death of the (alleged) debtor before bringing a case against his heirs for what he claimed was a long-overdue debt.50 Petitions with requests for registration were submitted at different stages of the process of disputing and litigation in order to evidence material facts. In a few cases, a dispute was at an advanced stage and litigation imminent. In two instances, the petitioner’s opponent had failed to appear in court at the appropriate time;51 in two others the petitioners wanted a court hearing and apparently had the petition registered to deal with the possibility that their opponents would fail to attend.52 More commonly, however, litigation was not imminent, but rather was a hypothetical possibility for which the petitioner was pre- paring. In twelve cases from the second and third centuries, the petitioners did not yet know the identities of the people who had wronged them, but evidently held out hope of someday discovering them.53 For example, in a petition from Oxyrhyn- chus dating to 258/9 the petitioners claim that papers and cer- tain other articles had been stolen from their deceased father’s 49 P.Oxy. II 237.viii.7–18, BL I 318. 50 P.Mil.Vogl. I 25.iii.12–14 (A.D. 127). 51 P.Ant. II 88; P.Ammon II 30. 52 BGU I 242 = Chrest.Mitt. 116 (A.D. 187–188); SB XXII 15608. For fur- ther discussion see Kelly, in Recht haben und Recht bekommen 441. 53 BGU I 35; 46 = Chrest.Mitt. 112; 72; II 651 = Chrest.Mitt. 111; III 731.ii; P.Fouad 29; P.Harr. II 200; P.Mich. IX 527; P.Oxy. XLI 2997; XLVI 3289; L 3561; P.Tebt. II 330 = Chrest.Mitt. 110. 122 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 house. They request the registration of the petition in case they discover the identity of the perpetrator in the future.54 There are also cases in which it is not even clear from the document that a wrong had been committed or that a dispute had emerged, and actual litigation was a distant possibility that lay over the horizon, so to speak. The petitioners nevertheless considered it prudent to register their version of the facts in case litigation did eventuate. Thus, in one example, the peti- tioner and his brother had taken over the cultivation of land previously leased by their deceased father. The petition states that they have now withdrawn from the land, having paid all the necessary dues. There is no mention in the document that there is a dispute with the owner of the land; nevertheless, the cultivators want the petition containing their factual claims registered by the strategos, apparently to protect them in the event of later litigation about the precise time and circum- stances of the withdrawal.55 In another petition, this one from the Oxyrhynchite and dating to 214, the overseer on the estate of Claudia Isidora, alias Apia, reports that some farm equip- ment was damaged in a fire, and asks for the document to be registered. The petitioner clearly did not know how the fire started, and the language of the document does not commit to the view that it was arson, but remains agnostic. The purpose of the registration was most likely to support a lawsuit if the blaze proved to have been deliberately lit and an offender was apprehended.56 A petition from two years later deals with a similarly hypothetical scenario: the petitioner’s father and her brother had set off hunting twenty-three days previously and had not been heard from since. The petitioner clearly suspects 54 P.Oxy. XLVI 3289. 55 PSI I 57, BL XI 242 (A.D. 51). For the goals of this document see too D. P. Kehoe, “Legal Institutions and the Bargaining Power of the Tenant in Roman Egypt,” ArchPF 41 (1995) 232–262, at 241–242. 56 P.Oxy. XLI 2997 = BASP 6 (1969) 55–56. Note especially the non- committal request formula: ὅθεν ἐπιδίδοµι (l. ἐπιδίδωµι) τὸ βιβλίδιον ἀξιῶν εἶναι αὐτὸ ἐν καταχωρισµῷ µὴ ἄρα τι ὕστερον ἀναφανῇ (lines 15–18). BENJAMIN KELLY 123 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 some kind of foul play, but does not seem to have any firm facts about what happened, let alone know who (if anyone) was in- volved. She mentions that she has already submitted a version of the petition to the strategos to be registered by him (πρὸς τ̣ὸ̣ ἐν καταχ[ωρ]ισµῷ γενέ[σ]θαι), and, although it is not said ex- plicitly, the goal of submitting the extant document to the centurion is probably the same.57 Again, then, the registration of petitions can be seen to be aimed at a hypothetical future need, and was submitted at a point when a dispute had not clearly emerged and when litigation was far from imminent. It is my contention that the petition of Aurelia Thaisous should be understood in light of this practice of submitting petitions for registration to create timely testimony of facts that would potentially be of future forensic use. The concrete act that she requests is that the petition should be deposited ἐν τῇ σῇ τοῦ [δια]σηµοτάτου τ[ά]ξ<ε>ι (“in your eminence’s office”); the phrase ἐν τῇ τάξει with an appropriate verb of deposit ap- pears in two other registration petitions, once to refer to the registration of a petition in the office of the prefect and once in the office of the katholikos.58 It should also be noted that in none of the registration petitions is there any hint that additional evidence was appended to the document, or that the official re- ceiving it made some kind of preliminary ruling on it. Rather, the value of the claims made in the petition would be that they were made as early as possible, and that they would be consis- tent with whatever additional evidence was actually presented at trial, if one eventuated. There is an explicit hint in P.Oxy. XII 1467 that there is no expectation that the prefect make any ruling on the factual veracity of the claim about the fecundity of Aurelia Thaisous: the request is that the petition should be archived in the prefect’s office ἀπροκρίτως τ̣ο̣[ῖς δι]καίοις µ[ο]υ, a phrase that should be translated “without my rights having been prejudged.” As Paul Meyer astutely recognized, 57 P.Tebt. II 333 = Chrest.Mitt. 115 = Sel.Pap. II 336 (A.D. 216). 58 SB XXII 15608.9; P.Ammon II 30.14. 124 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 there is a conceptually analogous formula in birth declarations for Roman citizens, citra causarum cognitionem, which was de- signed to indicate that the prefect had not enquired into the veracity of the professio, but had taken it at face value. The entry of the professio in the records created a presumption as to its veracity, but this could be rebutted.59 The critical fact that Aurelia Thaisous wanted to put on record at an early date was that she had borne the requisite number of children to qualify for the ius liberorum. It might strike a casual reader of the petition as odd that she does not mention the names of these children and their paternity, but these were not legally relevant facts.60 Besides, if there ever was a dispute over the petitioner’s entitlement to the ius liberorum and a judge had to make a ruling, specific details could be given then. What mattered for the moment was that she had recorded her claim, and had done so before entering into any transactions for which a guardian was legally required. In a world with no genetic testing, people entitled to the ius liberorum could produce the requisite number of children in court, but would have trouble conclusively dismissing the claim that these were not actually their children.61 Moreover, as Burger rightly pointed out, producing the children in person at some point in the future might even be impossible: one or more of them 59 Meyer, Jur.Pap. p.10. See too Schulz, JRS 32 (1942) 87, for the mean- ing of citra causarum cognitionem and FIRA III pp.72–73 n.2 for further discus- sion of lines 21–24 of the petition. 60 The absence of specific information about the children in P.Oxy. XII 1467 is a more serious objection to the theory that one had to apply to obtain the ius trium liberorum: surely the prefect would want some assurance that the children exist before making a definitive ruling, and specific details might reassure him. This absence of specifics no doubt has prompted the suggestion that there must have been other documents accompanying the petition (see 108 above). But there is no hint of this in the text. 61 Cf. Cod.Theod. 12.17 praef. (A.D. 321 or 324), which responds to cases in which doubts have been cast on the paternity of children allegedly fathered by men who were using the ius liberorum to gain exemption from compulsory public services. BENJAMIN KELLY 125 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 might be absent or dead, for instance. In difficult situations in which conclusive proof of a material fact from documents or witnesses was unavailable, Roman courts were often swayed by what have been called “global proofs”: i.e. general indications that a party was honest, of high social standing, or had consis- tently acted in good faith.62 If her maternity was challenged at a later date, therefore, Aurelia Thaisous evidently hoped that a court would be swayed by the fact she had consistently asserted the existence of her three children from an early date—even before she entered into the disputed transaction. One scenario that she perhaps feared was making a contract with someone who later attempted to argue that the contract was unenforceable as a guardian’s consent had not been given. There is also the possibility that she feared that a third party (for instance, a relative) would try to challenge some kind of legal instrument (for instance, a will) on the grounds that a guardian had not given his blessing to it. It is possible that she had a very specific individual in mind who was likely to cause her trouble. The text does not mention this, but it would have been irrelevant to the specific request that she makes—and perhaps even counterproductive to her long-term credibility— to rehearse her specific fears in this document. On the other hand, P.Oxy. XII 1467 could be like the cases discussed above in which a petitioner took steps to prepare for quite hypo- thetical future litigation by registering her factual claims before a dispute and litigation became imminent. 4. An unorthodox solution The suggestion that P.Oxy. XII 1467 was designed to have evidentiary value returns us to an earlier question: Why is this the only specimen of such a petition? This woman was surely not alone in needing a firm evidentiary basis for her claim to the ius liberorum. The answer to this question lies in the fact that 62 G. Pugliese, “La preuve dans le procès romain de l’époque classique,” in La preuve I Antiquité (Rec. Soc. Jean Bodin 16 [Brussels 1964]) 277–348, at 298. 126 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 there were three (or perhaps four) other ways in which a mother in Roman Egypt in the third century could prove the maternity of her children if challenged, so presumably most women used these in the event of a dispute. However, we can see clear reasons why a mother in the early 260s would not have been able to employ several of these—and thus might have resorted to the unorthodox tactic of submitting a registra- tion petition. The first and perhaps most conventional method that a mother might use to support a claim for the ius liberorum was the Roman system of birth declarations: as we have seen (112–113 above), this was apparently one of the purposes of these declarations. There are, however, serious doubts about the extent to which this system was operational in Egypt in the mid-third century.63 In the years immediately prior to the petition of Aurelia Thaisous in 263, there was at least one period in which approaching the prefect to register a birth would have been physically difficult, if not impossible. In late 261 the prefect L. Mussius Aemilianus attempted to supplant the emperor Gallienus.64 According to the Historia Augusta, the 63 It is perhaps no accident that a Tebtynis papyrus preserves an imperial constitution of A.D. 239 about the legal consequences of failing to register children: P.Tebt. II 285 = Chrest.Mitt. 379, BL III 241, IX 355. The text was found with a bundle of other family papers, and Arthur Verhoogt has made the attractive suggestion that some of them relate to a family inheritance dispute in the 260s in which the failure to register a potential heir led to questions about her legitimacy: A. M. F. W. Verhoogt, “Family Papers from Tebtunis: Unfolding a Bundle of Papyri,” in A. M. F. W. Verhoogt and S. P. Vleeming (eds.), The Two Faces of Graeco-Roman Egypt: Greek and Demotic and Greek-Demotic Texts and Studies presented to P. W. Pestman (Leiden 1998) 141– 154, at 147–151. 64 For this incident and its chronology see especially J. Schwartz, “L. Mussius Aemilianus, préfet d’Egypte,” BSRAA 37 (1948) 34–46; D. Kienast, Römische Kaisertabelle (Darmstadt 1990) 224–225; A. M. Demicheli, Rapporti di pace e di guerra dell’Egitto romano con le popolazioni dei deserti africani (Milan 1976) 136–139; J. R. Martindale, “Prosopography of the Later Roman Empire: addenda et corrigenda to Volume I,” Historia 23 (1974) 246–252, at 246–247. BENJAMIN KELLY 127 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 latter dispatched Aurelius Theodotus, who invaded Egypt, de- feated Aemilianus in battle, captured him, and sent him to Gallienus for execution.65 There are hints of fighting in both Alexandria and Hermopolis Magna.66 The usurpation was over by 30 March 262 at the latest, at which time Gallienus is recognized again as emperor in dating formulae.67 During such a period of civil war, it is hard to believe that the usual birth declaration system functioned. Even when Egypt was not embroiled in civil war, the system of Roman birth registrations must have been strained to the breaking point by the mass enfranchisement of the Egyptian population in 212. We can obtain a sense of the scale of this strain if we assume as the annual birthrate 44 per 1000, the rate that Bagnall and Frier suggest as part of their theoretical reconstruction of the Egyptian population.68 If we again use the low estimate of 4.5 million for the total population, the result is 198,000 births per year, which we might lower by 11% to just over 176,000 to reflect the fact that some children were born as slaves and therefore could not be registered as Roman citizens (see 115 above). It is scarcely any wonder that the last known tabula recording a professio liberorum from Egypt dates to 242.69 65 HA Gallieni duo 4.1–2, 5.6, Tyr.Trig. 22.1–8; cf. Epit. de Caes. 32.4. 66 Alexandria: Schwartz, BSRAA 37 (1948) 43–46; Hermopolis Magna: M. Drew-Bear, “Guerre civile et grands travaux à Hermoupolis Magna sous Gallien,” in B. Kramer et al. (eds.), Akten des 21. internationalen Papyro- logenkongresses I (Stuttgart/Leipzig 1997) 237–243. 67 P.Stras. I 7.1–2, with D. Rathbone, “The Dates of Recognition in Egypt of the Emperors from Caracalla to Diocletianus,” ZPE 62 (1986) 101–131, at 119. 68 Bagnall and Frier, Demography 105. 69 FIRA III 1 = SB VI 9200 = C.Pap.Lat. 163, BL VIII 341. The format of this document and two others like it (P.Oxy. VI 894, BL VI 98 = Chrest.Wilck. 213 = C.Pap.Lat. 158 = ChLA III 214, and P.Oxy. XXXI 2565 = ChLA XLVII 1412) suggests that from the 190s there was apparently a stream- lining of the process that allowed the professio to be submitted in writing: see Geraci, MEFRA 113 (2001) 690–691, citing earlier literature. But this sub- mission needed to be done in person by a family member (or perhaps an 128 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 One suspects that the system was abandoned altogether; even if it continued, it would have been logistically impossible for most parents to register their children before the prefect. There was a second conventional method that could be used to prove parentage in the absence of any regular records in public archives: witness testimony. There was no requirement for paternity or maternity to be proved by documentary evi- dence only. For example, a constitution of the emperor Probus suggests that the testimony of neighbors (vicini) or others with appropriate knowledge (alii scientes) might assist a man who could not demonstrate his daughter’s paternity because she had not been properly registered at birth.70 One can reasonably assume that some mothers needing to prove the maternity of their children would have been able to rely on similar witness testimony (or unofficial documents like private letters)71 in the absence of official documentary evidence. For reasons that were understandably not outlined in the petition (and therefore are opaque to us), such witness testimony was not available to Aurelia Thaisous, or at least she or her legal advisor(s) deemed it to be potentially inadequate. But it is not unreasonable to assume that some women were unable to produce adequate witness testimony, for instance if they were trapped in acri- monious relationships with their families or neighbors, or socially isolated in some other way. In the absence of Roman birth registrations or reliable wit- ness testimony, what else might a woman do to prove her maternity? A third option might be to use the house-by-house declarations of the provincial census. These were archived by ___ agent) and still had to be processed by the prefect’s staff, which surely would have been an impossibility if there were ca. 200,000 births a year. 70 Cod.Iust. 5.4.9. See too Cod.Iust. 7.16.15; P.Tebt. II 285 = Chrest.Mitt. 379. Note that in cases where the free status of someone was at issue, wit- ness testimony alone was apparently not enough: Cod.Iust. 4.20.2 (A.D. 223). 71 See Dig. 22.3.29 praef., an imperial constitution showing in cases of disputed parentage a preference for witness testimony to be supplemented by documents such as private letters, if these exist and seem reliable. BENJAMIN KELLY 129 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 the authorities and copies were kept by families. The parents of each individual in the house were recorded in the declarations. We know that as a result of this practice of recording paren- tage, census declaration were sometimes used to prove legally important facts concerning individuals such as parentage and status.72 Whilst house-by-house declarations were not intended to evidence the ius liberorum, it is not impossible that they could have been so used. The problem for Aurelia Thaisous, how- ever, was that the most recent census year had been in 257/8, which ended almost five years before. Thus, if any of her children had been born subsequently, their existence would not be evidenced in census documents. This brings us to a fourth and final method one could con- ceivably use to prove the maternity of one’s children. Aside from the birth registration system for Roman citizens, there was also a local, Greco-Egyptian system for registering children that parents could opt to use—although not necessarily at birth.73 This system continued after 212, which meant that Roman citizens began using it. Aurelia Thaisous could, there- fore, have registered her children using this system, even if the Roman birth registration system was difficult to use or defunct in the early 260s. However, although around forty Greco- Egyptian birth returns have survived, these have so far evaded scholarly attempts to explain the rationale for the system. The one certain thing is that the original purpose cannot have been to evidence the ius liberorum, since this system of declaring births 72 Bagnall and Frier, Demography 28–29. A case of census declarations being used to prove parentage (and hence the relationship of two half- siblings) is provided by P.Lond. II 324 = Chrest.Wilck. 208, BL I 245, III 93, X 98 (A.D. 161), discussed by H. C. Youtie, “ἀπάτορες: Law vs. Custom in Roman Egypt,” in Scriptiunculae posteriores I 17–34, at 17–19. 73 For discussion and references to earlier literature see A. Jördens, P. Bingen pp.389–391; C. Sánchez-Moreno Ellart, “ὑποµνήµατα ἐπιγεννήσεως: The Greco-Egyptian Birth Returns in Roman Egypt and the Case of P. Petaus 1–2,” ArchPF 56 (2010) 91–129. 130 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 only came to be employed by Roman citizens after 212.74 Thus, using such birth declarations to evidence one’s right to the ius liberorum would have involved using an administrative procedure for a purpose for which it was not initially employed. Instead, however, Aurelia Thaisous (and perhaps her ad- visor[s]) opted to use another administrative procedure—the registration of a petition—in an unusual way. The precise rationale for using this rather than Greco-Egyptian birth declarations is unclear: perhaps it was simply that one document rather than three would need to be drafted. In short, then, the decision to use a registration petition to evidence the ius liberorum was a somewhat daring use of a pro- cedure generally employed in other kinds of legal disputes. But one can see clear reasons why a woman in the early 260s might not have been able to use Roman birth registrations or house- to-house census declarations. One can also at least suggest plausible reasons why witness testimony and Greco-Egyptian birth registrations were deemed inadequate. Her unorthodox solution was either unique or uncommon, since there were other, more mainstream ways to achieve the same goal, and most women evidently were content to rely on these. 5. Literacy and status Reinterpreting the petition of Aurelia Thaisous as a some- what recherché legal maneuver may help to solve one of the other enduring problems raised by her petition: why does the text stress her literacy so strongly? As mentioned above, this cannot have been a legal requirement for the exercise of the right. More plausible was Herbert Youtie’s suggestion that the petitioner’s literacy would have made the prefect more inclined to grant the ius liberorum to her, owing to the practical advan- tage that literacy had for a woman wanting to conduct her 74 Sánchez-Moreno Ellart, ArchPF 56 (2010) 123–125, also raises and firmly rejects the hypothesis that the Greco-Egyptian birth registration system was reformed in the late-second century to help women to evidence the ius liberorum. BENJAMIN KELLY 131 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 affairs independently.75 But this explanation is only viable if we accept that the grant of the right depended on the prefect’s discretion. As we have seen, this is highly unlikely: Aurelia Thaisous had the ius liberorum from the moment her third child was born. If we are to avoid the assumptions that the reference to literacy was a case of pointless boasting or was inserted be- cause of the legal ignorance of the petitioner (or her advisor[s]), a new explanation is therefore required. A more sympathetic explanation is possible if we consider what literacy would have meant to a high-level imperial ad- ministrator like the prefect of Egypt. For him, the assertion that the petitioner was not only literate (ἐνγράµµατος) but could also write with ease ([ἐ]ς̣ τὰ µάλιστα γράφειν εὐ⟦β⟧ ̀κ̣ ́ό̣π̣ως [l. εὐκόπως] δυναµένη) would have been a coded claim to a high cultural (and also social) level. Illiteracy carried no heavy social stigma in Roman Egypt,76 but it is also true that members of the privileged stratum of the metropoleis would have been much more literate than the general population.77 Moreover, an important imperial administrator like the prefect could be expected to see a kind of cultural kinship between himself and an educated member of the local elite. The strategy of trying to place oneself within the same cultural universe as one’s judge is paralleled in other Roman legal contexts. In his Apologia (dating to 158/9), Apuleius of Madaurus defended himself on charges of magic before the governor of Africa Proconsularis. The speech strives to place Apuleius in the same rarefied cultural milieu as the governor. It is a treasure house of learning, crammed with recondite references ransacked from poetry, philosophy, history, and zoology. As Keith Bradley has argued, the speech is therefore a calculated attempt to appeal to what 75 Youtie, Scriptiunculae posteriores I 199 n.62, approved by Pomeroy, ANRW II 10.1 (1988) 721. 76 Youtie, Scriptiunculae II 620–623. 77 See Youtie, Scriptiunculae II 624–625, on the expectation of literacy amongst males of the gymnasial order. 132 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 Peter Brown has called “a common culture that was held to be the distinguishing mark of the diffused governing class of the empire”78—a culture in theory common to local notables and imperial government personnel.79 Closer to the world of Oxyrhynchus in the 260s is another remarkable document. Probably around 258 or 259, a public grammaticus of the town drafted a petition to the emperors Valerian and Gallienus.80 Although it related to a mundane dispute over remuneration of the petitioner by the metropolis, it is intended as a tour de force of the grammarian’s craft, packed with high vocabulary and complex, learned syntax. The motive behind this performance is made patent in the first of the two versions of the draft: Paideia, writes the petitioner, sits beside the emperors, and this fact emboldens the writer to request their assistance (line 9). The petition, in Brown’s words, represents an attempt “to cling to the great through the delicate osmosis of a shared culture.”81 Tantalizingly, the name of the grammarian was Lollianos, alias Homoios. Lollianos and its feminine equivalent Lolliane are exceedingly rare in the papyri. The editor of the papyrus, Peter Parsons, therefore raises the possibility that he was from the same family as our Aurelia Thaisous, alias Lolliane, in view of their geographical and temporal proximity, and their pride in their literary attain- ments.82 One could add that the tendency in Roman Egypt of names to be passed down through families—sometimes in both a masculine and feminine form—supports this suggestion.83 78 P. Brown, Power and Persuasion in Late Antiquity: Towards a Christian Empire (Madison 1992) 36. 79 K. Bradley, Apuleius and Antonine Rome: Historical Essays (Toronto 2012) 12–20. 80 P. Parsons, P.Coll.Youtie II 66 = P.Oxy. XLVII 3366 = New Docs. I 26. On the date see Parsons pp.418–419. 81 Brown, Power and Persuasion 35. 82 Parsons, P.Coll.Youtie II p.410. 83 Y. Broux, Double Names and Elite Strategy in Roman Egypt (Leuven 2015) 64–75, 254–259. BENJAMIN KELLY 133 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 Whether or not they were in fact related, Aurelia Thaisous, alias Lolliane, was employing the same strategy as Lollianos, just on a less spectacular scale. Whether this was one of the factors that helped convince the prefect to agree to the request to register the petition, we cannot know. The prefect in ques- tion was probably Aurelius Theodotus, who became governor himself after defeating the usurper L. Mussius Aemilianus. Nothing secure is known about him, beyond his military suppression of the revolt and his prefecture.84 It is entirely possible that governors from military backgrounds in the mid- third century did not always meet the lofty standards of paideia expected of imperial administrators. But Aurelius Theodotus’ actual attainments are beside the point. The petition of Aurelia Thaisous is attempting to exploit an ideological expectation of shared elite culture, an expectation to which an upwardly mobile prefect (if such he was) might be expected to subscribe. Another aspect of the petition would have helped to signal that the petitioner was a member of the local elite. The lost address formula that stood at the top of the document would have contained the petitioner’s full double name and her patro- nymic. From another document, we happen to know that her father also carried a double name: Sarapion, alias Agathos Daimon. As has recently been shown, double names con- structed with the formula ὁ καί (or ἡ καί) tended to correlate with elite status in metropoleis in the third century.85 More- over, it is quite likely that the fact that Sarapion was a former agoranomos of the town would have been mentioned in the address formula too, as daughters of municipal office-holders 84 For the man and his career see PLRE I 906 and PIR2 A 1617, to be read with the skeptical notes of T. D. Barnes, “Some Persons in the Historia Augusta,” Phoenix 26 (1972) 140–182, at 173–174. For the dates of his pre- fecture see G. Bastianini, “Lista dei prefetti d’Egitto dal 30a al 299p,” ZPE 17 (1975) 263–328, at 315. 85 Broux, Double Names 268–285. 134 PROVING THE IUS LIBERORUM ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 are found mentioning this fact in other petitions.86 The petition would therefore have conveyed a powerful sense of the elite status of Aurelia Thaisous within Oxyrhynchite society, an impression to which the statements about literacy were designed to contribute. If the petition was a routine ap- plication required by law, all this would have been utterly redundant. But if it was a slightly daring and novel reuse of a procedure generally used in other kinds of matters, it is entirely explicable: the petitioner needed to establish that she was the right class of person, entitled to the benefit of any doubt on the part of the prefect (or his staff). 6. Conclusions With P.Oxy. XII 1467, we are not seeing the tip of the iceberg—the only surviving trace of a widespread application system for women who wished to qualify for or exercise the ius liberorum. Rather, we can see the opportunistic use of a petition- ing procedure usually employed in other kinds of cases. Aurelia Thaisous, for reasons that we can partly comprehend and partly guess, could not rely on any of the conventional methods to prove that she had borne three children. The decision was therefore made to submit a registration petition—something more commonly done by would-be plaintiffs in assault, theft, and property damage cases, but sometimes also in other kinds of cases, and by people who feared that they might become defendants. This would provide evidence that she had con- sistently claimed from an early date to have borne three 86 Sarapion’s name and tenure of the post of agoranomos are known from the address formula of a sale contract in which Aurelia Thaisous appears as buyer: P.Oxy. XII 1475.7–8, 11–12: παρὰ Αὐρηλίας Θαϊσοῦτος τῆς καὶ Λολλιανῆς θυγατρὸς Σαραπίωνος τοῦ καὶ Ἀγαθοῦ Δαίµονος ἀγορανοµή- σαντος τῆς Ὀξυρυγχιτῶν πόλεως. For cases of female petitioners mentioning municipal offices as part of their patronymic see P.Turner 41.2–5 (ca. 249/50; cf. BL IX 361): παρὰ Αὐ[ρ]ηλ[ίας] Σ̣α̣ρ̣α̣π̣ιάδος τῆς καὶ Διονυσα- ρίου θυ[γα]τρὸς Ἀπολλοφάνους τοῦ καὶ Σαραπάµµων̣ο̣ς̣ ἐξηγητεύσαντος τῆς Ἀντινο̣έω̣(ν) πόλεως; BGU XV 2459.3–6 = P.Turner 42 (3rd cent.); Chrest.Mitt. 229.7–8 = P.Lond. III 908 (p.132) (A.D. 139). BENJAMIN KELLY 135 ————— Greek, Roman, and Byzantine Studies 57 (2017) 105–135 children, and would bolster her credibility if this detail was ever disputed in court. In an attempt to make the prefect (and perhaps his staff) more inclined to accept the petition and deposit it in their archive, in spite of its slightly novel use of the registration procedure, allusions to the petitioner’s high social status were included. Thus, the petition reveals nothing impor- tant about the architecture of the Augustan social legislation, but a good deal about the ingenious legal strategy employed by one woman in third-century Oxyrhynchus.87 November, 2016 Department of History and Programme in Classical Studies 2140 Vari Hall, York University Toronto, Ontario M3J 1P3 Canada benkelly@yorku.ca 87 Earlier versions of this paper were read in 2014 at the Classical Asso- ciation of Canada Annual Meeting in Montreal and at a Classical Studies Research Seminar at York University. I am grateful for the suggestions offered by the audiences on both occasions, and for those made by Angela Hug and Jonathan Edmondson in subsequent discussions.