Huebner - 2007 - 'Brother-Sister' Marriage in Roman Egypt a Curios

8/13/2019 Huebner - 2007 - 'Brother-Sister' Marriage in Roman Egypt a Curios

1/29

'Brother-Sister' arriage inRoman Egypt:aCuriosity ofHumankind or aWidespread Family Strategy?

SABINE R. HUEBNER

I INTRODUCTION

Although the range nd definition f relatives ho are considered o be ineligible s sexandmarriage partners aries from ociety o society, odern anthropologists, sychologists nd sociologists gree on a few ross-cultural haracteristics, bove all the niversality f the prohibition gainst sexual relations ithin the nuclear family, hat s betweenparents and children nd between full siblings. he avoidance of sexual relationshipsbetween full iblings sgenerally eld as a universal ocial custom, nd numerous heorieslike the o-called indifference heory' ave been proposed to explain this henomenon.1

The vaguely prurient leasure with which ancient historians nd papyrologists avepresented heir vidence f the revalence f full rother-sister arriage nparts f RomanEgypt to those outside their field s thus understandable.2 hile we know of severalsocieties nhuman history nwhich royal ncest mong the ruling amilies as practised,the ociety fRoman Egyptwould have been, after ll, the nly society nhuman historyinwhich marriages etween rothers nd sisters eem to have been celebrated n a regularbasis among commonpeople.Here we have an historical uriosity hat hallenges ellestablished theories nd raises the questionwhether universal ocial and moral normsregarding exual relations ithin thenuclear family eally xist.3

According to official ensus returns rom oman Egypt dating to the first o thirdcenturies .D.,more than ixteen er cent, r twenty-two f the one hundred nd thirtysixdocumented arriages nwhich the egree f kinship etween pouses s scertainable,

were celebrated between full brothers nd sisters nd four-fifths f thesemarriagesbelonged to the second century lone.4 In this century, eventeen ut of forty-ninemarriages in themetropolis of Arsinoe were between 'brothers' nd 'sisters', ratio

*This article grew out of a broader study on intergenerational relationships in the ancient Eastern Mediterranean.I am deeply grateful to Roger Bagnall, Walter Ameling, and Walter Scheidel for their comments, criticism andencouragement. Iwould also like to express my gratitude to Alison Sharrock and the anonymous readers of JRS fortheir thoughtful comments and valuable suggestions. My special thanks go to David Ratzan without whosegenerous help this article never would have reached its present form. Any remaining errors are of course all mine.1 E.

Westermarck,The

History ofHuman

Marriage (1921, 5th edn), 320;W.

Scheidel,'Ancient

Egyptian siblingmarriage and theWestermarck effect', in A. P. Wolf and W. H. Durham (eds), Inbreeding, Incest, and the IncestTaboo: The State of Knowledge at the Turn of the Century (2004), 93-108.2M. Hombert and C. Pr?aux, Recherches sur le recensement dans l'Egypte romaine (1952), 149-53;H. Thierfelder, Die Geschwisterehe im hellenistisch-r?mischen ?gypten (i960); J. Modrzejewski, 'DieGeschwisterehe in der hellenistischen Praxis und nach r?mischem Recht', Zeitschrift der Savigny Stiftung 81 (1964)

(= J. Modrzejewski (ed.), Status personell et Heus de famille dans les droits de l'antiquit? (1993), VII), 52?82;K. Hopkins, 'Brother-sister marriage in Roman Egypt', Comparative Studies in Society and History 22 (1980),303-54

3 cf. Hopkins, op. cit. (n. 2), 304-11 for a good summary and W. Scheidel, 'Brother-sister and parent-childmarriage outside royal families in ancient Egypt and Iran: a challenge to the sociobiological view of incestavoidance?', Ethology and Sociobiology 17 (1996), 319-40, at 323.4 Census returns discussed in R. S. Bagnall and B. W. Frier, The Demography of Roman Egypt (1994), 127 were

supplemented by those published later in the second edition (R. S. Bagnall and B. W. Frier, The Demography of

Roman Egypt (2006, 2nd edn), 313?23?

all following references refer to this edition), and in P.Oxy. Census(R. S. Bagnall, B. W. Frier, and I. C. Rutherford, The Census Register P. Oxy. 984 (1997)).

JRS 97 (2007), pp. zI-49. ? World Copyright Reserved.Exclusive Licence to Publish: The Society for the Promotion of Roman Studies 2007


2/29

22 SABINE R. HUEBNER

consistent ith an actual incidence f between twenty-three nd forty-eight er cent.5 nthe urrounding illages n the rsinoite nome, sixout of thirty-eight arriages, r aboutsixteen er cent, were between iblings.6 twas presumed hat hese igh rates ould havebeen even higher, ad not the lreadyhigh ncidence f mortality esulted nmany couplesnot having both a son and a daughter urvive o adulthood. n fact, nly forty er cent ofall families ould boast of having son and a daughter urvive o marriageable age.And,sincewomen were ordinarily he junior partner in these marriages, the percentage offamilies ho exhibited this particular pattern, .e. having an elder son and a youngerdaughter, as only about twenty er cent.7 herefore, espite the stablished ssumptionthat o amount of cultural ndoctrination ould fully ffset he nnate version to matureincestuous relationships, t yet appears that inmore than one-third f all families n

Roman Egypt men who had a sistermarried her instead f looking or bride outside thefamily.8 n fact, his ustom seems to have represented omething f a 'cultural orm', anorm 'completely nknown nany other time r place'.9

The only problem is that no one has ever advanced a convincing xplanation for this

phenomenon. or half a century istinguished cholars ave proposed theories hich haveattempted odescribe the ircumstances nwhich everyday oman Egyptianswould haveoverridden ne of themost fundamental aboos in human behaviour bymarrying fullsiblings oone another, endering hese elationships ot only cceptable, ut evennormal

infact, o different n egal nd social status han xogamousmarriages.'0While severalsuch theories ere brought orward n arlier ecades,scholars nrecent earshave ceasedto offer ewexplanations, ontenting hemselvesnstead ith discounting hose previousefforts hile openly dmitting heir wn aporia.Hopkins confesses: Ido not knowwhatIam looking or; do not know what factors hould ause this henomenon. ..) The endof this rticle isdisappointing. do not have an explanation. ' Again, Alston likewiseacknowledges: Therehas been no satisfactory xplanation for the opularity f brothersistermarriage inRoman Egypt and Ihave no startling olution to theproblem to offerhere. 12 cheidel concedes: 'The underlying ationale for incestuous marriages remainsobscure. 13 haw in the arly nineties as the ast to offer n explanation, ut his theoryof racism ikewise as not foundmuch approval.14 rother-sister arriage and its atalystis thus ne of themost intractable roblems n the ocialhistory fGraeco-Roman Egypt.

5W. Scheidel, 'Incest revisited: three notes on the demography of sibling marriage in Roman Egypt', BASP 32(1995), 143-55, at 149.6

Bagnall and Frier, op. cit. (n. 4), 127-9.7Hopkins, op. cit. (n. 2), 304; Bagnall and Frier, op. cit. (n. 4), 128; Scheidel, op. cit. (n. 5), 149; Scheidel, op. cit.

(n. 3), 322;W. Scheidel, 'Brother-sister marriage in Roman Egypt', Journal of Biosocial Science 29 (1997), 361-76,at 365.8

Hopkins, op. cit. (n. 2), 304; B. D. Shaw, 'Explaining incest: brother-sister marriage inGraeco-Roman Egypt',Man 27 (1992), 267?99, at 274?5; Bagnall and Frier, op. cit. (n. 4), 127-8.9

Scheidel, op. cit. (n. 5), 150; cf. Scheidel, op. cit. (n. 3), 323; Scheidel, op. cit. (n. 1), 93.10Thierfelder, op. cit. (n. 2);Modrzejewski, op. cit. (n. 2), 52-82; Hopkins, op. cit. (n. 2); J. Goody, The Oriental,

the Ancient, and the Primitive (1990), 319-41; Shaw, op. cit. (n. 8), 267?99; W. Scheidel, Measuring Sex, Age andDeath in the Roman Empire: Explorations in Ancient Demography (1996), 9-52; Scheidel, op. cit. (n. 3), 319?40;Scheidel, op. cit. (n. 7), 361-76; S. Parker, 'Full brother-sister marriage in Roman Egypt: another look', CulturalAnthropology n (1996), 362-76.11

Hopkins, op. cit. (n. 2), 327.12R. Alston, The City in Roman and Byzantine Egypt (1999), ^6.13Scheidel, op. cit. (n. 7), 363. Most recently there has emerged an interest in 'the biology of brother-sister

marriage', that is the biological consequences of endogamy with respect to the genetic fitness of those who practisedthis form of marriage (Scheidel, op. cit. (n. 5), 143-55 ^or tne spread of brother-sister marriage; idem, op. cit.(n. 10), 9-52 for the demographic consequences; idem, op. cit. (n. 3), 319-40; idem, op. cit. (n. 7), 361-71; idem, op.cit. (n. 1), 93-108). Parker, op. cit. (n. 10), 374, observed,

though hardlyon the basis of the source material, that

marriages between siblings did not lead to 'negative genetic effects' or 'unhappy marriages'.14Shaw, op. cit. (n. 8). Cf. replies and comments by R. Abrahams, 'Explaining incest in Graeco-Roman Egypt',

Man 28 (1993), 599 and R. Firth, 'Contingency of the incest taboo', Man 29 (1994), 712?13; see also Alston,op. cit. (n. 12), 97.


3/29

BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 23

Apart from oman census lists, rother-sister arriages are recorded n a handful fother documents. ost of our testimonies or rother-sister arriage date from hefirstand second centuries .D.Two, however, re dated as far back as the third nd secondcenturies .C.15 aken together, e have one birth ertificate,'6 woclaims for rivilegedstatus,17 our marriage settlements, 8 ne tax payment,'9 wo invitations o a wedding,20one divorce ettlement,2' ne sale of crop,22 ne lawsuit,2' nd one petition o an official.24It is little onder then that opkins suggests hatwe should potentially nderstand headdress f husbands to their ives as 'sisters' nprivate etters obe literal ruth nstead fmetaphor.25 ickey, however, as recently hown n convincing anner that hese amilyterms annot in general be taken s biological references ut as an Egyptian ustom ofaddressing lso unrelated ersons.26

For the purpose of this rticle t will suffice o give a short verview f the rejectedtheories o far ffered n xplanation f this henomenon.27he most popular and earliesttheoriesmade recourse to an indigenous gyptian tradition;28 ther theories roughtforward ere the avoidance of dowry,29 hemaintenance of family roperty,'0 r the

racism f Greek settlers.First, no one has been able tomarshall any significant istorical vidence emonstrating that brother-sister arriage was practised amongst ordinary people in PharaonicEgypt.The story hilo (2oB.C.-A.D.50) tells boutMoses, that hewas purportedly orrified o learn that the aw-giver f the gyptians i.e., in theNew Kingdom periodwhenthe sraelite xodus isassumed to haven taken lace)gave full iberty omarry any sisterof either arent or of both, isoften ited as evidence that rother-sister arriagewas anindigenous, ncestral ustom of the gyptians.'2 iodorus Siculus (90-zI B.C.) likewisereports hat ccording to his sources the ancient gyptiansmade a law that permitted

marriage to one's sister, ollowing he xample of Isiswho married her brother siris.However, both stories hould indicate o us that t pparently as not a custom widelypractised mong the gyptian population in theHellenistic period. In addition, neitherDiodorus' anecdote nor Philo's may serve s evidence that this form f marriage was acustom verpractised utside the haraonic family. n fact, here s no secure videncefor

15SB i2.11053 (from 267 b.c.): auyypacpTj guyo??ia? Ylpa?,ib?\ia Kai l cuo??. ?|io>x)y?? npa^t?aua? i8i]?)Tr|cx v TeX?axoD xfji [?]ai>TO? a?e^fcpfii Ia>aio?; P.Tebt. 3.1.766 (from 147 or 136 b.c.): if?icoaa ?? ?iaypayat urc?pEuT?p7ir|? Aiov?(a?ou) xfj?; ??e?xpii? jiod Kai yovaiKO? sic ttjv ?rcofiopav. See J.Modrzejewski, 'Droit de familledans les lettres priv?es grecques d'Egypte', JJP 9-10 (1955/6), 339-63, at 346; Modrzejewski, op. cit. (n. 2), 58.16

P.Oxy. 38.2858 (from A.D. 171); see Hopkins, op. cit. (n. 2), 321.17P.Amh. 75 (from A.D. 161?168); P.Tebt. 320 (from A.D. 181); see Hopkins, op. cit. (n. 2), 321?2.18SB 12.11053 (from 267 b.c.); BGU 1.183 (from a.d. 85); P.Mil.Vog. 85 (from A.D. 138); P.Vindob.Worp. 5 (fromA.D. 168); see Hopkins, op. cit. (n. 2), 322-3.19P.Tebt. 3.1.766 (from about 136 b.c.).

20 P.Oxy. 3.524 (from the second century A.D.); P.Oxy. lui (from the third century A.D.); see Hopkins, op. cit.(n. 2), 324.21

P.Mil.Vogl. 85 (from A.D. 138); see Hopkins, op. cit. (n. 2), 323.22 P.Tebt. 2.379 (from A.D. 128).23 P.Tebt. 2.317 (from A.D. 174/5).24 BGU 3.983 (from A.D. 138-161).25Hopkins, op. cit. (n. 2), 324.26 E. Dickey, 'Literal and extended use of kinship terms in documentary papyri', Mnemosyne 57 (2004), 131-76.27Hopkins, op. cit. (n. 2), 303-54; see also Shaw, op. cit. (n. 8), 274-7; Scheidel, op. cit. (n. 10), 9-52.28Thierfelder, op. cit. (n. 2), 7-9; Shaw, op. cit. (n. 8), 274-5; Modrzejewski, op. cit. (n. 2), 54.29Rejected by Hopkins, op. cit. (n. 2), 322-4; again revived by Goody, op. cit. (n. 10), 333-4; see also Shaw,

op. cit. (n. 8), 276.30Hopkins, op. cit. (n. 2), 350-2; Goody, op. cit. (n. 10), 334; Shaw, op. cit. (n. 8), 276-7.31Shaw, op.

cit.(n. 8).32 Philo, Spec. leg. 3.23-4.33Diod. Sic. 1.27:No(io08Tiiaai ?? q>aai to?? AiyimT?ooc napa t? koiv?v e6o? tg?v ?vGpcimcov yajie?v ??e^cp??

?ia x? yeyovo? ?v toutoi? Tfj? Iai?o? ?rc?TeDyjia- Ta?TT|v y?p aovoiKf|aaaav 'Ocripi?i T(p ?8e?,(p(p cf.Modrzejewski, op. cit. (n. 2), 55.


4/29


full brother-sister ouples in the demotic papyri.34 oreover, it is widely held thatDiodorus inhis account ofAncient Egypt borrowed heavily from ecataeus of Abdera,the early third-century reek historian from Asia Minor, and presumably lso from

Manetho, the gyptian priest nd historian from ebennytos, ho lived under Ptolemyand 11;35hilo also concerned imself ith their ritings.36 oth Hecataeus andManethowere the uthors f anAigyptiaka nwhich they ried o legitimize he tolemaic laim toEgypt, and did so in part by by helping to forge n historical ontinuity etween thePtolemies and the Pharaohs.37 t is thus not going too far to assume that these courthistorians ought historical over for the ncestuous ynastic marriage of Ptolemy I andhis full ister rsinoe II, an act known to have scandalized theGreek world, by seekingrefuge nan ancient gyptian law permitting rother-sister arriage.38 n fact, ontrarytocommon ssumption, e do not have secure vidence hat he haraohs themselves verpractised full brother-sister arriage. There are some possible cases in the EighteenthDynasty, more than millennium efore tolemy I re-introduced his ustom, but eventhe full rother-sister tatus f these haraonic couples isuncertain.39

Turning to economic theories, owman and Goody have suggested hat ndogamousmarriages would have been effective n reducing hedispersion f land, a phenomenonwhich appears to have assumed alarming roportions t the beginning f Roman rule.40However, Hopkins who does not believe in materialist xplanations in generalargued onvincingly hat uch onsiderations f economic dvantagewould not have beenunique toGraeco-RomanEgypt, nd itwould therefore e surprising f nly there id theylead to brother-sister arriage.41

Finally, with respect to the social explanations, while Patlagean believes thatendogamousmarriages reflected ocial insecurities,42haw goes so far s to suggest hattheGreek settlers ere induced oseparate themselves rom he gyptian opulation outof a concern for heir acial purity.43 ccording to this ine f thought, Lagermentalititborn of near-paranoiac ejection f contacts with immediate eighbours ed the Greeksettlers o turn to incestuous relationships, nsulating hemselves n a claustrophobicsocialworld .44 nRoman times, ccording toShaw,maintenance f the rivileged thnicpolitical class of Greeks inEgypt led to an overwhelming oncernwith blood ancestry norder tomeet the requirements f a pure Greek lineage o as to avoid tax disadvantages

34J. Cerny, Consanguineous marriages in Pharaonic Egypt , Journal of Egyptian Archaeology 40 (1954), 23?9, at

29; Thierfelder, op. cit. (n. 2), 7?9; P. W. Pestman, Marriage and Matrimonial Property in Ancient Egypt. AContribution to Establishing the Legal Position of Women (1961), 2-5; Hopkins, op. cit. (n. 2), 311.35 For Diodorus depending on Hecataeus see: E. Schwartz, s.v. Diodoros (38) , RE 5.1 (1903), 663-704, at 669?70;A. Burton, Diodorus Siculus: A Commentary (1972), 1?3; K. S. Sacks, Diodorus and the First Century (1990), 70-1;J. Marincola, Authority and Tradition in Ancient Historiography (1997), 108-9; J- Warren, Epicurus andDemocritean Ethics: An Archaeology of Ataraxia (2002), 152-4. For Manetho and his later reception see:R.

Laqueur,Manethon , RE 14 (1928), 1060-106;

H.-J.Thissen, Manetho , inH. W. Helck and W. Westendorf

(eds), Lexikon der ?gyptologie 3 (1980), 1180-1.36 P. Borgen, Philo of Alexandria: An Exegete for his Time (1997), 38-40; 44-5; 250-1.37 cf. Burton, op. cit. (n. 35), 88. See also Warren, op. cit. (n. 35), 154: The presentation of Egyptian monarchymight have offered to Hecataeus an indirect means of commenting on the new monarchy of Ptolemy, perhaps even

offering a positive model for him to emulate. (...) It is tempting, therefore, also to see Hecataeus work ofethnography functioning within this negotiation of a new order in Egypt.38On Greek reaction to this marriage see Paus. 1.7; Memnon, FrGH 3 B 434.8.7; cf. D. Ogden, Polygamy,Prostitutes and Death. The Hellenistic Dynasties (1999), 73-80.39G. Robins, The relationships specified by Egyptian kinship terms of the Middle and New Kingdoms ,Chronique d Egypte 54 (1979), 197-217; S.Whale, The Family in the Eighteenth Dynasty of Egypt: A Study of the

Representation of the Family in the Private Tombs (1989), 251-2; I. Shaw, The Oxford History of Ancient Egypt(2000), 408.40A. K. Bowman, Egypt After the Pharaohs (1990), 136; Goody, op. cit. (n. 10), 338.41

Hopkins, op. cit. (n. 2), 351; cf. also Shaw, op. cit. (n. 8), 276; Scheidel, op. cit. (n. 10), 49.42 E. Patlagean, Pauvret? ?conomique et pauvret? sociale, ^e-je si?cle (1977), 118-28.43Shaw, op. cit. (n. 8), 277-93.44Shaw, op. cit. (n. 8), 290-1. See for a more balanced view D. J. Thompson, The Hellenistic family , in

G. R. Bugh (ed.), The Cambridge Companion to the Hellenistic World (2006), 93-112, at 108.


5/29

'BROTHER-SISTER' MARRIAGE IN ROMAN EGYPT 25

and impediments o upward mobility. Accordingly, ull brother-sister arriage was theinstitution evised to maintain these ritical thnic-political istinctions.45 owever, thistheory olds only if brother-sister arriage was a purely Greek cultural practice in

Graeco-Roman Egypt. This does not appear to have been the ase: Bagnall and Frier conclude from he Roman census returns hat there s a fair dmixture f Egyptian names,and brother-sister arriages occur in smallvillageswith, as it appears, a predominantlyEgyptian opulation'.46

In all these theories cholars have sought to uncover the uniquematerial and socialconditions hat hey ould point to as the ne 'true ause or decisive atalyst'47 eading owidespread brother-sister arriage n gypt s nowhere lse in the ncient editerranean.The fact that these theories ave all failed to convince raises the questionwhether thismarriage form an really e traced ack to some unique social and cultural attern revalent nEgypt. n fact, wish to argue that f we had the ame form f documentation orother regions f the astern editerranean aswe have for gypt,we wouldmost probablyfind he ame kind of 'incestuous' arriage pattern hroughout.

As indicated t the beginning f this rticle, ur most important irect testimony orbrother-sister arriages re the ensus returns rom oman Egypt, which date to the firstto third enturies .D., most of them riginating rom he econd-century rsinoite nome.

Again, it is on the basis of thismaterial that agnall and Frier take theview that iblingmarriage was more characteristic f Lower andMiddle Egypt than f Upper Egypt, ndperhaps lso that the henomenon as lesscommon n the first entury han the econd,finally oming to an end in the arly third entury.48 he real problem is and thishasnever been fully cknowledged y the cholars who have dealtwith this uestion thatwe simply o not knowwhether this henomenon as limited o this eriod, this region,or even Egypt as a whole.49 he geographical nd chronological istribution f brothersistermarriages nour census returns oincideswith the istribution f census returns ngeneral. In other words, seventeen f the twenty-two ensus returns hat documentbrother-sister arriages ome from he rsinoite nome,50 rom here sixty er cent f allcensus returns riginate.5 he problem behind the seeming uniqueness of Egyptianbrother-sister arriage is that no one has ever connected this phenomenon with thedistribution f the ources nall its onsequences.52nce one realizesthatwe are dealingherewith a phenomenon hose entire hape coincidesperfectly ith the ontours f theevidence namely that t is documentedmainly by a source type mployed rimarily nthe first hree enturies .D.and which survives here it does purely because of specificclimatic onditions other olutions ecome possible, and it turns ut that his may notbe the odd, isolated phenomenon we take it to be at first glance. We just do not know howour datawould look ifwe had the amekind of evidencefor ellenistic Egypt, r the restof the reek East from ellenistic toRoman times.

45Shaw, op. cit. (n. 8), 292.

46Bagnall and Frier, op. cit. (n. 4), 129. Cf. also Scheidel, op. cit. (n. 10), 49.47Shaw, op. cit. (n. 8), 277.48Bagnall and Frier, op. cit. (n. 4), 129-30.49 Scheidel stresses that this conclusion is not tenable on statistical grounds since the sample is not big enough and

provenance across different nomes varies considerably (Scheidel, op. cit. (n. 5), 153-4).50We have to add to the twenty cases discussed in 1994 inBagnall and Frier, op. cit. (n. 4), 129-30, two more cases

published in the second edition in 2006: ii7-Ar-i3 (full or half-sibling marriage) and i73-Ar-2i.51 See the table of distribution in Shaw, op. cit. (n. 8), 278.52cf. however already Scheidel, op. cit. (n. 5), 145, who acknowledges that the proportion of sibling marriages inthe Arsinoite nome 'rises and falls with the overall amount of documentation of married couples'. He further states

(p. 147) that 'the apparent increase of sibling couples in the late second century (...) can easily be explained withthe small number of cases' and that 'nothing in the census returns suggests a significant increase or decrease of the

practice over time'. Similarly Bagnall and Frier, op. cit. (n. 4), 129.


6/29

z6 SABINE R. HUEBNER

All scholars ngagedwith the roblem f brother-sister arriages gree n certain asicpoints:

i. The practice tarted ometime fter he dvent fGraeco-Roman rule.53

z.Marriage between biological full siblings was not imported o Egypt by theGreeks,Macedonians, or Romans andwas not practised by indigenous gyptian society n aregular nd common basis.54 t has often been noted that societies in the Eastern

Mediterranean andNear East were more inclined owards lose-kin marriage, such ascousinor uncle-niece arriages,while theWestern Mediterranean was rather ostile toit.55ut both Greeks56 nd Romans57 learly rohibited exual relationships, nd aboveall marriages, between full siblings, regarding hem against the general custom ofmankind .58 onsequently, odrzejewski has drawn the nevitable onclusion that nRoman Egypt a tendency onveyedby the Greek traditions avourable oendogamy

I ,Mt 59(was) pushed to its extreme limit

3.Brother-sister arriages were much more common in the metropoleis than in the

villages,where the opulationwas predominantly gyptian.60 rother-sister arriagestherefore eem to have been more common among those of Greek than those ofEgyptian ncestry, hough twas apparently ractised by both ethnicities.

4. There were no specific nd compelling conomic circumstances nRoman Egypt thatcould have inducedwide swaths f the opulation to considermarrying heir hildrento one another, gainst Greek, Roman, and Egyptian cultural prohibitions. omanEgypt differs ubstantively rom he ther ontemporary rovinces nly in the eculiarjuridical regimen hich theRomans adopted. But despite all efforts o convincingexplanation has been offered hat links this individuality ith the phenomenon ofbrother-sister arriage. In other words, everyone grees that is difficult o explainbrother-sister arriage as a peculiar local tradition.

Where Roman Egypt distinctively iffered rom arlier eriods and other regions f theRoman Empire is in the kind and degree f documentation:i. Papyri did not survive lsewhere, ith few xceptions.

53Thierfelder, op. cit. (n. 2), 7-9; Shaw, op. cit. (n. 8), 274-5.54Cerny, op. cit. (n. 34), 29; Thierfelder, op. cit. (n. 2), 7-9; Pestman, op. cit. (n. 34), 2-5; Shaw, op. cit. (n. 8),

274-5.55J. Goody, The Development of the Family and Marriage in Europe (1983); L. Holy, Kinship, Honour, and

Solidarity: Cousin Marriages in theMiddle East (1989); Goody, op. cit. (n. 10), 342?96, 429-64; S. Treggiari, Roman

Marriage: lusti Coniuges from the Time of Cicero to the Time ofUlpian (1991), 105-18; Shaw, op. cit. (n. 8), 270.56 cf. Aristoph., Ranae 1079; Eurip., Androm. 173-5; Xen., Mem. 4.4.20-2; Plat., leg. 8.6.838b. Unions of this kindwere only allowed for their gods; cf. J. Rudhardt, De l inceste dans la mythologie grecque , Rev. Franc.

Psychoanalyse 46 (1982), 731-63, at 733-9 and 760-1; Shaw, op. cit. (n. 8), 271.57Roman law forbade marriage between full siblings, half-siblings, and adopted siblings (Gaius, Inst. 1.61; Dig.23.2.8; 23.2.39.1; 23.2.54; 23.2.68). The Gnomon of the Idios Logos informs us that Iulius Pardalas (head of theoffice in A.D. 123/4) confiscated the property of a Roman couple who had married despite being siblings {Gnom. Id.

23 {BGU 5.1210)). The term employed in this case (a?e^cpc?v) designates inGreek any sort of sibling : full siblings,half-siblings, and siblings by adoption {SB 5.7871; cf.M. Kurylowicz, Adoption on the evidence of the papyri , JJP19 (1983), 61-75, at 63; E. Dickey, Greek Forms of Address: From Herodotus to Lucian {1996), 88; and see above in

this section). It is thus unclear how we should interpret this passage: (1) as evidence of Romans attempting to availthemselves of a local succession strategy (as I shall argue below); or (2) as evidence of a half-sibling marriage; oreven (3) as a real (and uniquely documented) case of full brother-sister incest in the Roman population. Also, one

must bear in mind that the Gnomon was not a published protocol so much as an internal handbook for officefunctionaries when confronted with difficult or unusual cases (see P. R. Swarney, The Ptolemaic and Roman Idios

Logos (1970), 120).58Diod. Sic. 1.27.1: Tcap? t? koiv?v sBo? t?ov ?vGpCMtcov yau^?v ??etap??.59J.Modrzejewski, Greek law in the Hellenistic period: family and marriage , inM. Gagarin and D. Cohen (eds),

The Cambridge Companion to Ancient Greek Law (2005), 343-56, at 351.60Bagnall and Frier, op. cit. (n. 4), 129.


7/29

BROTHER-SISTER' MARRIAGE IN ROMAN EGYPT 27

z. The Roman census records required etails (e.g. regarding arentage f women) notfound ncensusdocuments f the tolemaic period.

3. The indication f the dentity f the mother inGreek papyrus ocuments id not come

into use until the first entury A.D.61These facts, however, ught to lead us to look for n explanation that sees Roman

Egypt s distinctive n its ocumentation ather han n its ocialpractice.SinceHopkins' famous tatement hat the formula t)v1l aLa6CX(p11ntoiatptOg ai

6PORfTPWoeaves ittle oom for mbiguity',62 o one has questioned that he pouses inthese rother-sister arriageswere natural full iblings. haw isadamant in supporting

Hopkins' position: 'There s no reasonable oubt that he ersons nvolved n the iblingmarriages recorded n these ensus reports ere genuine sisters nd brothers.'63 o besure, do notwant to revive he iscussion fwhether r not we are dealing herewith apurelymetaphorical denomination f one's spouse as 'brother' r 'sister'.64 he Romancensus returns ere official ocuments, nd they efinitely eft o room for thiskind of

sentiment. urthermore, do notwish todeny that hesemarried coupleswere siblings na legalsense. suspect, owever, hat they ere not so biologically. n short, will arguethatwe aremost probably not dealingwith 'real' brother-sister arriages, ut that hese'incestuous'marriages were in factmarriages between biological child and an adoptedone, a practice ttested for lassical andHellenistic Greece, as well as forRoman Asia

Minor, Syria, and Macedonia, and discussed in Roman, Byzantine, and WesternMediaeval law.65

II ADOPTION IN FOURTH-CENTURY ATHENS

Adoption practice in the ancient Eastern Mediterranean has not yet attracted muchinterest n historical scholarship.66 owever, adoption in fourth-century thens hasrecently een studied ndetail byRubinstein67 nd Lindsay.68 he practice f adoption was

widely used in fourth-century thens as a strategy oshape the family nd succession. hemost important ourcesfor doption in this eriod are the aw-court peechesby Isaios69and Pseudo-Demosthenes.70 doption of an heir took place when the adopter hadabandoned hope that he would produce natural offspring, in despair of his present

61 cf.M. Depauw, 'Do mothers matter? The emergence of metronymics in early Roman Egypt', in T. Evans andD. Obbink (eds), Buried Linguistic Treasure (forthcoming); see also Thierfelder, op. cit. (n. 2), 90-6; Modrzejewski,op. cit. (n. 2), 57-9. See also below, Section v.62

Hopkins, op. cit. (n. 2), 321.63Shaw, op.

cit.(n. 8), 274.64 Shaw, op. cit. (n. 8), 275; Dickey, op. cit. (n. 26).65 See below, Section in.

66R. Taubenschlag, Opera Minora II (1959), 261-321; Kurylowicz, op. cit. (n. 57), 61-75; J- Beaucamp, Le statutde la femme ? Byzance (4e-je si?cle). I: Les pratiques sociales. Travaux et m?moires du Centre du recherchel'histoire et civilisation de Byzance (1992), 48-52; J.-U. Krause, Witwen und Waisen im r?mischen Reich (1994/5),III, 80-1; B. Legras, 'L'adoption en droit hell?nistique, d'apr?s les papyrus grecs d'Egypte', in A. Bresson and

M.-P. Masson et al. (eds), Actes du colloque international 'Parent?, sexe et genre dans le monde grec, de l'antiquit?? l'?ge moderne' (Volos, 18-21 juin 2003) (2006), 175-88.67 L. Rubinstein, Adoption in IV. Century Athens (1993); idem, 'Adoption in Classical Athens', inM. Corbier(ed.), Adoption et fosterage (1999), 45-62.68H. Lindsay, 'Adoption in Greek law: some comparisons with the Roman world', Newcastle Law Review 91(1998/9), 91-110. Cf. now also E. M. Harris, Democracy and the Rule of Law in Classical Athens: Essays on Law,

Society and Politics (2006), 365-70.69Is. 1-7; 9-10.70

Dem., or. 41, 43 and 44. The evidence found in the Athenian forensic corpus of the fourth century is

supplemented by several of Menander's comedies (Dyskolos, Samia and Adelphoi); cf. Rubinstein, op. cit. (n. 67),93; see also L. Rubinstein and L. Bjertrup, 'Adoption inHellenistic and Roman Athens', C M 42 (1991), 139-51,at 141.


8/29


circumstances, ussinghis old age , as Isaios said of the ld Athenian Apollodoros.7 TheAthenianMenekles contemplated dopting n heir when he had not had a natural son andwas already pproaching ld age: Menekles egan to considerhow to avoid being childless, nd how tohave someonewho would look after im inhis old agewhile he was alive,and when he was dead,would bury him and, in the future, erform herites for him. 72The institution f adoption thus rovided n individual r a couplewith a sonwho couldcare for them n their ld age;who continued his father s usiness;who perpetuated hefamily ults; ndwho maintained the gnatic ineage, nd so the ontinuation f the ikos.

There was a strong reference or dopting n heir from ne s own agnatic lineage,73even though here as no legalbar that hindered the dopter from dopting completestranger.74 nAthens, a son of one s sister as generally he referred andidate for doption.75amilieswith more than ne childwould not have resisted hen an heirless relativerequested ne of their ons.Apparently, twas not considered ishonourable ogiveone schild away in adoption, especially as it was a win-win ituation for both sides; theadopterwould gain a legitimate hild nd heir, nd the dopted sonwould become the ole

heir to his adoptive father nd therefore o longer eed to share the state f his biologicalfather ith his biological siblings. n Isaios 11, anAthenian mother persuades her husband,with whom she had several ons, tohave one of their onsadopted into the state fher childless rother or his ery reason.76

If this ption, i.e., the doption of a nephew, was unavailable, then thenians usuallyreached ut to more distant relations n order to perpetuate their ines.Girls were alsoadopted, even if less frequently han boys.77 s a daughter s son was considered therightful eir to hismaternal grandfather s roperty,78 o adopting female relative s anepilleros could also secure the ontinuation f the ikos even though his ould mean anintermission f one generation.

Sometimes t happened, however, hat he dopterwas forced o lookbeyond blood kinif lose relatives ere lacking. or instance, he ld and childless enekles had been married twice ut both marriages had failed oproduce offspring ue tohis sterility, nd sohefirst onsidered adopting a relative.79 ince his only brother had only one son, and

Menekles did notwant todeprive him of his heir,80 edecided to ask the twobrothers fhis ex-wife, ith whom he had parted on friendly erms, f ne of them ould consent tohis being adopted. One of these brothers agreed, and Menekles adopted the young manand introduced im into his phratry nd deme.8 This father-son elationship astedhappily for twenty-three ears fter he doption untilMenekles death.82 s indicated nthis ase, the dopted sonwas introduced o the family, hratry, nd deme of his adoptivefather nd inscribed n the ublic register, n themanner of a natural son.

71 Is. 7.14 (trans. Rubinstein, op. cit. (n. 67), 21).72 Is. 2.10 (trans. Rubinstein, op. cit. (n. 67), 63).73 cf. D. M. MacDowell, The Law in Classical Athens (1978), 100. The preference of close-kin adoption was

apparently a cross-cultural phenomenon, from which only our modern Western society seems to differ.74A. R. W. Harrison, The Law of Athens I: The Family and Property (1968), 88; Lindsay, op. cit. (n. 68), 94.However, since Athenian citizens were only allowed to adopt another free citizen, adopting a foundling causedmany problems because his status was unknown (Lindsay, op. cit. (n. 68), 92).75 L. Gernet, Droit et soci?t? dans la Gr?ce ancienne (1955), 121-49, esp. 129?31; Rubinstein, op. cit. (n. 67),117-25; V. Hunter, Agnatic kinship inAthenian law and Athenian family practice. Its implications for women , in

B. Halpern and D. Hobson (eds), Law, Politics, and Society in the Ancient Mediterranean World (1993), 100-21, at103?8 and 117?19; C. A. Cox, Household Interests. Property, Marriage Strategies, and Family Dynamics inAncient

Athens (1998), 126.76 Is. 11.49.77 Is. 2.10; n.41; 12.8; Rubinstein, op. cit. (n. 67), 48; 89-90.78[Dem.],

or.46.20; Rubinstein, op.

cit.(n. 67), 89-90.79 Is. 2.10.

80 Is. 2.10 and 2.21.81 Is. 2.12; Rubinstein, op. cit. (n. 67), 33; Lindsay, op. cit. (n. 68), 95.82 Is. 2.15; 2.45.


9/29

'BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 29

Adoption severed ll legal ties to the iological father nd his oikos, and thus nnulledall original rights f succession.83 nd official registers nAthens apparently id notdifferentiate etweennatural and adopted children.84 hen the Athenian Thrasylloswasadopted by Apollodoros, hewas inscribed n the ublic register f his deme asOpa6nuAXsoAnoko o66pou, Thrasyllos, on of Apollodoros', and nothing more.85The form ofregistration ave no hint that Thrasyllos had been adopted by, rather than born to,

Apollodoros. The name of his natural father, ippolochides,86 as neither egistered otmentioned.87 verything as done in such away as to securefor he doptee a status thatreplicated hat f a natural son.The adopted sonof the thenian Menekles later laimedin court: o0K IXT1tV tni6LKOq KX1pOg 6MCVEK0,oU;, OVT05 cpOuDiou IKtiVou.88

Along with the legal relationship ame the expectation f normal filial nd familialrelations. ccording to the ourt peeches ivenby saios and Pseudo-Demosthenes, twasexpected that he dopterwould treat he dopted son just s hewould a natural son. Inreturn, he dopted sonwas obliged to act like real son, e.g., show filial ove and affection; care for his adoptive father nhis old age; provide himwith a proper burial; and

perform he ccustomed rites t his grave.89 n the father's eath, a son adopted duringhis adoptive father's ifetime ad the ameright s anatural on toenter nto his estate ndwas, in fact, on a par with a natural son.90 If a man had living sons, adoption was onlyallowed bywill and would only have been effective hen the biological sonsdied beforecoming f age.91 owever, if man had adopted a son and later ad biological sonsborntohim, the doption couldnot be revoked, orwas the dopteedeprived f the rights hatthe doption had conferred n him. notherwords, hewas still ntitled oinherit n equalshare f his adoptive father's roperty.92

Itwas not only childless thenians, however,who commonly ractised doption butalso those thenian fathers ho had only a daughter.93 doption inAthenswas no bar tomarriage, s itwas inRoman law and law inmodernWestern societies. n fact, ccordingto Isaios, the dopted sonwas not merely llowed tomarry his adoptive sister, utwasrather bliged todo so.94 f n onlydaughter as still oo young tomarry, er father ouldadopt a prospective usband for herwho was bound to marry herwhen she came of age.

83Only if the adopted son had produced a son in his adoptive oikos, was he permitted to return to his own family

(Is. 6.44; 9.33; 10.11; [Dem.], or. 44.64; 44.68; Rubinstein, op. cit. (n. 67), 57-8). The same practice holds inmodernIndia: the adopted child is excluded from succession to his biological father's name and inheritance (J.Goody andJ. A. Tambiah, Bridewealth and Dowry (1974), 81; H. Lindsay, 'Adoption and its function in cross-culturalcontexts', in S. Dixon (ed.), Childhood, Class and Kin in the Roman World (2001), 190-204, at 194).84 For Roman adoptive nomenclature and the retention of original filiation and tribe, see O. Salomies, Adoptiveand Polyonymous Nomenclature in the Roman Empire (1992).85Kai outgo

jLi?vm?

?covto? 87roif|0r|vKai

sict? koiv?v

ypawiaTe?ov sveypaqmv ?paor>M.o? ATroM-o?copou(Is. 7.17 (ed. P. Roussel, i960); Lindsay, op. cit. (n. 68), 93. Cf. Harris, op. cit. (n. 68), 365?70.86 Is. 7.23.87 Is. 7.17.88 IS. 2.2.89

Rubinstein, op. cit. (n. 67), 64-76. Again, this expectation has parallels: if an adopted child in ancient

Mesopotamia said to a chamberlain or epicene who was his adoptive father and had brought him up: 'You are not

my father', his tongue was cut out (M. David, Die Adoption im altbabylonischen Recht (1927); G. R. Driver and

J. C. Miles, The Babylonian Laws H: Laws of Hammu-Rabi (1955), 75-7; Lindsay, op. cit. (n. 83), 192).90[Dem.], or. 44.29; 44.42-3; 44.53-5; Is. 2.2; 2.17; 5.16; Rubinstein, op. cit. (n. 67), 40-1, 45. An heir designated

by will could not enter directly into an estate but first had to state his claim before the archon (Harrison, op. cit.

(n. 74), 95).91Rubinstein, op. cit. (n. 67), 57.92Rubinstein, op. cit. (n. 67), 56.93Is. 3.68; MacDowell, op. cit. (n. 73), 100; Lindsay, op. cit. (n. 68), 91-2.94 Is. 3.50; 3.69; 10.13. Harrison suggested that a man in Athens who had one or more daughters probably could

not adopt a son without marrying him to one of them (Harrison, op. cit. (n. 74), 23; cf. 85. Similarly, W. Erdmann,Die Ehe im alten Griechenland (1934), 188). Potentially, another possibility was to adopt

? either inter vivos or bywill ? one's daughter's husband if she was already married when her father came to draw up his will.


10/29


If the doptee failed to do so, the doption was, according to Isaios, declared invalid.95This strategy erved several aims: first, dopting a son in the absence of a male heirprovided abour nd support nold age; second,marrying he dopted son to the naturaldaughter meant that the daughter could stay in her parents' home, a situation morecomfortable orher and her parents, ho would otherwise ave twounrelated aregiversin their ome (i.e., the dopted son and his exogenouswife), neither f whom was boundby blood or reciprocal eelings; hird, he aughter's owry,which otherwise ould havebeen lost to another oikos, remained in the family; nd finally, nd perhaps mostimportantly, hedaughter's hildren ontinued the gnatic line.96 n illustration f thecustom outlined above is found in the personal history f theAthenian Polyeuktos,mentioned in emosthenes' speech gainst Spoudias. Polyeuktos ad only twodaughterswith his wife.Therefore, e adopted hiswife's brother, eokrates, nd married him tooneof his daughters.97 eokrates thus ecame the doptive son and son-in-law f his brotherin-law olyeuktos, nd thehusband of his niece. After doption and marriage, eokrates

moved into the household of his wife and her parents. In this way, although itmight take

a generation, he family ould survive hemisfortune f having no male heir.

III ADOPTION PRAXIS IN THE GRAECO-ROMAN EAST IN LATER PERIODS

As regards the period after the fourth entury .C., Rubinstein and Bjertrup, using theepigraphic vidence for doption from ellenistic and Roman Athens, showed that venthough t is impossible o estimate hefrequency f adoption, the nstitution f adoptiondid not decline nHellenistic times, utwas practisedwell into the oman period.99 s faras we know, the oncept f the ikos and the nstitution f the pikleros ost their mportance n the ourse of the fourth entury .C., but thewish to ensure upport none's oldage still rompted hildless ndividuals nd couples to adopt.100n fact, doption iswidelyattested or the ellenistic nd Roman Eastern Mediterranean.101 he epigraphic videncefor doption of both males and females eems to be concentrated nHellenistic and earlyRoman Rhodes and Roman Caria. ForRhodes alone we have about 550 instances f adoption recorded n inscriptions ating from between the third entury .C. and the firstcentury A.D.102We should wonder whether we are really dealing here with an unusuallyhigh ncidence f adoption onRhodes, or simply Rhodian tendency o record nd advertise adoptions on stone a localepigraphic abit103-which would mean that doptions

were probably as common nother regions.104

95 Is. 3.50; 3.69; see also Dem., or. 41. Cf. Gernet, op. cit. (n. 75), 136; Harrison, op. cit. (n. 74), 85,151; D. Schaps,Economic Rights of Women in Ancient Greece (1979), 32; R. Just, Women in Athenian Law and Life (1989), 95;Rubinstein, op. cit. (n. 67), 95-6.96Thus the daughter did not become an epikleros after her father's death, since she was already married to the

legal heir (Is. 3.64). See Harrison, op. cit. (n. 74), 82-5; S. Isager, 'The marriage pattern in Classical Athens: menand women in Isaeus', C&M 33 (1982), 81-96; Lindsay, op. cit. (n. 68), 92. The custom of the epikleros is found

only in Classical Athens and has no relevance for later periods.97 It is nowhere explicitly stated that a man was allowed to adopt only one son: if he had two daughters who werenot yet married and for whom he wanted to provide, he could probably adopt two future sons-in-law (Harrison,op. cit. (n. 74), 23).98

Dem., or. 41. See Cox, op. cit. (n. 75), 35.99Rubinstein and Bjertrup, op. cit. (n. 70), 139-51.100Rubinstein and Bjertrup, op. cit. (n. 70), 140.101cf. the statement of Isaios (2.24) that not only Athenians, but also all other Greeks and some barbarians

practised adoption in the absence of a natural son.102G. Poma, 'L'adozione a Rodi',Epigraphica

34 (1972), 169-305; E. Stavrianopoulou, 'DieFrauenadoption

aufRhodos', Tyche 8 (1993), 177-88.103cf. R. MacMullen, 'The epigraphic habit in the Roman Empire', AJPh 103 (1982), 233-46; E. A. Meyer,'Explaining the epigraphic habit in the Roman Empire: the evidence of epitaphs', JRS 80 (1990), 74-96.104cf. however Poma, op. cit. (n. 102), 185-91.


11/29

BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 3I

In any event, t is clear that the trategy f adoption with a view to marriage was notconfined o fourth-century thens, but was in fact practisedwidely across the Mediterraneanworld. This practice of calling in the on-in-law n the bsence of a natural son(with r without formally dopting im) can already e traced n ur ancient ear Easternsources from Nuzi and Ugarit.105 urthermore, hieratic papyrus from the reign ofRamessesXI (early leventh entury .C.) tells s that he hildless idow Rennufer married her younger rother adiu and the daughter f her housemaid to one another ndafterwards dopted the ouple as her children nd rightful eirs.Padiu thuswas not onlyRennufer's brother, but also her son and son-in-law, nd husband and brother toRennufer's aughter.106 nd we find he ame custom recorded nour epigraphic ourcesfrom oman Macedonia and Asia Minor, the yrian idascalia, and Roman law codes.Let me present he known cases: in a grave inscription rom etella inMacedonia datedto A.D. io6, Dioskourides, the adoptive son of Moukasos, and natural son of Paibos,erected gravemonument for himself, iswife, Soura, daughter fMoukasos, their on,and their randson.107is wifewas apparently he atural daughter f his adoptive father.

Turning to Asia Minor, Menodora, daughter of Apollonides, from mperial ydai inLycia, erected an honorific monument for her deceased husband Theougenes, son ofTheramenos nd the doptive onofApollonides. It seemsthat er father ad adopted herhusband.108 e have another tomb inscription, Christian epitaph from arly fourthcentury hrygia, or certain urelios Trophimos.'09 rophimos had only twodaughtersand thus had apparently dopted a son.Ammia and his adopted son-in-law elesphoroserected monument for him. The late father rophimos confirms iswish of successionat the nd of the memorial inscription, wish thatwas firmly mbedded n theClassicalGreek tradition f inheritance trategy: Mymonument was erected by my daughterAmmia, and my adopted sonTelesphoros towhom I leftmy daughter mmia for awfulwife. '0 Even though iXo~ov n this nscription srestored, herestoration ssecured y asubsequent pitaph nscribed n the amestone, nwhich Telesphoros and his wifeAmmia

mourn their ittle aughter.111 n addition, n ametrical inscription rom ourth-centuryLycaonia, a son-in-law, neAphthonios, escribes is father-in-law alerianos, for homhe dedicated the rave monument, s 'coKei6. 2 is father-in-law as therefore robablyalso his adoptive father. 3

Let us look more closely t one examplewhich has generated omediscussion:Attalosfrom alatia erected tomb for his adoptive father, emetrios, in the third entury .D.together ith his adoptivemother. 14 n ithe provided n inscription or emetrios which

105Y. Ben-Barak, 'Inheritance by daughters in the ancient Near East', Journal of Semitic Studies 25 (1980), 22-33,at 24-5; J. Paradise, 'A daughter and her father's property at Nuzi', Journal of Cuneiform Studies 32 (1980),189-207; cf. Goody, op. cit. (n. 10), 351.106A. H.

Gardiner, 'Adoption extraordinary', JEA26

(1940), 23-9;S.

Allam,'De

l'adoptionen

Egyptepharaonique', Oriens Antiquus 11 (1974), 277-95; C. J. Eyre, 'The adoption papyrus in social context', JEA 78(1992), 207-21.107SEG 30 (1980), $96: Aiocncoupi?n? MouK?aou, (p?aei ?? nai?ou, ?coo? ear/coo Kai Eoopa MouK?aou ODv?icp

T?êi)Tf|O aVT0? ?8 AlOVUG?Ol) 0i0? 8TU)VK?' Kai AlOCFKO?p?OOl) T?pKOt) ??OOU STG)V8', ?V80T|K8V.108TAM 2.1.148: Mnvoo?pa AtcoMxov?oo? Kpnvem? oeuy?vnv ?npa|n?vo? Ka0' ?(io0?o ?av) ??ATto^ vi?oi) Kpnv?a t?v eaircfjc ?v?pa iepaTe?aavTa Kn?XXcovoq Kai Ai?c Kai 0e?)v ?yp?cov (piôaTOpyia?8V8K8V TTJ? SIC ?ai)T[f|]V.109SEG 6 (1932), 137. For a discussion of Roman law and practice in the provinces see below, Section vu.110SEG 6 (1932), 137: Efjuu ?? jLioixevâv Au|iia GuyaTnp, 9p87tTO? ?? TeXeoxp?poc, (p A,i7t?|Lir|v Koupi?inv[?Xoxov] ?|iuiav ?uio 0oy?Tpa.111SEG 6 (1932), 139: A?)p. Te?eacpopo? K? Aup. ?unia xfj ?a[i)Tv]GuyaTpi Aû?a. Cf. J. Fraser, 'Inheritance

by adoption and marriage in Phrygia, as shown in the epitaphs of Trophimos and his relatives', inW. M. Ramsay(ed.), Studies in the History and Art of the Eastern Provinces of the Roman Empire (1906), 137-53, at 142.112

MAMA 1.232: yaji?p?c ?' ?|toi rc?vTa TeX?aaaTO r\ Ta%' arcavTec |AcpOovio? q> TOKsei y>,UKep?> auoi?f|cTe??aaa?.113 'T0K88I probably means that Aphthonios was u??? 08TO? as well as ya|i?poc.' {MAMA 1.232).114RECAM 2.303: AttoXo? Arjur|[T]pi[o]i) ui?c GeT?c y[a]u?p?c ?vecn;r|aanev; seeW. M. Calder, 'Adoptionand inheritance inGalatia', JThS 31 (1930), 372-4, at 373.


12/29


stated mong other things hat he,Attalos, was bothDemetrios' adopted son and his sonin-law Mi6g F'ct KaUy7CLOp6g).ttalos moreover bore the patronymic f his adoptivefather: vtakoq Arncptiou. Calder believes that he doption of a son-in-law ocumentedin this nscription s revealing n 'oldAnatolian custom','15 ut when seen in the ight fthe ontext f Athenian marriage patterns nd adoption law as discussed bove, it seemsmore likely hat we are dealingwith amarriage and inheritance attern hat was widelypractised throughout he Greek East. Fraser, in fact, rgues that no trace of any nonHellenic system f inheritance or ons isknown in thewide range f Phrygian pigraphy'and that this custom 'was necessarily on-Phrygian, nd was certainly dopted underGreek influence nd of the Greek legaltype'.116

The information bout adoption praxis thatwe can gain from he yrian idascalia(dated to the first alf of the third entury .D.) suggests hat this ustom fmarrying nadoptive child to a natural one was also known inRoman Syria.The Didascalia advised:'When ny Christian becomesan orphan, whether tbe a boy or girl, t s good that ome

one of the rethren ho iswithout a child hould take the oy, and esteem im in the lace

of a son; and he who has a son about the same age who has reached the age of marriage,shouldmarry the girl to him. For they hich do so perform greatwork, and becomefathers o the rphans, nd shall receive hereward f this harity rom he ord God. '17

Orphaned boys shouldbe adopted by families hat had nomale offspring; rphaned girls,however, ere to be adopted by families ho had a sonnear in ge.The natural on shouldlatermarry the doptive aughter, ho as an indotata ould have had difficulties nfindinga husband. Unlike our cases from ourth-century thens, the idascalia had the rphan'swelfare inmind, rather han he nterests f the doptive arents. ere it was especially headoptive daughter who was destined tomarry a natural son. 18 he evidencediscussedabove indicates hat we are not dealing herewith a specific hristian nnovation, utwitha centuries-old radition n the asternMediterranean, re-invented naChristian etting.

Finally, turning o Roman law,marriage of an adopted daughter to a natural son andmarriage of an adopted son to a natural daughter re both discussed by Gaius in hiscommentary n the rovincial dict. SinceRoman law considered doptive relationships

- as long s they ndured as equivalent to blood relationships, ntermarriage etweenan adopted child and a natural one was only possible if the natural child was emancipated.119he same procedure for doption cummarriage was later bligatory nByzantium nd themediaevalWest.120115

Calder, op. cit. (n. 114), 372-4.116Fraser, op. cit. (n.ni), 149.117Apost. Const. 4.1 (= Syr. Didasc. 17) (ed. B. M. Metzger, 1985-7): 'Opcpavo? ?? tivo? y^vouivou xpicrtuxvo?

t|toi 7tai?o? rj 7rap6?vou, KaX,ov u?v, iva xi? xr?v ??e?-cpr?v o?k sxcdv t?kvov TtpoaA-a?ofievoc to?to exfl ei??tai?o? Tcmov, tt^v 8e 7tap0?vov ? ?xcov ui?v ?uvaîevov a?Tfj xat? to? y?îou copai? auyxpoviaai au?eu^rvtodto

y?po?

7toio?vT?? ?pyov uiya87tiT8ô0atv, 'opipav v

TtaT?pe? ?7t?pâvT8c,Kai

rcap? Kup?ouxo? 0eo?

>.i?i|/ovTai tov (itaGov xfj? ?iaKOvia? Taircri?.118cf. Krause, op. cit. (n. 66), III, 82-3.119Dig. 23.2.17 (Gaius n ad ?d. provine): 'Per adoptionem quaesita fraternitas eousque impedit nuptias, donee

manet adoptio: ideoque earn, quam pater meus adoptavit et emancipavit, potero uxorem ducere. aeque et si me

emancipato illam in potestate retinuerit, poterimus iungi matrimonio. 1. Itaque volenti generum adoptare suadetur,ut filiam emanciparet: similiter suadetur ei, qui nurum velit adoptare, ut emancipet filium.' Cf. Dig. 23.2.55 (Gaius11 ad ed. provine); Inst. 1.10.2: 'Sed si qua per adoptionem s?ror tibi esse coeperit, quamdiu quidem constat

adoptio sane inter te et earn nuptiae consistere non possunt: cum vero per emancipationem adoptio dissoluta sit,poteris earn uxorem ducere: sed et si tu emancipatus fueris, nihil est impedimento nuptiis. et ideo constat, si quisgenerum adoptare velit, deber? eum ante filiam suam emancipare: et si quis velit nurum adoptare, deber? eum antefilium emancipare'; Inst. 1.19.2; cf. Lindsay, op. cit. (n. 68), 91. See also below, Section vu.120For Byzantium see O. Montevecchi, 'Ricerche di sociolog?a nei documenti dell'Egitto greco-romano. Il contrattidi matrimonio e gli atti di divorzio', Aegyptus 16 (1936), 3-83, at 18;R. Macrides, 'Kinship by arrangement: the case

of adoption', DOP 44 (1990), 109-18; Beaucamp, op. cit. (n. 66), 108; R. Macrides, 'Substitute parents and theirchildren in Byzantium', inM. Corbier (ed.), Adoption et fosterage (1999), 307?19, at 309. For the mediaeval West:M. C. Cohn, Breviarium Alaricianum: R?misches Recht im fr?nkischen Reich in systematischer Darstellung (1908),

102; G. Vismara, 'Adozione (diritto intermedio)', Enciclopedia del Diritto I (1958), 582-3; B. Jussen, SpiritualKinship as Social Practice: Godparenthood and Adoption in the Early Middle Ages (2000), 61.


13/29

'BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 33

The devolution f property o a daughter osed problems nd concerns for father nany society n which virilocality as the ideal.121p to one third f all men inRomansociety id not have a male heir upon their eath.'22 ince women regularly arried outand took their oney with them, ny inheritance ould be lost to her father's amily ndwould go instead o her husband's.Adoption of the on-in-law as an effective trategy norder to maintain the ncestral roperty nd continue hefamily ineage. nd indeed, efind any other societies mploying his trategy norder to solve the ame problem. orinstance, doption cummarriagewas practised n hina, Japan, ndia, nd Russia until farinto the twentieth entury. hese were all societies xhibiting irilocal marriage patternslike that f the ncient astern Mediterranean, nd, as inClassicalAthens, families ithoutmale heirs mported ons-in-law norder to ensure that young ouple was availableto support he lder generation nd continue he family ame and lineage.123he proportion f adopted son-in-lawmarriages in these ocieties, hich in some regions nd times

was nearly thirty er cent, declined nly over the first alf of the twentieth entury s thenumber f peasant families ho worked their ncestral and declined in tandem ith the

rise f wage employment, he ntroduction f social security rogrammes nd a decreasingemphasis n the duty to perpetuate hefamily nd the ncestral home.124inally, uxorilocalmarriage n the bsenceof a son is a common feature f many other re-modern nd

modern societies that re otherwise ominated by virilocalmarriage patterns. he inmarrying on-in-law snot explicitly dopted here by his parents-in-law ut sometimestakes ver his wife's family ame.125doption of a future aughter-in-law, he so-calledsim-pua,was practised nChina, Taiwan, and Japan until themiddle of the twentiethcentury.126ere, an unwanted r orphaned girlwas adopted into nother amily here shewould be raised tobecome the ventual ife of a son, thus voiding the ostof engagementandwedding presents, bviously bove all interesting orrather oor families.

Given the prevalence f this trategy nder similar family ystems nd demographicconditions, ould itnot bemore economical to readour Egyptian vidence n the ight fthis ustom f adoption cummarriage rather han o seeRoman Egypt s the nly nstanceinhuman history fwidespread institutionalized ncest?

IV ADOPTION IN GRAECO-ROMAN EGYPT

Due to lack of evidence adoption is usually not regarded s a widespread custom inGraeco-Roman Egypt,127 ven though the institution f adoption was already welldeveloped inPharaonic Egypt as a strategy or vercoming hildlessness.128ur earliest

121cf. e.g. M. L. Satlow, 'Marriage payments and succession strategies', in R. Katzoff and D. Schaps (eds), Law inthe Documents of the Judaean Desert (2005), 51-65.122R. P.

Sailer, Patriarchy, Propertyand Death in the Roman

Family (1994), 52: 'Male, ordinary ,Level

3West';and 58: 'Male, senatorial , Level 3West'.123For China see: J. Goody, Production and Reproduction: A Comparative Study of the Domestic Domain (1976),76; Goody, op. cit. (n. 10), 45f.; A. P. Wolf and C.-S. Huang, Marriage and Adoption in China 1854-1945 (1980).For Japan see: R. P. Dore, City Life in Japan. A Study of a Tokyo Ward (1999, 2nd edn), 147; C. J. Dunn, EverydayLife in Traditional Japan (1969), 71. For India see: P. Diwan, Law of Adoption, Minority, Guardianship and

Custody (2000, 3rd edn), 38-40. For Russia see: D. I.Kertzer and M. Barbargli, Family Life in Early Modern Times

1500-1789 (2001), 52. See also below, Section vu.124cf. Dore, op. cit. (n. 123), 147; Goody, op. cit. (n. 10), 46, 106-7.125See e.g. A. Shimizu, 'On the notion of kinship', Man 26 (1991), 377-403; J. R. Bowen, 'Equality, difference, andlaw in Indonesian inheritance practices: a Sumatran case study', PoLAR 19.1 (1996), 83-90; E. A. Hammel and

A. Gullickson, 'Kinship structures and survival: maternal mortality on the Croatian-Bosnian border 1750-1898',Population Studies 58 (2004), 145-59, at 151.126See A. P. Wolf, 'Adopt a daughter-in-law, marry a sister: a Chinese solution to the problem of the incest taboo',

Am. Anthropol. 70 (1968), 864-74; Goody, op. cit. (n. 10), 107.127cf. e.g. Scheidel, op. cit. (n. 10), 48.128Gardiner, op. cit. (n. 106), 23-9; Pestman, op. cit. (n. 34), 4; Allam, op. cit. (n. 106), 277-95; A. McDowell,

'Legal aspects of care of the elderly in Egypt to the end of the New Kingdom', inM. Stol and S. P. Vleeming (eds),The Care of the Elderly in the Ancient Near East (1998), 199-222, at 219.


14/29


testimony rom hePtolemaicperiod dates to z48 B.C.129n this ocument, the so-called'agenda list of Zenon', the author was interested n determining he identity nd

whereabouts f a particular erson nd sohe planned to inspect he ontracts fmarriageand adoption: '[See] the ontracts fmarriage and adoption.Who is the man and where ishe? 130t is indeed remarkable hat this arliest vidence for doption inPtolemaicEgyptmentioned the contracts fmarriage nd adoption' as somehow tanding ncloseconnection to one another. apyrus sourceson adoption for the following enturies re scarceand for he most part contain ittle nformation.13' e find everal remarks bout persons,men and women, who were adopted indiverse ocuments, ut the doptive status sonlymentioned here as part of the doptee's name.'32 ur earliest urviving ontract f adoption omes from hefourth entury .D.,'33nd it is only from his nd later vidence that

we are able togainmore information n the rocedure f adoption in gypt. et, this oesnot mean that adoptions necessarily occurred less frequently efore this point. Aavvypawpi as merely document f proofwithout any constitutive r dispositive ower,as oral contracts ere generally ufficient.134rom the fourth entury nwritten ontracts,

however, eem to have become more important.In reviewing ll the evidence, Taubenschlag,135 itteiS,136 nd Kurylowicz137 avedemonstrated hat he ractice f adoption inRoman Egypt, oth in its onception nd inits onsequences, ested pon Hellenistic legal principles.138he terminology or doptingand adoptee, aswell as the egal thinking ndergirding doption inRoman Egypt,werederived from ttic law.139he contracts f adoption provide uswith the most informationin this respect. rom these ocuments, e learn that doptionwas, as inAthens, considered binding nd permanent. he biologicalparents ost all rights egarding heir hild.Their sonor daughter ecame the egalchild of the dopting parents, ho were responsiblefrom hen n for upplying ll of his or her needs.140 he adoptee forewent ll kinship

with his biological father.'4' n adoption ontract rom arly fourth-century xyrhynchusstates: Weagree,Heracles and hiswife Isarion, n our side, thatwe have surrendered oyou,Horion, for doption our sonPatermouthis, bout two-years-old, nd I,Horion, onmy side, that hold him as my genuine on (yvfcntov i'v) as regards hemaintenance of

129P.Col.Zen. 3.58; cf. now Legras, op. cit. (n. 66), 175-7. F?r adoption in Pharaonic times see C. Seidl, ?gyptischeRechtsgeschichte der Saiten- und Perserzeit (1968), 54, 80; Allam, op. cit. (n. 106), 277-95.130P.Col.Zen. 3.58: T?8?V (...) T?? or>yypav ya^io?vTOW Kai tskvo0scu(X)v. t?? ? avGpc?Tto? Kai rcoO ?crriv;For the at)vypa(pf|, the written contract that was taken over from Greek notary practice, cf. H.-A. Rupprecht,'Greek law in foreign surroundings: continuity and development', in M. Gagarin and D. Cohen (eds), The

Cambridge Companion to Ancient Greek Law (2005), 328-42, at 330.131e.g. P.Heid. 4.329 from a.D. 105/6 (male); P.Mert. 1.18 from A.D. 161 (male); P.Oxy. 3.504 from the second

century A.D. (female); P.Stras. 1.4 from A.D. 551 (two sisters adopted by the same man); cf. Kurylowicz, op. cit.

(n. 57), 61; Ch. Kunst, R?mische Adoption.Zur

Strategieeiner

Familienorganisation (2005), 233132Kurylowicz, op. cit. (n. ^y), 6z.133P.Oxy. 9.1206 (from A.D. 335); P.Lips. 1.28 (from A.D. 381); P.Oxy. 16.1895 (from A.D. 554). However,

P.Cair.Masp. 3.67305 (from A.D. 568) cannot be regarded as an adoption contract.134Rupprecht, op. cit. (n. 130), 331; 335-6.135R. Taubenschlag, The Law of Greco-Roman Egypt in the Light of the Papyri (332 B.C.-640 A.D.) (1955, 2nd

edn), 261-321.136L. Mitteis, 'Adoptionsurkunde vom Jahre 381 n.Chr.', Archiv f?r Papyrusforschung und verwandte Gebiete 3(1906), 173-84, at 179.137

Kurylowicz, op. cit. (n. 57), 61, 72-5.138For an overview see Taubenschlag, op. cit. (n. 135), 261-321; Kurylowicz, op. cit. (n. 57), 61-75; Beaucamp, op.cit. (n. 66), 48-52; Krause, op. cit. (n. 66), III, 80-1. For the terminology used see Taubenschlag, op. cit. (n. 135),263; Kurylowicz, op. cit. (n. 57), 61.139

Kurylowicz, op. cit. (n. 57), 61. See also Mitteis, op. cit. (n. 136), 179; U. Wilcken, Urkunden aus der

Ptolem?erzeit I (1927), 124; Taubenschlag, op. cit. (n. 135), 263, no. 10;Harrison, op. cit. (n. 74), 84.140P.Oslo 3.114 (from the first to second century A.D.); P.Oxy. 9.1206 (from A.D. 335); P.K?ln 7.321 (from A.D. 335);P.Lips. 1.28 (from a.d. 381); P.Oxy. 16.1895 (from A.D. 554). For adoption in the papyri see Taubenschlag, op. cit.

(n. 135), 327; Kurylowicz, op. cit. (n. 57), 61?75; Beaucamp, op. cit. (n. 66), 48-52; Krause, op. cit. (n. 66), III, 80-1.141[Dem.], or. 44.21-2; cf. Rubinstein, op. cit. (n. 67), 22.


15/29


16/29


natural daughter utside the homemeant that er children ould belong to her husband'sfamily, ot to her father's.151

Itmust be stressed hat t does not seem to have been that ncommon n ntiquity orfather oapproach old age and fear yingwithout amale heir. The probability f amanover the age of fifty aving a living on never exceeded, according to Saller's tables ofproportion f living in, sixty-nine er cent in the rdinary roup,152nd in the enatorialgroup not even fifty-six er cent.153 ne-third to nearly ne-half f allmen above the geof fifty ccordingly id not possessamale heir, nd, abandoning ope that hey ould everbeget a natural son, they herefore ould have been likely oconsider doption.154hesecalculations f living in probability re not based upon the mpirical vidence from heRoman census returns, ut have been calculatedon the asis of model life ables.'55allershows at the ame time n a convincing anner that his computer imulations onstitute'a realistic epresentation f the ancientMediterranean experience' nd by comparison

with the ata from he oman census returns hat he atter re 'coherent nd demographically lausible'.'56While themodel life ables onsider ll living ons regardless f where

they ived, theRoman census returns ocument only those sonswho lived in the samehousehold as their athers. considerable ercentage f youngmen did not, however, iveat home due towork migration, sBagnall and Frier have shown.157or the ensus returnsfrom oman Egyptwe must therefore ssume that the roportion f men over the ge offifty ho did not register son in their ousehold shouldhave been correspondingly venhigher han the figures alculatedby Salleron thebasis ofmodel life ables.

Fifty-six en over the ge of fifty re documented n theRoman census returns ublished nTheDemography f Roman Egypt ncluding upplemental ensusdeclarations nthe second edition and further aterial from an Upper Egyptian city published inP.Oxy.Census.'58 ut only seven f these ifty-six en did not register son 159hat means

151For the same sentiments at Rome, see Krause, op. cit. (n. 66), I, 150. For China see Goody, op. cit. (n. 10), 45-7.152Sailer, op. cit. (n. 122), 52: 'Male, "ordinary", Level 3West'.153Sailer, op. cit. (n. 122), 58: 'Male, "senatorial", Level 3West'.154cf. M. Corbier, 'Divorce and adoption as familial strategies', in B. Rawson (ed.), Marriage, Divorce, and

Children in Ancient Rome (1991), 47-78, at 67: 'This [a high proportion of the male population withoutdescendants] can probably be attributed to numerous factors: men's late marriages, their prolonged absences, longperiods of voluntary celibacy after widowhood or divorce (...), the practice of taking a concubine, the high degreeof sterility in couples of the past, the mortality of young children, many of whom died before their parents.Adoption could also be preferred to remarriage.'155cf. A. J. Coale and P. Demeny, Regional Model Life Tables and Stable Populations (1983, 2nd edn); Sailer,op. cit. (n. 122), 44-69.156

Sailer, op. cit. (n. 122), 66.157Bagnall and Frier, op. cit. (n. 4), 160-9; R? S. Bagnall, review ofW. Scheidel, Measuring Sex, Age and Death inthe Roman Empire: Explorations in Ancient Demography {1996) in BMCR 97.8.17 (1997): 'It should be borne in

mind (...) that the census declarations list (...) only those surviving children who still lived in the parentalhousehold. Those who had married out are not included in their parents' declaration.'158Cited after Bagnall and Frier, op. cit. (n. 4): 11-Ar-i; 61-Ar-i; io3-Ar-7; 117-AP-5; ii7-Ap-6; 117-AP-7; 117-Ar6; 117-Ar-n; 117-Ox-i; i3i-Ar-8; i3i-Ar-i4; ii7-Ar-i3 (2men); 131-Be-i; 131- He-3, i3i-He-4; 131-Ox-i; 145-Ar9; I45-Ar-I2; i45~He-2; 145-Oa-i; i45-Oa-2; 145-OX-2; 145-Pr-i; 159-Ar-i; 159-Ar-n; 159-H111-3; 173-Me-i; 173

M3-3; i73-Pr-5; 173^-7; i87-Ar-4; i87-Ar-8; i87-Ar-i8; i87-Ar-22; i87-Ar-26; n?-Ox-i; 20i-Ar-6; 2oi-Ar-9;2oi-Ar-i4; 2i5~Ar-4; 2i5-Ar-6; 215-Ar-n; 2i5-He~3; 229-Hm-i; 229-Hm-2; ???-Ar-3; i3i-Ar-i2. Bagnall, Frier andRutherford, op. cit. (n. 4): 89-PM9; 89^-23; 89-Pt-26; 89-Pt-46.159

Bagnall and Frier, op. cit. (n. 4): i45-Oa-2 (the fifty-three-year-old Tithoes lived with his fifty-year-old wife andtheir two seventeen-year-old and sixteen-year-old daughters); i45-Ar-i2 (the seventy-six-year-old Petheus livedtogether with his seventy-year-old wife and his thirty-three-year-old divorced daughter); 173^-3 (tne fifty-six-yearold unmarried and childless Athas

lived togetherwith his

two,also

unmarriedand

childless, sisters); i87~Ar-26 (thefifty-seven-year-old Petsoraipis lived alone with his thirteen-year-old daughter and three female relatives); 2oi-Ar-6(the fifty-year-old Neilos and his wife had only a daughter who had married away); 2oi-Ar-i4 (the fifty-year-old

Alba lived alone with his twenty-seven-year-old daughter Teieus); ???-Ar-3 (the fifty-seven-year-old NN and histhirty-eight-year-old wife Thaesis had only three daughters).


17/29

BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 37

that nly a rate f around ten er cent ofmen above the ge of fifty ad no living on.160And given the ther vidence, erhaps this umber was even lower, s many young men,as mentioned above, had to migrate from their ometown due to work commitments.Bagnall and Frier propose that pproximately ine per cent of all fifteen-year-old illagemales migrated to metropoleis between the ages of fifteen nd twenty-four. 16 cheidelargues for even higher numbers due to the chronic under-reporting f juveniles n thevillages.162ive of our seven cases of elderly men without sons indeed originate fromvillages.163he sixth ase,which comes from hemetropolis ofArsinoe, records fiftyyear-oldman and his forty-four-year-old ife, humble peoplewho were only renting heplace they ere living n.They had an adult daughter ho had married the wner of thehouse.Due to their overty nd inability oprovide a legacy, he ouplewould have haddifficulty inding young man willing to be adopted.164n the eventh ase, the x-son-inlaw of the eventy-six-year-old etheus,who lived together ith his seventy-year-old ifeand his thirty-three-year-old ivorceddaughter, scalledNN, son of Petheus.He couldhave been Petheus biological, r rather, s Iargue bove, his adopted sonwho married his

adoptive sister nd later ivorced her.165 etheus thus an either be not counted amongthese lderly athers ithout a son,or if e believe n the doption theory utlined bove,had not accepted his fate ut adopted a son and married him to his daughter. n the ightof this vidence, t eems that nly a very mallpercentage fmen, in fact essthan en ercent, over the ge of fifty ecorded n theRoman census returns id not possess a livingmale heir.According toSaller s tables we should expect a rate f around thirty o fortyfive er cent ofmen who did not have a living on, and if ccupationalmigration s takeninto ccount, n even higher rate.Given thehighmortality ates thesenumbers from hecensus returns eem rather mpossible.

The most obvious explanation would be that lderly fathers ithout sons inRomanEgypt had taken ecourse o adoption.Adoption in the bsence of a natural onwould nothave distorted he verall fertility ates n the ensus returns which ave been proven tobe in linewith most populations before themodern fertility ransition166)utwould haveresulted na more even llocation ofmale offspring ver all households. owever, amongthe lmost ,500persons recorded n the oman census ists, ot even one declared that ehad adopted a childor that e had himself een adopted.167et, no scholar ngaged nthestudy f the oman census returns as evernoted that doptive status s never xpressedin these ocuments. iven thehigh percentage f elderlymen with a living ale heir, thisshouldgiveus pause for hought.

Our theory f adoption,moreover, finds urther upport n the bservation hat oevalsiblings re far ver-represented n the ensus returns.168 nly four f these oeval siblings

160The fifty-seven-year-old Papontas (117-Ox-i) is registered without wife and children in the household of hiselder brother but lived somewhere else. It is possible that he lived with his wife and children. The list of names thatfollowed the elderly couple of i3i-Ar-8 is lost; they could have had several sons. The census list inwhich only the

fifty-year-old Aurelius Theognostos and his wife are registered (229-Hm-i) is heavily mutilated and could have

comprised further names.161Bagnall and Frier, op. cit. (n. 4), 165.162W. Scheidel, Death on the Nile: Disease and the Demography of Roman Egypt (2001), 156, 172.163Bagnall and Frier, op. cit. (n. 4): i45-Oa-2 from Mesobe (Great Oasis); i73-Pr-3 from Thelbonton Siphtha; 187

Ar-26 from Karanis; 2oi-Ar-i4 from Narmouthis; ???-Ar-i2 from Soknopaiou Nesos.164Bagnall and Frier, op. cit. (n. 4): 2oi-Ar-6.165Bagnall and Frier, op. cit. (n. 4): i45-Ar-i2.166Bagnall and Frier, op. cit. (n. 4), 138.1671,100 persons in Bagnall and Frier (op. cit. (n. 4,1994), p. xv) plus about 100 persons in the supplement (Bagnall

andFrier, op.

cit.(n. 4, 2006)), plus

about another250 persons

documented inP.Oxy.Census (Bagnall,

Frier andRutherford, op. cit. (n. 4)).168

Bagnall and Frier, op. cit. (n. 4), 43;W. Scheidel, What s in an age? A comparative view of bias in the censusreturns of Roman Egypt , BASP 33 (1995), 2.5-59; W. Scheidel, Twins in Roman Egypt: postscript to BASP 33(1995) , BASP 34 (1996), 35-7. I am grateful toWalter Scheidel for calling my attention to this point.


18/29


were, however, xplicitly eferred o as 'twins';'69 n the ther nstances, he iblings eresimply ecorded s being of identical ge.170 agnall and Frier concede that these iblingscould be the ame agewithout being twins, ut adduce studies that how that uch closespacing s in generalvery rare nhuman populations.17' lso, the hances of twins urviving together o a certain ge is smaller han the hance of either f them;moreover, nfant

mortality ates re higher or twins hanfor ingletons. o the fact thatwe have even twotwins n their id-fifties'72 shighly uspicious. or this eason, cheidel suggests hat theattestation f twins in the ensus returns) annot reflect eality'.'73 n explanation for heunexpectedly igh number f coeval siblings, ore economical than ge exaggeration rage-rounding hich are usually considered negligible n the returns,174 ould again beadoption. That the coeval siblings recorded n the census returns re on averagemucholder that those coeval siblings xplicitly alled twins meshes with the suggestion fScheidel that atural twins ave amuch lower ife xpectancy han ingletons,175nd likewise points in the direction that these coeval 'siblings'were neither twins nor evenbiological siblings.

We may even have one direct hint in the ensus records hat doptions occurred. Wehave two consecutive ensus returns rom .D. I3i and I45 for family onsisting f thefather hentmouphis, iswifeDemetrous, their onAnikos, and his sister hamistis, whowas four years older thanAnikos.176n both returns hamistis isdeclared as the hild ofChentmouphis nd his wife Demetrous and also as the full sister f Anikos. In a laterchirograph fAnikos, however, ritten elow the returns nd dated to A.D. i6i, at a timewhen both parents were apparently lready dead, Thamistis is said to be only the halfsister fAnikos on themother's side with her father nknown. t is therefore ot going toofar to assume that hentmouphis, erhaps uponmarriage, had adopted the aughter hathiswife had from n earlier relationship, fact which Anikos as his father's atural childlater apparently ontested because itwould have diminished his inheritance xpectations.177e might therefore ssume that while adoptionwas not recorded n the ensusreturns, twas nonetheless egularly ractised nRoman Egypt.And adoption apparentlyoccurred not onlywith a view to old age support nd continuation f the family ineage,but also in order to provide male and female stepchildren nd related and unrelateddestitute hildren ith parents nd a home.

The fact that doptionswere not indicated n the returns s not particularly urprisingwhen we recall that unlike theRoman practice, Greek onomastic convention id notgenerally ecord doptive status.178I1 st a remarquer ue le droit de decider neglige esprescriptions trictes u droit romain, l'adopte ne gardant pas le nomen de son perenaturel comme cognomen. 179s attested for Classical Athens, so also in Asia Minor

169Bagnall

andFrier, op.

cit.(n. 4): i45-Ar-i7 (fourteen years old); 145-Ox-i (three years old); i73-Ar~9 (thirtyeight years old); i87~Ar-4 (one year old); cf. Bagnall and Frier, op. cit. (n. 4), 43.170

Bagnall and Frier, op. cit. (n. 4): 33-Ar-i (eighteen years old); iiy-Ar-2 (thirteen years old); 131-Ox-i (halfsiblings of identical age; twenty years old); 159-Ar^ (four years old); i73-Ar-i6 (twenty-four years old); 20i-Ar-9(fifty-six years old); 201-Ar-io (fifty-four years old); 2i5-Ar~4 (half-siblings of identical age; three years old).171

Bagnall and Frier, op. cit. (n. 4), 43, n. 41.172Bagnall and Frier, op. cit. (n. 4): 2oi-Ar-9; 201-Ar-io; Scheidel, op. cit. (n. 168), 53-4. Co-resident twins in their

mid-fifties are overrepresented in the Roman census returns up to 700- to 800-fold.173Scheidel, op. cit. (n. 168), 56.174See Bagnall and Frier, op. cit. (n. 4): 44-7 who conclude that age exaggeration and age rounding in the Egyptian

census are rather low compared with modern less developed countries' census data, and 'some trust' (p. 45) shouldbe placed in reported ages.175

Bagnall and Frier, op. cit. (n. 4): 2oi-Ar-9; 201-Ar-io; cf. Scheidel, op. cit. (n. 168), 53-4.176Bagnall and Frier, op. cit. (n. 4): 131-Pr-i; 145-Pr-i.177cf. H.C. Youtie, 'APATORES: law vs. custom in Roman Egypt', in J. Bingen, G. Cambier and G. Nachtergael

(eds), Le monde grec. Hommages ? Claire Pr?aux (1975) (=H. C. Youtie, Scriptiunculae Posteriores I (1981), 17-34),723-40, at 723-5.178

Salomies, op. cit. (n. 84).179Modrzejewski, op. cit. (n. 15), 349.


19/29

BROTHER-SISTER MARRIAGE INROMAN EGYPT 39

adoption had the consequence that the adoptee took over his adoptive father'spatronymic.'80 ermakotas, for nstance, doptive on of Aischylos from rneai in ycia,was called EppxaKotcaqX6X'kou.'8'The samepractice eems to have prevailed nRomanEgypt: adoptive status was not declared in the official registers. o, Aurelius Horionconfirmed n the doption ontract ited bove from arly fourth-century xyrhynchus: I,AureliusHorion, have adopted the oy and will register im (in the fficial ecords) s mygenuine on' ([At~p]Wioq 'Qpicov rapci2,'pc ov natLba i[qUio0ciuv Kali] noypa'VoiLacUTOvi; .g uco yvfio[v tiov]).'82he same seems to have applied for the Romancensus returns: dopting parents pparently eclared their dopted child like a naturalone,with the hild taking n both his or her new father's nd mother's name.

This brings s to the uestion ofwhether women inRoman Egypt were able to adopt,a precondition oexplain thephrase 7uvrjKtaLi6ckp'1 0jioiuftTptoqai OgOiTptoq satisfactorily. nClassicalGreece women probably could not adopt183 nd Roman law alsoruled t out.184nGraeco-RomanEgypt, however, ocial custom ccorded women a significantly igher tatus han oman law: a woman was allowed to giveher daughter nmar

riage ointly ith her husband or was even free odo so herself f shewas widowed ordivorced. n addition,we seemany women serving s guardians f their inor children,practice ondoned neither yAttic norRoman law.185nd our evidence oints also in thedirection that a woman was in a position to join her husband in the adoption of a child oradopt one on her ownwhen she waswithout a husband.186 papyrus from econd-centuryOxyrhynchus records that a widow had adopted the two natural sons of her latehusband.187n another ase from econd-century xyrhynchus, n adopted woman callsthewife of her adoptive father F?6GFjTflp.'88rom the ate third entury .D. we canfinally bserve a recognition y Roman law of the prevailing ocialpractice, by giving

women the official right o adopt: women did not obtain potestas over their doptivechildren s they id not have their atural children npotestate, but by adoption theadoptive child acquired the same rights o inherit rom is adoptivemother as natural

180Kai outgo \iev ?tt? ?r?vTO? 67toif|0r|v Kai sic t? koiv?v ypajijiaTe?ov ?veypa(pr|v ?paaoAAo? ATto^Xo? poi)(Is. 7.17 (ed. P. Roussel, i960)); Lindsay, op. cit. (n. 68), 93. Cf. Harris, op. cit. (n. 68), 365-70 and see above, Section

m.181TAM 2.776.182

P.Oxy. 9.1206 (from A.D. 335).183Harrison concluded from one passage in Isaios (7.25) that in Classical Athens adoption did not severe the legalties between the mother and her biological child nor was any relationship established between the child and the

adopter's wife (Harrison, op. cit. (n. 74), 94). Thrasyllos was adopted by his childless unmarried uncle Apollodoroswho was the brother of his mother. He stated:

Mr|Tpo?8'

o??ei?sgtw

SK7toir|TO?,?Xk'

?|AO?(0? ?m?pxei tt^va?rcf|v e?vai ur|T?pa, Kav ?v t(?> 7taTpQKpuivrj ti? o?KCpK?v 8K7toir|0fi. (Is. 7.25 (ed. P. Roussel, i960)). But we donot know how the situation would have been described if the adopting uncle had been married.184 Inst. 1.97 und 104 (Gaius); Clust 7.33.8 (from A.D. 294); Krause, op. cit. (n. 66), III, 83.185cf. O. Montevecchi, 'Una donna prostatis del figlio minorenne in un papiro del lia', Aegyptus 61 (1981)(= O. Montevecchi, Scripta Selecta (ed. S. Daris) (1998), 273-85), 103-115, at 113-15; T. J. Chuisi, 'Zur

Vormundschaft der Mutter', ZSSR.RA in (1994), 155-96, at 175-91; R. van Bremen, The Limits of Participation:Women and Civic Life in the Greek East in the Hellenistic and Roman Periods (1996), 228-30; Evans Grubbs,op. cit. (n. 150), 254-7; H. J.Wolff, Das Recht der griechischen Papyri ?gyptens in der Zeit der Ptolem?eer und des

Prinzipats. Vol. I: Bedingungen und Triebkr?fte der Rechtsentwicklung (ed. H.-A. Rupprecht) (2002), 78, n. 31;S. R. Huebner, 'Callirhoe's dilemma: remarriage and stepfathers in the Graeco-Roman East', in S. R. Huebner and

D. M. Ratzan (eds), Growing up Fatherless in Antiquity (2008), forthcoming. Roman law at the end of the fourthcentury officially granted mothers the right to administer their children's patrimony, but on the precondition that

they promised not to enter a new marriage {CTh 3.17.4 from A.D. 390 (=CIust 5.35.2)).186Kurylowicz, op. cit. (n. 57), 62-3.187P.Oxy. 3.583 (from A.D. 119/20). See Taubenschlag, op. cit. (n. 66), 264. Cf. the adoption by Arsinoe II of her

husband's children in Theoc, Id. 17.128; and the adoption of Herakles by Hera in Diod. Sic. 4.39.2; cf. Legras,op. cit. (n. 66), 183-5.188

P.Oxy. 2.504 (from the early second century A.D.).


20/29


childrenwould have had.189 he normalization f adopted children regarding omenclaturewithin the family xtended not only to father ndmother; adoptive children venreferred o thenatural children f their ew parents s brothers nd sisters .190

To sum up our conclusions from the foregoing iscussion: first, e have seen thatoutside gypt doption (of males aswell as females) as awidespread phenomenon n theancient astern Mediterranean. Second, the arliest vidence for doptions among Greeksettlers nEgypt dates to the arly Ptolemaic period. 9 Third, the percentage f elderly

men in the Roman census returns ho had a livingmale heir was unrealistically igh.Fourth, oeval siblings lso appear far oo frequently n the ensus returns iven the emographic robability. nd finally, n adopted sonwas registered s a yviatoq utio6qnaturalson) in the public registers nd we thuswould not be able to distinguish im frombiologicalchild.All these trands f this rgument, he irect vidence f the ources, thelogical onclusions ased on demographicmodelling, nd the omparative vidence fromother parts of the Hellenistic world, together uggest that doption inGraeco-RomanEgyptwas much more common than sdocumented nour sources.192

V BROTHER-SISTER MARRIAGE

We have seen thus far that there remany indirect igns which point to routine doptioninRoman Egypt. We have also seen that nClassicalAthens, Roman Macedonia, Asia

Minor and Syria adoptionswere often, fpossible, combinedwith marriage within thefamily, nd that n the bsenceof a natural male heir adoption represented traditionalstrategy o ensure labour, continuity, nd succession in severalpre-industrial ocieties,particularly hosewhich exhibit irilocalmarriage patterns omparable to those bservedfor the ncient astern Mediterranean.193 ll this hould lead us to question the llegeduniquenessof the gyptian situationwith respect obrother-sister arriage and, moreover, the earch for unique socialand cultural onditions that ould have produced thisphenomenon. herefore, when we look at the practice of brother-sister arriages inGraeco-Roman Egypt, s itnot likely hat e are dealingwith the amephenomenon swehave found nClassicalAthens,AsiaMinor, Syria, ndMacedonia, namely that heneverwe find a brother married to his sister we should understand that one of the spouses is anadoptive childmarrying henatural child of his/her doptive family?

We shallnow take closer lookat our evidence for pparent brother-sister arriage inEgypt.Our first estimony or thismarriage form s a heavilymutilated marriage contract

189Diocletian assured a certain Syra, who wanted to adopt her stepson, that the adoptive child gained the same

rights as if he was born from you {Clust 8.47(48).5 from A.D. 291). Cf. Inst. 1.11.10: Feminae quoque adoptare non

possunt, quianee

naturales liberos in potestate sua habent: sed ex indulgentia principis ad solatium liberorumamissorum adoptare possunt. See also Krause, op. cit. (n. 66), III, 83. In an adoption contract from early sixthcentury Oxyrhynchus, a married couple jointly adopted a nine-year old girl surrendered by her widowed mother{P.Oxy. 16.1895 from a.d. 554,1. 8-12: f|?icoaa ?[ii?? Ta?Trjv 7tapa>,a?ei]y rcap ?jio? ei? 0?yaTspa, Kai ?uo?,oy?>7iapa[?e?a>K8vai a?Tfjv ?uiv arco tod] io[v?v ei? t?v ??]fj? arcavTa xpovov ei? 0uyaT?pa vou?urjv, [coaTe ?fi??XopnyoOvTa? Ta ??ovTa x^pav yov]?cov ei? 0uyaT?pa ?7ro7cXr|pG)fjai ei? a[?Tf|v, Kai 8VTeC0ev \ir\ ovaa0ai ixeTaircnv ?]7toa7taaai a(p uux?v). We have, in addition, many testimonies from all over the early Byzantine world forChristian deaconesses and female ascetics adopting a child on their own (Eus., H.E. 6.2.12-13; Pallad., Hist. Laus.6.2 (Bartelink 32); Ennod., v. Epiphan. 1; 3-4 {P.G. 41, pp. 24-5; 28-9); Krause, op. cit. (n. 66), III, 75; Beaucamp,op. cit. (n. 66), 48-52; S. R. Huebner, Der Klerus in der Gesellschaft des sp?tantiken Kleinasiens (2005), 50).190SB 5.7871: ??e?upov o? TT] (puaei. Cf. Kurylowicz, op. cit. (n. 57), 63.191P.Col.Zen. 3.58 (from 248 b.c.). See also above, Section m.192Also Roman legal sources indicate that adoption was practised in the Eastern provinces {Dig. 45.1.132; see alsoClust 8.47.4?6 (from A.D. 290-293); Clust 4.19.13 (from a.d. 293)). Cf. Kurylowicz, op. cit. (n. 57), 65. Krause

concludes: Der Befund der Inschriften und juristischen Quellen (deutet) doch auf eine gro?e Bereitschaft hin, sichfremder Kinder wie der Eigenen anzunehmen (sie etwa auch am Erbe partizipieren zu lassen). In diesem Kontextscheint es durchaus denkbar, da? Waisen in gr??erer Zahl, als es die juristischen Quellen nahelegen, an Kindes Stattin fremde (bzw. verwandte) Familien aufgenommen wurden. (Krause, op. cit. (n. 66), III, 84).193cf. Goody, op. cit. (n. 10), 4.


21/29

BROTHER-SISTER MARRIAGE IN ROMAN EGYPT 4I

from hird-century .C.Tholtis in the xyrhynchite ome, drawn up by Pra