Hostname: page-component-586b7cd67f-2brh9 Total loading time: 0 Render date: 2024-11-23T20:03:40.679Z Has data issue: false hasContentIssue false
  • English
  • Français

Child-Exposure in the Roman Empire*

Published online by Cambridge University Press:  24 September 2012

W. V. Harris
W. V. Harris
Affiliation:
Columbia
Get access Rights & Permissions [Opens in a new window]

Extract

The exposure of infants, very often but by no means always resulting in death, was widespread in many parts of the Roman Empire. This treatment was inflicted on large numbers of children whose physical viability and legitimacy were not in doubt. It was much the commonest, though not the only, way in which infants were killed, and in many, perhaps most, regions it was a familiar phenomenon. While there was some disapproval of child-exposure, it was widely accepted as unavoidable. Some, especially Stoics, disagreed, as did contemporary Judaism, insisting that all infants, or at least all viable and legitimate infants, should be kept alive. Exposure served to limit the size of families, but also to transfer potential labour from freedom to slavery (or at any rate to de facto slavery). Disapproval of exposure seems slowly to have gained ground. Then, after the sale of infants was authorized by Constantine in A.D. 313, the need for child-exposure somewhat diminished, and at last — probably in 374 — it was subjected to legal prohibition. But of course it did not cease.

Type
Articles
Information
The Journal of Roman Studies , Volume 84 , November 1994 , pp. 1 - 22
Copyright
Copyright ©W. V. Harris 1994. Exclusive Licence to Publish: The Society for the Promotion of Roman Studies

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1 Boswell, J., The Kindness of Strangers (1988), 25Google Scholar, preferring the term ‘abandonment’, objects to ‘exposure’ on the grounds that it ‘conveys a sense of risk or harm’ which is in his view absent from the terms ἔχθεσις and expositio. But innumerable texts that associate exposure with dreadful deaths or with slavery make it obvious that the Greek and Latin terms have very unpleasant connotations.

2 Of the early literature Armaroli, L., Ricerche storiche sulla esposizione degl'infanti presso gli antichi popoli e spedalmente presso i Romani (Venice, 1838)Google Scholar, is particularly impressive. Among numerous later works note G. Glotz, in Daremberg-Saglio, s.v. expositio (1892), 930–9 (G. Humbert on Rome, 939), with the essay in his Etudes sociales et juridiques sur l'antiquité grecque (1906), 187–227Google Scholar. The most useful discussion of the subject in general is Eyben, E., ‘Family planning in Graeco-Roman antiquity’, Ancient Society 11–12 (1980–1981), 5–82Google Scholar; see also Brunt, P. A., Italian Manpower, 225 B.C.–A.D.14 (1971), 148–54Google Scholar. On exposure in the Greek world prior to the Romans see especially A. Cameron, ‘The exposure of children and Greek ethics’, CR 46 (1932), 105–14, Tolles, R., Untersuchungen zur Kindesaussetzung bei den Griechen (1941)Google Scholar.

3 Pomeroy, S. B., ‘Copronyms and the exposure of infants in Egypt’, in Studies in Roman Law in Memory of A. Arthur Schiller (1986), 147–62Google Scholar, Kudlien, F., ‘Kindes-aussetzung im antiken Roman: ein Thema zwischen Fiktionalität und Lebenswirklichkeit’, in Groningen Colloquia on the Ancient Novel II (1989), 25–44Google Scholar, Huys, A., ‘ἔχθεσις and ἀπόθεσις:the terminology of infant exposure in Greek antiquity’, AC 58 (1989), 190–7Google Scholar, M. Memmer, ‘Ad servitutem aut ad lupanar …’, ZSS 108 (1991), 21–93.

4 D. Engels (‘The problem of female infanticide in the Greco-Roman world’, CPh 75 (1980), 112–20; ‘The use of historical demography in ancient history’, CQ 34 (1984), 386–93) ignores almost all this evidence (for his demographic theory, see below, p. 18). Doubt is more legitimate about the extent of child-exposure in Greece before 400 B.C.: see, for example, Gallo, L., ‘Un problema di demografia greca: la donna tra la nascita e la morte’, Opus 3 (1984), 37–62Google Scholar.

5 Kaser, M., Das römische Privatrecht II 2 (1975), 204Google Scholar, says orientalistically that exposure was seldom practised at Rome but was widespread in ‘the East’. M. Golden writes that there can be ‘reasonable doubt’ that child-exposure was extensive (in the ancient world in general) (‘Did the ancients care when their children died?’, G&R 35 (1988), at 158). Kudlien's account (see n. 3) is balanced but seems quite doubtful (27). Parkin, T. G., Demography and Roman Society (1992), 97Google Scholar, concludes vaguely that ‘the evidence does not allow us to generalize on this practice [it is not clear whether he means exposure or infanticide] as a social phenomenon’. M. Schmidt has maintained, contrary to a widespread view, that many handicapped infants were allowed to live, and this may be right (‘Hephaistos lebt — Untersuchungen zur Frage der Behandlung behinderter Kinder in der Antike’, Hephaistos 5–6 (1983–1984), 133–61Google Scholar).

6 Motomura, R., ‘The practice of exposing infants and its effects on the development of slavery in the ancient world’, in Yuge, T. and Doi, M. (eds), Forms of Control and Subordination in Antiquity (1988), 410–15Google Scholar; Boswell, op. cit. (n. 1), 42, 128–31.

7 Riddle, J. M., Contraception and Abortion from the Ancient World to the Renaissance (1992).Google Scholar

8 ‘Towards a study of the Roman slave trade’, MAAR 36 (1980), 123–4 (cf. P. Veyne, ‘La famille et l'amour sous le Haut-empire romain’, Annales ESC 33 (1978), 46); with respect to Egypt: Biezuńska-Malowist, I., ‘Die Expositio von Kindern als Quelle der Sklavenbeschaffung im griechisch-römischen Aegypten’, Jahrbuch für Wirtschaftsgeschichte (1971), 2, 129–33Google Scholar, L'Esclavage dans l'Égypte gréco-romaine 11 (1977), 22–6Google Scholar.

9 Recent work: Enfance abandonnée et société en Europe, XIVe–XXe siècle. Actes du colloque … 1987 (1991); Kertzer, D. I., Sacrificed for Honor: Italian Infant Abandonment and the Politics of Reproductive Control (1993)Google Scholar.

10 Boswell, op. cit. (n. 1), 431.

11 Cappelletto, G., ‘Infanzia abbandonata eruoli di mediazione sociale nella Verona del Settecento’, Quaderni Storici 18 [53] (1983), 421–43Google Scholar, and F. Doriguzzi, ‘I messaggi dell'abbandono: bambini esposti a Torino nel ‘700’, ibid., 445–68.

12 Cameron, however, op. cit. (n.2), 105, wrote that ‘the cruelty involved in infanticide even by exposure is very slight’.

13 Dickemann, M., ‘Concepts and classification in the study of human infanticide’, in Hausfater, G. and Hrdy, S. B., Infanticide: Comparative and Evolutionary Perspectives (1984), at 428Google Scholar.

14 For a balanced view of Roman attitudes towards children, at least in one region, see Garnsey, P., ‘Child rearing in ancient Italy’, in Kertzer, D. I. and Sailer, R. P. (eds), The Family in Italy from Antiquity to the Present (1991), 48–65Google Scholar, who warns (49) against the pervasive assumption that all pre-modern societies shared callous attitudes towards children. As to when childhood was ‘discovered’ by the Romans (cf. Manson, M., ‘The emergence of the small child in Rome’, History of Education 12 (1983), 149–59CrossRefGoogle Scholar), there is still much to be said.

15 Dion.Hal. 11.15.2, as interpreted by Delcourt, M., Stérilités mystérieuses et naissances maléfiques dans l'antiquité classique (1938), 50Google Scholar.

16 Harris, W. V., ‘The Roman father's power of life and death’, in Studies in Roman Law in Memory of A. Arthur Schiller (1986), 81–95Google Scholar.

17 Hence some disagreement with C. Patterson, ‘“Not worth the rearing”: the causes of infant exposure in ancient Greece’, TAPhA 115 (1985), 104, who rightly observes that there were ‘complex social, economic and psychological variables’, — also, one might add, legal, religious, and medical ones. Cf. Dickemann, op. cit. (n. 13), 433–4, for a reasonable warning against being preoccupied with numerical aspects of the problem.

18 Dickeman, M., ‘Demographic consequences of infanticide in man’, Annual Review of Ecology and Systematics 6 (1975), 130CrossRefGoogle Scholar.

19 Gilbert Murray listed the evidence from comedy, ‘Ritual elements in the New Comedy’, CQ 37 (1943), 46–54.

20 cf. Kudlien, op. cit. (n. 3), 28.

21 cf. Bresson, A., ‘Démographie grecque antique et modèles statistiques’, Revue, informatique et statistique dans les sciences humaines 21 (1985)Google Scholar, at 15.

22 Glotz, op. cit. (n. 2), 932.

23 Plin., Ep. X.65–6, 72.

24 Soranus, ed. P. Burguière, D. Gourévitch and Y. Malinas, II (1990), p. 85. This reticence was apparently a Greek tradition: cf. Plato, Rep. V. 460c, Delcourt, op. cit. (n. 15), 66 n. 1.

25 Patterson, op. cit. (n. 17), Gallo, op. cit. (n. 4), 40–1 (referring to Athens).

26 Pol. VII.16.1335b 19–26, with the text and interpretation proposed by Viljoen, G., ‘Plato and Aristotle on the exposure of infants at Athens’, Acta Classica 2 (1959), 66–8Google Scholar (a doxography would be irrelevant, but note that Ross's text and J. Aubonnet's translation (1986) are not to accepted). See also Cameron, op. cit. (n. 2), 109.

27 Theopomp., FGrH 115 F204, Aristotle, Zoika fr. 283 Rose.

28 Fr. II Kock = 12 Kassel-Austin, cf. Ter., Heaut. 626.

29 Patterson, op. cit. (n. 17), 119–21, says that this text is ‘urban’, which seems a simplification and is not very relevant. Gallo, op. cit. (n. 4), 52, ignores its significance for the fourth century, relegating it to his Hellenistic section; but it is not very likely that exposure had very recently become common at Athens when Poseidippus wrote these lines at some date in the decades around 300. See further Eyben, op. cit. (n. 2), 17.

30 Petr., Sat. 116.7–8. For a somewhat similar exaggeration, cf. Sen., Contr. X.4.10 end (‘omnes omnibus…’)

31 e.g. Gallo, op. cit. (n. 4), 52–4.

32 For : Eyben, op. cit. (n. 2), 17, Pomeroy, S. B., ‘Infanticide in Hellenistic Greece’, in Cameron, A. and Kuhrt, A. (eds), Images of Women in Antiquity (1983), 207–22Google Scholar; against: Gallo, op. cit. (n.4), 37–62.

33 Cic., De leg. 111.8.19, where the MSS read legatus, which has been emended to necatus (Puteanus, followed by Ziegler) and delatus (Delcourt, op. cit. (n. 15), 51). Cicero's reference is incidental and he may not have been quoting verbatim.

34 cf. Sen., Contr. IX.3.11 — which is not to say that such an act of exposure would have been unusual.

35 For the archaic principle of exposing twins see Sen., Contr. IX.3, Delcourt, op. cit. (n. 15), 103–4, S. Scrimshaw, ‘Infanticide in human populations: societal and individual concerns’, in Hausfater and Hrdy, op. cit. (n. 13), 446, etc. The explanation of the name Vopiscus given by Plin., NH VII.47 may suggest that it had once been common to kill one of a pair of twins.

36 Plaut., Amph. 499–501, Cas. 39–44, Cist. 120–202, True. 399, Ter., Andr. 215–24, Heaut. 614–67, Hec. 400, Phorm. 647, Afran., com. 347.

37 It is not likely that Tertullian, Ad Nat. 1.15.3, is an allusion, though Mommsen considered this to be probable (Römisches Strafrecht (1899), 619 n.3). Much the best reading, incidentally, is not ae[ditui] (Klussmann, Borleffs) but ae[dilis] (Reifferscheid).

38 cf. R. Jackson, Doctors and Diseases in the Roman Empire (1988), 107.

39 It is Hellenistic according to E. Weiss, in RE s.v. Kinderaussetzung (1921), col. 466.

40 Harris, op. cit. (n.8), 123.

41 Cic., Att. XI.9.3: it was at least conceivable to Cicero that he might not have been susceptus. See also, e.g., II Verr. 3.161, 5.123, Cluent. 179, Post red. ad Quir. 2, De domo 34, 36, Phil. 111.17.

42 Suet., De gramm. 21. Another grammaticus who had been exposed in Gaul: ibid. 7.

43 This argument is advanced by Sailer, R. P. in Gabba, E. and Schiavone, A. (eds), Storia di Roma IV (1989), 537Google Scholar. Riddle, op. cit. (n. 7), 11, incorrectly attributes this position to me.

44 On Dio LIV.16 see below, n. 94. In the article ‘Abandonment and exposure’, in Hastings' Encyclopaedia of Religion and Ethics I (1905), 4Google Scholar, P. Giles argued that the Roman way of naming women presupposes the exposure of daughters.

45 This passage is an exaggeration but not a straight forward falsehood (pace Engels, op. cit. (n.4), 393), rather like the statement about population in Polyb. XXXVI.17. Cf. above on Petr., Sat. 116.7–8. It has to be repeated (cf. CQ 32 (1982), 116) that the Loeb translator of this passage, W. C. Helmbold, misunderstood it (‘when poor men do not rear their children, it is because…’; this is followed by Boswell, op. cit. (n. 1), 103), presumably because he was unwilling to face the unpleasant reality it describes.

46 Mosès, A., intro. to De Specialibus Legibus III–IV (1970), 30–1Google Scholar.

47 For Philo's opinions see more briefly On Virtues 131–3, which seems to be the earliest existing text to say that exposure is contrary to nature (132).

48 Note ‘inter eas provincias’, X.66.2. The law about alitnenta in causae liberates still needs discussion (see esp. Amelotti, M., Per l'interpretazione della legislazione privatistica di Diocleziano (1960), 132–9Google Scholar, M. Bianchi Fossati Vanzetti, ‘Vendita ed esposizione degli infanti da Costantino a Giustiniano’, SDHI 49 (1983), at 185–6).

49 On the Stoic background see van Geytenbeek, A. C., Musonius Rufus and Greek Diatribe (1963), 82–3.Google Scholar

50 Diodorus 1.80.3 (Egyptians), XL.3.8 = Hecataeus Abd., FGrH 264 F6 (Jews), Strabo XVII.824 (Egyptians), Tac., Hist. V.5 (the Jews consider it nefas to kill agnati, presumably meaning extra children), Germ. 19 (‘numerum liberorum finire aut quemquam ex agnatis necare flagitium habetur’ — which is explicitly contrasted with Roman practice), Dio LXXVI/LXXVII.12 (Scots). See also Dion.Hal. IX.22.2 on early Rome.

51 Philo, op. cit. (n.47); Jos., Contra Apionem 11.202.

52 Cameron, op. cit. (n. 2), 112–13. He concludes that ‘writers subsequent to Aristotle who report the custom [of not exposing infants] are drawing not on experience but on literary tradition’. But Strabo on the Egyptians may easily have had a good source, and the argument that Philo's and Josephus' assertions derive from Hecataeus is not credible.

53 A different story about how the Germans treated new-born children: Galen, De san. tuenda 1.10 (VI.51 Kühn).

54 Apul., Met. X.23, Longus 1.2–3, etc.

55 , Justin, First Apology 27, 29Google Scholar, Tert., Ad nat. 1.16, Apol. 9.17–18.

56 The famous texts are P. Oxy. 1.37–38 (A.0.49), IV.744 (1 B.C.). Many other papyri refer to exposure: O. Montevecchi, ‘I paragrafi 41 e 107 dello Gnomon dell'idios logos: implicazioni socio-culturali e demografiche’, Atti del XVII congresso internazionale di papirologia (1984), 111, 966 n. 3.

57 Above, n. 8.

58 Pomeroy, op. cit. (n. 3), arguing against the theory of P. Perdrizet, REA 23 (1921), 85–94. On Sterc- names in the African provinces see Lassère, J.-M., Ubique Populus (1977), 504Google Scholar. I am inclined to accept the argument of D. Hobson, ‘Naming Practices in Roman Egypt’, BASP 26 (1989), 157–74, at 163–5, that copronyms were ‘derogatory-protective’ names.

59 Sect. 41, which taxes Egyptians wh o adopted boys (probably Greeks for the most part ) who had been rescued ἐχ χοπρίας is easier to understand. See Montevecchi, op. cit. (n. 56), for bibliography and discussion.

60 Harris, MAAR 36 (1980), 123; Deonna, W., ‘Cimetières de bébés’, Revue archéologique de l'Est et du Centre-Est 6 (1955), 231–47Google Scholar; Plin., NH VII.72 (see also Juv. XV. 139–40).

61 Garland, R., The Greek Way of Death (1985), 78–9Google Scholar.

62 Justin, First Apology 27, asserts that it was practised χατὰ πᾶν ἐθνος.

63 Quoted by Sen., Contr. X.4.21.

64 Math. VII.2.9, 11, 12, 13, 20, 21; cf. Paus. 1.43.7.

65 See for example Aelian, VH XII.42 (a list); Longus 1.2 and 1.5; Redford, D. B., ‘The literary motif of the exposed child’, Numen 14 (1967), 213–14CrossRefGoogle Scholar. Some mocked the ‘melodramatic absurdity’ of the wolf that offered its dugs to Romulus and Remus: so says Dionysius 1.84.1. On reports of children raised by wild animals see Armen, J.-C., Gazelle-Boy. A Child Brought up by Gazelles in the Sahara Desert (1974)Google Scholar.

66 Festus 105 Lindsay. The reference to ‘loco celebri’ in Ps.-Quint., Decl. 306.24 has been bracketed by recent editors.

67 The meaning is not certain (cf. Eyben, op. cit. (n. 2), 54 n. 163, and E. Courtney's comm.), and possibly they were not outside the city.

68 Eyben, op. cit. (n. 2), 17, may have been right to hold that the infant was often abandoned ‘at a much-frequented spot’, but Boswell, op. cit. (n. 1), 25, and Memmer, op. cit. (n. 3), 23, go too far in claiming it as a general rule that infants were exposed where they could easily be discovered (Juvenal provides no evidence of this).

69 Longus 1.4, etc. The alert reader may have thought of Eur., Ion 938, 958.

70 Köves-Zulauf, T., Römische Geburtsriten (1990), 20–4Google Scholar.

71 Hähnle, A., Γνωρίσματα (1929)Google Scholar. He emphasizes their function as amulets.

72 cf. Dio Chrys. IV.25, with Eyben, op. cit. (n. 2), 19 n.48.

73 See Longus IV.24.1 (ἐντάΦια), Glotz, op. cit. (n. 2), 934, Cameron, op. cit. (n. 2), 107.

74 This hope is derided by Philo (see above). Hope for pity: Tert., Ad nat. 1.16.10, Paulus in Dig. XXV.3.4, Lactant., Inst. VI.20.22.

75 But concerning Isis see below, p. 16.

76 Glotz, op. cit. (n. 2), 934; but see also Dio Chrys. XV.8 (cf. Juv. VI.602–3), Boswell, op. cit. (n. 1), 74 n. 71.

77 In Boswell's view, op. cit. (n. 1), 128, ‘most expositi in fiction do not even know they were abandoned, which suggests a general presumption that exposed children were brought u p as adoptees’. But since fiction requires surprise, the expositi naturally cannot be in the know.

78 Nani, T. G., ‘ΘΡΕΙΙΤΟΙ’, Epigraphica 5–6 (1943–1944), 45–84Google Scholar.

79 Cameron, op. cit. (n. 2), 105 (‘extremely rare’). Similarly Brunt, P. A., Studies in Greek History and Thought (1993). 350 n. 18Google Scholar.

80 Other evidence that this really happened: Plin., Ep. X.65–6, Suet., De gramm. 21. But Boswell, op. cit. (n. 1), 74, goes much too far in saying that parents ‘frequently’ reclaimed exposed children; he makes insufficient allowance for the element of fantasy in the literary sources. Flavia Domitilla, wife of Vespasian, is cited as an instance by P. Veyne, Latomus 21 (1962), 50 n. 2, on the basis of Suet., Vesp. 3, but his alternative explanation of the text is better.

81 First Apology 27; 29 also gives the impression that most but not all survived.

82 Boswell, op. cit. (n. 1), 43; Sailer, op. cit. (n.43), 537, who adds that it would have suited Justin's case still better if the exposed could have been said to die in most cases. Kudlien, op. cit. (n. 3), 30 supposes that those who were exposed on rubbish dump s were more likely to be rescued than not; see also Memmer, op. cit. (n. 3), 22–3.

83 See also Clem. Alex., Paed. 111.3.21.5 (fathers ‘often’ have unwitting sexual relations with their sons and daughters; sexual promiscuity rather than child-exposure is at the centre of his attention here), Min. Fel., Oct. 31.4, and the references in n. 55 above.

84 cf. Glotz, op. cit. (n. 2), 934. Longus balances this against the statement of the other ‘exposing father’ in the case, who says in effect that he expected his child to die (IV.24). There is probably some intentional irony in the fact that it is the father of the boy who expected the infant to die and the father of the girl who expected his child to live.

85 Boswell, op. cit. (n. 1), 136, Memmer, op. cit. (n. 3), 23.

86 cf. Firm. Mat., Math. VII.2.14, with Cappelletto, op. cit. (n. 11), 421, Kertzer, op. cit. (n.9), 138–44.

87 cf. Ter., Heaut. 641–2.

88 Philo, On Special Laws 111.115, Longus 1.2.1, Aelian 11.7 (but this text seems to distinguish between forms of exposure: one involving a definitely lethal intention and a deserted place, the other not), Suda s.v. ἐχτίθεναι τὰ βρέΦη 11.233 Adler, etc.

89 cf. Dion.Hal. 11.15.2, as well as the sources referred to in the text.

90 The lois sacrées mentioned in n. 141 imply that child-exposure was fatal. Athenagoras, Supplicatto 35.6, shows that Christians equated exposure and infanticide, which would have made no sense if the exposed had often survived. In both Tac., Germ. 19 and Apul., Met. X.23 the killing is presumably thought of as taking place by means of exposure.

91 Ps.-Quint., Decl. 306.22.

92 Dig. XXV.3.4.

93 To support the view that it was unusual for an exposed child to die, Boswell, op. cit. (n. 1), 128–31, cites some texts that are not germane: Ps.-Quint., Decl. 306.24, Chariton II.8–10, and Tert., Ad nat. 1.16.10. The latter passage does not show that Tertullian thought the exposed were generally rescued: it is his reprise of the ‘unconscious incest’ topos, and I.15.4 indicates that he knew how dangerous exposure was (‘quod frigore et fame aut bes<tiis si exp>oniti s…’; Borleffs' text).

94 Brunt, op. cit. (n. 2), 151 (but I do not think that in Dio LIV. 16 τὸ εὐγενεῖς means ‘the free-born population’), Rousselle, A., Porneia: de la maîtrise du corps à la privation sensorielle (1983), 70Google Scholar, Pomeroy, op. cit. (n. 3), 161, Wiedemann, T., Adults and Children in the Roman Empire (1989), 36Google Scholar.

95 Gallo, op. cit. (n.4), 37.

96 W. V. Harris, ‘The theoretical possibility of extensive infanticide in the Graeco-Roman world’, CQ 32 (1982), 114–16.

97 Dion. Hal. II.15.2, discussed earlier, Lucian, Hetainkoi dialogoi 2.1. Ov., Met. IX.676–9 and Apul. X.23 refer to instructions to kill new-born girls, which is probably understood to mean exposure. See further Eyben, op. cit. (n. 2), 16n.43.

98 See Gallo, op. cit. (n. 4), 45, Bagnall, R. S. and Frier, B. W., The Demography of Roman Egypt (forthcoming, Cambridge, 1994)CrossRefGoogle Scholar.

99 The question is raised by Parkin, op. cit. (n. 5), 96.

100 Still less, with Engels, op. cit. (n. 4), 386 n. 1, that ‘in general, the practice of exposure was restricted to deformed infants’.

101 Cic., De leg. III. 19, on which see above, The ‘Law of Romulus’ referred to earlier simply said in Dionysius' version (1.15.2) that deformed children under three were to be put to death. Incidentally Cicero says nothing here about the involvement of the tribunes, in spite of den Boer, W., Private Morality in Greece and Rome (1979), 99Google Scholar.

102 Sen., De ira 1.15.2 (‘liberos quoque, si debiles monstrosique editi sunt, mergimus’). Cf. Tibull. 11.5.80, Philo, On Special Laws III. 114, and later Tert., Ad nat. 1.15.4 (but this is not very valuable evidence), Firm. Mat., Math. VII.2.10–11.

103 Liv. XXVII.37.5–6, Obsequens 22, 27a, 32, 34, 36, etc. There were variations: an apparent pair of Siamese twins were burned in 136 B.C., and their ashes thrown into the sea (Obsequens 25).

104 Rousselle, op. cit. (n. 94), 69.

105 Plin., NH VII.34. See further Marquardt, J., Das Privatleben der Römer I (1886), 152–3Google Scholar, Delcourt, op. cit. (n. 15), 59–61.

106 Schmidt, op. cit. (n. 5), Grassl, H., ‘Behinderte in der Antike. Bemerkungen zur sozialen Stellung und Integration’, Tyche 1 (1986), 118–26Google Scholar.

107 To summarize: the infant's mother must have enjoyed good health during pregnancy, the birth must not be premature, the infant must cry vigorously, all its limbs and organs must be sound, its sense organs must work, its orifices must all open, the movements of each part of the body must be neither sluggish nor weak, and the articulation of the limbs must be correct.

108 e.g. Ov., Her. XI.84; Heliod., Aeth. IV.8 (where the illegitimacy was merely apparent).

109 Syme, R., ‘Bastards in the Roman aristocracy’, Proc. Amer. Philosophical Soc. 104 (1960)Google Scholar, cited from Roman Papers II (1979), 511, 513Google Scholar. In Brunt's account, op. cit. (n. 2), 150, exposure and infanticide form part of the explanation.

110 Engels, op. cit. (n. 4), 386.

111 Patterson, op. cit. (n. 17), 117–18.

112 cf. Garnsey, P., Famine and Food-Supply in the Graeco-Roman World (1988)CrossRefGoogle Scholar.

113 See Plin., Pan. 26.5, Lactant., Inst. VI.20.24–5, as well as Aelian 11.7 on the Theban law. The Ephesian law (date uncertain): Proclus on Hes., Works and Days 496–7 (Plu., Mor. (Loeb edn) XV.165).

114 Pan. 26.5–7: ‘locupletes ad tollendos liberos ingentia praemia et pares poenae cohortantur, pauperibus educandi una ratio est bonus princeps… haec prima parvulorum civium vox aures tuas imbuit’, etc.

115 See also Ps.-Quint., Decl. 306.24.

116 This is commonly taken (e.g. by Kudlien, op. cit. (n. 3), 41) to mean that he wished to avoid splitting his fortune.

117 For debate see esp. Hopkins, K., ‘The age of Roman girls at marriage’, Population Studies 18 (1964–1965), 309–27CrossRefGoogle Scholar, Shaw, B. D., ‘The age of Roman girls at marriage: some reconsiderations’, JRS 77 (1987), 30–46Google Scholar (a some what different view), P. Morizot, ‘Remarques sur l'âge du mariage des jeunes Romaines en Italie et en Afrique’, CRAI 1989, 656–68.

118 Brunt, op. cit. (n. 2), 137–8, 151–2.

119 cf. Hopkins, op. cit. (n. 117), 314–15.

120 Brulé, P. holds that in Greece children born to slave women were generally exposed if the owner was the father (‘Infanticide et abandon d'enfants’, Dialogues d'histoire ancienne 18 (1992)CrossRefGoogle Scholar, at 84), but this is quite unproved.

121 Suet., Aug. 94. Cassius Dio (XLV. I) tells the tale that because of an evil prophecy Augustus' father thought of destroying him — perhaps not by exposure — very shortly after his birth.

122 Suet., Cal. 5 (‘quo defunctus est die, lapidata sunt templa, subversae deum arae, Lares a quibusdam familiares in publicum abiecti, partus coniugum expositi’). This is dismissed as rhetoric in the commentary of D. W. Hurley (1993).

123 It would have been children who were born that day who suffered (Delcourt, op. cit. (n. 15), 63).

124 For discussion see Versnel, H. S., ‘Destruction, devotio and despair’, in Studi in onore di Angela Brelich (1980), 541–618Google Scholar. The key question for a reader of Versnel's rich article must be whether Germanicus' death is likely to have produced real despair, as Tacitus also says, at length (‘nihil spei reliquum clamitabant’, Ann. III.4); see esp. Versnel, 542–55, 617.

125 Sen., De clem. 1.13.5. Tyranny: Rhet. Lat. Min.,ed. Halm, p.343.10–11.

126 See Scaevola in Dig. XL.4.29.

127 Gallo, op. cit. (n.4), 38–9, perhaps gave a higher estimate of the effectiveness of Greek contraception than was justified at the time he wrote. Parkin, op. cit. (n. 5), 126, already seems superseded.

128 There were serious impediment s of course, including the probably widespread notion that the most fertile time in the menstrual cycle was in the last days of menstruation (Soranus 1.36).

129 There are still unnoticed references to this subject in various authors. What else, for example, can Argentarius have been thinking of in Anth. Pal. V. 104.6 ( = Garland of Philip 1328 Gow-Page)?

130 Ov., Amores 11.13, Plu., Lyc. 3.

131 cf. Eyben, op. cit. (n. 2), 76.

132 Cameron, op. cit. (n. 2), 113, says that Greek ethnographical statements about peoples who bring up all their children are criticisms of Greek practice, but this is seldom if ever clear.

133 Watson, A., The Law of Persons in the Later Roman Republic (1967), 77–82Google Scholar. It hardly needs saying that many exposing fathers were still under their own fathers' potestas, or that many others were non-citizens.

134 Brunt, op. cit. (n. 2), 558–66. Is Aen. VI.428 (‘ab ubere raptos’) an allusion to expositi?

135 In this as in many other cases in which criticism is voiced, it is unclear whether all infants whatsoever, including the handicapped, are meant to be raised. Eyben, op. cit. (n. 2), 55, perhaps overstates the opposition to exposure in the Julio-Claudian period because he misunderstands Sen., Contr. IX.3 and 4, supposing that those who reacted with horror to the deliberate maiming of the exposed were critical of exposure as such, which as far as this text is concerned they clearly were not.

136 The former claimed that Epicurus had said ‘Let us not raise children’, and this has become part of a fragment (525 Usener).

137 Pan. 26.5–7, quoted in n. 114.

138 Another topic with a large bibliography: see esp. Duncan-Jones, R., The Economy of the Roman Empire: Quantitative Studies (1974), 288–319Google Scholar; Mrozek, S., ‘Die privaten Alimentarstiftungen in der römischen Kaiserzeit’, in Kloft, H. (ed.), Sozialmassnahmen und Fürsorge (1988), 155–66Google Scholar. The earliest evidence, from Atina in Latium, is Neronian or a bit earlier (ILS 977). For such benefactions in Greek cities in the second century, see Jones, C. P., ‘Eastern alimenta and an inscription of Attaleia’, JHS 109 (1989), 189–91CrossRefGoogle Scholar.

139 1.15.3: ‘vos quoque infanticidae, qui infantes editos enecantes legibus quidem prohibemini, sed nullae magis leges tarn impune tarn secure… eluduntur’. He must be talking about Roman statutes, pace Boswell, op. cit. (n. 1), 60 n. 16.

140 Vatin, C., Recherches sur le mariage et la condition de la femme mariée a l'époque hellénistique (1970), 235Google Scholar, who does not distinguish carefully between child-exposure and other practices. The evidence is LSCG Supplément No.119 l. 77 (Ptolemais, first century B.C.), and LSAM No.84 ll. 3–4 (Smyrna, second century A.D.).

141 On evidence of this type see Cameron, op. cit. (n. 2), 108.

142 Ov., Met. IX.685–701 (‘nec dubites… tollere quid-quid erit’), Plu., De Is. et Os. 14 = Mor. 356f. (cf. Maroi, F., ‘Intorno all'adozione degli esposti in Egitto romano’, in Raccolta di scritti in onore di Giacomo Lumbroso (1925), at 385Google Scholar). In Egypt itself: Ranke, H., Die ägyptischen Personennamen 11 (1952), 380Google Scholar, Pomeroy, op. cit. (n.4), 160.

143 Boswell, op. cit. (n. 1), 90.

144 Ter., Heaut. 626–52. Cf. Ov., Met. IX.680–4. Husbands may quite often have been deceived as to whether the child had been exposed: cf. Ps.-Quint., Decl. 306.4.

145 cf. G. Glotz in DS s.v. infanticidium, 490.

146 Inst. VI.20.20; cf. V.9.15.

147 On Jewish disapproval see, e.g., B. Schöpf, Das Tötungsrecht bei den frühchristlichen Schriftstellern (1958), 120–3.

148 See nn. 55, 83. See also Apocal. Petri 8 (Ethiopic, translated in E. Hennecke (ed.), Neutestamentliche Apokryphen2 (1924), 322)Google Scholar; Epist. ad Diognetum 5.6; Athenagoras, Supplicatio 35.6; Clem. Alex., Strom. 11.18, 92–3, V.14 (Eclog. proph. 41 explains how the exposed will be saved); Orig., Contra Celsum VIII.55; Min. Fel., Oct. 30.2; Orac. Sib. 11.282. Cf. Schöpf, op. cit. (n. 147), 124–42. Kikillus, G. T. T., De Invloed van het Christendom op de Romeinsche wetgeving ten opzichte van de zorg voorhet kind (1924)Google Scholar, contributed very tittle.

149 Christians prior to Constantine's time may sometimes have exposed infants, but (pace Boswell, op. cit. (n. 5), 3) there seems to be no evidence that they did.

150 Parkin, op. cit. (n. 5), 84, with the comments by B. W. Frier, Bryn Mawr Classical Review 3 (1992), 385.

151 This applies even to nineteenth-century western Europe: S. C. Watkins, ‘Demographic nationalism in western Europe, 1870–1960’, in J. R. Gillis et al. (eds), The European Experience of Declining Fertility (1992), 272–6.

152 See n.96, and also Bresson, op. cit. (n. 21), 7–34, Parkin, op. cit. (n. 5), 95. Cf. Golden, M., ‘Demography and the exposure of girls at Athens’, Phoenix 35 (1981), 316–31CrossRefGoogle Scholar.

153 Parkin, op. cit. (n. 5), 147.

154 B. W. Frier, ‘Roman life expectancy: Ulpian's evidence’, HSCPh 86 (1982), 245, adapted in Parkin, op. cit.(n.5). 144.

155 But some constituent populations of the Empire, Greeks for instance, may not have reproduced themselves.

156 Parkin, op. cit. (n. 5), 86.

157 See Parkin, op. cit. (n. 5), fig. 10 (p. 160), which derives from the concept employed by Wrigley, E. A. and Schofield, R. S., The Population History of England, 1541–1871 (1981), in their fig. 7.10 (p. 239)Google Scholar.

158 In colonial New England, with female age at marriage about twenty-one, average completed family sizes between 7.32 and 9.3 children are quoted: Marcy, P., ‘Factors affecting the fecundity and fertility of historical populations’, Journ. of Family History 6 (1981), 310CrossRefGoogle ScholarPubMed.

159 The U.N. Statistical Office's Demographic Yearbook 17 (1965)Google Scholar gives figures for two countries with a median female age at marriage lower than twenty for which it also provides a GRR of the same or virtually the same date (in the 1960s): Jordan, GRR 3.4, and Tunisia, GRR 3.1. In both countries the average age at marriage was far higher than in the Roman Empire. In both countries it can be assumed that there was a certain amount of fertility limitation.

160 Engels, op. cit. (n.4).

161 The rate of natural increase is the birth-rate minus the death-rate.

162 These are not the only possibilities. Child-exposure is quite likely to have increased fertility by curtailing lactation.

163 See above.

164 cf. Bradley, K. R., ‘On the Roman slave supply and slavebreeding’, Slavery and Abolition 8 (1987), 42–64CrossRefGoogle Scholar, who, however, scarcely deals with child-exposure.

165 No need to discuss here how much of the trade in slaves was in the hands of specialists. We can count as a slave-dealer anyone who acquires slaves with the intention of selling them, whatever other occupation he may have.

166 Masciadri, M. Manca and Montevecchi, O., have shown (‘Contratti di baliatico e vendite fiduciarie a Tebtynis’, Aegyptus 62 (1982), 148–61Google Scholar; I contratti di baliatico (1984), 14–16) that the nursing contracts from Tebtunis are probably disguised loan contracts, but those from more Hellenized places — in the surviving evidence this means mainly Alexandria and Oxyrhynchus — are agreed to be genuine nursing contracts.

167 cf. Glotz, op. cit. (n. 2), 935; Ter., Heaut. 640.

168 Engels, op. cit. (n. 4), 391, 393. On the economics of raising slaves in this fashion see Manca Masciadri and Montevecchi, op. cit. (n. 166, 1984), 19, who, however, take too little account of the inevitably heavy mortality of the children who were contracted out.

169 See the Delphic inscription published in BCH 1893, 383 No. 80 (first century B.C. ); Aelian 11.7.

170 Bianchi Fossati Vanzetti, op. cit. (n.48), 187, 199.

171 It is enough to cite Weiss in RE s.v. Kinderaussetzung(1921), col. 467, de Dominicis, M. A., ‘Satura critica sulle fonti postclassiche’, in Studi in on ore di Edoardo Volterra (1971), 1, 540Google Scholar, Kaser, op. cit. (n. 5), 11.204 n. 17, T. Mayer-Maly in KP s.v. Kinderaussetzung (1969), 214, Eyben, op. cit. (n.2), 31, Memmer, op. cit. (n. 3), 69. Authentic: A. Mau in RE s.v. Aussetzung (1896), col. 2589, M. Radin, ‘The exposure of infants in Roman law and practice’, CJ 20 (1924–1925), 339–40.

172 The absence of this sentiment from our other sources for the Sententiae is of no consequence since they are so brief. The more commonly invoked argument is simply that the text contradicts what we know about the law concerning exposure in Severan times. I leave aside here the related problem raised by Sent. III.4B.2.

173 cf. Volterra, E., ‘Sull'uso delle Sententiae di Paolo’, in Atti del Congresso internazionale di diritto romano, Bologna I (1934), 162–5Google Scholar.

174 In particular it is almost impossible to think that Dig. XXV.3.4 can have been written before the Constantinian edict contained in C.Th. V.10.1 (which is discussed later on in the text).

175 The MS date is 13 May 315. Seeck, O., Regesten der Kaiser und Päpste (1919), 54Google Scholar, moved it to 329, and since it has subsequently emerged (cf. Barnes, T. D., ‘Lactantius and Constantine’, JRS 63 (1973), 36Google Scholar, based on numismatic work by P. Bruun) that in 315 Constantine did not yet control Naissus, where the edict is supposed to have been issued, the case for 329 is still stronger. ‘IMP. CONSTANTINU S A. AD ABLAVIUM. Aereis tabulis velcerussatis aut linteis mappis scribta per omnes civitates Italiae proponatur lex, quae parentum manus a parricidio arceat votumque vertat in melius. Officiumque tuum haec cura perstringat, ut, si quis parens adferat subolem, quam pro paupertate educare non possit, nec in alimentis nec in veste inpertienda tardetur …’.

176 On this see Evans-Grubb, J. A., ‘Munita coniugia’: the Emperor Constantine's Legislation on Marriage and the Family (unpub. diss., Stanford, 1987), 183–4Google Scholar.

177 But still quite vague: ‘et universis, quos adverterint in egestate miserabili constitutos, stipem necessariam largiantur atque ex horreis substantiam protinus tribuant competentem’.

178 It is evident that there were no imperial foundations of the old kind: Evans-Grubb, op. cit. (n. 176), 185, and in Harries, J. and Wood, I. (eds), The Theodosian Code. Studies in the Imperial Law of Late Antiquity (1993), 135Google Scholar. Memmer, op. cit. (n. 3), 61, has this wrong, and in support of his view quotes an imaginary inscription from the Arch of Constantine.

179 The MS date of the latter is 18 August 329. Seeck, op. cit. (n. 175), 65, wished to move it to 319 or 320, but this remains speculative. The arguments of Vanzetti, op. cit. (n.48), 197–8, for supposing that this did not refer to child-exposure are trivial. See Memmer, op. cit. (n. 3), 62–4.

180 cf. above n. 175.

181 It could also be resolved by dating C.Th. V.10.1 before XI.27.1 (as Seeck, op. cit. (n. 175), 65 does for other reasons) and supposing that in the interval the emperor had decided to treat child-exposure as parricidium. But this seems much less likely, not least because C.Th. V.9.1 of 331, though it nullifies the rights of exposing fathers, shows no sign of regarding them as parricidae.

182 Vanzetti, op. cit. (n. 48), 200 and Memmer, op. cit. (n. 3), 67 respectively. Vanzetti, 202, 211, must be alone in thinking that there was no significant change in the law regarding child-exposure between 331 and 529.

183 Vanzetti, op. cit. (n.48), 199–201.

184 Harris, op. cit. (n. 16), 92.

185 It was Iustus Lispius who realized, and Gerard Noodt who first argued at length, that there was such a long delay: see van den Bergh, G. C. J. J., The Life and Work of Gerard Noodt (1647–1725) (1988), esp. 207–13.Google Scholar

186 The earliest evidence is Frag. Vat. 34 (FIRA II, p.469) and C.Th. V.10.1. See Buckland, W. W., The Roman Laic of Slavery (1908), 420–1Google Scholar, Humbert, M., ‘Enfants à louer ou à vendre: Augustin et l'autorite parentale (Ep.10* et 24*)’, in Les lettres de Saint Augustin decouvertes par Johannes Divjak (1983), 189–204Google Scholar, esp. 1195–6 (not convincing on all points, however), Evans-Grubb, op. cit. (n. 176), 192–202 (also in Harries and Wood, op. cit. (n. 177), 134).

187 See C.J. IV.43.1 (A.D.294).

188 Ruggini, L. Cracco, Economia e società nell' ‘Italia annonaria’ (1961), 72Google Scholar, Boswell, op. cit. (n. 1), 428 (after A.D.250).

189 C.J. VIII.51.2. It was addressed to Probus, PPO in command of Illyricum, Italy, and Africa. Note that after 374 patroni and domini were still allowed to expose: Memmer, op. cit. (n.3), 70–1 (they were, however, denied the right to recover the victims).

190 ‘IMPPP. VALENTINIANUS VALENS ET GRATIANUS AAA. AD PROBUM PPO. Si quis necandi infantis piaculum adgressus adgressave sit, erit capitale istud malum…’. Cf. Eyben, op. cit. (n. 2), 31. There is admittedly a slight gap in the story (discussed by Memmer, op. cit. (n.3), 69–70), since the edict of 5 March seems to take for granted, rather than plainly asserting, that exposure counted as necatio, the point supposedly made by Paulus (Dig. XXV.3.4).

191 Vanzetti, op. cit. (n.48), 214, cannot believe that child-exposure was a capital offence as early as 374 or indeed until Justinian. E. Herrmann-Otto, ‘Die Reproduktion der Sklaverei auf dem Wege der natürlichen Aufzucht’, in Kraus, O. (ed.), Regulation, Manipulation und Explosion der Bevölkerungsdichte (1986), 95 n. 44Google Scholar, claims that exposure as such was still straflos after 374. As for Valentinian, Ammianus says of him ‘nec enim usquam reperitur miti coercitione contentus’, XXX.8.3.

192 But the silence of the Canons of Elvira about child-exposure presumably means that those present at the council were not greatly concerned about it.

193 See in brief MacMullen, R., ‘What difference did Christianity make?’, Historia 35 (1986), at 326–30Google Scholar.

194 See recently Vanzetti, op. cit. (n. 48), 202–23, Memmer, op. cit. (n. 3), 67–82.

195 C.J. VIII.51.3 (529). Cf. N.Val. 33 of 451.