Raising and Killing Children: Two Roman Myths: Liberum. Recent Standard Treatments of The Roman Family Have

RAISING AND KILLING CHILDREN:
TWO ROMAN MYTHS
by
BRENT D. SHAW
Often the purpose of historical research is to create by explana-

tion and description; occasionally, however, it is destruction that is
required. In the present case, two intertwined historical myths are
the problem. One is easier to dispel than the other, but both are
so deeply rooted in modern scholarship and in popular consciousness
that the hope of eradicating them by this brief exposition is not
very great. The rst myth is the claim that the Roman father, espe-
cially as delineated in the legal model of the paterfamilias, maintained
in his hands a formal power, the so-called ius vitae necisque—the ‘right
of life and death’—by which he could legally kill his children.1) The
emphasis must be on ‘legally’ since no sensible historian of antiq-
uity has ever sought to deny the pervasive reality of infanticide or,
more commonly, the exposure or setting out of unwanted newborns.
Closely related is the second myth: the widespread acceptance, again
by both scholars and the informed laity, of a liminal ritual by which
the father formally accepted the newborn child into his possessions
and power, which is to say into his familia. The ritual, we are told,
consisted of the father ceremonially lifting the newborn up from the
ground after it had been placed at his feet, and then raising the
infant in his arms for all to see. The scene is dramatic and worthy
of a DeMille or a Mankiewicz, but not, alas, of mundane history
despite constant allusion, description, and detailed analysis by rep-
utable historians. The ritual, so it is claimed, was designated by the
technical phrase tollere liberum, or sometimes by its equivalent suscipere
liberum. Recent standard treatments of the Roman family have
1) W. V. Harris, The Roman Father’s Power of Life and Death, in: R. S. Bagnall &
W. V. Harris (ed.), Studies in Roman Law in Memory of A. Arthur Schiller (Leiden 1986),
81-95; at 93 he explicitly connects the two, with the one explaining the other: the
ius vitae is the father’s ceremonial ritualistic lifting up of the newborn infant.
© Koninklijke Brill NV, Leiden, 2001 Mnemosyne, Vol. LIV, Fasc. 1

32 BRENT D. SHAW
aYrmed, occasionally in intriguing detail, the nature of the liminal

ceremony itself.2)
Raising the Newborn: Tollere Liberum

By the turn of the century, eminent Romanists had come to
accept that a formal ‘lifting’ ceremony was necessary for a Roman
father to accept a newborn child into his family.3) Joseph Declareuil,
for example, argued that the father had the nal right of disposi-
tion over the infants produced for him by his wife. Iustae nuptiae or
formally recognized marriage was not the sole determinant of the
status of the children born to a legally married couple. Without
legal marriage, it is true, the father could not claim the infant as
his own, but iustae nuptiae did not impose on him the requirement
to receive into his familia the infants born from his wife. Declareuil
continues:4)
“Let us say, rather brutally, that the head of the domus engaged with
his uxor iusta in a labour over which he maintained the right either
to accept or to reject the results that were thereby produced. This is,
in fact, what he did. The newborn child was placed on the ground
in front of the father: if he raised the infant and took it up into his
arms in an act that was called liberum tollere or suscipere, the infant was
then admitted into his lineage; it became the heres suus. Then, after
the divinities favorable to infancy and living things within the house-
hold had been invoked, the paterfamilias made his wishes and inten-
2) For example, S. Dixon, Tollere liberos: the Birth of the Roman Child, appendix
1, in: The Roman Mother (Norman Okl. 1988), 237-40. Dixon quotes the standard
literary sources; cf. at 238: “Literary sources refer to the father as raising (tollere
or suscipere) the child in this way [i.e., from the ground] as a formal acknowledg-
ment of his paternity and, by implication, of his decision to rear the child”, cit-
ing Cic. Tusc. Disp. 3.1.2 in support; and E. Cantarella, La vita delle donne, 1. La
famiglia e il potere del ‘pater’, in: E. Gabba & A. Schiavone (ed.), Storia di Roma, 4:
caratteri e morphologie (Turin 1989), 557-608, at 558: “Al momento della nascita, i
neonati venivano deposti per terra, ai piedi di pater, che senza bisogno di giusti cazione
alcuna, poteva decidere se sollevarli prendendoli nella braccia (tollere o suscipere
liberos), ovvero lasciarli ove erano stati deposti . . .”.
3) J. Declareuil, Paternité et liation légitimes. Contribution à l’histoire de la famille légale
à Rome, in: Mélanges P. F. Girard: études de droit romain dédiées à M. P. F. Girard, vol.
1 (1912), 315-52; Silvio Perozzi, Tollere liberum, in Studi in onore di Vincenzo Simoncelli
nel XXV anno del suo insegnamento (Naples 1915), 213-38 = U. Brasiello (ed.), Scritti
giuridici: 3: famiglia, successione, procedura e scritti vari (Milan 1948), 93-126 [cited from
the latter text].
4) Declareuil, op. cit., 326 [my translation from the French].
RAISING AND KILLING CHILDREN: TWO ROMAN MYTHS 33
tions towards the newborn child clear in various ways, while at the
same time he presented the infant to his neighborhood and handed
it over to begin its feeding. Finally, the father incorporated the infant
into the religious society constituted by the family by means of the
ceremony of lustratio and by sacri ces made eight or nine days later.”
Quite colorful, but almost none of it underwritten by any evidence.

More elaborate versions could be quoted, but Declareuil’s will suYce.
Debate raged over the precise legal signi cance of the ceremony.
Some Romanists, like Sirio Perozzi, argued that the ceremony of
tollere liberum was not, in fact, necessary to the entrance of the new-
born child into the familia; but even he admitted that failure to per-
form the ritual left the child “legally dead.”5) The position that
eventually came to be accepted by most legal scholars, and not
a few historians, was precisely the one so clearly enunciated by
Declareuil in his general history of Roman law, in which he encap-
sulated the results of his earlier research on the subject:6)
“Modern legislations declare the child legitimate who is born under
conditions prescribed by them, and take no account of the parents’
wishes. It is the law, and not the parents, that determines his civil
status. But the ancient city did not, originally, intervene either in the
composition or the internal organization of the family . . . In order to
rank as a legitimate child it was not suYcient to be born in [legal]
wedlock; it was necessary, above all, to be acknowledged by the pater-
familias . . . the child is presented to him immediately after birth and
laid at his feet, and if he turns away from it (liberum repudiat, negat), it
5) Perozzi, op. cit. (1948), 128, followed by authorities like Pietro Bonfante, Corso
di diritto romano, vol. 1: Diritto di famiglia (Rome 1925; reprint: 1963), 15-8; more
arguments were added by Fabio Lanfranchi, Ricerche sulle azioni di stato nella liazione
in diritto romano, II: la c.-d. presunzione di paternità (Bologna 1964), 28 V. The signi cant
point made by Perozzi and Lanfranchi was to show that later legal texts (esp. D.
1.6.8.pr; 22.3.29.1; 22.40.4.9, and 28.3.6.1) demonstrate that legitimate birth alone
created patria potestas and that it was probable that this was already true for the
texts of the XII Tables (as they are reconstituted). The ceremony of tollere liberum
was therefore not essential to the creation of paternal power even by this time;
cf. L. Capogrossi Colognesi, L’ingresso del neonato nella familia, and Tollere liberos:
una cerimonia solo religiosa?, chs. 6.8-9 in: Proprietà e signoria in Roma antica, vol. 1 (Rome
1994), 224-37, at 234-5.
6) J. Declareuil, Rome et l’organisation du droit (Paris 1924), 126 [= Legitimate Children
Born in Wedlock, in: Rome the Law-Giver, transl. E. A. Parker (London-New York
1927), 114-7, at 114-5]; Declareuil’s entire statement, it must be noted, is under-
written by no primary source other than Dion. Hal. 2.15 [= FIRA I2, ‘Leges
Regiae’ no. 4]; 9.22. I know of no ancient authority for liberum repudiat or liberum
negat; they seem to be more modern inventions required by the plot.
34 BRENT D. SHAW
is exposed, sold or put to death. Even if it survives, it is excluded

from the family and from the city, and whoever nds it can make it
his slave as if it were an alien. On the other hand, if he lifts it up
from the ground and accepts it (liberum tollit, suscipit), the acknowl-
edged child is submitted eight or nine days later to the lustratio, pre-
sented to the gods of the house and given a praenomen. It is [only]
now capable of continuing the domestic cult, and becomes the heres
suus and necessarius of its father.”
Declareuil’s statement is cited primarily for the clarity with which

he expresses the concatenated series of assumptions in the argu-
ment, replete with notice of what he regards as the principal legal
technical terms that connected the ritual of the lifting of the child
with the creation of paternal power. Ideas such as those espoused
by Declareuil have now become thoroughly entrenched in our
received views of Roman history and society. The ritual is enshrined
as accepted doctrine in the handbooks and modern textbooks of the
Roman law.7) Indeed, it is so commonly accepted as an historical
reality that standard dictionaries of Latin de ne the term tollere liberos
as: ‘especially of a Roman father: to pick up a newborn child from
the ground in the process of recognition.’8)
It will be the simple contention of our investigation that no such
ritual ever existed—that it is a modern or early-modern creation
on the order of the sowing of the land of defeated Carthage with
salt or, perhaps more to the point given the story’s origins and func-
tion, like the mythical construction of the so-called colonate.9) Once
7) In standard works from Adolph Berger, An Encyclopaedic Dictionary of Roman
Law (Philadelphia 1953), 738, s. v. ‘tollere liberos’ (“To lift a child. According to
ancient custom when a married woman bore a son, the father [pater familias] lifted
him up from the earth, thus denoting symbolically that he was accepting him into
the family as his son”) to the magisterial Max Kaser, Das römische Privatrecht, 1: Das
altrömische, das vorklassische und klassische Recht, 2nd ed. (Munich 1971), 65 § 15 ‘Der
Eintritt in die Familie’: “Der Vater p egt das eheliche Kind nach römischem
Brauch durch Aufheben des Neugeborenen (tollere liberos) anzuerkennen. Er bestätigt
damit die schon mit der Geburt erworbene Vatersgewalt”. Kaser cites the stand-
ard studies by Perozzi, Lanfranchi, Declareuil, and Watson that will be referred
to below; he further treats the matter at length again in § 83: ‘Entstehung und
Beendigung der patria potestas’ (345-6, esp. 346, n. 22) where he accepts Volterra’s
arguments on the signi cance of the phrase tollere liberos, ac si ex duobus . . . found
in military diplomas.
8) P. G. W. Glare, Oxford Latin Dictionary (Oxford 1976), 1947, s. v. ‘tollo’ (2);
basically unchanged from C. T. Lewis & C. Short, A Latin Dictionary (Oxford 1879),
s. v. ‘tollo’ 2A (1876).
9) R. T. Ridley, To Be Taken with a Pinch of Salt. The Destruction of Carthage, CPh
the words tollere liberos had come to be construed in this context in

the literal sense of physically lifting an infant oV the ground, there
was almost no end to the speculative elaboration of the Roman rit-
ual that was thereby created.10) Over time, scattered literary references
were culled and pieced together into a standard picture of the lift-
ing ritual that assumed canonical form in the nineteenth-century
handbooks of Roman social life.11) By natural extension, the ritual
became a commonplace in legal and social histories of the Roman
family.12) It is therefore very diY cult for the modern researcher,
running so much against the current will to embrace the ritual as
reality, to prove a negative fact.13) A dispassionate investigation of
81 (1986), 140-6: J.-M. Carrié, Le ‘colonat du bas-empire’: un mythe historiographique?,

Opus: Rivista internazionale per la storia economica e sociale dell’antichità 1 (1982),
351-70. The infant-lifting myth is more analogous to the former, whereas the con-
volutions over the ius vitae necisque bear closer comparison with the more complex
historiographical myth of the colonate.
10) There is no sign in contemporary western European sources of mediaeval
date that the ritual existed, or any cognizance of its existence among their Roman
predecessors: S. Shahar, Childhood in the Middle Ages (London-New York 1992), chs.
1-3, 5; J. Boswell, Kindness of Strangers: the Abandonment of Children in Western Europe
from Late Antiquity to the Middle Ages (New York, 1988), 156, n. 63: reports that “dur-
ing the Middle Ages, tollere itself could be applied to ‘rearing children’ (Fragment
Gaudenzianus 20 [ MGH LL 1.1]) in a general sense, or even to ‘thinking of having’
a child”, as if this were diVerent from the Roman sources. Although he is aware of
the problems with the way in which the Greek and Roman sources have been inter-
preted, Boswell persists (e.g., 164, n. 92) in speaking of “the juridical term tollere”.
11) K. J. Marquardt, Das Privatleben der Römer, 2nd ed. rev. A. Mau (Leipzig
1886), 80-3 = vol. 7 in K. J. Marquardt and Th. Mommsen (eds.), Handbuch der
römischen Alterthümer [= La vie privée des romains (Paris 1892), 99]; E. Samter, Geburt,
Hochzeit und Tod: Beiträge zur vergleichenden Volkskunde (Berlin-Leipzig 1911), 1-3.
12) Some recent examples: J.-P. Néraudau, Être enfant à Rome (Paris 1984), 70,
272-3, citing A. Rousselle, Porneia: De la maîtrise du corps à la privation sensorielle II e-
IV e siècles de l’ère chrétienne (Paris 1983), 69-70 [= Porneia: On Desire and the Body in
Antiquity, transl. F. Pheasant (Oxford 1988), 50-1], who refers to the study by Nicole
Belmont (see n. 58 below); F. Dupont, Le citoyen romain sous la République, 509-27
avant J.-C. (Paris 1989; reprint: 1994), 262 [= Daily Life in Ancient Rome, transl.
C. Woodall (Oxford 1994), 220], with a description as detailed as that of Declareuil,
and citing, in turn, Y. Thomas, À Rome, pères citoyens et cité des pères (II e siècle avant
J.-C.—II e siècle après J.C.), ch. 5 in: A. Burguière, Ch. Klapisch-Zuber, M. Segalen
and F. Zonabend (ed.), Histoire de la famille, t. 1: mondes lointaines, mondes anciens (Paris
1986), 195-229 [= Fathers as Citizens of Rome, Rome as a City of Fathers (Second Century
B.C.-Second Century A.D.), ch. 6 in: id. (ed.), A History of the Family, vol. 1: Distant
Worlds, Ancient Worlds, transl. S. H. Tenison, R. Morris and A. Wilson (Cambridge
1996), 228-69, at 230-4]. She might also have referred to Paul Veyne, L’empire
romain, in: Histoire de la vie privé, vol. 1 (Paris 1985), 23 [= The Roman Empire, transl.
A. Goldhammer (Cambridge MA 1997), 9].
13) The singular consistent argument against the historicity of the ritual is that
36 BRENT D. SHAW
the individual literary sources (an ever increasing number of which

have been claimed over the last century), however, will show that
none of them, individually or collectively, can be regarded as accept-
able proofs of the existence of a formal infant-lifting ritual.
There were early warning signs that were noted by some. For a
social ritual that was deemed to be so important in Roman per-
sonal relations, one that was critical to the creation of patria potestas
and the formal acceptance of the infant into the father’s familia, not
a single reference to it could be found in the whole body of sur-
viving Roman legal sources. In an examination of the problem of
the father’s right to expose his newborn children, Fabio Lanfranchi
rather presciently noted that, despite the importance of the lifting
ceremonial, the phrase tollere liberum did not seem to be a technical
legal one. 14) The discovery of Roman birth declarations in papyri
from the province of Egypt published subsequent to Lanfranchi’s
disturbing observation con rmed trouble from this quarter: there
was no record in these oYcial documents of the phrase tollere liberum
or equivalent or, indeed, any allusion whatever to a liminal cere-
mony of recognition of a newborn infant by the paterfamilias. The
usual way in which the obvious implications of this striking absence
of evidence was circumvented—rather than focusing suspicion on
the existence of the ritual itself—was to postulate that the ceremony
had actually functioned in some hoary age of a pre-civic Roman
past, but that in time it had come to be superseded by the legal
‘rules of the city.’15) In this rigid linear evolutionary interpretation
recently made by T. Köves-Zulauf, Römische Geburtsriten (Munich 1990) [Zetemata,
87], esp. pt. I, Tollere infantem, 1-94; his decisive arguments on this part of the evi-
dence, however, seem to have made little impression on recent scholarship.
14) F. Lanfranchi, Ius exponendi e obbligo alimentare nel diritto romano classico, SDHI
6 (1940), 5-69, at 64 f.; Premesse terminologiche a ricerche sulle azioni di stato nella liazione
in diritto romano classico, Cagliari, Studi economico-giuridici dell’Università di Cagliari,
29 (1946), 3-51, at 10: “Sorprende solo che un critico così sagace come il Perozzi,
nel riportare l’elenco dei passi literari in cui questo cosidetto istituto compare, non
abbia rilevato che non in uno di essi c’è l’espressione tollere liberum tal quale, ma
sempre una perifrasi o un termine diverso de liber ”.
15) P. de Franciscis, Primordia civitatis (Rome 1959), 280-1, is representative of
this type of argument which is widespread and is still being elaborated by scholars:
L. Capogrossi Colognesi, Tollere Liberos, MEFRA 102 (1990), 106-27, and J. Cels-
Saint-Hilaire who, in her comments on Capogrossi Cologesi’s arguments (ibid., 123-
6), simply assumes the historicity of the ritual. Capogrossi Colognesi, Proprietà e
signoria (1994), 234-5, further accepts the view that the ceremonial was a ritual
deeply embedded in primitive Italic society long before the advent of the city, and
of the father’s power, the ritual of tollere liberum had a long ‘pre-
city’ history that had only gradually come to be controlled by the
institutions of the state and the law. The subsequent subverting of
the father’s power by the force of the state conveniently accounted
for the absence of the ritual in the later formal law. Then another
negative piece of evidence was noted that should have caused con-
cern. Although they outline in detail the various physical steps in
the birthing process, none of the medical writings of the Roman
age have a word to say on the subject.16) Again, the absence was
noted, rather perspicaciously; instead of confronting the logical impli-
cations of the aporia, however, its importance was dismissed.17)
A review of the primary evidence is therefore in order. Since the
number of texts claimed as evidence for the ritual has proliferated,
I have attempted, however crudely, to group them into broad cat-
egories of genre and type. First, there are a dozen or so texts that
have been culled from the plays of Plautus and Terence.18) None
of them refers to a Roman ritualistic act where the paterfamilias lifts
up a newborn infant from the ground. All are part of dramatic
scenes derived from Hellenistic ‘New Comedy’ where the problems
ensuing from the abandonment of a newly-born baby are contem-
plated, often when the results of an inopportune pregnancy are to
be concealed. The question is: should the woman, who has carried
her pregnancy to term, expose the newborn infant or should she
raise it as her own child? And, if the infant is to be abandoned,
should it then be picked up by someone else, thereby advancing
that its real signi cance therefore always lay in the realm of custom rather than
of law.
16) Köves-Zulauf, op. cit. (n. 13 above), 3-4, provides a convenient synopsis of
the steps in an ideal pattern that he has reconstructed from diverse medical writ-
ers (see his 3 n. 5).
17) Rousselle, op. cit., 69, n. 2 [= 51, n. 18 of the English translation]; since
Rousselle accepts the ritual as fact, however, she is then compelled to dismiss the
signi cance of the lack of reference in any of the medical sources (“aucun auteur
médical”): “car on pense ne trouver que ‘médecine’ chez les médecins”.
18) Plautus, Amphit. 1.3.1-3 = 499-501; Cist. 1.2.4-5 = 123-4; Truc. 2.4.48-9 =
398-9; Terence, Andr. 1.3.14 = 219; 2.3.26-7 = 400-1; 3.1.6 = 464; 4.4.20 = 759;
Heaut. 4.1.13-4 = 626-7; 4.1.51-2 = 664-5; Hec. 4.1.60-1 = 575-6; 4.3.81-3 = 703-5;
Phorm. 5.9.16-8 = 1005-7, are the usual passages that are cited in support; other
stray items have been culled from time to time, even more peripheral and less
relevant. For criticism of the prevailing misinterpretations of Terence, Andria, 219,
401, and 464, see further Köves-Zulauf, op. cit., 31-6.
38 BRENT D. SHAW
schemes of the plot? Such is the case in the Amphitruo (501) where
Jupiter addresses Alcumena: he cautions her to be careful because
of the advanced state of her pregnancy. He tells her that it is nec-
essary for him to go away, on business, but that she should nonethe-
less be sure to raise the child when it is born: mihi necesse est ire hinc;
verum quod erit natum tollito. Jupiter is certainly not asking Alcumena,
a woman, to engage in a ritualistic lifting of the infant ceremony,
but rather, simply, that she should raise the child when it is born
and not abandon it. In most of these cases, indeed, as with Mysis
in Terence’s Andria (464) or Sostrata, the matrona in the Heauton
Timorumenos (626-7; 664-5), or Myrrhina, the matrona in the Hecyra
(575-66), it is the case that it is the woman who is being ordered,
enjoined, or encouraged to raise a baby in the ordinary sense of
feeding and nurturing it, and otherwise providing for its needs as
an infant. From parallel scenes and diction in their Greek antecedents,
it is manifest that the meaning that is being transcribed in the plays
of Plautus and Terence is simply to raise or to rear a newborn
infant in the same sense that we would use the words.19)
Next there are sundry references from the large corpus of Cicero-
nian writings. In no case, however, is Cicero referring to a ritual-
istic lifting ceremony; rather it is to the mundane and, to Cicero,
agonizing task of raising children.20) In the year 47, in bemoaning
his own fate on the anniversary day of his birth (3 January), he
expresses the desire that he had never been born (quo utinam sus-
ceptus non essem). Cicero is surely using a standard phrase for the
fact that one had been reared as a child, with the wish that this
had never happened in his case.21) The same is true when he quotes
19) The terms in New Comedy are paidÛon naireÝsyai (vel sim.) and paidÛon
tr¡fein (vel sim.) and in the legal terminology of the time, see: A. Berger, Miscellanea
papyrologica, II: ƒAnalamb‹nesyai in the Epistulae Hadriani, BGU, I, 140, JJP 1
(1945), 13-40, at 28-40 = BIDR 55-56 (1951), 98-123, at 112-23: 114: “The ancient
tollere liberum had no legal importance, in particular it had no signi cance of recog-
nition of paternity or legitimacy of the child. Equally, the omission of that sym-
bolic gesture was without any legal eVects” (but he still accepts the ritual as fact).
See Menander, Sam. 354-5; Epit. 250-5; cf. E. Fantham, H[e]autontimorumenos
and Adelphoe: A Study of Fatherhood in Terence and Menander, Latomus 30 (1971), 970-
98, on elements of innovation in Terence, but none that aVect the arguments here.
20) Cicero, De div. 1.21.42; Ep. ad Att. 11.9.3; De har. resp. 27, De orat. 155-6;
and Phil. 3.17, 13.23 are the canonical references.
21) Cic. ad Att. 11.9.3; see, in detail, on the meaning of suscipere, Köves-Zulauf,
some verses from an unknown play on the Trojan war (Cic. De Div.
1.21.42) about a monitory vision that came to Priam concerning
his pregnant wife Hecuba. When Priam sacri ces to Apollo, the
god’s reply is:
. . . puerum, primus Priamo qui foret
post illa natus, temperaret tollere;
eum esse exitum Troiae, pestem Pergamo.
The dream message was not cautioning Priam to refrain from per-
forming the rituals of a Roman paterfamilias; rather, it was warning
him not to raise the child for fear of the dire consequences predicted.
Similarly, in a vitriolic attack on Antony and the father of Gaius
Trebonius (Cic. Phil. 13.23) whom Cicero labels a buVoon (scurra),
the orator jabs: Is autem humilitatem despicere audet, cuiusquam, qui ex
Fadia sustulerit liberos? Again, the meaning is clear: ‘Who is this man,
who had children by Fadia, to look down on someone else’s low
birth?’ Nothing in Cicero or in the standard passages quoted from
other prose authors to the end of his age oVers any hint of a ritual
or of a peculiar ceremonial attached to the recognition of newborn
infants.
In the two texts where the precise words liberos tollere rst occur
in extant Latin literature, not insigni cantly in the grand history of
Livy, there is enough circumstantial detail to show that it is pre-
cisely the ordinary task of raising children that is meant.22) The rst
of the two references is found in Livy’s account (4.54.7) of the ple-
beian-patrician con icts of the year 409 that resulted in the rst
election of plebeians to the quaestorship. The possibility of the acqui-
sition of undue power by the plebs is met with a bitter complaint
by the patres about the consequences of the diminution of status for
those who counted in Roman society:
Patres contra non pro communicatis sed pro amissis honoribus fre-
mere; negare, si ea ita sint, liberos tollendos esse, qui pulsi maiorum
op. cit., 28-29, n. 100; and, at length, 40-53, where he demonstrates, conclusively,
that in these contexts it signi es ‘to become pregnant, to give birth, to be born,’
or similar meanings.
22) Livy 4.54.7; 5.30.8. It is hardly without signi cance that they appear in an
author so closely identi ed with the formation of an Augustan perspective on
Rome’s past, and precisely in the most obviously ideologically laden of all his
books, the rst pentad.
40 BRENT D. SHAW
loco cernentesque alios in possessione dignitatis suae, salii aminesque

nusquam alio quam ad sacri candum pro populo sine imperiis ac
potestatibus relinquantur.
The early Fathers of the pristine Republic were not saying that they
were less likely to engage in a ‘lifting of the child’ ceremonial because
they had lost a few minor o Yces of the state to plebeians, but
rather, that if matters continued to degenerate it would be wrong
for them to raise children who would then be excluded from the
honors that were once held by their ancestors. The second Livian
passage (5.30.8) refers to events of 393 in which a severe stasis in
Rome was narrowly averted by a last-minute concession by the ple-
beians to patrician lobbying. In return, the senators rewarded the
people by decreeing a general distribution of Veientine land:
Adeoque ea victoria laeta patribus fuit, ut postero die referentibus
consulibus senatus consultum eret ut agri Veientani septena iugera
plebi dividerentur, nec patribus familiae tantum, sed ut omnium in
domo liberorum capitum ratio haberetur, vellentque in eam spem
liberos tollere.
It is signi cant that Livy has the land grants made not just to patres
familiarum, but to every household of free men so that ‘these men,
because the same prospect or expectation was held out to them,
would also wish to raise children.’ The meaning of Livy’s words is
clear. They do not mean that every citizen would now be encour-
aged to lift up newborn infants because of the receipt of seven iugera
of land. The land grant was material assurance for those who hoped
to start new families. It is the same aim that Tacitus later describes
was achieved for individual families in the senatorial aristocracy of
the Principate, only in the new order the end was accomplished by
the intervention of imperial patronage.23) And it was the same ‘hope’
of being able ‘to raise children’ (tollere liberos) that Pliny the Younger
claims Trajan’s public benefactions oVered to the Roman parents
of his own day.24)
23) Tac. Ann. 2.37: Tiberius gave money to impoverished senators, amongst
them Marcus Hortulus, to whom Augustus had already given 1,000,000 HS:
inlectus a divo Augusto liberalitate decies sestertii ducere uxorem, suscipere liberos, ne claris-
sima familia extingueretur.
The emperor was hoping not for an archaic ritual, but rather for actual children.
24) Pliny, Pan. 27.1: Super omnia est tamen quod talis es, ut sub te liberos tollere libeat
Words similar to tollere liberos are also found in the Augustan poets.
Despite modern translations and commentaries colored by the per-
vasive weight of the ritualistic interpretation, a close investigation
does not sustain strained reference to a birth ritual. For example,
in the verses from Ovid’s Metamorphoses (9.698-9) that are frequently
cited in this context, the ordinary meaning of raising children is
what must be meant, despite much scholarly insistence to the con-
trary.25) The Ovidian case is reminiscent of the ones from Hellenistic
New Comedy. Ligdus, a poor but honest man, has ordered his wife
Telethusa to raise their soon-to-be-born infant if it is a boy, but to
put it to death if it is a girl. The goddess Isis appears to Telethusa
in a dream and addresses her:
. . . Pars o Telethusa mearum,
pone graves curas, mandataque falle mariti.
Nec dubites, cum te partu Lucina levarit,
tollere quidquid erit . . .
Isis encourages Telethusa to raise the child to whom she gives birth,
whatever its sex, and not to destroy it. She is not encouraging
Telethusa, a woman who faces more serious decisions, to perform
a ‘raising of the child’ ceremonial. The passage, in any event, has
nothing to do with the actions of a putative Roman paterfamilias,
against which everything tells, including the mise-en-scène in the
middle of the island of Crete.
Without reviewing every possible literary citation that has been
used to sustain the existence of the ritual, we might nally concen-
trate on a group of rst century authors including both of the Senecas
and Quintilian. But before them, a brief excursus into Petronius. Much
has been made of the words liberos tollit in a scene late in the Satyrica.
Following the incineration of Lichas’ corpse, the novel’s anti-heroes
approach the city of Croton where, they are informed, ‘no one
raises children’ (liberos tollit) since whoever has his or her own heir
is not invited to banquets, to the theater, or to other spectacles; a
expediat. Nemo iam parens lio nisi fragilitatis humanae vices horret . . . Magnum quidem est
educandi incitamentum tollere liberos in spem alimentorum, in spem congiariorum; maius tamen
in spem libertatis, in spem securitatis.
25) F. Bömer, P. Ovidius Naso: Metamorphosen: Kommentar, Buch VIII-IX (Heidelberg,
1977), 486: “tollere ist eigentlich der Terminus für die Anerkennung des Kindes durch
den Vater (durch Aufheben vom Boden)”.
42 BRENT D. SHAW
person with children is left to languish in obscurity.26) The context

itself reveals what the words mean: that oVspring are not produced
by the wealthy persons of the city because they expect that their
childless state will encourage crowds of clients to court them with
favours. Quite apart from the ctional and greatly exaggerated
character of the scenario (including the so-called legacy hunters), it
is improbable that the words liberos tollere signify a ritual of recog-
nition. The mundane meaning is clearly the preferable one: the
childless who have chosen not to raise children have done so for
prospective social advantage.
In the eld of rhetoric, the orator Quintilian oVers only a few
examples of lium sustulit, lium suscepit, and facsimiles.27) The elder
Seneca, however, provides exact instances in model mock court dis-
putes, where he uses the precise words tollere liberos.28) In all the lat-
ter imaginary cases, including one where an angry father rails against
his son whom he has disinherited for having children from a pros-
titute, to maintain consistency and sense the words tollere liberos or
equivalents must be translated simply as to have or to raise children.
For example, in a ctitious case, the question is posed: ‘Is it pos-
sible to have a son from a slave woman?’ The scenario envisages
a father who has a legitimate son, but who ‘raises’ (or has) another
son from a slave woman. The whole case rests on the fact that the
latter son was not recognized legally (or, putatively, raised from the
ground by the father) but rather was born to a slave woman and
was only the natural son of the father by the brute facts of his
birth. Yet it is this son—the one who could not be recognized as a
citizen or legitimate oVspring—whom the father is spoken of as hav-
26) Petr. Sat. 116.7: In hac urbe nemo liberos tollit, quia quisquis suos heredes habet, non
ad cenas, non ad spectacula admittitur, sed omnibus prohibetur commodis, inter ignominiosos lati-
tat. Were it not for the impress of the prior construction of the ritual, no one
would understand these words otherwise; because of the assumption, however, even
the words in this passage have been subject to legal scrutiny to discern their tech-
nical signi cance: G. Gualandi, Tollere liberos in un passo di Petronio, RISG 89
(1952-53), 413-7, who thought (417), following a review of all of the existing
hypotheses on tollere liberum, that there must be a connection between the words
tollere liberos and suos heredes habere, and that the ritual therefore signalled the acqui-
sition of patria potestas by the father (an argument that Perozzi had already deci-
sively refuted).
27) Quintil. Inst. Or. 3.6.97; 4.2.42; Ps.-Quint. Decl. 358; cf. the parallel case in
Decl. 372.
28) Sen. Controv. 1.1.7; 2.4.5; 6.3.
ing ex ancilla sustulit (or) tollit. Every one of the cases described by
Quintilian has this same valence. For example, in cautioning against
excess verbiage when before a judge, Quintilian advises, just say:
Est mihi lius; do not say: Cupidus ego liberorum uxorem duxi, natum lium
sustuli, educavi, in adolescentiam perduxi. No ceremonial here; just the
drudgery and responsibility of raising a son.
Finally, the modern proponents of a ritualistic interpretation hold
that the critical rst step of the baby-lifting ritual was the deliber-
ate placing of the newborn child on the earth or ground. This pur-
poseful ceremonial deposition of the infant is seen to be heavily
freighted with symbolic signi cance.29) The few canonical texts repeat-
edly cited to prove this rst important step in the ritual, however,
seem to have perfectly ordinary meanings. Given the usual Roman
birthing procedures, the newborn infant did indeed sometimes end
up at ground level or close to it.30) There is nothing in any of the
descriptions of this part of the birthing process to show that it was
regarded as a religious ritual.31) A few of the texts refer to nothing
more than the banal fact that the general term for a human (homo)
was thought to be linked to the word for earth or ground (humus/
humilis).32) Therefore, ctive linkages to a presumed ritual have been
postulated in order to impute special ritual signi cance to words of
quite ordinary meaning. Almost the only passage that is taken to
state the fact in plain words is Suetonius’ description of Nero’s birth
(Suet. Nero, 6.1). It therefore deserves to be quoted in full:
Nero natus est Anti post viiii mensem quam Tiberius excessit, xviii
kal. Ian. tantum quod exoriente sole, paene ut radiis prius quam terra
contingeretur.
29) In earlier scholarly literature it was usually the anthropological and ethno-
graphic parallels that counted heavily (the whole approach of Nicole Belmont);
latterly, it is the idea of the ‘appropriation’ of female power by males in gender-
based theoretical approaches that has prevailed (the interpretation o Vered by Angela
Romano).
30) Köves-Zulauf, op. cit., 4, nn. 9-10.
31) Modern studies usually cite some combination of the following: Cic. De har.
resp. 27; Ovid, Trist. 4.3.45-6; Pliny, NH, 7.1.2; Suet. Nero, 6.1 (on which great
weight has been placed); and Tert. De anim. 39.2 (CCL, 2, 842); Macrob. Sat.
1.12.20.
32) Ovid, Trist. 4.3.45-6; Pliny, NH, 7.1.2, and Macrob. Sat. 1.12.20, for exam-
ple, all fall into this bland category; they tell us nothing about a ritual of placing
a child on the ground.
44 BRENT D. SHAW
This is, in fact, the proof text that is repeatedly cited in support of
a deliberate ‘placing the infant on the ground’ ritual. Shorn of the
a priori presumptions of a religious ritual, a straightforward transla-
tion of the passage indicates nothing of the kind. It states, rather,
the mundane fact that Nero was born on the 15th of December,
in the ninth month after Tiberius’ death, so early in the morning
that he was touched by the rays of the rising sun just before he
was touched by the ground.33) There is nothing in the words them-
selves to suggest that this is anything other than an elaborate way
of indicating the timing of Nero’s birth by reference to the physi-
cal mode of birthing in which the newborn Roman infant came to
be placed at ground level.
Since the words tollere liberos, or close equivalents, do occur in
Roman legal texts, the powerful image of the paterfamilias lifting
aloft the newborn child naturally gave rise to the assumption that
the words indicated a ritual and, consequently, to the question of
the relationship of the ritual to the creation of patria potestas. If the
Roman father had to perform this critical ritual to recognize the
newborn child as his own, then what was the connection between
the formative ritual act and the preceding legal fact of the birth of
the child from two Roman citizens who had been joined in legal
matrimony (iustae nuptiae)? An arti cial problem was thereby created:
was it birth from legally married parents or the ceremony of tollere
liberum that created patria potestas? The diVerent alternatives have
been upheld by opposing camps of Romanists. Many of them came
to sustain the position that the ritual was the liminal act that sig-
naled the nal legal acceptance of the baby into the father’s family.
33) Robert Graves, for example, oVers the translation: ‘The sun was rising and
his earliest rays touched the newly-born boy almost before he could be laid on
the ground’ (Suetonius: The Twelve Caesars, Harmondsworth 1957), 216; the Loeb
translator, J. C. Rolfe, is more explicit: ‘Nero was born . . . just as the sun rose,
so that he was touched by its rays almost before he could be laid on the ground’.
Rolfe, like Graves, works with the existing image of the baby-lifting ritual in mind,
since, in order to explain his translation, he refers the reader to Suet. Tib. 7.2:
sublatoque ex ea lio Druso (which does not refer to any lifting ceremonial, but sim-
ply means that ‘his son Drusus was born from her’); and Aug. 5: Augustus, also
just before sunrise ( paulo ante solis exortum), was born at a place later owned by
Gaius Laetorius, who could boast that he esse possessorem ac velut aedituum soli, quod
primum Divus Augustus nascens attigisset. But the phrase is nothing other than a grandi-
loquent way in which Laetorius could impute that a divinity such as Augustus
‘touched earth’ on Laetorius’ property at his birth.
According to this interpretation, the lifting up of the newborn infant

by the father was a social custom whose strength and power pre-
ceded the coming into existence of the law of the state. The cere-
mony therefore remained critical to the de nition and creation of
paternal power over the child. Up to the point where the infant
was lifted from the ground by the father, so the argument went,
the child was only potentially under the father’s power. Until the per-
formance of the ceremony, the pater still maintained an absolute
‘right of life and death’ over the infant which he would then exer-
cise either by picking the child up and lifting it above his head for
all to see—thereby oY cially recognizing the child’s existence—or
by refusing to do so. It was therefore the critical ritual of tollere
liberum or the ‘raising aloft of the infant’ that activated paternal
power over the newborn.34)
In an investigation of the problem published in 1915, Sirio Perozzi
established two basic facts. First, he showed that the Roman laws
of the classical period were clear and emphatic on the simple fact
that birth alone from a legal Roman marriage in and of itself cre-
ated patria potestas. Then, after a review of a number of the liter-
ary texts that had been used to explicate the nature of the lifting
ceremony, Perozzi concluded—without questioning the existence of
the ritual itself however—that there were clear cases of the ritual
not being performed, but with no discernible impact on the creation
of paternal power.35) Perozzi’s study perhaps ought to have provoked
a wider concern with the existence of the ceremony itself. It should
also have drawn attention to the fact that although the words tollere
liberum do occur in legal texts, in not a single case do they refer to
any ritual lifting of a newborn infant. Rather, in every context it is
manifest that they refer to the ordinary duty of parents to raise
their children in the sense of feeding, clothing, educating, and oth-
erwise caring for them.36) Equally signi cant is the fact that the
34) De Franciscis, op. cit., 280-1, and G. I. Luzzatto, Nota minima sul diploma mili-
tare del 306 rilasciato ad un pretoriano di origine italiana ( NSA 1959 8 ss.), in: Studi in
onore di Biondo Biondi, vol. 2 (Milan 1965), 95-110, both sustain this position.
35) Perozzi, op. cit., later con rmed as his basic position in his handbook: Istituzioni
di diritto romano, 2 vols., 2nd ed., vol. 1: Diritto delle persone. Le cose e i diritti delle cose
(Florence 1928 [reprint: Milan 1947]), 437, no. 3.
36) See, for example, D. 23.4.27, 29.2.92, 31.77, 34.4.24, 37.14.6, 37.4.14; CJ.
1.4.24, 5.27.11, 8.51.3.pr, and 8.51.3.2. Even a scholar as ardently committed to
46 BRENT D. SHAW
words, or their equivalent, do not occur in even one of the laws

dealing with the creation of patria potestas, with the de nition of per-
sons or their basic legal status, or with the composition of the
Roman familia.
Indeed, the contextual sense of the most extensive descriptive lit-
erary passage in which the words tollere liberos occur makes their
meaning particularly clear. In his De beneciis (3.11), Seneca oVers
an analysis of the social position and the rights accorded to the
parents of children. Since his words are part of a general discus-
sion of parental authority, they are particularly pertinent to our
inquiry. Unusual powers, Seneca states, have been granted to par-
ents by society. For example, when they are compared with the
powers held by patrons over clients, the powers of parents over
children can be seen to be rather exceptional. Why? Seneca answers:
‘We have made the status of parents sacred because it is bene cial
(to society) that they should raise children (quia expediebat liberos tolli )’.
He explains: ‘It is necessary to support parents in this hard work
because they have to confront unknown risks’. The situation in
which parents nd themselves is quite unlike that existing between
patrons and clients since, in Seneca’ view, there is no element of
free choice in their parental duties: ‘In the raising of children (in
liberis tollendis) there is no choice left in the matter of those who will
raise them (nihil iudicio tollentium licet); everything is a matter of hopes
and prayers. Therefore, in order that parents might undertake this
great risk with equanimity, it is necessary to concede a peculiar
type of power to them.’ In the argument that follows, Seneca pro-
ceeds to discuss the unique social position of parents with respect
to their children.
I have quoted the words from Seneca’s De beneciis at length
because they are part of the one literary passage that provides
enough circumstantial detail surrounding the words liberos tollere to
make their normal meaning reasonably clear. From Seneca’s com-
ments we see that whatever liberos tollere was, it was something that
both parents did and, further, that it was to encourage them in this
the ritual as Angela Romano, op. cit., 885-6, n. 17, admits that “in tali passi l’uso
del termine appare inequivocabilemente generico”, by which she must mean that
the words are in fact not ‘a term’ and that their meaning is the quotidian one of
raising children.
labor that unusual powers were conceded to them by Roman soci-

ety. Moreover, what the parents did with their children is described
as diY cult work that was done by both of them: hard labor in
which they had to confront risks and the unknown. For Seneca, no
choice or decision is involved in raising children since it is a mat-
ter in which the parents are committed to a particular kind of work
as if by fate. None of this description even remotely ts a putative
ceremonial act performed by a paterfamilias who made a decision
physically to raise an infant aloft in his arms. The only reasonable
meaning to be given to what Seneca is describing is the long, ardu-
ous, and indeed often unpredictable, everyday work of raising chil-
dren. The implications of this speci c case are con rmed by a
detailed examination of who does the ‘raising’ of the children: they
include, rather indiscriminately, mothers and fathers, uncles and
aunts, grandparents, other relatives, and even, on occasion, complete
strangers.37)
A survey of all of the available literary evidence therefore leads
to some simple conclusions. None of the passages individually or
the sum of all of them will stand close scrutiny as attesting to a
peculiar Roman birth ritual engaged in by the paterfamilias whereby
he literally lifted a newborn infant from the ground, raising the
child aloft in his arms, by which act he recognized the oVspring as
a legitimate member of his familia. Quite the opposite. The brief
phrases tollere liberos or suscipere liberos turn out to have rather ordi-
nary meanings, virtually the same as the English to rear or to raise
children, in the sense of ‘bringing them up.’ This is concordant with
the singular absence of any mention of the supposed ritual in the
legal texts where one would assume that it would at least be men-
tioned, given the great formative power conceded to it by histori-
ans and lawyers.
37) Köves-Zulauf, op. cit., 73-80, who demonstrates that in the surviving liter-
ary texts men outnumber women as the persons responsible for the ‘raising’ of
children by a ratio of 3:1, thereby giving the (false) impression to modern read-
ers that somehow formal imputations of patria potestas were involved. Naturally,
fathers as household heads were seen as responsible for raising children; but this
was a sociological fact, not a ritualistic or a legalistic one. Just as naturally, moth-
ers, as Köves-Zulauf demonstrates, also counted as parents who raised children,
and so could other relatives; his detailed investigation of who did the ‘raising’
clearly demonstrates that anyone who acted as a parent could do this.
48 BRENT D. SHAW
In a review of the whole question published in 1951, however,

Edoardo Volterra believed that he had discovered the Roman rit-
ual’s ‘holy grail’: irrefutable primary documentary evidence which
nally indicated that the words tollere liberos had a formal legal
signi cance that speci cally attested to the existence of the ritual.38)
The documents were the so-called military diplomata of the high
Roman empire or, to be more precise, special phrases found in cer-
tain of them.39) Volterra drew attention to a series of veteran diplo-
mas in which conubium with peregrine wives was granted by the
emperor to the retiring soldier: proinde liberos tollant, ac si ex duobus
civibus Romanis natos.40) In a supplementary treatment of the subject,
Volterra later noted a whole series of military diplomas that seemed
to be of this same type, and which, he claimed, con rmed his argu-
ments on the legal signi cance of the ritual act of raising the newborn
infant aloft.41) The central point that Volterra wished to make—
and which he took to be con rmed by subsequent discoveries of
more military diplomas—was that the soldiers to whom such grants
were made, unlike the recipients of the more usual auxiliary cohort
diplomas, were already Roman citizens. Volterra argued that the
formal words in the diplomas referred to a liminal ritual of tollere
liberos that could only be exercised by Roman citizens.42) Here, at
last, in a formal legal document issued by the Roman state was
con rmation of the ritual of the raising infants on high as the for-
mal ceremonial recognition of the status of the newborn. Volterra
therefore translated the phrase as ‘thenceforth they may engage in
38) E. Volterra, Un’osservazione in tema di tollere liberos, in: Festschrift für Fritz
Schulz, vol. 1 (1951), 388-98, at 398: in the end, he has to admit that for the puta-
tive critical ceremony “le fonti giuridiche non ci consentono di trovare una spie-
gazione certa alla frase perinde liberos tollant, ac si ex duobus civibus romanis natos”.
39) Volterra, ibid., 390 f.
40) Volterra, ibid., 390, referring to the diplomas then published in CIL III,
Suppl. I, nos. 12, 75, 81, 83, 84, 86, 88, 89, 92, 95, 96, and 97.
41) E. Volterra, Ancora in tema di tollere liberos, Iura 3 (1952), 216-7.
42) Volterra, op. cit. (1951), 392: “Comunque è certo che il tollere liberos è atto
riservato esclusivamente a cittadini romani nei confronti di gli nati da un iustum
matrimonium con una cittadina romana.” His further argument (398) that this was
diVerent than the case of the auxiliary troopers who were granted citizenship for
themselves and their children, and conubium with their wives, in the sense that these
latter Roman fathers did not have patria potestas over their children seems, to me,
quite perverse (although he emphatically repeats it in his review of Watson [n. 48
below], 169).
the ritual of the lifting of infants, and their children will be as if

born from two Roman citizens’.43)
Volterra’s argument has persuaded many, and has led to even
further elaborations on the signi cance of tollere liberos.44) One must
therefore ask if these documents, so far the most explicit drawn into
the argument, do in fact attest to a legal infant-lifting ritual as
claimed by Volterra. If interpreted in context, I would argue, the
phrase has a rather diVerent and more ordinary meaning than that
attributed to it by legal scholars. We might begin by considering
the full range of all military diplomas in which this phrase or a
close equivalent occurs.45) Whereas it was important to Volterra’s
argument that these men were Roman citizens, by a careful inves-
tigation of the contents of all such military diplomas, it can readily
be seen that this is not the signi cant criterion that sets them apart
as a category.46) First of all, ordinary auxiliary diplomas were often
issued to men who were already Roman citizens, but this fact did
not materially aVect the formulaic phrasing of the rights bestowed
on them by which citizenship was granted to themselves and to the
children whom they already had from their existing wives.47) Rather,
43) Volterra, op. cit., 398; I cannot, therefore, accept his bald statement that
“Liberos tollant avrebbe il signi cato giuridico indicato da Gaio I, 57, et qui ex eo
matrimonio nascuntur et cives Romani et in potestate parentum unt ”, since the two facts—
the raising of children [tollere liberos] and the legal status of children born from
legal Roman marriage as outlined by Gaius—are nowhere explicitly connected in
our surviving literature.
44) For those who accept Volterra’s argument at face value, it simply adds more
evidence for the continuing power of the tollere liberum ritual in the Principate; for
those legal scholars, from Perozzi and Lanfranchi onwards, who have rejected the
legal force of the tollere liberum ritual after the Twelve Tables, its presence in these
diplomas signi es the great ‘social value’ that the ceremony now had: so Capogrossi
Colognesi, Proprietà e signoria (1994), 236.
45) Based on a collation from CIL XVI and M. M. Roxan, Roman Military
Diplomas, 1954-1977 [= RMD 1]; 1978-1984 [= RMD 2]; 1985-1993 [= RMD 3]
(London 1978, 1985, 1994) [Occasional Publications, nos. 2, 9, 14]. The relevant
diplomas are: CIL XVI: 18, 21, 81, 95, 98, 122, 124, 133-5, 137, 139-40, 142-3,
145, 147-9, 151, 153, 155-6; CIL XVI Suppl.: 189; RMD 1: 1, 73-8; RMD 2: 124,
131-3; RMD 3: 139, 171-2, 188-9, 192, 194, 196, and 199. Neither the editors of
CIL XVI nor Roxan (nor, indeed, any others) allow the peculiar punctuation
inserted by Volterra.
46) For this speci c type of diploma in context, see M. Roxan, The Distribution
of Roman Military Diplomas, Epigraphische Studien 12 (1981), 265-86, at 265: her
category ‘l’diplomas: ‘praetorian and urban cohorts’; and 269-70.
47) On my computation, based on all those auxiliary diplomas where enough
of the name of the veteran who was receiving his honesta missio has been preserved,
50 BRENT D. SHAW
it is clear that the stereotypical phrase to which Volterra drew atten-

tion was a normal part of the imperial benecia granted to a speci c
group of soldiers upon their retirement: the soldiers of the Praetorian
Cohorts48) and of the Urban Cohorts.49) The usual phrasing, prac-
tically unaltered by a single word for a period of well over two cen-
turies, is as follows:
. . . nomina militum, qui in praetorio meo militaverunt in cohortibus
decem I. II. III. IV. V. VI. VII. VIII. IX. X, item urbanis quattuor
X. XI. XII. XIV subieci/subiecimus, quibus fortiter et pie militia
functis ius tribuo/tribuimus conubii dumtaxat cum singulis et primis
uxoribus, ut, etiamsi peregrini iuris feminas matrimonio suo iunxerint,
proinde liberos tollant ac si ex duobus civibus Romanis natos . . .
Alan Watson has already pointed out in a critique of Volterra’s

interpretation that the words liberos tollant cannot be interpreted in
the way that Volterra wished: namely, as a separate phrase to des-
ignate a peculiar ritual.50) As Watson argued, the additional clause
that follows the grant of conubium is legally unnecessary and point-
less, since the grant of conubium, when combined with the father’s
possession of Roman citizenship, was in itself suYcient to produce
children who would automatically be Roman citizens. The purpose
for the whole of the second century AD (the main time period of the diplomas
that we are studying), 31 of 80 (39% of the total) were granted to men who were
already Roman citizens. Indeed, of the whole known series, from Nero onwards, 38
of 112 (34% of all) diplomas were granted to men who were already Roman cit-
izens—a fact, therefore, of which the issuing authority could hardly have been
unaware (the gures were compiled from CIL XVI; Roxan, RMD 1, 2, and 3).
48) M. Durry, Les cohortes prétoriennes (Paris 1938), 289-98; and Alfredo Passerini,
Le coorte pretoriani (Rome 1939 [reprint: Rome 1969]), 130-9.
49) F. Mench, The Cohortes Urbanae of Imperial Rome: An Epigraphic Study (PhD
Dissertation, Yale University, 1968), esp. 101-9 on the diplomata; H. Freis, Die
Urbaniciani nach ihrer Entlassung, pt. (iv) in: Die Cohortes Urbanae (Cologne-Graz 1967)
= Epigraphische Studien no. 2), 49-50, reviews the problem of why the Urban Cohorts
and, by implication, the Praetorians, alone received this privilege. He accepts the
suggestion of K. Kraft, Zur Rekrutierung der Alen und Kohorten an Rhein und Donau
(Berne 1951), 122-3, that legionaries were not granted the same privilege in order
to preserve their ‘purity,’ to hinder intermarriage with peregrine ‘barbarian’ women.
I doubt this ‘racial’ explanation. It seems that the privilege was just that: a spe-
cial bene t conceded to the men in the Urban and Praetorian Cohorts precisely
because of their higher status; as such, it was concordant with all the other rewards
for these elite soldiers upon retirement, which were normally greater than those
of ordinary legionaries.
50) A. Watson, Patria potestas, in: The Law of Persons in the Later Roman Republic
(Oxford 1967), 77-101, at 77-81.
of the clause, in Watson’s view, “is to explain to the relatively une-

ducated Roman soldier what the eVect of the grant of conubium is”.51)
Despite Watson’s criticisms, Volterra persisted in his belief that tollere
liberos was “a material act” that had “a technical juridical signi-
cance”.52) Viewed from the perspective of all of the diplomas, and
their history, however, his persistence seems unjusti ed.
The reason why the description of the bene ts granted to the
veterans of the Urban and Praetorian Cohorts was worded diVerently
from that of auxiliary soldiers is straightforward. Since they were
already Roman citizens—a prior requirement for enlistment into
these elite units of the army—there was no purpose in granting
Roman citizenship to them. Rather, a diVerent and special right
was being conceded to these soldiers. This right was important since
many of these soldiers had been drafted or had entered the army
in frontier regions of the empire, homes to which they would return
following their retirement where it was likely that they would marry
local women who were not Roman citizens or who did not possess
Latin status or the conubium. This special right of proper marriage
had to be granted to these men because like legionaries, but unlike
ordinary auxiliary soldiers, they were held to the higher moral stand-
ard of not being permitted to marry during the period of their mil-
itary service.53) The emperor therefore granted these higher status
51) Watson, op. cit. (1951), 80, adding “that is why the clause has the oversimpli ed
ac si ex duobus civibus Romanis natos”. The obvious riposte is “were not other pere-
grine soldiers just as ill informed”, to which Watson answers that in the case of
the auxiliary troopers the men were being granted citizenship for themselves and
their existing children, and conubium with their wives, but in the case of these men
there is only one ‘right’ that is being conceded to them, that of conubium—in which
case it is more reasonable to expect a brief explanation to be oVered; in essence
accepted by Voci, op. cit. (1980, n. 62 below), 80-1, n. 196 who points to other
parallels: Livy 38.36.6: . . . nati . . . uti iusti sibi liberi heredesque essent.
52) E. Volterra: review of Watson, Law of Persons, in: Iura 19 (1968), 161-70,
at 168: “. . . non ho mai aVermato che per acquistare la patria potestas su un glio
fosse necessario compiere l’atto materiale del tollere questo, ma avevo rilevato come
la frase proinde liberos tollant . . . è ripetuta costantemente nei diploma militari . . . e
come pertanto essa debba avere un preciso signicato giuridico-tecnico . . .” [emphases
mine].
53) M.-P. Arnaud-Lindet, Remarques sur l’octroi de la civitas et du conubium dans
les diplômes militaires, REL 55 (1977), 282-312, at 296-8, oVers the most persuasive
explanation: non-Roman citizen men in the auxiliary cohorts were permitted mar-
riages (iure gentium) during their period of service, and so their retirement diplomas
granted citizenship to them and, retroactively, to their existing children; cf. Durry,
op. cit., 293-5.
52 BRENT D. SHAW
military men a special privilege: the right of legal Roman marriage

with the rst women whom they married following their discharge
so that their children might be raised as Roman citizens. The pas-
sage ought to be translated properly as follows:
‘. . . I/we have listed below the names of the soldiers who have served
in my headquarters in both the ten praetorian cohorts (I . . . X) and
the four urban cohorts (X . . . XII, XIV) and who have bravely and
loyally completed their military service. To these men I/we grant the
right of legal marriage, so long as it is for marriage with one woman
and with their rst wives only, so that, even if they take women of
peregrine right in marriage, thenceforth they will raise their children
as if they had been born from two Roman citizens.’
Far from signifying two diVerent facts—(a) the ritualistic rite of rais-
ing a child aloft from the ground and (b) having children born from
two Roman citizens—the phrase has a quite ordinary meaning. The
ac on which Volterra placed such great emphasis (placing a comma
before it so as to emphasize the separation of the tollere liberos from
the rest of the phrase) does not bear the meaning that he ascribes
to it: i.e., a full coordinate conjunction. Rather, the ac si signi es
something like ‘as if, indeed’ and it serves to emphasize and to
make speci c the conditions under which the new parents in their
legitimate marriage are to raise their children.54) The words tollere
liberos do not refer to some special ritual; they have the normal quo-
tidian meaning of ‘to raise children’.
A variant parallel phrase appears in a diVerent set of military
diplomas, those issued to marines and sailors from the Mediterranean
eets of the Roman state: lios suscipere. In this case, citizenship is
granted to marines in the praetorian eets of the imperial navy,
including the ones stationed at Misenum and Ravenna:
. . . ipsis liisque eorum quos susceperint ex mulieribus quas secum
concessa consuetudine vixisse probaverint civitatem Romanam dedit/
dederunt et conubium cum iisdem quas tunc secum habuissent cum
54) J. B. Hofmann, rev. A. Szantyr, Lateinische Syntax und Stilistik (Munich 1972),
478-9 (d): ac si = quasi; cf. ThLL, vol. 2, s. v. atque, 1083, lines 76 V., where a
plenitude of good examples are given, some of them rather close to our own: e.g.,
Paul. Fest. (247 L): senatores agrorum partes adtribuerant tenuioribus ac si liberis propriis
(perinde ac liberis). The use has parallels in the imperial Latin of the period, as fre-
quently in Tacitus.
est civitas iis data aut siqui tunc non habuissent cum iis quas postea
uxores duxissent dumtaxat singulis singulas . . .
Despite the slightly diVerent terminology, the meaning in context

is clear: the emperors were granting citizenship to the retiring men
and to their children. The nature of the grant is diVerent in the
case of eet diplomas, since the men usually entered service with-
out Roman citizenship, but since they were ‘lower caste’ they were
allowed to form ‘common law’ marriages (recognized not in Roman
law, but under ius gentium) with the women whom they married.
These diplomas therefore follow the same format as those of aux-
iliary soldiers who were also permitted ‘non-Roman’ marriages dur-
ing their service—marriages whose validity, including the children
produced from them, were recognized retroactively at the time of
the grant of citizenship that the men received upon retirement.
The bureaucratic modi cation of phrase did seek to signal a ne
status diVerence, but a simple translation does not sustain the claim
of a peculiar ritualistic act conducted by the paterfamilias:
‘. . . to the men themselves and to the children whom they have had
from the women with whom they can demonstrate that they have
lived according to long-established custom, he/they [i.e., the emperor(s)]
grant Roman citizenship, and the right of legal marriage with those
women whom they have with them at the time when the citizenship
is granted, or if they do not have any [i.e., wives] at this time, with
those women whom they later take as wives, so long as it is one
woman with one man . . .’
Moreover, a review of literary texts that make the meaning of the

words suscipere liberos explicit by context reveals that they are syn-
onymous with tollere liberos.55) For the sake of variation, authors, such
as the imperial bureaucrats who drafted documents like military
diplomas, could use the verb suscipere as an alternative for tollere.56)
55) A good case is Statius, Silv. 4.8.13-4: quaeque sibi genitos putat attulitque benigno
Polla sinu, on which see the translation and commentary of Kathleen Coleman,
Statius: Silvae IV (Oxford 1988), 45, 212-3. While still accepting that tollere liberos is
the technical term for the father’s recognition of the child’s legitimacy (citing Kaser),
Coleman nevertheless points out that in the context of Statius’ verse the words
signify nothing more than a “gesture of maternal aVection” and are equivalent to
sinu excipere or suscipere.
56) Arnaud-Lindet, op. cit., 297, n. 1, recognizes this: “tollo dans le sens de sus-
cipio que nous avons vu à propos des marins”, but she had earlier (295) included
54 BRENT D. SHAW
Nevertheless, the pervasive power of the mental image of the child

being lifted up has led one modern translator after another to
attribute to this verb as well a technical meaning of juridical recog-
nition of the infant by the father. Thus a quite mundane passage
in Cicero’s Tusculan Disputations (3.1.2) in which he discusses the abil-
ity of a human child to perceive moral categories like good and
bad from birth reads: nunc autem, simul atque editi in lucem et suscepti
sumus, in omni continuo pravitate et in summa opinionum perversitate versamur
ut, paene cum lacte nutricis errorem suxisse videamur. The words suscepti
sumus are almost always translated as ‘when we were lifted (or raised)
from the ground’ or ‘when we were legally recognized’. On the
rare occasion when the translator provides an explicit reason for
such a translation, it turns out to be the prior mental picture of
the ‘infant lifting’ ceremonial.57)
Once the ritual of tollere liberos was created, the myth itself was
powerful enough to generate misleading presumptions, which in turn
permitted further elaborations, sometimes quite egregious, of what
could be known of the supposed ceremonial. An elaborate study,
replete with visual images (all of them irrelevant), has oVered a
detailed anthropological explication. All the probative evidence, alas,
is adduced from societies other than that of ancient Rome itself.58)
How then did the ordinary Latin words come to be so misunder-
stood? By simple inadvertence? By too literal an understanding of
the verb tollere? Or otherwise—perhaps by some deliberate malfea-
sance? The former would argue for a late medieval or early-modern
fabrication by commentators and glossators who wanted to create
a harsh version of the Roman father’s power that they could claim
an element of legal recognition that is not necessarily present in suscipere: “de ces
unions peuvent être issus des enfants que leur père reconnait et élève, comme
témoigne l’utilisation du verbe suscipio”. [m.i.].
57) The Loeb translator, W. E. King, for example, provides the rationale for
his translation (226, n. 1): “If the father was prepared to ‘acknowledge’ a newborn
child, he lifted it from the ground and thus showed that he was willing to raise
it as his own”.
58) N. Belmont, Levana, ou comment ‘élever’ les enfants, Annales (E.S.C.) 28 (1973),
77-89: a few standard Roman literary sources—Ovid, Varro, Tertullian, Suetonius,
and Augustine—are quickly dispensed with in the rst note, after which one quickly
passes to ‘Mutter Erde,’ southeast Asia, central Mexico, the speculations of Sir
James Frazer, Freud, Grimms’ fairy tales, the Chinook and Ifago, and, necessarily,
Van Gennep.
as cultural, and legal, precedent. The latter might happen by a pur-

poseful misconstrual of Roman ‘pagan’ rituals. It is possible that
Tertullian, the earliest Christian author of substance in Latin, did
aver to birth rituals.59) He certainly mocked the multitude of Roman
gods and spirits linked with the procreation, birth, and raising of
the newborn. Tertullian’s words were later replayed at length by
Augustine in the De Civitate Dei where he parodies the pagan Roman’s
propensity to have a deity or spirit in charge of every act of human
behaviour, including the processes of procreation, pregnancy, and
birth. Amidst the urry of deities that oversaw everything from the
seminal uids of men to the infant in the cradle was one dea Levana
who ipse levet de terra.60) Alas, the original text of Tertullian, of which
this is a later reworking, is in such a state of disrepair that it is not
possible to say what sort of elaboration Augustine performed on
it.61) Even so, it seems reasonably certain from reading the whole
passage in context that Augustine is describing nothing other than
the purely physical stages of the birthing process and that he nowhere
refers to any ritual. At each step a deity watches over the progress
of the infant in procreation and birth, from conception to cradle.
Whatever Levana did had no explicit connection with a rite of recog-
nition, nor does Augustine ever suggest such. Rather, a direct read-
ing of the text indicates that each god or spirit was correlated with
actual steps in the birthing process.
Our best estimate, therefore, is that the creation of the myth of
tollere liberos as a legalistic family ritual must be located in a later
age and for other purposes. A review of all of the available literary
and epigraphical data therefore leads to one inescapable conclusion:
59) Other Christian sources do not seem to be relevant, since, like Lactantius,
Div. Inst. 1.16.10 (CSEL, 19.1, 62), they assume the ordinary meaning of tollere
liberos, as in this passage where Lactantius cites the Roman poets for the fact that
salacissimus Iuppiter desierit liberos tollere, etc., which, as the context makes clear means
nothing more, even in Jupiter’s case, than simply to raise or to procreate children.
60) Aug. Civ. Dei, 4.11 (CCL, 47, 108): Ipse praesit nomine Liberi virorum seminibus
et nomine Liberae feminarum, ipse sit Diespater, qui partum perducat ad diem; ipse sit dea
Mena, quam praefecerunt menstruis feminarum, ipse Lucina, quae a parturientibus invocetur; ipse
opem ferat nascentibus excipiendo eos sinu terrae et vocetur Opis; ipse in vagitu os aperiat et
vocetur deus Vaticanus; ipse levet de terra et vocetur dea Levana, ipse cunas tueatur et vocatur
<dea> Cunina . . .
61) Tert. Ad Nat., 2.11.1-13 (CCL, 1, 59: to be preferred to the CSEL version
[20, 114-5] which is a heavily ‘reconstructed’ text, mainly by Gothofredus, that
obscures the large lacuna in the original).
56 BRENT D. SHAW
a modern myth. In the texts that have been deployed to sustain

the existence of the ritual, the words tollere or suscipere liberum or
liberos bear the quite ordinary and normal meanings of to raise or
to have children. Replaced in their original context, they do not
refer to a ceremony performed by the Roman paterfamilias, and no
such ritual needs to be hypothesized to make perfectly good sense
of the texts at our disposal. Once the ctional ritual is abolished
from modern historical analysis, knotty but unnecessary legal conun-
drums are avoided and a great deal of spurious anthropological
speculation about the ‘deep past’ of the Roman family is vitiated.
The Power of Life and Death

More serious, perhaps, is another power attributed by historical tra-
dition to the stern Roman paterfamilias: the father’s ‘right of life and
death.’ The ius vitae necisque is ordinarily interpreted by modern his-
torians as the legal right of a father to kill his own children. A crit-
ical study of this lethal aspect of Roman fatherhood asserts that
“this was the central or at least the most remarkable element in
patria potestas” and that “the father’s right to kill his children con-
tinued in existence until the late empire”.62) Another analysis of this
power begins by quoting the authoritative textbook on the Roman
law to the eVect that the vitae necisque potestas was “the core (Kernstück)
62) W. V. Harris, op. cit. (1986), 81; C. Fayer, Ius vitae ac necis, ch. 2.4 in La
Familia Romana: aspetti giuridici ed antiquari (Rome 1994), 140-78 oVers a summary
of the legal scholarship; for some of the earlier studies, see E. Sachers, Tötungsrecht
(ius vitae ac necis), in: Potestas patria, RE 22.1 (1953), 1046-1175, at 1084-9; P. Voci,
Storia della patria potestas da Augusto a Diocleziano, Iura 31 (1980), 37-100, and his
Storia della patria potestas da Costantino a Giustiniano, SDHI 51 (1985), 1-72; and
A. M. Rabello, EVetti personali della ‘patria potestas’, I: dalle origini al periodo degli Antonini
(Milan 1979), 32-3, 117-23, 145-9, 179-81, 209-25; the latter two authors cite the
standard earlier works on patria potestas, almost every one of which contains an
assertion of the existence and centrality of this right. Amongst these, I should espe-
cially like to note Bernardo Albanese, Note sull’evoluzione storica del ius vitae ac necis,
in: Scritti in onore di Contardo Ferrini pubblicati in occasione della sua beaticazione, vol. 3
(Milan 1948), 343-66. A more recent study has been proVered by Y. Thomas,
Vitae necisque potestas. Le père, la cité, la mort, in: Du châtiment dans la cité: supplices
corporels et peine de mort dans le monde antique (Rome 1984), 499-548, but its perspec-
tives on this history of this power have been seriously questioned in the critical
analysis of patria potestas by Richard P. Saller, Pietas and Patria Potestas: Obligation
and Power in the Roman Household, ch. 5 in: Patriarchy, Property and Death in the Roman
Family (Cambridge 1994), 102-32, esp. 114-7. It will be clear from what follows
that my interpretations are much closer to those oVered by the latter scholar.
of patria potestas, denoting the widest extent of the power enjoyed

by the paterfamilias”.63) The same author continues: “The powers of
the paterfamilias were, legally, virtually unrestricted . . . The most
important of these powers was the ius vitae necisque, which gave the
father the right to put his children to death.”64) Only recently have
there been more systematic attempts to set the nature of the legal
and the ideological claims of the Roman father’s paternal powers,
including the linkage of the vitae necisque potestas and patria potestas,
in the context of a “sociological description of family relationships”.65)
More frequently, and usually, the power of life and death has
been interpreted as a great power originally vested in the father
and the kinship group of which he was part—a power that had
exceptional strength in the earliest periods of Roman history and
which was only gradually ameliorated by the infringements of the
state on the one hand and by customary practice on the other.66)
It should be emphasized, again, that we are not dealing here with
the tragic realities of infanticide or infant exposure that were com-
mon to many premodern societies, including that of imperial Rome,
but rather with a rmly established legal right (ius) or a formal
power ( potestas) possessed by adult male citizens over their oVspring:
the ius/potestas vitae necisque.67)
Even in the strongest ideological representations of them, these
paternal powers, including those imputed to the fathers of earliest
Roman times, were hardly untrammeled. In the strict linear evolu-
tionary interpretation of the Roman family favored by some modern
63) R. Yaron, Vitae Necisque Potestas, Tijdschrift voor Rechtsgeschiedenis/Revue
d’histoire du droit 30 (1962), 243-51, at 243, citing Max Kaser, Das römische
Privatrecht, 1 (Munich 1955), 52 (unchanged in the 1971: at 60-1); so too, Fayer,
op. cit., 140: “Il diritto di vita e di morte era la componente più caratteristica,
dura e sconcertante della patria potestas, che ne attesta la sua straordinaria esten-
sione ed intensità”.
64) Watson, op. cit., 98.
65) Saller, op. cit., 115-7, for the treatment of the ius vitae necisque under the sub-
section on ‘potestas’.
66) G. Franciosi, Famiglia e persone in Roma antica dall’età arcaica al principato (Turin
1989), 55, describing a standard view; he refers to Bonfante, Corso di diritto, vol. 1
(Rome 1925/63), 94-105.
67) Vita must signify the power to grant life, the right to support life or the
power to permit one to live, to pardon capital wrongs (so, Yaron, op. cit., 248-9).
Thomas, however, is surely right to argue that the words joined by the que or ac
form a kind of hendiadys, a way of describing a single spectrum of power (op. cit.,
508; so, already, Yaron, op. cit., 245).
58 BRENT D. SHAW
historians, the statements by Dionysius of Halicarnassus on the ‘royal

laws’ of Romulus are accepted as veridical historical fact.68) These
regal injunctions are supposed to oVer a credible description of the
great powers originally wielded by the Roman father. It is impor-
tant to note that even in the mythical construction of the Laws of
Romulus that were supposedly part of Rome’s ‘ancestral constitu-
tion’, a father was already constrained to raise all male children
and at least rst-born daughters.69) Modern historians who take the
account in Dionysius as brute fact, however, argue that the potestas
of the father in the evolutionary stage that preceded the rst restric-
tions imposed by King Romulus, a fortiori must have been more
fearsome and potent—indeed virtually unrestricted in any fashion.70)
As with the ritual of the raising aloft of the newborn infant, this
powerful right of the father to kill has become a textbook standard.
That some archaic avor was indeed intended by the words potes-
tas vitae ac necis (or variants) is signaled by the use of the term nex.
It is a word that is rarely found in authors of the middle and late
Republic but was one that was being glossed by antiquarians of the
Augustan age for its signi cance. Still avoided for common use by
Vergil and Livy, nex seems to have enjoyed a sudden eÞ orescence
in writers of the full Principate, beginning with Ovid and extend-
ing to Tacitus.71) The archaizing research of the antiquarians on
68) For example, Capogrossi Colognesi, Proprietà e signoria (1994), 231, wishes to
set aside much modern speculation and to restart afresh “dalle più antiche testi-
monianze esistenti sull’organizzazione dei rapporti fra pater e i gli neonati”. The
‘most ancient testimony’ turns out to be that of Dionysius of Halicarnassus, an
Augustan ideologue.
69) Dion. Hal. 2.15.2 [= FIRA I2 , ‘Leges Regiae,’ no. 4]; since E. Weiss,
Kinderaussetzung, RE 11 (1921), 466, it has been recognized that the ‘laws’ attrib-
uted to Romulus are most probably a rhetorical ctional reworking of the regu-
lations attributed to the Spartan ‘constitution’ governing the exposure of infants
(cf. Plut. Lys. 16 ); see also, J. Carcopino, Les prétendues ‘lois royales’, MEFR 54 (1937),
344-76.
70) Rabello, op. cit., 31-5 & 37 f., followed by Capogrossi Colognesi, Proprietà e
signoria (1994), 232.
71) Based on Festus (158 L): Neci datus proprie dicitur, qui sine volnere interfectus est,
ut veneno aut fame, it is often claimed (e.g., Dominique Briquel in: Thomas, op. cit.,
548, as well as the author himself at 509; Fayer, op. cit., 140, n. 37) that nex indi-
cates a ‘non-sanguineous’ death. One might also add the notice for occidere under
Festus (190 L): Occisum a necato distingui quidam, quod alterum a caedendo atque ictu eri
dicunt, alterum sine ictu. This might well be true in some arcane ritualistic or anti-
quarian sense, but ordinary usage indicates that nex usually signi ed a violent death
or one imposed coercively, not infrequently with a connotation of homicide. Only
early Roman history was also underwritten by anthropological reports

that explicitly understood more primitive societies to re ect in the
present-day ‘chronotypes’ of what Roman society had once been
like at its origins. So Julius Caesar in his ethnographic asides on
the nature of Gallic and German social relations refers to the potes-
tas vitae necisque to indicate the total power wielded by their magis-
trates.72) The claim is not surprising since, in the primitive power
relations that characterized Gaulish society, men exercized potestas
vitae necisque ‘over their wives in the same way as over their chil-
dren.’73) In both cases (as later in Tacitus) the statements are intended
as a manifest contrast with contemporary Roman practices. Apart
from ideological or literary uses such as these, however, what is the
evidence for the historical and legal status of this power of life and
death? The disturbing thing, given the gravity of the claims, is that
it is so slight.
The main problem with the evidence is not its total absence, as
in the case of the infant lifting ritual, but rather its relative sparse-
ness and late date. One can indeed assume that such a power
existed as part of the most basic primaeval powers vested in the
paterfamilias and that it only came to be diminished gradually with
the progress of time and the emergence of more lenient standards
of civil behaviour. If so, however, a manifest diYculty faces the
claimants for its historicity. As one legal historian has explained,
“the Roman sources, due mainly to the late date at which they
were composed, do not furnish the means for a proper and exact
understanding of the early, original import of vitae necisque potestas”.74)
In cold fact, there does not exist a single contemporary source ear-
lier than Cicero’s passing allusion to the power in his De domo sua
of 57 B.C.75) One must also remember that the only notice that
two uses are found in Livy and Horace, half a dozen in Vergil; the eVusion begins
with Ovid (65 instances) and continues through Seneca (36, the majority poetic),
Silius Italicus (12), Statius (12), and Tacitus (75-80 instances).
72) Caesar, Bell. Gall., 1.16: Vergobretes among the Aedui; see 6.23 for war
leaders among the Germans.
73) Caesar, Bell. Gall., 6.19, among the Gauls.
74) Yaron, op. cit., 245; at this point Yaron forgoes further analysis based on
the Roman sources, since they are so nebulous, late in date, and few in number,
and attempts a reconstruction of the meaning of the power by recourse to earlier
Near Eastern evidence.
75) There are a few late sources that are so derivative and peripheral that they
60 BRENT D. SHAW
Cicero provides is to a formulaic phrase, embedded in a ritual for

a species of adoption (adrogatio), that was purposefully designed to
convey an aura of antique authenticity and authority.76) Furthermore,
recent studies on the supposed use of this power by the paterfami-
lias (as used to discipline errant sons) have demonstrated how frag-
ile a construction has been built on the supposition that such cases
must represent the actual exercise of a legal power of life and death
that could be wielded by Roman fathers.77)
In fact, not one of the cases of fathers executing their sons retailed
by Livy and other similarly ideologically-driven historical narratives
will sustain the interpretation that the father was killing his son
under the aegis of a personal legal power that he held as a pater-
familias.78) Closer investigation of each of the cases reveals not actual
will not be considered further in my argument here, amongst them Dio Chrysostom,
15.20 (in his rst discourse on slavery and freedom) where Rome, although not
explicitly mentioned as one of the ‘many peoples’ amongst which fathers have the
same powers that owners have over their slaves, must be meant: ¤feÝ`tai gŒr aétoÝw
pokteÝ`nai m®te krÛnantaw m®te ÷lvw aÞtiasam¡nouw. Similarly, when Lactantius,
Div. Inst. 4.4.11 (CSEL, 19, 284) transfers this power to God ‘the father’, he is sim-
ply playing with received images: solus pater vocandus est qui . . . habet vitae ac necis
veram et perpetuam potestatem.
76) If Aulus Gellius (Noct. Att. 5.19.6) is to be believed, there is every reason
to hold that the formula was concocted by Q. Mucius Scaevola when he was
Pontifex Maximus, at the same time that he formulated the oath that was put to
the adrogator and adrogatee: ius iurandum a Q. Mucio pontice maximo conceptum dici-
tur, quod in adrogando iuraretur, and therefore it dates to no earlier than 89 B.C. It
was fabricated by a reputed systematizer of the law who, notably, was much con-
cerned with the legal standard set by the bonus paterfamilias.
77) See Sachers, Yaron, and, in detail, Harris. What these scholars have demon-
strated is that tales of fathers imposing capital punishment on their sons functioned
as stereotypical exempla of the father demonstrating his higher pietas to the state,
to its rules, and to the duties required of him as a magistrate (under which pow-
ers the executions were ordered or performed) as a loyalty superior to that of a
father to his son. It was a pattern of behaviour that was already understood in
antiquity (Polyb. 6.54, cf. Saller, op. cit. [1994], 115, n. 55). The whole meaning
of these stories therefore depends on the assumed nature of the bond linking fathers
and sons—that it was a very powerful one that would be expected to override the
dictates of the state—than it does on an assumption that fathers had a legal right
to kill their sons. All three scholars argue that there are some redeemable histori-
cal cases, but I cannot nd one amongst the whole lot. Wherever the fuller his-
torical narrative of any case is known, it turns out that the son is being executed
under an already existing formal sanction of the state.
78) Harris, op. cit., 82-6, presents ten enumerated cases. He dismisses all of these,
except his nos. 7 and 10, as demonstrating the lethal exercise of a potestas vitae
necisque by Roman fathers. But I would argue that even his exceptional cases (7
& 10) do not demonstrate the existence of the right. The rst of these is the case
acts of killing done by fathers who were exercizing speci c powers

that they held as patres, but rather ideological interpretations of their
actions. Typical is the case of the senator, a Fulvius (otherwise
unknown and unattested) who is said to have had his son killed for
attempting to join Catilina’s forces in the political frights and rural
jacqueries of 63.79) There is no source on the incident earlier than
a brief tendentious clause in Sallust. In the extremities of the crisis
as it developed in late 63, one cannot rule out the possibility that
the father was treating his son as one of the hostes publici—as those
who were labelled ‘conspirators’ had already been declared to be
some time before the father issued his fatal order (Sall. Cat. 36.2-3).
Furthermore, there is no explicit statement or even suggestion that
any paternal ius vitae necisque was involved. In his later moralizing
of the story, Valerius Maximus (5.8.5) simply indulges in his usual
tactic of deliberately misconstruing such cases as ‘historical exam-
ples’ of the extreme exercise of paternal power.
In Sallust, as in the other authors, the nature of the misrepre-
sentation has its roots in the writer’s ideology. The contrast is struc-
tural. In Sallust’s representation of the Catilinarian conspiracy, just
as ‘bad’ women are shown to be ready to murder their husbands
so, it is averred, ‘corrupt’ young men were ready to murder their
fathers. In earlier and better times, the reverse was true: in those
times, it was fathers who exercized the right of death over their
of a senator, Fulvius, the relevance of which I reject in the argument that follows
in my text above. The other instance is no more convincing. It concerns an inci-
dent in the reign of Augustus in which an eques Romanus named Tricho beat his
son to death (Sen. De Clem. 1.15.1). Once again, there is no sign that any potestas
vitae necisque was invoked. The story is told by Seneca in such a brief anecdotal
fashion that it is surely too dangerous, in ignorance of any of the circumstances,
to claim that Tricho’s actions were justi ed by a formal legal power that he held.
And the response by the people (a crowd assailed Tricho in the forum and stabbed
him with their writing instruments) makes this seem rather improbable. The most
probative of Harris’s cases, I think, is his no. 6, in which Quintus Fabius Maximus
(perhaps ‘Eburnus’ cos. 116, cens. 108) killed his son for an unspeci ed sexual
oVence (Quint. Decl. mai. 3.17; Val. Max. 3.17). The importance of this case—one
of the few in which we have some of the attendant details, and where it seems
probable that the father was not acting under the aegis of any state power—is that
the father was charged with murder and was convicted on the charge (Oros. 5.16.8;
Cic. Balb. 28: he went into exile at Nuceria). If he thought that he was acting
under some claim of a potestas vitae necisque (a presumption that has to be made in
most of these cases) then, as Harris himself says, the claim was rejected as valid.
79) Sall. Cat. 39.5; reiterated by Dio 37.36 at a much later date.
62 BRENT D. SHAW
sons. In his version of Cato’s speech on the fate of the conspira-

tors, Sallust has Cato refer to the early Republican example of Titus
Manlius Torquatus (apud maiores nostros) who ordered his own son
to be killed quod is contra imperium in hostem pugnaverat. It is no acci-
dent that the phrase is virtually a word-for-word repetition of the
historian’s own moral reconstruction of the distant Roman past, an
age of uncontaminated virtus before the annihilation of Carthage.
In that epoch, the foundation of good morals was demonstrated by
the fact that punishment was in icted on those qui contra imperium
in hostem pugnaverant.80) Whatever the version of the story, however,
Manlius Torquatus was not imposing the death penalty as a pater-
familias, but as a eld commander who was unwilling to exempt his
own son from the higher dictates of the state.81) The good senator
Fulvius perhaps did the same in the ‘conspiracy’ of 63. It is by sug-
gestion alone, however, that the reader is lead to believe in an
exemplary revival of the fatal powers of a father.
The most explicit attestation of the right itself, which is usually
described simply as a power or potestas, is found in the formulaic
words uttered in the ceremonial for the type of adoption known as
adrogation. In the course of discussing the two types of adoption
(adoptio and adrogatio), Aulus Gellius (Noct. Att. 5.19.9) reports that
the words were part of the formal request put to the Roman peo-
ple, assembled as a comitia curiata, in the form of a motion for leg-
islation—a rogatio con rming the most serious form of adoption
(adrogatio) that would result in the extinction of an existing family:
Eius rogationis verba haec sunt: ‘Velitis, iubeatis, uti L. Valerius
L. Titio tam iure legeque lius siet, quam si ex eo patre matreque
80) Sall. Cat. 52.30 (Cato’s speech); 9.5 (his own ideological construction of the
past, almost word for word); noted by Donald Earl, The Political Thought of Sallust
(Cambridge 1961), 97. Sallust’s text has the praenomen Aulus. The problem is
precisely that it was a moral exemplary tale on which no one, even then, could
agree as to its precise historicity: one tradition, including Cic. De OV. 3.112, De
Fin. 1.23; Aul. Gell. Noct. Att., 9.13 and Livy 8.7, has the name as Titus Manlius
and the incident in a war against the Latins in 340 B.C.; Dion. Hal. 8.79 is part
of the tradition that Sallust follows that placed the event in the war against the
Gauls in 361. The praenomina vary accordingly. The story bears all the hallmarks
of a mythical (i.e., structuralist) construction of the past.
81) According to the Livian tradition (8.7), he was consul of 340 and com-
manded the war against the Latins; for the other sources: see Broughton, MRR,
I, 135; even Livy’s rhetoric, attributed to the father, plays heavily on the contrast
between paternal love and the higher duty to the state.
familias eius natus esset, utique ei vitae necisque in eum potestas siet, uti
patri endo lio est. Haec ita uti dixi, ita vos, Quirites, rogo.’
It is these same formal words that are alluded to by Cicero in the

course of a lengthy and vitriolic attack on his personal archenemy
Clodius, including a scathing denunciation of his plebeian adoption:82)
Credo enim, quamquam in illa adoptione legitime factum est nihil,
tamen te esse interrogatum, ‘auctorne esses, ut in te P. Fonteius vitae
necisque potestatem haberet, ut in lio’ . . .
Once modern readers pass beyond these two references, however,

they nd themselves in the nebulous realm of later laws and legal
commentaries, most of them dating from the fourth to the sixth
centuries A.D., in which the writers believe that ‘they’ (that is, ear-
lier Romans) had attributed some such legal power to the father of
the family. All these later writers, however, either assume or explic-
itly state that this power was no longer in force at the time that
they were composing their own texts.
The problematic status of this power, however, was already appar-
ent at a much earlier date. By the late Republic and early empire,
the confusion amongst contemporaries was such that the potestas
came to be identi ed (mistakenly) with the formal ius occidendi that
was granted to a father in the case where he caught his daughter
in agrante in the act of adultery either in his own house or in that
of his son-in-law.83) A modern-day historian must suspect that the
general sentiment that fathers ought to have such a power was
somehow cathected onto the Augustan laws on adultery, since these
laws seemed to oVer a formal legal basis for the potestas vitae necisque.
That the Augustan laws did not in fact oVer any such legal grounds,
and were never intended to do so, was somehow lost in the later
82) Cic. De domo sua, 29.77; this is the only reference to the power in all of
Cicero’s public speeches; more signi cant, perhaps, is the absence of any refer-
ence at all to the power in the whole of Cicero’s personal correspondence, philo-
sophical, and rhetorical works.
83) The provision was contained in the second chapter of the Lex Iulia de adul-
teriis, the text of which we do not have; but based on reports of it in later legal
texts (e.g., Collatio 4.2.3), the following words would seem to be close to the orig-
inal: (Secundo vero capite permittit patri), [si in] lia sua, quam in potestate habet, aut in ea,
quae [eo] auctore, cum in potestate esset, viro in manum convenerit, adulterum domi suae gene-
rive sui deprehenderit isve in eam rem socerum adhibuerit, ut is pater eum adulterum sine fraude
occidat, ita ut liam in continenti occidat.
64 BRENT D. SHAW
process of identi cation. Since there was no other statutory basis

for the power of life and death, by default most subsequent dis-
cussions of the assumed potestas were framed in terms of the father’s
ius occidendi that was part of the Augustan legislation on adultery.84)
The two were constantly confused. Thus Ulpian’s opinion that a
father cannot murder his son unheard, but rather must accuse him
before the urban prefect or the governor of the province, has been
taken to re ect attempts to control the vast powers of life and death
wielded by the pater.85) But the fragment is a discussion of restric-
tions imposed on the speci c ius occidendi that was permitted to a
father by the Augustan adultery legislation, and nothing more.86)
It was as easy and desirable then, as it is now, willfully to con-
fuse the speci c ‘right of killing’ allowed to a father under the
speci c circumstances of the Augustan laws on adultery with some
general overriding power of the father to kill his children. For exam-
ple, a fragment of Paul that explains the position of the liusfamilias
as an automatic heir to the family’s goods has been taken to refer
to precisely such a sweeping power. Paul says that the son is an
automatic heir whether he has been formally instituted as such or
not: ‘For this reason, even if they [i.e., sons] have not been for-
mally instituted as heirs, they are still masters; and it is no objec-
tion that it is permissible to disinherit them or that it was once
permitted even to kill them.’87) Likewise, the author of the late
Collatio Mosaicarum et Romanarum legum, this time explicitly under the
84) A. M. Rabello, Il ius occidendi iure patris della lex Iulia de adulteriis
coercendis e la vitae necisque potestas del paterfamilias, in: Atti del Seminario Romanistico
internazionale, Perugia-Spoleto-Todi, 11-14 ottobre 1971 (Perugia 1972), 228-42.
85) D. 48.8.2: Inauditum lium pater occidere non potest, sed accusare eum apud praefec-
tum praesidemve provinciae debet. The conviction that no decent or sensible father would
act so rashly as to harm his own son without permitting him ‘a legal hearing’ was
already a common sentiment much earlier; compare Tac. Ann. 4.11 on Tiberius
and Drusus.
86) W. Selbe, Vom Ius Vitae Necisque zum beschränkten Züchtigungsrecht und zur
magistratischen Züchtigungshilfe, The Irish Jurist 1 (1966), 136-50, at 137-8, with ref-
erence to the modern discussions.
87) D. 28.2.11 (Paulus libro secundo ad Sabinum): hac ex causa licet non sint heredes
instituti, domini sunt: nec obstat quod licet eos exheredare, quod et occidere licebat. Accepted
by Selbe, op. cit., 143, but, again, the critical change in the tense of the verb licere
is surely signalling the diVerence between present practices and present beliefs
about what might have been true in the past.
heading de adulteriis, claimed that the jurist Papinian spoke of a ‘royal

law’ that conferred such a power on the paterfamilias:88)
Cum patri lex regia dederit in lium vitae necisque potestatem, quod
bonum fuit lege conprehendi, ut potestas eret etiam liam occidendi,
velis mihi rescribere: nam scire cupio.
Since it is known that the author of the Collatio is able to quote

accurately from original laws, there is some probability that words
like these did appear in Papinian’s treatise on the adultery law.89)
But it is most important to note that the precise words at issue are
not Papinian’s own.90) They are those of an interested dilettante
inquirer who has put the question to Papinian, the learned jurist.
The question seems to come out of a popular milieu where a ques-
tioner has assumed the existence of a father’s ‘right of life and
death’ and is therefore confused by the express provisions of the
adultery law with respect to a father’s right to kill his daughter
(called, notably, a ius occidendi). Papinian rightly corrects the inquirer
and states that the adultery law has nothing to do with any sup-
posed potestas vitae necisque, but that it is concerned with establishing
the conditions under which it was the father’s obligation to kill his
daughter. If the father killed the adulterer himself, then he was
duty-bound to ‘exercise his right’ and to kill his daughter as well.
The law has nothing to do with any general paternal right to kill
children.
88) Collatio, 4.8.1; probably dating to about A.D. 427, and almost certainly com-
posed in a North African milieu: E. J. H. Schrage, La date de la Collatio Legum
Mosaicarum et Romanorum étudiée d’après les citations bibliques, in: J. A. Ankum et al.
(ed.), Mélanges Felix Wubbe (Fribourg 1993), 401-17.
89) M. Hyamson, Mosaicarum et Romanarum Legum Collatio (London 1913), xxxvii;
Rabello, op. cit. (1972), 233 f. reviews the principal objections, mainly those of
Albanese, op. cit., 363, n. 3: the whole passage is an interpolation, it is entirely
spurious; and of Beseler, ZSS 45 (1925), 453: the diction is decidedly non-classi-
cal and not strictly juristic and so “the style and content are not those of Papinian,
but rather of a rhetor and commentator”. Rabello has shown that both objections
fail based on misunderstandings of the construction of Papinian’s Liber singularis de
adulteriis (written in a question and response mode) and the fact that the problems
of the ‘unjuristic’ diction are found in the words of the questioner, not of Papinian
himself.
90) Therefore it cannot be taken, as it is by Harris, op. cit., 82, and many oth-
ers, as evidence that “Papinian asserted that a lex regia gave this same power” (i.e.,
the right of life and death to fathers).
66 BRENT D. SHAW
Of the real existence of this lethal power of the father in the

later empire there is, to tell the truth, no direct evidence—none at
least that will withstand close scrutiny. In an imperial constitution
issued in May of 323, the emperor Constantine prefaced the fol-
lowing remarks to the main body of his constitutio which concerned
the possibility that free persons who had been enslaved could assert
their freedom or libertas:91)
Libertati a maioribus tantum impensum est, ut patribus, quibus ius
vitae in liberos necisque potestas permissa est, eripere libertatem non
liceret.
Constantine wished to assert that freedom was such a valuable thing

that even fathers, to whom the power of life and death over their
children had once been granted, were not permitted to remove it.
The possibilities are two. Interpolation of the phrase must be one.
But if that option is excluded, the next best reading—the one that
I would accept—is that Constantine himself accepted the traditional
ascription of such a right to fathers in the past as a striking exam-
ple of an extreme range of power.92) In his decisive refutation of
claims that this constitution attests to the continuing strength of the
‘right of life and death’ in Constantine’s own day, Bernardo Albanese
concluded with admirable clarity: “La quale non v’ha dubbio che
si presenti come un rilievo storico, onde se ne può trarre almeno
una legittima presunzione nel sense che il ius vitae ac necis in esse
menzionato sia considerato nell’altro che un istituto arcaico, carat-
teristico dei maiores, ma assolutamente desueto al presente”.93) The
conclusion that the notice in the constitutio is an historical aside
accords very well with the whole ideology of the Constantinian
restoration of the Roman state and with the emperor’s manifest and
91) CTh 4.8.6.pr [18 May 323]; the important parallel text in CJ 8.46.10 (the
title is de patria potestate) makes the temporal force of permissa est clear with the pre-
cise words olim erat permissa. That the permissa est of the constitutio must, because of
logical and historical reasons, be translated as an ordinary past tense was demon-
strated decisively by Albanese, op. cit., 343-50.
92) Which seems probable on the basis of the conscious archaizing force of the
law, and its expression, under Constantine, especially as regards his laws govern-
ing family relationships: J. Evans-Grubbs, Law and Family in Late Antiquity: The
Emperor Constantine’s Marriage Legislation (Oxford 1995), especially her summation at
338-41.
93) Albanese, op. cit., 350.
latent appeals to the past as justi cations for his parallel interest in
restoring the integrity of the family.
Last of all, and dating to a much later period, there is a detailed
legal commentary on the text of Gaius composed by a learned
scholar, probably a Gallic school master, of the late fourth or fth
century.94) The author does indeed mention the existence of this
power, but in a passage explaining the meaning of noxal surrender
that is so fragmentary that it is diY cult to know precisely what was
intended or understood:
Ego cum praetor dedere dom . . . parentem putes . . . iure uti . . .
[do]mino vel parenti et occidere eum et mortuum dedere in no[xam]
. . . patria potestas potest . . . [n] . . . cum patris potestatis talis est ut
habeat vitae et necis pot[estatem].
De lio hoc tractari crudele est, sed . . . non est post . . . r . . . [occ]idere
sine iusta causa, ut constituit lex XII tabularum. Sed deferre iu[dici]
debet propter calumniam.
The most that can be elicited from the passage is that the late
Gallic commentator, in glossing a text on the surrender of the delin-
quent son (or his body) himself believed that patria potestas had once
encompassed a power of life and death. But there is nothing to
show that this pedant is doing anything other than repeating a sort
of conventional wisdom on a power or right that was nowhere con-
ceded by the jurist Gaius himself. The only time that Gaius recog-
nizes such extensive powers is when he discusses the absolute power
of a slave owner over his slaves.95) Apart from the uncertainty over
94) Therefore not Gaius himself, as Harris, op. cit., 82, avers. Text: Fragmenta
Interpretationis Gai Institutionum Augustodunensia, 4.85-6: FIRA, II, 207-8, at 224. For
background see E. Chatelain, Fragments de droit antéjustiniens tirés d’un palimpseste d’Autun,
RPh 23 (1899), 169-84 [text: 177-81]; P. Krüger, Der Kommentar zu Gai Institutiones
in Autun, ZSS 24 (1903), 375-408 [text: 378-404]; C. Ferrini & U. Scialoja, BIDR
13 (1900), 5-31; ending with the standard edition by E. Seckler & B. Kübler,
Fragmenta Augustodunensia, in: Iurisprudentiae anteiustinianae reliquiae, 2.2 (Leipzig 1927),
454; cf. C. Ferrini, Sui frammenti giuridici del palinsesto di Autun, in: Opere di Carlo
Ferrini, vol. 2: Studi sulle fonti del diritto romano (Milan 1929), 421-36; cf. Voci, op. cit.
(1980), 74-6.
95) Gaius, Inst. 1.52: In potestate itaque sunt servi dominorum. Quae quidem potestas iuris
gentium est. Nam apud omnes peraeque gentes animadvertere possumus dominis in servos vitae
necisque potestatem esse. Gaius clearly distinguishes this extensive power from the pow-
ers inherent in patria potestas which he then proceeds to outline—and in such detail
that there is no chance that he would have omitted the ‘power of life and death’
if he had believed it to be a legal power actually wielded by the patres of his own
68 BRENT D. SHAW
what the commentator’s original statements were, none of his words

can be taken to re ect anything more than his own views, which
in turn were based on traditional assumptions and interpretations
that he accepted as fact, including the claim about rules supposedly
contained in the Twelve Tables.
It is diYcult to nd explicit references to such a power, other
than the supposed ‘real life’ cases that have been taken by modern
historians to exemplify its use. Dionysius of Halicarnassus, in a work
signi cantly entitled „Rvmaók¯ ƒArxaiologÛa, is the only author to
state explicitly that a law attributed to the rst king Romulus gave
fathers the right to enforce extraordinary physical punishments on
their sons, even to the point of killing them.96) Augustan and later
sources tend to agree that this was an ancient prescription of the
‘ancestral constitution’ of the Romans, a precept of Romulus that
was later embedded in the Twelve Tables.97) The problem is that
no reasonably dependable reconstruction of what are believed to be
the contents of the Twelve Tables accords any such legal power to
the father. An explanation for the appearance of this power in the
creation of foundational Roman Urgeschichte in the Augustan age,
therefore, seems to have more to do with the ideologically driven
reconstructions of earliest Roman social relations that were typical
of this later age.98)
Certainly, it was widely believed by fathers that they had or should
have such extensive powers. In the early empire, ‘the power’ was
averred to repeatedly in the rhetorical school exercises, as is demon-
strated by a school debate joined by various jurists, including the
eminent Labeo, over the legal position of a son who had raped a
girl. The young man was able to persuade the girl’s father to accept
a settlement for the crime, but he was unable to assuage the anger
time. The whole phrase is repeated in D. 1.6.1, again de ning the powers of mas-
ters over slaves (with nothing at all about patres familias and children).
96 ) Dion. Hal. 2.26.4 [= FIRA I2, ‘Leges Regiae,’ no. 1.8]: [ „O „Rvmælow]:
‘pasan . . . ¦dvken ¤jousÛan patrÜ kayƒ ußoè, kaÜ parŒ p‹nta tòn toè bÛou xrñnon
¤‹n te eàrgein, ¤‹n te mastigoèn, ¤‹n te d¡smion ¤pÜ tÇn katƒ gròn ¦rgvn kat¡xein,
¤‹n te poktinnænai proair°tai.
97) Accepted as such by A. Watson, Rome of the Twelve Tables: Persons and Property
(Princeton 1975), 42-3.
98) E. Gabba, Dionysius and the History of Archaic Rome (Berkeley 1991), who clearly
explicates this side of Dionysius’ work, although his main purpose is to understand
the work as a response to critical Greek readers.
of his own father. The son then retaliated against his father by
accusing him of insanity (dementia). In the various scenarios that are
canvassed on the legal problems involved, Latro, one of the lawyers,
wonders if a ravisher could summon anyone to a court during the
thirty-day period that he himself was awaiting trial and, even if that
was possible, could he begin a legal action against his own father
‘who has been granted the power of life and death’ over him’? In
the discussion that follows, it is clear that what Latro means to sug-
gest is not that the father literally had the legal right to kill his son,
but rather that he ideally possessed a species of total power over
the dependent son.99)
However improbable some of the imaginary scenarios of the
schools might seem, life and rhetoric sometimes did imitate each
other, although with diVerences that are highly signi cant for our
analysis. For example, one rhetorical exercise involving the power
of life and death imagined a lurid case in which a father, in an
uncontrollable t of rage upon discovering that his son was involved
in an incestuous union with his own (i.e., the father’s) wife, was
moved to kill his son. The emperor Hadrian actually had to decide
a real life case involving precisely the same criminal act: a father
had conspired to kill his miscreant incestuous son by making the
murder look like a hunting accident. In the historical case, the
emperor’s judgment was plain: it was murder. The killing was done
not iure patris, but rather iure latronis.100) In another model case from
the rhetorical schools, the debater, Gallio, felt that the diVerence
between a grandfather and a father was that the former could rep-
rimand or discipline a grandson in order to protect him, whereas
the father could kill his son.101) The right is proferred not as for-
mal law, but rather as part of ‘nature’s laws’; even so, during the
course of the debate the legality of the claim is refuted.
Further removed from the speci c references to a ‘right of life
99) Sen. Controv. 2.3.11: Etiamsi cum alio potest, an cum patre possit, quoi vitae mor-
tisque arbitrium datum est . . .?
100) Quintilian, Declam. mai. 19.2; cf. 6.14 (on which, see below); these ‘decla-
mations’ were composed perhaps during the second century: Michael Winterbottom
ed. & comm., The Minor Declamations ascribed to Quintilian (Berlin-New York 1984),
xiv-xvi; for the real-life case, see D. 48.9.5 (Marcianus).
101) Sen. Controv. 9.5.7: Habet sua iura natura, et hoc inter avum patremque interest,
quod avo suos servare licet, patri et occidere.
70 BRENT D. SHAW
and death’ in such rhetorical exercises, but eventually connected

with them, are collections of moral exempla made historical by Livy,
poetic by Vergil, and reduced to moral lists by Valerius Maximus.102)
On close scrutiny these vignettes fail to illustrate a primordial legal
power to kill children vested in the paterfamilias. But they are recounted
in a manner, a style, and in a rhetorical frame that is clearly meant
to suggest as much to the reader. Whatever their ‘historicity’, how-
ever, these exemplary stories were especially powerful since they
were one of the real recursive ways in which the moral norms of
the Roman élite were inculcated into themselves and their children.
As elaborated by Livy and Dionysius, the raw historical details of
fathers who had killed their sons were fed into the practice sessions
of the controversiae and suasoriae of the rhetorical schools. The pre-
cise historical details of such cases, which could not by themselves
attest a discrete power in the hands of the father, were therefore
attached to the loaded phrase ‘power [or] right of life and death’
that suggested such a genuine legal basis. For example, in the decla-
mationes attributed to Quintilian, one such story is taken as an
aYrmation of a general principle: ‘I, the father, have commanded.
This name (sc. pater) is greater than any law. It is we who escort
the tribunes, we who create magisterial candidates. To us has been
permitted the right of life and death.’103) The aggressive bragadoc-
cio is the imagined outburst of an authoritarian father who, it might
be noted, vaunts the power as extra legal. And in the subsequent
debate the bold statement itself is repudiated and is recognized for
what it was: a piece of paternal bombast. Even so, the recon guration
of the facts of the past as presented by historians into exemplary
stories of a lethal paternal power furnished raw materials for the
rhetorical practice sessions and for the recording of enunciations of
the type made by Quintilian’s pater. The discursive texts of the sua-
102) W. M. Bloomer, Valerius Maximus and the Rhetoric of the New Nobility (Chapel
Hill-London 1992), 11-7, 254-9. Given what follows, it is important to note that
Bloomer’s main argument (not one on Augustan ideology) is that the exempla were
produced in the main for direct exploitation by those training lawyers, rhetors,
and their kind.
103) Ps.-Quintilian, Declam. mai. 6.14: Pater iussi. Hoc nomen [sc. ‘pater’] omni lege
maius est. Tribunos deducimus, candidatos ferimus, ius nobis vitae necisque concessum est. Si non
fecerint quod iubes, non deferam illum ad sepulturam. For deducere see E. Maróti, Deducere—
deductores, Oikoumene 5 (1986), 236-2, pt. iii, Deductores, 241-2, citing Cic. Pro Mur.
70.
soriae and controversiae therefore both rhetoricized and reinforced desired

ideals of authority, although in the penumbra of formal laws that
established rather diVerent norms governing powers wielded by
fathers.
On a cold and unprejudiced reading, none of the above items
of evidence will sustain the claim of a primordial legal right found
in a primaeval age of Roman society when the powers of the father
were somehow “merged with those of the state”. 104) There are
diVerent ways in which the historical development of claims about
the Roman father’s power of life and death can be analyzed. One
historical model posits a strictly linear evolutionary view of the devel-
opment of the Roman family and of the powers exercized by its
constituent members. According to this view, the further back one
goes in time, the more severe and formidable were the real powers
of the paterfamilias. Thus the potestas that amounted to ‘life and death’
that was aYrmed in the adoption (adrogation) ceremonial is a ver-
bal ‘survival’ of realities from much earlier periods in Roman social
relations. A diVerent interpretation of the development of family
relationships, as essentially contested and constructed, however, oVers
an alternative model within which the sparse records of later date
would assume a diVerent valence. And such a model would be just
as congruent with the existing facts.
An interpretation of this type would emphasize the later devel-
opment of a formalized legal patria potestas in tension with the unremit-
ting forces bearing against and shaping the power of fathers. In
such a struggle for legitimacy, the rhetorical assertion of ‘pre-legal’
customary powers vested in the father could draw on a more gen-
eral rhetorical eld in which the phrase ‘power of life and death’
was used to give expression to the desired state. The expression of
total power over dependents embodied in the phrase potestas vitae ac
necis was in fact not only attributed to fathers, but also to magis-
trates of the state, to soldiers and, by the fearful and the suspicious,
104) This is the essence of Yan Thomas’ argument, op. cit., esp. at 524 V. (for
‘droit primordial’ see 508). In the light of his analysis, Thomas claims to have
opened up a hitherto unseen chapter of early Roman history: 528-9: “Il faudrait
ici ouvrir tout un chapitre sur le contenu politique de la patria potestas à Rome: la
vitae necisque potestas ne peut s’interpreter en dehors d’un droit paternel qui est pu-
blic autant qu’il est privé . . .”.
72 BRENT D. SHAW
to the power of Greek doctors over their Roman patients.105) In the

nal instance, any person who was thought to wield such regal
power over the lives of others was said to possess a vitae necisque
potestas. It was a powerful rhetorical image with which one could
conceive, think about, and aY rm sweeping and extensive authority.
The words were ones that could be used formally to assert the
extent of powers desired by Roman fathers, given the constant
threats to familial material goods and family authority which they
faced.
The same authority was attributed, equally symbolically, to the
father of all fathers, the princeps as pater patriae. Amidst the great
uncertainties in the city in A.D. 69, Vitellius, the sometime ruler
of the Roman world, attempted to resign his powers as emperor.
He tried to play the part expected of him in a piece of political
theatrics staged in the forum before the Roman people, heavily
laden with ritual signi cance. In his nal act, in a scene redolent
with pathetic appeals to his infant son and to family obligations,
Vitellius tearfully turned to the consul, Caecilius Simplex, who was
standing by him: exsolutum a latere pugionem, velut ius necis vitaeque civium,
reddebat.106) The consul refused to accept the oVer of the symbolic
dagger. Having failed even at resignation, Vitellius trudged with his
entourage back to the palace. The dagger was of particular signi cance
as a symbol of power in the uncertain events of that long year, but
only Tacitus takes the time to explain the meaning of the dagger
as representing the ‘right of life and death’.107) Needless to say, no
105) So, for example, Seneca, Ira, 1.19.8 (the power wielded by a proconsul);
Clem. 1.1.2 (Ego vitae necisque gentibus arbiter: the power of the emperor); Phoen. 103
(of the king Oedipus); Thyest. 608 (of a king, Atreus); Ep. 4.6.8 (vitae necisque arbi-
trium: held by a bandit); Ben. 3.23.3 (by a slave master); for doctors and patients,
see Pliny, NH, 29.11.
106 ) Tac. Ann. 3.68; Tacitus then describes how Vitellius departed from the
contio, intending to go to the Temple of Concordia to place the insignia imperii there,
amongst which must have been counted the symbolic pugio. Cf. Suet. Vitell. 15, on
the same incident: Tunc solitum a latere pugionem consuli primum, deinde illo recusante ma-
gistratibus ac mox senatoribus singulis porrigens, nullo recipiente, quasi in aede Concordiae posi-
turus abscessit.
107) Galba assumed the dagger, along with the military paludamentum, in order
to show that he was no longer merely governor of Spain, but rather emperor:
Suet. Galba, 11: deposita legati suscepit Caesaris appellatione iterque ingressus est paludatus
ac dependente a cervicibus pugione ante pectus; at some point, praetorian prefects also
assumed the symbolic dagger, if SHA, Comm., 6.13 is to be believed (i.e., that there
legal or ‘constitutional’ sources on the emperor’s powers describe a

ius vitae necisque as a formal right over his fellow citizens that stemmed
from his position as father of the fatherland.108) But it was a good
image with which to think and to describe. In the same year, indeed,
the dejected soldiers of a defeated legion could imagine their fate
at the hands of those who had subdued them: et omnia in arbitrio
eorum quos vitae necisque dominos fecissent.109) In a speech attributed to
Clodius Albinus (SHA, Clod. Alb. 13.8), in which the nascent emperor
both enacted a ritualistic refus du pouvoir and also asserted old
Republican virtues, he is made to claim that senators as late as the
reign of Nero had not been afraid to express their frank opinions,
even in the face of a ruler who vitae necisque potestatem atque imperium
tunc tenebat.
Such rhetoric begins with Sallust and continues with Livy.110) For
early Rome, Livy uses it to represent the full range of powers over
a citizen’s life that was held by a magistrate of the state. Later, in
the pathetic speech made by Sophonisba to her captor Massinissa,
who pleads for her own life apud dominum vitae necisque, it is used to
indicate the dependency of a war captive.111) Here, as in the Tacitus
passage on the rebel soldiers, the use of the loaded word dominus
seems to suggest the actual legal powers of the slave master more
than anything else. Further, the Livian phrase dominus vitae necisque
demonstrates that the words vitae necisque are an hendiadys for ‘life’
and that the phrase therefore amounts to not much more than a
ready and convenient way of expressing total power of one person
over another. That is what the African king Adherbal meant in his
was a freedman who was the a pugione of the prefect Cleander); cf. Val. Max.
3.6.3. When it was that the emperors rst assumed this symbol of power is not
known.
108) But it is rhetorically foreshadowed in Sall. Hist. 1.55.13 (Maurenbrecher)
= 1.48.13 (McGushin) in Lepidus’ speech to the populus: Leges, iudicia, aerarium,
provinciae, reges penes unum, denique necis civium et vitae licentia, as also in the later school-
piece Invect. in Cic. 3.5: erepta libertate omnium nostrum vitae necisque potestatem ad te unum
revocaveras.
109) Tac. Hist. 4.62; the power, however, is imagined as one in the hands of
a master gure, not a father.
110) Sall. Bell. Iug. 14.23, in Adherbal’s speech to the Senate, notably, as in
Livy, in a rhetorical piece.
111) Livy, 2.35.2: se dominum vitae necisque inimici factum videbat (the internal dis-
cord of 491) and 30.12.12, again, amidst ritual gestures: the grasping of knees and
the touching of the right hand. These are the only instances in Livy.
74 BRENT D. SHAW
futile appeal to the Roman Senate when he spoke of himself as a

miserable person whose life depended on the power of others.112)
This is exactly the sense that the words have in the astrological
writer Manilius. As he carefully explains to the reader at the begin-
ning of the fourth book of his Astronomica, our fate, our whole life
and death, is preordained at birth. The fates are the forces that
determine the entire course of our existence: they lay down the leges
vitaeque necisque.113) Similarly, the man born in the sign of Librae
(albeit in a certain convergence) will achieve great power over the
lives of others: as a judge ‘he will set up the scales weighted with
life and death.’114) From what follows, it is clear that Manilius is
‘predicting’ the total powers wielded by Augustus over the subjects
of his empire. Just as the word nex is used by prose and poetic
authors as a way of referring to death, and vita has the same valence
for life, the combination of the two indicates the sum of human
existence.
Whatever the symbolic force of the formulaic words embedded
in the legal ritual of adoption, therefore, they did not produce the
real eVect of conceding to Roman fathers the legal right to kill their
newborn. On this simple fact, the Roman law is about as plain as
it can be: the right legally to kill another human was severely lim-
ited by the state, especially in the aftermath of the Augustan
Restoration. This remained true, whatever symbolic signi cance the
words inserted into the ritual formula for adrogation might have
had for a father’s power. Concordant with all other known literary
usages, the words seem to be a formulaic way of signaling a basic
fact to those who heard them. They were not the basis, pragmat-
ically or legally, of any positive right or power of a paterfamilias to
kill his children.115) It is not, therefore, the fact that this supposed
112) Sall. Bell. Iug. 14.24: quoius vitae necisque potestas ex opibus alienis pendet.
113) Manilius, Astronomica, 4.14-23, at verse 23; the same words for the absolute
power of a king in Manilius’ near contemporary, Quintus Curtius Rufus, Hist. Alex.
Mag., 4.1.22: Et cum in regali solio residebis vitae necisque omnium civium dominus; as ear-
lier Cicero, De re pub., 3.23: sunt enim omnes, qui in populum vitae necisque potestatem
habent, tyranni . . .
114) Ibid., 4.547-52, at verse 549: iudex examen sistet vitaeque necisque . . .
115) It is therefore diYcult to accept Harris’ argument, op. cit., 94, that by the
high empire the ius vitae necisque “was a convenient anachronism—a sort of pro-
tective umbrella for the real killing of sons and daughters that was going on,
namely that of infants”. He himself (94-5) agrees with what must surely be the
power was hardly ever (if, indeed, ever) exercized that is the point
of my argument. Rather, it is to draw attention to the more signi cant
fact that the ritualistic assertion of this power was deliberately sus-
tained in response to countervailing pressures, powers, and rights
asserted by other members of the family, especially wives and sons.116)
The assertion that fathers had expansive powers that encompassed
the range between life at one end of the spectrum and death at the
other, was embedded not in formal law but in the realm of daily
social behavior and quotidian verbal enunciation.117)
In any period for which there exists detailed legal delineation of
paternal power, this power was restricted to normal discipline (even
if very harsh) within the household. Anything more serious in the
punishment of the father’s own children had to be referred to the
magistrates of the state.118) A whole series of imperial rulings from
Trajan to the constitution of Valentian and Valens of A.D. 365
make manifest the restricted range of the father’s punitive powers.119)
It is often claimed that a ruling of the emperor Constantine issued
legal view: that legally even expositio counted as murder. The problem, again, is
that this will only work if one accepts some point where the father was able to
exercise a distinct legal power of giving or denying life and for Harris (93) this
can only be explained by what we have shown to be a modern myth: “. . . there
was one short phase of the Roman oVspring’s existence during which the father
could easily be thought of as giving it life if he so chose, those rst few days dur-
ing which the infanct might be exposed. When the father recognized the child by
picking it up ( liam, lium tollere), he was giving it life”.
116) J. Crook, Patria Potestas, CQ 17 (1967), 113-22; and R. Saller, Patria Potestas
and the Stereotype of the Roman Family, Continuity and Change 1 (1986), 7-22.
117) J.-C. Dumont, L’imperium du pater familias, in: J. Andreau & H. Bruhns
(ed.), Parenté et stratégies familiales dans l’Antiquité romaine (Paris 1990), 475-95, after an
exhaustive study of the imperium of the father in context, concludes: “En réalité
l’imperium du père se fonde sur les mores et doit les respecter”. Any father who used
his powers beyond what was permitted by the general mores would automatically
undercut his own authority since, Dumont argues, fathers were not as much in
command as they would like to believe or, more important, as they presented
themselves in public.
118) Selbe, op. cit., 146-9.
119) See, e.g., D. 1.16.9.3 (Ulpian), 37.12.5 (Papinian, citing a ruling by Trajan);
CJ 8.46(47).5 (AD 287); CJ 8.46(47).1 (Antoninus Pius/Lucius Verus) and 3 (AD
227); 9.15.1.pr-2 = CTh 9.13.1.pr-2 (AD 365); the latter constitution has been
argued as one that brought a formal end to the vitae necisque potestas: for decisive
arguments against, see Albanese, op. cit., 352-4, with his important conclusion (353):
“La verità è che questa costituzione ci conferma nel più chiaro dei modi che il
ius vitae ac necis è un’ istituzione da gran tempo assolutamente dimenticato e che
la patria potestas è ormai da gran tempo una temperata autorità assai vicina a quella
dei tempi nostri”.
76 BRENT D. SHAW
in A.D. 318 by which fathers who killed their children were to be

held liable for parricide was instrumental in ending the ius vitae
necisque which had been in vigorous eVect up to that date. This
interpretation has been eVectively refuted. Careful attention to the
wording of the law demonstrates that it did not end any practice
still common at the beginning of the fourth century, but that it
forcefully restated existing legal norms.120) The history of the asser-
tion of this power was never a neat progressive evolutionary devel-
opment from a putative ‘ground zero’ point in the development of
the Roman social order when the father exercized a total legal
power over dependents which was then gradually encroached upon
by the state.121) The simple linear evolutionary model is a power-
fully attractive myth. This is the story that Roman men wished to
be accepted and to be repeated as if it were established fact; it is
also a story of how often we have acceded to their wishes.
The real history is a much more complex one in which the instru-
ments of the state, especially its legal apparatuses, abetted adult
males and household heads in countervailing the unusual power of
women, sons, and other dependents in Roman society. This grad-
ual legal fabrication of clearly de ned patria potestas expressed in law
was probably a later development—that is, a social construction—
rather than a primaeval given. Not only did Roman fathers not
exercise any such life and death decision by recognizing or not rec-
ognizing a newborn infant by ritualistically raising it (for which there
is no evidence), they also did not exercize a potestas vitae necisque
understood as a legal right to kill their own children. Whatever was
meant by the phrase in terms of pragmatic application, it seems to
have been a rhetorical assertion of the ‘great powers’ of a father.
120) CJ 9.15.1 (to Verinius, vicar of Africa: issued 16 Nov. 318; posted at
Carthage 14 March 319) = CTh 9.13.1; so, most recently, Evans-Grubbs, op. cit.,
325: “whereas in classical law, a father with patria potestas had the legal right [m.i.]
to put to death an errant child”—clearly refuted by Albanese, op. cit., 344-5, cf.
Th. Mommsen, Römisches Strafrecht (Leipzig 1899; reprint: Graz 1955), 617-20;
rather, the real place of the emperor’s ruling here is more in line with the rest of
Evans-Grubbs’ persuasive arguments that Constantine was more usually solidify-
ing existing legal and customary traditions.
121) R. Saller, Roman Kinship: Structure and Sentiment, ch. 1 in B. Rawson &
P. Weaver (ed.), The Roman Family in Italy (Oxford 1997), 7-34, oVers a provoca-
tive critique of the linear evolutionary view of the development of the structure
of the Roman family.
As such, it was contained in formulaic phrases and rituals created

in the process of the formation of the legal apparatuses that came
to de ne patria potestas in the middle Republic. Anthropological claims
for the hoary antiquity of this power is no doubt what such author-
itative statements meant to suggest, and which antiquarians—surely
Dionysius of Halicarnassus was not alone amongst them—intended
to picture in their foundational charter stories of Rome’s origins.
My argument, in short, is that the colorful stories, exemplary
vignettes, and rhetorical ourishes that imputed the wide-ranging
legal force of the father’s potestas vitae necisque are also a myth, but
one that is much more diYcult to deconstruct than the modern
baby-lifting ritual ascribed to the Romans. Unlike the purely ctitious
ceremony of tollere liberum, the father’s ‘power of life and death’ at
least has the legitimacy of being a genuine Roman construction—
one deeply believed in by its citizens, high and low. By anchoring
this power in their foundation myth as a primordial power regu-
lated by Romulus, they meant to guarantee its unquestioned status.
It was probably as powerful and persuasive as that other fabrica-
tion accepted by the most critical and eminent of its own histori-
ans—that from its origins the city of Rome had been ruled by kings.
Philadelphia , University of Pennsylvania

Raising and Killing Children: Two Roman Myths: Liberum. Recent Standard Treatments of The Roman Family Have

Uploaded by

Copyright:

Available Formats

Raising and Killing Children: Two Roman Myths: Liberum. Recent Standard Treatments of The Roman Family Have

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Raising and Killing Children: Two Roman Myths: Liberum. Recent Standard Treatments of The Roman Family Have

Uploaded by

Copyright:

Available Formats

RAISING AND KILLING CHILDREN:

TWO ROMAN MYTHS

Often the purpose of historical research is to create by explana-

© Koninklijke Brill NV, Leiden, 2001 Mnemosyne, Vol. LIV, Fasc. 1

aYrmed, occasionally in intriguing detail, the nature of the liminal

Raising the Newborn: Tollere Liberum

Quite colorful, but almost none of it underwritten by any evidence.

is exposed, sold or put to death. Even if it survives, it is excluded

Declareuil’s statement is cited primarily for the clarity with which

the words tollere liberos had come to be construed in this context in

81 (1986), 140-6: J.-M. Carrié, Le ‘colonat du bas-empire’: un mythe historiographique?,

the individual literary sources (an ever increasing number of which

loco cernentesque alios in possessione dignitatis suae, salii  aminesque

person with children is left to languish in obscurity.26) The context

According to this interpretation, the lifting up of the newborn infant

words, or their equivalent, do not occur in even one of the laws

labor that unusual powers were conceded to them by Roman soci-

In a review of the whole question published in 1951, however,

the ritual of the lifting of infants, and their children will be as if

it is clear that the stereotypical phrase to which Volterra drew atten-

Alan Watson has already pointed out in a critique of Volterra’s

of the clause, in Watson’s view, “is to explain to the relatively une-

military men a special privilege: the right of legal Roman marriage

Despite the slightly diVerent terminology, the meaning in context

Moreover, a review of literary texts that make the meaning of the

Nevertheless, the pervasive power of the mental image of the child

as cultural, and legal, precedent. The latter might happen by a pur-

a modern myth. In the texts that have been deployed to sustain

The Power of Life and Death

of patria potestas, denoting the widest extent of the power enjoyed

historians, the statements by Dionysius of Halicarnassus on the ‘royal

early Roman history was also underwritten by anthropological reports

Cicero provides is to a formulaic phrase, embedded in a ritual for

acts of killing done by fathers who were exercizing speci c powers

sons. In his version of Cato’s speech on the fate of the conspira-

It is these same formal words that are alluded to by Cicero in the

Once modern readers pass beyond these two references, however,

process of identi cation. Since there was no other statutory basis

heading de adulteriis, claimed that the jurist Papinian spoke of a ‘royal

Since it is known that the author of the Collatio is able to quote

Of the real existence of this lethal power of the father in the

Constantine wished to assert that freedom was such a valuable thing

what the commentator’s original statements were, none of his words

and death’ in such rhetorical exercises, but eventually connected

soriae and controversiae therefore both rhetoricized and reinforced desired

to the power of Greek doctors over their Roman patients.105) In the

legal or ‘constitutional’ sources on the emperor’s powers describe a

futile appeal to the Roman Senate when he spoke of himself as a

in A.D. 318 by which fathers who killed their children were to be

As such, it was contained in formulaic phrases and rituals created

Philadelphia , University of Pennsylvania

You might also like

loco cernentesque alios in possessione dignitatis suae, salii aminesque

acts of killing done by fathers who were exercizing speci c powers

process of identi cation. Since there was no other statutory basis