While the freedman procurators and heads of departments from early in the second century were systematically denied equality of status with their equestrian replacements or counterparts, it occasionally happened that individual freedmen were advanced to membership of the equestrian order itself and full parity of status with the equestrians. In even rarer cases senatorial honours, the ornamenta praetoria and quaestoria, were granted. The main legal barrier which the freedman aspiring to equestrian status had to overcome was that of free birth. Freeborn status (ingenuitas) was conferred by the grant of the anuli aurei, which was in the hands of the emperor, and emperors, both good and bad, in the first and second centuries were very sparing in conferring such free birth on their own freedmen and in admitting them formally to the ordo equester. The very paucity of such instances emphasises the difference in status between the Imperial freedmen in general and the equites, and in particular is an argument against any intermingling of the freedman and the equestrian procuratorial cursus.

Augustus (Cassius Dio, liii. 30.3) granted the anuli aurei to his successful physician Antonius Musa, but there is no need to suppose that he did the same for his notorious procurator in Gaul, Licinus. On the other hand, Antonius Felix was almost certainly an eques when he was appointed procurator of Judaea in 52 and given command of troops there. His brother, Antonius Pallas, was even more highly honoured under Claudius by the ornamenta praetoria decreed him by the Senate in 52.

It is not possible to establish precisely defined periods within the overall span of two and a half centuries of the Familia Caesaris' effective existence. The dating criteria applied to the inscriptions are most fruitfully based on aspects of personal nomenclature which are not susceptible of rigid demarcation, especially the Imperial nomina gentilicia. We usually have to be content with a chronological tolerance of twenty years or even more. This means that not only does the possible date of a particular inscription overlap the reigns of two successive emperors, but not infrequently of three or even four.

For this reason a reign-by-reign account of the development of the Familia Caesaris is to be generally avoided as being methodologically dangerous and misleading. Unless, of course, it is somehow based on the literary sources – a procedure which, although partly satisfying the desire for precision, is even more detrimental to truth. Uncomfortably often we have to be content with sensing general trends, not with seizing on dramatic reversals of policy traceable to a single decision. The influence of individual emperors on the general development of the Familia Caesaris, both in its administrative and in its domestic services, can only be felt within quite narrowly defined limits. On special issues or for particular appointments the workings of the system of patronage give the emperor a decisive voice. By virtue of his position in the state he was required to deal personally with an incredible amount of detailed, day-to-day business.

It is clear that personal slaves of Imperial slaves must have existed in considerable numbers. What was their social status and how does it compare with that of the vicarii above? To find out this we must examine the inscriptions of the slaves and freedmen of the Augusti liberti, as both personal slave and slave-master, and particularly the latter, had often gained manumission by the time of life at which their inscriptions were put up.

The nomenclature of the Aug(usti) lib(erti) libertus presents few problems. Normally in inscriptions the lib. libertus takes the nomen gentilicium while the freedman-patron is referred to by his cognomen only – no doubt to save space as both have the same nomen; e.g. VI 4281: Ti. Iulius Demetrius Anthi Aug. 1. lib., Ti. Iulius Anthi Aug. 1. lib. Hilario, where the status indication is exhibited in different positions within the same inscription. The terms ‘patronus’, ‘libertus’, etc. are used as normal. But ‘verna’ is frequently not an indication of slave status at all, being in fact used for lib. liberti who died early, or for slaves who may have been manumitted at the point of an early death, or indeed for freeborn children of liberti. As for the slaves of Imperial freedmen, not only are their inscriptions much fewer in number than those of the lib. liberti, but also they mention the nomen of their freedman-master much less frequently – in fact in less than a third of the examples (12/40), including several vernae who are presumably, but not certainly, slaves.

In the study of social structure in the Roman world of the first and second centuries ad nothing is more important or more complex than the slave and freed slave classes. Their numbers, although not absolutely determinable, were large and even predominant in many urban and some rural areas. Most urban slaves of average intelligence and application had a reasonable expectation of early manumission and often of continued association with their patron. They enjoyed a high rate of social advancement, which was often much greater than that of the freeborn proletariate. The fundamental social legislation of Augustus attempted to provide a stable social hierarchy based on legal status. But, at the same time, there was in the early Imperial period a degree of social mobility sufficient to prevent the structure breaking down in violence and social discontent. Among the mobile sections of society the slave-born classes played a significant role.

But the status ladder within these classes themselves is both long and complex. From the point of view of juridical status, there is not only the fundamental distinction between slaves (servi) who are without rights, and freed slaves or freedmen (liberti) who are citizens, but there are also further distinctions within each of these classes – formally manumitted freedmen with full citizenship, informally manumitted freedmen (Latini Iuniani) who did not acquire Roman citizenship, but only Latin status without full political rights, and dediticii who could never become Roman citizens.

The origin of this work was a footnote in an article by Professor A. H. M. Jones on the Roman Civil Service. Since 1955 the footnote has grown into a book and beyond, largely because neither he, as I suspect, nor I certainly, envisaged that the evidence would present itself in quite such formidable quantities or pose such challenging problems of interpretation and presentation. These factors also made necessary a long series of preliminary studies, which in turn have contributed to the long delay in putting this book into definitive form. The nature of the material and of the writing is intractable, and it has become increasingly clear to me that time and effort will not bring much improvement. Hence I have thought it better to present the subject all together as it appears to me now rather than to pursue an unattainable perfection at the cost of an indefinite delay.

This is also an appropriate moment for stocktaking. Until quite recently the Familia Caesaris was an almost totally neglected field, despite its obvious importance for early Imperial social and administrative history, as well as for onomastics. But in the last few years two major studies in particular have appeared by G. Boulvert and H. Chantraine. I have taken this opportunity to point out the areas where progress and agreement can be recorded and where the main problems still lie.

EARLIER CHRONOLOGICAL LIMITS

An Imperial nomen in itself only gives a fixed earlier chronological limit, or terminus post quem, for an Augusti libertus. Thus, when used in conjunction with the Imperial praenomina, the earliest dates possible for the following are: C. Iulius Aug. l. 27 bc; M. Iulius Aug(ustae) l. ad 14; Ti. Iulius Aug. l. ad 14; Ti. Claudius Aug. lib. 41; Ser. Sulpicius Aug. lib. 68; T. Flavius Aug. lib. 69; M. Cocceius Aug. lib. 96; M. Ulpius Aug. lib. 98: P. Aelius Aug. lib. 117; T. Aelius. Aug. lib. 138; M. Aurelius Aug(g). lib., L. Aurelius Aug(g). lib. 161; L. Septimius Aug. lib. 193, Augg. lib. 197, Auggg. lib. 198. The freedmen of the Aurelii Augusti: M. Aurelius (161–80), Commodus (177–92), Caracalla (197–217), Elagabalus (218–22), Severus Alexander (222–35), all have the same nomen and praenomen, except some with the praenomen ‘L.’ manumitted either by L. Verus or by Commodus at certain periods, probably 177–80 and 190–2. The separate freedmen of these emperors cannot be distinguished by nomenclature alone, except the cases of M. Aurelius Auggg. lib. which are post 198 when Geta became Caesar. It is sometimes possible to distinguish between the freedmen of Augustus and Gaius by means of the forms of the status indication characteristic of the former, e.g. Caesaris l., Caesaris Augusti l. But separation of the freedmen of Claudius and Nero, and those of Vespasian, Titus and Domitian is very difficult in the absence of other dating criteria.

If the Familia Caesaris is indeed an élite in the slave and freedman classes of society, we would expect its marriage pattern to be significantly different from that of the slaves and freedmen outside the Familia. These slaves and freedmen themselves need not be considered as economically underprivileged; the only material available for such a study is their tombstones or funerary inscriptions – and the mere fact of being able to afford a tombstone, with or without decoration, as many of them conspicuously had, or of having access to a burial monument of a patron, implies security of a kind somewhat removed from the poverty of the ghetto.

A detailed study of the marriage pattern of these classes has yet to be made. The following discussion aims merely to supply material for such a study and to examine a control group as a basis of comparison between the Familia Caesaris and the slaves and freedmen outside it. To obtain a sufficiently large and homogeneous sample the sepulcrales of CIL VI, Parts 1–4, have been used. From these have been extracted the cases of marriages where the slave or freedman status of one of the partners is specifically stated or can be inferred with certainty (e.g. from the occurrence of ‘collibertus’, ‘patronus’, etc.), and where both partners are named and the relationship between them is specifically that of man and wife.

TABULARIUS A RATIONIBUS

In the career-type inscriptions of the Imperial freedmen, senior clerical status is very rarely mentioned in the sequence of posts held, although it can be assumed that the vast majority of all freedmen who reached the senior rank of procurator in the administrative service must have held at least one post in the clerical grades, probably of senior clerical status, before rising to procuratorial status. But here mention should be made of one type of exception which is significant – the careers of those who held the post of tabularius a rationibus. The examples are:

1. (1) T. Flavius Aug. lib. Delphicus, tabularius a ratio[nibus, p]roc.ration(is)thesaurorum, hereditatium, fisci Alexandrini (AE 1888, 130 = d 1518).

2. (2) Martialis A[ug. lib …], tabulariu[s a rationibus?], proc. fiscorum [transmarinorum e]t fisci castr[ensis, pr]oc. h[ereditatium et fisc]i libe[rtatis et peculiorum] (VI 8515, with the restorations of Sanders, Mem. Amer. Acad. Rome 10 (1932), 81).

3. (3) T. Aelius Augg. lib. Saturnin(us), pr[oc. pro vine] Belgicae […], proc. fisci libertatis et peculior(um), tabul(arius) a rationibus, tabul(arius) Ostis ad annona(m) (VI 8450 = d 1521).

The prestige of a post held in the central department of the a rationibus was always high. This is confirmed by the fact that only in the a rationibus is the clerical grade of tabularius recorded as giving access to the procuratorial grade. But it is the chronological distribution of the tabularii a rationibus that is significant here.

The literary sources show clearly that the Familia Caesaris in general was an elite among the slave and freedman classes of Imperial society. But it is also clear that this was not equally true for all its members, nor for all periods. A detailed examination of the voluminous inscriptional material reveals a considerable degree of social differentiation within the Familia Caesaris.

In nomenclature, apart from the legal differentiation between vicarius, servus, libertus, etc., the use of status indication, agnomen and occupational title are all significant for status and, with some exceptions (e.g. aquarii) and some regional variations (e.g. Africa), are rarely used by those of low status. The occupational elite are the professional civil servants of the clerical and senior administrative grades who could rise to auxiliary procuratorships or to sole headships of departments, always distinct from, but within measuring distance of, the equestrian procurators themselves. Those in the clerical–administrative service, whether through background or connections within the Familia Caesaris, patronage, training or ability, began their professional careers early and were distinctly superior to those who spent their lives in sub-clerical or domestic occupations. With the exception of posts of special opportunity or responsibility within the Palace, there was little chance of crossing the occupational dividing line upwards into the administrative service. In terms of location, Rome held the advantage over the provincial centres for most administrative grades, with the exception of those higher freedman careers where movement from one centre to another was more frequent, and among the slave dispensatores.

‘Familia’ appears as early as the Twelve Tables (v. 4–5) in the sense of ‘household’ or all those persons under the potestas of the pater familias, including a wife in manu and a filius familias, as well as slaves. But in a narrower sense it referred specifically to slaves (famuli) in relation to their dominus. The potestas of the pater familias over his servus was the same as that which he exercised over his filius familias or wife in manu, and a qualitative distinction between patria potestas and dominica potestas should not be pressed. A pater familias is correctly so called even when he has no filius under his potestas (Dig. 1. 16. 195.2). The original household cult of the Lares presided over by the pater familias came to have a specifically slave connotation by the second century bc, and the slave vilicus came to represent the head of the household in the celebration of the cult.

By a natural extension of this meaning ‘familia’ also included manumitted slaves (liberti) in relation to their patron, e.g. Dig. l. 16. 195.1 (Ulpian): ‘ad personas autem refertur “familiae” significatio ita cum de patrono et liberto loquitur lex’. Cf. Pauli Sent. v. 1.3:’ ex officio fisci inter fiscalem familiam’. In inscriptions liberti are commonly assumed to be part of a given familia, especially in the collegium familiae and in columbaria for purposes of worship or burial, e.g. VI 479 = XIV 32 = d 6152: ‘corpori familiae public(a)e libertorum et servorum’; cf. VI 10107 = d 5212,10257,10258, etc.

From the end of the first century, and certainly from the time of Hadrian, recruitment into the various services of the Familia Caesaris was predominantly from within the Familia, from the children, and especially the sons, of the Caesaris servi and Augusti liberti themselves. This is indicated especially by the increasing use of the term verna in the status indication of Imperial slaves in the second century. Recruitment from within, it has been argued in a previous chapter, is compatible with the predominance of ingenuae among the wives of the Caesaris servi only because of the application of the SC Claudianum.

In the first century, however, slaves entered the Familia Caesaris in considerable numbers from outside. This inflow can be directly traced in all periods up to the time of Hadrian by the agnomina in -ianus derived from the previous owners. They are particularly apparent in the period of Augustus and Tiberius. The methods of acquisition were the same as those by which the emperor's patrimonium in general was increased – gift, bequest and legacy, and, no doubt exceptionally in the case of slaves, even purchase and confiscation.

A further source of recruitment was through the slaves of Imperial slaves themselves, vicarii, who were in law strictly always the property of the emperor – and thence, as what may be called ‘vicariani’, into the Familia Caesaris itself.

What was the role of women – the Augusti libertae and Caesaris servae – in the Familia Caesaris and how does their status compare both with that of the Augusti liberti and Caesaris servi on the one hand, and with that of women of the slave and freed classes outside the Familia on the other?

Libertae in general in Roman society show a much more marked upward mobility in status than servi or liberti. One important aspect of this mobility, marriage with one's patron, needless to say was denied to the Aug. libertae, but not to the Aug. lib(erti) libertae, the freedwomen of Imperial freedmen, who very frequently marry their freedman patron. The well-known case of Acte, the freedwoman concubine of Nero, belongs to the exceptional personal history of that emperor. But Vespasian also had a famous concubine, Caenis the freedwoman of Antonia Minor. He lived with her after the death of his wife Flavia Domitilla, and he maintained her without embarrassment as a de facto wife till her death (Suetonius, Vesp. 3: ‘paene iustae uxoris loco’; cf. ib. 21). Another freedwoman, Lysistrata, was concubine of no less revered an emperor than Antoninus Pius after the death of his wife Faustina. Lysistrata, a freedwoman of Faustina no less, indeed appears in an inscription after Pius' death with the actual title ‘concubina divi Pii’ (VI 8972 = d 1836: Galeria [Aug(ustae) liber]tae Lysistrate concubina Divi Pii).

It is now necessary to consider more closely the hierarchy and the system of promotion in the sub-equestrian regions of the Imperial administration. No bureaucracy can function efficiently without order and opportunity in its lower as well as its higher ranks. It is necessary to affirm this of the lower ranks of the Roman bureaucracy in the early Empire precisely because it has so often been denied or ignored. The impression gained from reading many works on the Imperial administration is that of fervid equestrian movement from post to post and province to province at the top, with a static substratum of Imperial slaves and freedmen providing stability and continuity below. Such an account is to some extent true. But that it is not true enough emerges from a detailed consideration of the inscriptional evidence.

To begin with, two general points of method in the use of the inscriptions need to be made, as they are fundamental to any reconstruction of the slave and freedman hierarchy.

In the first place, there is a difference between the manner of recording an equestrian career, where all the posts from that of, say, praefectus cohortis, held in the twenties, are recorded, and a freedman career, where in the clerical and sub-clerical grades only the highest post actually reached by the end of a career (e.g. tabularius, a commentariis, dispensator) is recorded. This is not to be interpreted as meaning that the slaves and freedmen spent their lives in a single post on a single grade.

Higher in legal and occupational status than either of the two preceding groups are what I have called the ‘vicariani’. The use of the term ‘vicarianus’ in the early Empire requires explanation.

Imperial freedmen and slaves in the period from Augustus to Trajan frequently exhibit in their inscriptions second names (agnomina) ending usually in -ianus, sometimes in -anus, and occasionally in -inus. These names are derived from the nomen or cognomen of a former master from whose familia they have passed either by gift, purchase or inheritance into the Familia Caesaris. Such agnomina are found in only a comparatively small minority of the total inscriptions of each class and period – not more than 6% of all Imperial freedmen with or without nomen from the Iulii to the Ulpii,and about the same proportion of all Imperial slaves from Augustus to Hadrian.

If every slave of a Roman master on being transferred to the ownership of the emperor regularly took a second name derived from his former owner, we would possess valuable information on the sources of recruitment of the Familia Caesaris for the first century and a half of its existence. Unfortunately this is not so. Even allowing for the incomplete nature of the inscriptional evidence, it is difficult to believe that under 10% of the Familia Caesaris – although the figure is nearer 12% for the Iulii and Claudii – was recruited from such sources in the first century ad.