¹ Both literary examples cited in TLL are Medieval: Schol. to Hor. Od. 3.24.18 (II p. 384k codd saec X and XI) and [Quint, ] Decl. Maior. 2 p. 359.5Google Scholar (from the late Excerpta Monacensis, included in Lehnert's, G. 1905 Teubner edition)Google Scholar. In both cases the word is synonymous with privignus.

² For numbers, see note 13 below. The earliest datable inscription is from the time of Hadrian (CIL 6.9041, an epitaph by P. Aelius Aug. lib. Telesphorus cf. also 6.3447), but the absence of status indication in most inscriptions indicates a date of the 2nd or 3rd century A.D.

³ e.g. TLL, OLD, Lewis, C. T. and Short, C., A Latin Dictionary (Oxford, 1879)Google Scholar, Ernout, A. and Meillet, A., Dictionnaire étymologique de la langue latine ⁴ (1959)Google Scholar, etc. TLL also gives the alternative meaning ‘granddaughter’ found in late glossae (see further below, note 71) and refers to De Ruggiero (see below note 8) but without further comment.

⁴ cf. Spanish hijastro, Portuguese filhastro, Roumanian fiastru; see Meyer-Lübke, W., Romanisches Etymologisches Worterbuch ³ (Heidelberg, 1935)Google Scholar, s.v. filiaster.

⁵ i.e. born from a relationship not legally sanctioned as iustum matrimonium.

⁶ Meyer, P. M., Der römische Konkubinat (Leipzig, 1895), pp. 39–48Google Scholar, cf. Wilmans, G. on CIL 8.2848Google Scholar (L. Cornelius Cato is probably called the filiaster of M. Cornelius Cato (a soldier) because, though his son, he was not the product of a legitimate marriage). On this see also Dean, L. R., A Study of the Cognomina of Soldiers in the Roman Legions (Diss. Princeton, 1916), p. 103Google Scholar; contra (= stepfather) Lambertz, M. (Glotta 4 [1913], 101)Google Scholar.

⁷ A man's privignus is his wife's child by a former relationship, whether legal marriage or some other kind of partnership: Dig. 38.10.4.6 ‘privignus est uxoris meae filius ex alio viro natus, ego sum vitricus’; Scaev. Dig. 38.9.7 ‘privignus etiam is est, qui vulgo conceptus ex ea natus est quae postea mihi nupsit, aeque et is qui, cum in concubinatu erat mater eius, natus ex ea est eaque postea alii nupta sit’ (Stiegler, H., ‘Konkubinenkind, “privignus”, “parricidium”?’ in Sodalitas: Scritti in onore di A Guarino [Napoli, 1984–1985], viii.3191–3214Google Scholar, has recently attempted to demonstrate that the words ‘eaque postea alii nupta sit’ are an interpolation by the compiler [writing after the inauguration of legitimatio per subsequens matrimonium] and that Scaevola's words mean that an illegitimate son born before marriage, either out of wedlock or in concubinage, can be termed privignus).

⁸ de Ruggiero, E., Dizionario Epigrafico di antichità Romane (Rome, 1922), iii.88fGoogle Scholar. s.v. filius.

⁹ e.g. Rawson, B., ‘Roman Concubinage and Other De Facto Marriages’, TAPA 104 (1974), 279–305Google Scholar only considers children illegitimate if they have the mother's name and/or Sp. f. or Sp. lib.; Weaver, P. R. C., Familia Caesaris: a Social Study of the Emperor's Freedmen and Slaves (Cambridge, 1972)CrossRefGoogle Scholar only uses the term in its sense stepson; cf. also the articles listed in note 65 below.

¹⁰ Humbert, M., Le Remariage à Rome: étude d'histoire juridique et sociale (Milan, 1972), p. 202 n. 46Google Scholar.

¹¹ Humbert (writing before the publication of the computer concordance to CIL 6) obtained his examples by checking words such as privignus, vitricus, noverca and filiaster in the indices to CIL and, in the case of Vol. 6, by examining nos 4600 to 29600 (see ibid. p. 200 n. 40).

¹² Humbert includes some examples of filiaster where the sense ‘stepson’ is probable, e.g. 6.9475, 11791, 7527; see below.

¹³ i.e. all I have found by consulting the indices to CIL, including the computer index to 6 (= CIL 6, part 7) by E. Jory and D. G. Moore (1974), and the TLL archives in Munich: a total of 39 inscriptions of which 25 contain the masculine form filiaster, 12 filiastra and 2 are unclear.

¹⁴ See Rawson, , art. cit. (n. 9), 284Google Scholar, Weaver, , op. cit. (n. 9), p. 84Google Scholar, Taylor, L. R., ‘Freedmen and Freeborn in the Epitaphs of Imperial Rome’, AJP 82 (1961), 113Google Scholar.

¹⁵ All inscriptions cited are from CIL unless otherwise specified.

¹⁶ In six examples, the filiaster and his two parentes are all named, in the rest only one parent (5 cases male, 2 female).

¹⁷ Illegitimate but freeborn children derived their nomen from their mother (cf. note 26 below); a slave would be illegitimate by definition, since conubium was only possible between free or freed citizens (see e.g. Treggiari, S., ‘Contubernales in CIL 6’, Phoenix 35 [1981], 42–69)CrossRefGoogle Scholar.

¹⁸ The fact that the filiaster has the same nomen as the mother does not necessarily mean he derived it from her: he may be the child of a former marriage of hers to a person of the same gens or to her patronus.

¹⁹ i.e. whether the filiaster is the illegitimate child of the individual or couple mentioned on the tomb or whether he is the stepson of the person who calls him filiaster and the child of their partner by a former illegitimate union.

²⁰ Unless he had been adopted by him.

²¹ In that case she either bore the child late or was manumitted early, perhaps in order to marry her patron Scribonius (in which case her son derived his name from his father and Spatalus is the stepfather).

²² See Rawson, art. cit. (n. 9).

²³ Révue archéologique 9 (1852), 196Google Scholar.

²⁴ Anthion is the nominative form used by Meyer, though the index to CIL 10 lists it as Anthio. In the inscription it appears in the dative Anthioni.

²⁵ Otherwise why did he not set up the tomb himself? They are unlikely to be divorced, since the son would normally have gone with his father (cf. Gardner, J. F., Women in Roman Law and Society (London/Sydney, 1986), pp. 146fCrossRefGoogle Scholar., Rawson, B., ‘The Roman Family’, in The Family in Ancient Rome, ed. Rawson, B. (London/Sydney, 1986), pp. 35ffGoogle Scholar. Humbert, (op. cit. (n. 10), pp. 295ff.)Google Scholar points out that the usual practice was not always adhered to, especially where the father re-married, though even here the child's education, and presumably also his burial, remained the financial responsibility of the father.

²⁶ cf. Rawson, B., ‘Family Life among the Lower Classes at Rome’, CP 61(1966), 76–7Google Scholar; Buckland, W. W., A Textbook of Roman Law ³ (rev. by Stein, P., Cambridge, 1963), p. 105Google Scholar; Meyer, , op. cit. (n.6), pp. 34, 52ffGoogle Scholar.; Inst. 3.4.3, D. 38.17.2.1, P 4.10.1.

²⁷ Also it must be borne in mind that the literal meaning of filiaster is not illegitimate son or stepson but ‘a sort of son’ (see below): would a mother describe her own child in this way?

²⁸ cf. Rawson, art. cit. (n. 9); Gardner, , op. cit. (n. 25), pp. 31ffGoogle Scholar. The impediment could have been, among other things, the youth of Egnatia or the existence of a close family relationship.

²⁹ Especially as he lacks filiation, whereas the others have it: cf. Rawson, , art, cit. (n. 26), 74Google Scholar.

³⁰ Marriages between male slaves and ingenuae were rare unless the man was an imperial slave (Rawson ibid., Weaver, op. cit. (n. 9), pp. 145ff.). But in such cases the woman's status was reduced to that of liberta or ancilla (following the Senatus Consultum Claudianum) while the children (at least until the time of Hadrian) were slaves; see Crook, J. A., Law and Life of Rome (London, 1967), p. 62Google Scholar, Weaver, P. in The Family in Ancient Rome [cited above, note 25], p. 151Google Scholar.

³¹ Wilkinson, B. (Rawson), ‘A wider concept of the term parens’, CJ 59 (1964), 358–61Google Scholar.

³² Given the popularity of the name Aurelius, it is not necessary to assume that the parents were related, though Aurelia may have been the liberta or the kinswoman of her husband. It is also possible that the son was illegitimate and derived his name from his mother.

³³ Though the coincidence may not be so great, given that Restitutus is a very common cognomen. Kajanto, I., The Latin Cognomina (Societas Scientiarum Fennica, Commentationes Humanarum Litterarum Vol 36.2, Helsinki, 1965), p. 356Google Scholar, has found 485 men called Restitutus and 175 women called Restituta in CIL (pagan examples) plus 45 men and 35 women in Christian inscriptions.

³⁴ cf. 2.871 (for Lucius Accius Reburrus by Accius Reburrus and Atilia Clara phvigno pio).

³⁵ Thylander, H., Etude sur l'épigraphie latine (Lund, 1952), pp. 89–91Google Scholar. Examples include 10.2247 (an epitaph by Clodius Fructuosus and Ceionia Helias for their filius, L. Ceionius Fructuosus); cf. also 9.5932, 10.3031, 3395.

³⁶ 9.5932: L Freius Saturninus dedicates a tomb to his parentes, L. Valerius L. F. Saturninus and Freia L. Liber. Chreste.

³⁷ Though Thylander is so certain that the son should have the praenomen of the father that he refuses to believe, in two cases, that the child is the man's son, even where he is called filius (e.g. he says that at 10.2046 L. Acilius Sosus must be the stepfather of M. Amulius Epinicus [son of Amulia Maximilla] because their praenomina differ. In that case, however, they do not have cognomina in common).

³⁸ For Severinus from Severus see Thylander, , op. cit. (n. 35), p. 111Google Scholar quoting IPOA 70 (Ti. Claudius Severus' son is Claudius Severinus) and cf. Leumann, M., ‘Lateinische Cognomina auf -inus und -ilia”, Romanica Helvetica 20 (1943), 150–72Google Scholar.

³⁹ cf. Birley, E., Roman Britain and the Roman Army (Kendal, 1953), p. 121Google Scholar; Cagnat, R., Cours d'épigraphie latine ⁴ (Paris, 1914), p. 78Google Scholar and Mommsen, Th. (Hermes 19 [1884], 88ff.)Google Scholar.

⁴⁰ cf. 6.17438 where the daughter of L. Caecilius Successus' contubernalis is called Successa after her father. Perhaps the father bought freedom for his wife and son; cf. Rawson, , art. cit. (n. 26), 74Google Scholar, who suggests this solution in the case of Q. Lucretius Q. f. Zeuxis, whose wife is Sornatia Arescusa and son C. Sornatius Indus (6.21650).

⁴¹ For another example of a union between a praetorian soldier and his freedwoman cf. 6.32678 (an epitaph for M. Varsilius Martialis the patronus and coniunx for 16 years of Varsilia Stacte): cf. Watson, G. R., The Roman Soldier (London, 1969), p. 212 n. 480Google Scholar.

⁴² The most famous case, of course, being Augustus' adoption of Tiberius. For less illustrious examples see 13.2036 (for Vireus Vitalis by Val. Maximus ‘vitricus qui eum sibi filium adoptaverat’); 6.38831 (a possible example – P. Rubrius Felicissimus [age 4] is called the alumnus of his vitricus Rubrius Soter: the use of alumnus and the common nomen might suggest adoption, though Rubrius Soter may be the tutor of his stepson and a relative of the mother's first husband).

⁴³ Though he could retain his own praenomen; e.g. M. Annius Verus when adopted by T. Aelius Aurelius Antoninus became M. Aelius Aurelius Verus.

⁴⁴ Note that Pliny's original cognomen Secundus was the same as that of his adoptive father. This could be a relevant parallel to the inscription under discussion, except that even if Sabinus' cognomen was his original one, he still might be expected to have as a second cognomen his former gentile name. On rules for nomenclature in cases of adoption see Mommsen, Th., ‘Zur Lebensgeschichte des jüngeren Plinius’, Hermes 3 (1868), 59–70Google Scholar; Cagnat, , op. cit. (n. 39), pp. 74–6Google Scholar.

⁴⁵ These coincidences would not be so great if Sabinus was the legitimate offspring of his mother and another Aurelius, as the name Sabinus might be in the family.

⁴⁶ Being a rural tribe, this would suggest that they were of superior status to Spurii put in the urban Colline tribe. Fo r examples, cf. 8.2565b2 (Cornelius C. f. Pol. Antulus Castr.); 8.3151, 3101.

⁴⁷ Cagnat, R., L'Armee romaine d'Afrique et l'occupation militaire de l'Afrique sous les empereurs (revised edition, Paris, 1912), pp. 300ffGoogle Scholar.; Meyer, , op. cit. (n. 6), pp. 110ffGoogle Scholar.; Sander, E., ‘Das Recht des römischen Soldaten”, Rh.M. 101 (1958), 164Google Scholar; Watson, G. R., op. cit. (n. 41), p. 39Google Scholar, with note 77. Meyer (followed by Sander) distinguishes between Lagerkinder born from a marriage where the partners would have had conubium if the man were not in the army and those born of a couple not so qualified: in the former case, a son who has entered the army has the name of his father, in the latter he bears the mother's name, though he still has the tribe Pollia.

⁴⁸ Meyer, , op. cit. (n. 6), pp. 47, 112fGoogle Scholar. Cagnat, , op. cit. (n. 47), p. 301Google Scholar n. 8 says the father was probably attached to the legion in some capacity and the son given the same privileges as sons of legionaries.

⁴⁹ Cagnat, (op. cit. [n. 47], p. 300)Google Scholar says the children took their father's name on recruitment, just as other provincials (whose fathers weren't in the army) took the names of the reigning emperor.

⁵⁰ Wilmans ad loc.; Cagnat, , op. cit. (n. 47), p. 301 n. 8Google Scholar; Mommsen, (Hermes 19 [1884], 11 n. 2)Google Scholar; Meyer, , op. cit. (n. 6), p. 113Google Scholar. Contra (= stepfather) Lambertz, M. (Glotta 4 [1913], 101)Google Scholar.

⁵¹ cf. Starr, C. G., The Roman Imperial Navy (Cambridge, Mass., 1941), p. 92Google Scholar, who points out that soldiers and sailors tended to regard their sons as legitimate even though they were not legally so. He quotes two examples where the young child of a soldier has been given its father's name (6.3109, 10.3547), though it is possible that these children were born before the father entered military service (they died at the ages of 6 and 4 years respectively).

⁵² 6.5600, 13317, 13441, 16934, 19412, 19462, 25471 and 34064.

⁵³ For another case where the stepfather and stepson have the same nomen, cf. 10.2862, where the privignus of Sextus Pompeius Demetrius is called Sextus Pompeius Mercurius.

⁵⁴ cf. 6.23342: Octavia Faustina's filiaster is Sextus Cornelius Verus; also 6.7527 (Antonia Fortunata's filiaster is M. Clodius Satyrus). Relevant here is also the argument I used previously, that a mother would not refer to her (illegitimate) son as filiaster.

⁵⁵ cf. 17243, 22308, 25726, 28519, Inschr. Christ. Diehl 2714 – in all these cases an incertus or slave dedicates a tomb to his filiaster (or filiastra) who bears a slave name.

⁵⁶ op. cit. (n. 9), p. 139 n. 1.

⁵⁷ As slave wives of imperial slaves were normally also from the imperial house (Weaver, , op. cit. [n. 9], p. 138)Google Scholar, Successa must belong to the familia Caesaris, and if Januarius is her son, he would belong to it as well. But it was rare for imperial slaves to be manumitted before the minimum age of 30 laid down by the Lex Aelia Sentia (see Weaver, , op. cit. [n. 9], pp. 97ffGoogle Scholar., Rawson, , op. cit. [n. 9], 285 n. 20Google Scholar; Rawson, , art. cit. [n. 26], 78fGoogle Scholar. has cases of early manumission but these are outside the familia Caesaris).

⁵⁸ For cases of broken families see Rawson, ibid.

⁵⁹ Weaver, (op. cit. [n, 9] [1972], p. 139 n. 1)Google Scholar suggests this solution in the case of Venuleius Proculus, son of Valentina (= Venuleia Valentina?) and the slave Proculus (3.4065).

⁶⁰ Contrast the example in the previous note, where the illegitimate child has its father's name as a cognomen.

⁶¹ It is theoretically possible that Ulpia Felicissima is freeborn and illegitimate (i.e. born after her mother had won her freedom but while the father was still a slave), or even slave-born and freed (but cf. note 57 above). In both these cases, however, we would need to explain why an illegitimate daughter is called filia but an illegitimate son filiaster.

⁶² In connection with (2) and (3), the anonymous referee points out that even the upper classes did not always stick rigidly to standard practice when naming their children, and that the lower classes who appear on most of the inscriptions discussed here may have been less likely to conform to the ‘rules’. In other words, arguments based on nomenclature, though strongly suggestive, cannot be entirely conclusive.

⁶³ I have found 25 inscriptions to or by a privignus (-a).

⁶⁴ Ti. Claudius Onesmus [sic] is the vitricus, Ti. Claudius Phoebus (aged 38), the filius. Here, since the mother's name is Claudia Cara, the similarity between the names of the two men is explained if we assume either that all three are ex-slaves of the same patron Ti. Claudius, or else that Claudia's second marriage was to her patron, who also manumitted her son, Phoebus, born of contubernium. Other cases are 10.2862 (Sextus Pompeius Demetrius sets up a tomb for his privignus, Sextus Pompeius Mercurius: the latter could be either adopted, or the libertus of his vitricus) and 6.38831 (see above, note 42).

⁶⁵ See Seek, F. and von Carolsfeld, H. Schnorr, ‘Das lateinische Suffix aster, astra, astrum’, Archiv f. lat. lex. 1 (1884), 390–407Google Scholar; Cooper, F., Word Formation in the Roman Sermo Plebeius (New York, 1895), pp. 192ffGoogle Scholar.; Leumann, M., Lateinische Grammatik (München, 1926–1928), i.220Google Scholar; Thomas, F., ‘Le Suffixe latin “-aster/-astrum”’, Revue des Etudes Anciennes 42 (1940), 520–8CrossRefGoogle Scholar; cf. Hiltbrunner, O., Gymnasium 74 (1967), 304fGoogle Scholar.

⁶⁶ Cic. Att. 12.44.3. Shackleton-Bailey ad loc. reads Favoniaster, meaning ‘a bad copy of Favonius’, with reference to a man of pretensions who tries to ape Favonius, the type of republican zealot.

⁶⁷ cf. Seek, , op. cit. (n. 65), 391Google Scholar; Cooper, , op. cit. (n. 65), p. 193Google Scholar. For the contrary view, see Thomas, , op. cit. (n. 65), 522Google Scholar (‘Ils ne comportent aucune nuance péjorative’).

⁶⁸ As well as filiaster, the terms patraster and matrastra are also occasionally found. Meyer, (op. cit. [n. 6], p. 48)Google Scholar includes one example of the former in his list of cases showing an unspecified extra-marital relationship: ‘D. M. Titinius Claeus Felix bixit ann. XVI. Pelacisanus(?) Daximia mater et Ammonius Sabinus patraster filio b.m.’ (10.3013). But the nomenclature offers no indication that Sabinus is anything other than stepfather. The same applies to the other examples of patraster that I have collected (Inschr. Christ. 2714; 6.14105, 13.1829 and 6.11753–4: in the last case [an epitaph by a patraster] the dedicatee is specified as the son of another man; Meyer [loc. cit.] takes patraster here as meaning fosterfather, as he does also at 10.6983 where the child [a slave] is referred to as alumnus).

⁶⁹ For matrastra, see 11.6730,4 (an inscription on an opus artis musivae beneath a picture of Hercules reads ‘Hic est Hirculis qui/ amatrastra sua perivit’); cf. Meyer-Lübke, op. cit. (n. 4), s.v.

⁷⁰ cf. Thomas, , op. cit. (n. 65), 520–2Google Scholar.

⁷¹ Further confirmation that the term was a loose one is provided by the fact that the term filiastra in later Latin was used also of a granddaughter (Gloss. II 329.41, III 254,14).