Bryn Mawr Classical Review 2003.01.03
Heikki Solin, Die stadtrömischen Sklavennamen. Ein Namenbuch I-III. Forschungen zur antiken Sklaverei, Beiheft 2. Stuttgart: Franz Steiner Verlag, 1996. Pp. xxiv + xvi + xvi + 727. ISBN 3-515-07002-8. EUR 100.00.
Reviewed by John Bodel, Rutgers University (firstname.lastname@example.org)
Word count: 2870 words
Slave names are inherently interesting. Conceptually, they fall between the names given to acknowledged children and those assigned to pets. Historically, they have served to distinguish chattel from persons who belong to recognized kin-groups, to separate property from progeny, and thus to support distinctions of status based on birth. In ancient Roman society, where the civil law distinguished the free from the unfree with exceptional clarity and where the onomastic system marked kinship with unusual precision, we might expect the ambivalence of slaves as human property -- res mortales, Ulpian calls them at one point (Dig. 126.96.36.199) -- to be reflected in the names they carried. Was it? If so, how? What do Roman slave names tell us about how the Romans regarded their slave familiae, and the freedmen who emerged from them? This book does not answer these questions, but it provides an invaluable tool for those who wish to explore them.
In the last half century no group of scholars has done more to advance our knowledge of Roman onomastics than the philologists and ancient historians trained at the University of Helsinki in Finland and the Institutum Romanum Finlandiae in Rome; and of these none has contributed more to our understanding of the Romans' use of Greek and foreign names than Heikki Solin.1 A preliminary study, Beiträge zur Kenntnis der griechischen Personennamen in Rom I (diss. Helsinki 1971), preceded by a decade the full publication, as an auctarium of the Corpus Inscriptionum Latinarum, of an onomasticon comprising the massive data on which it was based, Die griechischen Personennamen in Rom. Ein Namenbuch, 3 volumes (Berlin 1982; a revised and updated edition is in the course of publication). The former expounded many of the analytical principles upon which the division of material in the latter was based. Now the situation is reversed. With the work under review, we must await publication of a promised historical and philological study (vol. 1, p. xxii) to learn why some Roman slave names are registered here where they are and why others are excluded. In the meantime, we have a wealth of material to consider, all presented with the high standards of accuracy and thoroughness we have come to expect of the author.2 The onomasticon registers more than 28,000 attestations of some 5,500 names of slaves and ex-slaves (that is, the cognomina, including second cognomina, of freedmen, but excluding agnomina, supernomina, and signa), recorded from the earliest period down to the fifth century CE. The data are distributed into four sections: Latin names (volume 1: 8,579 examples, c. 32% of the total), Greek names (volume 2: 18,424 examples, c. 67%), "barbarian" names (506 examples, c. 1.8%), and a sizeable miscellaneous group (628 examples) comprising names of uncertain origin, corruptly transmitted names, names of freedmen without cognomina, and fragmentary examples (volume 3). The third volume concludes with a reverse index of nominative forms (with names attested only in Greek transliterated into Latin script) and an index of lemmata and selected irregular name forms.
The Latin names are divided into twenty categories according to their derivation from Roman praenomina; gentilicia; historical persons; mythology and the world of the gods; geography; the body and corporeal attributes; personality traits (the intellect, moral qualities, temperament, and character); personal circumstances ("wish" names and personal qualities); circumstances of birth; gender and familial and social relations; place of origin; titles, occupations, and positions in life; the calendar; animals; plants; natural elements; foodstuffs, clothing, tools, and man's products and affairs (e.g. "Calamus", "Sagitta"); collective groups (e.g. "Familia"); abstract concepts; and verb forms and verbal substantives. A final category registers fragmentary examples.
Following an alphabetical list of "Vollnamen" (187-238), the Greek names are distributed into twenty-three categories, which are for the most part identical with those for Latin names but which naturally omit slave names derived from Roman praenomina and gentilicia and which add entries for literary names, poetic appellatives, short names derived from the first elements of compound words (e.g., "Dama" from demo-, "Cerdo" from kerdi-, etc.), and a miscellaneous group consisting mainly of names of ambiguous or uncertain derivation. The "barbarian" names are arranged alphabetically within categories defined according to linguistic or geographical criteria into ten groups: Semitic, Iranian, Anatolian ("Kleinasiatische"), Thracian, Egyptian, African, Illyrian, Germanic, Celtic, and Iberian.
Beneath each lemma the individual attestations are distributed chronologically and are arranged alphabetically within groups broadly datable to periods of 50 or 100 years or to the reigns of individual emperors or groups of emperors, except in the case of names attested fewer than five times, for which dates are given in parentheses after each citation. In contrast to earlier onomastica, such as those of Kajanto and Bechtel,3 the names are usefully given in full in the grammatical cases in which they are attested and include any information provided in the text about the origin, ownership, or position of the person named (e.g., "Agathoni lecticar(io) Tauri", "Militis nat(ione) Marsac(i) = delicium nostr[um] Militem"). For evidence of the status of the person named or an indication of date one can in many cases resort to the full publication of the text in question but must at times rely upon the judgment of the author, which, though expert, is not, of course, infallible.4 Comparison of the entries for Greek slave names with Solin's earlier publication of Greek cognomina in Rome (1982) reveals numerous adjustments in the assignation of dates, owing primarily, it seems, to the greater attention now accorded to the dating of the monuments on which the names are recorded (the "text-carriers") and the elimination from the present volumes of the categories of "probable freedmen" and "sons or daughters of freedmen" included in the earlier work.5 In general, the new dates give the impression of greater accuracy, but, as always in the dating of inscriptions not securely dated by internal criteria, one must be wary of the fallacy that greater precision implies greater reliability.
More problematic are the criteria adopted for the inclusion of names. A number of ambiguities and gray areas, many acknowledged by the author, are inevitable. The distinction between second cognomina (included) and supernomina (excluded) is notoriously hazy, particularly in the period from the late second century CE on. Names attested at Rome need not have belonged to persons actually resident in Rome. Of the names recorded in stamps on certain types of instrumenta domestica, notably Arretine ware and the broken amphorae piled up at Monte Testaccio (both categories registered in the lists), most were borne by persons who almost certainly lived and worked elsewhere (vol. 1, p. xxi). More difficult to discern are the procedures followed with the slave names attested by ancient authors. Those belonging to historical persons mentioned in nonfictional sources (and therefore usually registered in RE) are regularly included, but at times the line between names "from Rome" ("aus der Stadt Rom", vol. 1, p. xxi) and those from elsewhere seems to be very finely drawn. Of two onomastically interesting freedmen of Cicero's friend Atticus, for example, M. Pomponius Dionysius finds a place in the lists (277), but T. Caecilius Eutychides does not (436), presumably because the former is known to have been active at Tusculum and Rome (Cic. Att. 4.8a.1, 4.11.2, 4.13.1, 4.19.2, etc.), whereas the latter is certainly attested in Cicero's company only in Epirus (Att. 5.9.1). This sort of onomastic gerrymandering is surely defensible on a narrow view of locality but often excludes, as here, evidence of obvious relevance to a full consideration of the phenomenon of slave names in use at Rome. Cicero, whose influence seems to have been instrumental in gaining Eutychides his manumission (and with it his unusual combination of praenomen and nomen: Att. 4.15.1), was using the name at Rome more than half a century before it is first epigraphically attested there (436).
In fact, in the case of literary (as opposed to inscriptional) sources, what the onomasticon seems to register are not names but persons -- and not all persons, but only those presumed to be real. With the names found in fictional or semi-fictional sources, both the rationale and the criteria for distinction are unclear. Ovid's Cypassis (Am. 2.7.17, 2.8.2), Horace's Davus (Serm. 2.7), and Persius's Dama (5.76) -- all unquestionably slaves and all represented as active at Rome -- do not find a place in the register, whereas both members of Martial's suspiciously congruent pair, Polyphemus and Scylla (7.38), receive listings (343, 357). As Solin well knows, thematically significant groupings of mythological names within actual slave familiae are not unknown (Namenpaare 25), but whether in this case Polyphemus and Scylla are any more real than Davus and Dama and Cypassis is at least debatable.6 The same can be said of Juvenal's praeco Machaera(s) (7.9), whose claim to servile or freed status is far from certain (539).
When we turn to the category of "literary names" (in Greek only, p. 360), we find only a single representative, Encolpius, the narrator of Petronius's Satyrica, whose own juridical status in the novel is ambiguous. Here a strict application of the criterion of certain (or near certain) servile or freed status has allowed a few interesting sprats to get away. A wider cast of the net in Solin's earlier Namenbuch pulls in (in addition to several more examples of names derived from "Encolpius") three interesting incerti: M. P( ) Ascyltus (whose cognomen duplicates that of Encolpius's erstwhile companion in the Satyrica), Bellia Neobul[e] (cf. Hor. Carm. 3.12.5), and Annaea Cypasis (after Ovid's hairdresser; Die griechischen Personennamen 564-65). Anyone considering the use of slave-names such as Encolpius or Cypassis at Rome will want to take account of related instances such as these in order to recognize the onomastic background against which such choices were made.7 At the same time it would have been useful and appropriate, in registering the slave names (as opposed to the persons with servile names) attested at Rome, to include all those represented in literary sources as occuring at Rome, if only to illustrate the creativity of Roman authors in assigning to their slave characters names that were common and perhaps typical (Dama, Davus), plausible but rare or unattested (Cypassis, Trimalchio), or patently invented for comic effect (e.g., Crurifragius, Plaut. Poen. 886).8
In fact, the invention of slaves' names was not restricted to the world of literary fiction. The peculiarly ambivalent status of slaves as "talking tools" -- instrumenti genus vocale, in Varro's memorable phrase (RR 1.17.1) -- rendered their names unusually fluid. According to Varro, slaves sold at the market at Ephesos were often renamed by a trader or buyer after the seller, or the region in which they were purchased, or the city where they were bought (Varro Ling. 8.9.21). More generally, slaves might undergo a change of name during the period of servitude, when transferred to a new owner or at the whim of a particular master: one contract for a slave sale from Side in Pamphilia identifies the merchandise, a ten-year-old Galatian, as "the slave girl Abaskantis, or by whatever other name she may be known" (P. Turner 22). We may suppose that most slaves eventually acquired a single name that marked their identity, and that this name is the one recorded in epitaphs, but the inscriptional record gives little indication of the undercurrent of mutability that only occasionally breaks the surface. After manumission, when they gained control of their identities, some ex-slaves (we do not know how many) exchanged their servile names for more respectable Latin cognomina. A famous example is the Tarentine grammarian L. Crassicius, who, upon winning his freedom sometime toward the end of the Republic, renamed himself (se transnominavit) from Pasicles (probably a personal name given at birth and retained when he was enslaved) to the more ennobling "Pansa" (Suet. Gramm. 18.1).9
For various reasons, then, the aggregated group of slave names known at Rome is not equivalent to the group of named slaves known at Rome, even when the latter are meticulously identified and categorized and counted, as they are here. The point seems obvious but bears emphasizing, since the impressive lists assembled by Solin might seduce the unwary into trying to extrapolate from them arguments that they cannot support. Even the very reasonable categorization of the Greek and Latin names according to philological criteria (useful for illustrating the semantic range of the material) may give a misleading impression of the distribution and popularity of the various types, since the categories vary greatly in size and chronological range, running from a single page with three lemmata and half a dozen citations, all probably from the first century CE (Latin names from collective groups, 168), to more than ninety pages with hundreds of lemmata and thousands of examples spanning a period from the early first century BCE to the third century CE (Greek names from mythology and the world of the gods, 265-359).
More fundamentally, we must accept that no lists of attested names can provide us with a representative, yet alone a complete, picture of Roman customs in naming slaves. Since inscriptions provide by far the most abundant source of information for onomastic practices, their limitations as sources need to be recognized. Not all slaves buried at Rome received inscribed markers recording their names; many (we cannot know how many, but the cemeteries at Isola Sacra near Ostia and under the basilica of St. Peter at Rome suggest that they may have been numerous) were consigned to anonymous graves.10 Inscriptions further underrepresent, considerably if immeasurably, the realities of rural slavery: what names the Romans gave to the slaves who worked the fields and cared for their rural estates -- a large majority, on any reckoning -- we can only begin to guess.11
The nature of this epigraphic bias is such that for many purposes we can only talk meaningfully about a subset of Roman slave names, those recorded in inscribed epitaphs. This group already represents a privileged segment of the slave population simply by virtue of its being commemorated in the epigraphic record. The record itself, of course, is incomplete. Not all inscribed tombstones are known, and the surviving sample is skewed by the hazards of preservation, discovery, and publication. For identifying slave names, the evidence is further restricted to the periods when the conventions of epitaphic commemoration favored onomastic formulae and groupings that allow the status of the persons named to be identified. These limitations are well-known to Solin, who does not claim more of the evidence than it warrants, and no doubt to most of those who will use these volumes, but the casual visitor to the serried columns of Roman name-lists will do well to keep in mind how much they do not reveal.
For all that, the data compiled by Solin provide invaluable material for exploring Roman practices in naming the slaves who ultimately earned the distinction of individual commemoration by name. Observations based upon a high number of attestations and properly couched in terms of commemorative behavior can reveal important truths about the social realities they reflect. It is worth considering, for example, in light of the common association in Roman contexts of servile origin with a Greek cognomen, that six of the ten most frequently attested slave-names at Rome (counting both the male and the female forms of the name) are Latin.12 That all six names also rank among the eighteen most common Latin cognomina in general use (those attested more than 1,000 times) suggests that in everyday address many Roman slaves and ex-slaves may have been onomastically indistinguishable from their free-born peers.13 Regarding Greek slave-names, Solin has recently compared the most popular types at Rome with those most frequently attested for slaves at Athens and generally in the Greek world to show that the Romans did not follow the fashion at Athens but instead drew upon their knowledge of Greek names from various sources (notably the trading centers of Delphi and Delos) to establish their own patterns of usage, avoiding the geographic slave-names popular in Attica and favoring names of historical figures such as "Antiochus" and "Alexander", which were seldom borne by slaves in the Greek world, and especially two "divine" cognomina, "Hermes" and "Eros" rarely found at all before the Roman period.14
Observations such as these enable us to begin to outline the contours of Roman practices in naming slaves and thus better to understand their place in Roman society. What considerations led the Romans to prefer Latin slave-names that facilitated onomastic assimilation into the urban population and Greek names that distinguished Roman slaves from Greek foreigners in Rome would be worth knowing, as would the motives that governed the choice of a Latin or a Greek name for any particular slave. Thanks to painstaking compilations such as this, we can begin to explore questions of broader historical significance such as these. In the future such investigations will be more easily pursued through electronic databases, which have the double advantage of permitting regular updating and of allowing the multiple tagging of names semantically relevant to more than one category (e.g. "Philargyrus", "Salvius") and thus of enabling individually defined searches across categories. For now, in providing a reliable, thorough, and reasoned catalogue of the material in traditional form, Solin has once again placed all students of Roman society in his debt.
1. In addition to the publications of Solin cited below, attention may be called to the fundamental works of Iiro Kajanto, Onomastic Studies in the Early Christian Inscriptions of Rome and Carthage (Helsinki 1963), The Latin Cognomina (Helsinki 1965), and Supernomina. A Study in Latin Epigraphy (Helsinki 1966); Olli Salomies, Die römischen Vornamen. Studien zur römischen Namengebung (Helsinki 1987), Adoptive and Polyonymous Nomenclature in the Roman Empire (Helsinki 1992), and (with Solin) Repertorium nominum gentilium et cognominum Latinorum (Hildesheim/Zürich/New York 1988; reprinted with addenda, 1994); Mika Kajava, Roman Female Praenomina. Studies in the Nomenclature of Roman Women (Rome 1994); and to Solin's engaging study, Namenpaare. Eine Studie zur römischen Namengebung (Helsinki 1990). Solin's numerous shorter contributions to Latin epigraphy and onomastics published in the journal Arctos between 1970 and 1997 are assembled (with excellent indexes and a bibliography of Solin's writings up to 1998) in a useful volume edited by Kajava, Analecta Epigraphica 1970-1997 (Rome 1998) (reviewed in BMCR 00.02.10 by Dennis Trout); and the series continues (recently in Arctos 34  149-92). In coming years we can look forward to a full-scale treatment of the nomenclature of upper-class Roman women from Kajava, a study of pejorative cognomina among the Roman aristocracy from Christer Bruun (see already Classica et Mediaevalia 49 (1998) 95-117 and Archiv für Kulturgeschichte 81 (1999) 259-82), and, from Solin, a general introduction to ancient onomastics, a monograph on the name-giving practices of the Roman aristocracy, an investigation of the history of the name "Petrus" in antiquity, and a revised edition of the Namenbuch of Greek cognomina in Rome.
2. Omissions (e.g., p. 171 s.v. "Spes, Sexus unbekannt", add "Spei 33703 (1./2. Jh.)") and duplications (e.g., p. 547 s.v. "Hymnis": CIL VI 38463 and HSCPh 20 (1909) 7 are two editions of the same text) are rare.
3. F. Bechtel, Die historischen Personennamen des Griechischen bis zur Kaiserzeit (Halle 1917, repr. Hildesheim 1982).
4. Solin outlines some general principles and criteria for the assignation of dates in Beiträge zur Kenntnis der griechischen Personennamen in Rom I, 35-38 and for the determination of slave or freed status in QUCC 18 (1974) 106-24.
5. Further details about the principles of organization and arrangement of the material in the onomasticon are clearly set out in the introduction to volume 1 (pp. xxi-xxiv).
6. A. Herrmann, "Sperlonga notes," AAAH 4 (1969) 27-32 plausibly argues that Martial here refers to the sculptural display at Tiberius's grotto at Sperlonga.
7. See the suggestive remarks of S. Priuli, Ascyltus. Note di onomastica petroniana (Collection Latomus 140) (Brussels 1975) 63-66.
8. W. Fitzgerald, Slavery and the Roman Literary Imagination (Cambridge 2000) provides a helpful introduction to the representations of slavery in Roman literature.
9. Registered in the volumes under review at 225 (s.v. "Pasicles") and 57 (s.v. "Pansa"); curiously (and perhaps misleadingly, since the former preceded the latter), the first is dated "Caesar", the second "spaetrepublikanisch". For Crassicius's "transnomination" and other similar cases at Rome, see T. P. Wiseman, "Who Was Crassicius Pansa?" TAPA 115 (1985) 189 (reprinted in id. Historiography and Imagination [Exeter 1994] 92). At Capua during the late Republic and early Empire some freedmen boasted two cognomina, one Greek , the other Latin: See M. Frederiksen, Campania, N. Purcell, ed., (Rome 1984) 301.
10. See W. Eck, "Aussagefähigkeit epigraphischer Statistik und die Bestattung von Sklaven im kaiserzeitlichen Rom," in P. Kneissl and V. Losemann, eds., Alte Geschichte und Wissenschaftsgeschichte. Festschrift für Karl Christ zum 65. Geburtstag (Darmstadt 1988) 130-39.
11. See R. Samson, "Rural Slavery, Inscriptions, Archaeology and Marx," Historia 38 (1989) 99-110.
12. In the "Top Ten" rankings Solin gives on p. 680, in which men's and women's names are recorded separately, Latin names occupy five places in the list. In my own list, which is based on Solin's revised figures in the article cited below, n. 14, and which combines men's and women's names, "Prima" and "Primus" (independent entries in Solin's rankings) reduce to a single entry and "Alexander" is displaced in favor of "Fortunatus/-a" and "Secunda/-us" at the end of the list. The figures are as follows: Felix/-ic(u)la (461 + 69 = 530); Hermes/-ia (392 + 37 = 429); Eros/-otis (357 + 60 = 417); Prima/-us (213 + 184 = 397); Hilarus/-a (248 + 143 = 391); Faustus/-a (184 + 117= 301); Onesimus/-e (242); Antiochus/-is (94 + 37) = 231; Fortunatus/-a (138 + 87 = 225); Secunda/-us (118 + 101 = 219). For the commonest Latin cognomina, see Kajanto 1965 (above n. 1) 29f.
13. Not surprisingly, the three most common Greek slave names at Rome (Hermes, Eros, and Onesimus) are also the three most frequently attested Greek cognomina there (Solin, below n. 14, 312). If Latin literary sources provide an accurate guide, by far the most common form of address in Rome was by a single name, three out of four times a cognomen: see E. Dickey, Latin Forms of Address from Plautus to Apuleius (New York and Oxford 2002) 50, 56.
14. H. Solin, "Griechische und römische Sklavennamen. Eine vergleichende Untersuchung," in H. Bellen and H. Heinen, eds., Fünfzig Jahre Forschungen zur antiken Sklaverei an der Mainzer Akademie, 1950-2000. Miscellanea zum Jubilaeum (Stuttgart 2001) 307-30.