Skip to main content

The Naked Reader: Child Enslavement in the Villa of the Mysteries Fresco


This article analyzes the naked boy who appears as a reader in the fresco cycle of the Villa of the Mysteries in Pompeii, ca. 60–40 BCE. Although this fresco and its many figures have received ample attention, few scholars have asked why the reading boy is naked. I mark the boy’s nakedness as proof of his enslavement, using iconographic and epigraphic evidence for child slaves in Roman-era Dionysiac cult. I also consider a Roman audience’s reception of this boy as decoration for the walls of a Late Republican villa. This image, I argue, worked to reinforce social hierarchies and the eroticization of child slaves, thereby perpetuating cultural systems of subjugation that organize the domestic sphere and the empire more broadly. By way of conclusion, however, I mark the painting’s unabashed acknowledgment of the lived experiences of a child slave as a subtle critique of slaveholding strategies, at least among enslaved viewers.1

Content warning: Readers are advised that this article discusses evidence for the sexualization and sexual exploitation of minors, especially enslaved minors.


The namesake fresco of the Villa of the Mysteries, painted ca. 60–40 BCE, is a veritable icon of Roman art, so much so that this Second Style painting—a continuous frieze stretched across the walls of a 5 × 7 m room in a suburban villa outside Pompeii—makes frequent appearances in, but also on the covers of, texts devoted to subjects as varied as the city of Pompeii, Roman painting, mystery cult and religion, and the Latin language.2 The ubiquity of this well-preserved and richly pigmented painting on book jackets, however, is hardly difficult to justify: the fresco cycle depicts no fewer than 29 megalographic figures: males and females, gods and mortals (fig. 1). Most are associated with Dionysus’ thiasos, and together they offer a tantalizing glimpse into ritual activities associated with the god’s mysteries.

Fig. 1. 
Fig. 1. 

Villa of the Mysteries frieze (looking east from the main entrance), Second Style fresco, mid first century BCE. Pompeii, Villa of the Mysteries (courtesy Parco Archeologico di Pompei).

On the east wall, opposite the main entrance to the room, for example, a kneeling woman lifts a veil from a possible phallus in a liknon (a winnowing basket). Above her, a winged female with exposed breasts raises a whip and directs it across the southeastern corner of the room, breaking the boundaries of two-dimensional painting. The object of this apparent flagellation is a woman whose back is exposed, her drapery falling around her; she cowers in another female’s lap on the south wall. Near this couple is yet another female figure who dances in the nude, raising crotala (cymbals or castanets) above her head and exposing her back, buttocks, and legs to the viewer. And on the long north wall near a small, secondary entrance to the hall (fig. 2), the viewer encounters a naked child of perhaps 6 to 10 years of age, the only nonmythological male figure in the fresco. This boy holds a scroll and appears to read to the mortal women around him, who are readying themselves for cultic practice (fig. 3).

Fig. 2. 
Fig. 2. 

North wall of the Villa of the Mysteries frieze. From left to right figures include: the veiled woman; the reading group; the serving girl; mortal women preparing for ritual; Silenus and satyrs; and the terrified woman (courtesy Parco Archeologico di Pompei).

Fig. 3. 
Fig. 3. 

Naked reading boy flanked by two women, detail of the north wall of the Villa of the Mysteries frieze (L. Horan).

The reader’s hair is centrally parted, with curly locks resting atop the head and along the part. He is naked save for a pair of cothurni (boot-like sandals) on his feet. These boots are traditionally worn by Dionysus’ followers, but here they draw attention to the boy’s nakedness; the absence of clothing accentuates his slim, effeminate, prepubescent figure and small penis.3 The boy stands with his legs pressed together, his right knee slightly bent, and his left hip raised. His shoulders are hunched and his bent arms, which clutch a scroll, are tight against his torso. His eyes stare down, anxiously but intently, at the scroll he holds up to his face. A seated woman behind the boy rests her right hand on his shoulder and directs a stylus toward the scroll he holds; another scroll rests on her lap in her left hand. Interestingly, this woman does not engage with the boy directly; rather she looks over his head to her right in the general direction of a veiled woman who enters what appears to be a domestic interior.

Since the discovery of this fresco in 1909,4 most scholars have identified the boy as a reader of sacred rituals or a child priest.5 In these syntheses, however, the boy’s nakedness is seldom remarked on, his nude physique tacitly accepted as a mark of his ritual significance. Recent work on this fresco, on the other hand, has been silent about the very inclusion of this boy and his function in any larger narrative. Some of this is likely due to the boy’s perceived insignificance as a child, together with changing attitudes about the fresco painting as a scene of ritual celebration versus an initiation; and yet modern scholarship’s reluctance to engage in detailed discussions of this boy may also underscore (conscious or subconscious) anxieties about the prepubescent nudity of the only mortal male in the fresco cycle.6

My work aims to confront the many unknowns that surround this boy by synthesizing the layers of meaning attached to his naked body, and the myriad identities that such a body suggested to Roman viewers. For although nudity is tacitly presented in modern scholarship as a condition of Graeco-Roman art and its idealization of the human figure, nakedness had variable meanings to a Roman audience.7 The reception of a naked body in art is, moreover, heavily dependent on context and situation, especially when that nakedness manifests on the body of a male child in a fresco painted for the private enjoyment of elite residents and guests in the probable triclinium of a Late Republican villa on the Bay of Naples.8

After a brief survey of previous interpretations of the Villa of the Mysteries fresco cycle, I use iconographic and epigraphic evidence to identify the boy as an enslaved reader. I review documentary evidence for the mysteries in the Early Empire, from which it is clear that enslaved children figure prominently in the cult. Pictorial allusions to the mysteries that include naked boys, I further argue, visualize the intersectional identity of these individuals as enslaved minors whose labors have ritual import.9 The significance of the work of these child slaves, however, fails to elevate their person, and previous arguments about their nudity as ritually sanctioned (versus culturally conditioned) are called into question.

I then consider the decorative function of this child slave, weaving in literary evidence for enslaved pueri (boys) in the Late Republic and Empire, and for deliciae (pet slaves) in particular. The depiction of a naked boy in a quasi-ritual scene on the frescoed walls of a private triclinium, I argue, worked on multiple levels to normalize and aggrandize the eroticization of child slaves, thereby perpetuating the cultural systems of subjugation that organize both the Roman house and the empire writ large. By way of conclusion, I present an extended discussion of the boy’s defensive body language and the varied reactions that this subtle display of resistance would have solicited from different audiences—the elite for whom the fresco was commissioned as much as the nonelite populations who worked alongside this image.

Previous Interpretations of the Villa of the Mysteries Fresco

The naked reader has received minimal attention in scholarship on the Villa of the Mysteries fresco. In general, scholars have focused on other figures in the painting or the narrative as a whole. In what follows, I therefore briefly survey previous interpretations of the fresco—its meaning and significance—to situate this study of the reading boy in a larger historiographic context. In doing so, however, I depart from earlier work by treating the boy in question as a heuristic. For indeed, his person gestures toward predominant themes in this fresco cycle: the cult of Dionysus, the life course of an elite Roman woman, and the convergence of the secular and the sacred in Roman culture.

The fresco’s concern with the experiences of an initiate in the Dionysiac mysteries has long dominated scholarship. Beginning in the 1910s, shortly after the discovery of this painting, many scholars interpreted it as a depiction of initiation in the god’s cult, with some granting the room itself the status of an initiation chamber.10 These hypotheses emphasized the central place of Dionysus and Ariadne on the east wall opposite the main entrance,11 and the inclusion of individuals whose actions appear ritually motivated, like the naked reader.12 The continuous frieze, notably, intimates a connection between various mythical members of Dionysus’ circle (satyrs, fauns, and maenads) who dominate the eastern half of the room, and mortals in the western half of the hall who identify themselves as followers of the god by their dress. For example, many of the women on the north wall—a serving woman with a tray and female attendants who assist a seated woman with the raising of a cloth and the washing of her hands—wear wreaths of olive leaves in their hair.

The boy’s placement on the north wall among mortal women readying themselves for Dionysiac ritual underscores another important theme in this fresco: the female experience. The only mortal male on these walls is a child who is physically dwarfed by the adult women around him. Indeed, the preponderance of women in this fresco has garnered the attention of (largely female) scholars since the early 20th century.13 Perhaps the most important analysis emerged in the work of Bieber and Toynbee in the late 1920s. Independently, both scholars advanced interpretations of the painting as a bride-to-be’s prenuptial initiation in the cult of Dionysus.14 Both Bieber and Toynbee based their arguments on the inclusion of a seated female figure in the southwestern corner of the hall whose hair is being styled by an attendant in the fashion of a Roman bride, while two Erotes look on (fig. 4).15

Fig. 4. 
Fig. 4. 

The bride and an Amor, detail of the west end of the south wall of the Villa of the Mysteries frieze (courtesy Parco Archeologico di Pompei).

Over the last 30 years, however, scholars have largely abandoned theories of the painting as a veritable initiation scene, whether of a bride or not. For although the fresco contains oblique references to ritual practices of various kinds (reading, whipping, the uncovering of the phallus, specialized hairstyling), nothing can be definitively marked as a revelation granted only to those initiated in the god’s cult. Secret rituals are not depicted in this fresco, and its tone is referential, as opposed to overtly ritualistic.16 Contemporary analyses highlight the spectacular and allusive character of the painting: its ability both to delight viewing audiences with dramatic (if unsurprising) references to the world of Dionysus and to celebrate the participation of his female followers in that same world.17 This is not to say that these painted figures and their actions failed to reward an initiated audience of viewers; the images may well have served as visual aids for recollection of personal experiences with the god and his cult. And yet the painting is also perfectly comprehensible to an uninitiated audience familiar with first century Graeco-Roman culture, judging from extant comparanda for many of these iconographies in the Bay of Naples region broadly.18

Thus, with its myriad references to Dionysus and mystery cult, to marriage and female rites of passage, we must accept the meaning of the Villa of the Mysteries fresco cycle as multivalent. The painting actively encourages its viewers to read it on many levels, layering and harmonizing divergent interpretations of its pictorial narrative as much as of individual figures within it. The rendering of the reading boy and his unclothed body in particular, I argue in the sections that follow, makes this aim patently clear.

Undressed Children in Roman Art

In the first publication of this fresco cycle, a brief archaeological note from 1910, Giulio de Petra identified the naked boy as a participant in a reading lesson with his mother (the woman behind him).19 This was quickly dismissed by successive scholars, but evidence discouraging the identification of a mother and her son merit brief consideration here; they highlight important distinctions across depictions of free or freed versus enslaved children in Roman art, and support the identification of the boy in the Villa of the Mysteries fresco as a child slave.

Although de Petra’s identification of the scene as a reading lesson would appear to explain the scroll in the child’s hand, neither the boy nor the women around him find parallels in extant depictions of paideia (classical education). Most of these come from Imperial-era funerary monuments, which depict clothed students and male teachers almost without exception.20 Furthermore, while the contemporary dress of the seated woman suits an elite female,21 the nudity of the 6- to 10-year-old boy who reads with or to her is inappropriate for a freeborn child; his naked body rather intimates enslavement.

In extant arts of the Late Republic and Early Empire, free or freed children wear clothing. Male children over the age of 3 or 4 years (i.e., beyond infancy)22 appear in the toga with the bulla around their necks in state and private funerary monuments alike (figs. 5, 6).23 This particular costume is visual metonymy for citizenship and social importance. As Hallett has demonstrated in his study of nudity in the arts of the Late Republic and Early Empire, gods and elite males were depicted nude in public art on occasion,24 but nakedness was more commonly employed to express the inferiority of an enslaved person, whether as captured booty in a triumph or as an object for sale at a slave market.25 Thus, for a Roman audience, although the unclothed body could mark a person as special (nude in a divine or heroic sense),26 nakedness was more frequently used to convey a mortal individual’s subjugation to viewers (whether as a slave, a prisoner of war, or a “barbarian” other). One might argue that the Villa of the Mysteries boy adorns the walls of a private residence and that his nakedness would have been received somewhat differently in this context, far from the public eye. But the house fulfilled a social function in the Roman world, whereby important distinctions about social class and nakedness were never far from the minds of its inhabitants.27

Fig. 5. 
Fig. 5. 

Funerary relief of the Servilii family, 30–20 BCE, marble. Depicted are Publius Servilius Globulus, the freeborn son of Quintus Servilius; Quintus Servilius Hilarus, father (pater), libertus of Quintus Servilius; and Sempronia Eune, wife (uxor), liberta of Gaius Sempronius (CIL 6.26410). The boy wears a toga and a bulla that, together with the inscription, proudly assert his status as a freeborn Roman citizen. Vatican City, Musei Vaticani, Museo Gregoriano Profano 10491, acq. 1838 (© Governorate of the Vatican City State – Directorate of the Vatican Museums).

Fig. 6. 
Fig. 6. 

Ara Pacis Augustae, Rome, detail of children in the imperial family, showing the boys wearing the toga and the bulla. Originally erected in the Campus Martius, 13–9 BCE, marble. Rome, Ara Pacis Museum (© Sovrintendenza Capitolina – Foto in Comune).

Previous scholarship, however, has failed to recognize the boy’s naked body as a sign of his enslavement because of his apparent participation in the cult of Dionysus; the most popular interpretation of the boy is as a sacred reader of some kind. As early as 1913, Mudie Cooke identified the boy as a child priest based on inscriptional evidence for such persons in Italy.28 Bieber and Toynbee took a slightly different approach several years later because of their interpretation of the mysteries as a prenuptial ritual; both marked the boy as a pais amphithales (πα̑ις ἀµφιθαλής), a child with both parents living who served a ceremonial role in marriage as a propitious omen for the fertility of the couple.29

Broader movement away from the fresco as an actual initiation scene, however, has discouraged more recent scholars from granting the boy a specific function in any narrative tied to the mysteries; most describe him simply as a nude reader with a scroll.30 This is not incorrect, but such a terse description leaves a modern audience responsible for making sense of his person and his nakedness. Glossing over the latter as ritual attire is especially problematic since it obscures different connotations that the naked body had for children who participated in Dionysiac cult, depending on their freedom or enslavement.

Children in the Mysteries: The Epigraphic Record

It is somewhat surprising that neither the statuses nor the functions of children in Bacchic cult are discussed at length in scholarship, given the ample evidence for their participation in the mysteries in the extant material record of the Italian peninsula.31 Children are documented indirectly in the cult by their inclusion in (or association with) paintings, funerary reliefs, and sarcophagi devoted to Dionysus (late first century BCE–fourth century CE).32 They also appear as ritual participants in the epigraphic record from the first century CE on. Taken together, this evidence points to the initiation of children in Dionysiac cult at a young age,33 but it also suggests they contributed (actively or by association) to the practice and maintenance of cult. To bring greater attention to the nature of child participation in the cult, and to review documentary evidence for the social gradations of such participants, I turn to the epigraphic record. Most of the extant testimony dates to the Imperial period but may be used to nuance interpretations of children who appear in the extant visual record.

Broadly, as far as the social class of Bacchic initiates in the Roman era is concerned, the ca. 165 CE Torre Nova inscription from outside Rome is especially illuminating (IGUR 1 160). The inscription, which is written in Greek and decorates the base of a statue to a Dionysiac priestess named Pompeia Agrippinilla, names 420 initiates together with their functions in the cult.34 They are listed in hierarchical order of their importance within the cult and within Roman society. That is, Pompeia and mostly male members of her family (by birth and by marriage, respectively) appear near the top of the list as priests.35 Their transliterated Latin names are followed by the mostly Greek names of more than 300 persons. Interestingly, nearly all individuals named in this inscription are identified by a single name; John Scheid has argued convincingly that the absence of the nomina gentilicia (the family name) for more than 400 initiates associates all those listed with Pompeia’s familia (which includes kin relatives and enslaved persons).36 A few—those with Latin names whose gentes are known through prosopographic study—are freeborn persons with high social standing, but the vast majority of individuals named on this statue base, those with Greek names, are clients and dependents of Pompeia’s family: freedpersons and slaves.

Two children whose names appear on the base must be assigned to the latter category.37 Shortly after the priests, the hierophant, and several other important persons, the inscription records two male amphithaleis. The exact function of these children in the cult is unclear, but the inclusion of their names near the top of the inscription signals their importance.38 And yet, the Greek names of both amphithaleis, Latrios and Menandros, intimate their enslavement. The first name in particular suggests the current or formerly enslaved status of the boy, as Latrios means “servant” or “slavish.”39 Moreover, because Latrios is named before the second amphithales, who goes by Menandros (also a common slave name), and because the inscription lists individuals hierarchically, we may safely assume that Menandros is also currently or formerly enslaved. According to this inscription, two boys of lower social status fulfilled important functions in the cult; servile identity and ritual importance, we may conclude, are not mutually exclusive.40

Intersectional identity also characterizes a 7-year-old boy named Herophilos, who is remembered in a funerary inscription as a young initiate from Latium. This epitaph, loosely dated to the first or second century CE, commemorates the boy’s short life and his accomplishments in Greek hexameters (in a Homeric mode). According to the inscription, Herophilos accomplished (ἐξετέλουν) the rites of Dionysus by speaking (βάζων).41

Θ(εο̑ις) Κ(αταχθονίοις). | µήπω γευσάµενος ἥβης | ὤλισθον ἐς Ἅδου (Ἅιδου) | δάκρυα | καὶ στοναχὰς λείψας αἰ̑ωνι γο||νε̑νσιν | δύσµορος, οὐδ’ ἐνόησα | βροτ̑ων ψα̑υσαι βιότοιο | ἑπτὰ | µόνους λυκάβαντας δύω | καὶ µ̑ηνας ἔζησα | ̑ὠν τρε̑ις | ἐξετέλουν Διονύσῳ ὄργια βά||ζων | Ἡρόφιλον δ’ ἐκάλουν µε | πατὴρ καὶ πότνια µήτηρ | ἔ|γνως, ̑ω παροδεῖτα, τίς ἤµην· | οὐκ ἐγενήθην

This was dedicated to the gods of the underworld. Not yet having tasted youth, I slipped into the realm of Hades, leaving behind tears and groans to my parents for my short time. Nor was I meant to reach the life span of mortals. I lived only seven years and two months, of which three years I accomplished and spoke the rites for Dionysos. My father and revered mother called me Herophilos. Oh passerby, you now know who I was. I was not even brought into being.

This epitaph is often discussed in conjunction with the Villa of the Mysteries since both feature young boys who speak or read (aloud) rites for Dionysus.42 But few scholars have highlighted the probable enslaved status of Herophilos. From the inscription, it is clear that he (and/or his parents) was well educated. Yet epigraphic habits of the Early Roman Empire are such that citizens are identified in funerary epitaphs by their three-part names, which testify both to family lineage and to social belonging. In Herophilos’ case, therefore, the single Greek name and the absence of any declaration of freed status signals the boy’s enslavement. This epitaph, taken along with the evidence for amphithaleis in the Torre Nova inscription, makes it clear that enslaved children regularly fulfilled important functions in ritual operations of Dionysiac cult.

Class differences that must have existed across the spectrum of young boys involved in Bacchic cult also emerge in the comparison of Herophilos’ epitaph with a later third- or fourth-century CE funerary inscription for a 7-year-old named Aurelios Antonios.43 This boy’s epitaph, when stating that he performed the mysteries (ἐκτελέσας µυστήρια), uses the same verb that appears in Herophilos’ inscription. But unlike Herophilos, Aurelius Antonius, a Roman citizen, is identified as a priest (ἱερεὺς). From the emphasis on this boy’s priestly office it would seem that he enjoyed a higher social standing than Herophilos, in the cult as much as in the world beyond it.44 Aurelius’ inscription, furthermore, does not task the child with speaking duties, the servile nature of which will be explored in greater depth in the next section together with the visual record for boys in Dionysiac cult.

Scroll, Thyrsus, Tray: Naked Boys in Visual Allusions to the Mysteries

The epigraphically attested enslavement of children with ritual import in Dionysus’ mysteries can and should impact interpretation of the naked bodies of young boys in the artistic record of the Late Republic and Early Empire. Given the enslaved status of ritually significant boys like Herophilos and the amphithaleis in the Torre Nova inscription, this section begins dismantling the perceived neutrality of ritual nudity by highlighting important distinctions in the ways that naked boys are represented in visual allusions to the mysteries. My analysis emphasizes the variety of objects placed in the hands or on the backs of these figures: thyrsi (staffs topped by a pinecone, carried by followers of Dionysus), scrolls, tablets, and trays. These objects, I argue, had different valences for viewers familiar with the mysteries, and helped them discern a free child from a slave.

The thyrsus, notably, appears to distinguish a young, freeborn child from a child laborer. A 2- or 3-year old boy in a small fresco panel (ca. 40–30 BCE) from an ambulatory passage in the Villa della Farnesina in Rome provides an excellent case study.45 This child is naked save for a wreath on his head and cothurni on his feet. He holds a thyrsus in his right hand (fig. 7).46 A draped woman behind him leans forward toward the boy, her arms stretched as though to protect and shepherd him. The pair stands in the center of the panel, framed by a tall, draped figure (sex is unclear) at left with a probable thyrsus and a female holding a tympanum (a drum or tambourine) at right. This panel revolves around this child and his experience, which strongly intimates his identity as a very young initiate.47 Significantly, the boy is not tasked with doing anything in particular in the fresco; he simply participates, which distinguishes him from the vast majority of older naked children in other images that reference the mysteries.

Fig. 7. 
Fig. 7. 

Watercolor after a panel painting of a naked boy holding a thyrsus, with an older woman behind him, Late Second Style fresco, ca. 40–30 BCE, from the Villa della Farnesina in Rome. Archival image (D-DAI-ROM-33160) of a late 19th-century watercolor by G. Massuero, now presumed lost (courtesy DAI Rom – Fotothek).

Most naked boys who appear in Dionysiac images do not hold a thyrsus; they are rather engaged in physical work for the cult. An Imperial-era marble relief now in the Louvre, for example, depicts a small naked boy in cothurni who hunches his back, balancing a heavy tray of fruit on his head (fig. 8).48 A bearded older man follows the boy and lifts his right hand out to stabilize the fruit on the tray (that is, he does not directly aid the boy or lessen his burden). The boy carries the tray to an altar, behind which stands a priestess with a thyrsus in her left hand. This same trio of figures—older man, laboring naked boy, and female priestess—appears on an ornate gladiator helmet from Pompeii, suggesting that this scene belongs to a standard repertoire of Dionysiac cult images.49 In both the relief and the bronze helmet, notably, the boys participate in the cult at a high level, judging from their proximity to the sacrificial altar, the priestess, and, in the Louvre relief, the revelation of a triptych on a column behind the draped cloth. And yet these boys do so as nude, young laborers. Their enslaved status is signified by their naked, hunched bodies, which pointedly contrast with the older, draped, and unencumbered bodies of the adults around them.50

Fig. 8. 
Fig. 8. 

Dionysiac relief of an older figure, a naked boy carrying a tray, and a priestess at an altar, 100–200 CE, marble. Paris, Musée du Louvre Ma 50, acq. 1803 (© RMN-Grand Palais / Art Resource, NY).

Perhaps the closest parallel to the Villa of the Mysteries’ reading boy (in terms of iconography but also geographical location) comes from a subterranean triclinium in the House of the Cryptoporticus in Pompeii (I.6.2), the walls of which were adorned with Dionysiac caryatids and small, fictive pinakes. A Late Second Style panel on the south wall dating to ca. 40–30 BCE depicts a naked boy who holds a large tablet in front of him (fig. 9).51 From his posture and gesture, he appears to be tasked with revealing this tablet (covered with indiscernible markings) to the other figures to his right, all of whom disregard the boy and look intently at the tablet. From left, these figures include a seated female who rests her chin on her right hand; a draped woman who stands behind her; and in the center of the panel but somewhat in the background, a winged goddess. The scene likely references some sort of revelation or instruction given to initiates in the cult, which appears on the tablet that the boy presents to the expectant women.52 The act of revelation may not seem like a particularly onerous task, but it is a labor nonetheless; that is, the contents of the heavy tablet that the boy holds are not offered up for his own person but by him for the older, presumably superior female figures in the scene. His nudity may therefore convey his participation in a ritual event, but it also marks him as a social inferior.53 As a child, moreover, it must also be stressed that his nakedness is not a choice made of his own volition.

Fig. 9. 
Fig. 9. 

Naked boy holding a tablet for two women and a winged goddess, Late Second Style fresco, ca. 40–30 BCE, from the House of the Cryptoporticus (I.6.2) in Pompeii, excav. 1911–29, Scavi Archeologici. Pompeii, Depositi del Parco Archeologico di Pompei 59469a. Print image (tricromie dal vero) by the Istituto Poligrafico e Zecca dello Stato, Rome, 1953 (after Spinazzola 1953, pl. XXX.b).

The general presentation of this scene is not dissimilar to that of the boy in the Villa of the Mysteries frieze, but in the latter the reader’s low status and lack of personal agency are even more pronounced. The text he holds may be less weighty than the previous boy’s, but the defensive posture of his body is telling. His inclined head, hunched shoulders, clenched arms, and closed legs reveal his discomfort (see fig. 3). Like the boy with the tablet in the panel at the House of the Cryptoporticus, the (much smaller) scroll is the only covering this reading child has. So too does this object mark his necessity as much as his enslavement.

Reading as a slavish activity in the Roman world has been proven by recent work on the social history of lectores and anagnostae: highly educated slaves trained in the art of reading aloud.54 Although previous scholarship has focused largely on the skills of these individuals—their elocution, intonation, expression—Howley’s work emphasizes their servile character. Elite Romans, he reminds us, “read” using the bodies and voices of the enslaved.55 That young boys served in this capacity, moreover, obtains in the literary record of the mid first century BCE: Cicero concludes a letter to Atticus with comments about his distress at the death of Sositheus, his puer festivus anagnostes (charming enslaved boy reader).56

The boys in the frescoes at Pompeii, moreover, make it clear that reading is a menial task even in the context of the mysteries. Further (if earlier) support for this argument is provided by one of the few literary references to the use of books in mystery cult, which comes from a fourth-century BCE speech of Demosthenes. Here the Greek orator insults his opponent Aeschines by mentioning his involvement in menial tasks as a youth, one of which was reading the mysteries (De Cor. 258–59):

δι᾿ ἣν πα̑ις μὲν ὢν μετὰ πολλ̑ης ἐνδείας ἐτράφης, ἅμα τ̑ω̖πατρὶ πρὸς τ̑ω̖διδασκαλείῳ προσεδρεύων, τὸ μέλαν τρίβων καὶ τὰ βάθρα σπογγίζων καὶ τὸ παιδαγωγεῖον κορ̑ων, οἰκέτου τάξιν, οὐκ ἐλευθέρου παιδὸς ἔχων, ἀνὴρ δὲ γενόμενος τῇ μητρὶ τελούσῃ τὰς βίβλους ἀνεγίγνωσκες καὶ τἄλλα συνεσκευωρο̑υ, τὴν μὲν νύκτα νεβρίζων καὶ κρατηρίζων καὶ καθαίρων τοὺς τελουμένους κἀπομάττων τ̑ω̖πηλ̑ω̖καὶ το̑ις πιτύροις, καὶ ἀνιστὰς ἀπὸ το̑υ καθαρμο̑υ κελεύων λέγειν

As a child you were brought up in extreme poverty, while you assisted your father at the grammar school, grinding ink, wiping down the benches, and sweeping the classroom, holding a domestic slave’s position, not that of a freeborn boy. And when you reached manhood, you recited the texts and organized everything else for your mother while she performed the rites, clothing initiates in the fawn-skin by night, mixing the libations, washing and wiping them with mud and bran, and then standing them up after the purification and administering the oath.

According to Demosthenes, one of the many signs of Aeschines’ poverty and depravity is assisting his mother with menial preparations for the practice of mystery cult by reading the service book.57 I acknowledge both that this speech is political invective, expressly designed to assail Aeschines’ character, and that it predates the Villa of the Mysteries fresco by three centuries. Furthermore, Aeschines has just become a man (ἀνὴρ δὲ γενόµενος) when he is tasked with reading the mysteries. And yet it is highly probable that this reference to Aeschines’ “advanced age” while following the dictates of his mother is an insult in and of itself. Thus, for our purposes, Aeschines’ participation in mystery cult is notable as early evidence for the recitation of cultic rites as a servile activity, an act undertaken (seemingly forcibly) to support a pleasurable experience for women in the cult, such as we see in the later Pompeian frescoes like that in the Villa of the Mysteries.

Naked Truths: The Trouble with Ritual Nudity in the Artistic Record

Although most naked boys appearing in visual references to the mysteries in the first centuries BCE–CE are slaves, it has been tradition to suppose that their nudity is a ritual costume.58 Admittedly, arguments for ritual nudity as the standard dress for child participants would provide an explanation for the nakedness of the young boy in the Villa della Farnesina fresco: a probable 2- or 3-year-old free or freed neophyte with a thyrsus (see fig. 7). Unlike the reading boy in the Villa of the Mysteries fresco and the other slaves performing labors noted in the previous section, there is nothing definitive to mark this small child as enslaved.59 If we accept nudity as ritually mandated for young boys in the mysteries (regardless of status), however, it would follow that the nakedness of enslaved children—nominally a visual advertisement of their inferiority—is elevated by their participation in ritual. But the sanitized notion of “ritual nudity” itself merits dis-assembling.

First and foremost, “ritual nudity” does not feature prominently in material depictions of religion and ritual practices in the Late Republic and Early Empire. Rather, mortal actors, especially those who serve as priests or attendants, are dressed and their genitals are covered without exception (enslaved or free, child or adult), regardless of whether they take part in state religion and initiated cults.60 Artistic depictions of naked mortals engaged in ritual activity—which would have surely garnered the attention of a viewing audience—are found almost exclusively in images that allude to Dionysus’ mysteries.

Yet some participants are more naked than others. Children are the only mortals who appear fully nude (see figs. 13). They may wear boots or hair ornaments, but their bodies are most often bare.61 These children are frequently joined by other nude or semi-nude figures, especially women, such as bacchantes and the god’s mythological followers, but their nude or semi-nude bodies, by contrast, are clothed with some type of garment or barely concealing drapery, which slips from their shoulders or swirls around them. The largely ineffective drapery draws attention to their semi-nudity, and yet the mere presence of clothing contrasts with the stark nakedness of many youth participants in Dionysiac tableaux. The ability to wear and remove clothing would suggest that adults enjoy agency within the cult whereas child participants do not. Their nakedness is compulsory.

One might therefore wish to interpret child nudity in the extant artistic record as a sign of the prepubescent minor’s innocence and innate purity, even beyond the age of infancy.62 But, as Johnston’s work on the use of children in mystic Graeco-Roman rituals has shown, children were not regarded as pure simply because they were young; rather, children were subject to the same ritual purification as adults insofar as (sexual) purity could not be assumed for participants, regardless of their age.63 As Johnston notes, “purity” is hardly a natural state; it is rather a temporary condition into which individuals of all ages had to be brought before they could participate in certain rituals.64 Thus, if it was tradition to purify individuals (via ritual baths, for example) for participation in mystic cults, and to convey that purification as having been completed by means of nudity, the extant artistic record would presumably be full of naked people in Roman religious settings. It is not. Artistic allusions to Dionysiac cult are the exception in the Late Republic and Early Empire, although as previously noted, only male children are commonly shown in full nudity.

Their nakedness in contrast to other, older participants appears to buttress literary and historical characterizations of the mysteries as licentious, with its youngest members subject to varying kinds of emotional and physical abuse. This is famously suggested by Livy’s discussion of the Bacchanalia affair of 186 BCE, which can be broadly summarized as follows.65 A formerly enslaved, female sex worker with intimate knowledge of Bacchic cult desperately tries to prevent her freeborn male lover’s initiation in the mysteries; she knows firsthand that the cult preys on youths, especially persons under the age of 20. In the course of things, the sexual promiscuity and corruption which characterize private Bacchic rituals comes to the attention of a consul. A formal investigation of the cult ensues, which results in its suppression by the senate on the grounds that it is fomenting a conspiracy against the state. Livy’s account is admittedly dramatized,66 but his comments on the involvement of young persons in Bacchic rituals and sexual depravity as characteristic of such rituals are telling—they reveal widespread knowledge of the cult’s exploitation of minors in the later Republic.67

From a sociohistorical perspective, however, the representation of young male participants as “ritually” nude in Roman-era art may evince Rome’s appropriation of the mysteries from the East, together with the aesthetic traditions of the Greeks. For indeed, apparently genuine ritual nudity is documented in literary evidence for select religious festivals involving children in the Greek world (e.g., Sparta’s gymnopaedia).68 Moreover, depictions of naked male youths in religious or ritual contexts, for example kouros statues and on grave stelae, are commonplace from the Late Archaic period onward.69 It is thus possible that Roman practitioners of Dionysiac cult borrowed certain Greek customs to preserve its “authenticity,” for instance the ritual nudity of youth participants and their frequent depiction as such in the artistic record.

Yet a ritual nudity lifted, borrowed, appropriated from the Greek world is nonetheless fraught. In the first place, the frequent display of naked youths in Greek art seems to have done little to desexualize their bodies. As Osborne has argued, the unclothed body of a male youth was titillating to the predominant viewing audience (the Greek citizen), whether in public or private contexts, in the gymnasium or the bedroom, in art or real life.70 We therefore cannot assume that the nakedness of a young boy in a religious setting diminished that figure’s potential to arouse and excite at least some audiences in the Greek-speaking world, simply because it was viewed in a ritual context.71 Ritual nudity in Greek art thus can never be wholly neutral, since naked boys were objects of desire seemingly regardless of context. Furthermore, if we accept that the ritual nudity of young boys in Roman-era images of the mysteries is indebted—at least in part—to traditions that were appropriated from the Greeks, we must entertain both the latent eroticism of this iconography, and the violence it conditioned in viewers towards minors broadly, but especially towards socially inferior, enslaved minors. Violence further permeates “ritual nudity” insofar as this artistic tradition itself comes by way of Roman capture and subjugation of the Greek East—its culture and its peoples.72

The baggage that such nudity carries may not have offended elite Roman viewers, but it undermines modern syntheses of ritual or Greek-inspired nudity as positive or neutral in the later Republic. At best, the nakedness of a young child in Dionysiac art is a testament to their inconsequentiality (as voiceless minors) in cult affairs and society broadly. However, judging from the extant evidence (see fig. 7), only the infans was afforded this luxury. At worst, and far more frequently, the naked bodies of working children marked their subservience to all others in the cult. Indeed, Roman sensibilities about nakedness as the quotidian costume of enslaved and sexually available persons are such that even the ritual context of a slave boy’s nakedness failed to convey his worth to viewers.

Decorating with Naked Pueri

But what might be made of the Villa of the Mysteries boy’s functional role as decoration? The hall that houses this fresco cycle is the largest in the villa (the starred room in fig. 10) with the exception of the atrium. It is frequently identified as a triclinium, and physical properties of the space affirm its social function.73 The room is located deep in the southern wing of the pars urbana (residential suite of a villa), which restricts access to inhabitants and invited guests.74 This, together with the painted decor and the curated views of gardened porticoes overlooking the coast provided by a large window on the south wall, establish the room as a premiere locus for banquets, reception, and various forms of private entertainment.

Fig. 10. 
Fig. 10. 

Plan of the Villa of the Mysteries, first constructed in the early first century BCE. The room with the fresco cycle is marked with a star (drawing by G. Longobardi, with modifications by A. Wallace-Hadrill and D. Brin; courtesy Parco Archeologico di Pompei).

At the most basic level, the reader and the other lavishly painted two-dimensional figures who stand alongside him project the wealth, culture, and social standing of the villa’s owners, who organized and controlled events in this hall, this residence, and the Roman world broadly. Admittedly, this interpretation is in line with previous scholarship on luxury domestic arts, which has tended to privilege analysis of these objects as their owners intended them to be viewed, as material conversation points in dialogues of status and display. But with an understanding that decoration in the Roman home reinforced social constructs and established hierarchies,75 let us explore the culture motivating the prominent placement of the enslaved child in the Villa of the Mysteries fresco cycle. For the earlier discussions of Dionysiac ritual and enslavement have barely scratched the surface of his naked body’s ability to invite and indulge the elite gaze.

Although the erotic character of the Villa of the Mysteries fresco is undeniable, with more than a half dozen nude or semi-nude figures,76 the boy is easily the most available. By contrast with the naked reader, other nudes and semi-nudes are adults or numinous beings, often with drapery to cover their genitalia.77 Take for example Dionysus and the winged goddess on the east wall (see fig. 1); the latter (whose breasts are exposed) is not emphatically sexualized, whereas the god’s open body position is more inviting, even if his genitals are hidden from the viewer. Two bacchantes in the southeast corner, the woman whose garment has been pulled off her back to prepare for a whipping and the dancing woman, tease viewers with the naked female body. Their draped garments fall in folds around them, removed or willingly cast off, yet neither woman’s pudenda are visible; a viewer must focus on other erogenous zones.78 All four of these figures and the Silenus on the north wall, whose drapery has slipped from his body revealing his genitals, are securely located in the Dionysiac realm. The nude figures in the western half of room, however, mark a departure. Here are three undraped childlike male bodies among a sea of dressed women in a nonmythological realm.79 The reader, the only mortal in this trio, stands opposite one of two Erotes who flank the bride in the southwest corner of the hall (see figs. 3, 4). The reader’s full frontal nakedness invites the viewer’s eyes to linger on his body, with his placement across from Amor—a childlike god of love, sex, and desire in the Roman era—further encouraging his eroticization.

Viewers likely found the enslaved boy’s soft and hairless physique attractive, especially when coupled with his deferential, downward glance.80 Latin literary sources from the second century BCE on manifest desire for the effeminate or androgynous bodies of boys, and they do so in ways that are intensely visual.81 According to Lucretius, for example, gazing upon the womanly limbs of a boy (puer membris muliebribus) aroused desire in viewers and in men brought on the urge to ejaculate.82 And significantly, although cultural mores in the Roman Republic cautioned against making sexual use of freeborn boys, there were no such stipulations against enslaved boys.83 The sexual availability of such persons made them an especially appropriate target for the erotic fantasies of elite slaveholders, whether those fantasies were brought on by someone immaterial (e.g., a painted boy) or someone real.

Indeed, in contexts like the frescoed room of the Villa of the Mysteries, the objectifying gaze was rewarded as much by images of attractive boys as by living, breathing, beautiful pueri. Sexy slaves of all ages were fixtures at the cena (dinner party) in the Late Republic and Early Empire.84 For pueri did not simply serve the host’s guests; their good looks also advertised the power and prestige of the owner and provided endless entertainment.85 According to Philo of Alexandria (40s/30s BCE), boy slaves at Roman-style dinner parties were so beautiful that their presence was explained less by their function as servers and more by their ability to bring pleasure to the eyes of voyeuristic guests.86 Dining in semi-repose was, furthermore, the premiere venue and position for ogling these slaves as they, by contrast, stood and moved around the room, occasionally surrounded by simulacra of themselves on the walls or in three-dimensional statues.87

This semiotic overlap of images, objects, and enslaved bodies at the Roman cena has received attention from scholars in recent years, although most have focused on three-dimensional representations of slave boys as opposed to painted depictions.88 For example, both Lenski and Bielfeldt have offered independent analyses of Roman-era bronze statuettes from Pompeii shaped as idealized male nudes.89 As art pieces but also physical supports for lamps or trays, these statuettes were multifunctional (fig. 11). Therefore, both scholars have placed them in conversation with ancient literature about pueri in the domestic sphere. Lenski’s work frames the statuettes as “working models,” and argues that they stood as material surrogates for the beautiful servers they reference.90 To an ancient viewer, Lenski further suggests, a pretty art object and a beautiful young slave have more in common than modern audiences might assume. Citing Varro on slaves as animate objects, Lenski reminds readers that enslaved bodies were speaking tools to Roman slave owners, human props that made elite life possible.91 So it follows that both art and enslaved persons were made to serve different but more or less comparably functional ends.

Fig. 11. 
Fig. 11. 

Statuette of a nude boy serving as a support for a candelabrum, first century CE, bronze, from the House of the Ephebe (I.7.10–12) in Pompeii. Naples, Museo Archeologico Nazionale 143753, acq. 1926 (G. Albano; courtesy Museo Archeologico Nazionale di Napoli).

As the physicality of these bronze sculptures is stressed in both Lenski and Bielfeldt’s scholarship, we must admit that painted pretty boys worked somewhat differently. Naked bodies in frescoes could not be handled (even if they could be stroked or petted), nor could they be put to work functionally in the same manner as serving statues. But their two-dimensionality does not diminish their ability to blur the lines between fantasy and reality. For example, the boy in the Villa of the Mysteries frieze is worth considering together with the description of a young slave who appears at Trimalchio’s dinner party (mid first century CE). At one point in the freedman’s interminable banquet, the narrator Encolpius is taken with a beautiful puer who enters the dining room dressed as Dionysus, wearing vine leaves and ivy in his hair.92 We are told that the boy parades around the room with a basket of grapes, impersonating the god and reciting Trimalchio’s verses in a high-pitched voice. He leaves to a round of applause and kisses from the diners. With relatively little difficulty, one can imagine such an entertainer bringing the nearly life-size lector on the wall of the Villa of the Mysteries’ triclinium to life: putting words into the boy’s mouth; animating his limbs; breathing life into his painted flesh; preparing him to entertain guests; imbuing him with import in the context of yet another sacred domestic ritual—the cena.93

The emphasis on gazing, viewing, and visual consumption at these banquets is such that painted images and enslaved pueri are virtually compelled to work in concert, distorting distinctions between fiction and the contemporary world, thereby generating the sort of private spectacle that delighted Roman audiences.94 Together they fulfill physical hungers of various kinds, generating multisensory experiences and sensations for many of the guests, whether male or female. Indeed, the naked reader appears to pay special attention to the needs of the women around him, thereby inviting the gaze of female viewers as much as male, whether their fantasies manifest as sexual encounters or more subtle but subversive challenges to the established hierarchies of a patriarchal society.95

The appeal of a painted puer like the one in the Villa of the Mysteries is rather obvious when one considers the special advantages this boy enjoyed over his living counterparts, even as a mere simulacrum of the latter. He did not need to be undressed by an interested viewer, nor was he able to put up any serious opposition to a guest’s wandering eye (nor elicit a host’s rebuke). Whether painted or sculpted, these naked boys could not lose the blossom of youth; nor did they face threats of stubble, body hair, or the deepened voice of postadolescence.96 Nothing stood to advance their age or lessen their beauty. A viewer, therefore, did not need to fear losing their chance to make sexual use of the boy at the right time.97 Naked boys in Roman domestic arts, we may conclude, avoided what the elite viewed as an inconvenient reality of their beloved sexy servers. As Martial laments, real pueri were doomed to grow older, at which point they were banished from their front-of-the-house role in the dining room forever.98

Age Ain’t Nothing But A Number

One might contend that at 6 to 10 years old, the Villa of the Mysteries boy is too young to be eroticized by a Roman audience.99 But as Richlin has argued in a recent article on child love, sexual attention, and Roman deliciae (pet slaves), age was no barrier to the sexual exploitation of the enslaved.100 Deliciae, prized for their good looks, garrulity, and general precociousness, were mere children in the modern sense of the word (toddlers to preteens), but they were eroticized from an early age and subject to abuses of all kinds.101 Modern historians working on this group generally follow Rawson, who wrote in 2003 that for these children “the line between indulgent affection and sexual exploitation must have been blurred.”102

Some deliciae received explicit sexual attention. Erotic language and elegiac phrasing in Statius’ Silvae 2.1 and Martial’s Epigrammata 5.37, both of which eulogize deliciae, are telling even if sexual relations with these children are not openly discussed by the authors.103 The routine eroticization of deliciae is, moreover, implicit in ancient comments about their comportment, manners, and behaviors. A passage in Cassius Dio about one of Livia’s boys is especially revealing. According to the historian, Livia and Augustus sponsored a banquet shortly after their marriage when she was pregnant with Drusus. Among the attendees was Livia’s former husband Nero, father of her unborn child. Drawing attention to the elephant in the room was none other than Livia’s puer, whom Dio describes as “one of those prattling, chatty boys that women kept around for their amusement.” The boy, he adds, was naked at the party, as was standard for these children (παιδίον τι τ̑ων ψιθύρων, ο̑ἱα αἱ γυνα̑ικες γυµνὰ ὡς πλήθει ἀθύρουσαι τρέφουσιν).104 Here Dio marks the nakedness of a deliciae as typical for a dinner party, but he also gives the boy credit for calling attention to his mistress’ sexual liaisons, a subject the child likely knew well, either as a witness or participant in some way.

Indeed, this passage calls to mind domestic paintings of slave boys observing their owners engage in sex acts.105 An example is furnished by yet another later first century BCE fresco from the Farnesina villa, which decorates a cubiculum (D) filled with erotic images.106 It features a man and woman on a bed, presumably in the early moments of a tryst (fig. 12). A viewer’s voyeuristic reception of this couple, however, is arrested somewhat by a naked slave at the far right of the panel. His body as well as the cup in his left hand call to mind images of Ganymede, which works to romanticize his nakedness in much the same way as the naked reader’s. Interestingly, although the body of the Farnesina boy faces the couple he attends, his head turns left towards the viewer. The nakedness of this puer begs his audience to acknowledge his sexual availability, his demure eyes inviting and ingratiating interest in erotic matters that require his participation.107

Fig. 12. 
Fig. 12. 

Male-female couple attended by a female slave and a nude slave boy, who looks out toward the viewer, Late Second Style fresco, ca. 40–30 BCE, from the Villa della Farnesina in Rome. Rome, Palazzo Massimo alla Terme 1188, acq. 1900 (S. Castellani; courtesy Museo Nazionale Romano).

That young slave boys were common attendants to their owners’ erotic encounters, and perhaps even facilitated them, is suggested by a contemporary poem of Propertius.108 As the drunken poet stumbles toward the house of his lover Cynthia, he is accosted by a crowd of small, naked boys (pueri minuti … nudi) with torches, arrows, and chains.109 The language—pueri minuti—calls to mind enslaved boys, but it is their dress—nudi and “armed” with Cupid’s weapons—that I wish to stress here. In Propertius’ imagination as much as his reader’s, it would seem that Cynthia’s pueri take on the character of Amores who help the poet find his way to his lover, their mistress.110

Cupid-like pueri who help their owners make erotic connections also appear in an episode of Plutarch’s life of Antony, when Cleopatra comes to Cilicia in 41 BCE to meet the triumvir in a most spectacular fashion. According to Plutarch, the Ptolemaic queen arrives on the river Cydnus in a barge, dressed as Aphrodite, flanked by a band of beautiful boys likened to Erotes in paintings (πα̑ιδες δὲ το̑ις γραφικο̑ις Ἔρωσιν; πα̑ιδες here meaning slaves).111 The author’s direct comparison of Cleopatra’s boys with painted images of Eros is a testament not only to the spectacle the ruler has put together for Anthony but to the power of visual suggestion.112

These literary vignettes of enslaved children taking on the personae of Amores call to mind George’s recent work on images of Cupid in chains, which are well documented in frescoes and sculptures from the Hellenistic era onward (fig. 13). George associates the popularity that the Cupid Punished trope enjoyed among the Romans with their corresponding predilection for deliciae.113 Her work identifies this artistic trope as a sentimentalized version of child punishment, which accommodated social and cultural practices of slavery. I would add that the routine overlap of Cupids with naked children and slave boys in the Late Republic and Early Empire is a less than subtle reminder that it was standard practice to eroticize young slaves of all ages.

Fig. 13. 
Fig. 13. 

Fountain statue of an Amor, first century CE, marble, from the House of the Vettii (VI.15.1) in Pompeii, excav. 1894–95, Scavi archeologici. Pompeii, Depositi del Parco Archeologico di Pompei 1158 (© Alinari Archives / Art Resource, NY).

The same mentality informs the presentation of the Dionysiac slave boy in the Villa of the Mysteries, who stands on the north wall opposite an Amor. Although the boy is not mythological, it is hardly difficult to imagine him taking on qualities of this same Cupid, should a besotted slaveholder wish him to do so (as in the poem of Propertius). Nor does his nakedness strike the well-dressed woman seated behind him, who rests her hand on his shoulder in a gesture that conveys not only her possession of the boy but her comfortable familiarity with his presence, as in the literary passages cited above (fig. 14). The frequent enslavement of children, moreover, is such that we should avoid characterizing her gesture as affectionate, which sentimentalizes her power over a prepubescent minor.114

Fig. 14. 
Fig. 14. 

The naked reader, north wall of the Villa of the Mysteries frieze (L. Horan).

As a final cue that the depiction of the Villa of the Mysteries boy dialogues with cultural attitudes toward Roman deliciae, I note the stylus that the woman behind our naked lector dangles over his right shoulder. It is difficult to connect this instrument to documented cultic activities or preparations associated with the mysteries and the reading of rites.115 With the cultural currency of deliciae and enslaved pueri in mind, however, we may read the stylus as serving a secondary function: a subtle reference to the training of these young slaves in the arts of reading, and the concomitant cultivation of their characteristic chattiness and impudence, all of which made them hits at Roman dinner parties.116 On whatever level a viewer layers their interpretations of this image, the boy in this fresco is being prepared for his appearance at an elite ritual that requires him to use his voice and expose himself to others, for their benefit alone.

Slater wrote in a 1974 article on deliciae that “the relation of art to real life in this case is very unclear.”117 But is it? Literary references to deliciae are surely in conversation with images of young, nude boys in the material record of the Roman house. Together they betray an elite mentality that authorizes and affirms the abuse of young enslaved bodies: everything from conversational microaggressions to voyeurism and rape. Especially in private contexts, images of naked boys were designed to grant pleasure to persons in places of authority, reinforcing their position over the marginalized.

Portrait of a Beloved?

Insofar as images of naked boys are powerful reminders of the systems of enslavement that support the status quo, the naked reader’s servile character merits consideration alongside Gazda’s argument for the Villa of the Mysteries fresco as a commemorative portrait of a mid first-century BCE household.118 Gazda’s recent work draws attention to the individualized physiognomies of several female figures in this fresco, especially the bride-to-be on the southwest wall (see fig. 4) and the veiled domina (matron) in the northwest corner of the room. She identifies both as portraits: the bride-to-be is the daughter of the woman represented as a matron, in whose honor the painting was commissioned posthumously.119

While attention to Bacchic cult in the larger fresco surely plays to the conceit of mythological portraiture,120 the depiction of a familia as participants in a Dionysiac thiasos is, for Gazda, a visual affirmation of the social bonds that held an elite Roman family together. Her argument here rests on the aforementioned Torre Nova inscription as documentary evidence for the full participation of both free and enslaved family members in private Dionysiac rituals.121 If we accept Gazda’s argument for this fresco as a commemorative familia portrait, it would follow that many of the painted faces on these walls are portraits of enslaved persons. Gazda does not discuss the naked boy in her analysis, but I note that even with part of his face hidden from view, his thin lips, the comparably wide mouth, and the sharp angling of the eyebrows near the bridge of the nose gesture in this direction (see fig. 14).

From an elite slaveholder’s point of view, the inclusion of a pet slave in a commemorative family portrait would represent a singular honor; although elite families had many slaves, few were so fortunate as to be preserved in perpetuity in a celebratory image of the gens. But as previously argued, sentimentality among slaveowners for child slaves does not negate their functional role as objects. Indeed, the depiction of the boy in the Villa of the Mysteries frieze alongside a probable slaveowner reflects the predominant understanding of deliciae as inconsequential speaking tools that sustain an elite world view.

Other portraits of deliciae, whether in the visual arts or literature, are similarly conceived.122 Consider the delicatus in a Fourth Style fresco from the House of the Triclinium (V.2.4) in Pompeii (fig. 15). This figure, an African boy of 6 or 7 years, has been identified by Clarke as a possible portrait.123 The child sits near the center of this panel to the left of the cena’s host. He turns toward this much older man and rests his left hand on the man’s chest. Visual parallels between this couple and that in the Villa of the Mysteries frieze suggest that a standard set of conventions informs representations of enslaved children and their owners.124

Fig. 15. 
Fig. 15. 

Dinner party featuring the host with his deliciae, Fourth Style fresco, ca. 62–79 CE, from the House of the Triclinium (V.2.4) in Pompeii. Naples, Museo Archeologico Nazionale 120009, acq. 1890 (G. Albano; courtesy Museo Archeologico Nazionale di Napoli).

First, the power of the adult is manifest in their seated posture.125 The manner in which their bodies make contact with the enslaved further signals their dominance. In the House of the Triclinium painting, the host’s left arm is draped around the child’s shoulders in a proprietary manner, enveloping the boy much as the woman does in the Villa of the Mysteries fresco.126 Yet neither adult looks directly at the enslaved child, an open indifference to their personhood and another assertion of their superiority.127

The parallels across young male deliciae in these couplings are further striking. Differences are readily apparent: the youth in the House of the Triclinium is clothed, whereas the boy in the Villa of the Mysteries panel is naked; the former dines with his owner in a contemporary setting, while the latter’s headwear and boots engender a fuzzy dissolution between fantasy and reality. And yet these are superficial distinctions, for both children are pretty playthings at work, demonstrably deferential to the adult figure under whose manus (literally “hand,” but also power) they operate. The child in the House of the Triclinium is, like a good delicatus, incredibly attentive—he looks up eagerly at his owner, ready and willing to satisfy any needs the latter might have presently or in the future. By comparison, the boy in the Villa of the Mysteries is charged with a different task—reading—but he exhibits a similar commitment to this role. His close consideration of the scroll in his hands may be read as evidence for his hyperfocused attention to his job. His person also suggests his sexual availability, but in a rather different manner than the boy in the House of the Triclinium—nakedness, as opposed to his body language, invites attention.

Compliance Versus Resistance: Elite and Nonelite Viewing Audiences

But what of the nonelite viewers? How might they have responded to this figure? Although the boy and the painting to which he belongs were painted for elite consumption, a diverse population encountered the Villa of the Mysteries fresco on a daily basis: the villa’s owners; their guests; and the enslaved workforce who were essential to the maintenance of both the residence and the owners’ sumptuous lifestyle. Although enslaved viewers have received relatively little attention in previous scholarship on luxury arts in the domestic sphere, Joshel and Petersen’s work has highlighted their constant, necessary presence in the Roman home.128 Enslaved persons of all ages, they remind us, frequently found themselves in elite spaces like the Villa of the Mysteries hall, whether serving diners at a banquet, assisting their owners during a formal reception, providing various forms of entertainment, or cleaning and preparing the room for social rituals. The enslaved staff of this villa likely spent no less time in its triclinium than the owners who commissioned this fresco cycle. These incidental enslaved viewers were also no less visually literate,129 and they would have identified the naked reading boy as a child slave. But because they looked on him with the eyes and experiences of the enslaved class, their reactions merit amplification, to the degree that such is possible.

For some enslaved viewers, the boy was a simple reminder of the subservience that was characteristic of a slave’s life and that permitted the use of their body in various ways. One basic purpose of not only the reader’s full frontal nakedness, but also and especially the casual attitude of the women who surround him, seems to have been the normalization of servitude, the conscious conditioning of all viewers (elite and nonelite alike) to accept the subordination of some humans as a fact of life.130 And yet, not all enslaved viewers would have welcomed this message, for dominated groups do not always look at art in the ways that the dominating class intends.131

Though we can only imagine how different slaves responded to the boy, I argue that, for some, their gaze ran counter to the establishment, silently undermining those same power structures. This argument is informed by recent work on slave resistance in the Roman period.132 For example, scholars have begun to synthesize comments made by slave owners about day-to-day frustrations with their slaves—protestations of laziness, untruthfulness, carelessness, and misbehavior—as evidence for challenges that the enslaved class regularly raised, whether by dallying on errands, shirking work, or attempting to flee.133 Richlin, Joshel, and Petersen have further explored the idea of a “hidden transcript” or a private discourse among the enslaved—ways of speaking and being—that further push back at the establishment.134 Richlin has examined slave resistance in the fabula palliata and the opportunities that the comic stage provided to enslaved actors (in costume and mask, i.e., disguised): the ability to speak aloud desires denied to the enslaved class; and the ability to speak out, critically and mockingly, against the slaveholding class.135 Joshel and Petersen have similarly studied the spaces frequented by slaves in Pompeian homes and villas, and the opportunities they afforded. Enslaved persons, they argue, were able to exert a degree of agency over their lives with actions as seemingly minor as lingering in, or retreating to, a particular area of the home that was known to be free of the owner’s gaze.136 The work of these three scholars necessarily locates slaves in a particular place (the stage, the home), and, by focusing on the spoken word or movement, characterizes resistance as an action that the enslaved undertook consciously.

The resistance painted into the presentation of the boy in the Villa of the Mysteries, I argue, suggests a different and subtler form of the same—less active and less conscious, perhaps, but no less present. Let us first consider the creation of this image, since the boy’s iconography cannot be connected with the pattern-book traditions informing other figures in this frieze and Dionysiac art more broadly.137 The naked reader, who is both a ritual participant and a delicatus, may well be a portrait, in which case his depiction is informed at least in part by the reality of slavery. And while the painters surely adapted existing iconographic motifs for his general presentation,138 they likely worked from direct observation.139

The probable use of a live model may underlie the body language of the naked reader, which is notably defensive.140 His shoulders are hunched, his neck invisible. He clenches his elbows to his chest and presses his legs closed. He looks down almost performatively at the text in front of him, as though to avert his eyes from the viewer’s. Although his nakedness invites his objectification, the spine curves and the body pulls inward upon itself in an effort, however feeble, to repel that gaze and whatever unwanted advances or demands accompany it.

The meaning of this defensive stance, I argue, is deliberately vague and open to interpretation.141 Some observers surely sensed his body language as a sign of his submission, his vulnerability—an interpretation that would necessarily perpetuate feelings of helplessness among enslaved viewers, while validating the power and dominance of the slaveowning class who doubtlessly observed similar postures among their own familiae. But alternatively, a viewer might recognize this defensive body language as a sign of the boy’s obvious discomfort with the situation he finds himself in. To this viewer, the child’s defensive stance may signal his willful, if futile, resistance, thereby engendering an image that is as potentially subversive as it is patently submissive.

Indeed, the boy’s posture and his person bear witness to the abuse that typifies the lives of the enslaved, a quiet but firm assertion of their humanity. This child, a possible avatar for the many who walked in paths like his, and the ready discomfort of his naked body openly acknowledge enslaved viewers as people. If such a simple and frank admission of a slave’s humanity seems inconsequential, rather than subtly powerful and provocative, it is because we as scholars have been accustomed to read the material record in the same way as elite ancient viewers, largely unaware of and untroubled by the horrors of natal alienation, forced labor, and relentless physical, emotional, and psychological abuse.

Furthermore, to muted groups like the enslaved, this boy may offer the smallest consolation. For he is not simply a representation of a life of abuse; he is a literate enslaved child. At a very young age, the boy is tasked with reading to others. A cultivated mind and a quick intellect here do not only mark the boy’s enslavement, they are also his primary, perhaps only, means for social advancement.142 Education does not spare him from the abuse that is so blatantly underscored by the anxiety on his face, the nakedness of his figure, and perhaps even the stylus in the hand looming over his shoulder. But the disconnect between the boy’s exploited body and his nimble mind underscores the latter as the only path forward—an almost positive message, if in different ways and for different ends, among both elite and nonelite audiences.


I close this article much as I began it, though with a more direct call to continue what Joshel and Petersen refer to as the work of seeing and un-seeing, and seeing again, vis-à-vis Roman slavery and material culture.143 The Villa of the Mysteries boy appears frequently on book covers, blog posts, and advertisements for lectures and courses. In antiquity as in the present era, he serves primarily as a “sexy” background for discussions that are largely unrelated to his identity as an enslaved child, “sexy” in a broad, figurative sense for us, but, as I have argued, in a far more literal sense for the original audience. As scholars and humanists, we owe the many children whom this boy represents our attention to their voices, however faint they may be. Going forward then, let us begin with openly admitting that this boy is a child slave and continue reevaluating and even rewriting the narratives of art in the Roman house that predominate—by looking for and listening to the stories of its most powerless inhabitants whenever possible, in the places we least expect to find them.


1 My sincere thanks to the many scholars and institutions that helped me bring this article to publication, and especially to Amy Richlin, who has been a sounding board for this article since its inception. Portions of this argument were presented at a colloquium for undergraduates at UCLA (2019), at the Annual Meeting of the Archaeological Institute of America (2020), and at a conference on child slavery organized by Ulrike Roth (2021); I thank those who provided feedback and encouragement. I am also grateful to Ann Kuttner, Sam Beckelhymer, Lydia Spielberg, and the AJA Editors-in-Chief Emma Blake and Robert Schon together with the journal’s anonymous reviewers, all of whom offered insightful comments and critique on the manuscript at various stages. Any remaining shortcomings are my own. Finally, many institutions helped me secure the rights to publish the images that appear in this article, and I am especially grateful to Anna Pizza (Museo Archeologico Nazionale di Napoli) and Daria Lanzuolo (Deutsches Archäologisches Institut Rom) for their help. Translations are my own unless otherwise indicated.

2 E.g., Leach 1988; Elsner 2007; Beard 2008; Bowden 2010; Hartnett 2012; Shelmerdine 2013. See also Henderson 1996, 235–40, for extended discussion of this image in popular media.

3 For possible gender ambiguities: Ling 1991, 102. See also Bartman’s (2002) discussion of this androgynous posture more broadly in Imperial-era idealplastik.

4 Excavations at the Villa of the Mysteries began in April 1909; the great frieze was found within two weeks of breaking ground: de Petra 1910; Maiuri 1947, 23–25; see also Bergmann (2007, 239–49) for handling of the fresco cycle in this period. Maiuri would later oversee the exhaustive excavation of the site from 1929 to 1931.

5 E.g., Mudie Cooke 1913; Bieber 1928; Toynbee 1929; Maiuri 1947; Herbig 1958, 34; Houtzager 1963, 85–86; Little 1972; Brendel 1980.

6 The boy is rarely discussed in, e.g., Clarke 1991, 98–105; Ling 1991, 101–4; Gazda 2000. Only Henderson (1996, 245–47) openly questions his social status and possible enslavement.

7 For adult nudity in the Roman world with emphasis on the male body: Hallett 2005. For recent work problematizing nakedness in the Greek world (again with emphasis on the adult male body): Bonfante 1989; Osborne 1997; Stewart 1997; Hurwitt 2007.

8 Esposito’s (2007) reinvestigation of the villa’s architectural remains identifies the residence as the product of a single, unified construction in the early first century BCE, contra Maiuri (1947), who identified two phases of construction in the late second and early first centuries BCE. Esposito’s argument is based primarily on construction techniques and the location of the villa between two roads that were laid out after the foundation of the Sullan colony at Pompeii. According to Esposito, the Dionysiac frieze belongs to a large campaign of decorative renovations (mainly Second Style wall paintings) added to the villa ca. 60 BCE.

9 Here I build on Crenshaw’s (1989) definition of intersectionality, which she (a professor of law) coined to highlight social dynamics that court rulings around discrimination laws were not taking into account, such as race, gender, and class. Admittedly, intersectionality as a term has taken on a life of its own in the 21st century, but in general, studies of intersectional identities address the overlap of different social categories and the way that these intersections empower some and marginalize others.

10 E.g., de Petra 1910; Mudie Cooke 1913. Rizzo (1918, 64–67) also emphasizes the role of the child qua Dionysus as the first initiate; see also Bastet 1974, 231–33; Sauron 1984; 1998. For the room as catering to initiation: e.g., Maiuri 1947, 169–73.

11 For arguments in favor of the god as an amalgamation of Bacchus and Liber: e.g., Swetnam-Burland 2000. For his consort as Semele: Boyancé 1965/66; Sauron 1984; 1998; and as Venus: Pappalardo 1982, 271–72; Longfellow 2000b.

12 Scholars have been particularly interested in figures on the east wall near Dionysus. On the faun who looks into a bowl perhaps reading something in it (lecanomancy), see, e.g., Mudie Cooke 1913, 167–68; Toynbee 1929, 74–76. On the woman who raises a veil from a phallus and the whipping goddess, see, e.g., Nilsson 1957, 66–76, 123–29; Brendel 1980, 107–12.

13 In addition to Bieber 1928 and Toynbee 1929 (discussed below), see, e.g., Bergmann 2007; Gazda 2021.

14 Bieber 1928; Toynbee 1929.

15 Arguments for a bride-to-be’s participation in the mysteries were upheld by many scholars through the 1980s (e.g., Little 1972; Brendel 1980), but others presented arguments for the frieze as a celebration of the life and mythology of Dionysus (e.g., Sauron 1984; 1998), building on earlier interpretations of the painting as a scene of initiation into the god’s cult.

16 For early comments about the painting as a reference to ritual, rather than a depiction of actual ritual events: Rostovzeff 1927, 45–46; Nilsson 1957, 74–75. Actual initiation rites of the Dionysiac mysteries are poorly understood, but see Turcan 2003; Bremmer 2014, 104–9. For recent discussion of the visual culture surrounding mystery cult and approaches to “reading” such images: Belayche and Massa 2020.

17 See, e.g., Kuttner (1999, 107–9), who focuses on the painting’s “performance armature” and its relationship to public spectacle. See also arguments for its allusive character in Clarke 1991, 104–5; Ling 1991, 103–4; Gazda 2000; 2021.

18 See, e.g., the late first-century BCE walls of a subterranean triclinium in Pompeii’s House of the Cryptoporticus (I.6.2) with Dionysiac caryatids and small rectangular panel paintings. Striking parallels with the Villa of the Mysteries fresco appear in a panel of Silenus looking into a satyr’s bowl, and another panel of a naked boy holding a large tablet for two women and a winged goddess (discussed in depth below); Spinazzola 1953, 488–525. See also the late first-century BCE decor of several rooms in Rome’s Villa della Farnesina (Bragantini and de Vos 1982; Dolciotti et al. 1998). For other Second Style paintings, cf. the megalographic images in the House of the Prince of Naples (VI.15.7–8) at Pompeii and the Villa of Publius Fannius Synistor at Boscoreale (e.g., Longfellow 2000b, 117–20). Recent work at the Villa dei Papiri in Herculaneum recovered a fragment of a megalographic half-nude woman with a thyrsus; Esposito 2011, 538, 541–43, pl. XXXVIIIa. On the compositional dissimilarities between the Villa of the Mysteries fresco cycle and other extant megalographic figural frescoes: Little 1972, 13–39.

19 de Petra 1910, 143: “una insegna a leggere suo bambino.” Other scholars have identified the woman as the boy’s mother; see, e.g., Little 1972, 10; Uzzi 2005, 184–86. Bieber (1928, 312), however, identifies the seated woman as the bride and the woman who enters to the boy’s right as his mother.

20 Marrou 1938, 27–45, nos. 1–15, for scenes of teaching, e.g., with a magister (teacher).

21 Note also that her contemporary dress (and that of others on the north wall) discourages identification of the boy as a young Dionysus; see supra n. 10.

22 Infants are commonly depicted naked in Roman art; their nakedness distinguishes them from speaking children and adults. The precise meaning of their nakedness is unclear—naivete perhaps. Huskinson (1996, 87) has argued that chubby infant bodies on Imperial-era sarcophagi served as powerful reminders of the vulnerability of children.

23 For children in Roman art: Rawson 2003, 17–92; Lovén 2013. See also Uzzi 2005; 2007, for contrasting depictions of Roman and non-Roman children and families in state art. For children in funerary art: Huskinson 1996; Mander 2013. For infants: Carroll 2018, 118–46.

24 As far as nakedness in the Republican era is concerned, Hallett (2005, 61–158) argues that heroically nude statues of conquering generals and public persons were less frequent in this period than we might assume, even as adoption of Greek arts became commonplace. On the dominant body language of the free male in nude statuary: Davies 2018, 100–10; also 5–60 on nudity in the Greek world and its impact on Roman arts. For a fuller treatment of Greek traditions: supra n. 7.

25 For young boys at slave markets in particular, Republican-period author Varro references a market for deliciae at Capua, which is preserved in Nonius Marcellus 141.13: Si venisses Capuam, quod et pueros minutos vides libenter et | maiores animadvertere non vis [multos vidisses?] (If you had come to Capua, insofar as you both like to look at small boys and you do not wish to pay attention to older ones [you would have seen many]); see Dahlman 1950.

26 Note that several scholars have identified this boy as a young Dionysus: Rizzo 1918; Bastet 1974; Sauron 1984; 1998. These scholars presume the young god is being prepared for initiation in his own mysteries, although there is no discussion of such an episode in Dionysiac myth until the High Empire; see Nonnus, Dion. 9.111–31.

27 On the social character of the Roman house: e.g., Wallace-Hadrill 1994; Hales 2003.

28 Mudie Cooke (1913, 170–71) mentions two funerary inscriptions for children associated with Dionysiac mysteries in Italy (IGUR 3 1169 and 1228); both inscriptions are discussed below. Mudie Cooke also cites a speech of Demosthenes in which the orator accuses Aeschines of reading ritual books (Dem., De Cor. 258–59); again, see below. For later discussions of these inscriptions and the Demosthenes passage: Bieber 1928, 312; Toynbee 1929, 71–73; Houtzager 1963, 86.

29 Bieber 1928, 311–13; Toynbee 1929, 71–73. For child attendants in Roman weddings: Hersch 2010, 159–62. For children with both parents living—[pueri] patrimi et matrimi—in Roman religious contexts more broadly: Hersch 2010, 159–62; Mantle 2002, 105. There is nothing to suggest these children were naked, however; e.g., Catull. 61.174–76 refers to a child attendant to the bride as a praetextus, which signals his young age but also implies he is dressed appropriately.

30 E.g., Ling 1991; Clarke 1991.

31 See the extended discussion of the child in the mysteries in Nilsson (1957, 106–15), who assumes that most are freeborn neophytes. Indeed, although many studies of mystery cult acknowledge children as participants, they fail to consider their social status and their roles (e.g., Rostovzeff 1927; Bowden 2010, 105–36). Children are also generally absent from discussions of slavery and religion, e.g., North 2020.

32 The paintings and funerary reliefs are discussed below, but for sarcophagi: Huskinson 1996, 30–35, 101.

33 This was likely true from at least the fourth century BCE in Magna Graecia (Kron 2019). See also comments in Livy about the initiation of persons less than 20 years old: 39.10.6; 39.13.14; and Turcan 2003, no. 26.

34 IGUR 1 160: The base is now in New York, Metropolitan Museum of Art 26.60.70a, b, acq. 1926.

35 For prosopographic study of high-ranking Romans associated with the gens Gavia or Pompeia: Alföldy 1979.

36 Scheid 1986, 286–87 and Nilsson 1957, 46–47, contra others who interpret the Greek names as evidence for their ethnic origin (e.g., Vogliano 1933, 217; Cumont 1933, 236–37), or as evidence for the mysteries’ elision of social status (Vogliano 1933, 218; Cumont 1933, 234). Cumont (1933, 234) admits some of the individuals named in the inscription might be slaves, but Vogliano (1933, 225) is adamantly opposed to this idea. The only exceptions to the single name of participants are two freedwomen: Iulia Eutychia (face I, col. C.20) and Strategia Valeria (face III, col. C.10).

37 The inscription also mentions a high-ranking Roman boy named Orfitus, who is a priest (IGUR 1 160, face I, col. A.11; see Scheid 1986, 279–80), and a male archineaniskos, whose job it was to instruct children in the rituals (face I, col. B.22–23). The inclusion of an archineaniskos presumes young initiates, even if neaniskoi (youths) are not identified here beyond the amphithaleis.

38 It is usually assumed that they had special powers (e.g., divine prophecy) or blessed status; see Cumont 1933, 250.

39 Latrios is derived from the Greek noun λάτρις, meaning servant or slave.

40 The same might also be said of many adults, e.g., the hierophant Agathopus (face I, col. A.16–17). What is remarkable in the case of Latrios and Menandros is their status as amphithaleis (children with both parents living) because slave children had no legal status or legal family. Yet this parallels the use of freeborn terms (e.g., vir, uxor, maritus, coniunx) in the epigraphic record for older slaves and freedpersons; Treggiari 1975; 1981; Joshel 2010, 141–52.

41 IGUR 3 1228:; trans. in Harland 2012.

42 This inscription was first discussed in conjunction with the Villa of the Mysteries boy by Mudie Cooke; see supra n. 28. The author identifies Herophilos as a priest, although the inscription does not name him as such.

43 IGUR 3 1169: Mudie Cooke 1913, 171; Toynbee 1929, 71–73.

44 This argument finds additional support in the Roman boy Orfitus named as a priest in the Torre Nova inscription; see supra n. 37.

45 I follow recent arguments for the 40–30 BCE date of frescoes in the Villa della Farnesina, a decade that Mols and Moorman (2008, 77) characterize as “seemingly liberal” with respect to aesthetic tastes, especially in comparison with later Augustan art. This date is largely accepted in recent scholarship; see La Rocca’s (2009) catalogue of the frescos at the Villa della Farnesina as well as the House of Augustus on the Palatine.

46 Rizzo 1918, 47, fig. 7; Nilsson 1957, 81, fig. 13; Bragantini and de Vos 1982, pls. 16, 30; Dolciotti et al. 1998, fig. 42. For extended discussion of the cryptoporticus, its painted decor, and iconographic allusions to Dionysiac mysteries, see Dolciotti 1982; Dolciotti et al. 1998, esp. 33–45.

47 Here I accept Nilsson’s (1957, 81) identification of the boy as a neophyte, although Rizzo (1918, 47) identified this child as Dionysus at the moment of his initiation into his mysteries based on the premise of divine nudity of the boy and the women, whom he identifies as the nymphs of Nysa. Others have followed suit, cf. Dolciotti 1982, 84–85; Dolciotti et al. 1998. There is, however, no comparable representation of a young Dionysus in this fashion in extant depictions of the god with nymphs (LIMC 3:479–81 s.v. “Dionysos”).

48 Rizzo 1918, 58–59, fig. 13; Rostovzeff 1927, 79–80, fig. 15; Nilsson 1957, 83–84, fig. 15.

49 For the helmet: Rostovzeff 1927, 94–97, fig. 20; Nilsson 1957, 84–88, figs. 17a, b.

50 We might also include the naked figure who appears in a Dionysiac relief scene on a glass amphora (printed in Nilsson 1957, fig. 14). This figure is short and stocky; he may be a boy or a dwarf. The iconography suggests an enslaved initiate: his head is covered with a cloth, and he holds the narthex (which may appear in place of the thyrsus), but also, he carries the liknon on his head.

51 Spinazzola 1953, fig. 578, pl. XXX.b. Stylistic and iconographic similarities between these triclinium paintings and the painted decor of the ambulatory passage in the Villa della Farnesina (see supra nn. 46, 47) suggest a date for the former.

52 Rostovzeff 1927, 77–78, fig. 14; but see Spinazzola 1953, 517–19, for the scene as a presentation of a trophy to Ariadne.

53 The same is true of another naked boy in a different panel of this room who sits below a couch and a dining trio of Silenus, a maenad, and a satyr: Spinazzola 1953, 507, pl. XXIX.

54 Starr 1991; Dozier 2008, 108–21. Note that lector is a Latin term for a (presumably male) enslaved reader (there were probably occasional women readers, but most that we know of are male), whereas anagnostes is a transliteration of the Greek word ἀναγνώστης, a slave trained to read aloud.

55 Howley 2020.

56 Cic., Att. 1.12.4. Latin puer may refer to a child, a boy, or a slave (possibly of an older age, though this is uncommon). In this particular passage, I am taking puer as a reference to the reader’s young age because of both the associated adjective festivus (charming, pretty, jovial) and the appositive anagnostes, which evinces his enslavement. See Nep., Att. 13.3, for Atticus’ “pueri litteratissimi, anagnostae optimi et plurimi librarii” (highly literate slave boys, excellent readers and very many scribes), who were trained in Atticus’ home. Here the young age of the pueri is not definitive, but the use of puer in conjunction with comments on these slaves as born and educated in Atticus’ house is suggestive. For the trope of a slaveowner’s lament for his deceased lector and paternalism as a form of mastery in Pliny the Younger’s letters, e.g., 8.1: Joshel 2011, 234–38; Blake 2017, 94–96.

57 Yunis 2001, 253–55.

58 See, e.g., Nilsson 1957, 106–15, where he discusses children but glosses over their nudity as unproblematic.

59 His nudity is better associated with his identity as an infans; see discussion above and supra n. 22.

60 For the dress of child attendants (e.g., camilli) in particular: Mantle 2002.

61 Child initiates are not always naked. See, e.g., another stucco relief from the Villa della Farnesina, which depicts a child of perhaps 6 to 10 years being led to an altar (possibly for his own initiation?) by an older, dressed woman: Nilsson 1957, fig. 11; Dolciotti et al. 1998, fig. 100. Initiates, moreover, are almost always dressed in extant images, e.g., a first-century CE painting of an adult initiate of unknown gender from the Domus Aurea (Nilsson 1957, fig. 16); a Campana relief, later first century BCE or Augustan era, of an older male initiate (Nilsson 1957, fig. 18).

62 The innate purity of a child is somewhat of a modern conceit, although it is true that infants in the Roman period are regularly naked in Roman art (see supra n. 22).

63 According to Johnston (2001, 106–7), “the child per se and purity per se must be kept separate”; children were not inherently pure and were likely subject to the same purification rituals as adults. For purification rituals in the context of Dionysiac cult (sexual abstinence and baths): Livy 39.9.4; Ov., Fast. 2.327–30.

64 See also Johnston’s (2001, 114) comments about practitioners of divination working with children drawn from pools that catered to other “unsavory professions.”

65 Livy 39.8–19.

66 See, e.g., Walsh 1996.

67 E.g., Livy 39.10.6; 39.13.14; 39.15.13. Hispula’s personal testimony also provides indirect evidence of initiates as both free and enslaved, female and male: Livy 39.10.5; 39.12.6.

68 For the ritual nudity of youths in Greek culture: Burkert 1985, 260–63; Bonfante 1989, 550–52. For children in Greek religion more broadly: Neils 2003.

69 Kouros statues remain foundational to our understanding of ritual and heroic nudity; see, e.g., Bonfante 1989; Stewart 1997.

70 Osborne 1997. The titillating character of the Greek nude is also discussed in Bonfante 1989; Stewart 1997; Hurwitt 2007. See Shapiro 1992 for Greek depictions on vases of young boys and pederastic relationships. For idealized young slaves in Greek art: Wrenhaven 2013, 76–80, 86–89. Connections between Greek nudity, the gymnasium in particular, and pederasty are also discussed by ancient Roman authors, e.g., Cic., Tusc. 4.70.

71 Roman attitudes toward young boys and pederastic relations are discussed in greater depth in the next section, but see Richlin 1983, 34–44; Williams 2010, 68–84. As Pollini (1999, 28) notes, Greek homosexual relations were not something Romans learned from the Greeks; rather Greek customs like pederasty “probably gave impetus to the practice … among Roman citizens” in the Late Republic because they were seen as fashionable.

72 See Uzzi 2015 for the sexual politics of the colonial narrative and eroticization of young non-Romans.

73 Called a triclinium: e.g., Wallace-Hadrill 2018. Others refer to the room more generally as an oecus (a Greek term used in Vitruvius for large rooms devoted to banqueting and reception): e.g., Ling 1991; Kuttner 1999. On the multipurpose nature of this room: Longfellow 2000a.

74 For spatial syntax analysis: Longfellow 2000a.

75 Previous scholarship on enslaved persons depicted in wall paintings and other arts has acknowledged the social hierarchies perpetuated by an elite viewing system (e.g., George 2013; Lenski 2013), but there are differing opinions about the degree of violence perpetrated on nonelite viewers and the complicity of such persons; see, e.g., Green 2015 on the normalizing effect of these images; Bielfeldt 2018 on object agency and subversions of the slave-as-tool rhetoric.

76 See discussion in Bergmann 2007 (250–61) of this image’s reception in the 20th century.

77 Kuttner (1999, 108) commented that “all the satyrs are oddly civilized in deportment and unusually fully or even conventionally draped.”

78 Female features deemed attractive in Latin love poetry and Imperial-era Greek epigrams include hair, breasts, and eyes; buttocks may receive praise, but genitalia are rarely discussed; Richlin 1983, 44–56.

79 Concerns about the room’s division into two pictorial halves (mythological versus mortal) are raised in Rizzo 1918, 91; Nilsson 1957, 74–75; but cf. Clarke’s arguments against a continuous narrative (1991, 99–100).

80 For the downward glance as an invitation: Shapiro 1992 (Greek boys); Bartman 2002 (Roman boys).

81 For young boys in early Latin literature: Williams 2010, 20–40; for Imperial-era literature: Richlin 1983, 34–44.

82 Lucr. 4.1052–56.

83 Different standards of sexual mores were set for freeborn men and women, but sexual contact with enslaved bodies was perfectly acceptable. Ancient authors (e.g., Hor., Sat. 1.2.114–19) advocate using slaves as sexual relief, thereby avoiding the disgrace that accompanied sex with a freeborn.

84 See, e.g., Philo, De vita contemplativa 50–52, for voyeurism at the cena and slave boys of different ages. On dining, slaves, and the archaeological record of Pompeii: D’Arms 1991; Joshel and Petersen 2014, 37–59. See also Roller 2006.

85 For slaves as prestige objects: e.g., Juv. 5. Juvenal compares his own waiter, a dark-skinned “Ganymede,” to his host Virro’s waiter, dubbed “the flower of Asia” (5.51–62).

86 Philo, De vita contemplativa 50: διακονικὰ ἀνδράποδα εὐμορφότατα | καὶ περικαλλέστατα, ὡς ἀφιγμένα οὐχ ὑπηρεσίας ἕνϵκα μ̑αλλον ἢ το̑υ φανέντα τὴν τ̑ων θεωμένων ὄψιν ἡδ̑ναι (Ready to serve are slaves of exceptional form and beauty, as though they have come not to provide service, but rather to appear and so to delight the eyes of viewers).

87 E.g., D’Arms 1991, 173–74. See Roller 2006 on “standing as a posture of instrumentality” (177) and child slaves at dining parties (162–63).

88 See Pollini 1999; 2010. For erotic paintings in private domiciles generally, though not necessarily with attention to the enslaved: Clarke 1998, 145–94; Elsner 2007, esp. 177–99.

89 Lenski 2013; Bielfeldt 2018.

90 Lenski 2013, 136–41.

91 Varro, Rust. 1.17.1, discussed in Lenski 2013, 130–31. See also Joshel and Petersen (2014, 27) on the identification of domestic slaves by their task, which characterizes them as furnishings of the house. For the paradox of slaves as objects with agency, see Joshel 2011.

92 Petron., Sat. 41.6. The age of the boy is not mentioned; we hear only that he is a puer speciosus (beautiful boy).

93 See Dozier 2008, 108–21, on the performance of lectores at dinner parties in the Imperial era.

94 Similar comments might be made about the triclinium paintings in Pompeii’s House of the Chaste Lovers (IV.12.6–7) and House of the Triclinium (V.2.4). A nude dancer and a male delicium appear in a fresco at the latter (Clarke 2003, 233–45, pls. 17, 18, 20–22).

95 Discussion of the boy’s relationship to these women follows below, but for recent work on female subjectivity and erotic paintings: Frederick 1995, 279–87. For female subjectivity in other contexts: Richlin (e.g., 1992) and Levin-Richardson (e.g., 2019).

96 Boys were held to be most attractive before their first beard; for depilatory treatments and efforts to ward off advancing age: e.g., Sen., Ep. 47.7.

97 As far as male viewers are concerned, “it is often implied [in Roman-era erotic poetry about boys] that the lover admires smaller boys and awaits the day when such a boy will be ready for anal intercourse, at the same time fearing the day when the boy will be too old for it” (Richlin 1983, 37).

98 Mart., Epigramata 10.66.

99 Many scholars accept 12 as the age at which a boy becomes sexually interesting to a Roman audience; Laes 2011, 275–77; but see supra n. 84 on enthusiasm for pueri of different ages.

100 Richlin 2015.

101 For the definition of deliciae and the range of relationships this term seems to cover (parent–child to slave owner–enslaved): Richlin 2015; Laes 2003 (who notes that positive associations with deliciae tend to be Late Imperial, 305); Rawson 2003, 261–63. Laes (2003, 304–14) in particular discusses the epigraphic evidence, which supports synthesis of deliciae (and related terms delicium; delicatus) as a term of endearment and genuine devotion; still, most deliciae named in inscriptions are (or were formerly) enslaved.

102 Rawson 2003, 261; quoted by others, e.g., George 2013, n. 31; Richlin 2015, 365; Uzzi 2015, 254.

103 On Melior’s boy Glaucias: Stat., Silv. 2.1.41–51, 60–64; Asso 2010; Laes 2011, 223–30. On the eroticization of Erotion, Martial’s five-year-old puella in Epigramata 5.37 (identified in 5.34 as his deliciae): Watson 1992; Laes 2011, 256–58.

104 Cass. Dio 48.44.3. See also Suet., Aug. 83, for the emperor’s own pueri. As these passages demonstrate, deliciae (and related terms) refer to the same group of child slaves as pueri minuti (pl., small boys) and Greek πα̑ιδες (pl., children broadly but also enslaved persons): Slater 1974; Laes 2003.

105 For slave boys as witnesses to sexual acts: Green 2015, 147–48; also Pollini 1999, 38–39, on the voyeuristic slave boy who appears on the Augustan-era Warren Cup.

106 Clarke 1998, 93–107, for discussion of erotic paintings in cubicula B, D, and E of the Villa Farnesina; also Bragantini and de Vos 1982, 128–233, 284–336. The paintings likely date to 40–30 BCE; see supra n. 45.

107 In this way the fresco affirms the needs, desires, and feelings of the elite viewer, working in much the same way as erotic poetry about deliciae; the child’s feelings are never discussed, let alone entertained. See, e.g., Laes’ comments on Stat., Silv. 2.1, and insensitivity toward Glaucias’ true “feelings,” which Laes compares to the treatment of rape victims (2011, 225–26).

108 Laes (2003, 302) argues that Cassius Dio’s comment about the routine nakedness of deliciae stretches the truth, but for later qualifications: Laes 2011, 234–36.

109 Prop. 2.29a.1–7: Hesterna, mea lux, cum potus nocte vagarer, | nec me servorum duceret ulla manus, | obvia, nescio quot pueri, mihi turba, minuti, | venerat (hos vetuit me numerare timor); | quorum alii faculas, alii retinere sagittas, | pars etiam visast vincla parare mihi. | sed nudi fuerant (Yesterday night, my love, when I was staggering drunk in the night, and no band of slaves was leading me, a crowd—some number of small boys—came and got in my way (fear forbade me from counting them); some of them seemed to be holding torches, others arrows, still another part seemed to be ready to put me in chains. Yet they were naked). Richardson (2006, 295) noted, “We need not suppose that the poet’s hallucinatory adventure was more than a drunken dream,” but what matters is the poetic elision of Cupids and pet boys.

110 For other elegiac women who own pueri: e.g., Catull. 56.5 on the poet’s rape of Lesbia’s boy; but see also Richlin 2015, 356–58.

111 Plut., Vit. Ant. 26: αὐτὴ δὲ κατέκειτο µὲν ὑπὸ σκιάδι χρυσοπάστῳ κεκοσµηµένη γραφικ̑ως ὥσπερ Ἀφροδίτη, πα̑ωιδες δὲ το̑ις γραφικο̑ις Ἔρωσιν εἰκασµένοι παρ᾿ ἑκάτερον ἑστῶτες ἐρρίπιζον (She [Cleopatra] herself reclined, under a gold-flecked canopy, adorned like she was Aphrodite in a painting, and boys, made to look like painted Erotes, stood on each side and fanned her). See also Pelling 1988, 186–89.

112 Although Cleopatra is a Hellenistic ruler, Plutarch’s second-century CE audience likely had little trouble associating her πα̑ιδες with Roman practices of deliciae and slave ownership.

113 George 2013. The author does not cite the ancient references to Cupids or Amores reviewed here; I do so independently to support her findings and encourage readers to acknowledge the erotic tone that characterizes images of nude young children in Roman art broadly. For earlier arguments in this vein see Birt 1893. See also Richlin 2015, 359–68, on Birt’s approach generally.

114 Here I echo Joshel’s (2011, 236–37) challenge to Pliny the Younger as a sympathetic slave owner, which takes into account Hartman’s (1997) work on slavery in 19th-century America.

115 Cf. Herbig (1958, 34), who argues that the woman holding the stylus and a scroll in her left hand may be a priestess with some function as recorder in the cult.

116 For the training of deliciae in the “art” of chattiness: Stat., Silv. 2.1.72–75 and 5.5.66–69; Sen., Constant. 11.3 notes the role of a magister (teacher) in training young boys in the art of impudent speaking. See Keegan 2013 for graffiti in Rome’s Palatine Paedagogium, which signals reading and writing as basic skills taught to these slaves.

117 Slater 1974, 135. Lenski (2013, 130) begins with a similarly cautious proviso.

118 Gazda 2021.

119 According to Gazda (2021, 136–38), the woman is represented in various stages of her life: as a veiled domina on the west wall; as a young bride in the guise of Ariadne in the now lost fragment of Dionysus’ consort on the east wall; and as a bride-to-be in the small portrait held by the cupid for her daughter, who prepares for her own nuptials on the south wall.

120 Bergmann 2017, esp. 154–60, on the “willful dissolution of boundaries between mortals and immortals” (156) in Roman art of the Late Republic and Early Empire.

121 Gazda 2021, 140–42.

122 See supra n. 103 for Statius’ poems about Glaucias and Martial’s Erotion; Pollini 2010 for portraits of slave boys in sculpture. It is more difficult to identify material examples as true portraits, but for portraiture in Roman painting and tropes versus particularized details: Bergmann 2017, 143–45.

123 Clarke 2003, 241–42; 2007, 130–31.

124 To my knowledge there is no comprehensive study of the visual depiction of child slaves with their owners in Roman art, but see Uzzi 2015; and see Davies 2018 for gesture and identity in Roman art more broadly.

125 For sitting as a mark of superior status: Davies 2018, 195–97.

126 Touching is frequently (though not always) a mark of higher status in Roman art: Davies 2018, esp. 16, 254–55.

127 In matrimonial scenes, women frequently look at their spouse, or look down in deference; men regularly fail to acknowledge or reciprocate these gestures; Davies 2018, 251–57.

128 See Joshel and Petersen 2014, 1–17, on un-seeing and seeing slavery in the archaeological record.

129 Romans are frequently discussed as a visually literate people (e.g., Elsner 2007); the use of art and architecture to project power, whether in the home or the public area, also presumes competent viewers.

130 Green (2015, 149), e.g., has interpreted images of enslaved women and boys observing sex acts as explicit reminders of their vulnerability and the things they were forced to endure. The despotics of the elite, primarily male, gaze also loom large in Severy-Hoven 2012.

131 Here I build in part on bell hooks’ (1992) theory of the oppositional gaze and Black viewers of cinema, especially female viewers.

132 For compelling studies of slave resistance in domestic contexts not discussed in this paragraph: Webster 2008; Bradley 2011; Bielfeldt 2018; Roth 2021. For nondomestic contexts: Levin-Richardson 2019.

133 Joshel and Petersen 2014, esp. 7–8.

134 The notion of a “hidden transcript” builds on the work of political theorist James C. Scott (1990).

135 Richlin 2017.

136 Joshel and Petersen 2014, 8–17, 59–78, 191–95. Their analysis of slave tactics that disrupt established power dynamics builds on Michel de Certeau’s concepts of “strategy,” which belongs to the dominant class and is backed by law, and “tactics,” which are chance offerings or opportunities that push back at dominant systems without necessarily altering or rejecting them.

137 For pattern books and the fresco’s use of Late Republican visual vocabulary: Clarke 1991; Ling 1991; Kuttner 1999; Gazda 2000.

138 For the iconography of child participants in the mysteries, and for nudity among nondominant groups, see the arguments above. See also Davies (2018, 82–118) on the female nude; her discussion of the Aphrodite of Knidos (2018, 82–89) entertains various reactions to the goddess’ naked body, depending on age, gender, and world view.

139 For the use of human models, see Plin., HN 35.37.119–20. For direct observation of live models as a strategy of painters when there was no model to be found in pattern books, see Clarke 2003, 156, 273.

140 For such body language as defensive in the Roman world but also in the animal world broadly: Davies 2018, esp. 97–98.

141 See, e.g., Clarke 2003, esp. 9–13, on viewing scenarios and “what if” scripts, which acknowledge varied reactions to art based on the status and gender of the viewer. My argument here responds to what we know of painters in the Late Republic. For the “import” of painters in this period, and the use of both free and enslaved painters in workshops: Clarke 1991, 45–47.

142 Joshel 1992.

143 Joshel and Petersen 2014, esp. 1–3.

Works Cited