In his 1981 CQ article ‘Dionysiac drama and Dionysiac mysteries’, Richard Seaford argued that both Greek mystery cults and Greek drama had roots in older, coming-of-age rituals.Footnote ¹ Focussing on initiatory themes in Euripides’ Bacchae, Seaford conjectured that ritual patterns of maturation rites were preserved in Dionysiac cult and thereby helped to shape the development of tragedy.Footnote ² In recent decades, however, separate developments in the study of tragedy, maturation rituals and mystery cults have all shifted the ground on which Seaford's case was built. At the time of Seaford's article, Arnold van Gennep's three-stage model of rites de passage Footnote ³ had already been applied to Greek maturation rituals,Footnote ⁴ and the same model has proven productive as a tool of analysis for Athenian drama.Footnote ⁵ At the same time, the category of ‘initiation’ has also received needed refinements, and scholars have scaled back its applications to much ancient evidence, including age-transition rituals.Footnote ⁶ Meanwhile, evidence for early Bacchic mysteries continues to appear from unexpected quarters. Further Orphic-Bacchic funerary gold tablets have been discovered, mainly from the Late Classical period;Footnote ⁷ these finds, together with the fourth-century Derveni PapyrusFootnote ⁸ and the fifth-century Orphic bone plates from Pontic Olbia,Footnote ⁹ have dramatically expanded our evidence for early Dionysiac τελεταί and left us far less reliant on later Graeco-Roman sources than was the case only a few decades ago.Footnote ¹⁰

Seaford's claims are now due a fresh look. Shifts in our understanding of age-transition rituals call for some theoretical adjustments. Rites of transition in the Greek world do not always follow an initiatory pattern. They were probably less prominent in Greek society than was once assumed; where practised, they existed alongside other kinds of ceremony, both public and private, that could mark passage from adolescence to adulthood.Footnote ¹¹ Most scholars today would hesitate to accept a developmental account of the evolution of Dionysiac mysteries from maturation rituals, and in any case the similarities between mystic initiations and age-transition rites do not necessarily indicate that they developed within a continuous tradition.Footnote ¹² Even if these points are conceded, however, Seaford's observations retain much of their merit, and my purpose is to revisit some of his basic insights that remain valid within the current evidentiary picture. In particular, I argue that a limited case for a parallelism and even direct interaction between Greek age-transition rituals and mystery initiations can still be sustained. Rather than revisiting Seaford's argument point by point, however, my focus will be on two pieces of fourth-century evidence: an Orphic-Bacchic funerary gold tablet from Hipponion and a Cretan pederastic mock-abduction ritual recounted by Ephorus of Cyme. Following in the trajectory of recent ‘neo-ritualist’ approaches to Dionysiac cult,Footnote ¹³ I suggest that these two sources point toward a basic homology between puberty rituals and mysteries in Greek religious thought during the Classical period. I especially focus on the word κλεινός (‘illustrious’), which was used in a quasi-technical sense to designate the elevated status of initiates in both maturation rites and Bacchic mysteries. I do not presuppose or argue for any direct genealogical development between age-transition rites and the mysteries of Dionysus. In view of their shared vocabulary and structural parallels, however, I contend that such private mysteries modelled themselves in certain respects after the authoritative language, norms and ideologies of maturation rituals in the Classical period, and perhaps earlier.

The Orphic-Bacchic gold leaves play only a marginal role in Seaford's article,Footnote ¹⁴ but the growing corpus of these objects constitutes our most important direct evidence for early Dionysiac mysteries. The Hipponion gold leaf (OF 474), first published in 1974, was excavated from a woman's cist grave near Locri Epizephyrii and has been dated to c.400 b.c.Footnote ¹⁵ The hexameter text runs sixteen lines and is still the best-preserved of the B-group that come from a range of locations in Magna Graecia, Sicily, Thessaly and Crete.Footnote ¹⁶ Like other texts from this group, the Hipponion inscription guides the initiate's soul through the Underworld. On entering Hades, she is warned to avoid the first stream she encounters (probably to be understood as the waters of Λήθη), where ordinary souls congregate. Instead, she is instructed to continue on toward the cool waters flowing from the lake of memory (πρόσθεν δὲ εὑρήσεις τᾶς μναμοσύνας ἀπὸ λίμνας | ψυχρὸν ὕδωρ προρέον, 6–7). There she is challenged by Underworld guardians to explain her presence. She declares herself a ‘child of Earth and starry Heaven’ (Γῆς παῖ⟨ς⟩ εἰμι καὶ Οὐρανοῦ ἀστερόεντος, 10) and asks to drink from the stream. The wish is granted, and her arrival is announced to the king of the Underworld. In the final lines, she is promised a place with her fellow initiates: ‘You, too, shall go along the sacred road that other illustrious initiates and bacchants tread’ (καὶ δὴ καὶ σὺ … ὁδὸν ἔρχεα⟨ι⟩ ἅν τε καὶ ἄλλοι | μύσται καὶ βάκχοι ἱερὰν στείχουσι κλεινοί, 15–16).

The word βάκχοι in the final line received the most attention when this text was first published, since it gave the first clear evidence of a Dionysiac background for the gold leaves.Footnote ¹⁷ The final κλεινοί, however, has enjoyed less notoriety. This word and the line containing it do not appear in any other gold leaves, but both of them likely belonged to the source-text from which the B-tablets derive. Despite its emphatic position at the end of both the line and the inscription,Footnote ¹⁸ the word has attracted little notice, and neither of the two standard commentaries offers any remark on it.Footnote ¹⁹ Those who have commented on κλεινοί have rightly noted its semantic connection with κλέος and poetic fame.Footnote ²⁰ Yet, while the texts of the gold tablets frequently draw on traditional poetic diction,Footnote ²¹ κλεινός/κλεεινός never appears in early hexameter epic (which uses κλειτός instead); it is therefore difficult to read κλεινοί in the Hipponion text as an epic borrowing. κλεινός does appear in early lyric, usually modifying toponyms and localities: Stesichorus refers to ‘illustrious Erytheia’ (κλεινᾶϲ Ἐρυθείαϲ, fr. 9.2 Finglass) and to the Achaeans as sackers of the ‘illustrious [plain?] of well-founded Troy’ (Τ]ρ̣οΐα̣ϲ κλεεν̣νὸ[ν … (ἐκ)πέ]ρϲαντεϲ ἐϋκτιμέ̣[νης, fr. 117.6–7 Finglass);Footnote ²² a brief papyrus fragment of Alcman speaks of ‘illustrious Sardis’ (κλ[ε]ι̣ναὶ Σάρ[διες, fr. 13d PMG), and Solon calls Cyprus an ‘illustrious isle’ (κλεινῆς … νήσου, fr. 19.3 W.). (Two additional attestations in Alcman that are connected more directly with age-group initiations—frr. 1.44 and 10b PMG—will be examined shortly.) Similar uses of κλεινός with wider applications are common in poetic genres of the Classical period, especially in epinician and tragedy. In epinician, κλεινός especially marks the nodes and conduits (including significant locations, but also people, local myths/heroes, songs and festivals) through which κλέος accrues to the victor, to his city and household, and to the poet.Footnote ²³ This poetic background, while relevant to any interpretation of κλεινοί in the gold tablets, cannot entirely account for the word's prominence in the Hipponion text or explain its meaning in the context of Bacchic mysteries.

One parallel from the ritual sphere, roughly contemporary with the Hipponion tablet, has passed largely unnoticed. The fourth-century historian Ephorus of Cyme, in an indirect fragment preserved in Strabo's Geography (10.4.16–22 = BNJ 70 F 149), describes male age-transition rituals of Crete that include a pederastic mock-abduction.Footnote ²⁴ In Crete, Ephorus relates, young boys were gathered in ‘herds’ (ἀγέλαι), and boys would serve in the men's dining hall (ἀνδρεῖον), first as wine-bearers and subsequently in pederastic pairings. Such practices among the Dorian aristocracy in Crete resembled Spartan age-transitions, as both Ephorus and Aristotle observed; but the process in Crete was shorter and culminated with the young men being married en masse to the corresponding female cohort (who would presumably have undergone an initiatory process of their own).Footnote ²⁵ Within this complex of interconnected age-transition rituals, Ephorus tells of a pederastic mock-abduction (ἁρπαγή) that conferred a distinctive social status on certain young men. He does not assign the ritual to any specific point in a boy's ‘development’, and he notes that not all boys underwent this abduction; but it must have taken place subsequent to the boy's introduction into an ἀνδρεῖον and shortly before marriage.Footnote ²⁶ In this practice, according to Ephorus, the abductor declared his intentions to the boy's older male relativesFootnote ²⁷ at least three days in advance, and it was expected that they would offer only token resistance. A mock pursuit followed, after which the boy was brought to the abductor's ἀνδρεῖον. The abductor presented the boy with gifts and took him into the wilderness, where the two would eat and hunt together for up to two months (the duration permitted by law). On their return to the city, the boy received additional gifts (again prescribed by law) that included warrior's clothes, a sacrificial bull and a drinking-cup. The boy's family contributed additional non-prescribed gifts of conspicuous expense. To be abducted was an honour, and privileges accrued to abductees: they received the most prominent seats at public dances and races, and they were entitled to wear the fine clothes given to them by their abductors—garments that distinguished them from other young men who had not undergone the ritual. Abductees at this stage were called ‘the ones placed beside’ (οἱ παρασταθέντες). Later, as adults, abductees continued to wear distinctive dress, ‘from which each one would be recognized as having become illustrious’ (ἀφ᾽ ἧς γνωσθήσεται ἕκαστος κλεινὸς γενόμενος). The abduction rite is integrated into the larger complex of Cretan age-transition rituals, and it follows a sharply delineated initiatory pattern of separation, liminality and reintegration with a change of status.

Aspects of Ephorus’ testimony in this fragment have been justifiably questioned. The fragment is an indirect testimony related in paraphrase by Strabo. Ephorus’ assignment of the abduction ritual to ‘Crete’ rather than to any specific city is suspect, and similarities with the Spartan ἀγέλαι and the Athenian ἐφηβεία suggest that his account of Cretan practices is influenced by familiarity with those rituals. Nevertheless, Ephorus in his fragment shows continuities with older (and specifically Cretan/Theran) material, and his account is consistent with social concerns that appear in age-transition rites in other parts the Greek world.Footnote ²⁸ The present argument, in any case, relies on Ephorus less for referential accuracy than as a witness to puberty rituals in Greek cultural imagination at the end of the Classical period, especially regarding their shared vocabulary and structural similarities with other ritual genres.

In both the Cretan ἁρπαγή and the Hipponion tablet, the word κλεινός marks as exceptional certain individuals who have undergone a rite of initiation. Some resonances of this usage are perhaps audible in poetic uses of κλεινός, especially in ceremonial contexts. Epinician victors, for instance, are marked as κλεινοί in the moment of proclamation (Pind. Pyth. 1.29–33; Bacchyl. 5.182–6), and Bacchylides says that victors become κλεινοί when they are crowned: ‘Illustrious among mortals are they who, from those famous contests in Nemea, cap their golden locks with a triennial crown’ (κείνων ἀπ᾿ εὐδόξων ἀγώνων | ἐν Νεμέαι κλεινο[ὶ β]ροτῶν, | οἳ τριετεῖ στεφάνωι | ξανθὰν ἐρέψωνται κόμαν, 9.21–4). The στέφανος of athletic victory, like the finery of Cretan κλεινοί, is a distinguishing ‘garment’ that makes the victor's status visible. The same sense is discernible even in non-ritual contexts where an exceptional status is acknowledged or conferred. Thus Sophocles’ most celebrated protagonist introduces himself as ‘the man called Oedipus, illustrious to all’ (ὁ πᾶσι κλεινὸς Οἰδίπους καλούμενος, OT 8), and Athena in Euripides’ Erechtheus confers upon the king's daughters ‘an illustrious name’ throughout Greece by proclaiming them Hyacinthid goddesses (ὄνομα δὲ κλεινὸν θήσομαι κα[θ᾿ Ἑλλ]άδα | Ὑακινθίδας βροτοῖσι κικλή[σκε]ιν θεάς, fr. 65.73–4 Kannicht).

Tellingly, however, the closest parallels for κλεινός/κλεινοί in Crete and Hipponion come from Alcman's Partheneia and the maturation rituals of Archaic Sparta. That these choral songs belong to a complex of initiatory rites for young women is now widely agreed.Footnote ²⁹ In the Louvre Partheneion (Alcm. fr. 1 PMG), the chorus praise their leader Hagesichora as ‘the illustrious chorus-leader’ (ἁ κλεννὰ χοραγός, 44; cf. [χο]ροστάσις, 84), pre-eminent (ἐκπρεπής, 46) in her visual beauty, with focus on her hair (50–4, 101?), face (55) and ankles (78). Along with her counterpart Agido, Hagesichora also offers prayers and performs rituals on behalf of the chorus (78–91).Footnote ³⁰ In a separate and much briefer fragment preserved as a quotation in an ancient commentary (fr. 10b PMG = P.Oxy. 2506), the male leader of a women's chorus is praised as ‘Hagesidamus, chorus-leader beloved of the gods, illustrious son of Damotimidas’ (σιοφιλὲς χο[ρα]γὲ Ἁγησίδαμε κλε[νν]ὲ Δαμοτιμίδα, 11–12).Footnote ³¹ No individual apart from these two χορηγοί in Alcman's extant works is called κλεινός, and Budelmann observes that the expression χοραγὸς κλεννά ‘may have had a formulaic ring’ for Alcman's audiences.Footnote ³² In Partheneia, the role of χορηγός seems typically to have been taken by someone close in age to the chorus, but of higher status, who facilitated transitional rites for the choral cohort. A female χορηγός, such as Hagesichora, was also to some degree a ‘model initiate’ for the young women of the chorus. It is probable, though uncertain, that the girl taking the role of Hagesichora had herself previously undergone the same choral initiation rite. In any case, the chorus praise Hagesichora and Agido especially for their ritual expertise and mature physical beauty: in the context of the choral song, these qualities distinguish the leaders from the chorus, but are also traits that the female initiands will possess as fully ‘initiated’ adult women who are prepared for marriage.Footnote ³³ For my purposes, it is notable that the word κλεινός in both Crete and Sparta is used to mark successful passage through initiatory maturation rites. It is an attractive conjecture that the recognition of certain individuals as κλεινοί in these two settings—the κλεννὰ χοραγός in Alcman, the pederastic κλεινός in Crete—reflects a common underlying ritual vocabulary, especially in light of the more extensive correspondences between the Cretan and Spartan ritual complexes.Footnote ³⁴

The availability of such a vocabulary, with roots in Archaic maturation ritual, would carry substantial interpretative implications for κλεινοί in the Hipponion tablet. Although a number of shorter B-tablets have been found in Crete, it is doubtful that the ritual sense of κλεινοί in the Hipponion text is due to Cretan influence.Footnote ³⁵ More likely these are parallel applications of a common ritual vocabulary in two different contexts. Ephorus notes that κλεινός is the Cretan term for an ἐρώμενος, and LSJ⁹ accordingly cites κλεινός as a Cretan equivalent for the Attic τὰ παιδικά.Footnote ³⁶ This identification, however, is potentially misleading removed from its larger context. For Ephorus, not only does the ἁρπαγή belong to a complex of different coming-of-age rituals, but the status of κλεινός is enjoyed into adulthood, literally worn as clothing, long after the abduction. The two terms for abductees, παρασταθέντες and κλεινοί, seem (at least notionally) to separate the ritual's temporary pederastic phase from the enduring status conferred on the abductee.Footnote ³⁷ Ephorus also stresses the ritual's social embeddedness: the boy's family, he says, are glad to hand him over (ἐπιτρέπουσιν ἄγειν χαίροντες), and it is considered shameful if an attractive boy ‘of good stock’ (τοῖς … καλοῖς τὴν ἰδέαν καὶ προγόνων ἐπιφανῶν) fails to attract a suitor. Several aspects of the ritual are legally prescribed or regulated, and the public privileges conferred on the abductee imply a wider communal interest in the ritual and its outcome. Becoming κλεινός, then, has less to do with the pederastic relationship itself than with the social setting in which the abductee's new status is conferred and recognized.

Seaford and others have argued that mystic initiation accomplished an analogous revision of personal identity.Footnote ³⁸ An initiate of Dionysus entered a ‘liminal’ state where ordinary markers of identity such as gender were suppressed or reversed, culminating in a ritualized ‘death’ and subsequent ‘rebirth’ into a new and elevated state. Transvestism is a characteristic device by which social identity is destabilized in Dionysiac cult, and the Hipponion text (like the texts of other gold leaves) omits the gender of its owner along with most other determinants of social identity, such as proper name, parentage, etc.Footnote ³⁹ The Cretan ἁρπαγή accomplishes something analogous, albeit at a public level that reflects more obviously public concerns. The pedagogical context of the ἁρπαγή, often overlooked in scholarly discussions, is crucial for a correct understanding of the practice: as Ephorus makes clear, the practice belongs to an interconnected system of rituals, culminating in a mass wedding, that facilitate the adolescent boys’ transition into adult roles. The ‘rape’, with its temporary reversal of sexual and gender roles, deprives the adolescent of his identity as a child (a ‘death’) in preparation for his ‘rebirth’ as a man, husband/father, and citizen.Footnote ⁴⁰ The pattern of ritualized death and rebirth, with the instrumental inversion of sexual norms, corresponds with the pattern of Dionysiac initiation. To be clear, this homology does not show that initiations developed directly from maturation rites: it does indicate, however, that aspects of age-transition rituals could be readily identified with features of the mysteries and vice versa.

The word κλεινός/κλεινοί suggests a further correspondence between the two ritual systems. In a brief but insightful study of the Hipponion tablet written shortly after its first publication, Domenico Musti suggested that its concluding word might be connected with the sought-after waters of μνημοσύνη and the role of memory in constructing the initiate's post-mortem identity. While the primary sense of ‘memory’ in the B-texts must be the activation of the individual consciousness of the deceased after death, this ‘individual’ memory is also entangled with two forms of ‘social’ memory: the preservation of the memory of the deceased among the living mourners whom she leaves behind, and the recognition of her initiate status among fellow μύσται καὶ βάκχοι with whom she is united in the hereafter.Footnote ⁴¹ Here, as in the Cretan abduction, becoming κλεινός presupposes a larger spectatorship: ‘This κλέος is … perhaps an allusion to the community where the memory and fame of the deceased reside.’Footnote ⁴² This imagined community of recognition must also include the guardians who compel her to disclose her mystic identity and the Underworld gods to whom she is announced and on whom she depends for post-mortem reception.Footnote ⁴³

In a more obvious sense, of course, it is the mourners who buried the Hipponion initiate who are the true ‘spectators’ of the deceased. Although we are ill-informed about the social realities of Orphic-Bacchic cults, they seem to have been propagated by itinerant initiators, and it is likely that such Orphic cult associations as existed would have been small, local and short-lived. Despite this social reality, however, the evidence of the gold leaves has made it clear that an idea of community among initiates was important to the ideology of the mysteries.Footnote ⁴⁴ In any case, it is ultimately among these real-life cult participants that the initiate and her fellows are recognized as κλεινοί. In the gold leaves, this imagined community always remains extra-textual, but they are reflected indirectly in the idealized Underworld society of initiates (μύσται καὶ βάκχοι, 474.16; cf. πρὸς μυστῶ⟨ν⟩ θιάσους, 493a), ‘heroes’ (ἡρώεσσιν, 476), ‘pure ones’ (ἕδρας ἐς εὐαγέων, 489–90) and ‘blessed ones’ (ὄλβιοι, 485). While poetic comparisons cannot entirely account for the use of κλεινοί in the gold leaves, it is no great stretch in the Hipponion tablet to see in this imagined network of initiates something ideologically similar to the ‘paths of praise’ mapped by Pindar and other epinician poets: both genres plug their subject into a nexus of ongoing ritual/poetic performance that assures immortality and an elevated post-mortem status.Footnote ⁴⁵ Poetic performance and initiatory ritual both imagine an audience for themselves and make performers into ‘self-spectators’, with the texts of the gold leaves promulgating an identity for the cult group itself as much as for the deceased.Footnote ⁴⁶

In the Cretan ἁρπαγή, it is the individual who is stamped κλεινός, whereas in Hipponion the entire mystic cohort are recognized as κλεινοί, even as markers of individual identity are elided or suppressed. Both rituals, however, agree in differentiating the initiate from non-initiates. The Cretan ἁρπαγή may be usefully redescribed in Pierre Bourdieu's terms as a ‘rite of institution’, whereby specific groups could be prepared and marked for membership in their appropriate classes, functions and occupations.Footnote ⁴⁷ As Bourdieu observed in an important modification of Van Gennep's model, the most meaningful distinction asserted in transitional rites de passage is not that between the individual's pre- and post-initiatory states but between those who undergo the ritual and those who do not. Such a description fits the Cretan ritual and other Greek coming-of-age practices very well, as Mark Griffith has shown.Footnote ⁴⁸ Age-transition rituals in other parts of the Greek world, such as the Arkteia at Brauron, were elite affairs, and not universally accessible even for elites.Footnote ⁴⁹ Pederasty in the Greek world is generally marked as an elite practice, and several features of the Ephorus fragment—the expense of the gifts, the prestige that accrues to the abductee and his family, the family's concern for his status and recognition—make clear that recognition as κλεινός was valued mainly among the upper class.Footnote ⁵⁰ As Ephorus suggests, the imposition of social distinction—and, implicitly, the ultimate acceptance and legitimation of such distinction—was an explicit function of the Cretan ἁρπαγή.

Bourdieu's model may also be applied, with modifications, to the evidence for Bacchic cults. Central to the ideology of all Greek mysteries, both private and institutional, was the basic differentiation of initiates from non-initiates. In the Mysteries of Eleusis, this distinction was observed through a pretence of secrecy: even though large numbers of Athenians were initiates and any initiatory ‘secrets’ would have been widely known, the conceit of secrecy was taken seriously, and violators (most famously Alcibiades) could be punished. The private Dionysiac mysteries stressed even more forcefully the ‘separate’ status of initiates. The incipit used in Orphic poetry (OF 1a–b)—attested in the Derveni Papyrus (P.Derv. col. VII) but also familiar enough for Plato to allude to it (Symp. 218b)—addresses ‘those who know’ (ξυνετοῖσι) or ‘those for whom it is permitted’ (οἷς θέμις ἐστί) while exhorting the uninitiated to ‘shut the doors’ (θύρας δ’ ἐπίθεσθε, βέβηλοι).Footnote ⁵¹ Orphic-Bacchic eschatology sharply differentiates initiates from non-initiates: in the Hipponion leaf, the initiate must first of all shun the company of the other souls who refresh themselves at the stream of Oblivion, seeking instead the waters of Memory that will (re)unite her with her exceptional mystic community. Similar motifs of differentiation appear in other gold leaves and testimonia of the Classical period about Bacchic and Eleusinian mysteries: the Eleusinian mystics of Aristophanes’ Frogs, for instance, dance in the light and feast on roasted pork, while their uninitiated counterparts wallow in mud and dung.Footnote ⁵² As in the Cretan ἁρπαγή, the initiate's final recognition as κλεινός reflects the social ideology of the cult group. Unlike the Cretan ritual, Dionysiac mysteries played no part within any larger complex of ‘public’ ritual. Quite to the contrary, mysteries at all levels frequently cut across social ordinary divisions and hierarchies: women and even slaves could be initiated at Eleusis and Samothrace, and the archaeological evidence of the gold leaves suggests that Dionysiac cults were especially popular among women.Footnote ⁵³ Both Marcel Detienne and Radcliffe G. Edmonds III have understood these cults as devices of protest (chemins de déviance) available to socially marginal groups.Footnote ⁵⁴ Adapting Bourdieu's terms, these might be described as rites of ‘counter-institution’, imagining a different ‘community’ from that of the polis and affirming the initiates themselves as κλεινοί.Footnote ⁵⁵

This typology of the Cretan and Orphic-Bacchic rituals, especially given their common use of κλεινός/κλεινοί to different purposes in their respective contexts, accords with many of Seaford's observations about the relation between maturation rites and Dionysiac mysteries. These parallels need not imply that the latter developed directly from the former; but the sharing of vocabulary and overall functions (‘institution’/‘counter-institution’) suggests that participants in the Classical period, and perhaps earlier, recognized an affinity of function between these ritual genres. Perhaps it is for similar reasons that the Arkteia at Brauron, one of the most widely discussed of the early Greek age-transition rituals, was characterized by scholiasts and lexicographers as an ‘initiation’ (τελετή) or a ‘mystery’ (μυστήριον).Footnote ⁵⁶ Dionysiac mysteries employ such initiatory ritual mechanisms within a ‘private’ social ambit that is more restricted than that of the ‘public’ Cretan ἁρπαγή, but within their own context they are no less socially meaningful than the outward-facing Cretan transition rites. The designation κλεινοί in the Hipponion leaf points us toward an all-too-neglected social dimension of a private religious practice in the Classical Greek world.